Gabbay 1999 Corporate Social Capital and Liability

Corporate Social
Capital and Liability

Corporate Social
Capital and Liability
Edited by
Roger Th.A.J. Leenders
School of Management and OrganiZOlion
University ojGroningen, The Netherlands
Shaul M. Gabbay
Davidson Faculty of Industrial Engineering and Management
Technicn - Israellflsli/utl! o!Tedllw{ogy
....
.,
Springer Science+Business Media, LLC
...
" Electronic Services <http://www.wkap.ol>
Library of Congress Cataloging-in-Publication Data
Corporate social capital and liability I edited by Roger Th. AJ.

Leenders, Shaul M. Gabbay.
p. em.
Includes index.
ISBN 978-1-4613-7284-4 ISBN 978-1-4615-5027-3 (eBook)
DOI 10.1007/978-1-4615-5027-3
1. Business networks. 2. Social nelworks. 3. Industrial

organizatian. 1. Leenders, Roger Th. A. J. II. Gabbay, Shaul M.
HD69.S8 C67 1999
302.3'5 - - dc21
99-28470
CIP
Copyright@I999bySpringerSciCIlCl'+RusinessMediaNewYork
Originali)' published by Kluwer Academic Publishers 1999
Softcover reprint of the hardcover 1st edition 1999
AII rights reserved. No part of this publicatian may be reproduced, stored in a

retrieval system or transmitted in any form Of by any means, mechanical, photo-
copying, recording, or otherwise, without the prior written permission of the
publisher, Kluwer Academic Publishers, 101 Philip Drive, Assinippi Park, Norwell,
Massachusens 02061
Printed an acid-free paper.

Contents
•
INTRODUCTION
CSC: The Structure of Advantage and Disadvantage
Shaul M. Gabbay & Roger Th.AJ. Leenders
Section I
CONCEPTUAL ISSUES - Theory, Models, and Measurement
Organizational Networks and Corporate Social Capital 17
David Knoke
2 Social Capital of Organization: Conceptualization. Level of Analysis.
and Performance Implications 43
Johannes M. Pennings & Kyungmook Lee
3 A Relational Resource Perspective on Social Capital 68
Luis Araujo & Geoff Easton
4 Social Capital by Design: Structures. Strategies. and Institutional Context 88
Wayne E. Baker & David Obstfeld
5 Corporate Social Capital and Liability: a Conditional Approach 106
to Three Consequenses of Corporate Social Structure
lIan Talmud
6 Dimensions of Corporate Social Capital: Toward Models and Measures 118
Shin-Kap Han & Ronald L. Breiger
7 Organizational Standing as Corporate Social Capital 134
Patrick Doreian
8 Customer Service Dyads: Diagnosing Emperical Buyer - Seller
Interactions along Gaming Profiles in a Dyadic Parametric Space 148
Dawn Iacobucci
Section II
STRUCTURE AT THE INDIVIDUAL LEVEL - Social Capital in Jobs and Careers
9 The Sidekick Effect: Mentoring Relationships and the Development of
Social Capital 161
Monica Higgins & Nitin Nohria
10 Social Capital in Internal Staffing Practices 180
Peter V. Marsden & Elizabeth H. Gorman
11 Getting a Job as a Manager 197
Henk Flap & Ed Boxman
12 The Changing Value of Social Capital in an Expanding Social System:
Lawyers in the Chicago Bar. 1975 and 1995 217
Rebecca L. Sandefur, Edward O. Laumann & John P. Heinz
vi - Corporate Social Capital and Liability
Section III
STRUCTURE AT THE INDIVIDUAL LEVEL - Social Capital and Management
13 Generalized Exchange and Economic Performance: Social Embeddedness of
Labor Contracts in a Corporate Law Partnership 237
Emmanuel Lazega
14 CEO Demographics and Acquisitions: Network Effects ofEducational
and Functional Background 266
Pamela R. Haunscbild, Andrew D. Henderson & Alison Davis-Blake
15 Public Service Organizations - Social Networks and Social Capital 284
Ewan Ferlie
16 The Dark Side ofSocial Capital 298
Martin Gargiulo & Mario Benassi
17 Social Capital, Social Liabilities, and Social Resources Management 323
Daniel J. Brass & Giuseppe Labianca
Section IV
STRUCTURE AT THE FIRM LEVEL - Social Capital in Collaboration and
Alliances
18 The Triangle: Roles ofthe Go-Between 341
Bart Nooteboom
19 The Management of Social Capital in R&D Collaboration 356
Onno Omta & Wouter van Rossum
20 Technological Prestige and the Accumulation ofAlliance Capital 376
Toby E. Stuart
21 Networks and Knowledge Production: Collaboration and Patenting in
Biotechnology 390
Laurel Smitb-Doerr, Jason Owen-Smith, Kenneth W. Koput & Walter W. Powell
22 Supply Network Strategy and Social Capital 409
Christine Harland
Section V
STRUCTURE AT THE FIRM LEVEL - Social Capital and Financial Capital
23 Choosing Ties from the Inside ofa Prism: Egocentric Uncertainty and
Status in Venture Capital Markets 431
Joel M. Podolny & Fabrizio Castellucci
24 Corporate Social Capital and the Cost ofFinancial Capital:
An Embeddedness Approach 446
Brian Uzzi & James J. Gillespie
25 Venture Capital as an Economy of Time 460
John Freeman
CONCLUSION
CSC: An Agenda for the Future 483
Roger Tb.A.J. Leenders & Sbaul M. Gabbay
REFERENCES 495
INDEX 545
CONTRIBUTORS 559
EDITORS 563
• Introduction
CSC: The Structure of Advantage
and Disadvantage
•
Shaul M. Gabbay
INTRODUCTION
Scholars of the firm have long concerned themselves with identifying the
differential characteristics that make some firms-and some of their members-
more successful than others. A recent approach to the study of success and failure in
the competitive marketplace is the theory of social capital. The theory of social
capital suggests that players gain access to various kinds of resources that accrue to
them by virtue of their engagement in various kinds of relationships. Social capital
theory is fundamentally concerned with the resources inherent within structures and
social exchange. Until now, social capital theory has mainly been applied to
individual actors-human beings. In this volume the central question is how is
social structure related to the attainment of goals of corporations and their
members (denoted below by the terms 'corporate players' or 'corporate actors')? We
suggest that Corporate Social Capital refers to the resources, inherent in the social
structure, that accrue to corporate actors. Social structure refers to a network of
actors who are in some way connected via a set of relationships.
In the current chapter, we will briefly discuss the concept of social capital in the
context of organizations and introduce the concept corporate social capital.
Subsequently, we will discuss the relationship and distinction between social
structure and social capital. In so doing, we will introduce the notion of corporate
social liability and focus our attention on the various levels of aggregation at which
social structure, social capital, and social liability reside. We will coin the acronym
CSC to denote the interplay of social structure and social capital/social liability , both
at the firm, intermediate, and individual levels. In the concluding chapter of this
volume, we will discuss the research and practical applications of further
developments. We will highlight critical questions that, in our view, should be
2 - Corporate Social Capital and Liability
resolved. In our final discussion, we will open new questions for future discussions.
At its initial stages, CSC is an emerging research agenda. This, of course, presents a
wide window of opportunity for future and further contributions.
SOCIAL CAPITAL AND ORGANIZATIONS

The term 'social capital' first appeared in scientific literature around a century ago,'
but it was not until the mid-1980's that a theory of social capital was developed and
applied in the context of work and organizations. James Coleman was the first to
develop a comprehensive theory of social capital and his work inspired the diffusion
of the use of the theory in relation to the study of actors who are pursuing interest
driven goals. Coleman's seminal work-Foundations of Social Theory (1990)
sparked a host of studies applying social capital theory.2 The work of Burt (1992)
was explicit in its focus on corporate actors. In his study of managerial mobility in a
high technology firm, Burt was the first to introduce a quantitative measure for
social capital. Studying White et al.'s (1976) suggestion that the absence of ties may
provide an advantage, Burt (1992) showed that actors who are connected to
disconnected others (structural holes) advance faster in the corporation under study.
These studies explicitly used the term social capital. Many social network
studies, that did not explicitly employ the term 'social capital' have also proven
important in the development of social capital theory. For example, the work of
Granovetter (1973), The Strength of Weak Ties, greatly influenced the study of
social capital as did the work of Lin (1982; Lin et al. 1981) and the work of
Laumann (1973). The theory of social capital has gained a prominent place in a wide
range of scientific fields, spanning social, economic, and political research agendas.
For a discussion of the history and use of the theory of social capital we refer to
Portes and Sensenbrenner (1993), Gabbay (1995, 1997), Nahapiet and Ghoshal
(1998), and Woolcock (1998).
CORPORATE SOCIAL CAPITAL

Although a large body of research has emerged on social capital, and a growing
group of researchers is now using social capital in their research, consensus on the
definition of social capital has yet to be established. Some authors equate social
capital with social structure, whereas others refer to the resources an actor can
mobilize through the social structure. In the current volume, we specifically refer to
the assets embedded within and available through networks of relationships. We
define social capital as the set of resources, tangible or virtual, that accrue to an
actor through the actor's social relationships, facilitating the attainment of goals.
Before moving on to the discussion to 'corporate' social capital, it is important to
discuss briefly three elements of our view of social capital. First, we view social
capital as goal-specific. A large number of social ties does not necessarily translate
itself into social capital. It only does so if these ties assist the actor in the attainment
of particular goals. An example of this is provided by the following event. In
February 1996, Ben Van Schaik, the CEO of Dutch aircraft builder Fokker, gave a
presentation to a potential alliance partner. 'We are the second largest aircraft builder
in the world,' he said, supporting his claim with a colorful bar graph-the bars
CSC: The Structure of Advantage and Disadvantage - 3
represented various aircraft manufacturers; their length represented the number of

the manufacturer's clients. With this graph, Van Schaik showed the audience that
Boeing/McDonnell Douglas had 846 clients, Fokker had 225, and the rest trailed
behind at a large distance.3 Unfortunately, the number of ties did not represent the
number of aircraft sold, nor did it represent the credit rating of the customers.
Fokker's clients were primarily small companies, leasing only one or two airplanes
(rather than buying them), and many of them 'forgot' to pay their bills. Fokker
declared bankruptcy only a few months later and was liquidated. Its 'much smaller'
competition is still alive and kicking. 4 Social network and social capital are different
things. A social network only conveys social capital if its social ties are beneficial
for the attainment of goals. In the case of Fokker, they clearly were not.
Second, an actor need not be conscious of the social capital he enjoys. The
social structure in which an actor is embedded may confer advantages to the actor,
without the actor even realizing it.
Third, the social structure that brings opportunities for the realization of
particular goals need not have been built in the pursuit of these goals-social capital
often is a by-product of other social activity.
Like other forms of capital, social capital is productive, making possible the
achievement of goals that would not be attainable in its absence. This goal
specificity has a number of implications.6 First of all, social structure may be
beneficial for the attainment of multiple goals, since the multiplex character of many
social relationships results in overlap of opportunity structures. Networks created for
one purpose may be employed for another-which often was not foreseen when the
actors initially engaged in a relationship with each other. In some situations, the
same social structure can be beneficial for the attainment of one goal, while
obstructing the attainment of others. 7 Social networks can have a positive effect in
providing network members with access to privileged resources, while lowering
transaction costs, but can, simultaneously, place high demands on these network
members, restricting their individual behavior and opportunities.
In this volume, we concern ourselves with corporations and their members. s
Corporate social capital, then, refers to:
The set of resources, tangible or virtual, that accrue to a corporate player
through the player's social relationships, facilitating the attainment of goals.
Although social structure and social capital are often equated in the social
capital literature, they are different entities. In this volume we therefore make an
explicit distinction between social structure per se and the outcomes of social
structure. When these outcomes are positive, helpful in attaining specific goals, we
say the social structure conveys social capital. But when social structure prohibits
and obstructs action, it produces social liability. Although the absence of social
structure precludes social capital from corning into existence, the two are distinct.
They are generated by related, but distinct, processes.
SOCIAL CAPITAL AND SOCIAL LIABILITY

Just as bank accounts can run dry and can have a negative balance, social capital can
vanish. Social structure that provided social capital in the past may not necessarily
do so in the present. The social structure required to sustain corporate social capital
may shift as transactions, activities, and conditions change and become more or less
complex. Relationships beneficial to the achievement of goals in the past may thwart
goal attainment in the future. In their study of how managers adapt to changing
environmental conditions, Gargiulo and Benassi (this volume) find that relational
structures that in the past had provided ample social capital for managers, later
increased the number of coordination failures for which these managers were
responsible. The network no longer provided resources, but had become a constraint,
impeding performance.
Brass (1984) and Blau and Alba (1982) found that relationships to the clique of
top executives in an organization had a strong positive relationship with power and
promotions. Top executives were likely to have more social capital in the form of
more (relevant) information to share with those who were connected to them. Since
men are more likely to maintain relationships to top executives than are women,9
women may be forced to forgo any preference for homophily in order to build
connections with the dominant coalition and share the social capital. Brass and
Labianca (this volume) therefore conclude that actor dissimilarity (in this case based
on gender) may affect interaction patterns and may consequently exclude some
people from sharing in the social capital. From the viewpoint of human resources
management, Brass and Labianca provide ample examples of the positive and
negative effects social relationships yield on social resource management outcomes
such as recruitment, socialization, turnover, job satisfaction, power, and conflict.
Leenders (1995) shows that the relations among social service workers may help
in preventing these workers from becoming burned out, but that feelings of burnout
are also contagious through these exact same relationships. Increases in burnout
experienced by social service workers also increased the level of burnout
experienced by their co-workers. Conversely, decreases assisted in decreasing
burnout among alters.
Gabbay (1995, 1997) shows that, for some actors, strong ties combined with
structural holes ('structural ports') were beneficial at the inception stage of their
business, but were detrimental for future expansion. Successful entrepreneurs
strategically changed their networks over time in order to maintain or build social
structures that would be rich sources of social capital.
Social structure translates into social liability in at least two different ways.
First, ongoing, strong social relationships may constrain the behavior of actors,
impeding their action and attainment of goals. IO For example, long-standing
relationships with customers may stifle the firm by monopolizing a disproportionate
share of its resources, inhibiting the firm from forming relationships with alternative
customers. Second, actors may be unfavorably affected in their opportunities by
negative ties in the social structure. In this volume, Brass and Labianca explore the
effects of this type of social capital. As an example of the social liability stemming
from negative ties, they argue that 'it is likely that an actor's negative ties within an
organization will prevent promotion, particularly if those negative relationships are
with influential others. Others may withhold critical information that worsens an
actor's performance or they may provide bad references in order to prevent a
promotion' (this volume: 324).
Organizations and their members continuously engage in social relationships.

Effective organization requires a constant balancing of the potentially opposing
forces inherent in social structures.
LEVELS OF ANALYSIS
Relational structures can be recognized at different levels of analysis and
observation. Four levels of analysis normally characterize organizations. \I The
individual human being is the basic building block of organizations. The human
being is to an organization as a cell is to a biological system. One step higher, we
find group or departments, where individuals work together on group tasks. Next is
the organization itself: a collection of groups or departments. Above this level,
organizations themselves can be grouped into an interorganizational network. 12
Social capital and social structure are relevant at all of these different levels. Classic
methodological approaches require a researcher to choose one particular level of
analysis as the primary focus of a study. However, as we will discuss below, the
very nature of social capital runs through all these organizational levels. Social
structures at the individual level translate into social capital and social liability for
individual actors, but also translate into social capital effects at higher levels of
analysis. Similarly, the structure/capital connection also works its way down the
levels of analysis. A full study of social capital should thus incorporate structure and
capitallliability at mUltiple levels of analysis.
Relations Among Individuals

Intraorganizational networks have been the locus of exuberant research efforts. The
effects of network structure on, among many other things, the distribution of power,
job satisfaction, career opportunities, and productivity of individuals have been
central on the research agenda. Undoubtedly the most attention has been given to the
sources and consequences of the distribution of power in organizations, focusing on
such concepts as prestige, status, and control. Next on the agenda has been the
related topic of information flows in organizations. This type of research addresses
questions related to 'who has what information when and how'? Considering the fact
that we are only at the start of the information age, and are pushed along by rapid
developments in information and communication technology, this latter question is
likely to remain among the most popular objects of study for a long time.
The distribution of power and information is both an outcome and generator of
network structure. As an outcome, it portrays (part of) the social capital (and social
liability) that employees extract from their relationships. As a generator, social
capital affects intraorganizational structure. For instance, by the very nature of his
work, a top executive has many relational ties, spread throughout the organization,
and is at the intersection of numerous flows of information. This network position,
characterized by high centrality and the presence of structural holes, provides the
manager with ample opportunity for control, power, and influence: the manager
draws social capital from his network structure. In turn, his power gives him the
opportunity to surround himself with a network of loyal subordinates and use his
network ties to stay informed about events throughout the organization. By placing
trusted friends in strategic positions, a manager gains power by being well-informed,
by having access to other (important) people in the organization, and by having

mUltiple people depend on him. Capitalizing on his social capital, the manager
effectively reengineers his position in the network. Ideally, this will then allow him
to harvest more social capital from his (new) network, or secure the capital drawn
from the old structure-social capital creates social capital.
Social structures among individual members of the firm relate to social capital
at various levels of analysis. First, intraorganizational networks yield social capital
for the individuals comprising these networks as individuals employ their direct and
indirect ties to fulfill individual goals. Occupying a central position in various
networks of relationships is often regarded as a source of power--power is social
capital to those who can utilize it for getting the job done. Having a tie with a
mentor can be both helpful and detrimental in securing jobs (see the chapter by
Higgins and Nohria in this volume). Social ties may help individuals to find (more
favorable) jobs both in the framework of internal market opportunities (Marsden and
Gorman, this volume) and in the context of external labor markets (Flap and
Boxman, this volume). In particular, having access to higher-status individuals is
often argued to be beneficial in the job search process.
Second, social structures at the level of individuals also impact social capital at
higher levels of analysis. For instance, Kratzer, Van Engelen, and Leenders (1998)
discuss how the structure of various types of relationships among members of R&D
teams affects their success. They contend that segmentation of the problem-solving
network among the team members is detrimental to the performance of the team, an
effect stronger for teams performing highly complex tasks than for teams
performing tasks of low complexity. R&D teams that agree on their basic product
development goals-but are characterized by a reasonable disagreement on how to
achieve these goals-tend to develop products with a much higher probability of
market success than teams whose members fully agree (or disagree) on these
issues.13 Based on the studies focusing on social contagion,14 it is possible to
conclude that the structure of social relationships between individuals affects their
level of consensus and, consequently, their group performance. The structure of
these relationships thus affects outcomes at the firm level.
Social structures at the individual level of observation can also have negative
effects at these higher levels of analysis. For instance, consider the so-called
grapevine-an informal, person-to-person communication network of employees
that is not officially sanctioned by the organization. Grapevines are sources of
rumors and gossip that spread quickly throughout an organization. Often,
management decisions circulate through grapevines days ahead of their official
announcement. 15 Because they feel threatened by it, managers often try to suppress
the grapevine, but find themselves confronted with a nearly impossible exercise.
Another example is related to the resilience of personal networks. Managers in
charge of (re)designing business processes often experience difficulties in breaking
through the power structures that exist between the firm's employees. 16 As Knoke
(this volume: 40) notes, 'power structures are highly stable and resistant to change.'
In addition to relationships with others within a particular firm, individuals also
maintain relationships with those outside of the firm. The relationship a lawyer
maintains with her clients often directly translates into status, financial turnover, and
profitability at the law firm level. In effect, the firm draws social capital from its
employee's contacts (who also draws social capital from the relationship herself).
Interlocking directorates connect firms-they lead to information sharing and policy
binding and can be a strong source of organizational power.17 In the chapters by
Knoke and by Pennings and Lee, the boundaries differentiating roles of individuals
are discussed extensively.
Relations Among Organizations

No organization really stands on its own. Almost every organization needs suppliers
and buyers, and deals with intermediaries. Ongoing exchange relations with other
organizations translate into more or less stable interorganizational networks.
Drawing from theories as diverse as contingency theory, resource dependence
theory, transaction cost theory, and population ecology, scholars are collecting data
and devising explanations of interorganizational relationships. Unfortunately,
however, these theories hardly devote any attention to how these interorganizational
relationships lead to a network of relationships. 18 They also ignore how these affect
opportunities and restrictions for the firms comprising these networks.
Recent trends influence the extent and intensity of the engagement in
interorganizational relationships. The unification of Europe has led to a surge of
mergers, alliances, cooperation, and licensing. More than ever before, European
companies are looking at the structures around them, actively investing in new and
reassessing old ties. Advantages are not necessarily in the creation of a large bundle
of ties, but in the creation and maintenance of relations that provide access to new
technologies, resources, and legitimacy.19 In other words, companies are actively
molding their networks so as to draw as much corporate social capital from them as
possible, with a long-term focus.
Europe is not the only stage where interorganizational relationships are crucial.
With strong trends toward globalization in virtually every sector, and with rapid
developments in information and communication technologies, it is possible and
necessary for many firms to enter into intercorporate relationships.
Just as individuals can tap social capital from their social networks in order to
facilitate some action or attainment of goals, so too can organizations extract
resources from their networks. Harland (this volume) shows how corporate social
capital is created through the buildup of strategic interorganizational supply
networks. Using the cases of Benetton and Toyota she illustrates how firms, that are
seeking to increase competitive advantage, actively pursue the role of a 'supply
network hub' to facilitate effective and efficient flows of materials and information.
Smith-Doerr et al. (this volume) examine the link between social structure and
intellectual output in the context of the biotechnology industry, in which
interorganizational collaboration is commonplace. Their interest is in the number of
patents organizations obtain as a result of their collaborative relationships. For
instance, they argue that firms that are more central in their collaboration network
will attain more patents in a timely manner, due to their deeper insights into the
knowledge-base of the field and their ability to combine and utilize a broad
configuration of partners' expertise in choosing promising lines of research and
framing patentable claims (this volume: 394). Not only does a favorable position in
the network (as measured by centrality) lead to the generation of more patents, but
receiving more patents, in turn, also leads to an even more favorable network
position. In other words, social capital is both an outcome and a generator of social
structure.
Stuart (this volume) also studies the social capital that can be gained from
alliances. In concert with Smith-Doerr et al. (this volume), patents playa central role
in his study. Stuart's interest is in whether the status of an organization (as a
function of the organization's position in the patent citation network) has an effect
on the rate at which the organization acquires technologies from other firms.
Technologically prestigious firms are found to be able to access the technological
assets of other firms at a higher rate than are firms of lesser prestige. Interestingly,
the rate at which a firm acquires technology from its competitors is more strongly
positively related to the firm's technological prestige than is the rate at which the
firm supplies technology. In this sense, high-status firms gain a positive balance of
social capital from alliances with lower-status partners by taking more technology
than they give. Being at the other end of the relationship, the lower-status firm sees
some of the status of its partner reflected on its own technology, and draws a
different type of social capital from the same relationship.
The question of how status relates to social capital is also the focus of the
chapter by Freeman (this volume). Freeman shows that start-up businesses that
receive support from centrally connected venture capital firms (VCs) have an
increased probability of an initial public offering. In Freeman's study, the structural
position of one organization (the VC) produces social capital for another (the start-
up business). But status also brings a risk. As the analyses show, VC centrality also
increases the probability of a startup business being acquired. Here, the structural
position of the VC creates (potential) social liability for the entrepreneur. The VCs
themselves also experience risks. Freeman suggests that more central, high-status
VCs behave like option traders by making high-risk investments and walking away
from the losers. This could undermine the deal flow advantages that centrality
affords alter the central position of the venture capital firm.
Uzzi and Gillespie (this volume) are concerned with the question of how social
structure affects the probability of a small business securing capital. In partiCUlar,
they examine the relationship between social structure and lending practices. They
show that small businesses garner more loans at lower interest rates by increasing
the duration and multiplexity of their relationships with financial institutions. Their
findings clearly show that by strategically engineering social structure and by
creating a proper mix and intensity of ties to financial institutions, small businesses
can gain much social capital (in the sense of securing capital at decent prices).
So far, we have only concerned ourselves with the relationship between
interorganizational networks and social capital at the organization level. But these
structures also yield social capital at the individual level. For example, alliances or
consortia can provide employees with the opportunity to be trained or temporarily be
located at other organizations. Employees can also more easily change employers,
moving between firms within one alliance. With employability becoming
increasingly important in the workplace, these interorganizational relationships
Structure Structure
firm level individual level
Figure 1. esc framework: Interdependencies between social structure and social capital at
various levels of analysis
provide the individual employee with the social capital of learning new skills and
the opportunity to display his skills beyond the boundaries of the ·firm he is
(currently) affiliated with. Obviously, this also creates potential social liability for
the organization, as it may lose valuable employees. In this volume, Higgins and
Nohria study career opportunities of employees moving between companies.
Mixtures of Levels
Employees of every firm are involved in social relationships that extend beyond
their own firms. Consequently, firms themselves are part of wider social and
economic networks that can be expected to influence their relationships with
potential suppliers and customers.
Many relationships between organizations are mediated by individuals. Lawyers
bring their clients to their law firms. Formal relationships between firms often start
out as informal, personal relationships between representatives of these firms and
then become institutionalized. But not all interorganizational relationships are
mediated by individuals. As Pennings and Lee (this volume: 50) note 'as a legal
entity, the firm is capable of contracting, of acting as a partner in any market
relationship, including the setting up of joint ventures, the acquisition of another
firm, or the shedding of a business unit to other firms, etc .. '
An important aspect of the institutionalization of ties maintained by the firm's
members is that it shifts the agency of the social capital to the firm level. A
consultant, whose social contacts bring in large revenues for his firm, may try to
keep the relationship 'to himself.' By retaining the relationship in his personal
portfolio, he has a powerful means to strengthen his position in the firm. A manager
whose contacts lead to a formal, contractually regulated alliance transfers his claims
to the relationship (and the social capital inherent in it) to his organization.
An organizational risk attached to the personal nature of relationships is that

these relationships may be lost when the employee leaves the firm. Consultants may
take their contacts with them to their new employer. 20 Organizations search for ways
to appropriate individual-level social capital, often by developing a reward system
that favors employees who bring in the most social capital. Salesmen often are
rewarded in proportion to their achievements in terms of sales and revenues;
successful consultants are offered the opportunity to become a partner in the firm.
Relational structure may have both positive and negative consequences for the
attainment of goals of corporate players at either level of analysis (see Figure 1).21
Strong inter-firm relationships (firm level structure) can provide the firm (firm level
social capital) with corporate social capital-they provide the firm with resources
while lowering risks and costs of opportunism. At the same time, these ties may
provide firm members (individual level social capital) with social capital by giving
them access to a number of other firms to which they can relocate to further their
individual careers. As a result, the firm is confronted with the corporate social
liability of seeing its valued employees leave (firm level social liability).
Relationships of individuals of firm members with others outside the firm
(individual level structure) can provide the firm with valuable clients (firm level
social capital). The ties among individuals in teams and departments (individual
level structure) affect their job satisfaction (indidividual level social capital) and
affects the productivity of the group (firm level social capital). As argued above, a
full study of social capital should deal with the association between structure and
capitalIliability at mUltiple levels of analysis.
CSC
To summarize, we have the following elements:
• Social structure and social capital are related but distinct entities;
• Social capital represents the resources that accrue to an actor through the actor's
social relationships, facilitating the attainment of goals. When social structure
hinders the attainment of goals, it yields social liability;
• 'Corporate social capital' refers to the social capital of corporate players: firms
and their members;
• Social structure and social capital can be distinguished at different levels of
analysis, at minimum at the levels of the firm and the individual;
• Social structure and social capital are not only associated within the same level
of analysis, but are outcomes and generators of each other at all of these levels;
• Social structure can provide social capital to one player, but social liability to
another. It can provide social capital for the attainment of one goal, but social
liability for another goal.
We have found it useful to use the acronym 'CSC' to denote the constellation of
{structure, capital, liability, level of analysis}. In this volume, we will employ this
acronym in the current chapter and in the concluding chapter.
CSC: The Structure of Advantage and Disadvantage - II
THE ADDED VALUE OF esc

The study of social networks has been the focus of the work of organizational
scholars for over a decade. A growing body of studies from various areas and
disciplines has focused on the study of social networks in the context of
organizations. However, Salancik (1995) recently noted that the wealth of social
network studies and methodology has not yet produced an encompassing theory of
organization. He therefore suggests that 'a network theory of organization should
propose how structures of interactions enable coordinated interaction to achieve
collective and individual interests.'22 Theories of corporate social capital present a
(partial) answer to this call. CSC theory attempts to explain how social structure is
connected to organizational outcomes. It explicitly relates to (the interplay of)
structures at the individual level and the firm level, and to how these affect
attainment of relevant organizational goals. This invokes questions such as: What
kinds of social structures bring what kinds of outcomes? Which networks are good
for what actors and under what conditions? When are networks vessels for social
capital and when do they render social liability? The fundamental difference
between the study of social networks and organizations in general and the
overarching agenda of CSC is that, within the framework of CSC, social network
structures are the focus of attention in their explicit productive or destructive
association with goal attainment of corporate players. With this approach, CSC
shifts attention to more daring questions. We move from a descriptive set of findings
to a theoretically based orientation with practical implications. A CSC theory
provides a different framework for understanding organizations.
Organizational Paradigms
The study of corporate social capital suggests an analytical framework that cuts
across other theoretical frameworks that are eminent in the literature on
organizations--for example, Scou's (1992) paradigmatic typology of organizations
as rational, natural, and open systems. 23
The rational paradigm perceives organizations as 'collectives oriented to the
pursuit of relatively specific goals.' The basic assumption of this paradigm is that of
goal-oriented rationality. Rationality suggests a cost-benefit analysis--a balance
between the interest of actors, costs, and payoffs. The rationality-based premise of
our definition of corporate social capital is straightforward because of the explicit
focus on the functional (negative or positive) aspects of social structures in their
relation to the goals of corporate actors. Social structure in the framework of CSC
has costs and benefits and is directed towards the attainment of goals.
The natural paradigm perceives organizations as 'collectives whose participants
share a common interest in the survival of the system and who engage in collective
activities, informally structured.' It is largely in the framework of social networks
that collectives operate. These social networks can be measured, mapped, and
related to outcomes-at the firm and individual levels. The development, creation,
and maintenance of these social networks to fit corporate goals are what ultimately
transform these networks into sustained corporate social capital. esc is thus
explicitly related to the notion of organizations as natural systems. In the natural
paradigm, the survival of the system is a goal-the challenge in studying corporate
social capital is to explain which social structures make the achievement of this goal
easier or more difficult. 24
The open system paradigm sees organizations as open systems 'embedded in-
dependent on continuing exchanges with and constituted by-the environment in
which they operate.' The open system perspective is a natural context for the theory
of corporate social capital, which emphasizes social structure in the relevant
environment of firms and their members. An important asset of corporations lies
outside and beyond their immediate boundaries. Firms are embedded in
interorganizational networks, affecting the goal attainment of corporate players. A
firm's environment is made of opportunities and potentially creative relationships
that may further the goals and achievements of corporate players. From a social
capital perspective, the boundaries of firms fade into their environment-
corporations are embedded in a networked environment. The relationships between
(members of) the firm and players outside the firm's formal boundaries affect the
efficiency and effectivity of the firm's internal organization and, at the same time,
can be modified to suit the firm's goals. The ramifications of the existence of
'external' ties on corporations are at the heart of CSc. 25
Placing the notion of corporate social capital in the paradigmatic discourse of
organization studies suggests that CSC contributes a new framework and an
additional dimension to the study of organizations and organizational processes.
THE VOLUME
It is common sense that relationships are important, both for individuals and for
corporations. It is also intuitively clear that not all relationships are (equally) useful.
Unfortunately, it is largely unknown which relationships, under what conditions, are
beneficial or obstructive to organizations or their members. It is often argued that
part of the explanation of which corporations (individuals) do better than others is
found in the relationships these corporations (individuals) maintain. But we are not
yet sure what exactly it is in those structures that benefit some and impede others.
The world is becoming increasingly complex and dynamic. Organizations must
incorporate even greater diversity to survive and thrive. More complexity compels
more organizations to (try to) increase their social capital. Organizations enter into
various forms of inter-corporate relationships. We are moving from core
competence to core capabilities, increasing the tendency toward hybrid forms of
organization. Changing labor contracts are now built on the notion of
employability--organizations offer less permanent contracts.
In a competitive and fast-paced world, the differences between the winners and
the losers will more strongly be determined by the way actors make use of their
opportunity structures. Social capital represents the pipeline to those opportunities.
Social capital theory has gained prominence in a wide range of academic fields,
particularly in sociology, economics, political science, and managementlbusiness
science and is appearing with increasing frequency in intellectual magazines and
professional reports.26 Woolcock (1998: 184) even claims that social capital is
'arguably the most influential concept to emerge from economic sociology in the last
decade.' He also points out that 'in social capital, historians, political scientists,
anthropologists, economists, sociologists, and policy makers-and the various
camps within each field-may once again begin to find a common language within
which to engage one another in open, constructive debate, a language that
disciplinary provincialisms have largely suppressed over the last one-hundred-and-
fifty years' (Woolcock 1998: 188).
Notwithstanding the substantial insights in the literature into the advantages
and-in a much smaller part of the literature-the predicament of social
embeddedness, researchers still lack a coherent theory for explaining how and when
social structure transforms itself into corporate social capital or into corporate social
liability. We were both disappointed and inspired by this. Trying to develop a more
coherent and encompassing framework for studying the interplay between social
structure and social capital in the context of corporate actors, we aimed at extending
the insights from various fields into a systematic exploration of the concept. We
therefore invited prominent scholars to contribute to this volume and suggested they
remain within their own field of expertise in writing their chapters. As a
consequence, some of the chapters have a slightly different take on corporate social
capital than we do in this chapter. Still, we believe that the chapters provide a
coherent overview of the field . Some of the chapters are of a theoretical nature,
some of them methodological, many of them empirical.
The first chapters in the book present conceptual issues dealing with CSC
theory, models, and measurement. These chapters delineate the pressing challenges
in these three domains. In the second and third section, CSC is discussed starting
from social structure at the level of the individual. The effect of social structure is
discussed as it facilitates or inhibits players in pursuit of their professional careers-
moving ahead in a competitive environment. The third section explicitly highlights
two basic characteristics of CSC: contingency and fragility-not every social
network conveys social capital and social networks carrying social capital at one
point in time can create social liabilities at another. Social structure at the firm level
is considered in the next two sections---highlighting ways In which
interorganizational networks bring social capital to the firm.
We had the pleasure of reading multiple versions of the chapters, moving from
rough outlines to the texts you find in this book. If, after reading several chapters,
the reader has an (increased) appreciation for the theory of corporate social capital, it
has all been worth it. We hope you will have as much pleasure reading the book as
we had putting it together. Enjoy!
We thank Laura Rittmaster and Judith de Kleuver for their help in producing this book. The assistance
and gentle pressure provided by Julie Kaczynski and Elizabeth Murry of Kluwer Academic Publishers
were instrumental for the project as well. We thank the School of Management and Organization (Cluster
of Business Development) at the University of Groningen and the Davidson Faculty of Industrial
Engineering and Management at the Technion for their support and resources.
NOTES
l. The first appearance of social capital in the scientific discourse (see soc-net discussion on the
genesis of social capital) was in Marshal (1890) and Hicks (I942)-1hey, however, used the term to refer
to different types of physical capital. Hannifin (1920) employed the term in the context of community
studies. Again in the context of community studies, social capital was used by Jacobs (1965) and Hannerz
(1969). Loury (1977, 1987) used the term in the field of child psychology. Bourdieu (1972, 1980)
developed the term in reference to cultural capital. The first researchers who used social capital explicitly
in the study of organizations were Flap and De Graaf (l986}-in their investigation of job mobility.
Coleman (1988, 1990) was the most influential in developing a general wide-scale theory of social capital
as it is mostly used by scholars today. The work of Bur! (1992) was important in its wide visibility among
students of organizations as well as for his explicit emphasis on actors ('players') who are described as
competing in the market-place. Putnam (1993ab) has been influential in his application of social capital to
macro development policy issues, some of which are used by macro world bank policy makers.
2. Even though most of Coleman's work did not deal with business organizations, it provided a lot of
inspiration to scholars of the firm who used his theories in their study of organizations.
3. Third largest was Airbus with 124 customers, fourth largest Bombardier with 107.
4. Koelewijn (1997).
5. Coleman (1990).
6. Also see Gabbay and Sato (1996)
7. During the Second World War, German soldiers were entitled to two liters of beer a week. The
Germans were not able to supply their troops that occupied The Netherlands, so they relied on Heineken
Company, a Dutch brewery, to distribute beer to the German troops. However, in order for Heineken to
distribute the yellow nectar to the Germans, the German army had to inform Heineken about the location
and movements of its troops. Heineken then provided the Dutch resistance with this information. The
business relationship with Heineken thus helped the Germans quench their thirst for beer, but also
negatively affected the German ability to protect strategically sensitive information.
8. With the term 'corporation' we refer to both for-profit and not-for-profit organizations (see, for
instance, the chapters by Ferlie and Doreian). Also, we both consider organizations that deal with tangible
combodities, and those that provide services (see, for instance, the chapters by Iacobucci, Lazega,
Sandefur et al.).
9. This finding is reported by Brass (1985a).
10. Gabbay (1997) calls this 'negative social capital.'
II. Daft (1995).
12. Gabbay and Stein (1999), studying infrastructural network projects and their effect on the Middle
East, develop the notion of 'regional social capital' inherent in country-to-country networks.
13. Van Engelen et al. (1999). Also see Amason and Schweiger (1994) and Dess and Priem (1995).
14. For an overview, see Leenders (1995b, 1999).
15. See Davis and Newstrom (1985), Hyatt (1989), Simmons (1985).
16. Exemplary is the resistance the Dean of a management school experienced when a new curriculum
was introduced that was completely different from the old one. In their fear of having to give up some of
their hobbyhorses and relinquish control over the new curriculum, a number of the professors mobilized
the social ties they had within the school. They were not fighting the introduction of a new curriculum per
se; they were merely fighting having to give up credit points offered for their personal favorite topics. The
formal organization set up for the (substantive) development of the outlines of the new curriculum only
provided these professors with limited influence. Still, at a meeting with the entire academic staff present,
one of them got up and said to the Dean 'why do you even think you can make these decisions? Let the
decisions be made where the real power is: with us!' He was (largely) right: the informal network still
maintained a lot of the power that, formally, belonged to the dean. Although the dean officially had the
authority, he would not be able to pass anything the informal network of professors wished to block.
17. See, Mintz and Schwartz (1985), Stokman et al. (1985), Useem (1979), and Pennings (1980).
18. Nohria, (1992a: 11).
19. Gabbay and Leenders (1999).
20. In some industries, contracts are occasionally signed that prohibit the leaving employee from
working with former clients for a set number of years.
21. The discussion of the interplay between social structure at the firm level and social structure at the
individual level is beyond the scope of this chapter.
22. Salancik (1995: 348).
23. We use here the most well known typology in the study of organizations. Other typologies can also
be connected to corporate social capital. For instance, see Allison's (1971) typology of organizations as
'rational' and 'political' actors, as discussed in Pennings and Lee (this volume).
24. Also see Walsh et al. (1999) on el~onic networks and Gabbay and Stein (1999) on country-to-
country networks.
25. Modem technology and the increasing use of computer mediated communication systems extend the
intuitive understanding of corporations far beyond these limited descriptions (Walsh et al. 1999).
26. See, for instance, World Bank (1997).
SECTION I
• Conceptual Issues
theory, models, and measurement
Organizational Networks
and Corporate Social Capital
1
•
David Knoke
ABSTRACT
Corporate social capital is defined as processes of forming and mobilizing social
actors' network connections within and between organizations to gain access to
other actors' resources. Following a brief overview of basic network concepts and
principles, I discuss alternative theoretical explanations for the origins, spread,
transformation, and erosion of social capital. Two sections next investigate how
network dynamics have reshaped corporate practices and changed the employment
contract between workers and their firms . In conclusion, researchers should conduct
more empirical investigations and construct better theories about the mechanisms
through which social capital networks change the fates organizations and their
participants.
INTRODUCTION
Michael Eisner, the imperious chairman of the Walt Disney Company, hired his
long-time friend Michael Ovitz to fill the media giant's presidency in August, 1995.
That position had lain vacant for months following Frank Wells' death in a
helicopter crash. Ovitz began his Hollywood career humbly, as a tour guide at
Universal Studios followed by a stint in the William Morris Agency's mailroom. It
soared after he co-founded the Creative Artists Agency in 1975, which soon grew
into the entertainment industry's premier talent agency, representing a thousand film
personalities including Tom Hanks, Barbra Streisand, and Tom Cruise. Ovitz
became the Hollywood's most-powerful and most-feared figure, personally
brokering such mega-deals as the Matsushita-MCA merger and SONY's acquisition
of Columbia Pictures. In 1995, he lured CBS news executive Harold Stringer to
head Tele-TV, a new video-program joint venture of three regional telephone

companies. Tele-TV was relying on the super-agent's connections to procure
production deals with the Hollywood studios. But Ovitz's surprise defection to
Disney, which was allied with a rival group of telecommunication companies, left
CAA scrambling to hold onto its business and entertainment clients.
Industry observers assumed that Disney hired Ovitz to strengthen its ties to top
Hollywood talent, to manage its recent $19 billion takeover of Capital Cities!ABC,
and eventually to succeed Eisner as chair: 'It provides the company with Ovitz's
wide network of contacts and skills,' said one stock analyst. 'Frank Wells was always
a sounding board for Eisner. Ovitz will playa similar role' (Wells 1995). Yet after
just fourteen acrimonious months, unable to agree with his boss on how best to run
the company, Ovitz left the Magic Kingdom 'by mutual agreement.' Denying rumors
of a personal feud, Eisner said, We have been doing business together while being
friends for many years, and I know that both our professional and personal
relationships will continue' (Van de Mark 1996). To salve his wounds, Ovitz took a
reported $90 million severance package.
Although exceptionally dramatic in its particulars, the Disney episode
exemplifies the use of social capital to achieve individual and collective goals,
especially the importance of intra- and interorganizational connections for employee
careers and corporate performances. Social actors-persons, groups, organizations,
nations--continually interact in pursuit of their interests, and through those
processes they generate complex webs of social relations that advance or thwart
their goals. Three key issues, regarding the significance of social relations as social
capital, permeate this chapter: 1) How do actors in organizations create, reproduce,
and change their network ties to other participants? 2) How do actors' positions in
multiplex network structures facilitate or constrain their acquisition of social capital
and affect personal and organizational performances? And 3) how do these network-
formation and performance-outcome processes interact and mutually shape one
another over time? That is, selection and influence may operate reciprocally:
connections may originate because well-performing organizations prefer ties to
other outstanding actors, while superior performances may result from the social
capital those network ties generate. Answering those crucial questions draws
research attention to temporal sequences of network formation and consequences for
corporate social capital at both individual and organizational levels of analysis.
The next section defines social capital as the processes by which social actors
create and mobilize their network connections within and between organizations to
gain access to other social actors' resources. Then I briefly develop some basic
network concepts and principles. I next review competing accounts of the origins,
proliferation, modification, and destruction of social capital. The following two
sections investigate how network processes are reshaping corporate practices and
changing the employment contract. The conclusion appeals for more research and
theory construction about how social capital networks alter the fates of organizations
and their participants.
Organizational Networks and Corporate Social Capital- 19
SOCIAL CAPITAL EMBEDDED IN SOCIAL RELATIONS

Following suggestions by Bourdieu (1980, 1986) and Loury (1977), James S.
Coleman defined social capital as social-structural relationships that an individual
could mobilize in actions 'making possible the achievement of certain ends that
would not be attainable in its absence' (Coleman 1990: 302; see also Coleman 1986,
1988). In contrast to other forms of capital (physical, financial, human, and cultural
capital), social capital resides neither in a person's skills and knowledge nor in such
physical instruments as tools and machines. It is embedded in social connections and
is 'created when the relations among persons change in ways that facilitate action'
(Coleman 1990: 304). Indeed, for many purposes, a network tie is itself a form of
capital. People and organizations actively shape their social relationships to obtain
better opportunities and benefits. By forging large volumes of connections to
numerous, diverse, and well-endowed contacts, a social actor gains potential access
to the assets controlled by those contacts:
Social capital is at once the resources contacts hold and the structure of contacts in a
network. The first term describes whom you reach. The second term describes how you
reach. (Burt 1992: 12)
An actor's social capital encompass many types of relations, including trust and
confidence, obligations and cooperation, and information. These 'moral resources'
ultimately depend on shared social norms that sustain and strengthen the cooperative
bonds among actors in a social system. The dynamics of social capital may parallel
other forms of capital growth, with initial investments in riskier relations
subsequently paying higher returns to the investors. For example, rural communities
and small towns are typically pervaded by extensive webs of mutual social and
economic assistance among the residents. Over time, these neighborly exchanges
generate high reservoirs of trustworthiness that outstanding obligations to
reciprocate will likely be repaid in future times of need. Persons in such
interdependent social systems benefit from greater amounts of social capital 'credit'
on which to draw, as demonstrated by heroic yet futile sandbagging efforts to
prevent the Red River from inundating North Dakota's and Minnesota's riverine
towns during the spring of 1997. Similar endeavors occur in The Netherlands,
involving periodic evacuations of thousands of residents from lowland areas.
Several social network analysts propose that people mobilize their chains of
direct and indirect social connections, particularly their 'weak ties' to higher-status
persons in stratified social systems, to accomplish personal goals such as finding
jobs and achieving upward social mobility (Granovetter 1973; Lin, Ensel, and
Vaughn 1981; Boxman, De Graaf, and Flap 1991; Lin 1995a; Bian 1997; see the
Higgins and Nohria chapter in this volume). Job-seekers obtain timely information
about openings and necessary qualifications primarily through informal
communication channels rather than from formal employment institutions (Rees
1966). An applicant's chances to secure a job or a promotion may depend less on
human capital, in the form of school credentials and examination scores signaling
achieved skills, and more on obtaining favorable assessments from key advocates
and gatekeepers who can vouch for the applicant's reliability. For example,
1
JANE DICK
BANKRIGHT SQUAREBILT
Figure 1. Schematic of a multi-level social capital network
university hiring and promotion procedures typically require letters of

recommendation from trusted assessors willing to testify about a candidate's
worthiness. Thus, people's chances for improved career outcomes are often
enhanced by their ability to tap into (,borrow') the scarce social resources held by
actors beyond their immediate social circles. Resources accessed through network
relations become an individual's social capital, regardless of whether those ties were
explicit investments made by the person in expectation of a future payoff, or were
overtly mobilized during instrumental actions (Lin 1995b). The wider the range and
the more diverse the contacts directly and indirectly available to a job-seeker, the
greater that person's chances for ultimate success. Thriving in a highly competitive
labor market favors candidates possessing heterogeneous networks over applicants
whose opportunities are restricted by their redundant, impacted networks (Burt
1992: 195; 1997).
The idea that social networks constitute an actor's social capital investments can
be readily extended to social groups and corporate entities. Just as an individual can
mobilize her personal contacts' social resources for purposive action, so can a
formal organization activate various resource networks to achieve its goals. Figure 1
illustrates a hypothetical social system involving complementary exchange networks
at macro- and micro-levels of analysis (for simplicity, I omit intermediate groups
such as teams and departments). The circles represent the formal boundaries of two
organizations: Bankright, a financial corporation, and Squarebilt, a construction
company. This example highlights two key employees: Jane, a Bankright
commercial loan officer, and Dick, Squarebilt's chief controller. Pairs of directed
lines (arrows) represent social exchange relations linking these workers and their
enterprises. Within each organization, the employees supply resources essential to
their firms' operations, specifically financial expertise and experience in loan-
making and acquisition. As returns on these human capital investments, their firms
reward Jane and Dick not only with salaries, fringe benefits, and job-training
opportunities, but also with increased social capital via their relationships with other
company employees, for example, authority, social status, and collegiality.
The two pairs of horizontal lines depict resource exchanges that span
organizational boundaries at both levels. As agents for their organizational
principals, Dick and Jane directly negotiate business loans, but apart from these
roles they may also maintain personal and professional ties, such as common school,
Organizational Networks and Corporate Social Capital - 21
church, and voluntary association affiliations. At the institutional level, the two
companies are linked primarily through their long-standing banker-client
relationship. Such financial transactions are often reinforced by solidary connections
between their chief executive officers, who belong to the same social clubs and
political parties, and whose families may even intermarry. These instrumental
networks payoff by turning one actor's social resources into the social capital
investments of other individuals and corporations. Squarebilt gains access to
Bankright's resources indirectly whenever Dick mobilizes his ties to Jane, who
possesses authority to make construction loans. Alternatively, were Jane to reject his
loan application, Dick could mobilize his connections to Squarebilt's CEO to tap
into the firm-level authority of Bankright's CEO and bring pressure to bear from
above on Jane to reverse her decision.
Actors often nurture and manipulate their social relations as deliberate strategies
for coping with uncertainties arising from dependence on external environments for
many critical resources needed to 'get the job done.' Explaining how people and
organizations actually behave requires considering factors beyond the purely arm's
length economic transactions occurring in spot-market exchanges between
anonymous buyers and sellers, where efficiency criteria are allegedly the paramount
determinants. Many economic relations are embedded within larger social, political,
and legal contexts (Granovetter 1985), which constrain participants' choices and
actions according to normative and political criteria transcending pure cost-benefit
calculations. These contexts can be conceptualized and empirically modeled as
multiple networks, connecting diverse social actors with varying interests and
resources, activated for individual and collective purposes. Hence, researchers
ignore network structures at the peril of providing incomplete insights into
organizational structures and processes.
Mixed competitive and cooperative modalities suffuse many types of network
interactions. At times actors mobilize their social capital to gain personal advantages
over their adversaries, while in other circumstances they jointly coordinate actions
for collective benefit. For example, corporate employees engage in self-serving
career strategies, seeking out mentors or networking with superiors to advance up
the promotion ladder. In contrast, new management practices and workplace designs
stress teamwork and collaborative responsibility for production, encouraging
workers to pool their skills and social capital to improve group performance.
Similarly, both modalities operate at the level of organizational strategy where plans
to achieve global corporate goals are implemented. Firms operating in the same
industry generally compete for customer loyalties and form exclusionary supplier
relations, yet they may also collaborate in strategic alliances and joint ventures with
expectations of mutual gains. An important task for network theory is to explain
under which conditions zero- and positive-sum interactions are more likely to occur.
SOME NETWORK FUNDAMENTALS

This section briefly reviews some fundamental issues facing network researchers. I
explicate these concepts and principles in a nontechnical manner, which should
suffice for understanding the chapters in this volume. Readers seeking deeper
knowledge should consult such didactic texts as Knoke and Kuklinski (1982) or
Wasserman and Faust (1994).
Identifying Actors and System Boundaries

Networks are social constructions arising from exchanges and joint activities among
participants in a social system, defined as a 'plurality of actors interacting on the
basis of a shared symbol system' (Parsons 1951 : 19). The actors in a network may be
designated at varying levels of analysis: individuals (children in a kindergarten
class); small groups (work teams on an automobile assembly line); formal
organizations (corporations in business association); coalitions (lobbying alliances);
even nations (members ofthe World Trade Organization).
Identifying a social system's boundaries, and hence its size, requires specifying
which potential members are relevant to the system's functioning. Investigators
using a nominalist strategy typically achieve conceptual closure by including all
actors possessing one or more key characteristics (Laumann, Marsden, and Prensky
1983). Nominal designations often restrict network membership to incumbents
occupying formal positions, for example, directors of Fortune 500 companies or
middle managers at Apple Computer. The alternative realist approach to boundary
specification assumes that system participants themselves can best identify who
belongs. Uncovering the network members' subjective meanings requires a
researcher to designate 'a substantively defined criterion of mutual relevance or
common orientation among a set of consequential actors' (Knoke and Laumann
1982: 256). Typically, potential network members carry out a reputational ranking
of other actors according to their importance to the system's performance. Actors
enjoying high reputations, indicating that their peers believe they must be taken into
account, are included but actors with low or no reputations are dropped because of
their marginal importance.
Total and Ego-Centric Networks

For a small social system, researchers may be able to obtain data on connections
among all system participants, comprising a total network. 'Small' in this context
means anywhere from a dozen to several hundred actors. To construct a total
network among G actors requires each actor to report on existence of ties to all other
system members, typically by checking a previously compiled name list (see
questionnaires for U.S. energy and health domains in Laumann and Knoke 1987:
401 -500).
Some social systems are either too large or too weakly connected to collect total
networks. In such instances, the only feasible alternative representative is to draw a
sample from the target population and to elicit the direct network ties of each
sampled actor (an ego-centric network). For example, a high-tech firm studied by
Burt (1992) employed 3303 managers who had few direct or indirect connections to
one another. He sampled 547 managers and ask them to describe their social and
work-related contacts. An ego-net procedure identifies a unique set of network
alters, typically through a two-step name-generator protocol. First, an informant ego
lists the most important people with whom specific types of interaction occurred.
Then, ego describes each alter's key attributes, such as age and gender, the nature
relations with each alter (e.g., closeness and frequency of contacts), and perceived
direct ties among the alters. Importantly, because the alters are not subsequently
interviewed, ego's self-reported information remains unconfirmed. The ego-net
approach is the only plausible network methodology for general population surveys
(Marsden 1987) and large samples of diverse organizations (Kalleberg, Knoke, and
Marsden 1995).
Network Relational Contents

The next task is deciding what substantive relationship to measure. Network' s
relational contents refer to some relatively homogeneous tie among actors.
Relational contents seem to fall into two general categories: I) A transaction
involves exchanges where one actor yields rights of control over some physical
commodity or intangible value to another actor, possibly in expectation of eventual
reciprocation; 2) Joint actions require actors to co-participate in an event located in
specific time and space, without relinquishing control over resources. A minimal
joint action is mere co-presence, for example, software firms attending the annual
Las Vegas COMDEX trade fair. More intense joint activity involves coordinated
efforts to achieve common goals, for example, lobbying the Congress on legislative
proposals (Knoke et al. 1996).
Deciding which types of network content to operationalize should be guided by
a project's theoretical objectives: what substantive relations are useful for
understanding important actor and system behaviors? In practice, multiplex ties link
social actors into complex webs, meaning that no such creature as 'the network'
exists for any system. Typically, more than one relational content may be studied in
a given project.
The following typology roughly classifies relations by increasingly formality:
Recognition
Minimal awareness, with actors reporting whether they 'know about' or 'have ever
heard of others in the system. Because recognition is typically unreciprocated,
mutuality in choices is rare, thus differentiating the 'stars' who enjoy high visibility
from their anonymous fans in sports, politics, and science networks. Among
organizations, corporate reputations for quality products and services similarly
differentiate the well-regarded from the invisible players (Fombrun 1996).
Co-Attendance
Common presence at the same events, or membership in the same collectivities,
disregarding direct interaction. Mass public assemblies, such as political rallies and
athletic contests, anchor one end of a continuum, while co-participation in restricted-
access enterprises, such as private schools and social clubs, implies the existence of
potentially cohesive social classes (Domhoff 1975).
Information Exchange
Routine and regular communication of data, whether about scientific-technical
matters (a supplier's current catalog of available products and prices) or socio-
political affairs (claimant organizations' positions on legislative bills). Such
information may be widely broadcast through press releases and email-server
networks, or more narrowly targted on specific recipients, for example, a marketing

unit providing the sales department with consumer survey results.
Advising
Confidential information intended to help a recipient gain advantage over
competitors. An advisor transmits her superior knowledge and experience to an
advisee, sometimes with no expectations beyond gratitude and deference.
Organizational superiors mentor their favored underlings (Keele 1986; Noe 1988b),
expert consultants familiarize their organizational clients with R&D opportunities,
and venture capitalists scrutinize start-up investment prospects. (See the chapter by
Freeman in this volume.)
Trust
Confidence in the dependability of another actor's promises, reducing the chances of
opportunistic behavior by one's partners. Trust relations run a risk that such reliance
will prove ill-founded, but their efficiency in reducing transaction costs is an
important precursor for building long-term exchanges between people and
organizations. (See the chapter by Nooteboom in this volume.)
Support
Expressions of sympathy, empathy, or commiseration in times of need, including
instrumental actions that demonstrate solidarity with an afflicted actor. At the
individual level, friendship and kinship bonds comprise the most obvious varieties
of support. Support relations among organizations involve public legitimation, for
example, testimonials regarding one company's willingness to rally behind another
enterprise confronting political or legal difficulties.
Financial Aid
Transfers of money, credit, or physical facilities, other than arm's-length market
purchases of goods or services. Some financial ties seem altruistic, as in
corporations' philanthropic donations to charities (Galaskiewicz 1985), while other
exchanges attach implicit quid pro quo strings, such as corporate contributions to
parties and candidates via political action committees (Mizruchi 1992). Other
financial exchanges entail explicit ownership and control connections, such as the
debt and equity transactions linking corporate and banking members of Japan's
famous keiretsu (Gerlach 1992a).
Authority
Legitimate power to expect that commands will be obeyed, backed up by sanctions
for failures to comply. Employment relations in corporate and public bureaucracies
typify intraorganizational authority networks, while a regulatory agency's power to
constrain an industry embodies interorganizational ties. The interlocking directorate,
connecting corporations and banks (Mintz and Schwartz 1985; Stokman, Ziegler,
and Scott 1985) and nonprofit organizations (Useem 1979), may reflect hybrid
authority patterns where agents of outside organizations help set policies binding on
participants inside another organization.
Alliances
Collaborative arrangements involving two or more organizations that combine
resources to pursue common or complementary objectives, often involving
applications of uncertain technologies or entry into risky markets. These cooperative
strategies vary in the substantive content of interorganizational cooperation and
types of governance structures designed to monitor and control the partners'

behaviors. (See the chapter by Stuart in this volume.)
The nine types of ties classified above are illustrative, but not exhaustive, of the
diverse relations that network analysts may find useful for their theoretical interests.
Magnitude of Ties
Network informants should also indicate the magnitude or value of their relations.
At a minimum, only a dichotomous coding-presence (1) or absence (O)-might be
recorded, for example, using name list checkoff. The unchecked alters yield
important data, since a network's structure depends as much on its gaps as on its
direct connections. More detailed magnitude codings assign scalar values reflecting
each tie's relative strength. For persons, tie strength usually refers to subjective
intensity of commitment, for example, a friendship study requesting egos to indicate
which people are their 'casual,' 'close,' or 'best friends' (Leenders 1996). For
organizations, relational magnitudes may involve objective data, for example, the
dollar amounts of loans from commercial banks to manufacturing firms, or the
numbers officers sitting on other companies' boards of directors.
Another important consideration is the time span observed: too short and
important but infrequent relations may be overlooked; too long and dormant ties
might mistakenly be treated as current. Unfortunately, the temporal dimension
hasn't been well-integrated into research procedures. Most network projects yield
static snapshots of a long-established network, without revealing their origins,
evolution, and ultimate fates. For example, we know little about whether informal
ties between employees of different companies subsequently generate organizational
alliances, or whether the opposite causal process occurs. Despite evident theoretical
payoffs from understanding network dynamics, data collection has not kept pace
with recent methods for investigating network changes over time (Wasserman and
Iacobucci 1988; Frank 1991; Zeggelink 1994; Snijders 1996; Leenders 1996).
Network Forms
Basic network forms describe the patterns connecting system actors regardless of
their specific relational contents. Figure 2 displays a hypothetical chooser-by-chosen
binary adjacency matrix and its associated graph. Think of the {ABCD} subset as a
production department located in one building, while the {WXYZ} subset is a
geographically dispersed salesforce. Actors A and W are these units' respective
heads. The relational content is regular communications about work. Each matrix
row represents a potential sender and each column a receiver of dichotomous social
ties. A 'I' entry indicates that the row actor communicated with the column actor,
while '0' means that no communication occurred. Graphs depict actors as labeled
points and their relations as arrows pointing from sending actors to receiving targets.
Because every communication tie is reciprocal, all ten arrows are doubled-headed.
The network volume is the total number of ties and its density is the proportion
of observed ties to the number of possible connections not counting self-ties (i - g
for a g-actor system). The example has a volume of 20 ties and a density of (20/56)
= .357. Actor connectedness counts the number of non-zero entries in a matrix row
PRODUCTION DEPARTMENT SALES FORCE
.y
X
/' /'
c A ... • W I(
'" D
A
A 0
B 1
B
0
1
C
1
1
D W
1
1
1
0
X
0
0
Y
0
0
'"
Z
0
0
Z
C 1 1 0 1 0 0 0 0
D 1 1 1 0 0 0 0 0
W 1 0 0 0 0 1 1 1
X 0 0 0 0 1 0 0 0
y 0 0 0 0 1 0 0 0
Z 0 0 0 0 1 0 0 0
Figure 2. Graph and matrix representations of a hypothetical eight-actor network
(out-degrees) and column (in-degrees). Thus, department heads A and W are the
best-connected actors, with out- and in-degrees = 4. Next come B, C, and D with
three ties each, trailed by the dispersed salespeople with just one connection to their
boss. Successively mUltiplying a matrix by itself reveals the minimal path length
required to connect pairs of actors. Visually, a path can be traced across directed
arrows between pairs, with the length being the number of steps needed to connect
that dyad. For Z to pass a message to C requires a path of length = 3: (ZW) + (WA)
+ (AC). The four production members are connected by one-step paths (direct ties),
but the salespeople require 2-paths to reach one another.
A and W enjoy unique and powerful roles in the system, an insight confirmed
by measures of actor centrality (see Freeman 1977, 1979). Basically, a central actor
participates in a large volume of social relations, with refined centrality concepts
differentiating among the type or 'quality' of connections. The simplest centrality
measures is an actor's in-degrees, measuring the sheer volume of ego-centric
contacts received from alters. Closeness centrality captures the extent to which ego
maintains connections to many alters who themselves have many non-overlapping
ties, thus enabling ego to reach many others by relatively short paths. A and W each
have the highest closeness scores (70), while B, C, and D enjoy somewhat higher
closeness (50 each) than the less-connected X, Y, and Z (43.8 each). Betweenness
centrality reflects an actor's ability to mediate many connections between
subgroups, thereby potentially leveraging greater impact on system activities.
Because the salespeople are less connected than are the production employees, W
has a higher betweenness score than does A (15 to 12), while the other actors'
betweenness centrality scores are O.
The final network forms cluster actors into positions, which simplifies the
system's social role structure. Two basic approaches involve cohesion and
equivalence criteria. A clique is a network sub-set in which all dyads are maximally
connected (all reciprocal direct ties occur, yielding a density of 1.00). By this
rigorous definition, only the {ABCD} cluster comprises a genuine clique. Two
actors are structurally equivalent to the extent that they display identical or very
similar patterns of ties to all other alters, regardless of their ties to one another
(Sailer 1978). For example, firms in an industry that buy from the same sources and
sell to the same customers are fundamentally interchangeable competitors from the
market's perspective. The four equivalent blocks are {A}, {W}, {BCD}, and
{XYZ}. Automorphic equivalence identifies actors i and j are automorphically
equivalent if, after removing the 'names' of the actors from the nodes, nodes i and j
are impossible to distinguish. In the example, the sets of automorphic ally equivalent
actors are {AW}, {BCD}, and {XYZ}. In contrast to structural equivalence, which
puts A and W into separate positions because they supervise different individuals,
they are automorphically equivalent because they are connected to corresponding
others-their work-unit subordinates.
INTERORGANIZATIONAL NETWORKS
By the 1980s, converging environmental pressures began restructuring forever the
ways organizations would relate to their competitors, employees, customers, and the
larger society. While every analyst offers a favor~te list of key factors driving
organizational change, the six master trends cited by the Hay Group (Flannery,
Hofrichter and Platten 1996) seem particularly concise and comprehensive: rapidly
expanding technologies; growing global competition; increased demand for
individual and organizational competencies and capabilities; higher customer
expectations; ever-decreasing cycle times; and changing skilled personnel
requirements. A seventh trend, at least in the U.S., is increased investor pressure on
companies to improve their short- and long-term financial performances (Useem
1996). As corporations grew increasingly exposed to international competition, they
sought new ways to remain viable by slashing costs and prices, improving
production performance, and responding rapidly to technological innovations and
fickle consumer preferences. With consumer demand simultaneously globalized and
fragmented, niche markets for specialty goods and services supplanted cumbersome
mass-production systems run by 'Fordist' principles. Firms perceived performance
gains from unbundling their internal hierarchical structures and deinstitutionalizing
the product-unrelated conglomerate form (Davis, Diekmann, and Tinsley 1994).
These incessant pressures to achieve corporate flexibility and specialization drove
organizations to restructure their internal employment systems along much more
participatory lines. They also compelled companies to reach outside their traditional
boundaries to form long-term collaborative relationships enabling them to stay afloat
in an increasingly cutthroat world economy.
One consequence of the strategic search for new competitive advantages was the
proliferation of many new interorganizational forms . Figure 3 presents an alliance
typology, modified primarily after Yoshino and Rangan (1995 : 8). At one extreme
are pure market relations, whose transactions require no enduring collaboration by
exchanging parties. At the other extreme are hierarchical arrangements in which one
firm assumes full authoritative control over the other, absorbing the participants into
a unitary enterprise. Between these extremes fall various 'hybrid' arrangements that
are neither clearly markets nor hierarchies but typically blend elements from both
types (Jensen and Meckling 1976; Williamson 1975; Powell 1987; Heydebrand
1989). An appropriate label might be the N-form or 'networked' organization, to
emphasize that relations are central to these mixed structures. For Yoshino and
Rangan (1995: 5), a strategic alliance's critical characteristics are partner firms that:
1) remain independent after the alliance is formed; 2) share benefits and managerial
control over the performance of assigned tasks; and 3) make continuing
contributions in one or more strategic areas, such as technology or products.
Based on this definition, they classified licensing and franchising as traditional
market contracts because one company grants another the right to use patented
technology or production processes in return for royalty payments. However, Figure
3 reassigns franchising under non-equity partnerships, since many distribution
Hierarchical relations Subsidiaries

Acquisitions
Mergers
Equity partnerships Minority equity investments

Equity swaps
Joint ventures
Business groups
Non-equity partnerships Franchising

Small firm networks
Joint R&D, production development, manufacturing,
marketing, distribution, service & other functions
Multi-participant alliances Data banks and information clearinghouses

Standards-setting consortia
Government-sponsored R&D consortia
Trade associations
Cooperatives for purchasing, marketing
Action sets for lobbying campaigns
Cartels
Market relations Spot exchanges

Arm's length buy-sell
Short-term subcontracting
Licensing
Figure 3. Types of interorganizational alliances
Modified after Yoshino and Rangan (1995)
franchisers such as McDonald's Corp. exercise centralized coordination to safeguard

their corporate interests while leaving ownership and operations to the local
entrepreneur (Reve 1990: 148; also Osborn and Baughn 1990).
While some alliances require only a bilateral (dyadic) relations, other forms
involve mUltiple participants structured into complex 'alliance networks' (Gomes-
Casseres 1996: 52). Resource and authority commitments also vary considerably
among the hybrid types, particularly the extent of equity exchanges among partners.
Space limitations prevent detailed discussions of each interorganizational form, but
brief distinctions among the most important types may be helpful. An action set is a
short-lived organizational coalition whose members coordinate their efforts to
influence public policy decisions, for example, the passage of legislative acts
affecting the coalition members' collective interests (Knoke et al. 1996: 21). Cartels,
or pools, are unstable alliances formed to constrain competition by cooperatively
controlling production and/or prices in specific industries (Fligstein 1990: 39). The
'trusts' flourishing in late 19th-century America - in railways, heating oil, steel,
aluminum, sugar, salt-were outlawed but cartels periodically arise elsewhere, for
example, OPEC.
Small-firm networks (SFNs), are a late 20th century innovation involving large
numbers of very small firms (often with fewer than 10 employees) interacting on a
long-term basis, 'sharing information, equipment, personnel, and orders, even as
they compete with one another' (Perrow 1992: 455). They are complex clusters of
raw materials suppliers, producers, financial service, marketing, and distribution
firms. SFNs arise primarily in clothing, toys, publishing, motion pictures,
construction, light machinery, and electronics industries rather than in heavy
manufacturing or extractive industries. For example, Bennetton, an Italian apparel
producer, owns very few facilities but parcels out almost all production tasks to
hundreds of firms employing thousands of workers (Clegg 1990: 120-125; Kanter,
Stein, and Jick 1992: 228). The regional economies of Emilia Romagna in northern
Italy (Brusco 1982; Lazerson 1988) and Baden-Wiirttemberg in southwestern
Germany (Herrigel 1996) are the most famous SFN exemplars, as is perhaps Silicon
Valley in northern California (Saxenian 1994). SFNs should not be confused with
the more prevalent small supplier networks dominated by a large industrial firm
such occur in Germany's Ruhr Valley (e.g., Grabher 1993) and the American and
Japanese automobile industries (Womack, Jones and Roos 1990).
Four main types of equity partnerships occur. In a minority equity investment,
one firm buys an interest in another by direct investment, perhaps only a 3-5% stake,
'although the active involvement of the management of the partner-company is
retained and the assessment of expertise of the company can be made without
complete integration' (Hagedoorn 1993a: 132). Examples abound in high-tech fields
where large corporations use minority shareholding to access start-up firms'
technologies. An equity swap involves mutual direct investments of partners. A joint
venture occurs when 'two or more legally distinct firms (the parents) pool a portion
of their resources within a jointly owned legal organization' (Inkpen 1995: 1) that
serves a limited purpose for its parents. For example, the well-known NUMMI
automobile plant in Fremont, California, enabled General Motors to learn about
Japanese management techniques while Toyota gained a U.S. foothold (Adler 1993).
Joint ventures may involve 50:50 ownership between two parents (Lewis 1990: 173-
192) or unequal equity shares among mUltiple partners.
Finally, a business group is a coherent collection of firms bound together at an
'intermediate' level between short-term strategic alliances and the unitary
corporation (Granovetter 1994: 454). East Asian partnerships among manufacturers,
suppliers, and financial institutions-such as the Japanese keiretsu (Lincoln, Gerlach
and Takahashi 1992; Gerlach 1992a) and Korean chaebol (Steers, Shin and Ungson
1989)-exemplify business groups spanning multiple fields that are integrated
through complex debt, interlocking directorates, and equity ownership patterns.
Although some observers argue that Chrysler Corp.'s comeback allegedly resulted
from an American weak-tie version of keiretsu, its cooperative relations with parts
suppliers involve neither the equity investments nor the management connections
that Toyota and Nissan have with their suppliers (Dyer 1996b). (See the chapter by
Pennings and Lee in this volume.)
The following subsections examine explanations of three aspects of
interorganizational alliances: why they occur, how they develop, and their
consequences. In seeking to understand network development over time, we should
ask whether any single theory can account for such diverse alliance phenomena or
whether distinct theories are required?
Theories of Alliance Formation

Two prominent theoretical explanations of why organizations engage in nonmarket
relations are transaction cost analysis and resource dependence theories, respectively
emphasizing economic and socio-political factors. Transaction costs determine
where to draw an economically efficient boundary between an organization
(hierarchy) and its environment (market), in other words, whether to make or buy a
particular function. Oliver Williamson (1975) argued that the most important factor
driving organizational efforts to economize is asset specificity, the extent to which
investments are specialized to particular recurrent transactions between buyers and
sellers. The greater the specificity, the more likely are the parties to 'make special
efforts to design exchanges with good continuity properties' (Williamson 1981:
555), thus effectively locking them both into prolonged bilateral exchanges. For
example, a corporation requiring only intermittent legal advice is likely to retain an
outside legal firm, while a company with persistent legal problems may create its
own in-house legal department. Thus, interorganizational ties arise from specialized
investments that would lose their value if transferred to another exchange partner.
Otherwise, market exchanges will be more cost-efficient. A second core assumption
of transaction cost analysis is that at least some actors are 'given to opportunism'
(Williamson 1981: 553), that is, dishonesty and dissembling about preferences and
information. Employing a different terminology, Williamson argues that
interorganizational relationships that convey social capital, are at risk of being
turned into social liability by opportunistic actions of one of the parties. The
necessity to monitor partners' performance and safeguard against duplicity can
increase interorganizational transactions costs. For example, Seagram's Universal
Studios successfully sued Viacom Inc., claiming the latter's launching of a
competing cable TV network violated their agreement on joint operation of the USA
Network (Shapiro 1997).
Resource dependence explanations of alliance formation emphasize inherent
tensions between organizational resource procurement needs and the desire to
preserve freedom of strategic decision making. Intercorporate relations arise from
interdependencies and constraints among organizations: situations where one
organization controls the critical resources or capabilities-such as money,
information, production and distribution skills, access to foreign markets-needed
by another organization. Alliances tend to occur more often among interdependent
than between independent firms , that is, where complementarity rather than
similarity prevails. However, organizational efforts to manage problematic external
interdependencies 'are inevitably never completely successful and produce new
patterns of dependence and interdependence' (Pfeffer 1987: 27). Dependence
theorists argue that network ties arise from managers' efforts to control the most
troublesome environmental contingencies through complete or partial absorption
(e.g., mergers or joint ventures).
In their drive to acquire critical resources from network partners, organizations
risk losing control of their own destinies (social liability). Resource dependence
generates interorganizational power differentials that constrain firms' opportunities,
since organizations tend to comply with demands from the more powerful actors in
their environment. 'Organizations seek to form that type of interorganizational
exchange relationship which involves the least cost to the organization in loss of
autonomy and power' (Cook 1977: 74). Given an opportunity set of potential
alternative providers, a company will optimally choose a partner that can best satisfy
its resource needs while imposing minimal constraints on its discretionary actions.
For example, confronted with many suppliers capable of providing equivalent-
quality inputs, a large manufacturing firm is likely to purchase from the smallest
supplier, thereby gaining power to impose terms and conditions. Similarly, a small
supplier would prefer to spread its business across many customers, thereby
avoiding the loss-of-control stemming from dependence on a single partner.
Few analysts have explicitly tested hypotheses about alliance formation drawn
explicitly from the transaction cost or resource dependence perspectives. Pfeffer and
Nowak (1976) found that resource interdependencies (high exchange of sales and
purchases) among companies in technologically intensive industries significantly
increased joint venturing at the industry-level of analysis. Zaheer and Venkatraman
(1995) tested hypotheses about interorganizational strategies drawn from transaction
cost economics and social exchange perspectives, using data from a mail survey of
329 independent insurance agencies. Their two dependent variables were 'vertical
quasi-integration' (the percent of total premiums handled by an agency's 'focal
carrier,' the company with which an agency conducted most of its business) and
Joint action' (a multi-item scale measuring planning and forecasting activities with
the focal carrier). Although transaction-specific assets predicted quasi-integration,
neither uncertainty nor reciprocal investments were statistically significant. Instead,
quasi-integration and joint action were both positively related to mutual trust
between agency and carrier, a relationship opposite to the transaction cost
hypotheses but consistent with social exchange theory.
Resource dependence principles seemed more helpful than transaction cost

concepts for understanding cooperative networks between new biotechnology firms
(NBFs) and established corporations in the 1980s (Barley, Freeman, and Hybels
1992; Kogut, Shan, and Walker 1992; Powell and Brantley 1992; see the chapters by
Smith-Doerr et al. and Stuart in this volume). Complementary resource needs drove
strategic alliances, primarily involving exchanges of financial support for technical
expertise. The small, innovative R&D laboratories typically lacked funds, public
legitimacy, and in-house capability to market their products and maneuver through
the regulatory maze. Hence they allied with diversified, resource-rich
pharmaceutical, chemical, and agricultural companies able to provide sustaining
resources. In turn, these established firms welcomed collaborative agreements as
means to acquire tacit knowledge and to learn new technological skills from their
NBF partners (social capital). As relationships accumulated and stabilized over time,
the network positions occupied by individual organizations constrained their access
to information regarding potential alliance partners (social liability). 'It is the
structure of the network, rather than attributes of the firm, that plays an increasingly
important role in the choice to cooperate' (Kogut, Shan, and Walker 1992: 364).
Two studies of changing networks patterns in other fields underscore the
importance of past ties on future actions. Leenders (1995b) reanalyzed dyadic data
from the social service networks of two Pennsylvania counties between 1988-90.
Informants named the organizations with whom their agencies maintained relations,
such as coordinating client treatments or sharing funds and personnel, including ties
mandated by the state. In both counties, estimated dyad-transition models revealed
that 'reciprocity both increases actors' inclination of creating and maintaining ties
and decreases the inclination of withdrawing ties' (Leenders 1995b: 193). Although
he did not use the term corporate social capital, the evolution of these
interorganizational networks clearly fits such an interpretation.
In a study of dyadic international corporate alliances, Gulati (1995a) found
evidence consistent with both resource dependence (which he called 'strategic inter-
dependence') and social structural explanations. Using a 1980-89 panel of 166
corporations operating in three worldwide sectors (U.S., Japanese, and European
new materials, industrial automation, and automotive products firms), he conducted
event-history analyses on a variety of dyadic alliances ranging from arms-length
licensing agreements to 'closely intertwined equity joint ventures' (1995a: 634).
Strategically interdependent firms (i.e., those companies operating in comple-
mentary market niches) formed alliances more often than did firms possessing
similar resources and capabilities. Previously allied firms were more likely to
engage in subsequent partnerships, suggesting that 'over time, each firm acquires
more information and builds greater confidence in the partnering firm' (l995a: 644).
Beyond a certain point, additional alliances reduced the likelihood of future ties,
perhaps prompted by fears of losing autonomy by becoming overly dependent on a
single partner. Indirect connections within the social network of prior alliances also
shaped the alliance formation process: previously unconnected firms were more
likely to ally if both were tied to a common third-party, but their chances of
partnering diminished with greater path distances. Gulati concluded that 'the social
network of indirect ties is an effective referral mechanism for bringing firms
together and that dense co-location in an alliance network enhances mutual

confidence as firms become aware of the possible negative reputational
consequences of their own or others' opportunistic behavior' (1995a: 644). His
results reflected a logic of clique-like cohesion rather than status-competition among
structurally equivalent organizations.
Trust as Corporate Social Capital

The formation of successful strategic alliances between corporations hinges on
creating and sustaining relationships among the partners based on mutual trust. At
the individual level, we consider a person trustworthy if 'the probability that he will
perform an action that is beneficial or at least not detrimental to us is high enough
for us to consider engaging in some form of cooperation with him' (Gambetta 1988c:
217). At the interorganizational level, trust provides a foundation for one firm to
achieve some degree of social control over another's behavior under conditions of
high uncertainty. From a transaction cost perspective, the social capital of trust
expectations may provide an efficient mutual deterrent to both partners' temptation
to opportunism or malfeasance, thereby reducing alliance costs relative to more
formal control mechanisms such as written contracts (Gulati 1995a: 88-91). Hence,
interfirm trust relations fall conceptually somewhere between the polar logics of
hierarchical authority and market price relations (Bradach and Eccles 1989: 104;
Sako 1991).
The business-risk view of trust stresses confidence in the predictability of one's
expectations hedged by formal contractual means such as insurance (Luhmann
1979). Ring and Van de Ven (1994) emphasized an alternative psychological
conceptualization of trust as confidence in another's goodwill, of faith in the
partner's moral integrity. In their approach, trust constitutes a fundamental type of
organizational social capital, a strong-tie relationship between an ego firm and the
alters comprising its organizational field . Organization attributes and network
relations interact over time. As a company builds a reputation among its peers for
fair dealing and impeccable reliability in keeping its promises, that reputation itself
becomes a prized asset useful for sustaining its current alliances and forming future
ones. Reputed trustworthiness signals to potential partners that an organization is
unlikely to act opportunisticly because 'such behavior would destroy his or her
reputation, thus making the total outcome of the opportunistic behavior undesirable'
(Jarillo 1988: 37).
The social psychological explanation of trust is rooted in basic social exchange
principles, including conformity to such norms as reciprocity, commitment,
forbearance, cooperation, and obligations to repay debts (Stinchcombe 1986;
Bradach and Eccles 1989: 105; Lewis and Weigert 1985). Because typical interfirm
transactions are widely separated across time, trust reinforces these ties by invoking
such principles as that exchange values should balance over the long run, and that
each partners' payoffs should be roughly proportional to their contributions to any
joint enterprise. As trust relations became historically institutionalized in modem
industrial societies (Zucker 1986), initial arms-length market transactions grew
increasingly suffused with many normative connotations, generating and upholding
STRucrURAL
CONDITIONS
TRUST ) ALLIANCE
FORMS
COMMUNICATION
NETWORK
Figure 4. Trust is an intervening factor in the alliance formation process
the moral communities within which trustworthiness conveyed great importance in

members' decisions whether to continue or break off relations.
What are the macro-level sources of trust among organization? Figure 4
proposes that interorganizational alliances emerge over time with trust occupying a
pivotal role between antecedent conditions and consequent alliance formations. Note
the feedback loop in which trust shapes the form of alliance, while events occurring
during the alliance may subsequently transform the interorganizational trust
relations, either reinforcing or weakening each partner's belief the other's
trustworthiness. Thus, trust and alliance relations mutually change one another as
interactions accumulate over time.
As suggested in Figure 4, communication networks structure an organization's
ability to screen and evaluate initial information about potential alliance partners.
These exchanges involve factual data about alters' interests and competencies, but
also provide indirect evidence about other organizations' trustworthiness via path
connections to knowledgeable peers in an organizational field. The more central an
organization's position within a field's communication network, the greater its
visibility and hence more informants are available to testify regarding its reliability
and integrity. Organizations located in peripheral positions have fewer opportunities
to become familiar with potential alliance partners and for their own trustworthiness
reputations to become vetted by the field.
A second set of antecedent factors fostering or thwarting trustworthiness are
macro-structural conditions. Imbalances in the resources controlled by each
organization (such as their financial size or market shares) may impede trust creation
because of unequal partners' inability to satisfy their reciprocity obligations. Pairs of
organizations that share similar or complementary characteristics are more likely to
develop strong trust relations. Tacit understandings and taken-for-granted
assumptions may be rudely violated when partners have little in common. For
example, many cross-border alliances, undertaken between foreign partners to gain
access to local markets, are fraught with pitfalls stemming from incompatible
national cultures (Lewis 1990: 253-278; Lorange and Roos 1992: 177-204; Bleeke
and Ernst 1993: 12-13; Gilroy 1993). Even domestic alliances can suffer from
clashing corporate cultures. A major instance was the office-network software
producer Novell Inc.'s disastrous effort to integrate its subsidiary WordPerfect

Corp.'s 'close-knit and insular' staff with the parent organization's profit-driven style
(Clark 1996). After two years of plummeting market share and stock prices, Novell
sold WordPerfect to Corel Corp. at one-tenth its original $1.4 billion acquisition
price.
The feedback loop between trust and alliance depicted in Figure 4 implies a
temporal dynamic to changing governance forms through accumulating
interorganizational experiences (Smith et al. 1995). Many alliances begin with
formal linkages that expose the partners only to small risks. Because the
organizations as yet have few bases for trusting one another, equity-based contracts
(,hostage-taking') predominate as legal protection against potential opportunism. But
after partners gain confidence in one another through repeated testing, then 'informal
psychological contracts increasingly compensate or substitute for formal contractual
safeguards as reliance on trust among parties increases over time' (Ring and Van de
Ven 1994: 105). This substitution process is succinctly summarized in Gulati's
(1995b) affirmative answer to the question 'does familiarity breed trust?' Because
strong-tie trust relations can counteract firms' fears of the partner's betrayal of
confidence, governing alliances through legal documents yields to relations
governed by interorganizational trust. Reduced transaction and monitoring costs
make informal social control the preferred cost-effective alternative to both market
pricing and hierarchical authority. Consistent with these expectations, Gulati's
(1995b) analysis of multi-sector alliances found strong evidence that formal equity-
sharing agreements decreased with the existence and frequency of prior ties to a
partner. Domestic alliances less often involved equity mechanisms than did
international alliances, supporting claims that trust relations are more difficult to
sustain cross-culturally.
Another crucial developmental issue concerns the relative agency of
organizations versus individuals qua persona in creating interorganizational trust. In
general terms, network analysis needs to resolve its quandaries about the role of
human agency in social action, 'the capacity of socially embedded actors to
appropriate, reproduce, and, potentially, to innovate upon receive cultural categories
and conditions of action in accordance with their personal and collective ideals,
interests, and commitments' (Emirbayer and Goodwin 1994: 1442). Applied to the
present context, a central question is whether trust relations operate at the
organizational level, or whether trust encapsulates purely interpersonal phenomena?
As noted above, some theorists emphasize that trust originates in the social
psychology of interpersonal interactions, and thus often evokes strong emotional
overtones of sharing and caring for the welfare of one's partner (McAllister 1995).
As the employees who occupy key boundary-spanning roles try to cope with their
organizations' environmental uncertainties, they socially construct strong bonds of
mutual confidence and trust with their counterparts in other organizations that may
affect interorganizational behavior. For example, a study of company decisions to
switch auditing firms found that the individual attachments of such boundary-
spanners as the company's chief executive, financial, and accounting officers
attenuated the pressures arising from changing resource needs (Seabright et al.
1992). If only people can manifest beliefs and emotional attachments, then trust may
reside wholly within the individual fiduciaries who establish and nurture trust
relations on behalf of the organizations they represent. The potential for
intermingling the reputational social capital of people and organizations spawns
some knotty dilemmas for intraorganizational control: exactly who legally and
morally owns the trust relations in which both employers and employees have
invested? This question is not a trivial concern for firms, as reflected by such
practices as 'noncompete' clauses restricting local television news personalities from
working for rival stations after severing their employment ties, and in law-suits
against lawyers and talent agents who defect to rival firms, taking along their client
lists (Tevlin 1997). In the most extreme instances of trust violation, agents may
pilfer major corporate secrets for their new employers, as in Jose Ignacio Lopez's
alleged transfer of General Motors procurement data to Volkswagen.
Alliances Outcomes
The belief that interorganizational networks offer corporate social capital in the form
of performance benefits superior to both markets and hierarchies is widespread
among social scientists and corporate managers. Networks are allegedly 'lighter on
their feet' than hierarchies (Powell 1990: 303). They enable organizations and their
agents to respond rapidly to emerging contingencies, particularly gaining timely
access to swiftly changing technological knowledge and data essential for survival
and prosperity. Yet the evidentiary basis for such claims remains remarkably slim.
Researchers have proposed numerous criteria for judging alliance 'success,' ranging
from mere organizational survival to economic performance levels above industry
norms. One difficulty in assessing performance outcomes is that most alliances
explicitly seek only limited purposes and are intentionally short lived, so duration
alone may be an inappropriate yardstick. When an alliance terminates in one
partner's acquisition ofthe other, as in the majority of cases (Bleeke and Ernst 1993:
18), does that outcome constitute a failure of the alliance? A success for one
organization but a failure for the other?
Embeddedness in interorganizational alliances seems to contribute to
participants' survival chances compared to organizations engaging only in arm's-
length market transactions. Uzzi (1996a) used both ethnographic and quantitative
methods to study the impact on firm failure of the mUltiple network ties among 23
New York better dress apparel ftrms. 'Social capital embedded ness' indicated
whether a contractor had a network tie to a business group, typically formed around
CEOs who were kin or colleagues from previous jobs. Other measures involved the
proportion of work exchanged between organizations and the degree to which the
ego firm maintained arm's-length or embedded ties with partners. Uzzi's logit
analyses showed that 'firms that connect to their networks have greater chances of
survival than do firms that connect to their networks via arm's-length ties' (1996a:
694). But optimal networks were a mix of both types of relations:
A crucial implication is that embedded networks offer a competitive form of organizing
but possess their own pitfalls because an actor's adaptive capacity is determined by a
web ofties, some of which lie beyond his or her direct influence. Thus a firm's structural
location, although not fully constraining, can significantly blind it to the important effects
of the larger network structure, namely its contacts' contacts. (Vzzi 1996a: 694)
Organizations enter alliances with many motives and strategic objectives, including:
speed of entry into new product or geographic markets; faster cycle times in
developing or commercializing new products; improved product or service quality;
gaining technical skills, tacit knowledge and competencies; sharing costs; spreading
risks and uncertainties; monitoring environmental changes. Bleeke and Ernst (1993)
relied on unpublished reports and interviews with insiders of 150 top companies in
the U.S., Europe and Japan to determine that, in 49 cross-border alliances, 51 % were
successful for both partners while 33% were mutual failures. Alliances were 'more
effective for edging into related business or new geographic markets' (1993: 18)
while acquisitions worked better for core businesses and existing areas. Other
conditions leading to success included alliances between equally strong partners,
evenly split financial ownership of the joint venture, and autonomy and flexibility
for the joint venture to grow beyond the parents' initial expectations and objectives.
Empirical evidence regarding the financial outcomes of strategic alliances is
scarce, with network studies of investment banking and the stock exchange a notable
exception (Eccles and Crane 1988; Baker 1990; Podolny 1993). For example Chung
(1996), analyzing cooperative exchanges among 98 top investment banks involved
in new stock issues in the 1980s, found that the best long-term performers (measured
by amounts underwritten) were involved in a strategy of exchange initiation, which
also led to subsequently higher popularity and expanded participation in stock deals.
However, few researchers have studied whether joint venture partners recover their
capital investments, or whether such collaborations yield a higher returns than
available from alternative resource allocations. Theorists tend to emphasize only the
social capital emerging from networks, while ignoring potential social liability
inherent in interorganizational relations, specifically that social embeddedness may
exert a drag on market efficiency. For example, Sako (1991: 239) speculated that a
major disadvantage of obligational contractual relations is '[r]igidity in changing
order levels and trading partners [and] potential lack of market stimulus.' Similarly,
the impact of trust on alliance success remains uninvestigated. Trust presumably
fosters goal attainment by facilitating the favorable resolution of conflicts inevitably
cropping up during joint operations. Given its subjective basis, high mutual trust is
likely to correlate with feelings of satisfaction about the partner's performance and
contributions. Researchers might inquire whether collaborators feel their venture is
worthwhile and whether they would repeat the alliance for other purposes or to
recommend their partner to other firms seeking to form strategic ventures. On the
negative side, trust and other obligational norms may attach organizations too
strongly to their partners, carrying relations beyond rationally efficient limits by
resisting swift dissolution of inefficient or inequitable situations. Clearly many
opportunities await for imaginative research on the outcomes of interorganizational
alliances. .
INTRAORGANIZATIONAL NETWORKS
The macro-change forces noted above that reshaped interorganizational relations
also wreaked enormous transformations inside factories, offices, and clinics. During
a prolonged and painful decade of downsizing, reengineering, and restructuring
exertions, more daring or desperate corporations implemented flexible new designs.
Rigid bureaucratic hierarchies yielded to experiments in cross-functional teams that

devolved increasing volumes of information, technical skills, and managerial
responsibilities down to the front-line worker level (Katzenbach and Smith 1993).
Employees were prodded to contribute to restructuring decisions by such schemes as
job enrichment, quality circles, job rotation, gain-sharing, and stock ownership
plans. The social control of organizational performance became increasingly
internalized through corporate cultures based on Deming and Juran's total quality
management principles, 'a set of powerful interventions wrapped in a highly
attractive package' (Hackman and Wageman 1995: 339). TQM emphasized the
never-ending collaboration between management and workers for continuous
learning and quality improvements, assessment of customer requirements, scientific
monitoring of task performance, and process-management to enhance team
effectiveness. These high-performance innovations were all intended to lower
supervisory costs and increase employees' work-life morale, thereby raising
corporate productivity, quality, and profitability (Levine 1990; Lawler 1992).
During the last half century, the implicit employment contract binding workers
and firms changed from a virtual guarantee of long-term job security to one
emphasizing employability (Cappelli et al. 1997). In the insightful words of Intel
Corp.'s vice president for human relations, 'You own your own employability. You
are responsible' (O'Reilly 1994: 47). By flattening managerial hierarchies and out-
sourcing formerly internalized staff functions, firms shortened or eliminated many
traditional internal labor markets that had provided career ladders for regular
promotions to ever-higher levels of responsibility, prestige, and pay. Instead, jobs
evolved from fixed positions into flexible bundles of tasks that were subjected to
periodic restructuring to grapple with organizational contingencies in tumultuous
world economic markets. Jobs metamorphosed into project-based appointments
through which multiply-skilled employees rotated in short-term assignments on their
way to newer projects inside the firm or with other employers. Temporary and
subcontracted workers became the fastest growing segments of the U.S. labor force
by the 1990s (Belous 1989; Parker 1994). The proliferation of computerized
communication (Internet and intranets) and production-control systems (CADI
CAM), coupled with escalating customer demands for made-to-order goods and
services, drove the relentless quest for continual upgrading of employees' technical
and interpersonal skills. Firms deployed a multi-track approach, searching for new
workers with requisite competencies, training current employees in-house, and
forging ties to external vendors of job-training services such as junior colleges
(Kalleberg, Knoke, and Marsden 1995).
From these gales of creative destruction emerged a new corporate form-the
network organization, whose external alliances were discussed in the preceding
section (Miles and Snow 1995). Its distinguishing internal features are multiplex
exchange ties among the firm's loosely coupled divisions, departments, work
groups, and the individual managers and employees. It breaks down hierarchical and
functional barriers, replacing them with task-specific units connected through
communication, advice, and trust networks (Krackhardt and Hanson 1993). The
network organization
creates autonomous units, but it increases the volume, speed, and frequency of both
vertical and horizontal communication within the organization to promote collaboration .
... The result is an organization with superior performance characteristics for the 1990s.
Network management is, in the end, management by empowerment. (Limerick and
Cunnington 1993: 61)
Intraorganizational networks operate according to a logic of economically efficient
asset allocation. Rather than transferring goods and services by centrally
administered prices, the quasi-autonomous units are subjected to internal market
discipline when buying and selling resources, thus assuring they will continually
seek to improve their performance (Snow, Miles and Coleman 1992: 11). But,
corporate networks also function politically and socially in ways that defy strict
economic utility maximization principles. In particular, network relations offer
employees a prime source of social capital for developing rewarding careers under
the new employment contract terms which stipulate greater personal responsibility.
Networks and Career Capital

Employees have always used networking activity as important strategies for getting
ahead in their companies. A worker's personal networks comprise crucial social
capital investments that are as essential for career development as her or his human
capital assets of knowledge, skills, and experience. Employees survive and thrive by
learning how to construct and manipulate ego-centric networks that provide
advantages in the competitive scramble for jobs, project assignments, promotions,
and rewards. Networking abilities assume an even greater significance for
employees of the new forms of network organizations, where formal positions are
ill-defined and perpetually changing.
In a study of senior managers of a computer firm, Burt (1997) examined how
social capital affected rapid promotion. He measured social capital as constraints on
personal networks, that is, concentrated on fewer contacts. Persons whose networks
span more 'structural holes' are well-positioned to broker the flow of information
and to coordinate and control interactions between unconnected people on opposite
sides of the hole. Not only were managers with less-constrained networks promoted
relatively early, but the effect varied with the number of competitiors. The
correlation between network social capital and promotion was stronger for those
managers with few peers compared to those in positions where many people did the
same work. The social capital payoff was higher for people in unique corporate roles
'because such managers do not have the guiding frame of reference provided by
numerous competitors, nor the legitimacy provided by numerous people doing the
same kind of work' (Burt 1997: 356). Thus, their entrepreneurial networks offered
access to more rewarding opportunities.
Research on gender differences in network dynamics sought to explain how
personal networks are converted into corporate advantages. In a New England
advertising and public relations firm, Herminia Ibarra (1992) found differential
patterns of homophily (tendency to form same-sex ties) among the 80 male and
female employees. Men tended to concentrated their ties across multiple networks
(communication, advice-seeking, support, friendship, and influence) primarily on
other men. Women employees differentiated according to network contents,

obtaining social support and friendship from their female co-workers and
instrumental access through ties to higher-status men. That is, expressive and
instrumental ties coincided for men, but were inversely correlated for women.
Consequently, men seemed to receive higher returns than women on their social
capital investments, in the form of greater network centrality.
Similar gender-differentiated network propensities occurred among 63 managers
of four large corporations (Ibarra 1993a), with men relying more on weak-tie
homophilous networks and women forging more strong expressive ties to other
women. The relationship between managers' ego-net strategies and their potential
for promotion, as judged by supervisors and human resources staff, were also
conditional by sex. High-potential women and low-potential men placed greater
relevance on expressive networks, such as trust and reciprocity, while high-potential
men and low-potential women stressed instrumental ties. Ibarra concluded that
women's preferred network strategies placed them at a disadvantage relative to their
male peers: The 'entrepreneurial' network pattern characteristic of successful male
managers is less effective for females who require stronger network ties to achieve
the same level of legitimacy and access to resources' (Ibarra 1993a: 27).
Networks as Power Resources

Social power is a structural property of the relationships among actors in a social
system, rather than inherent in individuals' formal roles or personalities (Pfeffer
1981: 3; Knoke 1990: 1). Hence, power and political action in organizations are
rooted in the multiple intraorganizational networks connecting the participants. Even
formal authority that assigns legitimate rights to control corporate human resources,
as exhibited in an organization chart displaying supervisor-subordinate positions,
should be viewed as just one of several networks conveying political implications,
including such informal ties as communication (Pfeffer 1992: 111-125), advice,
support, trust, friendship, and horizontal workflow (Brass 1984). Knoke (1990)
argued that the primary analytic power relations of every social system are reducible
to two basic exchange networks that follow differing logics under which actors
affect one another's behavior. Influence occurs when 'one actor intentionally
transmits information to another that alters the latter's action' (1990: 3), while
domination involves controlling another actor's behavior 'by offering or withholding
some benefit or harm' (1990: 4). For example, an employee may induce a co-
worker's collaboration by persuading the co-worker that cooperation is in the mutual
interests of both firm and employee (influence), or by offering resources essential
for the co-worker's project (domination). Over time, these informal political
relations become institutionalized as the company's intraorganizational power
structures, with the employees occupying the dominant and influential positions
affecting their less-powerful colleagues' perceptions, cognitions, beliefs, and
behaviors. Intra-organizational power structures are highly stable and resistant to
change as the persons in power seek to perpetuate their advantages. Structural
change occurs mainly as the result of major external shocks, for example, corporate
takeovers or technological innovations that drastically rearrange existing political
relationships (Burkhardt and Brass 1990: 105).
A core theoretical proposition, derived from resource dependence principles, is

that actors who occupy the more central positions in intraorganizational networks
can exercise greater political power. While centrality may generate power for an
employee, the reverse causal process may also operate over time: people seek to
establish connections to the most powerful organizational players, in expectation of
enhancing their own power through these contacts. Incumbents in central locations
enjoy a variety of advantages over peripheral positions: through their proximity to
others in communication exchanges they can acquire more timely and useful
information; can better control the flow of resource exchanges and mobilize support
for initiatives; can mediate and broker deals between interested but unconnected
parties; and, through boundary-spanning ties to external organizational actors, they
can direct the organization's strategic objectives (e.g., Kanter and Myers 1991). In
short, 'network centrality increases an actor's knowledge of a system's power
distribution, or the accuracy of his or her assessment of the political landscape . ...
Those who understand how a system really works can get things done or exercise
power within that system' (Ibarra 1993b: 494).
As noted above, network methodologists developed alternative measures of
network centrality, including in-degree, closeness, and betweenness scores (Brass
1992). Rather than treating all relations as making equal contributions to each
person's centrality, a prominence index assigns higher centrality scores to
employees who are also highly central within the organization (Knoke and Burt
1983; Bonacich 1987). Thus, network centrality as prominence extends the principle
that 'it's not what you know, but whom you know,' to emphasize that your own
power depends importantly on 'whether the whom that you know has power.'
A few empirical investigations have uncovered compelling evidence that the
attribution of power covaries positively with employee centralities in
intraorganizational networks (see overviews by Krackhardt and Brass 1994; Brass
1995b). In a study of communication, friendship, and workflow networks among
140 nonsupervisory employees of a newspaper, Brass and Burkhardt (1992) reported
numerous statistically significant correlations between three types of centrality
scores and reputations for power (as attributed by supervisors and by peers). The in-
degree measure proved to be a stronger predictor than closeness or betweenness,
suggesting that a large volume of direct contacts may be necessary for coalition
formation and also provide the best mechanism for 'learning the network' (1992:
211). Krackhardt (1990) investigated the effects of betweenness centrality and
cognitive perceptions of both friendship (trust) and advice-giving networks on the
power reputations of 36 employees of a small entrepenurial firm. Controlling for
formal position in the company, persons who were more central in the friendship
network and who had more accurate cognitions of the advice network were rated as
more powerful by others. Neither advice centrality nor friendship accuracy had
statistically significant bearings on reputed power.
Finally, network centrality appears to affect some work-related perceptions and
activities. Ibarra's (1993b) analysis of which employees adopted problem-solving
innovations a New England advertising firm showed that centrality (prominence in
five types of network relations--communication, advice, support, influence and
friendship) 'was the most significant predictor of administrative innovation roles ...
and mediated the effects of various of various non network variables on innovation
involvement' (1993b: 492). But centrality was not a statistically significant factor in
the adoption of technical innovations. Burkhardt and Brass's (1990) longitudinal
analysis of computer adoptions in a federal agency also found similar patterns, with
early adopters' power and centrality increasing more than later adopters. In further
analyses of the advertising firm data, Ibarra and Andrews (1993) showed that advice
network centrality and friendship network proximity to varying degrees each
affected perceptions of such organizational conditions as risk-taking, acceptance,
information access, interdepartmental conflict and autonomy.
CONCLUSIONS
The preceding review of research and theory construction about organizational
networks and corporate social capital suggests that we are collectively investigating
several critical issues. Researchers are probing the social forces that lead to the
formation of intra- and interorganizational ties, their persistence, and their
severance. We have fragmentary understanding of how global network structures
simultaneously facilitate ('social capital') and constrain ('social liability') the
opportunities available to people and organizations in pursuit of their interests. And
we now better appreciate corporate social capital as both a generator and an outcome
of strategic actions embedded in complex social structures. Still missing is a
comprehensive framework to coordinate and accelerate the efforts of numerous
scholars toward a more coherent and cumulative research program that could
integrate the diverse facets of these elusive phenomena. After decades of network
analysis developments, we have abundant conceptual and methodological tools with
which to forge such a synthesis.
Two generic tasks should be intensified in tandem. First, researchers should
track social capital across multiple levels within and between organizations. At the
intraorganizational level of analysis, research designs could examine the
concatenation of multiplex relations among employees, work groups, departments,
and divisions into complex yet reproducible assemblages that maintain the identity
and integrity of the corporation as a bounded social actor. At the interorganizational
level, investigators must examine the detailed mechanisms through which social
network investments yield individual and collective benefits to alliance members.
Until we gain a clearer picture of how relations between firms shape economic and
political outcomes, our perceptions of the emergent N-form organizations will
remain fuzzy. The second major task for corporate social capital researchers should
be to collect and analyze longitudinal data about changing network structures and
processes. Current knowledge is cramped by the cross-sectional nature of most
research designs. Many methodological advances promise boundless opportunities
to expand the temporal dimension of social capital dynamics. We need to learn how
seemingly minor changes in specific connections, involving a handful of critical ties,
can cascade rapidly through a network, radically transforming its shape and
functions. And we need to integrate unique events into the actor-relation dualism,
thereby increasing our capacity to capture the historical forces changing corporate
social capital.
Social Capital of Organization:
Conceptualization, Level of Analysis,
2
and Performance Implications
•
Johannes M. Pennings
Kyungmook Lee
ABSTRACT
In this chapter we explore the benefits of social capital and the harmful effects of
social liabilities. Following Allison (1971), two models of the organizations are
juxtaposed: those of the Rational and Political Actors. The issues of social capital
require different perspectives when its implications for performance are addressed.
The mediation through individuals takes a prominent place in the Political Actor,
and moves to the background in the Rational Actor. The issue of aggregation from
the member to the organization is primarily an issue when we view the organization
as a Political Actor in which the members' social capital aggregates to that of their
organizations. Two illustrative cases that fit the two models are then presented, the
industrial business groups in Japan and Korea on the one hand, and the popUlation of
professional services firms in the Netherlands on the other. In the case of business
groups we point to both the benefits of social capital and the drawbacks of social
liability. When we shift to the study of professional services firms, we demonstrate
that social capital as a distinct organizational resource diminishes the likelihood of
dissolution. The implications for social capital and social liability are exposed and
reviewed.
INTRODUCTION
Organizations are presumed to have boundaries. They are endowed with various
kinds of assets on which they make ownership claims, and which are protected with
isolating mechanisms such as patents and contracts. They are liable for their
products and services. Also, they have members whose inclusion in the organization
is usually beyond dispute. In fact, the firm as a collection of individuals is often
bracketed when considering the competitive game it is playing with other firms. Yet,
organizational boundaries are precarious and permeable. Organizations have
exchange relationships with suppliers and clients, collude with competitors, and
forge all kinds of alliances because they cover only part of the value added in their
value chain. In their positioning across the chain they face such decisions to 'make
or buy' components and supplies, whether to share or even outsource R&D efforts,
or to operate on a stand-alone basis. Their coherence and integrity might decline and
bundles of resources often unravel into discrete parts, but these resources might also
become combined-for example in divestments and acquisitions, respectively.
Organizations are embedded in a web of relational ties. In the present chapter,
the term social capital captures important aspects of this relational web. Social
capital of organizations constitutes a distinctly collective property that might be
mediated by individuals, yet is uniquely organizational. Social capital complements
financial and human capital as assets that are more or less valuable, scarce and
imperfectly tradable (Barney 1991). Social capital is even more unique and difficult
to appropriate than these other types of assets as it hinges on the continued
involvement of two or more parties. Firms, as repositories of unique resources
require complementary assets in order to compete successfully. Social capital is
crucial in bundling intangible assets and provides the absorptive capacity to merge
proprietary knowledge with that of others. Organizations need to coordinate their
interdependencies in the value chain and negotiate a position in their industry. By
forging external networks, the organization maintains optimal boundary conditions
and remains in tune with external trends and events. At the same time, its boundary
structures preserve an organizational modicum of identity and protection against
erosion of its assets.
The social capital benefits seem beyond doubt; less intuitive might be the cost of
social liability. Social embedded ness endangers a firm's appropriability regime, and
might also envelop the firm too tightly into a web of ties that stifles its ability to
change or impedes its innovative capability. While network relationship is often
viewed as conferring various benefits, we should therefore also examine its
undesirable consequences.
As numerous chapters in this book indicate, social capital refers to resources
inherent in sustained long term relationships and associations. The concept
originates in sociology, with two writers standing out: Bourdieu (1980, 1994) and
Coleman (1990). In this chapter we extend their representation of social capital by
treating it as a unique organizational resource. We will further reflect on the nature
of organizations, and ask how such human aggregates or their social organization
are capable of possessing social capital. As with human capital, we need to dwell on
the tension between individual and organizational levels of analysis. While it is
tempting to 'anthropomorphosize' the firm as a human aggregate and impute an
ability to mold its surrounding network, we need to ask how such semblance comes
about, who the agent is, and what collective motives are operating. After having
dwelled on these issues, we explore the implications for organizations of having
accumulated social capital. We do so by contrasting two contrasting settings, i.e.,
business groups and professional service firms, as these stylized forms might
respectively illustrate the firm as rational and political actor, and by implication, the
Social Capital of Organization - 45
sort of aggregation issues that color the reason we depict their social capital. Below,
we belabor these two metaphors to highlight aspects of corporate social capital. We
conclude by spelling out implications and future research opportunities.
CONCEPT OF ORGANIZATION AND ITS SOCIAL CAPITAL

It has not been customary to view organizations as embedded in a network of
relationships, although person based networks have been used to describe a firm' s
external linkages (e.g., Levine 1972). Much of the pertinent literature has focused on
individuals (e.g., Burt 1997; Coleman 1988; Granovetter 1985; Uzzi 1997a), their
place in some larger network, and the impact it has on their behavior and attitudes.
Many views stand in sharp contrast with an 'over-socialized' view of man.
Economists tend to couch transactions in personal, self-interest seeking terms. As
parties in a market, people engage in 'arm's-length' relationships and their
interaction is solely conditioned by the need for exchange. Contrary to a utilitarian
tradition, norm theory in modern sociology assumes that people are overwhelmingly
sensitive to the expectations of others (Wrong 1961). Sociologists often stress the
structural context within which parties meet, and such a context might give rise to a
small number of conditions in which actors develop personal bonds, based on trust
and mutuality. Uzzi (1997a, this volume) calls such links 'embedded ties.') Within
such bounds, utility maximization is often suspended for the sake of preserving
reciprocal, even altruistic relationships. The next issue involves the extension from
the individual as a party onto himself versus the individual as an 'office holder.' Size
also matters; for example a market with single proprietorships entails rather different
inter-firm networking than the US banking world in which firms are tied together,
for example, through interlocking directorates.
Entrepreneurs, new ventures, and small firms differ markedly from large
corporations in terms of the links they maintain. The links that bind them might vary
from those that are heavily endowed with trust to those that fit the arm's length
relationships. The large corporation is prone to have arm's length relationships with
external actors, but as we will see, they often invest in boundary spanning systems in
which personally mediated links are discernible. Small firms are more likely to
develop bonds of trust and mutual adjustment with external actors such as suppliers
and clients, although some conditions give rise to arm's length relationships.
We need to position these distinctions against the 'model' of the firm, which is
often implicit (Allison 1971; Simon 1957; Thompson 1982). Organizations have
often been viewed as 'rational actors' (Allison 1971) or have otherwise been treated
as unitary economic agents. As a singUlar agent, the firm might be embedded in a
multiplex web of inter-firm relationships as manifest in contracts, joint ventures,
stock cross-holdings, etc. As units with clear legal boundaries and other 'isolating
mechanisms' firms complement each other in the value chain. The ties that bind
them can be viewed as social capital for coordinating inter-firm activities. If we,
however, view organizations as human aggregates, as Allison (1971), for example,
stipulates in his organization as 'political actor,' we might attribute to that
organizational social capital by virtue of the aggregate social capital of its members.
The presumption of firms endowed with social capital appears non-problematic but
the implications are rather different in the two scenarios thus depicted. In this
chapter we visit the issue of firm as rational versus political actor in greater detail.
In this chapter, for the sake of the argument, we juxtapose the rational actor
caricature with its political actor counterpart and examine the social capital as an
integral part of these models. 2 In the case of the firm as rational actor, we treat
individuals as a component in what often appears to be a multi-layered network;
partly mediated by individuals and partly by other linking vehicles. In the case of the
firm as political actor, the link will often be personal and fit the characterization of
simple tie, based on trust and tacitness. 3
We want to stretch the concept of social capital such that it might become an
extension of the individual as an office-holder in an organization and, consequently,
become an accessory for his firm's functioning. For example, an early study by
Pettigrew (1974) on the 'politics of organizational decision making' narrates the
position of an information technology specialist as a boundary spanner between his
firm and external vendors. As office-holder his significance derives from the quality
of internal and external embeddedness. We might then ask whether the office
holder's network connections can be combined with that of others into an index of
organizational social capital. Furthermore, inter-firm links might also be discernible
beyond the IT specialist, for example, by the long term outsourcing of data storage
and retrieval services, or the presence of a hot line with the IT consultants. Such a
link is not 'simplex,' but what might be called 'multiplex.' The Pettigrew example
illustrates the transition from the firm as a human aggregate to the firm as a
coherent, singular entity where the issue of aggregation becomes bracketed, or
remains altogether outside the purview of the observer.
LEVEL OF ANALYSIS
It is problematic to move from the individual to the organizational level of analysis
when analyzing inter-firm networks. The issue of aggregation from the member to
the organization is primarily an issue when we view the organization as a Political
Actor in which the members' social capital aggregates to that of their organization.
Nevertheless, people associated with the organization as Rational Actor carry out
deeds on behalf of their firm, and while the model is agnostic about their integrity,
we could focus on their role as distinct linking mechanisms as well.
At face value, the individual-collective distinction seems merely conceptual, not
'real.' The issue oscillates between two frames: do individuals as agents or office-
holders connect organizations and other human aggregates? Or do organizations and
other human aggregates connect individuals? In this chapter, we are mostly
concerned with the first type of framing. Nonetheless, we recognize that many inter-
firm links condition the intermediation of individuals. In abstracting away from
individuals as mediators of inter-firm links we shift from the view of the firm as a
'political' actor to that of a 'rational' one (Allison 1971). The level of analysis
becomes moot and little need exists for acknowledging cognitive, cultural, or
strategic differentiation-whether in the organizational core or at its boundaries.
To the extent that aggregation surfaces as a salient feature, we should abandon
the neoclassical notion of the firm as a unitary actor with a well defined preference
ordering and whose strategy betrays a clear and unambiguous preference ordering.
Its membership has a singular identity. The challenge for firms is to consolidate
divergent identities into a coherent one such that they might even approximate the
firm as a unitary integrated actor. The members are assigned to interlocked sets of
roles and they develop informal sets of hierarchical and horizontal relationships with
other people inside and outside the organizations. A large chunk of organizational
social capital exists by virtue of the individuals whose relationships span
organizational boundaries.
Some organizational participants are more contributory in their social capital
than others, depending on their involvement in the focal firm and its transacting
partners. Indeed, not all members are equivalent in their ability to leverage their
social capital for the firm. Members vary not only in their contribution to external
ties but also in their participation in the organization (e.g., Cohen, March, and Olsen
1972). When aggregating the social capital of members to arrive at a stock index of
firms, there is also the issue of redundancy. A network link is redundant if the
marginal increase in benefits from acquiring or maintaining that link equals zero.
Redundant ties have been well documented at the individual level, e.g.,
Granovetter's (1995) 'weak' versus 'strong' tie and Burt's (1992) presence or absence
of 'structural holes.'
The aggregation of the networks of organizational participants is prone to have
redundant contacts. The number of members maintaining contact with
representatives of other organizations might produce 'stronger' ties that are
particularly beneficial for the transfer of sophisticated knowledge. For the
transmission of information or what might be called 'explicit knowledge,' such
strong ties are hardly efficient (compare Hansen 1997). Furthermore, not all social
capital of members aggregates to the social capital of the organization. The social
contacts of certain organizational members may have little or no instrumental value
for their organization.4 Only overlapping membership in groups and organizations,
that are operationally or strategically relevant, matter when aggregating individual
social capital to that of the organization; the most common example is interlocking
directorates (Pennings 1980; Stokman, Ziegler, and Scott 1985).
Boundary Spanner or Multiple-Group Membership

The concept of overlapping membership as a way to represent an individual's social
capital should also be invoked to revisit the issue of a firm's boundaries. If members
vary in their inclusion in the focal organization, their external contacts should vary
in value as well. Even if organizational members have valuable external ties, they
become a valuable component of the firm's social capital only if the members enjoy
access to certain peers-for example, those with power, information, and other
resources. If inclusion is highly partial, their individual social capital becomes
marginalized for the firm as well.
For simplicity's sake, organizational members might be stratified into a core
group, a regular or associate group, and temporary or marginal workers. The core
group consists of essential employees who are long-term employees and owners.
Their fate is usually tied to that of the organization. The regular or associate group
consists of rank-and-file employees who have been involved in the organization for
'some' time and face good prospects to join the core group. Many members who
participate in that tournament will 'plateau,' become sidetracked or might even be

terminated, however. The temporary or marginal category include temporarily hired
workers and employees of sub-contractors, i.e., workers who fill the jobs not
requiring firm-specific skills and who have little chance of moving into another
category of members. 5
It follows that the social capital associated with the core group is more
important for the organization than that of the regular group. The reason is two-fold.
First, members in the core group are more likely to use their social contacts on
behalf of the organization. Consistent with the garbage can model (Coh~!1, March
and Olsen 1972), these members have the highest 'net energy load,' as their fate is
closely tied to that of the organization. Second, they are likely to maintain more
valuable social contacts for the organization. They are more central to the access
structure, and enjoy higher positions with more power and authority. Many of the
firm-relevant social contacts are based on the job and title of individual members. A
CEO becomes a board member of a peer organization, supplier or some other
organization; a partner in a consulting firm befriends senior executives in the firm he
works for, etc. Compared to the employees in regular or temporary groups,
members in the core group tend to have social ties with people who occupy higher,
more visible and more prestigious positions in their organizations. In other words,
people who have social contacts with members in the core group of a focal
organization tend to have more valuable resources at their disposal for the focal
organization than do the people who have primarily social ties with members in its
more peripheral ones. Core members also stay longer with their organization such
that their organization stands to benefit more from their social capital. Overall, we
need to focus on the nature of the employment relationship to weigh an individual's
ability to link his firm with other ones.
Figure 1 provides a graphical display of organization stratification in terms of
magnitude of personal inclusion.
There are also other ways to compartmentalize the firm as a community of
people who are endowed with human capital, and who are differentiated by skill,
function, types of markets, products, or technologies. Firms have either a functional
Marginal: free agents, sub contractors,

contingent workers, strategic alliance
guest employees
Associate: temporary employees,

'in-transit' workers
Core: long term employees, partners,

owners, managers, residual claimants
Figure 1. Stratification of firms based on partial inclusion of their members

or divisional (and in many cases some hybrid) structure whose boundaries define
identities. In fact, although finns proclaim to be a hierarchy that economizes on
transaction costs (Williamson 1975), they in fact comprise numerous sub-cultures,
with their own identity and parochial interests. While hierarchy and lateral linkages
integrate disparate units, they often face major hurdles in consolidating their skills or
knowledge, or more generally in bundling their contributions to the common good
(Brown and Duguid 1997; Kogut and Zander 1996). A firm's internal networks such
as heavy duty project managers (Clark and Fujimoto 1991), overlapping teams, and
interdepartmental career paths become vehicles for knowledge migration, but such
networks are often comparatively deficient because specialization impedes
knowledge transfer, especially knowledge that is difficult to package. Ironically,
communities of knowledge within the firm have often easier access to like-
communities in other firms than they do with the sister departments within their own
firm. The implication is that such external networks are often more efficacious in
bridging the firm with external actors than do networks that embrace the total
organization. By way of example, we might consider a firm's participation in an
'invisible college' less problematic than its participation in a trade association
(Powell 1990; Lazega, this volume).
Multiplex versus Personal Forms of Organizational Boundaries

At the level of inter-organizational relationships, we could make an even stronger
argument about the individually anchored social capital of organizations. When the
vendor of a software firm leaves, he might appropriate the connections with clients
that he has built up during his tenure. One might thus argue that the social capital of
organizations is tied up in the individuals they employ.
Yet, as with all intangible assets, social capital can also be treated as an
intangible asset that is not exclusively buried in personal networks. Social capital is
often 'depersonalized' or is couched in mUltiplex forms. Interorganizational links
established through individuals might begin to lead a life of their own. Or such links
become embellished by other glue such as contracts, traditions, and institutional
arrangements. The members who are then a complement to a system will in fact also
be governed by the norms and beliefs that are endemic to local social arrangements.
When links become multiplex, they cease to be dependent on individuals who act as
brokers. By way of examples, patent citations signal proximity of knowledge among
organizations and can be examined as a conduit for inter-firm knowledge transfer.
Cartels amount to a clique with shared norms where the members are firms rather
than people. A set of firms might be tied through mutual share holdings. Affiliation
among organizations, such as keiretzu in Japan, chaebol in Korea, or business
groups in Sweden illustrate bundles of inter-firm connections that cannot be reduced
to middleman-members.
Strategic alliances such as joint ventures, R&D partnerships, and minority
investments embody nodes in webs of inter-firm networks in the telecommunication,
micro-electronic and biotechnology industries (e.g., Ajuha 1998, Hagedoorn and
Schakenraad 1994; Omta and van Rossum, this volume). Severing some of these
linkages might be impossible. For instance, Microsoft has extensive lock-in
agreements with PC makers and their suppliers and PC manufacturers in fact have
contracted for the pre-arranged installation of Microsoft's operating system in what

used to be called 'IBM-compatible' personal computers. Biotechnology firms'
entrenchment can be inferred from patent citation networks in which their
intellectual property is more or less linked with that of other firms ; the tightness of
their links is derived from the proximity as measured by relative citation frequencies
(Stuart, Hoang and Hybels 1997, see also the chapters in this volume by Stuart and
Smith-Doerr et al.).
All of this requires us to dissect the ingredients of inter-firm networks into at
least three categories: 6
Link
Any sort of association between two or more firms, including equity cross-holdings,
patent-ties, licensing agreements, R&D partnerships, equity joint venture
agreements, gatekeepers, or interlocking directorates.
Ties
Human mediated links, such as interlocking director or guest engineer. Ties can be
'neutral,' reflexive (Pennings 1980) or even universalistic versus parochial and
particularistic.
Relationships
Human mediated ties that are particularistic, as for example the guest engineer who
has an OEM employment status but resides on the premises of a supplier.
In short, corporate social capital bifurcates into personalized and depersonalized

forms, with relationships often augmented with ties and links; while in other
instances, the link might persist without the benefit of a relationship. This distinction
often corresponds to a simplex versus a multiplex web of network connections.
Multiplex 'links' appear to be more congruent with the rational actor metaphor of
Allison, while 'relationships' feature prominently in treatments of organizations as
political actors. Table 1 furnishes some examples. First, the organization itself can
have a link with other organizations that is instrumental for its functioning.
Affiliation among organizations, such as keiretzu in Japan or chaebol in Korea is a
social link of the organization itself rather than of organizational members. As a
legal entity, the firm is capable of contracting, of acting as a partner in any market
relationship, induding the setting up of joint ventures, the acquisition of another
firm, or the shedding of a business unit to other firms, etc. Indeed, independent of
their members, the organization often maintains social capital through the repetitive
exchanges with other organizations. The pattern of exchanges has stabilized, even if
the individual members who participate in the process have been changed (Chung
1996). Investment banks perpetuate their collective efforts when they syndicate
public offerings (Chung, Singh, and Lee 1995). Semiconductor firms joined
SEMATEe when they sought to acquire greater economies of scale.
Whether one assumes a personal or impersonal link (or a hybrid form
comprising both links and relationships) between organizations, links constitute the
ingredients of arrangements that govern the firm-environment interface. In some
cases the arrangements can be viewed in their own right, but their efficacy in
.........................................
...
Figure 2. Boundary transaction system comprising four individuals among two organizations
managing external dependencies depends critically on the quality of the relationship

with internal and external decision makers. Adams (1976) was one of the first
writers to review such arrangements. He refers to so called 'boundary transaction
systems.'
Boundary Transaction System

Social capital fits with the notion of more or less permeable boundaries of
organizations that become spanned by a 'boundary transaction system' (Adams
1976). Figure 2 provides a graphic representation. As Table 1 indicated, such
systems diverge into pairs of individual dyads such as the interlocking director or
guest engineer whose role in maintaining the firm's network depends critically on a
balanced overlap between the inside and the outside. Or boundary transaction
systems are larger and more elaborate entities-for example kereitzus and R&D
partnerships. In the latter case the inter-firm link is not nearly as dependent on the
presence of boundary-spanning individuals such that the significance of their
mediation is comparatively minor. The personal ties often complement non-personal
ones such as reciprocal ownership arrangements and R&D partnerships.
Furthermore, the relative salience of the system hinges on the duration of links
that are maintained by individuals that make up the system. The longer the tenure,
Table 1. Examples of social capital among organizations
Mediated by Individuals (Simplex) Mediated by Systems (Multiplex)

Interlocking directorates (Pennings 1980) Business groups (Acevado et al. 1990)
Guest engineers (Dyer 1996a) Chaebol (Kim 1997)
Social register (Useem and Karabe11986) Keiretzus (Gerlach 1987)
Revolving door syndrome (Pennings, Lee Investment bank syndicates (Chung, Singh, and Lee
and Witteloostuyn 1998) 1995)
Alumni (McKinsey) Joint ventures
Double agent R&D partnerships
Gatekeeper (Tushman 1978) Guanxi-chia-jen (Tsui and Fahr 1997)
Emissary Electronic clearing house (Pennings and Harianto
1992)
the more distinct the boundaries of the transaction system and the greater the
likelihood that its members 'go native,' i.e., acquire an identity almost different from
the firms they span. Consider boards of directors, or executive councils of Japanese
business groups who over time might become closely knit teams. Employees
originate from leading universities, where they have already formed friendship
networks, and synchronically move upward through equivalent organizational
ladders, such that the 'old boy network' remains intact from university years until
retirement. The implication is that succession patterns further strengthen the
boundary system's identity (Yoshino and Lifson 1986).7
The boundary transaction system is useful in that it points to the role of
member's social capital in producing organizational social capital. Likewise, by
recognizing that the system often evolves into a system that cannot be reduced to the
participating members, social capital might become depersonalized. The system
might become part of a business group, cartel, a joint venture, a long term licensing
agreement, or R&D partnership. Such systems are bound to become semi-
freestanding entities when three or more firms decide to participate. For example,
SEMATEC and ESPRIT are consortia of semiconductor firms that joined forces at
the behest of the US and European Union governments respectively to create what
we might call a boundary transaction system.
A key difference between a simplex and multiplex boundary system involves
the notion of trust. In a simplex system, trust is anchored in a dyad of trustor and
trustee who maintain a form of personal trust of what Simmel calls 'mutual
faithfulness.' Bradach and Eccles (1989) refer to expectations that the other side will
not behave opportunistically. It accords with the definition of trust by Mayer, Davis
and Schoorman (1995: 712)-a willingness of a party to be vulnerable to actions of
another party based on the expectations that the other party will perform a particular
action important to the trustor, irrespective of the ability to monitor or control the
other party. This definition excludes the social context of the dyad.
In multiplex systems, the social context becomes central and will in fact color
the nature of the relationships between individuals who are part of that system. The
context includes not only traditions, ties inherited from individuals who are no
longer present, contracts and financial leverage, but also forms of institutionalized
trust (Luhmann 1979; Shapiro, Sheppard and Cheraskin 1992; Zucker 1986). The
institutionalization evolves both temporally and spatially. Firms have often recurrent
contacts with other firms, and the history of their relationship provides a platform
for the current boundary system. Firms are also entrenched in larger entities, most
notably business groups. The firms that make up a business group share norms about
inter-firm transactions, have developed routines for contracting, and enjoy a group-
derived reputation that molds the dynamics of interpersonal relationships within a
boundary transaction system between two member firms. And history matters here,
too: the member firms have collectively gone through actions that resulted in shared
practices, mutual stock ownership, exclusive supplier-buyer relationships, or
investments in transaction specific assets (Dyer 1996a). The historical and spatial
context for two individuals who span their respective firms is therefore critically
important in comprehending corporate social capital.
The fact that building up social capital requires time was nicely illustrated in the
recent difficulties between Ford and its suppliers. Ford sought to redesign its Taurus
model, while at the same time redesigning its boundary transaction system (Walton
1997). For example, the firm attempted to move from multiple, arm's length ties
with suppliers to single source relationships. Having made few investments in social
capital, its 'relational competencies' (Lorenzoni and Liparini 1997) for managing
such supplier relationships were grossly inadequate. The boundary system included
individuals such as Taurus project managers and representatives from 235 suppliers.
The project's social architecture was to embrace a Japanese-style long-term
cooperative relationship with suppliers. Yet, the culture of the system could be
described as 'You could not trust them.'
The boundary transaction system should not be confined to individuals who
gave rise to the system or were involved in its perpetuation. It ranges from dyads of
individuals to complex social, economic, and technological arrangements. It evolves
from individuals who interact frequently so that the firms become familiar with each
other. Familiarity alleviates transaction costs, improves coordination across
organizational boundaries, and reduces agency problems-in short the familiarity
that comes with organizational networks confers benefits. Familiarity also produces
group-think, cuts the firms off from important external stimuli, and renders it
increasingly inflexible. More specific benefits of social capital and the harmful
effects of social liability are discussed next.
PERFORMANCE IMPLICATIONS OF SOCIAL CAPITAL

At the onset of this chapter, social capital was mentioned as an integral part of the
organization'S intangible assets. The reference to assets suggests a rent producing
potential. However, social capital as such cannot produce rents, but it contributes to
greater rent maximization of other resources that complement social capital.
Burt (1992) points out that social capital is owned jointly by the parties to a
relationship whereas financial and human capital are the property of individuals or
firms. In other words, social capital is embedded in the positions of contacts an
organization reaches through its social networks (Lin, Ensel, and Vaughn 1981).
Second, social capital is related to rate of return in the market production function
whereas financial and human capital pertain to the actual production capability. We
should ask: What is the role of social capital in economic transaction? Under perfect
competition, social capital cannot generate any economic rents (Burt 1992). The
market however is hardly perfect and information is not costless. The member's
social capital strengthens his firm's ability to retain clients, perform market
intelligence, and learn about new technologies. This is particularly true in our
knowledge economy where many industries are characterized by abstract products
or services, whose quality and other dimensions are difficult to articulate and where
delivery of output is highly coupled with reputation (cf. Burt 1992). Clients resort to
their social contacts to screen their suppliers because assessment criteria for quality
might be hard to come by. While social capital is not part of the production function,
it has profound impact on the benefits that firms derive from their productive
capabilities. Putting it differently, social capital brings the opportunities to exploit
financial and human capital at a profit.
In the next two sections, we belabor these implications by reviewing two

examples with rather different manifestations of interfirm ties: industrial firms that
make up business groups such as Keiretzus and Chaebol and professional services
firms that comprise the audit industry. We have hinted that these two examples
present different manifestations of a firm's external networking. Firms that belong to
a business group are typically depicted as ('rational') actors in a conglomerate-type
setting with mutual equity ownership, long term supplier-buyer transactions, and
shared directorships. The relational structure that business groups have is assumed to
furnish social capital to member firms. We impute such benefits to the firm without
confronting aggregation issues or delving into internal factions. Individuals are
merely one of the threads that make up the fabric of networks of business groups.
Thus the member-firms of business groups are depicted as integrated, unitary actors
who might benefit from their inclusion. The groups furnish interesting data on the
benefits of social capital and costs of social liabilities among firms that come close
to the stylized Allison-type Rational Actor.
In contrast, professional services firms belong to a sector that resembles a
cottage industry, where individual professionals appear to be the most salient
participants. While many professionals join a partnership and thus become co-
owners of the firm, these organizations are very flat and by dint of the
professionalization comprise members whose loyalty might be as strong to their firm
as it is towards the profession. The social capital of the firm might in fact be the
social capital of individual professionals. Even if we aggregate their social capital to
that of the firm they belong to, there always remains the issue as to whether it is the
partner, his peers as co-owners, or his firm who can make claims on the social
capital that is mediated by the professional. The professional has his own roster of
clients and might feel more loyal to those clients than to his brethren with whom he
makes up the partnership. His ties, and by implication his firm' s links, often fit the
notion of embedded ties. Arm's length transactions are incompatible with the
rendering of services, although some emotional distance with the client is often
deemed appropriate. Since partnerships often break-up, or witness an exodus of
partners, the caricature of Allison's political actor might sometimes be quite
appropriate as a general descriptor. Yet, as we will see we often have to qualify this
caricature.
Business Groups
Social networks have been a pervasive feature of Asian socIetIes in general.
According to Hofstede's (1980) landmark study, Asian societies stress collectivist
values and cherish loyalty and commitment to family, organization, and community.
At the corporate level we also discern a preponderance of networking-most visibly
in business groups. Business groups include Japanese keiretzus, or their pre-war
predecessors, called zaibatzus, and Korean chaebols. These groups contain a myriad
of firms held together by ownership links, supplier-buyer relationships and mutual
guarantee for each other's bank loans. Other countries, most notably Sweden (e.g.,
Hakanson and Johanson 1993; Sundqvist 1990; Berglov 1994) and Argentina (e.g.,
Acevado et al. 1990), harbor business groups, but take on a local, idiosyncratic form.
Therefore it is prudent to limit ourselves to a relatively homogeneous class of cases
(cf. Guillen 1997). Furthermore, some other Asian countries manifest distinct forms
of social capital among organizations; we could mention bamboo networks that are
depicted as a guanxi (relation)-based cluster of Chinese firms (cf. Tsui 1997;
Weidenbaum and Hughes 1996). In these cases, the individual as family member
performs a primary role in forging inter-firm links, and the family rather than the
firm appears to be the most salient unit of analysis. Unlike more centrally coupled
business groups in Korea and Japan, these Chinese forms of organization are
octopoid and opportunistically diversified (Tam 1990). In this section, we restrict
ourselves to keiretzus and chaebols.
Chaebols
Korean business groups manifest several features that set them apart from Western-
style business groups (Kim 1997). They display family ownership and management,
controlled by a powerful chair. The chair's power derives from stockholdings and
from being the father or senior family member who are heads of member companies.
Kim (1997) even refers to unquestionable filial piety and patriarchy based family
control within modem multinational firms. A founder' s descendents actively
participate in the top management of the chaebols. When the founder dies, his
descendents succeed as heir. When the founder with multiple descendents dies, 'his'
chaebol sometimes divides into several mini-chaebols as the case of Samsung
indicates. Still, the kinship and family networks link the member firms of those
mini-chaebols.
Chaebols also exhibit high flexibility in mobilizing financial capital, technology
and human resources. Unlike keiretzus and zaibatzus (although the same Chinese
character is used to denote this extinct type of Japanese business group as well as
chaebol!) that are governed through consensus building and psychological
commitment, chaebols are nimble in their deployment of resources and the patriarch
can implement strategic decisions without consulting others. There is widespread
rotation of key personnel, R&D efforts are pooled across companies and transfer of
cash can be arranged through financial services firms, and the member companies
can guarantee each other's borrowings from financial institutions.
Finally, the complex set of networked firms that make up a chaebol are
exceptionally broadly diversified. Kim (1997) shows that a chaebol like Samsung
operated in light and heavy manufacturing as well as in financial and 'other' (e.g.,
construction, media, hospitality, and advertising) services. Presumably, such
diversification allowed chaebol to offset lack of high-tech skills by exploiting semi-
skilled and unskilled labor in a way that would not be feasible to a non-networked
competitor (Amsden 1989), while at the same time produce products that are price-
competitive rather than quality-competitive in the global markets. Compared to
keiretzus, chaebols are basically shaped on the basis of the founding family. The
financial institutions are less utilized to form the relationship within a chaebol's
member firms, because chaebols are blocked from owning more than 8 % of shares
in commercial banks.
Keiretzus
Chaebol should thus not be confused with keiretzus or even with their name sake
zaibatzus although the degree of contrast is a matter of controversy. After the second
world war, zaibatzus were dismantled but reappeared in a different form called
'keiretzus.' As a result of the transformation, the founding families of zaibatzus lost
their shares and power and thus were no longer a source of connections. The
insurance companies are at the keiretzus' apex, and from them cascades a transitive
pattern of equity cross holdings-the implication being that the insurance firms and
their executives are the ultimate center of power and influence (Nishiyama 1982).
Keiretzus' governance is much more decentralized with decision making among
firms by consensus rather than through fiat by the keiretzus' insurance firm's
executives. The zaibatzus provided a template and became mimicked by Korean
entrepreneurs and in any event evolved into a prominent form during Korea's
industrial revolution. Zaibatzus and chaebols share characteristics such as family
ownership, management by patriarch, and unrelated diversification. However, unlike
the chaebol, the zaibatzu also controlled commercial banks, giving them access to
capital markets.
Keiretzus are laterally federated with transitive stock ownership arrangements
that induce minimal interference in between-firm interactions, rather than
resembling a chaebol-like holding with a vertically arranged governance structure. s
Gerlach (1987: 128) refers to them as 'business alliances,' which he defines as the
'organization of firms into coherent groupings which link them together in
significant, complex long-term ownership and trading relationships.' They are
distinct in the manner in which they have established coordinative mechanisms to
govern their relationships. These include high level councils of executives, the
shaping of exchange networks, and the external presentation as a coherent social
unit, for example, through advertising and product development activities.
Prominent, but largely invisible in the structuring of network links is the role of
financial institutions, which unlike the chaebol are an important component of the
Japanese style alliance. The member firms are heavily indebted to the keiretzu's
main life insurance company and bank. The cross equity holdings constitute an
important link over and beyond the relationships that could be uncovered if one
were to have access to their inner circles. Unfortunately, no research exists on the
power structure within such circles, and the sort of collective decision making
processes that ensue. Thus we are also deprived from making strong conclusions
regarding the stock of social capital among keiretzus firms. These links are not
merely leverage tools, but in fact might acquire a significant symbolic meaning on
their own and complement other media of networking such as exclusive R&D
projects. The keiretzu as a somewhat hierarchical network is therefore multiplex-
debt holdings, cross-equity holdings, supplier-buyer links, and personnel bonds are
part and parcel of the connections that bind the firms into a tight and relatively
unified alliance.
Social Capital of Business Groups

Firms that are part of chaebols and keiretzus are presumed to benefit from the social
capital that ensues from their membership in these alliances (e.g., Kim 1997;
Lawrence 1991). Social capital is manifest in two ways: First, business groups
provide its member firms with access to resources from other firms. As a quasi-
holding or federation of businesses, they can furnish superior access to financial
capital through a member financial services firm and cross guarantee each other's
bank loans. Similarly, business group specific suppliers and their Original
Equipment Manufacturers belonging to the same group display shorter lead times in
new product development because they circumvent transaction costs, for example,
by making significant asset specific investments that in the absence of a business
group context would incur significant hold-up problems (e.g., Dyer 1996a; Gerlach
1992b). The inclusion in the keiretzu reduces the outsourcing to one or at most two
suppliers, and the relationship is typically based on trust and mutuality. By way of
contrast, Toyota relies often on a single, keiretzu-anchored supplier, while US auto
manufacturers such as GM usually rely on as many as six suppliers, with whom they
interact opportunistically and at arm's length (Dyer 1996a; Nooteboom, this
volume). The suspension of the hold-up problem results also in joint R&D and in the
geographic clustering of OEM and their suppliers, thus economizing on value chain
coordination costs, transportation distance, and inter-firm transfer of tacit knowledge
(e.g., Hansen 1997; Nooteboom, this volume).
The social capital of business groups, however, is not confined to intra-group
relationships. Since their boundaries are also salient at the group level, they have
enjoyed scale advantages, not unlike those accorded fully vertically-integrated firms.
Such assertions question the saliency or distinctiveness of boundaries, and in
particular the issues associated with vertical integration, governance, and transaction
costs (Williamson 1996; Powell 1990). Even though inter-firm links are not
exclusively mediated by individuals-as we have argued they are mUltiplex, to say
the least-the links that bind them might be so strong that the focal attention often
shifts to that level of analysis when discussing social capital. They maintained levels
of flexibility in moving around human resources and other assets, and because of
superior access to cheap and unskilled labor, were able to claim cost leadership
positions in their world of multi-point competition. (Kim 1997: 180-195). Yet, on
the next higher level of analysis, these groups commanded clear benefits that
surpassed inter-firm arrangements, as reviewed by Powell (1990).
Empirical Evidence
In Korea, there is the often documented 'cozy' chaebol-government interface.
Chaebols as groups are often endowed with a good deal of social capital because of
the support they have extracted from the South Korean government. Compared to
non-chaebol firms, chaebols have had better access to state-controlled resources, and
were thus able to exploit governmental powers for their own benefit (Kim 1997).
The chaebol dominated segment of the economy grew much faster than the economy
as a whole.
The reasons that chaebols have received a great deal of governmental support
are two-fold. First, the sheer size of chaebols has made them very important for the
development-oriented Korean government. For instance, the value added by the 30
largest chaebols has been around 15 % of GNP and their sales volume has been
around 80 % of GNP (Cho, Nam, and Tung 1998). Since chaebols have been used as
a tool for the government's industrialization policy, the Korean government has
provided a great deal of favors including soft loans, import prohibition, tax breaks,
etc. Second, the relationships of elite university graduates strengthened the
relationship between the Korean government and chaebols. People who graduated
from elite universities have occupied major positions in the Korean government,
banks, and parliament. As a result, chaebols appointed elite university graduates as
CEOs to lubricate their relationship with external entities. For instance, 62% of
CEOs of the seven largest Korean chaebols in 1985 graduated from Seoul National
University (Steers, Shin, and Ungson 1989).
There is also some provisional evidence that member firms within a chaebol or
keiretzu might encounter the adverse effects of 'over-embeddedness.' In Korea we
have the case of the Kukje chaebol and recent bankrupcies of major chaebols, while
in Japan the differential learning of keiretzu versus non-keiretzu suppliers provide
testimony to the harmful effects ('social liability') of social embeddedness. The
Kukje case emerged in February 1985 and evolved from an ordinary bankruptcy into
a scandal when the Chun government disbanded the chaebol due to 'reckless
management, and exceedingly high debt rates.' It is most relevant for our argument
because of 'nepotic management by the sons of the founder' (Kim 1989). The
bankruptcy case is somewhat ambiguous and opinions varied as to whether it was
over-embeddedness among member firms or deficient external social capital that
accounted for the disbanding of Kukje. Yang, the chaebol president, claimed
favoritism on the part of the Chun government. In any event, further research should
identify whether it was social capital at the group level or at the group-state level
that explains the demise of Kukje.
Due to the risk-sharing role of chaebols, Korean chaebols enjoyed very high
survival chances and thus only a few chaebols experienced bankruptcy. During the
period of January 1997-January 1998, however, nine chaebols among the 30 largest
chaebols experienced insolvency. The mutual guarantee of bank loans made whole
member firms rather than some of them insolvent. In some cases, the failure of one
member firm became the reason of the bankruptcy of the chaebol. Over-
embeddedness to other member firms rendered profitable and financially sound
member firms bankrupt, thus revealing the 'dark side of social capital' (Gargiulo and
Benassi, this volume).
Keiretzus in Japan also function as a tool for risk sharing among member firms
(Nakatani 1984) and thus they enjoy a lower bankruptcy rate (Suzuki and Wright
1985). However, criticism has surfaced regarding their traditionally claimed
advantages. Gerlach (1992b) sees the potential unraveling of keiretzus now that their
benefits have appeared to wane. Nobeoka and Dyer (1998) have recently completed
a survey of OEM-automotive supplier relationships and produced evidence
indicating that suppliers that diversify away from a single keiretzu based OEM are
more profitable compared with firms who are locked in a close single-source
relationship. They interpret this finding as being due to either superior bargaining
power, or to a broader exposure to technological know-how; such firms diminish
their dependence on a single OEM or they witness learning benefits in that their
know-how is likely to be more generic and less firm-specific.
Similarly, Lincoln, Ahmadjian and Mason (1997) provide evidence of Toyota
the auto manufacturer and Toyota the keiretzu member, which diversified away
from keiretzu-based automotive suppliers. These authors report that intra-keiretzu
knowledge was not only limited, but that Toyota did not even attempt to elevate its
'internal' suppliers to the standard that would meet its needs. The implication is that,
in spite of trust and inbred capabilities, the firm begins to question the benefits of
traditional arrangements. Such precedents might lead the keiretzu on a path of
further unraveling its stale social capital and the substitution of a fresh one.
Summarizing, business group's endowment of social capital should be
differentiated into that social capital that is discernible at the group level versus that
which resides at the interface between the business group and external actors. The
beneficiary of social capital is the firm or a group of firms who are portrayed as
unitary actors, operating in their economic-political arena. The evidence so far has
focused on the social capital inherent in social structure, but more recent evidence
shows also that over-embeddedness might lead to social liability.
Audit Firms
The accounting sector presents another setting in which the costs and benefits of
social embeddedness are evident. Unlike markets with industrial firms, as is the case
with industries comprising business groups, the accounting sector produces largely
intangible and abstract services. The measurement of product quality is elusive, the
production flow is exposed to the client who is often an active co-producer of the
services rendered. The firm has some degree of hierarchy but is usually much flatter.
In fact most firms are stratified into partners (i.e., owners) and employees, some of
whom expect to join the partnership. Their close exposure to the market place and
their intense involvement with clients makes social capital a central feature of
operations and a key driver of organizational performance. This sector resembles
numerous cottage industries where personally mediated ties predominate, not unlike
the settings of garment district members (Dore 1983; Uzzi 1997a), or investment
bankers (Burt 1997).
Ironically social capital can be viewed as a substitute for objective criteria of
quality, reliability and consistency. In the absence of objective, verifiable and
measurable product attributes, clients might rely on their networks to select auditors
or to remain loyal to them even after the honeymoon period has passed (cf. Podolny
and Castellucci, this volume). The endowment of social capital is therefore a critical
resource in such sectors. Absent social capital, the firm might not extract much rent
from its human capital. Furthermore, social capital allows the firms to leverage their
human capital thus extracting more quasi rent from that asset. Social capital is not
only valuable as rent producing potential, but is also scarce and difficult to
appropriate. These aspects suggest social capital as a resource not unlike brand
equity, reputation and goodwill, and should be further explored here.
As we indicated at the onset of this essay, social capital fits Barney's (1991)
criteria of the resource-based-view of the firm. Resources that provide a competitive
advantage should be valuable, rare, hard to imitate, and imperfectly substitutable.
Applying these conditions to accounting firms and other professional service
sectors, it appears therefore obvious that the social capital of an audit firm forms a
major source of competitive advantage in this 'knowledge' sector. Social capital of
audit firms has a rent-producing potential, in that it is valuable and scarce (product
market imperfectness) as well as imperfectly tradable (factor market imperfectness).
Araujo and Easton (this volume) employ a similar list when they conceptualize
social capital through a 'relational' lens. Let us review these aspects of social capital
in closer detail in order to reveal their role in explaining the benefits of embedded
ties.
Valuable
As far as the value argument is concerned, a substantial number of studies in
sociology have shown that social ties transfer influence and information (e.g., Burt
1992, 1997; Coleman, Katz, and Menzel 1966). At the individual level, the benefit
of having supportive relations has been welI established. Supportive relations
contribute to getting a job (Granovetter 1995), high compensation (Boxman, De
Graaf, and Flap 1991), and promotion (Burt 1992). We argue that this argument
pertains to the (audit) firm level as well. Burt and Ronchi (1990) and Burt (1992)
applied the notion of social capital to organizations. Burt (1992: 9) pointed out that
'the social capital of people aggregates into the social capital of organizations.'
Social capital amassed in the organization's members is among the firm's most
valuable productive assets (Burt and Ronchi 1990). Unlike the setting of business
groups, in this sector we can define an organization's social capital as the aggregate
of the firm members' social capital. An individual member's social capital is
captured by his connectedness with client sectors.
Why would audit firms with social capital enjoy competitive advantages and
higher survival chances? That is, what is the role of social capital in the economic
transaction of providing audit services? Under perfect competition, social capital
cannot generate any economic rent (Burt 1992). However, the market for auditing
services is hardly perfect, and information about audit services is not costIess. The
owner's social capital strengthens his firm's ability to retain and attract clients. This
is even more true in the audit industry, where information with respect to qualities of
professionals is hardly perfect (cf. Burt 1992; Polodny and Castelluci, this volume).
Clients resort to their social contacts to screen their service providers, because
assessment criteria for auditing quality are hard to come by. Crucial contacts include
those that involve the client sectors that an audit firm serves. There are three reasons
why network ties with client sectors may well facilitate the building and retention of
clientele.
First, people tend to rely on their current social relations to alleviate transaction
cost (Ben-Porath 1980). A stranger who does not anticipate an enduring exchange
relationship, has an incentive to behave opportunistically. To curb this malfeasance,
ill-acquainted exchange partners typically rely on elaborate, explicit, and
comprehensive contracts. These contracts, however, are difficult to write and hard to
enforce (Williamson 1975). Mutual trust between the actors, developed through
repetitive exchanges, obviates the need for writing explicit contracts. If the creation
of trustworthy social relations were costless, however, the existing network ties
would not confer benefits to those who nurtured them. In reality, individuals and
organizations have to invest substantial time and energy in forging durable relations
with others (Burt 1992). Variations in networking among firms should then
contribute to differences in the firms' ability to attract clients. Second, trustworthy
relations produce information benefits for the linked actors (Burt 1992). Information
is not spread evenly across all actors. Rather, its access is contingent upon social
contacts (Coleman et al. 1966; Granovetter 1985). An actor cannot have access to all
relevant information, nor can he process and screen all important information
single-handedly. Being embedded in a network of relations allows a particular actor
to economize on information retrieval. Second hand information, at least, serves to
signal something to be looked into more carefully (Burt 1992). Personal contacts
also make it possible for the involved actors to acquire the information earlier than
others. Third, trustworthy relations enhance the possibility for an actor to refer his
contact person ( for example, an auditor, physician or management consultant) to a
third party (i.e., 'tertius'). Burt (1992: 14) puts the benefit this way: 'You can only be
in a limited number of places within a limited amount of time. Personal contacts get
your name mentioned at the right time in the right place so that opportunities are
presented to you.' The counterpart in a dyadic relation can playa role as a liaison to
link the social actor to third parties.
Scarcity
The argument as to the scarcity issue is, again, specific to the CPA profession. The
CPA profession is there to attest financial outlets of organizations. In effect, this was
the very reason for the origination of the profession. In away, this is comparable to
other public professions. For example, police officers are trained to perform their
public, and legally protected, role of preventing and bringing action against
violations of the civil order. In a similar vein, CPAs are expected to prevent and
bring action against violations of the 'financial order.' Therefore, CPAs are trained to
perform their public attesting role-this is the core of any CPA education program.
This very nature of the profession implies that the majority of CPAs are employed in
public practice, working within audit firms rather than client organizations. Only a
minority is attached to internal control jobs within client organizations. Hence,
social ties that come with current (or previous) partners or associates with previous
(or current) employment outside the audit industry - i.e., through jobs in
governmental bodies or private enterprises - are not abundant. For example, in 1920
roughly 80 % of Dutch CPAs worked in public practice. In the period from the
1960s up until the 1980s, this percentage dropped to slightly above 50 %. Hence,
there is much room for audit firm heterogeneity in this respect, both in time as well
as over time.
Nontradability
Apart from market imperfection (resource value and scarcity), nontradability is
needed to guarantee the sustainability of rent appropriation. Social capital is
tradeable, however, though all but perfectly. Within audit firms, an individual CPA
handles a set of client accounts. That is, from the perspective of the client there is a
double tie to the audit service supplier-i.e., to both the audit firm and the individual
auditor. For one, client loyalty to the audit firm is rather high. This is particularly
true for large companies, which rarely switch from one audit firm to the other
(Langendijk 1990). Among small and medium-sized client firms, audit firm
switching may well be common, though. Additionally, however, a client's financial
reports are attested by an individual CPA. This introduces a tie to the individual
auditor, too. In many cases, the auditor' s position involves confidentiality and trust.
In a way, the auditor develops into a mediator who plays an advisory role in a wide
array of financial and even non-financial issues. So, social ties are partly linked to
the audit firm, and partly to the individual auditor. This implies that by moving to
another firm, an auditor only depreciates part of this social network, because client
sector ties are both an integral part of the firm as well as linked to the trust
relationship with the individual CPA. Of course, the partner-associate distinction is
relevant from the observation that ownership is associated with limited mobility.
Finally, we should mention that during the last half century partnership
contracts have further diminished the portability of social capital. In both the US.
Europe and elsewhere, partnership agreements typically contain a clause that blocks
partners from taking clients with them in the event they leave the firm. Needless to
say, such contractual constraints bolster the non-tradability assumption of a firm's
social capital. Such clauses have also become standard since the second world war
and diminish the mobility of a partner's roster of clients.
In sum, a firm of which partners are tied with potential clients is better
positioned to build clientele since a potential client can 1) actually become the
firm's client, 2) provide valuable information about potential markets, and 3) refer
the firm to other potential clients. These aspects should strengthen a professional
service firm's survival chances
An Empirical Test
An empirical study of the Dutch accounting industry over a period of 110 years
(1880-1990) was used to test the proposition that social capital diminishes the
likelihood of firms getting dissolved. Social capital was proxied by various
measures. For example, 'partner from client sectors' was the proportion of partners
who worked in client sectors (i.e., other industries or governmental agencies). They
are assumed to have more valuable network ties with potential clients than partners
without job experience in client sectors. When departing partners find employment
in client sectors, they are likely to have an affiliation that can utilize their
professional knowledge. 'Controller' and 'chief financial officer' are examples. As a
result, they are likely to be in a position to choose a professional service provider.
Because they have strong incentive to take advantage of their social capital, they are
likely to choose the professional service firm they worked for (Maister 1993; Smigel
1969). To reflect this effect, the study included a 'partner to client sectors' variable.
This is the proportion of partners who left the firm within the previous ten years in
order to work for other industries or governmental agencies. A ten-year span was
adopted for two reasons. First, the strength of network ties may decrease over time,
as the departed partners develop new network ties. Second, the departed partners are
ultimately bound to retire from the business world and thus no longer provide
economic opportunities to the firm. Note that these proxies of social capital derive
from the mobility of professionals who move through a revolving door between two
firms. Much of the social capital literature assumes stationary individuals who link
two or more organizations through overlapping membership, for example,
interlocking directorates.
To test the hypothesis regarding social capital and dissolution, a hazard analysis
Table 2. Complementary log-log regression of firm dissolution

Variables Model 2
b (s.e.)
Intercept -.450 (.934)
Partners 'from' client sectors -.090*** (.033)
Partners 'to' client sectors -.0\3*** (.001)
Heterogeneity in partners' origin -.061 (.062)
Heterogeneity in departed partners' destination .024 (.076)
Partners' industry-specific human capital (Graduate school -.138*** (.040)
education)
Partners' industry-specific human capital (Industry tenure) -.\06** (.050)
Partners' industry-specific human capital (Industry-tenure)2 .145*** (.032)
Partners' firm-specific human capital -.236** (.107)
Partners' firm-specific human capital 2 .226*** (.034)
Log-likelihood (Degrees of Freedom) -2060 (39)

chi-square compared with previous model (d.f.) 118*** (9)
Notes: *p < .10, **p < .05, and ***p < .01 (two-tailed test)
Data: 1851 firms, 8696 firm-intervals, and 1164 firm failures .
Regulatory, historical, industry level (e.g., density), firm level (e.g., size, age) and control
variables not displayed (compare Pennings, Lee, and van Witteloostuyn 1998). Model 2
includes control variables, but not the variables involving associates' human and social capital.
was conducted on these firms, while controlling for numerous other variables (e.g.,
industry level variables such as density, size distribution, history and regulation; and
firm level variables such as firm age and size). The rent producing potential of
human capital is conditional on the firm possessing social capital. Further details are
provided in Pennings, Lee and Witteloostuyn (1998). A partial display of the results
is provided in Table 2.
The results were supportive of the hypotheses. In Table 2 we present the results
involving the human and social capital of owners, i.e. the partners without showing
the simultaneous effects of numerous control and other variables, including those
that are associated with firm and industry characteristics. Consistent with the
hypothesis, all coefficients of the social capital proxies were negative, indicating
that a firm's social capital statistically significantly decreases firm mortality. The
effect was statistically significant for two classes of social capital: ties that derive
from the recruitment of professionals out of the accounting firm's client sectors, as
well as ties that are associated with a firm's 'alumni' who after the tenure in the firm
have moved to client sector firms . The heterogeneity in bundles of social ties, as
derived from the interfirm mobility of professionals did not statistically significantly
affect the firms. When we add the proxies of human and social capital involving
associates to the model of Table 2, it was found that associates' social capital does
not seem to benefit their firm. In short, this study provided some important findings
regarding the beneficial effects of social capital.
What Further Implications Regarding Social Capital?

This chapter illustrates the benefits and drawbacks of social capital that are either
mediated by individuals or become formed through an array of linking vehicles such
as cross-stockholdings and long-term buyer-supplier relationships. We have
suggested that the model of the firm conditions our conceptualization and
operationalization of social capital and the consequences associated with social
capital. Firms are conceived of as unitary actors that interact with other actors (e.g.,
peer firms in business groups), or they can be conceived of as a community of
practices and aggregates of individuals with their distinct objectives and unique
agendas (e.g., professional services firms). Allison's (1971) labels of rational and
political actor correspond with these stylized forms of organization. In the former
case, social capital inheres in the multiplex arrangements that bind a firm to other
actors. In the latter case, we focus on individuals and their ties that aggregate to
organizational social capital. We then set out to review the benefits of social capital
as a distinct organizational (intangible) asset.
Mediated by individuals, social capital nonetheless can be viewed as an
organizational property. The individuals might be stationary (as illustrated by the
linking pin (Likert 1961) or double agents) or they might migrate between firms (as
illustrated by the revolving door syndrome). The relative inclusion of the individual
defines its functionality for information and knowledge transmission: the personal
needs to be available for external linking, yet requires also sufficient proximity to
internal members and groups who can convert the flow of knowledge and other
resources into some competitive advantage.
Individuals can also mediate social capital in the case of business groups. In
fact, some of the pertinent literature has focused on individuals as transmitters of
knowledge between firms they span-for example, so called guest engineers who
are employed by the OEM or its supplier and are assigned to work in the partners'
site, or civil servants who have been recruited by a chaebol firm and join their ranks.
For example, in the above Kukje bankruptcy, it has been suggested that the chaebol
management shunned participation in semi-public sectors such as the I1hae
Foundation, thus depriving themselves from individually mediated social capital.
The Pusan based chaebol neglected to maintain part of its boundary transaction
system. What sets business groups apart from partnerships, among others, is that
business group links are typically mUltiplex, comprising both personal and
impersonal means for maintaining durable links.
In spite of such differences, this chapter has indicated that network
embeddedness can have both positive (social capital) and negative (social liability)
consequences. The links that bind provide access to competitively critical resources,
but they can also be so binding that they are stultifying and rather harmful. The case
of Kukje illustrate the deleterious effects of embedded ness that becomes fractured as
a result of governmental interventions. The inclination of Toyota to reduce its
embeddedness within its keiretzu signals a desire to increase the flow of novel
information that current links cannot furnish; its conventional supplier links might
be too limited in contributing potentially innovative ideas. The negative first order
effect and positive second order effect of social capital on performance in the
apparel industry might be the most robust finding to date regarding the paradox of
embeddedness (Uzzi 1997a).
Uzzi (1997a, this volume) makes the important observation that embeddedness
is a two-edged sword. Embeddedness ranges from 'under-embedded,' via 'integrated'
to 'over-embedded networks.' As was shown, this distinction hinges largely on
whether links are 'arm's length'(i.e., contacts based on selfish, profit seeking
behavior) versus 'embedded' (i.e., contacts based on trust and mutual intimacy). A
firm's network that comprises largely arm's length links does not confer much
advantage in knowledge transfer, coordination, or strategic alignment. Conversely, a
firm that is strongly entrenched in embedded networks might become so insular that
it suspends exposure to markets and technologies that reside outside its immediate
environment.
It appears that these distinctions do not readily map on the two contrasting cases
we presented in this chapter. The partnerships in a professional services sector fit the
conceptual distinctions between arm's length and embeddedness, together with their
functionality such as trust, tacitness of knowledge being transferred, and mutual
adjustment (Thompson 1967) as coordination mode. At face value, partnerships are
internally personalized and anchored in trust, and so we would expect some of the
relationships to be among professionals and their clients. Uzzi's (1997a) case
involves similar Gemeinschaft-like firms, i.e., small entrepreneurial firms, mom and
pop, a trade making up a cottage industry-in short, organizations in which face-to-
face relationships predominate and which often become extended externally. The
apparel world resembles the Chinese 'bamboo network' (Tsui 1997) and Dore's
(1983) description of the Japanese textile industry, which he labels as 'cottage
industry' and in which goodwill becomes the central feature in describing the
prevailing trust and mutuality. The network ties are largely mediated by individuals.
How do we map these descriptions onto the social capital of firms in business
groups that tend to be mUltiplex? Are such links more Gesellschaft-like in their
appearance and functionality? What sort of processes can we envision in a boundary
transaction system in which personal ties complement contracts, equity cross-
holdings, and traditions that outlive their instigators? We should ask such questions
particularly when the individuals in the boundary transacti0n system 'do not go
native,' and continue to link up with people and groups in the firms they span,
together with other elements that define their inter-firm context. The issue is
germane to our earlier review of the firm as a layered entity in which the boundary
spanning system resides largely in the more peripheral bands. Such networks abound
with actors possessing 'structural autonomy' (Burt 1992) and creating opportunities
for opportunism, information asymmetry, and knowledge hoarding--opportunities
which Uzzi considers antithetical to embedded ties.
The implication of these observations is to recognize the two faces of
organizations and to develop divergent frameworks for capturing the performance
implications of network embeddedness. Without forcing us onto a meso-level of
research, by artificially integrating face-to-face and small group dynamics with large
scale firm-interface arrangements, we might develop a middle range theory of social
capital that fits the specific questions we might ask. Whether organizations have at
least two faces, or whether we invoke two cognitive models of organizations might
be an issue left to philosophers and epistomologists. Empirically, we might envision

a continuum in which organizations range from highly cohesive, well bounded
aggregates that are tightly coupled and present few if any intrafrrm hurdles for
coordination, knowledge sharing, and strategic positioning. We can also envision
organizations that are loosely coupled, with permeable boundaries and few isolating
mechanisms, barely holding themselves together and maneuvering on the brink of
dissolution. In either case, the firm is part of a larger context. How they position
themselves onto this continuum, and what image we impose on them remains a
never-ending challenge. The research on social capital will shed further light on how
they negotiate their embeddedness, and what sort of advantages and shortfalls they
derive from that social structure.
We appreciate the comments from Jon Brookfield, Jeff Dyer, Giovanni Gavetti, Jim Lincoln, Lori
Rosenkopf, and Brian Uzzi.
NOTES
1. Embedded ties could have two (if not three) rather divergent meanings: I) ties that are reinforced by
mutual feelings of attachment, reciprocity, and trust; and 2) ties that are a link within a larger set of links
and nodes. Since Uzzi's work is confined to dyads, the first meaning applies. When members of a dyad
become affected by third parties who envelop their tie, as in Burt (1992), the second meaning applies. In
both cases, the concern is with a focal person. If one moves to an even higher level of analysis 3), as for
example the internet, transactions among textile traders in 15th century Florence and Flanders, or
community power structures, then the network takes primacy over the ties between individuals who are
embedded in those networks. A person's or firm's 'centrality' conveys relative access to other actors in the
network such that a focal actor's social capital hinges partly on the direct and indirect ties that the tied
partners possess (e.g., Levine 1972). Empirically the effect of centrality on firm behavior or performance
has not been studied adequately (an exception is Freeman, this volume).
2. Note that the rational model of the firm does not presume anything about its embeddedness here. In
either the rational or political scenario, we do not assume organizations to behave as if they are atomized
from the impact of their relations with other organizations, or from the past history of these relations. If
we were to extend methodological individualism to the embeddedness of firms, we would not be able to
furnish an adequate account of how firms' actions combine up to the level of the value chain, markets or
institutions. We only make the analytical distinction based on the relative saliency of aggregation when
examining social capital as a firm-specific asset.
Hence, our reluctance to include Allison's second model, the 'organizational actor model' in our review.
In the extreme, over-socialized individuals would reduce to mentally programmed automatons who
mechanically replicate the routines that the organizational socialization process has imprinted onto them.
As role incumbent, they would have no discretion to embellish their position or protect personal interests,
nor could they be construed as the personal authors of their social network.
3. Some examples might illustrate the issues at hand. Firms are tied to each other through trade
associations, business groups, consortia, cartels, joint ventures, and directors who sit not only on their
board but on the boards of other organizations as well. They are locked into licensing agreements and
long term supplier-buyer arrangements, and might have made significant investments in specific inter-
firm relationships. The presence of such links and their benefits seem obvious, when that capital is treated
as firm level or individual level phenomena.
For example, Boeing's 747 aircraft requires the input from numerous contractors and sub-
contractors-only certain chunks of the cockpit and wings are developed and produced by Boeing. Such
inter-firm transactions result in long term links that become independent of the members who forged
them originally. Many firms occupy positions in the value chain with interdependencies so dense that one
might consider the value chain to be a more salient unit of action than the firms that exist within the value
chain. A simple illustration from the computer industry might further illustrate this observation.
During the main-frame computer era, it was common for firms like mM and Hitachi to control all
the steps in the value chain, from silicon, computer platform, system software, application to distribution
and service. The firm was the value chain, and competition between corporations matched competition
between value chains. In the late nineties, we observe a fragmented horizontal competition between finns,
but vertically dense complementarities have surfaced. Microsoft competes with Apple and Unix, but is
symbiotically linked with upstream PC manufacturers and their suppliers, such as Intel. Downstream, the
finns relate to distribution and service finns such as computer stores and mail order firms. Microsoft has
been a shrewd exploiter of network externalities: the various technologies require complementary
products, lead to the fonnation of virtuous cycles such as software developers writing more Microsoft
Windows applications, and when these become available, more customers adopting Microsoft Windows.
Increasingly all firms in the value chain become 'locked-in' (or locked-out!) resulting in a complex string
of links that are straddled around a dominant computer design (e.g., Yoffie 1996). In such a value chain,
links are often de-personalized and it is the organizations that become the salient unit of the network. The
ties in such networks are critical for the finns involved as their products and technologies become heavily
intertwined with those of others.
Much of the social capital literature has an individual slant (e.g., Burt 1997) and finn attributes have
often been examined as an individual manifestation. Burt's (1997) recent study examines investment
banks but really focuses on its traders and the 'structural holes' that benefit the size of their perfonnance
based bonuses. One might also focus on their banks' tombstones and the social capital that could be
inferred from them. Coleman's (1988) classic example involves the tight social circle of diamond traders
in New York whose smooth and paperless transactions hinge on the social ties that they maintain with
other traders. The trust that is sustained within such a network results in a substantial reduction of
transaction costs. Likewise, he (Coleman 1988) shows that children whose parents know other parents
and teachers are better embedded in their school community and show lower dropping-out rate. Finally,
Uzzi (1997a) recounts the linkages among individuals who make up the New York apparel industry. In
such instances, the issue of aggregation and presumption of finn as a unitary actor is rather moot: the
entrepreneur is the finn . In these and many other contributions, social capital is a resource that belongs to
the networking or interacting individuals and that might affect the venture with which the embedded
individual is associated.
4. By the same token, an individual who is neutral to the bridging between two finns cannot easily be
incorporated in the organization's social capital. Referee, arbitrator, or mediator roles are sharply
different from those we associate with ambassador, spy, or guest engineer. The fonner's neutrality might
depreciate or sanitize whatever infonnation or knowledge the 'middle-man' furnishes to the linked
organizations. His neutrality also precludes intimacy and creates social distance. We assume that
organizations have discrete bundles of knowledge and infonnation whose rents will be augmented by the
development of 'proprietary' social capital.
5. Sherer (1995) identifies three major types of employment relationships. The first is the employment
relation coupled with ownership. It includes employees who share the risk of organization via various
incentive systems which link their earnings to the perfonnance of the organization. Employees in that
relation constitute the core group in our analysis. The second is the traditionally described employment
relation in which employees receive a fixed amount of earnings, provide a fixed length of time, and
perfonn work based on the direction from the supervisor or job description. Employees in these types are
designated regular group in the present discussion. The third embodies relationships that involve
temporary employment or contracting out. Employees in this type fonn the temporary or marginal group.
Note that with the rise of temporary employment agencies, outsourcing and sub-contracting, this latter
group has acquired huge proportions. Analogous distinctions have been made by Jensen and Meckling
(1976) and Milgrom and Roberts (1992).
6. The classification was suggested by Jon Brookfield.
7. For example, Toshiba and Tokyo Power maintain close buyer-seller relationships; they both draw
graduates from Tokyo University who get promoted in their respective companies, and they move in
tandem, their roles might change but their mutuality stays intact. The demography of the system co-
evolves with that of the respective organizations. Such evolutionary arrangements ensure network
continuity throughout the firms' history.
8. The tenn transitive cross-equity holding refers to a string of keiretzus finns between which
ownership is mutual yet unequal. Nishiyama (1982) reports the pattern of large block holdings in the
Sumitomo Business Group, with Sumimoto (S.) Life Insurance owning a larger percent of shares in S.
Bank, S. Metal, S. Chemical, S. Electric, etc. than vice versa; it augments its power over these finns
because these finns in tum own shares in each other, such that cumulatively, S. Life Insurance scores
highest on the 'comprehensive power index.'
A Relational Resource
3
Perspective on Social Capital
•
Luis Araujo
Geoff Easton
ABSTRACT
This chapter reviews the notion of social capital from a resource based perspective.
We argue that the notion of social capital relies on a metaphorical mapping of
features associated with economic notions of capital or assets into the social domain.
We start from the notion that not all economic resources can be classified as assets
in the way the term is deployed within legal and accounting language, and argue that
social capital shares many features with other less understood and intangible
resources. By employing a framework to examine the multifaceted and relational
dimension of resources, we examine in detail the entailments of the social capital
metaphor and relate to current applications within the business and management
literature. We conclude by reflecting on the characteristics of social capital as an
economic resource and caution against the dangers of engaging in facile
prescriptions based on a cursory understanding of the logic of accumulation and use
of social capital.
INTRODUCTION
The notions of embedded ness (Granovetter 1985), social capital (Bourdieu 1986;
Coleman 1988, 1990), and social resources (Lin 1990) have, in recent times,
contributed to a rekindling of interest in the interaction between the economy and
society. At the heart of this revival is an attempt to trace the mutual influences
between economic exchange and the social structures in which the economy is
embedded.
The notion of embeddedness relies on the insight that economic life is shaped
and constrained by norms, social networks, institutions, and a variety of motives
A Relational Resource Perspective on Social Capital - 69
other than the unconstrained pursuit of self-interest. The revival of the interest in the
social has also got to do with the increasing awareness that modern economies,
while relying on impersonal forms of exchange and complex forms of contracting
with third-party enforcement, cannot dispense with other forms of support in the
guise of moral, social rules or codes of conduct (Platteau 1994ab).
The emergence of market economies does not diminish the need for social
solidarity and trust. On the contrary, as Macneil (1986: 592) argues, market
economies have acute problems regarding social solidarity. The embeddedness of
economic exchange in social structures very often dictates complex legal structures
remote from, though essential to, the exchange relations themselves. Thus markets
cannot be regarded as a spontaneous order or a primitive state of nature, but a
convergent network of actors and institutions mixing different forms of exchange
and where order is generated through translation processes and rules that are
reproduced across exchanges and over time.
Market order is partially generated by institution building, to establish and
enforce sanction systems and solve coordination problems involving the risk of free-
ridership and dilemmas of collective action. But, as Bates (1988) argues, formal
rules enforced through third parties can also be subjected to free-riding, and the role
of institutions as impartial rule makers and enforcers can be questioned. The
problem cannot be resolved through appeal to a further tier of institutions to monitor
the performance of the first tier. In short, order can emerge only in the presence of
both institutions promoting and enforcing formal rules, and informal norms such as
a generalized morality that draws on a society'S social fabric and culture. These
informal norms can thus act as substitute for or a reinforcement of formal rules and
control mechanisms, with the consequence that coercive enforcement of formal
norms becomes either redundant or of secondary importance. 1
The objective of this chapter is to revisit some of the issues concerning the ways
social structures impact on economic exchange. In particular, we are concerned with
the ways the notion of social capital, residing both in concrete, interpersonal
relationships inside and outside formal organizations as well as in wider social
structures, can be deployed to break down some of the artificial divisions between
the economy and society. The structure of the chapter is as follows: in the first part
we look at the notions of embedded ness and social capital and the way they have
been used to understand the coordination of socioeconomic life. In the second part
of this chapter we use a framework we developed to dimensionalize economic
resources (Easton and Araujo 1996), to look at the characteristics of social capital as
an economic resource. We conclude with some speculations on the role of social
capital in the coordination of socioeconomic life.
The Notion of Embeddedness

The key issue that arises in discussions of embeddedness and that is germane to the
notion of social capital is the extent and form of the embedded ness that we might
expect to find in a modern economy. Below we review briefly some of the more
recent offerings to give a flavor of the debate.
The embedded ness argument is given short shrift in traditional neoclassical,
undersocialized conceptions of human action and also in neoinstitutionalist
arguments that, although acknowledging the role of social networks and private
orders borne out of repeated social contacts, still regard the role of abstract, formal
rules as solely responsible for market order (Platteau 1994a). Greifs (1994) analysis
of the contrasting solutions adopted by Maghribi and Genoese traders in the eleventh
century to trade expansion demonstrates the advantages and limitations of
embedded ness as a mechanism for governing economic life. The collectivist system
adopted by Maghribi traders, where order was enforced through moral sanctions,
worked well in the case of intraeconomy agency relationships, but was incapable of
supporting intereconomy relationships and of allowing for the division of labour
necessary to take advantage of new trade opportunities. By contrast, the Genoese
introduced formal enforcement institutions to support impersonal forms of economic
exchange and promote further division of labour, thus enabling their society to
capture the efficiency gains stemming from the expansion of trade.
Hardin (1993: 510) regards the thick relationships that the embeddedness
argument prescribes as yielding only a part of the knowledge we have of others. But,
of course, one might learn from the experience of others, through reputational
effects and a variety of other indirect means. As Hardin (1996: 31) argues, there are
two modal categories of controls operating in society. There are geographical
associates-the group of friends, family, and associates with whom one is inevitably
bound up in repeated interactions and long-term relationships. And there are the
elaborate large-scale controls associated with institutions, such as the legal system,
relying on formal rules and coercive enforcement. But between these two modal
categories there are a variety of mixed devices such as broad social norms, mixing
elements of both modal categories, that provide important elements of social control
(Hardin 1996).2
In his oft quoted critique of conceptions of human action in economics and
sociology, Granovetter (1985) rejects both notions of undersocialized actors as
behaving atomistically and oversocialized views of human action, where actors
simply follow a script attached to the intersection of the social categories to which
they belong. Instead, Granovetter (1985: 490) revives Polanyi's (1957) notion of
embeddedness and stresses 'the role of concrete personal relations and structures (or
'networks') of such relations in generating trust and avoiding malfeasance.'
Granovetter's argument revolves around the notion that the production of trust in
economic life is mainly accounted for by concrete social relations rather than
institutional arrangements or norms of generalized morality.
He further argues, however, that the existence of strong social relationships may
contribute both to the production of trust and trustworthy behavior as well as,
perversely, mistrust and malfeasance. In short, embedded ness can both contribute to
the resolution and the collective dilemma implied by the Hobbesian position as well
as introduce the possibility of disruption on a larger scale than the one that is
possible in a truly atomized, state-of-nature social situation (Grano vetter 1985: 493).
Zukin and DiMaggio (1990: 15) elaborate on the notion of embedded ness by
defining it as the contingent nature of economic action with respect to cognition,
culture, social structure, and political institutions. Cognitive embedded ness is
defined as the equivalent of bounded rationality, the set of heuristics and biases that
pervade all forms of reasoning. Cultural embedded ness refers to yet more limitations
on economic rationality imposed by the constraints of shared values and

understandings in shaping economic goals and strategies. Structural embedded ness
is for Zukin and DiMaggio (1990: 18) what embedded ness tout court is for
Granovetter-the contextualization of economic exchange in ongoing patterns of
social relationships. Finally, political embeddedness refers to the power struggles in
which different types of actors (e.g., business firms, institutional actors, and the
state) vie for the power to shape the rules that govern economic life.3
Uzzi (l996a, 1997a) elaborated the concept of structural embeddedness
describing it as a specific logic of exchange or coordination in which trust, borne out
of repeated interactions and the prospect of a continuing relationship, pushes the
logic of calculativeness and monitoring to a secondary role. In this mode of
coordination, thick relationships provide more fine-grained, tacit, and holistic
information transfer with a significant and positive impact on problem solving and
conflict-resolution arrangements as well as innovation. 4 Uzzi's conclusions are
largely unsurprising given the plethora of studies over the last twenty years on
buyer-supplier relationships in Europe (Hakansson 1982; Axelsson and Easton
1992; HAkansson and Snehota 1995), the U.S. (Helper 1991), and Japan (Sako 1991;
Smitka 1991 ; Nishiguchi 1994). Furthermore, the literature on flexible specialization
and industrial districts has long since underlined the role of embeddedness in
promoting flexibility, innovation, and adaptability to changing demands as
alternatives to mass production of standardized goods (Lazerson 1988, 1995).
One central conclusion that can be drawn from this brief review is that the form
and extent of embeddedness seen in an economy is strongly influenced by the
discipline and the school of thought the writer subscribes to. While this phenomenon
is a commonplace, as any sociologist of knowledge would acknowledge, what is less
usual is that one line of attack on the issue is the result of the borrowing of a concept
from one of these disciplines for use in another.
Economics has been described as the study of the allocation of resources,
sociology as the study of human aggregations and their behavior. Clearly, these
definitions already imply an overlap. Resources cannot be allocated in an economic
system without the intervention of human agency. Nor can the behavior of social
systems be unaffected by the resources they create, consume, and exchange, directly
or indirectly. However the crucial link here is that of resource. What has happened is
that sociologists have appropriated the concept of resource, or more narrowly
capital, from economics and have used it to capture the notion of the embeddedness
of economic behavior, itself described in terms of resources, in social structures by
way of the notion of social capital. While the basic idea has been around for a long
time, it is only more recently that it has become popular.
Social Capital
The notion of social capital is an intriguing one and raises interesting possibilities as
a counterintuitive metaphor, mapping an economic domain where the notion of
capital is well established and tying accounting and legal conventions to a social
domain, where accounting for value, investment, depreciation, and so on poses a
number of difficulties. One of the functions of metaphors in theory building is an
exploratory one (Easton and Araujo 1993). This is not much removed from an
explanatory function, but it emphasizes research directions and possibilities rather

than assumes a comprehensive mapping of the base into the target domain. Before
proceeding to explore the relationship between the base and target domains, we
focus briefly on the notion of social capital that has been used in the social sciences
and the business literature. Portes and Sensenbrenner (1993), while recognizing the
value of Granovetter' s embedded ness argument, make a case for focusing on the
more manageable concept of social capital as developed by Pierre Bourdieu in the
Francophone world and by James Coleman in the Anglo-Saxon literature.
In Bourdieu's hands the notion of social capital is deployed in a rather broad
and undifferentiated sense. Bourdieu (1986: 249) defines social capital as 'the
aggregate of the actual or potential resources that are linked to a possession of a
durable network of more or less institutionalized relationships of mutual
acquaintance and recognition ... a 'credential' which entitles them to credit, in the
various senses of the word.' Further on, Bourdieu characterizes these connections or
obligations as the product of investment strategies aimed at establishing or
reproducing social relationships that can be leveraged either in the short or long run.
Different forms of capital (e.g., cultural or political) derive their forms from the
fields in which they are deployed, but the notion of economic capital remains the
master metaphor for all other forms of capital.
Coleman (1988) starts from the rationalist assumption that each actor has
control over certain resources and that social capital is one type of resource available
to an actor. Coleman (1990: 302) distinguishes between physical, human, and social
capital as three separate forms of capital, all of them having a potential to facilitate
action, or as Coleman puts it 'social capital is productive, making possible the
achievement of certain ends that would not be attainable in its absence.'5 If physical
capital is regarded as wholly tangible, being embodied in observable physical forms,
and human capital is less tangible but still embodied in the skills and knowledge
possessed by an individual, social capital is even less tangible and nonlocalizable for
it is a property of social relationships and social structures.
On closer inspection, Coleman's notion of social capital encompasses both
properties of dyadic relationships and the social structures in which these
relationships are embedded. Social capital takes the following forms : obligations,
expectations and trustworthiness of structures, information channels, norms, and
effective sanctions. So, for example, when in a dyadic relationship one of the parties
does a favor for the other, there is an expectation of reciprocity at some unspecified
future date. Coleman (1990: 306) refers to this situation as one of the parties issuing
a credit slip that can be called on at a future date. But other forms of social capital
are clearly properties of social structure that impinge on and constrain specific,
dyadic relationships. An example of this would be a moral norm prescribing the
foregoing of self-interest in favor of collective welfare.6
Coleman (1990: 317) also argues that the public good aspect of most forms of
social capital is, perhaps, its most distinctive feature vis a vis other forms of capital.
As in Olson' s (1965) collective action dilemma, self-interested individuals free-ride
or seek to reap the benefits of public goods while evading the costs of participation
in their production, and so rational actors will consequently underinvest in the
production of these public goods.7 Many of the benefits flowing from actions that
create social capital can be enjoyed by people other than the ones who have
contributed to those activities in the first place. In short, those who produce them
cannot necessarily appropriate the benefits flowing from investing in the creation of
social capital.
Smart (1993: 393) elaborates on the problems concerning the use of the term
capital and offers a few interesting avenues to preserve the distinctions Bourdieu
makes while clarifying the interrelationship between different forms of capital.
Smart notes that social capital by its very nature is vague and unmeasurable: it lacks
a currency and a space of calculation where debits and credits can be accumulated
and compared. An obligation, for example, becomes concrete only once it is
liquidated, and until then there is no certainty that will ever be reciprocated. For
Smart (1993: 393), if social capital cannot be possessed but can be converted into
other forms of capital, then it is entirely contingent on the reproduction of the social
structures in which it is embedded. If social debts-such as obligations-can be
recovered through enforcement by third parties then, according to Smart, we are
talking about economic and not social capital. 8
Smart relies on a similar logic to argue that social capital is a resource that
resides in dyadic, specific social ties and that more generalized resources such as
honor or reputation that are valued within society or subgroups within it, are best
characterised as symbolic capital. Smart goes on to distinguish between Bourdieu's
notion of capital-in-general and the notion of power. Whereas power is seen as
resting on the authority to command the actions of others, capital in the sense
Bourdieu uses it, is the ability to induce others to act in one's interests through the
leverage of resources available to the agent. Smart (1993: 394) distinguishes
between the different forms of capital invoked by Bourdieu in the following way:
Economic capital involves ownership of objects, but property ownership entails claims
that others may not interfere with your property without your permission, and exclusive
ownership may be used to induce others to act in particular ways (such as hiring or firing
them). Cultural capital is a claim to having the ability to engage in certain types of
practices, and in the strongest forms it accords a monopoly over such practices (for
instance, medical doctors or accountants). Symbolic capital involves claims by the
possessor that he or she be treated in particular ways by classes of others. Social capital
consists of claims to reciprocation and solidarity from particular others. What is
fundamental to social capital. however. is that explicit claims are normally excluded
from the performances within which they are made, so that power over the action of
others is radically distinct from exercises of power utilizing the discourse and apparatus
of command. (emphasis added)
Smart uses the example of gift giving and guanxi in China as an instance in where
explicit recognition of instrumental goals is excluded from the performance, and
incompetent performances results in loss of face and dissipation of the very outcome
that the performance intended to achieve. Gift exchange must, by its very nature, be
devoid of any explicit reference to a calculative logic, lest they be devalued as failed
gift performances or understood as a bribe. In essence, Smart's reinterpretation of
Bourdieu's social capital consists of locating it within concrete, dyadic relationships
and insisting on the absence or accomplished concealment of explicit claims by the
possessors of social capital when attempting to cash in their credits as well as the
nonexistence of third-party enforcement to reclaim bad debts. Although we are less

keen on the fine distinction that Smart adopts between social and symbolic capital,
his definition has at least the advantage of relating social capital to contextual
performances that have to be interpreted in the light of experience and further cues.
What we conclude from this discussion is that while there is agreement that
there are different forms of social capital, there is disagreement as to what
conceptualizations of these forms might be most useful. One of the ways we can
explore the notion of social capital is by examining in more detail how the base
domain of the metaphor (economic capital, or more generally economic resources)
map into the object domain (social interaction). In the next section we briefly review
the notion of economic resource before proceeding to examine the properties of
social capital as an economic resource.
The Notion of Economic Resources

The notion of resource has recently been deployed extensively within the strategic
management literature under the guise of the resource-based theory of the firm.
These developments can be traced back to the pioneering work of Penrose (1959)
and pursued in different guises since Wernerfelt (1984).
Resources, in the traditional language of business strategy, are strengths, and
there is little attempt here to suggest how and why they become so and how that
might relate to one another. Despite its growing popularity, the resource-based
theory of the firm has come under attack for its lack of a clear conceptualization of
what constitutes a resource. In the absence of a clear definition, we are left with a
tautology: firms must discover which resources contribute most to their competitive
advantage and develop or acquire more of these (Porter 1991).
This chapter draws on a specific notion of economic resources, inspired by the
industrial networks research program into interorganizational relationships-see
Axelsson and Easton (1992) and Hakansson and Snehota (1995) for recent
examples. We define economic resource as any entity that can be deployed by an
actor that is capable of continuing independent existence, has futurity, and can, or
may, meet an economic need. In this sense, an economic resource is a stock that can
be drawn on either through ownership and control or through indirect access.
Resources stand in contrast to activities. Activities comprise current actions and
have no direct continuity. Resources have a durable character, even if they are the
products of past activities, and acquire their meaning and value when mobilized
within specific activity structures.
We make a further distinction between the generic notion of economic resources
and the terms asset or capital as used in management accounting theory. We reserve
the use of the term asset or capital to describe a resource whose ownership, control,
and market valuation can be determined. Thus assets are simply a subset of
economic resources protected by property rights enforceable by third parties, whose
valuation has some claims to be measurable which, in turn implies the construction
of calculable spaces and technologies able to enforce standardized measures of value
(Miller 1994).
An obvious corollary to our distinction between resources and assets is that we
are not constrained by legal and accounting definitions of the firm. We draw on the
early work of Chamberlain (1968) to emphasize the distinction between the firm as
an entity defined for legal and accounting purposes and the firm as an entity for
coordinating a set of business activities, a key aspect of the industrial networks
approach. Chamberlain employs a broad notion of asset and argues that most of the
assets that constitute the firm's instruments of action are not the ones described on
its balance sheet. In other words, Chamberlain does not assume that a firm's
capability for strategic action resides within the boundaries of what it owns and
controls. Instead the firm's capabilities are seen as embodied in evolving networks
of interdependence both within the firm and with aspects of its environment. A
firm's strategic capability depends on two factors: 1) its capacity to generate
resources from the its current operations and 2) its capacity to mobilize support and
resources from entities and institutions within its environment. More recently,
Barney (1991) argues that a wide variety of the firm's resources may be complex to
the extent that they reside in relationships among people and may be socially
constructed. Among these Barney includes the interpersonal relations among
managers, a firm's culture, and the firm's reputation among suppliers and customers.
And, from an industrial networks perspective, we would add social bonds developed
in the course of economic exchange relationships between two organizations
(Hakansson and Snehota 1995).9
The above discussion has some relevance to the use of the concept of capital-in-
general and, in particular, the distinctions between different forms of capital.
Economic capital, in the cursory way most authors in the social capital literature
seem to understand it, is only a small subset of all the economic resources that are
necessary to conduct economic life. Social capital patently lacks most of the
characteristics that define economic capital. In particular, the absence of accounting
conventions, property rights enforceable through third parties and a system of
exchange devalue the metaphor somewhat. But, on the other hand, it shares many
other features with other economic resources that cannot be qualified as capital, in
the way defined above.
In the following section we attempt to compare the notion of social capital with
that of an economic resource by relying on the framework adapted from Easton and
Araujo (1996). The adaptation consists in changing the focus from resources as an
organization-based phenomenon to resources as a property of intra and
interorganizational, individual-based dyads and social networks.
Social Capital as an Economic Resource

This framework dimensionalizes resources in four separate categories relating to
their existence, evaluation, relationship to actors, and relationship to other activities
and resources. These dimensions are further subdivided into separate categories each
(see Figure 1). Pennings and Lee (this volume) employ a similar list in their
discussion of the benefits of embedded ties for audit firms.
•
Creation
Existence Depreciation
Durability
•
Controllability
Relationships to actors Accessibility
Tradeability
Valuation
Evaluation
• Evaluability
Scarcity
Value
Integrity
Versatility
Relationships to other activities and resources
Complementarity
Understandability
Figure 1. Dimensions of resources (adapted from Easton and Araujo 1996)
The Existence of Social Capital

Starting with the existence group, the processes involved in the creation,
depreciation, and durability of social capital are difficult to understand, and there is
no obvious causal mechanisms that relate investment to a logic of accumulation of
social capital. The same is true of other intangible resources such as organizational
competencies in producing innovative products, for example.
Coleman (1990: 317-318) argues that much social capital arises or disappears as
a by-product of other activities and without anyone's willing it into or out of being.
Furthermore, some of the conditions that may foster the creation of social capital can
also contribute to its destruction. The closure of social networks, for example, where
judgements of trustworthiness may depend on intersecting patterns of relationships
and transitive judgments (A trusts C, because A trusts Band B trusts C) may lead
both to inflationary and deflationary spirals of trust placement in a community
(Coleman 1990: 318). Stability in social structures may become an important
precondition for the creation and maintenance of social capital, although strangely
Coleman makes an exception for formal organizations where social capital is
assumed to be vested in role occupancy rather than on the identity of the occupant. 10
Finally, Coleman (1990: 321) argues that all factors that help to minimize social
interaction and increase individuals' self-reliance, such as affluence or public aid,
tend to contribute to the depreciation of the stock of social capital, through
bypassing the very activities-e.g., the exchange of obligations and favors-through
which social capital is produced and maintained. Like other economic resources
such as reputations or brand names, social capital is depreciated through the lack of
use, and only through use can its stock be maintained and rejuvenated.
Coleman's account of the creation, destruction, and maintenance of social
capital undervalues instrumental actions directed at creating social capital and elides
some of the dilemmas we alluded to earlier-namely the relationship between social
capital as a property of dyads, small-knit face-to-face groups, or wider social
structures. For example, Burt's (1992, 1997) notion of structural holes describes
how social capital is created and leveraged as a result of brokerage opportunities in a
social network. Uzzi's (1996a, 1997a) account of the New York apparel industry
documents how social capital is created by one-sided investments in business

relationships anticipating future reciprocity.
Pennings and Lee and Lazega (this volume) discuss the relationship of dyad-
specific social capital to wider structures at intra- and interorganizational levels.
Putnam (I993a), in his celebrated work on modern Italy, elaborates on this point in
the case of trust at the societal level. lI For Putnam (1993a: 171-176) the
generalization of trust from closed social networks to the wider society can be traced
back to two sources: norms of generalized reciprocity and networks of civic
engagement. Norms of generalized reciprocity, or diffuse reciprocity to use
Keohane's (1986) expression, are seen as a catalyst for the development of social
capital. As in Gouldner's (1960) account of reciprocity, social capital is built not
solely on the basis of exchanges of obligations that are perfectly balanced but on
mechanisms that induce further debts and credits of obligations and avoid the
expectation that all exchanges will ever return to a zero state in which none of the
parties is indebted to the other.
Keohane (1986) makes an important connection between the norm of specific
reciprocity in dyadic relationships, where the exchange of obligations and duties is
roughly equivalent and performed in a strictly delimited sequence, and the norm of
diffused reciprocity, where equivalence is less precise and sequence of events less
clearly bounded. The repayment of debts and obligations in specific reciprocity
situations may lead actors to take a broader view of their interests and engage in
diffuse reciprocity. Conversely, the decay of diffuse reciprocity may lead actors to
revert to exchanges narrowly bounded by specific reciprocity. As Keohane (1986:
25) puts it: 'specific and diffuse reciprocity are closely interrelated. They can be
located on a continuum, although the relationships between them are as much
dialectical as linear.'12
In short, the creation and maintenance of social capital as an economic resource
is often a by-product of distributed activities and actor orientations that may be
reinforced or hindered by the institutional context in which these activities are
embedded. Whereas the role of institutions and formal rules has been underlined in
helping to supersede collective action dilemmas, by providing a stable background
of enforceable norms and duties and thus lowering the risks involved in engaging in
cooperative relationships, less has been made of their role in destroying social
capital.
Recently, Pi Ides (1996) has argued that the intervention of the state through the
provision and enforcement of legislation can destroy social capital through an
insufficient appreciation of the role of social capital that underwrites the successful
enforcement of laws. Pi Ides (1996) advances an argument on the role of the state in
contributing to the destruction of social capital that can be extended to all attempts at
formalizing rules. There are three ways in which social capital can be destroyed
through state intervention: 1) indirectly, by attacking the structural foundations on
which social capital can be created and maintained-e.g., by prescribing spaces in
which social interaction mayor may not take place, 2) directly, by attacking norms
of reciprocity through efforts to rationalise conventions, and 3) indirectly, by
attempting to incorporate social norms into legislation without taking due account of
the role of remedial flexibility in the enforcement of social norms.
The Evaluation of Social Capital

In this section, we attempt to relate the notion of social capital to four dimensions
associated with the evaluation of economic resources: valuation, evaluability,
scarcity, and value (positive/negative) (Easton and Araujo 1996). All economic
resources are generally held to possess a value, even if that value cannot easily be
expressed in the metric of money. More often than not, it is the intangible and more
difficult to evaluate resources that constitute the most valuable resources a firm
possesses (Itami 1987). Furthermore. the value of a resource is intrinsically tied to
the context in which it is used. or to put it differently, to describe the content and
value of a resource is also to describe its context of use.
Social capital is usually hailed as an important resource for promoting both
effective government and prosperous economies. In particular, trust has been singled
out as a key resource in promoting economic growth and competitiveness and as
having a positive impact in intraorganizational relationships (Breton and Wintrobe
1982; Miller 1992) as well as in interorganizational contexts-from buyer-supplier
relationships (Barney and Hansen 1994; Nooteboom 1996) to industrial districts
(Lorenz 1993) or the relationship between capital market institutions (Neu 1991).
In general. as Gambetta (1988a: 221) observes: 'societies that rely heavily on the
use of force are likely to be less efficient, more costly, more unpleasant than those
where trust is maintained by other means. In the former. resources tend to be
diverted away from economic undertakings and spent in coercion. surveillance, and
information gathering and less incentive is found to engage in cooperative activities.'
Social capital can thus be regarded as a valuable. infrastructural resource to
economic life but one that cannot be evaluated easily. Sabel (1995) contrasts two
opposite views on this issue, one inspired by Hayek and the other by Durkheim.
Whereas the Hayekian view stresses the distributed and partial nature of knowledge
in society. and the way norms evolve as a spontaneous. unintended outcome
(Bianchi 1994), the Durkheimian perspective stresses the role of institutions in
curbing the freedom of rational. selfish agents to act as they see fit. From a
Durkheimian perspective, the economy can perform well only if is embedded in a
well-integrated society and that society is willing and able to impose normative
constraints on the pursuit of self-interest (Streeck 1997: 199).
The role of social capital as an economic resource is clearly less important under
a Hayekian perspective. For Hayek (1978) the market with the price mechanism,
sets of property rights. and enforceable rules is the best example of an evolved
institution. Social institutions are instrumental, problem-solving mechanisms that
evolve spontaneously under the pressure to find ways to coordinate the activities of
individuals facing a complex world equipped with limited knowledge. Under a
Durkheimian perspective the value of social <;apital to economic life is much clearer.
For example. social constraints on the pursuit of self-interest in the form of moral
norms or formal laws can reinforce trust and thus increase the efficiency of
economic transactions by reassuring potentially suspicious parties of continued
adherence to norms of reciprocity and normal business practice. As Streeck (1997:
202) puts it: 'Credible information that the other side has noneconomic in addition to
economic reasons not to defect accelerates and consolidates the growth of trustful
relations. '
But if a Durkheimian perspective exalts the value of social capital in promoting
economic efficiency, it also warns about the dangers of attempting to design social
institutions with the sole purpose of economic efficiency in mind. Streeck (1997:
217) warns that 'the kind of social embedding good economic performance requires
can be built only for reasons other than good economic performance, enabling it to
support rational-economic action by containing it.'
These considerations lead us to reflect on two other dimensions of the
evaluation of social capital: scarcity and value (positive/negative). The issue of
scarcity of social capital is important and related to the issue of valuation. As a
resource, most forms of social capital are not scarce in the sense that it is a stock
depleted through use. On the contrary, social capital depreciates through lack of use
and appreciates through extensive use and imbalanced exchanges. For Putnam
(1993a: 178-179) both reciprocity/trust and dependence/exploitation can become
stable equilibria in particular communities depending on initial conditions-namely,
stocks of social capital and paths of evolution. As in North (1990), path dependence
is deemed responsible for the self-reinforcing nature of a particular equilibrium.
Distrust, for example, once set in has the capacity to become self-fulfilling and lock
a particular community into a low-level equilibrium of authoritarian government,
patron-client relationships or putting them at the mercy of criminal organizations
(Gambetta 1988a).
Sabel and Zeitlin (1997) effectively reject the notion of path-dependence and
claim instead that actors are able through reflexivity and choice to transform the
conditions under which they live and to be less discriminating between the two
equilibria Putnam describes. Economies often contain what Sabel and Zeitlin (1997)
describe as moral borderlands where moral rules are only partly observed or
significantly relaxed. In short, the cycles of trust and mistrust that Putnam describes
as being the resultant of initial endowments of social capital and path-dependent
evolution may be less sharply defined and delimited in historical epochs and equally
dependent on stocks of social capital.
The issue of the scarcity of social capital as an initial condition to foster
economic development has been hotly debated for some time. Recently, Evans
(1996) reflects on the role of social capital in promoting economic development and
suggests that endowments of social capital cannot be seen as the major obstacle to
foster development projects in the Third World. In most Third World communities
prior endowments of social capital, in the form of social networks, norms of
solidarity and trust at the micro-level are seen as adequate, but the difficulty lies in
scaling up such capital to generate solidarity and diffuse norms at a politically and
economically efficacious level.
Lastly, we must consider the valence of social capital as a resource. Throughout
most of this chapter we have implied that social capital is an asset to political and
economic life, rather than a liability. But, as Portes and Sensenbrenner (1993)
remind us, the same social mechanisms that account for useful resources
appropriable at the individual, dyadic, or community level may also set important
limits to action and economic development.
Gargiulo and Benassi (this volume) illustrate how social capital can, in the face
of changing circumstances, turn into social liability constraining opportunities for
organizational change. Recent research on entrepreneurship and immigrant
communities in the U.S. documents both the negative and positive aspects of social
capital and embeddedness. Portes and Sensenbrenner (1993) argue that the greater
the social capital produced by solidarity and community ties, the more likely that
particularistic demands will be imposed on successful entrepreneurs, thus limiting
the possibilities of individual expression and the expansion of opportunities outside
the community boundaries. Waldinger (1995), in his study of immigrant
communities in the New York construction industry, argues that in this instance, the
embedded ness of economic behavior in ongoing social relations among a myriad of
social actors impedes access to outsiders. Embeddedness contributes an exclusion
effect from social networks, breeding a preference for established players with track
records. However, the overlapping of economic and ethnic ties has a further
undesirable effect, since outsiders also fall outside those networks that define the
industrial community.
Lastly, Gambetta (l988b) reminds us that there are situations where the public
interest might be better served if trust and the social capital built through exchange
of obligations and favors is collapsed rather than reinforced. Baker and Faulkner's
(1993) study of social networks involved in collusive practices in the American
heavy electrical equipment industry demonstrates the point.
The Relationship of Social Capital to Actors

The third group of dimensions to classify economic resources is labeled relationship
to actors. In this case, actors might be individuals, groups, business firms, or other
collective actors such as business interest associations. Three dimensions
characterize the relationship of a resource to actors: controllability or the extent to
which actors can appropriate and control a resource, accessibility or conversely the
existence of barriers to use a resource and tradeability, or the ability of actors to
convert resources through exchange processes.
Starting with controllability, we recall that social capital can be regarded as a
public good in the way Coleman (1990) defines it. Public goods are usually
characterized by jointness of supply and nonexcludability of benefits (Olson 1965).
But clearly this is not always the case with social capital. For example, Portes and
Sensenbrenner (1993) describe an informal financial system operating in the Cuban
immigrant community in Miami in the 1960s, whereby Cuban bankers were
prepared to provide 'character loans' to newly arrived refugees, based on the
personal reputation of the recipient in Cuba. This system worked well until a new
wave of immigrants, no longer known to the Cuban banking community in Miami,
arrived in 1973 after which character loans were discontinued.
Hechter (1987) discriminates between public and collective goods, on the basis
of the degree to which benefits are excludable. Most forms of social capital such as
trust cannot be characterized as public goods, in the sense that its benefits are
distributed homogeneously within a population. One might describe generalized
trusting within a community as a form of social capital but trust itself, is a specific
property of a dyadic relationship: A trusts B to do X (Hardin 1993). My ability to
deepen trust within a dyadic relationship--A trusts B in a range of situations, rather

than in a specific instance X--or to extend it to other relationships, is a learned
response. As Hardin (1993: 508) puts it: 'Trust has to be learned like any other
generalization. Insofar as my trust is a generalization in the face of new persons, this
merely means that the capacity to trust, the optimistic Bayesian estimate of
trustworthiness, is learned perhaps from long experience.'
But my propensity to trust is also dependent on factors outside my own
experience-namely, my family upbringing and the community in which I have
been brought up. The amount of generalised trusting I have encountered will
undoubtedly affect my estimates of the benefits of trusting and others'
trustworthiness. Whereas Hardin is right to contend that trust is not a form of human
capital, it is also true that the generalized trusting that is so critical to foster trust in
dyadic relationships could itself be seen as a form of social capital. My ability to
trust someone is partly dependent on my experience with similar people and
possibly also the experience of people I trust with that third party. Thus, Uzzi's
(1996a) study of the New York apparel economy found that, as in third-party
referral networks, the stock of social capital existent in a business relationship was
often used in a new, connected relationship by applying expectations and norms
from an existing relationship.
In short, social capital cannot be owned, controlled, or appropriated by an actor.
But that doesn't mean to say that social capital is a simply a resource uniformly
distributed and awash in the social structure of a particular community. It is at once
located in concrete, identifiable dyads and as a generalized resource that actors can
selectively draw for their social interactions.
Accessibility is an associated dimension to controllability. It denotes the ability
of actors to access a resource even when they are unable to secure control over it.
Information, for example, is one type of economic resource that presents specific
problems in terms of control and access. Access to social capital requires active
participation in the social networks and engagement in specific relationships.
Krackhardt (1992), Lazega and Lebeaux (1995), and Lazega (this volume) provide
good examples of how in formal organizations social capital is differentially
distributed among different social networks and how individuals might leverage
their social capital differently depending on specific circumstances. Krackhardt
(1992) is concerned with the strength of strong ties or what he describes as phi/os
relationships that have are based on long-standing friendships and affection. In a
casestudy, exploring an unsuccessful attempt to unionize an electronics firm in
California, Krackhardt attributes the failure to unionize the plant to the power of
individuals, based on their friendship networks, in opposing the move. Whereas the
advice networks reflected routine and workflow knowledge and influence, the phi/os
network became more important in a crisis context, when advice was sought from
those one trusted (friends).
Lazega and Lebeaux's (1995) study of a law firm uncovered partly overlapping
social networks based on different factors (friendship, advice, coworker) and
showed that actors leverage their social capital rather selectively, depending on the
particulars of a situation. Lazega and Lebeaux (1995) research focused on how a
focal actor used an intermediary to influence a target actor in a range of situations.
The results of the research demonstrated how focal actors drew selectively on their
social capital by using different social networks for different purposes and were
concerned about limiting the negative effects of leveraging that social capital.
However, this research also showed how the focal actors were often forcing
intermediaries to draw on their social capital in less discriminating and selective
ways.
Smart' s (1993) study of the use of guanxi in China is at once a good description
of how social capital is produced and reproduced within close-knit networks of
relationships, and also a good example of how it takes time and investment in
relationships to build up social capital in a particular network. As knowledge of the
rules of guanxi in China are difficult for an outsider to comprehend, many Hong
Kong investors use social connections as intermediaries to contact local brokers with
reputations for being able to solve problems (Smart 1993: 403).
Lastly, social capital is not a tradeable resource in the sense that it can be easily
exchanged for other resources. Although it is tempting to use the language of debits
and credits in the exchange of obligations, favors or trust, these exchanges are
characterized by the absence or accomplished concealment of explicit claims by the
possessors of social capital when attempting to cash in their credits as well as the
nonexistence of third-party enforcement to reclaim bad debts. The logic of social
capital is thus the logic of making calculativeness, self-interest, profit, accumulation,
and so on taboo (Bourdieu 1994).
So one party might build social capital in a particular relationship by a gift
presentation or trusting gesture, but there is no guarantee that the debt will ever be
repaid or, if repaid, the exchange will be equitable. But, to the extent that one invests
in gift giving and disinterested gestures, this will have positive outcomes via
building up status and reputation in a particular community or group. If
reciprocation is not direct, benefits can still be accrued by others' recognition of
one's generous disposition. But as Smart (1993: 396) remarks, we have had
experiences of unreciprocated gifts and invitations or noted the absence of
acknowledgment of social debts.
In economic life, the notion of gift presentation and exchange of obligations is
more nuanced and less clear cut than in the examples given above. Often, the
presentation of gifts in the social exchanges that accompany economic exchange are
part of specific, particularistic relationships between roleoccupants (e.g., salesman-
purchasing agent). The social capital built in these social relationships may be
recognized as institutional as well as personal social capital. However, this form of
social capital may only be recognised when individuals leave roles or take with them
a range of relationships (e.g. , customers) to a different organization. 13
In practice the picture may be more complicated since organizations are related
not simply through economic exchange but also through business-interest
associations, joint suppliers or customers, and their members' activities in
professional associations, industry forum, and technical standards committees
(Hakansson and Snehota 1995). Studies of information transfer and social networks
demonstrate how social capital is created and consumed in a variety of industry
contexts. Saxenian' s (1994) study of the evolution of the computer and
semiconductor industries in Silicon Valley and Route 128 ascribes the comparative
success of Silicon VaHey to informal and densely patterned communication

networks between communities of practice, based on professional aHegiances and
reciprocity of information exchanged, that provide an efficient and rapid source of
learning.
In short, social capital as a resource in economic life is characterized by a
continuum ranging from areas where the absence of calculativeness and self-interest
is paramount to areas where the tradability of favors, obligations, and reciprocity in
information exchange is more readily accepted. In all cases though, the borderlands
between specific reciprocity and equity in the exchange, and performances where
instrumental ends and calculativeness are either absent or successfully concealed,
are contingent performances subject to continuous negotiations within concrete
relationships, rather than fixed and socially prescribed rules.
Relationship to Other Resources and Activities

The first dimension in this set is that of integrity. It measures the extent to which
resources are simple or compound; that already comprises a mix of resources or else
representing some fundamental unit of resource that cannot be subdivided. Social
capital provides an excellent example of a compound resource that is impossible to
disentangle from the context in which it is produced and consumed. As Smart (1993 :
393) argues social capital is social in the sense that it cannot be possessed and its
existence is contingent on the reproduction of the concrete social relationships
within which it resides. At the same time, social capital in the form of obligations or
connections constitutes the very raw material through which social relationships
within groups or communities are reproduced and maintained.
The compound nature of social capital is intrinsically linked to the nature of its
creation and maintenance. The creation of some forms of social capital such as trust
can be regarded as fortuitous by-products of experiences over which the individual
may have little control or even did not undertake (Hardin 1993: 525). An individual
may grow up trusting and reproducing trust in his or her social relationships, as a
result of growing up in a supportive environment and learning through repeated
experience that trusting pays off handsomely. Similarly, Putnam's (1993a) argument
is that social capital is created and reproduced in communities where associations
proliferate, membership overlaps, and multiplex relationships are formed. 14
A dimension widely acknowledged to be important in understanding resources
is that of versatility. Resources may be versatile or specific in nature. In this context
versatility refers to the ways in which a resource may be combined with other
resources or activities. In practice, versatility is unequally distributed among
resources and, in any case, is not regarded as an intrinsic property of resources but a
function of their use-namely, in the context of exchange activities.
Social capital provides an excellent example of a resource that may be both
specific and versatile, depending on the context of use and its form. For example,
the existence of an obligation in a dyadic relationship is narrowly confined to the
context of that relationship and the ways in which it can be repaid forms part of the
atmosphere and tacit understandings prevailing in that relationship. Similarly,
placing trust in somebody to do something is often a highly particularistic judgment
that is embedded within the joint history and the atmosphere of a relationship.
However, belonging to a particular community, group, or clique and drawing on

the social capital accorded to members may be a highly valuable and versatile
resource. Examples of uses to which social capital acquired in this manner include
getting a job, landing a contract, or getting a client's trust (Waldinger 1995).15 In
particular, social capital built across overlapping relationships between the same
individuals (e.g., as buyer and seller, friends, professional colleagues) may provide a
highly versatile resource that can be leveraged across a range of contexts. Coleman
(1990) argues that social capital is often the by-product of membership in voluntary
associations and the social capital thus formed can be harnessed for other collective
action ends. Taylor and Singleton (1993) claim that the endogenous solution of
collective action dilemmas is greatly facilitated by the existence of social capital
lodged in multiplex relationships.
Similarly, clear-cut typifications of an individual as member of a group or
organization, may shut off the very same opportunities it affords generously to
insiders. Thus Waldinger' s (1995) study of the New York construction industry and
its ethnic niches showed that black contractors often had no choice but to restrict
themselves to public works contracts since all the other contracts were stitched up in
social circles and activities (e.g., golfing, boating) from which they were excluded.
The third dimension in this set is that of complementarity. Some resources are
easier to combine with other resources, and with activities, than others are.
Resources may be competitive as well as complementary even within the same firm.
Complementarity is a measure of the processes of combination. It is concerned with
the ease with which combinations with a specific resource may, in general, be
carried out. The complementarity of human capital and physical capital, for
example, is well documented in the resource-based theory of the firm, starting with
the pioneering work of Penrose (1959).
However, the complementarity of social capital with both physical capital and
human capital has received less attention. Coleman (1988) uses family upbringing
and relationships within the community in which the family is located as examples
of how social capital influences the creation of human capital. Waldinger (1995:
560) argues that Coleman' s examples only illustrate the indirect effects of social
capital on the development of human capital and that a more compelling case could
be made of how social capital is constitutive of all human capital. Learning through
socialization and participation in a community of practice such as the building
workers that Waldinger studied, are good examples of how the formation of human
capital is enmeshed with social capital. Ostrom (1994) provides a rare example of a
study where the interaction of social and physical capital is explicitly considered in
collective action problems.
Burt (1992) provides the most forceful argument for the belief that social capital
is both complementary to other forms of capital as well as being capable of being
leveraged for the benefit of an individual or organization. For Burt (1992) in firms
that provide services, for example, there are people valued for their competencies in
providing expertise and ability to deliver a quality service and those (the rainmakers)
who are valued for their ability to deliver and handle clients. The former represents
financial and human capital, whereas the latter represent social capital. In short,
different forms of capital are highly complementary and divisible on the basis of a
sharp division of labor and endowment of competencies even if, as Burt states,
'social capital is the final arbiter of competitive success.' In a world where players
can easily be matched on their endowments of human and financial capital, social
capital narrows down the pool of individuals who can compete successfully for new
opportunities. However, Burt (1992, 1997) privileges the notion of structural holes
or a relationship of nonredundancy between two contacts, and thus, in his view,
social capital is effectively reduced to the instrumental use of social relationships for
the pursuit of entrepreneurial opportunities to transform network structures to one's
own advantage.
The final dimension in this set is understandability. It is a measure of how easy
it is to comprehend the nature of a resource or even to recognize a resource at all.
Some resources are easily recognized as such and are well understood. Most forms
of physical capital come into this category. They are tangible, controllable, and
easily evaluated. Other resources, human capital included, do not share these
properties. They are intangible and cannot easily be valued. As mentioned
previously, organizational competencies such as routines fall into this category.
Exchange relationships and managerial skills are examples of this kind of resource.
If a resource can readily be understood, then it is more likely that the ways in
which it can be created and maintained as well as combined with other resources
will be well understood and hence exploited successfully. Virtually all forms of
social capital fare badly on this score. Coleman (1990: 312-313) argues that there
are few cases where social capital is well understood as a resource and can be
created as a direct result of investment of actors who have the aim of receiving a
return from their investment. The creation of formal organizations and the associated
authority structures with well defined obligations and expectations associated with
role occupancy is presented as a case of investment in social capital, in the same
way investment in human capital is associated with the appointment of individuals
to specific roles. But Coleman's formulation seems to conflate notions of power,
vested in authority structures, and social capital as the ability to induce others to act
in one's interests through the leverage of resources available to the agent. Whereas
power can ultimately be seen as resting on the authority to command the actions of
others and enforce sanctions for noncompliance, social capital tends to rely on
implicit rather than explicit claims to reciprocation and cannot resort to formal
sanctions on recalcitrant targets.
In hierarchies, as Miller (1992) cogently argues, managers do not spend their
time writing and enforcing contracts defining obligations and expectations
associated with specific roles anymore than employees spend their time maximizing
self-interest within the constraints imposed by those contracts. Norms of co-
operation and trusting relationships are as critical to the functioning of an
organization as a supplement, but not necessarily as a substitute for incentive
systems based on formal contracts. In short, notions of power and social capital are
as crucial to the functioning of hierarchies as private and public orders are vital to
the functioning of markets.
In summary, the character of most forms of social capital makes it a poorly
understood resource. Its creation is often dependent on by-products of other
activities, mostly escaping the direct control of actors, its leverage is dependent on
contextual performances, it cannot be easily transferred from context to context and

its linkages with other resources cannot easily be disentangled. Moreover, the
concealed or implicit nature of leveraging social capital, often attached to the history
and atmosphere of particularistic relationships, makes it even less understandable as
an economic resource.
CONCLUSIONS
The increasing attractiveness and popularity of the notion of social capital can be
understood by reference to the appealing logic of the capital-in-general metaphor
and for the possibilities it brings to breach some of the artificial divides between the
economy and society. This appeal rests as we have argued, in the often insufficiently
understood mechanisms of investment, accumulation, and benefits accruing from the
use of capital. In this sense, as others have noted (see, e.g., Tarrow 1996) social
capital has an appealing normative component. Creating, maintaining, and growing
social capital, so we are told, is the key to healthy societies, high-performing
organizations, and prosperous economies. In this sense, social capital can be seen as
important resource since its existence obviates the need to allocate other resources to
the formalization of rules, coercion, surveillance, and information gathering to
supplement private norms.
On the other hand, social capital provides a vehicle to breach the divide between
the social and economic worlds and its attributes make it an all-encompassing
notion, connecting dyads to wider social structures and back again. Thus social
capital-although conceived as a property of wider social structures in the form of
organizational structures, occupational communities, trade or civic associations, and
so on or based on categories such as ethnic groups-is manifested through concrete
outcomes in ordinary social practices and relationships. In this sense, social capital
can provide an antidote to more generic notions of embeddedness, understood
simply as a static backcloth of social structures encapsulating economic exchange
but themselves immune from reciprocal influences.
The argument in this chapter has been that neither the normative nor the
structural attractiveness of social capital as a notion can mask the dangers of
engaging in too cursory an application of the capital-in-general metaphor to the
study of the interaction of social structures and economic action. Social capital
understood as a resource that can impinge on economic action is perhaps not unique
in some of its attributes, but its heterogeneous manifestations, intangibility if not
purposeful invisibility, its compound nature, and poorly understood logics of
accumulation and use make it a counterintuitive metaphor.
In this chapter we have attempted to be more specific about the nature of
economic resources and subject the social capital metaphor to a more systematic
appraisal of its attributes as a resource impinging on economic action. In so doing,
we effectively moved the notion of social capital away from the notion of capital-in-
general and instead stressed its intangible, dynamic, and relational nature. This
allows us to move beyond notions of social capital as an abstract property of social
structures to explore its concrete outcomes and contextual performances and link
these to the logic of production and reproduction of those social structures where
social capital is deemed to reside. This last point is perhaps worth stressing anew,
even if a full discussion of this point is outside the scope of this chapter social
capital consists of claims to reciprocation and solidarity contained in concrete
episodes and contextual performances. What governs these episodes, however, are
that explicit, calculative claims are normally excluded from the performances within
which they are made and thus the leverage of social capital has to be read and
interpreted anew in every single episode.
Finally, we wish to issue a cautionary note to ourselves as much as to others,
regarding the need to resist the facile prescriptions that inevitably spillover from too
many of the romantic accounts of the role of social capital in socioeconomic life.
Even if we begin to understand a little better the notion of social capital, the
operation of policy instruments directed at building, maintaining, or destroying
social capital remain as slippery, open-ended and uncertain as any other policy
instrument. 16 Hopefully, this volume will contribute to advancing our understanding
of these issues.
NOTES
I. Recent work in the sociolegal field has recently taken an interest in the interaction between the law
and infonnal social nonns in law making and enforcement (see, e.g., Pildes 1996; Posner 1996).
2. On this topic see also the belated discovery of the existence of a variety of blends of trust and
contract in buyer-supplier relationships in some sociolegal and economics literature (e.g., Burchell and
Wilkinson 1997).
3. See F1igstein (1996) for a comprehensive treatment of the role of political embeddedness in the
evolution of markets.
4. See also Podolny (1994) and Uzzi and Gillespie (this volume) on this topic.
5. Ostrom (1994: 527-528) defines social capital simply as the arrangement of human resources to
improve flows of future income for at least some of the individuals involved in the production of that
capital. Similarly, Portes and Sensenbrenner (1993: 1323) define social capital as the relevant expec-
tations concerning economic action within a collectivity, affecting the goal and goal-seeking behaviour of
its members.
6. For a parallel and insightful distinction between social capital as lodged in personalised, dyadic
relationships and depersonalized, institutional relationships, see Pennings and Lee (this volume). See also
Knoke (this volume).
7. For a similar argument on social capital as a public good see Putnam (1993).
8. On the role of third parties in exchange see Nooteboom (this volume).
9. See also Pennings and Lee (this volume) on the importance of relationships that span organizational
boundaries for corporate social capital.
10. See Pennings and Lee (this volume). See also Burt (1997) for an argument on how social capital is
becoming increasingly important in organizational life and managerial perfonnance.
II. For critical reviews of Putnam's perspective on the impact of social capital on political and
economic life see Levi (1996), Tarrow (1996), and Kenworthy (1997).
12. See Lazega (this volume) on the notion of the finn as a multiplex, generalized exchange system and
on the role of reciprocity in this exchange system. Scharpf (1993: 153-154) makes a related point
concerning generalized trust whose existence 'pressuposes a generalized willingness to cooperate even in
constellations where cooperation is not advantageous and is easily destroyed by the pursuit of self-interest
at the partner's expense. But where it exists, generalized trust is enonnously advantageous. It will enable
rational actors to enter into vulnerable positions, and to engage in high-risk (and potentially high gain)
mixed motive transactions under conditions of incomplete infonnation.'
13. See Pennings and Lee (this volume) for a more comprehensive discussion of these issues.
14. The importance of multiplex relationships for the emergence and maintenance of social capital is
highlighted in several chapters in this volume-see, in particular Pennings and Lee, Knoke, and Lazega.
15. See also Flap and Boxman (this volume).
16. See also Leenders and Gabbay (this volume).
Social Capital by Design:
Structures, Strategies,
4
and Institutional Context
•
Wayne E. Baker
David Obstfeld
ABSTRACT
We examine social entrepreneurship from a structural perspective, distinguishing
between two structures of social capital and their associated entrepreneurial
strategies: structural holes and the 'disunion' strategy versus social cohesiveness and
the 'union' strategy. These two strategies represent alternative ways social
entrepreneurs access and mobilize the resources inherent in the structure of a social
network. The disunion strategist exploits structural holes between alters by keeping
them apart; the union strategist creates value by bringing together disconnected
alters. The frequency, legitimacy, and success of each strategy depends on the
'design' of the institutional context in which social entrepreneurs operate. Disunion
strategies tend to occur in organizations and markets characterized by sparse,
disconnected, and differentiated networks, coupled with competitive rules of
exchange, opportunism, and an individualist orientation; union strategies tend to
occur in organizations and markets characterized by dense, connected, and
undifferentiated networks, coupled with cooperative rules of exchange, norms of
reciprocity, and a collectivist orientation. We illustrate the distribution of triadic
strategies in a specific institutional context by taking a triads census of alliances in
the global automobile industry and testing the structural hypothesis about the use of
disunion and union strategies.
INTRODUCTION
Ever since Schumpeter (1934: 156) identified entrepreneurship as a 'vehicle of
continual reorganization of the economic system,' entrepreneurship has been
recognized as playing a key role in catalyzing change, promoting innovation, and
Social Capital by Design - 89
enhancing productivity in the economy at large. For example, the continuous

process of 'breaking away' from existing firms to create new ones is a prime engine
of the growth and economic development of cities (Jacobs 1965, 1970). The role of
entrepreneurs as intermediaries between corporate actors (firms, associations,
governmental bodies) is well known (e.g., Coleman 1990: 180-188). However,
entrepreneurship is also recognized as a creative activity that occurs inside
organizations, an activity that is critical to their survival and health (e.g., Burgelman
1983a; Kanter 1983a). Entrepreneurial action within organizations is called intra-
preneurship (Pinchot 1985), corporate venturing, or corporate entrepreneurship
(Burgelman 1983a, 1983b).
Structural sociologists shifted the study of entrepreneurial behavior away from a
focus on traditional entrepreneurial activities, such as using economic capital to start
new ventures, to the analysis of the strategic use of 'social capital' both inside and
between organizations (e.g., Burt 1992; Bourdieu and Wacquant 1992; Coleman
1988, 1990). Our chapter follows in this structural tradition. We offer a new
theoretical distinction between two structures of social capital and their associated
entrepreneurial strategies-structural holes and the 'disunion' strategy versus social
cohesiveness and the 'union' strategy. These two strategies represent alternative
ways social entrepreneurs access and mobilize the resources residing in a social
network.
We use the phrase 'social capital by design' in two senses that together capture
the essence of our argument. First, social capital is related to the 'design' of an
institutional context. A context characterized by sparse, differentiated, and
disconnected networks, for example, yields social capital in the form of structural
holes. A context characterized by dense, integrated, and connected networks, in
contrast, yields social capital in the form of social cohesiveness. Second, social
capital can be created by intention, that is, by the strategic moves of an individual
entrepreneur or the deliberate manipulation of organizational and interorganizational
structures. Change agents, for example, can change the structure of social capital
(and the entrepreneurial strategies used to access it) by altering the design of the
institutional context.
STRUCTURES AND STRATEGIES

The Structural Approach to Entrepreneurship
Most attempts to understand and promote entrepreneurship have examined
entrepreneurs and their behaviors at the individual level, striving to define the key
traits and characteristics of successful versus unsuccessful entrepreneurs (Gartner
1989; Low and MacMillan 1988). For example, McClelland (1967) argued that the
'need for achievement' is a key psychological characteristic of the successful
entrepreneur, but empirical research has not supported a link between the need for
achievement and, say, the decision to start a new business (Sexton and Bowman
1985). Locus of control and propensity for taking risks have been proposed as
possible distinguishing characteristics, but research has not provided much empirical
support of these hypotheses (e.g., Brockhaus 1982; Sexton and Bowman 1985;
Gasse 1982). Tolerance of ambiguity does appear to distinguish entrepreneurs from
managers (e.g., Schere 1982; Sexton and Bowman 1985). In general, the concerted
attempt over the past decades to build a personality profile of successful

entrepreneurs has not yielded insights into the unique personalities of entrepreneurs.
Based on their review, Low and MacMillan (1988: 148) conclude that, '... at
a... fundamental level, it can be argued that the wide variations among entrepreneurs
make any attempt to develop a standard psychological profile futile. One is struck by
the appropriateness of Gartner's (1985) observation that 'there is as much difference
among entrepreneurs as between entrepreneurs and non-entrepreneurs."
The failure of personality research to identify key characteristics suggests that
the individual level of analysis may be inappropriate for understanding
entrepreneurial action (Gartner 1989). These efforts suffer from the neglect of social
structure and the complex relationships between the individual, corporate actor, and
environment (Martinelli 1994). Others have proposed a behavioral focus for
entrepreneurial research, emphasizing the processes associated with entrepreneurial
behaviors (e.g., Gartner 1989, 1990). The structural approach, which we use here, is
consistent with such a behavioral focus (e.g., Aldrich and Zimmer 1986; Burt 1992;
Krackhardt 1995; see also, Freeman, this volume, for a similar perspective).
The entry of structuralists into the study of entrepreneurial behavior shifted the
focus away from traditional entrepreneurial behaviors, such as starting new
businesses or founding firms, to the examination of the strategic creation and use of
'social capital' both inside organizations and among organizations (e.g., Burt 1992;
Bourdieu and Wacquant 1992; Coleman 1988, 1990). 'The central proposition of
social capital theory,' summarize Nahapiet and Ghoshal (1998: 243), 'is that
networks of relationships constitute a valuable resource for the conduct of social
affairs .. .. ' We use the term 'social entrepreneur' to describe individuals or corporate
actors who access and mobilize the social capital inherent in organizational
networks, as opposed to the 'traditional entrepreneur' who uses economic capital to
create new businesses and firms (of course, traditional entrepreneurial activities
often involve the use of both social and economic capital). The social processes we
discuss apply to activities both inside organizations and between organizations (what
we call 'institutional contexts' below). We focus on 'corporate social capital'-social
capital that resides inside, between, or among organizations-in contrast to other
locations of social capital, such as families and communities (see, e.g., Coleman
1988; Putnam 1995a).
The definition of social capital, along with the role and activities of the social
entrepreneur, have been the subjects of considerable debate. l Clarifying the
dimensions of social capital is a high research priority (Putnam 1995a). Nahapiet
and Ghoshal (1998), for example, make useful distinctions between three
dimensions of social capital: the 'structural dimension' (the configuration of social
networks), the 'cognitive dimension' (shared systems of meaning, narratives,
language), and the 'relational dimension' (norms, trust, obligations). Our structural
view of social entrepreneurship emphasizes the first dimension, exploring the basic
structures of social capital (structural holes versus social cohesiveness), their
corresponding entrepreneurial strategies (disunion versus union), and the
relationship between institutional context and strategy. Our distinctions, as we
elaborate below, help to clarify the debate and resolve some of the confusion about
the structural approach to social capital and social entrepreneurship. Though we
emphasize the structural dimension of social capital, we acknowledge the

importance of the cognitive and relational dimensions.
Two Structures of Social Capital

The triad of social actors, composed of 'ego' and two 'alters,' is the basic unit of
analysis in social entrepreneurship. Structures larger than the triad are possible, of
course, and often occur; however, they are based on the triad as the fundamental
building block. For example, 'communities of trust' are generalizations of the three-
actor structure (Coleman 1990: 188-189). Social actors can be persons or corporate
actors: people acting as individuals, people acting as agents or representatives,
organizational subunits (such as teams or departments), organizations, and even
governmental bodies, states, and nations. (For example, the triad is a basic unit used
in analysis of geopolitical relations.) Given our focus on corporate social capital,
however, we are interested primarily in two types of social actors: individuals who
are members of organizations and organizations themselves. Our concepts apply to
both types.
The structural basis of entrepreneurial action was suggested by Simmel (1950:
154-162), who stressed the importance of the 'third element' in group dynamics.
Simmel argued that the introduction of a third party fundamentally alters the social
dynamics of dyadic ties (see also Nooteboom, this volume). Of particular interest is
the triad type Simmel (1950) called tertius gaudens-'the third who enjoys' benefits
by his or her position between two disconnected parties. These two parties, because
of their unfamiliarity with each other, can be manipulated to the third party's
benefit. Simmel's tertius gaudens is the basis of Burt's (1992) influential theory of
structural holes. Burt argues that a sparse egocentric network with few redundancies
(few members of the network know each other) is a social structure rich in structural
holes. A hole exists between two people (alters) if they are not connected to each
other but share a tie with a common third party (ego). This structural arrangement
puts the third party in the role of the tertius gaudens who can take advantage of the
two disconnected persons (or corporate actors) for private gain. Burt (1992) found,
for example, that managers in a large corporation who have networks rich in
structural holes were promoted faster and at earlier ages, compared with otherwise
similar managers whose networks lacked structural holes. Burt (1992) also found
that firms with interorganizational networks rich in structural holes earned a higher
rate of profit, compared to firms without these structural advantages.
Some argue that Burt's (1992) structural holes theory is an alternative to, rather
than an example of, a theory of social capital (e.g., Walker, Kogut, and Shan 1997:
112). These arguments emphasize the 'relational' dimension of social capital
(Nahapiet and Ghoshal 1998), which is absent or undeveloped in Burt's theory.
From the 'relational' view, social capital exists in a relationship between two people
(or two corporate actors) if they develop personal bonds, attachments, and trust. '[A]
close working group of [graduate students] working on the same problems constitute
social capital for each of them for his [sic] graduate training' (Coleman 1990: 170).
This 'relational' view also stresses the restraints on opportunism maintained by social
capital that allow cooperation take place (Walker, Kogut, and Shan 1997; see also,
Granovetter 1985).
Rather than pitting structural holes theory against a relational definition of social
capital, we believe it is more theoreticalll productive to consider these as alternative
views of social capital (Obstfeld 1997) and to concentrate instead on delineating
and analyzing their characteristic social structures. Burt's (1992) theory emphasizes
only one of several structures of social capital. Coleman's (1988: 98) original
definition is broad enough to encompass other structures: 'Social capital is defined
by its function . It is not a single entity but a variety of different entities, with two
elements in common: they all consist of some aspect of social structures, and they
facilitate certain actions of actors-whether persons or corporate actors-within the
structure. Like other forms of capital, social capital is productive, making possible
the achievement of certain ends that in its absence would not be possible.' Similarly,
Bourdieu's (Bourdieu and Wacquant 1992) definition is broad enough to include
multiple structures of social capital. For Bourdieu, social capital is 'the sum of the
resources, actual or virtual, that accrue to an individual or a group by virtue of
possessing a durable network of more or less institutionalized relationships of
mutual acquaintance and recognition' (Bourdieu and Wacquant 1992: 119).
Social capital 'inheres in the structure of relations between and among actors'
Coleman (1988: 98). Structural holes theory emphasizes a structure of social capital
characterized by sparse networks and few redundancies. For Burt (1992), social
capital resides in the patterned absence of ties. This view is consistent with the
argument that social structure is defined more by the patterned absence than
presence of ties (White, Boorman, and Breiger 1976). What is the alternative? We
argue that social capital also inheres in the structure of social networks as the
patterned presence of ties. The alternative to the holes view of social capital is what
we call social cohesiveness, where the structure of social capital is characterized by
dense networks and mUltiple redundancies. This structure of social capital
corresponds to another of Simmel's triad types, the third party who acts as a
mediator or 'non-partisan' to create or preserve group unity: The non-partisan either
produces the concord of two colliding parties, whereby he withdraws after making
the effort of creating direct contact between the unconnected or quarreling elements;
or he functions as an arbiter who balances, as it were, their contradictory claims
against one another and eliminates what is incompatible in them' (Simmel 1950:
146-147).3 Also see Nooteboom (this volume), who describes six roles of the third
party (the so-called 'go-between').
Common examples of this structure of social capital include real estate
brokerage, literary agency, and political mediation. 4 For example, some Washington
lobbyists specialize in the introduction of corporate actors to public officials (such as
executive agency officials or congressional representatives) (Coleman 1990: 180-
182). Coleman (1990: 180) calls the third parties in these social structures
'intermediaries in trust.' The third-party intermediary is able to bring together the
other two parties because each one trusts the intermediary. This structure of social
capital is pervasive in society; as Coleman (1990: 184) describes:
This form of intermediary exists in all areas of social life. For example, professors write
letters of recommendation to prospective employers about students, and persons seeking
a job or a loan list other persons who will recommend them. The acceptance of a
recommendation by a prospective employer or creditor is a placement of trust in the
judgment of the intermediary, which allows a placement of trust in the ability of the
prospective trustee to perform as expected. If the latter defaults, then the trustor's trust in
the intermediary's judgment is reduced.
A social actor can advertise its position in the social cohesiveness structure of social
capital as a valuable resource. Consider, for example, the language used by
Mayfield, a venture capitalist firm, in its promotional brochure; as quoted by
Freeman (this volume): 'Because of our long association with a large number of
successful companies and entrepreneurs, a relationship with Mayfield is highly
regarded. It can enhance the credibility of a young company with potential
customers, vendors and employees, and with other financial institutions.' Similarly,
professional service firms, such as advertising agencies and investment banks, 'sell'
access to the social capital inherent in their networks. Indeed, advertising agencies
that occupy a central position in the market are likely to be kept by their corporate
clients, indicating the value clients place on this structure of social capital (Baker,
Faulkner, and Fisher 1998).
The level of trust in the social cohesiveness structure is probably higher, on
average, than the level of trust in the tertius gaudens arrangement. This is one reason
why some critics argue that structural holes theory is not a theory of social capital
(see above). Because the tertius gaudens exploits the structural hole between two
alters, the level of trust is presumed to be low. This is not necessarily so. Each alter
may trust ego, even though the alters are unaware of each other's existence; in other
cases, the alters may prefer to remain out of contact with each other, relying instead
on their trust in ego. Moreover, the issue of trust in the social cohesiveness structure
is not as unambiguous as it might seem. Sometimes an 'intermediary in trust' runs
the risk that he or she will be circumvented or 'cut out' by opportunistic alters. For
example, the risk that the principals in a real estate transaction might consummate
the deal in secret, saving the commission owed to the broker, is so high that the
standard legal contract between a real estate broker and seller contains protections
against such actions. 5
The distinguishing feature between the two structures of social capital is not
trust, but the answer to this question: What does the social entrepreneur do with the
gap between alters? While one social entrepreneur exploits the structural hole,
keeping alters apart, another social entrepreneur may choose to close the gap,
bringing together the two alters. These actions represent the two basic
entrepreneurial strategies for accessing and mobilizing the social capital inherent in
social networks. We next describe and illustrate these strategies, followed by our
analysis of the relationship between entrepreneurial strategies and institutional
context.
Two Entrepreneurial Strategies

Strategy is used here to refer to a social entrepreneur's plan of action for using
network structure to access and mobilize social capital. Burt's (1992)
conceptualization of social capital as the structured absence of ties favors what we
call 'disunion strategies.' In this strategy, the social entrepreneur generates 'profit' by
taking advantage of the disconnection of the two parties. The disunion strategy is
Alter I Alter 1
Ego Ego
Alter 2 Alter 2
Disunion Union
Figure 1. lIlustration of disunion and union strategies
illustrated in Figure 1. As shown, ego (the tertius gaudens) is linked to two

disconnected alters. The hole is represented by the blank space between alter 1 and
alter 2, and enables ego to play the two alters against each other or to secure a
valued resource from one and provide it to the other, extracting a profit in the
exchange. In such a case, ego benefits from the absence of a connection between the
two alters, may act to keep the alters apart, and at the very least, chooses not to
introduce the two alters. For disunion strategies, value is created by the exploitation
of these structural holes; value, therefore, is a 'private good'-the benefits of social
entrepreneurship accrue to the third party (ego).
The view of social capital as social cohesiveness leads to the alternative
entrepreneurial strategy we call 'union strategy'-illustrated in Figure 1. In this case,
ego (Simmel's non-partisan or arbiter) is linked to two alters who are disconnected
or in conflict. Ego 'closes' the gap between alters by bringing them together or by
resolving their differences. This suggests a sharing or exchange of resources. The
union strategy produces what Coleman (1988: 107) calls 'closure'-a social structure
that creates the conditions for the enforcement of norms through 'sanctions that can
monitor and guide behavior.' The combination of norms and trust that emerge under
these structural conditions facilitates additional use ofthe union strategy.
Disunion Strategies In Action

Disunion strategies are common in situations where the formal differentiation of the
organization presents the social entrepreneur with many temptations to play one
person (or department) against another person (or department). Burt's (1992)
original study of managers and structural holes was conducted in one such
organization. Consider, for example, an entrepreneur with ties to a person in the
sales department who collects current information about customer needs, and to
another person in the marketing department who is desperate for customer input for
a new product. The entrepreneur can position him or herself to sales as someone
with influence over the company's product development process and to marketing as
a source of new information (Obstfeld 1997). With this disunion strategy, the
entrepreneur seeks to benefit without ever introducing the two alters, a move that
would eliminate the advantageous position. A similar case is the classic role of the
boundary spanner who links two disjoint groups (Friedman and Podolny 1992).
Competitive markets are principal locations of disunion strategies at the
interorganizationallevel. Simmel (1950) notes that the market is a prime example of
the tertius gaudens strategy writ large. Here, disunion strategies are characterized as
'rivalry,' where two or more sellers vie 'for opportunities of exchange' with a buyer
(Weber 1978: 63; Swedberg 1994: 271). For example, Coca-Cola Company
maintains relationships with six different advertising agencies (Baker, Faulkner, and
Fisher 1998: 149), playing one advertising agency against the other in an elaborate
disunion strategy. The practice of competitive bidding is based on the disunion
strategy, where multiple sellers are pitted against each other. For example, disunion
strategies are evident in the garment industry studied by Uzzi (1996a) in cases where
dress manufacturers (buyers) select contractors (sellers) on the basis of price alone
to effect discrete, nonrecurring exchanges. Of course, sellers in an industry suffering
from intense competition caused by the buyers' relentless use of the disunion
strategy may become motivated to collude, employing an illegal union strategy to
counterbalance the power of buyers (Baker and Faulkner 1993).
Disunion logic drives the avoidance of ties in a competitive market. Competing
companies will not use the same supplier because doing so would put the supplier in
the structural position of the tertius gaudens. For example, General Motors avoids
using the investment bank Goldman Sachs because its chief American rival, Ford
Motor Company, uses Goldman as its main bank (Baker 1990). Similarly, corporate
competitors avoid using the same advertising agencies, citing 'conflict of interest' as
their rationale (Baker, Faulkner, and Fisher 1998).
The social structure of disunion strategies can be fluid, particularly in dynamic
markets. When AT&T and China began negotiations to install an undersea cable
system to provide a telecommunication link between China and the U.S., the
entrepreneurial leverage associated with the disunion strategy shifted from one actor
to another. AT&T initially approached the negotiation as the owner and operator of
international marine cables, enjoying an advantageous negotiating position (Glain
1997). Deregulation and other changes in the telecommunication industry, however,
unleashed a variety of competitors (such as Baby Bell SBC Communications,
Nynex, Britain's Cable & Wireless, and Japan' s Kokusai Denshin Denwa Co.) that
the Chinese invited into the negotiations. One of China's key negotiators indicated,
'Our general policy is to not engage in projects that exclude other parties. We want
to engage as many companies as possible on an equal basis' (quoted in Glain 1997).
This approach nullified AT&T's disunion strategy by leveraging confidential
information provided by the competing companies and the disconnections between
them. Ultimately, the Chinese insisted on a 14-member consortium that included all
of AT&T's major rivals.
Finally, direct exploitation of disconnected parties may represent only a fraction
of the activities associated with disunion strategies. Some disunion strategies require
constant effort and vigilance (perhaps even subterfuge) to keep alters apart and to
maintain the structural conditions of disconnection, secrecy, and concealment. Price-
fixing conspiracies, for example, require conspirators to deliberately maintain the
ignorance of their corporate customers (Baker and Faulkner 1993). Some corporate
actors attempt to prevent competitors' actions or the development of new regulations
that would reduce or close the structural holes in markets, fighting to maintain the
structural conditions that favor disunion strategies. For example, the strategic
alliance of IBM and Apple was opposed by many competitors (Baker 1994a) who
viewed the potential union as a threat that would reduce their ability to compete for
customers-that is, to operate using the disunion strategy.
Union Strategies In Action

The formal differentiation of an organization does not always lead to disunion
strategies. For example, Kanter (1983a: 141) describes the union strategy used by a
corporate entrepreneur who created an internal alliance between sales, service, and
product development:
He first wrote a memo to all of the sales people in his area, copying the district managers
for service and products.... He then held a series of sales meetings, inviting commercial
and service staff too .... [He] explained and reexplained the benefits of cooperation across
the sales/service/products boundaries to people from each function. (Ashkenas et al.
1995: 17)
Similarly, Burgelman (1991) depicts how a corporate entrepreneur at Intel

collaborated with two other product champions to develop RISC processor
technology and to line up a customer base in advance of Intel's entry into the new
market. Ashkenas et a!. (1995) describe union strategies used to transform General
Electric's Retailer Financial Services into a 'boundaryless' organization. For
example, a systems manager employed union logic to streamline a business process:
'Nastasi then brought together a group of systems, marketing, finance, and customer
service people and challenged them to complete new customer conversions in a
matter of days, not weeks.' In many cases, union strategies are used by senior
managers targeting major organizational change. Union strategies are also initiated
by relatively low-ranking employees with more modest objectives.
Union strategies exhibit variation in official support. General Electric CEO Jack
Welch, for example, officially sponsored GE's Corporate Executive Council to
stimulate the exchange of information and collaboration (Baker 1994a,b). In
contrast, Hutt, Reingen, and Ronchetto (1988) describe the emergence of a new
product through the collaboration of actors from multiple divisions along with key
customers-all acting outside prescribed corporate guidelines. Similarly, the
successful Intel intrapreneurs described above went against explicit corporate policy
and had to disguise the RISC project as just another 'co-processor' until sufficient
progress had been made (Burgelman 1991).
Union strategies also occur in markets. Some union strategies involve the
introduction of initially unconnected suppliers or customers via common third
parties. Starr and MacMillan (1990: 86) describe a case of an entrepreneur who
introduces two suppliers:
A corporate entrepreneur saw an opportunity to connect two of his major suppliers while
developing a new medical products business. One manufacturer, in the south, had cheap
labor costs and good employee morale, but was nonetheless losing money due to lack of
business. The other, in the north, had high labor costs and an employee shortage. By
introducing these two manufacturers the venture manager reduced the costs of one and
increased the sales of the other.
Uzzi (1996a: 679) describes a similar union strategy used in the New York garment
industry:
One CEO explained how [a] tie formed between him and a manufacturer named 'Diana.'
He said that his contact with Diana began when Norman, a close business friend of his
and Diana's, asked him 'to help Diana out' in a time of need (cut her fabric at a special
price and time), even though he had no prior contact with her.... [The CEO said] 'So why
did I help her out? Because Norman asked, 'Help her out.'
The formation of strategic alliances-joint ventures, technology sharing,
marketing arrangements, product development, and others-{;an result from union
strategies. For example, Coming introduces alliance partners to each other, fostering
creation of ties between them. AT&T Global Information Solutions convenes an
annual conference in which its strategic allies meet in 'alliance fests' used to generate
new alliances and associations (Baker 1994b). Digital Equipment Corporation
(DEC) invites its alliance partners to a four-day conference to share information and
foster alliance formation (Gulati 1995a). Similarly, the monthly meetings of the 128
Venture Group were convened by an entrepreneur and venture capitalist to establish
a place where venture capitalists, entrepreneurs, consultants, and management team
candidates could meet and explore collaborations (Nohria 1992b).
The introduction of disconnected parties is only a fraction of the activities
associated with union strategies. Union activities include investing in already
established ties, cultivating ongoing collaborations, and maintaining the general
structural conditions that facilitate union strategies. Some union strategies are self-
sustaining, such as those associated with the trading groups in the industrial districts
of north central Italy and southwestern Germany (Powell 1990). In other cases,
however, such as real estate brokerage, the repeated cultivation of new alters is
necessary to stay in business. Of course, the real estate broker with a good reputation
and lots of contacts enjoys the benefits of union strategies in reverse, as satisfied
customers refer new alters to the broker.
These examples illustrate the general principle that union logic drives the
selection of ties, whereas disunion logic drives the avoidance of ties (as described
above). Disunion strategies proscribe the use of common third parties, such as
suppliers, while union strategies prescribe the use of common third parties, such as
alliance partners. These strategies may exist side by side in the same institutional
context, which is a topic we take up in the following section.
INSTITUTIONAL CONTEXT
So far, we have not been very specific about the context in which social
entrepreneurs operate. This was intentional, so that we could be clear about the
distinctions between the two structures of social capital and their corresponding
strategies. However, the relationship of structures and strategies to institutional
context is critical. First, we argue that the nature, level, and forms of social capital-
and therefore the strategies social entrepreneurs employ~epend on the structure
and culture of the institutional context. This relationship holds in the institutional
context of an organization as well as that of organizational fields, business sectors,
industries, and markets. Second, we argue that the frequency, legitimacy, and
success of an entrepreneurial strategy depends on its 'fit' or compatibility with the
institutional context in which it is used.
Table 1. Institutional context and entrepreneurial strategies
Institutional Context Disunion Strategy Union Strategy

Structural Conditions
size large size small size
density sparse networks dense networks
connectivity disconnected networks integrated networks
formal differentiation many formal boundaries few formal boundaries
Cultural Conditions
rules of exchange competition cooperation
norms opportunism. distrust reciprocity. trust
orientation individualist orientation collectivist orientation
Nahapiet and Ghoshal (1998) argue that the 'firm' is a better institutional setting
than the market for the development of high levels of social capital; because the firm
is a 'social community' (Kogut and Zander 1996: 503), it enjoys an 'organizational
advantage' over markets. However, some firms are organized and operated as
'markets,' and some markets are organized and operated as 'firms' (Eccles 1981;
Eccles and White 1988; Stinchcombe 1985).7 Therefore, we make the bold
assumption that the characteristics of 'institutional context' can be specified in such a
way that they apply to both firms and markets. An advantage of our approach is the
generalization of the concepts of corporate social capital across institutional levels-
that is, within firms and between firms. The relationship between institutional
context and strategy is summarized in Table 1.
Structural Conditions
By structural conditions, we refer to the 'network configuration' of an institutional
context, consistent with Nahapiet and Ghoshal's (1998) definition of the 'structural'
dimension of social capital. Standard network measures can be used to represent a
configuration. Nahapiet and Ghoshal (1998) propose density, connectivity. and
hierarchy. We substitute 'formal differentiation' for hierarchy. because an
organization can be differentiated along three dimensions-spatial. horizontal, and
vertical. Markets, too, are organized along these three dimensions, as economic
geographers have documented. Formal divisions such as these, whether in firms or
markets. create structural holes. We add 'size,' since the number of social actors
influences the fragmentation of an institutional setting, and along with it, the number
and extent of structural holes.
The structural conditions associated with disunion strategies are large size,
sparse and disconnected networks, and many formal boundaries. The design of most
large-scale, traditional organizations favors disunion strategies because it creates so
many structural holes between departments, across levels, and between spatially
separated operations. Formal differentiation, for example, hampers collaboration
across boundaries (Lawrence and Lorsch 1986). Large size disfavors the integrated
'network' organizational design (Baker 1992a). Similarly, markets with many players
are more fragmented than markets with few players, producing many structural
holes. The alert tertius gaudens in a fragmented market generates profit by
arbitraging across these gaps in the social structure of trading (Baker 1984), just as
the tertius gaudens in a fragmented firm generates profit by exploiting the
disconnections inside the organization (Burt 1992).
The structural conditions favoring union strategies are small size, dense and
integrated networks, and low differentiation (see Table 1). Union strategies are
common in network organizations, for example, because these organizations are
characterized by flexibility, lateral ties, and a high degree of integration across low
formal boundaries (Baker 1992a). A dense network of customers, producers, and
suppliers provides a structural basis for cooperation (Perrow 1992), such as the
dense social networks of firms, local universities, community colleges, research
institutes, financial institutions, trade associations, and regional governments in
Silicon Valley (Saxenian 1991, 1994).
The biotechnology industry features dense and well-connected networks of
interfirm cooperation as well (Walker, Kogut, and Shan 1997; in this volume see
also Stuart, Smith-Doerr et al. and Omta and Van Rossum). These structural
conditions are conducive to the formation of new alliances based on the union
strategy. For example, the likelihood that two previously unallied firms will form an
alliance increases with the number of third-party ties they have in common (Gulati
1995a). Similarly, the greater the number of research and development alliances and
other types of collaborations a biotechnology firm has at a given time, the more
diverse its future portfolio of ties will become (Powell, Koput, and Smith-Doerr
1996).
Cultural Conditions
Cultural conditions can be defined in many ways. Indeed, the definition of culture
itself is a subject of considerable debate and rival interpretations (DiMaggio 1994;
Scott 1995). Rather than trying to resolve this debate, we propose a simple definition
of cultural conditions-the institutionalized rules of exchange, norms, and social
orientation in an institutional context-and focus primarily on the relationship
between strategy and cultural context. s
Rules of exchange are shared social understandings about 'who can transact with
whom and the conditions under which transactions are carried out' (Fligstein 1996:
658). For example, exclusivity (sole-source) is a rule of exchange governing buyer-
seller relationships in the advertising industry (Baker, Faulkner, and Fisher 1998:
151). Competitive rules of exchange prohibit union triads. The avoidance of
common suppliers is a good example (Baker 1990). Disunion strategies are
legitimate and successful in firms operated as 'markets' (Eccles and White 1988). In
settings such as markets operated as 'firms' (Stinchcombe 1985), however,
cooperative rules of exchange discourage disunion strategies (which would be
interpreted as self-serving, opportunistic behaviors). Similarly, organizations in
small-firm networks share information, establish long-term relationships, and
support each other's efforts, in opposition to the competitive rules of exchange in
classic markets (Perrow 1992). The search for greater efficiencies through closer ties
between customers and suppliers can displace traditional rules of exchange that pit
suppliers against one another. New developments in just-in-time (JIT) inventory
control, for example, call for close coordination between customers and their
suppliers. Honeywell orchestrates close cooperation between its internal buyers and
five suppliers who would ordinarily compete for business (Bleakley 1995).
Norms are shared expectations that regulate behavior (DiMaggio 1994),
including choice of entrepreneurial strategy. Norms of openness, teamwork, trust,
and reciprocity favor union over disunion strategies. For example, Putnam (1993a)
argues that norms of reciprocity in the industrial districts of northern Italy foster
interfirm cooperation and limit opportunistic behavior-that is, the use of disunion
strategies. Similarly, norms of reciprocity support the union strategies evident in
such business groups as the Japanese kieretsu and Korean chaebol.
Social orientation refers to the distinction between individualism and
collectivism (Triandis 1995: 2). Individualism, for example, is 'a social pattern that
consists of loosely linked individuals who view themselves as independent of
collectives; are primarily motivated by their own preferences, needs, rights, and the
contracts they have established with others; and emphasize rational analyses of the
advantages and disadvantages of associating with others' (Triandis 1995: 2).
Individualism favors disunion strategies; collectivism favors union strategies. For
example, collectivist practices such as cross-functional teams, multi-level
management networks, group-level reward systems, and team-building programs,
facilitate union strategies within organizations (Baker 1994a). Recent research has
shown that a collectivist orientation often increases the odds of alliance formation
(Dickson and Weaver 1997). Private economic associations and political
organizations foster cooperation in the industrial districts of northern Italy (1993a).
The robust collaboration found in these industrial districts is supported by technical
colleges, vocational training, supportive banks, and extended kinship ties (Powell
1990).
Distribution of Strategies
We argue that the frequency, legitimacy, and success of an entrepreneurial strategy
depends on its 'fit' with the 'design' of the institutional context in which social
entrepreneurs operate. This suggests the following proposition: The ratio of disunion
to union strategies varies according to the structure and culture of the institutional
context. At one extreme, disunion strategies dominate in settings characterized by
sparse, disconnected, and differentiated networks, along with competitive rules of
exchange, opportunism, and an individualist orientation; at the other extreme, union
strategies dominate in settings characterized by dense, connected, and
undifferentiated networks, coupled with cooperative rules of exchange, reciprocity,
and a collectivist orientation. A mix of strategies occurs in an institutional context
located between these two extremes.
The distribution of strategies in a given context can be determined by a triads
census (Wasserman and Faust 1994: 556-602). The pattern of strategic alliances in
the global automobile industry 9 illustrates one empirical distribution (Baker 1992b),
though the triads approach can be applied in any institutional context. For strategic
alliances, a triads census includes only four possible triad isomorphism classes: 1)
the null triad, composed of three unallied firms; 2) the dyad, composed of three
firms of which only two are connected by an alliance; 3) the disunion triad,
g.
~
Table 2. Triads census of alliances in the world automobile industry

Null Triad Dyad Disunion Triad Union Triad
[003]" [102] [201] [300] Significanceb
Type of Alliance observed expected observed expected observed expected observed expected 'tD 'tu
All Alliances 5616 5510 1814 2010 314 241 26 9 5.972 5.540
joint venture 7045 7024 681 722 43 24 I 0 4.463 1.538
manufacturing/assembly 6631 6607 1057 1103 78 60 4 1 2.767 2.939
technology sharing 6877 6861 844 873 46 36 3 0 1.981 3.752
supplier relationship 6529 6513 1158 1186 77 70 6 1 .976 4.079
marketing/distribution 7030 7024 712 722 26 24 2 .555 3.573
10
equity investment 7257 7256 502 503 10 11 I 0 -.278 3.466
aNumbers in square brackets [MAN] refer to triad isomorphism classes using standard labeling, where M = number of mutual dyads, A =number of CIl
asymmetric dyads, and N = number of null dyads. on
btau test statistic from TRIADS (Walker and Wasserman 1987): 'tD = disunion triads; 'tu = union triads. E
Source: Baker (1992b). Data from Wards International, 1985 (N = 37 automobile companies). (')
.§
[
r:T
'<
~
til
~.
::s
o
-
composed of two alliance dyads; and, 4) the union triad, composed of three alliance
dyads. 10 The first class, the null triad, represents a classic competitive situation-the
complete absence of strategic alliances among three firms. The second class, the
dyad, represents an isolated alliance between two firms . The third class, the disunion
triad, represents the tertius gaudens arrangement (as illustrated in Figure 1). The
fourth class, the union triad, represents the social cohesiveness structure of social
capital (see Figure 1).
The empirical distribution obtained in the triads census of the automobile
industry is shown in Table 2. For our purposes, we focus here on the two triad types
of particular interest, disunion and union triads. Both triad types occur much more
often than expected by chance alone, considering all alliance types combined. 11
However, if we examine the triads census for each type of alliance, we find an
interesting pattern: 1) Disunion triads (but not union triads) occur more often than
by chance in joint ventures; 2) Disunion and union triads occur more often than by
chance in two types of alliances-technology sharing and manufacturing/assembly;
3) Union triads (but not disunion triads) occur more often than by chance in three
types of alliance-supplier ties, marketing/distribution, and equity investments.
This pattern suggests that the structure of social capital varies by alliance type.
The structure of joint ventures implies that social capital is accessed and mobilized
by exploiting structural holes, using the tertius gaudens strategy. The structure of
supplier ties, marketing/distribution, and equity investments implies that social
capital is accessed and mobilized by closing structural holes. The mixed structure of
technology sharing and manufacturing/assembly suggests the presence of both types
of social capital. However, the statistically significant use of both disunion and
union strategies may indicate a competition of strategies in which neither dominates.
The general relationship between entrepreneurial strategies and institutional
context is illustrated in Figure 2. This stylized representation implies the possibility
of change or movement, of the transformation of an institutional context and its
corresponding strategies. The intersection of the curves in Figure 2 represents a
balance of disunion and union strategies. This point is an unstable state in which
neither strategy dominates, such as the mixed structure of triadic strategies for
technology sharing and manufacturing/assembly in the automobile industry (Table
2). If so, then the automobile industry may fall back on earlier, simpler, and
overlearned strategies (arms' -length competition), as people and organizations are
prone when faced with uncertainty and ambiguity (Weick 1995: 102).
For an organization, the point of intersection in Figure 2 indicates a particularly
risky stage in an organization's transition from one institutional design to another.
For example, a firm attempting to improve collaboration and cooperation must foster
both the structural and cultural conditions that favor the union strategy (moving
from left to right in Figure 2). Some consultants claim that a hierarchy can be
converted into a network organization simply by adding links (Lipnack and Stamps
1994: 72; see, also, Mueller 1986), but structural change is not enough. Failure to
change both structural and cultural conditions endangers a change effort. For
example, the effort to transform Industrial Computer and Control Group failed
because change agents altered only organizational structure (Nohria and Berkley
Entrepreneurial
Strategy
Disunion
-~----
.- .- -- --
"
/
/ ""
/
I
- - Disunion
SO/50 - - - - Union
-----
Union
Small, dense, integrated Large, sparse, disconnected

networks; cooperation, networks; competition,
trust, collectivism Institutional Context opportunism, individualism
Figure 2. Entrepreneurial strategies by institutional context
1995). By replacing hierarchy with a network design, they may have induced more
union strategies, but by itself this structural change was not enough. Without a
corresponding cultural change, the effort to foster cross-divisional collaboration was
doomed. Employees did not fundamentally change their strategies for action (Nohria
and Berkley 1995). The change effort may have collapsed at or near the point of
intersection in Figure 2.
CONCLUSION
Our chapter attempts to specify and clarify dimensions and structures of social
capital. We offer a set of concepts that encompasses different views of the structural
sources of social capital, the basic strategies used to access social capital, and the
relationship between institutional contexts and strategies. We argue that social
capital 'inheres in the structure of relations between and among actors' (Coleman
1988: 98) in two fundamental ways-one based on the patterned absence of ties, the
other on the patterned presence of ties. Structural holes theory (Burt 1992)
emphasizes the absence of ties, where social capital resides in a social structure
characterized by sparse networks and few redundancies. We maintain that social
capital also inheres in social structure as the presence of ties, which we call social
cohesiveness, where social capital is found In dense networks with multiple

redundancies.
Each structure of social capital represents opportunities to access and mobilize
the resources inherent in a social network. The first structure calls for a 'disunion'
strategy that exploits the structural holes between alters by keeping them apart. The
second structure calls for a 'union' strategy that creates value by bringing alters
together, closing the hole between them. Our qualitative and quantitative examples
provide ample evidence of the use of these strategies in a wide range of institutional
contexts, both inside and between organizations. The concepts of two types of social
capital (structural holes versus social cohesiveness) and two types of entrepreneurial
strategies (disunion and union) can be generalized and applied to understand the
structure and use of social capital in many different institutional settings.
The actual distribution of strategies in a particular institutional context depends
on the 'design' of the institutional context in which social entrepreneurs operate. An
entrepreneurial strategy that 'fits' the structure and culture of a given institutional
context occurs more frequently, enjoys greater legitimacy, and will be more
successful in the long run, compared to a strategy that does not fit. Disunion
strategies dominate in organizations and markets characterized by sparse,
disconnected, and differentiated networks, coupled with competitive rules of
exchange, opportunism, and an individualist orientation; union strategies dominate
in organizations and markets characterized by dense, connected, and
undifferentiated networks, coupled with cooperative rules of exchange, norms of
reciprocity, and a collectivist orientation.
Of course, a mix of triadic strategies falls between the pure disunion and union
extremes. Thus, one avenue of additional research is to explore further the precise
connection between institutional context and entrepreneurial strategy. This line of
work calls for a demography of social entrepreneurial relationships (see, also, Baker,
Faulkner, and Fisher 1998: 173). The statistical methods available for taking a triads
census (Walker and Wasserman 1987), which we used in our analysis of the
distribution of strategies in the global automobile industry (Baker 1992b), can be
used to determine the distribution of strategies in any organization, market, or
organizational field. In addition, organizational change can be tracked by measuring
shifts in the distribution of strategies over time, using methods designed to model
longitudinal change in networks (e.g., Leenders 1995a, 1996; Wasserman and Faust
1994). By coupling new and better measurement methods (e.g., Borgatti 1997; Han
and Breiger, this volume; Doreian, this volume) with more precise concepts of social
capital and their corresponding entrepreneurial strategies, it is possible to take
another step forward in the exploration of corporate social capital.
We are grateful to the editors of this volume for their helpful comments and suggestions. Direct
correspondence concerning this chapter to Wayne Baker, University of Michigan Business School, 701
Tappan Street, Ann Aibor, MI 48109 (wayneb@umich.edu).
NOTES
I. The debate takes place in published works (see citations herein) as well as in infonnal discussions
on the Internet, such as the recent interchanges on social capital in SOCNET between January and June,
1997. We cite the published literature to support our points, but we also acknowledge the contributions to
the debate in SOCNET made by Xavier De Souza Briggs, Robert Putnam, Barry Wellman, and others.
2. This point was also made in SOCNET by Barry Wellman and Robert Putnam (January 1997).
3. Simme1 (1950) describes another triad type, 'divide and conquer,' in addition to the tertius gaudens
and 'non-partisan' types. In 'divide and conquer,' the third party deliberately introduces conflict between
the other two. This type is not our concern in this chapter.
4. Real estate brokerage and literary agency qualify as this type because the ego is not a principal in a
transaction, but instead brings together the two principals (alters) who consummate the transaction. In the
first case, the real estate broker (ego) matches a buyer and seller (alters) who exchange property for
money; in the second, a literary agent (ego) matches a publisher and author (alters) who exchange 'the
property' (manuscript) for royalties. Some types of brokerage do not qualify as this type. For example, in
a wholesaler arrangement, the broker is a principal in one transaction (buying goods from a manufacturer)
and in a second transaction as well (selling the goods to the consumer).
5. Consider, for example, the protections contained in the standard 'listing agreement' (officially, the
'cooperative selling contract') used in the state of D1inois to define the legal obligations of the seller to the
real estate broker: 'SELLER SHALL: Cooperate fully with Broker; refer all inquiries to Broker; conduct
all negotiations through Broker; ... and pay a real estate brokerage commission of X % of the sale price; if
1) Broker provides a purchaser ready, willing, and able to purchase in accordance with this Contract; or
2) if the property is sold, exchange, gifted, or optioned by Broker or by or through any other person
including the Seller during the period of this contract; or 3) if it is sold directly or indirectly within six (6)
months after termination of this contract to a purchaser to whom it was offered during the tenn thereof.'
6. It is possible, of course, that the individual pursuit of entrepreneurial profit as a private good
produces value at the collective level. This line of reasoning is consistent with traditional management
and economic theories. For example, the tournament model of mobility assumes that competition among
managers yields better ideas, more satisfied customers, and greater shareholder value. The theory of the
market similarly assumes that individual striving maximizes social welfare. Burt (1992) does not
emphasize the public-goods aspect of social entrepreneurship, though he mentions the possibility: An
entrepreneur is 'a person who generates profit from being between others. A nonprofit player, pursuing
entrepreneurial opportunities just for the pleasure of being the one who brings others together to build
value, could choose to reinvest it all' and strengthen existing relationships (Burt 1992: 34-35). Of course,
social capital can be considered a public good in and of itself. Putnam argues, for example, that
Coleman's (1988) original concept of social capital incorporates such a view. He adds, however, that the
views of social capital as a private and public good are complementary (see discussions in SOCNET; see,
also, Putnam 1993a, 1995; Knoke this volume).
7. Nahapiet and Ghoshal (1998: 261) briefly mention that some interorganizational networks may
develop an institutional context that is conducive to high levels of social capital, and point to this as an
avenue of future research.
8. We acknowledge, but do not elaborate here, the importance of the 'constitutive' as opposed to the
'regulatory' view of culture (DiMaggio 1994). In addition, we acknowledge but do not discuss the state as
part of the institutional context (Fligstein 1996).
9. We obtained data for the triads analysis from a special issue of Ward's Automotive International
(1986). This issue reported the alliances known to exist among 37 major automobile companies, classified
by the six alliance types reported in Table 2. We created a square, binary, symmetric matrix for each
alliance type, where each company is assigned a row and corresponding column, and an entry indicates
the presence or absence (0,1) of an alliance between two companies. Thus, we created six 37 X 37
matrices. We analyzed the triads structure of each matrix using TRIADS (Wassennan and Walker, 1987).
These data are treated as symmetric because alliances are naturally mutual ties, and we adjusted the
weighting vector in TRIADS accordingly.
10. A triads census includes sixteen isomorphism classes when asymmetric ties are considered
(Wassennan and Faust 1994: 566). Since alliances are mutual ties, however, there are only four classes of
interest: triad types 003 (null), 102 (dyad), 201 (disunion), and 300 (union ). These three-digit numbers
[MAN) refer to triad isomorphism classes using standard labeling, where M =number of mutual dyads, A
= =
number of asymmetric dyads, and N number of null dyads.
II. Structural hypotheses are tested using the TRIADS program (Walker and Wassennan 1987) in
which the empirical distribution of triads is compared against a theoretical distribution of triads assuming
random assignment of ties. The weighting vector used in these tests reflects the number of disunion
configurations contained in the disunion triad, and the number of union configurations contained in the
union triad. The test statistic, tau, is interpreted using the standard nonnal distribution (Wassennan and
Faust 1994: 595). We use a significance level of .05 in this analysis.
Corporate Social Capital and Liability:
a Conditional Approach to Three
5
Consequences of Corporate Social
Structure
•
Dan TaImud
ABSTRACT
This chapter defines and illuminates three aspects of corporate social capital which
are created by different aspects of corporate social structure. The chapter also shows
that corporate structure which generates social capital may become later a liability.
The chapter briefly reviews the literature regarding the contingent value of corporate
structure in creating competitive social capital. Then two other kinds of corporate
social capital, political and cognitive, are illustrated via a case study of an Israeli
Armament firm. Finally, the chapter briefly discusses the relations between the three
kinds of corporate social capital for business policy and strategy as well as for future
studies on corporate advantage.
INTRODUCTION: VARIETIES OF CORPORATE SOCIAL CAPITAL

Certain corporate social structures provide comparative advantage to the firm. The
term 'corporate social capital' is a heuristic device, assists us in comprehending how
corporate social structure benefits certain business organizations at the expense of
others. Corporate social capital is any positive, goal-specific outcome, originating
from corporate social structure. 1 It includes any means of corporate control or
business leverage, embedded in social relations, thus assisting a firm in promoting
internal assets and resources. 2 Excess profit or economic rent, which is an outcome
of the operation of corporate social capital, is depicted as 'any advantage or surplus
created by nature or social structure over a certain period of time' (Sorenson 1996:
1344). Corporate advantage may be explained then by the differential capacity of
firms to create, promote and take advantage of corporate social capital (Nahapiet
and GhoshaI1998).3
Corporate Social Capital and Liability- 107
I argue that one can differentiate analytically between three kinds of corporate
social capital, each derived from a particular dimension of corporate social structure
(Zukin and DiMaggio 1990) [see Table I]: I) Competitive corporate capital is as a
function of a firm's structural location in imperfect competition (Burt 1992). The
term expresses the extent to which the social organization of the competition, as
measured via transaction networks, systematically benefits certain market players at
the expense of others (Burt 1992; Talmud 1992, 1994; Talmud and Mesch 1997).
For example, a particular form of competitive corporate capital is the manner by
which suppliers or buyers are dependent on a player in an exchange system. 4
Another example is the ability of a firm to engage in exclusive relations with sub-
contractors, thus increasing its capacity for complex adaptation and economy of
speed (Uzzi 1997a).
By contrast, 2) political corporate capital indicates the manner in which the
political organization of the economy, institutional and legal regulation of corporate
behavior, and the collective struggles over economic options extend the strategic
possibilities of corporate executives. By virtue of membership in larger political
networks, certain firms have more institutional leverage and can buy political
protection from state and local institutions. For example, firms with direct
institutional ties to the political center may gain more collective resources than such
lacking those ties (Talmud 1992; Talmud and Mesch 1997).
In their strategic considerations, firms attempt to influence and take advantage
of legal arrangements, political ties, and state policies at various levels (Zukin and
DiMaggio 1990: 20-21; Talmud 1992; Talmud and Mesch 1997). Political
arrangements have uneven effects on market players and therefore may be defined
as political corporate capital. This is measured by strong linkage to the
institutionalized political center, as well as by institutional regulation and political
arrangements favoring the focal firm or industry. However, political arrangement are
dynamic and high reliance on political capital in one period may be detrimental in
another time. Continuing benefits from political corporate capital compel corporate
players to calculate and rearrange their corporate capital for sudden shifts in the
political sphere or to politically-directed changes in the corporate task environment. 5
Paradoxically, those corporations rich in political corporate capital which have
learned to rely on institutionalized political arrangements are among those firms,
highly prone to crisis and mismanagement following transformation of political
agencies. These firms often enjoy rewards such as lower tax rates, tariff protection,
interest-free loans, business tips, favorable environmental conditions, and direct and
indirect state subsidies (Talmud 1992). In economic terms, firms that concentrated
their efforts on using political corporate capital are engaged in rent-seeking conduct,
or in 'directly unproductive, profit-seeking behavior' (Bhagwati 1982). Therefore,
their strategic repertoire is not geared toward efficiency but toward an effective
usage of their political leverage.
Finally, only recently the organizational literature has characterized a less
known form of corporate social capital: 3) cognitive corporate capital.6 While
competitive and political corporate capital are instrumental for the external economy
of the firm, cognitive corporate capital is chiefly relevant for the internal economy
of business organization. It is defined as the extent to which a firm can facilitate
advantageous business action by using internal networks which reconstruct shared

understanding, disciplinary power, mental models, organizational identity, tacit
knowledge, corporate norms, the ability for corporate foresight, reflexivity, and self-
awareness. 7 Cognitive corporate capital then is facilitated by relations, formal and
informal, inside the firm (Podolny and Baron 1997; Nahapiet and Ghoshal 1998). It
assists the organization in accomplishing goals, to speed production process, to
improve quality according to clients' requirements, to 'move things ahead' and to
fortify its competitiveness, effectiveness, or efficiency (Uzzi 1996a).
The importance of cognitive corporate capital cannot be overstated. Several
perspectives have currently acknowledged this unique form of corporate social
capital. In sociology and psychology, organizational self-command has been
perceived as an integral part of 'relational management' and 'knowledge
management' of the firm as a social community.8 Resource-based business strategy
has shifted organizational theory from an emphasis on value appropriation to the
exploitation of the firm's social structure for value creation. Similarly, Human
Resources Management have recently put more weight on employees' long term
relational contracts and 'employability' (Ghoshal and Moran 1996).
These cognitive assets can be constructed and reproduced because any corporate
firm is first and foremost a social community, a nexus of social relations (Nahapiet
and Ghoshal 1998). Intra-organizational social networks thus yield and convey
organizational memory and shared narratives and norms. In fact, social networks can
create or modify trust in the organizational system (Kramer, Brewer, and Hanna
1996). Social networks also convey intangible corporate assets such as socially-
produced tacit knowledge, which is particularly useful when and product complexity
and quality are at stake (Polanyi 1958; Dore 1983; Jones, Hesterley, and Borgatti
1997).9 Moreover, cognitive corporate capital, like any other social capital
discussed in this chapter, is a firm-level phenomenon. It is endowed in corporate
social relations, and not in individual players, and thus the firm can take advantage
of these nested relations (Monge, Cozzens, and Contractor 1991). Cognitive
corporate capital then may speed a firm's reaction to market cycles, enhancing
complex adaptation and reducing learning cost and time-to-market (Uzzi 1997a). It
may, furthermore, promotes collaborative production of social knowledge capacity,
which is especially relevant for the construction and transfer of socially tacit
knowledge (Nahapiet and Ghoshal 1998). It may also increase learning speed and
effectiveness, and promote smooth solutions to problem solving (Dore 1983; Uzzi
1996a, 1997a; Nahapiet and Ghoshal 1998). Still, cognitive corporate capital is a
firm-level virtual asset which decreases the economic cost of monitoring, learning,
adapting, and also diminishes agency problems. Like any other form of corporate
social capital, it has diminishing marginal utility, and thus could be accumulated up
to a certain point (Uzzi 1997a; Talmud and Mesch 1997). If external conditions
significantly change, the very same corporate social structure producing cognitive
corporate capital in one circumstance, may in another instance create a corporate
liability, thus inhibiting the firm's adjustment to new environmental circumstances.
Because cognitive corporate capital is an intangible corporate, relation-specific
asset, it cannot easily be imitated by competitors (Barney 1991). It is transmitted via
Table 1. Three types of corporate social capital

Competitive
Social Structure Producing Structural Embeddedness: Corporate Network Position
Corporate Social Capital
Examples Non-redundant Market Ties (Burt 1992)
- Rewards Profit squeezing capacity (Burt 1992; Talmud 1994); sustained
industry leadership (Talmud and Mesch 1997)
Political
Social Structure Producing Political Economy and Institutional Regulation
Examples State regulation; political institutional ties (Talmud and Mesch 1997)
- Rewards State subsidies (Talmud 1992b)
Cognitive
Social Structure Producing Normative and Cognitive Embeddedness: Shared narratives,
Corporate Social Capital normative role expectations and commitment, corporate identity, vision
and discourse enacted via intra-organizational social networks.
- Examples 'Self-command capital' (Lindenberg 1993); 'Strong culture' (Kotter and
Heskett 1992); tacit knowledge (Polanyi 1958); 'combinational
capability' (Nahapiet and Ghoshal 1998)
- Rewards Corporate innovation (Monge, Cozzens, and Contractor 1991);
complex adaptation, 'economy of speed' (Uzzi 19%a; 1997a);
provision of product quality (Dore 1983)
a complicated web of social relations. The creation, configuration and maintenance

of cognitive corporate capital is costly, consuming time and corporate resources.
An organization needs to maintain a careful balance, therefore, between various
modes of corporate social capital, but in a similar vein to those business
organizations relying upon political corporate capital, firms which have
advantageous cognitive corporate capital tend to remain stationary. Corporate
networks conveying identities and normative expectations are inclined to become
inflexible over time (White 1992). Accordingly, ideal cognitive capital should not
prevent a firm from being flexible.
All these three forms of corporate social capital have a few common attributes:
• They are not a property of any individual agent;

• They can be accumulated over time;
• They are relation-specific and cannot easily be transformed from network to
network;
• Therefore they are expensive, bearing investment-specific sunk cost;
• If they are not properly maintained, they tend to fade out or become
irrelevant to the extent that a firm is located in a turbulent and uncertain
environment, and to the degree that the corporate social structure producing
corporate social capital is rigid, they tend over time to turn from a corporate
capital into a corporate liability.
There are also some significant differences between the three forms:
a. Competitive corporate structure creating corporate social capital is the most

measurable as precisely operationalized by network models of competitive
advantage (cf. Burt 1992).
b. Political corporate capital is more vague, composed of many sub-forms, and is
molded corporate actors' ties as well as by historical and institutional
circumstances.
c. Cognitive corporate capital is the most 'sticky' and vague form. It is a product of
both past and present relational composition inside the firm, and of interface
with other agents, such as competitors, consultants, suppliers and clients.
The following part has three sections, each of which addresses a particular
dimension of the corporate social capital. The first section summarizes the literature
on the contingent sources of competitive corporate capital. The next section
demonstrates how social structure producing political corporate capital becomes a
corporate liability. Finally, the chapter discusses to inability of a manufacturing firm
to renovate its cognitive corporate capital.
A CONDITIONAL APPROACH TO CORPORATE SOCIAL

STRUCTURE AND CORPORATE SOCIAL CAPITAL
Recent literature on corporate social capital and competitive advantage has stressed
the role of weak ties and sparse social networks in determining corporate
performance (Burt 1992).\0 Burt (1992) and Talmud (1992, 1994) found that firms
using 'structural holes,' spreading their ties with unconnected market segments are
more profitable than those connecting to 'redundant market areas.'
In contrast, other studies on corporate social capital have found significant
impact of strong ties and dense networks in creating opportunities. New and
sensitive business information and opportunities are enhanced through cohesive
contacts (Aldrich and Zimmer 1986; Gilad, Kaish, and Ronen 1989). Moreover,
Krackhardt (1992), Podolny (1994) and Gabbay (1995, 1997) showed the
importance of enclosed networks for managing corporate uncertainty.
It seems, accordingly, that there is no such thing as a universal optimal network
structure (dense or sparse); as Coleman phrased it 'social relationships that constitute
social capital for one kind of productive activity may be impediments for another'
(Coleman 1994: 177).
Furthermore, the effect of strong ties seems to be a conditional one. Burt (1992)
and Han (1993) revealed that, contrary to the logic behind the theories of resource
dependence and structural holes, those American firms having high recourse
dependence on the government survive at higher rates. Also, Israeli firms that are
'locked in' by political ownership are more profitable (Talmud 1992), and are more
stable over time (Talmud and Mesch 1997). In uncertain conditions, or where
identity and role expectations are important for performance, embedding economic
transactions in strong relations, or constructing the economic deal in terms of strong
ties, could serve as remedies for survival (Dore 1983; Podolny 1994; Jones,
Hesterly, and Borgatti 1997; Podolny and Castellucci, this volume).
Organizations develop indirect corporate leverage via social networks to

manage their dependence, while protecting their advantage in strategic positions
(Galaskiewicz 1985; Baker 1990). Juggling between contrasting business pressures,
corporate firms attempt to maintain their access to reliable and non- standard
information, while also striving to reduce trade dependency and uncertainty. Thus,
corporations are bound into repeated transactions, while also preserving other
opportunities at hand (Powell 1990; Podolny 1994). Uzzi (I996a, 1997a) found in
his study of the New York apparel industry that organizational failure increases to
the degree that the focal firm's network tends to be composed of either all arm's-
length ties (i.e. 'under-embedded,' disconnected, non-redundant contractors), or all
embedded ties (i.e. 'over-embedded,' dense, semi-integrated contractors). Uzzi
(I997a) showed that organizational survival is positively associated with a 'mixed
model' network, which provides a combined advantage. On the one hand, market
embeddedness provide benefits such as trust, joint problem-solving arrangements,
complex adaptation, reduced bargaining and monitoring costs. On the other hand,
arm's-length contacts provide the firms new and novel information outside the
immediate ties (Uzzi 1996a, 1997a). Similarly, Talmud and Mesch (1997) have
demonstrated that the probability of the leading Israeli firms to sustained their
industrial leadership is a function of the 'mix embedded ness' type of relations.
Similarly, Uzzi (1996a, 1997a) claims that each type of controlled embedded ness
'yields positive returns up to a certain point,' and 'embedded networks offer a
competitive form of organizing, but possess their own pitfalls.' Competitive
pressures create complex, dissimilar, and distinct network forms, rather than a singly
optimal embeddedness strategy (Uzzi 1997a). As a result, 'a paradox appears:
optimal networks are not composed of either all embedded ties or arm's-length ties
but integrate the two' (Uzzi 1996a: 694). In sum, the conditional approach to
competitive corporate social structure argues that the corporate value behind
network forms is contingent upon contextual features of the environment. II Next, I
will apply this insight to political and cognitive corporate capital.
THE CASE OF TA'AS: ISRAELI ARMEMENT CORPORATION I2

Political Corporate Capital: The Demise of Corporate Leverage
This section illuminates how political corporate capital diminishes due to a
combination of organizational rigidity and changing environmental conditions.
In many cases, competitive corporate capital is not sufficient to understand the
construction and maintenance of corporate social capital. In most countries,
corporate political embedded ness of the market, institutional regulation, and
governmental restrictions on competition augment firms' market positions and
elevating barriers to entry (Levacic 1987: 118-138; Talmud 1992). In other words,
political corporate capital assists a firm in its capacity for 'price making.'
Ta'as (The Hebrew acronym for the Israeli Military Industry) was incorporated
in 1947 as an institutionalized organ of armament. It operated first in the Jewish
underground against British rule and later on for the Israeli Defense Forces, which
had been its chief client over the years. Ta'as became then a state agency which
gradually transformed into a state-owned enterprise with very low autonomy and
without financial reserves or structural leverage. State regulation and laws prevented
private investors to compete with Ta'as in the local market, but also to from loaning
finances to the company. Ta'as' business policy and strategy was effected much
more by political considerations than by sheer economic issues (see also Samuel
1998). This situation has facilitated a monopoly position for many Israeli Defense
Industry firms. Yet, the emergence of the newly globalized weapons market, the end
of the cold war, and the peace process in the Middle East brought about a radical
change in the political embeddedness of the weapon's industry, as demonstrated in
the case of Ta'as. If political corporate capital can be measured by a premium paid
for (or excessive revenue earned by) a given investment in political ties, then-
because it is a relation-specific investment-dearly, any drastic modification of the
political structure can alter the fate of the firm, from a revenue to a loss.
Until the 1980s, the firm could use the dramatically leaking bucket of the State
of Israel's defense budget. Accordingly, its business strategy was typified by the
overarching principle of autarky and freedom from functional constraints.
Structurally, it had one main buyer (monopsony), but Ta'as was that buyer's main
supplier as well (monopoly). These structure of bi-Iateral monopoly had been
maintained and fortified by the firm's identification with the heroic Jewish
underground under the British Mandate in Palestine, to be superseded later on by its
strong cultural affiliation with the dominant security symbols of the State of Israel.
Already heavily invested in political corporate capital, the firm rationally preferred a
strategy emphasizing effectiveness over efficiency. As a result, the organization
forged an ambitious business strategy: it tended to 'make' rather than cheaply 'buy'
various components of its product; it stressed, moreover, activities of research and
development function and employed a disproportionately large, permanent and
expensive workforce (increasing from 1000 employees in 1948 to 14,500 in 1985;
Talmud and Yanovitzky 1998). Maintaining its political corporate capital. the
organization possessed an excessive production capacity. accompanied by useless
vertical integration along many of product lines including logistics, maintenance and
support systems (Kleimann 1992; Talmud and Yanovitzky 1998) to overcome
prospective shortage or demand fluctuation. Additionally, the firm diversified into
about 500 different products in various market areas, at the expense of achieving
competitive advantage by focusing on core advantage. Ta'as chose, in fact, to
focusing on technological push (imposing scientific innovation and engineering
excellence) rather than by market pull (imposing product's fit to customers'
preferences) (Samuel 1998; Talmud and Yanovitsky 1998). Additionally,
competition in the global market was considered to be only secondary to the focus
on the local Israeli market. Many governmental restrictions on military equipment
export were compensated by state agencies. Still, the industry was booming until
the 1980s, as a result of its political corporate capital. In the 'heroic' period of
1947-1985, Ta'as was socially conceived as strategically important, and thus
enjoyed close institutional linkage with the Ministry of Defense, which was formally
its owner and main buyer (Talmud and Yanovitzky 1998). The fact that Ta'as
lacked competitive corporate capital (its transaction network was concentrated on a
single buyer for any single product) was compensated for by the higher value of its
political corporate capital. which was deemed instrumental for corporate solvency.
Following 1985, there have been increasingly severe cuts in the Defense budget.
The State of Israel thus has become more sensitive to armament prices. In addition,
global prices have decreased, as suppliers' competition in the world market has
intensified. The sharp modification of Ta'as' political corporate capital created a
series of acute structural crises, accompanied by a severe financial toll, layoffs, rapid
managerial turnover, and cultural shocks to employees. Moreover, it exposed the
structural weakness of the firm's transaction networks and its high dependency on
suppliers and customers (Talmud and Yanovitzky 1998). Ta'as' political corporate
capital has been eroded because it was based merely on a social ties with political
agents which were no longer relevant.
Additionally, it is important to note that political and competitive corporate
capital may also be linked. The case of Ta'as shows that the political structure
affects the industry competitive structure, and thus provide regulated competitive
leverage for the firm. Moreover, state regulation and laws prevented private
investors to compete with Ta'as in the local market, but at the same time enforced
the company to lean on political backup. Barriers-to-entry and barriers to exit were
fixed by the political organization of the industry. The competitive structure was
therefore strongly affected by the political web which Ta'as was a part of. Upon
state lessening its grasp on Ta'as later on, other kinds of structures could have been
nurtured and capital could be drawn from them. Indeed, this was a prerequisite for
survival and solvency. The state ceased to be highly dependent on the firm, as the
global market emerged, and the sense of closure of politicians, public officials and
CEOs has been also damaged because of the eroding symbolic importance of the
security sector (Talmud and Yanovitzky 1998). Nonetheless, the firm was unable to
embark on developing other forms of compensating corporate capital (neither
competitive nor political), partially because of the transformation of its cognitive
corporate capital.
Cognitive Corporate Capital: From Corporate Superiority to Corporate

Liability
In its formative period, Ta'as enjoyed a relatively safe, prestigious and certain
environment with a comparatively assured clientele (Evron 1992). The organization
was composed of relatively stable intra-organizational networks, producing
consensus regarding the corporate mission and working procedures. The firm's
managers were production-oriented, lacking market orientation and marketing
skills. \3 The firm, leveraged by its political corporate capital, emphasized production
capacity and product quality rather than marketability. The formal and informal
networks transmitted this shared language throughout the organization, and the kind
of novel knowledge developed inside the firm in the 'heroic period' (especially
regarding ammunition), reflected this socially accepted value. The fact that the
corporate vision was to assist the Israeli Defense Force regardless of cost brought
about a shared sense of commitment to a special 'community-organization.'
Ta'as consisted of four basic structural elements deemed necessary for the
creation of its cognitive corporate capital: 1) shared symbols and narratives, 2) a
sense of closure, being a part of the 'security sector,' collective history, 3) repeated
interactions between experts, and 4) interdependence between customers and experts
(Nahapiet and Ghoshal 1998; Talmud and Yanovitzky 1998). Consequently, many
risky innovations, devised inside the firm by engineers and technicians-some of
them lacking formal training-were made possible by mitigating socially-connected
trust within the organization using technical experts' commitment and tacit know-
how (see also Kramer, Brewer, and Hanna 1996).
By contrast, in the post-1985 'competitive era,' Ta'as has attempted remedial
measures. Nevertheless, these measures have been ineffective. As I will show, Ta'as
social structure~eemed necessary for creating its older cognitive corporate
capital-has become inappropriate for the new environment, thus resulting in
corporate liability.
Following the radical changes in environmental conditions, the firm undertook a
partially successful restructuring program including decentralization, the
establishment of relatively autonomous profit centers, implanting export orientation,
introduction of (failed) products' modification for the civil markets, layoffs, and
cutting administrative costs.
Moreover, the new organizational vision, imposed by new Chairs of the Board,
CEOs and General Managers, which emphasized market competition, stood in sharp
contrast to the old one, emphasizing priceless product quality, conveyed via existing
intra-organizational networks. The management did not trouble to pass on its new
vision via social networks since its image of the firm was formal and hierarchical.
Yet relations inside the rank and file were persistent, thus assuring a highly shared
resistance to change. In other words, the newly introduced uncertainty hampered the
'psychological contract' and the trustworthy social relations inside the firm,14 and
was not accompanied by management's proper awareness of informal processes
inside the corporate units. Furthermore, uncertainty and a lack of trust were
magnified by contradictory messages relayed from the government to the firm, and
by high turnover rates of staff and line managers, including the CEO's. This, in turn,
diminished social cohesion inside the organization, and restricted the ability of the
new managers to set up effective ties with unit managers and other influential
players. Consequently, this impaired Ta'as' capacity to implant new-fashioned
corporate identity, fresh strategic vision, and more important, trust in the corporate
headquarters (Talmud and Yanovitzky 1998).
The literature on corporate social capital underscores the fact that socially
embedded exchange may prevent agency problems and opportunistic behavior
which resulting in sub-optimal performance (cf. Podolny 1994; Jones, Hesterly, and
Borgatti 1997). Yet the lack of social linkage and trust between the new CEO's and
the unit managers created a dual agency problem, originated by relational
discontinuity inside the firm. This discontinuity, in turn, was created by a lack of
durability in identity-making ties (Podolny and Baron 1997), which is required
especially in times of crisis (Mishira 1996; Webb 1996). It was also found that both
the owner (the State) and the corporate directors were not receiving reliable
information on the business and financial operation of Ta'as from their respective
subordinates. As a result, they could not make relevant decisions and could not
install monitoring devices regarding the fate of various profit centers within the firm
(Talmud and Yanovitzky 1998). The corporate ship, consequently, began to drown.
In a highly diversified corporation mitigating a dual-level crisis (at the industry

level and at the firm level), suffering from a lack of internal linkage and trust,
bureaucratic formal control is not as effective a device as that of corporate social
control transmitted via social networks. The new management ignored the veterans'
social networks. Layoffs did not solve the problem, as recruitment was virtually
non-existence. Thus the old, durable, identity-formation networks inside the firm
have created, in fact, corporate paralysis (Crozier 1960). In Ta'as, the 'old guards' of
the veteran employees and unit managers gave credence to norms such as
'effectiveness,' 'quality,' 'security,' while the new corporate managers were more
attuned to catchwords such as 'efficiency,' 'market testing,' and 'cost-benefit
analysis.' Hence, a lack of social linkage accompanied by a deficit in shared
narratives and corporate trust inside the firm has diminished the organizational
ability to create and foster cognitive corporate capital (Nahapiet and Ghoshal
1998).15 The erosion in the cultural position of Ta'as, marked by the shift from the
'heroic period' into the 'competitive era,' has eroded to a certain extent the
employee's self-image and sense of belonging to a close-knit 'special community,'
solely dedicated to Israel security's needs; a community inducing special codes of
conduct, whose members are trustworthy.16
Moreover, a lack of effective ties, and resulting lack of social trust, have
prevented the profit-centers' managers from taking risky decisions. Furthermore, the
importance of social networks in blocking corporate vitality is demonstrated by the
finding that even though individual managers' financial compensation was
moderately associated with their center's profitability, they lack the cognitive
corporate capital to facilitate such profitable action. They were unable to take
initiatives and risks due to the lack of trust within their corporate constituencies
(Talmud and Yanovitzky 1998). Consequently, informal social control prevented
managers from initiating business ventures and from excelling in their profit centers'
performance. Because the relations inside the firm were constructed in the pre-1985
period and were significantly sustained without corporate ability to modify them,
Ta'as lost its ability to create cognitive corporate capital. Moreover, the very social
structure which had created Ta'as' cognitive corporate capital and was beneficial to
its purposive action, become detrimental to the organization, impeding it from
attaining its corporate goals, turning a once-beneficial attribute into a corporate
liability.
MANAGING CORPORATE SOCIAL CAPITAL

The case of Ta'as corroborates Gargiulo and Benassi's conclusion (this volume) that
strong reliance on cohesive networks, which had been beneficiary to the firm,
prevented its executive officers from managing their corporate networks'
composition, and therefore compromising Ta'as' adaptability.
To the extent that a firm has structurally equivalent competitors in a market
niche (Burt and Talmud 1993), it needs to develop other compensating forms of
corporate assets to promote its comparative advantage. Burt et al. (1994)
demonstrate that for those corporations weakly situated in the exchange network,
'strong organizational culture' is moderately associated with profitability. One can
conclude, then, that where competitive corporate structure cannot be advantageous,
imposing unmanageable constraints on the firm, it is especially useful for its

managers to develop other kinds of more manageable corporate social capital,
facilitated by other dimensions of social structure.
Nevertheless, investing in corporate social capital is not risk-free. Precisely
because corporate social capital is relation-specific, it is highly costly. The keyword
in minimizing these costs is balance. A firm must 1) balance its investment in three
modes of corporate social capital; and 2) carefully determine how to construct each
investment so its sunk cost would be more sensitive and flexible to changing
conditions. The case of Ta'as reveals that concentrated, over-excessive investment
in one kind of corporate social capital may later impede corporate goal attainment.
This inference is consistent with Uzzi's conclusion that 'a firm's structural location ...
can significantly blind it to the important effects of the larger network structure'
(Uzzi 1997a).
Political corporate capital and cognitive corporate capital may be, in principle,
more flexible, indeed even more elastic, and thus may seem to be more change-
prone, because they are not merely technologically-driven. In other words, they are
more affected by the social construction of the firm (Berger and Luckman 1967;
Cicourel 1973; Nahapiet and Ghoshal 1998). Yet the paradox is that the social
structure creating political and cognitive corporate capital is one of the most rigid
dimensions of the firm. Moreover, political and cognitive corporate capital are
costly because they are relation-specific and indivisible. The litrature on corporate
social capital clearly indicates that value is created through exchange and by
combining assets. This process takes both time and relations-specific investment
enabling business organizations to develop, nurture and exploit political and
cognitive corporate structures (Nahapiet and Ghoshal 1998: 257-259). Yet the
'sticky' nature of the social structures producing them induces the alternative
hypothesis that to the extent that a firm holds crucial investments in corporate
capital, its barriers-to-exit are costly as well. The above-mentioned case of Ta'as
indicates that sometimes these two rather neglected aspects of corporate social
capital can be highly rigid precisely because they are socially situated.
DIRECTION FOR FUTURE RESEARCH

The growing number of managerial perspectives on knowledge-based firms, flexible
specialization and the high-technology sector have raised the awareness of all three
aspects of corporate social capital among managers, business scholars and
practitioners. I would hypothesize that firms investing in cognitive corporate capital
(embedded in intra-organizational relations) may be found among:
a. Mature technological business, having lengthy learning curve;
b. Service industries who focus on a market niche (Porter 1980), wishing to edge
out their competitors (Burt and Talmud 1993);
c. Potential entrants developing competence-destroying technology (Tushman and
Anderson 1986).
I also would hypothesize that 1) to the degree that the content of relations
revolves around identity formation, shared narratives and tacit norms, the relational
management is (after evolution) more rigid, and 2) to the extent that a network
Corporate Social Capital and LiabiJity- 117
circulates strategic options, entrepreneurial opportunities and material resources, its

relational management is more flexible. I would put forward the proposition,
accordingly, that one can order the degree of managerial flexibility of each form of
corporate capital as the follows: competitive corporate capital is the most flexible,
then political, and the last in order is cognitive corporate capital. Future research
should examine relational management and possible tensions between these three
forms of corporate social capital.
I would like to thank Paul Ritterband and the two volume editors for their valuable suggestions.
NOTES
I. This definition is slightly different from those of Coleman's (1994: 170) and Nahapiet and Ghoshal
(1998: 243). In this chapter I use the term corporate social capital in general, and for simplicity, I refer to
each kind of corporate social capital as 'corporate capital' (i.e. 'cognitive corporate capital' means
'cognitive corporate social capital').
2. The term 'control' is used here to refer to what Alfred Marshall calls 'the external economy' of the
firm, including its relational management, and its 'internal organizational economy,' including governance
structure (see also Simon 1991).
3. I limit my discussion here only to kinds of corporate social capital-tangible or virtual-which
benefit corporate performance.
4. Competitive social capital is a network-oriented conceptualization of economic action (see Gabbay
1997: 17-20).
5. On the difficulty in shifting social capital see Gargiulo and Benassi (this volume).
6. My use of the term 'cognitive corporate capital' is akin to the 'cognitive dimension' of social capital
used by Nahapiet and Ghoshal (1998). It is a generic term, and I do not differentiate here between sub-
forms of cognitive corporate capital (intellectual corporate capital, cultural or symbolic corporate capital),
as all of these sub-forms are intangible assets of the firm, embodied by social relations (Barney 1991 ;
Nahapiet and Ghoshal 1998).
7. This is a structural form of social capital (Gabbay 1997: chapter I), which is akin to Coleman's
closure theory of social capital (1988), and to Lindenberg' s definition of 'self-command capital'
(Lindenberg 1993).
8. See, for example, Weick (1979). See Moran and Ghoshal (1996) and Nahapiet and Ghoshal (1998)
for its relevance to organizational advantage. A useful manifestation of the self-awareness of firm-based
knowledge is the emergence of ERP systems. The sociological version of rational choice terms it 'self-
command capital' (Lindenberg 1993), while economic theory labels it 'self control,' involving two
contradictory forces in the organization: farsighted planner and myopic doer (Thaler and Shefrin 1981).
Yet economic theory seldom treats its problem setting in terms of relational management, apart from
Williamson (1994).
9. On the relations between networks and shared narratives, see especially White (1992).
10. As I limit my summary here to the firm-level of analysis, I exclude the rich literature relating to
individual and market levels of analysis. I specifically treated these levels elsewhere (Talmud 1992, 1994;
Talmud and Mesch 1997; Izraeli and Talmud 1997, 1998; Talmud and Izraeli 1998).
11. For more on the conditional nature of social capital, see Han and Brieger (this volume).
12. A complete, in-depth description of the case study is in Talmud and Yanovitzky (1998). I include
here only a few examples drawn from the case in order to illuminate my analytical substance.
13. Their marketing was performed solely by an Armament Export Unit at the Ministry of Defense.
14. All Ta'as' employees were tenured and unionized. In 1997. Ta'as outsourced its computation center
to a British-owned company which hired Ta'as former employees.
15. On the relationship between personal relations inside the organization and the creation of cognitive
and even intellectual corporate capital, see Nahapiet and Ghoshal (1998).
16. On the importance of a sense of community to the creation of corporate capital see Nahapiet and
Ghoshal (1998: 258).
Dimensions of Corporate Social Capital:
•
Toward Models and Measures
6
Shin-Kap Han
Ronald L. Breiger
ABSTRACT
Despite an emerging consensus on the importance of corporate social capital, little
work has been done on the analytical problem of which aspects, precisely, of a
corporate network might be identified as manifesting the concept. Where in a
specific configuration of network ties is the corporate social capital located? Is
network capital a unitary phenomenon or are there various ways to conceptualize it?
In addressing these questions, we formulate models for corporate networks that
produce counts for the expected number of ties between each pair of actors on the
basis of sets of parameters which are themselves measures of network capital. The
model we prefer decomposes a network into separable dimensions comprising
status, volume, and proximity. We apply the models to a network of 'doing deals' in
which billions of dollars of finance capital was raised by syndicates of major U.S.
investment banks, data of Eccles and Crane (1988). We show that the model
performs well with respect to empirical validity. The modeling framework can be
applied and extended to other corporate network settings, and provides measures
appropriate for theoretical analyses of markets and corporate relations concep-
tualized as embedded within social fields.
INTRODUCTION
Analysts as diverse as Coleman and Bourdieu put forward virtually identical
definitions of 'social capital' as denoting the resources for social attainment that
individuals acquire through networks of mutual acquaintance, obligation, and
information channeling. Bourdieu defines social capital as the sum of the resources,
actual or virtual, that accrue to an individual or a group by virtue of possessing a
Dimensions of Corporate Social Capital- 119

acquaintance and recognition (Bourdieu and Wacquant 1992: 119).
Coleman (1990: 302) too emphasizes that, unlike physical capital and human
capital, social capital 'inheres in the structure of relations between persons and
among persons.' Relevant here is the considerable body of cross-national research on
the impact of social networks and the transmission of job information on income and
occupational attainment (reviewed in Coleman 1990: 302; Lin 1990: 250-52; Burt
1992: 11-13), productivity (Bulder, Leeuw, and Flap 1996), and the organi-zational
side of job searches (Marsden and Campbell 1990). Studies such as these illuminate
concrete mechanisms by which individuals are linked to larger structures through
organizations and labor markets (see the review in Breiger 1995).
Much of the recent excitement generated by research on corporate networks
results from a focus on how the structure of their ties both affects and results from
the interests, resources, and positions of firms in the network-i.e., their social
capital. Galaskiewicz (1985) studies business philanthropy as relations among
corporations creating 'a grants economy.' Baker (1990) examines the interface
between corporations and investment banks with relation to power-dependence
concepts. Leifer (1990) seeks to explain authority and market relations among
organizations as embedded within the multiplex relations among key actors. White
(1992) emphasizes the mutual relations of corporate identity and intercorporate
structures and processes of control. Focusing on profit and a typology of markets,
Burt (1992: 82-114) shows how firms in production markets can use gaps in social
structure-'structural holes'-to their advantage in negotiating transactions with
suppliers and customers. Podolny (1993, 1994) theorizes and studies how status
orders arise from market and network relations among firms. Haunschild (1994)
investigates the effects of interorganizational relationships on the decision of how
much to pay when acquiring another company. Han (1994) shows how the interplay
between inclusion and exclusion among firms in a market yields a status dimension
affecting networks of imitation leading to isomorphism in the selection of audit
services. Many chapters in this volume also deal with the issue in a variety of
contexts.
NETWORKS AND DEALS

The new work on corporate social capital, such as the studies cited above, has
resulted in an explosion of new thinking and new knowledge about social networks.
Nonetheless, little work has been done on the analytical problem of which aspects,
precisely, of a corporate network might be identified as manifesting the concept.
Where in a specific configuration of network ties is the corporate social capital
located? Is network capital a unitary phenomenon or are there various ways to
conceptualize it? If multiple meanings exist, how might each of them be measured?
These are the questions concerning which we seek to contribute clarification and
analysis.
To illustrate the scope of our contribution, including its limitations, consider the
network of co-management relations among the major investment banks in the u.s.
in the 1984-86 period. Eccles and Crane (1988) show that, in order to do their job
effectively, investment banks create a complex network of ties to other banks, in the
form of syndicates containing 'lead manager banks' and 'co-manager banks' with a
separate syndicate organized around each specific 'deal' that is successfully put
together. Within the syndicates. lead managers and co-managers do the bulk of the
distribution. and work most closely together in the underwriting. In the aggregate
these 'deals' generated billions of dollars of financial capital in the mid-1980s in the
U.S.
The data in our Table 1 are taken from Eccles and Crane's appendix (1988: 230-
31). Rows and columns list the major investment banks in the identical order. Rows
index each bank in its role as 'lead manager' of a 'deal' in the capital market. that is.
as a bank that works with a group of co-managers to form a syndicate to underwrite
a security issue. A lead manager normally 'runs the books' (manages the
underwriting and determines distribution allocation) and is usually the investment
bank that originated the 'deal' (Eccles and Crane 1988: 237). Columns index the
same banks in their role of 'co-manager': banks that work with the lead manager and
often a group of other co-managers in the syndicate. These are the top investment
banks in the country.! Entries off the diagonal in the Table are frequency counts of
joint participation in deals. For example. during the study period Salomon Brothers
served as the lead bank in 161 deals in which First Boston was a co-manager (see
[row 1. column 2] of Table I). whereas First Boston served as lead manager in 118
deals in which Salomon Brothers was a co-manager in a syndicate ([row 2. column
1]). Entries on the diagonal report the number of times that each bank served as lead
manager without any other top bank serving as a co-manager (either because it was
sole lead manager or because no co-managers were among the nineteen top firms).
For example. during the study period Salomon Brothers led 609 deals without
participation of other banks listed in the Table (see [row 1. column 1] of Table l)?
THE SIMMELIAN PROBLEM OF ORDER

Table 1 is a network of ties among corporations. These network ties are at once
collaborative and adversarial. The ties are adversarial in that a fixed number of
securities must be allocated among the members of a syndicate. and also due to the
struggle among banks for recognition and position. The ties are collaborative in that
the offering must be distributed to desirable investors in a timely manner (Eccles
and Crane 1988: 93-94). This world of 'doing deals' is not a Hobbesian war of all
against all. which necessitates external control. but rather an instance of what we
might term the Simmelian problem of order. Georg Simmel. a theorist of sociology's
classical period. identified within certain forms of conflict a 'fight of all for all'
entailing the internal social control of an intrinsically ordered interweaving of
relations based on 'the possibilities of gaining favor and connection' (1955: 62). The
Hobbesian problem of order. as it is described critically by Markovsky and Chaffee
(1995: 255; Macy and Flache 1995) portrays the structure as seeking to extract from
its components something that they would not otherwise provide. We believe that
recent work on solidarity. with a new focus on reachability and relative unity of a
structure (Markovsky and Lawler 1994; Markovsky and Chaffee 1995) and on
models of macro-structure (Breiger and Roberts 1997). enables contemporary
g
0-
-
Table 1. Data of Eccles and Crane (1988, pp. 230-31) on ties among investment banks, 1984-1986
Co-Managers: I 2 3 4 5 6 7 8 9 10 Il 12 13 14 IS 16 17 18 19
Lead Managers:
1 Salomon Brothers 609 161 123 23 86 243 34 73 36 19 23 20 17 IS 7 4 4 7 18
2 First Boston 1I8 331 159 15 100 1lI 20 122 35 19 6 Il 27 2 7 2 4 2 3
3 Goldman, Sachs 98 64 441 Il 46 57 10 49 30 4 5 10 1 18 4 IS I 3 7
4 Drexel 9 12 7 699 17 21 7 1 17 5 5 12 23 IO 0 5 8 2 2
5 Shearson Lehman 78 99 75 8 309 59 17 28 IS 24 25 33 8 21 5 10 8 2 2
6 Merrill Lynch 123 98 47 16 63 249 15 32 24 27 17 Il 20 13 2 14 9 5 7
7 Paine Webber 28 18 6 8 30 19 420 2 10 8 Il 6 4 2 1 I 0 3 0
8 Morgan Stanley 93 65 57 32 21 56 6 92 28 12 16 IS 18 8 4 7 9 9 3 I::'
~.
9 Kidder, Peabody 22 Il 22 16 13 31 3 7 322 7 17 19 2 4 3 I 7 2 0
:;,
10 Pro-Bache 2 7 2 2 5 3 I 0 3 274 4 4 5 5 0 0 4 4 0
'"o·
II E.F. Hutton 4 2 4 5 2 I I 0 I 3 227 6 5 II 1 I 2 1 2 :;,
12 Smith Barney 9 2 3 3 8 4 7 4 4 I 9 120 6 9 0 3 2 0 0 o'"
.....,
13 Bear, Stearns 2 8 2 4 4 6 5 3 1 1 I 1 131 1 I 1 I I 1 (j
14 Dean Witter 3 9 6 3 8 7 3 0 6 5 12 9 5 86 2 0 4 1 5
15 Dillon, Read 3 9 4 0 3 2 0 I 2 0 2 2 I 2 86 I 2 0 0
.8o
16 Alex. Brown 3 2 2 4 3 3 I 2 5 0 2 1 0 0 0 80 1 3 3 ~
17 DU 1 1 0 1 2 0 2 2 0 1 7 0 0 0 0 1 83 0 0 til
on
18 L.F. Rothschild 2 0 0 0 6 I 1 0 I 0 1 2 2 4 0 3 2 71 0
19 Lazard Freres 1 3 5 0 4 7 1 0 3 3 1 0 1 0 1 0 1 1 14 ~
(j
.g.
tv
-~
researchers to address the alternative, Simmelian, problem of order: that 'the

structural membership of the individual in his group always means some mixture of
enforced limitation and personal freedom' (Simmel 1959: 47; for further exposition
see Breiger 1990). We seek formal models and quantitative measures that allow us
to tease apart the subtle intermixture of various kinds of organizing principles
(hierarchy, reciprocity, and so forth) that constitute the intercorporate network.
The technical infrastructure for our modeling project is already in place. Within
the class of multiplicative and related models for contingency tables of frequency
data, we focus on quasi-symmetry and its special cases and generalizations.
Maximum likelihood techniques for estimating parameters are well known and
widely available (see Sobel, Hout, Duncan 1985; Goodman 1984; Clogg and
Shihadeh 1994). Thus freed from the necessity of creating new models, we
concentrate instead on the call of Sobel, Hout, and Duncan (1985: 371) for
researchers to 'attach new meanings to already existing models of this type and to
generate new models for the square table.' Related work of ours includes Breiger and
Roberts (1998), Breiger and Ennis (1997), and Han and Breiger (1996); see note 6
below.
MODELS AND MEASURES

To establish notation, we will use fij to refer to an observed count in Table 1. For
example, fl2 is the count in Table 1 at the intersection of row 1 and column 2; as we
noted earlier, this entry reports that, during the period studied by Eccles and Crane,
the first-listed bank, Salomon Brothers, was the lead manager in 161 deals for which
the second-listed bank, First Boston, served as a co-manager. The set of fij defines a
quantitative network of ties among these investment banks. We seek to model this
network so as to bring out aspects of social capital. We will use Fij to refer to the
expected count that we derive for cell (i,j) on the basis of a model. We assume the
table is square and of size gxg, with the same entities (in our case, a set of
investment banks) indexed by rows and columns.
We consider several models, all of which satisfy or are special cases of quasi-
symmetry, a model which we will explicate. Other kinds of models might also be
considered for these data, but we have found quasi-symmetry models to be
particularly helpful in studying corporate social capital (cf. Friedkin 1998). These
models are distinctive in that they decompose a network of counts into separable
aspects comprising status, volume, and proximity, as we will demonstrate. The
existing literature on corporate relations is rather vague on whether these aspects are
conceptually separate, and their meanings have often been blurred. In fact, they
might be strongly related to one another; however, this should be an empirical
question. We seek explicit definitions of these concepts as features of networks, thus
enabling us to jointly model and measure the concepts in relational terms.
Following the path-breaking work of Sobel, Hout, and Duncan (1985), we para-
meterize the quasi-symmetry model as follows:
(1)
Terms on the right-hand side are parameters to be estimated. Several

stipulations are imposed. An <X is estimated for each of the g actors indexed in the
table, and the product of the <Xj is 1. It is understood that ~i = ~j if i =j. The o's are
symmetric, Oij = Oji. Maximum-likelihood procedures for estimation of the
parameters and expected cell frequencies under this log-linear model of quasi-
symmetry are well-known; interested readers are referred to Bishop et al. (1975),
Goodman and Clogg (1984), Agresti (1990), Clogg and Shihadeh (1994).
It will be useful to consider the pair of cells Fij and Fji . For example, Figure 1
indicates expected frequencies among three pairs of banks. 3 We see for example
that, according to the model, Goldman Sachs is the lead manager in an estimated
45.618 deals in which Shearson Lehman is a co-manager. In the opposite direction,
Shearson Lehman is lead manager in a larger number of deals, estimated at 59.245,
in which Goldman Sachs is co-manager.
The ratio between the two, then, indicates and measures the extent to which the
pairing between the two actors is asymmetric. In pairing Shearson Lehman (i) and
Goldman, Sachs (J), for example, it is the former who is more likely to be the lead
manager, as shown in the first panel of Figure 1 (59.245 versus 45.618). The ratio
between the two, R ij , then, is 1.299, showing that Shears on Lehman (i) is 1.299
times more likely to be the dominant partner (i.e., lead manager) vis-a-vis Goldman,
Sachs (j). In other words, Rij is a measure of the dominance of i over j.
We may form an entire g x g matrix, R, of such ratios, with
R-~
ij=T. m
)1
Although the equation above is definitional of dominance, we now turn to a

result. The matrix of R is a function of one dimension: the <X parameters in our
model. As may be seen by substitution of equation 1 into this definition of R:
F. a.
R.. =-2..=_) (3)
F
I)
ji ai
Each actor in the network thus has a unique value of alpha, and the asymmetry
in any pairing between two actors can be described in terms of the ratio between the
two alphas. With respect to the tie in network R from Shearson Lehman to
Goldman, Sachs, for example, the ratio of the alphas (shown in panel b of Figure 1)
is .667/.513, which equals 1.299, the value of Rij given previously in the text. The
vector of alphas forms a linear ordering and in this sense captures the dominance
hierarchy among all the actors in the network. 4
Furthermore, if we define a column vector a to have 1/ <Xj as its j-th element, we
can trace the relationship between our alpha and a conventional measure of
centrality in the social networks literature (Bonacich 1972, 1987; Wasserman and
Faust 1994).
The centrality of an actor is very often operationalized as the sum of an actor's
social connections, weighted by the centrality of the others to whom the focal actor
is tied. A natural implementation of this concept of centrality is that centrality is an
eigenvector of the matrix representation of the social network (see also the
applications of eigenvector centrality measures in studies of corporate interlocks
Shearson
Lehman
45.618 ) \ 20.665
9.245 5.96~
Goldman, ~ Smith
( sa:hS ~ Ba~ney \
~ 3.860 ~
Panel a: Expected frequencies (Fij) from Model 3 in Table 2.
Shearson Lehman
'1(513)~
Goldman, Sachs _ _ _ _.~ Smith Barney
(.667) 2.668 (1.778)
Panel b: Ties between firms (Rij = Fij I Fji). Values in parentheses are Clj.
Shearson Lehman
Goldman, Sachs Smith Barney

.027
Panel c: Oij; see equation 5.
Figure 1. Modeling the network of investment banks: a three bank example

such as those referenced in Mizruchi et al. 1986).

With respect to the network R defined above, we have
Ra=A.a (4)
with A. = g, the number of actors in the network. This equation establishes that the ex;
parameters of the quasi-symmetry model define an eigenvector of matrix Rand
thus, in our modeling context, represent the dimension of networks that is typically
captured by the family of centrality measures including prestige, status, and
popularity.5
This interpretation of the ex; parameters illustrates the most fundamental feature
of our modeling approach: We formulate models for the network that produce
expected cell counts for the number of ties between any two actors on the basis of
parameters (a., ~, 8) which themselves are measures of network capital.
The models and measures are duals to each other, as roles and positions are
duals to the structure. The models themselves (such as quasi-symmetry) are well-
known. Our distinctive contribution is to develop their relevance to the analysis of
network data on counts, such as the co-manager ties of Table 1. From a statistical
point of view, the fit of the models to observed networks may be assessed by means
of standard maximum-likelihood chi-square procedures, making them feasible to
apply to square tables of frequency data in many different substantive contexts. 6
In models of quasi-symmetry for network data, ~i is the average volume of ties
sent and received by actor i, controlling for the other parameters in the model. In the
present context, this parameter indexes a bank's total involvement in deals (whether
as a lead- or as a co-manager, without distinguishing between these two roles but
focusing only on the extensiveness of its ties). In this sense, ~i indexes the
(relational) volume of an actor.
The third set of parameters, the 8ij' measures how strongly or closely i and j are
related to each other, net of a. and ~. The raw expected count, F ij , is by postulation a
function of all three parameters and thus is not a proximity measure. The geometric
mean of Fij and Fji could be a proximity measure but it also suffers from a built-in
dependence on all three parameters. However, if we norm this product appropriately,
then the following is a measure of proximity between the two actors that is net of the
asymmetric status (a.) and relational volume (~) effects, as may be seen by
substitution of equation 1 (see also Sobel et al. 1985: 364):
(~X~)=5, (5)
Although it is usual to interpret the 8 ij parameters with reference to the social

process of reciprocity (e.g., Sobel et al. 1985), in the context of our corporate data
we prefer the more direct interpretation given by the equation above: 8 is a measure
of the social proximity of two actors. The relations themselves may not be
symmetric, whereas 0 is an average of their intensities, the degree to which two
investment banks are likely to encounter each other in a syndicate formed to put
together a deal. In Figure I, panel c reports the estimated 8 parameters (computed as
in the equation above) for three illustrative banks. It is seen that Goldman Sachs and
Shearson Lehman are much closer to one another (on the basis of the average net
intensity of their relations) than either bank is to Smith Barney. The 0 parameters
give us a network of proximity coefficients. The oij might well be related to
differences in status (<Xj I Clj) and in volume (/3i I /3j), as they indeed appear to be in
the illustrative example of Figure 1. Such relations among the parameters can be
investigated on the basis of a model that decomposes the expected cell counts into
just these three components.
A variety of models simpler than quasi-symmetry may be postulated; see Table
2. The simplest model we consider consists solely of the /3 parameters for volume;
that is to say, the model imposes that all <Xj = 1 and all Oij = 1. This model is taken as
the baseline model for analysis of square tables by Hope (1982), who terms it the
'halfway' model, and it is discussed by Goodman (1985) and by Hout, Duncan, and
Sobel (1987: 152), who note that the model imposes both independence and
marginal homogeneity.
The conventional model of statistical independence for rows and columns of a
square contingency table may be obtained from the 'halfway' model by adding to the
halfway model estimates of the <Xj parameters measuring dissimilarity among the
average counts in column j and in row j; see Table 2. In the independence model no
=
pairwise interactions are allowed (all oij 1). An alternative generalization of the
'halfway' model is to allow symmetric pairwise interactions (Oij) but to preserve
marginal homogeneity (<Xj = 1); this alternative defines the usual model of (full)
symmetry. Putting together, so to speak, the model of independence and the model
of full symmetry yields the model of quasi-symmetry that we have been discussing
in this chapter (see Table 2). In other words, we relax the marginal homogeneity
condition, specified in the full symmetry model (<Xj = I), while allowing the pairwise
interactions, prevented in the independence model (all Oij = I), with some
constraints.
The remaining models of Table 2 provide more parsimonious representation of
the symmetric interaction parameters (the oij), portraying these pairwise terms as one
or more dimensions of interaction, rather than requiring one parameter for each pair
of actors. Notice for example that the 'homogeneous RC(I) model' requires
estimation of only (g-l) parameters more than the model of independence (the
difference in degrees of freedom is 305 - 287 = 18; see Model 3 in Table 2) in order
to represent all the pairwise proximities in terms of a single dimension (Jl). The
'RC(2)' model uses two dimensions in preference to estimating a parameter for each
pair of actors. All the models in Table 2 are well known and are discussed in detail
by Goodman (1984, 1985) and others (Sobel et al. 1985; Hout et al. 1987; Agresti
1990). As shown by the parameter specification in Table 2, all these models are
models of quasi-symmetry.
The fit of any of these models to data on frequency counts may be assessed in
the usual way by means of a comparison of the chi-square and the degrees of
freedom left by the model (see columns labeled G2 and df in Table 2). Or, one may
measure improvement in fit relative to a baseline model. The last column in the
--l
~
0'
IV
Table 2. Models of quasi-symmetry *
Parameter Specification Clj Pi I)ij df G/ bie 1-{Gk2/Gl}

kModel:
1 Halfway Pi 323 1624.4 -1099 0.0%
2 Independence Clj Pi 305 1248.0 -1324 23.2%
o
3 Homogeneous RC(I) Clj Pi exp{-~¢{u; -ll Y} j 287 779.1 -1641 52.0% ~.
::s
en
o·
::s
en
4 Homogeneous RC(2) Clj Pi ex p{-.!. r.¢k{ut; -Ilkjr} 270 644.6 -1632 60.3% o......
2 k=l
('}
5 Full Symmetry Pi I)ij 171 777.3 -664 52.1%

.a
~
CD
6 Full Quasi-Symmetry Clj Pi I)ij 153 330.9 -959 79.6% CIl
o
~.
* As implemented in this paper, all models fit the diagonal cells exactly. Models are ordered by their degrees of freedom. a
('}
{l
IV
-..l
-~
Table provides such a measure relative to Modell, the 'halfway' model.

Alternatively, because the number of deals indexed in Table 1 is large, and to avoid
overfitting (that is, the inclusion of terms in a model for which we have little or no
substantive interpretation) we may apply the Bayesian information criterion (BIC) of
Raftery (1986; see also Raftery 1995); values of this criterion are given in Table 2
for each of our models. In applying BIC one chooses the model with the lowest BIC
value.
In fact, all the models of Table 2 are relatively more or less parsimonious efforts
to represent quasi-symmetry. For example, the oij values estimated from the one-
dimensional model of Table 2, RC(l), correlate .82 with the Oij parameters estimated
for the full quasi-symmetry model, and the single dimension estimated for the one-
dimensional model in Table 2 correlates .98 with the first dimension of the two-
dimensional model, RC(2).
According to the BIC criterion the best balance of parsimony and substance is
attained by the one-dimensional model, RC(I), with a single dimension
characterizing scores for both the row and (identically) the column categories of our
data (see Model 3 in Table 2). This one-dimensional model differs from (full) quasi-
symmetry only and precisely by modeling the pair-wise Oij parameters by means of a
single dimension of proximity, as specified by the l1i parameters in the column of
Table 2 labeled 'Oij'. The illustrative example of Figure 1 above is based on expected
counts and estimated parameters from this RC(I) model.
NETWORKS AND OUTCOMES

In this section, usil1g the model discussed earlier, we examine the social structure of
investment banking industry by analyzing the data on the ties between investment
banks. Our preferred model (Model 3 of Table 2) fits the data with three dimensions:
status (a), volume (~), and proximity (11). The estimated parameters are reported in
Table 3 for each of the nineteen investment banks in the dataset.
The model decomposes the relational structure among the investment banks into
three components. The first one, a, captures the disparity between playing the role
of the lead manager and the role of the co-manager. In other words, it measures the
propensity of an actor to be on one side of this asymmetric relationship versus the
other. In particular, the reciprocal of a measures the propensity to be the lead
manager in a deal rather than a co-manager, and indicates each actor's position in
the linear ordering of the dominance hierarchy among all the actors in the network.
The estimated a's closely reproduce the well-known 'bracket' structure in the
industry, the strong and elaborate status hierarchy among the investment banks,
typically shown in the 'tombstone' advertisements (Hayes 1979; Eccles and Crane
1988; Podolny 1993). All of the six special (or bulge) bracket firms-First Boston
(#2); Goldman, Sachs (#3); Morgan Stanley (#8); Salomon Brothers (#1); Merrill
Lynch (#6); and Shearson Lehman Brothers (#5)-are found at the top of the list of
estimated a's, with Paine Webber (#7) being an exception in the ordering. Drexel
(#4) and Dillon, Read (#15), follow closely after these six.
Table 3. Estimated parameters for Model 3 of Table 2

ID Investment Bank Special Parameters
No. Name Bracket In(1/a) In(13) In(~)
I Salomon Brothers Yes 0.9099 3.6608 1.4346
2 First Boston Yes 0.8301 3.3161 1.7183
3 Goldman. Sachs Yes 0.4056 3.2473 1.4390
4 Drexel No 0.3369 3.4433 -0.5509
5 Shearson Lehman Yes 0.6670 3.2002 0.6637
6 Merri11 Lynch Yes 0.4041 2.9608 1.0601
7 Paine Webber No 0.5593 3.2998 0.1743
8 Morgan Stanley Yes 0.8607 2.6913 1.1855
9 Kidder. Peabody No 0.2063 2.9904 -0.0587
10 Pru-Bache Securities No -0.6952 2.4590 -0.4910
11 E.F. Hutton No -0.9994 2.2128 -1.5178
12 Smith Barney No -0.5757 2.1059 -1.0161
13 Bear. Stearns No -0.8767 1.9993 -0.4697
14 Dean Witter No -0.2102 2.1221 -1.2014
15 DiUon. Read No 0.3217 2.3880 0.3272
16 Alex. Brown No -0.3679 2.0071 -0.3828
17 DU No -1.2533 1.5828 -1.4986
18 L. F. Rothschild No -0.4190 1.9218 -0.9267
19 Lazard Freres No -0.1043 1.2674 0.1109
2.v 2.u
1.0'
»
i.
~
1
1.0' , i
; 1,
1:
:c."
~5 i
j
j 4 a j
~5 i ~
~ 0.0' Yl 0.0 ~9
V:'" ~4 ~4
,... ~d.6 111 ~6
"c
.r.
If If
3-1.0 ~o -1 .0 ~o
::: ~3 ~1 ~1
~3
....c
-2.
11 _2.n
~7
-2.0 · 1.0 0.0 1.0 2.0 -2.0 -1.0 0.0 1.0 2.0
In(Beta). Standardized In(I1). Standardized
Figure 2a. Status (lIa) by Volume (13)· Figure 2b. Status (lIa) by Proximity (p).
• ID numbers in the Figure are keyed to the name of the investment banks in Tables I and 3.
Parameters are standardized by taking Z-scores.
The second parameter, p, taps into the volume effect, the extensiveness of a
firm's involvement in deals either as the lead manager or the co-manager. Salomon
Brothers (#1) is on top, followed by Drexel (#4). These two parameters, a and p, are
highly correlated with each other. Eccles and Crane (1988) observed that a firm's
hierarchical position is based partly on its volume, and the volume of securities it
gets to underwrite and sell is based in turn on its hierarchical position.
Consider, however, the joint distribution of the two. The scatterplot in Figure 2a
shows a pattern of deviation from the expected association between the two sets of
parameters, lIa and p. Although most of the firms are along or near the regression
line describing the expected linear relationship between status and volume, there are
two groups of firms that fall far outside the expected range.
1. On the one hand, there is a group of firms-Morgan Stanley (#8);

Dillon, Read (#15); and Lazard Freres (#19)-known to be Estab-
lishment firms that enjoy high status relative to their volume (Podolny
1994; cf. Stuart in this volume). Morgan Stanley is known to be 'the
bluest of the blue-chip investment bankers' (Hayes 1971: 139; Eccles
and Crane 1988); Dillon, Read used to belong to the special bracket
(Hayes 1979); and Lazard Freres represents the old Wall Street
(Stewart 1991). This provides an example of how the inertia of the
status hierarchy (a) mitigates the raw influence of the market as
indexed by overall volume (P).
2. On the other hand, another group suffers low status despite high
volume. Drexel (#4) is the most prominent case among the latter group
of outliers, which also includes Prudential-Bache (#10) and E. F.
Hutton (#11). Drexel's reputation as an aggressive newcomer and its
strong association with Junk bonds' provide partial explanations for
this discrepancy,? with Eccles and Crane (1988: 115-16) noting in
some detail that 'during our project the firm was described to us ... in
the most unflattering ways' (see also Stewart 1991). This is an example
of how the hierarchy, defined and legitimized by the participants
themselves, acts as a conservative mobility barrier to firms trying to
shift their position.
Volume is important, and it is highly correlated with status, yet the two are
separate dimensions, and one does not necessarily translate into the other. The
contrast between Lazard Freres (#19) and Drexel (#4) clearly illustrates this point
(also see the more recent case of DU-Donaldson, Lufkin and Jenrette, #17-in
Doherty 1997). The model we propose is precisely suited to such a setting, for it
allows the two conceptually independent components to be separated out from each
other. Although the two are rather highly correlated in this case, for instance, the
underlying structural dimensions they tap are distinct. As shown in Table 4, it is p
that best captures the overall volume effect, as in total market share and number of
issues. By contrast, lIa does better at explaining the dimension that is associated
Table 4. Correlations between I/a and ~ and other select variables *
Variable InOla) In(S)

Total Market Share, 1984-1986 .689 .857
Number of Corporate Securities Issued, 1986 .775 .843
Market Share in Investment Grade Bonds .687 .684
Market Share in Non-Investment Grade Bonds .274 .478
Status Score (Podolny 1993) .621 .535
* The sources for the other variables are Eccles and Crane (1988, Tables 5.4
and A.6) and Podolny (1993).
with prestige or status, an excellent example being the status scores computed by
Podolny (1993, 1994) applying a widely used measure of centrality to the placement
pattern of investment banks in the tombstone advertisements. Also revealing is the
way they correlate with the market shares of two different financial products,
investment grade bonds and non-investment grade bonds (see note 7). With the
former both a and ~ are correlated to the same degree, while for the latter, a high
risk product (Podolny 1994), the correlation with a is not statistically significant.
The third dimension, J.l, is about how close or distant the firms are from each
other net of a and ~. The more deals a pair of firms do together, the closer they are
to each other vis-a-vis the others. The vector of J.l'S thus forms a one-dimensional
space along which each firm can be located, and the proximity between any pair of
firms can be obtained from the distance between the two. The smaller the difference
between J.li and J.lj' the closer firm i and firm j are, and vice versa. For example,
Salomon Brothers (#1; J.li = 1.4346) and Merrill Lynch (#6; J.lj = 1.0601) are
relatively close to each other (J.l; - J.l:i = 0.3745), and they frequently serve as
partners. s In contrast, Morgan Stanley (#8; J.li = 1.1855) and Dean Witter (#14; J.lj =
-1.2024) are relatively far apart (J.l; - J.lj =2.3869), and they rarely do deals together.
Plotting relational proximity (J.l) against status (l/a) reveals a very important
social dynamic that occurs among the investment banks, that of status homophily.
The firms that are close to each other on J.l are likely to be near to each other on a as
well. The smaller (J.l; - J.lj), the smaller (<Xi - CX;). Or, to put it otherwise, the
investment banks that put together the deals tend to be status equals, which is the
central finding in Podolny (1993) based on somewhat different data for these same
firms.
This result also precisely matches what Eccles and Crane observed. The top six
firms, those in the special bracket, are clustered to the right in Figure 2b. They are
close to one another as a result of the security issues they do together. Yet they can
be broken down into two groups. The first one consists of Goldman, Sachs (#3),
Salomon Brothers (#1), and First Boston (#2). These three rely most heavily on
other special bracket firms as co-managers. The second group-Shearson Lehman
(#5), Merrill Lynch (#6), and Morgan Stanley (#8)-<10 so less often, bridging
instead to banks further down the status ordering. The relative location of the two
groups on J.l, i.e., the second group being closer to the rest of the banks to the left,
bears out Eccles and Crane's account. In particular, these authors report that 'a
partial explanation of the difference between the two groups is that a larger share of
the deals led by the second group, particularly Merrill Lynch and Shearson Lehman,
were security issues in which regional firms were used as co-managers to get more
retail distribution' (Eccles and Crane 1988: 94).9
IMPLICATIONS FOR RESEARCH AND THEORY

An emerging consensus on the importance of corporate social capital spans a broad
spectrum of theorists and researchers, as aptly illustrated in this volume. Although
the core ideas seem to be shared by many and have been shown to be operative in a
variety of settings, little work has been done to parse out analytically the mUltiple
and conceptually distinct aspects of social capital. In examining a network of 'doing
deals' constructed among the major U.S. investment banks (Eccles and Crane 1988)
from the analytic perspective of log-linear models, we rely on a set of well-
established techniques. Yet at the same time we heed the call to 'attach new
meanings' (Sobel et al. 1985: 371) to the models, and we expand their interpretive
horizon by means of our novel application.
We formulate models for the networks that produce counts for the expected
number of ties between each ordered pair of actors on the basis of a set of
parameters which themselves are measures of network capital. More specifically, the
model we prefer decomposes a network into separable dimensions comprising
status, volume, and proximity. We show that the model and its associated estimated
parameters, which situate the actors within a social space in relation to one another,
performs well with respect to empirical validity.
'Something about the structure of the player's network and the location of the
player's contacts in the social structure of the arena,' Burt argues, 'create a
competitive advantage in getting higher rates of return on investment' (1992: 57).
That 'something' is social capital, upon which economic constraints and
opportunities that confront a producer are very much contingent (Podolny 1993).
The approach proposed in this chapter makes it possible to formulate this intuitive
notion in a more explicit and structural manner. Providing measures appropriate to
analyze the intercorporate relations, competitive or otherwise, and a way to map the
configuration of the market as a whole, the model can easily be applied and
extended to other corporate network settings, such as joint ventures, strategic
alliances, or R&D collaboration (e.g., Freeman, Pennings and Lee, Smith-Doerr et
aI., and Stuart in this volume). The model thus should allow better informed
strategic deliberations in a variety of management settings (e.g., Porter 1980).
Furthermore, the model and its parameters offer a way to portray a social field
in general, beyond this specific framework. Consider, for instance, different
configurations of actors in <X-~ space, as in Figure 2a. For the investment banking
industry, there is a high correlation between a firm's volume of underwriting
business and its place in the industry'S hierarchy. The two, however, are
conceptually distinct. The model we propose is precisely suited to such a setting, for
it allows each dimension to be separated out from the other. It should be possible,
nonetheless, to find a case completely opposite to ours, one in which the two are
correlated negatively-as in luxury goods. Or, one might be able to find a setting in
which the mid-sized firms are the most prestigious (cf. White 1988, 1992). In <X-Il
space, shown in Figure 2b, we showed precisely a pattern of status homophily. And
yet, other configurations are plausible as well. There are settings that exhibit social
proximity between status unequals, as for example between managers and
secretaries (Kanter 1977). Curvilinear relationships are also possible, as for example
among U.S. Supreme Court justices, where we have argued that those justices 'in
between' the coalitions of liberal and conservative ideologies are highest in the status
hierarchy (Han and Breiger 1996). Such an inverse-U shape is itself in contrast to
the U-shape for plotting status against social proximity that is often postulated in the
cases of the middleman minority (Bonacich 1973).
In sum, the theoretical and research work that is necessary to further develop the
concept of corporate social capital requires structural and systemlltic measures on
networks. In this chapter we have introduced relationally based measures which
appear to be most promising for future work.
For comments on an earlier draft we are grateful to the editors and to Robert Faulkner, Noah Friedkin,
Joseph Galaskiewicz, Philippa Pattison, and John M. Roberts, Jr.
NOTES
1. That these nineteen investment banks are the major players seems to be a robust finding. Eccles and
Crane (1988: 228) report that, 'since nineteen is not a round number, we attempted to chose a twentieth
but could not. All the candidates were firms strong in narrow product categories, purely regional firms, or
varied in the strength of their performance through the three-year period' of 1984-86.
2. Of these, 488 were deals in which Salomon Brothers was the sole lead manager, and 121 others
involved deals with banks other than those listed in Table I.
3. The expected frequenCies in Figure 1 are derived from Model 3 in Table 2 of this chapter. This
model is a special case of quasi-symmetry, to be discussed below.
4. This discussion of R;Jis somewhat analogous to related formulations, including those of R. D. Luce
(who developed implications of the choice axiom for the scaling of preferences) and of S. E. Fienberg and
K. Lamtz (who showed that the Luce model implies a linear preference model in the logit scale, and who
formulated aspects of Luce's model as well as other models for paired comparison experiments in terms
of the quasi-symmetry model). See Agresti (1990: 370-74) and the brief review that appears in Breiger
and Roberts (1998).
5. Equation 4 may be obtained by substituting equation I into the definition of R;J.
6. For example, Breiger and Roberts (1998) and Han and Breiger (1996) examine networks of joining
in one another's opinions on the part of members of the U.S. Supreme Court. Breiger and Ennis (1997)
examine as a social network a cultural field of relations among writers in KOln, Germany, data of Anheier
et al. (1995).
7. A 'junk bond' or noninvestment-grade bond is a certificate of debt promising a high rate of return on
investment but carrying a high risk; these securities are often used to finance corporate takeovers.
8. Eccles and Crane (1988: 96) note that 'the tie between Merrill Lynch and Salomon Brothers was
particularly strong.'
9. See Doherty (1997) for a detailed account of DU' s (#17) relative location vis-a-vis the others and
its recent movement in the parameter space.
Organizational Standing as
7
•
Patrick Doreian
ABSTRACT
Organizations in social service delivery networks interact in order to provide many
services to a wide variety of client populations. In the course of these interactions,
staff and directors of these agencies form assessments of the utility of working with
other agencies. These assessments, as social network information, can be used to
operationalize a measure of network generated corporate social capital.
Organizations well regarded by other organizations have higher social capital.
Further, organizations well regarded by well regarded other organizations have
higher social capital. Input-output methods are used to generate measures of
standing. These assessments are also disaggregated by sector in a way that permits a
comparison of the relative contributions to social capital by sector. Data from the
SSDURC project are used to illustrate these methods.
INTRODUCTION
Social service organizations are distributed across a variety of sectors. Scott and
Meyer (1991: 117) define a sector as '1) a collection of organizations operating in
the same domain, as identified by the similarity of their services, products or
functions, 2) together with those organizations that critically influence the
performance of the focal organization.' This seems too inclusive and it is useful to
modify the first characteristic to focus exclusively on the primary functions served
by organizations in their sectors. Two organizations in the same sector need not have
the same services and products: a residential facility providing community living for
the mentally ill will differ from a multifunction agency and a psychiatric ward in a
hospital will differ from both.
Organizational Standing as Corporate Social Capital - 135
The sectors considered here are defined in simple functional terms. Units in the
education sector educate students (primarily in schools); units in the health sector
provide services to promote good health and to treat ill-health; units in the judicial
sector process perpetrators and victims; mental health agencies provide services to
deal with people with mental illnesses andlor those who are developmentally
disabled while units in the poverty sector attempt to reduce poverty and lessen the
adverse consequences of being in poverty.
Scott and Meyer (1991: 120) argue that 'the concept of a societal sector suggests
the presence of organizational systems that are, to some degree, functionally
differentiated.' The five sectors listed above can be viewed as differentiated systems.
Yet there are many overlaps and, with them, potential conflicts. If social service
agencies, having differing charters, mandates, philosophies and technologies, are
distributed across a variety sectors, there is limited consensus concerning the core
technologies used by these agencies. The presence of 'multi-problem clients' makes
this even more acute. This is particularly the case when there are poor clients with
health andlor mental health problems and who break the law or are victims of others
who break the law. Distinct agencies will have claims, with differing degrees of
legitimacy across sectors, concerning their organizational domains. Further, the
conditions are created where there can be both duplication of services and clients
who 'fall through the cracks.'
Social service organizations have to interact under ambiguous conditions in
order to provide services and coordinate their activities. It is clear that organizations
working together with, or fighting over, specific clients have impacts on each other.
Or, if a sequence of services is provided for clients in different agencies, there have
to be clear referral pathways linking the organizations. The second defining feature
of sectors, as proposed by Scott and Meyer, has limited utility. On one hand, if
agencies are linked in networks, they cannot help but have impacts on each other
and the condition is satisfied trivially. On the other hand, even if they belong to
different sectors they do have impacts on each other and the second feature is
contradicted. In the following, Scott and Meyer's second criterion for defining a
sector is discarded here.
Regardless of these distinctions, organizations still have to work with each
other. This is not a straightforward process and takes place in a web of inter-
organizational relations (IORs). In part, these relations are generated by funding
streams and, in part, they are shaped by local interactions between organizations in
some geographical location.
CORPORATE SOCIAL CAPITAL

Social service organizations are embedded in many lOR networks. It is a truism that
these networks provide access to needed resources for some organizations and that
they place constraints on the actions of other organizations in the network. Network
ties are created, changed, maintained or dissolved. In short, the lOR network (ION)
changes, and with this change, the opportunities and constraints shift also.
Organizations negotiate their way through their changing environments. As these
agencies interact, histories of interactions are generated across the network. Some
working arrangements are successful while others fail. Organizations build histories
of which ties work for them and which ties do not work. As personnel in these
agencies come to form assessments of other organizations, these other organizations
get reputations. Shrum and Wuthnow (1988) argue that the reputational status, in
part, is driven by past performance.! Such performances-and the shared histories
they create-form foundations for the reputations that agencies acquire. Agencies
acquiring positive reputations acquire also social capital while those acquiring less
positive, or even negative, reputations lose social capital or build up social liability.
Given sector memberships and uses of different treatment procedures, the worth
of different technologies becomes ambiguous. There is limited agreement on what
constitutes an 'effective' treatment modality and assessing the merits of technologies,
together with the organizations employing them, becomes subjective. Knowing the
reputation of another organization makes this evaluation simpler. Consider an
organization and suppose, for whatever reason(s), there is a need to interact with one
or more other organizations. Some criteria are needed to judge if future joint action
with another unit is worth pursuing.
One is suggested by institutional arguments where procedures and forms come
to be accepted as 'legitimate' in some sense (Meyer and Rowan 1991). Part of this
stems from which agencies and programs are funded. Getting funded is a profound
legitimating force, assuming the funder is reputable. These funds have three primary
sources: 1) extra-local governmental (Federal, State and County level) funding
sources; 2) legitimate charities, and 3) agencies purchasing services from one
another.
In the mobilization of relations for the delivery of services, agencies build social
capital through the quality of the relations that are established and used. Substituting
'agency' for 'person,' Coleman's (1990) insight that social capital is generated
through the 'structure of relations between agencies and among agencies' is
pertinent. Bourdieu and Wacquant define social capital as 'the sum of the resources,
durable network of more or less institutionalized relations .....(emphasis added)'
(Bourdieu and Wacquant 1992: 119). Han and Breiger (this volume) also use this
idea as a point of departure for establishing measures of social capital. Burt (1992:
8) views social capital of an actor as the relations the actor has with (relevant)
others. Lin (1982) regards an actor's relationships as resources for instrumental
action. Coupling these ideas, an organization's network generated social capital
comes from the number of ties with other organizations and the evaluated quality of
those ties (where the evaluations are made by the relational partners). Further, the
network generated social capital will be higher if the favorable evaluations come
from highly regarded organizations. The proposal here is to use the reputational
standing of an organization as an indicator of its (network generated) social capital.
A MEASURE OF SOCIAL CAPITAL

The foundations for doing this in a social network context can be traced back to
Hubbell (1965) who used input-output models to generate a measure of standing for
the nodes in a network. 2 Doreian (1985, 1987) applied these ideas to actors in small
groups and journals in citation networks. For this application, a more important
development is the method provided by Salancic (1986) to disaggregate such

standing measures, a procedure of particular use in studying IONs.
Let C be a (gxg) matrix containing the evaluations of agencies (by directors
and/or staft) in a square matrix form. In C, cij is the evaluation of agency j by
(representatives) of i. Let Tbe a (gxl) vector containing all of the evaluation activity
where T=C'u+Cu and u is a unit (gxl) vector. Let diag(1) be a diagonal matrix
where T is in the main diagonal. With v a (gxg) matrix of 1's and D=diag(1)v, a
matrix of weights, W, is created by the elementwise division of C by D. Next, let Sj
be the standing of the jth agency, then with ei an exogenous status measure:
(1)
which in matrix form is
s=Ws+e (2)
leading to
(3)
where I is a (gxg) identity matrix. This measure is akin to the measure of status of
Podolny and Castellucci (this volume) and, as noted by Han and Breiger (this
volume), to eigenvector based measures of centrality.
The operation in equation (3) is used to generate a set of measures of standing
for each organization in the network. However, these organizations also belong to
sectors and it is useful to disaggregate these measures into contributions to standing,
as social capital, from sectors. In the context of a journal citation network, Salancik
(1986) has provided a method for doing this. Let M be a matrix with a column for
each sector and mik is defined as:
mir 1 if organization i is in sector k

o otherwise
The intrinsic value of an organization in the ION can be viewed as stemming from
the sector to which the organization belongs. Let ek be this contribution for sector k.
and let E=diag(ek) be a diagonal matrix dimensioned by the number of sectors.
Equation (3) is then replaced by:
(4)
and if each sector has the same intrinsic value, E is the identity matrix. The
measures of standing generated, through the use of equations (3) and (4)
operationalize the idea of network based social capital.
METHODS AND DATA

The data used here come from the Social Service Delivery Under Resource
Constraints (SSDURC) study. For details see Doreian and Woodard (1992),
Woodard and Doreian (1994) and Woodard (1995). An expanding selection
procedure was used to identify all of the relevant agencies in three Pennsylvanian
rural counties. We use the Fall County network from the first wave of the study. An
expanding selection procedure, using both director and staff respondents, was used
to identify all organizations in the network (Doreian and Woodard 1992). The data
used here come from the director responses to the following question 'To what
extent do you feel your relationship with this organization is productive?' with
response characteristics: 1, no extent; 2, little extent; 3, some extent; 4, considerable
extent and 5, great extent.
A network with sixty five (65) agencies in a social service delivery network is
considered here. These agencies are located in a variety of sectors. The education
sector is represented by the Intermediate Unit (lUI) providing special educational
services and five schools (AGSCH, BSCH, CNSCH, LHSCH, USCH). From the
health sector there are two hospitals (BVILL, UHOSP), a provider of human
services (GALL), Easter Seals (ES), Family Planning (FP), the local office of the
State Department of Health (PH) and the Association for the blind (BLIND).
The judicial sector has 20 agencies in the network. The core agency is Children
and Youth Services (CYS). There are four local police departments (CPOLC,
GPOLC, MPOLC, UPOLC) and the State Police (SPOLC). Two agencies, the
Women's Resource Center (WRC) and the Family Abuse Council (FAC), provide
services, usually for women and children, to deal with domestic and marital
problems. Present also are many units of the court system-the court office
(COURT), Domestic Relations (DR), the District Attorney Office (DA), JUDGEs,
Legal Aid (LAID), magistrates (MAGRT), the Mental Health Review Officer
(MHRO), the Public Defender's Office (PDEF) and the Victim Witness Program
(VICTW). Included also is the Juvenile Probation Office (JPO), the Drug and
Alcohol Office (DnA) and a residential treatment home (ARTH).
The mental health sector is dominated by the Mental HealthlMental Retardation
Office (MHlMR) and the Community Mental Health Center (CMHC). The CMHC
is the largest mental health agency, with a large array of programs. The MHlMR
Office does not provide direct services but is the largest conduit of mental health
funds into the local network of mental health organizations. There a second (and
geographically peripheral on the county boundary) community mental health center
(DHSI) and a psychiatric hospital (HPSYC). The Base Service Unit (BSU) provides
intake and case management services for the MHIMR. Two units providing
community living arrangements for the developmentally disabled (CLRI and CLR2)
are present. CMHC runs two partial hospitalization programs: STEP is for 8-13 year
olds and PHASE is for teenagers. SRRMI provides social and residential
rehabilitation for mentally ill patients. THOME provides domestic counseling, foster
placement and adolescent pregnancy foster placement services. The Fall County
Association for the Developmentally Disabled (CDC) provides a variety of services
and the Child Development Care Center (ECDC) provides daycare and preschool
services. 3 Finally, the Child and Adolescent Social Service Program (CASSP) was
established to create coordination and referral services for agencies in different

sectors.
Fourteen units are included from the poverty/social welfare sector. The
dominant agencies are the County Assistance Office (CAO) and the Community
Action Agency (COACT). Catholic Charities (CATH) and Community Ministries
(CMIN) are religion-based service agencies for the poor. The Housing Authority
(HAUTH) provides the poor with information and resources to obtain subsidized
housing they otherwise could not afford and the Food Bank (FOODB) provides food
for the poor. The Red Cross (RC) and the Salvation Army (SARM) provide
resources for people in poverty. The Women, Infants and Child (WIC) program
provides nutrition resources and screening to identify people who are
developmentally disabled. Social Security (SS) runs financial support programs and
provides Supplemental Security Income. Head Start (HEADS) provides educational
opportunities for the poor. Both the Private Industry Council (PIC) and the Jobs
Training Partnership Act (JTPA) provide job training, remedial counseling and
drop-out prevention programs. A transportation service unit (TRANS) provides
transportation for the poor.
6 CYS
5 CMHC
DnA
MHMR
4
lUI
3 CAO
CAOCT WRC
2
o~ ______~====~ _________
Figure 1. Distribution of overall organizational standing
6
CYS
5
CMHC
4 -r- MHMR
3 lUI CAO
COACT
=
2
o
e h
bJ
m p
Figure 2. Side-by-side boxplot of standing by sector
Table 1. Measures of standing for all agencies in network

Rank Agency Sector Standing Rank Agency Sector Standing
1 CYC j 5.010 34 UWAY f 0.501
2 CMHC m 4.040 35 HAUTH P 0.489
3 DnA j 3.896 36 CMIN p 0.477
4 MHMR m 3.651 37 PIC P 0.475
5 lUI e 2.771 38 PDEF j 0.462
6 CAO p 2.510 39 VICTW j 0.455
7 COACT p 2.376 40 CLRI m 0.453
8 WRC j 2.343 41 BLIND h 0.424
9 JPO j 2.214 42 JUDGE j 0.420
10 COURT j 2.189 43 CASSP m 0.395
11 FAC j 2.059 44 FP h 0 .381
12 SPOLC j 1.739 45 ARTH j 0.380
13 CDC m 1.253 46 PHASE m 0 .339
14 HPSYC m 1.253 47 DREL j 0.306
15 DA j 1.190 48 RC P 0.272
16 MAGRT j 1.142 49 THOME m 0.268
17 UHOSP h 1.078 50 WIC P 0.250
18 CNSCH e 0 .936 51 CATH P 0.242
19 CCOM f 0.832 52 BSU m 0.227
20 PH h 0.825 53 JSERV P 0.227
21 GALL h 0.780 54 MPOLC j 0.213
22 BVILL h 0.776 55 FOODB P 0.200
23 BSCH e 0 .762 56 UPOLC j 0.191
24 LAID j 0.734 57 ECOC m 0.182
25 YMCA 0 0.724 58 CLR2 m 0.166
26 USCH e 0.717 59 DHSI m 0 .160
27 ES h 0.716 60 GPOLC j 0.152
28 AGSCH e 0.670 61 MHRO j 0. 139
29 SRRMI m 0.650 62 CPOLC j 0.130
30 HEADS P 0.644 63 TRANS P 0.117
31 SARM P 0.569 64 STEP m 0 .116
32 SS P 0.540 65 CHCTR 0 0 .047
33 LHSCH e 0 .519
MEASURES OF CORPORATE SOCIAL CAPITAL

As argued above, the measure of standing can be used to operationalize corporate
social capital. These measures are given in Table 1 where the organizations are
ranked in terms of their stocks of social capital. These measures are displayed in
Figure 1 in the form of a boxplot. Figure 2 contains a side-by-side boxplot of these
measures by sector.
Organizational Standing as Corporate Social Capital- 141
Table 2. Contributions to standing in the education sector

Percentage Contribution by Sector
Agency Education Health Judicial Mental Health Poverty
lUI 17.2 11.7 8.4 37.0 15.9
AGSCH 2.5 17.0 7.5 Sl.8 19.2
BSCH 3.4 16.0 12.8 48.2 16.7
CONSH 3.0 13.8 32.5 25.9 15.9
LHSCH 12.6 4.0 17.1 38.5 24.1
USCH 4.7 3.3 40.0 28.0 20.3
There are eight agencies whose social capital makes them stand out as high
outliers in Figure 1. The overall dominant agency is CYS, a large multi-purpose
agency, which, despite the high variability of social capital among judicial sector
agencies, remains a high outlier in that sector. The Drug and Alcohol Office (DnA),
another agency from the judicial sector, is prominent also in Figure 1. The Rape
Crisis Center (WRC) is the third prominent agency from the judicial sector, one that
was particularly active in Fall County with a high profile director who sat on many
boards. Neither DnA nor WRC stands out when attention is confined social capital
measures of agencies in the judicial sector.
Both the CMHC and the MHlMR are prominent overall-consistent with their
domination of the mental health sector. Similarly, the County Assistance Office
(CAO) dominates the poverty-social welfare sector (with one of the largest budgets
state wide due to the high incidence of poverty in Fall County). The Community
Action Agency (COACT) is large also with many programs. The Intermediate Unit
(lUI) is prominent in the education sector. However, as argued below, its
prominence has less to do with dominating the education sector and more to do with
its linking role between sectors. There are no such prominent agencies from the
health sector. The two primary funding agencies (CCOM and UWAY) are
unremarkable with regard to social capital generated through the provision of
services. 4 Having CCOM well above UW AY reflects it being the larger funding
agency. From the view of service provision, it is surprising that the Fall County
Health Center is mentioned as it is the facility within which many agencies are
located. Having it rank last seems veridical as it provides no direct services.
Disaggregation of Corporate Social Capital by Sectors

The social capital measures are more interesting when they are disaggregated into
contributions coming from sectors. These are shown in Tables 2 through 6 and were
generated by the use of equation (4). To aid the interpretation of these contributions
the following labels are used. A contribution (from a sector) is labeled primary when
it contributes 50% or more of the total social capital of an organization. A
contribution is secondary if it is between 25% and 50% and tertiary if it is above
10% and below 25%. The primary contributions to social capital are bolded and the
secondary contributions are emphasized in the following tables. Attention is
confined to contributions from the education, health, judicial, mental health and
poverty/social welfare sectors. 5
Table 3. Contributions to standing in the health sector

BLIND 24.3 5.5 5.5 37.7 21.9
BVILL 8.9 24.8 22.1 26.8 12.4
ES 10.2 8.0 6.8 32.7 36.0
FP 1.8 21.9 8.1 42.4 19.4
GALL 2.5 17.2 7.0 34.2 22.9
PH 10.5 29.5 5.9 13.4 24.4
UHOSP 4.3 12.7 18.5 35.8 24.7
The mental health sector, as a secondary contributor (37%), is the largest

contributor to the social capital of the lUI . The education, health and poverty sectors
are tertiary contributors. See Table 2. Mental health agencies were seen as useful by
the Intermediate Unit for providing their services. When the school districts are
considered, the mental health sector is a primary contributor (52%) to one district
(AGSCH) and a secondary contributor to all of the rest with contributions ranging
from 26% to 48%. This makes a lot of sense. Given the nature of this ION, the
schools are linked to other agencies through the treatment of 'unusual' students
(highly gifted, developmentally challenged and disruptive).6 The gifted and
developmentally disabled children are referred often to the Intermediate Unit 7 or to
mental health agencies . Hence the high contribution from the mental health sector
as agencies in that sector saw lUI as effective. Often, these children come from poor
households and poverty agencies are mobilized. The poverty sector is a tertiary
contributor to the social capital of all agencies in the education sector. While there
are referrals to judicial sector agencies, not all of these are seen as effective.
However, the judicial sector is a secondary contributor to the two largest schools
(USeH at 40% and CONSH at 32%). The location of the largest schools is the
location also of the largest police departments. The judicial sector is only a tertiary
contributor to another two schools. As judicial agencies process children and youth
as either victims or perpetrators they are seen as more useful when the disruptive
students are removed from the school system. That the education system does not
contribute much to the social capital of schools is not surprising. Their primary
interactions are with the lUI and agencies in other sectors and not with each other.
As shown in Table 3, no sector is a primary contributor to the social capital of
health sector agencies. The mental health sector is a secondary contributor for all but
one of the health agencies with contributions ranging from 27% (to BVILL) to 42%
(FP). It is a tertiary contributor to the remaining health unit (PH). The poverty sector
is a secondary contributor to one health unit ES at 36% reflecting the perception of
ES as an agency whose' members believed in charitable sevice. In this network, ES
worked with poverty ageffcies and was seen as having value in the provision of
subsidized service. This sector is a tertiary contributor for all of the remaining health
sector units. The health sector is a secondary contributor to the social capital of PH
and a tertiary contributor to another four units in its sector. The education sector is a
high tertiary contributor (25%) to BLIND which is consistent with the low
Table 4. Contributions to standing in the judicial sector

ARTH 6.3 3.1 46.2 11.8 29.4
COURT 6.3 1.3 69.6 10.0 11.0
CPOLC 1.7 0.8 85.2 3.2 8.1
CYS 6.5 8.4 34.5 22.9 23.1
DnA 11.4 6.4 26.5 22.6 24.8
DREL 1.6 3.8 58.4 6.1 26.4
DA 2.4 2.0 82.8 5.8 5.6
FAC 2.0 2.6 55.6 11.4 20.4
GPOLC 0.0 0.0 100.0 0.0 0.0
JPO 14.5 2.6 52.6 16.7 11.4
JUDGE 5.8 3.5 57.4 17.2 13.6
LAID 3.2 6.5 53.9 5.8 28.4
MAGRT 15.9 1.8 62.0 10.1 8.4
MHRO 4.3 0.9 79.2 68 7.6
MPOLC 0.0 0.0 100.0 0.0 0.0
PDEF 4.0 1.0 84.0 5.1 4.9
SPOLC 12.2 1.7 67.1 11.8 5.8
UPOLC 7.7 3.6 63.2 12.4 10.3
Vlcrw 4.4 3.4 70.2 10.2 9.6
WRC 2.4 12.0 32.8 20.8 23.4
contribution of the education sector to the social capital of this agency. The relations
of education sector agencies to BLIND are educational services for the blind and
relations to the health sector are less consequential for BLIND. With the exception
of the hospitals (BVILL and UHOSP), the judicial sector does not contribute much
to the social capital of agencies in the health sector-a veridical result.
The primary feature of the social capital of agencies in the judicial sector is that
they are generated primarily in that sector. There are sixteen judicial sector units for
which the judicial sector is a primary contributor. See Table 4. These contributions
range from 53% to 100% (with the latter figure holding for two local police
departments). Many of these units can be viewed as being in the court system of the
sector: Public Defender Office (84%), District Attorney (83%), Legal Aid (53%),
Mental Health Review Officer (79%), the Court (70%), Magistrates (62%),
Domestic Relations (58%), Judges (57%), Legal Aid (56%) and the Juvenile
Probation Office (53%) all have most of their social capital generated within the
judicial sector. For the remaining four units of the sector, it is a secondary
contributor to their social capital with the contributions ranging from 26% to 46%.
This is an inward looking subsystem within the overall social service system. When
children and youth enter as perpetrators they tend to remain and judicial agencies are
useful for each other within the subsystem boundaries with most judicial agencies
lacking strong ties elsewhere.
Table 5. Contributions to standing in the mental health sector

BSU 0.5 1.4 2.7 85.2 5.5
CASSP 16.1 4.8 12.1 54.7 8.6
CDC 7.6 10.6 3.9 49.4 24.2
CLRI 1.7 4.4 8.6 52.5 17.5
CLR2 2.5 6.4 12.4 17.1 29.6
CMHC 8.8 6.8 15.5 30.1 32.9
DHSI 6.6 4.9 55.5 19.0 9.6
ECOC 0.5 \.3 12.0 2.5 92.5
HPSYC 5.9 16.0 37.2 24.7 11.4
MHMR 5.2 12.9 31.8 18.6 20.5
PHASE 0.9 0.7 34.2 57.2 4.1
SRRMI 2.0 3.1 12.7 56.4 6.2
STEP O. 0.3 0.5 97.2 1.0
THOME 0.9 1.5 7.7 13.6 9.6
The poverty sector is a secondary contributor to the social capital of four

judicial agencies-ARTH (29%), DnA (25%), with CYS and the Rape Crisis Center
(WRC) at 23%. These are exactly the four agencies for which the judicial is 'only' a
secondary contributor to their social capital with contributions at 46% (ARTH), 34%
(CYS ), 33% (WRC) and 26% (DnA). All four of these agencies have significant
ties outside the judicial sector. The mental health sector is a tertiary contributor for
CYS, WRC and DnA. These are exactly the three high social capital agencies in the
judicial sector and while the judicial system overall is an inward looking system,
these three agencies have social capital generated from three sectors. The poverty
sector is a secondary source of social capital for Legal Aid (28%) and Domestic
Relations (26%) both units that provide legal services for the poor. The mental
health sector is a tertiary contributor for nine judicial agencies. With the education
sector being a tertiary contributor to only four judicial agencies and the health sector
being a tertiary contributor to only one unit in the judicial sector, it is clear that these
sectors are not a source of social capital for units in the judicial sector.
Just as there are agencies in the judicial for which social capital is generated
largely within their sector, there are corresponding (but fewer) agencies in the
mental health sector as shown in Table 5. They are STEP (97%), the BSU (85%),
PHASE (57%), SRRMI (54%), CASSP (55%) and CLRI (52%). The first three are
internal to the CMHC. Both SRRMI (mental illness) and CLRI (developmentally
disabled) provide residential treatment and are internal to the mental health sector. It
is not surprising, therefore, that most of their social capital is generated within the
sector. CASSP was a unit set up (with funds through the mental health sector) to
promote coordination of services between sectors. Having most of its social capital
generated within the mental health sector, it is not surprising that it had little success
in establishing between sector coordination. The mental health sector is a secondary
Table 6. Contributions to standing in the poverty sector

CAO 2.3 9.5 26.0 14.1 36.4
CATH 0.5 0.9 7.5 6.9 38.4
CMIN 1.5 2.6 29.8 11.4 37.4
COACT 4.3 14.0 15.6 12.2 39.2
FOODB 1.8 6.7 7.0 6.0 72.0
HAUTH 3.0 20.3 7.2 5.4 60.8
HEADS 2.0 30.4 1.7 19.6 25.8
JSERV 1.6 5.2 13.8 11.0 62.5
PIC 2.8 4.9 10.4 15.8 61.0
RC 2.1 26.0 3.2 4.6 59.4
SARM 1.7 10.8 9.5 10.7 57.8
SS 2.3 6.4 17.9 39.8 28.7
TRANS 3.7 33.3 7.6 7.6 38.7
WIC 0.9 47.6 5.1 8.1 34.5
source of social capital for the CMHC (30%) and CDC (49%). It is a tertiary source
of social capital for another five agencies. This includes MHMR for which only 19%
of its capital is generated within the sector. However, the judicial sector is a
secondary source of 32% of its social capital and the poverty sector is a source of
another 33%. As with CYS in the judicial sector, MHMR has a significant amount
of its social capital generated in multiple sectors. The CMHC has 33% of its social
capital generated in the poverty sector. The only mental health agency for which
social capital is not generated within the mental health sector is ECDC. Virtually all
of its social capital (92%) was generated by its ties to the poverty sector. As noted
above, a strong case can be made for viewing this agency as part of the poverty
sector.
Six of the agencies in the poverty sector have most of their social capital
generated in that sector. See Table 6. The Food Bank at (72%) heads this list
followed, in order, by JSERV (62%), HAUTH (61%), PIC (61%), RC (59%) and
SARM (58%). The poverty sector is a secondary contributor to the social capital of
the remaining agencies in the sector. This suggests, rightly, that this is another
inward looking sector. However, only the education sector is not a contributor to the
social capital of poverty sector agencies. The health sector is a secondary contributor
to the social capital of four poverty agencies-WIC (48%), TRANS (33%), HEADS
(30%) and RC (26%) underscoring the interdependence of poverty and health
problems. This is particularly the case for the nutritional program (WIC) for women,
infants and children. TRANS provides transportation services for health provision
services for the poor. The health sector is a tertiary contributor for another three
poverty agencies. Mental health agencies are a secondary contributor for the social
capital of two agencies and a tertiary contributor to another seven, with
contributions ranging from 40% (SS) to 11%. Finally, the judicial sector is a
secondary contributor to two poverty agencies (CMIN with 30% and CAO at 26%)
and a tertiary contributor for another four.
DISCUSSION
With organizations occupying niches located in inter-organizational networks, they
generate ties in order to coordinate their actions and deliver social services. The
services are located in different sectors and are directed at a variety of client pools.
As organizations have scarce resources, issues of how to allocate them arise. While
it is clear that some agencies cannot and do not work together, the range of
collaborative efforts is quite remarkable. Over time, collaborative (and, on occasion,
adversarial) relations are explored.8 Organizations come to acquire reputations.
(Indeed, we came to have a working hypothesis that agencies giving us trouble also
gave trouble to other agencies in these networks. It was confirmed.) The evaluations
they form of each other can be used to provide a measure of network generated
corporate social capital.
While this measure is useful and a veridical image of standing as social capital,
there are limitations to the materials presented here. One is the use of a unit vector e
for the overall measure of standing and E for the disaggregated measure. With
mUltiple waves for these networks, it is possible to use standing, s(t-l) at a prior year
as e for the estimation of social capital s(t) in the following year. In a similar
fashion, E(t-l) can be used as E to generate E(t). While it is not surprising that the
dominant agencies of sectors have high social capital, a next step will be to seek
predictors of high social capital other than size.9 Clearly, funding from extra-local
sources and the nature of the working relations will be relevant. A secondary gain
from using these measures of social capital is measurement of the relative
contributions to social capital from different sectors. The overall dominant agencies
have their social capital generated from multiple sectors. Examining the relative
contributions by sector permits images of where cracks form in the overall system.
Of particular importance is the idea that the judicial and poverty sectors are the most
inward looking sectors. It is clear that collaboration must involve agencies from
multiple sectors and, given the nature of these counties, even if the focus is on health
and mental health issues, both judicial and poverty agencies will be involved. This
will not be easy and a reasonable speCUlation is that these service delivery systems
will work best when the social capital of enough key agencies is generated from
multiple sectors. Exploring the generation of social capital in multiple sector
networks remains another important next task.
This work was supported by NIMH Research Award #ROI-MH44-1948. Comments by the editors and
Katherine L. Woodard on an earlier draft are much appreciated.
NOTES
I. They argue that another two factors-organizational structure and network location-operate. In
their framework, all three factors can be used to 'explain' the variation reputational standing. As my focus
here is measuring standing, as social capital, such explanations are not the primary focus . Also, if I am
using the network structure, within which organizations are located, to define standing, network location
cannot be a non-redundant predictor of standing. Of course, this is my problem and not theirs.
Organizational Standing as Corporate Social Capital- 147
2. Viewing the structure of an economy as a social network, this lineage can be traced back further to
Leontief (1951)
3. As this agency is primarily funded by the County Assistance Office, a strong case can be made for
including it in the poverty sector. This receives support when the contributions to social capital by sector
is examined.
4. Again, there is an emphasis on the provision of services which tends to make those agencies
providing resources to fade into the background
5. While the funding agencies (CCOM and UWAY) and the other agencies (CHCfR and YMCA)
were used in the estimation of social capital, they are not used in these comparisons. Thus the row sum
for lUI in Table 2 is 90.2% leaving slightly less than 10% coming from the 'non-sectoral' agencies.
6. These categories are not all mutually exclusive
7. This accounts for the 17% as the highest contribution to the social capital from the education sector
to an agency in that sector.
8. Han and Breiger's (this volume) empirical example considers ties that are always both cooperative
and adversarial.
9. Podolny and Castellucci (this volume) propose a status measure that seems similar to the one
proposed here. Also, Han and Breiger (this volume) are persuasive in their decomposition of social
capital into (related) components.
Customer Service Dyads:
Diagnosing Empirical Buyer-Seller
8
Interactions along Gaming Profiles
in a Dyadic Parametric Space
•
Dawn Iacobucci
ABSTRACT
This chapter proposes a methodological approach to studying buyer-seller
interactions to understand corporate social capital in consumer services. The method
combines the theoretically strong game theory tradition with newer dyadic modeling
that should provide fruitful means of characterizing buyer-seller relations.
INTRODUCTION
Managers and corporate leaders acknowledge that their satisfied customers are an
essential asset to their firms. Those customers can be other firms, in the roles of
supplier, distributor, partner, etc., or the customers can be the end-users, the
consumers. While many of the chapters in this book (in sections 3 and 4) focus on
the corporate social capital that resides in inter-firm relations, this particular chapter
focuses on the social capital inherent in the relationship of a firm with its consumers.
Within the class of purchases consumers can make, this chapter further focuses
on services industries (e.g., hotels, airlines, health maintenance, legal advice) rather
than manufactured goods (e.g., shampoo, toothpaste, blue jeans, etc.). One of the
primary distinctions between services and goods is that the former is comprised of a
greater interpersonal interaction between the buyer and seller (Iacobucci 1998): for
services, the customer engages in a transaction with the service provider, a frontline
representative of the service firm. By comparison, for goods, a customer selects the
desired packages from shelves, and deals rather minimally with a retail
representative, and not at all with a manufacturer representative. Thus, services
would seem to offer a better environment (than goods) in which to study corporate
Customer Service Dyads - 149
social capital at the consumer level (e.g., also see the chapters by Ferlie and Doreian,
this volume, who study networks of relationships in the health services setting).
Indeed, Reichheld and Sasser (1990) describe the importance of consumer
loyalty, drawing the analogy between the manufacturer's goal of 'zero defects' and a
service provider's goal of 'zero defections.' Heskett et al. (1994) take the
implications of loyalty further, demonstrating the long-term profitability of satisfied
customers; their life-time customer value. The long-term phenomenon of
relationship management (Baker and Faulkner 1991) is enhanced by positive
employee-customer relations. Strong positive interactions add value and contribute
to customer satisfaction (Crosby and Stephens 1987; Koelemeijer 1995; Price,
Arnould, and Tierney 1995) and future sales opportunities (Crosby, Evans, and
Cowles 1990). Thus it is clear that the consumer quite directly provides corporate
social capital, and that the process by which this occurs is at least in part social or
interpersonal, involving the interactions between the consumer and the provider.
Schlesinger and Heskett (1991) emphasize the interdependent nature of a firm's
employees and its customers. They illustrate the importance of a firm keeping its
employees happy and competent (i.e., well-paid and well-trained) so that they may
in turn keep the customers happy. For example, Marriott is known for relatively
satisfied employees (high retention rates) and customers (high room occupancies),
and the argument is that each drives the other. Given that customer and employee
satisfaction are positively correlated, the converse is also true; i.e., customer
dissatisfaction and poor customer service can also be attributed in part to
dysfunctional customer-employee interactions. The service provider's inability to
anticipate customer needs or recover from service failure (i.e., resolve the customer
complaint to the customer's satisfaction) is frequently seen as the source of customer
disgruntlement (Bitner et al. 1990; Brown and Swartz 1989; Solomon et al. 1985;
Surprenant and Solomon 1987). There are, of course, also other classes of
explanation for poor customer service, e.g., systemic industry problems, poor
organizational cultures, etc., but the dyadic exchange is of primary importance to the
present investigation.
In addition, we know that services purchases are typically characterized as being
more intangible, and accordingly the quality of the purchase is more difficult for the
customer to assess. Hence, customers rely upon cues to quality, including the
interpersonal nature of the service provider. Customers place credence in those
providers whom they trust, so the relational aspects that exist in the customer-
provider dyad are critical. Positive relationships, conveying trust, thus bring the
service provider the corporate social capital of drawing and keeping customers.
Researchers are coming to recognize that repeat purchases do not necessarily
imply loyalty. However, if the service organization provides quality and value and
satisfaction, perhaps with additional perks such as customization and friendly front-
line staffs, customers will be more inclined to return to these providers, given that
there are always some search and start-up costs in switching to alternatives. These
provisions, e.g., quality and value, are not monumental requests on the part of the
consumer-which is not to say that many service organizations provide such. And
certainly there are customers who are over-demanding, constituting social liability,
and firms are also coming to recognize that they do not necessarily wish to support
the efforts of building relationships with all consumers, rather presumably those
segments projected to be more profitable.
We might also offer the caution that while firm-to-firm and consumer-to-
consumer relations are those between matched entities, there is a peculiar
asymmetry inherent to a firm-consumer relation (Iacobucci and Ostrom 1996), and
such imbalances in power tend not to favor (Hibbard and Iacobucci 1997) long-term
relational development. A dissatisfied customer may vow to never return to a
provider, but there are masses ready to take his or her place, thus usually voting with
one's purchases going elsewhere rarely feels like a satisfactory solution. On the
other hand, many firms are now talking about the relationships they wish to develop
with customers, e.g., through interactive marketing, data-base and direct marketing,
etc., yet inquiries to customers suggest that their frequent user accounts do not make
relationships. Clearly the relationship with a customer is precious, and firms that
manage customer interactions can enhance these ties, strengthening the social
bonding capitals with their target segments.
In sum, provider-customer relationships do not necessarily yield social capital.
Overdemanding customers create social liability, and service providers are better off
without them. In addition, providers can not take the maintenance of profitable
customer-relationships for granted and are starting to actively intensify these ties.
If we have established thus far that consumers contribute to corporate social
capital, e.g., in the form of their relational purchasing behavior, and that services
sectors may be an ideal business environment in which to study buyer-seller
interactions, we might also query as to strategies for studying these transactions. As
in studies of inter-firm relations, most researchers survey one side of the dyadic
relationship, primarily for ease of data collection. Critics of one-sided perspectives
would say the approach is suboptimal for studying the entire relational phenomenon,
and that dyadic methods are inherently superior: both points of view in a dyad must
be integrated to understand the individual actors in the relationship, as well as the
gestalt effect of the relationship itself. For example, Swartz and Brown (1989)
studied the interaction between medical professionals and their clients as dyadic:
they obtained both parties' perspectives on perceived quality and service encounter
expectations. Researchers studying household purchase decisions must also integrate
sometimes conflicting dyadic perspectives (e.g., Corfman and Lehmann 1987;
Davis, Hoch and Ragsdale 1986). Perhaps not surprisingly, Menon et al. (1995)
found that accuracy of proxy-reports improved with greater partner communication.
This chapter will provide a dyadic analysis of buyer-seller interaction. Two
methods will be used. First, we will draw from elementary game theory, a highly
evolved theoretical framework that is fundamentally dyadic in nature. Second, we
will explore the utility of dyadic interaction models that have been created within
the social network paradigm to model data that would arise in a customer-provider
engagement. We begin with a brief description of game theory. This introduction
may be skipped for the reader familiar with the area.
ELEMENTAL GAME THEORY

We should begin by acknowledging that the area of game theory comprises a huge
Table 1. Diagram of outcomes for the prisoners' dilemma game

Suspect B:
Do Not Confess C Confess 0
Suspect A Do Not Confess C 3,3 1,4
Confess 0 4,1 2,2
Note: The first number in each pair corresponds to the outcome for suspect
A, the second number corresponds to the outcome for suspect B. Higher
numbers represent better outcomes.
literature, and developments are far more complex than we need for the current
research purposes. Many good references exist for additional information. Recent,
lucid introductions include Zagare (1984) and Van Lange et al. (1992).
This class of modeling refers to 'games' because two or more parties engage in
interactions, each player has goals to achieve, and they play by given rules-
together these properties characterize gaming. The simplest games require only two
players. Certainly researchers have extended games to more than two players, as in
studies of social dilemmas and coalition formation, but we can focus on the two-
party game, given that it most closely resembles the simple customer-employee
dyad. Even with only two players, Rapoport and Guyer (1966)' presented a
taxonomy of 78 classes of games, the most familiar of which is the 'prisoners'
dilemma game.'
The intuitive motivations of the players in that game follow (Dawes 1980: 182):
two prisoners have jointly committed a felony and have been apprehended by a District
Attorney who cannot prove their guilt. The D.A. holds them incommunicado and offers
each the chance to confess. If one confesses and the other doesn't, the one who confesses
will go free while the other will receive a maximum sentence. If both confess they will
both receive a moderate sentence, while if neither confesses both will receive a minimum
sentence.
These outcomes are diagrammed in Table 1. One suspect's choices (i.e., to confess
or not) form the rows of a 2x2 matrix, the other's comprise the columns. The Table
contains the valuation of the outcomes for each party. For example, note that if A
confesses, and B does not, A receives the best outcome (a '4') and B the worst a ('I');
and vice versa if B confesses and A does not. If both confess, the outcomes for both
are worse than if both resist confessing.
As this game illustrates, each player chooses how to behave. For simplicity,
games may be restricted to decisions between two choices. Depending on the
research context, the choices take on different guises, but one choice can generally
be referred to as 'cooperative' (commonly labeled 'C') and the other as 'competitive'
(labeled 'D' for defect from cooperation). In this game, the prisoners who choose to
not confess are said to be cooperating with their partners.
Both parties obtain their outcomes as a function of the joint dyadic choices, thus
the players are interdependent because the actions taken by each affects the
outcomes of the other. Each player is expected to maximize his or her outcomes, but
note that in this game, there is a conflict of interest between maximizing one's own
outcomes versus maximizing the joint dyadic outcome; i.e., individualistic versus
collectivistic rationality. For both A and B, confessing yields better outcomes for
themselves than not confessing, when considered independently of what the other
chooses to do. Thus, confession is said to be a dominant strategy for the individuals
involved. Unfortunately, if both players select their dominant strategy, a nonoptimal
collective outcome results, because both players could have done better as
individuals under a different choice (i.e., had they not confessed). Individual
rationality prescribes noncooperation, whereas collective rationality prescribes
cooperation (Van Lange et al. 1992).
DYADIC INTERACTION STRATEGIES

In our context, cooperating may be defined as the customer and provider interacting
pleasantly and helpfully, whereas defection may represent a break-down in smooth
interpersonal transactions. More extensively, cooperation may be the provisions of
positive elements in the marketing exchange, e.g., providing good service at
reasonable prices perhaps with some customization on the part of the provider, and
repeat purchases, positive word-of-mouth, and creative input on the part of the
customer. Defection may be more broadly defined as well, e.g., providers who over-
charge, act impudent and uncaring, are unresponsive to customer requests, or
tempermental customers who take much of the providers' time and effort without
repeat purchasing, and so forth.
Given the superiority of the collective outcomes for mutual cooperation, and the
implications for social interactions and welfare more generally, much research has
been devoted to determining how cooperative behavior might be enhanced (Pruitt
and Kimmel 1977). Three strategies seem especially effective. First, the payoff
structure might be modified; i.e., the incentive to cooperate can be increased or the
incentive to not cooperate can be decreased. Second, if the parties are allowed to
communicate, they can promise to cooperate with each other (Misumi 1989). Third,
for games with multiple trials, each player can adopt the so-called 'tit-for-tat'
retaliation strategy whereby a player chooses 'C' or 'D' at time t to mimic the other's
choice at time t-I (Axelrod 1980). Thus, a cooperative choice is answered by
subsequent cooperation, and a competitive choice is reciprocated with competition.
This responsive strategy is thought to encourage cooperation because neither party is
taken advantage of, and neither party makes a first strike; accordingly, each player
shapes the other's behavior. The retaliation strategy requires that the dyadic
interactions iterate over multiple trials.
Table 2 contains several patterns of dyadic interactions that will illustrate
different gaming structures. The leftmost column depicts the 10 trials or time
periods over which the pairs of actors interact with each other, learning how best to
respond to the dyadic partner to optimize individual or collective outcomes. The
next three pairs of columns of numbers depict three patterns of dyadic interactions
that may be described as exemplar, in that they may not be likely in real data, but
they depict particular dyadic structures clearly, at the extreme. The first of these
patterns is the 'mutual cooperation' pattern in which both parties, actor i and partner
j, make the cooperative choice, defined in these data as a '0.' In contrast, the second
pattern illustrates 'mutual competition' in that both parties choose to continually
strike the other competitively, defined by the 'l's. The third dyadic pattern is also a
Table 2. Dyadic interactions in gaming data
trial 3 exemplar strategies ~timal data interactions

mutual mutual chump retaliate random rand-comp
cooperation competition
ij ij ij ij ij ij
00 11 01 00 00 10
2 00 11 01 00 00 10
3 00 11 01 01 10 10
4 00 11 01 11 00 01
5 00 11 01 10 01 10
6 00 11 01 01 00 10
7 00 11 01 10 00 10
8 00 11 01 01 10 10
9 00 11 01 11 00 00
10 00 11 01 10 00 10
clear, but probabilistically unlikely extreme, that we may tenn, 'chump,' due to the
fact that one actor, i continues to choose to cooperate even though the other actor, j,
always responds with competition;j always takes advantage of i.
The fourth dyadic pattern is labeled 'optimal,' because it depicts the
aforementioned tit-for-tat retaliation strategy. For purposes of illustration, the
choices by j were chosen randomly (literally by the flip of a fair coin), and whatever
choice j selected at time t, actor i chose at time t+ 1.
The last two data patterns are termed 'data interactions.' The data in the pattern
designated 'random' were selected by subsequent coin tosses, resulting in a 50-50
chance for the 0-1 datum for each of the actor i and partner j at each of the ten trials.
The data in the pattern designated 'rand-comp' were selected similarly, except that
for actor i, there was a 25% chance of obtaining a '0' and a 75% chance of obtaining
a '1 '; i.e., it was a random pattern that favored the competitive choice.
The proposal in this chapter is as follows. We will model each of the three
exemplar patterns, and the optimal retaliation pattern, and thereby obtain their
parametric description via dyadic modeling. The two data interaction patterns will
then also be analyzed using the same modeling, and the resulting parameters will be
used to classify the data streams as resembling more or less one of the previous four
theoretical patterns. While the first four dyadic patterns are indeed exemplars and
therefore highly unlikely in real data, we can use them to anchor the parametric
space that can be used to characterize dyadic interactions in real data. As the
parameters describing the real dyadic interactions tend toward the ideal bounds of
the exemplar space, it will become spatially clear what sorts of structures underly
the real dyadic interactions. We turn now to the dyadic models.
DYADIC INTERACTION MODELS

The dyadic models that will be useful in this chapter have been described elsewhere
in detail, so we are brief here (Iacobucci and Hopkins 1992; Iacobucci and
Wasserman 1988; Iacobucci and Wasserman 1987). Typically, the dyadic model:
(1)
is fit via the hierarchical log linear model:
U +Ul(i) +U2(j) +U3(k) +llt(1) +UI2(ij) +UI3(ik) +U24(jI) +U23(jk) +UI4(il) +U34(k1) (2)
to the four-dimensional y-array defined as Yijkl=1 for actor i sending relational ties of
strength k (i.e., 0 or I) to partner j and receiving at strength I (also 0 or 1 for these
data)? The a, ~, and p parameters are usually of particular interest to network
researchers. The first two parameters reflect actor-level behavior called
'expansiveness' and 'popularity,' tendencies for actors to send and receive relational
ties at strengths k and I, respectively. The p parameter is dyadic in nature, reflecting
mutuality or the extent to which relational ties are reciprocated. The a parameters
would represent an actor's tendency to act cooperatively, ~ would represent the
likelihood that a party typically elicits cooperation, and p would represent
tendencies for mutually cooperative interactions. We say more shortly about these
parameters in terms of which ones would expected to be statistically significant in
the presence of different gaming strategies.
For the strategies described by the data in Table 2, however, a modification of
the dyadic model (1) may be of greater interest. In real data, we would presumably
have replicate dyads, or even groups of dyads that differed on theoretically
interesting properties (e.g., games occuring with and without communication
between parties, say). For the purposes of our illustration, Table 2 contains only a
single dyad, for each of six patterns. Furthermore, the data in Table 2 occur over ten
trials, so model (1) must be modified to allow for longitudinal structural effects. In
particular, we may begin by examining adjacent time points, a time lag of two; i.e.,
collapse over the ten trials to obtain pairs of data for times t and t+ 1. We may
aggregate Table 2 over the individuals comprising the dyad and focus on the
relational structure (note that in the very act of aggregating, the model examines
relational structure with no regard to the particular identity of the actors, but rather
the typical interaction stream, e.g., for a class or segment of customers):
V kl ,II,k2,12 = :Ei:Ej Y i,j,kl,II,k2,12. (3)
We may now fit the model:
In P{YkI ,II ,k2,12=Ykl,l1,k2,12}= 91c1+911+9k2+912+Pkl,l1+Pk2,12

+'¥kI,12+fu,l1 +$kl,k2+$II,12 (4)
via the log linear terms:
u +Ul(kl) +U2(11) +U3(k2) +U4(12) +UI2(kl.lI) +U34(k2.12)

+UI4(kl.12) +U23(k2.II) +UI3(kI.k2) +U24(1I.12)' (5)
The model (4) includes the main effects for the relational ties, the S's, which
essentially reflect the volume of 'l's (vs. 'O's). The p's are the reciprocal effects, as
per model (2), but now depicted for time 1 (or t) and time 2 (or 1+1).
The terms new to model (5) are the y's, which express a multivariate exchange:
what one actor does at time one (kl) determines with greater than random likelihood
what they receive from their partner at time two (12). Similarly, what the other actor
does at time one (11) determines what their partner does at time two (k2).
The cj>'s are also new terms. These parameters reflect multiplexity, the constancy
of behavior on the part of the actors. The term reflects an actor's autocorrelative
behavior; what an actor does at one point in time (e.g., kl) is related to what they do
at the following point in time (e.g., k2)'
INTEGRATING THE GAME THEORY STRUCTURES AND

PREDICTED DYADIC PARAMETER PATTERNS
Cooperation and Competition. We can now reexamine the six dyadic interactive
patterns in Table 2 with an eye toward these model parameters. The data patterns for
mutual cooperation and mutual competition appear similar in structure, in that both
actors' streams of relations to their partners are constant, either O's in the case of
cooperation or 1's in the case of competition. Accordingly, the parametric structure
for these two cases should also appear similar in the reciprocal structures. Both
patterns depict mutuality, albeit (1,I),s for competition and (0,0) for cooperation, so
we may test the proposition that for the cooperative and competitive interactions, the
reciprocity parameters should be statistically significant:
PI: both Pkl.1I and Pk2.12;tO.
On the other hand, the pattern for competition clearly depicts mutual strikes (1,1)
and that for cooperation is its mirror image (0,0), so the baseline volume parameters
should appear in the opposite direction (i.e., positive for competition and negative
for cooperation, given the data coding):
P1 : Skit SII, Sk2, and SI2 will be negative for cooperation, and positive for
competition.
Chump. The chump pattern should exhibit high degrees of autocorrelation because
both parties continue to do what they had previously done without having been
affected by their partner's past behavior:
P 3 : cj>kI.k2 and cj>1I.I2 will be statistically significant and positive.
In addition, the reciprocity parameters should be statistically significant, but
negative in direction to express the asymmetry of these dyadic interactions-all
dyads are of the class (0,1):
P 4: Pkl.1I and Pk2.12:;t0 and are negative.
• Retaliation
•
Chump
• Competitive • Competitive
• Cooperative • Cooperative
p p
• Random
• Rand-Ccmp
• • Random
Rand-Comp
• Retaliation
.Chu np
Figure 1. Gaming strategies in 3-dimensional parametric space
Retaliation. The retaliation dyadic exchange should not result in statistically

significant reciprocity parameters, given thatj's behavior was determined randomly,
without consideration for i's behavior:
Ps: Pkl,lI and PU.12 = 0; i.e., not statistically significant.
And yet i's behavior at time t is perfectly predicted by knowingj' s behavior at time
t+ 1. Thus, we should expect to see statistically significant multivariate exchange
parameters:
P,: {'Yk1.12} ={fu.lI} ;t O.
Random Interactive Behaviors. The final two dyadic interactions were the randomly
generated pattern, and that which was a cross between random and competitive
action patterns. For the wholly random data, we would expect all parameters to be
statistically non-significant, unless the modeling approach were hypersensitive to
erratic structures, or unless the particular random stream were odd. For the random-
competitive data, we would expect to see parameters between the near-zero random
parameters and those describing the competitive interactions. Notice too, that the
random-competitive data look fairly similar to a mirror image of the data for the
'chump' stream.
P,: All parameters should be near-zero for the random data.
P s: Parameters for the random-competition should moderately resemble those
for the competitive data.
Let us examine the results of the dyadic modeling and evaluate these propositions.
RESULTS
Figure 1 contains the results of the dyadic models applied to the data in Table 2,
presented in two plots. It represents a p-dimensional space, where p=3 is the number
of parameters being used simultaneously to create diagnostic profiles to characterize
interaction data. The first plot in Figure 1 represents the estimates Pkl ,I2 and 'YkI ,12,
and in the second, Pld.l2 is plotted against chl .u (1 and 2 represent times t and HI)?
The values plotted in Figure 1 represent the parameter estimates, though plots of the
fit statistics (i.e., essentially representing statistical significance values) confirm the
predictions regarding statistical significance. (Points far from the origin are
statistically significant; e.g., competition and cooperation strategies did not yield
statistically significant yor $ values.)
Let us first focus on reciprocity parameter, Pld.lI, comprising the first dimension
in these plots. These parameters were predicted to be statistically significant for the
cooperative, competitive, and chump patterns (PI and P4), and indeed these are all
large. The cooperative and competitive estimates are positive indicating symmetry
(O,O's and 1,1 's), and the chump reciprocity is negative indicating the asymmetry of
the (O,I),s. The retaliation dyadic interaction was expected to exhibit a statistically
insignificant reciprocal pattern (Ps), and indeed, it is near zero.
One prediction was made about the multivariate exchange parameter, the 'YkI,12'S
and one about the mUltiplex autocorrelation, $t1.U' P6 predicted that for the
retaliation data, 'Yk1.12 should be statistically significant and positive, and indeed it is.
P3 predicted that for the chump data, $11.12 would be statistically significant and
positive, and it is.
For the random data patterns, P7 had predicted that all parameters would be
near-zero and indeed none of them are statistically significant-the point
representing the purely random data maintains a position near the origin in both
plots. The prediction for the random-competition data, Pg, stated that parameter
estimates should appear somewhere between the random and competitive data
patterns, and that they may resemble the opposite of the chump parameter results.
The latter characteristic appears to have dominated; evidently the asymmetry of the
random-competition dyadic interactions drove the model parameters more than the
symmetric-appearing competition element. Finally, the main effect parameters, 9k i>
911 , 9 u , and 912 are not included in the three-dimensional plot in Figure 1, but P2 had
stated that these effects should be negative for cooperation, and positive for
competition, and indeed they were.
In terms of a brief summary, we might offer a profile of the gaming strategies in
terms of the main parameters investigated: Pk\,\i> 'Yk1 .12, and $11.12. That is, rather than
examining the plots for distances between empirical profiles and the exemplar
gaming profiles, we might alternatively find just as diagnostic an exercise by which
we compare the parameters' profiles more discretely. For example, cooperation, as
per the plots in Figure 1, might be characterized qualitatively along the three-
dimensional parameter vector as: high, low, low. Competition would be
characterized similarly, except that a fourth parameter, the 9's would be required to
differentiate these patterns. The remaining profiles are also included in Table 3.
Table 3. Profile of Kaming strategies

Ptl,l1 "(kI,12 ~1.t2 Btl
Cooperation high (+) low low
Competition high (+) low low +
Chump high (-) high (-) high (+)
Retaliation low high (+) high (-)
Random low low low
Rand-Camp high (-) medium(-) low
Table 3 also shows how the data pattern labeled 'random competition' fits none
of the exemplar row-profiles perfectly, presumably because it is more errorful, as
would be real data. Nevertheless, it is fairly straightforward to identify the closest
candidate case as the 'chump' profile for these random competition data.
CONCLUSION
This methodological approach, the combination of gaming strategies and dyadic
modeling, could be used somewhat like a discriminant analysis in that any number
of dyads may be modeled, and their resulting p-dimensional vector of parameter
estimates used to plot the dyad in space. The distance between the empirical dyads
and the exemplar gaming strategies may be interpretable like ideal point
characterizations, in that dyads lying nearer the retaliation extreme would
presumably be predicted to be more optimal in outcomes than dyads lying nearer
any of the other points. Dyads lying nearer cooperation than, say, chump or
randomness could be predicted to yield more social capital, extended longevity,
even greater profitability depending on the research context, etc. The pure data
patterns have been simulated and serve as functional anchors to the p-dimensional
space, and are therefore useful in describing the presumably more errorful data
interactions.
It should certainly be noted that this demonstration is only a beginning; both
game theory and dyadic modeling are further developed than what is suggested by
the fundamental tools used in this chapter. However, this chapter is an initial
combination of a theoretically strong gaming tradition with a newer sophisticated
dyadic modeling heritage that should provide fruitful means of studying buyer-seller
interactions.
From the models we presented, it turns out that some buyer-seller interaction
patterns lead to more intensive and long-lived cooperation than others. With the use
of empirical data, we can describe (and perhaps predict) which patterns are likely to
bring sufficient social capital to make the interaction worthwhile to both parties.
Finally, we might also suggest that while we presented this approach as ideal to
study buyer-seller interactions in a consumer services framework, the approach is
purely methodological and could certainly be applied to inter-firm relations as well.
The requisite data need only be dyadic and temporal, though surely these qualities in
data will be more easily obtained on persons rather than firms.
NOTES
1. Cited in Zagare 1984.
2. If multiple dyads were to contain common buyers or sellers, the dyads would not be wholly
independent units, but researchers have demonstrated fair robustness of this class of models to such
circumstances (e.g., Frank and Strauss 1986; Strauss and Ikeda 1990).
3. For these simple data, the results for the reciprocity parameters at times 1 and 2 were similar in sign
and magnitude, so we use only the first set for these plots. This finding is not a general result, and the
plots could be extended to four-dimensions, or any p parameters the researcher wishes to model and
examine simultaneously, e.g., including higher order terms as say, 0%1.11.>2.
SECTION II
• Structure at the Individual Level

social capital in jobs and careers
The Sidekick Effect:
Mentoring Relationships and the
9
Development of Social Capital
•
Monica Higgins
Nitin Nohria
ABSTRACT
This chapter examines the benefits and pitfalls of mentoring relationships with
respect to a protege's ability to develop social capital, measured here as ties across
multinational subsidiary boundaries that might produce access to information and
resources. The results indicate that early mentoring relationships are negatively
related to a protege's stock of social capital and that later mentoring relationships
are positively related to a protege's stock of social capital. We call for a contingency
approach to studying how mentoring relationships affect career outcomes and
discuss implications for future research.
THE SIDEKICK EFFECT

Mentoring is not a new concept. The word 'mentor' may be traced back to Homer's
Odyssey, in which a guardian named Mentor took on the role of adviser and teacher
to Odysseus' son Telemachus. Building on this ancient tradition, research on
mentoring has flourished of late. According to the International Center for
Mentoring in Vancouver, British Columbia, academic research on mentoring
includes more than 500 published articles, 225 conference papers, 150 doctoral
dissertations, 65 books and 150 mentoring program descriptions (Caruso 1992).
Most of this literature concludes that the effects of mentoring relationships are
quite positive. Mentoring has been shown to enhance career development (Kram
1985; Phillips-Jones 1982), career progress (Zey 1984), rates of promotion and total
compensation (Whitely, Dougherty, and Dreher 1991), and career satisfaction
(Fagenson 1989; Riley and Wrench 1985; Roche 1979). Given these generally
positive findings, much of the recent research has focused on two important
questions. First, what are the major factors affecting the amount of mentoring a
protege actually receives and/or initiates? And, second, what types of help do these
mentoring relationships actually provide, or, put differently, what functions do
mentoring relationships serve?
In addressing the first question, scholars have uncovered important debates
around individual difference factors such as proteges' gender (Scandura and Ragins
1993), socioeconomic status (Ragins and Scandura 1994), and personality
characteristics (Turban and Dougherty 1994). In addressing the second question,
scholars have demonstrated that mentors provide career and psychosocial help and
other functions, such as protection or role-modeling (Turban and Dougherty 1994;
Scandura 1992). These questions and answers have brought us much closer to
understanding the antecedents and consequences of mentoring relationships. Yet we
still lack a good understanding of how these relationships actually work. As Turban
and Dougherty (1994: 699) put it, 'we now need to more closely examine how
mentoring influences career success.' For example, how do career mentoring
functions such as providing exposure and visibility enable proteges to obtain
promotions?
We suggest and adopt two means of better understanding how mentoring
relationships have an impact on career progress. First, we suggest that, in order to
understand what makes mentoring relationships work, it is also important to
consider the conditions under which mentoring relationships might not work. To do
this, we consider both early and late mentoring experiences and generate hypotheses
regarding both when mentoring should work and when it should not work with
respect to individual career variables. Second, we suggest that, in order to develop
theory in this area, it is important to examine specific mechanisms through which
mentoring might facilitate career progress. To do this, we focus our research on one
career factor that has been empirically linked to upward mobility-the development
of beneficial social ties (social capital), or ties with others in an organization. PuIling
these two approaches together, then, we examine the benefits and the pitfalls of
career mentoring with respect to a protege's ability to develop social capital.
Researchers have already cautioned that there may be liabilities associated with
having a mentor (also see Brass and Labianca, this volume). Scholars have warned
that having a mentor creates the possibility for overdependence (Fagenson 1988;
Zey 1984), smothering (Kram 1985), anxiety that can turn to dependence (Clawson
1980), and clinging (Clutterbuck 1985). Still, most of the empirical work on
mentoring has focused on positive career outcomes. And, when negative outcomes
have been explored, the research has tended to focus on detrimental effects to the
mentor-protege dyad itself and not to relationships that extend beyond the mentoring
alliance (for an exception, see Fagenson 1994).
The career variable we study here, social capital, measures an individual's
potentially beneficial relationships beyond a particular dyad. We define social
capital here as the number of ties an individual has across multinational subsidiary
boundaries. Research on networking and on multinational organizations has
demonstrated that the volume of social relationships is related to upward progress in
individuals' jobs, higher paying positions, and career performance (Burt 1992; Lin
1982; Nohria and Ghoshal 1997). The volume of a protege's cross-subsidiary ties
The Sidekick Effect - 163
thus reflects social capital by providing the protege with access to information and
resources which are beneficial to his or her individual performance. Although
scholars of the mentoring literature have not directly tested the effect of mentoring
on this career variable, many have suggested that mentoring relationships may
enhance career progress because these relationships enhance the protege's ability to
develop ties with others. Specifically, researchers have espoused that career
mentoring enables a protege to overcome barriers of access to elite groups, to bypass
hierarchies (Fagenson 1994), and to have a 'special form of entry into important
social networks' (Dreher and Ash 1990: 540).
In this article, we empirically test the link between mentoring and social capital
by testing hypotheses regarding when mentoring should and should not work to
increase a protege' s stock of social capital. We do not take the blanket view that
mentoring is either generally beneficial or harmful, but rather argue that mentoring
may have its own time and place in helping an individual develop social capital.
THEORY AND HYPOTHESES

Early Career Mentoring
In general, scholars concur that mentoring relationships provide at least two major
types of assistance: career help and psychosocial help. Career help includes
sponsorship, coaching, protection, exposure, and challenge, while psychosocial
functions include role modeling, counseling, acceptance, confirmation, and
friendship. During the 'initiation' stage of a mentoring relationship, usually during
the period from six months to a year on the job, mentors are said to perform career
helping functions for the protege, such as providing exposure, visibility, and high-
profile work assignments (Hill and Kamprath 1991; Krarn 1985). At the beginning
of the second (or 'cultivation') stage, a period of two to five years, both the mentor's
career development functions and psychosocial functions emerge-with the career
development functions emerging first. Thus, at the initial stages, the mentor is said
to provide the protege with a maximum level of 'challenging work, coaching,
exposure-and-visibility, protection, and/or sponsorship' (Krarn 1985: 53).
What are the short term benefits for proteges receiving early career help from
their mentors? Research and practical experience provide us with at least three
possible answers. First, career mentoring signals the value of the protege to the
organization and, hence, facilitates the newcomer's socialization process. Second,
career mentoring provides the protege with quick access to important information
and resources by lending the mentor' s social capital to the protege. Third, career
mentoring not only brings the protege 'to the racetrack' by providing access, but
actually places him or her 'on the starting blocks' by providing concrete
opportunities to display valuable skills and talents. We review each of these benefits
in turn, paying particular attention to how these factors affect the protege's stock of
social capital over the course of his or her career in an organization.
Mentors as Signaling Agents

Having a mentor early in a protege's organizational career I can signal to others that
the protege is a valuable player in the organization and, hence, can smooth his or her
socialization into the organization (Van Maanen and Schein 1979). Because
organizational members can't fully assess the qualities of newcomers, they rely upon
information that reflects the protege's prior work experience, such as a resume or
recommendation, and they rely upon available signaling devices, such as who hired
the individual or who the individual's mentor is. Both actions taken on behalf of the
protege by the mentor and the protege's mere association with the mentor can
facilitate this socialization process. Specifically, providing exposure and visibility
by introducing the protege to high status others in the organization grants the
protege 'reflected power,' (Fagenson 1994; Kanter 1977), allowing the protege to
reap positive short term rewards in two ways. First, the protege's reputation may get
a boost as he or she rides on the mentor's coattails of success (Whitely, Dougherty,
and Dreher 1991). Second, people's first impressions of the protege may be
enhanced since others may assume that the mentor has, to some degree, selected this
person above others to be his or her protege.
However, beyond these immediate rewards, there are also liabilities associated
with developing one's reputation in an organization through an early association
with a mentor. Kram (1985: 26) describes such liabilities as 'the risk that others will
question the extent to which the younger individual can thrive on his or her own
without a particular senior's support.' The protege's reputation can become
inextricably linked to that of the mentor such that his or her competence, apart from
that touted by and associated with the mentor, is called into question. Therefore,
while short-term benefits may exist, the protege may, over the long run, be viewed
as the mentor's sidekick rather than as a worthy individual contributor with whom to
develop an independent relationship. Consequently, we expected that having a
mentor early in a protege's organizational career would be negatively associated
with the amount of social capital the protege was able to develop later on.
Mentors as Access-Providers
Having a mentor who provides career help early can also assist a protege by giving
him or her quick access to important information and resources. In effect, the mentor
lends the protege his or her social capital (Burt 1997) by introducing the protege to
others in the organization. This enables the protege to bypass traditional hierarchies
(Fagenson 1994) and barriers to information and enables the protege to tap more
directly needed resources. Therefore, not only do career functions such as exposure
and visibility grant the protege reflected power in the organization, they grant the
protege 'legitimacy' as well (Burt 1997)-Iegitimacy that is needed to get things
done or to get others to get things done for the protege.
In the short run, borrowing a mentor's social capital provides the protege with
the resources he or she needs to get specific tasks accomplished. It may also increase
the protege's returns of human capital (Burt 1997), since it positions the protege in
spots in which his or her knowledge can increase. However, there are long term
costs associated with borrowing someone else's social capital. Relying upon a
mentor for access to social networks may, over time, reduce one's own motivation
to develop instrumental relationships with others in the organization. As Fagenson
(1994: 56) warns, mentoring functions may 'prompt [proteges] to be less attentive to
and concerned about relationships external to the mentor-protege union ...
[suggesting] that proteges will have less favorable peer, superior, and department
relationships than nonproteges.' Thus, we again expect that having a mentor early
will be negatively associated with a protege's social capital over the longer run.
Mentors as Opportunity-Providers
Having an early mentoring relationship can also benefit a protege in the short term
by giving a protege opportunities to prove him- or herself to others in the
organization. Providing high visibility work assignments is beneficial beyond
merely introducing a protege to important others; it gives a protege the chance to
demonstrate unique skills and competencies to those who might not yet know him or
her. The risk of providing such opportunities early is that, despite the mentor's best
intentions, the protege might not be ready for such assignments. Noe (1988a: 475)
has implored scholars to develop a 'readiness for mentoring' scale to ensure that
proteges who are selected for mentoring programs in organizations actually benefit
from their mentoring experiences.
We agree with Noe that advising is not necessarily advisable. There are indeed
long term risks to providing a protege with high-exposure work opportunities before
he or she has built sufficient individual-level competence and confidence at work.
From an outside-in perspective, organizational members might view this early
exposure as 'jumping the gun' or providing the protege with too much, too soon.
Jumping the gun could result in early career failure or, at a minimum, could result in
the perceived risk of failure, such that the protege has neither the capability nor the
referent power (French and Raven 1968) needed to develop substantial social
capital. From an inside-out perspective, providing career opportunities early, when
the protege has not yet developed a sense of self-efficacy or confidence in his or her
ability to perform (Bandura 1977) might backfire in building alliances with others.
The social support research suggests that people with high self-efficacy are better
able to obtain effective support from their established social networks than people
with low self-efficacy (Wills 1991). Without such confidence and competence, we
expected that, over time, proteges would be unable to maintain the social capital
they had borrowed from their mentors and/or that they would be unable to develop
substantial social capital on their own:
HI: Having a mentor early in a protege's career in an organization will be
negatively associated with his or her long term stock of social capital. 2
Later Career Mentoring

The aforementioned liabilities associated with receiving early career help from a
mentor do not apply to the protege who has spent more than a couple of years in an
organization-i.e., once the socialization process is over. By then, he or she should
have established some sort of reputation and identity in the organization so that there
should be substantially less risk that the mentor would smother the protege and,
hence, prevent him or her from developing ties with others. And, with his or her own
professional identity established, there should be less risk that others would view the
protege as his or her mentor' s sidekick. Rather, we expected that the mentor's role
as signaling agent would shift from socializing the protege to confirming the
protege's value to the organization. Now, more than before, the mentor's affiliation
with the protege should signal to others that the protege is a valuable organizational
member-that the protege has actually demonstrated potential. Therefore, the
mentor should accentuate, rather than overshadow, the protege's accomplishments.

Hence, we expected later career mentoring to be positively associated with a
protege's stock of social capital.
Regarding access to important resources and information, earlier concerns
should subside once again with respect to later mentoring relationships. After
several years in the organization, the protege is no longer an organizational member
seeking legitimacy who needs to borrow another person's social capital in order to
gain organizational legitimacy. The protege should have his or her own ties with
others in the organization and, hence, should be attentive and motivated to develop
ties with several organizational members. Moreover, at this later stage in the
protege's career in the organization, the mentor may serve as a linchpin to facilitate
the connection of clusters of subnetworks of individuals in the organization. In this
way, the mentor's introductions should build upon or enhance the protege's own
stock of social capital. Therefore, we again expected a positive relationship between
having a mentor later in a protege's organizational career and having a large number
of intraorganizational ties.
Finally, we do not believe that later mentoring relationships that provide high
visibility work opportunities run the same sort of risk of jeopardizing the protege's
ability to develop social capital as is the case for the new protege. At this point, there
should be less risk that the protege would fall on his or her face by jumping the gun,
since he or she had already joined the race, so to speak. Although the stakes may
still be high for individuals later in their organizational careers, we presume that the
protege has amassed sufficient self-efficacy to benefit from potential helping
situations and to view them as challenges rather than as threats. Therefore, we
expected that the opportunities provided by a mentor to a protege later in his or her
organizational career would facilitate the development of social capital and, hence,
would be positively associated with a protege's stock of social capital:
H2: Having a mentor later in a protege's career will be positively associated with
his or her long term stock of social capital.
METHODS
Respondents and Setting
Data were collected from 323 managers in the consumer electronics businesses of
three major multinational corporations (MNCs), Philips of Holland, Matsushita of
Japan, and Nippon Electronics Corporation (NEC) of Japan. Managers were either
department heads or general managers of subsidiaries in 24 countries in which all
three organizations had national subsidiaries. The subsidiaries in our sample were
selected in consultation with corporate managers of the MNCs and constituted, in
their view, a representative sample of all their subsidiaries. Questionnaires were
mailed to all departmental managers in each of these subsidiaries. The response rate
to the survey was 87% in Philips, 93% in Matsushita, and 75% in NEC, representing
an overall response rate of 83%. In no subsidiary did we get a response rate of less
than 60%.
The population from which our sample was drawn was a well-defined group of
upper middle-level managers in all the firms. Respondents had worked for an
average of 13.6 years in their current organization and had careers spanning more
than 20 years across different organizations. Although we did not collect data on
their ages, we estimate that the median age of the managers in our sample was 40+
years. All the managers were men, reflecting the gender bias of the population
surveyed. There were no female senior subsidiary managers in the particular MNCs
we surveyed.
Because of the uneven distribution of responses across countries (there were
several small countries in which there were only a few subsidiary managers at the
level we had selected), we narrowed the sample to include respondents from only
the top ten countries (on the basis of the total number of respondents in all three
firms) for this analysis. After cases were eliminated because of missing data, the
final number of managers included in our analysis was 177. The composition of our
final sample was not statistically significantly3 different than the population
surveyed (based on t-tests comparing the means of all the variables included in our
survey).
Measures
Social Capital. An individual's social capital is generally determined by the
breadth or range of his or her social networks as well as the position that the
individual occupies in them. People are presumed to have greater social capital if
they have greater range (i.e., they have contacts that span the different clusters
within the network) and they occupy a more central position or are positioned to
take advantage of the structural holes in the network (Burt 1992; Nohria 1992b).
In this study, we operationalized social capital as the range or the number of
national subsidiaries of the firm in which the respondent had contacts. Range was
calculated as the number of different subsidiaries in which the respondent reported
having a contact with whom he communicated, regardless of the frequency.
Given that each national subsidiary represents a relatively separate network,
these contacts can be viewed as 'bridging' ties as they allow the individual to tap into
otherwise disconnected social clusters (Granovetter 1973). Indeed, Burt (1992) has
suggested that the number of bridging ties possessed by an individual may be one of
the best measures of social capital. Our focus on ties across subsidiaries was also
influenced by previous research on multinationals that shows that such ties are
crucial conduits for information and resource flows and have a positive impact on an
individual's performance (Nohria and Ghoshal 1997). Research by Brass (1984) and
Blau and Alba (1982) provides additional support to the notion that ties across
organizational subunits are important to individual career outcomes.
We recognize that-although it is an important measure of an individual's social
capital-range (as we have operationalized it here) is, nevertheless, a limited
measure of social capital. It does not fully capture the strength or weakness of a
person's network position. Ideally, we would have liked to measure the individual's
centrality as well as autonomy in various networks. However, that would have
required data on the ties among all the actors in the multinationals we studied-a
data collection effort that was impractical because of the very large number of
individuals who would have to be surveyed.
Mentoring Relationships. To measure whether respondents had early and later
mentoring relationships, we asked them if they had developed a close personal
relationship with any senior manager in the company during their first two years of
employment with the organization and also if they had such an informal advisor or
sponsor at present (again, after working an average of 13.6 years at the company).
These questions were based upon Kram's (1985) work on mentoring relationships
and are consistent with more recent definitions of mentors as high-ranking,
influential members of an organization who are supportive of their proteges' career
development (Ragins and McFarlin 1990). These questions allowed us to examine
the existence of an informal mentor at two distinct points in time to see whether, and
when, such mentors affect an individual's formation of social capital.
Controls
We included several control variables in our analysis in order to account for factors,
other than the individuals' mentoring relationships, that might influence their range
of intersubsidiary contacts. A brief description and rationale for including these
controls follows.
Firm. We included dummy variables to control for the possibility of firm-
specific variations across the three firms. One might expect that being a part of a
firm that placed greater emphasis on career development might have an impact on
the effectiveness of early and/or later mentoring relationships, as well as a protege's
stock of social capital. Since we lacked information regarding the relative emphasis
placed on career development for these three firms, we had no expectations that one
firm would be more important with respect to a protege's ability to develop social
capital than another.
Subsidiary. Within any multinational system, some subsidiaries may be more
important and central than others in the internal resource-flow network (Bartlett and
Ghoshal 1989; Gupta and Govindrajan 1991). In such subsidiaries, managers may
have more opportunities to form intersubsidiary contacts. To control for this source
of internal variation within any MNC, we included a variable that measured the
relative importance of the respondent's subsidiary in the multinational system. This
measure was based on ratings of each subsidiary' s relative strategic importance and
resource capabilities provided by five headquarters-level respondents in each
multinational. The final score was the average rating of the subsidiary on an additive
scale comprising both dimensions. The scale ranged from subsidiaries that scored 1
(low importance) to 5 (very important). Given that our final sample was composed
of responses from the top ten countries across these multinationals, we ended up
with subsidiaries that were all quite important in each multinational, as reflected by
the high mean and low variance of the subsidiary measure for our sample of
respondents.
Function. Because our respondents were either functional department heads or
general managers in their respective subsidiaries, we controlled for their job
responsibility. Several dummies were included to control for the possibility that
being a general manager as opposed to a functional head of marketing-or
manufacturing, research and development, finance, purchasing, legal affairs, or
administration--could have an influence on an individual's social capital. We
expected general managers to have a broader range of contacts than the heads of
functions by virtue of their more integrative organizational role.
Career History. We also controlled for variations in the career histories of our
respondents that might influence their social capital. We included variables that
measured their tenure in the firm, the number of years they had worked in other
organizations, the number of other functions they had performed within the
organization, the number of other subsidiaries in which they had worked, the speed
of their promotion to their current position, and whether they were expatriates. We
expected each of these measures to have a positive influence, if any, on an
individual's range of intersubsidiary contacts.
Networking Opportunities. We controlled for differences in social capital that
might arise as a result of activities in which the respondent participated that could
provide him with unique opportunities to network or make contacts. In keeping with
work done by Ghoshal, Korine, and Szulanski (1994), we controlled for the time
(days per year) the respondent spent attending meetings and conferences;
participating in committees, teams, or task forces; attending training programs; and
whether or not the manager had received any formal training upon joining the
company. We expected all of these networking opportunities to have a positive
influence, if any, on the range of contacts developed by the managers in our study.
Perceived Constraints. Finally, we tried to control for the extent to which the
respondents perceived they had autonomy in their organizations and the extent to
which they perceived that they were not restricted by formal organizational rules and
procedures. Both measures were based on scales of perceived autonomy and
informality proposed by Ghoshal and Nohria (1989). We expected that managers
who perceived their jobs as more autonomous and/or informal would be more likely
to develop more extensive social networks than those who felt more constrained.
Analyses
We used multiple linear regression (MLR) initiaIly to explore the effects of the
various independent variables on range. The output of this first regression suggested
that there are two clear groups of respondents: those with many ties, or high range,
and those with some ties, or low range. We defined 'many ties' as having 11 to 17
ties with other subsidiaries, while those 'some ties' as having from 1 to 10 ties with
other subsidiaries. No individual responded that he had no ties at all with other
subsidiaries. Given the results, it was appropriate to recode the subjects into these
two groups and to perform subsequent analyses based on this new dichotomous
dependent variable. 3
Following correlational analyses, two different levels of analyses were
conducted. First, using logistic regression, we tested the initial hypotheses: having a
mentor early hurts one's development of social capital, while having a mentor later
helps a manager develop a broad range of intersubsidiary contacts. Second, using
analyses of variance, we examined more closely the differences among individuals
falling into one of four categories: 1) those who had a mentor neither early nor late;
2) those who had a mentor late but not early; 3) those who had a mentor both early
and late; and 4) those who had a mentor early but not late.
Table 1. Means, standard deviations, and correlations a
Variables Mean s.d. 2 3 4 5 6 7 8

1. Mentor early .46 .50
2. Mentor late .65 .48 .18*
3. Employee of .60 .49 -.03 .20**
Matsushita
4. Employee of Philips .17 .38 -.02 -.14 -.55 **
5. Relative importance of 3.71 .76 -.13 .06 -.09 .10
subsidiary
6. General manager .14 .34 .17* -.12 -.21 ** .22** .05
7. Marketing function .22 .42 -.13 -.01 -.04 .01 .06 -.21 **
8. Manufacturing function .19 .39 -.00 .05 -.16* -.10 -.19* -.19* -.25**
9. Purchasing function .07 .25 .02 .01 -.01 -.00 -.03 -.11 -.14 -.13
10. Research & .06 .23 -.03 -.03 .10 -.11 -.07 -.10 -.13 -.12
development function
II. Finance function .15 .36 -.03 .07 .01 .03 .05 -.16* -.22** -.20**
12. LegaVother .10 .30 .09 -.00 .03 -.10 .09 -.13 -.17* -.16*
administrative functions
13. Tenure at current finn 13.58 7.69 -.09 .09 .18* .21** -.04 -.03 -.07 .18*
14. Years at other 7.10 8.58 .06 -.17* -.17* -.02 .12 .08 .03 -.12
employers
15. Number of other 2.20 1.98 .18* .01 -.06 .05 .02 .22** -.05 .08
functions worked in
16. Number of other .98 1.56 .02 -.00 -.03 .13 .05 .03 .03 .13
subsidiaries worked in
17. Speed of promotion .23 .12 .11 -.11 -.11 -.11 -.11 .12 -.05 -.12
18. Expatriate status .41 .49 -.24** .23** .12 .05 .06 -.10 -.03 .13
19. Time spent in cross- 13.69 15.04 .09 .04 .03 -.02 .25** .13 -.22** -.10
subsidiary meetings
20. Time spent in cross- 7.99 12.01 .07 .12 .00 -.15* .17* .07 .11 -.01
subsidiary teams
21. Time spent in cross- 4.39 10.22 .15* -.04 -.04 .22** -.06 -.01 .03 -.01
subsidiary training
22. Received initial training .47 .50 -.12 .32** .41** -.22** .01 -.11 -.12 .16*
23. Perception of 3.58 .74 -.09 .14 .21 ** -.10 .07 -.18* -.05 .04
organizational
24. Perception of 2.29 .57 -.19** -.13 -.22** -.17* .09 -.04 .12 -.12
organizational
informality
25. Social capitalb 9.95 6.91 -.23** .15* -.14 .05 .18* -.07 .14 -.17*
*p<.05, ** p<.OI, aN = 177, Social capital is defined as the range or the number of national
subsidiaries of the firm in which therespondent has contacts.
9 lD 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-.07
-.11 -.lD
-.09 -.08 -.14
-.04 -.02 .07 -.14

-.01 -.11 -.09 .12 -.55**
-.04 -.05 -.24** .08 .18* -.00
-.05 -.06 -.03 -.06 .18* -.08 .49**
.18* .06 -.05 .04 -.42** -.34** -.16* -.07

.04 -.06 .11 -.20** .46** -.46** -.07 .17* -.11
-.lD -.01 -.08 -.12 .09 -.11 .08 -.04 .02 .15
-.02 .02 -.11 -.01 .00 -.02 .06 .02 .00 .06 .21 **
.12 .14 -.09 -.03 .04 -.07 .02 -.01 .08-.15* .05 .08
.06 .01 -.04 -.08 .33 ** -.36 ** .02 .13 -.lD .51 * .07 .10 -.01
.06 .13 .lD -.01 .24** -.20" -.17* -.14 -.12 .20** .lD .03 .17* .25**
.05 .02 -.lD .11 -.16* .lD -.02 -.01 .07 -.02 -lD .03 -.04 -.06 .03
.04 -.00 .12 -.02 -.06 .04 -.13 -.00 -.03 .18* .04 -.lD -.16* -.04 .03.18*
RESULTS
Correlational Results
Table 1 presents the means, standard deviations, and correlations of the variables.
Intercorrelations between the control variables were of low to moderate strength.
The correlations between the control variables and the two mentoring variables were
all low, ranging from 0.00 to 0.24, indicating no problems of multicollinearity for
our regression analyses.
Logistic Regression Results

The results from the logistic regression are shown in Table 2. The logistic model fits
the data nicely. Overall, 72.32% of the respondents were correctly classified by our
model. The goodness-of-fit statistic (chi-square of 171.00 with 151 degrees of
freedom) further confirms that the model fits the observed data. We observed a
statistical significance level of p=0.13, indicating that our model does not differ
statistically significantly from a 'perfect' model in which the observed probabilities
exactly match those predicted by the model. Finally, similar to the overall F-test for
multiple linear regression, the logistic model chi-square (chi-square = 50.7, 25
degrees of freedom, p=O.OO 17) allows us to reject the null hypothesis that all of the
coefficients in our model are 0, except the constant. In sum, our model fits the
observed data quite well.
Our results show that the only variables that have a statistically significant effect
(at the p<0.05 level) on the respondent's range of contacts are having a mentor late,
having a mentor early, and the amount of time respondents spent in teams. These
results confirm hypotheses 1 and 2. Specifically, having a mentor early decreases the
odds-whereas having a mentor later increases the odds-of an individual being in
the high-range group. Counter to our expectations, the time a manager spent in
meetings actually had a negative impact on their range of contacts. With respect to
the other controls, the only other variable that had any statistically significant
influence was expatriate status, which had a positive impact on range if we relaxed
the confidence level (p<O.1 0).
The values shown in the last column of Table 2 allow us to gauge the magnitude
of these effects. They indicate the amount by which the odds are increased that a
respondent would fall in the category of those with a broad range of contacts, given
a unit increase in the explanatory variable. Our results indicate that if a respondent
had a mentor early in his career (the 'mentor early' variable changes from 0 to 1), the
odds of the respondent's being in the high-range group falls statistically significantly
(by a factor of .41). In short, an individual with an early mentor is only 40% as
likely as someone who does not have an early mentor to fall into the high-range
group. In contrast, we find that if a respondent has a mentor later in his career (the
'mentor late' variable changes from 0 to 1), the odds of the respondent being in the
high-range group increase dramatically (by a factor of 3.81). These odds ratios
confirm our main hypotheses.
In addition, expatriate status had a sizable effect on range. If the respondent was
an expatriate, he was much more likely (by a factor of 2.67) to be in the high-range
group. This result makes sense because expatriates are likely to have a cosmopolitan
Table 2. Results of logistic regression analysis for social capitala,b

Regression Standard Significance Odds
Variables
Coefficients Errors Levels Ratios
Mentor early -0.90 0.41 0.03 0.41
Mentor late 1.34 0.46 0.00 3.81
Employee of Matsushita -0.61 0.53 0.25 0.54
Employee of Philips 0.06 0.73 0.94 1.06
Relative importance of subsidiary 0.30 0.26 0.25 1.35
General management 0.52 0.95 0.59 1.68
Marketing function 1.13 0.93 0.22 0.31
Manufacturing function 0.10 0.93 0.91 1.11
Purchasing function 1.16 1.11 0.30 3.18
Research & development function 1.19 1.16 0.30 3.29
Finance function 1.02 0.97 0.30 2.77
LegaVother administrative functions 0.74 0.99 0.45 2.09
Tenure at current fmn -0.01 0.05 0.92 1.00
Years at other employers 0.03 O.OS 0.48 1.03
Number of other functions worked in -0.09 0.12 0.44 0.91
Number of other subsidiaries worked in 0.04 0.16 0.79 1.04
Speed of promotion 0.53 3.13 0.86 1.71
Expatriate status 0.98 0.52 0.06 2.67
Time spent in cross-subsidiary meetings 0.01 0.01 0.45 1.01
Time spent in cross-subsidiary teams -0.03 0.02 O.OS 0.97
Time spent in cross-subsidiary training -0.03 0.02 0.27 0.97
Received initial training -0.48 O.SI 0.3S 0.62
Perception of organizational autonomy 0.01 0.28 0.98 1.01
Perception of organizational informality 0.S4 0.38 O.1S 1.72
a Social capital is defined as the range or the number of national subsidiaries of the fmn in which
the respondent has contacts. For this analysis, a manager may have 'many ties' (contacts in 11-17
subsidiaries) or 'some ties' (contacts in 1-10 subsidiaries).
b Goodness of fit:
Chi-Square df Significance
Model Chi-Square 50.71 2S 0.00
Goodness of Fit 171.00 lSI 0.13
orientation (Ritti)5 as well as contacts in their home country. Research has also
shown that expatriates often serve as gatekeepers to the outside world in many
subsidiaries, which also partially explains the results we observe (Edstrom and
Galbraith 1977).
Finally, although participation in teams was a statistically significant variable in
the overall model, the size of the effect is negligible; the odds of being in the high-
range group versus the low-range group remain largely unchanged as participation
increases.
Analyses of Variance Results

Having confirmed our initial hypotheses, we performed a second level of analysis to
explore further differences among respondents who fell into one of four groups: 1)
those who had a mentor neither early nor late; 2) those who had a mentor late but
not early; 3) those who had a mentor both early and late; and 4) those who had a
mentor early but not late. In this way, we examined if having a mentor both early
and late or neither early nor late had an equivalent effect on an individual's ability to
develop social capital. In addition, we strove to understand individual differences
between respondents that may have influenced the mentoring relationships they
developed.
We used one-way analyses of variance with a dichotomous dependent variable
to assess differences between the four groups. In accordance with our regression
analysis, the results (see Table 3) show that respondents in each of these groups
differed statistically significantly in their range of intersubsidiary contacts. More
specifically, the range of contacts for respondents having a mentor late but not early
(Group 2) was significantly greater than each of the other groups (i.e., having a
mentor early but not late (Group 4), having a mentor both early and late (Group 3),
and not having a mentor at all (Group I)). Interestingly, although not statistically
significantly so, having no mentors at all was better than having an early mentor, in
line with recent research that has suggested that those who take personal
responsibility for building their social capital do better than those who rely on others
to build contacts (Gabbay 1995, 1997). Also, having a mentor only early was clearly
the worst scenario because those who had mentors both early and late (Group 3)
appeared to have had somewhat higher range than those who had only an early
mentor. All of these results are consistent with our hypotheses: having a mentor
early creates a sidekick effect, whereas having a mentor later in one's organizational
career helps a manager develop social capital.
Beyond differences in their range of contacts, there were a few other noteworthy
differences among these groups. For instance, expatriates were more likely to have
formed a mentoring relationship later in their career than earlier. We speculate that
one reason might be that these individuals moved to a subsidiary late in their
careers-thus becoming expatriates-and, at that stage of their careers, sought out
mentors to help them perform better in the context of the broader multinational
network.
Another difference to note is that initial training can be a substitute for initial
mentoring. Those who received some formal training on joining the organization
were most likely not to have an early mentor and most likely to have one later on
(Group 2). In sharp contrast, 90% of those reporting having a mentor early but not
late (Group 4) did not receive initial training. This finding has interesting
implications for the design of career development programs. Rather than focusing on
forming early mentoring relations, as many career development programs are
increasingly attempting to do (Caruso 1992), firms may do better by providing
initial training as a way of socializing new employees into an organization and
letting individuals develop mentoring relationships on their own later in their
careers. Indeed, research has suggested that some of the greatest benefits of the
~
0-
VJ
§. ~ g- ; s. ~ Description of Variable Groups ;'

!l '0g'
5:_. ~E¥ -l9@ 8.e?..;' Different· ..S!:
~. (1) g n {:, §. !'"
n 101 (l.. C» • CI5
~ 8: 5' Mentor early No No Yes Yes ~
c:: "'I!:S ~
;~ s..... "s.[ Mentor late No Yes Yes No
::r _. '" !!! iii'
Ef ~ ~ ~. g Dependent Variable o
...,
~ ~ >-j
g g- Social capitala 9.73 12.69 8.57 7.19 (1,2) (2,3) (2,4) §
j!i Control Variables
(f,I U (f,I
g. g a. '<~
.....::: !Z el ",- a Employee of Matsushita 0.41 0.76 0.58 0.57 (1,2) fg,
g 61 g- 8~ Employee of Philips 0.27 0.11 0.15 0.19
~ g. § .€I.. Relative importance of subsidiary 3.72 3.86 3.64 3.52 <
~.
;. iii' <Ill a· ~ General management function 0.17 0.02 0.18 0.24 (2,3) §
_ ~ Sl g ~ Marketing function 0.29 0.25 0.18 0.10 ~
- ~
!.. "'" (1)
- if 0
3 j!i Manufacturing function 0.20 0.18 0.22 0.10 ~
0.
-; p;' .. ~ ~ Purchasing function 0.02 0.09 0.05 0.14
g. g- ~ § ':-' Research & Development function 0.07 0.05 0.05 0.05
e; ii" .., ~ Finance function 0.10 0.20 0.13 0.14
~
~. g~ ~ Legal lother administrative functions 0.07 0.07 0.12 0.14
r
'"
~
o g~.
_ 0 !!!
~
Tenure at current Firm 13.20 14.98 13.22 11.71
:- < i ~. Years spent at other employers 8.68 5.03 6.88 10.00
~ ~. ~ ~. Number of other functions worked in 2.27 1.60 2.80 1.95 (2,3)
:! if go ~ Number of other subsidiaries worked in 0.98 0.95 1.02 1.00
f
~ ~
(1)
.g C/.l
~~ iii'
~ ~ Speed of promotion to current position 0.23 0.22 0.23 0.29 '"
8' ~ ii' ci. Expatriate status 0.32 0.67 0.33 0.14 (1,2) (2,3) (2,4) 0..:
(1)
S- ~ ~ ~ Time spent in cross-subsidiary meetings 11.83 13.02 15.07 15.14 t::

n
~ 3 e- g Time spent in cross-subsidiary teams 5.54 8.45 9.58 7.05 ?I"
s· ~ ~ ;;;, Time spent in cross-subsidiary training 4.49 1.89 6.12 5.67 trl
! ;,; 8 ~ Received initial training 0.34 0.67 0.52 0.10 (1,2) (2,3) (3,4) ~
~
o § S· ~ Perception of organizational autonomy 3.42 3.81 3.51 3.50 (1,2)
.....
J1; ~ ::r Perception of organizational informality 2.46 2.35 2.14 2.28 (1,3) -.l
!l ::r U.
mentoring relationships are derived through informal means and pertain to informal
aspects of the company (Collin 1979). Alternatively, early training may serve as one
of many mechanisms through which individuals can begin to develop helping
relationships that, in the longer run, could blossom into mentoring relationships.
Future research is needed that investigates the relationship between formal
organizational initiatives, such as training programs, and the development of
valuable informal relationships, such as mentoring.
Finally, the results suggest that those who form no mentoring relationships
(Group 1) may be 'prisoners of their own perceptions.' These individuals scored
lowest on perceptions of both organizational autonomy and informality. Because
these people view the organization as being constrained, they may undervalue the
benefits of mentoring relationships that can serve as pathways to accessing the
informal side of the organization. This finding supports Turban and Dougherty's
(1994) view that we need to examine more closely how an individual's personality
and attitudes might affect the mentoring relationships he or she seeks and initiates
and the benefits derived from them.
CONCLUSIONS AND DISCUSSION

Our objective in this chapter was to better understand how mentoring relationships
actually work in organizational settings to affect specific career outcomes. We
pursued this goal by examining both when mentoring relationships do work and
when mentoring relationships do not work with respect to one career variable that
scholars have linked to career progress-social capital, or beneficial ties with others
in an organization. Our results suggest that the timing of having a mentor affects a
protege's ability to develop social capital over the course of his or her organizational
career. Specifically, having a mentor early in one's organizational career can
produce a sidekick effect in which individuals are overshadowed by their sponsors
such that their development of independent social capital is difficult. However, this
sidekick effect disappears once the protege has developed a reputation as an
individual contributor within the organization. At this later time, a mentor may be
quite helpful in facilitating the formation of new ties since the protege has already
developed some form of professional identity and reputation and, therefore, is not
viewed as walking in anyone's shadow.
In addition, two other findings from this research were interesting and consistent
across our analyses: expatriate status had a positive effect, and integrating
mechanisms (such as participation in team work), had a negative effect upon the
development of social capital. First, it appears that being an expatriate increases the
likelihood that a protege will have many ties across multinational subsidiary
boundaries. This seems plausible, given that an individual who has lived outside of
his or her home country may have a better understanding of the cultures, history,
language, and customs of other nations. Overall, expatriates may develop a more
cosmopolitan orientation (Ritti)5 that later serves to facilitate the development and
maintenance of international ties.
Second, integrating mechanisms, such as the amount of time spent engaged in
team work or participation in training activities seems to have a negative impact on
the ability to develop ties with others across subsidiaries. Although the specific type
of integrating mechanism varied somewhat across our analyses, this general finding
seems worthy of further investigation. Drawing on the work of Louis (1990), we feel
that these types of interactions may lead to densely clustered subnetworks which
impede one's ability or interest in fostering a broad range of connections with others
outside the subnetwork. As Louis' work suggests, those who develop strong 'buddy
relationships' are less apt to negotiate relationships beyond their local work
environments. Thus, peer group interaction can lead to fraternizing that impairs
newcomer acculturation into the organization. In contrast, 'isolates' are more likely
to gain a better understanding of their work environment simply because they have
been forced to rely on multiple and varied sources of information. Thus, while
interaction with others may provide social support, it does not necessarily lead to a
broader network of contacts across intraorganizational boundaries.
Our study has several limitations that would be important to address in future
research. First, and most importantly, our work is based on a limited amount of data
regarding the nature of the mentoring relationships we studied. All we asked our
respondents was whether they had a mentoring relationship or not. Factors not
known to us-such as the identity of the mentor, the mentor's structural position in
the firm, the process by which the mentor was chosen, and the nature and strength of
the mentoring relationship--should all be studied to see whether and how they
mediate our main findings. For example, we did not know whether the early mentor
was also the protege's direct boss, and so could not compare supervisory and
nonsupervisory mentoring relationships-a topic Scandura and Schriesheim (1994)
have suggested for future research. Indeed, as Burt's (1992) work showed,
hierarchical networks are most effective when they revolve around sponsors who are
distant from the individual's immediate work group--that is, when the sponsor is
not the manager's direct boss. In short, we expect that some mentors may be better
than others at preventing their proteges from being seen as sidekicks.
Second, we studied just one issue of importance to an individual's career-the
formation of social capital. Although social capital plays an important role in
shaping such career outcomes as advancement, mobility, and performance, the
amount of human capital (educational qualifications and skills) and physical capital
(money) possessed by an individual can play an equally important role. Thus, even
if we find that early mentoring relationships have a negative effect on the
development of an individual's social capital, we cannot conclude that such
relationships necessarily harm the individual's career because they may have a
countervailing positive effect on the development of the individual's human capital.
Third, although we attempted to control for several factors that might influence
the range of intersubsidiary contacts possessed by our managers, we recognize that
there are other factors for which we did not control. For instance, recent research
suggests that an individual's personality (on which we collected no data) can
influence the extent to which they seek out and make use of mentors (Turban and
Dougherty 1994). It has also been argued that gender and race can playa significant
role in mentoring relationships (Ibarra 1993c; Thomas 1993). We were not able to
assess gender effects in this study since all of our respondents were male; neither did
we have data on race and, hence, could not control for such effects.
Finally, the group of managers we studied are only a subset of the population of
all managers. They all worked in subsidiaries of foreign multinationals and were all
male. Thus, our findings have limited generalizability. It would be useful to explore
these issues in other organizational settings, especially in U.S. firms, and with
managers of more diverse demographic characteristics.
Despite several limitations, this work has many important implications for those
conducting research on mentoring relationships and on helping relationships more
generally. First, by examining the impact of mentoring relationships on a protege's
stock of social capital, we began to unravel possible explanations of how mentoring
functions actually work to affect proteges' career outcomes. We proposed that there
are several roles that mentors can take in providing career help to proteges; mentors
may serve as signaling agents, providers of access, and providers of opportunities.
Although we expect that such roles may lead to near term benefits for the protege,
our findings suggest that there are also long term costs associated with having such
relationships. Specifically, having a mentor early in one's career can produce a
sidekick effect such that the development of ties with others in the organization
becomes difficult.
Second, our findings and the rationale that supports them challenge assumptions
in the mentoring literature. For instance, much mentoring work asserts that different
mentoring functions work together to influence career outcomes, and that the more
functions provided by a mentor, the more beneficial the mentoring relationship
(Kram 1985; Noe 1988a). The research presented here, however, points to a
contingency model in which we consider how different types of mentoring
assistance provided at different times during a protege's career can affect career
outcomes. This raises important questions regarding the appropriate sequence and
ordering of specific career mentoring functions over the span of a protege's career.
For example, is it beneficial for a mentor to provide access before the protege has
established confidence and competence in his or her ability to excel in an
organization? On a related note, this research calls for future work to examine
whether there is a trade-off between which sets of mentoring functions best serve the
protege early versus later in his or her organizational career. For example, is
psychosocial help early and career help later the best combination of help for a
protege who is trying to advance in an organization?
Third, this work suggests a host of opportunities for future research on both the
timing and types of interventions designed to help in general (Higgins 1997a). For
example, since our questions tapped whether or not a respondent had a mentor at
different points in time, regardless of whether this was the same person doing the
mentoring, our research opens up the possibility for future work on the optimal mix
of help-providers over the course of one's career. Stretching this notion of timing
even further, one could also consider how a protege's career is affected by the mix
of mentors he or she has across the stages of adult development (Levinson et al.
1978) or during the transitions between them (Daloz 1986; Kegan 1982). Thus,
rather than explore the timing of different types of help that are given by one help-
provider over the course of one mentoring relationship, researchers might also
consider the combination of different types of help-providers an individual has over
the course of his or her career and life.
In short, we need to consider not just knowing how mentoring relationships

work, but also what effects the who and when of such relationships have on a
protege's ability to progress in his or her career (Higgins and Thomas 1997). This
approach would capture the inherently more contingent nature of careers and
helping relationships. Most importantly, it would allow our thinking to move beyond
the 'myth of the perfect mentor' (Hill and Kamprath 1991) to be more in line with
recent research that has suggested that career developmental relationships are not
monodyadic, hierarchical, and intraorganizational, but rather exist in constellations
(Kram 1985) or portfolios composed of both superiors and subordinates and can
exist both inside and outside the firm (Higgins 1997b; Thomas and Higgins 1996).
Indeed, the sidekick effect itself is likely to be affected under such a perspective, for
a person may be less likely to be overshadowed if he or she has several people from
whom to draw career support rather than just one prominent superior.
In sum, the sidekick effect highlights the potential negative implications of early
mentoring relationships with respect to one important career factor, the development
of social capital, within a literature that has characterized such relationships as
primarily positive. From a practical perspective, when initiating informal mentoring
relationships, this work suggests that it is important to consider the possibility of a
sidekick effect which could hinder a protege's ability to develop ties with others.
And, from a theoretical perspective, when trying to understand how mentoring
relationships actually work in organizational settings, this research suggests that it is
important to recognize the inherently contingent and complex nature of giving and
receiving help--in particular, to consider how the effectiveness of the type and
timing of mentoring relationships can affect individual career outcomes.
We would like to thank Chris Bartlett and Sumantra Ghoshal for providing us with the data used in this
study and Jack Gabarro, Richard Hackman, Linda Hill, Herminia Ibarra, David Thomas, and Douwe
Yntema for their helpful comments on this research.
NOTES
1. We use the term 'organizational career,' rather than 'career,' because this research focuses on the
impact of mentoring on an individual's career in one organization, as opposed to an individual's career in
more than one organization.
2. 'Early' is defined as within the first two years of joining an organization. 'Later' or 'long term' is
defined as anytime thereafter.
3. Statements on statistical significance refer to the .05 level.
4. We reached this conclusion after examining the frequencies table of the dependent variable and after
reviewing the plot of residuals versus predicted values, which showed two parallel lines of scattered
points. Note that this procedure is not equivalent to simply dichotomizing the range variable using a
median split. Instead, the logistic regression analysis is conducted because the residuals violate the
random distribution assumption in ordinary least squares models.
5. As referenced by W. Humphrey (1987).
Social Capital in Internal
10
Staffing Practices
•
Peter V. Marsden
Elizabeth H. Gorman
ABSTRACT
This chapter examines the information sources that U.S. employers use in the course
of internal staffing, that is, when promoting or transferring employees. We focus on
the use of methods involving informal ties: referrals and direct approaches to
candidates for promotion or transfer. Such ties may produce 'social capital' by
providing employers with information about the qualifications and abilities of
personnel; at the same time, they provide employees with information about
opportunities for mobility within the workplace or firm. Data from a representative
sample of work establishments indicate that informal methods are widely used in
filling vacancies with internal candidates, often in combination with more
formalized procedures such as job posting and seniority systems. Differences in
internal recruitment procedures across types of employers and jobs suggest that they
are selected in light of both efficiency concerns and pressures for equity and
procedural rationality in the treatment of employees.
INTRODUCTION
Organizational employment policies and practices playa central role in structuring
the inequality of rewards among individuals (Baron 1984). Careers consist of
sequences of positions, which may involve moves-promotions, lateral transfers, or
even demotions-within as well as between organizations. Promotion practices, in
particular, shape upward career mobility for individuals (Baker, Gibbs, and
Holmstrom 1994; Rosenbaum 1979a, 1979b; Stewman 1986; Stewman and Konda
1983) and influence earnings (Cappelli and Cascio 1991; KaUeberg and Lincoln
1988; Le Grand, Szulkin, and Tahlin 1994). One strand of research on
Social Capital in Internal Staffing Practices - 181
organizational promotion practices has examined structural, institutional, and market

correlates of the prevalence of job ladders and other structured mobility channels
within organizations (Baron, Davis-Blake, and Bielby 1986; Bridges and Villemez
1991; Kalleberg, Marsden, Knoke, and Spaeth 1996; Pfeffer and Cohen 1984).
Other studies have explored the shape of formal job ladders and their interplay with
informal career trajectories (DiPrete 1987; Miner and Estler 1985; Kanter 1983b).
Still others have examined the criteria applied when rewarding individuals with
promotions (Baker, Gibbs, and Holmstrom 1994; Bills 1988; Kanter 1977;
Nicholson 1993).
In light of the centrality of intraorganizational mobility to careers, it is
surprising that researchers have given little attention to the processes through which
it occurs, that is, the manner in which organizations identify and select candidates
for promotion. Both formal and informal methods exist for accomplishing these
tasks. Seniority systems and job posting are the principal formal procedures. They
identify and select from candidate pools using abstract, universalistic criteria. As
such, they can be viewed as part of a larger process of bureaucratization within the
employment relationship (Baron, Dobbin, and Jennings 1986; Baron, Jennings, and
Dobbin 1988; Bridges and Villemez 1991), along with other features of internal
labor markets. Indeed, the dearth of research on internal recruitment procedures may
reflect an assumption on the part of researchers that promotion systems necessarily
rely on formal selection procedures.
Yet even relatively bureaucratic organizations can and do make use of informal
channels when identifying and selecting candidates for promotion (Pinfield 1995).
Such methods include referrals as well as direct contacts between the selecting
official(s) and candidates for promotion and transfer; they limit the pool of
candidates to those with direct or indirect social ties to the selecting official(s).
Studies examining individual differences in mobility within organizations have
shown that particular configurations of social ties are associated with promotions
(Burt 1992; Podolny and Baron 1997), and it is therefore of interest to examine them
from the organization'S side.
In related research on methods of hiring from external markets, researchers have
demonstrated that organizations frequently rely on recruitment and selection
procedures that make use of candidates' social ties (Fernandez and Weinberg 1997;
Fevre 1989; Jenkins, Bryman, Ford, Keil, and Beardsworth 1983; Kalleberg et al.
1996: chapter 7; Marsden and Campbell 1990). A cognate body of work has
demonstrated that the form and content of social ties is linked to individual success
in finding jobs (e.g., Boxman, De Graaf, and Flap 1991; Flap and Boxman, this
volume; Granovetter 1995; Lin, Ensel, and Vaughn 1981; Marsden and Hurlbert
1988; Wegener 1991). This chapter focuses on similar phenomena in the
intraorganizational setting.
In the next section, we discuss the manner in which we view organizational
information channels as avenues for the operation of social capital. We then describe
the various methods that organizations use in disseminating information about
opportunities for promotion and transfer. Our theoretical account of variation in the
use of informal methods is based on efficiency and strategic concerns, together with
external pressures for equity in the allocation of opportunities. After introducing the
data base for our analysis, the 1991 National Organizations Study (NOS), we present
findings, and conclude by discussing their implications for organizations and
employees.
INTERNAL STAFFING PROCEDURES AND SOCIAL CAPITAL

We view recruitment procedures that draw on preexisting social ties as settings in
which the operation of social capital within organizations may be observed. Because
the concept of social capital has been used in diverse ways, however, as illustrated
by discussions in Foley and Edwards (1997), Greeley (1997), and Newton (1997),
analyses invoking it must be explicit about its denotation. Most or all variants on the
concept posit that social ties may serve as resources for individual or collective
action; we take this as the essential core of social capital, as distinct from related or
additional features such as norms of trust that may be associated with the presence
of social ties. The ties in question need not be channels deliberately developed to
facilitate purposive action; indeed, much is made of the unplanned or adventitious
use of social ties originating outside of a given context. Thus, Granovetter (1985)
contends that economic transactions are often 'embedded' in pre-existing social
relations, while Coleman (1988) stresses the 'appropriability' of social
organization-i.e., its use for purposes other than those for which it was
established-as a fundamental characteristic of social capital.
Different conceptualizations of social capital highlight its role as either a
property of collectivities and communities or an individual resource. Both
conceptual strands are evident in Coleman's (1988) influential early statement. As a
collective property, social capital inheres in the web of social relations within a
collectivity, and the set of obligations, information channels, and norms that those
relations support and enforce (Coleman 1988). Like clean air or municipal policing,
social capital at this level is a collective good, one that can facilitate both collective
and individual actions (Coleman 1988; Putnam 1993b, 1995a). As an individual
resource, social capital resides within the direct and indirect social ties giving the
social location of a focal individual. Like human capital, social capital at this level
provides individuals with differential competitive advantages (Burt 1992).
In this study, we emphasize the second conceptual strand; we conceive of
informal channels within organizations as vessels for social capital that enable
employers to gather information of superior quality about candidates for promotion
or transfer. Individuals, of course, benefit from being in social locations permitting
access to such networks. As well, the staffing decisions reached by making use of
the social capital contained within informal channels may result in collective
benefits such as strengthened organizational performance, but we do not stress this
here. We direct our attention to the nature and determinants of organizational
staffing procedures that draw on individual social capital and thereby enhance its
influence on individual mobility outcomes.
INTERNAL RECRUITMENT AND SELECTION METHODS

The principal contrast we draw in studying internal staffing methods parallels the
formaVinformal distinction used in existing literature on job matching in external
markets. We begin by discussing informal methods involving referrals and direct

approaches to employees. We then mention two formal methods, seniority and job
posting systems, more briefly.
Informal recruitment involves information transfers in social networks, and may
take several distinct forms. Within a work unit, supervisors usually monitor the
performance of employees on an ongoing basis. Using this information, they may
directly approach employees they deem qualified when higher-level vacancies occur
or new positions are created. Information about employees who might be promoted
or transferred may also be shared through word-of-mouth referrals from other
supervisors, other employees, or even people outside the firm such as clients or
customers. Such channels contain social capital because they can convey relatively
rich, qualitative information about the capabilities of potential candidates for a new
position; reciprocally, they may also provide candidates with 'realistic job previews.'
A prototypical case in which one would observe informal staffing is that of
'drive systems' involving direct, personal control (Edwards 1979; Jacoby 1985).
Such systems delegate substantial discretion to supervisors who may be capricious,
even despotic. Pinfield's (1995) description of staffing as an 'administered process'
in his case study of a forestry firm involves important elements of informal
recruitment, though it occurred in a context where formal procedures also were
present. He observes that managers preferred the administered appointment process
because it took less time, and permitted them to 'manage lines of progression' for
their subordinates. With administered appointment processes, managers could use
the promise of promotion as a means of motivating lower-level employees within
their units; they were also able to reduce their risks because informal channels
yielded trustworthy information about the personal reputations of candidates.
Of course, as drive systems and personal control remind us, by using informal
channels to disseminate and gather labor market information, organizations provide
an opening for not just supervisory discretion and good jUdgment, but also prejudice
and favoritism, to enter promotion and transfer decisions. There may, then, be
liabilities as well as benefits to the use of social ties. This highlights the omnipresent
distinction between form and content in the study of social networks; it is
problematic to infer content from form. The fact that personal ties may be used not
only to transmit subtle, local 'impacted' information, but to exclude and/or unduly
advantage some candidates at the expense of others, raises concerns about equity in
the treatment of employees and calls for limitations on the discretion of supervisors.
Formal methods of internal recruitment have arisen in response to these
concerns. Seniority-the allocation of promotion opportunities on the basis of length
of service-represents one approach to limiting managerial caprice and thereby
increasing universalism within the promotion process. Kochan, Katz and McKersie
(1994) include seniority among the central elements of the Job control' industrial
relations system that was widespread in the U.S. in mid-century, particularly in
unionized settings. Freeman and Medoff (1984) identify several ways in which the
seniority principle may enter into promotion decisions, ranging from seniority as the
dominant factor in promotions (irrespective of ability) to seniority as a basis for
setting priorities among candidates whose qualifications are otherwise equivalent.
Job posting constitutes a distinct approach to formalizing procedures for

promotions and transfers within internal labor markets. A posting system publicizes
internal job opportunities through written notices on bulletin boards or employee
newsletters; see Pinfield (1995) for a generic description of such a system.
Announcements of vacancies include job descriptions and specifications of
qualifications; current employees are encouraged to put themselves forward as
candidates to fill vacancies. Posting systems introduce open competition into the
allocation of job opportunities to current employees.
Pinfield (1995) observes that posting systems provide a high degree of
procedural rationality and legitimacy to promotion and transfer decisions. Jacoby
(1985) notes that unions view posting as an alternative or complement to seniority as
a mode of developing job ladders and providing advancement opportunities for their
members. Additionally, many seeking policy levers with which to reduce sex
segregation in employment (e.g., Kanter 1977) advocate job posting as one approach
to increasing equality of opportunity within organizations by publicizing vacancies
beyond the reach of informal ties that are often confined within work units.
THEORETICAL CONSIDERATIONS IN THE SELECTION OF

INFORMAL STAFFING METHODS
Our concern in this chapter lies in the factors that make it more or less likely that an
organization will use informal channels in internal staffing decisions. Some factors
of interest refer to the organization per se, while others pertain to differences in the
nature of work done in positions within organizations. We argue in this section that
these factors operate through several distinct intervening mechanisms. From a
rationalist standpoint, some factors are linked to the benefits to be gained by using
informal channels or the costs (including opportunity costs) of doing so. Others
instead reflect institutional pressures toward equity or procedural rationality in the
management of staffing, which may emanate both from sources internal to an
organization and from the environment in which it is situated.
Benefits: Information QUality

From the standpoint of rationality and efficiency, one would expect organizations to
adopt staffing methods that rely on interpersonal connections under circumstances in
which the benefits of such methods are greater. The principal benefit of informal
methods drawing on social capital is that they may provide richer, more extensive,
and more situated information concerning candidates than can formal, impersonal
methods.
Informal methods involving social ties should be especially likely to yield
higher-quality information when interpersonal skills and the ability to form networks
of contacts are central to effective performance. Such skills are important in jobs
that involve frequent communication with clients or customers; for example,
investment banks rely on their employees' networks of external ties to bring in new
business and maintain client relationships (Eccles and Crane 1988). We therefore
expect to find greater use of informal internal staffing methods in professional, sales,
and service occupations.
Interpersonal skills and social networks are also important in flexible

organizations with staffing strategies that adapt goals and objectives to the
capabilities of their personnel, rather than seeking people who fit the requirements
of standardized jobs. In such organizations, staffing can be seen as equivalent to
strategy formation (Snow and Snell 1993). They make less use of formal job
descriptions and rules to clarify and coordinate tasks, and instead expect the
contours of jobs to be altered to suit the talents and initiative of their incumbents. In
such situations, networks of internal ties serve as a coordination mechanism (Eccles
and Crane 1988) and convey clear normative expectations associated with each
employee's role in the organization (Podolny and Baron 1997). Given the need for
rather nuanced, local knowledge of someone's capabilities in such systems, we
anticipate that organizations without formal job descriptions will tend to make
greater use of informal staffing procedures.
Informal selection methods should also provide better-quality information when
the skills involved in the position to be filled are difficult to measure objectively. If
the responsibilities of the position to be filled are nonroutine and involve the
exercise of discretion, objective standards are difficult to apply. Subjective
evaluations obtained through social networks may provide the best available
information concerning a candidate's potential (Pfeffer 1977). This argument
suggests that organizations will rely more heavily on informal selection methods for
managerial, professional, and skilled-craft positions. For managers, this claim is
supported by a study of supervisory selection in plants within eight manufacturing
industries (Northrup et al. 1978) which found that 'most procedure is highly
informal.'
Finally, informal selection methods should provide better-quality information in
circumstances where the costs of a selection error are relatively high. In such cases,
employers are more likely to be interested in obtaining the fine-grained information
concerning candidates' personal characteristics that only informal procedures can
provide. The costs of a selection error are higher when an employer invests in the
training of promoted employees, because such investments will be lost if the
subsequent performance of the employees is poor. Selection errors can also be costly
when promotions occur within multiple-level job ladders, because incumbents in
such positions are likely to have long tenures in the organization. Due process
guarantees may make it difficult to discharge them, and selection for a position on a
job ladder is likely to enhance an employee's eligibility for further promotions. We
therefore expect to find greater use of informal methods in occupations that receive
formal training and that have mUltiple levels (i.e., that form part of job ladders).
Administrative and Opportunity Costs

The intrinsic costs of administering informal procedures for identifying and
selecting promotion candidates are generally low. Because substantial informal
information is acquired passively in the course of undertaking other tasks, it can be
viewed as essentially costless. Even when invoked actively, the use of such methods
may involve nothing more than picking up the telephone or wandering down the hall
to discuss a candidate; of course, intensive pursuit of closely-held information
through informal ties may be costly, especially for confidential matters.
One reason for relying on relatively low-cost informal methods is that

alternatives are too expensive. Small, single-site organizations may not find it cost-
effective to administer formal staffing procedures, for example. Small
establishments that are part of multi-site organizational systems may, however, be
required to use centrally prescribed formal procedures for making promotion and
transfer decisions, and hence be less apt to use informal methods of doing so.
Because the range of candidates that can be identified using informal ties is
limited to those who can be tapped through the interpersonal networks of selecting
officials, a principal cost (or social liability) of informal methods is the foregone
opportunity of missing talented candidates who are not identified through the social
ties used for an internal personnel decision. In small, single-site establishments, this
opportunity cost should be low, because selecting officials are apt to have direct or
indirect social ties to most or all employees. In larger establishments and
establishments within multi-site organizations, on the other hand, the pools of
candidates that can be defined on the basis of social ties will usually consist of much
smaller fractions of the potentially eligible workforce, creating a greater opportunity
cost. Accordingly, we expect to find less use of informal procedures in large or
multiple-site establishments.
Constraints: Equity and Rationality Pressures

As noted above in the discussion of seniority systems and job posting, constituencies
both internal and external to organizations exert pressures on them to allocate
rewards, including promotion and transfer opportunities, in an equitable and
procedurally rational manner. Informal methods of identifying and selecting
promotion candidates are more vulnerable to charges of prejudice and favoritism
than are formal procedures that rely on objective, universalistic criteria. Informal
procedures are also less consistent with norms of rationality and bureaucratization,
which many see as institutionalized myths or societal values that infuse
organizations (Meyer and Rowan 1991; Bridges and Villemez 1991; Baron, Dobbin,
and Jennings 1986).
External pressures emanate from unions and government regulators. Differences
in personnel practices between unionized and nonunion workplaces have been
stressed by a number ofresearchers (e.g., Jacoby 1985; Baron, Dobbin, and Jennings
1986; Dobbin et al. 1988). Cohen and Pfeffer (1986), for example, argue that unions
advocate formalized staffing procedures because such procedures can prevent
employers from penalizing employees with pro-union attitudes. Institutional
arguments hold that exposure to the public sphere places organizations under special
pressure to conform to evolving norms about legitimate employment practices
(Dobbin et al. 1988). Larger establishments and establishments that are part of
multi-site organizations are more visible to regulators, and consequently such
establishments should be more reluctant to make use of informal promotion
procedures. Public sector establishments, in particular, must demonstrate high levels
of fairness, objectivity, and openness in their employment practices; they are also
subject to civil service laws and regulations that mandate the use of certain formal
procedures (DiPrete 1989; Tolbert and Zucker 1983). Thus public sector
establishments should be especially likely to avoid the selection of personnel

through informal methods.
Internally, personnel departments are likely to favor formal methods of
identifying and selecting promotion candidates. One reason is that personnel
professionals are especially aware of the constraints and sensitivities of external
constituencies (Jacoby 1985); a second is that they enhance their intraorganizational
power through the possession of specialized knowledge about the conduct of
personnel actions (Pfeffer and Cohen 1984). We therefore expect that establishments
having personnel departments will make less extensive use of informal staffing
procedures.
THE NATIONAL ORGANIZATIONS STUDY

The data presented in this chapter are drawn from the National Organizations Study
(NOS), conducted during 1991. Telephone interviews were completed with
informants for a multiplicity sample (see Parcel, Kaufman and Jolly 1991)
representative of U.S. work establishments. l When contacting establishments, NOS
interviewers were instructed to speak with 'the head of the personnel department or
the person responsible for hiring.' Overall, the NOS attempted to contact informants
for 1,067 establishments; it successfully conducted interviews with 688 of them, a
completion rate of 64.5 percent. 2 For additional details about field procedures used
in the NOS, see Spaeth and O'Rourke (1994) or Kalleberg et al. (1996: chapter 2).
To take into account possible between-occupation, within-establishment
variation, the design of the NOS interview schedule specified that several question
sequences, including that having to do with internal staffing, were to be repeated for
up to three different occupations in each establishment. One of these was the job
title of the employees 'most directly involved' with the main product or service
provided by the establishment; below we refer to this as the 'core' occupation. A
second was the occupation given by the General Social Survey (GSS) respondent
who had given the name of the establishment. Finally, questions were also posed
about 'managers or other administrators.' This multiple-occupation design permits
the separation of establishment- and occupation-level influences on several NOS
outcome variables.
MEASURING INTERNAL STAFFING METHODS

To measure the procedures used when filling a job internally, the NOS used this
question sequence:
When you fill this job with a person already in the organization, how often do you
1. Consult a seniority list?
2. Inform current employees by posting or circulating a vacancy notice?
3. Ask the person leaving the job to recommend other current employees?
4. Ask others at your workplace for recommendations?
5. Go directly to specific employees and encourage them to apply?
Table 1. 'Frequent' use of internal staffing methods by NOS establishments
Percentages of Establishments
Referrals Any
Seniority Job from Referrals Direct Informal
Occupation Lists Posting Incumbent from Others Approach Method
'Core' employees
Unweighted 38.5% 67.8% 8.1% 24.1% 19.4% 34.8%
Weighted 26.9 42.2 12.7 25.0 24.1 38.1
(N) (387) (388) (382) (386) (386) (388)
'GSS' employees
Unweighted 34.4% 64.7% 8.3% 16.9% 16.4% 29.0%
Weighted 39.0 45.3 12.8 22.6 25.6 34.0
(N) (221) (221) (218) (219) (219) (221)
Managers and
Administrators
Unweighted 18.4% 59.5% 12.4% 25.2% 28.4% 43.0%
Weighted 29.8 41.6 11.7 22.4 33.1 41.1
(N) (446) (449) (442) (445) (447) (449)
Note: Questions about internal staffing were asked only of informants who stated that current
employees were 'sometimes' promoted or transferred to fill vacancies in a given occupation.
Informants were asked to indicate whether they used each method 'frequently,'
'sometimes,' or 'never.' The sequence of questions was repeated for each of the three
occupations, if the informant stated, in response to a filter question, that the
establishment ever fills vacancies in that occupation with people it already employs.3
Table 1 presents the percentages of informants who answered 'frequently' to the
items in the sequence, separately for the three occupations studied in the NOS. It
shows that formal staffing procedures-seniority and job posting-are most widely
used. The un weighted percentages, which reflect the experience of the typical U.S.
employee (see note 1) show that between three-fifths and two-thirds of internal
vacancies are frequently advertised through posting of lists or circulation of vacancy
notices. The seniority principle is frequently used for almost 40 percent of vacancies
in core jobs, and even for 20 percent of managerial positions.
There is, nonetheless, appreciable use of informal methods in internal staffing.
Only about a tenth of positions are frequently filled by asking the previous occupant
of a position to recommend possible successors, but recruiting for nearly a quarter of
them often draws on other interpersonal referrals. Employers make frequent direct
approaches to candidates for promotion and transfer for about a fifth of core
positions, and for almost 30 percent of managerial jobs. The final column shows that
at least one of the three informal methods is frequently used for more than a third of
core jobs, and for 43 percent of manageriaVadministrative vacancies. Results not
displayed in Table 1 indicate that informal methods were 'never' used for less than
15 percent of jobs studied in the NOS.
The weighted results in Table 1 estimate the distribution of internal recruitment
methods across establishments (see note 1). Percentages for job posting are lower in
the weighted results, indicating that this procedure is less often used in the smaller
establishments which become more numerous when the data are weighted.
Conversely, weighted percentages generally exceed the unweighted ones for
seniority and the three informal methods, suggesting that these methods may be
more often used in smaller establishments.
Crosstabulations of the frequency with which establishments use formal and
informal methods of staffing (not shown) reveal negligible associations. The fact
that these associations are not negative indicates that formal and informal methods
do not constitute mutually exclusive approaches: informal channels are sometimes
used alone, and sometimes as a supplement to posting or seniority. Establishments
using one type of informal channel also tend to use others: there are substantial
positive associations linking all three pairs of informal methods in Table 1. 4
ORGANIZATIONAL AND OCCUPATIONAL CORRELATES OF

INFORMAL STAFFING PROCEDURES
We have argued above that staffing methods are chosen in light of their anticipated
information benefits and both administrative and opportunity costs, as well as
internal and external pressures for equity and rationalization in the management of
internal personnel actions. We suggested that as a result of the operation of these
mechanisms, we should observe variation in the use of informal methods across
establishments according to size, affiliation with a larger, multi-site organization,
sectoral location, and the presence of personnel departments, specific job
descriptions, and unions. At the occupational level, our three mechanisms predict
variation in the use of informality across types of occupations, and according to the
presence of training and job ladders. This section presents our findings; descriptions
of our measurements of these explanatory variables appear in the appendix.
Bivariate Differences in Internal Staffing Methods

Table 2 provides a preliminary indication of the relationships between our
organizational and occupational covariates and the use of informal methods of
identifying and selecting candidates for promotion or transfer. The first column
reports percentages of establishment-occupation observations S of a given type for
which informants reported that at least one informal method was used 'frequently.'
The other columns give similar percentages for the three specific informal methods.
The results in Table 2 are broadly consistent with the theoretical ideas
developed above. As establishment size increases, informal methods are less often
used to fill occupational positions; nearly half of observations from establishments
size 10-49 use at least one informal method frequently, but only about 30% of those
from establishments with more than 500 employees do so. About 44% of
observations from independent establishments make frequent use of an informal
method, as compared to a third of those from within multi-site organizations. These
differences reflect the greater visibility of larger and multiple-establishment firms,
as well as size differences in the information benefits, opportunity costs, and
administrative efficiency of staffing methods.
Table 2. Bivariate differences in internal staffing methods (all occupations)
Percentage Using Method Frequentlya

Any Incumbent Referrals Direct
Variable Informal Referrals from Others Contact Nb
Method
Size
1-9 38.9% 14.9% 22.7% 30.0% (87-90)
10-49 46.0 9.7 26.5 32.4 (247-250)
50-99 38.2 9.1 22.9 23.2 (142-144)
100-499 34.0 10.6 21.9 19.8 (255-259)
500+ 31.3 8.9 22.0 14.1 (302-307)
Multi-site organization
No 43 .6 9.5 20.9 27.1 (346-351)
Yes 33.8 10.2 27.3 20.4 (696-707)
Auspices
Private 40.4 12.2 24.3 26.6 (615-622)
Public 32.2 6.9 21.3 17.5 (334-342)
Nonprofit 33.0 6.5 21.3 14.9 (93-94)
Union presence
None 41.0 11.0 25.1 20.6 (520-531)
Some 33.3 9.1 21.2 24.8 (518-523)
Personnel department
No 41.5 9.9 25.1 28.8 (523-530)
Yes 33 .2 10.5 21.3 16.5 (497-506)
Job descriptions
No 43.2 13.2 22.6 34.5 (136-139)
Yes 36.1 9.5 23.1 20.8 (906-920)
Training
No 34.6 9.6 18.7 23.5 (239-243)
Yes 38.1 10.1 24.6 22.5 (791-803)
Multiple levels in
occupation
No 39.8 9.9 22.6 24.6 (322-327)
Yes 35.8 10.0 23.3 21.7 (720-731)
Occupation
Managerial 42.3 12.0 25.0 28.0 (518-525)
Professionalffechnical 34.7 9.2 26.7 15.2 (141-147)
Sales/Service 43.0 8.4 25.2 26.2 (107)
Administrative Support 27.6 10.5 18.6 14.0 (86-89)
Craft 43.3 20.0 26.7 30.0 (30)
Unskilled 22.2 3.1 13.7 13.0 (160-162)
a Based on unweighted NOS data.
b Observations are occupational groups within establishments. Owing to missing data, Ns for the
different methods vary slightly.
Informal methods are less often used frequently when an organization has a
personnel department, when most positions have written job descriptions, when an
establishment operates under public or nonprofit auspices rather than in the private,
for-profit sector, or when unions are present. These findings substantiate the
arguments we offered above about staffing strategy and constraints as factors
shaping the selection of staffing practices.
Turning to occupational characteristics, Table 2 shows slight tendencies to make
more use of informal procedures when those selected to fill a vacancy are to receive
formal training. Contrary to our expectations, informal methods are slightly less
likely to be used in occupations with multiple levels, which offer the prospect of
future promotion. As we anticipated, though, the use of informal methods differs
appreciably across kinds of work. Internal staffing decisions for managers,
professional and technical occupations, sales and service occupations, and craft
occupations make use of informal methods substantially more often than those for
clerical (administrative support) and semi- or unskilled occupations. This pattern
suggests that organizations rely on informal methods when filling autonomous,
structurally unique positions in which performance is not readily metered.
Multivariate Analyses
We use ordinal logistic regression to examine the partial effects of organizational
and occupational variables on the frequency with which informal staffing methods
are used. This model is the appropriate choice for the study of ordered dependent
variables like those studied here (Long 1997; Winship and Mare 1984). A positive
coefficient indicates that an increase in the value of the explanatory variable is
associated with a rise in the odds of using a given internal staffing method at all
rather than 'never,' or 'frequently' rather than less often. 6 Table 3 presents the results
of these analyses. 7
Beginning with organizational factors associated with the use of any informal
method of internal staffing, we find statistically significant8 tendencies for
establishments linked to multisite firms, those having personnel departments, and
those in the public rather than the private sector to be less likely to use informal
processes in internal staffing. The odds of using an informal method in multisite
establishments, net of other factors, are 0.69 times as large as in independent
establishments; in public sector establishments, they are 0.56 times as large as in the
private sector. Of particular interest is the finding that the bivariate negative
association of establishment size with informal methods in Table 2 becomes
statistically negligible when we adjust for the other independent variables included.
Thus, the size differences in staffing methods seen in Table 2 are attributable to the
fact that size is associated with factors that have statistically significant coefficients
in Table 3, including the presence of personnel departments, multi site affiliation,
and occupational mix. Contrary to our expectations, the presence of a union does not
seem to reduce the use of informal staffing procedures, net of the other factors we
have controlled. Written job descriptions are negatively associated with the use of
informality in promotions and transfers, but this association is statistically
nonsignificant.
Table 3. Organizational and occupational correlates of informal internal staffing: ordinal logit
coefficients a
Internal Staffing Method
Explanatory Any Informal Incumbent Referrals
Variables Method Referrals from Others Direct Contact
Organizational
characteristics
Estabishment size (log) -.002 (.047) -.014 (.048) -.007 (.046) -.002 (.047)
Multisite organization -.373** (.135) .069 (.138) -.190 (.133) -.511 ***
(.134)
Public secto{ -.573*** (.149) -.520*** (.153) -.421** (.146) -.621***(.150)
Nonprofit sectorC -.240 (.228) -.039 (.227) -.202 (.225) -.154 (.224)
Union presence -.030 (.107) .083 (.109) .038 (.104) .134 (.108)
Personnel department -.345* (.172) -.082 (.176) -.406* (.168) -.549** (.175)
Job descriptions -.213 (.199) -.201 (.204) .030 (.194) -.413* (.200)
Occupational
Characteristics
Formal training .500** (.158) .210 (.163) .476** (.154) .395* (.158)
Multiple levels -.260 (.142) .051 (.146) .217 (.140) -.138 (.141)
Managerial C 1.042*** (.188) 1.222*** (.206) .366* (.181) 1.015*** (.190)
Professionalffechnicalc .810*** (.241) 1.139*** (.257) .731** (.233) .584* (.238)
Sales/Servicec .850*** (.255) .529 (.273) .683** (.244) .835*** (.253)
Administrative SupportC .216 (.262) .946*** (.282) .052 (.258) .281 (.263)
CraftC .574 (.398) .914* (.412) .346 (.378) .486 (.403)
Threshold parametersb
First cutpoint -1.821 .822 -.370 -1.194
Second cutpoint .644 3.082 1.667 1.230
Log-likelihood -966.89 -918.59 -1046.91 -987.48

N 1014 998 1006 1010
Source: 1991 National Organizations Study. Standard errors given in parentheses.
* p<.05 ** p<.Ol ***p<.OOl
aBased on unweighted NOS data. Observations are occupational groups within establishments.
Owing to missing data, Ns for the different methods vary slightly.
bConstant term is constrained to zero.
cReference category for auspice dummy variables is the private for-profit sector; reference
category for occupation dummy variables is unskilled occupations.
The odds of using an informal staffing method grow by a factor of more than
1.6 when those in an occupation receive formal training, suggesting that employers
seek more detailed information when considering internal candidates for such jobs.
The presence of job ladders (as measured by multiple levels), on the other hand, has
no statistically significant association with informality in internal staffing.
Consistent with our hypotheses, informal procedures are used much more often for
some types of occupations than for others. The findings show that the use of any
informal channel is most common for managerial employees; however, such
methods are also more common for professionaVtechnical and sales/service

occupations than for semi- or unskilled ones. As anticipated, the coefficient
contrasting craft and unskilled occupations is positive, but it does not reach
conventional levels of statistical significance.
Several differences in the effects of certain organizational variables across
specific informal methods of identifying and selecting internal candidates are worthy
of note. Differences are more strongly patterned for direct approaches than for either
type of referral. Affiliation with a multi-site organization primarily reduces the
frequency with which direct approaches to candidates are used, as does the presence
of formal job descriptions. In contrast, location in the public sector has a statistically
significant negative coefficient for all three informal methods. The presence of a
personnel department is negatively associated with relying on referrals from
employees other than the position's current incumbent, and with making direct
approaches to candidates; but having a personnel office does not reduce the
frequency with which employees leaving a job are asked to recommend potential
successors.
DISCUSSION
This chapter examined the organizational side of promotion and transfer events, and
the extent to which they involved individual-level social capital, in the form of
interpersonal networks linking selecting officials to candidates. Using data from a
national study of U.S. employers, we found that informal methods of recruitment
and selection-direct approaches to candidates as well as referrals-play a
substantial part in internal staffing actions; they were 'frequently' involved in
promotions and transfers into more than a third of the occupations studied, and
'never' involved in only 15 percent of them. While formal procedures for allocation
of internal vacancies, particularly job posting, were even more widespread, this
nonetheless indicates that many U.S. employers find it advantageous to rely on
interpersonal as well as impersonal sources of information when making internal
staffing decisions.
Our multivariate analyses demonstrated that variation in the use of informal
methods for internal staffing is patterned across types of organizations and types of
work in a fashion to be anticipated, given the differential information benefits,
administrative expenses and opportunity costs we associate with such methods in
distinct situations. Particularly notable are the between-occupation differences,
which suggest that employers seek to activate the social capital embodied within
informal ties to obtain information about candidates for positions involving training,
autonomy, and interpersonal skills, in which performance cannot be readily
assessed. We found it interesting, however, that scale per se was not linked to
informality in staffing, given our covariates.
Our data also indicate that an employer's use of social capital to inform
promotion and transfer decisions is strongly constrained by forces both internal and
external to an organization. The presence of a personnel department substantially
reduces reliance on informal ties in internal labor markets; externally, location in the
public sector has a similar effect. These findings are consistent with institutional
theorizing (e.g., Dobbin et al. 1988) which stresses the maintenance of
organizational legitimacy as an important priority for organizations. Here,

legitimacy is presumably enhanced through the use of staffing methods that assure
equity and procedural rationality. We found it somewhat curious, therefore, that the
presence of unions was not similarly associated with a decline in the use of informal
networks in internal staffing decisions.
In closing, we offer two further observations about the part played by social
networks in promotion and transfer processes. The first is that there is some reason
to think that networks are even more involved in internal staffing processes than our
data indicate. We have examined the procedures that establishments use to
disseminate knowledge about vacancies and assemble information about candidates
for intraorganizational mobility. Candidates need not necessarily hear about
opportunities via these channels; even formally distributed information quickly
enters and diffuses through workplace social networks. Suggestive evidence on this
point comes from the 1991 ass, which provided the sampling frame for the NOS.
ass respondents who had held more than one job with their current employers were
asked about how they heard about their present jobs. The most common answers
cited interpersonal contacts: more than half (51 %) of the respondents said that they
had been approached and asked to take their new jobs; nearly a third (31 %) learned
of their new jobs through a supervisor, while 15% referred to a coworker. Less than
a fifth heard about their jobs by virtue of job posting or a vacancy list (17%) or as a
result of unions or seniority (16%).9
Secondly, theoretical understandings of social capital stress the superior
capacity of interpersonal ties to convey certain types of information. Taking this as a
premise, we have invoked the usual 'norms of rationality' (Thompson 1967) to make
predictions about which employers will make use of particular staffing methods. We
enter the caveat, however, that informal methods may be used in ways that diverge
from received theory. Indeed, this may occur for formal methods as well: Reskin
and Hartmann (1986) review cases in which bidding restrictions or seniority
provisions undermined efforts to broaden gender equality in access to desirable jobs
through job posting practices. It seems especially likely, however, that informal
methods of staffing might be used for reasons other than their superior information
benefits. Though as social capital informal ties have the capacity to convey subtle
and nuanced information about the capabilities of candidates, as social liabilities
they simultaneously can convey other, less meritocratic information-about, say,
interpersonal loyalty rather than technical ability. By using interpersonal channels-
that tend to link socially similar persons-when diffusing information about
promotion and transfer opportunities and assembling information about candidates
for them, employers run the risk of excluding those lacking such social capital from
consideration. This at once raises equity concerns (from the individual's side) and
opportunity cost worries (from the employer's).
We think that organizations face a continuing dilemma about how to balance the
information benefits and risks associated with the use of individual social capital in
personnel actions. To the extent that selection into jobs in contemporary
organizations requires information that is accessible only through interpersonal ties,
employers must also provide safeguards against supervisory malfeasance in using
networks, guaranteeing that this takes place in an atmosphere of procedural
rationality. An intriguing direction for future research would be to determine what

differences exist between employers relying on pure 'network staffing'-that is,
those that use informal methods alone-and those using networks as an adjunct to
more impersonal methods.
APPENDIX: MEASUREMENT OF EXPLANATORY VARIABLES

This appendix describes the measurements of independent variables that appear in
Tables 2 and 3. Descriptive statistics reported are for the set of 1091 organization-
occupation observations examined in this chapter, for which internal vacancies were
ever filled using current employees. More details on measures in the NOS appear in
Kalleberg et al. (1996).
Size. Natural logarithm of the number of full-time employees in the
establishment (mean, 4.98; standard deviation, 1.98). Table 2 uses actual size.
Multi-Site Organization. Dummy variable identifying the 67% of observations
from establishments that are part of larger, multiple-establishment organizations.
Personnel Department. Dummy variable identifying the 52% of observations
from establishments that have a separate department or section responsible for
personnel and/or labor relations.
Job Descriptions. Dummy variable identifying the 87% of observations from
establishments in which there are written job descriptions for most jobs.
Union Presence. Scale combining four items indicative of the presence of
unions (mean, 1.50; standard deviation, 0.64).
Public Sector. Dummy variable identifying the 32% of observations from
establishments operated by federal, state, or local governments.
Nonprofit Sector. Dummy variable identifying the 9% of observations from
private, not-for-profit establishments.
Training. Dummy variable identifying the 76% of observations in which those
in an occupation had received formal training within the past two years.
Multiple Levels. Dummy variable identifying the 69% of observations in which
an occupation has more than one level.
Occupational Categories. Core and ass occupations were classified into three-
digit 1980 Census codes by the NOS. We subsequently grouped them into the six
broader classes used in this chapter. No specific occupational title was used by the
NOS when asking about 'managers and administrators,' so we classified all
observations for these occupations into the 'managerial' group. Overall, 50% of our
observations are managerial, 14% are professional, 10% are in sales or service
occupations, 8% are in administrative support occupations, 3% are in craft
occupations, and 15% are in semi- or unskilled occupations.
Data collection and writing were supported by National Science Foundation awards SES-89-11696 and
SBR-95-l17l5. For helpful comments, we are indebted to the editors, Roger Th.AJ. Leenders and Shaul
M. Gabbay.
NOTES
I. A work. establishment refers to a specific geographic site or address. Some establishments are part of
larger, multi-site firms or organizations. The sample of establishments was drawn as part of a topical
module on 'Organizations and Work.' included in the 1991 General Social Survey (GSS; see Davis and
Smith, 1996). In 1991, the GSS interviewed a random sample of 1,531 English-speaking U.S. adults. At
the end of the interview, each employed respondent was asked to give the name, address, and telephone
number of herlhis workplace; married respondents were asked to provide the same information about the
workplaces of their employed spouses. This generated a multiplicity sample in which work
establishments have known, but unequal, probabilities of inclusion; the probability that an establishment
is included in the NOS is proportional to its number of employees. Thus, there are more large
establishments in the NOS than would appear if workplaces were to be drawn at random from some
listing of establishments. The unweighted NOS sample describes work settings from the standpoint of a
typical U.S. employee, since it gives each GSS respondent equal weight. To instead describe the
population of U.S. work establishments, the data must be weighted inversely to workplace size. Most
figures presented here are for the unweighted sample.
2. Owing largely to the clustering entailed in the area probability design of the GSS, some
establishments were sampled more than once. The data reported in this chapter include only one record
for such duplicated cases. Including duplicates, there were 1,127 interview attempts and 727 completions.
3. It did not seem sensible to ask organizational informants about practices that they do not enact. The
fact that questions about internal staffing practices were posed only for occupations that were ever staffed
internally raises the prospect of sample selectivity. The selection criterion-internal filling of a position-is
central to definitions of internal labor markets, and we observe that many of the covariates in our
empirical analyses below (especially size, formalization, and affiliation with multi-site firms) are also
strongly linked to the presence of ILMs (Kalleberg et aI. 1996). Net of the effects of these common
covariates (especially establishment size) on selection into our sample and on staffing methods, we
believe that any association between a propensity to fill vacancies with internal candidates and the use of
staffing procedures is weak, and therefore that any residual selection bias is slight.
4. For incumbent and other references, Goodman and KruskaI's gamma is 0.62; for both types of
references and direct approaches it is 0.51.
5. That is, there may be up to three records for each establishment in this Table-<>ne for each
occupation (core, GSS, or managerial) for which the establishment's informant answered the sequence of
questions about internal staffing.
6. The coefficients can be understood in multiple ways. For an ordinal dependent variable with J
categories, ordinal logistic regression models the natural logarithm of the odds that an outcome will be in
category j or lower, rather than above j, as '1:J - pX, where '1:j is a 'threshold parameter' or 'cutpoint' and X is
a set of independent variables. The log-odds of being in a category above j, rather than j or lower, are
therefore -'1:J - pX. The proportional change in the odds of being in a category above j, rather than in
category j or a lower one, that is associated with a one-unit increase in a given explanatory variable Xi is
the same at any value of j and is equal to expC/3;). Thus, for example, the coefficient of -.573 for public
auspices and the use of any informal method indicates that for public-sector establishments, the odds of
using any informal method 'frequently' rather than 'sometimes' or 'never' are exp(-.573)=O.56 times as
large as they are for private, for-profit establishments. The same proportional factor applies to the odds of
using at least one informal method 'frequently' or 'sometimes' rather than 'never.'
7. Because of the clustering of observations on occupations within establishments, we estimated
multilevel models including between-establishment random effects on the likelihood of using a given
staffing method. The results of these analyses indicated that, net of our covariates, only small and
statistically insignificant components of variance are associated with organizational differences.
Accordingly, we elected to present findings using a simpler and more easily interpretable single-level
approach.
9. The percentages reported here total more than 100% because respondents were permitted to mention
more than one source of information when responding to the question. Case bases (Ns) for the
percentages range from 238 to 251.
Getting a Job as a Manager
•
11
HenkFlap
Ed Boxman
ABSTRACT
In his seminal study, Granovetter (1995) demonstrated how the job-attainment process
is embedded within social networks. The ensuing research effort and theoretical
discussion left two points unclear. To what extent do people with higher societal
position use informal channels to find a job? Do they receive positive returns in terms
of income? A replication of Granovetter's analysis in a large sample of Dutch
managers at larger companies (n=1402) in 1986-1987 shows that Dutch managers
generally rely on their social contacts to find a job, and they do so more frequently at
higher executive levels. Moreover, using informal job-finding methods leaves them
with higher earnings. Granovetter' s weak-ties argument has been refuted: although
they are the most widely used, finding a job through weak ties does not produce a
higher income level. It is not true that managers rely more on informal contacts later in
their career.
Our own more general hypotheses on social capital have been confirmed.
Managers with more social capital (association memberships and external work
contacts) find a job more frequently through some informal channel. Moreover, they
earn a higher income independent of their human capital. Burt's hypothesis that social
ties enlarge the returns of human capital has been refuted. Human and social capital do
interact in that social contacts help workers to earn more income at any level of human
capital, but the returns of human capital decrease at higher volumes of social relations.
INTRODUCTION
Many members of the workforce in Western societies have found a job through some
kind of informal contact. In the 1980s roughly a third of the employed people in the
Netherlands did so, and in the 1990s as many as half did so (Moerbeek, Flap, and
Ultee 1997). Comparable figures for former West Germany and the U.S. in the 1980s
are 42 percent and 59 percent respectively (De Graaf and Flap 1988). According to a
recent study of the French labor market, 32 percent of all the employees used an
informal channel to find their job (Forse 1997), and around 1990 this was also true of
nearly half the employees in Spain (Requena 1991). These facts are at odds with the
official universal ideologies in these countries.
In contrast to the official ideology, popular opinion holds that if you want to get a
job, it helps to mobilize your networks. Networking also gets you a better job.
Furthermore, social relations are important, especially at the top of the societal ladder,
since the people there have business relations who bring them financial and technical
information that can be useful in their current job and also labor market information
that can be useful, inter alia, in learning about job-openings. Because of the often
influential positions of these contacts, they can be instrumental in helping those at the
top of the ladder get similar jobs. However, general social surveys in industrial
countries such as the U.S., Germany, France, Spain, and the Netherlands have
demonstrated again and again that it was precisely the people who are lower on the
ladder who found their jobs through informal channels. In addition, this research has
shown that the use of social contacts rarely brings a better job than the use of more
formal channels or applying directly.
A number of empirical studies of specific occupational groups do, however,
indicate that people with better jobs make extensive use of informal contacts. In his
seminal study Getting a Job, Granovetter (1995) noted that over 60 percent of the
people in higher positions (he studied a group of persons with technical, professional,
or managerial occupation in Newton, Massachusetts, a suburb of Boston) found their
jobs via someone they knew. That was also how they found the better jobs.
Preisendorfer and Voss (1988) described how almost three-quarters (74 percent) of
200 employees at universities and vocational colleges in the former West Germany
found their jobs through some informal channel, although networking did not help
them find a better job-for example a steady full-time job.
Findings like these suggest a U-shaped association between a person's social
status and the use he or she makes of informal contacts in finding a job. This type of
U-shaped association probably remains hidden in random sample surveys of the
general population because those surveys usually contain only a small number of
people in higher positions. Findings like these also make us wonder about the
circumstances under which it is profitable to mobilize contacts to get ajob.
A small research literature has recently emerged on the social networks of
managers, their determinants, and the returns of their networks while doing their job as
a manager. There is a special interest in the occupational group of managers, since
their actions are supposedly more decisive for the fate of enterprises and of the
economy at large. Managers are in the business of entreprendre, bringing people and
other production factors together (Burt 1992: 274}-that is, they are in the business of
networking. One can assume that if they are good at this job, their social networks will
capture not only individual social capital but also ftrm-Ievel social-capital-and that
the firm will benefit-for example, in terms of better survival chances, larger sales
volume, or greater profit. Moreover, these managers will probably also benefit as
Getting a Job as a Manager - 199
individuals and receive a promotion, a higher salary, or other compensation because of

the advantages they bring to their frrm.
This research concentrates more on the specific characteristics of managers'
networks and less on the use and possible social capital returns of these networks,
since data on actual use and outcomes for individual managers-and to an even greater
extent on frrms-are hard to come by. Our contribution adds to this research by
concentrating on how managers and would-be managers use their own social network
in the job-finding process and it deals with the returns of their networks during their
careers as managers.
Case studies with anecdotal material (Mintzberg 1973; Kotter 1982) have
demonstrated that managers usually have large work-related networks. Managers talk
with numerous people at their frrms to keep informed about upcoming business
opportunities as well as potential threats. It is also clear that managers execute their
agendas by mobilizing their social networks.
Other, more quantitative studies allow for more solid conclusions. l Ibarra (1992,
1997) has demonstrated that women often experience a sex-specific division within
their network at work: for friendship and emotional support they turn to their female
colleagues, and for work-related advice and other instrumental help they calIon their
male colleagues. This probably puts them at a disadvantage if vacancies arise at a
higher level, since they then have to compete with men who have more strong-that
is, more multiplex-ties to those men who decide about promotions and filling
vacancies. The result of these processes is that women who aspire to higher positions
within their organizations, come up against a glass ceiling they are unable to break.
Brass (1985a) has shown that for women, this sex-specific division is indeed not
conducive to being promoted into a managerial position.
Yet there do not seem to be large differences in the networks of managers at
different positions or of different sexes (Moore 1992; Burt 1996). Even the kind of
work contacts cited by managers as being the most important or most troublesome are
stable across kinds of managers. Managers as compared to ordinary employees,
however, do have larger and more open networks, with more work colleagues. They
also have stronger ties to their alters, but the alters are less strongly linked to each
other (Carroll and Teo 1996). The fact that there are clear differences between the
networks of managers and employees and no clear differences between the networks
of managers themselves suggests that managers (including female ones) are selected,
inter alia, for their network characteristics, making them more similar to their
colleagues.
Once they have a job as a manager, what are the returns of their network in their
job as a manager? Burt (1992) has demonstrated that male managers with a more
autonomous position within their networks are promoted earlier, whereas female
managers and males who have just entered the higher managerial ranks need a more
cohesive network or even a high-prestige sponsor if they are to succeed in getting
further promotions. In another study Burt (1996) described how these structural holes
(meaning that one's alters are disconnected from each other) contribute to higher
returns for executive managers at an investment banking division of a large American
financial organization. They receive larger bonus compensation payments since their
alters have no alternative for them, but they do have alternatives for their alters.
Boxman, De Graaf, and Rap (1991) have showed that a larger, more diverse
external network of Dutch managers at large companies makes for a higher income,
quite independent of the managers' human capital and number of subordinates. In a
replication, Meyerson (1994) claims to have found that strong ties bring higher
compensation for managers at Swedish public firms. Carroll and Teo (1996) have
discovered that a greater number of colleagues in a person's network does not generate
more income for managers, but it does so for other employees. They account for this
by arguing that all managers have a sizable number of colleagues in their core network
and that if they did not, they probably would not last long as a manager.
Apart from Granovetter's small study, the job-finding process of managers and the
role that is played in job finding by their social networks has not been studied
extensively. This state of affairs makes it interesting to inquire more deeply into the
job-finding process of managers and the instrumentality of social networks in their
getting a job. We test a number of hypotheses (mainly taken from Granovetter's study,
as it is the landmark study on getting a job) in a large data set of Dutch managers. We
describe the job-finding process of the managers and how frequently they found their
current job through informal relations and, more specifically, through weak ties. In
addition, we examine whether managers with more social contacts do indeed use these
contacts to get their job. For a number of job-outcomes we also analyze whether there
are positive returns to using or having a better network, i.e. greater social capital.
In the last section of this chapter, we discuss whether an individual's social
network will also provide the firm with social capital.
RESEARCH TRADITIONS ON THE RELATION BETWEEN LABOR

MARKETS AND SOCIAL NETWORKS
Within the sociological and economic research literature on labor markets, at least four
traditions have paid attention to the influence of social networks on labor-market
positions. The theoretical background of the research literature on the networks of
managers is similarly heterogeneous, since it has been influenced by human capital
theory, status-attainment theory, structural-organizational ideas, and social capital
theory.
The status-attainment research studies the extent to which an individual' s chances
of upward or downward mobility depend more on his or her achieved characteristics
than on ascribed features such as family background. In the course of research, the
awareness has grown that the chances of finding a job depend not only on social
origin, education, and work experience but also on access to the kind of labor market
information that is provided by social networks. The main findings of the general
surveys incorporating these ideas are that the status of the contact person contributes to
the occupational attainment independently of human capital and that part of the
original effect of human capital is in fact to be attributed really to the social resources
used in the job-finding process (Lin, Vaughn and EnseI1981).
The structuralist tradition assumes that there are barriers between different parts of
the labor market, including cleavages between networks of relevant people. To find a
job in more attractive segments, like an internal labor market, individual capacities are
often not enough. One also needs social contacts that feed into this segment.
According to this idea of social closure, outsiders are excluded if company recruitment
occurs along lines of preexisting social networks linked to the workplace-for

example, by way of employee referrals (Grieco 1987; Windolf and Wood 1988). In
addition to the social networks being instrumental as a channel of information,
institutional and organizational factors thus influence social networks at the work
place. Also see Marsden and Gorman (this volume) on internal hiring practices.
The neoclassic economic tradition of labormarket analysis has a number of
varieties-human capital theory, job-search theory, and transaction cost theory.
Human capital theory does not provide much information on the relation between
networks and labor market position. It states that human capital, as indicated by
education and work experience, constitutes the best explanation for attained
occupational position. Job-search theory does offer a suggestion (Stigler 1961, 1962;
for a review of job-search theory, see McKenna 1985; Devine and Kiefer 1991). It
assumes that the collection of information on labormarket opportunities entails certain
costs, like loss of time, money and other opportunities forgone. Various search
strategies have different costs. The people who are the highest achievers are those who
are able and willing to invest in the search for information. The job-search theory does
not devote much attention to the embeddedness of search processes by employees and
employers in the informal social networks of employees and employers (Granovetter
1985). Although the theory starts from the assumption that the use of personal contacts
lowers the search costs if people exchange their information with each other (pooling),
it does not specify the conditions under which the exchange will occur and with what
result. In reality, search processes do not emerge at random in markets of anonymous,
disconnected actors, but within the confines of preexisting networks. Interesting
enough, job-search theory provides a mechanism that could explain the effects of
social networks, but its weakest point is the lack of empirical testing.
A recent addition to this line of thought is transaction cost theory (Williamson
1994). It stresses that most of a [trm's costs are not production costs but transaction
costs. They include not only the ex ante costs of the search for attractive exchange
partners but also those of arriving at an agreement and the ex post costs of having to
enforce an agreement. Networks lower not only the search costs but also those of
contracting and enforcement (Granovetter 1985; Raub and Weesie 1990). Since they
engender mutual trust, denser networks will make the management of business and
labor relations less cumbersome, stabilize cooperation and indirectly improve
performance (see, however, Burt and Knez 1995).
The three traditions described above focus on the matching process between
people and positions. Neo-classic human capital theory, search cost theory and
sociological status-attainment research focus mainly on the role of individuals with
different characteristics. Structuralist theories look for the possible effects of industries
and organizations. Transaction cost ideas are only just beginning to influence
empirical research on the returns of work-related networks in the occupational career.
Network research has accepted suggestions from each of these three traditions in
research on networks and their returns in occupational attainment, and elements of
each can be recognized in studies on the networks of managers and their returns that
were discussed above. Insights from these three traditions come together in
Granovetter's Getting a Job. He explicitly links individual and positional factors. He is
interested in the latent structure of social networks and its influence on labormarket
behavior. Although according to Granovetter information is mainly a byproduct of

other social interactions that come without costs, he assumes that informal search
efforts decrease search costs. This is why an employee's and an employer's interests
are served if they search through informal channels, especially weak ties, since they
provide access to information that is more likely to be new (Granovetter 1973).
Granovetter's ideas have constituted an important source for the recent fourth
tradition focused on networks and labormarket chances-social capital theory
(Bourdieu 1981; Flap 1988; Burt 1992). According to this theory, social capital
produces a better life. People who have access to more people and to people who have
more resources (including contacts) and who are more prepared to lend a helping hand
are more successful in goal attainment, including in the labormarket. What is different,
though, is that Granovetter sees networks mainly as a by-product of other kinds of
interaction, while social capital theory suggests that networks are also the result of
people investing in each other. Provided that there is a common future, people' invest
in particular ties and types of networks to the degree that they are instrumental. In
short, social capital is the present value of future help. Social capital theory specifies
the strength of the weak-ties argument in that weak ties do give access to people with
better second-order resources, but those people usually are less prepared to help.
We now present Granovetter's major hypotheses on the relation between
networks, job search, and the attained position, as well as three additional hypotheses
that have been the product of social capital theory proper. We specify Granovetter's
hypotheses and test them consecutively for our large data set of Dutch managers.
GRANOVETTER'S HYPOTHESES AND THREE OTHER ONES

BASED ON SOCIAL CAPITAL THEORY
HI: Search through social networks decreases search costs.
People are more apt to find a job and to find a job with less effort if they search along
informal lines instead of formal ones, since it enables them to hear about vacancies
earlier. In addition to the advantage of timing-an advertisement informs potentially
everyone at the same time-informal contacts provide more in-depth information on
the particularities of the job in question and everything that goes with it, a kind of
information that cannot be found in advertisements. Employees can also make their
productivity and other capacities known through their social contacts to potential
employers and get an introduction via them. Contact persons feel an obligation not to
recommend people they know for bad jobs or refer poorly qualified people to
employers who are looking to fill a vacancy. Looking at the process from the other
side, employers can lower the costs of recruitment by searching informally for
candidates. They get more in-depth information on the true productivity of potential
candidates more quickly. This more intensive information prevents disappointment on
both sides. Moreover, it saves on the costs of control and training (see Grieco 1987).
H2: Better jobs in terms of income and prestige are found through social contacts.
People who enter the labor market want the best job they can get, and they will
mobilize their resources accordingly, including their social resources. Since social
contacts bring extensive and intensive information and possibly also other kinds of
support, better jobs are found using contacts.
H3: Better jobs are found through weak ties.

Weak ties-ties that are less intensive and less frequent, like those with acquaintances
and colleagues-lead to the best jobs because weak ties can serve as bridges between
different social circles (cf. Granovetter 1973). Information travels across these bridges
over horizontal and hierarchical cleavages between different occupational groups and
levels and leads sooner to people in higher social positions, who are better able to
provide the relevant information on a vacancy or help in some other sense.
H4: Employees at later stages of their careers make more use of their informal ties,
especially their weak ties, to attain jobs.
This hypothesis builds on the above three hypotheses by adding the auxiliary
assumption that people acquire contacts, especially weak ties, on the job and even
more so by changing jobs. Under this assumption, it is logical to expect informal
channels to playa greater role in the later stages of an employee's career (see also
Granovetter 1988: 193).
H5: Employees at higher occupational levels make more use of informal channels
in the job-finding process.
Another plausible auxiliary assumption is that people in higher social positions have
larger networks and networks containing relatively more weak ties. They consequently
make more use of informal ties, especially weaker ones, to find ajob.
H6: Employees with better networks make more use of their contacts in finding a
job.
Hypothesis 1 can be specified using the notion of social capital. If having a greater
number of contacts with alters who are higher placed is instrumental, then people will
employ these resources more frequently in finding a job.
H7: Employees with more social capital will have higher returns of their social
capital.
There is no reason to assume the advantages of social relationships stop after a person
has acquired a particular job. The basic idea is that informal social relations to relevant
others not only help get a job but also help while doing a job, and improve the returns
from this job (Han 1996; Flap, Snijders. and Van Winden 1996).2
H8: Employees with more social capital succeed in getting larger returns of their
human capital.
This hypothesis was formulated by Burt (1992), who argues that social capital
provides opportunities for applying one's human capital and thus promotes the returns
of human capital.
DATA AND MEASUREMENTS

Our data set on Dutch managers is unique. Generally, data in the public domain on the
recruitment of managers and their ensuing careers are scanty. To our knowledge there
is no other sample of this size on the labormarket behavior of managers at many
different firms that includes multiple indicators of human and social capital.
The actual research, a survey, was conducted in 1986-1987. A questionnaire was

sent to 4000 large companies with fifty or more employees. These 4000 companies
were randomly selected from the total population of 8746 large Dutch companies.
Packets were sent to the companies, each containing one questionnaire. Above the
company's address we printed the job title of a particular manager, for instance: To
the Managing Director of General Motors Netherlands.' We included job titles in five
categories of the most important decision-makers at Dutch companies: 1) managing
director, 2) personnel manager (human resources), 3) commercial manager (sales
manager or purchasing manager), 4) manager production/automation, and 5) financial
manager. For each category, we mailed 800 letters. The letter of introduction requested
that, if the company did not have the job mentioned, the questionnaire be completed
by the personnel manager or the commercial manager. Since at some firms, someone
in a middle-management position answered the questionnaire, these managers were
classified as an additional category. The response rate was 35.2 percent (n=1402). Due
to missing information for some respondents on specific variables, the number of res-
pondents used for the analysis in this chapter is 1369.
The nonresponse is rather high. Although we expected managers of larger
companies to be overrepresented in our sample because--e.g., smaller companies do
not always have a personnel manager-we cannot exactly determine whether this is
the case because we have information only on the size of the establishment our
respondent works at. However, based on information on the whole sample of
companies with more than fiftly employees, we know that 65 percent of them work at
companies with more than 100 employees; 55 percent work at establishments with 100
or more employees. This was to be expected, since it is possible to have large
companies with small establishments but not the other way round.
The key concepts in our theoretical arguments were operationally defined as
follows:
Job-Finding Method. The channel through which the current job was attained is
classified in two categories: 1) informal contacts (via relatives, coworkers, employers,
acquaintances, or somebody working at a headhunter's office), and 2) formal channels
(through advertisements and employment agencies). Of the three categories usually
used in labormarket research-formal, informal, and direct application-the third was
unfortunately not included. Rightly or wrongly, the reason for this omission was our
fear of inviting socially desirable answers on the part of the respondent. Saying you
found your job by yourself makes a much better impression than admitting that
somebody helped you. Formal and informal means of finding a job together account
for seven subcategories: 1) employment agency, 2) newspaper advertisement, 3)
information from a relative, 4) information from an acquaintance, 5) information from
a work contact, 6) being asked by the employer, and 7) being asked by someone from
a headhunter's office.
Strength of Tie with a Contact Person. The strength of tie categories were: 1)
strong tie (contact with a relative) and 2) weak tie (contact with a coworker,
acquaintance, employer, or a headhunters). Regrettably, the questionnaire did not
contain the category 'information from a friend,' so we had to work with the
assumption that every relationship with a nonrelative is weak. Although this is
unfortunate, the hypotheses on the differential effects of strong and weak ties can still
be tested since family is generally considered to be the strongest of all the existing
types of ties.
Human Capital. Human capital was measured by two indicators-years of formal
education and work experience. The first indicator simply consists of the total number
of years at school. The second consists of the number of years worked after school.
Social Capital. Social capital was proxied by 1) the amount of work contacts with
people in other organizations, particularly people with the same level of education, the
same position at another company, and with people with many subordinates, 2) the
number of association memberships-for instance, Rotary or Lions, professional
associations, frequency of attending receptions (in five categories, ranging from no
memberships to four or more memberships), and 3) the number of family contacts in
managerial positions. The number of work contacts was actually measured on a
Mokken-scale (H=0.61; rho=O.79) (Niem611er, Van Schuur, and Stokman 1980). To
test hypothesis 7, we used the first two indicators-work contacts and memberships-
computed factor scores on social capital for all the managers, and divided the resulting
scale in four groups of comparable size (339 with hardly any social capital, and then
420, 304, and 296 managers with increasing amounts of social capital).
Income. Income was measured in gross annual income in Dutch guilders, with the
following categories: 1) less than Fl. 50000, 2) F1. 50000-70000, 3) F1. 70000-
100000, 4) Fl. 100000-150000, and 5) more than Fl. 150000. One Dutch florin or
guilder (1998) is worth about U.S. $0.50. For the analyses, we used the logarithm (In)
of the average category income to create a normal distribution. Except for the function
level variable, all other variables are quite normally distributed.
Function Level. Function level was measured by the logarithm (In) of the number
of a manager's direct and indirect subordinates, which takes into account differences
in the positions of managers in organizations of different sizes. It provides more
information than just the job title. For example, a commercial manager at a large
multinational is usually much higher functionally than a director at a smaller firm.
Job Satisfaction. Job satisfaction was measured by constructing a scale asking for
level of satisfaction with several aspects of the current job (H=0.41; rho=O.86).
Perceived Chances of Mobility. Perceived chances of mobility were measured
with a scale consisting of three items (H=O.51; rho=O.66).
Quality of Information on the Labor Market. This variable was established with a
scale of items on information on relevant vacancies, working conditions, job contents,
and possibilities of doing similar work at other organizations (Mokken-procedure on
four items produced a scale with H=O.60; rho=O.75).
Importance of Advertisements. The measurement of importance of
advertisements as a source of information on vacancies is straightforward.
THE ROLE OF INFORMAL CONTACTS FOR THE DUTCH

MANAGER
Our research shows that personal contacts are a very important instrument for
managers and would-be managers not only to get a job, but also to get a better job.
0-03 N
o
0\
~
- ~
-ti<::l
Table 1. Way offmding currentjob for Dutch managers oflarger companies
i3
~
Managing Personnel Commercial Manager Financial Middle
Directors managers managers production/ managers management ~
<")
automation §:
1. Employment bureau 1% 1% 1% 1% 3% 4% Q
'1::1
~.
2. Respond to an advertisement 18% 28% 34% 32% 46% 38% ~
;::
in daily /weekly
-\::>..
t"-<
t;.
3. Advertisement in professionaljournal 7% 7% 5% 11% 5% 9"10 ~
~.
-
4. Company-intemal advertisement 0% 3% 1% 0% 1% 0%
5. Family 12% 3% 7% 6% 4% 4%
6. Acquaintance 3% 7% 4% 4% 5% 4%
7. Work contact 6% 9% 4% 6% 4% 4%
8. Asked by employer 49% 36% 40% 40% 27% 30%
9. Asked by an employment agency 6% 9"10 4% 1% 4% 7%
Total percentage: 100% 100% 100% 100% 100% 100%

N (1369) (517) (167) (232) (278) (104) (71)
Table 2. Network characteristics and information on the labor market among Dutch managers of
larger companies, 1987
Importance of
Information about advertisements for Perceived
labormarket information about mobility
opportunities the labormarket chances
Number of Tc= .08** Tc=-.l0** Tc= .06*
memberships (n = 1394) (n = 1347) (n = 1162)
Number of Tc= .25** Tc=-.l0** Tc= .13**

work contacts (n = 1380) (n = 1340) (n = 1149)
Number of Tc= .07** Tc= .OI Tc= .00

family-contacts (n = 1371) (n = 1330) (n=1143)
Kendall's Tau-test: * statistically significant < .05; ** statistically significant < .01
Sixty-two percent of the managers in our study did find their current job through some
kind of informal tie, and up to 75 percent of the managing directors found their current
job that way (Tables 1 and 4). The numbers are somewhat higher than those found by
Granovetter in his research (58 percent), but his figures relate to another country and
another time. Moreover, his respondents are not a random sample (Granovetter 1995:
7-10) and he did not inquire into the category 'asked by employer,' which is how,
according to our research, most managers (40 percent) in the Netherlands get ajob.
Granovetter (1995 : 19) reports that 14.8 percent of the managers he studied found
their job through direct application. The kind of frrms worked in by the managers who
were studied by Granovetter is unclear. Our study focuses on managers of firms with
fifty employees or more. Four out of ten of our respondents, and half of our managing
directors, stated that they were asked personally by their current employers to take the
job. Although around one-third of Dutch managers got their jobs through
advertisements, advertisements do not seem very important to Dutch managers, as 66
percent of our respondents do not think they are relevant as a source of information
(see Tables 1 and 2). In the recent National Organizations Study, 50 percent of all the
employers in the U.S. said they used advertisements when they recruit managers from
outside (Marsden 1995: 138). Referrals from current employees as well as referrals
from business or professional contacts are used less often to recruit new managers
from the outside. One difficulty interpreting the figures from that study is that
employers were not responding to a question about an actual hiring but about what
they typically do when they want to fill a vacancy through external hiring. Moreover,
employers often use two or more recruitment methods simultaneously.3
To determine whether social networks lower the costs of search (Hypothesis 1),
we looked into the influence of networks on the amount of labormarket information,
the perceived importance of advertisements, and the perceived chances of upward
mobility. If it is cheaper to acquire information on the labor market through informal
channels, then people with more extensive and diverse social networks should have
more labormarket information, and they should also perceive more opportunities for
mobility. Advertisements would also probably be less important to them. Table 2
shows that people who have more work-related contacts, attend receptions and other
social gatherings more frequently, are member of more associations and have more
family members in managerial positions do indeed have more information about the
labor market. In addition, they also have a lower opinion of advertisements as a source
of relevant information on vacancies, and they see better chances for themselves on the
labor market. The associations found also make clear why financial managers and
middle managers have a greater use for formal channels than informal ones (see
Tables 1 and 4); they have fewer contacts at other companies and they attend fewer
receptions and the like than do managing directors or commercial managers. Although
the indicators are only proxies for search costs, our results nevertheless confirm the
hypothesis that networks lower the costs of job search.
The hypothesis on better placement through social networks (Hypothesis 2) has
been corroborated by our findings. Table 3 shows a statistically significant4 positive
association between informal ways of finding a job and income earned, level of the
attained job, number of subordinates, and job satisfaction. If a managerial job is found
through some kind of informal contact, this job will be at a higher level, bring a better
income, entail responsibility for more subordinates, and give greater job satisfaction.
Multiple regression analysis of income on education, work experience and use of
personal contacts for finding a job has resulted in the following statistically
significantly beta-coefficients: education = .42, work experience = .35, and use of
=
informal contact to find the job .13 (R2 .20). =
Hypothesis 3 on the strength of weak ties fared less well when confronted with
the data. According to Table 3, it does not find support in our data. There is no
statistically significant relation between the strength of a tie with the contact person
and the income acquired (Tallc = 0.03). Moreover, managers who found their job
through a weak tie are also not more satisfied with their job, nor do they have more
subordinates or higher-level management positions. Yet, as can be seen from Tables 1
and 4, strong ties are rarely used to find a managerial job. Strong ties are more often
used at smaller firms (see Table 4), a fact that is probably related to the fact that small
firms are more often family finns.
The test of the career-cycle hypothesis (Hypothesis 4) shows no statistically
significant associations between a more advanced career and a greater use of informal
ties in general or weak ties in particular. The third panel in Table 4 shows that work
experience also does not have a statistically significant effect on the use of informal
ties in general nor on the use of weak ties in particular.
Table 3. Information channels used by Dutch managers of larger companies and

characteristics of the job
Attained Number of Work

PQsition Income subordinates satisfaction
Informal contacts Tc = .21 ** Tc = .15** Tc=.l2** Tc= .09**
vs. formal channels (n = 1364) (n = 1364) (n = 1359) (n = 1368)
Non-family Tc= .01 Tc = .03 Tc=.OO Tc= .02

vs. family (n= 843) (n = 843) (n= 840) (n = 844)
Kendall's Tau-test: ** statistically significant < .01.
Concerning Granovetter's hypothesis on the effect of the job level on search

behavior (Hypothesis 5), more positive findings can be reported. Table 4 demonstrates
that managing directors usually make more ample use of personal contacts to find a
job than lower-level managers. Personnel managers and commercial managers are
special in that, together with managing directors, they have more work-related contacts
at other companies and professional associations. For example, 47 percent of the
personnel managers are members of a professional association and 44 percent have
contacts with a professional recruitment agency. Most of the commercial managers (67
percent) state that they find advertisements unimportant as a source of information
because they hear that information from others. The relation we have noted between
the executive level and the use of weak ties is contrary to Granovetter's hypothesis.
Although weak ties are more important to most managers than strong ones as an
avenue for locating a job, they seem to be especially important to people in lower
management positions.
We have also examined some special categories of managers. Managers with
more human capital in the form of education generally make more use of informal ties.
This human capital effect is even more clear in their use of weak ties. No statistically
significant differences have emerged between male and female managers in how they
acquired their present position. But since women are a very small part (4 percent) of
our research group, not much weight can be given to this statistically nonsignificant
difference. Of greater importance seem to be our findings on how a job was found by
managers in large and in small establishments, combined with what was found on
internal and external recruitment. Table 4 shows that the size of the establishment
combined with whether recruitment has been internal or external makes a statistically
significant difference. Managers who change firms and start at a smaller establishment
of another frrm make more use of social contacts than those who are going to work at a
smaller establishment of their own frrm. One explanation could be that companies with
large establishments have ample relations with small suppliers or organizations
servicing them. These contacts probably promote informal mobility among firms (see
also the chapter by Higgins and Nohria).
The next question about the job-finding process is whether people with more
social relationships have also more often found their present job through an informal
channel. The answer to this question is important, not only as a preliminary test of the
network-as-resource argument (Hypothesis 6) but also because a positive finding
would indicate that our implicit assumption on the causal order of social capital and
job outcomes has some truth to it-ties are used to find jobs and attain a decent
income, rather than that ties are the result of a job and a particular labormarket
position). This validation is needed because the research design was cross-sectional,
meaning that the respondents' social capital was not measured before they entered
their current job but that their social capital, as well as their human capital and income,
were all measured in the same period.
Table 5 shows that managers with an average or large number of work contacts at
other frrms have more often reached their current positions via informal channels.
The same can be observed of managers with relatively numerous club and association
memberships. Although the differences are not very large-the extreme categories
differ only by 10 percent-they are statistically significant. This does not hold true for
Table 4. Group-specific differences in employment of social contact in finding a job among Dutch
managers oflarger companies
Percentage employing social contacts Percentage

non-family
Total external internal total
Population recruittnent recruittnent population
(N=1402) (N=776) (N=618) N=1402)
Gender
Women 63 62 67 94
Men 62 54 71 88
Formal education
Primary-extended primary 63 63 64 79
Grammar 67 **) 54 **) 79 **) 83 **)
Higher vocational 54 48 64 91
University 71 69 75 94
Work experience
0-10 years 58 50 73 88
11- 15 years 48 36 67 89
16-25 years 62 55 72 90
26-50 years 63 58 70 86
Size offirm or establishment
< 100 employees 64 64 **) 74 84 **)
> 100 employees 60 52 70 92
Function
Managing director 74 72 76 84
Personnel manager 62 **) 55**) 69 89
Commercial manager 59 50 69 94
Manager production/automation 57 49 73 92
Financial manager 45 40 60 90
Middle manager 49 49 52 91
** statistical significance of chi -test <.01.
2
the association between the number of managers in one's own family and the use of
informal job entrance means (this absent association is not shown here). So, we have
found some support for the network-as-resource argument. These results reinforce the
confidence we have in our indicators of social capital.
The final question on the returns on social capital on the job (Hypothesis 7) and
the related one on the relative returns of human compared to those of social capital, is
answered in Table 6. This Table presents the average annual incomes of managers in
four categories of human and social capital. More social capital clearly produces a
higher income. There is also a clear positive income effect of having a better
education.
Burt's hypothesis that social capital enlarges the returns of human capital
(Hypothesis 8) was refuted. An examination of successive rows of the same Table
Table 5. Proportion of Dutch managers finding a job through social contacts by two indicators of
social resources, 1987 (n=1359)
Percent Number of F-test

informal valid cases (significance)
Work contacts with managers
in other organizations
1. none 54.6 240
2. very few 55.4 276
3. moderate 67.9 274
4. many 65.0 546
(1336) 5.57 (.0008)
Memberships in clubs.
professional organizations. etc.
1. none 63.3 128
2. one 54.9 446
3. two 63.8 423
4. three 66.3 252
5. four or more 71.3 87
(\336) 3.80 (.0043)
shows that human capital and social capital interact: social relations help to earn more
income at any level of human capital. The returns of social capital are about equal for
all the educational categories but, as is shown in the successive columns of Table 6,
the returns of human capital in terms of income decrease at higher volumes of social
relations. The same pattern emerges, although somewhat less clearly, if human capital
is measured by years of work experience: social capital does not multiply the returns
of human capital. This is particularly the case for the managers with the most social
capital. The difference in mean annual income between managers with primary
school education and those with an university education is smallest for the managers
with the lar~est social network (for more intricate analyses, see Boxman, de Graaf, and
Flap 1991).
CONCLUSIONS AND DISCUSSION

Our questions addressed the extent to which managers employ informal contacts to
find their job and whether informal contacts benefit managers in terms of income and
the like. Our research among Dutch managers made it clear that managers generally
rely on social contacts to find their jobs, especially at the higher executive levels. This
finding gives further credence to the conjecture of a U-shaped association between job
level and the use of social contact, a conjecture that explains the apparent contradiction
between general social surveys that establish a negative association between job level
and the use of informal job-finding methods, and our finding that managers, a
well-defined group of high-status occupations, mainly get their jobs through social
contacts.
Table 6. Average gross yearly income of Dutch managers of larger companies in Dutch guilders (x
1000) by social capital and formal education. 1987 (n=1359)
Social capital
low high
Formal education (I) (2) (3) (4)
1. Primary/low vocational 73 (n=45) 86 (n=41) 88 (n=26) 128 (n=18)
2. Extended primary-grammar 81 (n=llO) 100 (n=107) 104 (n=80) 114 (n=75)
3. Higher vocational 91 (n=14O) 99 (n=206) 104 (n=127) 117 (n=136)
4. University 108 (n=44) 132 (n=66) 132 (n=71) 142 (n=67)
Dutch managers do receive higher earnings if they have used informal job-finding
methods. and their work satisfaction is higher as well. Moreover, the ones at higher
executive levels have more frequently found their job informally. Contrary to
Granovetter's hypothesis, weak ties do not bring better jobs, although they are the
most important way managers have found their present job. His ideas that weaker ties
are used most widely by higher-level managers and that managers at more advanced
stages of their career make more use of informal job finding methods have also been
refuted.
The specifications of Granovetter's hypotheses with the help of the social capital
theory have held up well in our empirical analysis. Managers equipped with greater
number of social ties (more memberships and external work contacts) make more
ample use of their social contacts to find a job. In their jobs as managers, these
contacts provide them with social capital as they earn a higher income if they have
more social ties. Social capital and human capital both contribute to a higher income.
Burt's argument that social capital multiply the returns of human capital has, however,
been refuted. Human capital and social ties interact in their contribution in a particular
way to income: social ties bring the social capital of higher returns at any level of
human capital, but for managers with a larger volume of social ties, more human
capital does not add to their income. Human capital and social capital are only partly
interchangeable as regards earning an income as a manager: social capital can act as a
substitute for human capital, but not the other way round. Contrary to the hypothesis
forwarded by Burt (1992), human capital is not most valuable for people with the most
social contacts but for those who are practically without any social resources at all; it is
least valuable for managers with many social resources. Elsewhere we have
established that human capital produces social capital, but the effect is not very strong
(Boxman, De Graaf, and Flap 1991).
Apart from the confirmation that informal searches lower the costs of collecting
labormarket information, there is more direct evidence to support the idea that is at the
basis of most of the arguments on networks and labor market chances: managers with
larger and better networks are better informed about the labor market and have a lower
opinion of the value of advertisements.
A number of our findings call for further comments. The refutation of the weak
ties argument stands out, or rather the question why. Of course our operationalization
of strong and weak ties may be held responsible for the absence of any positive effects
of weak ties in our study. But one should bear in mind that according to other studies,
including general social surveys among the workforce, informal ties do not always
lead to better jobs for every group, and weak ties hardly ever do. For a review, see
Flap (1991), Breiger (1995), and Granovetter's (1995) afterthoughts in the second
edition of his 1974 book.
Probably the answer should be sought in the fact that there are structural and
institutional restrictions that have influenced the search process as well as in the search
outcomes. Important examples of disturbing factors include the situation of contact
persons who act as go-betweens, the search and recruitment behavior of companies
that search for candidate managers, and the situation in a particular sector of the labor
market. As to the first factor, it is an empirically well-established relationship in
studies on networks and social stratification that contact persons in higher social
positions bring better returns. So perhaps the determining factor is not the informal tie
as such or the strength of a tie but the person at the other end-what he stands for or
the resources he or she provides access to. For example, individuals of higher social
standing signal higher value: perhaps they are more trustworthy as judges of quality in
others, or perhaps they do indeed have the necessary contacts with the employer who
does the hiring and is looking for a candidate.
Furthermore, recruitment behavior differs between organizations and sectors of
the labor market, depending on tradition or on what is needed on the job. A particular
kind of search behavior by a job candidate-for example, an informal search-might
have quite different outcomes, especially when a job is such that the person who
occupies that job can do a lot of damage, employers recruit informally and preferably
through stronger ties, thus enlarging the returns on informal search. For these jobs,
they need more subtle information on the quality of a candidate to lower the risks of
damage (Flap and Boxman 1996).6 By hiring someone informally, the employer has
some assurance that the contact person will somehow vouch for the individual he has
recommended and that the newly hired person will get along with sitting personnel.
Tacit skills needed at the job will also be more easily transferred to the newcomer if
there is some preexisting tie (cf. Grieco 1987; Marsden 1995). This explanation seems
appropriate for the job-finding process of managers, especially given that 40 percent of
the managers and about 50 percent of the managing directors in our sample were asked
to apply for the job by their employer.
The value of informal search, or for that matter the value of a tie is contingent.
Employers should recruit informally. If contact persons are involved, the contacts
should have relevant information or contacts to still other people who do, and they
should be prepared to share that information. Granovetter sometimes gives the
impression that news flows along the ties in a network like water in a system of canals
(cf. Frenzen and Nakamoto 1993). But contact persons are not always prepared to pass
on job information to anyone they know because in doing so they vouch for the person
they give this information to.
The career hypothesis has also been refuted. As to why managers at later stages of
their career do not make greater use of their social contacts to find a job, it is
conceivable that the number of relevant contacts does not increase with the number of
years at the job. This can be the case when, for example, the job does not bring with it
many contacts, the organization operates in a market segment with fewer contacts, or
the employee has been working too long at the same company thereby ossifying his or
her network so that fewer persons elsewhere know about the employee's qualities (cf.
Granovetter 1988). Yet another possibility is that for certain higher-level occupations,
the specific labor market is small and circumscribed. The people in that market all
know each other and the readiness to provide information about a particular vacancy to
others might decline if a would-be information provider fancies that job himself
(Spector 1973).
Much still remains to be explored in research on the instrumental role of networks
in the life of managers. For example, longitudinal studies could establish in greater
detail the extent to which a particular network comes with the job or is a prerequisite
for getting and doing the job. Another question that might be answered with
longitudinal data is whether individuals with more social capital hop faster from one
job to another than those with less social capital.
A second item on the research agenda is the existence of organizational and
institutional variation, which might make for different returns of social capital. The
instrumental value of social relations is contingent partly on the institutional and
organizational context and partly on the strategy chosen by the management of the
firm. Our analysis has been practically devoid of any institutional context. Elsewhere
(Boxman, De Graaf, and Flap 1991) we did check for company size, number of
subordinates, and market sector (manufacturing or service), but that did not greatly
alter the results presented above. Nevertheless, an important question we only touched
on is: how do institutional conditions alter these effects? Bauer and Bertin-Mourot
(1991) argue that in France, state-controlled firms and financial institutions often
perform more poorly than do private banks because they parachute into top positions
managers who are from outside and who lack ties to the people working at the firm or
to relevant business partners outside the firm. In fact, specific institutional
arrangements, such as appointment procedures and rules of succession, are responsible
for this.
Most importantly perhaps, future efforts should concentrate on the mechanism
that accounts for the social capital returns on social structure. What is the mechanism?
There are many ways for social connections to provide advantages resulting in a
higher income and not just at the moment of job-finding. Is it scarce information, for
example, about the situation in the particular market the company operates on? Is it
learning from the good and bad examples of other managers at other companies? Or
should we look at the recruitment process? How do employers recruit managers? Do
they 'buy' managers who are rich in social networks? Do they have more trust in the
quality of a managers if referrals come from particular others? Do certain networks
keep others from opportunistic short-term profit-making (Meyerson 1994)? Do
managers indeed experience higher rates of return on their social contacts as a reward
for the higher returns they and their contacts bring to the firm? Or are some relations
more helpful in doing a job while others are needed to get a better compensation, at
least partly independently of how the job is done (cf. Belliveau, O'Reilly, and Wade
1996; Lazega, this volume)?7
All this brings up the question we referred to above-the question of the
relationship between individual-level and firm-level social capital. Managers usually
use their own networks to promote company goals. D'Aveni (1991) demonstrated that
managers with an extensive informal network in the business community can help
ward off business failure and bankruptcy. But the agency problem of managers who
act opportunistically and misuse corporate social structure to advance their own goals
to the detriment of the goals of the company they work for (cf. Meyerson 1994) makes
it clear that individual-level and fIrm-level social capital are not identical. The latter is
also not a simple aggregate of the former (see Leenders and Gabbay, this volume;
Pennings and Lee, this volume) because the collective good aspect of corporate social
capital creates the danger of underinvestment in corporate social networks. Managers
sometimes do not share business and other work contacts with each other because they
are competing for status, money, or other company bound rewards. Property rights to
social capital are unclear, as can be seen when employees are lured away by a
competing fInn, and take along their contacts with clients or suppliers. This is done,
for example, by account managers at investment banks (cf. Eccles and Crane 1988), or
by editors at publishing houses, who take along the authors whose work they have
edited in the past (Powell 1984). Many companies have formal and informal rules
forcing managers and other employees to share their contacts with the other employees
(cf. Lazega, this volume).
It cannot be denied that companies do benefIt from the private part of a manager's
network as well. It was recently demonstrated by Podolny and Baron (1997) that, for
managers, having a tight core network of persons with whom to consult in the event of
personal problems is conducive to promotion. Earlier research has shown that married
managers are evaluated by their superiors as being more productive and are given
higher functions (Korenman and Neumark 1991). A wife who does not have ajob and
can devote more time to supporting her husband is especially advantageous to
individual managers (Pfeffer and Ross 1982}-and probably also to some extent to the
company he works for.
In a wider sociological context, it is also interesting to contemplate the
possibility that the networks of managers help them reproduce the positions of their
family group or class. Wright and Cho (1992) demonstrate that the lines of authority
are quite permeable, at least as far as friendships between managers and
nonmanagers are concerned. These lines are far less divisive than those of property
and expertise. Wright and Cho account for this by noting the many opportunities for
interaction that are forced on managers and other employees while at the workplace.
In what could be called a semiquantitative, comparative study, Bourdieu and De
Saint Martin (1978) describe how different factions of the French elite-either
connected to state-owned or state-controlled enterprises or to large private fIrms
largely controlled by a small number of families-try to preserve and reproduce
their status and social capital in the next generation by way of various strategies. For
the former, educational credentials plus friendly connections with other highly
placed offIcials are their main resource. The latter typically try to hold on to their
position by having more children and having them marry children from other
families in the private faction of big business. 8
We would like to express our gratitude to P. de Graaf and 1. Schmidt for contributions to earlier versions
of this chapter. Vedior Personnel and Advice 8.V. at Almere Haven allowed us to collect the data. Please
address all correspondence to H. flap. ICSlDepartment of Sociology. Utrecht University. Heidelberglaan
1.3508 TC Utrecht. Email: h.flap@fsw.ruu.nl.
NOTES
I. A large literature exists on the networks of managers filed under the heading 'interlocking
directorates.' With few exceptions, the numerous quantitative studies of interlocking directorates barely
touch on the importance of networks of individual managers. Stokrnan, Van der Knoop, and Wasseur
(1988) is such an exception. Their study of large corporations in the Netherlands suggests that the pattern
of interlocking between large firms and banks is not only the outcome of corporate actors looking after
their interests by using their directors and CEOs, but also the outcome of the actions of individual
managers, who, looking after their own interests, create and use their own network.
2. As was noted above, our dataset does not include information about direct applications. In his study
on the job-finding process of professional, technical, and managerial workers, Granovetter (1995: 19)
reports that 14.8 percent of the people in the managerial category (there were only 81 managers in his
sample of 282) succeeded in finding their current job through direct application. For a discussion of the
difficulties under the heading 'direct application,' see Granovetter (1995: 154-156). There are indications
that direct application brings in less income compared to the use of informal channels (see Faase 1980).
3. If we combine informal job-finding methods, human capital, and social capital into one multivariate
analysis, all three of them have a statistically significant effect on the income managers attain.
4. Statements about statistical significance refer to the .05 level.
5. Their National Organizations Study in the U.S. showed that traditions within a particular firm seem
to be a stronger detenninant of recruitment methods than characteristics of the firm or the occupation
(Marsden 1995: 149).
6. There is also the possibility that the strength of a tie is not measured correctly if one equates
colleagues with weak ties or for that matter friends with strong ties. In our research on how people got
ahead in their occupational careers in former East Germany, people classified 24 percent of their ties to
workrnates as strong and 15 percent of their friendship ties as weak (VOlker and flap 1998).
7. These questions lead to a wide range of other questions. For example, what do the networks of
managers mean to the goal achievement and performance of their department or of the company as a
whole (e.g., O'Aveni 1991)? Are different types of networks needed to accomplish different company
goals (Briider! and Preisendorfer 1997)? And is social capital also managerial capital-that is, can it be
willfully created and can it be steered (Kanter and Eccles 1992)? For a review of these questions, see
flap, Bulder and Volker (1998) and Leenders and Gabbay (this volume).
8. A somewhat similar idea of Bourdieu is that there is a division between different types of jobs.
Higher jobs, especially, differ according to whether they belong to the economic domain or the cultural
domain. Earnings advantages are likely to be found among people who choose education and occupations
in the sector of their origin and in occupations where the criteria for measuring work performance are
unclear. Recent research by Hansen (1996) in Norway demonstrates that this indeed the case and is also
true for managers in the cultural and in economic domains.
The Changing Value of Social Capital
in an Expanding Social System:
•
Lawyers in the Chicago Bar, 1975
and 1995 12
Rebecca L. Sandefur
Edward O. Laumann
John P. Heinz
ABSTRACT
Social capital is 'some aspect of a social structure' (Coleman 1990: 302) that acts as
a resource that individuals may appropriate and use for their own purposes. In this
paper, we examine the economic value of ties to local professional elites:
specifically, the income returns to Chicago lawyers of contacts among the elite of
the Chicago bar. Contact with the elite of the bar represents a channel through which
rank and file lawyers may 'tap in' to the social structure of the bar and acquire
valuable resources. The information and influence lawyers access through contacts
with notables are properties of the corporate organization of the bar, and, as such,
are benefits of corporate social capital. We briefly outline a theory of social capital
and suggest ways in which elite ties may act as social capital. We then discuss
changes in the social organization of the bar and suggest how these changes may
affect the value of the social capital represented in elite contacts. We reason that
acquaintance with elites will become more valuable because of the relative scarcity
of such contacts in larger social systems. In analyses of factors affecting lawyers'
incomes, we find evidence consistent with the hypothesis that ties to elite system
members are more valuable in a larger system.
INTRODUCTION
In addition to the obvious reasons for hiring a lawyer-i.e., to acquire expertise
about the law and its procedures-lawyers are sources of a broader range of
information about the community or the business environment, and they help
corporate clients span the boundaries between their own organizations and other
organizations. Lawyers are not the only agents employed by corporations for the
latter purposes, of course, but lawyers make an important contribution to these

functions. When a business hires an attorney or a law firm, it gains access to
courtroom skills and legal knowledge, which are aspects of the lawyers' human
capital, and it also gains access to the lawyers' social capital. One component of this
social capital is lawyers' connections to other lawyers, both within and outside their
own firms (cf. Lazega, this volume). Contacts with other lawyers provide useful
information about opportunities, contingencies and strategies, and they give access
to persons who may be willing to use their influence to benefit the lawyer, whether
by dropping the lawyer's name in tones of approval and regard or by intervening
more directly on the lawyer's behalf. These connections may benefit the lawyers'
clients, as well as the lawyers themselves.
This paper considers the returns that accrue to a particular form of social
relationship, ties to local professional elites, held by lawyers in the bar of a major
American city. 1 Our analysis builds on the fact that these urban lawyers constitute a
large, loosely-bounded social system. The social system of the bar differs from the
social system of a large formal organization such as a large corporation in the
distinctness of its boundaries; but, the bar is not unlike many large corporations in
the contemporary United States in its size and in the diversity of work that is
conducted within its boundaries (cf. Burt's (1992) study of a U.S. high technology
firm that is 'the size of a small city'). At the same time that the private practice
lawyers we consider in this paper are members of the corpus of the bar, they are
themselves members of corporate formal organizations: law firms ranging in size
from a solo practitioner and her secretary to large firms employing hundreds of
lawyers in offices around the country. The social capital we will consider is social
capital from sources outside the lawyer's office, but it is social capital indigenous to
the social system of the local profession.
Lawyers who possess valuable social capital, providing access to information
and ties to influential others, should benefit from this capital in two ways that will be
reflected in the rewards that accrue to their work. First, lawyers with superior social
connections should have an advantage in attracting clients, since they have the
benefit both of access to information about opportunities and of contacts potentially
willing to refer business to them. Second, lawyers with superior social capital should
be more valuable to their clients. Access to well-placed contacts gives lawyers
access to resources useful not only in finding and courting new clients, but also in
dealing with the legal matters of existing clients. Successful legal strategy requires
not just technical knowledge of the law, but also other salient knowledge, such as
knowledge of others' personalities, of others' likely strategies, of others' history of
past behavior in similar circumstances. Such extra-legal knowledge aids lawyers in
outwitting and outmaneuvering their opponents in the adversarial contests carried
out in the texts of legal briefs and in conference rooms and court rooms. Well-
connected lawyers whose associates can pass along such useful extra-legal
knowledge should be more valuable to their clients than equally competent lawyers
without access to such useful contacts.
Below, we briefly outline a theory of social capital that identifies important
differentiations in its forms by examining the mechanism through which it has its
effects and the way in which it is acquired. In light of these observations, we suggest
The Changing Value of Social Capital in an Expanding Social System - 219
specific ways in which ties to the elites of one's profession may act as social capital.
We investigate whether ties to elites are valuable to lawyers in increasing their
incomes net of other characteristics and organizational contexts which might aid
them in attracting clients and in being superior lawyers. We then ask whether the
value of the social capital of elites ties changes when the social system in which it is
embedded changes.
SOCIAL CAPITAL
Our conceptualization of social capital draws from the work of James S. Coleman
(1988, 1990). In Coleman's formulation, social capital is appropriable social
structure; it is 'some aspect of a social structure' that acts as a resource that
individuals may use for their own purposes (Coleman 1990: 203). Such social
structure may exist in relatively discrete forms, such as organizations, or in more
diffuse forms, such as extended families, communities, or other loosely bounded
social systems. Always, it consists of relationships. These relationships may be
components of formal organization, such as the relationships of classmate,
department head, co-worker, and instructor; or, the relationships that constitute
social structure may be defined by other criteria, such as the relationships of
neighbor, lover, uncle, co-conspirator, and friend-of-a-friend. These relationships
may be characterized by both their structural form and the content that inheres in
them; and, aspects of both their form and their content will condition their
productivity as social capital.
An individual's stock of social capital consists of the collection and pattern of
productive relationships in which she is involved and to which she has direct access,
and further of the location and patterning of her associations in larger social space.
That is, her social capital is both the contacts she herself holds and the way in which
those contacts link her in to other patterns of relations. Social capital thus exists in
an amazing multitude of forms; nevertheless, any form of social capital may be
productive through one or more of three types of benefits: information, influence
and control, and social solidarity. Information benefits arise when social
relationships provide access to relevant, timely and trustworthy information of use to
the actor in question (Burt 1992; Laumann and Knoke 1987). The influence and
control benefits of social ties are obverse sides of one coin: the ability to influence
others (Parsons 1963; Coleman 1990) and the ability to be free of others' influence.
Solidarity benefits arise when there is some degree of mutual trust and commitment
among a group of individuals. In this paper, we confine our discussion to the first
two types of benefits. 2 Our conception of social capital differs from the definition
given by Gabbay and Leenders (this volume), who emphasize a distinction between
social ties and social capital. In their view, social capital comprises the beneficial
resources actors draw from their social networks, rather than the relationships that
constitute those social networks. In our view, the aspects of social structure and
social relationships that are potentially valuable through their provision of beneficial
resources are forms of social capital; beneficial resources provided by forms of
social capital are what we term benefits (Sandefur and Laumann 1998b).
As presented by Coleman (1988, 1990), social capital is defined by its function;

thus, any form of social capital has a quality that may be called its valence. That is,
forms of social capital which provide one benefit or are useful for one purpose may
not provide a different benefit and may be useless or actively harmful for other
purposes. This variable characteristic of benefit or liability, of positive or negative
'charge,' may change given the context in which a particular form of social capital is
activated and the purpose for which it is employed. Among lawyers, for example,
the solo practitioner is free of formal collegial entanglements and so has the
opportunity to be a self-determining entrepreneur (Seron 1996). At the same time,
however, he is a relative isolate, which means that, while he is relatively free of
others' control, he may lack others on whom he can rely when he faces difficulties.
Thus, Arnold and Kay (1995) in a study of lawyers' professional misconduct, find
that solo practitioners are more likely than lawyers who work in law firms both to he
charged with acts of professional misconduct, and to be convicted of them. The
misbehaving solo practitioner might have been saved from disgrace if he had been
prevented by vigilant colleagues from committing his crime. Post hoc, he might
hope to mobilize others to collude to hide his misdeeds or to try to sway the
decisions of members of the disciplinary committee considering his disbarment.
Thus, social structure that is social capital for some purposes is social liability under
other circumstances.
The third characteristic of social capital to note is the variation in its mode of
acquisition. Some forms of social capital are acquired relatively passively (such as
family connections acquired by birth), while others may be actively sought (such as
acquaintance with influential colleagues). The acquisition and cultivation of such
key relationships can be part of a deliberate strategy of personal advancement, as
well as the unintended by-product of ordinary informal social contact.
The Social Capital of Ties to Professional Elites

Lawyers who work in private practice are rewarded with income accruing as a result
of evaluation by two audiences, their employers and their clients. Lawyers can
increase their incomes by being superior performers and by appearing to be superior
performers. Contacts who are knowledgeable about the bar or its clients can be
valuable by providing a lawyer with information that helps make her human
capital-her talents, abilities, and experiences-more productive (Burt 1992).
Contacts who are influential in the bar or with its clients can promote a lawyer's
reputation with clients and among other lawyers. Elite lawyers in a city bar can act
as such informed and influential contacts.
The elite of the bar are not simply prestigious or respected members of the
profession; they are leaders of various constituencies within the bar and liaisons with
people and organizations outside it (cf. Laumann and Pappi 1976; Heinz et al. 1982;
Heinz et al. 1997). As leaders, they are consulted because they are knowledgeable
and their acquaintance is sought because they are influential. A notable lawyer is not
just then a prominent member of his local profession, he is also a 'relay point' among
his associates. Considered from a network analytic perspective, the notable lawyer's
prominence reflects his advantageous structural position as a leader of some
segment of the bar or the bar at large-his position at the center of a pattern of
branching relationships. Through his ties to constituents and his relations with other
prominent members of the bar, he can synthesize a picture of the bar and its clientele
that a less prominent lawyer would not have the information to construct. Further,
his status as a leader may give him access to particular, possibly sensitive items of
information--either about the larger bar or about matters of interest to those whose
work intersects with his own. Thus, contacts with elites can provide a rank and file
lawyer with information about the general 'lay of the land' and about specific
opportunities.
Acquaintance with a member of the elite can also benefit a lawyer through the
influence elites may use on her behalf. Such sponsorship by an elite or 'notable'
lawyer may benefit a rank and file lawyer in three ways. First, the judgment by a
respected member of the profession that one is a valuable colleague or a key player
enhances one's reputation among peers and superiors. Second, the elite lawyer can
circulate his favorable evaluation among his other constituents and other prominent
lawyers; in such a way, his position in the network of relations in the bar can
augment the benefit of his favorable evaluation.3 Thus, one's reputation benefits
both from the seal of approval of a respected judge and from the ramifying nature of
the elite lawyer's social network. The third way in which an elite lawyer may use his
influence to benefit a constituent is by a specific intervention. For example, 'making
a call' is a common practice in which elites use both their social connectedness and
their influence to benefit their constituents by intervening in another's decision
(about, for instance, whom to hire, whom to promote, or how much to reward
someone).
Even quite unintentional actions by elites, such as the passing mention of a
name to a potential client or a gossipy story about a competitor can benefit a rank
and file lawyer. There is no doubt great variation in the intensity and type of elite
sponsorship and support. In the analyses to follow, we do not have measures of the
specific mechanisms through which the social capital represented in elite contacts
has its effects. We simply argue that, on the whole, such contacts should provide
access to information and to someone potentially willing to use his influence on
one's behalf.
Social Capital and Changes in the Social System

By this reasoning, elite contacts would be a form of social capital in a city bar with
1500 lawyers or a bar with 15,000 lawyers. But, we argue, in addition, that the rate
of income return on such capital will change as characteristics of the system change.
That is, since social capital is an aspect of social structure, as the social structure
itself changes, the value of its appropriable components may change. In particular,
we hypothesize that the value of elite ties will be greater in larger systems than in
smaller ones because the social capital of elite ties is likely to be more scarce in
larger systems.
Let us consider elites for the moment as groups with some degree of internal
social cohesiveness (rather than a priori designated proportions of the population,
such as the top 1%). For example, consider an elite such as the generally recognized
group of respected scholars in some academic field, for instance the top researchers
of poverty in sociology or of Anglo-Saxon in English. In addition to being aware of
one another's work, these scholars meet at conferences and on review panels and
they correspond about their students, their research and other concerns of their field.
Each of them may not personally know all of the others, but each is acquainted with
a (probably very high) proportion of her colleagues. Consider a situation in which
such more or less cohesive elites are relatively open to communication with rank
and file system members. The hypothetical respected scholars exemplify such an
elite, as do the leaders of a community or town (Laumann and Pappi 1976) and
prominent lawyers in a city bar (Heinz et al. 1982; Heinz et al. 1997).
As a social system expands, such elites are unlikely to grow at a corresponding
rate. Two different approaches suggest support for this proposition. Mayhew (1973)
presents a formal proof of the proposition that the rate of elite expansion will be
slower than the growth of the social system the elites inhabit. Employing Mosca's
(1939) definition of a ruling elite as some numerical minority of the popUlation,
Mayhew shows that the ruling elite of a social system will comprise a decreasing
proportion of the popUlation as the system increases in size. A similar prediction
follows from a network analytic perspective. Again, let us assume that a certain
degree of social connectedness among members of the elite is necessary for the elite
to persist as a relatively defined group. As an elite grows in size, the ability of
individual members of the elite to maintain ties of a specified intensity with other
members will decrease. Any individual can maintain only a certain number of
relationships; as the number of potential interaction partners increases, the average
proportion of those potential partners known by individuals will likely decrease.
Individuals' 'carrying capacity' for relationships thus places an upper bound on the
size of an elite with a specified degree of internal cohesion.
From the perspective of 'ordinary' system members, the question may be framed
somewhat differently. In a smaller social system, acquaintance between a rank and
file member and a member of the elite is more likely than in a larger one, if only
because acquaintance between any two randomly selected members is more likely.
Because elites comprise a smaller proportion of the population in larger systems, the
chance probability of a tie between a randomly selected member of the population
and a member of the elite is lower. To the extent that members of the elite possess
special influence or other resources that they may use on behalf of their
constituencies, access to such resources is then more restricted (in the sense of being
less widespread) in larger systems. Thus, the resources are more scarce and
consequently may be more valuable.
Changes in the Social Organization of the Bar and the Value of Elite Contacts
Since 1970, the number of lawyers in the United States has roughly doubled; the rate
of expansion of the bar has far outstripped growth in the population (Abel 1989:
Table 23). This tremendous growth in the number of lawyers has been attended by
changes in the way law practice is organized. Lawyers and their work have become
increasingly incorporated into organizations such as large law firms. Many lawyers
now occupy work roles more similar to that of employees than to the tradition of
free-standing professionals. Job placement has become increasingly mediated by law
school placement offices and by groups of firms that join forces in their search for
qualified students (Abel 1989: 224). Law firms themselves have grown larger, and
have experienced attendant organizational changes that arguably erode the

professional autonomy once characteristic of work in the law (Galantar and Palay
1991). Thus, lawyers' opportunities to recruit clients and to increase their own
incomes have become increasingly structured by the formal organizations that train,
sort, employ and reward them.
Reflecting national trends, the Chicago bar has doubled in size in the past
twenty years and law practice has become increasingly concentrated in large
organizations. In a random sample of Chicago lawyers taken in 1975, 19% of
licensed lawyers were solo practitioners, while a similar survey conducted in 1995
found that solo practitioners constituted only about 13% of lawyers. The twenty
years saw no change in the proportion of lawyers working in law firms (67%), but a
great increase in the average size of the firms in which lawyers were employed (37
lawyers in 1975, 178 lawyers in 1995). As these firms have grown, they have in
many cases formalized procedures for hiring, evaluation, and promotion. In such a
context, Chicago lawyers' income attainment chances may have become
increasingly 'bureaucratized: in the sense that a greater proportion of lawyers are
subject to formalized procedures applied with supposedly less partiality.
One could argue that the increasing formalization of lawyers' income attainment
chances should lead to a decrease in the importance of elite contacts and other
particularistic factors, in favor of the increasing importance of formal qualifications,
such as education. Such would be a classic Weberian argument about bureaucracy
(Weber 1946). However, one might also argue that key contacts would remain
important or in fact increase in value with the increasing bureaucratization of
mobility contests. Even in more formalized settings, the information accessed by ties
to elites should still be valuable to lawyers in increasing their incomes by providing
notice of opportunities, contingencies or complications. And, clients and employers
will continue to encounter situations in which there is some uncertainty about
someone's past or future performance that available information 'in the file' cannot
resolve. In such cases, elite influence or the 'reference' provided by association with
elites can help resolve this ambiguity in favor of the lawyer who counts elites among
his acquaintance. Further, since employers (and, likely, some clients) have
implemented formal procedures intended to reduce the role of non-productive
factors in hiring and reward decisions, they will find themselves faced with several
formally equivalent candidates. Here again, elite influence can be a crucial factor in
determining who among formally identical lawyers is to receive scarce and valuable
opportunities and rewards. In sum, it is not clear that increasing rationalization will
lead to a reduction in the value of elite ties.
Below, we present a preliminary test of the two hypotheses that we have
outlined above. First, we ask whether, net of more evident productive
characteristics, elite contacts are associated with higher income. Second, we ask
whether the value of elite ties increases as a social system expands. We consider the
specific case of the Chicago bar as a large, loosely-bounded social system in which a
group of elite lawyers is identified as leaders of the bar and of various constituencies
within it (Heinz et al. 1997). Employing two surveys of the bar, the first conducted
in 1975 and the second in 1995, we investigate changes in the value of contacts with
elite lawyers that coincide with changes in the social organization of the bar.
DATA AND METHODS

The 1975 study of Chicago lawyers includes 777 randomly selected respondents
with offices within the city limits (Heinz et al. 1982). The 1995 survey includes 788
randomly selected lawyers drawn from the same area (Heinz et al. 1997). The
response rates for both surveys were just over 82% of the target samples. The
lawyers in each survey were asked a series of questions about their career history,
their educational and family background, the nature and structure of their current
work and work experiences, their social and political attitudes, their social
participation, and their relationships with specified other groups of lawyers.
Analyses presented below are for private practice lawyers (those working in law
firms or as solo practitioners). The sample includes 522 such lawyers in 1975, 35 of
whom were dropped from the analyses due to missing data, resulting in a sample of
487 respondents. In 1995, the sample includes 521 such lawyers, 33 of whom were
dropped due to missing data, resulting in an analysis sample of 488 respondents. 4
In consultation with informants knowledgeable about the Chicago bar, the
principal investigators identified lawyers who appeared to be representative of
several salient elites. The elites selected were drawn from a large range of social and
professional categories: bar association officers, political partisans, type of work and
kind of clientele, practice setting, gender, race, and ethnic background (Heinz et al.
1997). Lawyers who were currently in public office were excluded from the
identified elites, to avoid conflating official power or influence with prominence
among colleagues. We refer to these selected elite lawyers as the 'notables.' The
1975 questionnaire included the names of 49 notables, 6 of whom were dropped
from subsequent analyses. 5 The 1995 survey included the names of 68 notables,
three of whom were dropped from subsequent analyses because they proved to be
relatively unknown (for details, see Heinz et al. 1997). The notables of 1995 are
more diverse demographically than those included in the 1975 study, as is consistent
with the increase in the demographic diversity of the Chicago bar. The 1995 list of
notables also includes lawyers who work as the internal counsel of non-legal
organizations and those who work for legal aid organizations, while the 1975 list did
not.
As part of each survey, respondents in the random sample were given a list of
the notables and asked to indicate which notables they knew at two specific levels of
association. On their first pass through the list, they were asked to mark those
notables with whom they were 'personally acquainted.' Next, they were asked to
mark the notables of their acquaintance who they believed would take the time to
advise them if asked. Analyses in this paper will focus on the notables nominated at
the second, stronger level of connection.
A variety of aspects of lawyers' personal and work characteristics affect their
income attainments (d. Hagan and Kay 1995; Abel 1989; Sandefur and Laumann
1998a). In this analysis, we are interested in the relationship between ties to elites
and income net of these other characteristics. In our models, therefore, we include
measures of respondents' law school quality, achievement in law school, practice
setting, practice type, and position in their firm. We include a control for age, which
is here a proxy for experience in the law. The relationship between age and income
is modeled by a term for age and a term for age-squared. This latter term permits the
positive effects of age (and, by proxy, experience) to decrease over time, and so acts
as a statistical control for the obsolescence of skills acquired early in one's
professional career. We also include a control for gender; female lawyers have been
shown to receive lower pay than male lawyers, on average and net of certain
productive characteristics (cf. Hagan and Kay 1995). We are unable to include a
measure of hours worked, since this question was not asked of respondents in 1975.
The distribution of ties to notables is quite skewed in both periods, and we have
no reason to believe that the relationship between the number of notables known and
income is linear; therefore, we constructed dichotomous measures-whether the
respondent reported knowing 1,2,3, or 4 or more notables well enough to ask them
for advice. In the regression analyses, nomination of 0 notables is the omitted
category. Age is the respondent's age calculated from her self-reported year of birth.
As a measure of the lawyer's ability (or human capital) as indicated by her law
school performance, we include a pair of dummy variables indicating whether the
respondent was in the top 10% of her class and/or on the law review of her law
school or in the top 11-25% of her class. In the regression analyses, class ranks
below the top 25% are the omitted category. Whether the respondent attended an
elite law school (e.g., Harvard), a prestigious law school (e.g., Northwestern
University), or a local law school (e.g., John Marshall Law School in Chicago) is
indicated by a trio of dummy variables; regional law schools (e.g., University of
Illinois) are the omitted category in the regression analyses.6 Practice setting is
measured by a dummy variable indicating whether or not the respondent is a solo
practitioner; effects presented in the regression analyses have as their referent
respondents working in law firms.
Practice type is measured slightly differently in the two surveys. In the 1975
survey, respondents were asked what proportion of their income they earned from
work with clients who were businesses. Two dichotomous variables indicate
whether 25% or less of their income or 75% or more of their income comes from
work for businesses (as opposed to non-profit corporations or personal clients). In
the regression analyses, the omitted category is lawyers receiving 26-74% of their
income from clients who are businesses. In the 1995 survey, respondents were asked
what proportion of their clients were businesses; using this information, we
constructed a pair of dichotomous variables parallel to the indicators constructed for
1975.7 A lawyer's position in her firm is modeled by a dummy variable indicating
whether or not the respondent is a partner. The respondent's gender is modeled by a
dummy variable indicating whether or not the respondent is a woman. 1975 income
is converted to 1995 dollars using the Consumer Price Index, and its natural log is
taken before computation of the regression equations. Modeling the natural log of
income reduces the impact of outlying observations (in this case, respondents with
very high incomes) on the coefficient estimates, and it permits straightforward
comparisons across time. In the regression equations, the estimated metric
coefficient of a variable predicting income may be interpreted as the proportionate
change in the dependent variable given a one-unit increase in the predictor. The
magnitude in dollars of the effect of a change in the value of an independent variable
will depend upon the value at which it and the other independent variables in the
model are evaluated (Hauser 1980; Stolzenberg 1980).
Table 1. Descriptive statistics for analysis variables

1975 1995
n 487 488
Logged Income (st.dev.) 11.6 (.73) 11.4 (.83)
I Notable Advisor 18% 11%
2 Notable Advisors 14% 13%
3 Notable Advisors 10% 6%
4 + Notable Advisors 27% 17%
Age (st.dev.) 44.7 (14.0) 42.0 (10.4)
Top 10% of ClasslLaw Review 33% 29%
Top 11 - 25% of Class 29% 32%
Elite Law School 23% 14%
Prestigious Law School 16% 14%
Local Law School 45% 43%
Solo Practitioner 27% 20%
Clients 0-25% businesses 31% 24%
Clients 75-100% businesses 47% 58%
Partner 42% 40%
Female 3% 21%
Table 1 presents means and standard deviations for continuous measures and
percentages for dichotomous measures used in the regression analyses. Recall that
this sample is restricted to private practice lawyers. It is worthwhile to discuss
briefly some of the changes in the Chicago bar between 1975 and 1995. The most
striking change is in the proportion of women. In 1995, 29% of all and about 21 % of
private practice lawyers in Chicago were women, nearly ten times their presence in
1975. (Currently, women are over-represented in government employment and
underrepresented in private practice.) Businesses constitute a larger proportion of
the client base of private practice lawyers in 1995 than in 1975. Solo practitioners,
who serve mostly individuals and small businesses, have declined as a proportion
both of lawyers in general (see above) and of lawyers in private practice. The share
of Chicago lawyers who graduated from elite schools decreased. This is likely
related to two factors. As law firms grew during the 20 years between surveys, they
broadened recruitment beyond graduates of top schools. Further, as the number of
applicants to law schools increased during the same period, elite schools held their
enrollments relatively constant while lower-ranked law schools increased the size of
their incoming classes. Thus, changes in both supply and demand are reflected in the
educational backgrounds of Chicago lawyers.
Acquaintance with notable lawyers is clearly much more common in 1975 than
in 1995. Fifty-three percent (53%) of lawyers know no notables in 1995, while only
31 % of lawyers lack such connections in 1975, even though the list of notables was
longer (50% longer) and more inclusive in 1995. As we suggested earlier, ties to
elites should be a rarer commodity in a larger system. A different and convenient
summary measure of the degree of interconnectedness in the bar is a simple measure
Table 2. Regressions for logged income. 1975 and 1995. Metric coefficient
estImates. standard errors In parentheses.
1975 1995
Intercept 8.49*** (.37) 7.73*** (.49)
1 Notable -.08 (.08) .03 (.09)
2 Notables .00 (.08) .25** (.09)
3 Notables .02 (.10) .13 (.11)
4 Notables .35*** (.08) .43*** (.08)
Age .11*** (.02) .13*** (.02)
Age2 -.00*** (.00) -.00*** (.00)
Top 10% of Class .08 (.06) .25*** (.07)
Top 11-25% of Class .06 (.06) .21*** (.09)
Elite Law School .02 (.09) .21* (.09)
Prestigious Law School -.09 (.09) .10 (.09)
Local Law School -.00 (.08) -.15* (.07)
Solo Practice .15 (.08) -.38*** (.09)
Clients 25% or less business .01 (.07) .02 (.09)
Clients 75% or more business .18** (.07) .24** (.08)
Partner in firm .52*** (.07) .45*** (.08)
Female -.21 (.15) -.14* (.07)
Adjusted R2 .40 .50
*p < .05. ** P < .01. *** P < .001.
Italic-face type indicates a statistically significant difference between the
two periods at the level of p < .05. Bold-face type indicates a statistically
significant difference between the two periods at the level of p < .10.
of the density of ties between elites and rank and file lawyers. In this case, the
density is simply the ratio of the number of observed nominations of elites by rank
and file members to the number of possible nominations; as such, it represents the
degree to which possible connections have been realized. In the Chicago bar, the
density of ties between elite lawyers and the rank and file has declined considerably.
In 1975, the sample density of such ties at the level of simple acquaintance was
.117; the density at the stronger level of connection was .065. In 1995, the sample
density of ties of simple acquaintance was .059, while the density of advisor ties was
.029.8
ANALYSIS AND FINDINGS

Table 2 presents results for OLS regressions of logged income in 1975 and 1995.
Estimated metric coefficients and their standard errors are presented separately for
each model in each year.9 In the body of Table 2, statistically significant differences
between the coefficients in the two periods are indicated by differences in typeface;
statistical significance within each period is indicated by the familiar asterisk
designations.
The findings support our first hypothesis that the social capital represented by
ties to elite lawyers would be positively associated with higher incomes net of other
characteristics. In both periods, counting four or more notable members of the bar as
advisors is positively associated with higher income (p < .001); in 1995, counting
two notable members of the bar as advisors is positively associated with higher
income (p < .01). The pattern of findings is also consistent with our hypothesis that
elite ties would be more valuable in the larger bar. The income returns to the highest
degree of elite connection appears to have increased slightly, but this increase is not
statistically significant. The income returns to less extensive connections to the elite
have increased to a degree that is statistically significant (p < .05).10 In neither
period does 1 tie to a member of Chicago's legal elite evidence an appreciable
relationship to income.
To get a sense of the 'real money' value of ties to notable lawyers, let us
compare the predicted incomes of similar lawyers with and without ties to local
legal elites. In 1975, the predicted income of a thirty year-old male graduate of the
elite Harvard Law School, who graduated in the top 10% of his class and worked as
an associate in a law firm where his clientele consisted of a roughly equal balance of
businesses and persons would be essentially the same whether or not he counted two
notable lawyers among his closer acquaintances. In 1995, an identical lawyer would
gain roughly $20,000 by acquaintance with two notable members of the bar. To take
a different example, consider a fifty year-old man who graduated in the middle of
his class from a local law school and operated a solo practice where he served
largely personal clients. In 1975, if he counted four or more elite members of the bar
among his acquaintance, his predicted income would be 44% higher (in 1995
dollars) than that of an identical lawyer with no notable contacts. In 1995, a similar
lawyer gains 58% more income by acquaintance with four or more elite lawyers.
The increase in income associated with elite contacts varies in size depending on
characteristics of the lawyer and her practice, but the returns to notable ties are
considerable and, on the whole, larger in the larger system than in the smaller.
Specific other findings in these models are interesting and worthy of discussion,
both as they inform us about changes in the determinants of lawyers' income and as
they suggest the workings of other forms of social capital. Earlier, we suggested that
work in the law have become increasingly structured by organizations such as law
firms and law schools, and that procedures for hiring, promoting and rewarding
lawyers have become more formalized. This observation is borne out by inspection
of the effects of class rank and law school prestige on lawyers' incomes. Changes in
the relationship between legal education and income are striking. When partnership
status, client base, law school, practice setting and gender are controlled, class rank
has no direct effect on lawyers' incomes in 1975. In 1995, however, class rank is a
statistically significant predictor of income, net of all controls (p < .001). The effects
of law school prestige are more than twice as large in 1995 as in 1975 (p < .10 for
the test of statistically significant difference between the two periods for the effect
of elite law school attendance). It appears that income attainments may have become
more keyed to the human capital signaled by law school performance and attendance
at a well-regarded institution.
Net of age, elite ties, legal education, client type and partnership status there is a
negative relationship between solo practice and income in 1995 (p < .001), while the
positive coefficient in 1975 does not attain statistical significance. In both periods,
solo practitioners have, on average, lower incomes than lawyers in law firms
(Sandefur and Laumann 1998a). Part of the deficit experienced by solo practitioners
stems from the fact that their clients tend to be individuals and small businesses. The
most lucrative of private practice legal work, work for large corporations, is
performed by lawyers in large law firms. Thus, even when solo practitioners have
businesses among their clients, these clients tend to be small, locally owned
businesses that require neither the volume nor type of work done for large
corporations. In addition, solo practitioners must rely on referrals and advertising to
attract clients,l1 while firm lawyers have the benefit of association with an
organization with an established reputation and of formal organizational contacts
with other lawyers who can share their expertise and who may make referrals that
benefit them (cf. Lazega, this volume).
Given these general considerations about the different situations faced by solo
practitioners and lawyers in firms, there are two possible reasons for the difference
between the two periods in the effect of solo practice on income. It may be the case
that, in 1975, the best lawyers in Chicago were more evenly distributed between
positions in firms and solo practice, while in 1995 the best lawyers were more
strongly attracted to work in firms. Our measures of lawyers' ability are restricted to
measures of their education, and so do not capture all the attributes that make for a
successful lawyer. Thus, stronger selection of talented lawyers into firms in 1995
than in 1975 might account for the different relationships between solo practice and
income. An alternative explanation has to do with the increasing concentration in
law firms both of opportunities to do lucrative work and for professional
advancement. Solo practitioners may be at a greater disadvantage in the larger, more
formally organized bar not only because of the greater organizational resources and
more lucrative client bases of larger firms, but because the career mobility pathways
provided in large firms are now more remunerative than those open to solo
practitioners.
Unsurprisingly, partners in law firms make more money than associates in both
periods. As we noted above, work for business clients is more lucrative than work
for personal clients and non-profit organizations; thus, those lawyers whose practice
consists largely of work for businesses have higher incomes. In 1975, the negative
effect on income of being a woman does not reach statistical significance (there
were very few women in the Chicago bar in 1975). In 1995, the negative effect on
income of being female is statistically significant, net of controls for practice setting,
client base, education and elite contacts.
Further Considerations
The results from our statistical analyses of the relationship between social ties to
elites and income in the two periods are consistent with our suggestion that such ties
are useful in increasing income and with our hypothesis about the increasing value
of elite contacts in expanding social systems. Nevertheless, competing explanations
for our findings deserve attention, and we turn now to discussion of three classes of
competing explanations.
The first set of alternative explanations has to do with causal ordering. We have
discussed our findings as though elite contacts were a 'cause' of income attainments.
Of course, the relationship is most likely non-recursive; that is, the causal arrows run
in both directions. More successful lawyers are likely more attractive associates for
notable lawyers and are perhaps more likely to come into contact with them in the
course of their work and work-related socializing. At the same time, acquaintance
with notable lawyers likely affects income attainments in the ways we have
discussed, and perhaps in other ways which our theory did not address. Further, it is
unlikely that most of the lawyers in our sample became acquainted with the elites
they nominated only in the year of the survey; thus, for many (if not most) lawyers,
we are investigating the relationship between contemporary income and
relationships formed some time before income was measured. It is highly unlikely
that the opposing hypothesis about causal order-success in law as measured by
income affects the likelihood of elite acquaintance, but the social capital of elite
acquaintance has no effect on income-is the truth. Nevertheless, if such were the
case, these findings would then tell a story about the social stratification of the bar
that would be more about the social distribution of success than about the resources
which contribute to success.
The second set of alternative explanations concerns the issue of selection or
omitted variables in an analysis. This concern is ubiquitous in social scientific work.
Any or all of our findings which we have interpreted as the 'effect' of one factor on
another may in fact be the result of selection into certain roles and circumstances on
the basis of some other factor we did not model or did not measure. For example, if
the most motivated lawyers go to work for firms, while the less motivated lawyers
become solo practitioners, then all or part of the effects of firm practice which we
have argued are due to organizational factors are in fact due to unmeasured
characteristics of the individuals who enter firm practice. Likewise, if the most
motivated and consequently successful lawyers are those who are most likely to be
associated with members of the elite, then what we interpret as the effects of social
capital are simply the social organizational results of individuals' traits. If such
selection processes differ in the two periods under study, then effects that we have
argued are due to specific changes in the social and formal organization of the bar
could in fact be due to other changes in the bar's organization that affect the ways in
which particular individuals are selected into particular circumstances. In this paper,
we can only appeal to the plausibility of our theory to counter this type of alternative
explanation. We note that, if selection were the only operative factor, the findings
would remain an interesting descriptive characterization of changes in the social
organization of the bar, though our explanation would require reconsideration.
The last set of potential competing explanations for our findings has to do with
the identification of the notables themselves. We have operated under two
assumptions about the notables identified by the survey directors. First, we have
assumed that the notables were correctly identified in each study, or at least that any
error in identifying them was not systematic in either period. Our arguments do not
require that the notables be an exhaustive or perfect sample, but they do require that
they comprise an accurate and reasonably representative characterization of the elite
of the Chicago bar. If any error in identifying the notables is systematic-for
instance, if the leaders of some sector of the bar are excluded-then our findings
may be compromised by the invalidity of the notables as representatives of the bar's
elite in one or both periods.
It is difficult to know for certain how the exclusion of some proportion of the
bar's elite would affect our findings. However, let us consider the case in which this
exclusion did occur in both periods. Some substantial proportion of respondents in
each period then possess elite contacts which we have not measured. In the
regression analyses, these respondents who do have the unmeasured advantage of
notable ties are included in the reference category against which the effect of notable
ties is presented. This should bias the regression coefficients for notable contacts
downward in both periods, a bias which is not fatal to our hypothesis about different
effect sizes in the two periods.
The second assumption about the notables is specific to the second period under
study. While the Chicago bar doubled in size between 1975 and 1995, the number of
notables that respondents were permitted to nominate was only increased by an
additional 50% or so (65 notables in 1995 versus 43 notables in 1975). We argued
above that the elite would not grow at the same rate as the bar; however, even if
such is the case, that does not mean that the list of notables for 1995 was increased
by a factor corresponding to the true increase in the size of the bar's elite. If the
process by which elites were selected in the second period more strongly excluded
relatively unimportant prominent lawyers, then our findings could be an artifact of
the greater importance of the individuals chosen for the 1995 study elite, rather than
a result of the greater importance of elite contacts.
We are very doubtful that the second survey's group of notables results from
more stringent skimming of the cream of the bar. Neither group of notables was the
result of an 'attempt to create a list of the most notable, successful or influential
lawyers in Chicago. Rather, the list[s] include[s] a selection of lawyers, of varying
types, who are prominent in one respect or another, but not necessarily more
prominent than others' (Heinz et al. 1997: 447). We find no reason to believe that the
selection procedures resulted in a systematically more influential and informed
group of notables in the second period. The second possibility, that the selection
procedure did not result in skimming, but did fail to increase the list of notables by a
sufficient amount in the second period, is less cause for concern. If such were the
case, we would again have a number of respondents who have the unmeasured
benefit of notable ties, but are included in the reference category against which the
effects of notables ties are estimated. Such an occurrence would bias the coefficients
downward in the second period, in a direction unfavorable to our hypothesis. Thus,
if such bias were present, the relationships we report in support of our hypothesis
would appear weaker than they actually are.
CONCLUSION
Contact with the elite of the Chicago bar represents a channel through which rank
and file lawyers may 'tap in' to the social structure of the bar and acquire valuable
resources. The information and influence lawyers access through contacts with
notables are properties of the corporate organization of the bar, and, as such, are
benefits of corporate social capital. The business firms and individuals who retain a
lawyer gain not only from the lawyer's human capital-her skills, abilities, talents,
and industry-but also from their social capital, which she may use as a resource in
her work for them. The findings of this paper suggest that the social capital of ties to
legal elites is associated with valuable income rewards for the lawyers who possess
it. Changes in the organization of a social system-in this case, a tremendous
increase in the size of the Chicago bar-are linked to changes in the value of given
forms of social capital. Our findings are consistent with our hypothesis that the
greater scarcity of contact with elites would lead to increases in the value of elite
ties.
As formal organizations have become increasingly salient in structuring
Chicago lawyers' access to opportunities and rewards, formal qualifications, such as
educational credentials and law school performance, have become more important
predictors of lawyers' incomes. Employment in law firms, which provides access to
organizational resources, such as office equipment, support staff and colleagues, and
to a more lucrative client base, seems to confer greater benefits in increasing income
for contemporary lawyers than for lawyers of 20 years ago. Yet, at the same time
that the distribution of rewards appears to have become in some ways more
'universalistic' and structured by formal organization, the economic value of ties to
specific others, such as those between a rank and file lawyer and legal elites, has not
declined and, in fact, appears to have increased. In the contemporary urban bar, the
extra-organizational social capital provided by ties to local professional elites
remains valuable, both to the lawyers who have such capital, and, by implication, to
the firms in which they work and to the clients who pay for their services.
This research was supported in part by grants from the American Bar Foundation and the National
Science Foundation (#SBR-9411515). Additional support for Heinz was provided by Northwestern
University's Institute for Policy Research. We thank Charles E. Bidwell. Ronald S. Burt. Jeffrey A.
Hayes, Ray Reagans. Ross M. Stolzenberg. Christopher B. Swanson. Jeffrey Y. Yasumoto. Ezra
Zuckerman. and an anonymous reviewer for helpful comments and useful discussions. This paper is a
revision of a section of a paper presented at the Conference in Honor of James S. Coleman. Mannheim.
Germany. November 2. 1996. The discussion of the benefits of social capital draws heavily on Sandefur
and Laumann (l998b). The views presented in this paper are the authors' alone. and neither the
supporting agencies nor supportive colleagues are responsible for any errors.
NOTES
1. The term 'bar' refers to the lawyers licensed to practice in a particular community. not to any
particular formal bar association or affiliation of lawyers.
2. For a more thorough discussion of this paradigm for social capital. see Sandefur and Laumann
(l998b).
3. Of course. it could also augment the impact of his negative evaluation and so have negative effects
on the rank and file lawyer's attainments (cf. Brass and Labianca. this volume). In this paper. we measure
relationships with elites which are. on the whole. likely to be of a positive character.
4. In both samples. the largest number of cases was lost due to missing data on income (33 cases in
1975. all cases in 1995).
5. Two notables were never intended to be included in analyses; the investigators knew them to be
relatively unknown. and included them on the list only for purposes of gauging a baseline probability of
chance acquaintance. One notable died during the collection of the data, and the other three were elderly
and relatively unknown to respondents (see Heinz et al. 1982. Ch. 9).
6. The principal investigators coded law schools into categories indicative of their relative prestige and
selectivity. For details. see Heinz et al. (1982).
7. Since work for businesses is generally more lucrative than work for personal clients. the
incomparability in measurement between the two periods probably results in an underestimation of the
increase in business clients as sources of lawyers' income.
8. Sample densities were calculated in the following manner:
Density = observed ties/possible ties
Observed ties = total number of nominations from rank and file to notables
Possible ties = (sample n)*(number of notables).
Thus. the sample density of 'advisor' ties in 1975 is 1923/(43)*(692) = 1923129756 = .065. where 692 is
the number of practicing lawyers (in all practice settings. including private practice. internal counsel and
government in employment) in the 1975 sample. The number 1923 is the number of advisor nominations
of notables made by rank and file practicing lawyers in the 1975 sample.
9. The dependent variables are in a common metric, logged 1995 dollars. In comparison of the metric
coefficients for each period, the test of statistically significant difference is as follows:
[Coefficient(75) - Coefficient(95)]/[std err(75)2 + std err(95)2]112, which is approximately distributed as
Student's t.
10. In previous analyses (Sandefur, Laumann, and Heinz 1997), we estimated similar regression
equations in which we modeled elite ties with a pair of variables indicating whether the respondent
nominated 1-3 notables or 4 or more notables. In these models, the effect of knowing 1-3 notables was
statistically significantly larger in 1995 than in 1975. In the present paper, it is likely that the coefficient
for 3 notables in 1995 fails to achieve statistical significance because of the small number of observations
in that category (29 respondents).
11. In fact, the advertising of lawyers' services became legal only in 1977. Lawyers in the earlier period
had to rely on referrals by existing clients, contacts made through membership in local voluntary
organizations, such as the Kiwanis or the Knights of Columbus, and the strength of their local reputations
as ways of attracting new business.
SECTION III
• Structure at the Individual Level

social capital and management
Generalized Exchange and Economic
Performance: Social Embeddedness of
•
Labor Contracts in a Corporate Law
Partnership 13
Emmanuel Lazega
ABSTRACT
This chapter examines the relationship between social structure and economic
performance at the intraorganizational level. It attempts to identify a few conditions
under which individual social relationships are most productive for the firm in
collegial organizations-where the production process is difficult to routinize, where
professional expertise and advice cannot easily be standardized, and therefore where
internal transaction costs for the firm as a whole can be assumed to be a large part of
total costs. An empirical study of a medium-sized northeastern u.s. corporate law
firm is used for that purpose. In this firm, attorneys are shown to be bound by a
labor contract that is difficult to sustain on pure economic terms: partners can easily
free-ride, and associates can threaten the quality of work. Against this damage
potential, a social system sustains their commitment. Using network data collected
in the firm, social capital is described and measured at the individual, workgroup,
and structural levels to show that the more constraining the member's coworkers
network, the easier it is for the firm to extract higher economic performance,
including from tenured partners, by controlling the time put into work. With regard
to partners, such teams represent an element of self-entrapment compensated by
status and professional recognition. Examples of low and high economic performers,
and their respective combinations of social resources, are provided as illustrations. A
locally multiplex generalized exchange system is then described as providing firm-
level social capital. Its existence is viewed as a precondition for individual-level and
group-level social structure to be productive because it maintains the circulation of
social resources in the firm. A multilevel form of embedded ness is thus revealed
here and shows the importance of taking into account a meso level when measuring
the relationship between social capital and performance. In this particular case, the
notion of multilevel embeddedness advances our understanding of economic

performance: the latter is rooted in individual social capital, that is itself rooted in
workgroup and firm social capital, which in tum helps individual members in being
more productive. Finally, however, this virtuous circle is shown to be fragile. First,
because group-level social capital also threatens the cohesiveness of the firm: well-
knitted teams can defect and take away with them valued members and clients.
Second, given the dependence of economic performance on firm social capital, the
relative contribution of individual ties and collective social structure to economic
performance is also a highly politicized issue.
INTRODUCTION: STRUCTURAL SOLUTIONS TO STRUCTURAL

PROBLEMS
Current managerial conceptions of corporate social capital often build on basic
tenets of transaction-cost economics. They contend that economic growth
increasingly depends not only on production costs but also on a critical factor, the
so-called transaction costs. Transaction costs arise when market situations produce
information asymmetry and therefore need visible signs for building confidence
between economic actors. In fact, such a perspective stems from a more general
tradition that focuses on social mechanisms supporting and enhancing economic
performance, beginning with Durkheim (1893) and by now strongly established
(Macaulay 1963; Bourdieu 1980; Coleman 1990; see Flap, Bulder, and VOlker 1998,
and Gabbay 1997 for a review). Following this general perspective, social networks
of ties within organizations are assumed to enhance performance because they
reduce transaction costs by decreasing information asymmetries and opportunism.
Thus, maximizing performance means not only improving technology, product and
organizational innovation, managerial coordination, or financial management but
also changing informal social restrictions, improving solidarity, and investing in
social capital. These means introduce a long-term time frame in exchanges and
make it possible for people to invest in a starting or an ongoing relationship. In this
view, large amounts of social relationships are a basis for efficient cooperation and
for solidarity within organizations.
This chapter examines the question of the relationship between performance and
social structure at the intra-organizational level. It attempts to identify a few
conditions under which individual social relationships are most productive for the
firm in collegial organizations (Waters 1989)-where the production process is
difficult to routinize, where professional expertise and advice cannot easily be
standardized, and therefore where internal transaction costs for the firm as a whole
can be assumed to be a large part of total costs. In this organization, social capital is
located at different levels: individual level, group level, and firm level. Individual
social capital is defined as the resources flowing through an individual's contacts. At
the workgroup level, social capital is a function of high density and cohesiveness in
strong work relationships. Firm social capital is provided here by a locally multiplex
and generalized exchange system-a pattern of ties among members that helps them
exchange various resources indirectly, supports cohesive work ties, and maintains a
specific form of solidarity.
Generalized Exchange and Economic Performance - 239
To address the issue of the relation between performance and social structure, I
look at the effects of these levels on one another. Following other scholars' more
general work (Burt 1992; Flap 1990; De Graaf and Flap 1988; Flap and De Graaf
1986; Lin 1982, 1995a; Lin and Durnin 1986; Lin, Ensel, and Vaughn 1981;
Campbell, Marsden, and Hurlbert 1986; Marsden and Hurlbert 1988), I do so by
reporting an empirical network study of a corporate law firm focusing on
intraorganizational performance data. This study shows that one of the conditions
under which individual social capital is most productive for the firm is precisely that
such a locally multiplex and generalized exchange system constrains some of its
members-partners and associates with specific relational patterns and workgroup
membership--in reaching higher economic performance (defined in terms of
number of hours worked and dollar amounts brought into the firm). In other words,
the members' labor contract (the partnership agreement for partners and the
employment contract for associates) is combined with other social ties-such as
strong collaboration, advice, and friendship, both at the dyadic level and the
structural level.
This approach makes particular sense in collegial organizations (Waters 1989,
Lazega 1998b). In effect, in the economic conditions of 1990, it can be assumed that
when performance of partners and associates is good, firm performance is also good.
However, in this firm, getting practically tenured partners to work well is sometimes
a problem: there are enormous incentives to free-ride. Getting associates to work
well is also a problem: although they are well paid, there is a very low chance for
them to become partners. If partners can free-ride and associates threaten the quality
of work, members have a labor contract that is difficult to sustain on a purely
economic basis. Because, as Durkheim (1893) pointed out, a contract is always
incomplete, members need the expectation that this contract will be fulfilled, and-
in case this exchange does not work-the flexibility that multiplexity provides to
enforce these contracts. Getting cooperation and keeping production going are also a
result of exchanging these resources in a mUltiplex way. Therefore, a generalized
exchange system can be assumed to relate the three levels of capital to one another
by maintaining the circulation of social resources in the firm. Several steps are taken
to show that this exchange system, a component of this firm's corporate social
capital, increases the efficiency of individual performance and collective action.
First, I show that a social system sustains members' commitment to their labor
contract: members who are strongly socially integrated perform well economically;
others perform less well. On the one hand, formal dimensions of structure, in
particular hierarchical status and seniority, have the greatest influence on economic
performance. In general, partners put in less hours but collect more dollars than
associates because they charge more; in this firm, the more senior attorneys are, the
higher their hourly fees. Associates collect less although they put in more time than
partners. On the other hand, however, the nature of the individual social network
also has an influence on economic performance: attorneys informally sought out for
advice and for collaboration by many others tend to bill and collect considerably
more than others. In addition, when social capital is proxied in terms of Burt's
(1982) constraint scores, results confirm that members (both partners and associates)
with a constraining coworkers network are pressured to put in more time, thus
collecting more dollars. The more constraining one's coworkers network, the higher
one's economic performance. Thus, position in relational structure and social capital
accumulated in this position do count for explaining performance, although these
effects are weaker when compared to the weight of institutionally defined hourly
rates.
Second, the effect of individual-level social capital on economic performance is
decomposed at the dyadic level by looking at specific combinations of ties that
provide a decisive push in performance increase (or that represent a liability that
decreases performance). For example, specific configurations of social ties, such as
mutual triplex ties, are strongly correlated with high performance. This fleshes out
the positive effect of constraint scores in the coworkers network. Examples of
partners with low and high economic performance (in terms of dollars brought in),
and their respective combinations of social resources, are provided as illustrations.
High performers draw heavily on their social resources. In that respect, the firm as a
whole benefits from the networks that some individuals have (dense multiplex
networks, especially constrained at the group level in the coworkers network) and
suffers from the networks that other individuals have (sparser networks and
especially weakly constrained in the coworkers network).
Third, strong and stimulating (that is, constrained) coworkers ties combined
with advice ties or with both advice and friendship ties have a chance of being
economically more productive than other ties and more likely to happen in dense
workgroups. At the structural level, a locally mUltiplex and generalized exchange
system encouraging the emergence of workgroups is then described as a
precondition for individual social network to be productive because it maintains the
circulation of social resources in the firm. This structure integrates work and social
ties in a way allowing strongly knitted positions to perform better by extracting and
facilitating higher efforts from their members. By doing so, it makes it possible to
improve performance for members whose work ties are embedded in a way offering
access to advice (which, in this firm, is a form of free collaboration), creating
economies of time, providing flexibility in exchanges by allowing a resource of one
type to be engaged for a resource of another type, and helping in foregoing
immediate self-gains for a smoother and longer-term collective action. This social
structure, to put it in Uzzi's (1997a) words, governs the intervening processes that
regulate performance outcomes, both positive and negative. This system is a feature
of the firm as a whole: it is a component of its corporate social capital. It can thus be
asserted that if economic performance is rooted in individual social capital, the latter
is itself rooted in collective social capital or firm social capital.
In sum, firm social capital, that requires many kinds of contributions, is also key
to maintaining collective action and production. It is made more visible in the
description of the firm as a locally multiplex generalized exchange system.
Generalized exchange is strong at the level of workgroups. In particular, cycles
characterizing local (that is, group-level) and multiplex generalized exchange can be
found in this system, especially in the flows of resources among teams within the
same office and specialty. This form of embeddedness is shown to relate three levels
of social structure and help enforce the labor contract between partners and
associates. It makes them more productive by relying on chains of mutual
obligations and debts. In particular, partners are especially well positioned to play on
resource dependencies to get associates' commitment to their labor contract (which
is not necessarily in their narrow and short-term self-interest).
I then argue that, in many ways, this exchange system is good for both the firm
and the individual. Individual social ties help the individual perform and this
network transforms itself into firm-level social capital because it produces an
exchange system that makes the firm as a whole more successful in billing and in
helping members maintain their commitment to their labor contracts and solidarity.
In other words, the firm has found in this exchange system a structural solution to
the structural problem of cooperation and commitment to the labor contract, a sort
of partial equilibrium in the circulation of resources needed to fulfill it. I By allowing
such a system to exist, this firm maintains certain forms of resource circulation, a
precondition for group solidarity.2
However, this virtuous circle is also fragile. First, group-level social ties also
threaten the cohesiveness of the firm: well-knitted teams can defect and take away
with them valued members and clients (Lazega 1992a, 1999). Second, the relative
contribution of individual ties and collective social structure to economic
performance can be a highly political issue. Allocation of credit in teams (that is,
clearly disentangling members' contributions) is often difficult. This study therefore
confirms that looking at how an organization, especially a collegial one, extracts
economic performance from its members requires a sociological conception of
performance that must pay attention to the micropolitical context of members'
action and to their strategic behavior in this context (Weber 1978 edition; Coleman
1990; Granovetter 1985; Hechter 1987; Lindenberg 1990, 1996; Raub and Weesie
1990; White 1981). This approach therefore entails a micropolitical conception of
economic performance. Narrow conceptions of performance ignore the fact that-as
Crozier and Friedberg (1977), Friedberg (1993), or Meyer (1994) put it-no
measurement of performance in organizations ever goes unchallenged. To some
extent, criteria used to measure efficiency of actors are negotiated by members
themselves. This negotiation means that measurements of efficiency are strategic
and politicized. In effect, they are always multidimensional; research in this area is
complex, and multivariate approaches rarely conclusive. Practitioners know that it is
impossible to find simple measures of performance for organizations with mUltiple
and often conflicting goals. Meyer (1994), for example, shows that performance
measures can be considered to be temporary constraints to which members of the
organization adjust. It is thus impossible to define absolute measurements of
performance, outside of a strategic context or institutional conventions. In the
context of the law firm examined here, team work, autonomy and flexibility in
selection of coworkers, and a weak hierarchy unable to force partners to cooperate
all make it harder to provide one best measurement of performance by individual
employees and workteams. This is why a study of the relationship between
generalized exchange and economic performance necessarily leads to issues of
fairness and to an examination of the policitized nature of performance
measurements.
In conclusion, the forms of embeddedness discussed in this chapter help
members perform and reach a form of solidarity, but do not necessarily produce, by
themselves, a self-sustaining social order. The structural solution provided by this

exchange system raises new problems related to the balance of power within the
fIrm (Lazega 1992a, 1999): it is a politicized issue, all the more easily politicized,
because members can make their relative contribution to fIrm performance
unmeasurable-for example by bringing more resources into the exchanges-which
is made possible precisely by the existence of a multiplex and generalized exchange
system. This topic of politicization of performance measurements, however, is
beyond the scope of this chapter.
PROFESSIONAL LABOR CONTRACTS IN A CORPORATE LAW

PARTNERSHIP
Fieldwork was conducted in a northeastern corporate law fIrm (71 lawyers in three
offIces located in three different cities, comprising 36 partners and 35 associates), in
1991. All the lawyers in the fIrm were interviewed. In Nelson's (1988) terminology,
this fIrm has characteristics of both a traditional and a bureaucratic law fIrm. 3 It is
closer to the traditional type, especially because the tensions between bureaucracy
and partners' participation in management are open and explicit and because it does
not have formally defIned departments.
The fIrm services mainly corporations and other organizations prepared to bear
its fees. Interdependence among attorneys working together on a fIle may be strong
for a few weeks and then weak for months. As a client-oriented, knowledge-
intensive organization, it tries to protect its human capital and social resources, such
as its network of clients, through the usual policies of commingling partners' assets
(clients, experience, innovations) (Gilson and Mnookin 1985) and the maintenance
of an ideology of collegiality. Informal networks of collaboration, advice, and
friendship (socializing outside), are key to the integration of the fIrm, that has to deal
with many centrifugal forces created by differences in status, geographical location,
and specialties (Lazega 1992a, 1992b, 1999).
The partnership agreement applies to the various aspects of a fIrm's life the
prevailing fIrm philosophy regarding legal practice and how it should be undertaken.
In doing so, it brings an element of predictability to fIrm operations and minimizes
the room for disputes regarding issues such as fIrm management, compensation
decisions, and withdrawal terms. The agreement accomplishes this by setting clear
ground rules as to each partner's rights and responsibilities in connection with these
issues and with the operation of the fIrm itself (Rowley and Rowley 1916 and 1960).
The law fIrm is a relatively decentralized organization that grew out of a
merger, but it does not have formal and acknowledged distinctions between profIt
centers. Although not departmentalized, the fIrm breaks down into two general areas
of practice: the corporate litigation area (half the lawyers of the fIrm) and the more
strictly speaking corporate area (anything other than corporate litigation). Work is
supposed to be channelled to associates through a specifIc committee of partners,
but this rule is only partly respected. Sharing work and cross-selling among partners
is done mostly on an informal basis. Given the classical stratifIcation of such fIrms,
work is supposed to be channelled to associates through specifIc partners, but this
rule is only partly respected.
A weak administration provides information but does not have many formal
rules to enforce. The firm has an executive committee made of a managing partner
and two deputy managing partners who are elected each year, renewable once,
among partners prepared to perform administrative tasks and temporarily transfer
some of their clients to other partners. This structure was adopted during the 1980s
for more efficient day-to-day management and decision making. The current
managing partner is not a rainmaker who brings in important clients and does not
concentrate strong powers in his hands. He is a day-to-day manager who makes
recommendations to functional standing committees (finance, associate, marketing,
recruitment, and so on) and to the partnership as a whole during partnership
meetings.
This specific law firm is very much a stratified organization, in spite of a set of
rules that tries to smooth the hierarchical nature of its business. Consistent with
Nelson's (1988) more general terminology, its authority system is based on a
distinction between finders, minders, and grinders. With a few exceptions, the
finders are partners who find new and lucrative clients and bear the greatest
responsibility for them. Their governing authority is not as formal as that of their
analogues in corporations. Directives are reached by a form of gentlemen's
agreement. The minders are partners with managerial roles and responsibility for
long-established clients. The managerial role in large firms arises from the necessity
of coordinating diverse practice areas, promoting an efficient organization of work,
and decentralizing control over a large professional staff working on highly
specialized matters. The grinders are other lawyers~ither partners who function as
little more than salaried staff or associates-who are subject to the demands of
partners and perform the actual legal work.
Partners' compensation is based exclusively on a seniority lockstep system
without any direct link between contribution and returns. The firm goes to great
lengths-when selecting associates to become partners-to take as few risks as
possible that they will not contribute enough to firm's revenue or pull their weight.
Partners may argue informally about what contribution might fairly match one's
benefits, but the seniority system mechanically distributes the benefits to each once
a year. Great managerial resources are devoted to measurement of each partner's
performance (time sheets, billing, collecting, expenses, and so on), and this
information is available to the whole partnership.
A low performance cannot be hidden for long. However, such firms usually
make considerable profits, which may help partners overlook the fact that some
voluntary contributions to shared benefits may not always be consistent with the
successful pursuit of long term self-interest:
Our compensation system has no built-in peer review process. There is no committee
meeting with each partner, no interview devoted to pulling out from that individual his or
her state of affairs. The peer review that we have right now is everyone sits down in the
partners' meeting, and you have in front of you the printout that shows how many hours I
worked, how many hours I billed, how many hours I collected, and how outstanding my
account receivable is, and then you get people grumbling at the meeting about the
account receivable going up and not coming down. (The managing partner at the time of
the study)
The firm does not have a formal peer review system that could provide intermediate
steps between informal control and formal court procedures:
With the compensation system there is no built-in financial incentive for people to do
things. If you have people who are motivated by other things, like self-respect, pride in
craftsmanship, intellectual curiosity, competitiveness-whatever those different personal
attributes ar~hat's not a problem. There are people who aren't as motivated by those
other things as certain other people and may wind up resting on their laurels, sitting on
their hands, whatever euphemism you want to come up with for becoming lazy both
intellectually and how much they are willing to work. (The managing partner at the time
of the study)
Before expulsion, partners have the power to punish each other seriously by
preventing one of their own from reaching the next seniority level in the
compensation system. As mentioned above, a partner can be expelled only if there is
near-unanimity against him or her. Buying out a partner is very difficult and costly:
The rule is 90 percent affirmative votes of all partners to expell a partner. Abstention
counts against expulsion. We have expelled only two partners in twenty years. Both were
extremely serious situations. For one of them, there was a unanimous vote: this person
went way around the bend with a number of things. The second guy didn't want to be
part of us. A petition was circulated, saying 'I would vote to expell him.' He was
presented with that and resigned. But expUlsion is extremely hard to exercise. Three
persons can block it. At least two of our partners are good examples of that. One of them
is a decent guy on a personal level. He says that under pressure he does things but that he
is a corporate lawyer, and there is no pressure on him usually, so he just doesn't do it.
The other partner is in a different situation. He was at the low end (of the performance
scale) for a long time. Partner 5 went to talk with him. He claims that there isn't anything
to do. He says: 'I am a corporate lawyer, my kind of work has dried up, 1 am out there in
the bushes, hustling, doing everything 1 can.' That is difficult to check. But there is also
the fact that part of the people he has worked with do not want to do it again because
they think his competence is in doubt. Partner 17 is extremely good in his field. He once
volunteered to go see this partner to share some work with him. He went to see him, but
he says this partner did a horrendously poor job. So with our 90 percent rule, we don't
cover for that. Apart from these examples, all the other partners do carry their weight. Of
course, you're hot this year, you'll be down next year. But there is no need for a
compensation committee that would just do what it wants to do, with all the subjectivity
involved. (The managing partner at the time of the study)
It should also be mentioned that, since partners cannot be forced out unless there
is near-unanimity against them, most partners in this firm manage to have at least
one safety partner-a friend who would presumably side with them unconditionally
and become their insurance policy against this consequence (Lazega 1992a, 1995b;
Lazega and Lebeaux 1995; Lazega and Krackhardt 1998). Therefore, despite the
existence of direct financial controls, the firm does not have many formal ways of
dealing with free-loading. The harm that a single partner can inflict on others might
become very substantial in the long run. Conversely, partners can try to isolate one
of their own informally by, at the very least, not referring clients, not lending
associates, and not providing information and advice. This strengthens the
suggestion that performance depends on the social circulation of resources in the
firm.
A few general indications about the relational climate and structural tendencies
in this firm are needed here to understand the way in which the notion of
embeddedness of labor contracts is used below. The climate and tendencies can be
summarized by two separable forms of interdependence describing the interlocking
of social relations such as being strong coworkers, advice relations, and socializing
outside work. One of these forms concerns the interplay of friendship and advice,
wherein an individual's friends may be a source of further friendship and advice ties
(the friends of friends are often friends, and the advisors of one's friends are often
friends or advisors, or possibly both). Advice ties, that will be shown to be key to
economic performance, are thus often driven by personalized ties. The second form
concerns the interdependence of advice and coworker ties, with the possibility that
advice ties play an important role in generating further advice ties and, in
conjunction with coworker ties, further coworker ties. Finally there is a weaker
association between coworker and friendship ties in that the configurations
comprising two friendship and one coworker tie are unlikely. There is a taboo with
strongly personalizing work ties, particularly for partners with regard to associates
who might leave the firm before coming up for partnership, as well as with
associates who will eventually come up for partnership. Partners often feel that such
relationships would tie their hands on the day of the vote. Associates know that they
are being kept at arm's-length and express it more openly. Associates 58 and 51
express this in the following ways:
I am a big believer in keeping your job and personal life separate. You have to have your
job to fall back on if you have a personal problem in your life. I would respect that
privacy. I don't think friendship could matter here. I think the partners share their
personal life with other partners, and associates with associates. But I don't believe that
partners and associates share their personal life. It is probably also the associates' fault.
Their goal is to make partner, and you don't want to show your vulnerability. You think
it will affect your chances to become a partner--especially during the first four years,
when you are not secure about your job, and your knowledge is limited. (Associate 58)
An interesting change in relationships comes when some people who were senior
associates two years ago are now partners, and we don't spend as much time together
outside work any longer. Since they will have to make a decision concerning us, we
don' t socialize anymore. But I still consider them to be friends because I feel close to
them. (Associate 51)
Associates also feel that partners don' t let them in more generally:
Between partners and associates, there is a gulf-regardless of how friendly we are. As
an associate, I am left in doubt. I know that some friends I have were not made partners.
There have been people who have gone all the way to the eight years and were not made
partners. I will feel less constrained to talk about finn matters with other associates.
There are a few partners, however, to whom I would talk about sensitive finn matters. I
would actually never bring up the subject myself. But when they bring up the subject, I
listen. It is difficult to realize how associates grapple and work just to be able to
understand where they are, the kind of finn they really are in. (Associate 48)
These networks are thus partly dependent on each other, even though they each do
have a life of their own, mainly because they each solve different problems of
cooperation (Lazega 1992a; Lazega and Pattison 1998).
WHAT MAKES TENURED PARTNERS WORK?

In any organization, measurements of performance are intrinsically difficult to
interpret and their informative value can change from one year to another. 4 As seen
above, performance data are never as hard and undisputable as one often expects
them to be. In fact managers know and learn that such data must be handled with
great care. A narrow conception of organizational efficiency ignores the fact that no
measurement of an actor's performance goes unchallenged within the organization.
Therefore, using performance measurements is not easy and rarely provides
spectacular results. In spite of correlation between some forms of human capital (for
example, years with the firm) and performance, there is no easy and clear-cut
correlation between performance and social networks.
Members of this firm know that measuring economic performance describes
only one aspect of contribution to collective action. This is why managing partners
often emphasize spreading credit around generously. Many factors account for
members' individual performance. These factors can be external or environmental
(some areas of practice provide more work, some markets are currently more
lucrative) and individual (some attorneys are personally more motivated or hard
working). For these reasons, the following analyses link information on members'
economic performance-narrowly understood as the amounts of dollars brought into
the firm at the end of the year-with information on social status and relations
among them. Differences in such performance may be explained, in part, in terms of
relationships within the firm-for instance, because relational factors can help gain
access to needed resources, reduce transaction costs with coworkers, or help
pressure colleagues back into more productive behavior. To examine such effects on
performance, I used information collected in early 1991 about each attorney's
(partner and associate) relationships within the firm and combined it with
information on their individual performance for the year before the study was
conducted.
Variables and Analysis

Specifically, to study the effect of position in firm structure on this type of economic
performance, I look at the correlations between measurement of economic
performance and various social factors related to firm structure, work process, and
members' ties in 1991. I first use as covariates three dimensions of formal structure
of this firm that were expected to be the most important (status, office, specialty), as
well as two attributes of members defined from outside the firm (gender and law
school attended). Table 1 presents the distribution of lawyers in this firm per
variable.
Table 1. Distribution of lawyers per variable
Status Partner Associate Total

Seniority
Levell 14 7
Level 2 13 10
Level 3 9 5
Level 4 7
Level 5 6
Office
I (Boston) 22 26 48
II (Hartford) 13 6 19
III (Providence) 3 4
Specialty
Litigation 20 21 41
Corporate 16 14 30
Gender
Man 33 20 53
Woman 3 15 18
Law school
Ivy League 12 3 15
New-England non-Ivy League 11 17 28
Other 13 15 28
Total 36 35 71
The first covariate is status, a variable with two levels-partners and associates.
We can hypothesize that status matters for economic performance in the sense that
firm rules require associates to put in more time than partners. This variable is
elaborated on in the second covariate-seniority. We can hypothesize that seniority
matters for economic performance in the sense that the more senior members are,
the higher the hourly rates systematically charged to clients. This second covariate is
a variable with eight levels, indicating the three possible levels of seniority for a
partner,S and five levels of seniority of associates. For associates, seniority has the
meaning of being member of a cohort recruited the same year. We can thus look at
gradual effects of numerical rank on economic performance. Office membership and
practice are the third and fourth covariates. Office is a variable with three levels-
Office I, II, and III; practice has two levels-litigation and corporate. They are
expected to have an effect on economic performance as indicators of variations in
market demand. The next covariates are other actors' attributes-gender and
lawschool attended. These attributes are included as control variables representing
two characteristics of the outside world that could have an influence on economic
performance. In this firm, women attorneys are mostly associates and often feel that
they need to work harder than their male colleagues to reach the same economic
results-for example, because they mostly have to deal with male clients or partners.
Law school attended, a typical human capital variable, has three levels, indicating
whether a lawyer went to an Ivy League law school, to a New England non-Ivy
League law school, or to another law school. This variable is introduced in the
model to look at the extent to which a form of prestige acquired outside the firm
may have an effect on performance-for example, via more lucrative clients.
To locate members in the informal structure of the firm and proxy their
individual social capital, I use six variables based on standard sociometric
information on three types of relations collected in this firm in 1991: coworkers,
advice, and friendship (for name generators, see Appendix A). From these data, I
derived proxies of individual social capital by looking at two types of
measurements. The first was their individual indegree centrality scores in these
networks. 6 Indegree centrality represents a measurement of the extent to which
members are popular in these networks and therefore accumulate work-related
resources circulating in them (Wasserman and Faust 1994: 169-219). One can
therefore hypothesize that they will be in a better position to perform economically.
The second was their individual constraint scores as defined by Burt (1992) in
the same networks. For Burt (1992), network constraint measures social capital as a
form of network structure. Specifically, constraint is a function of network size,
density, and hierarchy (that measures the extent to which relations are directly or
indirectly concentrated in a single contact). A contact in which relations are
concentrated is a knot in the network, making it difficult for negotiations to proceed
independently in separate relationships. Constrained networks leave little
opportunity for individual initiative and little chance to withdraw from difficult
relationships. Difficult relations persist because they are interlocked with coope-
rative relations. The higher the constraint, the fewer opportunities for alternatives
offered by one's contacts or contacts' contacts, and the lower the performance. In
our case, constraint represents a measurement of the extent to which colleagues can
exercise unobtrusive but insistant pressure on a member. High constraint in a
specific network means that clique members in that network have high investments
in each other and high expectations from each other. The denser a member' s
personal network of coworkers, the more his coworkers can coordinate their
informal efforts at prodding him or her back into performing more. They can, for
instance, try to increase their own collaborations with him or her, and exercise
unobtrusive but insistant pressure to put in more time. One can therefore make the
hypothesis that, in the case of this firm, high constraint facilitates high economic
performance. 7
In other words, in this type of collegial organization, a constraining network of
strong ties, in which members overinvest, tends to create teams of partners and
associates who rely on each other, at least for work. With regard to partners, such
teams represent an element of self-entrapment compensated by status recognition:
teams allow partners to create emulation and redefine what is an acceptable
performance within their workgroup, including themselves. But coordination is also
made easier in such workgroups: they bring associates together productively. The
following is an example of how partners in this firm typically like to talk about their
associates and the kind of intellectually challenging attitude they encourage within
their teams:
Ours is a fascinating structure built on, to some extent, maximizing a certain type of
efficiency. All are encouraged to think hard. You look good as an associate if you can
convince a partner that he is wrong about something. You have the freedom of thought
within legal problems. There is a great intellectual freedom here. An associate yesterday
told me that she didn't think that a decision of mine in a file was correct. She stuck to her
guns, and fifteen minutes later I called her to tell her that she was right. In other places, if
the boss says something, everone says 'good idea, boss.' Not here. (Partner 19)
Thus proxies of social capital add a set of covariates to the baseline model. The
effect of centrality and of constraining ties on economic performance is expected to
be positive.
Using these covariates, several models were estimated to explain economic
performance measured as the amount of dollar fees brought to the firm (managing
partner not included) in 1990. It is important to realize that not all the covariates
representing various dimensions of position in firm structure can be used at the same
time because of strong dependency between them. This is typically the case for
status and seniority; in the models, the most refined covariate, seniority, is used. In
addition, status and seniority overlap with the number of hours worked and hourly
rates as explanatory variables. The more senior attorneys charge more per hour.
Associates work longer hours than partners.s Therefore, to avoid this problem,
analyses below test the robustness of social capital effects using three different
models. This multicollinearity will be taken into account in the interpretation of
results. In terms of economic and relational variables, the best overall models
achievable with this data set predicting the number of hours worked and the amount
of dollar fees brought in are presented in Table 2.
Coworkers' Constraints and Economic Performance

As predicted, model 1 contains an important effect confirming our expectations. It
shows that, in this firm, constraint by teams of coworkers has a positive and
statistically robust effect on the number of hours that members put in. The density of
members' network of coworkers helps the latter control and increase the amounts of
effort invested into hard work. In that respect, the effect of members' social network
on their performance is confirmed. However, beyond this confirmation, an
additional and unexpected dimension of social structure also emerges here.
Interesting negative effects of dense friendship ties (and centrality in the friendship
network) on performance are apparent. One can speculate about the reasons for
these effects, such as spending more time on personalizing work relationships, or
providing role-distance, outside work than on actual work. 9 But this shows that
individual and group-level social ties can also take less productive forms-forms
that decrease performance. In terms of economic outcome, structure that brings
social capital to the individual may thus bring social liability to the firm. Social
capital can be a double-edged phenomenon.
Focusing back on the positive effect of constraint in one's network of strong
coworkers, models 2 and 3-where the weight of economic variables such as hourly
rates is more visible-show that such an effect is robust. Indeed, it is also present as
a good predictor of the amount of fees brought into the firm at the end of the year. It
is thus relatively independent of the way performance is measured.
Table 2. Variables explaining economic performance measured by the Number of

hours worked (Model 1) and by the amount of dollars brought into the firm in fees in
1990 (Models 2 and 3) in 1990
Effects Standardized Estimates

Modell Model 2 Model 3
Seniority 0.01 0.76***
Hourly rates 0.78*·*
Time input' 0.40***
Office 0.24** O.IS* O.OS
Specialty -0.16* 0.01 0.07
Gender -0.03 0.00 0.02
Lawschoolattended -0.14 -0.03 0.02
Centrality
Friendship -0.27* -0.11 -0.01
Coworker 0.17 0.01 -0.02
Advice -0.02 0.27* 0.23*
Constraint
Friendship -0.81**· -O.IS O.ll
Coworker 0.23* 0.16* 0.13*
Advice -0.04 O.OS 0.04
*p<O.OS, **p<O.OI, **·p<O.OOI. Statistical significance levels are purely indicative,
since observations were collected for the population, not for a sample. Adjusted R-
squared are 0.66, 0.86, and 0.89, respectively (and 0.63, 0.84, and 0.87, respectively,
without the coworker constraint variable). The managing partner, who concentrates on
firm policy and administrative work and is not a timekeeper during his tenure was not
included in the computations of these parameter estimates.
'. Including the interaction effect of time input and hourly rates does not provide
additional insights here because senior partners who charge high rates are not among the
members who put in the greatest number of hours.
In model 2, status and seniority count. Partners do collect more money than
associates, and the more senior lawyers are, the more money they tend to collect. In
general, partners put in fewer hours but collect more dollars than associates because
they charge more. Associates do collect less money although they put in more time
than partners. The more senior associates are, the more hours they bill and the more
they collect; senior partners charge more per hour, put in less time, and collect more
money. Market conditions have a mixed effect on economic performance: in terms
of office membership, Hartford attorneys put in more time than Boston attorneys,
and collect more; but in terms of division of work, no specialty (corporate or
litigation) seems to be systematically more lucrative than the other-in spite of a
general advantage for litigation in the early 1990s in U.S. corporate law firms.
Given our interest in the effect of informal relationships on performance, it
makes sense to look into the latter effects in more detail for associates and for
partners separately. The following results are then obtained. For partners only,
effects considered here for hours billed are at best unstable. Practice has a very weak
effect: litigation partners bill slightly more than corporate partners. Again, partners
popular as friends tend to bill slightly less (as for associates mentioned below). For
associates only, centrality in terms of friendship affects collection negatively. This

means that associates very active socially and providing on average more emotional
support or role-distance to others may end up being handicapped in terms of
collecting. Obviously, the link between the two phenomena is indirect and remains
to be explained, but it is nevertheless persistent. The more central associates are in
terms of their number of coworkers (that is, the more colleagues they work with),
the more hours they bill. The only associates who collect more hours are those who
are also central in the coworkers network (they tend to be senior associates). For
associates, seniority is the best predictor of performance in terms of hours collected,
but seniority and centrality as a coworker are the best predictors of performance in
terms of hours billed.
In this context, the analysis of the effect of individual social ties on economic
performance also shows that attorneys informally sought out for advice and for
collaboration by many others tend to bill and collect more than others. This extra
effect is partly due to the fact that senior partners are central advisors. As already
mentioned, the effect of social capital measured in terms of Burt's constraint scores
is still present in model 3, in which seniority is replaced by hourly rates and time
input. This confirms our expectations: members with a constraining coworkers'
network put in more time, bill more hours, and collect more dollars. The more
constraining one's coworkers' network, the higher one's economic performance.
Thus, overall, multivariate analyses show that the effect of position in informal
structure and social capital accumulated in this position do count for explaining
performance, but these effects are weaker than the weight of seniority or hourly rates
as defined by the institutional setting. to The next sections provide a more detailed
illustration of such effects.
STATUS, SOCIAL CAPITAL, AND ECONOMIC PERFORMANCE: AN

ILLUSTRA TION
As suggested, in particular, by the effect of constraint in the coworkers' network,
economic performance is indeed rooted in the structure of individual social ties in
the firm. This raises the question of the relative weight of specific combinations of
ties in providing a decisive push in performance increase (or representing a liability
that decreases performance). In this section, I look at similar effects from a different
angle by showing that a specific configuration of social ties is required, at the dyadic
level, for high economic performance in this firm. This is due to the fact that
maintaining production (collective action) requires many kinds of contributions and
resources. Describing this configuration is equivalent to understanding how social
resources combine with one another to increase or decrease individual performance.
Results obtained above are translated into a description of specific configurations of
ties that are associated with economic performance. These associations can be
measured by the correlations between the frequency of more or less multiplex
combinations of dyadic ties for each actor and his/her performance measurements.
~
~ N
;-
(M
~
-tio
Table 3. Correlation table for performance measurements in 1990 and most frequent types of combination of ties among dyads of actors a
i:l
Hourly rates Time Input Time Input Fees Billed Fees Billed Fees Collected Fees Collected ~
Dollars Dollars Hours Dollars Hours Dollars Hours ..,~

No ties -.34" -.31" -.08 -.31" -.15 -.31" -.17 5'
iFj andjFi -.36" -.20 .11 -.26' .05 -.26' .01

-
~
~
[
JAi .47'" .36" .11 .42'" .19 .41'" .20 I::>
;:s
I::>..
t--o
jCi andjAi .36" .15 -.08 .21 -.01 .20 .01 5'
[
iCj andjCi andjAi (i.e. 'Blau tie') .54'" .10 .43'" .18 .45*** .24' ~.
.44'"
iCj and jCi and iAj -.52'" -.34" .04 -.40'" -.05 -.39'" -.04
iCj and jCi and iAj and jAi .22 .28' .10 .24' .10 .25' .16
iCj and jCi and iAj and jAi and iFj and jFi .39'" .43'" .19 .46'" .25' .48'" .30'
a Only combinations of ties with a significant correlation were kept in the table. (Statistical significance levels are purely indicative, since observations were
collected for the population, not for a sample).
·p<o.OS, ·'p<O.OI, "·p<O.OOl. 'F' refers to Friendship, ' C' to Coworker, 'A' to Advice. 'iFj' represents a friendship tie from i to j . ' iFj andjFi' represents a
reciprocated friendship choice. For example, having many ties of the 'jAi' type --which means that many people G) seek out your advice-- is strongly,
statistically significantly and positively correlated (0.41"') with collecting high fees measured in dollar amounts.
Configurations of Dyadic Ties and Economic Performance

As mentioned in the presentation of the firm, many combinations of ties are possible
at the dyadic level in three networks. The most frequent type of combination
(besides the 'no tie' possibility) between two persons in the firm is a mutual cowork
tie with one unreciprocated advice tie (from j to i)-I will call this compound a
Blau-tie in reference to Peter Blau's (1964) work on exchange of advice for status.
Many partners have such ties with themselves as advice distributors and associates
as advice seekers. In contrast, there are no duplex mutual cowork and friendship ties.
The complete absence of such ties is interesting. As seen above, although friendship
and work can go hand in hand when an advice component is also present, the two
resources seem to have distinctly different natures.
Table 3 presents correlations between specific configurations of dyadic ties for
each individual in the firm and several measurements of performance. This
exploratory analysis provides several results. First, there is a negative correlation
between having many 'no ties' (or many 'empty' ties-potential ties not actually
realized) and putting time in, billing, and collecting. The less one exchanges
resources with others, the lower one's performance. Second, there is a positive
correlation between having many ties such as 'being sought out (unreciprocated) for
advice exclusively' (no cowork or friendship component) and dollar amounts
invested (time worked by hourly rate), billed (time billed by hourly rate), and
collected (fees actually collected). The same correlation holds when the tie includes
an unreciprocated friendship component. This is partly related to the fact that senior
partners, who are often sought out for advice and cited as friends (without
reciprocating), bill and collect more given their higher hourly rates. Third, and more
suprisingly, correlations between having many ties that are exclusively reciprocated
work ties and any performance index are all really small. 'Work only' relationships
are quite neutral in terms of their association with economic performance. Such ties
are more frequent for associates than for partners. However, adding one component
to this combination changes this result. Given that partners work with associates
more than with other partners, there are positive correlations between having many
such ties with the added component 'being sought out for advice' (unreciprocated)
and performance indexes. Such Blau-ties are much more frequent for partners than
for associates. In my view, this confirms that self-entrapment by partners in teams of
coworkers is compensated by Blau-type status and professional recognition. When
adding friendship components to such ties (such as citing j as a friend or citing each
other as friends), positive correlations become stronger, although there are less
occurrences of that type. 11 Fourth, and conversely, there are mostly negative
correlations between having ties with components such as 'having a reciprocated
work relationship and seeking out advice (unreciprocated)' and dollar amounts put
in, billed, and collected. Such ties are more frequent for associates (particularly
nonsenior associates). Finally, there are strong and positive correlations between
having individual triplex mutual reciprocated ties and dollar amounts put in, billed,
and collected. Exchanges of that type are much more frequent for partners. In short,
the results translate our previous statements into relational terms. Partners' higher
economic performance, for example, shows in the correlation between having many
Blau-ties and the amounts of dollars collected.
To illustrate this analysis, the data provide examples of low and high economic
performers and their specific combinations of social resources. In these examples,
each attorney has a relatively different profile, but important common characteristics
are related to low or high performance.
Low and High Performers: The Importance of Blao-ties

Among low performers (still in terms of dollars brought in in 1990), one partner has,
in terms of compounds of ties, a higher than average proportion of persons that he
cites as strong coworkers who do not reciprocate, and he has a lower than average
proportion of Blau-ties, and of triplex reciprocated ties. He considers many people to
be his friends, but they do not reciprocate. His partners do not listen much to his
opinions about firm management and policy issues. In terms of his configurations of
dyadic ties, this partner-who is the least productive or performant one-has the
following combinations. Four types of ties in which he says that he is friends with
others who do not reciprocate, four types of ties in which he is sought out for
collaboration and that he does not reciprocate, four types of ties in which he is
sought out for advice (lower number than the average partner), and five types of ties
in which he seeks out others for collaboration but is not cited by them. The strong
specificity here is also the absence of direct reciprocity. There are many relations
with a transfer of resources, but few relations where there is direct and multiplex
exchange. Notice that none of the configurations associated with strong economic
performance is present in this partner's profile. Recall the managing partner's
comment about this partner:
[Jack] is a decent guy on a personal level. He says that under pressure he does things but
that he is a corporate lawyer, and there is no pressure on him usually, so he just doesn't
do it.
A roughly similar pattern characterizes a second low-performing partner. He has
a much higher than average proportion of 'no ties' and a higher than average
proportion of persons that he cites as strong coworkers who do not reciprocate; he
also has a lower than average proportion of Blau-ties, and of triplex reciprocated
ties. He nevertheless has an average number of reciprocated friendship ties. Most of
his partners do not listen much to his opinions about firm management and policy
issues. Very few come to him for advice, and he has a relatively low number of
coworkers. At the dyadic level, the specificity of this partner's relational profile is
the contrast between his very few ties with others in the firm and two very strong
(triplex reciprocated) ties with unconditional partners (his insurance policy against
expulsion). He also claims three friendships with other partners, but they are
unreciprocated. Recall the managing partner:
[Frank] was at the low end (of the performance scale) for a long time. Partner 5 went to
talk with him. He claims that there isn't anything to do. He says: 'I am a corporate
lawyer. My kind of work has dried up, I am out there in the bushes, hustling, doing
everything I can.' That is difficult to check. But there is also the fact that part of the
people he has worked with do not want to do it again because they think his competence
is in doubt. Partner 17 is extremely good in his field. He once volunteered to go see this
partner to share some work with him. He went to see him, but he says [Frank] did a
horrendously poor job.
In sum, the two low performers are less involved than others in exchanges of
production-related social resources in the firm. They do not create teams of
associates who can rely on them for work and provide status recognition.
Among high-performing partners, there are broadly speaking two of types of
relational profiles. The first type includes some of the most senior partners,
especially given their high hourly rates. Their ties often seem to combine
unreciprocated transfers toward j more than exchanges with j. They are more often
sought out than seeking someone out-providers rather than beneficiaries. This
makes economic sense, since they are in charge of running the organization that
ultimately produces their larger share of the pie. But it also makes sense socially.
For example, Partner 1 is highly sought out for advice. He cites few coworkers
because he mainly worked, during previous year with a junior partner acting as a
'foreman' (Partner 26) concentrating coworkers' exchanges. Partners 2 and 4 have
the same relational profile, except for their more frequent mutual involvement in
mutual coworkers' ties and a higher than average proportion of Blau-ties. In addition
to the work-oriented components, Partner 4 has an exceptional five triplex mutual
ties with other partners. Thus, in general, unless they work with a 'foreman,' senior
high performers are strongly work-oriented persons intensely involved in
coworkers' ties and highly sought out for advice. Their compounds do not include
many friendship components, except with a very few select other senior partners
with whom reciprocation is taken for granted. 12 The taboo concerning friendship
with associates seems to be extended to young partners with a few exceptions for
contemporaries.
The second type of high-performing partner also includes work-oriented persons
with an above-average proportion of Blau-ties. They do, however, diversify their
exchanges of resources-their types of ties-much more than their senior partners.
Their compounds are more personalized and include more friendship components. \3
Typical of this relational profile is Partner 26. He is hard working, heavily involved
in the business of the firm, one of the persons with the most ties in the firm
altogether, among the record holders for the Blau-ties and for the triplex mutual ties
(with his Boston contemporaries who-for reasons linked to the history of the
firm-form a very cohesive group of partners). But in addition to many mono-
resource uniplex ties (such as being often sought out for advice and only for advice,
although not as much as Partners 1 and 6), he is involved in many duplex or triplex
ties with colleagues who work with him but also like him (often unreciprocated) and
come to him for advice (unreciprocated). He is thus more relaxed about
personalizing work ties and shows more social openness (than senior partners) to
colleagues working with him. He is a work-oriented all-around exchanger and
investor of resources.
The contrast between low and high performers' relational profiles shows that
high performers (in terms of dollars brought in) draw heavily on their social
resources and have common specific relational characteristics, in particular those
involving them-not surprisingly-in task-related exchanges with associates. In that
respect, the firm benefits from the networks that some individuals have (dense
multiplex networks, especially constrained at the group level in the coworkers
network and rich in Blau-ties) and suffers from the networks that other individuals
have (sparser networks and especially wealdy constrained in the coworkers'

network).
However, as suggested by the effect of constraint scores in the coworkers'
network and by the illustrations above, such configurations of ties (described at the
dyadic level) favoring high economic performance are not distributed randomly in
the firm. Their distribution depends on a wider pattern of social ties, a pattern that
represents a multiplex generalized exchange system. Recall that, as described above,
the structural tendencies in this firm can be summarized by two separable forms of
interdependence: first, the interplay of friendship and advice, wherein an
individual's friends may be a source of advice ties; second, the interplay of advice
and coworker ties, with the possibility that advice ties play an important role in
generating further advice ties and coworker ties. Thus mutual strong work ties
occuring in conjunction with either unidirectional (partner to associate Blau-tie) or
mutual advice ties-which are positively correlated with economic performance in
Table 3-have a higher chance of occurring in workgroups with constrained
coworkers ties. In that respect, productive coworker ties are embedded in advice
ties. In tum, advice ties are often driven by more personalized ties. While there is a
very weak association between duplex coworker and friendship ties, there is a strong
one between advice and friendship ties. This configuration also has a higher chance
of occurring in cohesive workgroups, and it is especially consistent with the
strongest positive correlation in Table 3 between triplex reciprocated ties and high
economic performance. Thus strong and stimulating (that is, constrained) coworkers
ties combined with advice ties or with both advice and friendship ties have a chance
of being economically more productive than other ties and more likely to happen in
dense workgroups. Mentor relationships, for example, exemplify one way in which
closely selected associates can work with some partners, combine work and personal
ties, and create teams.
The existence of such workgroups and the multiplex generalized exchange
system in which they are themselves embedded are structural features of the firm as
a whole: it is a component of its corporate social capital. It can thus be asserted that
if economic performance is rooted in the structure of individual social ties and in
cohesive workgroups, the latter are themselves rooted in collective social structure,
yielding firm-level social capital. This firm-level social capital, that requires many
kinds of contributions, is also key to maintaining collective action and production. It
is made more visible in the following section.
CIRCULA TION OF RESOURCES AND ECONOMIC PERFORMANCE

Firm social capital is partly reflected by a multiplex generalized exchange system. 14
The existence of this system in this firm has already been suggested by an
abundance of density effects and scarcity of direct reciprocity effects at the dyadic
level across the three networks (see, for example, Lazega and Van Duijn 1997).
Such a general pattern of relations is made possible, in part, by reliance on the
institutional framework of the firm as a guarantee that resources contributed will
eventually come back (although not necessarily in kind) and by the fact that the
longer members stay in the firm, the more personal relationships develop among
Figure 1. Relationships between positions of approximately structurally equivalent actors in

three networks of relations (advice, coworker, and friendship) in the law firm (adapted from
Lazega 1998)
Thick lines represent reciprocated choices. Grey line represent advice ties, black lines
coworker ties, and dotted lines friendship ties. Roman figures represent the number of the
position. Letters have the following meaning: H for Hartford, B for Boston, P for partners, As
for associates, L for litigators, C for corporate. Thus position One represents BLP-Boston
litigation partners. The bigger circles are positions of partners; the smaller circles are positions
of associates.
them. For this pattern to encourage partner self-entrapment in their own workgroups.
for example. it has to be shown to reflect and sustain the tendencies described above.
Productive members can share several types of resources without immediate
reciprocity. The system involves forms of indirect reciprocities that take into
account several resources.
A 'Locally Multiplex' Generalized Exchange System

When staying at the dyadic level. it is difficult to get a sense of how exchanges use
different resources at the same time. and of the overall pattern of exchanges of the
three resources in the firm. Recognizing this more generally. Levi-Strauss (1949)
distinguished two forms of exchange: direct or restricted exchange (dyadic) and
indirect or generalized exchange (structural). An analysis approximating structural
equivalence among members of the firm across the three networks provides such an
overall view, represented in Figure 1. A detailed presentation of this Figure and the
density tables and image matrices on which it is based are available from the author.
To summarize, it shows that the three networks stacked together break down into
nine positions of approximately structurally equivalent members. Asymmetries in
the transfers of resources, along with the dependencies attached to them, create a
mUltiplex generalized exchange system as shown in Figure 1.
Positions One, Two, and Three are positions of partners: all the others are
positions of associates. The thick gray lines (reciprocated advice ties) reflect the
backbone of the firm: three partners positions and their senior associates. Notice that
requests for advice converge toward them. Many of the relationships between
positions are not symmetric. Position One is a socially dominant group of Boston
litigation partners who get advice, strong collaboration and friendship from
Positions Two and Five. Its members get almost what they want from the people
they choose. Many positions of associates are directly indebted to it for advice and
collaboration, but it is not the top-performing position (third in average individual
dollar collection). Position Two is a group of Boston corporate partners almost in the
same dominant situation as Position One, except that it has an exchange of advice
for friendship with Position Five. Many positions of associates are directly indebted
to it for advice but not for friendship. It is nevertheless the top economically
performing position. Position Three is a group mixing Hartford corporate and
litigation partners, in the same category as Positions One and Two in terms of
dependence on others for resources. It claims strong collaboration from Position
Two but is unreciprocated in kind: Position Two members tend not to rely on
Position Three for strong collaboration but do so for advice and friendship. Here
reciprocity tends also to be direct but not necessarily in kind. Note that Position
Three has direct exchanges of strong collaboration only with two positions of
associates, Positions Six and Seven, and not with other positions of partners-but
cross-selling can still take place, for example, since it is a mixed position (in terms
of specialties). As for Position Two, many positions of associates are directly
indebted to it for advice but not for collaboration or friendship. It ranks second in
average individual dollar collection.
If we come back to the individual and dyadic levels, this picture confirms that
partners' positions are characterized by a comparatively very high proportion of
Blau-ties-whereas associates' positions are characterized by the corresponding
high proportion of ties in which they are the advice-seeking party-and are highly
cohesive given their high representation of mutual triplex ties. Thus, at the aggregate
level, all three resources tend to circulate within positions of partners and among the
two Boston positions. Two additional forms of embedded ness indicate that the
partnership agreement is enforced through both economic interest and social ties.
Partners from Position One are in a socially advantageous situation to remind
partners from Position Two of their commitment, or to calm down status
competition among them. Partners from Position Two, in tum, are in a socially
advantageous situation for pressuring Position One and Position Three partners back
to good conduct. Although they are not often asymmetrically indebted to each other
and dependent on indirect reciprocity from each other, economic relations among
partners are overembedded for Position One and underembedded for Position Three,
with Position Two members playing a key role of balancing the two forms.
This also confirms that the forms of embedded ness of the labor contract for
partners and associates are radically different. Here, multiplexity is also more
extensively used to enhance the productivity of this economic tie. Associates tend to
feel indebted to partners for strong collaboration, advice, and sometimes friendship
(mostly unreciprocated). Two senior associates positions, Positions Five and Six,
have a less clear profile-in terms of direct reciprocity-than partners' positions.
Positions Five exchanges the three resources with Position One partners. In that
sense, it has almost a partners' profile. It has other uniplex and directly reciprocal
exchanges and does not have to exchange one resource for another (with the
exception of friendship for advice with Position Two). Two positions of associates
are directly indebted to it for collaboration, advice, and friendship. Positions Six is
in a different situation. It has more direct exchanges of strong collaboration with
positions of partners and has also a direct exchange of advice with a position of
partners. But it does not exchange friendship with partners (remember that it is made
of lateral associates who did not grow up through the ranks but were hired away
from other firms). It claims friendship with Position Two but this is not
reciprocated-unlike Position Five's access to Position One's friendship. Position
Six thus gets friendship from colleagues (Positions Five, Seven and Eight)
sometimes different from those to whom it provides it.
Small cycles characterizing local and multiplex generalized exchange can be
found in this system. Multiplex and local cycles reflect the existence of highly
embedded strong coworkers' teams, for example, between Positions One, Five and
Nine, or between positions Two, Six and Eight, or One, Two, and Four. This is due,
in particular, to the fact that indirect reciprocity is at its strongest with less senior
associates who are never in a position, for instance, to reciprocate for advice. They
most often have to reciprocate in Blau-type status recognition and strong
commitment to work. This reciprocation in commitment to work is not necessarily
directly directed to sources of advice but to others. For example, Position Eight gets
advice from Position Three but is only indirectly involved in strong work ties with
Positions Three through Six. Another example is Position Seven, that gets the three
resources from Position Five but is not in a position to reciprocate directly at all, and
thus remains indebted but provides Position Three with strong commitment to work
and friendship. A similar short cycle is also present between Positions Three, Six,
and Seven. Position Four is in the same dependent situation as Position Seven with
regard to the three partners' positions it has ties to. Note that it does not get
friendship at all from other positions, not even indirectly, which shows that the
generalized exchange system fails to provide this resource to some of its members.
A large spectrum of forms of embedded ness is present here to enforce the labor
contract for partners and associates. It would be too simple to summarize the
situation by saying that performance of individuals is affected by these forms of
embedded ness because high performers are always involved in very multiplex
cycles and low performers in less mUltiplex cycles, or because resistance to the
general exchange circulation rule always produces low performers. For some
associates high productivity goes hand in hand with strong friendship ties with
partners: Positions Five and Seven overbill and overcollect compared to the other
associates. Associates may often feel kept at arm's-length and that 'partners don't let
them in,' but here we also see that some partners are nevertheless in a favorable
social situation to extract high commitment from these associates. And at the other
extreme, Position Four associates are also highly productive but much less socially
connected to the partners with whom they work: requests for advice still indicate
social embedded ness through status recognition, but partners are less in a position to
extract work for friendship. In contrast, junior associates are less productive
economically and claim friendship with one another and with senior associates. Here
partners can play on resource dependencies to get commitment from associates to
their labor contracts in different ways, sometimes indirectly through the dependence
of junior associates on senior ones. Senior associates can also play this resource
dependence game with strongly socially embedded junior associates (in Position
Eight, pure friendship ties are highly overrepresented, and Position Seven sends a
record proportion of unreciprocated triplex ties) but not so much with partners.
Back to the dyadic level, we have seen that there is very often-in Positions
Two and Seven, One and Nine, Two and Four, and Three and Six-a Blau-tie. This
compound reflects one of the most frequent types of embeddedness. Partners also
concentrate requests for advice and unreciprocated citations as friends and advisors
or as coworkers and advisors. Such citations converge towards all partners' positions
but are strongly overrepresented for Position One. Also associates ' positions are
much less cohesive, which is confirmed by a low proportion of mutual triplex ties.
In Position Three, empty ties and mutual duplex (cowork and advice) ties are
overrepresented.
Multilevel Embeddedness: A Virtuous Circle?

Exchanges and transfers of resources, along with asymmetries and dependencies
attached to them, create a local and multiplex generalized exchange system. This is a
social ecological system that has grown around the formal dimensions of the
organization and constraints imposed by interdependence in production. It is part of
firm-level corporate social capital: it helps members accept terms of the labor
contract that they do not necessarily have a narrow interest in accepting. In other
words, the firm has found in this exchange system a structural solution to the
structural problem of cooperation and commitment to the labor contract, a sort of
partial equilibrium in the circulation of resources needed to fulfill it. Given informal
constraints guiding the choices of coworkers, advisors, and friends in this firm, and
given the existence of such multiplex and local cycles reflecting the existence of
highly embedded strong coworkers' teams, givers have a guarantee that they will
become receivers, although not necessarily in kind. By allowing such a system to
exist, this firm maintains certain forms of resource circulation, a precondition for
group solidarity.
Thus, individual commitment to labor contract and individual economic
performance are increased by membership in dense workgroups that themselves
need a wider and more multiplex exchange system to operate. Individual
performance that benefits the firm as a whole is driven by workgroup pressure and
the social system of the firm : the organization helps its members perform, thereby
helping itself through aggregated performances. In many ways this mechanism is

good for the firm as a whole as weII as for the individual. It aIIows very different
individuals to carve out a place for themselves in the organization, based on the
exchange of various resources. Social networks of individuals make them more
productive in the organization-creating individual-level social capital; but it also
transforms itself into firm-level social capital because the firm as a whole is more
successful in billing and maintaining labor contracts and integration.
However, the embeddedness of individual economic performance in a social
exchange system also raises the issue of the relative contribution of individual social
capital and firm social capital to collective action and performance. 15 When several
resources circulate, exploitation is present but not easily measurable. For example, if
firm-level social structure helps individual members perform, then members who
participate in building up this social capital are also entitled to some of the credit
that goes to high performers. In this situation, the restraint of members (both in
keeping track of all their contributions and returns, and in politicizing the use of
measurements) is a micropolitical condition for the efficiency of this system. This is
especially the case with teamwork, where others' behavior can easily be perceived
to be opportunistic. In fact, members' restraint is often weak: many conflicts in
corporate law firms or other private professional services firms are disputes about
fairness in compensation to individual partners. But many cultural aspects of such
organizations, such as an ideology of collegiality, are explicit exhortation to such a
restraint.
Thus measuring only economic performance tells only one side of the story of
contribution to collective action. This limiting condition for the efficiency of such a
virtuous circle is the reason for managing partners often to emphasize 'spreading
credit' as widely as possible. Nevertheless, however, recognizing and measuring the
relative and specific importance of corporate social capital (as compared to other
forms of capital) cannot be done without understanding this political negotiation
that enables members to evaluate their contributions. This means that the relative
value of social resources is negotiated, and that this negotiation is political.
FoIIowing these politicized negotiations and how they try to disentangle the merits
of the firm as opposed to that of its individual members in the production of a
specific performance is beyond the scope of this chapter, and perhaps beyond the
possibilities of structural analysis.
CONCLUSION
In summary, this chapter examined the question of the relationship between social
structure and economic performance at the intraorganizational level. I have
identified a few conditions under which individual social ties are most productive
for the firm in collegial organizations-where the production process is difficult to
routinize, where professional expertise and advice cannot be standardized, and
therefore where internal transaction costs for the firm as a whole can be assumed to
be a large part of total costs. An empirical study of a medium-sized northeastern
U.S. corporate law firm was used for that purpose. In this firm, attorneys are shown
to be bound by a labor contract that is difficult to sustain on pure economic terms.
Partners can easily free-ride; associates can threaten the quality of work. Against
this damage potential, a social system sustains their commitment. Using network
data collected in the firm, I describe and measure social capital at the individual,
workgroup, and structural levels to show that the more constraining the member's
coworkers' network, the easier it is for the firm to extract higher economic
performance, including from tenured partners, by controlling the time put into work.
Thus, position in the relational structure and social capital accumulated in this
position do count for explaining performance, although these effects are weaker
when compared to the weight of hourly rates as defined by the institutional setting.
With regard to partners, such teams represent an element of self-entrapment
compensated by stimulation, status and professional recognition.
Second, the effect of the individual social network on economic performance is
decomposed at the dyadic level by looking at specific combinations of ties that
sustain this commitment and provide a decisive increase in performance (or that
represent a liability decreasing performance). For example, specific configurations
of social ties, such as mutual triplex ties, are strongly correlated with high
performance. This fleshes out the positive effect of constraint scores in the
coworkers' network. Examples of low and high economic performers, and their
respective combinations of social resources, were also provided as illustrations.
Third, back to the structural level, a locally multiplex generalized exchange
system is described as a precondition for individual and group-level social structure
to be productive because it maintains the circulation of social resources in the firm,
thus making it possible to improve performance for structurally well located
members and workgroups who can benefit from this circulation. A multilevel form
of embedded ness is thus revealed here, which shows the importance of taking into
account a 'meso' level when measuring the relationship between social structure and
performance. In this particular case, the notion of multilevel embeddedness
advances our understanding of economic performance: the latter is rooted in
individual social capital, which is itself rooted in workgroup and firm social capital,
which in tum helps individual members in being more productive. Individual social
ties can be most productive-for the individual and/or for the firm-when they are
part of a favorable system accumulating collective social capital. By favorable, I
mean that the individual social network is located at the right place and mobilized at
the right time in a wider production and exchange system.
Finally, however, the politicized nature of performance measurement prevents
us from claiming that a permanent virtuous circle is produced by the fact that social
capital is located at the individual, group, and structural levels. Exploitation and
behavior perceived to be opportunistic are also corollaries of this multilevel
embeddedness. A final example illustrates one possible implication of these results.
As many managing partners in law firms know, the importance of constraint at the
group level is not necessarily an encouragement for management to create dense and
permanent workgroups in collegial organizations. The existence of such groups is
risky for the firm. They can threaten the firm with disintegration when entire teams
consider themselves exploited (relatively to others in the firm or to others outside
the firm), decide to defect, and take away with them part of the firm's human and
social capital. As shown elsewhere (Lazega 1992, 1999), what the managing partner
of this firms says about individuals is also true about workgroups:
There are client loyalties to individual lawyers within the firm, but among ourselves we
view all clients as clients of the firm. And indeed, if you are an individual to whom the
client has demonstrated a great degree of loyalty, one of your responsibilities is to make
sure that there are other partners to whom that client may also look and rely on, not
necessarily on an ongoing basis. But if for some reason, for example, I am away, if I
were suddenly to decide to go pump gas for the rest of my life-any number of things-
that client loyalty is not an asset that belongs to me. If I were to go to another firm, if I
have done my job well here at S,G&R, if I call my client and say 'I want you to know that
I am in firm X,Y &Z now,' that client's response should be 'Whom at S,G&R should I call
now?' It shouldn't be 'What's your new number?' Whether that would be the case in all
cases, who knows? That's what ideally it should be. (Managing partner at the time of the
study)
Social structure can thus produce a large amount of social capital for collegial
organizations, but its manipulation is double-edged.
APPENDIX A
SOCIOMETRIC NAME GENERA TORS USED TO ELICIT COWORKERS,
ADVICE, AND FRIENDSHIP NETWORKS
Here is the list of all the members of your firm.
Coworkers network: Because most firms like yours are also organized very
informally, it is difficult to get a clear idea of how the members really work
together. Think back over the past year, consider all the lawyers in your Firm.
Would you go through this list and check the names of those with whom you have
worked with. (By 'worked with' I mean that you have spent time together on at least
one case, that you have been assigned to the same case, that they read or used your
work product or that you have read or used their work product; this includes
professional work done within the Firm like Bar association work, administration,
etc.)
Basic advice network: Think back over the past year, consider all the lawyers in
your Firm. To whom did you go for basic professional advice? For instance, you
want to make sure that you are handling a case right, making a proper decision, and
you want to consult someone whose professional opinions are in general of great
value to you. By advice I do not mean simply technical advice.
Friendship network: Would you go through this list, and check the names of those
you socialize with outside work. You know their family, they know yours, for
instance. I do not mean all the people you are simply on a friendly level with, or
people you happen to meet at Firm functions.
I would like to thank Henk Flap. the November 1997 ICS-seminar participants in Groningen. The
Netherlands, and the book editors for comments on a previous draft.
NOTES
1. Elsewhere I have provided other examples of the fact that this finn finds structural solutions to
structural problems (Lazega 1992a, 1995a, 1995b, 1997; Lazega and Vari 1992; Lazega and Lebeaux
1995; Lazega and Krackhardt 1997).
2. This is not to say that this ecological system, which makes the partnership agreement enforceable,
disciplines all the members equally strongly. Some pay a higher price to be part for it. For example, some
associates are put in a better position to try to build their competitive advantage on the use of these
embedded ties. Close demonstration of this is, however, beyond the scope of this article.
3. Nelson (1988: 91-92) defines traditional management as characterized by 'I) ad hoc and reactive
policy-making, with little long-range planning; 2) direct administration by leading lawyers, aided only by
a part-time managing partner, with no regular monitoring of internal perfonnance measures or financial
infonnation; and 3) infonnally defined and shifting work groups.' Bureaucratic management is defined by
'1) a specialized policy-making group that actively engages in strategic planning; 2) a developed
administrative component consisting of a managing partner and a mechanism for collecting and analyzing
data on the financial perfonnance of individual lawyers and work groups; and 3) well defined work
groups (usually taking the fonn of departments) with recognized heads who supervise the group and
report to the central policy-making group.'
4. For example, given the way a partner is compensated in the finn, looking at the dollar amount
actually collected in 1991 does not indicate exactly how productive this attorney was in 1991. Work done
in 1990 can be compensated in 1991 (or perhaps even later), and such overlaps make it difficult to
disentangle an attorney's productivity in one year as opposed to his or her productivity in another year.
Simultaneously, looking at the number of hours billed in 1991 gives an idea of an attorney's productivity
in 1991 but does not mean that all the work was done in 1991.
5. Seniority is defined by the rank of partners in the letterhead, which is mainly based on age and years
with the finn (with the exception of four partners who were hired away from other finns). Coding of
seniority levels in senior, medium seniority, and junior partners is based on cutoffs between Partners 14
and 15 (a difference of eight years in age) and between Partners 27 and 28 (a difference of nine years in
age). These categories were explicitly used by the partners themselves.
6. For more infonnation on these networks, see Lazega (1992a, 1992b, 1993, 1994, 1995a, 1995b), and
Lazega and Van Duijn (1997).
7. This is partly at odds with Burt's (1992) general statement about association between low constraint
and high perfonnance and more consistent with Coleman's (1990: ch. 12) ideas on the benefits of closure
and embeddedness of ties. In this chapter, I mainly analyze economic perfonnance understood as the
amount of fees brought into the finn at the end of the year. Such amounts depend minimally on the
amounts of time worked and on hourly rates. Nevertheless, the more members work, the more they
perfonn in that sense. My point is that extracting work from them should be easier in a constrained
network of work ties. Analyzing the determinants of other types of individual perfonnance, such as
promotion to partnership, could presumably yield different results and an opposite sign to the
association-which would be more in line with Burt's results. About this issue, see also Gabbay (1997).
8. There are many reasons for the inability of hourly rates and numbers of hours worked to explain all
the variations in financial perfonnance. First, billing partners do not bill all the hours worked by their
team. There are various fonns of nonchargeable time, and very often there is a negotiation between the
finn and the client as to what is an acceptable price for the services rended. The billing partner then writes
off a considerable proportion of hours worked before the bill is sent. Second, corporate law finns have
notorious difficulties collecting what was billed, and many partners choose to live with high account
receivables rather than antagonize a client from whom they expect more business in the future.
9. However, causal links are difficult with non longitudinal data; it is impossible to know here whether
members are low performers because they establish different types of relationship with their colleagues or
whether they establish these relationships to try to mitigate the effects of their low perfonnance and carve
out a different place for themselves in the group.
10. A longitudinal approach could bring more insights into these effects and especially enable
researchers to identify causal relations. A look at perfonnance data for the next year (1991) suggests the
possible existence of a cyclical and infonnal mechanism in which attorneys who work and collect a lot on
big cases tend to bum out temporarily, to attract much social approval, to slow down a bit, and work with
fewer colleagues, until they are ready to pick up again, thus triggering a new phase in which they start
working with more colleagues again, increase their work load, and bill more until they collect again.
Needless to say, the existence of such a mechanism remains to be proven, and, if so, the cycle should vary
from one attorney to another.
II. The type of tie in which i and j consider each other as strong coworkers and in which j seeks advice
from i whereas i seeks j for socialization outside the fmu is a frequent type of compound.
12. There are exceptions, of course: Partner 12, for example, has more than ten ties including
unreciprocated (by j) friendship ties. Declaring more friends than one may actually have nevertheless
characterizes associates' profiles much more.
13. Practicing very diverse and more complex-less task-oriented-types of exchanges also
presupposes more flexibility with rules of exchange of resources. Description of such normative games,
however, is beyond the limitations of this chapter.
14. For a discussion of the relationships between networks and generalized exchange systems, see for
example, Levi-Strauss (1949), Ekeh (1974), Bearman (1997), Breiger and Ennis (1997).
15. The fact that this system helps some individual members reach high performance does not mean that
it is egalitarian in the distribution of resources and in the provision of structural solutions to individual
problems. This can also be illustrated by looking at the relative chances of senior associates to become
partners. In their competition for the attention of partners, associates with the right connections to the
right partners-with a specific position in this pattern and a specific type of individual social capital-
have a clear structural advantages in the highly selective race to partnership. They are in the fast lane
because these connections, among other advantages, allowed them to play with organizational rules in an
rewarding way, in particular to cross internal boundaries (for example, seek advice from very senior
partners), provided that such 'infractions' are limited and well localized (Lazega 1995a).
CEO Demographics and Acquisitions:
Network Effects of Educational and
•
Functional Background
14
Pamela R. Haunschild
Andrew D. Henderson
Alison Davis-Blake
ABSTRACT
This study investigates the effects of CEO educational and functional background on
corporate acquisitions. Educational and work-related functional backgrounds are
likely to come with interorganizational networks, networks that stay with individuals
over long periods of time and have the potential to affect acquisitions. We argue that
these networks constitute a form of interorganizational social capital, which directs
acquisition activities along certain channels. Hypotheses are tested on 449 firms and
their acquisitions during the 1986-1993 period. We find evidence that the networks
that come with different CEO education and functional backgrounds are related to
the type of acquisition completed by that CEO's firm. Obtaining a degree from an
elite school. for example. is related to engaging in acquisitions in unrelated
industries. We also find functional background effects are strengthened under
conditions of uncertainty and educational background effects are weakened with
tenure. These results suggest the importance of personal networks in affecting major
firm strategic actions. and highlight the contextual nature of acquisition decisions.
INTRODUCTION
What factors affect a firm's choice of strategic actions? This question lies at the
heart of research on strategic management, and factors external and internal to the
firm have been proposed as answers. Several recent studies have shown that key
strategic actions such as engaging in acquisitions. adopting a poison pill. and
implementing the multidivisional form are affected by external factors. These
factors include the actions of other firms in the same industry (Fligstein 1985. 1991;
Palmer, Jennings. and Zhou 1993). the actions of interlock partners (Davis 1991;
CEO Demographics and Acquisitions - 267
Haunschild 1993; Palmer et al. 1993), and the constraints imposed by resource
dependencies (Pfeffer and Salancik 1978).
Although external factors undoubtedly affect strategic actions, there has been a
renewed interest in internal factors as determinants of strategy. In particular, several
recent studies have investigated the effects of CEO demographics on a firm's
strategic actions (Fligstein 1990; Finkelstein and Hambrick 1990; Palmer et al.
1993; Davis, Diekmann, and Tinsley 1994). Most of this work has focused on the
relationship between functional background and strategy. Early foundations for this
work came from the Carnegie School theorists. They argued that, by creating a
cognitive map or way of thinking about issues, functional background affects
managerial decision making in complex situations (Dearborn and Simon 1958; Cyert
and March 1963; but see Walsh 1988 for an alternative point of view). Later,
Hambrick and Mason (1984) advanced the upper echelons perspective in which they
argued that organizational strategies reflect the cognitive bases and preferences of
the top management team. Various demographic variables, including functional
background, were proposed to represent these cognitive bases. The relationship
between various demographic variables and strategy has been empirically
demonstrated by several researchers (e.g., Norburn and Birley 1988; Bantel and
Jackson 1989; Grimm and Smith 1991; Wiersema and Bantel 1992). Demographic
variables have also been proposed to be indicators of power, and power affects
strategic action in predictable ways. For example, Fligstein (1990) showed the wave
of conglomerate acquisitions that occurred in the 1960s can be partially explained by
finance CEOs that rose to power and executed acquisitions consistent with their
functional specialty.
Past research on the link between managerial background and strategic action is
important because it shows that the backgrounds of top managers can have an effect
on major aspects of corporate strategy. The theoretical foundation of prior work,
however, has been limited to arguing that managerial background affects managerial
cognition and power. Yet certain educational and work-related backgrounds are
likely to come with interorganizational networks, networks that stay with individuals
over long periods of time and have the potential to affect major strategic decisions.
The impact of these networks, in turn, is likely to be affected by other contextual
factors. The purpose of this study, then, is to examine the network effects of CEO
education and functional background on a major firm strategic action: corporate
acquisitions. We argue that the networks that come with functional and educational
background provide a form of interorganizational social capital, which directs
acquisition activities in predictable ways. By studying the networks that come with
education and functional background, we provide a link between the internal (CEO
characteristics) and external (interorganizational factors) perspectives on strategic
action. A second purpose of this study is to investigate contextual conditions likely
to moderate these network effects. We study two such conditions: decision
uncertainty and CEO tenure. We choose uncertainty and tenure because each have
been shown to moderate the impact of networks in other settings.
Corporate Acquisitions
Corporate acquisitions are an interesting and important arena in which to explore the
network effects of CEO educational and functional background. Acquisitions
represent a critical strategic choice, one that typically involves the commitment of
substantial resources and the expenditure of political capital. Acquisitions are also
highly complex events (Haspeslagh and Jemison 1991) that can generate a great deal
of decision uncertainty (Haunschild 1994). Some researchers (Dearborn and Simon
1958; Cyert and March 1963; Hambrick and Mason 1984) have proposed that
demographic indicators like CEO educational and functional background are
particularly likely to be associated with complex, uncertain decisions. Thus,
corporate acquisitions are an important setting in which to test whether CEO
background affects strategic action and whether these effects vary with contextual
conditions such as decision uncertainty. Although the effects of functional
background on acquisitions have been the subject of a few previous studies (Song
1982; Fligstein 1990, 1991; Davis et a1.1994), these studies have been limited to
proposing that the relationship between functional background and acquisitions are
due to its effect on managerial cognition and in determining who has power in
organizations. By examining network effects, we present a very different view of
acquisitions from what has been presented before.
What causes firms to do acquisitions, and do acquisitions of certain types? This
is an interesting and important question, especially in light of evidence indicating
that acquisitions often reduce the value of the acquiring firm (Ravenscraft 1987).
Much of the work on acquisition motivations is financial and efficiency-based,
proposing that acquisitions are driven by such things as the search for synergy (e.g.,
Jensen 1984; Ravenscraft 1987). Resource dependence theory presents an alternative
perspective, proposing that acquisitions are a response to the constraints imposed by
organizational interdependence (Pfeffer and Salancik 1978). The managerial and
agency theories propose that top managers, especially CEOs, have a great deal of
influence over acquisition activities, and thus their motivations and incentives can
playa significant role in predicting corporate acquisition activity. Some managerial
theories, for example, propose and find evidence that acquisitions are driven by
managerial desire for the power, prestige, and financial rewards associated with
managing large companies (e.g., Marris 1964; Baumal 1967). Agency theory
proposes that poorly-monitored managers and managers whose incentives are out of
alignment with shareholder interests will engage in acquisitions consistent with their
own interests and inconsistent with shareholder interests (e.g., Morck, Schleifer and
Vishny 1990).
While these latter theories present evidence for individual effects on corporate
acquisition activity, they are largely asocial, considering only the impact of
individual incentives and desires. The same is true for the few studies looking at the
relationship between functional background and acquisitions. This means that most
existing theories of acquisitions do not consider the social context in which firm
managers are embedded. Firm managers exist in a social world, and carry with them
interorganizational networks that provide them with social capital in the form of the
information they receive, and their ability to engage in acquisitions. This study
provides social context by investigating whether the interorganizational networks
associated with CEO backgrounds affect their firms' acquisition activities.
We examine CEO background rather than the background of the entire top
management team (TMT) because we expect that the number and strength of
interorganizational network ties is greater for CEOs than for other members of the
top management team. CEOs are more likely than other top management team
members to be on corporate boards and to participate in other arenas in which they
might learn about acquisition targets (Useem 1982). Thus, examining only the CEO
is adequate to test the background effects in which we are interested. Our approach
is also consistent with several recent studies of the effects of managerial background
on firm action. For example, studies of M-form adoption (Fligstein 1985; Palmer et
al. 1993), diversification activities (F1igstein 1990, 1991) and acquisitions (Davis et
al. 1994) have looked at the effects ofthe CEO, not the entire top management team.
Our theoretical arguments are organized as follows. Educational and functional
backgrounds are demographic indicators of managerial networks. We start by
discussing the effects of CEO educational background on corporate acquisitions,
focusing on the network effects of an elite graduate degree. Second, we discuss the
effects of CEO functional background on acquisitions, focusing on the network
effects of an output and peripheral function background. We conclude by discussing
how these background effects are likely to vary with decision uncertainty and
power.
HYPOTHESIS DEVELOPMENT
CEO Educational Background and Corporate Acquisitions
We argue that the interorganizational networks that come with graduate degrees
from elite schools provide executives with information about acquisition targets and
access to financial resources necessary to engage in acquisitions. These networks
thus provide corporate social capital, allowing these CEOs to engage in certain types
of acquisitions. Graduate degrees from elite schools have been said to provide
individuals with lifelong business contacts among the corporate elite and there is
evidence that these contacts affect firm actions (Useem 1982; Palmer, Jennings, and
Zhou 1993). Only one study, however, has empirically examined the relationship
between elite education and firm strategic actions. Palmer et al. (1993) proposed
those graduates of elite business schools are especially likely to maintain contacts
with each other and thereby learn about accepted business practices such as the
adoption of the multidivisional form (M-form). They found a positive relationship
between a CEO having an elite business school degree and a firm's likelihood of
adopting the M-form.
Contacts formed during an elite graduate education may also affect acquisitions.
As discussed by Palmer et al. (1993), these ties may provide information about
business trends. Elite ties are thus likely to provide information about acquisitions,
either general information about the appropriateness of acquisitions as a strategy or
specific information such as information about the availability and suitability of
particular acquisition targets. Elite education also shapes corporate social capital that
may provide access to the financial resources and social networks necessary to
engage in acquisitions. Useem and Karabel (1986) showed that a degree from an
elite graduate school facilitates corporate ascent. They also found that holders of
elite degrees were more likely than graduates of nonelite schools to become
directors of multiple corporations and leaders of major business associations. Thus,
elite education provides social ties with others that control large firms (Useem and
Karabel 1986; see also Useem 1979; D'Aveni 1990; Palmer, Jennings, and Zhou
1993).
The information and interorganizational corporate social capital provided by
elite ties are likely to be especially important for domain-expanding (conglomerate
and vertical) acquisitions. Conglomerate acquisitions occur when firms buy
unrelated firms in industries that are neither potential buyers, suppliers, competitors,
nor complements to the acquiring firm's existing business. Vertical acquisitions
occur when firms acquire their suppliers or distributors. Conglomerate and vertical
acquisitions are domain-expanding acquisitions, because the firm is entering entirely
new lines of business, or moving up or down the supply/distribution chain.
The contacts provided by an elite education are likely to be helpful for CEOs
engaging in domain-expanding acquisitions for two reasons. First, because these
contacts are likely to cross industry boundaries (Mintz and Schwartz 1985), CEOs
with these contacts are probably more able to identify acquisition targets outside of
the industries of their current business than CEOs without these contacts. Second,
consolidating firms across sectors is likely to require more power and control in the
intercorporate network than consolidating firms within sectors, and CEOs with elite
degrees are particularly likely to have this power (Useem and Karabel 1986). We
thus hypothesize:
H1: Firms in which the CEO received an elite graduate degree will be more likely
to engage in domain-expanding acquisitions than firms in which the CEO did
not receive an elite graduate degree.
CEO Functional Background and Corporate Acquisitions

Following Hambrick (1981), and Hambrick and Mason (1984), functional
backgrounds can be classified into four general categories. Output junctions, which
include marketing, sales, and product R&D, involve an emphasis on growth and the
search for new opportunities. Throughput junctions, which include production,
process engineering, and accounting, involve a focus on improving the efficiency of
the input-to-output transformation process. Peripheral junctions, which include law
and finance, are areas that are not integrally involved in a firm's core activities (cf.
Hayes and Abernathy 1980). Finally, general management involves a concern with
the functioning of the firm as an integrated whole and a focus on what businesses
the firm should be in (Andrews 1971; Asnoff 1965).
There are several possible mechanisms through which CEO functional
background might affect patterns of acquisitions. To date, researchers have focused
on the effects of functional background on managerial power and managerial
cognition. According to the power perspective, CEO functional background affects
acquisition strategies because individuals who have the skill to engage in
acquisitions gain power when acquisitions will help the firm deal with critical
strategic contingencies (Pfeffer 1981; Fligstein 1990, 1991). Consistent with this
perspective, Fligstein found that the functional background of the CEO varied with
critical contingencies that firms faced over the past century (Fligstein 1990).
According to a cognitive perspective, functional background causes individuals
to interpret an organization's problems and solutions to those problems in
characteristic ways (Dearborn and Simon 1958). Fligstein (1991), for example,
proposed that executives with finance backgrounds possess a 'finance conception of
control,' in which the firm is viewed as a bundle of assets to be bought and sold, and
the finance conception of control has been prevalent since the 1960s. Consistent
with this, Fligstein (1991) found that firms with finance CEOs engaged in more
conglomerate acquisitions in the 1960s and 1970s than firms with CEOs without
finance backgrounds. Using similar logic, Mizruchi and Stearns (1994) found that
firms with finance CEOs tended to use more external borrowing as a financing
strategy than firms whose CEOs had other backgrounds.
However, we propose that the effects of CEO functional background go beyond
power and cognition. Functional background may also lead to systematic patterns of
external corporate ties, resulting in opportunities to engage in certain strategies. The
functional backgrounds that are most likely to be associated with external ties that
affect acquisitions are output and peripheral functions. Yet, these two backgrounds
are associated with different types of ties: the social capital inherent in CEO
networks varies with the CEO's functional background. Peripheral function CEOs
are likely to have wide-ranging networks, including somewhat extensive contacts in
the investment banking and legal communities. These contacts are likely to provide
access to acquisition opportunities that are based more on capturing financial value
than on creating operating synergies. Thus peripheral function CEO are more likely
to do conglomerate acquisitions since their networks produce information and access
to target firms in unrelated industries.
H2a: Firms in which the CEO has a peripheral function background will be more
likely to engage in conglomerate acquisitions than firms having CEOs with
other backgrounds.
CEOs with output function backgrounds have spent much of their careers in
marketing and sales oriented jobs. These jobs require extensive interactions with
suppliers and distributors. Thus, output function CEOs are likely to have many more
contacts with firms in distributor (also possibly supplier) industries than CEOs with
other backgrounds. These contacts in distributor industries are likely to facilitate
vertical acquisitions--drawing attention to targets in these industries and producing
contacts that facilitate acquisitions in these industries. Thus, the interorganizational
links associated with an output background are likely to affect the number of vertical
acquisition targets noticed and to improve access to those targets through social ties.
If they do, then we expect that CEOs with output function backgrounds are more
likely to engage in vertical acquisitions than CEOs with peripheral backgrounds.
H2b: Firms in which the CEO has an output function background will be more likely
to engage in vertical acquisitions than firms in which the CEO has a
peripheral function background.
Moderating Effects of Decision Uncertainty

The relationships between networks and acquisitions outlined above are likely to be
moderated by an important factor: uncertainty. While other possible moderators
exist, we chose to study uncertainty because it is suggested by current theory, yet
has not been studied in this context. For example, uncertainty has been shown to
moderate the impact of interorganizational networks on strategic action (e.g.,
Haunschild 1994).
Several theories support the idea that uncertainty surrounding a decision is
likely to affect that decision in two ways. First, uncertainty causes individuals to rely
on cognitive shortcuts and heuristics (Cyert and March 1963; Hambrick and Mason
1984), such as those produced by one's educational and functional background.
Second, uncertainty increases the extent to which individuals engage in social
comparison (Festinger 1954). This means that managers experiencing uncertainty
are likely to look to others outside their organization for ideas and information about
business practices. Consistent with this latter idea, Haunschild (1994) found that
uncertainty about acqUisItion targets causes firms to establish more

interorganizational relationships, and to be more influenced by the information
obtained through these relationships. Both the cognitive shortcut and social
comparison arguments suggest that the interorganizational networks that arise from
educational and functional background should be more influential under conditions
of uncertainty. Thus, the acquisition decisions of CEOs experiencing a high level of
uncertainty should be more affected by the CEO's educational and functional
backgrounds than the decisions of CEOs experiencing a low level of uncertainty.
H3: The effects of educational and junctional background on acquisition activity
are strengthened by firm uncertainty.
Moderating Effect of CEO Tenure

CEO tenure should weaken the relationship between CEO background and
acquisition strategies. As tenure increases, a CEO's management paradigm is likely
to become more fixed and less open to diverse input (Finkelstein and Hambrick
1990; Miller 1991; Hambrick and Fukutomi 1991; Davis, Tinsley, and Diekmann
1997). Consistent with this, studies have found that CEO tenure is associated with
strategic persistence (e.g., Finkelstein and Hambrick 1990). One reason strategic
persistence occurs is that information accessed for decisions tends to become
routinized over time, reflecting past experience more than new information (Katz
1982). This means that long-tenured CEOs are likely to become more parochial in
their decision making, relying more on internal organization factors, and less on
external factors. If this is true, then the interorganizational networks associated with
an elite education and functional backgrounds should have weaker effects on long-
tenured CEOs.
H4: The effects of educational and junctional background on acquisition activity
are weakened with CEO tenure.
METHOD
Sample
The sample for this study consisted of all firms completing an acquisition between
111186 and 7/15/93. The acquisition had to meet the following four criteria: 1) the
acquiring firm bought a controlling interest in the target firm; 2) both acquirer and
target were U.S. based, publicly held companies; 3) data about the acquiring firm's
CEO was available; and 4) at least ten acquisitions occurred in the acquiring firm's
industry during the period studied. The fourth criterion was used to allow us to
control for industry effects. The following 13 industries were included in the
sample: oil and gas extraction; food and kindred products; chemicals and allied
products; fabricated metal products; electrical and electronic machinery; measuring,
analyzing and controlling instruments; lum6'er and wood products; transportation
equipment; communication; electric, gas and sanitary services; insurance; banking;
and business services. Including only acquisitions for U.S. based, publicly held
companies was necessary to insure the availability of data on CEO background and
several of the control variables. The data were obtained from the Merger and
Corporate Transaction Database maintained by Securities Data Corporation
(hereafter, SDC). This produced a sample of 271 acquisitions, completed by 228
firms. To rule out the possibility that results of our tests for different types of
acquisitions were affected by the fact that only firms completing acquisitions were
in these analyses, we also collected a sample of nonacquirers. The 221 nonacquiring

firms were randomly selected from the 1985-1992 Business Week Corporate Elite
survey, producing a final sample of 449 firms, 221 of which did not complete
acquisitions, and 228 of which completed one or more.! We then tested for effects of
sample selection bias using methods described in Berk (1983). The results of these
tests show no evidence of selection bias due to including only the acquirers in the
analyses (results available from the authors).
Measures
A dummy variable for each firm was coded one if the firm completed one or more
acquisitions during the study period, zero otherwise. For each acquisition, three
dummy variables were created to indicate whether the acquisition was related,
vertical, or conglomerate. Each variable was coded one if the acquisition was of that
type, zero otherwise. A classification scheme similar to those used in other studies
of acquisitions (e.g., Davis, Diekmann, and Tinsley 1994) was used to classify
acquisitions. An acquisition was coded as related when the two-digit SIC code of the
acquiring firm matched that of the acquired firm. SIC codes were obtained from the
SDC database. An acquisition was coded as vertical when the industry of the
acquiring firm either sold more than 5% of its output to, or received more than 5%
of its input from the industry of the acquired firm. Input-output numbers were
obtained from the U.S. Department of Commerce' s Survey of Current Business. All
remaining acquisitions were coded as conglomerate.
Explanatory Variables
CEO biographical data were obtained from Business Week's Corporate Elite survey,
which includes CEOs of the Business Week 1000 (Business Week 1986-1993).
These data were obtained by questionnaire from either the CEO or a designated
corporate representative. The biographical data include the CEO's self-reported
education, primary career path, and tenure as CEO. While there are limitations to
this type of self-report data, any errors in such reporting are unlikely to be
systematically related to acquisition activity.
The educational data from the Business Week survey were used to construct
dummy variables for education level (undergraduate, masters, PhD), education type
(MBA, other), and whether the CEO has a graduate degree from one of the eleven
elite schools identified by Useem and Karabel (1986). As preliminary support for
the idea that CEOs with elite graduate degrees have more network ties than CEOs
with nonelite degrees (see HI), we obtained data on the interlocks of the CEOs
included in this study. CEOs with elite graduate degrees participate on more boards
than CEOs without elite graduate degrees (F(1,245)=3.98, p<.05). As interlocks are
only a subset of all possible interorganizational networks, this data does not provide
conclusive evidence that elite degrees are associated with more interorganizational
networks, and thus we use the elite degree dummy variable rather than the number
of interlocks in the analyses.
The primary career path data in the Business Week survey were used to
construct dummy variables for output function, throughput function, peripheral
function, and general management backgrounds. Each CEO's primary career path
was used to construct the dummy variables. For example, if a CEO's primary career
path was finance, the peripheral variable was coded 1, and the output, throughput,
and general management variables were each coded O. In all analyses, peripheral
function background is the omitted category. As preliminary support for the idea
that CEOs with output function backgrounds have more network ties to firms in
vertically related industries, we obtained data on the interlock ties of CEOs. We then
classified these interlocks according to whether they were to firms horizontally
related, vertically related, or unrelated to the CEO's own firm, using the
classification scheme described earlier. We then tested to see whether CEOs with
output-function backgrounds sit on more boards of firms that are vertically related to
their own firm than CEOs with other backgrounds. Results show CEOs with output
function backgrounds do tend to sit on boards of firms that are vertically related to
their own industry (F(1,487)=3.80, p<.05). They also tend to sit on boards of firms
that are unrelated to their own industry (F(1,487)=4.21, p<.05). They tend to not sit
on boards of firms that are horizontally related. Thus, output function background
seems to be associated with sitting on boards of firms that are not closely related to
one's own firm. We also found that CEOs with peripheral backgrounds tend to sit on
boards of firms that are horizontally related to their own firms (F(1,487)=4.23,
p<.05). This seems somewhat inconsistent with our idea that CEOs with peripheral
function backgrounds have more ties to the investment banking and legal
communities. Since interlock ties are only a subset of all possible ties among
members of different firms, however, this result may reflect that investment banker
and legal ties are not captured well in interlock data.
CEO tenure was measured as years as CEO. Uncertainty was operationalized as
the variance of opinion regarding the value of the acquiring firm. While this cannot
be directly observed, a proxy measure was obtained from the IIBIFJS database of
Lynch, Jones and Ryan. Lynch, Jones and Ryan monitor the earnings per share
(EPS) estimates produced by research analysts from leading brokerage firms for
over 2000 companies. The variance of these estimates reflects the dispersion of
opinion among analysts regarding the future performance of a company. If analyst
estimates vary, then there is a lack of agreement or clarity about the underlying facts
affecting the firm, which is likely to reflect the uncertainty experienced by their
management. Similar to Haunschild (1994), we measured each firm's uncertainty as
the coefficient of variation of the analyst's projected acquirer EPS estimates for that
firm.2
Control variables
Six variables were included to control for factors related to the likelihood of a firm
completing acquisitions of various types. These variables were the size, profitability,
and industry of the acquiring firm, CEO age, CEO education level, and the year the
acquisition occurred. The size and profitability of the acquiring firm might affect
either the likelihood of an acquisition or the type of acquisition undertaken. Size and
profitability are important indicators of the existence of resources needed to
undertake an acquisition. In addition, there is some evidence that size and
profitability are related to the likelihood of conglomerate acquisitions (Davis et al.
1994; Haunschild 1993). Size was measured as a firm' s annual sales. Profitability
was measured as return on equity. Both size and profitability were obtained from the
Business Week listing. Industry controls were included to control for the fact that
patterns of acquisition vary by industry (Pfeffer and Salancik 1978), and
~
CD
Table 1. Descriptive statistics and correlations among key study variablesa

Mean S.D. Min Max ~Q ~22 P2 ~42 ~52 ~62 ~72 ~82 ~92 (102 ~112 ~122 ~132 ~142 ~15) ~16) ~17)
I. Acquirer (I =yes, .555 .497 0 I
0=00)
2. Conglomerate .219 .414 0 n1a
Acquisition
3. Vertical .102 .303 0 n1a -.179
Acquisition
4. Related .670 .471 0 n1a -.756 -.482
Acquisition
5. Domain-Expanding .323 .468 0 n1a .756 .490 -.983
Acquisition
6. MBA (nooelite) .112 .315 0 .184 -.180 -.134 .164 -.164 ~
7. Elite Grad. Degree .290 .454 0 .033 .105 .128 -.078 .078 -.227 ~
(incI.MBA) g
8. Elite Grad. Degree .083 .276 0 I -.070 .129 .202 -.222 .222 -.107 .471
(cxcl.MBA)
9. Elite MBA .207 .405 0 I .085 -.068 .009 .044 -.044 -.181 .799 -.154'
~::r
10. Output Background .376 .485 0 I -. 151 .063 .164 -.168 .168 -.036 -.128 -.022 -.128
n·
II . Throughput Bkg. .065 .248 0 .036 .088 -.002 -.072 .072 -.068 -.025 -.050 .007 -.206
'"
12. Peripheral Bkg. .359 .480 0 I .141 -.139 -.114 .192 -.192 .043 .136 .065 .108 -.582 -.198 8-
13. General Mgmt Bkg. .198 .399 0 I .112 .035 -.047 .008 -.008 .034 .007 -.021 .022 -.387 -.132 -.373 ~
.0
14. CEO Tenure 8.14 6.80 0 38 -.012 .125 .028 .124 -.124 -.128 .103 .091 .053 -.066 -.032 -.057 .168 5.
15. Uncertainty 10.66 27.18 o 336.5 n1a .018 .157 -.114 .114 -.037 -.091 .145 -.105 .172 .035 -.144 -.045 .170 ~.
16. CEO Age 55.41 7.18 34 72 -.128 .047 .068 -.101 .101 -.133 -.040 .077 -.097 .050 -.001 -.017 -.041 .298 .012
a.
o
17. Sales 3434.46 5885.07 19.00 71600 .137 .184 .026 -.172 .172 -.568 .059 -.012 .074 .090 .026 -.068 -.043 -.007 -.039 .145 :s
18. Profits 219.59 421.38 -1100 4600 .162 .105 .005 -.088 .088 -.019 .029 -.060 .074 .108 .013 -.068 -.058 -.049 -.101 .124 .830
'"
!j
a Correlations greater than .11 are statistically significant at the .05 level. U\
CEO functional background may also vary by industry. Thus, in the absence of
industry controls, an observed relationship between CEO background and patterns
of acquisitions may be spurious. Industry was measured by a series of thirteen
dummy variables for the 2-digit SIC codes described earlier. CEO age controls for
the possibility that age is correlated with interorganizational networks or CEO
tenure. CEO age was obtained from the BW listings. CEO education level
(undergraduate degree, graduate degree, MBA) was included to control for
correlations of CEO networks with education level. Year of acquisition controls for
any macroeconomic factors, e.g., tax changes, GNP, that may affect the likelihood
of acquisitions, or certain types of acquisitions.
RESULTS
Table 1 presents descriptive statistics and correlations for all study variables except
year and the industry of the acquiring firm.
Table 2 presents the results of logistic regressions on whether a firm engaged in
a domain-expanding acquisition during the study period. For these models, the
dependent variable was coded 1 if the firm engaged in a domain-expanding (vertical
or conglomerate) acquisition, 0 if they engaged in a related acquisition. Modell of
Table 2 presents the effects for the control variables alone. As can be seen from this
model, larger firms are more likely to do domain-expanding acquisitions, less
profitable firms are more likely to do domain-expanding acquisitions, and there were
fewer domain expanding acquisitions during 1987, 1989, 1990, 1991, and 1992 than
1986 (the excluded year). Model 2 adds the effects of functional background. There
are no statistically significane relationships between CEO functional background
and the likelihood of a firm engaging in a domain-expanding acquisition.
Model 3 reports the results of the analysis of whether firms whose CEOs have
an elite graduate degree engage in more domain-expanding acquisitions than firms
whose CEOs do not have an MBA (HI). As predicted by Hypothesis I, CEOs with
elite graduate degrees were more likely to engage in domain-expanding acquisitions
than CEOs without elite graduate degrees. Model 3 reports results for elite graduate
degrees excluding elite MBAs. When we include elite MBAs with the other elite
graduates, the results are the same-the effect of an elite graduate degree on
domain-expanding acquisitions is positive (see Model 4). We thus combine elite
graduate degree and elite MBA in all further analyses. Additionally, supplementary
analyses reveal that an elite undergraduate degree produces the same effects-CEOs
with elite undergraduate degrees were more likely to engage in domain-expanding
acquisitions than CEOs with nonelite undergraduate degrees (details of these
analyses available from the authors). It appears that something unique to obtaining
an elite degree, which is not restricted to an elite graduate degree, has an effect on
engaging in domain-expanding acquisitions. This 'something' is likely to be the
social capital that resides in networks that come with elite education. Such networks
may be more powerful in graduate schools, but should also come with
undergraduate education.
While we had not hypothesized any effects for education level (undergraduate,
MBA), Model 3 also shows that firms whose CEOs have a non-elite MBA engaged
in fewer domain-expanding (and thus more related) acquisitions than firms whose
CEOs do not have an MBA.4 While we report all three educational variable results
Table 2. Logistic regression analysis of determinants of domain-expanding acquisitionsa
Variable (I) (2) (3) (4) (5) (6)

Output Background -.00 -.12 -.12 -.11 -.19
Throughput Background .20 .\3 .09 .18 .01
Peripheral Background -.01 -.10 -.12 -.05 -.11
Elite Grad. Degree (exc!. MBA) 1.49**
Elite MBA .35
MBA (nonelite) -.98* -1.00* -1.32* -1.32*
Elite Grad. Degree (inc!. MBA) .81** .84 1.29*
Uncertainty -.00
Uncertainty x Elite Grad. -.03
Uncertainty x MBA -.06
CEO Tenure .02
Tenure x Elite Grad. -.08*
TenurexMBA JJ7**
Control Variables"
CEO Age .02 .02 .01 .01 .00 .01
Year-1987 -1.31* -1.30* -1.25* -1.32* -1.58* -1.42*
Year-1988 -.63 -.62 -.57 -.67 -.58 -.79
Year-1989 -1.24* -1.23* -1.22* -1.35* -1.31* -1.47*
Year-1990 -1.37* -1.38* -1.22 -1.33* -1.25 -1.30*
Year-I991 -1.46* -1.45* -1.23 -1.36* -1.57* -1.44*
Year-1992 -2.45** -2.46** -2.32** -2.49** -2.41** -2.56*
Year-I993 -1.01 -.99 -.64 -.75 -.65 -.84
Sales (thousands) .13* .13* .12* .11* .12* .11*
Profits (thousands) -1.70* -1.70* -1.60* -1.50* -1.40* -1.50*
Intercept .06 -2.36** .38 .50 .91 .27
N 273 273 273 273 262 273

Chi-squared 97.94** 97.87** 107.31** 105.52** 99.65** 118.29**
-2 Log Ukelihood 248.4 248.3 238.8 240.6 229.1 227.8
% Cases Correcdy Classified .83 .83 .85 .85 .84 .86
*p<.05;**p<.OI, I-tailed tests for independent variables, 2-tailed tests for control variables.
aDependent variable is whether the focal firm engaged in a related acquisition (l=yes, O=no)
bResults for industry control variables are not reported, but are available from the author
in the same model (Model 3), separate analyses show that the statistical significance
of the results does not change when these variables are run separately. MBA is
negatively related, elite degree is positively related, and elite MBA has no
statistically significant effect on the likelihood of engaging in a domain-expanding
acquisition. Thus, we find evidence supporting a network perspective on education
(our elite graduate degree results). The elite MBA case, where cognitive and
network predictions oppose each other (elite education dictates more domain-
expanding acquisitions and MBA dictates more related acquisitions) produces no
statistically significant effect. This may be due to the opposing effects cancelling
each other in this situation. For our purposes, it is important to note that the network
effects are independent of the cognitive effects, supporting the validity of our
network perspective.
Model 5 of Table 2 examines whether uncertainty moderates the effects of an
MBA and an elite graduate degree on acquisitions (H3). Model 5 shows, contrary to
Hypothesis 3, that the main effects of education on acquisitions are hardly
moderated by uncertainty. In support of Hypothesis 4, however, Model 6 shows that
while firms whose CEOs have an elite graduate degree are more likely to engage in
domain-expanding acquisitions, this effect is weaker for long-tenured CEOs. Tenure
also weakens the negative relationship between a nonelite MBA and domain-
expanding acquisitions found in Model 3. Thus, it appears that tenure, but not
uncertainty, moderates the relationship between education and domain-expanding
acquisitions.
Models 1-3 of Table 3 present the results of analyses of vertical acquisitions.
For these models, the dependent variable was coded 1 if the acquisition was vertical,
and 0 if it was related or conglomerate. Model 1 shows, as predicted by Hypothesis
2b, firms having CEOs with output function backgrounds are more likely to engage
in vertical acquisitions than firms having CEOs with other backgrounds. Model 1
also replicates the effect of elite educational background (HI) (elite graduate
degrees are positively related to vertical acquisitions).
Models 2 and 3 examine whether uncertainty and CEO tenure moderate the
effects of output background on vertical acquisitions (H3 and H4). Model 2 shows,
in support of Hypothesis 3, the main effect of an output background on vertical
acquisitions is strengthened by uncertainty. While CEOs whose social capital is
shaped by output backgrounds are more likely to do vertical acquisitions, they are
even more likely to do so under conditions of uncertainty. Model 3 shows, contrary
to Hypothesis 4, that tenure has no statistically significant effect on the relationship
between output function background and vertical acquisitions. Thus, functional
background effects on vertical acquisitions are affected by uncertainty, but hardly by
tenure.5
Models 4-6 of Table 3 present the results of analyses of conglomerate
acquisitions. For these models, the dependent variable was coded 1 if the acquisition
was conglomerate, and 0 if it was related or vertical. Model 4 tests whether firms
whose CEOs have social capital that is shaped by peripheral function backgrounds
engage in more conglomerate acquisitions than firms whose CEOs have other
functional backgrounds. Model 4 shows, contrary to Hypothesis 2a, firms with
CEOs that have peripheral function backgrounds are no more likely to engage in
conglomerate acquisitions than firms with CEOs that have other backgrounds.
Model 4 also replicates the effects of elite educational background (HI) (elite
graduate degrees are positively related to unrelated acquisitions). Thus, the positive
effects of an elite graduate degree apply to both vertical and conglomerate
acquisitions. Models 5 and 6 of Table 3 show that uncertainty and tenure have no
statistically significant effect on the relationship between peripheral background and
conglomerate acquisitions. Thus, Hypotheses 3 and 4 get no support when looking at
conglomerate acquisitions alone.
Overall, then, Hypothesis 3 gets mixed support from these results. Uncertainty
does not appear to weaken the effects of educational background. It does not weaken
Table 3. Logistic regression analysis of determinants of vertical and conglomerate acquisitions
Vertical Acquisitions ConglOlrerate Acquisitions

(I) (2) (3) (4) (5) (6)
Output Background .19" .18" .17"
Peripheral Background .07 .34 .38
Elite Grad. Degree (incl. MBA) .91" 1.33" .n" .OS" . 14 .08"
MBA (nonelite) -1.24 -1.25 -1.26 -.66"· -.66* -.67*
Uncertainty .01 .00
Uncertainty x Periph. Bkg. -.04
Uncertainty x Output Bkg. .05"
CEO Tenure .03 .01
Tenure x Output Bkg. -.01
Tenure x Peripheral Bkg. -.02
Control Variables'
CEO Age -.02 -.02 -.65 -.01 -.00 -.01
Year - 1987 -.52 -.85 -.67 -.66 -.92 -.71
Year- 1988 -.53 -.66 -.58 .09 .08 .04
Year- 1989 -.39 -.62 -.66 -.79 -.86 -.75
Year- 1990 -.45 -.51 -.66 -.66 -.80 -.69
Year- 1991 -.60 -1.04 -.70 -.61 -.72 -.70
Year- 1992 -1.27 -1.18 -1.29 1.05 -1.77 -1.04
Year - 1993 -.62 .82 .39 .63 -.86 -.63
Sales (thousands) .01 .02 .OJ .11* .12 .12*
Profits (thousands) 1.70 2.00 1.90 1.40 1.20 1.10
Intercept 2.59 .57 .84 .26 .25 .32
N 273 262 273 273 262 273

Chi-Squared 44.98"* 40.86** 45.94*" 62.10"* 58.42** 61.86·
-2 Log Ukelihood 135.6 123.2 134.6 225.4 216.0 225.7
% Cases Correctly Classified .83 .83 .84 .80 .80 .79
*p<.05, * *p<.O I, 2-tailed tests.

IResults for industry control variables are not reported, but are available from the authors.
the relationship between functional background and conglomerate acquisitions_ It

does, however, weaken the relationship between functional background and vertical
acquisitions. Hypothesis 4 also gets mixed support. Increasing CEO tenure weakens
the effects of educational background on domain-expanding acquisitions. Tenure,
however. does hardly seem to affect the relationship between functional background
and vertical or conglomerate acquisitions.
DISCUSSION
Overall. the results provide support for the idea that the networks that come with
educational and functional backgrounds affect acquisitions. These background
effects are net of many control variables. We found evidence that the networks that
come with elite graduate degrees and certain functional backgrounds affect
acquisitions. These network effects are independent of the cognitive effects of such
backgrounds. We expected that an elite graduate degree would provide CEOs with
access to financial resources and far-reaching contacts in the intercorporate network.
In turn, we argued that these contacts would aid firms in completing domain-
expanding acquisitions (conglomerate or vertical acquisitions that take firms into
entirely new areas). Consistent with these arguments, we found that firms having
CEOs with elite degrees were more likely to engage in domain-expanding
acquisitions than those without such degrees. When transacting across industries
with unfamiliar actors of unknown reputation, an elite degree may act as an
important signal, both to target firms and potential investors, that the acquirer's CEO
is competent (D' Aveni 1990). The social capital inherent in the networks that come
with an elite degree may also provide information about acquisition opportunities in
different industries, information not available to those without such networks.
Alternatively, this finding may suggest that something about CEOs that obtain elite
degrees is related to such acquisitions, e.g., an ability to think more globally about
possible opportunities in other industries.
We argued that the social capital that resides in the networks of CEOs with
different functional backgrounds, would result in different types of acquisitions.
Indeed, we find support for the idea that output function work history is
systematically related to external networks and acquisitions. We expected that
output background would increase a CEOs awareness of and access to targets in
distributor and also possibly supplier industries. Consistent with this idea, we found
that CEOs with output backgrounds engage in more vertical acquisitions than CEOs
with other backgrounds. Such CEOs also tend to sit on boards of firms that are
unrelated to their own firms. We are not aware of any studies looking at the
relationship between functional background and interorganizational networks. Our
results suggest that such networks may have important effects on firm strategic
actions.
We had also, however, predicted that CEOs with peripheral function
backgrounds would be more likely to do conglomerate acquisitions than CEOs with
other backgrounds. We had predicted this on the assumption that peripheral
backgrounds would be associated with interorganizational networks that provide ties
to firms very different from the CEO's own firm, including ties to others in the
investment banking and legal communities. Our network analyses, however, showed
that finance CEOs tend to have ties to firms similar to their own firms.
Many of the CEOs with peripheral function backgrounds come from finance.
Taking a cognitive perspective on the relationship between functional background
and strategy, Fligstein (1990) argued that a finance background produces a 'finance
conception of control,' in which a firm is viewed as a bundle of assets to be bought
and sold. Because of their experience with financial analysis, peripheral function
CEOs prefer to evaluate and manage unrelated businesses, which requires running
businesses as a portfolio of assets. Fligstein found that finance CEOs tended to
complete more conglomerate acquisitions in the 1960s (Fligstein 1990). We, on the
other hand, find no evidence that CEOs with peripheral function backgrounds
completed more conglomerate acquisitions in our sample, even when we break out
finance backgrounds from law and other peripheral background categories. So
peripheral backgrounds are not associated with networks to firms in different

industries, and peripheral and/or finance backgrounds seem not to be associated with
engaging in certain types of acquisitions. We may have obtained different results
from Fligstein because the institutional pressures against conglomerate acquisitions
in the 1980s overcame any tendencies toward acting in accordance with a finance
conception of control, suggesting that perhaps the finance conception of control has
declined in importance in recent years.
Moderators of Background Effects

Our hypotheses for the effects of uncertainty on the relationship between
background and acquisitions were supported for output functional background, but
not for educational background. It may be that the reliance on cognitive shortcuts
and heuristics proposed to result from uncertainty only applies to functional
background because, when faced with uncertainty, functional background is more
salient to decision makers. Education occurred in the more distant past than the
experiences provided by working at a functional specialty. The social ties and norms
decision-makers have been exposed to on the job, which have been reinforced over
the years, are more salient than those gleaned from formal educational experience.
When faced with uncertainty, decision makers may indeed fall back on cognitive
shortcuts and social comparison processes (e.g., Festinger 1954; Cyert and March
1963; Haunschild 1994); however, recency and salience may dictate the set of
referents chosen.
We found that tenure weakened the effects of educational background, but had
no effect on functional background. The attenuation of educational effects over time
is consistent with our earlier arguments about long-tenured CEOs becoming less
sensitive to external cues (like external trends, norms, and ties with others), and
more attuned to their own views of a particular firm and industry than short-tenured
CEOs (see Hambrick and Mason 1984; and Davis, Tinsley, and Diekmann 1997 for
similar arguments). The attenuation of educational effects may also be due to a
weakening of the effects themselves-network contacts from one's elite education
weaken over time. This does not appear to be the case, however, for the contacts
from one's functional experience. Such contacts may be more frequently used,
resulting in effects that are more stable over time. Future research investigating
these different forms of ties and their stability and influence over time would be
valuable.
Limitations and Future Research

Some of our most striking findings concern the relationship between education and
acquisitions. An alternative explanation for our education effects, however, is that
firms that are likely to engage in acquisitions, or certain types of acquisitions (for
whatever reason), hire or promote CEOs with MBAs or elite degrees. However, we
think that this alternative explanation is unlikely because we controlled for several
factors that affect both acquisitions and the selection of CEOs with certain
backgrounds. For example, industry is likely to be correlated with the tendency to
pursue acquisitions and also with the tendency to select CEOs with certain
backgrounds. Additionally, if a firm were to hire (or promote) a CEO because the
CEO's background matched the acquisition desired, there should be a relatively
short period between the time the CEO was hired and the time the acquisition was
completed. This means that we should find effects only for short-tenured CEOs. Yet
we control for the interaction of CEO tenure and background, and still find effects
for background, making this alternative explanation less likely.
One limitation of our study is that top managers do not always have discretion
to act in accordance with their preferences. Although we find evidence that top
manager demographics affect acquisitions, there may be situations where these
demographic factors are more or less important. For example, the agency theory
literature shows that incentives can induce managers to act in accordance with
shareholder interests. Thus there may be situations where, for example, CEO
networks or normative preferences would result in a related acquisition, but the
CEOs incentives overcome such preferences and result in a conglomerate
acquisition. Boards of directors may likewise deter a CEO from acting in accordance
with his or her preferences.6 Thus, the moderating effect of factors like CEO
incentives and board power on the relationship between cognition, networks, and
acquisitions is an important topic for future research.
A second limitation of this study is that we have somewhat indirect indicators of
managerial networks. While we go further than many demographic studies in
actually measuring the intervening process variables between demographics and
firm outcomes, our interlock measures can only capture a subset of all managerial
networks. Showing that output background is associated with more ties to firms in
vertically related industries is important in supporting the validity of our network-
functional background relationships. However, interlock ties are probably better
indicators of the output background networks than they are of peripheral background
networks. This is because the presence of investment bankers and attorneys on firm
boards is relatively rare in our sample, so such ties are probably imperfectly
captured in the interlock data. It would be useful, therefore, to be able to more
directly measure the relationship between demographic backgrounds and their
associated networks.
Our results show that the relationship between functional background and firm
outcomes is strengthened with uncertainty. The relationship between educational
background and firm outcomes, on the other hand, is unaffected by uncertainty and
weakened by tenure. The moderators studied, uncertainty and tenure, have different
effects on education and functional background, suggesting that studying additional
moderators would be useful.
CONCLUSION
We sought to assess the effects of educational and functional background on
corporate acquisitions. Educational and functional backgrounds are demographic
variables, and demographic variables have been proposed to influence
organizational outcomes by affecting managerial cognition (e.g., Dearborn and
Simon 1958; Hambrick and Mason 1984; Fligstein 1990, 1991) and managerial
power (Fligstein 1990, 1991). Overall, our data are consistent with an alternative
view of the demographic variable-organizational outcome relationship: that
demographic variables are indicators of interorganizational networks and access to
information, which in turn affects outcomes like corporate acquisitions. Networks
give rise to interorganizational social capital, which directs acquisition activities
along certain channels. This is an important finding for both demography and
network research, demonstrating the link between demographic variables, networks,

and firm outcomes.
Our results also add to the growing literature showing that social context affects
acquisition activities (e.g., Fligstein 1990; Haunschild 1993; Davis et al. 1994). We
find evidence that firm managers carry with them interorganizational social
networks that affect their attention to information, and ability to engage in
acquisitions. Returning to our original question of what affects a firms choice of
strategic actions, we find evidence for the effect of both external (interorganizational
networks) and internal (managerial preferences) factors. We thus link the internal
and external perspectives on strategic action: external factors in the form of
interorganizational networks are linked to internal factors (CEOs) and predict major
strategic actions (acquisitions).
NOTES
I. Some of the finns in the sample completed multiple acquisitions. To test whether the
nonindependence inherent in these transactions affected the results, all analyses were conducted twice:
once with all acquisitions done by the multiple-acquisition finns included, and once where only one
randomly-selected acquisition was included for these multiple acquisition finns. The results do not vary
by whether one or all acquisitions for the multiple-acquisition finns was included, so results with all
acquisitions are reported.
2. The measure's strengths lie in the fact that it is I) specific to the finn; 2) involves multiple
respondents (analysts); and 3) allows uncertainty to be measured at the proper time prior to the
acquisition.
4. This result may be consistent with a cognitive argument for the effects of educational background on
acquisitions. Graduate business schools tend to socialize students into accepted business practice (Schein
1968; Van Maanen 1983), resulting in cognitive preferences for some practices over others. Individuals
with business degrees are likely to learn to be environmental scanners or monitors and thus prefer to
adopt practices that are currently popular. Since related mergers and acquisitions were popular and
nonnatively accepted during the period we study (Haspeslagh and Jemison 1991; Davis et al.l994), finns
having CEOs with MBAs should be more likely to engage in related acquisitions than finns having CEOs
without MBAs.
5. Uncertainty does not moderate the effect of an elite graduate education on vertical acquisitions,
either (details of this analysis available from the authors).
6. See Walsh and Seward (1990) for a discussion of situations where boards might have more or less
influence on management.
Public Service Organizations:
Social Networks and
15
Social Capital
•
Ewan Ferlie
ABSTRACT
There is evidence cited in the literature of a growing trend to network based styles of
management, especially in high-technology or expert based organizations. For such
network based organizations, the effective management of many external ties is a
major invisible asset or form of social capital. It is unclear whether this social capital
is owned by leading individuals or by organizations. The question of whether
corporate forms of social capital can be created and sustained across the
organization is in part an empirical one and not to be assumed through theory.
This chapter adds to the extant literature is two ways-first, by extending this
form of analysis to public-service settings such as health care organizations and,
second by providing recent qualitative empirical data. Public service organizations
remain of substantial scale and significance in many OECD countries, and they need
to be fully considered in the management literature.
Data are presented from U.K. health care organizations. An empirical study was
undertaken of a set of health care purchasing organizations rated as 'leading edge' by
peers. The data may not then reflect typical practice but rather as it is developing in
an innovative form. A series of interviews and visits took place; full notes were
taken and subject to contents analysis.
The study confirmed the importance of networks in health care management, so
that the possession of rich ties was a key invisible asset in 'getting the business
done.' The data does not confirm the conventional picture of total domination of
professional networks as more managerial networks were also in evidence. Social
capital belonged to individuals or autonomous groups at least as much as to the
corporate organization, although there were instances cited of attempts to create
more corporate forms of social capital.
Public Service Organizations: Social Networks and Social Capital - 285
However, the depth of the shift to network-based management can be

questioned. The transition was only partial, recent in origin, and often mandated by
the top. It is still premature to talk of a wholescale transition to network based
management in health care, as managerial practice in health care in faddish and
characterized by short life cycles.
INTRODUCTION
Markets, hierarchies, and networks are three ideal types of organizing (Thompson et
al. 1991). As these are ideal types and not empirical descriptions, in practice mixed
styles may be evident. There is evidence of a gathering trend toward more network
based styles of management within private sector firms (Axelsson and Easton 1992;
Nohria and Eccles 1992), especially in high-technology or expert-based
organizations. Fenton's (1996) literature review suggested that the New Competition
in the private sector may be resulting in widespread organizational transformations.
Associated with this is a predicted emergence of the N form (or network form)
organization with such 'signs and symptoms' as a flatter structure, self management,
diffuse decision making, accelerated information flows, and an emphasis on
learning.
For such network based organizations, the effective management of many
external ties creates a major invisible asset or form of social capital. The quantity
and quality of interorganizational ties between actors becomes a key factor of
production (Coleman 1988). Of course, this social capital is often created and
maintained by individuals rather than organizations. Effective organizations are
those that create a degree of cohesive unity and corporate contribution out of these
individually based networks. The depersonalization of network ties and their
aligning with corporate objectives then becomes a form of corporate social capital
creation (for a further discussion of the difference between social capital at the level
of the individual level and the firm, see Pennings and Lee, this volume) resulting in
a competitive advantage.
A reliance on interpersonal relationships may be supplemented by more formal
links such as joint board membership. However, interpersonal exchanges may
remain pervasive irrespective of formal structure, so we need to establish whether
these are motivated by corporate or particularistic objectives (also see Smith-Doerr
et al. this volume). Moving the focus of analysis from the individual to the
organization poses not only theoretical difficulties but also management questions
that need to be resolved empirically. Corporate social capital creation may be
especially problematic within certain organizational forms; for example, those of
very large size, those that display a range of occupational subgroups, where there is
much conflict between management and other employees, or where front-line
workers retain much autonomy over work practices, such as professional service
firms (see Pennings and Lee' s discussion of social capital within audit firms in this
volume). The organization may here be no more than a loose arena in which a large
number of subunits and subcultures coexist. The question of whether such corporate
forms of social capital can be created and sustained across the organization is in part
an empirical one and not to be assumed through theory.
This chapter adds to the extant literature in two ways-first, by extending this
form of analysis to public service settings such as health care organizations and,
second, by providing recent qualitative empirical data.
The Public Services Setting

First, the question arises as to whether similar shifts to network based forms of
social capital are evident within public service organizations as in private firms.
Public service organizations typically account for between 35 and 50 percent of
GDP (this definition includes transfer payments) in many OECD countries. They
discharge societally strategic functions such as the provision of health care,
education, income support, and law and order. While these organizations represent
important sites for analysis, they have not been fully explored in the managerial
literature on networks, certainly in Anglo American contexts. The chapters in
Nohria and Eccles (1992), for example, do not explore whether such shifts are
specific to the private sector settings analyzed.
One argument is that public service organizations have long been operating
along network based lines because of such factors as the historic lack of developed
market mechanisms; the existence of a range of strong professional groupings and a
relatively weak 'command and control' managerial spine; and the presence of
parallel organizational mandates, jurisdictions, and agencies that have required
interorganizational alliances. Corporate skills in managing these interorganizational
ties effectively represent a core form of social capital management as external actors
may be able to block key change objectives. larillo's (1993: 127) analysis noted
somewhat sharply: 'all the first writings on networks as forms of coordinating stable
relationships among different organizations were produced by sociologists working
outside the field of business. The networks they described had to do with
universities, health care centers and other non-profit making organizations. Business
writers did not find it a useful concept for many years.'
Some key questions emerge: to what extent is the management of public service
organizations network based? Are the issues and problems described in the private
sector networks literature also applicable to public sector organizations? Is there
evidence that the networking activity evident displays a corporate cohesiveness and
is related to corporate performance objectives?
Qualitative Empirical Analysis

This chapter also presents qualitative empirical evidence from one public service
setting (health care). Sometimes the concept of the N-form organization remains at
the level of a fashionable slogan, and detailed studies of how N-form organizations
function are still rare. The concept remains faddish, with the risk of
overgeneralization from restricted data. The focus is often on organization rather
than management, so that the managerial problems associated with N-form status
may be unexplored (Miles and Snow 1992).
Alongside a turn to theory (Salancik 1995), more finely grained empirical
analyses of particular organizational subpopulations are also needed. Two
approaches may be used in such empirical analyses. Quantitative sociometric
techniques have been used to model the formal properties of networks (Leenders
1995b). Other scholars have used a mix of quantitative and qualitative techniques
such as Provan and Milward (1995)'s close empirical analysis of network
effectiveness in community mental health systems. They identified features of
networks in that population associated with differential effectiveness, notably a
higher degree of network centralization through the presence of a core agency, high
direct fiscal control by the state, and high general system stability. Their results
questioned the effectiveness of highly decentralized systems without sources of
overall coordination.
We here focus on the subpopulation of health care purchasing organizations
(HMOs in American parlance) and are aware that conclusions should not be
extrapolated to other public service settings.
PUBLIC SERVICE SETTINGS: THE CASE OF HEALTH CARE

A rather separate literature on policy networks has developed within the public
administration and political science literatures. This literature sees government as
having a unique status in terms of its power resources (e.g., monopoly on the use of
legitimate force) and societal mandate so that public sector networks may well be
different in character (Kickert et al. 1997: 177).
Marsh and Rhodes (1992) explore the concept of policy networks (see also
Laumann and Knoke 1987; Bueno de Mesquita and Stokman 1994) within the
operation of government, drawing on pluralist and corporatist political theory. They
propose a typology of network types, ranging from more corporatist and highly
integrated policy communities to more pluralist and loosely integrated networks.
The strength of this political science literature is that it relates variation in network
structure to an analysis of the underlying distribution of power across the network. It
is less concerned than the managerial literature with identifying tactics that facilitate
effective action within the network.
An intermediate category of professionalised networks (Ham 1981) was seen as
applicable to health care settings, where narrow professional networks were highly
dominant. The implication was that they may serve professional rather than
corporate objectives. Wistow (1992) argued that the integration of the health care
network had historically been provided through dominant clinical professionals
insulated from other networks such as the managerial.
This chapter argues that there have been significant recent changes to the
network structure of U.K. health care purchasing organizations. Managerial
networks must increasingly be taken into account as well as clinical networks. There
is also preliminary evidence to suggest a modest opening up of health care networks
to influence from nongovernmental actors and civil society. Similar broad trends
toward cost containment, managerialization, and greater consumerism have been
reported in other American and European health care systems, and it would be
interesting to ascertain whether similar organizational shifts are evident there.
Recent Developments in Health Care: A Mixed Mode of Management

The present evidence from U.K. health care suggests that a pure paradigmatic shift
to a network-based ideal type of organizing is less likely than a period where mixed
modes of management coexist. Diversity is more likely than consistency, and
ideological pluralism more likely than hegemony. This is not an unusual situation in
management practice, but there are two schools of thought as to whether such mixed
modes represent a sustainable compromise or a set of internal contradictions.
O'Neill and Quinn (1983) argue that different models should not necessarily be
seen as contradictory and that apparent opposites could be reconciled. Rather they
broaden the repertoire of approaches available to managers as they move from one
situation to another, given that responses may be context specific. However, Miles
and Snow (1992) argue that a patchwork move to a networks mode of organizing
will fail to provide an organization with the coherent logic and set of skills needed to
make this distinctive approach effective.
The managerial agendas and practices of local health care organizations in a
publicly funded and accountable system such as the U.K. National Health Service
(NHS) are to a large extent mandated by central government, although enacted
locally. Recent legislatively driven changes (the NHS and Community Care Act
1990) introducing a quasi-market in health care were firmly driven from the top
(Ferlie et al. 1996). In the early 1990s, many NHS managers adopted market-led,
price-based and entrepreneurial modes of management. 'Macho purchasing' emerged
in some localities (Flynn, Williams, and Pickard 1996), so that traditional local and
informal networks dried up and transaction costs escalated. As the internal market
matured in the mid-1990s, crude classical models of contracting increasingly gave
way to more relational models of contracting.
This emergence of relational contracting began to highlight once again the
importance of informal interorganizational networks in the NHS. In the mid 1990s,
the center launched new policy initiatives consistent with the adoption of more
network styles of management. This marked a retreat from more aggressive market
based models of management and a turn back to a public health orientation. Recent
policy (Cm 3807 1997) signals the formal abandonment of the quasi-market model
and a stress on new core values of cooperation and the integration of care.
Network based initiatives have included the adoption of targets for improving
the public health that require widespread interorganizational working, the
development of enhanced primary care capacity in addition to hospital care, a
growing awareness of the need of health and social care agencies to work together to
provide long-term community care, and an increased stress on user involvement and
a more outward-facing orientation.
There appears to be no long-term or unambiguous shift to a pure new mode of
network based management. Rather, network, market, and line management
principles coexist and compete for influence. Each of the three models waxes and
wanes over time, with the networks approach finding most favor recently. However,
the life cycle of each favoured management approach seems to last no more than
about five years and is heavily dependent on central governmental sponsorship.
Skeptics argue it is still much too early to conclude that the adoption of a more
network-based model is an enduring trend, as it may yet tum out to be no more than
yet another passing managerial fad (Abrahamson 1991).
NETWORK BASED MANAGEMENT: SOME RECENT EVIDENCE

FROM HEALTH CARE
Methodology
Informed by these important developments, the author and a colleague were
commissioned by the National Health Service Executive (NHSE) to undertake an
empirical study of the management by networks phenomenon (1994-1995) (Ferlie
and Pettigrew 1996). An initial review of the literature was used to construct a
semistructured interview pro forma to test arguments. Visits were undertaken to nine
Health Authorities (purchasing organizations) across the country which were not
selected randomly but were nominated by experts in the Department of Health as
having the reputation for being at the 'leading edge' of development. The sample
included a mix of inner-city, urban and rural localities.
A typical program consisted of four to five interviews at headquarters in the
morning, combined with visits to one or two innovative localities or projects (e.g.,
local multiagency projects) in the afternoon, where further interviews took place. A
spread of respondents from different functions and agencies was accessed, resulting
in about seventy interviews in all. Both managers and clinicians were interviewed in
each authority. Full notes were taken at the interviews. Categories were then
developed from an analysis of the mass of interview notes, and documentation
collected. All data are thus qualitative in nature.
As access was accorded to 'interesting people in innovative localities' as judged
by expert reviewers, the data may not be typical of practice but rather reflect its
leading edge. Note also that data were gathered solely within purchasing
organizations, and it would be interesting to explore any differences with providers.
Empirical Evidence from Health Care Settings
The Networking Phenomenon Exists

The first conclusion was that network based styles of management are of substantial
and rising importance within the health care organizations studied. Specifically, we
found that the health care managers and clinicians interviewed were spending much
time in negotiating, persuading, and influencing partners in organizations outside
health care to increase their contribution to joint goals. Successful implementation of
a broad agenda for public health requires the support of many stakeholders far
beyond the formal health care system to work together. Another set of network-rich
settings has been uncovered empirically and in a different sector from that
conventionally studied in the management literature (Nohria and Eccles 1992).
Moreover, purchasing organizations were described as varying in their ability to
construct interorganizational alliances, and this skill was an important requisite for
social capital and organizational performance.
Network based management represented however a shift of emphasis rather than
a total change, and respondents talked of the need to operate in different modes in
different circumstances in order to create and maintain social capital. The caution
expressed by Miles and Snow (1992) in relation to the difficult coexistence of
different logics of managerial action needs to be recalled.
Most respondents agreed that networking with external stakeholders was now a
key managerial skill within health care organizations. However, some felt that this
had always been the case, citing traditional negotiations with other autonomous
public service agencies about boundaries, or the historic need to persuade
professionals (such as clinicians) or politicians before managerial action could take
place. Influencing had always been a more functional tactic than the exercise of
positional power. So constructing the 'right ties' in order to create social capital had
always been important and had become part of the managerial repetoire.
Nevertheless, there was more emphasis on networking than in the recent past, as
direct line management had fragmented. Reliance on the quasi-market and signals
from prices was still seen as marginal, and even as declining in importance as major
strategic change required an alliance between all key stakeholders to succeed.
When asked to characterize the interpersonal attributes and skills needed within
network based forms of management, 'trust,' 'reciprocity,' 'understanding,' and
'credibility' all emerged as important basic concepts in use. It is clear that there is an
important interpersonal component to network based forms of management, which
can make it vulnerable to the turnover of key people. What is interesting is whether
it is possible to institutionalize these characteristics at a collective level, so that they
become a form of corporate social capital, as some organizations were trying to do.
The alignment of incentives at an agency level is also important so that a
'win-win' situation can be established. Achieving such interorganizational trust may
be more difficult than building inter personal trust, yet vital if alliances are to
survive the departure of key individuals.
Changing Internal Management Structures and Styles

The networking phenomenon has had internal as well as external consequences.
Specifically, we found that a number of respondents reported moves internally away
from reliance on established functional hierarchies and toward looser and network
based or (cautiously) matrix management styles of working. Active organizational
development (OD) function was at work in some of the sites, trying to reshape
organizational style in this direction. A move from organizing along uni functional
or professional lines to more team-based and mUltidisciplinary approaches was
widely reported. This had important implications for clinicians who were working in
a more corporate and managerially orientated mode than previously. They were
increasingly immersed in the corporate core of the organization as well as
undertaking their traditional negotiating role with clinical subgroups.
As a result, respondents reported that they were being operating with a more
mixed portfolio of tasks (,matrix hoppers') and ad hoc teams. Some managers
reported problems with the lack of a clear structure, particularly at middle
management level where the pace and complexity of work has been 'upped.' Such
matrix management approaches depend on strong channels of informal
communication to make them work and a movement away from the 'meetings
culture' that had previously been dominant in these organizations. The introduction
of e-mail is seen as having a helpful decentralizing effect. Critics argued that
sometimes these organizations were changing their internal management styles less
in practice than they proclaimed in theory.
A growing number of organizational mergers were providing an opportunity to

move away from the old organizations, based on strong chairs and CEOs and
displaying a 'command and control' culture. It was commented that the new
organizations felt more 'diffuse' and 'looser' than the old. Nevertheless, these
mergers were generally imposed from above rather than growing out of locally
based ties and alliances.
Evolving CEO Role and Style

What is the role of the CEO in a more network based form of organization? Clearly,
the role is different from the role played in a 'command and control' organization as
the apex of all reporting lines and sole crossover point for a number of vertically
organised functions. However, the distribution of power and lines of upward
accountability are not entirely diffuse, and so CEOs were grappling with how to
retain a shaping, influencing, and indeed accountable role while also 'letting go.'
One role is to build and maintain networks through making external contacts
and then retain an overview across the multiplicity of networks that emerge. It is the
CEO who above all attempts to transform social networks into corporate social
capital. The CEO might be expected to take a view about the value added from each
network, deciding to boost involvement in one network, while withdrawing from
another. The consequence may be less CEO time spent in the office and a lessened
ability to react immediately to the short-term crises that constantly flare up in health
care. The CEOs interviewed felt themselves to be 'outward facing'-one estimated
that she spent only 25 percent of her time at HQ.
A second key role of the CEO is to give the collection of networks strategic
direction, to reinforce core purposes and tasks, to ensure corporacy and to assess
performance against key objectives. The CEO may then embody the performance
management orientation within the network. There was a concern expressed by
respondents that network based forms of management might degenerate into the
social liability of perpetual 'talking shops' and that key purposes could be lost.
A third role of the CEO could be to institutionalize strategic alliances by moving
on from a reliance on interpersonal trust to a deeper level of interorganizational
trust. CEOs may act to construct win-win situations, to broaden channels of
communication and joint working, to extend and deepen the group of personnel
crossing boundaries, and to build an internal culture more receptive to alliances. At
present, however, many of the alliances studied remained at the interpersonal level
and have not been progressed onto the interinstitutional level. Clinical networks and
alliances still remain distant in many cases from the corporate agenda.
In essence, the CEO plays a key role in corporate social capital creation and
aligning sources of human capital with the corporate agenda. Can clinicians be
drawn into the corporate agenda? Can cooperative relations be established with
other care agencies, particularly their senior management? Can research and
education alliances be created with senior personnel at local universities? Such
alliances are so complex and disparate that they may require committed group
leadership from within the health care organization to be progressed effectively.
Creating and maintaining this networking team is then a strategic role for the CEO.
Clinical Networks: Less Dominant But Still Important

Contemporary health care networks are more polycentric than the traditional
political science literature suggests. A variety of networks were found to be in
operation, some of them clinically based and difficult for management to penetrate.
Nevertheless, the networks contained more of a managerial component than
assumed in the earlier literature (Wistow 1992), with a (lay) CEO sometimes
assuming a metamanagement role.
Within the surviving clinical networks, the public health function displayed an
important coordinating role, although its strength at the core of the organization
varied. It often emerged as the natural focus for maintaining relationships with
clinicians in the provider units. Directors of Public Health (DPHs) also reported the
need to operate as 'persuaders,' communicating effectively with a variety of different
groups. They need to manage and secure change in practice, not only within the
public health function but more broadly. DPHs were seen in this respect as needing
generic skills, flexibility and an ability to switch networks and tasks in a 'reedlike'
way as well as narrow technical skills. They themselves wore two hats, as clinicians
and as executive directors of a corporate board.
These findings indicate a substantial change in network composition in health
care when set against earlier work, and in particular the emergence of newly created
managerial networks needs to be set alongside traditional clinical networks.
Some Opening Up: Networks from Below

One of the criticisms of network based management is that it operates in a closed
manner, excluding potential stakeholders, and privileging powerful groups such as
clinicians as the historically dominant elite, perhaps now joined by management as a
rising corporate elite. Community interests may, however, still remain absent from
the decision-making networks (to use Alford's 1975 categories). Closed networks
may militate against social change, especially of a discontinuous nature (Gargiulo
and Benassi, this volume, term this 'the dark side' of social capital).
Interestingly, there was-against this proposition-evidence of a revival of
community development ideas, projects and models and an attempt to open up
networks to influence from below. Influential in the 1970s, community development
work lost ground in the 1980s but now seems to be climbing back up the agenda.
These experiments were often well resourced and backed by the top even if
primarily dependent on local product champions. Initiatives were concentrated in
localities where there was evidence of deprivation, social isolation, and lack of
access to health care. Initiatives were associated with the presence of a strong public
health agenda or champion and could generate considerable conflict between very
different stakeholders. The proposition that a network based mode of management
would result in a bland or consensual management style was not sustained in these
case examples, as the social liability of sharp conflict between the governmental and
non governmental sectors could arise.
This very early evidence suggests some modest broadening of networks may be
occurring, as health care purchasing organizations increasingly face outward to local
civil society. This broadening was of a partial nature, and its force should not be
overstated.
A Web of Interagency Alliances

The diffuse, shifting, and sometimes transient nature of many of the health care
networks uncovered was striking. This contrasts with some studies of the private
sector that sometimes stress a narrower range of strategic alliances, often driven by
resource dependency (see Barley et al.'s 1992 analysis of strategic alliances in
commercial biotechnology).
The health care organizations studied were actively engaged-with varying
degrees of success-in establishing alliances with many other organizations,
including public agencies, private firms, and civil society. These ties took the form
of formal interorganizational linkages as well as interpersonal networks (as in the
analysis of the biotechnology sector by Smith-Doerr et aI., this volume). These
linkages were not so much driven by a search for financial flows as a desire to
influence policy implementation. We are not so much dealing with a single strategic
alliance directed from the top but rather a dense web of interagency alliances, with
different leaders in different relationships.
Managerial Problems and Challenges of Network Based Organizations
Networks and Performance

A major criticism encountered in the field was that network based forms of
management were highly time consuming and did not lead to tangible output. What,
in other words, are these networks for? There was a perceived danger that the means
of managerial process could drive out the ends of delivering better services. Some
alliances (for example, joint planning with social care agencies) are mandated by
legislation, and therefore there is a debate as to whether participation may be
ritualistic or whether such imposed structures may nevertheless reflect objective
requirements of work processes such as inter agency referral flows (Leenders
1995b).
Managers in the study wanted to be clear about the social capital that the
network was intended to deliver. The combination of network based forms of
management and a strong performance management orientation-with its focus on
quantification and short-term target setting-now evident in health care
organizations could represent an uneasy and volatile combination. Networks could
easily proliferate, soaking up a lot of time while delivering very little.
There was also a concern about lack of clarity in present network forms. While
senior management often welcomed moves to network based forms of management,
middle management was reported as feeling under threat and finding it difficult to
cope without a structured framework. A high level of turnover at middle
management level was apparent in some of the health care organizations visited.
Network Sustainability
Another question relates to the sustainability of some of these networks that were
time consuming to create and maintain but could collapse if they were not seen as
delivering by participants. Participation was usually instrumental rather than
altruistic and depended on what kind of social capital was on offer to each player.
Networks could also rapidly disintegrate if known players were pulled out,
destroying the social capital that was contained within them: 'a face that you know
and recognize is key' said one respondent. So networks may be seen as a volatile
mode of organization, with short life-cycle effects, especially where there is no
foundation of early successes or 'quick wins' to institutionalize the level of
commitment.
Many of the networks studied needed a focus or a broker to animate and sustain
them (Thorelli 1990), given the continuing need for network mobilization. This is an
influencing rather than a hierarchical role, so that the question of who is filling these
broker roles is critical. Are people with the networking skills coming forward, or is
the pack of existing personnel simply being reshuffled without thought being given
to selection against job and skill specifications?
Networks and Change

A heavy reliance on preexisting or closed networks is often associated with
continuity and maintenance management rather than innovation or change. Network
based management may here tend toward conservatism (see Garguilo and Benassi
this volume), while organizations are now judged in performance terms on their
ability to deliver an ambitious change agenda. The formally declared purpose-and
in at least some cases informal intent--of many of the health care networks studied
was to create service and indeed behavioral change, so a reliance on closed and
traditional networks would be dysfunctional.
So players may need to engage far more openly than hitherto in network
building activity to open up networks and bring in new players. Initially, they may
need to sell participation in the network to partners not previously involved.
Achieving behavioral change in local communities may be dependent on an ability
to 'script in' the nongovernmental sector. As reciprocity is a fundamental norm of
network life (Leenders 1995b: 193-197), official organizations may here need to
give before they can expect to receive from an initially suspicious nongovernmental
sector.
HRM Implications and Issues

The HRM implications of the move to network based forms of management have
not been fully addressed. Brass and Labianca (this volume) have recently argued
that the HRM function may have a strategic role as human and social capital brokers
and reconfiguring human resources in the light of rapidly changing demands. So
what sort of skills and attributes are needed to make these forms of organization
work? When asked, respondents were able empirically to identify key attributes and
skills, including strong interpersonal, communication and listening skills, an ability
to persuade, a readiness to trade and to engage in reciprocal rather than manipulative
behavior, and an ability to construct long-term relationships.
A mutual orientation and reciprocity emerge as key concepts in network based
organizations. The exercise of positional power gives way to influencing skills, such
as an understanding of how best to manage group decision making (Kanter 1989). It
is important to foster relational sentiments (e.g., indebtedness) so that exchanges are
not seen as one off transactions (Powell 1990). Such norms of obligation and
cooperation imply the continued existence of a community of shared values across
the health care field. Within such clannish organizational fields, contractual
opportunism is unlikely.
In addition, respondents pointed to the need for successful networkers to possess
the ability to cross a variety of occupational, organizational, social, and political
boundaries, an ability to speak different languages, and an ability to act as an
interpreter between different groups and be credible with a range of different groups.
A single master culture is not apparent within the health care sector, but rather it
continues to display very different subcultures. If anything there is a trend away
from a dominant professional network to a (modestly) broader and inclusive set of
networks. In Hofstede's terms (1994), participants in such polycentric organizations
need to cross conventional boundaries to manage intercultural encounters. It also
implies a need for highly developed analytic and diagnostic skills so that networkers
are aware when they are crossing subcultural boundaries and are able to switch
language and behavior patterns as appropriate.
A third cluster of desired attributes was identified in the following areas:
tolerance of high levels of ambiguity and uncertainty, a long-term as well as a
short-term view, a good strategic sense, an ability to reflect on experience and
conceptualise, a capacity to learn quickly and to adapt in new situations.
Managers in such network based organizations will need to cope with high
levels of uncertainty and ambiguity. They will face a greater pressure to innovate
and hence have to acquire an accelerated capacity to learn (Powell and Brantley
1992 examine the relationship between network forms and organizational learning in
relation to biotechnology).
Is accelerated learning explicitly recognized as an objective in network based
organizations? Learning patterns may remain at an individual level, but where there
is a corporate effort to develop individuals so that they are able to learn and adapt
more effectively, then it is legitimate to talk of a corporate social capital creation
initiative. Extensive corporate development programmes were evident in some of
the sites visited in an attempt to remould the working styles of key personnel.
However, effective learning may also depend on the capacity of individuals to
reflect on their experience, to conceptualize experiential knowledge, and to adopt
new skills and behaviors: an ability to impart knowledge to others, to act as a teacher
or a mentor, an ability to transfer knowledge from one setting to another, and an
ability to convey requisite standards and attitudes to others inside and outside the
organization (norm setting).
CONCLUDING DISCUSSION
Whose Networks? Whose Social Capital?
This study confirmed the importance of networks in health care management, so that
the possession of rich ties with actors in parallel organizations was seen as a key
invisible asset in 'getting the business done.' Markets and hierarchies both remained
weak modes of organizing, and so network based assets were of key importance.
Our data did not confirm the conventional picture of total domination by
professional networks (Wistow 1992). Clinical networks retain great importance and
centrality but have been somewhat broadened by the creation of other networks.
One view is that the introduction of the internal market in health care was
significant in beginning a process of deinstitutionalisation where the old professional
networks no longer have quite the same power as hitherto (Morgan and Maddock
1997). Managerial networks emerged as of greater importance than predicted. In
addition, some community-based networks were also emerging in a bottom-up
fashion.
The question is: who owns this social capital? Key individuals were seen as
particularly effective as boundary spanners so that in these cases capital accrued to
particular people who could well be head hunted elsewhere, on the grounds that they
had a track record in creating effective alliances. Other networks were contained
within particular subgroups (e.g., clinical networks) that it was difficult for
management to penetrate. Such social capital belonged to a group rather than an
individual, but hardly to the organisation.
However, there were also instances cited of attempts (to put it no higher) to
create forms of social capital at the level of the organization. Senior management
played a strong role in creating and managing a number of the networks uncovered,
including relationship building with managerial personnel in other organizations, in
the interests of the corporate business agenda. Clinical networks could be managed
through the public health function, which stood at the interface between managerial
and clinical systems. The fusion of managerial and clinical networks is a difficult
area that has been handled through the creation of special clinicaVmanagerial hybrid
roles. The Director of Public Health thus had a key role in taking the corporate
agenda to clinical networks. Many of the board level personnel had been subjected
to extensive management development and team-building interventions in an
attempt to shift them into more corporate modes of thinking and doing.
How Significant is this Shift?

The depth of the shift to network based modes of management in health care
purchasing settings can be questioned. The transition to network based forms of
management was found to be only partial, recent in origin, and often mandated by
the top. It is not yet clear what its long-term significance may be, whether there is
substantial ownership at local level, whether it is no more than a passing trend, and
whether the distinctive skill requirements of network based management are being
seriously addressed. It is premature to talk of a wholescale transition to a N-form of
organization in health care as managerial practice in health care is faddish and
characterized by short life cycles.
Currently the health care system is in mixed management mode. The problems
of such mixed modes of management have been commented on by Miles and Snow
(1992) as they are seen as creating an idiosyncratic system, highly dependent on a
few key individuals or units to perform. Rhodes (1997) also argues that mixing
different governance structures may be like trying to mix oil and water. The
competing values framework, by contrast (O'Neill and Quinn 1983), takes a more
optimistic and pluralist perspective. Further longitudinal studies are required to
establish whether the present mixed pattern continues and what the associated
management challenges may be.
The breadth of the shift across public service settings should also be examined
further. Developments in one set of organizations (health care purchasers) may be
specific, and others elsewhere may not be experiencing the same shifts. Further
detailed empirical work is required across different public service organizations.
Network Based Management: Not an Alien Practice

Can managerial theories and practices developed in one sector be usefully
transferred across to another sector, or are radically new concepts required? Public
administrative theory argues that the nature of the tasks undertaken in the two
sectors is fundamentally different (Pollitt 1990; Hood 1991; Ransom and Stewart
1994). Some have seen the rise of the New Public Management as a attempt to move
the public services 'downgroup' (Dunleavy and Hood 1994): that is, making the
public sector less distinctive as a family grouping of organizations. Other
organizations in essence converge on the model of the private corporation, reducing
intersectoral variety.
Our position (Pettigrew et aI. 1992; Ferlie et aI. 1996) is that there are both
differences and similarities between the two sectors that need to be disaggregated.
Up to the mid- 1980s, the dominant problem was one of isolation: there was a too
ready presumption of difference. Since the mid 1980s, the problem has been the
overmechanistic transfer of practice from private to public sector. There is here too
ready a presumption of similarity. Useful theories are those that can take account of
intersectoral differences as well as similarities. Indeed, those that do not are harmful,
resulting in imposed change that fails to root and that has to be unpicked.
So are we to see network based management as no more than the latest private
sector managerial fad, inappropriately exported into the public sector? It is rather a
very special case of a managerial practice originally evident in public service
settings (Jarillo 1993), later diffused into the private sector, only to be reexported
back. It can therefore 'fit' public service work to a greater extent than other
contemporary managerial 'products' (e.g., BPR). Some of the forces favoring the
emergence of the network form identified by Castells (1996) in the private sector-
including but going beyond the rapid proliferation of informational technologies-
are also sector neutral and apply equally to public service organizations.
Our data confirmed that many health care managers saw a networks based
approach as always strongly developed within their sector. Flynn et aI.'s (1996)
analysis of community health services analyzed them as naturally organised through
multidisciplinary networks, a form of coordination eroded by the importing of alien
concepts of contracting and the market. Morgan and Maddock (1997) similarly point
to the erosion of historic professional networks and the introduction of 'hard'
contracting styles, which may however fail to institutionalize themselves.
So a network based form of management could be appropriate within public
service as well as private sector settings: in a sense, it was invented there. This
would predict rapid diffusion of the networks approach within public services as it is
seen as 'fitting.' This is an empirical proposition that requires testing. We also need
to know more about who owns these invisible forms of social capital and in
particular, the degree to which they adhere to corporate tearns rather than
individuals.
16
The Dark Side of Social Capital
•
Martin Gargiulo
Mario Benassi
ABSTRACT
Research on social capital has stressed the advantages that networks can bring to
managers and other economic actors. The enthusiasm with this 'bright side' of social
capital, however, neglects the fact that social bonds may at times have detrimental
effects for a manager and produce social liability, rather than social capital. This
chapter tries to correct the optimistic bias by looking at the 'dark side' of social
capital. Continuing benefits from social capital require that managers can adapt the
composition of their social networks to the shifting demands of their task
environment. This often implies the ability to create new ties while lessening the
salience of some of the old bonds-if not severing them altogether. Available
evidence, however, suggests that this ability may be encumbered by the same
relationships purportedly responsible for the prior success of the manager. When and
how this may happen is the central question we address in this chapter. We argue
that strong ties to cohesive contacts limit a manager's ability keep control on the
composition of his network and jeopardize his adaptability to changing task
environments, which damage the corporate social capital of the organization. We
test our ideas with data on managers working for a special unit of a high-technology
firm operating in Europe.
INTRODUCTION
Social capital has become a ubiquitous notion in the study of organizations. Despite
differences in emphasis adapted to specific research agendas, most treatments
coincide with the general definition of social capital as ' ... the sum of resources,
The Dark Side of Social Capital - 299

acquaintance and recognition' (Bourdieu and Wacquant 1992: 119; see also
Bourdieu 1980, for an early formulation). Behind this compelling metaphor, a rich
stream of. research highlights how social networks can enhance the ability of
individuals and organizations to attain their goals. Scholars have identified two main
mechanisms that account for the positive impact of social networks on individual
and organizational action. First, networks facilitate access to information, resources,
and opportunities (Grano vetter 1995; Campbell, Marsden, and Hulbert 1986; Flap
and De Graaf 1989; Coleman 1990; Burt 1992, 1997; Podolny and Baron 1997).
Second, social networks help actors to coordinate critical task interdependencies
and to overcome the dilemmas of collective action (Blau 1955; Pfeffer and Salancik
1978; Kotter 1982; Gargiulo 1993; Gulati 1995a). The resources and information
embedded in networks of interpersonal and interorganizational relationships can
help managers and organizations to attain their goals, and hence they can be
discussed as 'corporate' social capital (Gabbay and Leenders, this volume).
Faithful to an agenda that vindicates the importance of social ties in modern
economic transactions (Granovetter 1985), sociologists have largely stressed the
advantages that embeddedness in social networks can bring to actors engaged in
such transactions (e.g., Coleman 1990). This enthusiasm with the 'bright side' of
social ties has led sociologists to disregard the possibility that social bonds may at
times hinder, rather than help, an actor's ability to pursuit his interests (Portes and
Sensenbrenner 1993). The instrumental value of social ties rests on the good match
between the resources needed by an actor and the resources he could access through
his contacts. To maintain this match, and thus maintain his stock of social capital,
the actor should be able to adapt the composition of his social network to fit his
changing needs. Available evidence, however, suggests that an actor's ability to
maintain his social capital may be encumbered by the particularistic demands posed
by the same relationships purportedly responsible for the initial success of this same
actor.
This chapter tries to correct the optimistic bias of the existing research by
looking at the 'dark side' of social capital. Building on existing theory, we argue that
the network that provided a manager with social capital might also limit his ability
to change the composition of this network as required by his task environment. In
these circumstances, the network that provided the manager with social capital
would no longer be an asset but rather a liability that impedes successful action.
When and how this may happen is the central question addressed in this chapter.
Our answer focuses on the obstacles that strong, cohesive social bonds may pose on
the adaptation of social capital to changing task environments. Evidence comes from
managers operating in a special unit of a European subsidiary of a high-technology
firm. The analysis emphasizes how the structural characteristics of the managers'
social networks thwarted the adaptation of their networks to a fluid organizational
context.
THE TWO SIDES OF SOCIAL CAPITAL

Managers bring to their job more than the skills they have accumulated through
years of education and experience. They also bring the assets they can procure
through their social networks. Economists refer to the assets embodied in the
manager's skills as his 'human capital' (Becker 1964). Sociologists coined the term
'social capital' to designate the assets tied to the manager's social network (Bourdieu
1980; Coleman 1990).
The compelling metaphor embodied in the notion of social capital has brought
together a number of research streams that focus on the beneficial effects of social
networks in economic exchanges. Different researchers have emphasized different
aspects of social capital, albeit not always making explicit reference to the concept
as such. In his comprehensive review of the literature, Gabbay (1995, 1997)
distinguishes between approaches that focus on the resources controlled by the alters
from those that stress the relationship between ego and the alters. Resource-based
approaches are dominant in studies of individual success in labor markets (Lin,
Ensel, and Vaughn 1981; Lin and Durnil 1986; Flap and De Graaf 1989), but these
studies have been also approached from a relational perspective (Granovetter 1973).
The focus on the structure of the relationships dominates analysis of managerial
careers (Burt 1992, 1997; Podolny and Baron 1997). The relational approach is also
dominant in studies of inter-organizational cooperation. Networks can help
organizations seeking to form partnerships by providing efficient access to
information on the availability, competencies, and reliability of potential partners,
thus lowering searching costs and alleviating the risk of opportunism (Granovetter
1985; Powell and Brantley 1992; Gulati 1995a). Closely related to this last point,
scholars have also demonstrated how social networks have a crucial role in helping
managers and organizations to reduce the uncertainty that results from task
interdependence (Blau 1955; Pfeffer and Salancik 1978; Kotter 1982; Gargiulo
1993).
The interest in the role of social capital as relationships in overcoming
coordination problems has been recently fueled by the growing importance of flatter
structures, teamwork, entrepreneurial initiative, and decreasing reliance on authority
relations to handle interdependence (Kanter 1983a; Ditcher 1991; Baker I 992a;
Sheppard and Tuchinsky 1996). This new environment poses an important challenge
for managers. While the new organizational blueprint gives managers considerable
more latitude to perform their role, it also increases the uncertainty surrounding their
task and careers. Managers have to handle a growing number of task
interdependencies that link them to people outside their line of authority. Since the
formal structure offers little assistance in dealing efficiently with these 'lateral'
interdependencies, managers have to find alternative ways of coordination to
accomplish their task effectively. Social capital becomes a strategic tool to add value
in this new organizational environment (Nohria and Eccles 1992).
The distinction between 'resources' and 'relationships' helps clarify a key aspect
of social capital, often ignored in the literature (see Araujo and Easton, this volume).
In fact, social capital encompasses both the resources controlled by the actors and
the relationships linking ego to those actors. As Gabbay (1997) adequately points
out, resource-based approaches implicitly assume the alter's motivation to assist
ego, while relational-based approaches assume that ego would choose the 'right' alter
to include in his network (e.g., Burt 1992: 13), restricting the causal force to the
strength and the structure of the relationships. Yet, effective social capital requires
The Dark Side of Social Capital- 301
both alters with the resources needed by ego and a social structure that facilitates
ego's access to those alters. Although analytically distinguishable, both resources
and relationships are essential to the notion of social capital. Indeed, the main theme
of this chapter is that the interplay between the inertial tendencies of some social
structures and the dynamic character of the resources needed by a manager may turn
social capital into a liability that hinders, rather than help, managerial action.
By demonstrating the positive role of social relationships in modern economic
life, sociologists have contributed to a time-honored research program revitalized by
Granovetter (1985) with his seminal essay on the social embeddedness of economic
transactions. While economists have typically ignored social relations or have
treated them as an obstacle to attain economic rationality, sociologists have shown
that economic rationality can be actually enhanced by embedding transactions in
social networks that facilitate trust and diminish the risk of opportunism. In pursuing
this research program, however, sociologists have often disregarded a well-known
duality also stressed by Granovetter (1985) in his essay: social structures can be both
a source of opportunities and a source of constraint for individual behavior. As one
might expect from the positive tones of the metaphor, the bias towards the favorable
effects of social structures is especially apparent in the discussions of social capital
(e.g., Coleman 1990: 300). This enthusiasm with the 'bright side' of social capital
has led sociologists to disregard the 'dark side' of social relationships economists
like to emphasize-that is, that social bonds could be an obstacle to pursuit
economic interests.
Despite this pervading optimism, the evidence on the constraining effects of
social bonds on individual action is not missing. While most of this evidence may
come from studies that portray the supporting and yet oppressive effects of dense
social networks in small communities (e.g., Fischer 1982), the dark side of social
capital has been also stressed by sociologists analyzing economic behavior. Thus,
Portes and Sensenbrenner (1993) review various studies describing ethnic
entrepreneurs suffocated by the particularistic demands posed by the same strong
social ties purportedly responsible for facilitating their initial access to essential
resources. Summarizing their analysis of these findings, Portes and Sensenbrenner
(1993: 1341) propose a direct relationship between the amount of social capital
initially available to successful entrepreneurs on one hand, and the level of
particularistic demands and restrictions to freedom of action placed on those
entrepreneurs, on the other.
The theme behind these warnings is one familiar to economists that emphasize
the negative effects of social bonds: the maintenance of social capital entails
honoring obligations that may conflict with the pursuit of self-interest. Confronted
with this problem, some sociologists have come to accept that actors may at times be
better off with the impersonal, 'arm's-length' transactions glorified by economists
than with the embedded exchanges advocated by their discipline (e.g., Uzzi 1997a,
this volume). Yet, the critique to the one-sided view of social capital goes beyond
the dichotomy between 'embedded' and 'arm's-length' transactions that typically
confronts sociologists and economists. The restrictions that strong social bonds may
pose on an actor may hinder not only his attempts to substitute arm's-length
transactions by previously embedded ones, but also attempts to initiate new
embedded transactions. In other words, the network that provides the actor with
social capital may also encumber his ability to change the composition of this
network as required by changes in his task environment. The tension here is not
between embedded and arm's-length transactions, but rather between the currently
available social capital and the social capital required to cope with the new demands
or opportunities perceived by the actor.
Perhaps one of the reasons for which scholars have largely ignored the dark side
of social capital is related to the specific conditions in which its negative effects may
become apparent. The dark side of social capital is not always consequential; rather,
its effects will be noticed only after the actor reaches a point beyond which the
resources available to him through his current contacts are no longer adequate. The
entrepreneurs discussed by Portes and Sensenbrenner (1993) did benefit from the
social capital initially available to them, but the obligations that resulted from those
benefits curtailed their subsequent ability to pursue other business opportunities. It
was the change in the entrepreneurs' opportunities-which may have resulted from
their initial success-that made the constraining effects of social bonds
consequential. Since the instrumental value of social capital lies on the match
between the resources needed by an actor and the resources provided by the actor's
contacts, changes in the actor's task environment may require changes in the
composition of his social network. At least in this respect, social capital is similar to
human capital. Changes in task environments may render obsolete current skills and
capabilities, forcing people to make further investments in education.
Investments in social capital, however, are substantially more complex than
investments in human capital (Coleman 1990). While a person can typically acquire
new skills without having to discard previous ones, the same is not always true for
social capital. The maintenance of social capital requires investing time and energy
in one's contacts. Since people-and managers are not an exception-have a limited
amount of time and energy, pressures to maintain relationships that are no longer
advantageous may hinder the ability to cultivate other relationships necessary to
renew the managers' social capital. The ability to withdraw from business
relationships that are no longer advantageous has been often recognized as an
important factor in the adaptability of managers and organizations to changes in
their environments (Miles and Snow 1992). Paradoxically, the better the fit between
an actor's social network and his current environment, the harder will be to adapt
this network to a new environment (Uzzi 1997a).\ Changing the composition of
social capital often implies creating new ties while lessening the salience of old
bonds-if not severing them altogether. The more intense and productive the ties
with the old contacts were, the more difficult will it be to part with those
relationships. Such is the paradox of social capital: the brighter its bright side, the
darker the potential effects of its dark side.
The mechanism behind this paradox is rooted in the very logic of reciprocity
that turns relationships into the assets that form social capital (Coleman 1990).
Reciprocity is more than repayment or the last favor received: it is a general norm
that prescribes a certain type of behavior towards relevant others: takers ought to be
givers. The norm of reciprocity may oblige a manager to assist a contact, even if he
expects few benefits from future exchanges with this particular contact. This is
especially the case if this contact has been rather helpful to the manager, or to
somebody he is indebted to. Failure to reciprocate may result in strong sanctions and
in a serious damage to his reputation as a trustful contact, a damage that can be
consequential for the manager's ability to create new ties. It is precisely this failure
to reciprocate what we often disapprove of the 'instrumental' individual, the person
who cuts the ties of obligation to the group once he sees no further benefit from
exchanges with its members. In a culture of 'generalized exchange' (Ekeh 1974;
Bearman 1997), reciprocity is not restricted to localized exchanges within cohesive
clusters: favors can be repaid to people other than the ones in the group (see Lazega,
this volume). Although most organizations probably aspire to have a culture of
generalized exchange, the majority of them operate within a system of restricted
exchange: favors ought to be repaid to the givers, or to the giver's friends. In these
organizational contexts, the norm of reciprocity, combined with the capacity to
impose sanctions to defectors, may force a manager to attend demands even when he
expects no further benefit from the contacts posing those demands. 2
The effects of reciprocity are likely to be compounded by a second mechanism
that may keep managers tied to contacts that have lost their value as social capital,
even without the managers being aware of the problem. This second mechanism is
relational inertia. People get used to dealing with their long-term partners. This
familiarity breeds strong bonds of mutual understanding and trust that greatly
facilitates cooperation (Gulati 1995b). Unfortunately, the same strong bonds may
also serve as a filter for the information and the perspectives reaching the actors,
generating a 'cognitive lock-in' that isolates them from the outer world (Grabher
1993). In addition, the easiness of cooperation with familiar partners raises the cost
of making the investments that are necessary to initiate and to consolidate new
relationships. This cost is even higher once the uncertainty associated with the
formation and with the probability of maintenance of new ties is taken into account.
These inertial factors can make established relationships extremely resilient to losses
in their instrumental value due to changing task environments. If the change in task
does not come with tangible signs such as a physical relocation or new a new
reporting relationship, the manager may fail to recognize the need for adapting the
composition of his social capital to fit the new task. Since failure to develop the
adequate relationships is likely to have an impact on performance, the manager may
eventually learn the hard way about the inadequacy of his network. Late adaptation,
however, comes at a cost. In some cases, the cost may be failure in the new task,
with potentially serious consequences for the manager and for the organization.
The tension between the forces of reciprocity and relational inertia on one hand,
and the changing nature of the resources needed by a manager on the other is
intrinsic to social capital. One simply cannot have the benefits of social capital
without the corresponding obligations and the risk of relational inertia. This tension
is particularly apparent for managers facing transitions that affect their task
environment without simultaneously introducing clear-cut changes in the conditions
under which they have to perform their jobs. This situation is rather frequent in the
dynamic contexts that characterize today's flat, entrepreneurial organizations. Social
capital is key to coordinate task interdependencies in these organizations. At the
same time, the fluid character of the managerial task makes task interdependencies
shift, following the changes in the resources and expertise managers have to
coordinate to add value to the organization. Indeed, the fluidity of the
interdependencies is one of the main arguments against attempts to coordinate them
through formal mechanisms and standard procedures. In these contexts, the ability to
adapt the composition of his social capital is perhaps one of the most valuable
attributes of a good manager.
For a manager, the problem is how to have the benefits of the 'bright side' of
social capital while minimizing the potentially deleterious effects of its 'dark side.'
How to enjoy the advantages of resourceful social networks while maintaining some
degree of control over the composition of this network? Our approach to this
question is to look at factors that may boost the impact of the forces of reciprocity
and relational inertia, hence hindering the manager's ability to adapt his network.
We argue that the structure of the network that defines social capital is one such
factor. More specifically, we argue that a manager's ability to adapt the composition
of the network that carries his social capital is a function of the structure of that
network. The issue is to specify how network structure may curtail the manager's
ability to decide how to allocate his time and energy across a set of actual or
potential contacts. For this purpose, we turn to the control implications of structural
hole theory.
THE EFFECTS OF NETWORK STRUCTURE

Structural hole theory (Burt 1992) asserts that the ability of an actor to have control
over his environment is a function of the structure of this actor's network. The more
a manager depends on parties who are difficult to substitute and who can coordinate
their behavior, the less his ability to negotiate his role in the network. Conversely,
lack of contact among those parties creates 'structural holes' that enhance the
manager's ability to control his environment.3 Players who occupy brokerage
positions between disconnected alters have comparative advantages in negotiating
relationships, which allows them to secure more favorable terms in the opportunities
they chose to pursue. According to Burt (1997: 343), 'managers with contact
networks rich in structural holes know about, have a hand in, and exercise control
over more rewarding opportunities. They monitor information more effectively than
it can be monitored bureaucratically. They move information faster, and to more
people, than memos.'
The focus of the research inspired by structural hole theory has been on the
outcomes of social structure. The information and control advantages that accrue to
managers occupying brokerage positions in networks are deemed responsible for the
managers' early promotions or comparatively higher bonuses, which are seen as a
recognition to their ability to add value to their organizations (Burt 1992, 1997).
Structural hole theory, however, may be also used to explain the process through
which social capital is created and renewed by the managers. According to structural
hole theory, a manager strongly tied to contacts that are in turn connected to one
another is expected to have little autonomy to negotiate his role vis-a-vis his
contacts. Likewise, we can hypothesize that such a manager should have difficulty
in renovating the composition of his network when the pressures of the environment
may require him to do so. More specifically, we expect this manager to be less able
to develop new relationships or to have control over the strength of the ties with
existing ones.
The reasons behind our hypothesis lie on the interplay between social structures
on one side, and the forces of reciprocity and relational inertia on the other. As
Granovetter (1985) has argued and empirical studies have confirmed (e.g., Gulati
1995a) the probability of cooperation between two actors is enhanced by the
presence of third parties that bring 'closure' to the social structure (Coleman 1990).
Burt and Knez (1995) dissect the mechanism behind this effect in their analysis of
the impact of third parties on trust, which they define as 'anticipated cooperation' (cf.
Coleman 1990). Indirect connections create a reputational lock-in that ensures
cooperation. Non-cooperative behavior by either partner may be reported to-or
worse, observed by~ommon partners. This would typically have serious negative
effects on future relationships entered by the defecting partner, hence serving as an
effective deterrent to defection (Raub and Weessie 1990; Burt and Knez 1995;
Nooteboom, this volume). Consequently, managers linked through ties embedded in
third-party relationships are more likely to conform to pressures of reciprocity. The
strong ties to his contacts, and these contacts' ability to coordinate their demands on
the manager, magnify the impact of the norm of reciprocity on the manager's
behavior as well as the effects of relational inertia, hence affecting the manager's
ability to invest in the development of new ties.
In line with our argument on the two sides of social capital, these potentially
harmful effects may remain latent until the manager actually needs to develop those
new ties due to changes in his task environment. Moreover, there is evidence that at
least in some situations the short-term effects of strong ties to coordinated contacts
may be actually beneficial to the manager. 4 As Podolny and Baron (1997) have
recently argued, dense ties among the key people a manager depends upon may
actually facilitate performance, especially for low and middle-level managers. The
manager with such a network is likely to face a well-defined and consistent
normative framework within which to perform his role, thus avoiding the tensions of
having to respond to conflicting demands. Our argument, however, suggests that this
consistent framework may turn against the manager when he, prompted by changes
in his task environment, needs to change the composition of the network. Although
the peril is perhaps negligible when the new task is accompanied by changes in
reporting line or physical location, this may not be always the case. If such tangible
changes in the conditions under which the manager performs the new job are absent,
the dense network that was hitherto a major source of support for the manager may
now curtail his ability to develop the social capital needed to succeed in the new
task. The old network is no longer a resource for the manager, but a liability that
encumbers his performance.
From a structural perspective, the negative effects of lack of structural holes on
the manager's autonomy to adapt his social capital can operate in two different
ways. First, restricted autonomy may result from the combined pressure posed by all
contacts in the manager's network. The stronger the ties to those contacts, and the
stronger the connections among the contacts, the more they will be in the position to
collectively enforce reciprocity from the manager towards the members of the
network (Lazega, this volume). Group cohesiveness is also likely to drive
members-including our focal manager-to a passive acceptance of the ongoing

patterns of interaction as the normal way to carry out their business, hence
promoting relational inertia. Second, restricted autonomy may also result from the
dominant role of one or a few contacts that monopolize a large proportion of the
time and energy the manager devotes to his network. The 'hierarchical' role of these
contacts in the manager's network puts them in the position to force their demands
upon the manager, leaving him little time to invest in new relationships.
Alternatively, the manager may perceive these prominent contacts as all the social
capital he needs, thus discouraging further investments in developing this social
capital (Higgins and Nohria, this volume).
To summarize our argument: we expect managers whose contacts networks lack
structural holes to face more difficulties in adapting their networks according to the
changing demands of their task environment, thus leading to a higher rate of failure
in the renewal of their social capital. Lack of structural holes may result from strong
ties to the various members of a cohesive cluster or from very strong ties to few
players that are well connected within the rest of the manager's network. Managers
lacking structural holes should be more likely to fail in adapting the content of their
social capital in line with changing environmental demands. To test this proposition,
we examine managers operating in a special unit of a European subsidiary of a high-
technology firm. We focus on their failures to adapt the composition of their social
network to the changing demands placed on them by a flat organizational
environment. We describe this organizational environment, then explain the data and
methods used in our analysis.
THE ORGANIZATION
Our study focused on a unit within the Italian subsidiary of a leading multinational
computer company in the early nineties (see Benassi 1993, for detailed description).
Like most firms in the industry, the company was dealing with difficult market
conditions. Impressive price-cutting and accelerating competitive dynamics driven
by dramatic advances in chip technology made profit margins plummet, forcing
firms, and especially large firms, to reshape their activities. In this context, the
search for more effective organizational configurations was a major endeavor for
large computer manufacturers (The Economist 1993). Our firm was not an
exception. At the time of our study, several initiatives of organizational change were
under way. Headquarters explicitly initiated some of these initiatives, while others
were emerging out of the everyday practice of organizational transformation. Our
study focused on one of these emergent strategies. A small unit operating in one of
the Italian plants was promoting alternative forms of voluntary cooperation among
business units within and outside the organization. Although our unit was not
originally conceived as a change agent, its style of work resulted in an emergent
strategy that could lead towards a radical change in the processes and the culture of
the Italian organization. A good part of that change was already observable inside
the unit.
The Direzione Processi Industriali (Direction of Industrial Processes, or DPI)
was a small unit formally created in January 1991 and staffed by 19 members, all
but one of them male. Its origins can be related both to a process of organizational
change and to individual initiatives. Early in 1991, the existing functional structure
was substituted by a business unit organization and new centers of competence were
created to deal with the emerging issues the company was facing. At the same time,
an autonomous research initiative undertaken by a group of people working at the
Italian plant focused on new technological strategies and on alternative models of
conceptualizing the overall activity of the firm, both at the national and the
international levels. The creation of DPI gave an institutional form to the new,
disperse expertise that resulted from these previous initiatives. Out of the 19
members of DPI, 10 came directly from a unit that launched some of these
initiatives and continued to work on their projects until these were completed or
terminated. Another group of 6 people also came from a single unit within the firm.
Thus, from the inception, the largest majority of DPI managers had worked before
with some of their current colleagues. At least initially, several of them even
continued with the same projects that they have started in their previous job,
although new colleagues might have joined the teams.
DPI's scope of activity was very broad, but is overarching mission was to
promote horizontal linkages within the firm and between the firm and its suppliers
and customers. These linkages were essential to implement the new business unit
structure. To this end, the unit collaborated with-and promoted collaboration
among-actors inside and outside the company, providing solutions to internal
business units, top management, and international functional managers, as well as to
external clients. Its competencies included devising manufacturing strategies for the
Italian plants, developing a market-driven quality approach, promoting marketing-
manufacturing cooperation, and creating tools and methods to implement the
different initiatives. It also coordinated activities of managers in charge of setting
long-term strategies and represented the Italian plant in international company
hearings. To an important extent, these tasks resulted from autonomous initiatives
that became institutionalized in the new unit and were then formally recognized by
the company. In this sense, DPI clearly resembles the entrepreneurial image
associated with flat organization. This image also corresponded with the internal
structure of the unit. Although formal hierarchy and task differentiation did exist,
barriers among task groups and individuals were insignificant. Personal initiative
was not only strongly encouraged within DPI: it was a prerequisite to do the job.
The managers knew that solutions were driven by multiple contributions and that
external and internal relationships were crucial for getting things done. They
believed that traditional organizational mechanisms were no longer valid to
accomplish the complex task they had at hand. Instead of relying on the formal
hierarchical structure, the managers sought to leverage horizontal relationships and
to promote cooperation among internal departments. These initiatives often led to
new business opportunities for the firm, both inside and outside the formal
boundaries of the organization. Although DPI had the support of top management,
its location in the formal structure made it a peripheral part of the organization. The
head of DPI reports to the plant manager, who is two steps below the top
management.
In line with the organizational blueprints adapted to its style of work, DPI
favored working through project teams by which the unit pooled resources from
within and outside the organization to provide specific solutions to both internal and
external customers. Between January and October 1991, DPI was directly involved
in 73 of such projects. At the time of our survey, 70 percent of the projects were still
under way, 20 percent were recently completed, and 10 percent abandoned. One
third of these projects was a continuation of those carried out by previous units. The
remaining ones were either a direct initiative of DPI (43 percent) or were launched
upon internal or external customers' demands (57 percent). In more than 60 percent
of the projects, DPI acted as leading unit. Only 8 of the 73 project teams were
formed exclusively with people from DPI. On average, project teams had 6.6 people
representing 3.7 different units. DPI contributed an average of 2.8 team members
per project. Typically, one of these people was the team leader, often paired with a
leader from another unit. On average, each of the 19 DPI managers was the team
leader in 3.8 projects, ranging from 1 to 7. In addition, he would participate in other
project teams. The average DPI manager participated in 10.2 different projects, with
a range from 2 to 21. Although managers did have initiative regarding their
participation in the projects, the composition of the teams was largely driven by
reasons of technical expertise and experience, and was ultimately the responsibility
of the head of the unit. In this sense, DPI managers were not truly self-selected into
the project teams they participated.
The managerial approach and the internal structure of DPI makes the unit a
good example of the fast pace, relationship-driven environment where adaptability
of ones' social capital is key to individual and corporate success. The managers
developed and leveraged relationships both inside and outside the firm, actively
seeking to promote cooperation ties that cut across the formal structure and the
boundaries of the organization. To attain that goal, DPI managers must also
coordinate the task interdependencies created by their joint participation in teams
and by their combined attempts to mobilize resources from independent areas of the
organization. This setting makes the unit a fruitful laboratory to analyze the factors
that may thwart the renewal of the content of social capital.
DATA AND METHODS

We argue that managers lacking structural holes should be more likely to fail to
adapt the composition of their social capital in line with changing environmental
demands. To test this hypothesis, we examined mismatches between the level of task
interdependence and the level of work-related consultation between the nineteen
managers working for DPI. Specifically, we focused on failures to maintain
consultation ties at levels deemed adequate for the intensity of the respective task
interdependence between the managers. Drawing on resource dependence theory
(Pfeffer and Salancik 1978), we assume that effective managers should shape their
social network to maximize their capacity to handle the task interdependencies
affecting their jobs. In an extreme case, this would mean a perfect mapping of the
consultation network on the interdependence network. Individual managers,
however, are likely to deviate from this extreme case in two ways. First, a manager
may keep ties with people he does not depend upon-Qr people with whom he has
only weak task interdependence. We may refer to these as cases of 'surplus'
consultation. Second. a manager may have a low level of consultation (or no
consultation at all) with people he strongly depends upon. We call these situations of
missing consultation 'coordination failures.'
From the standpoint of resource-dependence theory, the impact of each type of
deviation on managerial performance should be different. Although ties with people
a manager does not depend upon to carry his task may be seen as an inefficient way
to allocate effort, these ties do not necessarily have a negative impact on the
manager' s performance. Moreover, such ties may be instrumental in two ways that
go beyond the immediate task interdependencies affecting the manager. First,
managers may have to keep weak ties with people that were key for their task in the
past and who may be important again in the future. Indeed, Granovetter's (1973)
analysis of labor markets vindicated the instrumental value of those ties. Second,
managers may use those ties as bridges to access resources or information controlled
by people the manager has been unable to reach directly, or to effect indirect
political influence on people they depend upon but with whom they cannot establish
a good working relationship (Gargiulo 1993). Thus, consultative ties not coupled
with task interdependence may still perform an instrumental role for the manager.
The consequences of the second type of deviation are different. Failure to consult
with a colleague a manager clearly depends upon should make coordination more
difficult and time-consuming, which in turn should affect the manager's
performance. Our focus on coordination failure captures the theoretical difference
between the two types of deviation.
The data analyzed in this chapter comes from a self-administered questionnaire
distributed to all members of DPI in October 1991. This questionnaire was tailored
using the extensive information gathered through ethnographic observation. From
June 1991 to December 1991 , one of the authors was allowed to engage in daily
observation of the DPI operations. He also had easy access to the unit members for
interviews, could consult written communication and formal documents, and
attended team and unit meetings. All 19 managers working for DPI responded to the
survey. Our questionnaire comprised two parts. First, a general part covered
information on the manager's involvement in projects, as well as an evaluation of
the functioning of the unit. Each member was also presented with a roster of his or
her colleagues within DPI and asked to rank his or her level of consultation on
problems relevant to his or her work with each of these colleagues, in a scale from 0
(no consultation) to 3 (strong consultation). Second, those managers who were
responsible for the coordination of at least one of the 73 projects implemented by
DPI were asked to fill out a booklet with information about each project under his
supervision. The booklet surveyed detailed information on the project and on the
people involved. It also included a matrix in which the manager responsible for the
project had to list all the people that participated in the project team, including
himself or herself, and to rate the level of cooperation between all pairs of members.
Based on this information, we constructed independent measures of task
interdependence and consultation among the nineteen managers working for DPI.
Task Interdependence
We measure task interdependence among DPI managers as a linear function of their
joint involvement in the projects implemented by the unit. Two characteristics of our
measure are worth noting. First, the measure is an objective, rather than a perceptual
indicator of task interdependence among our managers. It is also exhaustive, since
the managers' task was largely confined to their participation in the projects initiated
or facilitated by the unit. Second, the measure captures both direct and indirect task
interdependencies between the managers. By working with one another in a project,
the managers' tasks become directly interdependent. Managers working on the same
projects are also indirectly interdependent, since they are demanding resources from
the same units in the firm, that is, from the ones that jointly participated in these
projects. For any two managers {i, j}, we can define task interdependence tij as the
number of projects in which managers i and j were jointly involved during the
period covered by our research. This raw measure is affected by the number of
colleagues with whom a manager jointly participated in project teams. To control for
this factor and to focus purely on the manager's pattern of interdependence-rather
than on its volume-we use a proportional measure of task interdependence.
Manager i's task interdependence with manager j can be thus represented in
proportional terms, as the ratio between the number of joint projects with manager j,
tij' and the sum of all joint projects across all managers q, including j:
i"# j .
Structural Holes
We expect that lack of structural holes in the managers' consultation networks will
restrict their autonomy to adapt this network to the requirements of their task
interdependencies, hence making them more prone to coordination failures. To
measure consultation, all 19 DPI managers were presented with a complete list of
their unit's colleagues and asked for the extent to which they routinely consult with
each colleague regarding matters that concerned their work in the unit. Managers
could rate their answers from 0 (no consultation) to 3 (strong consultation). The raw
response data generated a 19-by-19 matrix of consultative ties, with zeros along the
main diagonal. The consultation network reveals a relatively flat structure, with little
variation in the consultation centrality of the managers and a dense web of ties: Any
member of the unit could reach any other member in a maximum of three steps.5
For a variety of reasons not related to their task interdependencies, managers may
vary in their tendency to be the source or the target of consultation regarding work-
related matters. To control for this idiosyncratic variation, we use a proportional
measure of consultation rather than the raw scores; this allows us to capture the
pattern of how a manager has to allocate time to the different people he actively
consults or to the people who consult him for work-related matters. This allocation
is a function of both the attention the manager seeks from others and the demands
other colleagues pose on him. Thus, we measure manager i's consultation with
manager j, Sij' as the proportion of attention manager i has to allocate to manager j,
both as a result of his seeking out j and of being sought out by j (being sij the self-
reported raw measure and 0 ~ sij ~ 3):
and Li Sij = 1 for all managers. The inclusion of the demands posed on the manager
by his colleagues, even when the manager does not seek these colleagues out, is
relevant in this context. Indeed, our reasoning assumes that the manager will have to
pay some attention to these contacts, even if he does not seek them out to discuss his
own task. The measure is sensitive to this difference. If the manager simply
responds to demands on his time (Sji > 0), but does not seek out the particular contact
=
(sij 0), his proportional consultative tie with the contact (Sij) would be weaker. As
we shall see, such weak tie need not be detrimental. The tie, however, would
become stronger if the manager also seek this particular contact out for consultation
(Sij> 0).6 A manager's autonomy may be restricted by the combined pressure posed
by all his contacts in the consultation network, or from the dominant role of one or a
few contacts that monopolize a large proportion of the time and energy he devotes to
this network. We define two structural variables that respectively capture these two
conditions: network constraint and network hierarchy. Following Burt (1992), we
measure network constraint as the sum of the constraint posed by each of the
contacts in the network. This constraint is in turn a function of the direct joint
consultation between i and j and of the extent to which j consults with the other
contacts q in i's network (see Burt 1992: 50-71, for detailed discussion of these
measures):
where Sij is the proportional measure of consultation ties discussed before and Ljcij,
or network constraint, captures the lack of structural holes in manager i's network
resulting from having strong consultative ties to members of a cohesive cluster.
Network hierarchy, in turn, captures the fact that restricted managerial
autonomy can also result from a consultation network built around strong ties to one
or few players who occupy a central position in the manager's network. We measure
this informal hierarchy in the managers' consultation networks as a function of the
distribution of the constraint across the manager's contacts. Concentration of
constraint in dominant contacts is captured by the Coleman-Theil disorder index. 7
The closer the index gets to 1.0, the more a single contactj dominates the manager's
consultation network, and the higher his network hierarchy score. Contact j's
dominance is a function of the strength of the consultation between the manager i
and j (Sij), and of the extent to which j is also consulted by the other contacts q with
whom i also consults (SjtPqi). Network hierarchy captures the lack of structural
holes in manager i's consultation network resulting the dominant role of one or a
few contacts in this network.
Coordination Failures
The number 0/ coordination/ai/ures committed by each of the 19 DPI managers is
the dependent variable in the analysis. We define as coordination failures those
cases in which high (above average) task interdependence Tij between two DPI
managers is coupled with low (below average) consultation Sij between these
managers.8 As we noted before, DPI managers had a large number of consultation
ties, which in all but one case exceeded the number of their task interdependencies.
In this setting, the relative strength of the consultative tie is relevant. By focusing on
the joint occurrence of high interdependence and low consultation, we capture cases
of strong task interdependence that are coordinated by a weak consultative tie-that
is, a tie weaker than the expected level of consultation among any two managers in
the unit.
Following this criterion, each of the 342 dyads in the network was coded as
either failing (1) or not failing to coordinate (0). On the basis of this coding, we
computed an individual failure measure for each manager, defined as the sum of
coordination failures across all the people with whom he jointly participated in at
least one project team. This is the dependent variable in our study. The measure may
vary from 0 when the manager had successfully allocated above average
consultation ties to all his strong interdependencies to a maximum equal to the
number of strong (above average) task interdependencies if all these strong
interdependencies are coupled with weak (below average) consultation. For
example, one of the managers worked with 13 different colleagues and he was
strongly (above average) task-interdependent with 7 of those colleagues. His level of
consultation with those 7 colleagues, however, was below average in 4 of the cases,
hence our manager committed 4 coordination failures. On average, DPI managers
committed 3.7 failures in coordinating 10.6 task interdependencies.
Coordination failures were detrimental for DPI's corporate social capital and
had a notieeable negative impact on the level of cooperation attained in the project
teams, hence affecting DPI's ability to attain its goals. The main goal of the DPI was
to promote cooperation among the other members of the project teams, who often
came from different areas of the organization. Yet, coordination failures between
DPI managers jointly participating in a team prevented them from attaining
satisfactory levels of cooperation within that team. The proportion of coordination
failures within DPI dyads in a project team was negatively correlated with the
=
average level of attained cooperation among all team members (r -.308; p < .01).
The negative impact of coordination failures was also statistically significant for
both the average cooperation between DPI managers and other members of the team
(r = -.494; p < .01), as well as for the cooperation among non-DPI team members
(r=-.396; p < .01).9
Control Variables
We have controlled for factors that may affect a manager's tendency to incur in
coordination failures. Specifically, we looked at the number of weak consultative
ties in a manager's network and to his workload. High failure scores may result from
a manager's tendency to have a large number of 'weak' consultative ties-that is, ties
that are below the average strength of consultation in the unit. A larger number of
weak consultative ties can result from having a large network of contacts, which in
turn may translate into a larger number of coordination failures. Failure, however,
was not statistically associated with the size of the manager's consultation network
(r = .200; P = .206), nor with the number of ties to colleagues the manager did not
work with in a project team (r = .197; P = .209). Thus, failures were not the simple
effect of having a large consultation network or from a tendency to consult with
colleagues not working in their project teams, but rather from how managers
allocate consultative ties to their task interdependencies. The effect of weak ties may
be more troublesome for managers who score high in network hierarchy. By
definition, such managers have a large number of weak ties and a few very strong
ones. Unless the distribution of task interdependence is equally skewed, this profile
may automatically result in a large number of coordination failures. The mechanism
at work, however, would not be related to the dominant role of the strong-tie
contacts, as proposed in this chapter. A statistically significant association between
consultation network hierarchy and the number of coordination failures could be
simply an artifact of the number of weak ties in the consultation network, rather than
a test of the impact of network structure on the manager's ability to freely allocate
his consultative ties.
To rule out the possibility that the correlation between coordination failures and
network structure was simply an artifact of the number of weak ties in the
consultation network, we removed the effect of the number of weak ties from the
original variables and re-estimated the models using the adjusted scores. We
computed adjusted measures of number of coordination failures (our dependent
variable), network constraint, and network hierarchy. The adjusted measures are the
standardized residuals obtained by regressing coordination failures, constraint, and
hierarchy on the number of weak ties in the manager's consultation network.
Consultation ties whose proportional strength (Sij) was below the criterion level
used to define coordination failures were coded as 'weak.'10 The adjusted measures
are thus free from the spurious influence of the number of weak ties in a manager's
consultation network. If our theory on the encumbering effects of network structure
on the ability to match consultative ties to task interdependence is correct, both
network constraint and network hierarchy should still have an impact on
coordination failures once the spurious effect of weak ties is removed. On the
contrary, if the alleged impact of the structural variables on coordination failures is
simply an artifact of the presence of weak ties, the impact of these structural
variables should disappear after the effect of weak ties is removed from the
measures.
A manager's failure to maintain adequate consultation ties with people he
depends upon can also be affected by the particular workload conditions in which
this manager has to perform his task. The larger a manager's workload, the higher
the probability of failing to keep an adequate level of coordination for all the task
interdependencies defining his role, and thus the higher the number of failures. Two
workload dimensions are particularly relevant here: the number of project teams in
which the manager participated and the number of different colleagues with whom
the manager jointly participated in at least one project team, which is the number of
task interdependencies. As one might expect, these two variables are strongly
correlated (r = .744; P < .(01), since the number of task interdependencies typically
increases with the number of project teams in which the manager participates. This
g. w
0' -
"""
- ~
-tlc
a~
~
~
is'
Table 1. Means, standard deviations, and zero-order correlation matrix -
~
'1::i
Variables Mean (J Failure 2 3 4 5 6 7
[
O. Number of coordination failures 3.684 2.136
1. Level of network constraint .219 .006 .516 ~
2. Network hierarchy .043 .015 .661 .630 t-<
is'
3. Number of task interdependencies 10.632 3.862 .436 -.136 .088 ~
4. Number of project teams 10.158 5.786 .566 .014 .315 .744 ~
5. Adjusted coordination failures • .000 1.807 .846 .302 .352 .490 .529
6. Adjusted level of constraint' .000 .005 .292 .873 .340 -.179 -.109 .346
7. Adjusted network hierarchy' .000 .Oll .406 .406 .732 .081 .223 .481 .465
8. Number of weak consultative ties 8.895 1.823 .533 .488 .682 .042 .223 .000 .000 .000
a Adjusted scores are the standardized residuals of regressing the respective raw measure on the number of weak ties in the manager's network.
Table 2. Structural hole effects on the number of individual coordination failures (Standardized
OLS coefficients)
Structural Hole Effects Beta t-value R:l F

Network Constraint
Controlled by project participation .508*** 3.128 .578 10.978***
Controlled by task interdependencies .586*** 3.337 .527 8.927***
Removing weak tie effects:
Controlled by project participation .408** 2.176 .444 6.394**
Controlled by task interdependencies .448** 2.343 .434 6.136**
Network Hierarchy
Controlled by project participation .536*** 3.135 .579 11.007***
Controlled by task interdependencies .627*** 3.863 .581 11.098***
Removing weak tie effects:
Controlled by project partiCipation .382* 1.950 .418 5.745**
Controlled by task interdependencies .444** 2.353 It 1.435 6.167**
*p < .10; **p< .05; ***p<.OI; N=19
correlation creates multicollinearity problems that make it impossible to have both

controls in the equation simultaneously. To circumvent that problem, we estimated
two separated models. We expect both variables to have a statistically significant
positive impact on failure rates. We are not, however, interested in this impact, but
only on the changes in the effect of the structural variables after controlling for
workload conditions. Table 1 presents descriptive statistics and zero-order
correlations between all these variables.
RESULTS
Table 2 presents standardized regression coefficients measuring the impact of the
structure of the managers' networks on the number of coordination failures incurred
by this manager. The results furnish evidence supporting our predictions. The effects
are statistically significant for both network constraint and network hierarchy.
Network constraint increases the number of coordination failures committed by
managers. The impact of network constraint is still strong after removing the effect
of weak ties, which suggests that the number of weak ties in the managers'
consultation network play a marginal role in accounting for coordination failures.
Rather, these failures are the result of how those weak ties are allocated to the
different task interdependencies. Managers with a network lacking structural holes
were more likely to allocate weak consultation ties to strong interdependencies and
strong consultation ties to weak or non-existent interdependencies, which prompted
coordination failures. Similar results are obtained using network hierarchy. The
tendency to have a consultation network built around strong ties to one or few
players makes managers more likely to allocate weak consultation ties to strong task
interdependencies. Workload conditions were an important factor contributing to
managerial failures. The effects of the structural variables, however, were not
substantially affected by the introduction of the control variables. This suggests that
those effects are largely independent of the managers' workload. II
Both network constraint and network hierarchy have comparable effects

increasing coordination failures. The high correlation between these two
independent variables (r = .661 for the standard measures and r = .465 for the
adjusted measures; p < .01), coupled with the small number of observations, create
multicollinearity problems that made it difficult to evaluate the relative impact of
each of these structural properties on coordination failures . Attempting an
approximation to such an evaluation, we obtained instrumental measures of
constraint and hierarchy by removing the effect of the network constraint from
network hierarchy, and vice-versa. Regressing the number of coordination failures
on network constraint and on the instrumental hierarchy measure, we obtained
estimates for the impact of each variable on coordination failures. A similar
estimation was done using the instrumental measure of constraint and network
hierarchy. The results, however, do not allow for an empirical evaluation of
differences in the impact of each structural property on coordination failures. A
comparison of the corresponding standardized regression coefficients shows that the
impact of the instrumental measure of hierarchy is similar to the impact of the
instrumental measure of constraint. 12
Going one step further in the analysis, we explored the sources of variation in
the shape of the manager's consultation networks. More specifically, we
investigated individual and organizational sources of constraining relationships. To
this end, we examined constraint scores at the dyad level and sought to identify
correlates of highly constrained relationships. We used the marginal constraint
posed by each contact relatively to the average constraint in the manager's network.
Contacts posing high constraint are strongly tied to the focal manager and to this
manager's other contacts. 13
We found no statistically significant association between relative constraint and
similarity in background variables such as educational level, major, seniority, or age.
The average DPI manager was neither more nor less likely to be constrained by a
colleague like him or herself, when similarity is defined along these 'human capital'
attributes. Relative constraint was also unrelated to formal reporting relationship.
There was, however, a clear association between relative constraint and
organizational background, with average relative constraint higher among people
who were together in the same unit before being assigned to DPI (2.66 t-test; p <
.01). Similar results were obtained using raw constraint scores (cij) or the relative
strength of the tie (Sij). The dominant players in the managers' networks were more
likely to be the people the managers used to work with in their previous assignment
in the organization. As we mentioned before, ten of the nineteen DPI managers
came directly from a unit that launched some of these initiatives and continued to
work on their projects until these were completed or terminated. Another group of
six people also came from a single unit within the firm. The relatively smooth
transition between those previous assignments and DPI, which in some cases even
implied a continuation of the same project work, with the same people, did not
always signal a change in the task environment of these managers. As DPI started its
own projects, the task interdependencies of the managers shifted, but the change
went unnoticed for a number of them, still faithful to consultative relations built over
years in the firm.
This finding confirms some well-known ideas about the origin of strong
relationships. Granovetter (1973: 1361) has pointed out that the strength of a tie' ...
is a (probably linear) combination of the amount of time, the emotional intensity, the
intimacy (mutual confiding), and the reciprocal services that characterize the tie.'
Looking at the origin of ties, Feld (1981) stressed joint participation in similar
organizational contexts as one of the main sources of relationships. Common
organizational history puts people in contact, prompts the exchanges of advice and
services, and allows for repeated exchanges that are the basis for a strong
relationship. Our analysis suggests that these relationships are likely to outlive the
specific context in which they arise. If the change of context is not drastic enough-
such as physical relocation-the old relationship may remain strong. This was
particularly so in the case of DPI. A large number of the observed managers were
previously working together in the unit that preceded DPI and initially continued to
do so in the new unit, often in similar projects, even if new colleagues often joined
the teams. The smooth transition did not signal a clear change in the task
environment for these managers, who continued to consult with their old colleagues
even when the new projects created interdependencies with managers coming from
other parts of the organization.
Our analysis of the DPI managers furnished evidence supporting the
propositions discussed in this chapter. Lack of structural holes in the network that
defines social capital makes it difficult for the manager to renew this capital as
required in a changing task environment. The results are independent of the
potentially spurious influence of the number of weak ties in the consultation
networks and from differences in workload conditions across managers. The results,
however, did not shed light on the relative importance of the two ways in which a
constraining network structure may affect a manager' s ability to renew his social
capital. Both strong ties to cohesive cliques and the presence of dominant contacts in
the manager's network equally jeopardize his ability to adapt his social capital. An
examination of the sources of the constraining relationships responsible for the lack
of structural holes in the managers' networks revealed that those relationships
typically corresponded to ties forged through years of work in similar organizational
environments. The strength of those bonds, and the lack of tangible signs of change
in the managers' task environment, galvanized the mechanisms of reciprocity and
relational inertia and increase the risk of coordination failures.
DISCUSSION AND CONCLUSION

This chapter draws upon the control implications of structural hole theory to explore
the dark side of social capital. We explain how the structure of a manager's contact
network may affect his ability to adapt the content of his social capital. Faced with a
growing number of task interdependencies that are mostly beyond his administrative
control, a manager must develop informal mechanisms to effectively coordinate the
uncertainties surrounding these tasks. In this context, a manager's social network is
a critical resource to amass the necessary corporate social capital. Not any kind of
social capital, however, will be equally useful in such an environment. Effective
managers should be able to strategically adapt the composition of their social
network to help coordinate shifting interdependencies. To do so, however, they must
have the autonomy to develop, maintain, and leverage relationships, thus keeping
control on how they allocate time and energy to their contacts. Factors that reduce
such autonomy increase the likelihood of failures in coordinating critical task
interdependencies, which in turn decrease the available corporate social capital and
may have a negative impact on performance. Our research illustrates the impact of
two such factors. A consultation network formed by strong ties with a set of
cohesive contacts, or a network dominated by a small set of contacts, constrain the
manager's ability to adapt the composition of his social capital according to the
changes in his task environment. Conditions that facilitate the creation of strong ties
with densely connected clusters, such as a common organizational history, increase
the probability that managers will build constraining networks, hence affecting their
subsequent ability to renew their social capital.
Our research has consequences for the debate on social capital, as well as for its
effects on managerial and organizational performance. By focusing on factors that
may hinder the renewal of social capital in dynamic task environments, we showed
how the network structure that carries social capital could have an effect on the
manager's ability to renew the composition of this social network. Network
structure, network composition, and outcomes are interdependent, but the
relationship between these three variables is problematic. Early studies of social
capital showed how the composition of an individual's contact network had an
impact on the benefits accruing to this individual (Lin, Ensel, and Vaughn 1981;
Lin and Durnil 1986; Flap and de Graaf 1989).
More recently, scholars have focused on network structure as the motor behind
those benefits, assuming implicitly or explicitly that network composition is a
simple correlate of network structure (Burt 1992; Podolny and Baron 1997). By
focusing on the renewal of social capital, our research clarifies the relationship
between network structure and network composition, elucidating the mechanisms
through which this relationship operates. Our results suggest that both ties to a
cohesive cluster and strong, exclusive ties to few contacts are conditions may turn
the assets of social capital into social liability that damage a manager's adaptability
and performance in changing task environments.
The significance of our findings is enhanced by the fact that the relationship
between network structure and favorable outcomes is not stable across contexts. If
lack of structural holes was always detrimental for the performance of a manager,
our results simply would add another spin to this deleterious effect. Yet, this is not
always the case. Existing research suggests that the effects of structural holes on
managerial performance may depend on the particular situation of the manager. In
his analysis of managers in a U.S. high-tech firm, Burt (1992: 147-153) found that
strong ties to prominent contacts in a manager's network was the best route to fast
promotion for women and entry-level men. He suggests that such ties may be a
necessity for a manager in need of a legitimating sponsor. This positive effect of
what amounts to an informal 'network hierarchy' on promotion captures the weIl-
documented influence of mentoring in managerial careers (e.g., Kram 1988). With a
different emphasis, Podolny and Baron (1997) argued that strong ties to cohesive
contacts linked to a manager by authority relations may also be beneficial in early
stages of their career, when defining an identity and obtaining support from
powerful players is critical for the manager. A manager with strong ties to a
cohesive group may attain the necessary legitimacy to be recognized as a player. He
may also profit from the additional social capital he can 'borrow' from his sponsors.
These benefits, however, may come at a price. Our evidence suggests that strong ties
to a cohesive group may eventually curtail a manager's ability to adapt the
composition of his social capital, even if those ties were initially beneficial for the
manager. This result is congruent with Higgins and Nohria's research (this volume)
on the negative effects of early mentorship on the ability to develop social capital at
a later career stage. The difficulty in renewing the composition of his social capital
affects the managers' ability to coordinate task interdependencies with and among
other members of the organization, hence having a deleterious effect on his
performance.
Our findings, therefore, pose an interesting dilemma for managers as well as for
organizations. On the one hand, managers at the early stages of their career may
need to obtain decisive informal sponsorship to become legitimate players (Burt
1992) or to assert their identity in the organization (Podolny and Baron 1997). This
legitimacy requirement is compounded by the fact that the lack of structural holes is
less detrimental for entry level managers, where the benefits of a diverse contact
network may be negligible (Burt 1997). A small, cohesive core of supportive
contacts may be the best form of social capital for these managers. On the other
hand, the managerial career eventually requires autonomy to allocate time and
energy to different contacts to successfully cope with the shifting task
interdependencies associated to the role. Yet, the same cohesive social network that
was instrumental in asserting the manager's position in the organization may be an
obstacle to develop the type of social capital required to further his professional
growth and his capacity to add value within this organization. Our argument suggest
that this peril may be more consequential when the new task is not accompanied by
tangible changes in reporting line or physical location that naturally weaken the
manager's previous relationships.
Ideally, managers should be able to go through a smooth transition between
these two stages in their careers, adapting both the structure and the composition of
their social capital to the changes in their task environment. The human resource
practices and the culture of the organization should also support this transition.
Although the observed differences in network structure between senior and junior
managers suggest that a good number of people does succeed in adapting their social
capital to the growing complexity of their task environments, the transition may not
be always easy. Moreover, the focus on 'success' stories may hide the fact that a
considerable amount of managerial talent could be wasted in the process. The
initially supportive, legitimating sponsorship of a cohesive cluster or a strong mentor
may translate into a liability that hinders the manager's ability to adapt his social
capital in later stages of his career (Higgins and Nohria, this volume). For the young
manager, there is no easy way out of this dilemma. For the organization, the solution
lies in efficiently combining the benefits of sponsorship with the benefits of
autonomy. Factors that make the transition more noticeable, such as physical or
departmental relocation may help avoiding the perils of the dark side of social
capital discussed in this chapter. Yet, these systemic solutions may not be always
feasible or even desirable. Another, perhaps complementary way to deal with the
problem is to clarify the role of organizational sponsors and to build in the ability to
promote young 'entrepreneurial' managers into their performance evaluation. Like a
wise father, a wise sponsor must know not only how to help and defend his protege,
but also to allow him the necessary autonomy to grow independently.
Finally, our research has implications for the study of the managerial role in the
dynamic, flat organizational environments that increasingly penetrate today's firms.
We suggest that the attenuation of authority embedded in the idea of 'flat'
organizational structures may not suffice to promote entrepreneurship within a
traditional organization and to create the type of corporate social capital that is key
to its success. This is especially so in context of rapid organizational change. Like
the bird that remains inside the cage despite the open door, managers adapted to a
command-and-control structure may fail to use the freedom created by flatter
structures, even when they are convinced of the need to do so. Scholars have
recognized the role of previous socialization in this inability to take advantage of the
new freedom. Our argument suggests that the freedom itself may be sometimes a
mirage. Powerful control mechanisms embedded in informal structure and rooted in
the organizational history can create subtle bonds of interpersonal dependence that
may severely curtail the organization's ability to move away from a command-and-
control managerial style and to create the corporate social capital required in an
entrepreneurial firm. Although necessary, the attenuation of authority is not a
sufficient condition to promote managerial entrepreneurship. Hierarchies ingrained
in the informal organization may still create an effective obstacle to flexibility. The
unobtrusive nature of these ties may make their effect less apparent than the
conspicuous impact of the formal hierarchy. Such an effect, however, may be
equally consequential.
This warning is especially relevant in contexts of the transformation into a flat,
entrepreneurial organization. Moving away from the well-established command-
and-control structures is a task that requires more than administrative decisions. Top
management decisions of reshaping the overall organizational through a general
reduction of managerial levels are important. However, such decisions may not be
enough to promote entrepreneurship. Past organizational characteristics can survive,
even if formal structures have been revised or the chain of command reduced. A
manager cannot act as a flexible network entrepreneur if the informal organizational
structure does not allow it. Yet, the emergence of this structure is largely the result
of effective managerial initiative. Informal structures that result from established
managerial practices can have a direct impact on the behavior of those managers,
which in turn may affect the organization's ability to effectively enact the change.
The research reported in this chapter shed light on the 'dark side' of social
capital by discussing when and how social networks may become an obstacle to
managerial and organizational performance. We show how strong, cohesive social
bonds may hinder the necessary renewal of social ties to fit changes in the
manager's task environment, hence diminishing the level of corporate social capital
available to the firm. We suggest that the risk can be higher if those strong ties have
initially produced substantial advantages for the manager, a condition that defines
the effectiveness of social capital. This paradox reminds us that social capital, like
The Dark Side of Social Capital- 321
other properties linked to social structures, may have both positive and negative
consequences for individual and organizational action. Like most human things,
social capital has a bright side, but also a dark one.
This paper was written while the first author was a Visiting Scholar at Columbia University's Graduate
School of Business (1997-98 academic year). We would like to thank helpful comments by Ron Burt,
Ranjay Gulati, David Gibson, Roger Leenders, Andrej Rus, Harrison White, and the participants in the
Network Analysis Workshop at the Paul Lazarsfeld Center for the Social Sciences at Columbia
University.
NOTES
I. This phenomenon has been documented in studies of organizational adaptation, that show how the
same processes that help organizations to be well adapted to their current environment can curtail their
ability to adapt to a new environment (Grabher 1993). A similar idea plays a central role in ecological
models of organizations (Hannan and Freeman 1989). These models argue that 'specialists' firms-that is,
organizations highly adapted to specific environments-are less likely to survive drastic environmental
changes than 'generalists' organizations, since the latter are less dependent on specific resources for their
survival.
2. Leifer (1988) and Bearman (1997) note how the norm of reciprocity, when applied to dyadic
exchanges, may lock the players into endless exchanges due to ambiguity on the valuation of the 'gifts'
exchanged. A similar argument can be made regarding exchanges with a cohesive group. In this case, the
presence of third parties reinforces compliance, thus increasing the likelihood of people entering such
endless exchanges.
3. Structural hole theory builds upon a key intuition of exchange theory, which states that control in a
relationship is a positive function of the availability of alternatives (Emerson 1962; Blau 1964). A series
of simulation and experimental studies confirm the adequacy of this basic intuition. Players with access to
several exchange partners who themselves lack such alternatives enjoy competitive advantage (Cook and
Emerson 1978; Marsden 1982, 1983; Cook, Emerson, Gillmore, and Yarnagishi 1983).
4. It is worth noting that our treatment of social capital differs from that of structural hole theory,
which makes social capital a sole function of the structure of the network: the more structural holes, the
larger the social capital. Burt (1992: 13) defends his analytical choice by assuming that actors who know
how to structure a network to provide access to opportunities would also know whom to include in that
network. This assumption provides an elegant solution to the tension between the constraint and
opportunity aspects of social capital, but it also restricts the applicability of the concept. An individual
strongly tied to a dense cluster of resourceful people has little social capital in Burt's sense. Yet, in some
situations, strong connections with few strategic players may be actually beneficial for a manager (Burt
1992: 147-153; Podolny and Baron 1997). This effect is predicted by the more traditional definition of
social capital adopted here, but it does not fit easily within structural hole theory. More importantly for
this chapter, Burt's assumption dissolves the problem of adaptation of social capital that is central to our
discussion. Paradoxically, this drives the attention away from the fruitful control implications of
structural hole theory developed here.
5. In the consultation network, centrality scores varied from 1.00 to .594, with an average of .791 and a
standard deviation of .118. Centrality scores reflect a manager's tendency to be consulted by colleagues
who are also preferred targets for consultation (Bonacich 1987). The average strength of the ties in the
network is 1.32, with an effective range from 0 to 3. The average DPI manager reported some level of
consultation with II different colleagues. Out of the 342 possible consultation ties, 62.6 percent are
present; among these, 23.1 percent are strong, 23.7 percent medium and 15.8 percent weak.
6. The inclusion of both types of consultation is relevant in this context, since a manager may still have
to allocate time to colleagues who consult him, even if he does not seek these people out. Granted, when
two managers meet informally to discuss work-related matters at the initiative of one of them, the other
may still bring up issues that are of his specific concern. Yet, if this second manager does not report the
interaction-or reports a weaker interaction-he is probably less likely to initiate these exchanges, which
in tum should lead to a weaker overall relationship.
7. The Coleman-Theil index is calculated as follows:
Ai=
L ICfN
--.9L [In --.9L]
CfN
NIn(N)
where C is the lack of structural holes (aggregate constraint) on actor i (LjCij) and N is the number of
colleagues i consults with andfor is consulted by. Note that the quotient CIN is the average constraint
posed by i' s alters, and thus ci/(CIN) expresses the constraint posed by actor j in terms of the average
constraint in i' s network.
8. The mean strength of task interdependence between managers is .056, with a median of .038 and a
skewness coefficient of 2.035. Sixty percent of the interdependence ties fall below the mean and hence
could not prompt a coordination failure as defined in this chapter. Thus, a low level of consultation with a
sporadic project partner need not be a coordination failure. Consultation ties are normally distributed
(.056 mean, .057 median; .033 skewness). The proportional measures of task interdependence (Tij) and
consultation (Sij) used in this chapter removes differences in the volume of interaction of individual
managers, making average figures a good indicator of the socially expected levels of interdependence and
consultation between any two managers in the unit. Given the distribution of interdependence and
consultation ties, managers clearly had enough 'strong' consultation ties to adequately coordinate all their
strong task interdependencies. Failures resulted from the way they allocated their strong consultation ties.
9. At the project level. we defined a coordination failure index as the ratio between the number of DPI
dyads in the project team who failed to coordinate and the total number of DPI dyads in the project. For
this analysis, projects with a single DPI participant were coded as having zero failure rate. There were 65
projects in which at least one member of a unit other than DPI participated and 58 with 2 or more other
units. Cooperation was measured as the average level of cooperation between team members as reported
by the manager coordinating the project, on a scale from 0 (no cooperation among the team members) to
3 (strong cooperation). Reported values have a mean of 1.7 and a standard deviation of 0.8. As one might
expect, average cooperation in project teams was highest between DPI managers (2.1), intermediate
between DPI and other team members (1.8) and lowest between members of other units (1.3).
10. On average, DPI managers had 8.9 weak ties, ranging from 4 to 12. As expected, the number of
weak ties in the consultation network was significantly correlated with the three variables, that is, number
= = =
of failures (r .533; P < .05), network constraint (r .488; p < .05), and the hierarchy (r .682; p < .01).
These high correlations ruled out an alternative approach to control for the effect of weak consultative ties
on coordination failures by simply including the number of weak ties into the regression.
II. We also estimated models where the dependent variable is the rate of coordination failures-that is,
the number of failures divided by the total number of interdependencies-and obtained similar, albeit
slighty weaker results. Using the rate of coordination failures as a dependent variable, standardized beta
coefficients using the instrumental variables that control for the number of weak ties in the network are
.366 for constraint (p = .06) and .451 for hierachy (p = .03). Note that the rate of failure automatically
control for workload effects, which thus become statistically insignificant.
12. These results were computed using the adjusted measures of our variables and thus are independent
of the number of weak ties in the manager's consultation network. None of the coefficients for the
instrumental measures were statistically significant at the .10 level.
13. Relative constraint, given by the ratio ci/(CIN), is largely responsible for the level of hierarchy
measured by the Coleman-Theil index (see note 7). Thus, consultation networks were often built around
co-workers who had similar careers in the firm.
Social Capital, Social Liabilities,
17
and Social Resources Management
•
Daniel J. Brass
Giuseppe Labianca
ABSTRACT
This chapter explores the role of social capital in human resources management. We
suggest that the recent interest in social capital has neglected the possibility that
social networks may contain negative ties, and that attention to these social
liabilities may provide additional insights into relationships and social networks in
organizations. Research focusing on the antecedents and consequences of social
networks in organizations is reviewed. We consider the effects of social capital and
social liabilities on 'social' resources management outcomes such as recruitment,
selection, socialization, training, performance, career development, turnover, job
satisfaction, power, and conflict.
INTRODUCTION
Human resources management, as the name implies, has persistently defined its task
and focused its energy on developing methods of measuring individuals' human
capital. To focus on the individual in isolation is. at best, failing to see the entire
picture. Rather, people are embedded within a network of interrelationships with
other people. These networks of relationships provide opportunities and constraints
that make up the social capital of the individuals and the larger system. Social
capital inheres in the social relationships that can potentially confer benefits to
individuals and groups (i.e., Bourdieu 1972; Burt 1992; Coleman 1988, 1990;
Fukuyama 1995a; Gabbay 1995, 1997; Putnam 1995b). Social capital can be
contrasted with human capital (one's knowledge, skills, and abilities), financial
capital (money), and physical capital (physical property such as land, buildings,
machinery). We do not mean to suggest that individuals do not differ in their human
capital, their skills and abilities and their willingness to use them. Nevertheless, we
will try to nudge the study of human resources management toward the study of
social resources management; from human capital toward social capital.
Coleman (1988) suggests three forms of social capital. First, obligations and
expectations, and the trust that facilitates them, arise when actors are willing to do
something for other actors because they expect and trust that the recipients will
honor the obligation to reciprocate in the future. The second form of social capital
refers to actors obtaining information that may be useful to them from others,
indirectly taking advantage of others' knowledge, skills and other forms of human
capital. The third form of social capital, norms and sanctions, allows for the
reduction of transaction costs.We expand the notion of social capital by adding to it
the role of negative relationships-relationships in which at least one person has a
negative affective judgment of the other. Although relationships create opportunities
and benefits, the current focus on social capital emphasizes only the positive aspects
of social networks while neglecting the potential liabilities that may be associated
with negative relationships. Although the early social exchange theorists and
network researchers considered both the positive and negative aspects of
relationships (e.g., Homans 1961; Tagiuri 1958; Thibaut and Kelley 1959), recent
network research has focused almost exclusively on the positive aspects of social
structure, with little exploration of any possible negative aspects. The question of
where and when the potential benefits outweigh the potential liabilities of expanding
one's network has been left largely unanswered. Rather than solely investigating
social capital, we attempt to consider the 'social ledger'-both the potential benefits
as well as the potential liabilities of social relationships (Labianca and Brass 1997).
Just as a financial ledger records financial assets and liabilities, the social ledger is
an accounting of social assets (social capital) and social liabilities (negative
relationships). We also suggest that the liability side of the social ledger, negative
relationships, may have greater explanatory power than positive relationships.
THE SOCIAL LIABILITIES OF NEGATIVE RELATIONSHIPS

In suggesting that relationships have benefits, researchers have found that one's
social network relates to such outcomes as job attainment, job satisfaction, power,
and promotions in organizations (e.g., Brass 1984; Burt 1992). The consideration of
social liability may add to our understanding of these outcomes. For example, the
underlying assumption is that an employee's friends and acquaintances help the
employee obtain promotions by providing such forms of social capital as critical
information, mentorship, and good references. However, it is also likely that an
employee's negative relationships with others in an organization might prevent
promotion, particularly if those with whom the employee has negative ties are in
influential positions. Those people may withhold critical information or provide bad
references in order to prevent promotion. Thus, it may be equally important to
consider the negative side of the social ledger: social liabilities as well as social
capital.
We define negative relationships as relationships in which at least one person
has a negative affective judgment of another person. Prior research has viewed the
negative aspects of personal relationships in two ways. One perspective assumes that
Social Capital, Social Liabilities, and Social Resources Management - 325
every relationship contains both positive and negative aspects and that these aspects
are independent (e.g., Bradburn 1969; Diener and Emmons 1985; Russell 1979;
Watson and Tellegen 1985). We adopt a second perspective that acknowledges the
first, but further assumes that people form a global judgment of others that can be
captured by such terms as 'like' and 'dislike,' that are opposite ends of a continuum
(e.g., Berscheid and Walster 1969; Newcomb 1961; Tagiuri 1958).
We distinguish between social liabilities (negative relationships) and the
opportunity costs of building social capital. We note that the process of developing
social capital may include opportunity costs as well as benefits. For example,
developing strong, trusting relationships may interfere with the opportunities for
bridging diverse groups via weak ties (see Gargiulo and Benassi in this volume). As
Granovetter (1985) noted, the obligations and expectations of strong, long-lasting
relationships may prevent a person from realizing greater economic opportunities by
constraining the search for, and development of new trading partners. Gabbay (1995,
1997) and Gabbay and Leenders (this volume) have noted that some actors may be
constrained by the same social structure that benefits other actors.
In their chapters in this volume, Gabbay and Leenders the term 'social liability'
to denote the constraining effects of social structure. Social liability is the opposite
of social capital. In the current chapter, we use the term to explore the potential
liabilities that can result from negative relationships.
NEGATIVE ASYMMETRY
The importance of considering the social liabilities of negative relationships is based
on our hypothesis of negative asymmetry: negative ties may have greater
explanatory power than positive ties in organizations. This hypothesis is based on
indirect evidence from research in the area of social support in health care (see
Labianca and Brass 1998 for a review), and a diverse psychological literature
indicating that negative events elicit greater physiological, affective, cognitive, and
behavioral activity and lead to more cognitive analysis than neutral or positive
events (see Taylor 1991 for a review). In addition, Burt and Knez (1995) found that
third parties amplified the effects of gossip, and that the amplification effect was
stronger for negative gossip than positive gossip. In an organizational field study,
Labianca, Brass, and Gray (1997) found that negative relationships increased
perceptions of intergroup conflict, but strong positive relationships had no
counterbalancing effect. Why do negative events and relationships have more
impact than positive events and relationships? Evolutionary psychologists explain
the negative asymmetry by noting that it is adaptive to respond quickly to negative
events in order to enhance survivability (cf. Cannon 1932). Developmental
psychologists suggest that negative events are discriminated and evaluated earlier by
children than are positive events because negative events are more likely to interrupt
action. Children learn the rules governing negative behavior before those governing
positive behavior; children are punishment oriented (cf. Piaget 1932). Nature and
nurture combine to make humans risk-averse (Kahneman and Tversky 1984).
In seeking to theoreticaIIy explain negative asymmetry, Skowronski and
Carlston (1989) summarize a number of theories. For example, negative events
dominate social judgment because of the contrast effects with positive and moderate
events that people typically experience and expect. Since people expect positive,
moderate information, negative, extreme information is weighted more heavily in
impression formation. Negative information is also attended to because it is more
unambiguous than positive information; it allows people to make social judgments
more easily.
In addition to direct social information, indirect social information can also
result in negative relationships. Second-hand information is filtered and simplified
so as to be unambiguous, and amplification of negative aspects is often more
pronounced that positive information (Burt and Knez 1995).
ANTECEDENTS OF SOCIAL CAPITAL AND SOCIAL LIABILITIES

Although previous discussions of social capital have focused on outcomes, In
particular, the beneficial outcomes of social relationships, it is important to
understand the antecedents of social relationships and how they might affect social
capital and social liabilities. We now consider the factors that affect social
relationships, and social resource management in organizations.
Organizational Structure
Positions in organizations are formally differentiated both horizontally (by
technology, workflow, task design) and vertically (by administrative hierarchy), and
means for coordinating among differentiated positions are specified. Because
communication is a fundamental means of coordination, it follows that social
relationships are influenced by the prescribed vertical and horizontal differentiation
and the resulting need for coordination. Formally differentiated positions locate
individuals and groups in physical and temporal space (workers assigned to different
time shifts), and at particular points in the workflow and hierarchy of authority.
Organizational structure therefore restricts opportunities to interact with some
people, and facilitating interaction with others. As Festinger, Schacter, and Back
(1950) found, physical proximity is related to amount of interaction. Despite the use
of telephones and electronic mail, proximate physical and temporal ties are easier to
maintain and more likely to be strong, stable links (Monge and Eisenberg 1987).
The restrictions on social interaction imposed by the formal organizational
hierarchy and workflow requirements are particularly important when considering
negative relationships. In everyday, non-work activities, actors can easily avoid or
decrease interactions with others whom they dislike. However, the formally
prescribed, required interactions in organizations create the possibility of required
negative relationships; relationships with disliked others that cannot be avoided.
Thus, the effects of negative relationships on social capital and the social ledger may
be particularly relevant in organizational settings.
Size
The size of an organization (number of employees) may also affect the social
relationships and social capital. As the size of the organization increases, network
density naturally decreases (assuming that actors can maintain only a limited
number of ties), and the possibility of fragmentation (individuals forming
sub-groups) increases (Shaw 1971). Similarity and increased interaction result in
strong ties forming among sub-group members, and decrease the probability of
strong, positive connections across groups. Thus, decreased density may make it
more difficult to maintain the 'closure' needed for effective norms and sanctions
(Coleman 1988), and the trust necessary for obligations and expectations may be
more difficult to establish in larger systems.
Actor Similarity
Although social networks likely shadow the formal structure of the organization,
employees may informally modify the prescribed interaction. That is, we do not
ignore human agency. We do not assume that organization structure completely
constrains individual action. Instead, we adopt a bi-directional perspective (Giddens
1976; Leenders 1995b)-structure and behavior are mutually causal. Interactions
that occur within the constraints of structure can gradually modify that structure.
Individuals break relationships and build new ones. Thus, it is important to
understand why individuals choose some relationships and not others.
A good deal of research has noted a tendency for similar people to interact and it
is a basic assumption in many theories (Homans 1950; Davis 1966; Granovetter
1973; Blau 1977). Homophily (interaction with similar others) results from ease of
communication and increased predictability of behavior, and it fosters trust and
reciprocity. Brass (1985a), Ibarra (1992), and Mehra, Kilduff, and Brass (1998)
have found evidence of homophily by gender in organizations. For example, Ibarra
(1992) found that men had homophilous ties (with other men) across multiple
networks, whereas women had social support and friendship network ties with other
women, but they had instrumental network ties (e.g., communication, advice,
influence) with men. In addition, perceived similarity (religion, age, ethnic and
racial background, and professional affiliation) among executives has been shown to
influence interorganizationallinkages (Schermerhorn 1977; Galaskiewicz 1979).
In combination with age, sex, status, etc., we would expect similarity of human
capital characteristics such as personality and ability to be related to the
interpersonal network patterns of interaction and social capital. We also would
expect the characteristics of the links between actors to be related to the degree of
actor similarity. Communication between two dissimilar actors is likely to be
infrequent, not reciprocated, less salient to either, asymmetric, unstable, uniplex
rather than multiplex, weak, or even negative. Similarity of actors also may be
positively related to the density or connectedness of the network.
Similarity may be a necessary precondition to social capital that results from all
three forms noted by Coleman (1988). Individuals are more likely to trust similar
others in reciprocating obligations and expectations, similar others are more likely to
share information, and dense networks of similar others are more likely to develop
and maintain norms and sanctions. For example, Coleman has argued that 'closure'
of the network is essential for the effectiveness of norms. We emphasize actor
similarity, and later, attitude similarity, because both are key to such human
resources practices as recruitment and selection.
It is important to note that similarity is a relational concept; an individual can
only be similar with respect to another individual, and in relation to dissimilar
others. That is, interaction is influenced by the degree to which an individual is
similar to other individuals relative to how similar he or she is with everyone else.
Due to culture, selection and socialization processes, and reward systems, an
organization may exhibit a modal demographic or personality pattern. Kanter (1977)
has referred to this process as 'homosocial reproduction.' Thus, an individual's
similarity in relation to the modal attributes of the organization (or the group) may
determine the extent to which he or she is central or integrated in the interpersonal
network, and the extent to which he or she shares in the social capital of the
organization.
Conversely, actor dissimilarity may be a source of negative relationships and
exclusion from the benefits of social capital. Actors are less likely to trust dissimilar
others, share information with them, or include them in the norms of the system. For
example, research has shown that relational differences on such demographic
variables as age and tenure are related to commitment and turnover (Tsui, Egan, and
O'Reilly 1992; Tsui and O'Reilly 1989; Wagner, Pfeffer, and O'Reilly 1984;
Zenger and Lawrence 1989). Although all people possess some similarity, and some
capacity to identify with others, research on 'moral exclusion' suggests that
dissimilar others may be excluded from social relationships and the targets of
unethical behavior (Brass, Butterfield, and Skaggs 1998; Opotow 1990; Smith
1966).
The exclusion of individuals from the social network can be detrimental to the
overall social capital of the system. The diverse, non-redundant information that
may come from relationships with dissimilar others is lost (Burt 1992), and the
overall creativity and productivity of the collective may be diminished (Brass
1995a). Thus, exclusion of dissimilar actors, or negative relationships with dis-
similar others, decreases the amount of system-level social capital that is available to
anyone.
Attitude Similarity
In addition to noting the propensity for similar actors to interact, theory and research
have also noted that those who interact become more similar in their attitudes (see
Krackbardt and Brass 1994 and Leenders 1995b for a detailed discussion). The
relational basis of attitudes (Erickson 1988) assumes that people are not born with
their attitudes, nor do they develop them in isolation. Attitude formation and change
occur primarily through social processes, such as communication and comparison
(Leenders 1995b). As people attempt to make sense of reality, they talk with others
and compare their own perceptions with those of others, in particular, similar others.
For example, Kilduff (1990) found that M.B.A. students made decisions similar to
their friends' regarding job interviews with organizations.
We build upon Erickson's ideas by suggesting that negative relationships may
foster attitude dissimilarity. The dislike of another actor may prompt one to take
opposing positions on issues. Disagreements can be attributed to dissimilarity, and
may even be used to reinforce one's own attitudes. Thus, our understanding of
attitude similarity and social capital may be enhanced by accounting for negative
relationships as well as positive relationships. Negative relationships may interfere
with attitude similarity and disrupt social norms, expectations, and obligations,
especially when several negative relationships exist in a network. The presence of a
relative few negative relationships and dissimilar attitudes may actually strengthen
group norms, just as the overt intolerance of deviant behavior reinforces norms.
Although attitude similarity has typically been considered an outcome of social
interaction (Erickson 1988), we view it as an antecedent to social interaction as well.
That is, actors may seek out others with similar attitudes, and avoid those with
dissimilar attitudes, thereby reinforcing their own perceptions of reality. As
Leenders (1995b) notes, it may be as easy for actors with similar attitudes to seek
and find each other as it is for one actor to influence the attitudes of another. For
example, it is likely that actors with similar demographics, attitudes, and behaviors
will meet in similar settings. Activities are organized around 'social foci' (Feld
1981). Football fans meet in the stadium to watch a football game, school children
of similar age meet at school to be educated, accountants meet in the accounting
departments of organizations. Actors who gather around a particular social focus are
likely to be similar, and interaction with each other will likely enhance that
similarity (Leenders 1995b).
Just as actor similarity (or dissimilarity) may affect social capital, attitude
similarity may have similar effects. As with actor similarity, the trustworthiness of
similar expectations and obligations, as well as norms and sanctions, may depend
upon similar attitudes among organizational members.
SOCIAL CAPITAL AND SOCIAL LIABILITIES: THE OUTCOMES OF

SOCIAL RELATIONSHIPS
We now turn our attention to the outcomes of social relationships. We consider both
social capital and social liabilities in relation to typical human resource management
outcomes.
Job Satisfaction
The early laboratory studies of small groups found that central actors were more
satisfied than peripheral actors (see Shaw 1964 for a review). Since that time,
research in organizations has found that relative isolates (zero or one link) in the
communication network were less satisfied than participants (two or more links)
(Roberts and O'Reilly 1979), that job characteristics (such as autonomy and variety)
mediated the relationship between workflow network centrality measures and job
satisfaction (Brass 1981), and that betweenness centrality in the friendship network
was negatively related to job satisfaction (Kilduff and Krackhardt 1993). Kilduff and
Krackhardt reasoned that conflicting expectations and stress may result from
mediating the relationships between actors who are not themselves friends. Thus,
when linking actors who are not connected themselves, central actors may not enjoy
the common expectations and obligations of social capital.
Kilduff and Krackhardt's (1993) findings remind us that interaction is not
always positive. Since Durkheim (1897) argued that social integration promotes
mental health, there has been a long history of equating social interaction with the
positive aspects of social support (Wellman 1992). Yet we have all experienced
negative interactions with others in the workplace. We try to avoid these negative
relationships whenever possible, thereby producing a positive relationship between
interaction and friendship. However, prescribed interactions are not always positive
or avoidable. As the previously cited literature on negative asymmetry indicates,

negative relationships may overwhelm the job satisfaction resulting from positive
social relationships.
Power
It is also possible that the effects of centrality on job satisfaction are indirectly
related to power. Network studies of power have centered on the social capital of
information. The finding that central network positions are associated with power
has been reported in a variety of settings (see Brass 1995b for a review).
Theoretically, actors in central network positions have greater access to, and
potential control over relevant resources, in particular the social capital of
information. Just as collective information represents a type of social capital for the
entire system, individual actors can acquire social capital via access to information.
In addition, simple degree centrality measures of the size of one's network (number
of direct links) have been found to positively relate to power (Brass and Burkhardt
1992, 1993; Burkhardt and Brass 1990).
Knoke and Burt (1983) have argued that being the object of relations (chosen by
others) rather than the source (choosing others) is an indication of prestige, and a
possible source of social capital. For example, Burkhardt and Brass (1990) found
that all employees increased their closeness centrality (symmetric measure)
following the introduction of new technology, but the early adopters of the new
technology increased their (in-degree) prestige and their power significantly more
than the later adopters.
The social capital represented by access to information may be related to who an
actor knows as well as how central the actor is. For example, Brass (1984) and Blau
and Alba (1982) found that relationships beyond the prescribed workgroup and
workflow requirements were related to influence. In particular, links to the dominant
coalition in the organization were strongly related to power and promotions. The
dominant coalition was represented by a clique of the top executives in the
company. These top executives likely had more social capital in the form of more
information, and more relevant information, to share with those connected to them.
Brass (1985a) also found that men were more closely linked to the dominant
coalition (composed of four men) and were perceived as more influential than
women. Assuming that power positions in most organizations are dominated by
men, women may be forced to forgo any preference for homophily in order to build
connections with the dominant coalition and share in the social capital. As
previously noted, actor dissimilarity (in this case based on gender) may affect
interaction patterns and consequently exclude some people from sharing in the
social capital.
In addition, Brass (1985a) found evidence of the constraints of organizational
structure on social relationships and power. Women who were part of integrated
formal workgroups (at least two men and two women) were more closely linked to
the men's network (only male employees considered) and were perceived as more
powerful than women who were not in integrated workgroups. Similarly, men who
were closely linked to the women's network (only women employees considered)
were also perceived as more influential than men who were not. These findings
Social Capital, Social Liabilities, and Social Resources Management - 33 t
suggest that social capital in the form of information can be enhanced, or

constrained by formal organizational structure, and links to dissimilar others.
None of these studies considered the possible effects of negative relationships
on power. It is likely that an actor's enemies, friends of those enemies, and the
enemies of that actor' s friends, determine the amount of power an actor holds in an
organization. These enemies might actively attempt to thwart an actor's efforts by
withholding, or distorting important information in the hope of diminishing that
actor's power. The negative relationships represented by enemies are examples of
social liabilities.
Recruitment
In order to be recruited for a job, both parties, the individual and the organization,
must know of each other. The use of social networks, as contrasted with
employment agencies or job listings, can be very valuable in both searching for jobs
and recruiting workers. In the classic example of the strength of weak ties, people
were able to find jobs more effectively through weak ties (acquaintances) than
strong ties or formal listings (Granovetter 1982). Granovetter argued that an actor' s
acquaintances (weak ties) are less likely to be linked to one another than are an
actor's close friends (strong ties). An actor's set of weak ties will form a low
density, high diversity network, one rich in non-redundant information. A set of
strong ties will be densely interconnected and will likely represent a high degree of
redundant information. Thus, individuals may have greater access to more and
different job opportunities when relying on weak ties. Later findings (Flap and
Boxman, this volume; Granovetter 1995; Lin, Ensel, and Vaughn 1981) have both
modified and emphasized the strength of weak ties.
Although weak ties may be effectively used to increase the diversity and
nonredundancy of recruiting information, strong ties may be the mechanism behind
homosocial reproduction in organizations (Kanter 1977). As previously noted, actor
similarity is an important antecedent in building strong relationships, and repeated
interaction can lead to attitude similarity and, in tum, the further strengthening of
relationships. Organizations have for many years used referrals by current
employees to aid in recruiting (Granovetter 1995; Neckerman and Fernandez 1997).
Using the social networks of previously-hired employees represents a relatively
easy, low-cost mechanism for linking the organization to potential employees
(Marsden and Gorman, this volume). While these links may facilitate recruiting and
decrease the turnover of newly hired applicants (Neckerman and Fernandez 1997),
they may also promote homosocial reproduction. That is, current employees and
recruiters seek out those whom they believe will 'fit in' well in the organization. Fit
is often based on actor similarity. Actor similarity may ease the development of
social capital (sharing information, establishing common expectations, obligations,
and norms), but it may also decrease the diversity of information available for
sharing. Thus, the benefits of social capital in recruiting may also involve costs.
Selection
As Pfeffer (1989) and Ferris and Judge (1991) have noted, selection in organizations
is not entirely the result of human capital (abilities and competences). Pfeffer (1989)
argued that credentials and hiring standards are often the result of political contests
within organizations, rather than purely rational decision making. Those in power
seek to perpetuate their power by setting criteria and selecting those applicants most
like themselves. As in the case of recruiting via the use of networks, selection may
also depend on network ties, and indirectly, actor similarity. In particular, we expect
these effects when the qualified applicant pool is large, or when hiring standards are
ambiguous. In such cases, similarity between applicant and recruiter, or indirect
similarity through a common friend , may be important, but unstated, selection
criteria.
Just as positive relationships may prove helpful in recruitment and selection,
negative relationships may be particularly harmful. When applicant pools are large,
any negative information, or any dislike of an applicant may eliminate him or her
from further consideration. Thus, applicants are well advised to be risk averse with
interviewing for jobs or acquiring references. Negative asymmetry is particularly
apparent in letters of reference. Because typical letters are positive, any negative
information is attended to and weighted more heavily in diagnosis and decision
making.
Socialization
Once an employee is selected for a job, socialization of that employee may be
dependent on network involvement. Jablin and Krone (1987) and Sherman, Smith,
and Mansfield (1986) both found that network involvement was a key process in the
socialization of new employees. Similarly, Roberts and O'Reilly (1979) and
Eisenberg, Monge, and Miller (1984) found that network participation was
positively related to organization commitment.
However, due to the cross-sectional nature of these studies, it is impossible to
know whether integration into the network leads to socialization and commitment,
or vice versa. It is likely that early connections in the organizational network lead to
socialization of expectations, obligations, and norms, and enhanced social capital,
and that commitment leads to further integration in the network. It may be especially
important that new employees do not form negative relationships early in their
employment tenure, because such relationships may deter socialization and lead to
turnover.
Training
In traditional human resources management, the focus of training is typically on
human capital-acquiring new and innovative ideas and skills. A social capital
perspective on training might focus on the role of social networks in acquisition and
diffusion of these skills, thus paralleling the research on adoption and diffusion of
innovations (cf. Rogers 1971; Tushman 1977; Tushman and Anderson 1986; Burt
1982). Innovation is generally thought to require diverse and novel information. As
previously noted, weak ties may provide more diverse and novel information than
strong ties. A member of a closely knit, dense clique of strong ties is less likely to be
exposed to diverse, novel perspectives than an actor with weak ties to a number of
different social groups. Thus, Burt (1992) argued that the diversity of one's contacts
is more important than the size of one's network. The findings that cosmopolitans
(i.e., actors with external ties which cross social boundaries) are more likely to
introduce innovations than are locals (Rogers 1971) supports this diversity
argument. Conversely, central actors, sometimes identified as 'opinion leaders,' are
unlikely to be the source of innovations that are not consistent with the established
norms of the group (Rogers 1971).
Once training is introduced, the diffusion of the training (or the spread of new
ideas and skills) can be predicted by social network relationships. Some controversy
exists over whether diffusion is best predicted by a cohesion (communication
through direct interaction) approach or a structural equivalence perspective
(observation and comparison of others in similar roles in the network). As Leenders
(1995b) notes, these two theoretical approaches are practically impossible to sort out
empirically. Theoretically, the direct communication, cohesion apprOl1ch suggests
that the most central actors should be the first to experience training. Having the
most direct and indirect contacts, central actors will quickly diffuse the skills.
In their investigation of the introduction, training, and diffusion of a
technological change in an organization, Burkhardt and Brass (1990) found that
early adopters were not highly central prior to the introduction of the system.
However, the early adopters increased both their centrality and power in the
organization as the technology was implemented. The diffusion process closely
followed the new network patterns following the change, with structurally
equivalent employees adopting at similar times.
In a similar study of the introduction of a new computer technology, Papa
(1990) concluded that training programs can provide basic operating information,
but that much of the learning about a new technology occurs after training as
employees attempt to apply the training. He found that productivity following the
change, and the speed at which the new technology was learned, were positively
related to interaction frequency, network size, and network diversity (i.e., number of
different departments and hierarchical levels contacted). Both studies (Burkhardt
and Brass 1990; Papa 1990) support the idea that learning, whether through training
or networks, is an active process of information exchange, and an important
outcome of social capital.
Training can also be viewed as a social focus (Feld 1981) that provides an
opportunity to build social connections among participants. For example, for
organizations such as police forces and military units, training is intentionally
designed to develop strong ties among participants (Van Maanen 1975). Likewise,
corporations have recently emphasized teamwork and building strong, trusting
relationships among team members in intensive training programs where executives
spend prolonged periods of time in survival settings. Deep and lasting relationships
can develop as people go through the rites of transition together (Trice and Morand
1989). For example, relationships established as cohorts proceed through college
together can provide social capital in later life. Even when training experiences are
brief, network connections are formed and may develop further. When viewed from
this network perspective, training can be effectively used to build connections across
diverse, heterogeneous groups in anticipation of the future formation of
cross-functional teams (Krackhardt and Hanson 1993). In the context of training,
connections may be based on similar training experiences, rather than actor

similarity on attributes such as race, sex, or age.
Performance
The human resources management literature has focused on methods of increasing
the accuracy and reliability of appraisal of individual performance. However,
performance evaluations may also depend on actual and perceived similarity in
attitudes, values, and demographics between the supervisor (evaluator) and the
subordinate (See Ferris and Judge 1991 for a review). While actor similarity may
positively bias performance evaluations, negative relationships may result in lower
evaluations. Due to negative asymmetry, negative events may be more easily
recalled, perceived as more diagnostic, and receive greater weights in evaluations
than positive events (Taylor 1991).
The social capital perspective on performance invites us to analyze the pattern
of relationships rather than view individuals' performance in isolation. As is the case
with interdependent tasks in organizations, relationships with others affect
performance (Brass 1981; Roberts and O'Reilly 1979). For example, Brass (1985b)
found that performance varied according to combinations of technological
uncertainty, job characteristics, and interaction patterns. The results are consistent
with small group laboratory network studies of the early 1950's (see Shaw 1964 for
a review). These studies consistently found that performance was better when the
communication network matched the information processing requirements of the
task. Centralized communication networks resulted in more efficient performance
when tasks were simple, and decentralized networks were better for performing
complex, uncertain tasks.
As Lazega (this volume) concludes, social structure can result in social capital
and increased performance. Lazega found that the existence of a multiplex,
generalized exchange system of sharing information resulted in better performance
in a law firm. In addition, Uzzi (1997a) found that the social capital inherent in a
system of strong ties can provide economic benefits for small firms in the garment
industry (cf. Uzzi, this volume).
Career Development
From a social capital perspective, getting ahead in organizations may be a function
of 'who you know, not just what you know.' This perspective is contrary to the
individualistic values that drive much of the research in human resources
management: achievement and rewards should be contingent on individual effort
and abilities (human capital). Yet, most managers' careers are contingent on what
they can effectively accomplish in connection' with others. The myth of managerial
work is that it occurs in isolation (Mintzberg 1973). Most of a manager's roles
involve social relationships.
The popular press has noted the importance of 'networking' as well as the
advantages of having mentors in organizations. The advantages of building a large
network have been extolled and seldom questioned, although little systematic
research has actually addressed this prescription. But large networks require a large
amount of time and effort in maintaining relationships. Rather than simply building
relationships randomly, some strategy may be useful. For example, should actors
develop close personal relationships with mentors or highly connected others (i.e., a
strong tie strategy) or attempt to develop many weaker relationships with
disconnected others (i.e., a weak tie strategy)?
Assuming a limit to the number of direct links that an employee can maintain,
being linked to central others is more efficient than being linked to peripheral others.
This strategy allows an employee to be central by virtue of a few direct links to
others who have many direct links. The employee has access to social capital
(resources such as information) via the indirect links of the highly-connected other,
such as a mentor. However, the reliance on indirect links creates a dependency on
the highly- connected other (mentor) to mediate the flow of resources. Thus it is
important to form a strong, trusting tie to the highly connected other. We refer to
this as a strong tie strategy.
Alternatively, an employee may choose a strategy of creating weak ties to
diverse groups. Burt (1992) has extolled the advantages of linking groups that are
not themselves linked-taking advantage of 'structural holes.' A structural hole is
defined as the absence of link between two others who are both linked to an actor.
Because the others are not themselves linked, the actor gains non-redundant
information from the others (Le., the strength of weak ties argument), and is in a
position to control the information flow between the two (Le., broker the
relationship), or play the two off against each other.
Burt (1992) found that structural holes were associated with early promotions
for a sample of managers in a high-technology firm, except in the case of women
and newly-hired managers. For women, connections to highly-central others (the
strong tie strategy) worked best. However, because the network data were not
longitudinal, it is difficult to assess whether the networks were the result of early
promotions or the cause of early promotions. Kilduff and Krackhardt (1994) found
that strong ties also carry advantages. They found that a friendship link to a
prominent person in an organization tended to boost an individual's performance
reputation. In addition, they found that the perceived network, rather than the actual
network, statistically significantly predicted reputations. Thus, it appears that
individuals may improve their reputations by 'basking in the reflected glory' of
prominent others. Similarly, Brass (1984, 1985a) found that connections to
supervisors and the dominant coalition in an organization were related to
reputational measures of power and actual promotions over a three-year time period.
To the extent that acquiring power and influence is related to upward mobility and
success, much of the previous discussion of power applies.
Connections to powerful others may result in 'basking in the reflected glory,' but
they may also result in being perceived as 'second fiddle.' In the latter case, one's
own talents are diminished in the presence of a powerful other (i.e., one is perceived
as 'riding the coattails' or as a 'side kick'). These different perceptions may be
contingent on the stage of one's career, boundaries to entry, and/or the type of
organization. Early in one's career, strong connections to a powerful other may be
perceived as an indication of potential success. However, later in one's career, one is
expected to successfully perform on one's own, and to mentor others. Continued
attachment to a powerful other may not be beneficial to one's career (but see
Higgins and Nohria, this volume, for contrary evidence).
As the previous discussion of actor similarity suggests, women, minorities, or
newly hired managers may face barriers to entry in established networks, and may
be excluded from the social capital of such networks. A strong connection to a
powerful, well-connected mentor may overcome such barriers, as Burt's (1992)
analyses suggest.
Avoiding or resolving negative relationships may be equally, if not more
important to career development. As in the case of selection, negative asymmetry
may overwhelm any of the social capital of either strong or weak tie strategies. It is
likely that an actor's negative ties within an organization will prevent promotion,
particularly if those negative relationships are with influential others. Others may
withhold critical information that worsens an actor's performance or they may
provide bad references in order to prevent a promotion. Employees adopting a
strong-tie, mentor strategy must also be aware of any negative relationships between
the mentor and others. Being perceived as the 'side-kick' of a highly-disliked other
may be detrimental to career success. Conversely, forming the many diverse
connections involved in a weak-tie strategy may increase the possibility of negative
relationships. Avoiding negative relationships may be particularly important for
women and minorities in organizations. Any evidence of negative relationships may
confirm negative stereotypes and quickly interrupt career development.
Turnover
Krackhardt and Porter (1985, 1986) found that turnover was a function of an actor's
position in the social network. Turnover did not occur randomly, but in structurally
equivalent clusters in the perceived interpersonal communication network. In this
longitudinal study, the closer the employee was to those who left, the more satisfied
and committed the remaining employee became, cognitively justifying their own
decision to stay.
When turnover occurs in large numbers, such as layoffs connected with
downsizing in organizations, social networks, and any of the benefits of social
capital, may be disrupted or destroyed. Shah (1996) examined survivor's networks
and reactions following downsizing in a firm where 42% of the workers were
discharged. Although the advice network was restored in six months, the friendship
network remained depleted. Survivors responded negatively to the loss of the social
capital provided by friendships, but responded positively to the promotional
opportunities provided by vacancies due to the layoffs of structurally equivalent
others.
Actor similarity may also affect turnover. Research has shown that similarity in
age and tenure among group members is negatively related to turnover (Tsui et al.
1992; Tsui and O'Reilly 1989; Wagner et al. 1984; Zenger and Lawrence 1989). As
McPherson, Popielarz, and Drobnic (1992) found, similarity leads to increased
communication which, in tum, is negatively related to turnover in voluntary
organizations. They found that network ties within a group were associated with
reduced turnover, while ties outside the group (weak ties) increased turnover.
These findings suggest that employees who share in social capital may be less
likely to quit. We can further speculate that negative relationships may be a primary
cause of turnover, especially when these social liabilities involve supervisors.
Conflict
In a study of intergroup networks in twenty organizations, Nelson (1989) argued that
the interaction networks were significantly different for high and low conflict
organizations. He found that low-conflict organizations were characterized by a high
number of strong ties (measured as frequency of communication) between members
of different groups. Similarly, Krackhardt and Stern (1988) found that during a
crisis, strong ties across groups (friendships) provided the links necessary for
effective coordination in a simulated organization.
However, Labianca, Brass, and Gray (1997) found that strong friendship ties
between members of different groups in an organization had no statistically
significant effect on perceptions of intergroup conflict. Rather, they found that
negative interpersonal relationships between members of different groups strongly
predicted perceptions of intergroup conflict. The findings illustrated the effects of
negative asymmetry; strong positive relationships did not dampen, or counterbalance
the effects of negative relationships. They also found evidence of indirect effects.
When an actor's friend had a negative relationship with a member of another group,
the actor perceived increased conflict between groups.
CONCLUSIONS
Much of the progress in human resources management research and application has
been achieved via the traditional emphasis on the identification and measurement of
individual attributes. Yet, it has been estimated that the average adult maintains
more than 1000 informal ties (mutually recognizable others). Research on the 'small
world' phenomenon (Travers and Milgram 1969) has shown that two randomly
selected people can 'reach' each other through a path of a surprising few number of
links. We are a network of social interdependencies. In attempting to move toward a
focus on 'social resources management,' we have outlined some of the important
antecedents of social networks, and tried to show how social networks may affect
such human resource practices as recruitment, selection, training, socialization,
performance, careers, and turnover.
We have also attempted to expand the focus on social capital by resurrecting
attention to the negative relationships noted by earlier social network theorists such
as Homans (1961) and White (1961). It appears that the negative asymmetry of
negative relationship may destroy the possible benefits of social structure for
individuals. These social liabilities may also deter the system-wide benefits of social
capital. Negative relationships within a system may destroy the trust necessary for
common expectations and obligations, and the closure necessary for norms and
sanctions. In addition, negative relationships may prevent information sharing and
even lead to intentional distortions of information in a system. Negative
relationships cannot always be avoided in organizations where workflow and
hierarchy require interactions. Thus, it is important that we expand our research on
social capital to include negative relationships and a consideration of the entire

social ledger.
What does the future hold for organizations and 'social' resources management?
Many see increased acceleration of change in the environment, increased
uncertainty, and increased information processing requirements. They have
suggested the emergence of a new organizational form-the 'network' organization
(Miles and Snow 1986; Baker 1992a; Nohria and Eccles 1992). Managing social
resources in this new network organization may be as important as managing human
resources. Managers will likely be involved in identifying, locating, and organizing
employees across organizational and international boundaries. Social resource
managers will likely become human capital 'brokers,' bringing together the right mix
of people to successfully offer a product or service. And that mix may be used only
temporarily as environments and technologies rapidly change. Social resource
management will require identifying and nurturing potential relationships that may
change quickly. Given the diversity of people and perspectives, managing social
liabilities will be particularly important. Conflict management skills may help
prevent or resolve negative relationships. Overall, the network organization will
place additional importance on relationships, and the social structure needed for
social capital. As one executive said, 'Our business is one of relationships' (Business
Week 1986: 64).
SECTION IV
• Structure at the Firm Level

social capital in collaboration and alliances
The Triangle:
Roles of the Go-Between
•
18
Bart Nooteboom
ABSTRACT
Social capital is seen here as part of the overall resource base of a frrm. One part of
social capital is positional advantage in a network. It is established on the basis of
relational competence. Third parties can help in the development of social capital by
offering their relational competence, in playing one or more of six roles: the roles of
the go-between. Transaction cost economics recognizes that in an inter-firm relation
the inclusion of a third party can economize on the setting up and operation of a
governance mechanism ('trilateral governance'). The third party acts as a go-between
in monitoring and controlling compliance to agreements, thus eliminating the need of
intricate and costly forms of 'bilateral governance.' A second role of the go-between is
to serve as a repository of hostages. A third role is to help in the judgement of the
value that partners have for each other, i.e. to solve the 'revelation problem.' A fourth
is to serve as a filter against spill-over. A fifth is to mediate in the building of trust. A
sixth is to act as a boundary spanner: to offer a link between an established network
and outside sources of innovation, while maintaining the integrity of the network.
These roles are especially important in relations that are aimed at innovation. By
performing these roles, the third party also increases the flexibility of networks of
frrms. In sum, third parties may form an important part of the social capital that
supports networks of frrms. The analysis opens opportunities, or new perspectives for
fulfilment of their roles, to governmental agencies, such as innovation transfer
agencies, municipalities or provinces, and market agencies, such as banks, and
suggests that a new market is opening up for professional go-betweens.
INTRODUCTION
In studies of inter-organizational relations (lOR's), there is a tendency to look at dyads
of flrms, and to consider networks as aggregates of such dyads. But there are several
roles for a third party; a go-between. This chapter looks at a go-between not in the
sense of a middleman who intermediates in existing production or trade, such as an
agent, wholesaler, retailer, and not in the sense of an entrepreneur who intermediates
in the realization of new potential in connecting supply and demand. It looks at a go-
between in the sense of a relationship counsellor for the development and maintenance
of social capital; providing support in setting up, adapting and ending cooperative
relations between others. Or, in yet different terms: to help in the embedding of relati-
ons, in Granovetter's (1985) sense (Uzzi 1997a). Such roles may be performed by
middlemen or entrepreneurs, but also by specialized agents who do not playa direct
role in linking stages in a chain of production and distribution, as middlemen and
entrepreneurs typically do. Indeed, some of the roles require an independence that is
served by not having a direct stake in the relations that need to be developed.
I propose that the analysis yields a perspective for looking at the roles of trade and
industry associations in European business systems, and of banks and trading houses
in Japanese enterprise groups (Kigyo Shudan). Some of the roles are played by the
Innovation Centres that were instituted in the Netherlands to act as intermediaries in
the transfer of technology to small ftrms, without actually serving as suppliers of
technology. More importantly, perhaps, I propose that there is an emerging market for
the specialized services of the go-between. Some of the most important roles of the go-
between are associated with problems of information, knowledge and learning. These
are particularly important in lOR's that are aimed at innovation; at the development of
products, processes and competencies.
One theoretical perspective employed in this chapter is transaction cost economics
(TCE). I grant that TCE lacks a perspective on innovation and learning and fails to
deal with trust next to opportunism, and more generally fails to deal with
embedded ness of transactions. But in my view a more encompassing theory can be
developed which retains what is valid and useful in TCE and incorporates notions
from social exchange theory. I have attempted to contribute to such development
(Berger et al. 1995; Nooteboom 1992, 1996; Nooteboom et al. 1997). Especially the
'resource' or 'competence' view in present non-mainstream economics (Foss 1993;
Foss and Knudsen 1996) is well suited for such integration of perspectives. In the
spirit of this endeavor I will try to position the notion of social capital in the
framework of the resource perspective.
TCE retains the assumption from mainstream economics that technology is
accessible to all, and this conflicts with the competence view that competencies are
embodied in people, teams, procedures, organization and organizational culture, need
to be built up, and enable as well as restrict the cognitive and technical repertoire of
the flrm. TCE also does not incorporate issues of spill-over. When competence is
associated with tacit knowledge and is embedded in the flrm, risk of spill-over may be
limited. Nevertheless such risk often remains to a greater or lesser extent, and spill-
over control is another aspect of the governance of lOR's that must be taken into
account.
The Triangle: Roles of the Go-Between - 343
One might say that by introducing the competence perspective, innovationllear-

ning, spill-over, trust and the social embedding of transaction relations TCE is trans-
formed beyond recognition, and that its core assumptions are cut out, so that it is
misleading to call it transaction cost theory. But the point is that although trust is made
part of the theory, opportunism also remains part of the theory, next to trust:
trustworthiness restricts opportunism, but opportunism also puts a limit to trustworthi-
ness. Even the most loyal partners may succumb to pressures of survival or golden
opportunities. Repeated cheating quite rightly leads to a loss of trust (see also Uzzi
1997a). The core idea from TCE that dedicated investments, combined with
uncertainty concerning environmental conditions (market and technology) create
dependence and a potential hazard of hold-up, is retained. When it works, trust can
serve as a substitute for legal contracting, and thereby yields a form of governance
with lower costs, greater speed, more flexibility, less of a straightjacket for innovation
and a better basis for exchange of ideas and learning. But arms-length legal contracting
still is an instrument of governance that may fit the conditions and the type of
exchange. Other instruments from TCE also retain their potential use: relational
contracting on the basis of balanced ownership of dedicated investments, exchange of
hostages or mutual dependence; trilateral governance. Indeed: the notion of trilateral
governance yields the frrst role for the go-between.
SOCIAL CAPITAL AS A RESOURCE

Social capital is defined as the resources inherent in social structural arrangements,
facilitating the attainment of goals (Coleman 1990; the chapters of Gabbay and
Leenders in this volume). I would like to elaborate on this notion, and to embed it in
the wider notion of 'resources' available to frrms.
The 'resource' or 'competence' view of the frrm (Prahalad and Hamel 1990; Foss
1993; Foss and Knudsen 1996) goes back to the work of Penrose (1959). According to
this view the main cause of a frrm's profit is its unique competencies, embedded in the
firm. This view complements the 'positioning' view from Industrial Organization that a
frrm's profit is due to market conditions: market concentration, oligopoly, collusion,
entry barriers (Porter 1985). The two views meet in their joint recognition of the
importance of product differentiation to evade pure price competition, and the need to
protect one's differentiated product by basing it on competencies that are at least
temporarily unique and inimitable. The competence view implies the rejection of any
notion of the 'representative frrm': the essence of markets is precisely the heterogeneity
of frrms; frrms seek to be different, as the main source of profit. Competencies are
cumulative and path dependent, and that is why they cannot always be identically co-
pied (cf. Nooteboom 1992). In particular, the focus of the present chapter is on the
development of competencies: on learning and innovation, and the role thereby of
lOR's.
Following Stoelhorst (1997), I take resources as a general concept that can be
classified into assets, competencies and positional advantages. Assets are characterized
by the fact that they are subject to legal ownership and contracts. Competencies and
positional advantages are not easily subject to property rights. Resources cannot all be
instantly copied by other frrms because they are to some extent inscrutable or subject
to 'causal ambiguity' (Lippman and Rumelt 1982): even if a would-be imitator can
observe activities, that does not yet imply that he can understand them, let alone
implement them. Resources, and especially competencies, can be difficult to
understand and imitate, because the knowledge involved is to a greater or lesser extent
tacit (not documented) and embodied in the heads and hands of people, and embedded
in teams, organizational structure and procedures, and organizational culture.
Competencies can reside on the personal level, in the form of knowledge, skill and
relational competence, but related to the latter I would also include motivation and
morality. 'Morality' includes norms and values of conduct that the individual holds, his
degree of commitment to them and susceptibility to ethical appeals (concerning loyal-
ty, justice, truthfulness). On the aggregate, interpersonal level of 'communities of
practice' (Brown and Duguid 1991) within an organization, entire organizations and
even networks of organizations there are assets and positional advantages, but also
competencies. Competencies on the level of an organization or network would include
institutions and patterns of knowledge exchange and transformation. Institutions are
defined as environments and arrangements which limit and guide conduct (North
1990; North and Thomas 1973). They include practices, procedures, rules, technical
standards as well as cultural entities such as prevailing norms and values of conduct,
goals, role models, rituals. Among other things, they may serve to guide relations with
other organizations, and are then part of fmn-Ievel relational competence. Organi-
zational competencies in the form of 'patterns of knowledge exchange and conversion'
refer to the way in which knowledge is converted from tacit to documented
knowledge, absorbed from documented into tacit knowledge, transmitted, pooled,
shared and recombined in novel combinations (cf. Nonaka and Takeuchi 1995).
Relational competencies on the individual level enhance learning. Relational
competencies on the individual level, supported and guided by relational competencies
on the organizational level in the form of guiding institutions, yield positional advanta-
ge in the form of efficient access to resources of other organizations. Positional
advantages further include product-technology-market combinations, access to materi-
als, distribution channels, political acceptance, brand loyalty and reputation.
In this framework, social capital, according to the wide definition of 'social
structural arrangements which facilitate the attainment of goals,' would include
positional advantages achieved in networks, as well as structures of communication,
knowledge conversion and learning within the fmn. It this framework it is seen as the
product of relational competencies on both the individual and the fmn level. The
central point of this chapter now is that third parties can contribute relational
competence in the building of social capital in the form of positional advantages.
SYSTEMS OF COMPETENCE
From the perspectives of competence and dynamics, a basic hypothesis underlying this
chapter is that under present conditions fmns need other fmns in order to cope with
the need and opportunity to innovate and differentiate products. The need follows from
increased competition, due to globalization, and the opportunity follows from
increased prosperity, yielding more differentiated needs and wishes of consumers,
technologies of flexible production and information- and communication technology
(lCT). But this differentiation, innovation and technical development yield great
complexity and variability of opportunities and threats. For a single firm, a full grasp
of the detailed and fast changing stock of knowledge concerning market and techno-
logical opportunities and full command of all relevant competencies is not 'sustaina-
ble,' as Zuscovitch (1994) put it. One needs partner fIrms that are close to specifIc
markets, and partner fInns close to specifIc sources, who specialize in the required
knowledge and make it profItable by sharing it with partners, in networks of partial
cooperation.
Elsewhere, this has been called 'cross-fInn economy of learning' and 'external
economy of cognitive scope' (Nooteboom 1992). In a constructivist theory of
knowledge, knowledge, including perception, understanding and evaluation, is based
on categories that have been fonned in interaction with the physical and social
environment. Having different categories, different people are able to see, interpret and
evaluate different things. The underlying categories make knowledge tacit to a greater
or lesser extent, and the fact that categories are fonned in past experience makes
knowledge cumulative and path-dependent. While such theories of cognition apply to
individuals rather than fIrms, fIrms have similar properties because of the alignment of
perception, interpretation and evaluation between people in the fIrm, established in the
fIrm's procedures and its culture. The implications for transaction costs theory, going
beyond 'classical' TCE (Williamson 1985), are as follows. First, an additional reason
for cooperating with others is that one lacks the requisite categories, and cannot simply
buy and install them. You cannot buy understanding, for example. Second, to perceive
and understand opportunities and threats, one may need a diversity of external sources
with suffIcient variety of categories to grasp them, to the extent that conditions are
complex and changing.
This cognitive view stands in contrast with the view of Teece (1986, 1988) and
Chesbrough and Teece (1996) that, particularly in the case of 'systemic' innovations,
integration into a single fIrm is required to deal with connections between elements of
the system, to control spill-over and to ensure appropriability of profIts from the
innovation. That perspective does not deal with cognitive issues of understanding and
learning. I have argued that there is an argument for integration only under the
following limited conditions: the innovation is indeed systemic, innovation is not
radical but incremental, and it has not yet progressed so far as to generate standards
across the interfaces between elements of the system (Nooteboom 1999). In all other
conditions, but particularly when innovation is radical, a disintegrated structure is
better, to ensure suffIcient flexibility of novel combinations and suffIcient variety of
elements to make novel combinations of.
The view of fIrms as not only rivals, but also potential collaborators in
development, distinguishes European and Japanese enterprise systems from Anglo-
American ones. The fonner have been called 'voice' systems, with corporate social
capital in the fonn of networks or other groups of frrms, versus the Anglo-American
'exit' system, with more autonomous and more integrated frrms (Gelauff en den
Broeder 1996; Groenewegen 1997; Nooteboom 1997). The characteristics of the two
types of capitalist systems are summarized in Table 1 (adapted from the comparison
between the Anglo-American and the German system proposed by Gelauff and den
Broeder 1996).
Table 1. Exit and voice systems (adapted from Gelauff en den BrtJeder 1996)
Exit system Voice system
General characteristics market orientation network orientation
short-term relations long-term relations
rivalry cooperation
autonomy of fIrms embeddedness of fIrms
Culture/institutions individualistic group oriented
legalistic community ethics
Corporate governance
Important shareholders individuals fIrms, banks
Corporate control exit (sell shares) voice
take-over mutual interest
Shares managers many few
Creditor control withdraw loan co-owner, monitoring
Regulation supports markets obstacles to trade in shares
banks not owners (US) banks co-owners
no cross-participation cross-participation
Contractual governance
Relations market, arm's length networks, close
Contracts formal relational
Contract execution courts reputation
Organization of work
organizational form vertical integration networks
labour market competitive protection
contracts formal relational
motivation wage, profIt sharing wage, job security
labour participation low worker councils
TRUST
There is an important difference between competence trust and intentional trust. The
first pertains to someone's ability to perform, and the second to his intentions to do so.
Their breach has different implications for action. When trust in ability is breached an
appropriate response is to help to improve the partner's performance. When trust in
intentions is breached repeatedly, the response is more likely to be retaliation or the
breaking of the relation. But it is not easy to determine which of the two is the case.
When the breach is intentional, one can claim force majeure, i.e. a freak break of
competence, to prevent retaliation. Most discussions in the literature focus on
intentional trust, because that is the most difficult to deal with.
Intentional trust is a slippery notion, and this is the cause of dangerous misunder-
standings. A key question is whether trust is based on coercion and self-interest or
goes beyond it. A wide definition of trust is that one expects that no harm will occur.
But this may be due to the fact that contractually the partner has no opportunity to do
harm, or that it is not in his material interest to do so. According to a much narrower
definition, trust means that one expects that the partner will do no harm even though
he has both the opportunity and the incentive to do so (Nooteboom 1996, 1999b). If
someone says that you can trust him, which is meant? If people say they trust each
other and one means the one and the other means the other, unpleasant surprises are
likely to occur.
I agree with Williamson (1993) that if the notion of (intentional) trust is to yield
any added value in theory, it must go beyond the calculative self-interest that is already
covered in established economic theory. For instance, it must go beyond fear of
punishment by the law. It must go beyond loyalty based on the fact that it is in one's
material interest to maintain the relation, taking into account the 'shadow of the future':
the discounted value of future rewards in an ongoing relationship that one would
obtain by remaining loyal. It would need to go beyond reputation if that is based on
calculation of potential rewards in future relationships with others than the present
partner. I also agree with Williamson that trust should not be and rarely is blind. But I
disagree with Williamson that if trust is not to be blind it must be calculative. Trust can
be based on routine, or habituation: after a while, if things go well, actions and
procedures may be taken for granted; behavioral risks may no longer be part of 'focal
knowledge' (Michael Polanyi), as long as conditions and observed actions stay within
some range of tolerance. Only unexpected events trigger awareness and attention to
possible opportunism. As explained by Herbert Simon, such routinization is rational,
given bounded ness of rationality. When a certain pattern of action goes well, it is
conducive to survival to routinize that behaviour, and focus limited capacity of
rationality on new challenges. Thus routine can be non-calculative and yet have a
rational basis, in the sense of being conducive to survival, and in that sense it is not
blind. Uzzi (1997a) made the interesting proposal of using the notion of a 'heuristic' in
this context: some quick, efficient method of assessment that is neither calculative nor
blind. I suspect that it comes close to my notion of routine: the heuristic has arisen
from its survival value in the past.
Furthermore, trust can go beyond self-interest, in the realm of decency, loyalty,
solidarity or self-respect. Here we are dealing with trust in the narrow sense: the
expectation that the partner will not utilize opportunities for opportunism, even though
it is in his material interest to do so (Nooteboom 1996, 1999b). Trustworthiness in that
sense can be based on an intrinsic valuation of reputation, i.e. not as an instrument of
utility, or on self-respect, or on moral or social obligation. These can be related to
friend- or kinship. When related to membership of a social group it is in place prior to
the transaction. When it is not already present, it has to be built up in a relationship.
Habituation then plays a double role: in developing shared values and norms, and in
the sheer decay of awareness of what might go wrong, when things have consistently
been going right (Nooteboom 1996; Nooteboom et al. 1997).
But I accept that trust should not be blind, because trustworthiness has its limits: it
may be breached by a 'golden opportunity,' and some people are more resistant to such
temptation than others. Here I retain some of the perspective of TCE. But the
conclusion of the analysis is that no governance is needed when the advantage of
defection, or opportunism, does not exceed the resistance to temptation. And here I go
beyond TCE.
Trust between organizations can be based on inter-personal relations between staff
of those fIrms (see Pennings and Lee, this volume). In the resource view set up before:
relational competencies and competencies of motivation and morality. Motivation may

be based on personal friendship, kinship, empathy, habituation, shared norms and
values. Firm-level relational competencies can be found in institutions (rules, habits,
role models) that guide contact with other firms. Inter-organizational trust may further
be based on reputation and on survival conditions. When an organization is under
great pressure to survive, its loyalty may come under stress.
TRILA TERAL GOVERNANCE

Already in classical TCE the possibility was indicated of using a third party, in
'trilateral governance,' as an alternative to integration of the activity within the fIrm or
bilateral governance between fIrms, for solving the hold-up problem that arises in the
case of dedicated investments. This is recommended when the transactions involved
are infrequent, so that the volume of transactions does not warrant the cost of setting
up bilateral governance, while there are significant advantages of not integrating the
activity within the user firm. Thus the argument for trilateral governance is one of
efficiency: it is cheaper. Only limited explicit agreements between the partners are
made, and the most important one is procedural: If disagreement arises, the third party
will be called on to arbitrate. An important part of this role is to help the partners to set
realistic goals with balanced advantages and risks. A classic example is the role of an
architect as third party in transactions between a builder and a supplier of building
materials.
Clearly, a condition is that the third party must be trusted by both partners, in both
the competence and the intentional sense. He must be knowledgeable on the
technologies, markets and strategies involved, and he must be fair in judgement and
adjudication. The basis for this trust may be self-interest: the third party has an interest
in ongoing relations with both partners, and therefore has an interest in doing his best
and being fair, in order to maintain his reputation. The basis may also be more
intrinsic, stemming from desire, self-esteem or moral or social obligation. The
question then arises: if there is such a basis for trust in the go-between, why does it not
also arise between the partners themselves, so that the go-between is superfluous? The
answer is that between the partners there is the temptation to defect, because it yields
advantage, while for the go-between there is no such temptation. But note that the go-
between is not needed if between the partners the advantage of defection does not
exceed their resistance to temptation.
HOSTAGE KEEPING
Another instrument of governance, recognized by TCE, is the exchange of hostages, in
order to guarantee that agreements are kept. The characteristic of a hostage is its
asymmetry in value: it has value for the giver but not for the taker. This condition
serves to keep the taker from the temptation of keeping the hostage permanently, even
if the other side has kept his end of the bargain.
In relations between fIrms, people could be hostages: in intermarriage between
family businesses or the detachment to each other of key executives. It can also take
the form of cross-participation between fIrms, or pieces of technology or knowledge
that are sensitive to competition. The use of cross-participation as hostage taking is
admitted by the Japanese Agency for Economic Planning in a document in 1992

(Scher 1996: 17).
In hostage taking there is a well-known riddle. Suppose a king must supply a
hostage to another king with whom he is concluding a treaty, and must choose
between a charming, attractive daughter and a nasty, unattractive one, whom he both
loves dearly. The standard answer is: the latter, because that better satisfies the conditi-
on of asymmetric value. The more attractive daughter could more easily assume value
for the hostage taker, and thereby tempt him never to return her. But what if the first
king does not really intend to honour the treaty? Then it would be better to give the
attractive daughter, because then he has the greatest chance that the hostage taker will
not sacrifice her, in spite of the breaking of the treaty. But the second king might see
through this, and become suspicious when the attractive daughter is offered. But the
first king might trump this by signalling that he does not really love his unattractive
daughter, and the second king might therefore demand the attractive daughter, because
the unattractive one would not be an effective hostage, and thereby he would fall into
the trap of the first king.
Such problems can be reduced by having a third party keep the hostages. This
solution is antique; it was used in treaties between medieval rulers (De Laat 1996). It
works only if the third party is less tempted to retain the hostages after completion of
the agreement than the party in whose interest the hostage is taken. But note that the
third party would jeopardize his relation with both partners by not keeping the
agreement, and is thus likely to be disciplined by this double jeopardy. There would no
longer be the temptation to play the game discussed above. If the hostage giver
supplied his attractive daughter with the aim of breaking the treaty, and then did break
the treaty, his partner in the treaty would demand from the go-between to sacrifice her.
The ploy would no longer work.
REVELATION
Particularly when frrms cooperate for innovation, the knowledge they are pooling will
tend to be tacit, and to exchange tacit knowledge one must set up intensive interaction,
in which mutual understanding must be built up. Note also that the setting up of such
interaction and mutual understanding constitutes a dedicated investment. How do you
judge whether such investment is worth while before you commit it? This is the
revelation problem: how can value of knowledge or competence be revealed before it
is transmitted? Here we are dealing with Arrow's paradox: if you give information that
is sufficient to judge the value of the information, then little may be left to trade or
hold back. Here lies a third role for the go-between. Having ongoing relations with
both partners, it has already made investments in ability to understand and judge
competence. Thus it can act as judge of value and relevance for both partners before
they invest in their relationship. Having the trust of both partners, and knowing them
well enough to assess both on the value and relevance to each other, it can inform
them on that without actually supplying the underlying information.
SPILL· OVER CONTROL

Closely related to this is the role of the go-between in spill-over control. Partners do
not only want to have an indication of value before they commit dedicated
investments. They also face the problem that by giving information about their
knowledge they reveal at least part of that knowledge, and that may spill over to
competitors. The fourth role of the go-between is to serve as a screening device: the
go-between does not make the knowledge available to the partner, but assesses it for
him. Next, when the partners proceed to pool and exchange knowledge the go-between
can perform or monitor spill-over control. Spill-over control might otherwise require
mutual in-house monitoring of streams of information by the partners themselves, to
check that it does not leak to competitors, but such mutual monitoring may increase
rather than control information exchange, and hence aggravate rather than relieve the
spill-over problem. Especially at the beginning of the relationship such in-house
monitoring is likely to be unacceptable to the partners. Both the risk of spill-over and
the unwillingness to grant monitoring to guard against it will be greater to the extent
that the partners are actual or potential competitors, or have intimate linkages with
competitors.
Note that the whole issue of spill-over is relevant only under certain conditions.
The first is that the nature of the knowledge involved indeed enables spill-over. This is
not the case to the extent that the knowledge is tacit or embedded in the organization.
When knowledge is tacit, its transfer requires intensive mutual interaction, or a buy-
out of the people in whom the knowledge is embodied, and even that may not be
effective, to the extent that the performance of those people is contingent upon
supporting knowledge, procedures or organization that stays behind. The second
condition for the problem of spill-over is that the partner is himself a potential
competitor, or has connections with competitors. The third is that development is not
so fast that spill-over no longer matters. If change is so fast that by the time a competi-
tor has imitated it, the knowledge or technology has already changed, then the problem
of spill-over simply disappears (Nooteboom 1998a).
BUILDING AND MAINTAINING TRUST

Related to the previous roles, there is a fifth role in the building of trust. This was
recognized also by Uzzi (1997a), who described the role of a go-between in creating
the embedded ness of partnerships. It was also recognized in mechanisms in the
transmission of trust discussed by Deutsch (1973). One aspect of this is the transitivity
of trust: If X trusts Y and Y trusts Z, then X can be disposed to trust Z. Here Y is the
go-between. One often initiates relations with partners who are trusted partners of
one's own trusted partners. If X and Z both trust Y, then they may be disposed to also
trust each other. Here Y again is the go-between. Such bringing together of partners in
a disposition towards trust is only a beginning. Next, intermediation may help to set up
initial small steps of cooperation, and to ensure that they are successful and do not
raise misunderstanding that might antagonize or raise suspicion. This is important in
the building of both competence trust and intentional trust. It was noted before that
breach of competence trust calls for very different action from breach of intentional
trust. In the first case one might help to improve competence, but that would not help
in the second case, where it might be more appropriate to signal a warning or even
retaliate. But it can be difficult to tell which of the two is the case. The third party can
act as an independent observer, with access to inside information. He can eliminate
misunderstandings that might otherwise destabilize the relation. This role is
particularly important where trust is not pre-existent, as part of a group culture or set
of norms and values and has to be built up in the relationship.
The beginning of relationships tends to get most attention, but the ending of
relationships is at least as important, and often more difficult (Nooteboom 1996).
When in an existing relationship one partner runs into a more attractive alternative,
there is substantial risk of fierce antagonism, whereby the other partner tries to evade
loss by keeping the first partner from cutting loose, and the first partner retaliates with
nasty behaviour to badger the second partner to let go. In expectation of this, the fust
partner may keep his wish to exit secret, and prepare his defection in silence. But this
makes the problem worse for the second partner, because he has less time to adapt, and
may react all the more fiercely in trying to tie the fust partner down. One can observe
this in both fum relationships and marriages. The better road may be for the fust
partner to announce his intentions at an early stage, but stay on for the time being to
help the second partner to find an acceptable way out. But such an announcement
breaks trust, and the destructive spiral of mutual antagonism is difficult to prevent.
Careful counselling by a trusted third party may be needed to guide and control this
process.
An alternative is for the fust partner to give the relationship another chance, and
help the second partner to increase its attractiveness. But this entails that the second
partner further increases its relation-specific investments, while a threat has appeared
that the relation may break. Here also, trusted counselling may be needed.
BOUNDARY SPANNING
An argument against embedded, voice based network systems is that the relations
between fums, oriented towards the longer term, and based on relation-specific invest-
ments and trust, create exit barriers and entry barriers and thereby limit the efficiency,
flexibility and adaptiveness of the system. Uzzi (1997a) called this the 'paradox of
embedded ness.' This leads to a sixth role for the go-between: to act as a boundary
spanner between an existing network and potential outside sources of innovation. It
may be threatening for partners who are active in ongoing exchange relations to
maintain outside exploratory relations for scanning novel opportunities. In fact, one of
the most threatening and potentially destabilizing events in a trust relationship is the
appearance on the stage of a new player who might present a more attractive substitute
for one of the partners involved. The fear of this may create a taboo of outside
scanning in order to maintain the integrity of the network, and this can be disastrous
for innovation on the basis of outside sources. The problem of losing positional
advantage to an outsider, and being left with idle dedicated assets, is aggravated by
spill-over risk: one may also lose part of one's core competence. A go-between who
does not himself participate in exchange can perform outside scanning under less
suspicion, since it is not in his own direct interest to defect to more promising
outsiders, while it is in his, interest to maintain his reputation for confidentiality and
Table 2. Fonns of governance

exit system voice system
Elements of governance
specific investments low high
switching costs low high
tailored value partner low high
contractual restrictions high low
opportunism high low
Results
product costs low high
transaction costs high low
productdUferentiation low high
incremental innovation low high
creative destruction high constrained
fair dealing. Summing up: outside scanning is less threatening when done by a go-
between without a direct stake in exchange, and this opens up the network to potential
novelty from outside.
In fact, this role is related to some of the previous ones: the role of revelation,
spill-over control and the management of trust. One can bring in judgements of the
potential of outside novelty while maintaining confidentiality of its content and of its
source, without which one would not have obtained that information. And conversely:
one can give an assessment of the value of what the network could offer to outsiders
without giving it away and without specifying the precise source in the network
(which may tempt the outsider to pry that loose from the network). When novel
opportunities for redesigning the network have thus been identified and assessed, the
go-between can next help in the phasing out of the 'losers' in the old network, by
providing ways out, in exit to another network, or helping them to break up and
redistribute their resources. The go-between might even administer some insurance
developed by the network to deal with such contingencies.
INCREASING FLEXmILITY
When translated to specific aspects and outcomes of governance of inter-firm relations
(Nooteboom 1996), the differences between the exit based and voice based systems of
Table 1 are as summarized in Table 2.
As already indicated, the main argument against the voice based network systems
of capitalism is that the relations between firms, oriented towards the longer term, and
based on relation-specific investments and trust, create exit barriers and limit the
flexibility of the system, and thereby obstructs especially radical innovation. However,
it should be noted that in the Anglo-American system due to lack of network structures
between firms activities are more integrated within larger firms. That system must
derive its flexibility from easy break-up of firms, which also has disadvantages:
shorter contracts and assignments of labour, with less job security, yielding less
willingness to engage in firm-specific learning, less scope for cooperation in teams,
and a short term orientation which hinders innovation. However that may be, it cannot
be denied that the flexibility of voice based systems could be constrained. It then is
important to make relations as flexible as possible, while still being consistent with the
orientation towards mutual advantage and cooperation.
To make the voice-based system more flexible, relations between frrms should be
long enough to ensure pay-back of specific investments, but short enough to allow for
sufficient flexibility of re-aligning relations. When trust is not already present, and has
to be built up within relations, this also takes time and creates a danger of rigidity. Set-
up and maintenance costs of governance should be low enough to facilitate mUltiple
and flexible relations. Spill-over control should not constrain the formation of multiple
relations too much. The third party can help here, to offer trilateral governance,
keeping of hostages, revelation, spill-over control, the building of trust and boundary
spanning. In other words: The third party can help to make the system as flexible as
possible. Thus the role of the go-between is part of the social capital of voice based
systems, and helps to make it flexible.
EVIDENCE
I propose that the roles of the go-between explain the important role of banks and trade
and craft associations in European systems, and of banks and trading companies in
Japanese enterprise systems (Kigyo Shudan, cf. Scher 1996).
There is evidence that in the Netherlands the recently instituted 'Innovation
Centres' play such roles. Not entirely consciously, and largely by trial and error. This
evidence comes from my own experience in giving lectures along the lines of this
article to the centres, and informally testing to what extent they play the roles
indicated, by asking them to give specific instances. The role of trilateral governance
was recognized clearly: the coaching of partners in lieu of contracts; particularly the
setting of realistic goals, monitoring their achievement, eliminating misunderstandings
and making adjustments along the way. This role was clearly integrated with the role
of building and managing trust, as one might expect. The role of revelation was also
clearly recognized: here lies the central mission of the centres. Hostage keeping and
spill-over control were not recognized directly, but part of the experience was that the
care and confidentiality with which information was treated was enhanced by the fact
that the centre acted as a go-between. This could be interpreted as a hostage
mechanism: players were careful with information received from their partner because
sensitive information from themselves was held by the go-between. It could also be
interpreted as a reputation mechanism: they want to maintain a good reputation with
the centre, not to jeopardize future dealings and potential future partnerships. The role
of the boundary spanner was not relevant here because the networks in which the
centres operated had not yet reached the degree of consolidation in which such a role
becomes relevant.
Evidence is also found in a longitudinal case study by Klein Woolthuis (1996) of
cooperation between eleven frrms, in various configurations, in the development of
medical products, based on an initiative by an Innovation Centre and a Regional
Development Centre (ROC) in the Dutch province of Overijsel. Development projects
were submitted for subsidy in the EU EFRO programme (when awarded, it yields a
50% subsidy of product development costs). The Centre and the RDC were
particularly important in the start-up stage. An important factor for the firms was that
they trusted the initiators, and hence also the partner firms that they brought together.
Yet, 'parties were reluctant to reveal information and were hesitant to engage in any
relationship before it was completely clear what the other's motives were and what
role they would fulfil... This made the start difficult and time consuming.... Therefore
it was important that the Centre and the ROC took on a guiding role ..... they could ease
the negotiations ....this included chairmanship at meetings, encouraging and guiding
the first contacts.' This gives evidence of the roles of the go-between in trilateral
governance and the building of trust.
An external committee was instituted, with one member from the Ministry of
Economic Affairs, and two managers from local hospitals, to help in the assessment of
project proposals and to watch over fair distribution of work and subsidies from
EFRO. It is noted that 'It is very important that the committee should make fair
evaluations. Parties should be able to trust the board ...The perception of equal
treatment is therefore of crucial importance.' The task of the 'umbrella' of the go-
betweens, consisting of the Centre, the ROC and the advisory council, was
summarized as follows: 'In the first stage the umbrella performed an important role
because it provided firms with a platform to get to know the others. The umbrella
became of decreasing importance when direct links between firms evolved .... frrms
started to know each other and envisage future potential for more dense relationships.'
This again indicates the role of trust building, and also suggests the role of revelation.
There is no mention of the roles of hostage keeping and spill-over control. They may
have been present but unnoticed by the researchers reporting the case.
It is interesting that the go-betweens who were the initiators (the Centre and the
RDC) instituted a separate advisory council to guide the development of projects. This
makes sense, for two reasons. First, it was noted that the go-between requires expertise
to generate competence trust, and it may be too much to expect relevant and
sufficiently deep expertise of a single agency in a wide variety of relations. With a
separate council the go-between can seek specialized support that is tailored to the
situation. That is why in this case hospitals were included. Second, by creating a
special council, the go-between insulates itself to some extent from the risk of losing
reputation when something goes wrong in the judgement and guidance of cooperation:
then the advisory council takes the direct blame, and since it is unique to the situation,
its loss of reputation is less serious.
DISCUSSION
Six roles have been identified for the go-between: trilateral governance, keeping of
hostages, revelation, spill-over control, trust building and maintenance, and boundary
spanning. All roles except the first are especially important in the building of new
relations in innovation. It is especially there that the use of complementary
competencies is important, while they are not always visible, and exchange of
sensitive information is important. By playing these roles, the go-between can
contribute to the flexibility of network systems of frrms. Thus the go-between forms
an important ingredient of the social capital on which the networks of the voice-based
system of capitalism are based.
However, the roles of the go-between are not easy. They require thorough
knowledge of the technologies involved in the firms involved, as well as the
commercial interests and strategic interests and risks at stake. A sharp insight is
required of the factors which play a role in relationships, instruments of governance
and skill to handle them (Nooteboom 1996). Reputation, in the both competence and
fairness, is crucial, is difficult to build up and easy to destroy. The demands and risks
involved may be too much for any single agency. The go-between can solve this
problem by using special committees or councils to guide specific relations, which are
tailor-made in their expertise, to generate competence trust, while the go-between
itself monitors intentional trust. In this way, the go-between also insulates itself to
some extent from the risk of failing to perform all the balancing acts without a flaw:
part of the blame is then taken by the committee, that can be modified or replaced
while maintaining the position of the go-between himself.
The roles need not all be performed by a single agent, as the above analysis and
the preceding case study show. They can be distributed over several agents. One
reason for this can be to distribute risks of loss of reputation due to error. As I
indicated from the start, the roles are not the traditional roles of middlemen or
entrepreneurs, but that does not preclude that some of the roles are in fact played by
parties who are directly involved in production and distribution. In any network or
team, any participant may help to arbitrate, keep a hostage, reveal competence, control
spill-over, build trust and act as a boundary spanner for other participants. Some of this
is found in the notion of the 'closure' of social structure (Coleman 1990). But some of
the roles require some degree of independence; no direct involvement in exchange
between the others.
The analysis indicates that there is an emerging market for go-betweens. The roles
may be played by government or semi-government organizations: municipalities,
provincial development associations, innovation or technology transfer centres, but
also by banks, commercial consultants. Government agencies tend to be weak in
generating competence trust but strong in generating intentional trust. Thus
combinations of local semi-government agencies coordinating and monitoring
professionals from business may yield the optimal form.
The Management of Social Capital
in R&D Collaboration
•
19
OnnoOmta
Wouter van Rossum
ABSTRACT
Rapidly developing technologies, increased global competition, and more stringent
customer demands are compelling companies to improve the pace and quality of
their product and process innovation. As a consequence, companies are increasingly
pushed to form technological partnerships to share development costs and to reduce
time-to-market. This chapter focuses on the 'dark side of cooperation' by
concentrating on the complications and pitfalls encountered by thirty general and
R&D managers from ten leading companies in R&D collaborations in the
automotive, electronic, energy, heavy, and pharmaceutical industries. The partnering
relationships described include strategic R&D alliances, supplier-customer
partnerships, and cooperative relationships with knowledge institutions. The
empirical research shows that these relationships are extremely vulnerable. For
example, the partners may worry about releasing too much confidential information
and technology. The occurrence ofthese problems can be explained by social capital
theory. Network relations may enhance the social capital of a company by making it
feasible for it to get easier access to information, technical know-how, and financial
support. But, at the same time, these relationships may lead to social liability, e.g.,
by reducing the possibilities for relating to companies outside the network, risk of
spillover, and high coordination costs of the networkrelations. R&D relationships,
especially, are, for the most part, not very tightly knit, and hence their problems
relate to lack of information and to opportunism. This chapter then focuses on the
methods managers can use to minimize these problems by reducing social liability
and hence enhancing social capital. Partnership management is considered to be, to
a great extent, management of trust and goes far beyond signing confidentiality
agreements and agreeing to follow guidelines. This chapter ends with possible
The Management of Social Capital in R&D Collaboration - 357
solutions for every phase of the collaborative lifecycle. It includes an outsourcing

and partnering matrix as a management tool for the strategic foundation of make-
buy-or-collaborate decisions.
INTRODUCTION
Rapidly developing technologies, increased global competition, and more stringent
customer demands are compelling companies to improve the pace and quality of
their product and process innovation. Huge R&D investments may be lost when a
final product does not measure up to the required standards or when a similar
product of a competing company reaches the market earlier. The pressure to do more
with less inexorably pushes companies to focus on few, unique, hard-to-imitate, and
distinctive core competencies by continually nurturing and enhancing them, while
abandoning activities in which they do not possess distinctive competencies. These
R&D-intensive companies, while emphasizing their core competencies, are
increasingly seeking new upstream and downstream partners to share development
costs, accelerate product and process development and maximize commercialization
opportunities. The ability to build and maintain interorganizational network
relationships-such as joint ventures, license agreements, supplier-customer
partnerships, and strategic alliances-is increasingly viewed as the key factor to
sustained competitive advantage. Several authors refer to such R&D networks as
virtual corporations in which a number of firms create flexible linkages to attain
common or complementary objectives (e.g., Davidow and Malone 1992; Campbell
1996; Upton and McAfee 1996; Yoshino and Srinivasa Rangan 1995).
Innovation within technology networks is extensive because of the sustained
interaction between institutions and commercial organizations of different size,
capabilities, and expertise. In Silicon Valley for instance, various forums exist for
high-tech companies where relationships can be formed and maintained, information
exchanged, contacts made, and new ventures formed (Saxenian 1990). Forrest and
Martin (1992) note that in biotechnology alliances are frequently successful, since
the peculiar strengths of small firms (entrepreneurship and innovational climate) are
complementary to those of the large pharmaceutical firms (critical mass for
development activities and scaling-up expertise). As Gambardella (1992) concludes,
based on an extensive study of the relations between in-house scientific research and
external scientific knowledge in the U.S. pharmaceutical industry: To be part of a
network, and to be able to effectively exploit the information that circulates in the
network, has become even more valuable than being able to generate new
knowledge autonomously.'
As a result, in some high-technology areas, such as biotechnology, electronics,
and software, a large portion of the knowledge base for new products is shared,
rather than proprietary, knowledge (e.g., Cabo et al. 1996; Chatterji 1996). R&D
collaboration between firms and knowledge institutions is rapidly becoming the rule
rather than the exception (e.g., Biemans 1992; Millson et al. 1996). Currently, more
than 50 percent of Du Pont's new product leads in agriculture stems from university
laboratories (MacLachlan 1995). In the pharmaceutical industry, too, knowledge
institutions are critical partners in all aspects of new drug development. Taylor
(1994) estimates the innovative pharmaceutical industry to spend ten to twenty
percent of the total R&D budget on collaboration and sponsoring of universities and
government laboratories. This current trend of forming formal and informal R&D
partnerships is likely to continue. CEOs of fifty leading companies believe that by
the year 2000 close to 50 percent or more of their technological competitiveness will
derive from external sources and partnering (Jonash 1996).
These external relationships might be a reasonable response to the business
pressures, but, at the same time, they may create new long-term dependencies and
vulnerabilities, as companies are becoming increasingly dependent on outside
sources for their technological advances (e.g., Millson et al. 1996). For instance, if
industry is going to entrust critical parts of its research to outsiders, there must be
confidence that timing to produce results will be respected. Business heads worry
about security, cost-effectiveness, and relevance of results. They have the view that
only R&D under their direct control can be held to a focus on business needs
(Jonash 1996).
The occurrence of these problems can be explained by social capital theory.
Network relations may enhance the social capital of a company, by making it easier
to get access to information, technical know-how, and financial support. But, at the
same time, these relationships may lead to social liability (by reducing the
possibilities to relate to companies outside the network, risk of spillover, and high
coordination costs of the networkrelations). R&D relationships, especially, are, for
the most part, not very tightly knit, and hence the problems relate to lack of
information and to opportunism. This chapter focuses on social liability by
concentrating on the complications and pitfalls encountered in ten R&D
collaborations and then on the methods managers can use to reduce social liability
and hence enhance social capital.
TECHNOLOGICAL COLLABORATION
In management literature a variety of terms describe the phenomenon that
organizations are no longer defined in terms of their assets alone but increasingly by
their relationships with other organizations, as well. In reviewing the literature,
Campbell (1996) mentions the virtual corporation, the virtual organization, the
network organisation, the modular cooperation, and the web as examples of this
notion. The terminology used to describe technological collaboration is variable, as
well. The present study starts from the general definition given by Dodgson (1993),
who describes technological collaboration: 'as any activity where two or more
partners contribute differential resources and know-how to agreed complementary
aims.' Chatterji (1996) defines this broad terrain of business relationships to include
acquisitions of small companies, exclusive licensing of specific technologies, joint
ventures, minority equities, options for future licenses and joint development, R&D
contracts, and seed funding of exploratory research at universities, independent
research organizations, and start-up companies. This study concentrates on
technological collaborations (joint ventures, joint developments, and R&D
contracts) that continue for a definite period to achieve a common objective. These
might or might not shift into long-term alliances. The focus is on the complications
and pitfalls encountered by the management of the leading company. The leading
company brings in the original concept for the cooperation, where it is then further
developed in cooperation with the partners, which add varying degrees of value
(Lorenzoni and Baden-Fuller 1995).
Motivation to start and factors undermining the success of a collaboration

A number of reasons for starting a collaboration are identified in literature (e.g.,
Dodgson 1993; Hagedoorn 1993; Bruce et al. 1995; Bailey et al. 1996; Chiesa and
Manzini 1996). These include improving the efficiency of the developmental
process through spreading costs and reducing risks of product development,
allowing speedy access to new technology, widening the firm's breadth of
technological competence, and providing a window for monitoring technological
advances. In addition, technological collaboration can play a key role in
implementing corporate strategy-for example, in diversifying, globalizing, and
overcoming trade barriers.
Bruce et al. (1995) conducted a survey into the advantages and risks of
technological collaboration in 106 UK-based companies in the information and
communication industry. More than half of the respondents expressed as their view
that collaborations do not accelerate product development but make product
development more costly, in general. Bruce et al. expect that some of these problems
may be overcome by organizational learning because the companies in their sample
with much experience in collaboration (more than 25 percent of R&D-effort
stemmed from collaborative effort in the preceding two years in 37 percent of the
sample) noted these problems less often. Bruce et al. extract a number of risks in
technological collaboration from the management literature:
• Alliances are unstable and difficult to manage because of reduced direct
management control (MacLachlan 1995);
• There is a risk of creating a rival (Lorange et al. 1992);
• There is a risk of creating a dependency on a key supplier or partner;
• There is a risk of leakage of skills, experiences, and competencies that may
form the basis of the competitiveness of the firm;
• There is a risk of leakage of information and insights into possible markets and
future possibilities;
• There may be hidden administering costs of setting up and monitoring of a
collaboration;
• There may be costs of time in harmonizing different management styles and
budgeting processes (Farr and Fischer 1992).
Although most authors don't refer to social capital theory, the reasons for starting a
collaboration and the pitfalls and complications encountered refer to corporate social
capital and liability.
Technological collaboration as a process

Technological collaboration should not be considered a discrete event but a process
(e.g., Hamel et al. 1989; Chatterji 1996; Millson et al. 1996) characterized by a
sequence of phases that require the allocation of tangible and intangible resources,
and need a structure of organizational procedures and routines for managing them.
Chiesa and Manzini (1996) describe the process of technological collaboration
through a phase model, as shown in Figure 1.
Define the Select the Establish the Implement the

firm's goals ~ partners f+ partnership f+ partnership ..... Outputs
t f f
f
I Implement corrective actions
I'-
Measure network's
results
Figure 1. Phase model of technological collaboration (Chiesa and Manzini 1996)
An R&D co-operation is composed of the following phases:
1. Goal definition In the goal-definition phase the firm defines the goals for the
collaboration in terms of its short-term objectives (such as exploitation of a
definite innovation or development of a new product) and long-term objectives
(such as building skills in a technology area). This phase is extremely
important, because by doing its homework carefully, a company may avoid a lot
of trouble later in the collaboration.
2. Partner selection In the-phase one or more potential partners are identified.
This includes collecting information about and exploring the strengths and
weaknesses of potential partners and carefully assessing the most appropriate.
3. Establishment of collaborative agreement In this phase the partners get to know
each other and define the guidelines for the collaboration. Agreements are made
concerning aim, scope, time, type (such as strategic alliance or joint venture)
and organization of the collaboration. In this essential phase the basis of trust on
which a successful technological collaboration stands is built. Despite its
obvious importance, it is seldom distinguished as a separate phase in
management literature.
4. Implementation In this phase the partners start to collaborate in line with the
guidelines defined in the previous phase. Network activities should be
stimulated to maximize integration, communication, and learning.
5. Performance assessment and implementation of corrective actions During and
at the end of the collaboration, the partners evaluate whether the objectives are
achieved, measure the results, and verify whether the guidelines are actually
respected. Decisions on corrective actions should be taken collectively by the
partners if they relate to modifying the rules for the collaboration.
Research Questions
The present study was designed as an explorative study regarding factors
influencing R&D collaborations. An explorative approach was chosen because
there is still a mismatch between the literature, which emphasizes the importance of
R&D collaboration and indicates ideal typical solutions for starting and developing
partnerships, and empirical reality in which a range of many different forms of
Table 1. Partners and objectives 0 f the collaborations in different industry sectors

Partners Objectives
Energy Dutch power plant and knowledge institution Simulation-based power plant
and U.S. software company controlling system
Dutch power plant and knowledge institution Controlling systems for small-scale
and manufacturer of energy controlling power generation (two collaborations)
systems
Dutch power plant and knowledge institution Environmental friendly wood- and
and U.S. software fmn and Swedish wood coal-based energy generation
supplier
Dutch gas distributor and three U.S. power Advanced motor management
plant manufacturers Systems
Automotive Two automotive companies (Gennany, France) New coating for motor bodies
and Dutch steel company and Chemical
conglomerate
Six auto motor body companies Conversion of aluminium motor
bodies
Office Dutch copier manufacturer and suppliers New generation of high-end copiers
systems
Software Irish software company and major clients Tailor-made software solutions
Electronics Japanese e.Iectronics company and four High-perfonnance parallel and optical
universities (UK, France, Gennany, Italy) computing
R&D collaborative efforts can be found, each with its own problems and pitfalls. In
this situation it seemed sensible to first collect information on different forms of
R&D collaborations, then try to explain the findings, and finally formulate the
practical consequences for managing R&D collaborations.
Study Sample
The present study was conducted in 1996. Thirty half-structured interviews were
conducted with general and R&D managers of ten leading companies in
technological collaboration about the complications and pitfalls encountered. I All
managers had been actively involved in the preparation and/or the execution of the
technological collaborations at issue.
Five collaborations took place in the energy sector (see Table 1). Three of them
were initiated by different Dutch power plants and a central research institution
concentrating on energy and environmental issues. The other two were initiated by
the largest gas distributor in the Netherlands. They were all co-developments. The
first collaboration was directed toward the development of a new simulation-based
controlling system for power plants. Two others were directed towards the
development of advanced controlling systems for small-scale power generation. The
fourth concentrated on the development of environmental friendly wood- and
coal-based energy generation. The fifth was concerned with advanced motor
management.
The partner in the first collaboration was a U.S. software company, in the
second a Dutch manufacturer of energy controlling systems, in the third (another)
U.S. software firm and a Swedish wood supplier, in the fourth, three U.S. power
plant manufacturers; and in the fifth a Dutch manufacturer of motor systems.
Two collaborations took place in the automotive industry. The first was a co-
development of a new coating for motor bodies between two automotive companies
in Germany and France, a large Dutch steel company, and a Dutch-based
multinational chemical conglomerate. The second was a long-term R&D
collaboration of six motor body companies, currently concentrating on the
possibilities of converting aluminium bodies. The seventh collaboration was a co-
development between the largest copier manufacturer in the Netherlands and its
main suppliers to develop a new product family of high-end copiers. The last two
were networks-the first in the field of software development and the other in the
electronics industry. A worldwide operating software company, based in Ireland,
daughter of a large multinational company, works together in supplier-customer
partnerships with their clients to provide them with tailor-made software solutions.
The last collaboration was a university-industry collaborative network in the area of
high-performance parallel and optical computing in four university departments in
the United Kingdom, France, Germany, and Italy with a Japanese electronics
company.
RESULTS OF THE EXPLORATIVE STUDY

Motivation to Start a Technological Collaboration
From the interviews it became apparent that the general motives to start a
partnership were the following:
• To accelerate time-to-market,
• To improve the cost-effectiveness of R&D,
• To develop stronger technology competencies, and
• To broaden the scope of technology reach and/or geographic coverage.
Unfortunately, in none of the cases were all these objectives met because of
communication and coordination problems, misunderstandings and hidden agendas,
which led to overrunning of budget or loss of interest by one of the partners. From
the interviews it became apparent that technological collaborations are extremely
vulnerable, even more than other (supplier-customer) partnerships.
Complications and Pitfalls

The following complications and pitfalls were encountered in the different phases of
the technological collaboration.
Goal Definition
R&D partnerships are mostly directed toward new technologies and markets.
Therefore, careful upfront study is essential for a well-balanced assessment of the
technological and business opportunities. In two cases the partners were so
enthusiastic about the collaboration, that without further study the business
opportunities were considered to be good. The respondents stated that, although no
market turned out to exist for the developed product, there was gain of beneficial
experience in terms of organizational learning and technical know how. MacLaghlan
(1995) comments that where termination of an R&D project can create trauma inside
the R&D organization of an individual company, this is even more so in a
collaborative effort. Bruce et at. (1995) point in the same direction, by commenting
that the collaboration as such might establish its own agenda. The overriding desire
of the partners to ensure that the collaboration will be successful may cause the
partners to become blind to technological and market reality.
Partner Selection
Insufficient monitoring of the R&D environment may lead to obvious partners being
overlooked. This is often caused by insufficient management commitment to finding
a partner. In one case it took five years to find a partner with the required
competencies. The resulting partnership missed the business opportunity because the
market had changed in the five-year period. In another case, a leading university
department in the country of the lead company was ignored for more than half a year
because the company was searching worldwide, ignoring the possibility of finding
excellence around the corner. In two cases the technological capability and the
financial resources of the partner turned out to be insufficient to conduct its part of
the collaboration successfully.
Establishment of Collaborative Agreement

One of the problems that emerged from the interviews was that the technological
collaboration had evolved in a merely ad hoc way, while the interests of the
company were insufficiently secured. In five collaborations inconsistency with
corporate interests was mentioned to have caused (severe) problems later in the
collaboration. This included lack of clear agreements about the division of the
financial and R&D efforts among the partners and insufficient clarity about the way
the collaboration should be organized and managed. The omission with the most
severe consequences, however, was the lack of a clause into the contract about the
distribution of the potential gains. The partner stepped out of the collaboration to use
the knowledge competitively.
A number of respondents stated, however, that R&D partnership gains are often
too uncertain, and gains too unexpected to capture them all in contracts. Too
detailed contracts are also contraproductive. In two collaborations the respondents
indicated that their (American) counterparts showed up at the constitutional meeting
with their corporate lawyers and fist-sized contracts. This deterred the Europeans
and extended the negotiations considerably.
Implementation
The most important problem encountered was fear and distrust. Six out of the ten
technological collaborations suffered from this at any time during the collaboration.
The partners worried about releasing too much confidential information and
technology because they feared a hidden agenda and opportunistic behavior of the
partner. Two collaborations failed, because one of the partners was more interested
in the short-term exploitation of strategic information than in the success of the joint
program. In the pharmaceutical industry, where concentration tendencies are

widespread (Omta 1995), fear was expressed that the current R&D partner might
later merge with a competitor, which could lead to a drainage of sensitive
information. One of the respondents of a university-industry collaboration expressed
a fear that strategic information would leak out because a number of the graduate
students of the university department got jobs in competitors' R&D departments.
Furthermore, respondents warned researchers about asymmetric technological
collaborations between large and small companies. The small partner may fear (or
hope) that if the technological collaboration turns out successfully, it will lose
independence and be taken over by the financially stronger partner.
In three cases intercultural problems were encountered. In one case, language
problems and cultural differences hindered open communication between Japanese
and European companies. But problems also arose with American companies. The
problems caused by differences in negotiating culture were already mentioned. In
two cases there were also complaints about lack of openness and the provision of
scanty, not up-to-date information by the American counterparts.
Peiformance Assessment and Implementation of Corrective Action

The partners of large R&D networks tend to underestimate the communication and
coordination problems and costs that are encountered in large collaborations. Four
large R&D partnerships overran their budgets. Obviously, management still
encounters difficulties in coping with the coordination of networks. The central
problem in this respect is the absence of a steering agency. The network, or rather
the structure of the network, to a large extent determines how actors work together.
Consequently, executing R&D within the confines of a network requires different
steering solutions than the execution of in-house R&D. Moreover, the fit between
the nature of the research problems to be addressed in the network and the type of
the structure of the network is not always self-evident. Quite often the structure of
the network is not explicitly chosen but merely accidental-for example, based upon
previous experiences, as was found in a recent study of upstream R&D cooperation
between firms and knowledge institutions within the BIOTECH program (Cabo et
al. 1996).
A Social Capital Explanation of Problems in R&D Collaboration

The explorative study reveals that R&D collaboration is widespread but that many
relationships are characterized by social liability. More in particular, the study
shows that fear and distrust-that is, the risk of opportunism by the partner and the
lack of adequate information on (possible) partners-seem to be attributes of many
relationships. Apparently many of the firms investigated have difficulties adequately
managing their R&D partnerships and hence enhancing their social capital. How do
we explain this high frequency of problematic R&D relationships?
A first explanation refers to the nature of R&D and its role with respect to
corporate goals. The results of R&D are difficult to appropriate, patenting only
covers part of these results and hence firms will be secretive about their R&D
activities. According to this view it is considered to be rational to perform R&D
in-house and refrain from collaborative R&D activities. The empirical evidence
indicating complications and pitfalls is then an indication of the nonrationality of

R&D collaboration. However, there are several reasons why this explanation does
not suffice.
First, R&D ranges from fundamental scientific research to concrete
development of products and processes. No firm can encompass all these R&D
activities and thus it has to at least cooperate with the firms and research institutes
having the competencies that are not available (and are also not economic to have)
in the firm itself. The recent tendency to outsource those activities that are not
central to the central mission of the firm results in an increase of collaborative
activity in this realm, as well.
Second, some R&D activities are simply too expensive for the single firm; in
these cases firms are forced to cooperate, sometimes even with competitors.
Third, in some fields oftechnology, such as biotechnology, a large generic body
of knowledge is shared by firms and research institutes. Here it is economic to
engage in collaborations instead of developing a proprietary knowledge base. In
other words, if R&D collaboration is more the rule rather than the exception, one
needs to look for other explanations, instead of pointing to the nonrationality of
R&D collaboration.
The social capital argument assumes that the behavior and expectations of actors
are constrained by the degree to which the relationship between the actors is
embedded in the network structure. Consequently, one may distinguish between a
situation in which network structure is closely knit and relationships are redundant
(Le., actor A has relationships with actors Band C, and Band C also have a mutual
relationship) and a situation in which more non-redundant relationships prevail.
Coleman (1988) describes this phenomenon as the degree of closure of a network. In
a similar vein, Granovetter speaks about strong versus weak ties in a network
(Granovetter 1985). Burt (1992) formulates it slightly different as the occurrence of
structural holes. Conversely, it can also be argued that, considering the role of social
relationships in technological fields such as biotechnology, the absence of adequate
social relationships can be seen as a liability of research laboratories in such fields,
as well.
Networks with relatively many nonredundant relationships provide less social
constraints. Granovetter (1985) argues that this network effect occurs via the
mechanism of reputation (see also Raub and Weesie 1990). If a network comprises
many redundant relationships, relatively many actors will have information on the
actions of an actor. The actor's reputation (as assigned by the others in the network)
depends on this information. Accordingly, actors will refrain from (otherwise
individual profitable) opportunistic behavior because they anticipate that it will
affect their own reputation in the network. As Raub and Weesie (1990) show in a
series of game-theoretic experiments. this effect is diminished when the degree of
embedded ness decreases.
In our view. the network of R&D relationships is a prime example of a network
with relatively few redundant relationships. Actors do not look for similarity in the
attributes of their partners in the case of R&D relationships but, instead, attempt to
acquire partnerships with actors having complementary rather than similar attributes.
This will result in less densely knit networks. In fact. one may expect that firms
looking for R&D relationships will attempt to acquire partners not in their direct
vicinity but in remote locations of the network. Firms will, especially in the case of
R&D relationships, strategically invest in their (R&D) social capital by emphasizing
nonredundant relationships or, in Burt's words, relationships that span structural
holes.
If this argument is correct and the number of nonredundant relationships is
relatively high in networks of R&D relationships, one may expect a lack of social
constraints and as a result the occurrence of problems as indicated by the
exploratory study. These problems will be characteristic of all R&D relationships:
R&D networks are less dense and therefore do not generate the social norms
facilitating consensus on mutual expectations.
Although the exploratory study reveals a list of various different types of
pitfalls and complications, they nevertheless all point to inadequacies of network
structure. The problems indicated by the respondents are either problems related to
their own behavior toward partners (such as the uncritical acceptance of partners
without an adequate check of their relevancy for a project), the behavior of the
partner (such as a short-term exploitation of strategic information), or, and
predominantly, problems of inadequate mutual expectations resulting in fear and
distrust (suc as the fear of leaking of strategic information). Nooteboom's (this
volume) argument about the importance of including a trusted third party (the
go-between) in an R&D collaboration, also refers to the importance of increasing the
information level about the partner(s), enhancing trust by a party from outside the
network.
In the exploratory study we did not systematically collect information on the
nature of the respective R&D network structures. This will be the topic of a follow-
up project. In the remainder of this chapter we want to focus on the practical
consequences for the management of such R&D partnerships.
In the concluding chapter in Nohria and Eccles' collection of articles on
networks and organizations (Nohria and Eccles 1992), Moss Kanter and Eccles
correctly observe that there is a large discrepancy between the academic discussion
on attributes of networks of organizations and the completely different perspective
from which managers deal with their networks (Kanter and Eccles 1992: 521-527).
Accordingly, these authors request more practically inclined network analyses.
Our intentions in this chapter are, practical, although less far-reaching. If pitfalls
and complications of R&D partnerships are related to attributes of the R&D network
structure, then management tools should be developed that enable managers to
adequately reduce social liability and hence enhance the social capital of such
loosely knit networks where they cannot assume adequately developed expectations
and mutually attuned behavior.
GOAL DEFINITION
A number of problems could have been avoided if the technology forecasting was
more carefully assessed. Companies need to establish robust technology forecasting
systems, that express technology needs for the next five to ten years, and monitor
and interpret developments in different technologies, the emerging trends in
customer needs, and competitor actions. Based on these forecasting activities
Table 2. Competitive impact of technologies

Technology Description
Emerging Has not yet demonstrated potential for changing the basis of
competition.
Pacing Has demonstrated potential to change the basis of competition, but has
not yet been embodied in products or processes.
Key Is embodied in products and processes. It has a major impact on
product or process value-adding (in terms of cost, quality or
performance); and allows proprietyl patented positions.
Base Is required to be in business. It offers little potential for value-adding,
because it is widespread and shared among competitors
technology road maps can be set up that link future product plans to the technologies
required to achieve them. Many authors (e.g., Hamel et al. 1989; Roussel et al.
1991) have emphasized that firms should develop only a few strategic technological
capabilities and should outsource the other ones. The emerging role of R&D
management is to balance internal and external technological capabilities by
identifying the projects that are feasible to take outside, and match these with
external sources. Via management tools such as the outsourcing and partnering
matrix (discussed below), firms can decide which technologies should be developed
in-house, which in collaboration with one or more partners, and which technologies
are better outsourced.
Outsourcing and Partnering Matrix

The first step in framing the outsourcing and partnering matrix is to establish the
competitive impact of the company's technologies by dividing them in emerging
(embryonic), pacing, key, and base (existing) technologies (Roussel et al. 1991), as
shown in Table 2. To this end, for each technology the following questions must be
answered:
• Does this technology have the potential for competitive differentiation?
• Could it become critical to the company?
• What is its market value?
The second step is to assess the internal technological capability for each of the
technologies using the categories weak, moderate, and strong. By combining the
competitive technology impact with the internal technological capabilities, we get
the outsourcing and partnering matrix, that can serve as a comprehensive
management tool for the strategic foundation of make-buy-or-collaborate decisions
(see Table 3).
Emerging Technologies
An emerging technology may have a competitIve impact in the future. If the
technological capability is strong, optimizing the technological capability to
reinforce the potential competitive advantage is called for. If the internal
technological capability is moderate or weak, catching up may be necessary.
However, uncertainty demands that the R&D environment be scanned-for many
Table 3. Outsourcing and partnering matrix (Harris et a1. 1996)

Competitive impact of Internal technological capability
Technology Weak Moderate Strong
Emerging Scan Scan/collaborate Collaborate
Pacing Collaborate Share risks In-house
Key Optimise Optimise In-house
Base Outsource Outsourcelexchange SelVexchange
partners and flexible relationships, preferably in strategic partnerships and alliances

or via contract research and sponsoring of knowledge institutions. In all cases,
adequate patent protection strategies need to be considered.
Pacing Technologies
Pacing technologies may have a strong competitive impact on the short or
medium-term. If one's technological capability is strong relative to the rest of the
world, the bias should be towards doing the work in-house. Extra investments might
be required for research into the application of the technology in new products and
markets. If one's technological capability is moderate, sharing the risk by strategic
alliances with partner firms makes the most sense. If one's technological capability
is weak, acquiring licences or joint development may be viable alternatives. Pacing
technologies need utmost management care, especially if the technology is maturing
rapidly, because these might become essential for tomorrow's business. It is
therefore necessary to scan research efforts of competitors and potential technology
sources intensively. Furthermore, the technologies need to be protected carefully.
Key Technologies
Generally speaking, key technologies, being critical to current competitiveness,
should be owned by the company. If one's technological capability is weak or
moderate in the technology area at issue, one should acquire extra technological
capability for building in-house R&D strength by acquisition or by introduction of a
substitute technology.
Base Technologies
For noncritical base technologies outsourcing might be the appropriate choice if
one's technological capability in the field is weak. If it is moderate, it may serve as a
means of exchange in a partnership. If it is strong, it either may serve as a means of
exchange or may be sold to focus the internal technological capabilities on key
technologies.
It is important to remember that the outsourcing and partnering matrix does not
render a static model. For instance, by acquiring extra technological capacity in a
field of key technology where one's technological capability is weak, one should
gradually shift from the left-hand to the right-hand side of the third row in the
matrix.
Partner Selection
Erens et al. (1996) conclude, based on a study of fifty companies in Europe, U.S.
and the Far East (including Airbus, Boeing, Canon, Hitachi, IBM, Matsushita
Philips, and Toyota), that many leading companies are too selfish in the search for
an appropriate partner. They emphasize that companies should not only look at what
they need and want from a potential partner but also what they can deliver to a
partner in terms of skills, market access, and economies of scale. In addition, they
emphasize that the companies are too much oriented toward the hard aspects of the
collaboration, whereas a good match of the soft aspects, including business culture
and chemistry of (top) managers are far more important for successful cooperation.
Bailey et al. (1996), based on a recent study of seventy UK-based companies in
different industry sectors, also conclude that selecting partners for collaboration on
technical merits alone is clearly a suboptimal solution. Relying on the partner's track
record in previous collaborations has turned out to be a poor basis for collaborator
selection, as well. The authors even call this a recipe for disappointment. In short, a
company has to be very careful in the selection of the potential partner(s}. An ideal
partner should:
• Have an interest in and expect equal advantages of the collaboration,
• Have complementary technological capabilities and knowledge,
• Be capable and willing to share financial risks,
• Have no record of opportunistic behavior in former collaborations, and
• Have a business culture that favors collaboration with open communication and
a quality vision. Its management should not be afraid of losing some of its
authority.
Less suitable partners are:

• Partners with comparable core business or geographical markets (security
conflicts may easily arise where partners operate as direct competitors in other
markets),
• Partners with business cultures that differ too radically (Lorange et al. (1992)
also identify compatible organizational cultures as important, encompassing
similar perception of the environment, organizational values, and operational
routines), and
• Dependent companies (this finding is in line with that of Saxenian (1990);
high-tech Silicon Valley companies receive no more than 20 percent of any
supplier's output to ensure that they do not become too dependent on external
partners).
Supporting the Outsourcing and Partnering Decisions

Choosing a technology field in order to sustain long-term competitive advantage,
choosing outsourcing or not, and choosing of partners are-in the concrete practice
of managing a firm-decisions that are closely related and difficult to disentangle.
Consequently, if managers do not sufficiently separate aspects of these three
decisions, decision making will be less than optimal.
For example, suppose that a firm wants to develop a high-tech product that
includes different technologies. Further presume that some of the technologies still
have to be developed and are unrelated to key technologies within the firm. In this
case, it would be advantageous for the firm to find a partner with the relevant
competencies. Unfortunately, the most adequate, or perhaps the only feasible,
partner for this activity is one of the firm's main competitors in other fields. This
being the case, the firm will probably refrain from considering this potential partner,
even when the collaboration would be necessary from the point of view of
developing the new product.
But if the partner search would initially be confined to identifying the relevant
attributes of a potential partner, and then assessing the relative importance of these
attributes, then the first selection would be configurations of attributes rather than
concrete firms. By later adding names to the relevant configurations of attributes, the
search would probably include the competitor. In this way, the firm can
systematically scan the social network for potential partners in the periphery of the
network, which, as we indicated earlier are in the case of R&D relationships, can
sometimes be very significant in terms of reinforcement of R&D competencies.
In our view, this implies that managers should consider the set of decision
problems as one complex decision problem that first has to be analytically unpacked
in various aspects, after which the aspects should be synthesized in a meaningful
way. A powerful way of supporting this activity is the application of multicriteria
decision analysis. In a related project (see Hummel et al. 1998) we are using Saaty's
Analytic Hierarchy Process (AHP), and the related Expert Choice software, to
develop a support system for such decisions.
The Analytic Hierarchy Process method comprises four steps: 1) decomposition
of the complex problem into dimensions, criteria, aspects, and so on; 2)
hierarchization in terms of different levels at which the components can be seen; 3)
pair-wise comparison of the dimensions, criteria, and so on, at each level, using a
nine-point quantitative scale, resulting in ranks of importance of the dimensions,
criteria, and so on for each of the levels; and 4) synthesizing the respective rankings
by combining the scores of each of the levels, resulting in eigenvectors for each of
the possibilities. We refer to Saaty (1990) for a further elaboration of the method.
Although different types of AHPs can be developed regarding decisions about
R&D management (for example, a cost benefit analysis of the R&D portfolio, see
Liberatore et al. 1992), it is our contention that R&D decisions involving other
actors should be supported by a more elaborate model. This model combines two
different approaches: a forward and a backward planning approach. The forward
planning approach should start with the assessment of the actual R&D capabilities
and existing collaborations and consider the potential objectives that can be reached
with these capabilities and collaborations (also including the forces in the
environment and the behavior of other actors). In contrast, the backward planning
approach should take as its starting point the desired future (that is, the improvement
of the competitive position of the firm), considering various potential strategies, and
indicate the necessary R&D capabilities and collaborations in order to reach this
objective (taking into account forces in the environment and the behavior of other
actors). Discrepancies between the two approaches are diminished by subsequent
iterations of both approaches. In Figure 2 the two models are exemplified.
Establishment of a Collaborative Agreement

Too extensive contractual arrangements should be avoided because they are likely to
be interpreted as distrust of the good intentions of the partner. However, we don't go
as far as Wolff (1994), who states that contractual arrangements are important only
insofar as they define the terms under which a partner may exit the collaboration.
According to Chiesa and Manzini (1996) the partners should agree on the objectives
of the collaboration (expected results), the time required for the project, the expected
contribution of each partner in terms of time and resources, the organizational
structure, the role of each partner (allocation of tasks), the coordination mechanisms,
and criteria for assessing the collaboration's performance and evaluating results. In
accordance with Chiesa and Manzini's conclusions, our respondents indicated that a
contractual arrangement should minimally codify the following:
• The financial and personal responsibilities of the partners,
• The division of the possible gains among the partners,
• The way that knowledge will be protected, including patent and trade secret
rights and confidentiality agreements,
• Criteria for measuring and monitoring progress, so that deviations can be
identified and potential problems can be overcome (this includes milestones and
deadlines of the project, responsibilities and accountability of the project team
and the founding of a steering committee; for example, a contract between the
biotechnology company Immunex and the pharmaceutical company Smith Kline
Beecham included a list of the principal scientists who would be responsible, a
detailed schedule of at least weekly telephone conferences, and a provision for
at least quarterly joint meetings) (Leonard-Barton 1995), and
• Penalty clauses to discourage opportunistic behavior.
Implementation
We are convinced that partnership management is to a great extent management of
trust and goes far beyond signing confidentiality agreements and agreeing to
guidelines. To be effective, a collaboration requires the bridging across different
business cultures and lines of responsibility in the participating companies. Lewis
(1990) states that lack of trust is the major reason why many R&D managers don't
think their alliances are working as well as they should. Based on a study of 84
alliances, Lorange et al. (1992) conclude that trust and commitment are necessary
conditions for long-term collaboration. Alliances have to be designed to create
win-win situations rather than some form of a zero-sum game, otherwise they will
certainly fail (Rai et al. 1996). Bruce et al. (1995) comment that the creation of a
climate of trust might appear to be in direct conflict with the notion of establishing
limits to the knowledge exchanged. It is the challenge for the partners to find the
critical balance of openness and confidentiality. It seems feasible that over time, as
trust is built, the need to limit the scope of the collaboration might decrease. Trust is
built up by ensuring that partners receive suitable rewards for their efforts. To show
their interest in the venture, each partner should contribute high-quality R&D staff.
During the whole cooperation it is critical that the partners keep each other informed
about what they are doing. Frequent communication in building up mutual
understanding and in checking on the progress of the collaboration saves time and
Focus
Objectives
Strategy
Environment
Technological Fields - saturated • beginning . stable

- growth · mature . changing potential
• emergent • declining - fast changing
Policy
Technologies
Type of technologies · emerging

· pacing
• key
· base
Location
Partners • horizontal
• supplier
• buyer
Figure 28. A forward planning approach on R&D collaboration decisions: asssessment starts
from the current technology portfolio and partnerships
Focus Competitive position of firm
Scenario
Strategy
Technology focus on - Emerging

- Pacing
- Key
- Base
Location
Partners
Figure 2b. A backward planning approach on R&D collaboration decisions: asssessment

starts from the desired technology portfolio and partnerships
costs by preventing far more costly adjustments later in the collaboration. To

communicate frequently, exchanging all relevant memos and team reports helps in
creating a climate of trust.
Alliances can't be run over the telephone (Wolff 1994). The R&D staff of each
partner should visit and be allowed to work temporarily in the partner's laboratory to
gain a hands-on understanding of the situation. If geographical colocation is not
feasible or too expensive, the use of advanced telecommunication systems
(videoconferencing) might be the best alternative that allows direct face-to-face
contacts. In most cases, successful cooperative relations were established gradually
and were started with small and uncomplicated projects. It is important that the
partners learn as much as possible about each other before the partnership starts.
Leonard-Barton (1995) notes that the more managers understand the classes of
problems that may occur and develop the ability to anticipate issues through
preagreement diagnosis, the greater the likelihood of success. Managers have to map
possible differences in business culture among countries and differences in
management style among companies and try to foresee and anticipate possible key
issues of disagreement. Ideally, an in-depth organizational report should be made of
the business culture of the partner in advance of the collaborative agreement. This is
often problematic because differences are more likely to emerge as the collaboration
proceeds (Bruce et al. 1995). Nooteboom's (this volume) argument about the
importance of including a trusted third party (the go-between) in an R&D

collaboration, also refers to the importance of increasing the information level about
the partner(s) and thereby enhancing trust by a party from outside the network.
Performance Assessment and Implementation of Corrective Action

The partners should strive to clear accountability through performance measures
equivalent to those used for measuring internal R&D performance. A joint steering
committee with enough authority should meet periodically to review goals and
progress against schedules. This is very important because it creates a feeling of
urgency. In the home companies of the partners there are always R&D activities that
seem more urgent than a collaborative project with a far-away partner. Stringent
deadlines help to avoid arrears and budget overruns. Bruce et al. (1995) also
underline the importance of a steering committee as the management's visible
commitment to the collaboration.
CONCLUSIONS
As companies strive for more sustainable technological competitiveness, they must
rethink a number of critical technology management processes. New strategies,
processes, and organizational roles are required to develop and utilize the wider
range of external resources and competencies needed to optimize the social capital
of the company. According to this study, partnerships to be effective, should meet
the following requirements:
• The partnership is based on a good business opportunity,
• The goals are equally beneficial to and considered important by both partners.
• The scope of the partnership is well defined,
• Each partner makes an equal contribution, based on complementary strengths,
• There are no strategic conflicts between the partners, and no side dominates the
partnership,
• The partners show their interest in the venture, and each partner contributes a
high-quality R&D staff,
• A joint steering committee meets periodically to review goals and progress
against schedules, and
• The partners anticipate and respect the inevitable differences in business
culture.
These requirements align with the key attributes for strategic alliances as
distinguished by Hampson and Kwok (l996)-trust (reliability), commitment (a
win-win attitude), (acknowledged) interdependence, cooperation (self-interest to
achieve mutual goals), communication (accurate, timely and relevant), and (open)
joint problem solving.
It should be kept in mind that the survey reported here focused on social liability
occurring in single cooperative efforts. Bruce et al. (1995) rightly point to the
potential danger of a set of rules for success on collaboration that is based on
information about single collaborations only. Cooperation is basically an
evolutionary process. It may start as a highly structured project with clearly defined
responsibilities and objectives in terms of developing a single product, and it may

gradually develop into a profitable but less structured long term alliance as trust and
mutual confidence grow. There are also many intangible and unpredictable factors
that might affect the manner in which collaborations develop. The partners may
have more reasons to enter the collaboration (for example, turning potential rivals
into allies or getting access to foreign markets), and circumstances may change. The
collaboration may have unintended side-effects and present unanticipated
opportunities, as well. As was shown before, even unsuccessful cooperations may
lead to extensive organizational learning, and hence may enhance the social capital
of the company.
As a consequence, the relationships between the various aspects mentioned
above are not self-evident. For instance, the outsourcinglpartnering matrix might
indicate that a particular competitor is potentially an adequate collaborator for a
specific R&D project, whereas previous experiences with the competitor may
hamper the firm's ability to see this possibility. Consequently, it is paramount that
the various aspects are addressed in a systematic way, by weighing the various
aspects against each other. In this chapter we showed the potential usefulness of
Saaty's Analytic Hierarchy Process to support this process.
NOTES
The interviews were conducted by 26 students of Technical Business Administration. School of

Management and Organization, University of Groningen. The Netherlands.
Technological Prestige and the
Accumulation of Alliance Capital
20
•
Toby E. Stuart
ABSTRACT
In this chapter, I argue that the prestige of an actor is a primary determinant of its
ability to access resources held by others. The reason for this is that relationships
with prestigious actors are inherently valuable because they convey status to
affiliates, and so prestigious actors have many opportunities to form new
relationships in which they exchange status for other kinds of resources. The chapter
contains an empirical analysis that shows that high prestige semiconductor firms
establish many license alliances in which they gain the rights to produce and sell the
proprietary technologies of competing organizations. If we conceive of a portfolio of
interorganizational access relationships as a component of corporate social capital,
the findings show that social capital accrues at a high rate to high status
organizations.
INTRODUCTION
Social capital has been defined in many different ways, but it is always presumed to
inhere in the relationships that bring actors together (Coleman 1988) or the capacity
of an actor to bring others together where no relationships exist (Burt 1992). When
social relationships are completely absent, social capital is nonexistent. For this
reason, one must understand why and where relationships are formed in a social
structure to understand how and how much social capital is accumulated.
Sociologists have posited numerous behavioral tendencies that limit the
incidence of contact between actors, and so affect the accumulation of social capital
in a system. These tendencies are the mechanisms that create highly circumscribed
Technological Prestige and the Accumulation of Alliance Capital - 377
patterns of relationship formation; their effect is to cluster relations within and

around particular social structural locations. The behavioral proclivity that has
garnered the greatest attention in sociology is homophily: if strong homophilous
tendencies exist within a population, then the demographic composition (broadly
defined) of that community of actors will dictate the frequency and pattern of
relationship formation within it (Homans 1950; McPherson, Popielarz and Drobnic
1992). Marsden (1983) refers to factors that confine relationship formation under the
rubric of 'restricted access,' which suggests the operation of social structural and/or
social psychological mechanisms that sharply delimit patterns of interaction. While
access restrictions create the structural conditions necessary for the brokerage role
(Burt 1992; White, Breiger, and Boorman 1976) and therefore give substance to the
weak tie metaphor (Granovetter 1973), they can also limit the aggregate amount of
social capital in a community. Broadly speaking, access restrictions imply that social
capital will accrue unevenly and at the greatest rate to the occupants of positions
that, for one reason or another, convey access to a diverse array of others.
The general phenomenon that I investigate in this chapter is the association
between actor prestige and the ability to accrue social capital by building exchange
relations with one's alters. The specific instance of this more general phenomenon
that I explore here concerns whether the level of technological prestige of a firm
influences its propensity to engage in strategic alliances that convey access to the
technological resources of alliance partners. My thesis is that high prestige creates
opportunities for access: organizations with high-status generally possess many
opportunities to build new intercorporate coalitions. Not only are they able to pick
and choose among potential associates, high prestige firms are able to negotiate
access to the key resources of their competitors on appealing terms. In its essence,
the argument that I propose is that high-prestige firms exchange a social resource-
their imprimatur-for a technological resource-access to their competitors'
innovation. Hence, this chapter investigates the effect of an organization's level of
prestige on its accumulation of material and alliance capital through a portfolio of
strategic alliances.
NETWORK-BASED PERSPECTIVES ON ACCESS RESTRICTIONS IN

STRA TEGIC ALLIANCE NETWORKS
In organization theory and sociology, a number of studies have examined the
structural foundations of alliance formations. Perhaps the most widely influential
work in this area has been the group of 'embeddedness' studies (Granovetter 1985).
Scholars working from this view have argued that the pattern of new alliance
formations depends upon from the configuration of previously-established
coalitions. The reason that existing (or past) alliance ties are important for the
pattern and volume of formation of new coalitions is that cheap information about
the quality and the trustworthiness of potential alliance partners diffuses across
existing inter-organizational relations. Information on the attributes of possible
alliance partners is available from previous, direct interactions with them (Gulati
1995a; Podolny 1994; Powell, Koput, and Smith-Doerr 1996; Granovetter 1985;
Walker, Kogut, and Shan 1997) and from the experiences conveyed by other parties
who have had relations with them (Burt and Knez 1995; Nooteboom, this volume).
However, because firms differ in terms of how they are positioned in the
interorganizational alliance network, they also differ in the extent to which they are
able to gather information about potential alliance partners.
Possessing knowledge of the characteristics of potential exchange partners is
important because it influences the transactions costs associated with forming new
relationships. There are two reasons for this. First, intimate knowledge of a potential
partner eliminates the cost of investigating the quality of that organization, and
thereby removes one of the more significant expenses of alliance formation. Second,
when a potential partner is known to be reliable and trustworthy, a focal firm may be
willing to enter into a business relationship with that partner without contractually
specifying a priori all of the terms of the association (see the discussion in Knoke,
this volume). Conforming to Marsden's (1983) notion of restricted access,
embeddedness perspectives on inter-firm strategic alliances have argued that
existing alliance ties structure the search for new associates, directing attention
toward previous associates and thereby influencing the rate of accrual of social
capital.
From this very brief review, it is apparent that much of the sociologically-
informed literature on interorganizational relationships explains alliance formations
according to the influence of previously-established intercorporate relations on the
search for and selection of new partners. In contrast to the existing sociological
literature, I will use this chapter to investigate how the technological structure of the
industry in which I locate the empirical analysis influences the formation of new
(technology trading) alliance ties. In particular, I will investigate the prediction that
prestige differences between firms have a systematic effect on their proclivities to
enter into new intercorporate coalitions. The argument that I make is that prestige
creates invitations for association because of the benefits that others derive from
relationships with well-known actors. Therefore, high-prestige raises the ability of
organizations to develop 'alliance capital'-a collection of relationships with other
firms that can be a valuable source of information, know-how, intellectual property,
and new customers (see Smith-Doerr et aI., this volume).
Because the empirical analysis will focus on strategic alliances between high-
technology firms, I will use a measure of corporate technological prestige. In
general, organizations acquire prestige when they make significant contributions to
the communities to which they belong. In the context of a high-technology market,
firms attain prestige by forging the paths for new technologies and opening up
possibilities for follow-on inventions (Podolny and Stuart 1995). Hence, developing
inventions that other innovators recognize as important technological achievements
is the underpinning of high technological prestige. Just as the members of any
community of actors can be ranked by relative prestige, every industrial community
can be described as an ordering of status positions (Podolny 1993). Consider any
high-technology arena at a particular instant in time, and the names of its most
prestigious members will come quickly to mind.
In this chapter, I specifically focus on the relationship between corporate
technological prestige and the incidence at which firms establish strategic license
alliances with competing organizations. License alliances are inter-firm agreements
in which one organization gains the rights to produce or sell the proprietary
technology of a second firm, usually in exchange for a royalty on units sold or a

lump sum payment. Therefore, firms that serve as the licensee in these asymmetric
alliances do so to gain access to the proprietary innovations of a competitor. I My
thesis is that organizations with high prestige will be the most frequent users of
license alliances to acquire the rights to utilize technologies developed by their
competitors. A number of factors produce the association between actor prestige and
license alliance opportunities, but for the objectives of the present analysis, one
determinant merits particular emphasis.
High prestige firms encounter many opportunities to enter alliances because
status is implicitly transferred across inter-corporate ties (Podolny 1994), which
creates benefits for the affiliates of high-status organizations (Stuart 1998a).
Moreover, the advantages of affiliations with high-status firms are significant
enough that other organizations are likely to concede generous contractual terms in
order to secure the organization. This means that high-status firms are able to
negotiate favorable contract terms when they establish alliances.
Relationships with high-prestige firms create value because external resource
holders such as potential customers pay close attention to the strategic initiatives of
prestigious firms. Because of the presumption that prestigious firms will eschew
collaborations with those of mundane quality, prestige is transferred to affiliates
when associations with prestigious actors are publicized. Therefore, associations
with prestigious actors are status-enhancing: they can substantially improve the
reputation of the 'connected' actor. An enhanced ability to attract resources is likely
to follow the establishment of an exchange relationship with a prestigious affiliate,
and this implies that prestigious associations are a form of alliance capital that can
create a meaningful competitive advantage (Podolny 1994; Baum and Oliver 1991;
Stuart, Hoang, and Hybels 1999).
In high-technology industries, the transfer of status from high-prestige actors to
other organizations or to new initiatives is an every-day phenomenon. One very
notable instance of this dynamic occurred when IBM, a high-status manufacturer of
mainframe computers, chose to enter the fledgling personal computer industry in the
early 1980s (Anderson 1995). IBM's entry into the computer business is widely
thought to have brought legitimacy to the industry as a whole. A similar dynamic
occurs when a prestigious enterprise backs the initiatives of an unknown firm by
establishing an alliance with it, such as when three well-regarded pharmaceutical
companies (Pfizer, Glaxo Wellcome, and Smith-Kline Beecham) recently formed
technology alliances with Cantab, a young biotech company on the brink of
financial distress. Because there is a presumption that prestigious organizations
perform a thorough evaluation (due diligence) before entering into a cooperative
venture, a technology alliance with a prestigious partner is precisely the type of tie
that raises public regard for the initiatives of the lesser-known firm. Particularly
when a prestigious firm licenses technology from a different organization, attention
is directed to the endeavors of the affiliate and greater interest may be expressed in
the exchanged technologies. A license alliance conveys an implicit endorsement,
and when the licensee is a highly-regarded enterprise, it is a particularly valuable
certification.
High prestige firms are willing to purvey endorsements for technology access
for two reasons. First, as previously noted, prestigious firms possess the ability to
influence the terms of trade so that they can strike favorable alliance deals. Because
they are distinguished players, other actors will realize an advantage from
associating their initiatives with prestigious firms. Given that they can significantly
raise the value of their associate's endeavors, prestigious actors will enjoy an
advantage in negotiating the terms of an alliance contract.
Second, it is inexpensive to govern alliances when prestigious organizations are
participants in them, particularly when their affiliates possess lower status. The
reason for this is that high-prestige firms-central and influential actors in the
communities to which they belong-have a unique ability to enforce contract terms
and to block attempts on the part of their strategic partners to act against their
interests. From the vantage point of high-prestige firms, the transaction costs
associated with alliance formation are low because a positive reference from them
will be of great future benefit to their alliance partners, while negative referrals from
them will be particularly damaging to the reputations of young, small, and low status
organizations. The importance of staying on the good side of a prestigious firm
creates a strong disincentive on the part of their partners to defect against the terms
of an alliance contract or to otherwise behave in a manner that is disadvantageous to
the interests of a high prestige partner.
For the reasons just outlined, my central contention is that prestigious
organizations, by virtue of their standing in an industry's status order, are able to
choose among potential alliance partners and probably at favorable terms. Therefore,
I expect that technologically prestigious firms will have a relatively high number of
options to establish strategic coalitions on terms that appeal to them, and so are
likely to form many alliances. In particular, I investigate the extent to which its level
of prestige affects the likelihood that a firm will establish strategic license alliances
in which it receives (in-licenses) technology from a collaborator. I test this
prediction in an analysis of the effects of technological prestige on the rate of
formation of licensing alliances in the worldwide semiconductor industry.
THE SEMICONDUCTOR INDUSTRY

The sample that I analyze consists of all semiconductors firms for which I was able
to collect annual semiconductor sales during the analysis window (1986-1992)?
Because many semiconductor producers are broadly diversified (e.g., IBM, Siemens,
Fujitsu), it is typically not possible to collect semiconductor revenues for many of
the firms in the industry from securities filings. Therefore, I acquired sales data from
Dataquest, a high-technology consulting firm. Dataquest tracks all merchant
semiconductor firms with annual sales in excess of $10 million; given their data
collection procedure, the sample consists of the firms in the industry with the largest
sales volume. The Dataquest database consists of 150 companies, although some are
not present in the data during all of the years. The firms in the sample are
headquartered in the US, Europe, Japan, and other Southeast Asian nations. As a
group, the firms in the sample accounted for over 90 percent of the total, worldwide
semiconductor production volume in 1991.
Technological Prestige and the Accumulation of AlIiance Capital - 381
ESTIMATION
I treat the organization as the unit of analysis in the alliance formation process and
estimate the rate at which a focal organization forms new relations. 3 Assuming a
continuous-time event history, the hazard of alIiance formation, denoted rit), is
expressed:
(1)
where Pk(t,t+At) represents the probability that a firm will experience an alliance of
type k during the interval from t to t+L1t (Tuma and Hannan 1984). In this chapter, I
will only examine the determinants of one-way (i.e., directional) technology license
agreements. These are agreements in which technology is transferred from one firm
to another. Therefore, the data are asymmetric ties and it is possible to designate
which partner sends and which one receives technology. 'Receive' alliances are
considered to be those in which a focal firm is the licensee or recipient of
technology from a competitor. 'Supply' alliances are those in which a focal firm is
the licenser or source of technology. which is transferred to a competitor. Hence, I
specify the state space to allow for the occurrence of two event types.
Estimating equation 1 requires an assumption regarding the form of duration
dependence in the alIiance formation process. I have selected a piecewise
exponential model to estimate the hazard because it is extremely flexible with
respect to the form of duration dependence. The model assumes that the baseline
transition rate, given by the coefficient estimates on a vector of m time periods, is
constant within each time period. However, the baseline rate may change across
time periods. The piecewise exponential can be written as:
rk (t) =exp(am + f3Xt) (2)
where <Xm represents a vector of m constant coefficients designating the baseline

hazard (one coefficient is estimated for each time period), XI is an nxq matrix of q
covariates, some of which are time-varying and updated annualIy, and ~ is a vector
of q time-constant coefficients to be estimated. Note also that the formation of a
technology alliance is a repeatable event. Consequently, firms remain in the risk set
for the formation of an alliance even after they have established previous alliances.
Repeatable events are unproblematic in event history analysis: in practice, the
episodes between alliance events are separate observations, all spells are split at the
end of each calendar year to update the covariates, and all observations are then
pooled across units and time (Tuma and Hannan 1984).
Because I distinguish between technology supply and receive alliances in the
analysis, I can contrast the effect of prestige on the rate at which a focal firm is a
licensee (recipient) versus licenser (source) of technology. Hence, k indexes two
types of events in (1) and (2) above.
VARIABLES
Dependent Variable
The database for the analysis includes all publicly-reported technology alliances
involving the firms in the semiconductor industry between 1981 to 1992. An event
occurs on any month during which a firm in the sample announces the formation of
a license alliance with a different semiconductor firm (the partner firm can be inside
or outside of the sample of 150 that I analyze). The sources for the alliance data
were the electronics trade press, the business press, company press releases, annual
reports and 10Ks, and periodical indices.
Technological Prestige
Following Podolny and Stuart (1995), I define a firm's technological prestige as its
indegree score in a network of patent citations consisting of all semiconductor
inventions patented in the US.
Patents grant inventors exclusive property rights for the commercial use of their
inventions for a fixed period of time (see Smith-Doerr et aI., this volume, for an
extensive disucssion on patents). Patents are issued by the government in exchange
for the inventor's consent to publish a detailed description of an invention. To
receive a patent, an inventor must file an application that fully discloses a non-
obvious and industrially useful invention. The Patent Office publishes descriptions
of all patented inventions in successful (granted) patent applications.
Patent applications must contain a list of citations to the existing patents that had
made technological claims similar to those claimed in the application. Hence,
citations are issued to the patented inventions that are nearest in technical content to
the proposed invention. When the Patent Office receives a patent application, it is
assigned to a patent examiner who is knowledgeable in the pertinent area of
technology. The examiner then searches through a database of existing patents to
verify that the list of references in the patent application, known as the 'prior art,' is
complete. If the patent application is ultimately approved, it is published with the list
of prior art.
Citing a preexisting patent does not protect a patent holder from a legal action
initiated by other inventors: any party may contest the validity of a patent, even if
the action is filed by the holder of a patent that is cited in the disputed claim. For this
reason, the patentee accomplishes nothing by including superfluous citations in its
application. On the other hand, failure to list known citations can delay the issue
date (the date when the application is approved). Therefore, it is generally assumed
that the applicant possesses the incentive to cite relevant prior art, and the patent
examiner's prior art search safeguards the integrity of the process (Office of
Technology Assessment 1976).
Recently, social scientists have taken advantage of patent citation data because
they manifest technological similarities between inventions. Researchers have used
citations to identify pairs or groups of firms have been investing in similar
technologies. For example, Podolny, Stuart, and Hannan (1996) used patent
cocitations (when two patents cite a common, third patent) to measure the
technological overlap of pairs of organizations. In a similar vain, economists have
traced patent citation patterns to document technological knowledge spillovers

between firms (Jaffe, Trajtenberg, and Henderson 1993).
The US semiconductor patents assigned to the 150 firms in the sample are used
to compute technological prestige scores. Semiconductor firms routinely fIle for
patent protection for their inventions in the US. The importance of patenting in the
industry is underscored by the fact that the six firms awarded the greatest number of
US patents during calendar 1996-IBM, Motorola, NEC, Hitachi, Canon, and
Mitsubishi-all have sizable operations in semiconductors.
The US is the world's largest technology marketplace. For this reason, it is
routine for non-US-based organizations to patent in this country (see Albert et al.
1991; also, note that four Japanese companies were among the six largest holders of
US patent with a 1996 issue date). Thus, I collected all US semiconductor patents in
2400 distinct patent classes that contained semiconductor product, device, and
design inventions. After assembling all US semiconductor patents, I then
constructed detailed corporate family ownership trees for the 150 firms in the
sample. Using these ownership trees, I was able to assign the patents of subsidiaries
and divisions to the level of the corporate parent. Finally, after identifying the
corporations that owned the 48000 patents in the database, I configured the data into
a network to compute a measure of prestige for the statistical model.
To operationalize technological prestige, I follow the definition of Knoke and
Burt (1983), who stated that an actor is prominent to the degree that its network
position makes it visible to other actors, and it is prestigious when it is the object of
relationships from other actors in a network of directed ties. Among the many
measures of prestige developed in the network literature, the simplest is an actor's
indegree score, which is a simple count of the number of relations directed to a focal
actor.
Adopting this measure to a directed network of patent citations, a prestigious
innovator would be a firm with a patent portfolio that is highly cited by other
innovators (Podolny and Stuart 1995). Because patent citations are akin to building
relationships between inventions, highly-cited patents are those which have been
important building blocks for the efforts of other innovators (see above). In this
sense, patent citations are similar in meaning to citations between journal articles:
they are deference relations because they implicitly acknowledge the importance of
a contribution of a competitor (Podol ny, Stuart, and Hannan 1996). Therefore, just
as highly-cited papers are the origin for the prestige of a scientist, I argue that
highly-cited patents engender prestige for their corporate developers. Accordingly,
the technological prestige of semiconductor firm i is defined to be the proportion of
all patent citations directed to its patents. Specifically:
LCjit'
D. =_1_'__ i:1! j (3)
,t 4,
where Dil denotes the prestige of firm i at time t, Cjil , is coded as '1' if a patent of
firm j cites a patent of firm i during the interval t' (and '0' when there is no
citation), and 4' is a count of all patent citations accruing to the firms in the sample
during the interval t'. Thus, Djt is the proportion of patent citations made during the
interval t' which are directed to the patents in the portfolio of each firm i. The
restriction i :F j is included so that self-citations do not contribute to a fIrm's prestige.
I have chosen to normalize the indegree score with the total number of patent
citations made in the industry to correct for changes in the total volume of citations
accruing to the sampled firms over time. Because of the normalization, the prestige
measure maintains a consistent meaning across time periods (it is the proportion of
all patent citations that accrue to the portfolio of each firm in the sample within a
fixed period of time).
Prestige levels for the firms in the sample can be expected to change over time
as firms develop new inventions and as the characteristics of old ones lose their
relevance for a firm's current-day initiatives. Therefore, I have used a five-year,
moving window to compute technological prestige (t' in equation 3 designates the
five year interval that precedes year t). Five years was chosen because this is the
approximate length of the product life cycle in the semiconductor industry.
Following equation 3, Texas Instrument's prestige in the year 1991 is the number of
patent citations that were received by all of its inventions during the 1986-1990
interval, divided by the total number of patent citations received by all firms in the
sample during the same time interval.
Control Variables
A number of control variables are included in the analysis. First, firm size, measured
as annual semiconductor sales, is included in the models. This variable is added
because there is likely to be a positive correlation between size and prestige. Hence,
without size in the regressions, any effect of prestige may be spurious because of the
correlation between prestige and this omitted regressor. Second, the models include
the age of firms' semiconductor operations, defined as the number of years since
initial entry into semiconductors for diversified semiconductor producers and to be
the time since founding for dedicated producers. Third, the models include a dummy
variable coded as '1' if a firm is publicly traded.
I have also constructed a 'no patent' dummy variable, coded as '1' if a firm had
no semiconductor patents before the start of a year. Based upon equation 3, firms
without patents have '0' technological prestige because they are not at risk of
garnering patent citations. I have included these firms in the analysis with zero
prestige, but allow for an intercept adjustment for them (reflected in the coefficient
on the 'no patent' dummy). It should be noted that results are comparable when firms
without patents are omitted from the sample.
The final control variable in the hazard rate models is an endogenous occurrence
dependence term defined as the total number of alliances formed by each firm in the
sample during the previous five years. It is understood that unobserved
heterogeneity across observations produces occurrence dependence in event data
(Barron 1992). One strategy to control for unobserved heterogeneity is therefore to
include a variable that indicates the number of times that each actor has previously
experienced the event being modeled. Including the previous alliance count as a
variable in the event history analysis should therefore help to control for the effects
of unobserved factors that produce variance across firms in their proclivity to form
alliances.
The alliance data I possess are from 1981 to 1992, but because of the lagged
alliance count variable, I have modeled alliance formations during the period from
1986 to 1991, while using the first five years of the data to construct the occurrence
dependence variable. All firms in existence in 1986 are considered to be at risk of a
new alliance formation on January 1 of that year. For all firms, spells have been split
and covariates are updated on the first day of calendar 1987, as they are in each
subsequent year. All time-varying covariates have been entered into the models as
one year lags. The waiting time clock for each firm's initial spell has been set to be
the amount of time which has passed since the firm formed its last alliance (prior to
1986; this is defined as the time since founding for firms that had no pre-sample
alliances). Guo (1993) has demonstrated that this approach yields unbiased estimates
when using the exponential model. Therefore, because of the distribution, the results
are untainted by left censoring, even though some firms had formed alliances prior
to the time at which I begin to observe them.
RESULTS
Table 1 reports the results from the hazard rate alliance formation models. As
discussed in the 'Estimation' section, technology supply and receive license alliances
are treated as 'competing risks' in the analysis. Therefore, Table 1 reports a separate
set of coefficients for the effects of the variable on each of the two outcomes.
Beginning with the control variables, it is no surprise to find that firms without
patents ('no patent' indicator ='1 ') are far less likely to enter alliances, particularly
technology supply agreements. As one would expect, the count of previous alliance
formations (the occurrence dependence variable) has a positive and statistically
significant4 effect on both events. Among the remaining control variables, firm size
has a statistically significant effect only on the rate of entry into technology supply
alliances, while the age of the firm in semiconductors has a statistically significant
effect only on the rate of entry into technology receive alliances. It is interesting to
note that older firms, perhaps more inert and lagging behind their younger
counterparts in the adoption of the latest technological developments, have a greater
proclivity to utilize the social capital of license alliances as a strategy for sourcing
technology.
Turning now to the effect of technological prestige on the alliance formation
rate, the results demonstrate that the variable has a large and statistically significant,
positive effect on the rate at which a firm receives technology from its competitors
(the magnitude of the effect through the 95 th percentile of the distribution of the
variable is plotted in Figure 1). In contrast, while the effect of prestige on the rate of
supplying technology is also positive, the magnitude of the coefficient is small and it
is not statistically different from zero. This pattern of results is consistent with the
hypothesized status transference process. By the signing of a license contract, a
prestigious innovator publicly certifies another organization's technology. If (as I
have hypothesized) firms attempt to affiliate their endeavors with prestigious firms
Table 1. Maximum likelihood estimates of the technology supply and

receive alliance formation ratea
Receive Supply
Firm is public .4184* (.2291) .2528 (.2125)
Firm has no patents -.5828* (.2220) -.7828* (.2268)
Lagged alliance count .0237* (.0041) .0214* (.0038)
Firm age -.0122* (.0062) .0160 (.0057)
Firm sales ($8) .0039 (.0065) .0117* (.0056)
Technological prestige (DiI) .0237* (.0041) .0038 (.0033)
aAIl models include II unreported period effects and unreported calendar year
effects. Receive alliances are those in which a fmn is a technology licensee and
supply alliances are those in which a firm is the source of licensed technology.
*p<.05.
Log-Likelihood 905.84
Episodes 3127
Number supply alliances 338
Number receive alliances 409
2,5
~
.s
~
1,5
'C)
~C\,
':::1
"3
::E
0,5
0
0 0
0 N N .., .., ..,0 .,., .,.,
~ ~ ~ ~ ~ ~ ~ ~ 0~
~ 0 ~
0 0 0 ~ ~
0 0 0 0
Technological Prestige
Figure 1. Multiplier of the receive alliance rate

Legend: The Figure portrays the technology receive alliance rate multiplier plotted against the
level of focal fmn technological prestige. The curve is based on the parameter estimate for the
prestige in the first column of Table I.
to attract attention to their undertakings, we would expect to find that prestigious

firms are frequently on the receiving end of directed alliances.5 As the 'Implications'
section elaborates, this finding may reflect the changing roles of up-start and
prestigious innovators in high-technology marketplaces.
The statistically significant effect of prestige on receive alliances coupled with
the weak effect of the variable on supply alliances demonstrates the empirical
leverage that can be gained by distinguishing between alliance types in a competing
risks formulation of the alliance formation process. High prestige firms are the
industry'S leading innovators. As I have operationalized the variable, firms gamer
prestige by making important additions to the corpus of semiconductor technology.
Because high prestige firms control many of the industry' s important innovations, it
might have been expected that they would frequently serve as the source of
technology in asymmetric license alliances. Indeed, competence-based explanations
of the alliance formation process have emphasized that the most accomplished
innovators are likely to enter into alliances because they possess capabilities that
other firms will attempt to access through intercorporate alliances (Eisenhardt and
Schoonhoven 1996). As the results in Table 1 demonstrate, however, high prestige
only augments the baseline rate of licensing technology from other organizations.
This finding is in accord with the predictions of a status-based model of alliance
formations, but it is inconsistent with a capabilities-based explanation of the
phenomenon.
Before concluding, it is useful to present some descriptive data on the degree to
which prestige asymmetries between alliance partners appear in the patterns of
affiliation among semiconductor producers during the decade from the early 1980s
through the early 1990s. These data bear upon the arguments in this chapter that
governance considerations and the ability to exchange technology access for
endorsements often create the incentive for firms at different prestige levels to form
strategic coalitions. Because this suggestion stands in contrast to the prevailing
thinking on the effects of status differences on the probability of inter-actor
associations, I present some descriptive data to document the extent to which there
have been alliances in the industry involving firms at different prestige levels. 6
To explore the extent of status-based homophily in alliances between
semiconductor firms, I have sorted the 150 firms in the sample into three prestige
categories: low, medium, and high. To construct these categories, I simply rank-
ordered the firms in the sample by their technological prestige scores and then
trifurcated the sample into three groups of equal size (using cut-points at the 33rd and
67 th percentile of the prestige distribution to demarcate group boundaries). Table 2
presents the distribution of alliances within- and between these three status levels.
The data for the Table add joint product development, technology exchange
alliances, and joint ventures to the license alliances that were analyzed in Table 1.
Each cell in Table 2 reports the percentage of all innovation-oriented alliances
established between firms at the prestige levels denoted on the rows and on the
columns of the Table. Relationships among firms of comparable status (within-
prestige level ties) appear on the main diagonal of the Table because the cells on the
diagonal denote the proportion of all affiliations between firms in particular status
Table 2. Proportion of alliances between finns in different prestige categories a

Prestige Level Low Medium High
Low 0.02 0.07 0.27
Medium 0.07 0.02 0.21
High 0.27 0.21 0.41
a The Table reports affiliation patterns based upon 1000 strategic alliances between
semiconductor firms during the period from 1987 to 1992.
categories. Alliances between organizations at different levels of prestige fall in the

off-diagonal cells.
A number of patterns appear in Table 2. First, at least one high prestige firm was
involved in the vast majority of the strategic coalitions in the industry: only 11
percent of the alliances were between two low, one low and one medium, or two
medium prestige firms. In other words: 89% of the alliances included at least one
high-prestige firm. Second, a substantial proportion (27 percent) of the alliances
were established between firms with significantly different levels of prestige (one
low and one high prestige firm). Third, less than half of the alliances occur between
firms in the same prestige category. Thus, while high prestige firms were the
dominant players in the industry's alliance network-a fact that alone would seem to
implicate technological prestige as a consideration in the selection of alliance
partners-there were a significant number of coalitions between firms at different
prestige levels. While the Table 2 proportions are only suggestive, it does appear
that alliances were neither randomly spread across prestige levels nor did they occur
exclusively between firms of comparable prestige.
IMPLICATIONS
The results have demonstrated that technologically prestigious firms accrue alliance
capital at the greatest rate: they frequently utilize license alliances to access the
technological assets of their competitors. While I have only looked at the effect of
technological prestige on the accrual of alliance capital in a population of high-
technology organizations, it is likely that the findings are representative of a general
association between status and access. I anticipate that the prestigious members of
most populations will be able to gain access to the endeavors of other members of
their community, due to the benefits that derive from associating with high status
actors.
One of the traditional advantages of prestige in technology-intensive industries
has been that those who have possessed it-firms such as AT&T and IBM-were
able to attract top-quality engineers and scientists. Recent developments, however,
may have threatened this benefit and promise to alter corporate roles in high-
technology industries. The strong market for initial public offerings and the opening
up of new financing options for startup companies are among the developments that
have brought sometimes extreme wealth to managerial and scientific talent who
were early backers of successful new ventures. With sufficient private financing,
young companies can offer salaries near to those paid by established firms. Of
course, they can also offer equity, stock options and their concomitant: an outside
chance to make a small fortune. Because of these changes, prestigious companies

now face greater competition for top-quality human resources.
I speculate that the profusion of startup firms in high-technology industries will
mean that the social capital benefits of prestige, rather than dissipating, will appear
in the unique capability of the prestigious enterprise to certify the initiatives of
young or unknown organizations (see Stuart 1998a; Stuart, Hoang, and Hybels
1999). It is because of their ability to certify the initiatives of other organizations
that high prestige firm will gain access to the endeavors of others. The
correspondence between prestige and access implies that prestigious firms enjoy a
powerful positional advantage.
NOTES
I. I refer to license alliances as asymmetric deals because technology travels in only one direction:
from the licensee to the licenser.
2. Elsewhere (Stuart 1998a, 1998b) I discuss the reasons why the semiconductor industry is an
appropriate context to study alliances and I report on the incidence of alliance activity in the business.
3. I have chosen to model the rate of alliance formation, rather than to model the probability that pairs
of organizations form an alliance in some arbitrary time window (i.e., a panel dyad model; see Leenders
1995b, 1999). Treating the organization as the unit of analysis avoids issues of network autocorrelation (a
problem of dyad models given currently available estimation procedures), but at the same time does not
allow one to test hypotheses about which other organizations serve as a focal firm's alliance partners.
Although I have estimated dyad models and find a prestige effect somewhat analogous to the results that I
will report in this chapter (Stuart 1998a), the prediction is really about the effect of firm-level status on
the rate of firm-level alliance formations, independent of the identity of alliance partners.
4. Statements on statistical significance refer to a .05 level.
5. Given the modeling strategy, it is possible for prestige to have a positive effect on both types of
alliances, a negative effect on both types of alliances, or any combination thereof.
6. Podolny (1994), generalizing the argument that inter-individual as well as inter-corporate exchange
relationships tend to be among demographically similar actors, asserts that high status organizations will
exercise exclusivity in their selection of exchange partners and transact only with alters of similar status,
particularly in uncertain situations.
Networks and Knowledge Production:
Collaboration and Patenting
21
in Biotechnology
•
Laurel Smith-Doerr
Jason Owen-Smith
Kenneth W. Koput
Walter W. Powell
ABSTRACT
We examine the link between social capital and intellectual output in the context of
the biotechnology industry, a knowledge-intensive and expanding field. Due to a
range of economic, social, and scientific factors, interorganizational collaboration is
commonplace in the industry. Staying on top of fast-breaking research developments
is critical, thus patents are an important indicator of intellectual output and property.
But patents also provide a signal to organizational partners that a dedicated
biotechnology firm (DBF) is a worthy collaborator, in effect providing a basis for
building collaborative capital. We argue that collaborative capital will lead to
patents and that this intellectual capital leads to subsequent collaboration, finding
support for these claims. More centrally connected DBFs subsequently produce
more patents; and DBFs with more patents subsequently become more centrally
connected. Additionally, we look at the types of patents held, focusing on the effects
of principle (the most intellectually broad) patents. DBFs with more principle
patents become more centrally connected, and come to have more total patents.
Firms rich in collaborative capital have greater opportunities for producing ideas
with potential for timely payoff. We suggest that patents are a generative form of
intellectual capital which provide scientific visibility that further enhances a firm's
collaborative social capital.
INTRODUCTION
We examine the link between social capital and intellectual output in the context of
the biotechnology industry, a knowledge-intensive and expanding field in which
interorganizational collaborations are among a firm's most important activities. As
Collaboration and Patenting in Biotechnology - 391
an illustration, in public records such as annual reports and websites, biotechnology

firms commonly list their most prominent collaborations with external
organizations. For example, Biogen, a leading dedicated biotechnology firm, lists
nine drugs at various stages in its research pipeline on its website
(http://www.biogen.comI997). Four of these major projects are under joint
development with other biotechnology firms. Research and development (R&D)
collaborations are but one form of alliance, albeit a critical one, in an industry where
firms manage interorganizational relationships for a wide range of organizational
tasks, from financing to clinical trials to marketing and distribution (Powell, Koput,
and Smith-Doerr 1996). Biogen, our exemplar above, has more than 50 formal ties,
as of 1997 (Bioscan 1997). Of course, counts of formal collaborations exclude the
many informal arrangements that are even more commonplace in this field. Thus,
because the biotechnology industry has a strong basic science core and numerous
collaborative arrangements, it provides an ideal setting in which to study the
relationship between social and intellectual capital. We examine the effects of
collaborative social capital on the intellectual capital in patenting, then the
subsequent effects of patenting on social capital among biotech firms.
In the biotech industry, we have found evidence that network centrality is
closely related to productivity (Powell, Koput, and Smith-Doerr 1996). The most
centrally-connected firms have the highest market values and largest R&D budgets,
have been the first to bring new drugs to market, and are the most visible in the
scientific community. While network centrality brings many organizational benefits,
here we are concerned with capturing the intellectual returns. In high-tech industries,
patents are often used to indicate an organization's intellectual prowess. Our focus,
then, is on the intellectual returns of social structure. That is, how does network
centrality influence the assignment of patents? We begin with a discussion of
research on collaboration and patenting, briefly reviewing the relationship between
social capital and network position. Then, we characterize patenting as both an
intellectual and organizational activity, leading us to formal hypotheses about the
connection between centrality and patents. We then describe our data and methods
and present results. In the final section, we discuss the implications for how
organizations learn to develop and use social capital.
COLLABORATION AND SOCIAL CAPITAL

Networks and social capital are closely intertwined concepts. I Social capital has
been characterized by Burt (1992, 1997) as the ability to exploit vacant positions in
social networks, termed 'structural holes,' or potential relationships between non-
redundant contacts. Those rich in social capital are commonly found at the center of
social networks. Thus, network centrality provides an observable correlate of social
capital.
In studies of inter-organizational relations, social capital has at least three
variants? First, an individual may be a member of more than one organization. To
the extent that these memberships link otherwise unconnected organizations, social
capital would accrue to the individual (and the organizations) as a result. Interlocks
among corporate boards of directors are a primary example of this form of
interorganizational social capital. Second, individual members of organizations can
have social ties to members of other organizations, arising through friendship or

common membership in trade associations or social clubs. These ties often facilitate
goal attainment for these individuals. Third, organizations can have formal,
contractual agreements to engage in collaborative activity.
The third variant of social capital, arising from formal organization-to-
organization ties, is our concern here. In developing arguments about social capital
associated with formal interorganizational relationships, we stress that organizations
are not typically born with social capital on the basis of their parentage or bloodline
(aside from that due to individual founders). Instead, firms must learn to manage
relationships, build reputations, and ally with high-status partners by developing
routines for collaborating with other organizations (see Omta and Van Rossum, this
volume). For this reason, we refer to this variant of social capital as an
organization's collaborative capital.
The collaborative capital reflected in a firm's central location in a network of
organizations goes beyond the straight exchange advantage of having many partners
(who in tum also have many partners). Centrality not only affords access to existing
partners, but creates visibility and the opportunity to benefit from others'
relationships. Firms steeped in collaborative capital are able to combine partners and
resources in novel ways, fostering inclusivity and widening avenues of competition
(Koput, Smith-Doerr, and Powell 1997). In industries where technological
innovation is paramount, such as biotechnology, science-based start-up firms tum to
collaborations to assemble a broad array of skills and resources needed to sustain
their R&D efforts. In so doing, biotechnology firms gain experience at managing
interorganizational relations that can be used to support a diverse portfolio of
collaborative activities. Firms that utilize their initial R&D alliances as an admission
ticket to the field's network, drawing on their experience to exploit diverse
opportunities for collaboration, subsequently become more centrally-connected
(Powell, Koput, and Smith-Doerr 1996), an effect magnified by the eliteness of a
firm's partners (Koput, Powell, and Smith-Doerr 1998). Once centrally-placed in an
interorganizational network, a firm is positioned to garner knowledge spillovers,
generating further collaborative research opportunities. Hence, the more a firm
learns how to position itself centrally in industry networks, increasing its
collaborative capacity, the better positioned it is to generate patents.
PATENTS AS AN ADMISSION TICKET TO INTER-

ORGANIZATIONAL COLLABORATION
In this section we describe what patents are, explain why centrally-connected firms
in the biotech industry are more likely to patent, and argue that patents provide
intellectual capital leading to further interorganizational collaboration.
What is a Patent?
In the United States, patent protection for inventions is guaranteed by the
Constitution. Article I, section 8, clause 8 states:
Congress shall have the power to promote the progress of science and the useful arts by
securing, for a limited time, to authors and inventors, the exclusive right to their
respective writings or discoveries.
The patent system in the u.s. is based on a policy that promotes the growth of
technological knowledge and its public use by rewarding inventors for disclosing
discoveries. The reward takes the form of exclusive rights to a discovery for 17
years. Patent rights were designed to insure that useful innovations would be
transferred to the public domain despite considerable economic incentives to
maintain secrecy.
To be patented an innovation must pass through a rigorous examination process,
usually lasting at least 18 months. Patentable discoveries must pass three critical
tests: they must be useful, novel, and non-obvious to 'a person of ordinary skill in
the art' (Gregory et al. 1994). In the case of biotechnology, a person of ordinary skill
would hold a Ph.D. in molecular biology or a related field and be practicing in that
field at the time the innovation was developed. Such a criterion suggests that an
innovation cannot be patented, even if it was not mentioned anywhere in prior art, if
an ordinary practitioner could have arrived at the innovation by simply combining
existing ideas at the time it was developed (Gregory et aI. 1994: 28). In order to
meet the novelty criterion, an innovation must not have been 'anticipated' by public
knowledge or use, description in print, patents outside the United States, or
description in another U.S. patent application. The United States differs from other
nations in following a 'first to invent' patent policy (Kinston 1992), which means
that an innovation may not have been anticipated or invented by another individual
prior to the patent applicant's discovery.
Drafting the patent application requires both intimate knowledge of the idea and
of the prior art. The claims section of the specification sets forth the scope of
protections claimed in the patent, defining the rights of the inventor. In essence, a
claim is a one sentence definition of the invention (or specific aspects of the
invention). 'An unauthorized product infringes a patent if it includes the equivalent
of each and every element of any of the claims' (Schecter 1995: 71). The claims
section and no other part of the specification is consulted to determine whether a
patent has been infringed. Patent examiners may reject any or all of the claims in an
application and the scope of protection afforded an innovation is defined solely by
the claims that make it into the final patent. Writing claims is a strategic process by
which an inventor or his representative attempts to draft claims that will afford the
widest scope of protection possible while keeping each individual claim simple
enough that infringement could be proven. Drafting a patent, particularly the claims
section, has important implications for the success or failure of the application.
The Importance of Collaboration to Patents

A wide range of competencies and knowledge, both technical and institutional, must
be brought to bear in drafting a viable application. Undertaking research with the
aim of patenting implies both a knowledge of the patent process and of the relevant
literature and patent classes. Not only is it a challenge to determine what claims will
avoid infringement challenges, but there is an anticipatory decision to be made about
what to study in the first place. These decisions require a firm to be abreast of the
latest research being conducted by others in the field. Centrally-connected firms, we
argue, are best positioned to frame their findings in light of others' work, and to
choose research projects that mark sufficient departures from existing work.
The patent-examining procedure also holds pitfalls for the inexperienced. The
discovery of relevant prior art during an application requires a response by the
applicant. Conflicts over priority of invention, which are really claims about
infringement, take place during the examination process. The decisions of examiners
can be appealed but knowledge of when and how to make the appeal is necessary.
Patent applications provide no actual protection, but the examination process can be
expedited at the request of an inventor in cases where infringement might occur
before an innovation is patented. Each of these possibilities, and many others, can
further the scope of protections guaranteed by a patent and mean the difference
between a successful or unsuccessful application.
Thus, once an application is filed, a firm needs an array of competencies,
beyond just the scientific, to successfully navigate the patent examination process.
One way firms learn how to patent is by patenting a lot. But firms do not patent
frivolously-it simply costs too much. The examination process generally takes at
least 18 months and the cost of the application itself can run into the tens of
thousands of dollars (Emanuel 1995: 195). Given the difficulty of establishing that
an innovation is non-obvious, novel, and useful, as well as meeting the cost of the
process, there should be significant returns to knowledge of the patenting process.
Powell, Koput, and Smith-Doerr (1996) argue that one reason that ties to large
pharmaceutical firms are beneficial for small dedicated biotechnology firms (DBFs)
is that those ties allow the DBF to take advantage of the pharmaceutical's hard-won
knowledge of the U.S. Food & Drug Administration's approval processes. In much
the same way, a biotechnology firm turns to partners for accumulated knowledge of
the patent examination process and patent races more generally.
We stress that we are not arguing that a firm with more overall ties or more
R&D alliances will apply for more patents. We contend that firms with greater
collaborative capital, as evidenced by their network centrality, will subsequently
obtain more patents in a timely manner, due to their deeper insights into the
knowledge-base of the field and their ability to combine and utilize a broad
configuration of partners' expertise in choosing promising lines of research and
framing patentable claims. Put formally, we propose that collaborative capital is a
resource that builds up through participation in networks, which enables a firm to
patent more:
HI: The greater a firm's network centrality in a given year (controlling for its
overall number of ties and R&D alliances), the more patents it will obtain in
the subsequent year.
Patents as Intellectual Capital

Social scientists estimate a patent's value in at least three ways. Patents can be
conceived of as: 1) indicators of an unmeasurable variable, 'the accretion of
economically valuable knowledge' (Pakes and Griliches 1980: 377); 2) a set of
property rights to that knowledge (Griliches 1990); or 3) a unique form of
intellectual capital (Grindley and Teece 1997; Podolny and Page 1998). The first
two conceptions imply a static view of patents as valuable commodities that firms
hold; thus, a patent's value would decrease with use. The third conception takes a
more dynamic approach. While patents are still valuable commodities and the
carriers of exclusive proprietary rights, they are also a generative form of intellectual
capital that is distinct from the collaborative capital of network centrality. Under this
view, patents are valuable because they can be traded. As with other forms of social
capital, a patent's value increases with use.
Griliches (1990: 1672) reminds us that it is useful to distinguish between the
value of a patent and of patent rights. The distinction is conceptually important.
While many economists interested in innovation and technological change are
concerned with placing a value on individual patents, common measures of that
value depend on patent rights (i.e. Mansfield 1984; Lanjouw 1993; Lanjouw et al.
1996). However, patent law does not confer the right to make, use, or sell a patented
innovation. Patents only confer the right to exclude rivals from using the innovation,
even if that use is an outcome of independent R&D (Gregory et al. 1994: 8). Thus,
the important impact of a patent may not be in the creation of a monopoly on its use
but in the implied ability to affect the choices made by rival firms (Waterson 1990).
If a patent is defined as the right to extract rents from a new idea, then the value
of a patent is simply a function of its variable maintenance costs and rate of return.
Seen in this light, the decision to secure and maintain a patent is based solely on its
economic benefits (pakes and Schankerman 1984; Lerner 1995; Lanjouw et al.
1996). Consequently, patents would be viewed as relatively static measures of
innovative activity, or as the outcomes of zero-sum innovation races (Dasgupta and
David 1987, 1994). Firms would shy away from collaboration in order to secure the
exclusive award of intellectual rights. In this view, patents are associated with the
restriction of proprietary information, and thus are assumed to deter, rather than
foster ongoing collaboration between firms.
Furthermore, under this property-focused perspective, patents can be 'invented
around,' infringed, or invalidated (Mansfield 1984; Meurer 1989; Gallini 1992).
Meurer (1989: 77) suggests that patents are 'lottery tickets' whose value is uncertain.
Patents also 'spill over' technologically and economically from the firm that holds
them to its competitors. Jaffe defines 'technological spillover' as the influence of
other firms' R&D on a firm's R&D program (Jaffe 1986: 989) and estimates that
every million dollar R&D investment by a firm has a return of two patents to that
firm but also a return of .06 patents to rivals. To remain valuable under this
conception, patents must be held and closely guarded. When they are used publicly
their value decreases because the chance that rivals will invent around them
increases (Gallini 1992; Mansfield 1984). In industries where technological
knowledge 'spills over' and rivals wait to invent around or challenge patents, a
'lottery ticket's' value is predicated on a firm's ability to appropriate rents from an
innovation while keeping it from others.
We take a different perspective on patents, however. Recent analyses of the
patenting activities of universities (Slaughter and Leslie 1997; Powell and Owen-
Smith 1998) and its effects on academic culture (Etzkowitz and Webster 1995;
Packer and Webster 1996) emphasize instead the mUltiple uses and complex
ramifications of patenting. Rather than being solely the end product of an
organization's research efforts, patenting may also spur interorganizational
collaboration.
Patents represent not only economic value, but the intellectual competencies and
capabilities of a firm. As such, they may serve as advertisements to potential
collaborators. Dasgupta and David (1987: 533) make this point when noting the near
impossibility of transferring tacit technical knowledge through patents: '...the
purpose of filing the patent may be less that of seeking to deter imitators than of
signaling the availability of trade secrets for sale by the patentor.' We take their
claim a step further, arguing that patenting is a simultaneously backward- and
forward-looking activity. In addition to securing rents from completed research, a
firm's patents also advertise technical skills and areas of expertise to potential
partners. When a firm obtains more patents, it has more assets to trade or license.
Patent activity forms a key strand in the multiplex and ongoing web of relations that
characterize the biotechnology industry.
Grindley and Teece (1997) analyze patenting and licensing practices in several
high-technology industries, employing an intellectual capital view of patents. They
emphasize that rents can be appropriated from patents by licensing-trading the
right to infringe on a patent for royalty payments-and cross licensing-trading the
use of patented innovations for equivalent access to another firm's patent portfolio.
Strong patent portfolios both serve as advertisements-firms with strong portfolios
are more appealing partners than firms with weak portfolios-and provide access,
through cross-licensing, to innovations that might have remained unavailable
otherwise. A compelling quote from IBM's Assistant General Counsel supports this
idea: 'You get value from patents in two ways, through fees and through licensing
negotiations that give IBM access to other patents. Access is far more valuable to
IBM than the fees it receives from its 9000 active [U.S.] patents. There is no direct
calculation of this value, but it is many times larger than the fee income, perhaps an
order of magnitude larger.' (Grindley and Teece 1997: 15)
Note two points about this passage. First, IBM is not afraid that rivals will
'invent around' their patents. Their stance is not at all exclusionary. Second, the
strength and size of the firm's patent portfolio allows them to leverage access to
patents they need in cross-licensing negotiations. Here patents are valuable because
they can be traded for access to more patents. Access to the same patents can be
traded numerous times to multiple partners. Their value increases with use.
Furthermore, access to others' portfolios becomes a basis for further R&D, which
results in more patents making IBM a more appealing partner and opening avenues
to an even wider range of patents.
The fundamental difference between intellectual capital and intellectual
property views of patents stems from conceptions of their value. Intellectual
property views see patents as things to hold because their value decreases with use.
In contrast, an intellectual capital view sees patents as advertisements and things to
trade. Under this conception a patent's value increases with use. Where intellectual
property views highlight appropriability and proprietary rights we argue that
licensing, cross-licensing, and collaboration are the ultimate sources of patents'
value. An intellectual capital view of patenting links network connections and
collaborative capital with innovative R&D and further patenting. Thus we contend
that not only do network ties foster intellectual output, but in tum, network position
is enhanced as a result of extensive patenting:
H2: The more patents a firm obtains in a given year, the more centrally connected
the firm subsequently becomes (controlling for the overall number of ties and
R&D alliances).
Our view that patents are generative and lead to increased firm centrality in
networks may apply more to certain types of patents than others. Following Merges
and Nelson (1990), we recognize that patents vary significantly in the scope of
protection they claim. Thus we differentiate between prospect-defining patents and
patents on relatively discrete inventions. We attempt to capture a 'topography of
prospects' (Merges and Nelson 1990) by categorizing patents as principles,
processes, or products. Principle patents are those of the broadest scope, such as
Cetus' early patent for polymerase chain reaction (PCR) amplification. Rabinow
(1996) characterizes the development of peR at Cetus as a bricolage of multiple
competencies, technologies, and scientists. Kary Mullis received a Nobel Prize
largely on the strength of his work at Cetus on PCR, demonstrating the importance
of that principle patent. Process patents protect rights to particular uses of items,
such as the use of specific reagents or steps in the treatment of a patient. Processes
are considered less broad than principles but more broad than products, the most
narrow type of patent (e.g., Alza's patent on an innovative pipette tip). To assess the
effects of these three different types of patents, we measure the number of each
patent type held by a firm.
Of the three types of patents, we expect that the number of principle patents will
be a stronger predictor of future collaborations and patent activity. Principle patents,
as prospect defining, may be characterized as having a larger tacit component than
more narrow patents. The difficulty in transferring tacit knowledge through patents,
suggested by Dasgupta and David (1987, 1994), implies that principle patents may
advertise a firm's capacity for broader and more innovative know-how than do other
kinds of patents. The breadth of protections claimed by principle patents also
suggests that more diverse external organizations will seek to pursue collaborations
with the holder of the principle patent. And in an industry that maintains strong
reputational affinities with academic science (Powell and Owen-Smith 1998; Smith-
Doerr 1997), principle patents may have a greater effect on centrality because of the
scientific credibility provided by principles relative to other patents. Thus, we
hypothesize:
93: The more principle patents a firm holds in a given year (controlling for its
overall number of ties and R&D alliances), the more centrally connected the
firm becomes in the subsequent year.
And because we propose that network centrality leads to later patenting activity
(from HI), we further hypothesize that:
94: The more principle patents a firm holds in a given year (controlling for its
overall number of ties and R&D alliances), the more total patents it will hold
in the subsequent year.
DATA AND METHODS

Data Sources
Our data cover dedicated biotechnology firms (DBFs) over the period 1988-1996.
DBFs are defined as independently-held, for-profit firms involved in human
therapeutic and diagnostic applications of biotechnology. We do not include large
pharmaceutical corporations, international multi-business enterprises, agricultural or
veterinary biotechnology, or government or non-profit research institutes. These
organizations enter our database as partners that collaborate with DBFs. The
restricted nature of the firms included reflects our effort to assess the activities of
dedicated, independent firms in the most research-intensive sector of the field.
The data on firms and interorganizational agreements are taken from Bioscan,
an industry publication that reports on firms and the formal agreements they are
involved in. Bioscan covers nearly the entire population of dedicated biotechnology
firms in existence between 1988 and 1996, though sometimes it does not pick up
newly-formed, privately-held companies with few employees. By having nine years
of data, we can, in almost all cases, recapture such missing firms by tracing them
back from when they first appeared in Bioscan to their actual year of founding.
Bioscan initially published complete firm activity only in its April supplement; we
continue to use the data from April for consistency. Patent data are extracted from
CASSIS for the years 1987-1996. CASSIS is a government document, made available
on CD ROM by the U.S. Patent and Trademark Office, listing patent activity in all
scientific and technological areas. We sampled patents in CASSIS by recording those
assigned to firms on our list of DBFs. This method captures all patents in which our
DBF sample formally holds an interest, and allowed us to develop a relational
dataset that contains patent-level data for every DBF in our records. 3
By gathering patent data from the Assignment files (CASSIS/ASSIGN), we focus
on information about the organizations assigned and maintaining patents, rather than
on patent classifications (from CASSISIBIB). Because we begin with a list of biotech
firms, we assume that all their relevant patents fall in the area of human therapeutic
and diagnostic applications, and related instrumentation. Information from
CASSIS/ASSIGN reports was coded for characteristics of assignees, assignors
(inventors), and of the patents themselves. ASSIGN includes data on the assignment
of inventor's interest in patented innovations to other individuals or organizations. In
choosing assignment data, we may miss patents retained by individual scientists,
which are not assigned to their biotech firm employer. This potential problem is not
a key concern because its effect would mean we undercount patent activity, thus
weakening our findings rather than artificially inflating them. Our control of fixed-
firm effects insures that a particular firm with less of a propensity to obtain patent
assignments in the organization's name will not unduly influence results. We
suspect that if an idea is potentially valuable enough, a firm will file for a patent.
This data collection strategy captures every patent formally assigned to a biotech
firm. Our analyses focus on industry-level networks of firms, rather than on
networks of individual scientists employed by firms. Another potential problem with
patent assignment data is that firms may buy patent assignments from others rather
than doing the science in-house; but our suspicion is that this activity is rare in
biotechnology because of the emphasis on maintaining scientific credibility in the
industry (Smith-Doerr 1997). Moreover, if a firm becomes capable of knowing

where and how to acquire important patents because it is well-connected in the
industry, then the essence of our argument remains supported.
Measures
Patent productivity is measured by the number of patents granted to a DBF in each
year, 1987-1996. While a simple patent count is not a perfect measure of an
organization's intellectual output, it is a widely accepted proxy (Schmookler 1966;
Griliches 1990; Trajtenberg 1990). In their chapter in this book, Han and Breiger
classify volume as one dimension of social capital. Their focus is on the volume of
ties between organizational dyads; relatedly, we argue that the volume of patent
activity is an important dimension of intellectual capital in the biotechnology
industry. Other scholars argue for measurements of the 'value' or 'quality' of patents
(Lanjouw 1993; Lanjouw et al. 1996), but our focus on the generative aspects of
patent activity allows us to avoid this thorny data issue. By measuring the volume of
biotech firm patent activity we emphasize the 'advertising' role patents can play in
attracting potential collaborators. And in de-emphasizing patent economic value and
technological distance, our analysis avoids the necessity for either renewal data
(Lanjouw et al 1996) or co-citation4 and utility class data (Podolny, Stuart, and
Hannan 1996). In addition to volume, or number of patents in each year, recall that
we measure type of patent to capture variation in the effects of different types of
biotechnology patents. Above we described three types of patents: principles,
processes, and products. Patents with the broadest and most open prospects are
principles, while patents with the narrowest and most specific prospects are
products. The intermediate category, processes, captures patents that protect specific
methodological or therapeutic processes.
In computing centrality, we need to account for the fact that we do not have a
dosed network. In this respect, our measure of interfirm networks is somewhat
unconventional. We wished to examine the structure of the network linking our
sample of DBFs, but we need to define a closed set of firms to compute measures of
connectivity. Yet many of the ties that structure the field involve parties outside the
scope of our definition of a DBF-the overall universe of partners is open, diverse,
and expanding. We counted a connection between two DBFs: 1) when there was a
direct tie (degree one), and 2) when the DBFs were linked (at degree distance two)
through a common partner, to capture the information that may flow between DBFs
through a non-DBF partner. In measuring centrality we do not differentiate among
connections involving different business functions, i.e., R&D, investment, clinical
trials, manufacturing, licensing, distribution, joint venture, and complex ties (see
Powell, Koput, and Smith-Doerr 1996 for a more detailed discussion of data
coding). The various types of collaborative activities each playa comparable role in
creating a firm's overall set of relationships. Measures of central connectivity were
computed using UClNET IV (Borgatti, Everett, and Freeman 1992). Centrality is a
measure of how well connected, or active, a firm is in the overall network. We
gauged centrality of a firm locally rather than globally, in network parlance, such
that a firm's centrality is the number of other firms connected to that firm, ignoring
how well those partners are connected.
Included in the models are other variables that might be expected to have
statistically significant effects on patenting and centrality. These controls include
size (number of employees), firm age, and other network measures aside from
centrality. We controlled for alternative explanations that involve firm age or size as
predictors, rather than as outcomes, of network behavior. Age appears as a predictor
in ecological and life-cycle theories of organization, while greater size, indicating a
more extensive hierarchy, is seen as an alternative to alliances in the transaction-cost
literature, and as an outcome of learning in the network literature. Age and size are
not logged; skewness is shared by our dependent variables (see methods section for
discussion of skewness).
Other measures of network activity include number of R&D ties, network
portfolio diversity, and collaborative R&D experience. The number of research and
development ties a firm has captures the extent of its involvement in the core
activities of the industry, particularly activities driven by discovery efforts. As noted
earlier (Powell, Koput, and Smith-Doerr 1996), R&D has a unique status in theories
of organizational learning, and so we treat it separately. The range of ties that a firm
is engaged in at any given time reflects a firm's portfolio of collaborative activities.
Network portfolio diversity is computed for each firm in each year using Blau's
index of heterogeneity (Blau 1977).s Collaborative experience at time twas
measured as the time since inception of a firm's first alliance. Descriptive statistics
for each variable used in the analyses may be found in appendix Table AI.
Methods
Our data consist of nine years of cross-sectional records of firm-level variables. To
test the predictions, we used a panel-regression model. The selection of this
technique involves two primary theoretical considerations and the need to address a
number of statistical issues that stem from these concerns. The first theoretical
consideration is that social capital resides within firms and develops over time. We
argue that while social capital accrues through network relationships, firms are both
the actors and the recipients of the skills and expertise that social capital brings. To
eliminate any spurious effects due to differences between firms, we included fixed-
firm effects by entering a dummy variable for each firm. Thus we avoid, for
example, potential effects resulting from firms perhaps having more R&D ties and
more diverse portfolios because of patent activity prior to the period investigated.
We use fixed rather than random effects because we essentially have the population
of dedicated biotech firms over our observed time period, and not a random sample. 6
We wanted to estimate a dynamic model, in which the explanatory variables are
lagged one year to allow us to make causal inferences. But an important factor to
consider is that changes over time within firms will result in autocorrelated errors
and may bias estimates of the parameters in which we are most interested. One way
of breaking the correlation over time, so as not to overestimate the effects of our
hypothesized independent variables, is to include a lagged dependent variable,
Yi,t-l' as a predictor. When a lagged dependent variable appears as an explanatory
variable, however, the fixed effects estimator of the parameter on the lagged
dependent variable may not be consistent, because our dataset involves a large
number of firms observed over a short period of time. We used an instrumental

variables estimator (Hsiao 1986: Ch. 2) to remedy this potential problem.
The dynamics of social capital accumulation involve the co-evolution of firms
and networks. This process leads to an additional source of statistical non-
independence across our observations: For each firm we measure properties of its
position in the industry's network. but each firm's position can only be assessed in
terms of the network positions of others. For instance. a particular firm might have a
high centrality score because a number of outside firms act as partners to multiple
firms in the network. Hence. we need to control for effects that vary over time but
are constant across firms. such as the overall number of outside partners. the density
of the industry's network. government budgets for medical research. or the
economic circumstances of pharmaceutical companies. To do so. we included fixed-
year effects: a dummy variable for each year.
In using what is termed a fixed-effects specification (Judge 1985) for both the
firm and year controls. with a lagged dependent variable as one of the predictors. the
dependent variable. Yi.t' is modeled as:
J
Yi,l =a i + Y + ;"(Yi,l-t)+ L f3 j (Xi,t-t,j)+ Ei,l .
I
j=t
In this equation. a.I is the effect of firm i: i=I ....N; Yt is the effect of year t: 1=1 ....9;
f3.J is the within-firm slope for x .• pooled over all firms and years; and E.
J I,t
is a
normally distributed error term.
The strict assumptions of the normal regression model are violated. because our
primary dependent variables are skewed and only approximately continuous. The
number of patents and centrality can only take on non-negative integer values. Their
ranges are fairly large (approximately 25 and 130. respectively). however. making
the continuity approximation reasonable. While the truncation and skewness are
potentially problematic. these features are shared by the explanatory variables.
allowing us to make the a priori working assumption of symmetric disturbances. We
confirmed the validity of this assumption with diagnostic plots in a post-hoc residual
analysis (not reported here). Also. correlations among predictor variables are
displayed in the appendix.
We checked the robustness of our results in several ways. First. we applied a
square-root transformation to the dependent variables and re-estimated the normal
regression model. Second. for those dependent variables that are integer numbers.
we also estimated panel models for count data. Both of the additional models
confirmed the findings of the normal regressions. We report the results of the normal
regressions for all variables to ease interpretation and allow comparison of effects
across the models. We also conducted logistic regressions for the onset of network
ties and initial patents. The goal of the logistic models was to tease out the temporal
priority of explanatory variables in HI (network centrality) and H2 (patent
obtainment). In keeping with our learning argument. we expect that collaborative
capital increases success in navigating a patentable stream of research. and that
patents are not required for initiating alliances. but enhance network position
through a recursive feedback loop.
Table 1. Panel regression models of the effects of network centrality and patenting activity
Dependent variables at time t+ I
Explanatory variables at t Number of patents Degree Centrality

Degree Centrality .0288"'·· (.0052) .8356··· (.0285)
Number of patents .6553··· (.0273) .1375· (.0695)
Control variables at t
Total number of ties .0461·" (.0094) .2791·" (.0659)
Number of R&D ties .0071 (.0299) .3370··· (.1100)
Diversity of network activity .0251 (.2473) .0843··· (.0173)
Collaborative experience .1388· (.0629) .2486 (.1506)
Size .0735·· (.0221) .1945 (.1598)
Age -.0074 (.0101) -.3383·" (.0709)
Full R2 (within-firm R2) .6372 (.3535) .7871 (.2031)

N 993 992
·p<.05, ··p<.ool , ···p<.OOOI
Note: all models include fixed firm effects (dummy variables for firm ID), and dummy
variables for year effects as controls.
RESULTS
The results of regression models are displayed in Tables 1-3, which present
unstandardized regression coefficients, with standard errors in parentheses.
Table 1 presents results of panel-regression models addressing the hypotheses
that greater centrality leads to more patents (HI), and that more patents lead to
greater centrality (H2). Both hypotheses are supported. Centrality at time t increases
the number of patents obtained in time t+ 1, and receiving more patents at t leads to
increased centrality at 1+ 1. Note that the statistically significant effects7 of centrality
and patenting occur while controlling for the total number of ties held by a firm, as
well as number of R&D ties, diversity of collaborative activity, experience with
network ties, size, and age. Recall that the models in Table 1 also control for fixed-
firm and year effects, thus ensuring that unmeasured firm or time characteristics are
not skewing the results.
While we do not look at controls as dependent variables, 8 we found some
interesting implications in comparing the effects of control variables on patenting
and centrality. Experience and size have statistically significant effects on patents,
but not on centrality. Age is negatively related to centrality so that more central
firms are younger, a finding consistent with our earlier work (Powell, Koput, and
Smith-Doerr 1996). Holding constant earlier patent activity, other network
measures, size, and age, centrality predicts subsequent patenting at the .0001 level.
The effects of patenting on later centrality, taking controls into account, are also
notable.
R&D Ties
Centrality
Network Diversity
Figure 1. Visual representation of the relationship between network measures and patent
activity
The effects of network control variables speak to a more complex relationship

between network activity and patenting than might be expected. The number of
research and development ties does not predict patents at a later time, but does
predict increased centrality. Likewise, diversity of network activity leads to
increased centrality, but is not meaningfully related to patent activity. These findings
suggest the relationship between variables illustrated in Figure 1.
While R&D ties and diversity of activity lead a firm to a more central position
in the industry, they do not lead directly to patent activity. Prior to increased
patenting, a firm needs central connectivity, which indicates that social capital from
collaboration precedes the propensity to patent. Collaborative experience mediates
between centrality and patent activity, in that experience has a statistically
significant effect only on number of patents. Experience with ties increases
patenting, but only imperceptibly affects centrality, perhaps because centrality often
precedes extensive collaborative experience.
Table 2 provides further evidence of network centrality'S temporal priority to
patent activity. In a set of logistic regression models, we examine whether centrality
better predicts a first patent or whether patenting is a better predictor of an initial tie.
Models 1 and 2 in the Table clearly show that patents scarcely predict formation of
an initial tie, while network ties do, with statistical significance, predict obtaining a
first patent, even controlling for the size and age of firms. But as Table 1 and Figure
1 show, patenting provides intellectual capital that enhances a firm's collaborative
capabilities.
Table 2. Logistic regression models predicting initial patents, and initial network ties
Dependent variable at time t

Explanatory variables at time t-I patent assigned (1) tie initiated (2)
firm has tie(s) .8223** (.3277) 5.5660**** (.3865)
firm has patent(s) 1.6034**** (.1475) .7146 (.4527)
Size .2932**** (.0547) .1343 (.0842)
Age -.0301 * (.0139) -.0630** (.0240)
Chi-Squared 266.97**"'* 401.71 ****
N 1086 1086
*p<0.05, **p<O.01, ***p<.OOI, ****p<.OOOI
Note: models include dummy variables to control for flflD effects and year effects.
Table 3. Panel regression models including the effects of patenting scope
Dependent variables at time t+ I

Explanatory variables at t Number of patents Degree Centrality
Degree Centrality .0292*** (.0053) .8385*** (.0299)
Number of Principle patents .6203*** (.1541) .2471*** (.1153)
Number of Process patents .1348* (.0701) -.9254* (.4585)
Number of Product patents .5666*** (.0451) .6103*·* (.2952)
Control variables at t
Total number of ties .0461**· (.0094) .2791*·* (.0659)
Number of R&D ties .0071 (.0299) .3370**· (.1100)
Diversity of network activity .0251 (.2473) .0843·*· (.0173)
Collaborative experience .1388* (.0629) .2486 (.1506)
Size .0735** (.0221) .1945 (.1598)
Age -.0074 (.0101) -.3383*** (.0709)
Full R2 (within-firm R2) .6378 (.3692) 7877 (.2111)

N 993 992
·p<.OOI, ··p<.OOOI, ·**p<.OOOOI
Notes: All models include fixed firm effects (dummy variables for firm 10), and dummy
variables for year effects as controls. Partial models were run, but as each statistically
significant variable adds to the overall R2, only full models are presented.
In fredicting number of patents, prior patents of each type have positive

effects. Table 3 indicates that the effects of types of patents act additively, that is,
there are no interactive or multiplicative effects. Principle patents are the best
predictors of greater subsequent patent activity, and product patents have the next
largest effect. Degree centrality is also predicted by greater numbers of prior
principle and product patents. Process patents have a negative effect on subsequent
centrality, however. This finding seems to be due to a small group of firms that have
only process patents and no other types. These 'isolated' specialists also have a
higher than average failure rate, underlining the importance of network centrality to
biotech success.
As evident from the R2 values reported in the tables, the models are explaining
much of the variance in DBFs' collaborating and patenting. Each statistically
significant coefficient, when added hierarchically, adds to the model fit so that the
change in R2 is statistically significant at the .05 level. For all of the results
presented in tabular form, it is important to note that it is the within-firm R2 that is
driving the between-firm (full R2) effects rather than vice versa. The between-firm
collaborative and intellectual capital differences appear because within-firm learning
occurs at different rates. Thus our contention that the firm is still an important unit
of analysis even in a highly connected industry is supported by the importance of
within-firm variation to the models.
DISCUSSION
All told, we expect that collaboration and patenting constitute complementary
activities rather than cross-purposes in biotechnology. Of course, keep in mind that
we hypothesize the importance of patents as intellectual capital in the context of a
knowledge-intensive, expanding field, rather than one dominated by incumbent
organizations vested in a technological paradigm half a century old. Our results
show that, as hypothesized, centrality predicts subsequent patenting, and patenting
predicts subsequent centrality in the biotechnology industry. When controls for other
measures of network activity are entered into the relationship between centrality and
patenting, an interesting wrinkle appears. In earlier analyses (Powell, Koput, and
Smith-Doerr 1996), we had found that R&D ties are critical to accessing networks of
learning in the industry. While R&D ties continue to affect centrality here, they are
not a direct predictor of patent activity, suggesting that R&D ties may still be the
admission ticket to learning races, but collaborative capital (assessed by centrality,
controlling for R&D and other ties) provides the fuel for patenting activity. In order
to choose research projects that can be developed to the stage of obtaining a patent, a
firm appears to need relationships that place it at the center of the industry's
activities, rather than just R&D ties. 10
To some extent, this effect of centrality on patents suggests the importance of
other skills besides the capacity to do research in the patent process. For example, a
central firm may have access to more savvy information about the state of research
in the field as a whole. Our focus on data that look at patent activity in the year
following a firm's centrality score suggests that a central firm is more likely to have
the necessary connections to organizations with knowledge on how to get patents
through the examination process in a timely manner. Thus, patent activity is a result
not just of fragmented scientific knowledge and R&D, but of a broader awareness
and ability to combine knowledge-from otherwise unconnected collaborative
activities-about what to study and what to claim.
The results of the effects of patent types on subsequent centrality and patent
activity are generally supportive of our predictions. Principle patents have the
greatest magnitude of effect on centrality and further patenting. The finding that
process patents have a statistically significant negative effect on centrality is
somewhat puzzling, but a closer look at the data provided some initial answers.
Apparently, process patents are actually more closely related to specialization than
product patents, contrary to our expectations. A small group of firms whose only
patents are processes have a high failure rate, and not surprisingly, low centrality.
Firms that have a majority of their patents as processes also appear, based on their
annual reports, to be more specialized in their project collaborations than other
DBFs. But having a process patent alone does not result in less centrality-those
firms with the most process patents have the most patents generally, and are
centrally connected. A specialization in only process patents appears to detach a
firm from core networks. We plan further analyses of this issue and will use more
fine-grained coding of our process patents in future work.
To summarize, we stress the learning aspects of our argument and results. We
developed an organizational aspect of social capital which we termed collaborative
capital. Whereas individual social capital is often characterized as the ability to
connect otherwise disconnected social actors, collaborative capital includes the

ability to combine, in novel ways, the activities of diverse parties in an
interorganizational network. We assessed collaborative capital by measuring an
organization's network centrality, controlling for its number of overall ties and R&D
collaborations, time in the network, and the range of activities in its network
portfolio. Hence, collaborative capital is distinguished from collaborative capacity,
research intensity, experience, or diversity-it is an extra dimension that both
combines and extends these other aspects. Collaborative capital builds over time as a
firm participates in research, gains experience, develops capacity, and assembles a
diverse profile of activities that moves it toward the center of the network and gives
it greater access to knowledge spillovers. Once centrally-placed, collaborative
capital enables an organization to create opportunities with the greatest potential for
timely impact and payoff. Our results demonstrate a link between network centrality
and patent obtainment, which we take as evidence of collaborative capital in action.
We document that patents are a generative form of intellectual capital that, in
knowledge-intensive fields such as biotechnology, once obtained bring scientific
visibility to further enhance a firm's collaborative social capital.
Viewed more broadly, our results suggest that, in industries where knowledge is
either expanding rapidly and/or distributed broadly, firms that adopt a view of
patents as an admission ticket to an information network will be rewarded. As we
have argued previously (Powell and Smith-Doerr 1994) access to knowledge and
benefit-rich networks is a critical resource in an information-intensive economy.
Collaborative capital, a critical component of social capital, enables organizations to
expand their reach well beyond the capabilities and limitations of a single firm.
Collaborative capital provides not only deeper resource pools, but opens up the
possibility of combining ideas, people and resources in novel ways.
APPENDIX
Descriptive Statistics and Correlations ll
Table AI. Means, standard deviations, and range of variables analyzed
Mean Standard Minimum Maximum Number of

deviation Cases
Age 7.65 6.28 0 46 2005
Centrality 15.49 22.65 0 132 2062
Size 144 360 1 6000 1396
R&D ties 1.23 2.09 0 16 2063
Diversity .44 .28 0 .85 1562
NPatents 1.05 2.53 0 25 2733
Principle patents .06 .33 0 5 2733
Process patents .40 1.07 0 12 2733
Product patents .57 1.63 0 22 2733
Total ties 7.38 9.84 0 82 2063
Experience 3.78 3.80 0 32 2063
Table A2. Between-finn correlations of predictor variables. including patent types
Age Centrality Size R&D Diversity Total Exper. Patents Principle Process
ties ties patents patents
Age
Centrality .1574
Size .3752 .4480
R&D ties .2559 .67% .3002
Diversity .1745 .5091 .2834 .4593
Total ties .2190 .6651 .5699 .6214 .5568
Experience .3793 .2721 .3517 .2466 .1634 .2920
Patents .3049 .4812 .6054 .3680 .3044 .5877 .2216
Principle .1354 .4573 .3980 .2672 .2595 .5042 .1519 .5055
patents
Process .2918 .4393 .5994 .3119 .2992 .5788 .2097 .9044 .5270
patents
Product .2752 .3907 .4981 .3272 .2404 .4597 .1837 .9363 .3050 .7117
patents
Table A3. Within-finn correlations of predictor variables. including patent types
Age Centrality Size R&D Diversity Total Exper. Patents Principle Process
ties ties patents patents
Age
Centrality .1224
Size .0725 .1127
R&D ties .0862 .5278 .0540
Diversity .0890 .2967 .0348 .2353
Total ties .1171 .6042 .1286 .6180 .3979
Experience .2787 .3287 .1255 .1403 .1408 .2942
Patents .1547 .0730 -.0908 .0757 .0811 .1343 .0667
Principle patents .0571 -.0037 -.0l2l -.0038 .0222 .0423 .0256 .1839
Process patents .1199 .0920 -.0307 .0560 .0459 .0765 .0380 .2830 .2299
Product patents .1671 .0428 .0865 .0268 .0454 .0473 .0587 .4446 .1269 .3502
Research support provided by NSF grant #9710729. W.W. Powell and K.W. Koput, Co-P.I.'s. Direct
correspondence to Laurel Smith-Doerr at: Department of Sociology. Social Sciences 400. University of
Arizona. Tucson. AZ 85721. USA; or bye-mail: ldoerr@u.arizona.edu.
NOTES
1. For example, in the index to the Handbook of Economic Sociology (Smelser and Swedberg 1994),
under social capital the reader is directed to 'see also networks.'
2. These three interorganizational definitions of social capital illustrate common usage of the concept
but are not meant to constitute an exhaustive list of the fonns of social capital. See the chapters by
Gabbay and Leenders and by Pennings and Lee in this volume for further definition of social capital at
both individual and organization levels.
3. CASSIS data are counted at year end. Hence, between 9 and 24 months separate year t centrality and
year t+ I patents, for example. (See methods section for further details on statistical issues).
4. See Stuart's chapter in this volume for an analysis of patent citation data.
5. Blau's index is computed for each finn in each year as follows. For finn i in year t, denote the
number of ties of typej as nit,j and the total number of ties aggregated over all types (j=J •..•8) as nit'
The proportion of finn i's ties of type j, out of the total number of ties. is denoted as Pit.j and given by
. . =n.It,}. / n.It . Each p.It.). is
PIt,} squared and then the sum is taken over all j and subtracted from I,
8
resulting in the index of heterogeneity. y. , as so: y .
II It
=1- "" p2 . .
-'-' ",)
j=1
6. DBFs are our physical population. while our statistical population is all possible realizations of the
stochastic processes of patenting and collaborating-thus the use of probability statistics is warranted.
7. Statements on statistical significance refere to a .05 level.
8. Thus, the following discussion of control variable effects is meant as suggestive rather than
definitive.
9. Dummy variables to indicate whether patents of each type were held or not were also run in models
to predict further patenting and centrality, but had statistically nonsignificant effects.
10. Also recall that explanatory variables are only lagged one year. perhaps not allowing enough time
for the measure of R&D ties to affect patent activity in the data analysis. The fact that collaborative
experience is positively related to patenting would tend to support this interpretation.
II. The between-finn correlations are presented for inspection. Our models are within-finn. however.
The within-finn correlations are all modest, posing no coIlinearity problems.
Supply Network Strategy and
Social Capital
•
22
Christine Harland
ABSTRACT
The focus ofthis chapter is on interorganizational supply networks and, specifically,
the social capital and liability within them. Supply networks are sub-networks nested
within interorganizational networks that include the actors, resources and activities
associated with value adding processes converting resources to goods and services.
To describe the nature of supply networks and the concept of supply, the literature
review within the chapter traces the development of the concept through the
different systems levels of supply within the firm boundary, within dyadic
supplier-customer relationships, within chains of relationships, and within whole
focal firm supply networks, providing examples of social capital and social liability.
A normative, rational view of supply network strategy is presented, using
examples such as Benetton and Toyota to support the argument that social capital
may not be an emergent feature of supply networks but may actively be pursued by
firms seeking to increase competitive advantage; this view is in opposition to much
of the work of the Industrial Marketing and Purchasing (IMP) group that maintains
the coping nature of players in interorganization networks. Firms such as Benetton
and Toyota appear to have taken on the role of supply network hub, facilitating and
coordinating more effective and efficient flows of materials and information. They
also appear to have focused strategically on end customers and their requirements.
Findings are also presented to support the view that social capital in networks,
rather than being a tangible, measurable asset, is a function of actors' expectations
and perceptions of performance and is therefore a subjective notion.
INTRODUCTION
There has been increasing interest in interorganizational networks notably since the
early 1980s when original work by the Industrial Marketing and Purchasing group
(IMP) provided models of interaction and a language to discuss aspects of networks
(Hakansson 1982). One of the main contributions of the IMP work on networks is
the actor, resource, activity model that inspired the specific notion of economic
resources discussed by Araujo and Easton (this volume). However, to date there has
been only limited empirical work that might provide managerial guidance. Most of
the field research evident has been conducted in dyadic relationships, not networks.
Within interorganization networks exist other subnetworks; the focus of this
chapter is on supply networks. Supply is defined here from the operations
perspective as the value-adding transformation processes involved in the conversion
of input resources to provide goods or services.
Supply processes occur within firms and between firms. Different subject areas
have undertaken research into these supply activities within the firm, in
supplier-customer relationships, in chains of firms, and, more recently, in
interorganizational networks of firms. Findings to date at each of these system levels
are presented here, with particular attention to benefits gained (social capital) and
disadvantages suffered (social liability).
Before supply networks, and any consideration of them (such as social capital
within them) can be addressed, the concept of supply, including its boundary and
content, requires explanation and clarification, as do the observed trends in business
practice that have lead to the development of the concept. The next section,
therefore, examines the concept of supply.
THE CONCEPT OF SUPPLY

Research in supply integrates various existing bodies of knowledge and concepts, to
form an holistic, strategic perspective of management of operations, stretching
across interorganizational boundaries. Core to the concept of supply is the
procurement, use, and transformation of resources to provide goods and service
packages to satisfy end customers. End customers are defined as the ultimate
decision makers that take a product/service differentiation decision (Harland 1996a).
The concept relates to the integration of supply activities within firms, in dyadic
relationships, in chains of firms, and in interorganizational networks. These have
been expressed as different systems levels of supply (Harland 1996a) as shown in
Figure l. Common to all these levels is the flow of supply and the activities and
decisions associated with that flow.
The development of the concept and the business practice of managing supply
strategically has progressed over time through these levels whereby there is now
emerging research and documented business practice relating to supply strategy in
interorganizational networks; therefore, it is useful to discuss theory and practice
progressively through each of these levels.
Supply Network Strategy and Social Capital - 411
LEVEL 1: Supply within the firm boundary
LEVEL 2: Supply in a dyadic relationship

o
LEVEL 3: Supply in an interorganizational chain
LEVEL 4: Supply in an interorganizational network
Figure 1. Levels of supply (adapted from Harland 1996a)
SUPPLY WITHIN THE FIRM BOUNDARY

The management of supply within the firm boundary involves the integration of the
processes from the input side of the firm to the output side. In a manufacturing
organization this requires integration across the functions dealing with physical and
information flows, as shown in Figure 2.
This relates closely to the preexisting concepts of materials management
(Ammer 1968; Lee and Dobler 1965) and the value chain (Porter 1985; Johnston
and Lawrence 1988; Kogut 1985). Early work in supply chain management (Oliver
and Webber 1982; Houlihan 1984; Stevens 1989) identified logistics benefits of
reduced lead times and costs through integrating of the internal chain.
The span of the supply activities internal to the firm are decided through
make-or-buy decisions that determine which activities will be vertically integrated
or owned within the boundary of the firm and which will be procured through
transactions with other firms. The new institutionalists theory of transactional
economics (Williamson 1985; Van de Ven et al. 1975) provided a rationale for the
make-or-buy decision. Ellram (1991) related these decisions specifically to supply
by identifying advantages and disadvantages of vertical integration (Table 1).
A trend toward vertical disintegration has been highlighted in various industries
(Thackray 1986; Porter 1987), partly to reduce risk of being locked into
inappropriate technologies (Abernathy 1978; Harrigan 1983; Miles and Snow 1987)
as markets changed to demand more variety of products and services. To meet this
Table 1. Supply chain management: the industrial organization perspective (derived from Ellram
1991)
Advantages of Vertical Integration Disadvantages of Vertical Integration

Control: uncertainty reduction of costs, quality LimitIng competition: inability to replicate
and quantity of supply, convergent market incentives, and distortion of internal
expectations, reduced probability of infonnation.
opportunism, reduced probability of
externalities caused through dependence on Diseconomies: balancing scale economies,
monopoly suppliers and ability to protect inability of management to control large
important proprietory or competitive organization effectively, limits on span of
knowledge, ease of conflict resolution. control, and increased difficulty in
communication.
Communication: improved coordination of
processes and greater goal congruence. Risk: asset concentration, exit barriers,
perpetuation of obsolete processes, and
Cost: economies of scale through avoidance of exaggerated synergies.
intermediaries-notably procurement, sales
promotion and distribution, process integration,
avoidance of switching transaction costs.
Business management
Business
support
Operations management
Human resource management
Financial management
Marketing
Information systems management
Materials Purchasing Inventory Operations planning Sales-order Physical

rnngt. management and control management distribution
management
Physical Goods inward Stores Operation WIP Finished goods Transport
Figure 2. Internal supply cham

SUPPLY
•
change in market demand, supply activities within firms increasingly concentrated

on a limited, manageable set of tasks based on the principles of focus (Skinner
1969). These focused tasks were designed to meet the order winning criteria of
customer groups (Hill 1985; Christopher 1992). The trend of focusing and vertical
disintegration is continuing through outsourcing. Outsourcing noncore activities,
such as catering, security, and information systems, has been performed by many
manufacturing and service organizations. Recently, this has extended to some supply
activities such as logistics, purchasing, and sales. Reve (1990) argued that only core
skills with high asset specificity that significantly contribute to competitive
advantage should remain as internal activities. However, the less tangible these core
skills and assets are, the more difficult it is to quantify them to decide what is core.
Furthermore, what is viewed as noncore today may become core in the future; the
dynamics of business are such that the boundaries of core will ebb and flow
(Macbeth and Ferguson 1994). Therefore, significant change has occurred in supply
within firms, and this change has brought with it complexity of decision making
relating to what should remain internal to the firm and what should be supplied
through dyadic relationships.
SUPPLY IN DYADIC RELATIONSHIPS

Original work in industrial economics by Marshall (1923) and Coase (1937)
identified the existence of alternative types of relationship to vertical integration and
competitive, arm's-length market transactions. The intermediate types of
relationship were later defined by Richardson (1972) and Blois (1972),
differentiated by the closeness of the bonds between actors in terms of equity and
commitment. The shift to relationship rather than vertical integration has lead to the
industrial organization and contract view of the firm as a nexus of contracts (Aoki,
Gustafsson, and Williamson 1990; Cox 1997). Relating specifically to supply,
Christopher (1992) and Ellram (1991) positioned supply chain management as the
management of these intermediate types of relationship.
Authors and practitioners from many different disciplines and functions have
highlighted the increasing dependence on relationships with suppliers (see, for
example, Sabel et al. 1987; Christopher 1992; Slack 1991; Schonberger 1986).
Closer, longer-term relationships are evident in some industries, reported notably in
the Japanese automotive industry (Lamming 1993; Womack et al. 1990), the
Japanese textile industry (Dore 1983), craft-based Italian industries (Lorenzoni and
Ornati 1988), and various Swedish manufacturing industries (Hakansson 1987).
These closer relationships were enabled through the movement away from
multi-sourced, adversarial trading and toward single or dual sourcing. For example,
Rank Xerox reduced its supply base from almost 5000 suppliers in 1981 to 300 by
1987 (Morgan 1987). Lamming (1989) reported that Japanese lean producers
involved less than 300 suppliers in the development of new products, compared to
their Western counterparts, who dealt with 1000-2500. By transacting with
substantially reduced supply bases, these particular firms were able to collaborate
more deeply with suppliers to achieve improvements, notably in speed and
reliability of delivery, quality and cost through waste reduction, thereby deriving
much social capital through supply relationships. Smith-Doerr et al. (this volume)
discuss an upward spiral effect of collaborative capital glvmg rise to greater

innovation, giving rise to greater collaborative capital, and so on.
Frazier et al. (1988) and Lascelles and Dale (1990) identified that traditional,
adversarial transacting with competing suppliers was not conducive to generating
good quality, thereby giving rise to social liability. The embracing of waste
elimination concepts such as just-in-time (lIT) and total quality management (TQM)
necessarily involved and depended on suppliers. What was common to all these
developments in collaborative relationships is evident in Mowery's (1988) definition
of interfirm collaboration in that they all sought to provide mutual medium- or
long-term benefit and that they involved substantial contribution by partners in
capital, technology, know how or other assets. However, the fact that some
long-term, interfirm collaborations, notably in relatively high-volume, low-variety
situations, gave rise to social capital should not be extended, without substantially
more empirical evidence, to other contexts.
Recently these close, collaborative relationships with suppliers have been
termed partnerships. Partnership sourcing has been defined as 'where customer and
supplier develop such a close and long term relationship that the two work together
as partners to secure the best possible commercial advantage. The principle is that
teamwork is better than combat. If the end customer is to be best served, then the
parties to a deal must work together-and both must win. Partnership sourcing
works because both parties have an interest in each others success.' (Hornby 1992)
This introduced the end customer, and better serving the end customer, as a driving
force for forming collaborative relationships within which partners would increase
their commercial advantage over and above that derived through independent
working. Contractor and Lorange (1988) asserted the superiority of a strategy of
cooperating over independent operating. Some authors have expressed the
motivation for collaboration more strongly than a strategic choice taken for
competitive advantage, but rather as mandatory to compete in global markets (Kogut
1988; Ohmae 1989; Porter and Fuller 1986).
However, the nature of partnerships is more than just teamwork. Teece's (1986)
view of collaboration is one of supplier and customer providing complementary
assets, which Contractor and Lorange (1988) stated would provide synergy. A
substantial part of this synergy has been identified as learning (Hamel et al. 1989;
Macbeth and Ferguson 1994; Imai et al. 1985; Lamming 1993) providing social
capital to the relationship. However, the social exchange leading to social capital
requires an appropriate atmosphere (Hakansson 1982). Social exchange within
routinised, long-term collaborative relationships reduces uncertainty between the
partners leading to an interlocking of supplier and customer through developed trust
(Ford 1978). Thus trust is important in the building of social capital in relationships.
Trust in relationships relates in part to expectations (Hakansson 1982) and
perceptions of performance. Ring and Van de Ven (1992) suggest that trust evolves
partly from each party's perception of the other's fairness. Research on expectations
and perceptions of performance in interorganization relationships has been
contributed, notably, by the fields of service management and consumer behavior.
Supply Network Strategy and Social Capital- 415
Expectations
Both parties to dyadic relationships may have different expectations (Voss et al.
1985); this may be due to real differences in opinion of supplier and customer or
because they receive different information (Sheth 1973). Expectations may change
over time as the parties learn through the process of interaction and as the
relationship develops; satisfaction relates more to present needs, wants, and desires
than to those prior to interaction (Cadotte et al. 1987; Miller 1976; Swan and
Trawick 1980).
Expectations may be of different types. Boulding et al. (1993) classified
expectations as ideal, should, and will expectations. Ideal expectations reflect
enduring, possibly ambitious wants and needs that are relatively stable over time.
Should expectations are those assumed should happen. Will expectations are
tempered by experience and are a realistic assessment of what will happen and are
affected by recent or critical incidents (Bitner 1985).
Therefore, to achieve the required trust that enables the synergistic acquisition
of social capital in supply relationships, clarity and agreement on expectations is
vital.
Perceptions of Performance
Perception of performance replaces reality in the eyes of each party to a relationship
(Sasser et al. 1978; Parasuraman et al. 1985; Gummesson 1987; Berry and
Parasuraman 1991; Cronin and Taylor 1992). Perception is influenced by capability
to judge tangible and intangible aspects of performance (Haywood-Farmer and
Nollet 1991). Perceptions of performance have been identified as determining
success or failure in the development of long-term relationships (Carlisle and Parker
1989).
Satisfaction and dissatisfaction arise as a result of comparison of expectations
and perceptions of performance (Berry and Parasuraman 1991; Brogowicz et al.
1990; Gronroos 1990; Haywood-Farmer and Nollet 1991). Relationships where the
gap between expectations and perceptions of performance is small are more likely to
yield social capital in the form of satisfaction; however, where the gap is large, this
gives rise to dissatisfaction or social liability. The social capital within relationships
is vulnerable to critical incidents (Bitner et al. 1985), but distinction can be made
between short- and long-term satisfaction. Attitudes in longer term relationships are
more of a global nature and are a function of previous attitudes and satisfaction and
dissatisfaction with current transactions. Boulding et al. (1993) argue that behavioral
outcomes such as loyalty and positive word of mouth are a function of overall
cumulative perceptions rather than short-term satisfaction with individual
transactions.
Harland (1995) investigated gaps in perceptions of requirements and
performance in dyadic relationships between players in the automotive aftermarket.
A mismatch tool (see Figure 3) was developed to investigate these gaps.
In the relationships studied, customer dissatisfaction was highest where the
supplier and customer had the largest difference in perception about performance in
the relationship (mismatch 2). Substantial effort and investment was being made by
suppliers in understanding what ·was required. However, little attention or
investment was devoted to understanding and communicating with customers about

operational performance; this questioned the appropriateness of a marketing focus in
favor of a supply focus. A further finding related to the closeness of the relationships
and its impact on satisfaction and dissatisfaction. Dyadic relationships conducted
through very close interpersonal relationships with depth of social exchange were
compared to those conducted in more distant, nonfriendly relationships with more
systematized information exchange than social exchange. The close relationships did
not yield any higher satisfaction than the arm's-length relationships. This finding
reflected differences in expectations-in some relationships the parties wanted
closeness, in others they did not. These expectations then set the benchmark against
which performance was compared by the parties. The different expectations typified
cultural, territorial variation; parties to UK-based relationships expected more
distance than those in Spanish relationships studied where friendship between
supplier and customer was expected. These findings put into perspective some of the
social capital literature through highlighting that closeness in relationships, per se, is
not always a provider of social capital; if closeness is·not expected or welcomed by
one of the parties to the relationship, actions by the other party to become closer
may even give rise to social liability. Conversely, if closeness is not appreciated by a
partner, nonclose relationships will yield more social capital to the parties than will
close ties.
It is evident, therefore, that social capital is an important feature of longer-term,
cooperative dyadic supply relationships though its presence is affected by parties'
expectations, perceptions of performance, and resulting satisfaction and
dissatisfaction and trust.
SUPPLY IN INTERORGANIZATIONAL CHAINS

A supply chain is defined here as a connected string of dyadic relationships that
form a route for material-product-service flow. There are various terms to represent
the same route. Hayes and Wheelwright (1984) termed this a commercial chain
involving flow from raw material source, through manufacture, distribution, and
ultimate sale to a consumer. Farmer and Ploos von Amstel (1991) referred to a
supply pipeline. Marketing provides the term channel to refer to the route from a
manufacturer downstream to consumer through distribution (e.g., Stock and Lambert
1987).
Supply chains have been studied in some research merely as a convenient unit
of analysis of interorganizational features, notably logistical features. Studies in
industrial dynamics (Forrester 1961; Burbidge 1961; Towill 1991) have observed
what has been termed the Forrester or bull-whip effect. These studies have observed
information flows and ordering patterns in supply chains that have given rise to
cumulative error, or noise, that distorts real demand at the end of a supply chain as it
is manipulated and transmitted in dyadic relationships and within firms in the chain.
The distortion is amplified as demand is progressively transmitted up the supply
chain. This cumulative error-arising, in part, from second guessing, lack of trust
between dyadic relationships in the chain concerning forecasts of anticipated
demand, batching of information, and delay in its transmission-is a form of social
liability.
Supplier's Customer's
perception of perception of
requirements requirements
mismatch 1
mismatch 2
Figure 3. Mismatch tool (Harland 1996a)
The principles of the Forrester effect have been tested in relation to behavioral
factors in supply chains. Harland (1996a) showed that satisfaction and
dissatisfaction increased upstream in supply chains and was positively correlated to
the size of gaps in perceptions of performance. This provides evidence of other
forms of social liability in supply chains that varies according to physical position
(upstream or downstream) in the chain of supply.
However, supply chains have been considered more strategically than mere
convenient units of analysis. Hayes and Wheelwright (1984) and Hill (1985)
included strategic consideration of the direction, extent and balance of vertical
integration in the commercial chain and subsequent sourcing decisions within their
manufacturing strategy formulation approaches. Eberling and Doorley (1983),
Jarillo (1993), Kogut (1984), and Porter (1985) discussed the strategic design of the
value-added chain to implement different firm strategies, notably low-cost or
differentiated strategies. Kogut (1985) examined international location of links in
the value chain to gain comparative and competitive advantage. In the field of
operations strategy, Shi and Gregory (1994) differentiated between a domestic and
multidomestic orientation to international operations strategy and a global
orientation and their impact on domestic, regional, multinational, and worldwide
configuration of operations.
Evidence of normative, rational design decisions relating to differentiated
strategies in supply chains is apparent in some documented case examples. Johnston
et al. (1997) described how differentiated demand for blood products had lead to
purposefully designed supply chains to satisfy that demand. For example, neonatal
patients requiring fresh blood were served through the bleeding of donors located
close to the point of use and a rapid testing and delivery route to the patient. The
infrastructure of their supply chains was also differentiated; blood was collected in
different designs of blood bags that matched the requirement of the procedure that
would use the blood. In the automotive sector, Harland (l997b) identified that
supply chain design differed in some instances but not in others to meet the
differentiated demands of particular original equipment manufacturers compared to
aftermarket spares providers. Where a rational, differentiated approach to designing
the supply chain had not been taken, certain critical links in the chain were less
effective at coping with varying demands from downstream.
It appears, therefore, that individual supply chains can be designed and managed
to meet differentiated demands of end customers. The nature of this differentiation
gives rise to the need to understand and manage the complexity of the total set of
chains that any firm supplies in. Extending the concept of focus (Skinner 1969)
greater complexity and variety of these chains will impact on a firm's (and its dyadic
relationships) ability to learn and benefit from the social capital that this learning
provides. Therefore, a higher system level-the total supply network-has to be
understood. Harland (l996b) proposed that strategic supply decisions should be
considered at the level of the interorganizational network.
. SUPPLY IN INTERORGANIZATIONAL NETWORKS

Every firm is a nexus in its own unique network of upstream and downstream
chains, forming a structure. Within that structure supply activities use the resources
of the internal supply chain within each node and of the dyadic supply relationships
that connect the nodes; this has been termed the infrastructure of the network
(Harland 1996a).
Supply network infrastructure will not be discussed here but instead the
structure of supply networks and strategies for them will be examined using theory
and practice.
Supply Network Structure

Nodes within supply networks contain those resources associated with supply.
Bounding supply networks in this way distinguishes them from interorganizational
networks, which become their network context; a supply network can be viewed as a
slice, or plane, through the interorganizational network. Supply network structures
differ according to their shape, where a firm is physically positioned in the network
(which relates to type of player) and the size, power, and number of players
contained within. There are other differences such as the international location of
players and the capacity of the network, though these are not examined here.
Supply Network Shape

Supply networks differ in shape, reflected in the breadth and length of the network,
as shown in Figure 4.
In this volume, Uzzi and Gillespie highlight that the overall structure of the
network within which a firm is embedded influences both the behavior of the firm
and of the network as a whole. This chapter deals more specifically with supply
network structure and its impact on supply performance.
Breadth o/the Supply Network

The trends in dyadic supply relationships described earlier, notably supply base
reduction, have resulted in the breadth of many supply networks decreasing.
Supply Network Strategy and Social Capital- 419
Network
breadth
Network length
....1 - - - - - - (number of levels or echelons) ------t~~
Figure 4. Supply network shape
Turnbull and Valla (1986) and Hakansson (1987) suggested that most inter-
organization networks are now more concentrated and structured.
There have also been changes in breadth of the downstream supply network
structure in many cases. Womack et al. (1990) reported Japanese automotive
distribution and retailing supply networks contained fewer dealers compared to
Western counterparts-about 1700 in Japan compared with 17600 in the U.S. in a
market only 2-3 times as large.
Operational performance benefits have been reported in narrower networks.
Closer, longer-term relationships present in Japanese distribution networks provided
ample social capital as they allowed more efficient sales and distribution and
captured important customer feedback on design and manufacture (Womack et al.
1990), transmitting this knowledge upstream to facilitate learning and improvement.
Easton and Quayle (1990) proposed that single-sourcing networks would be more
rigid and strong in that there would be dense flows of exchanges within them. Also
it would be easier to retain confidentiality in narrower networks.
However, it has been argued that there may be liabilities in narrower supply
network structures. Concentrating sourcing on a smaller number of suppliers can
increase risk and reduce learning from other networks (Sabel et al. 1987); Podolny
and Castellucci (this volume) discuss the negative effects of being involved with too
few others vis-a-vis learning. Easton and Quayle (1990) identified that narrower
networks reduce the ability to adapt to changes in the environment through
switching. Where uncertainty existed in the upstream supply market, Puto et al.
(1985) advocated mUltiple sourcing as an important strategy.
Therefore, it can be seen that, while there is support in the literature of a trend
toward narrower supply networks providing social capital benefits, there are
possible social liabilities. The relative merits of broad and narrow networks are
summarized in Table 2.
Table 2. Relative merits of broad vs. narrow networks

Advantages of Broad Networks Advantages of Narrow Networks
Adaptable to change Collaborative innovation
More switching opportunities Rigid and strong
Wider access to knowledge Dense flows of information
Hedge against uncertainty Higher confidentiality
Cost competitive Shared destiny
To date most of the empirical research on supply network breadth has been
performed in the automotive industry, and its particular operational characteristics
should be understood. Most automotive production is relatively high volume, low
variety, as reflected in the documented cases on supply chains and networks, such as
Toyota (Womack et al. 1990) and Nissan (Nishiguchi 1994). Product life cycles are
such that a reasonable amount of stability can be maintained in the supply networks,
which assists learning through collaborative working. Product design can be
modularized, enabling automotive manufacturers to focus operations on assembly,
and immediate suppliers on assembling modules, thereby causing breadth reduction
through a tiered structure. Most automotive production competes primarily on price
and quality; the previous discussion identified that having fewer, closer, longer-term
relationships is more suitable for collaborative improvement of costs and quality.
However, recent research has provided an initial indication that supply network
breadth varies across industries and across supply networks with different
performance criteria, which may well imply that different types of supply network
give rise to different forms of social capital and social liability. Harland (l997a)
reported the findings of structured interviews of 200 firms across a range of sectors
and at various positions in their supply network (see the later discussion on supply
network position). Regarding supply network breadth, questions were asked about
concentration of immediate suppliers and customers and about future plans for
deliberate change to the supply network structure. Using Ward's method, the
findings were clustered with other variables relating to the category of operation
(type of operation, sector, volume, variety, size, and purchase spend), the business
strategy of the firm, and supply strategy (whether one was perceived to exist, who
was driving it, and upstream and downstream operational requirements). The
following significant clusters were identified.
Tightly focused niche players (including manufacturers in yarns and fabrics,
electronic components, cigarettes, soft furnishings, office and shop furniture fixtures
and fittings, household goods, computers and communications, and toys/sports!
leisure goods) had highly concentrated upstream supply networks, with 1 to 3
immediate suppliers representing 80 percent of purchase spend, whilst the
downstream customer network was less concentrated, with 250-1000 customers
representing 80 percent of revenue. These supply networks were driven by delivery
speed, delivery reliability, and range requirements from the customer end of the
network. In this cluster, limiting suppliers to such a small number increased social
capital in learning how to respond quickly, through closer collaboration and
dependence.
Manufacturers of food ingredients, printing and packaging, workwear apparel,

and medical and surgical items with survival strategies were driven by service and
flexibility by their customer network. The upstream supply network was relatively
unconcentrated and in a state of continuous change, responding to informal, short
notice demands from the downstream customer network. The shape of the supply
networks of food processors, wood component manufacturers, and fashion apparel
manufacturers were almost a mirror image of the tight, focused niche players. These
networks were highly customer dominated, with 1-3 customers representing 80
percent of revenue. These customer dominated networks, while possibly deriving
social capital benefits in the downstream, customer end of the network, appeared to
give rise to social liability in upstream relationships, in terms of inefficient flexing to
meet frequent, informal and unplanned changes, as well as stress within the
operations of those relationships.
An example of social capital derived in broad networks was provided by the
electrical goods manufacturers surveyed. These were in broad networks both up and
downstream; they were driven by demands for constant innovation by customers. In
order to satisfy the requirements of constant change from the customer end of the
network, it suited these firms to retain breadth, and therefore choice and opportunity,
in the upstream part of the network.
Variation in network breadth across industries requires more investigation
before a taxonomy of network shapes can be developed. However, these initial
indications of substantial differences do alert attention to the need for more research
of supply network shape outside the automotive sector, particularly where different
operational performance requirements exist. These differences are likely to impact
nature and location of social capital and social liability in the total supply network.
Having examined issues of network breadth, the next section considers the
number of echelons or levels in a network that determine its length.
Length of the Supply Network

Nishiguchi (1994) reported that Toyota and other Japanese companies organized
their suppliers into hierarchies; first-tier or primary suppliers provided systems
rather than components. This significantly reduced the number of suppliers dealt
with on a direct supply basis, but reorganization of the echelons in itself did not
necessarily reduce the number of supply sources in the network in total, but instead
imposed more levels in the network. Tiering made each firm more dependent on its
suppliers, whereas traditional broader networks attempted to reduce dependency to
suppress prices through competition to aid switching. Increasing the number of
echelons necessarily placed intermediate relationships between any player and other
players in the network. This may have reduced their visibility of operations in the
network and may have impacted on learning resulting from direct interaction.
Therefore, the motivations for supply base reduction must be understood before
the likely impact on social capital and social liability can be assessed. If supply base
reduction, and hence a narrowing of the network, occurs through the removal of
inefficient competition and adversarial bargaining across suppliers, this may provide
social capital through more intense development of the same parts or services.
However, if supply base reduction occurs through tiering, and hence lengthening of
the supply network, the remaining relationships no longer transact for the same
goods and services but rather for modules or complete packages previously
assembled in house. In this situation it is possible that social liability may occur
through increasing the number of intermediaries involved in the operational flow.
Physical Position in the Supply Network and Type of Player

In supply networks for manufactured products the types of players contained in the
network relate to their physical position. At the upstream end of the network there
are raw material extractors and processors whose materials are manufactured by
component manufacturers; these components are assembled and then pass through
distribution channels to reach their destination with the end customer. The supply
network context for each of these players therefore varies according to their physical
position. Each type of player has its own different role to play in the network; for
example, most of product quality value is added in the upstream part of the network,
whereas range, service, and delivery required by the end customer is usually
satisfied by the downstream part of the network (Harland 1997b). Raw material
extractors and processors often serve diverse downstream networks; for example, a
foundry may serve a large range of industrial sectors including automotive,
construction, capital equipment manufacturing, and any other sector where steel is
used. These sectors may exhibit very different operational performance
requirements. Relating this back to the idea of focus and role in networks (Harland
1997b), upstream players may have to learn about a wide spread of sectors and their
requirements compared to players further downstream. It is not yet understood
whether social capital in supply networks is most evident between the players in the
network contained within one sector, rather than being derived through relationships
with the more generalist upstream players.
Most of the research to date in supply networks has been conducted in networks
for the provision of manufactured products. Service provision is conducted in very
different network shapes. Typically a service provider is surrounded by a cluster of
immediate suppliers of component services, people, and information used to provide
a service package. For example, a travel agent deals with tour operators, transport
providers such as airlines and railways, insurance brokers and currency providers,
each providing a service element that the agent assembles into a service package. To
date little research attention has been paid to service supply networks. Documented
examples, such as in health care (Harland 1996b), tend to concentrate on the supply
network for the tangible elements of the service package (in the case of health care
provision, all the equipment and consumables required to perform health care
services). As service networks appear to be shorter in length than most supply
networks, there are fewer intermediaries to act as barriers to the transfer of
knowledge and learning in the network; therefore, greater social capital may be
evident in more of the network than in long, complex manufacturing networks;
however, there is little evidence of empirical research to support this, as yet.
Size, Power, and Number of Players in the Supply Network

There are only a limited number of documented cases on supply networks and their
structures; interestingly they tend to be of large, powerful, high-volume operations
residing in the center of the total supply network-for example, Toyota (Womack,
Jones, and Roos 1990; Fruin 1992), Nike (Lorenzoni and Baden-Fuller 1995),
Benetton (Jarillo and Stevenson 1991), Coming (Lorenzoni and Baden-Fuller 1995),
Nissan (Nishiguchi 1994), SweFork (Dubois 1994), Volvo (Kinch 1992) and Apple
(Jarillo 1993; Lorenzoni and Baden-Fuller 1995). Each of these large players
controlled the upstream and downstream parts of the network, apparently benefiting
from the social capital they derived in these closely managed relationships.
Sako (1991) highlighted that while the Japanese manufacturing industry
employed just over twice the number of people compared to British manufacturing,
the industry had nearly four times the number of firms, with significantly more
smaller businesses. However, Mitsui (1990) drew attention to structural forms in
Japan other than the highly pyramidal vehicle networks which tended to have larger
first-tier suppliers dealing with smaller lower-tier subcontractors, again highlighting
that care should be taken in generalizing findings from limited cases in one sector.
Highly flexible and organized networks of very small engineering companies existed
in textiles, printing, construction, and information technology sectors where the
smaller firms were indispensable partners to the large companies. Currently there is
inadequate research to identify if there is any correlation between the number and
size of players in supply networks, their resulting performance, and evidence of
social capital or social liability. However, the observed differences indicate potential
for such research.
Research into the size, power and number of players in downstream parts of
supply networks has been provided in the marketing literature on channel
management. Stock and Lambert (1987) described how goods are physically
transported between the place they are produced and the place they are consumed
through channels. If direct contact is made between every manufacturer and every
customer, a large number of contacts occur. Selling through an intermediary reduces
the number of contact points made in the market and also improves service to the
customer as they deal with one point of contact. Channels have therefore evolved to
fulfill this intermediate role between manufacturer and consumer, reducing
transaction costs in the downstream part of the supply network; therefore, the
creation of new supply network players, such as an intermediary, may provide social
capital in some circumstances because of specialization of relationship management.
Not all channels have evolved in the same way. Different channel structures are
apparent in different markets and for different products. Stock and Lambert's (1987)
figure of alternative structures of channels of distribution of industrial goods is
shown in Figure 5.
These variations of structure showed differences in the number, ownership, and
role of intermediaries. It does not appear that anyone structure, or the presence of
absence of an intermediary, necessarily gives rise to the presence of social capital or
social liability. Rather, it appears that it is more an issue of 'horses for courses,'
where different structures are appropriate to different contextual situations.
The traditional marketing literature, based on an industrial economics
perspective of channel structure with the underlying belief that producers organize
channels to minimize costs, has been criticized (Gattoma 1978), leading to the
emergence of behavioral models. These suggest that channels and channel structures
Manufacturer Manufacturer
Manufacturer Manufacturer
Figure 5. Alternative channel structures (Stock and Lambert 1987)
evolve as a consequence of the actions of all the parties involved-namely

manufacturers, agents, distributors and retailers. Behavioral models lead to the
conclusion that channel behavior is based on power, conflict, leadership and
cooperation. Notable studies on power in marketing channels have been provided by
Etgar (1976), El-Ansary and Stern (1972), and Lusch and Brown (1982); these
studies concentrate on identifying the nature and sources of one channel member's
power over another. The impact of power on conflict in channels has been
identified, notably by Rosenburg and Stem (1971), Lusch (1976), Etgar (1979), and
Wilkinson (1981). It appears from these studies that abuse of power may give rise to
social liability in terms of detrimental impact of conflict within the network.
More recently, models of channels have combined both economic and
behavioral theories; these have been termed managerial models. Managerial models
ambitiously attempt to incorporate a wide variety of variables including
consideration of the channel as a political economy affected by individual firm
dynamics and the economic environment of the channel. Extension of managerial
models into strategic models (see, for example, Knee and Walters 1985) support the
belief that channel evolution is a consequence of strategic choices made by
manufacturers, distributors, agents and retailers within the channel.
While the research on supply network structure has been conducted by
academics in different fields, evidence has been provided that shows differentiation
of supply network structure according to industrial sectors, operational performance
priorities, and territory. The examples provided in the discussion appear to indicate
that some firms have taken a rational, strategic approach to design of supply
networks or part of them. The next section examines the current state of knowledge
on supply network strategy.
Supply Network Strategy

To date work in supply network strategy has taken a rational, normative approach to
strategy formulation following Chandlerian principles that structure follows strategy
(Harland and Wensley 1996). A supply strategy formulation approach provided by
Harland (1996b) extended the rational operations strategy formulation process
provided by Hayes and Wheelwright (1984) and Hill (1985) and is shown in Table
3. This approach is in contrast to the IMP approach to interorganizational network
Table 3. Extension of operations strategy elements to supply network strategy

Manufacturing Strategy Supply Network Strategy
Competitive Priority Of the operation: Of the end customer and each
supply network actor:
Price (cost) Price (cost)
Dependability Delivery speed
Flexibility Flexibility
Quality Product quality
Innovation
Range
Service quality
- reliability
- responsiveness
- competence
- access
- courtesy
- communication
- credibility
- security
- understandinglknowing
the customer
- tangibles
Structure Capacity-size, volume, timing Capacity-size, volume, timing

Facilities Supply network actors
configuration
Production equipment and systems Supply network facilities
configuration (e.g., fleet, buildings,
materials handling systems)
Internal or external sourcing Do-or-buy
Infrastructure Human resource policies Supply network human resource

policies
Quality systems Supply network quality systems
Production planning and control Supply network operations
planning and control
New product development New product or service
development
Organization Network organization
Performance measurement Performance measurement
strategy, which takes a more emergent, Mintzbergian approach, claiming that

players in the network 'merely cope' (Hakansson and Snehota 1995). The differences
in approach might be explained by the tangibility in supply networks that enables a
more planned approach to supply strategy. A further justification for the continuance
of rationality lies in the role of operations, and supply, in meeting the focused,
measurable range of operational competitive priorities.
Some firms appear to have designed and managed their supply networks
strategically and successfully, embracing the rational principles of focus and
segmentation (both upstream and downstream). For example, Benetton retained
control over color quality and design within its firm boundary but formed
relationships for most of manufacturing and retailing. Toyota focused its
manufacturing and distribution operations according to models; model type reflected
differences in competitive priorities.
However, their success may lie in their management of information in the
supply network and the resultant social capital in the form of flexibility and speed of
supply network learning, much of which was conducted through social exchange.
Benetton was the first fashion retailer in its market to be able to respond to
variations in demand within the season. To do this they ensured quick response to
information on what was being sold in the franchised retail outlets by connecting
point of sales information back to the factory. They changed their production
operations to leave dyeing as the last process to be conducted, thereby increasing
their speed to response. But their upstream supply network was quite distinct; it
consisted of a large number of contractors and sub-contractors who were connected
through familial and social relationships with Benetton employees. The sub-
contractors only worked, and were paid through informal means, when there was
demand. In this way Benetton did not have to take the risk or responsibility for the
upstream supply network and were able to use it flexibly and quickly when required.
Toyota's use of door-to-door selling maintained social contact with end customers.
Not only did this provide information flows about likely future demand, but, more
important, it provided access to design and quality feedback from customers over a
vehicle's life; this information was fed back to Toyota manufacturing and to the
upstream supply network (Y'fomack et al. 1990).
Benetton and Toyota appear to have acted as hubs in the supply network,
facilitating and coordinating quick flows of information, connecting the upstream
supply network with downstream demand. It is in this hub role that they differed to
their competitors. Burt (1992) identified the importance of access, timing and
referrals relating to information flows in networks. Burt identified that information
does not flow easily through interorganization networks; the network helps to screen
the large volume of information held by players, providing some focus on key pieces
of information. He highlighted that benefit-rich networks 1) have contacts
established in the places where useful bits of information are likely to air and 2)
provide a reliable flow of information to and from those places. It is this
establishment of contact with the right parts of the supply network, the filtering of
relevant information and focus and communication of critical pieces of information
that constitutes the core role of a supply network hub.
CONCLUSIONS
Supply networks are a special case of inter- and intraorganizational networks,
containing value-adding flows of materials, products, and services. To date there has
been limited research on supply networks; most of the research has concentrated on
automotive networks and attempts to generalize from this work should be viewed
with caution. Recent research has provided initial indications of the substantial
differences between supply networks relating to whether they contain flows of
products or services, which sector they are in and the operational competitive
priorities driving demand in the network. These differences are evidenced in
different network structures.
While the feasibility of interorganization network strategies and their
implementation is contended, it appears that the tangibility of supply networks may
enable a more planned, rational approach enabling implementation of supply
network strategies. Increasingly it appears that these strategies address
differentiation to meet different operational competitive priorities of end customers
groups.
A feature of successful interorganizational supply networks that warrants more
research attention is that of coordination, management, and communication of
critical information. Firms that appear to have strategically managed their supply
networks have created social capital by acting as hubs, capturing important
information from one part of the network and transmitting it to where it required.
• SECTlONY
Structure at the Firm Level

social capital and financial capital
Choosing Ties from the Inside of a Prism:
Egocentric Uncertainty and Status
in Venture Capital Markets
23
•
Joel M. Podolny
Fabrizio Castellucci
ABSTRACT
This chapter draws an analytical distinction between altercentric and egocentric
uncertainty. Altercentric uncertainty refers to the uncertainty that buyers face about
the product quality of a focal producer (ego). Egocentric uncertainty refers to the
uncertainty that the producer itself faces about the resource allocation decisions that
will result in a product that is regarded as high quality by buyers. This chapter then
argues that the value that a firm derives from its own status is positively related to
altercentric uncertainty and negatively related to egocentric uncertainty. That is,
status is valuable when buyers can use it as a signal of quality, but status is not
valuable when a producer does not know how to 'spend' its status to produce quality.
As a consequence, high status producers should seek out markets or market
segments where egocentric uncertainty is low. This argument and hypothesis are
tested in an examination of the venture capital markets.
INTRODUCTION
Sociologists have traditionally regarded networks as the plumbing of the market.
That is, networks are the channels or conduits through which 'market stuff flows.
'Market stuff encompasses information about exchange opportunities as well as the
actual goods, services, and payments that are transferred between buyer and seller.
Granovetter's (1995) work on job search is one example of research emphasizing the
flow of information about exchange opportunities (see also Flap and Boxman, this
volume), as is recent research on the role of interlocking directorates in the
dissemination of information and perceptions of legitimacy (Haunschild 1993; Davis
1991). Perhaps the most noted example of work stressing the role of networks in the
transfer of goods and payments is Burt's (1992) work on intersectoral flows in the
US economy. Recently, however, sociologists have started to perceive a second
function of network ties within the market. Network ties between a delimited set of
actors are not only relevant as pipes for the flow of market stuff, but as
informational cues from which others outside that set make inferences about
qualities possessed by those actors (also see Nooteboom, this volume). As an
empirical example, consider the syndicate relations formed among investment banks
when underwriting corporate securities. When one bank agrees to be in an
underwriting syndicate led by another, the former invariably must give the latter the
right to list its name in the 'tombstone advertisement' announcing the offering in
prominent trade publications, such as The Wall Street Journal. While a number of
scholars have discussed in detail how the location of a bank's name in the
advertisement reflects and determines its status (e.g., Hayes 1971; Eccles and Crane
1988; Podolny 1993), it is sufficient here to note that a low vertical placement is
indicative of low status and a high vertical placement is indicative of high status. In
accepting a subordinate position in the advertisement, a bank exhibits deference to
those banks that are listed higher. A syndicate relation between bank A and B thus
has a dual significance. On the one hand, it is a conduit of resources or information
passed between A and B. On the other, it is an informational cue affecting third
party perceptions of the relative quality of the services that A and B offer in the
market. Raised to the level of a metaphorical abstraction, this example illustrates
that networks are not only pipes carrying the stuff of the market; they are prisms,
splitting out and inducing differentiation among a set of economic actors.
An increasingly broad array of sociological work can be grouped into this
prismatic perspective. For example, Baum and Oliver (1992) show that potential
consumers regard a day care center as more legitimate if it possesses a tie to some
prominent organization in the community, such as a governmental agency or church
group. Similarly, Stuart, Hoang, and Hybels (1999) demonstrate that the investment
community is more receptive to biotechnology firms that possess an affiliation to a
prominent alliance partner (also see Stuart, this volume). At a theoretical level, both
pieces of work show that the affiliation with a high-status or trustworthy third party
signals quality. As a third example, Zuckerman (1999) illustrates that a firm's
pattern of exchange relations with financial analysts affects how favorably that firm
is evaluated by the investment community. In all of these pieces of work, the
presence or absence of an exchange relation between two actors constitutes an
informational cue that relevant third parties use in evaluating those actors.
Considered at the more macro-level, exchange relations between some subset of
market actors yield perceptual distinctions along some dimension, and these
perceptual distinctions inform the market decisions of some other subset.
In this chapter, we build on the research within this broad, prismatic perspective
that has employed the sociological conception of status to understand this
differentiation. Prior research (Podolny 1993; Podolny, Stuart, and Hannan 1996)
has argued that there exist a diverse array of ties among competitors on one side of a
market with the same differential significance that we noted in the tombstone
advertisement. For example, patterns of alliance formation, personnel transfer,
resource flows, and directed technological imitation between competitors often
Choosing Ties from the Inside of a Prism - 433
convey deference and, in so doing, induce a status ordering among those

competitors. A firm's position in this status ordering is a signal that the firm's
constituencies-such as potential buyers, possible alliance partners, and third party
observers (e.g, financial analysts, investors, media, governmental agencies)-utilize
in making inferences about the firm's quality (Podolny 1993; Han 1994).
If a firm's relevant constituencies rely on status to make inferences about the
quality of the firm's product or service, then the value of status should vary
inversely with those constituencies' ability to observe and assess quality directly.
For example, as a firm's buyers become more certain of the quality of the firm's
service or product, then the value of the firm's status should decline. Conversely, as
the buyers become less certain of the quality of service or product that the firm
offers, the value of the firm's status should rise. There is some evidence that the
value of a firm's status does hinge on the difficulty that the firm's constituencies
have in directly perceiving or evaluating quality (Podolny 1993; Podolny 1994;
Podolny, Stuart, and Hannan 1996).
However, constituency uncertainty regarding quality is not the only form of
uncertainty that a focal firm faces. As numerous organizational scholars have
emphasized (Thompson 1967; Cyert and March 1963; Stinchcombe 1990), a firm
also confronts uncertainty in making its own resource allocation decisions. In fact,
from the perspective of a focal firm, we can distinguish between two types of market
uncertainty. The first type is the uncertainty that the firm has regarding market
opportunities and the set of resource allocation decisions that will best enable the
producer to realize those market opportunities. We label this type of uncertainty
'egocentric uncertainty' because the focal firm is the actor who is uncertain. The
second type is the uncertainty that consumers or possibly other constituencies such
as potential alliance partners have about the quality of goods or services that the firm
presents to them in the market. That is, a producer of a good or a provider of a
service may be very aware of the quality of that which it provides on the market, but
potential exchange partners can still be unaware of this quality due to lack of
previous contact, knowledge of, or experience with the firm offering the good or
service. We label this type of uncertainty 'altercentric uncertainty' because the firm's
alters-the consumers or potential alliance partners-are the actors who are
uncertain.
We know that the value of a firm's status increases with altercentric uncertainty,
but how is a firm's status affected by egocentric uncertainty? In the next few pages,
we argue the following: whereas the value of status increases with altercentric
uncertainty, the value of status declines with egocentric uncertainty. Though we
justify this claim in some detail in the following pages, we can at this point layout
the broad contours of the argument: the greater the egocentric uncertainty
confronted by a firm, the less that the producer knows how to leverage its status to
acquire resources in the pursuit of superior resource allocation decisions. Stated in
terms of social capital, status manifests itself as a productive asset and hence as
corporate social capital to the extent that a producer's constituencies are uncertain
about the quality of the good that the producer brings to market. However, the more
that the producer faces uncertainty about what resource allocation decisions yield a
product that will be regarded as high quality, the less value the producer derives
from status.
If the value of status declines with egocentric uncertainty, high-status firms
should attempt to move away from markets or segments in which there is a high
amount of egocentric uncertainty and toward markets or segments in which there is
a low amount of egocentric uncertainty. After developing this argument in more
detail, we test it in an examination of the venture capital markets.
STATUS AND EGOCENTRIC UNCERTAINTY

To understand why there is a negative relation between a firm's status and
egocentric uncertainty, we begin with a simple observation: a firm's quality is a
function of the feasible set of resource allocation decisions that a frrm can make, and
the feasible set of resource allocation decisions is largely determined by the
exchange partners to which the firm has access. Concretely, a firm will not be able
to make a high quality product if it lacks access to suppliers of high quality labor or
raw materials.
The set of exchange partners to which a focal firm has access is, in turn, a
function of the firm's status. There are two reasons why this is so. First, a high-
status firm will generally have greater access because potential exchange partners
will be more likely to believe that the firm's output is above some quality threshold.
For example, the higher a focal firm's status, the more that potential alliance
partners will positively evaluate the prospects of an alliance with the focal firm.
Second, a high-status firm will generally have more access because at least some
potential exchange partners will value the association with status either as an end
itself or as a means toward some other end. Consider consulting firms recruiting
MBAs. High-status consulting firms will presumably find it easier or less costly than
low-status firms to hire the MBAs of their choice. The MBAs may either value the
association with status as an end in itself or value the high-status affiliation as a
resource that can be leveraged to pursue future opportunities.
Some research demonstrates that if a high-status firm enters into exchange
relations with lower-status actors, then the firm risks the dilution of its status
(Podolny and Phillips 1996). Our point is, therefore, not that the high-status firm is
completely unconstrained in its ability to choose exchange partners. Rather, our
point is simply that at any particular time, a high-status firm has more exchange
opportunities than a lower-status counterpart. The high-status firm will find it easier
to enter into exchange relations with other high-status actors. And while the high-
status firm may not wish to risk the dilution of its status by entering into exchange
relations with lower-status actors, the high-status firm at least has the opportunity to
pursue such relations if it believes that the tangible benefit from the exchange
outweighs the potential loss of status.
Since part of the value of status lies in the breadth of exchange relations to
which it grants a firm access, the value of status necessarily declines when a firm
does not know which exchange partners it should be pursuing. Or, stated in terms of
egocentric uncertainty, the higher the egocentric uncertainty confronted by a firm,
the lower the value of status.
Were the discussion of status and egocentric uncertainty to end here, then we
would be left with the following conclusion: high-status firms should make better
resource allocation decisions when there is low egocentric uncertainty, and they
should make approximately the same quality of resource allocation decisions in a
situation of high egocentric uncertainty. That is, the foregoing discussion suggests
no reason why high-status firms should actually make worse resource allocation
decisions than low- status firms in a condition of high egocentric uncertainty.
However, if we adopt a behavioral conception of the firm (Cyert and March
1963; see also Uzzi, this volume) and assume that firms engage in problemistic
rather than profit-maximizing search, then it is possible to conclude that low-status
firms may make better resource allocations decisions than their higher-status
counterparts in a condition of high egocentric uncertainty. The reason is the
following: status is a source of positional advantage in markets, and by definition,
positional advantage insulates a firm from the competitive pressure that would lead a
firm to engage in problemistic search and therefore learning (Barnett 1997). When
egocentric uncertainty is low, this learning disability is not problematic because
there is, by definition, little need to learn. However, as egocentric uncertainty
increases, this learning disability becomes much more problematic for the high-
status firm.
Such a learning disability does not mean that a high-status firm will obtain
fewer market rewards than a low-status competitor. The positional advantage
deriving from status may still outweigh or at least counterbalance any weakness in
capabilities resulting from inferior learning. Rather, the learning disability of status
only means that when egocentric uncertainty is high, there is a sharp trade-off
between the positional advantage afforded by status and the capability-based
advantage that can be afforded by learning. As a consequence of this trade-off, high-
status firms may make worse resource allocation decisions than their low-status
counterparts.
Yet regardless of whether high egocentric uncertainty causes high-status firms
to make worse resource allocation decisions than low-status firms, the fact remains:
the competitive advantage of status declines with egocentric uncertainty. As a result,
we expect that high-status firms will avoid market segments in which there is high
egocentric uncertainty and move toward market segments in which there is low
egocentric uncertainty. That is, high-status firms will avoid market segments or
niches in which there is considerable uncertainty about which exchange partners will
help yield a product or service valued by consumers, and they will actively seek out
niches and segments in which the opposite holds. For example, when Drexel
Burnham Lambert-a relatively low status-bank in the early 1980s-pioneered the
market for non-investment grade or ~unk' debt, the high-status banks were initially
reluctant to enter this market despite the extremely high margins to that went to
underwriters. Undoubtedly part of the reason that high-status banks avoided this
segment was that they simply did not wish to be perceived as affiliates of the low-
status corporate issuers that populated this market (Podolny 1994). But part of the
reason may also have been that they did not see how they could leverage their status
in a market in which there was such high uncertainty about which issuers should be
sought out as exchange partners.)
Or, as another example, consider cultural organizations, such as film studios. As

Hirsch (1972) observes, the markets in which cultural organizations find themselves
are characterized by relatively low altercentric uncertainty, but relatively high
egocentric uncertainty. That is, after a movie is made, a typical movie-goer has
relatively inexpensive ways to gauge whether she will like the movie. She can rely
on the opinion of either trusted critics or trusted friends; that is, a relatively small
number of opinion leaders can effectively bear the search costs for a large number of
movie-goers. However, before the movie is made, the studios face considerable
uncertainty as to what exchange relations with writers, actors, and directors will
yield a product valued by the consumers. While the movie-going public may not
perceive salient status distinctions among film studios, investors clearly do. Based
on the foregoing discussion, we would expect that the high status studios avoid film
genres or projects in which there is relatively high uncertainty about which
combination of artistic inputs will yield a product valued by the movie-going
pUblic. 2
Thus, our central hypothesis is the following: the higher a firm's status relative
to its competitors, the more that it will retreat from market segments in which there
is high egocentric uncertainty and seek out market segments in which there is low
egocentric uncertainty. To test this hypothesis, we now move to a consideration of
the empirical context for our analysis: the venture capital market. 3
VENTURE CAPITAL
In the market for venture capital, venture capitalists occupy a broker role between
investors and entrepreneurial companies in need of financial capital (see Freeman,
this volume for a more extensive discussion of the venture capital industry). A
venture capital frrm enacts this brokerage role by first raising money from the
investors and placing the money into a fund. The fund is organized as a partnership,
with the senior members of the venture capital firm serving as general partners and
the investors having the role of limited partners. The venture capital firm invests the
fund's money in entrepreneurial companies in exchange for an ownership stake. At
the end of a fixed period of time, usually 7 to 10 years, the fund is dissolved. The
venture capital firm takes a fraction of the proceeds, usually about 20%, and
distributes the remainder to the limited partners in proportion to their original
investment in the fund.
Venture capital firms are evaluated on their ability to generate high returns for
their limited partners. A venture capital firm may manage more than one fund at a
time, but the returns to the limited partner derive solely from the performance of the
fund in which she or he invested. The greater the limited partners' return on the
money that they invested in the fund, the easier it is for a venture capitalist to raise
money for subsequent funds.
For the purpose of this analysis, we regard investors as 'consumers,' venture
capital firms as 'producers,' and the allocation of funds to entrepreneurial firms as
the most critical resource allocation decision made by venture capital firms. While
this particular resource alIocation decision is the most critical, it is not the only
resource alIocation decision that venture capital firms make. In addition, venture
capital firms often assist the entrepreneurial firms they back in the selection and
recruitment of board members and critical managers, such as the CEO or the CFO.
More generally, venture capital firms must decide how to best allocate the time and
energy of their own employees across different activities.
From the perspective of the venture capitalist, the venture capital markets are
characterized by both high egocentric uncertainty and high altercentric uncertainty,
where investors constitute the critical 'alters.' Neither the investors nor the venture
capitalists have a high degree of certainty about the benefit that they will derive
from any particular investment. Even though some of the uncertainty associated with
investments in this market can be 'priced away' and thus transformed into risk, there
still remains a large amount of uncertainty in these markets that is difficult if not
impossible to price. Venture capital emerged as an institution largely because
traditional financing institutions were unwilling to invest in entrepreneurial firms
lacking collateral. The origins of venture capital are thus indicative of the generally
high level of egocentric and altercentric uncertainty in this particular market. In
personal conversations with venture capitalists, we have found them to be quite open
about the fact that there exist no reliable quantitative formulas for evaluating the risk
associated with their investments.4 The quality of a venture capital firm is a function
of its ability to make superior investment decisions in the context of this uncertainty.
The venture capital firm must learn to identify which attributes of an entrepreneurial
firm make that firm more likely to emerge as a success and go public or be acquired
at a high price. In making an assessment of an entrepreneurial firm's chances for
success, the venture capitalist will often consider a diverse range of factors. For
example, the venture capital firm must evaluate the entrepreneurial firm's
technology, the managerial ability of the firm's founders, the dynamics of the
market(s) in which the entrepreneurial firm hopes to compete, and the potential
responsiveness of the financial markets to a public offering of the entrepreneurial
firm's equity.
While it should now be clear that there is a high level of altercentric and
egocentric uncertainty in the venture capital markets, it is equally important to note
that there is variance in the level of egocentric uncertainty across investments. To
understand one basis for this variance, we must discuss in slightly more detail how
entrepreneurial firms raise capital from venture capitalists. Entrepreneurial firms
generally raise money in 'rounds.' That is, entrepreneurial companies do not receive
a continuous stream of payments from venture capitalists. Rather, they seek
financing over some discrete time period. At the end of the time period, the process
is then repeated. In some rounds, a firm may require more money than anyone
venture capital firm is willing to invest. In these cases, multiple firms may invest in
a given round.
Rounds can be categorized into stages. For the purpose of this analysis, we
divide rounds into three stages. Start-ups that do not have a viable product are
regarded as being in the first stage. At this stage, the entrepreneurial firm may have
little more than an idea or concept for a product. At least some of the financing in
this stage is referred to as 'seed' or 'start-up' financing. Once an entrepreneurial firm
proves that its product is viable, it may seek financing for the commercial
manufacturing and sales of its product. We denote such financing as second stage
financing. Second stage financing takes place after a company has started to produce
its product but typically before the company has become profitable. Finally, the
company enters third stage financing when it is profitable but is seeking capital for
further expansion. As a company progresses through the various stages, its value
increases. From the perspective of the venture capital firm, that is from the
perspective of an investor purchasing equity in a company, the highest returns are to
be found when a company that was financed in the early stage demonstrates a
promise of high future earnings and goes pUblic. Because there is less uncertainty in
the later stages, the returns from a late stage investment that goes public are
generally smaller than the returns from an early stage investment that goes public.
An entrepreneurial company does not necessarily begin to receive financing in a
seed or a start-up round, and a company need not go through all stages of financing
before being acquired or going public. For example, because they have high capital
needs in the product development stage, biotechnology firms will typically go public
long before they have a marketable product. Given this categorization of stages
above, biotechnology firms frequently go public before they conclude first stage
financing.
For our purposes, what is important about these stages is that they provide a
basis for categorizing investments in terms of the egocentric uncertainty faced by the
venture capitalist at the time of the resource allocation decision. The later the stage
of the investment, the less uncertainty that the venture capital firm confronts
regarding the outcome of an investment decision.
Whereas the altercentric uncertainty of investors may also vary with the stage of
company in which their money is invested, we believe that it is reasonable to assert
that the variance in altercentric uncertainty within the venture capital markets is less
than the variance in egocentric uncertainty. There are at least two reasons why this is
so. First, as noted above, investors do not typically make decisions to invest in
particular companies; rather, they make the decision to invest in a fund. While some
funds may be targeted to the investment of companies at a particular stage, most are
not. Therefore, an investors' decision to place money in one fund rather than another
is associated with less variance in uncertainty than a venture capitalist's decision to
invest the fund's financial resource in start-ups at one stage rather than another.
Second, investors invariably have only part of their financial resources in the
venture capital markets; they diversify their risk by including the equity of more
established firms in their investment portfolio, and they often invest in other
financial products. As a result, investors will generally devote less time and energy
than the venture capitalists to being knowledgeable participants in these markets.
The less knowledgeable one is about existing entrepreneurial firms in general, the
less that one's insight into the future performance of a firm can vary with the firm's
stage. Thus, whereas egocentric uncertainty will be strongly associated with the
stage of an investment, altercentric uncertainty is more constant across these stages.
While this distinction between the variance in egocentric and altercentric
uncertainty may seem to represent an analytical digression, it is critical to our
empirical analysis. Because egocentric uncertainty varies across market segments
more than altercentric uncertainty, we can now move from our general hypothesis in
the previous section to a more specific prediction in this particular context. To the
extent that high-status firms seek out market segments in which there is high
egocentric uncertainty, we expect the following: as a venture capitalist's status

increases, it should gradually shift its portfolio of investment from the early rounds
to the later rounds.
If altercentric uncertainty revealed considerable variance across segments, then
it would not be possible to develop a straightforward prediction. While status would
still be of less benefit in making resource allocation decisions in the early rounds,
status would potentially be more valuable as a signal. If the signaling value of status
were noteably higher in the early stages than the later stages, then there would be
counterveiling effects on a high-status venture capitalist's decision to invest in early
or late stages. On the one hand, the higher egocentric uncertainty would push the
high-status venture capitalist to invest in later stages, but the higher altercentric
uncertainty and thus signaling value of status would push the venture capitalist back
in the direction of early round investments. However, because altercentric
uncertainty displays little variance across market segments, it is possible to make the
straightforward prediction that an increase in status should lead a venture capitalist
to invest more of its financial resources in later stage start-ups.
Model
To test the hypothesis, we propose the following model:
m
R il+1 =BISil + LBkX k +O'i +1", +£il+l
k=2
where Ril+1 denotes the average stage of firm i's investments during year 1+1, Sit
signifies the status of venture capitalist i in year t, X2 through Xm represent a set of
control variables, 0'; indicates a firm-specific effect for venture capital firm i, and 1",
reflects a time-specific effect for year t. The model can be estimated using
conventional OLS techniques for the analysis of panel data. Our central hypothesis
is that BI > o.
DATA AND MEASUREMENT

To test this hypothesis, we obtained data from the Securities Oata Corporation's
(SOC) Venture Economics data base. The SOC collects information on numerous
financial markets and then sells this information to the various financial
communities. The SOC obtains the information in the Venture Economics data base
from both public and private sources. Venture capitalists use this data to obtain
benchmarks for their performance, and entrepreneurs use this data to better
understand venture capitalists' investment preferences.
Though the SOC has information on entrepeneurial start-ups dating back to the
early 1970s, data from the 1970s is extremely scant, leading us to believe that much
of the data from this time period is missing. Moreover, it was not until probably the
mid-1980s that venture capital emerged as a clear industry whose provision of
financial resources was regarded as a critical resource for entrepreneurial start-ups,
especially in the high technology sectors. Because we wish to isolate the egocentric
uncertainty that is due to variance in the stages of investments from the variance that
may be due to the maturity of the venture capital markets and because of our
concerns about the scantness of the data during the early period, we limit our
analysis to the time period between 1986 and 1996, inclusive.
In addition to deciding on an appropriate time window for the analysis, one also
needs to decide which actors are to be considered venture capitalists. The SOC
venture capital database includes information not only on the investments of venture
capital firms, but also information on the investments of key individuals and
organizations, such as universities, that make investments in entrepreneurial firms.
These individuals and organizations might often be prominent actors in the venture
capital community; in fact, the partners in venture capital firms wiIl not infrequently
make investments in entrepreneurial firms as private individuals apart from the firm
for which they are partner. These individuals and organizations should be able to
derive the same access benefits from status that a venture capital firm would derive,
and these individuals and organizations should therefore experience similar
pressures to focus their investments in particular segments. Therefore, it seemed
important not to exclude them from the popUlation of venture capitalists that make
investments in entrepreneurial start-ups. That being said, there are hundreds of
individuals and organizations reported in the data that may make only a couple
investments over a several year period, and it seemed unreasonable to consider these
transitory investors as venture capitalists with pretensions to having some status in
this industry. As a selection rule, we exclude all individuals and/or organizations
from the population of venture capital firms that make less than 10 investments over
the 5 year period prior to the year for which the status matrix was constructed. After
employing this selection rule, we obtained a population of 387 venture capitalists
over the 1986 to 1996 time period. These 387 venture capitalists yield 2386 firm-
year observations. The average number of years that a firm was in the data was
2386/387 =6.16 years.
Dependent Variable: Average Investment Stage

Measuring the average stage of a venture capitalist's investments in year t+ I is
straightforward. The SOC Venture Economics database categorizes investments into
the three investment stages discussed above. To reiterate, an entrepreneurial firm is
considered a first stage investment when it does not yet have a viable product. It is
identified as a second stage investment when it has a viable product but is not yet
able to profitably manufacture and distribute the product. Finally, the entrepreneurial
firm is categorized as a third stage investment when it is generating a profit. We
assign first stage investments a value of I, second stage investments a value of 2,
and third stage investments a value of 3. We then add up the values associated with
a firm's investments in a given year and divide by the total number of investments.
The range of this variable is accordingly between 1 and 3, inclusive.
ExplanatoryVariable: Status
To measure the status of the venture capital firms in year t, we construct a matrix
based on the joint involvement of venture capitalists in financing entrepreneural
start-ups. That is, we construct a matrix R, in which cell Rij, denotes the number of
times venture capitalist i and j jointly financed a start-up between years t and t-4,
inclusive.
Given R" we then calculate status scores based on Bonacich' s (1987) measure
(see the chapters by Doreian and by Han and Breiger, this volume for alternative
ways to measure status). A venture capitalist's status is a function of the number and
status of the firms with which it jointly finances start-ups; the status of these
financing partners is in tum a function of the number and status of their syndicate
partners, and so on. Particularly given that we include the number of deals in which
a venture capital firm has participated as a control variable in the analysis (see
below), this status measure reflects the extent to which a firm has financing partners
who are 'players.'
Conversations with those in the industry provide at least anecdotal support for
the use of this particular measure. Lower status venture capitalists express a strong
desire to be included on deals financed primarily by higher-status firms, and higher
status venture capitalists occassionally refuse to finance a venture if that venture is
receiving financing from a lower status venture capitalist. In effect, to be high-status
is to be an insider; to be low-status is to be an outsider, and joint financing
constitutes a symmetrical form of deference in which each venture capitalists
acknowledges the standing of the others.
Control Variables
In addition to status, we include two control variables in the analysis. First, we
include the number of investments made by the venture capitalist between years t
and t-4, inclusive. There are several reasons to include ths variable as a control
variable. First, to the extent that we find evidence consistent with our hypothesis, we
would like to disentangle the effects of relationally-based status from simply the
volume of activity in which a venture capitalist is engaged. That is, we would like to
distinguish the effect of status from the effect of size. There is a second reason for
including this variable. The number of deals in which a firm is engaged is a measure
of a firm's experience, and to the extent that a firm learns from experience, a firm' s
total amount of investments should reduce the uncertainty that it confronts in
making subsequent investment decisions. Since there is more egocentric uncertainty
and thus more to learn about investing in the early rounds than in the late rounds, we
would expect that such learning from experience would have a greater effect on a
firm's ability to make judicious early stage investments than judicious late stage
investments. In effect, whereas status should push firms to make more late stage
investments, experience should push firms to engage in a high proportion of early
round investments.
Of course, it seems reasonable to expect that there is a strong positive
correlation between status and the number of investments made by a firm. In fact, as
indicated in Table 1, the correlation is .79 if one includes both within-firm and
cross-firm variance and .66 if one includes only within-firm variance. 5 Thus, in
arguing that one of these variables has a particular effect, we are implicitly assuming
that the value of the other variable is held constant.
A second control variable is the number of funds from which the venture capital
firm makes investments in year t. As noted above, a venture capital firm can have
more than one fund from which it makes investments at any particular time. There
Table 1. Descriptive statistics for explanatory variables
Variables Mean Standard Correlations

Deviation (1) (2)
(l) Status, t to t-4 0.685(0) 0.971 (0.228)
(2) Number of Deals, t to t-4 91.876(0) 111.600(45.982) 0.793(0.645)
(3) Number of funds in year t 2.198(0) 1.683 (0.754) 0.608(0.233) 0.720(0.370)
Values outside of parentheses are based on within-firm and between firm-variance; values within
parentheses are based on within-firm variance only.
are three reasons for including this control variable. The first two reasons are similar
to the reasons for including the number of deals over a five-year time interval as a
control variable. As with number of deals, number of funds is likely to be an
indicator of both size and experience. To the extent that number of deals is an
indicator of size, we wish to distinguish the status effect from the size effect. To the
extent that number of funds is an indicator of experience, number of funds should
have a negative effect on the average stage of a venture capital firm's investment.
The third reason for including this control variable is to rule out an alternative
explanation for our hypothesized effect of status. We have argued that high-status
firms should be more likely to make late-round investments because status is most
valuable in those market segments in which a firm knows how to best leverage its
status in forming exchange relations. However, if higher-status firms generally have
access to more financial resources than lower-status firms, high status firms may be
forced to make more late-stage investments because-with the noteable exception of
biotechnology start-ups-early stage investments generally have minimal capital
requirements. For example, the capital requirements for a first-stage software
company, which is engaged in the development of a product, are typically much less
than the capital requirements of a second- or third-stage software company, which
must either put together or at least outsource manufacturing and distribution
functions. Put simply, the high-status firms may simply have too much money to
invest in early stage firms. We thus include number of funds as a measure of the
amount of financial capital to which a venture capital firm has access. If the effect of
status is spuriously related to access to financial capital, then the effect of status
should disappear when number of funds is included in the analysis. This
interpretation of number of funds as a proxy for access to financial capital suggests
that the variables should have a positive effect on average stage of investment. On
the other hand, to the extent that number of funds is simply another indicator of
market experience, it should have a negative effect. Since we include this variable
only as a control, we do not hypothesize as to whether the negative or positive effect
should be stronger.
Finally, in addition to these two control variables, we include firm-specific
effects, O'i, and period-specific effects, 't'" to denote effects that should be either firm-
specific or year-specific. These effects simply control for unobserved heterogeneity
that would be common either to all of the observations drawn from a particular firm
or all of the observations drawn from a particular year.
Table 2. Determinants of average investment round
Variables (I) (2) (3) (4)

Status 0.054 (0.031)* 0.098(0.042)** 0.098(0.043)**
Number of dealS/IOO, 0.016(0.017) -0.033(0.020) -0.016(0.021)
t through t-4
Number of funds -0.026(0.010)** -0.026**(0.010)
R2 for within-firm variance .048 .050 0.049 0.052
N= 387; NT = 2386
* = p < 0.1; ** = p<0.05
Fixed-effects for fnms and years are not reported.
ANALYSIS
Table 1 presents descriptive statistics for the explanatory variables in the analysis,
and Table 2 presents the regression analyses. The reason that we report correlations
based on within-unit variance only is that the inclusion of fixed-effects for firms
essentially removes cross-firm variance when estimating the effects of the other
coefficients. Accordingly, in trying to interpret the substantive effect of the
coefficients reported in Table 2, it is helpful to have the descriptive statistics based
on within-firm variance only.
As we see in Table 2, the effect of status is positive and statistically significant,6
as hypothesized. Interestingly, the effect becomes stronger-almost doubling in
magnitude-when the number of deals is included as a control variable. There is
thus evidence of at least a mild supression effect; the status effect is weaker when
the measures of learning through experience are not included in the model. This
suppression effect lends at least tentative support to the view that higher-status firms
are on average worse learners.
Comparing the effects of the two control variables, we see that number of funds
making investments in year t has a stronger effect than number of deals undertaken
between years t and t-4. Moreover, the negative effect for number of funds clearly
indicates that scarcity of early-stage investments does not seem to be an issue for the
firms with more financial resources.
The positive effect for status is particular noteworthy in light of the fact that
each model includes firm-specific effects. Because these firm-specific effects
remove all cross-firm variance, the regression actually shows us that within-firm
changes in status are accompanied by within-firm shifts in the average stage of
investment.
CONCLUSION
In this chapter, we have tried to distinguish between two types of market
uncertainty-the altercentric uncertainty that a producer's constituencies confront
when evaluating the producer as a potential exchange partner and the egocentric
uncertainty that the producer confronts in making its own resource allocation
decisions. Previous research has demonstrated that the value of status increases with
altercentric uncertainty.
While this work does not provide direct evidence that social capital induced by
status necessarily declines with egocentric uncertainty, it does show that the higher a
producer's status, the more it seeks to avoid market segments that would be
associated with high egocentric uncertainty.
A related paper (Podolny and Feldman 1997) takes the next step and analyzes
how variance in egocentric uncertainty affects the ability of high-status firms to
make resource allocations that are superior to those made by low-status firms. This
related paper investigates the outcomes of the investment decisions made by the
venture capitalists. Specifically, it considers the investments of venture capitalists at
each stage and examines how a firm's status relates to the likelihood that the
entrepreneurial firms in which it invests go public. The paper finds that the higher a
venture capital firm's status, the higher the likelihood that its third-stage investments
go public, but the lower the likelihood that its first-stage investments go public. In
other words, in the latest stage, when egocentric uncertainty is low, the higher-status
firm make resource allocation decisions that are superior to those made by their
lower-status counterparts. But in the earliest stage, when egocentric uncertainty is
high, the higher-status firms apparently make resource allocation decisions that are
worse than their lower-status counterparts. By demonstrating that status yields the
greatest social capital when egocentric uncertainty is low and that status seems to
create social liability, as reflected by a learning disability, when egocentric
uncertainty is high, the findings in Podolny and Feldman (1997) complement and
reinforce the findings in this one.
Before concluding, we would like to briefly note some implications of this
chapter for future research. First, because new markets or new market segments are
likely to be characterized by especially high egocentric uncertainty, the results of
this chapter suggest that high-status producers will generally be more conservative
than their lower-status counterparts in moving into new markets or market segments
Earlier we briefly alluded to some evidence consistent with this conjecture: high-
status investment banks were extremely reluctant to enter the junk market despite
the apparent profitability of doing so. However, more systematic evidence is
obviously necessary in order to validate this conjecture.
More broadly, the results of this chapter suggest a new direction for research
that falls into the prismatic perspective on market networks. At the beginning of the
chapter, we invoked the metaphor of a prism to characterize work that highlights
how the pattern of ties among a set of actors on one side of a market induces a
differentiated social structure. A number of scholars (Baum and Oliver 1992; Stuart,
Hoang, and Hybels 1999; Zuckerman 1999) have illustrated how the affiliations and
positions that arise as part of this structure affect the decisions of those seeking to
evaluate and discriminate between the actors within this structure. However, we
know much less about how an actor's location within the prism affects the actor's
own ability to make decisions. Work following from the behaviorial theory of the
firm (e.g., March 1988) has identified a number of dysfunctional learning behaviors
in which firms engage. Moreover, there seems good reason to believe that decision-
making in some organizations is more political than decision-making in others,
where the term 'political' refers to the realization of individual self-interest through
bargaining and informal social exchange within the organization. Given variance
across organizations in the efficaciousness of learning and the political character of

decisions, it is valuable to ask how an organization's location within the prism
affects the decision-making process. For example, how does the presence or absence
of ties to legitimate actors affect the nature of decision-making within
organizations? In pursuing research along these lines, sociologists will provide
further insight into the functionalities and dysfunctionalities of the social structural
underpinnings of the market.
We wish to thank Michael Hannan and Ezra Zuckerman for helpful conunents on this chapter.
NOTES
I. Following Knight (1921), it is important to distinguish the risk associated with the underwriting of
securities that have a higher default rate than investment-grade debt from the uncertainty associated with
underwriting in a new type of security market. Investment banks can in a rather formulaic way set the
prices of the bonds to reflect the risk associated with underwriting and owning bonds with a higher
default rate. However, because the junk market was a new securities market in the 1980s and the demand
for and performance of bonds in these markets was not well-understood, investment banks also
confronted some uncertainty in these markets that could not be simply 'price away' as risk.
2. Of course, in conducing an empirical test of this hypothesis, it is essential to control for the fact that
higher-status firms will typically have larger reserves and thus are bener able to take on high-risk
projects.
3. Importantly, a high-status firm need not be aware of its learning disability in order for it to act on the
basis of that disability. Simple trial-and-error in segments with high and low egocentric uncertainty
should lead the high status firms to those segments in which their learning disadvantage is less
problematic.
4. Along these lines, it is worth noting that an interesting distinction between the way that economists
and sociologists study markets is that economists will typically seek to expand the set of market decisions
that can be framed in terms of risks, whereas sociologists will generally seek to expand the set of market
decisions that can be framed in terms of uncertainty.
5. Arguably, the within-firm correlation is more relevant to the analysis because the inclusion of fixed-
effects for firms removes cross-firm variance from the analysis.
6. Statements on statistical significance refer to a .I level.
Corporate Social Capital and the
Cost of Financial Capital:
•
An Embeddedness Approach
24
Brian Uzzi
James J. Gillespie
ABSTRACT
Using a structural embeddedness approach, we present argument and evidence on
the ways social capital affects the operation of financial capital markets in the
context of the small business loan market. We posit that the quality of a relationship
between a bank and a corporate borrower, as well as the network structure of ties
between the borrower and its bank(s) influences the cost of capital firms pay on their
loans. Specifically we examine two dimensions of structural embeddedness at the
dyad level and two at the network level. At the dyad level of analysis, we find that
the duration of the relationship and relationship multiplexity are associated with a
lower cost of capital (i.e., paying lower interest rates). At the network level, we find
that firms that have ego-networks composed of a mix of embedded and arm's-length
ties obtain a lower cost of capital then firms with either a ego network composed of
arm's-length ties or an ego-network composed of only embedded ties. We find no
effect for simple ego-network size on the cost of capital. The implications of our
embeddedness perspective on corporate social capital are discussed.
INTRODUCTION
Economic sociology is concerned with questions of how organizations acquire
resources and the mechanisms by which social structure influences the allocation of
resources in a market. In a capitalist economy, the key resource is financial capital,
and consequently, the connection between a firm and its lender(s) is equivalent to an
organizational umbilical cord. The purpose of this chapter is to apply a sociological
approach to the study of financial market intermediation with special attention to
how social structure affects an organization's cost of borrowed capital.
Corporate Social Capital and the Cost of Financial Capital - 447
Currently, financial economic theory has developed a widely used model to

explain how the market for capital operates and how the cost of capital is determined
for organizations. The theory predicts that, in a frictionless capital market,
individuals or firms with a positive net present value of investment opportunities
will always have access to funds and that the riskiness of the borrower determines
the cost of capital (Mintz and Schwartz 1985; Mizruchi and Stearns 1994; Petersen
and Rajan 1994). This approach focuses on the use of objective financial criteria in
evaluating the creditworthiness of a borrower and on how financial market
characteristics affect lending practices. The theory posits that firm level financial
statistics adequately measure the organization's ability to service debt through future
revenues, liquidity of assets, or both. Thus, high performing firms or firms with high
liquidity are prime candidates for receiving loans at competitive prices. In addition,
the age and size of the organization are also viewed as important measures of the
firm's ability to bear credit. Old and large firms are expected to receive better
financing terms because they have positive reputations and a more diversified
portfolio of assets.
The financial economic approach also focuses on how financial market
characteristics influence lenders' bargaining power and historical practices. A
significant characteristic of the market is the level of bank concentration in a given
region: higher bank concentration is thought to increase the cost of capital to
borrowers because the decrease in competition among banks can permit each to
bargain aggressively for a premium (Petersen and Rajan 1994). Also regional
characteristics are important. Certain regions may have structurally embedded
financing and production cultures that increase access to resources relative to other
geographic regions (Romo and Schwartz 1995). A California software firm may find
it easier to acquire capital than a Mississippi software firm. Industries vary in their
growth rate, which can provide organizations in these industries with an advantage
in acquiring capital. A biotechnology firm may have more favorable cost of capital
than a firm in a declining heavy manufacturing industry (Powell et al. 1996).
While evidence in support of this theory has been accumulating, particularly at
the level of large banks and large corporations (Uzzi and Gillespie 1998), a recent
critique argues that it fails to account for how social structure (e.g., lending
relationships, discrimination, and bias) affects the cost of that capital (Mintz and
Schwartz 1985; Podolny 1993; Petersen and Rajan 1994; Abolafia 1996). For
example, financial economics generally regards relationships as peripheral to the
operation of capital markets or as adding inefficiencies to the system (Blackwell and
Santomero 1982; Baker 1990; Podolny 1993; Mizruchi and Stearns 1994). Yet,
research has long recognized that relationships are an integral part of the banking
system and highly valued by entrepreneurs and bankers (Baker 1990; Hoshi,
Kashyap, and Scharfstein 1990; MacKie-Mason 1990; Diamond 1984). For
example, consistent with the argument that relationships matter, Hoshi, Kashyap,
and Scharfstein (1991) found that long-term ties between Japanese firms and banks
was associated with fewer liquidity constraints on a firm's investments and a greater
capacity to make investments when financially distressed. Petersen and Rajan
(1994) found that the number of banks from which a firm borrows and the number
of services the firm uses at the bank are associated with a lower cost of borrowing.
They concluded that ties between banks and borrowers can increase information
flow and the bank's control over the firm's actions, thereby addressing problems of
adverse selection (e.g., high interest rates attracting riskier borrowers) and moral
hazard (e.g., applicants choosing higher risk investments).
This chapter uses an embedded ness approach to extend the above work and
examine how social structure affects an organization's cost of capital, where cost of
capital is reflected in the interest rate on long-term financing. The interest rate on
long term financing is an appropriate measure of the cost of capital because 1) it is
clearly measurable and 2) it is the cost index most widely used in financial
reporting. Specifically, we examine the relationship between social structure and
lending practices, with particular emphasis on the quality and structure of ties
between small and medium sized organizations and their banks. Small and medium
sized organizations have annual sales up to 500 million, with the medium size in
terms of annual sales being ten million.
The embeddedness approach gives social structure a central role in explaining
lending practices and outcomes (Granovetter 1985; Portes and Sensenbrenner 1993;
Uzzi 1996a, 1997a, 1997b). It explicates how the substance of ties, as well as the
ego-network of ties in which an organization is situated, affects exchange between
organizations. The decisive factor is that particular types of social ties can mitigate
opportunism, increase resource pooling, and motivate actors to seek Pareto superior
outcomes rather than selfish gains. This theoretical approach extends previous work
in economics and sociology by developing more fully the social mechanisms by
which relationships benefit the firm and by furnishing more exacting measures of
embeddedness than the current literature.
Integrally related to our embedded ness arguments is the role of social capital in
the creation of financial capital (Gabbay 1995, 1997). 'Unlike other forms of capital,
social capital inheres in the structure of relations between actors and among actors.
It is not lodged either in the actors themselves or in the physical implements of
production' (Coleman 1988: 98). Social capital consists of the social relationships
between actors, and it importantly affects the operation of financial capital markets.
Banks are obviously repositories of financial capital. In addition, much like venture
capital firms, they are sources of experience, information, and personal contacts for
young firms (Freeman, this volume; Podolny and Castellucci, this volume). Banks
gain social capital by bridging structural holes and disconnects in the social structure
of the financial market, principally serving to connect savers with borrowers. As this
chapter shows, loan interest rates are partly a function of the strength of the
relationship (i.e., the degree of existing social capital) between a bank and a
potential borrower and the architecture of the firm's network of bank ties.
STRUCTURAL EMBEDDEDNESS: THEORETICAL ARGUMENTS

The structural embeddedness approach extends the work of classical sociological
theory on the economy and combines it with organization and social network theory
(White 1981; Granovetter 1985; Powell 1990; Portes and Sensenbrenner 1993;
Romo and Schwartz 1995; Uzzi 1996a). The basic argument is that the nature of
relationships between and among firms, as well as the overall structure of the ego-
network within which the firm is embedded, influences individual firm behavior and
the behavior of the network in its entirety. The type of network in which an
organization is situated defines its potential store of strategic opportunities, while the
quality of its relationships with other actors demarcates its capacity to access and
implement those opportunities.
The embeddedness approach assumes that actors' interest and motives are
variable and follow predictably from social structural parameters (Granovetter
1985). These differences in the microbehavioral foundations of embeddedness and
the macro structural conditions of exchange are what distinguish the logic of
embeddedness from other approaches (see Uzzi 1997a: 61). A key feature is that
actors operate under what has been referred to as the 'logic of embeddedness'
because ongoing social ties shape actors' expectation, motives, and decision-making
processes in ways that differ from the logic of market behavior (Portes and
Sensenbrenner 1993). According to this logic, actors use heuristic decision rules
rather than intensive calculation to make decisions, and they aim to cultivate
cooperative ties rather than narrowly pursue self-interest.
Structural embeddedness refers to the concrete social ties between and among
actors and focuses on material exchanges of resources and information as the basis
of the exchange. The argument posits that different structural conditions set in
motion either self-interested or cooperative interests and motives among banks and
borrowers, which in tum affect the cost and availability of capital. In addition, our
arguments recognize that social structure can either facilitate (social capital) or
derail (social liability) economic action. Therefore, 'social capital should be treated
as a context-dependent concept calling for the definition of the conditions in which
it has productive outcomes' (Gabbay 1997: 13-14). In an extreme case of
overembeddedness, social structure can be more constraining than beneficial and
constitute 'social liability' (Gabbay and Leenders, this volume).
We begin by discussing four dimensions of structural embeddedness:
relationship duration, relationship multiplexity, ego-network size, and network
coupling. These four dimensions all affect whether a given configuration of ties
turns into social capital or into social liability for to corporate borrower. Working at
the levels of dyads and ego-networks, we then examine how these four dimensions
of structural embeddedness affect the pricing of loans to small business
entrepreneurs. Important dimensions at the dyadic level are relationship duration and
relationship multiplexity (Coleman 1988; Baker 1990; Podolny 1994; Uzzi 1996a).
Important dimensions at the network level include ego-network size and ego-
network coupling (Baker 1990; Burt 1992; Uzzi 1996a), with the unit of analysis
being the set of ties between a firm and one or more banks.
Length of Bank-Firm Relationship

Relationship duration refers to the elapsed time in a relationship from the point of its
inception. According to financial theory, a firm's age could provide prospective
lenders with a gauge of the its ability to service debt by providing a record of the
firm's creditworthiness with past employees, suppliers, and lenders (Blackwell and
Santomero 1982). In contrast, recent arguments hold that the information learned
about a borrower through a long-term relationship may include information about a
borrower's creditworthiness that is not contained in past dealings with others. Of
particular importance is that the information may not be readily transferred to new
banks that have different lending policies, practices, or experiences with borrowers.
Thus, the longer a relationship between a borrower and a bank, the more likely it is
that the bank will view the borrower as creditworthy relative to other borrowers
(Petersen and Rajan 1994). Similarly, if a borrower has had only short-term
relationships with many different lenders, banks could interpret that as a signal of
credit unworthiness.
The embedded ness approach concurs with the insights of financial economics
but argues that a relationship is not only the source of specific information about an
exchange partner, but is the source of unique resources that would not be generated
in the absence of the relationship. 'Social capital inheres in the structure of relations
between actors and among actors. It is not lodged either in the actors themselves or
in physical implements of production' (Coleman 1988: 98). For example, several
studies of various types of exchange relationships have shown that ongoing
interaction fosters trust between exchange partners as they learn one another's
expectations and values (Coleman 1988). As trust accrues, the resources that were
dedicated to monitoring an exchange partner can be redeployed, increasing the value
of the tie. In a study of interfirm relationships, Macauley (1963) found that, the
longer two firms transacted with one another, the less detailed were their contracts
and the greater the organizations' ability to work out problems of transacting. Baker
(1990) reported that long-term relationships between investment bankers, and
investment bankers and their corporate clients, permits both parties to put faith in the
contractual pledges of other parties. Repeated interaction has also been found to
increase liking and the formation of 'business friendships' which can lower the
likelihood of opportunistic behavior and increase the search for mutually beneficial
outcomes (Homans 1950; Batson 1990). Baker (1990) showed that investment
bankers are more inclined to look for Pareto improved outcomes when business
friendships have developed, even when immediate or long-term payoffs are not
apparent or guaranteed. The important outcome of these processes is that ongoing
ties can lower costs, as well as the threat of opportunistic behavior by either
exchange partner, because both parties are motivated to preserve the tie and to first
look for Pareto-improved rather than self-interested outcomes.
In the case of banks and borrowers, this may mean that the bank can lower its
monitoring costs and contract writing costs for long-term borrowers, and firms with
ties can acquire cost of capital advantages relative to those lacking enduring social
ties. Indeed, in our conversations with bankers, the motivational benefits of
relationships are viewed as more important than the informational benefits because,
in many cases, more than enough information on the credit history of the firm and
the entrepreneur can be readily accessed through credit raters such as TRW (Uzzi
and Gillespie 1998).
Multiplexity of Bank-Firm Relationship

Another aspect of dyadic relationships that is important for understanding exchange
dynamics is the degree of multiplexity. Multiplex ties are relationships in which
persons are linked by more than one type of role (e.g., buyer and seller, business
partner, friend, etc.). Coleman (1988) argued that multiplexity increases the overall
level of resources available to actors because resources from one dimension of the
relationship can be appropriated for use in others. Multiplex ties develop when
transacting partners can enact a set of relationships in addition to the immediate
relationship of buyer and seller. In the context of banking-borrower relationships,
multiplex ties are likely to form when the borrower relies on the bank for multiple
services that span the roles of borrower and lender. Typically, these broader
relationships include financial planning, personal credit card issuing, retirement
planning, pension or payroll account services, lock boxes, letters of credit, etc. The
presence of multiplex ties of this form constitute a type of social capital that should
lower the cost of financial capital because resources from one dimension of a
banking relationship can support other dimensions either through direct resource
allocations or by expanding the possibility for finding compromise solutions to
gaining credit. For example, in return for a lower interest rate or access to more
credit -the primary resource desired by the firm-a business may use a bank's new
service division or expand its use of services already offered by the bank.
Consistent with this argument, there is evidence that when a new relationship is
added to a multiplex tie, the new ties rely on self-enforcement rather than external
constraints to manage interdependence (Gimeno and Woo 1996; Nooteboom, this
volume). This process has the effect of building interpersonal trust in multiple
contexts and roles (e.g., norms of reciprocity as benefactor and as beneficiary). For
example, research on automaker-supplier relationships suggests that, as US parts
suppliers and Big 3 automakers moved towards more cooperative relationships, the
level of contractual oversight over new relationships (e.g., supplier as co-designer or
investor) decreased (Helper 1990; Dyer 1997). Building on Coleman's (1988) initial
insight, Uzzi (1997a) argued that multiplex ties build redundancies that reinforce
relationships and reduce the risks associated with exchanges. In risky situations,
multiplex ties enable resource pooling and adaptation to random events by creating
or increasing the level of slack resources in the relationship. This increases the
likelihood of risk taking and investment on the part of both exchange partners.
These arguments suggest two hypotheses:
HI: The duration of the relationships between the borrowing organization and
banks is inversely associated with the organization's cost of capital.
H2: The degree of multiplexity in the relationships between the borrowing
organization and banks is inversely associated with the organization's cost of
capital.
Structural embeddedness also operates at the network level of analysis. Two key
measures are network size and network coupling. The logic behind these effects is
that an organization's overall ego-network of ties affects the value of each dyadic
relationship possessed by the firm.
Network Size
Network size measures the quantity of ties possessed by an actor. Several
perspectives argue that a large network of contacts is beneficial to an organization
because it increases the organization'S bargaining power and access to alternatives.
Transaction cost economics predicts that firms increase credit availability and lower
the cost of capital by maintaining many ties to many financial institutions. From the
perspective of transaction costs theory, the more trading partners a firm possesses,
the greater its probability of finding a prospective lender and the greater its
bargaining power vis-a-vis each bank (Milgrom and Roberts 1996). From the
perspective of resource dependence theory, a large ego-network of trading partners
should lower the power asymmetry between financial institutions and corporate
borrowers (Mintz and Schwartz 1985).
While we agree that a large ego-network of contacts may provide more
opportunities to acquire capital for the reasons outlined above, we predict that it will
have a negative rather than a positive effect on the cost of capital, particularly under
the credit rationing conditions that small businesses typically face. Our argument is
that large ego-networks work against the building of close relationships between a
borrower and a lender for several reasons, and thus lower the bank's incentive to
offer attractive rates. First, large ego-networks limit interaction because time and
resources are spread across a large set of partners. This reduces opportunities for
repeated interactions that can cause arm's length relationships to blossom into
business friendships (Granovetter 1993). Similarly, small ego-networks signal to
network members that the organization has enacted a strategy of cooperative
exchange and problem solving by consciously restricting its bargaining alternatives
(Kollock 1993; Jackson and Wolinsky 1996). If the number of banks that a firm uses
is small, then it is likely to attempt to maintain a close tie with its bank in order to
support the continuity of the relationship (Baker 1990). As Leenders (1995b) has
noted, this dependence cuts both ways because there is a mutual interdependence
between individual actors and network structure: Firms depend on banks as a key
source of financial capital but banks depend on firms to provide a market for the
purchase of their capital. We expect this effect to be strong in a well developed
banking market like the US because lenders use their knowledge of the going rate of
capital to bargain aggressively with borrowers who are shopping around since they
know that other lenders will also bargain aggressively to maintain their spreads.
Consistent with this argument, several studies have shown that firms with large
ego-networks were more likely to play their partners against one another (i.e.,
whipsawing) in an opportunistic manner (Helper 1990; Dyer 1997). Second, a large
network reduces the economies of time advantages typically found in the close
relationships that follow from small networks (Freeman, this volume). Smitka
(1991) showed that the development time of new models, as well as speed to market
was higher for Japanese automakers than for US auto makers partly because the
smaller contractor networks of the Japanese firms enabled tacit knowledge to
develop, which was crucial for faster decision making. In the context of our study,
we anticipate that small businesses with larger ego-networks of banks will pay
higher interest rates on their loans.
Network Coupling
Our last argument relating ego-network structure to the cost of capital concerns the
effects of the portfolio structure of an organization's network of ties. An ego-
network's portfolio structure differs from ego-network size in that networks of like
size can have a dissimilar portfolio structure. An ego-network can have a dispersed,
consolidated, or mixed-mode structure (Baker 1990; Uzzi 1996a). A dispersed
portfolio structure occurs when an organization spreads its banking business out
among many banks in small parcels. For example, the organization may use six
banks, each of which gets about fifteen percent of the organization's business. This
would represent a disbursed structure. Conversely, an organization with the same
overall size of banking business could use one bank for a sole-source relationship, or
it could give one bank 95 percent of its business and another bank the remaining five
percent. This structure would represent a consolidated structure. In between the
dispersed and consolidated structures is a mixed or dual mode structure, wherein the
organization sends the lion share of its business to one or two banks and then
distributes the remaining share to two or three banks.
An organization with a consolidated portfolio gains the benefits of close ties, yet
runs the risk of becoming insulated from new and novel information that is
circulating outside its network (Uzzi 1997a). For example, firms that use only one
bank may be unaware of innovations in banking or financial services or of new
competitive loan rates or instruments that other banks in the industry are adopting or
are first-movers in adopting. Over time, the accumulated effects of the social
liability of this weak network position can put the firm at a strategic disadvantage
for gaining timely market information about capital availability and cost, even if
their primary bank is motivated to find Pareto-improved outcomes, because there is
a lack of knowledge of alternatives.
Conversely, an organization with a dispersed portfolio can optimize a firm's
access to new and novel information (Granovetter 1973; Burt 1992) but lacks the
benefits of collaboration and resource pooling that are associated with embedded
ties (Powell, Koput, and Smith-Doerr 1996). Under these conditions, firms are likely
to be put at a disadvantage in garnering favorably priced loans because they lack the
embedded ties to a bank that promote integrative bargaining and the search for
Pareto-improved outcomes. The trade-off between consolidated and disbursed
networks suggests that a middling level of network coupling provides the benefits of
markets ties and embedded ties, while minimizing their disadvantages (Uzzi 1997a).
When a borrower's ego-network contains an integrated mix of arm's length and
embedded ties, the borrower is in a position to scan the market widely for
innovations in banking and financial services, while remaining in close collaboration
with a principal lender or two. For example, a borrower may learn of a new financial
service (e.g., processing of credit card receipts, revolving credit arrangements, or
employee retirement accounts) through its weak ties and then work with lenders it
has close relationships with to develop these services for the firm at competitive
rates. Consistent with these arguments, firms that maintain an ego-network with a
dual mode structure have been found to gain efficient access to market information
and to equalize power differences in the investment banking industry (Baker 1990)
and to minimize the probability of failure in supplier-manufacturer networks (Uzzi
1996a). In the context of the capital lending market, we expect mixed ego-network
coupling to have a similarly beneficial effect on the cost of capital.
These arguments suggest two additional hypotheses:
H3: The ego-network size of the borrowing organization is positively associated
with the organization's cost of capital.
H4: Organizations with a dual mode ego-network will obtain financing at a lower
cost than organizations with either a dispersed or a consolidated ego-network.
DATA AND METHODS

We use data from the 1987 National Survey of Small Business Finances, which was
administered by the Board of Governors of the Federal Reserve and the Small
Business Administration on a one-time basis in 1988-1989. The purpose of the study
was to investigate the sources of borrowing of small businesses and how
characteristics of the market, the firm, and the lending relationship affect the cost,
availability and conditions of credit. The face-to-face administered questionnaire
surveyed a random sample of 3,404 non-financial, non-agricultural small businesses
operating in the US in 1989. Sample range covered firms with 50 to 500 employees
and $1,000 to $154,000,000 in asset value; 1,875 firms were corporations and 1,529
were partnerships or sole-proprietorships. Nearly 90 percent of the firms were owner
managed. The response rate was between 70 to 80 percent, depending on the item.
This reduced the sample size to approximately 2400 cases. Respondents answered
questions about the characteristics of their firms, including the quality, number, size,
and duration of their lending relationships, sources of financing, and the conditions
of their loans. In addition, the survey administrators collected some financial data on
each firm for the previous year where applicable.
Variables
Our dependent variable is the cost of capital, which we operationalize as the interest
rate on the most recent loan received by the organization. Approximately 1300 firms
reported having secured long-term financing from one of their banks in the form of a
loan over the period from mid 1987 to 1989, the observation period of the survey.
Length of bank-firm relationship is measured as the number of years the business
has had done business with the bank. Multiplexity of bank-fIrm relationship is
operationalized as the count of the number of services the business engages in with
the lender. These include: Brokerage services, capital leases, cash management
services, checking accounts, equipment loans, letters of credit, lines of credit,
mortgages, motor vehicle loans, night depository, pension fund, processing of credit
card receipts, retirement accounts, revolving credit arrangements, savings accounts,
supplying money/coins for operations, trusts, and wire transfers.
Ego-network size is a count of the number financial institutions a firm uses for
any of the above financial services. Some firms reported non-banks as possible
sources of financial capital but did not report having a loan from them. Because the
potential for receiving capital was possible from these non-banks, we included these
reported non-bank sources in the ego-network size variable because some theories
view these potential sources of capital as an important dimension of price
competition in the banking market for small businesses (Pfeffer and Salancik 1978).
Ego-network coupling measures the level of consolidation in a firm's ego-
network of ties to banks they do business with. Consistent with previous studies, we
operationalize this measure using a modified Gibbs-Martin index of heterogeneity
(Baker 1990; Uzzi 1996a). It is calculated by summing three sources of business a
firm dedicates to its banks: the amount of cash in checking, the amount of cash in
savings, and the size of the line of credit. For each firm, we summed these three
sources and then added the sums across all banks. This permitted us to calculate the
percentage of each firm's business that is dedicated to each of its banks. For
example, a firm may have a total of $100,000 of 'banking business' to distribute

across the three banks it interacts with. Suppose it makes the following allocation:
The first bank gets $50,000 of its business (i.e., 50% because $50,000/$100,(0) and
the second gets $30,000 (i.e., 30%) of its business, and the third gets the remaining
$20,000 of its business (i.e., 20%). Our consolidation index is calculated by taking
the sum of the s~uared percentages. For our hypothetical example, the index would
be .38 (i.e., 52+.3 +.2 2). This Gibbs-Martin index ranges from just above zero to one.
A value close to zero indicates a dispersed network; a value of one indicates a
perfectly consolidated network; and values between .4 and .6 indicate a dual mode
network (Uzzi 1996a). Our prediction is that a middling level of coupling will result
in the best cost of capital for firms. In our model, we represent this by adding
network coupling and network coupling squared into the equation. A negative and
statistically significant value on the linear term and a statistically significant and
positive value on the squared term would suggest support for the hypothesis.
Following financial theories, we control for important firm, loan, and market
characteristics that affect loan interest rates (Petersen and Rajan 1994). Financial
ratios are widely used to determine the credit worthiness of a business. Creditors are
primarily interested in the firm's short-term liquidity (i.e., the ability to quickly
convert assets and other resources into cash) and its long-term ability to service debt.
We use two standard financial ratios to operationalize a firm's credit worthiness, the
acid (quick) test ratio and the debt ratio (Gitman 1979). The acid test is computed as
the firm's current assets minus inventory, divided by current liabilities, and the debt
ratio is computed as firm's total liabilities divided by total assets (i.e., the proportion
of total assets provided by the firm's creditors). Other firm-level factors controlled
for are organization age (log transformation) and organization size (log
transformation of number of employees). For the subs ample of firms analyzed, the
mean age was 14 years (with a range from 6 months to 105 years), and the average
number of employees was 25 (with a range from 1 to 475). We include two controls
for the characteristics of the loan. Collateral measures whether a firm pledged
physical assets as security in the loan agreement. In case of default, the bank can
seize the collateral, sell it, and apply the proceeds towards satisfaction of the firm's
obligation. Term spread controls for differences in interest rates attributable to
different loan maturities. It is calculated by subtracting the Treasury bill yield from
the yield on a government bond of the same maturity (Peterson and Rajan 1994).
Finally, we include four controls for financial market characteristics. The first is the
prime rate. The prime rate is the interest rate banks charge to their best customers
and serves as the pegging rate that banks use in pricing commercial and consumer
loans. The second control is the level of bank concentration in the local area: Areas
with high concentration contain one or very few financial institutions; areas with
low concentration contain many financial institutions. The Federal Reserve provided
this variable in an ordinal form (3=high, 2=medium, 1=low concentration). The
higher the level of concentration of banks in a region, the less competition there is
among banks and the more power they have to set rates (Peterson and Rajan 1995).
Finally, we include indicator variables to control for the census region where the
small business is located (Northeast, North Central, South, and West) and the
industry in which it operates (using two-digit Standard Industrial Classification

codes), since both of there variable are thought to affect interest rates.
We employ a Tobit regression model to analyze the effect of lending
relationships on the cost of capital because the interest rate variable cannot take on
values below zero percent or above the value set for usury in our sample. Tobit
analysis is appropriate for estimating models on this type of limited dependent
variable because it will not estimate values out of the range of truncated values as
would OLS, and because it produces unbiased and efficient estimates (Maddala
1983; Baba 1990; Roncek 1992).
RESULTS
Table 1 presents the results of our Tobit regression analysis. The overall model was
statistically significant at the 0.00 1 level. As expected, we find that many of the firm
and market level control variables are predictive of a firm's cost of capital. Older
firms received loans at lower interest rates. This finding suggests that banks value
older, more established firms. Age of the firm appears to carry important
information in a market context that is not fully substituted by the duration of the tie
between the organization and its bank.
Not surprisingly, the prime rate was a good predictor of the loan interest rate.
The lower (or higher) the prime rate, the lower (or higher) the interest rate on the
loan. Firms that pledged collateral with their loan received lower interest rates on
those loans. Firms located in areas with a high bank concentration (i.e., areas with
less competition among financial institutions) had higher capital costs. Two of the
four regional indicator variables were statistically significant; none of the seven SIC
indicator variables were statistically significant.
The results from the exogenous variables are broadly consistent with our
expectations. Consistent with hypothesis I, relationship duration is inversely related
to the cost of capital. The longer a small business and a financial institution have
been interacting, the lower the interest rate tends to be on the firm's loan. (The
average duration of relationships was 13 years, with a range of I to 95 years).
Consistent with hypothesis 2, the greater the degree of multiplexity in the
relationship between the bank and the firm, the lower the the cost of capital. Thus,
businesses maintaining mUltiple connections with their financial institution
performed well in the competition to secure capital at favorable interest rates. (The
average number of multipie x ties was 2.6, with a range of 0 to 14 ties).
Hypothesis 3 was not confirmed: Ego-network size did not have a statistically
significantly effect on the cost of capital. (Average ego-network size was 2.2 banks,
with a range of 0 to 12 banks). One possible reason for this may be that an average
size of just over two banks does not give firms much bargaining power or the ability
to shop the market widely enough. It may also be that the effect of network size has
no net effect once the quality of the relationship and the distribution of ties with the
network has been controlled for (Uzzi 1996a).
Consistent with hypothesis 4, a dual mode network of embedded and arm's-
length ties is positively related to a lower cost of capital. Small businesses
maintaining either only arm's-length ties or only embedded ties put themselves at a
Table 1. Tobit Analysis Predicting Interest Rate on Most Recent Loan, National Survey of
Small Business Finances, 1989
EXOGENOUS VARIABLES Coef. (s.e.)
Structural Embeddedness
Length of bank-finn relationship -.012* (.006)
Multiplexity of bank-finn relationship -.043* (.017)
Ego-network size .003 (.122)
Ego-Network coupling of lending ties -3.264** (.985)
(Ego-Network coupling of lending ties)2 2.828** (.874)
Firm Characteristics
Acid ratio -.012 (.011)
Debt ratio -.120 (.117)
Log of age -.154* (.069)
Log of employment .001 (.005)
Market Characteristics
Collateral -.343* (.170)
Term spread .095 (.060)
Prime rate .286** (.031)
Bank concentration in MSA .246* (.100)
Reg I -.332 (.189)
Reg2 -.362* (.185)
Reg3 -.379* (.184)
Sic1 -.882 (.631)
Sic2 .133 (.211)
Sic3 -.310 (.212)
Sic4 .252 (.360)
SicS -.269 (.161)
Sic6 -.445 (.284)
Cons 9.700** (.539)
*p < .05, **p <.01 (all tests two sided) n = 1308
Prob > X 2 = 0.000
Log-likelihood = -2890.08
disadvantage in the competition to secure capital at favorable interest rates. These

results suggest that the distribution of exchange within a network plays an important
role in determining which actors garner the potential benefits of a network of
relationships. In comparison to dual mode networks, networks that are overly
dispersed or overly consolidated are relatively less effective in shaping market
exchanges with trading partners than are dual mode networks.
DISCUSSION
This chapter asked the question: What logic governs economic exchange between
financial institutions and small businesses? Seeking to broaden our understanding of
capital market dynamics, we proposed an integrated social capital and network
embedded ness approach, arguing that the quality of the relationship between a bank
and an organization and the architecture of organization's network of lending
relationships shapes the cost of financial capital. In this view, the cost of capital is
not based solely on general market conditions, firm-specific financial ratios, or net
present values of investment opportunities (although these remain important).
Rather, the quality and mix of arm's-length and embedded ties between a bank and a
business create new value in the relationship and increase the flow of information.
Specifically, we advanced four hypotheses to test our arguments. We found that
small businesses garner loans at lower interest rates by increasing the duration and
multiplexity of their relationships with a financial institution. We also found that
businesses can most successfully lower their cost of capital by constructing an ego-
network portfolio that includes the proper mix and intensity of ties to financial
institutions. Finally, we found that a simple measure of network size had no effect
on the firm's cost of capital. These results offer evidence in favor of the
embeddedness thesis and suggest that economic exchange is not only embedded in
ongoing social ties but that such ties produce outcomes that add to the benefits of
market transacting.
The implications for corporate social capital theory are evident. Ties with
financial institutions need not yield favorable interest rates. In this chapter we have
shown that the mix of duration, multiplexity, and volume of these ties and a medium
amount of network coupling derive social capital from these ties (lower interest
rates). But we have also shown that, for example, a high level of network coupling
can create social liability. This type of research can shed light on the question of
under what conditions social structure produces social capital or social liability.
Our results also have implications for the economic sociology of financial
markets. Consistent with Podolny (1994), we found that relationships offer an
alternative way for firms to manage uncertainty and improve upon pure market
outcomes. In comparison to simple market ties, embedded relationships appear to
transfer useful information about the firm's ability to service credit. The idea that a
mix of consolidated and disbursed network connections may be highly functional is
expressed in Baker's (1990) research. He found that 'hybrid interfaces' are most
effective at exploiting power advantages and reducing resource dependence because
they combine the advantages of'relationship interfaces' and 'transaction interfaces,'
without many of their disadvantages. Our work differs slightly from Baker's on the
issue of whether parties are motivated more by the tradeoff between power and
efficiency or by the informational and bargaining relationship benefits of dual mode
ego-network structures. Baker stresses power considerations, as opposed to the value
of relationship building. Consequently, we would reverse Baker's order of priority:
Dual mode network coupling offers a way to maintain high-quality relationships,
while retaining the important benefits of atomistic markets. In our study, this
theoretical reversal is logical given the improbability that small firms can ever gain
true power advantages over the large banks. Another difference is the strategic
intentionality and agency implicit in Baker's idea. Our approach argues that network
coupling is an emergent property of actors' attempts to balance social and market
imperatives. And that the consequences of attempts to balance ties is always
imperfect because actors are in a web of ties, some of which are beyond the actor's
direct control. Nonetheless, creating a network with the proper degree of network
coupling requires sufficient knowledge of the market. Yet, mixed coupling is only
feasible when lenders trust the borrower enough so that the borrower can maintain a
non-consolidated network and when the borrower trusts the lenders enough to not
have to rely on disbursed ego-networks (Uzzi 1996b; see Uzzi and Gillespie 1998 on
the formation of bank-firm relationships).
This chapter also fits into recent efforts to develop a broader understanding of
the social processes surrounding entrepreneurship. The success of small businesses
and start-up firms is dependent on more than the personal traits of the entrepreneurs
or the financial characteristics of their business (Becker 1964). Success crucially
depends on gathering resources and information via networks extending beyond the
boundaries of a particular firm or individual entrepreneur (Gabbay 1995, 1997).
Similarly, network ties provide informational cues that outsiders make inferences
upon. When a struggling small business gets capital at a competitive rate, this can
serve as a signal of legitimation to other exchange partners who rely on banks to
evaluate the financial wherewithal of firms. Our analysis shows how structural
embeddedness plays an integral role in the process by which social capital (e.g., a
strong bank-firm relationship) is used to acquire financial capital. We further
speculate that the acquisition of this financial capital can contribute to building
social capital (e.g., increasing the firm's legitimacy and reputation). Thus, there is a
cumulative, reciprocal relationship between financial capital and social capital.
Future research might examine the processes of building and deconstructing
lending ties and how these processes are shaped by market characteristics. One
might expect that borrowers start with a large ego-network (trying to maximize the
probability of securing a loan) and then, as time passes and loans are acquired, they
gradually decrease the size of their network. A firm's ego-network size may
decrease over time because trust can only be gained through enduring and repeated
relationships. To address these propositions, there is a need to track the evolution of
specific firm ego-networks over time. Too frequently, social network analysis
consists of static snap-shots where network structure is assumed to be unchanging
(Leenders 1995b). Our chapter suffers from this weakness, so future research should
examine how bank-firm networks change over time. In addition, because we
examined lending relationships from the perspective of small businesses, as opposed
to financial institutions, future research could profit from a more in-depth analysis of
the supplier side of the loan market.
In the past twenty years, the pace of change in the financial services industry has
been without parallel-typical forms of market exchange and bank control have
diminished in importance or been supplanted by other exchange logics (Davis and
Mizruchi 1997). Although more research is needed on the economic sociological
dimensions of capital markets, this chapter has attempted to specify the mechanisms
and forms of embeddedness that shape lending relationships.
We express appreciation for the valuable comments offered by Roger Leenders and Shaul Gabbay on an
earlier version of this chapter. We also gratefully acknowledge the assistance of Mitchell Petersen,
Woody Powell, John Wolken, and the Board of Governors of the Federal Reserve System. Please direct
correspondence to Brian Uzzi.
Venture Capital as an
Economy of Time
•
25
John Freeman
ABSTRACT
Entrepreneurship is a social activity in which resources used to build new
organizations are acquired through the social relations that combine to create a
community structure. An important participant in this community is the venture
capital firm. These organizations provide funding for new ventures and also help
build the new ventures' social capital. They do this by making social connections to
other important actors and by providing advice. Venture capital organizations are
constrained by shortages of time that rise in severity as the venture capital firm's
centrality in the community increases. As a result of these shortages of time, there is
an inverse correlation between central location in the community and willingness to
work with the entrepreneur whose venture is struggling. The most central venture
capitalists are expected to display greater impatience with their portfolio ventures
leading to different outcomes for those ventures defined in terms of three liquidity
events: acquisitions, failures, and initial public offerings.
THE SOCIAL NATURE OF ENTREPRENEURSHIP

The purpose of this chapter is to present the beginnings of a sociological analysis of
entrepreneurship. Starting a business organization, the defining act of an
entrepreneur in this treatment, is a social act. Organizations by definition include
more than one member, so of course starting an organization involves attracting
these participants and organizing them. A second social aspect of the phenomenon is
a consequence of the resource gathering activities of the entrepreneur. These
resources usually come from other organizations.
Venture Capital as an Economy of Time - 461
Such observations are commonplace, yet many writers ignore the social context
of entrepreneurship, taking an heroic, individualistic view of the subject. The
problem with this approach is that when people do things in groups, and rely on
actors outside those groups for support, individual acts of heroism (and villainy) are
difficult to discern. So the social nature of the activity tends to distract attention
from a romantic story.
This individualistic perspective on entrepreneurship has led many scholars to
focus on the psychological properties of entrepreneurs. They are described as less
risk averse than most people, as having higher needs for achievement, and they are
often found to have a more internal locus of control (Brockhaus and Hurwitz 1986).
This is consistent with the treatment one usually finds in the business press, which
tends to focus on how some entrepreneur overcame various obstacles to build a great
enterprise. A sociological treatment of entrepreneurship does not deny that
entrepreneurs do great things. Rather, sociologists tend to look elsewhere for causal
explanations in order to preserve the falsifiability of the explanation. Much
apparently depends on the social capital members bring to those teams and on how
well people work together. The second observation noted above is that cooperative
activity spans emerging organizational boundaries. The entrepreneurial team almost
always receives support from non-members who themselves usually work for other
organizations. All business organizations require such mundane resources as raw
materials, power, water, waste disposal, mail and banking services. They also
require customers or clients. Most also need sponsorship and endorsement. All of
these are provided by other organizations and by individuals who are members of
those organizations. So the entrepreneurial effort is embedded in a network of social
relations that provide material support, information and legitimacy to the nascent
organization (Birley 1985; Aldrich and Zimmer 1986). The resources that comprise
this support are not randomly allocated in society and there is much to be learned by
studying the allocation and accumulation processes of this social capital
(Stinchcombe 1965; Aldrich 1979).
The channels of allocation are important mechanisms structuring opportunity at
the societal level and producing varying rates of social mobility. They are especially
important for ethnic minorities and immigrant groups. They channel capital and
human activity into areas of economic expansion. They affect the speed with which
society responds to changing technology and demography. Finally, of course, they
affect the pattern of organizational variability. Entrepreneurs struggle to solve
problems of organizing so as to commercialize new technologies, or market new
products or services. Successful organization invites imitation. New forms of
organization are thus created and their populations proliferate.
Successful new organizations grow but how much they grow and the time path
their growth follows is not well understood. Some become giant corporations. Most
fail. Some organizational forms include smallness as a design characteristic. Such
organizations often reach their maximum size at opening (e.g., diners) while others
grow slowly but may achieve very large size (e.g., insurance companies). At some
point, organizations outlive or outgrow the state of disorganization that usually
follows startup, or they fail. Sooner or later they stop being entrepreneurial in the
sense of being 'young' or 'new.' When this happens is difficult to define but almost
certainly varies between organizational forms.
Bounding the trailing edge of entrepreneurship is difficult no matter what
definition one employs. This is partly because defining when a pattern of change
slows down, and how much it has to slow down before one decides to stop paying
attention, is problematic. It is also difficult because different forms of organization
become institutionalized at different ages and at different developmental stages. For
some kinds of organizations, an initial public offering of stock (IPO) constitutes a
sea change that is irreversible and organizationally significant. After an IPO,
companies must be audited and report the results publicly. This constrains their
internal information and decision processes. They are forced to become more
orderly. But some kinds of organizations go through an IPO when they are quite
young, before they begin to sell products much less generate profits. Biotechnology
companies are such organizations. These companies often retain their useful
disorganization, looking very much like Mintzberg's (1983) adhocracies long after
they have 'gone public.' They may grow quite large without developing the
managerial functions that are fundamental to most corporations. For example, since
biotechnology companies usually have no products in the market, they have no
marketing function. Their financial officers may be scientists working as part time
fund raisers. Financial controls are often primitive. This is because their
organizational models are university research laboratories. The point here is that the
problem of defining ending dates for the period of entrepreneurship is theoretically
challenging, and this is true whether or not the research defines entrepreneurship as
we do here or in some other way.
In the research reported below, three terminal events are used to bound the
phenomenon. These events would probably be deemed worth studying by most
scholars even if they do not accept the definition of entrepreneurship we have
chosen here because they are the most common paths to liquidity. That is, they are
events through which investors are able to recapture invested funds: acquisitions
(ACQ), initial public offerings (IPO), and failures (FAIL). The first two are
theoretically important because they involve the sale of the company, not just its
assets. The organized entity has value above and beyond the desks, computers,
inventories and other tangible assets. This value is partly dependent on having
previously solved the organizational problems that confront anyone trying to employ
the technology those tangible assets were assembled to exploit. When those assets
are sold off, and the organization disappears, we simply refer to these organizations
as failed. These three events are ordered according to their usual payoff to those who
own equity in the company. An IPO produces higher returns than an acquisition,
which produces higher returns than a failure.
VENTURE CAPITAL FIRMS AND THE ENTREPRENEURIAL

COMMUNITY
As organizations move through time, from a gleam in someone's eye, to an
established company, an internal process occurs through which roles gain definition,
routines for operating the business are developed, authorities for acting on the
organization's behalf are specified. Physical plant and equipment are acquired and
put into operation. Nelson and Winter (1982) call this 'learning by doing.' Along the
way, with hard work and good luck, the company generates revenues from its
operations (rather than purely from equity financing and borrowing), begins to
generate positive cash flow, and finally, turns profitable. Of course, many business
organizations never get this far, but if they last, most go through some internal
evolution of this kind. Since this is not the main subject of study here, we will
continue noting as an aside that the speed with which this process occurs, the stages
that can be discerned and the problems managers and others encounter along the
way vary from one form of organization to another. The more technical and
organizational innovation is involved, the more challenging this process is for
managers and the greater the hazards associated with failing to negotiate the process
successfully. Starting an innovative organization is more difficult, and hazardous,
than starting an organization following a well-established form (Hannan and
Freeman 1984; Freeman 1986).
At the same time, organizations go through a process of developing relations
with their environments. No organizational theorist would deny this. They would
certainly debate how it is done, who does it, and why. These relationships vary in
longevity. Given the time scale that is appropriate for the organizational form in
question, these relationships usually last long enough to take on the properties of
relational contracts (Williamson 1985). Following the language of social network
analysis, any business entity, which we can call Ego, develops around itself a set of
relations with Alters. In real time, these relationships develop serially. Through them
flow money, information, and access to still other relationships. When strung
together, these ego nets form a network that is the structural basis of a community.
The flows of money, information and access constitute channels of opportunity. The
structure of the network, then, constrains opportunity (Larson 1992). It makes
entrepreneurial activity easy for some and nearly impossible for others (Van de Ven
1993). The advantages these social structures provide to an entrepreneur are 'social
capital' in Coleman's (1990: 300-321) sense of the term.
It is almost certainly true that the structures of some of these communities
facilitate the process of starting new business organizations while other structural
arrangements impede it. Indeed, Silicon Valley in California is famous for providing
the support that technology-oriented entrepreneurs require (Saxenian 1994). When
these structures operate efficiently, the internal process described above can be
expected to work faster as well (Florida and Kenney 1988). Some of the information
that flows through this network describes the routines for organizing and managing
new ventures. Some of the access involves finding key people who know what
routines to employ and how to make them operate when required (Bygrave 1988).
This chapter focuses on just one form of organization found in such
communities-Venture Capital firms (VCs). VCs are management firms set up to
invest capital in young ventures. This capital is provided by investors who
contribute to a venture capital 'fund.' The fund is organized as a limited partnership
that liquidates at a known date, distributing the proceeds to the partners. Since the
funds have a planned life cycle, the VC often raises multiple funds, staggering them
over time so as to balance the payouts and manage the workload they generate. The
classic definition of a VC is based on their investing activities in young, rapidly
growing companies (Bygrave and Timmons 1992; Barry 1994). But some VCs
finance corporate restructuring processes such as Management Buyouts. In this
chapter, we focus on the more classical definition and confine attention to their
activities with entrepreneurs.
Young ventures often develop through a sequence of investment events called
'rounds' of financing. The VCs involved in these rounds often change. This serial
quality to investment practices encourages careful and sometimes sober reviews of
the company's progress, or lack of progress. One of the important contributions of
venture capital investors is this periodic review. It forces the entrepreneurs to look
back to their previous claims and promises, and focuses attention on the dynamics,
not just the current state of affairs.
Venture capitalists frequently state that their consulting services are just as
important to the new venture as is the funding. Typical is the following statement,
taken from a promotional brochure provided by one of the most famous venture
capital firms, the Mayfield Fund:
We Offer A Global Network of Contacts
Mayfield has close working relationships with technology leaders, universities, other
venture capitalists, financial institutions, consultants and corporations throughout the
United States, as well as an international network... Mayfield's contacts provide a key
resource for developing the relationships critical to a growing technology-based
company, including potential corporate partners, both here and abroad.
Because of our long association with a large number of successful companies and
entrepreneurs, a relationship with Mayfield is highly regarded. It can enhance the
credibility of a young company with potential customers, vendors and employees, and
with other financial institutions.
In other words: Mayfield not only provides financial capital, but also provides social
capital to the entrepreneur.
Community structure develops like any social structure out of the differential
tendencies for some to associate with others. These relationships are implied when
venture capitalists and other actors participate in rounds of financing together. In the
vernacular, they 'do deals' together. When such actors provide services to an
entrepreneurial venture, they usually seek information on the identities of the other
parties. Syndicates of venture capital firms are formed in much the same way Stuart
(this volume) finds in the alliances developed among semiconductor firms. For
example, some venture capitalists have strong preferences for working with certain
accounting firms and will refuse to invest if another accounting firm provides
consultation and auditing services to the entrepreneur. Venture capital firms select
other venture capital firms. Those currently investing in a young company often
recruit and select other VCs for participation in subsequent rounds of financing.
Since their currently illiquid investments are made more or less valuable by the
actions of subsequent investors, they care who these investors are and they care
about the terms of the deals that are negotiated for their entry to the group of
investors.
One of the organizing principles underlying these associational preferences
among VCs and other members of the community is faith in the reliability of others
to perform, one of the important functions identified by Nooteboom (this volume) as

characteristics of individuals and organizations acting as go-betweens. This
attribution of reliability is a form of trust, and is based in part on having done
business together in the past. This is not the same thing as endorsing the other's
ethics, though some minimal level of ethical behavior is surely implicit. The point
here is that having been involved in the creation of a large number of ventures
establishes any actor in the community better than if that actor has only been
involved in a few. There are obvious status distinctions among venture capital firms
much like those Podolny has studied among investment banks (Podolny 1993). If it
is true that the community structure channels opportunity, then the more centrally
located actors should provide better access to entrepreneurs they support. Implicitly,
they deny easy access to others. The chapter by Podolny and Castellucci (this
volume) argues that a high status organization has more and better access to
resource providers than does a low status organization. A point completely in
harmony with their view is that being connected to a high status ally has this same
effect. Below, we will argue that when the high status other is a venture capital firm,
its ability to provide such access is limited by the time it takes to do so. For these
organizations, time is a very scarce resource.
In addition to these relationships between organizations, it is useful to note that
the entrepreneurial community is a very personal society. Managers, board members
and other individuals working for or with these ventures bring their own resources to
the organizations (Fried and Hisrich 1995). In particular, the venture capitalist
brings personal experience and contacts to the young venture. This is an example of
Granovetter's embeddedness concept (Granovetter 1985). These personal contacts
are a form of social capital that provides value to the entrepreneur by making it
easier to move the venture through the organization-building process. So we would
expect better connected venture capitalists to provide services that improve the
performance of the portfolio ventures. We need to qualify this statement, however,
because the strategic position of the venture capital firm affects its relations with
entrepreneurs.
THE ECONOMY OF TIME

Venture capital firms are highly competitive organizations. They are structured to
provide very powerful incentives to their managers, incentives that encourage them
to seek out young ventures that have the potential to grow rapidly. These incentives
also encourage the venture capitalist to seek out investors who are willing to take
risks and remain illiquid for lengthy periods of time. An early name for venture
capital was 'patient money.' In the industry's early years (from 1946 to about 1975),
VCs received the bulk of their funds from wealthy individuals and families such as
the Rockefellers and Mellons. More recently, they have drawn from pension funds,
insurance companies, and endowments. They succeed in drawing new ventures,
called 'deal flow; on the basis of their prior records of success with other young
ventures. These great successes compensate for the many unsuccessful ventures they
also back, deals that draw substantially less attention from the press.
The other resource that forms the basis for competition is the capital invested in
the funds VCs raise and manage. Investors evaluate the VCs on the internal rate of
return (lRR) , the annually compounded rate at which the proceeds from liquidity
events among their portfolio companies accumulate. The more famous the VC, the
higher its IRR tends to be. Investors willingly invest in the funds managed by people
who have been highly successful in the past. VCs whose last fund produced a low
IRR often fail to attract capital to their next fund.
Both of these resource flows depend on the company's past performance which
in turn depends on a minority of investments in very successful young ventures
(Bygrave and Timmons 1992). A successful VC firm produces an IRR of about 30%
per year. Performance tends to be approximately log-normally distributed. Many
ventures fail or are barely profitable (from the VC's point of view). A few are
enormously successful. When such successes occur, they make the person who
managed the investment famous. The attribution of a Midas touch draws the two
resources under discussion here. Such enhancements to market position are
sometimes called the 'Mathew Effect.' (Merton 1964; Podolny 1993) Working
against such advantages are the limitations of time available to the ventures
capitalists themselves. The more famous the VCs are, the easier it is to attract
money, but generating very high performance requires time for analysis, time for
monitoring the young venture's progress, and time to provide the mentoring and
contact-building that are so essential to the nonmonitory contributions VCs make.
The younger the venture, the more extensive these services are likely to be and
the greater the risk. Risk emanates from the fact that many strategic and
organizational problems wait to be solved and also from the illiquidity of the
investment. Failure to solve these problems can lead to slow growth or even to
outright failure. Slow growth is as bad as failure for the VC because of the fixed
time cycle for their funds (described above) and because their performance is
evaluated in annualized terms.
Only some of these activities can be delegated to staff members. Investors are
paying for the judgement of the prominent ventures capitalists who lead the venture
capital firm. Salaried staff can work through the analysis required by due diligence
but the judgment required to chose high potential ventures is not subject to
routinization. It often depends on extensive personal contact with the entrepreneurs.
Similarly, the investors conduct their own due diligence analysis on the VC and its
senior managers. Such analysis often consumes great amounts of the VC's time. So
the size of the venture capital firm is limited by the time of its senior managers.
The VC is paid in two ways, neither of which involves fees from the
entrepreneur. First, there is a management fee charged on capital invested, usually
between 1.75% and 2.25% per year. When a fund is raised, limited partners commit
funds. As ventures are identified, those commitments are drawn upon and the VC
starts charging a management fee on them. If there is no flow of deals, there is no
management fee. So there is pressure to invest the fund quickly. The second way in
which the VC makes money is by collecting a share of the profits when the fund
liquidates, called the 'carry rate,' usually about 20%. Notice that operating expenses
for the VC do not figure in these reward mechanisms. Costs are deducted from the
revenues generated by these two forms of payment and are not captured in IRR
calculations. So venture capital firms do not win competitively by being more cost
efficient. Rather, they win competitively by being more time efficient.
The VC looks for entrepreneurial ventures that have the potential to grow
rapidly, seeking liquidity that precedes the terminal date of the fund. The preferred
source of liquidity is an IPO, a public sale of stock. Often, the IPO is followed by a
secondary offering in which the VCs liquidate their holdings. So the VC is not
interested in many of the ventures seeking funding, even when they are quite viable
businesses. The venture has to have the potential to grow big enough, fast enough to
warrant a public offering. The investment banks who organize those offerings by
underwriting the stock sale have fixed costs that virtually prohibit offerings of less
than $35 million. So the venture has to be big enough to justify such a valuation.
The younger the venture, the more difficult it is to assess its growth potential.
Consultation to support the growth at young companies is also more time
consuming. Once the VC is committed, illiquidity creates dependence and the VC
responds by doing whatever is required to make the venture grow. This includes
providing managerial services, such as negotiating with lenders and major
customers, recruiting management and technical talent, and finding facilities (Bruno
and Cooper 1982). The VCs often serve on the young venture's board of directors.
Sometimes, they fire the entrepreneurs. When venture capitalists do this, they have
to find replacements. So most VCs shun ventures with glaring management
weaknesses. Fixing such weaknesses costs them time they would rather spend
identifying high potential prospects.
Venture capital firms are not expensive organizations to run. Their capitalization
is not high. They do not require costly equipment. Their staffs are often modest in
size. Since the profits of the firm are divided amongst the partners, there is an
incentive to expand the funds relative to the number of general partners. This
process is limited by the time these general partners have for analyzing new ventures
and helping them grow. Therefore, it is not money that is in short supply for most of
them, it is time. The larger the fund, the more critical this problem is because the
number of deals is correlated with fund size.
The evaluation of venture capital performance by the internal rate of return
exacerbates the time problem because the IRR is time-valued. That is, the return is
compared with the rate of return that would have been produced if the same funds
had been committed to some easily managed highly secure investment such as
government bonds or bank certificates of deposit. Given the risks involved, venture
capitalists are expected to produce much higher returns. So evaluation of the VC's
performance is not based simply on whether the fund cashes out at the end of its
time with a handsome surplus, evaluation depends on how long funds were under
management and what the return was over that time. So there is intense pressure on
venture capitalists to bring ventures to liquidity quickly.
There are two ways to deal with this shortage of time. The VC can restrict its
investments to older ventures that are closer to an IPO. Because the risk is lower and
the information is better, such an investment strategy allows fewer, larger
investments. Of course, such investments produce more modest returns. The second
way of dealing with this shortage of time is to follow the lead of other VCs who
have spent the time required to perform the due diligence. This strategy requires that
the focal VC trust the judgment of the lead VC. Trust here means faith in the ability
of this other VC, or set of VCs, to perform such analysis. In this sense, the followers
use exactly the same logic as their investors employ-they commit capital based on
the sagacity of some other investor. Why doesn't everyone just do this? The answer
is that not everyone can get a piece of every deal. One VC informs another about
ventures that need more funding, and they have influence over the prices that are
charged for stock in that venture. This argument follows those developed by
Podolny and Castellucci (this volume) and Stuart (this volume) who both argue that
superior social status increases the choices organizations have for establishing
relationships with others. So it is with deal flow. VCs that are central in the
community and have high status can reciprocate with access to ventures for which
they have performed the due diligence. This involves technical expertise as well as
time. So one VC may specialize in biotechnology, and include in its ranks general
partners who have advanced training in fields such as biochemistry, but still invest
in ventures that are not biotechnology companies by trading access to deals with
VCs knowledgeable in other areas. In other words: social capital generates social
capital.
Because of the greater time pressure on VCs managing multiple large funds, we
can expect less patience from them. They are looking for 'home runs,' huge
successes. Modest successes are not disasters, but these central players in the
community cannot maintain their elite status earning modest rates of return. To earn
the 30%, compounded annually, that is expected of the elite, they need great
successes. So ventures that are moving forward at a slow pace are often abandoned.
It is the time to monitor the ventures' activities and guide them through the
entrepreneurial community that is the main cost to the VC. The more deals the VC
does, the more connections it has with other organizations in the community. If this
is true, we can expect network centrality of VCs to accelerate the liquidity process.
This leads to the following hypothesis:
The more central a venture capital firm is in the entrepreneurial community, the
more impatient it is likely to be with its portfolio ventures.
Impatience is displayed in a higher rate of acquisition and failure among the

ventures in the portfolios of the most prominent VCs.
MODELS AND ESTIMATION

As indicated above, we are interested in three kinds of events that constitute
outcomes for entrepreneurial ventures: initial public offerings, acquisitions, and
failures. Ventures do not experience more than one of these simultaneously, but they
can experience them sequentially. That is, a company may sell stock to the public
and then be acquired or fail. Of course, the scope of the study changes
fundamentally if one focuses on the entire life courses of such organizations. Our
interest here is on their period as entrepreneurial ventures. In addition, it is also
important to note that when any of these events occurs, the relationships of the
venture with its associates in the community are altered. If a company is acquired,
its relationships become subordinate to its parent corporation. If it conducts an IPO,
it must be audited and its strategic relationships with key customers or suppliers
must be disclosed. In general, then, these events endogenize the community
relationships of greatest interest here. So terminating the analysis when one of the
three events occurs makes sense. The three events represent changes between
discrete states and changes can occur between these states at any point in time.
Accordingly, we study the probability of their occurrence using continuous time
models. We want to estimate the time-dependent process by which young ventures
move from startup to a liquidity event. We do this by estimating the hazard rate, as a
function of duration (organizational age). The methods for conducting such analysis,
often called event history methods, are by now well known (Tuma and Hannan
1984). We note here that the hazard rate can be defined as
. Pr~ ~ T < t1T ~ t)

r (t ) =hm--=----.,,-...l--..:..
1'-1 t -t
the ratio of the transition probability to the length of the duration in a state such as
illiquidity. It can be described (following Blossfeld and Rohwer 1995) as the
propensity to change from the state of illiquidity to the state of liquidity at time t.
We will estimate such a rate for each of the three liquidity events conceptualized as
'competing risks. ' For each kind of event, the other two are treated as censored
observations.
To assess the effects of exogenous variables on such a rate, we need to specify a
parametric model that considers time observed as duration. organizational age in this
formulation . Such an analysis requires the assumption of some form of the time-
dependence in the process. When previous research and theory produces good bases
for specifying this time dependence. a wide variety of models with different patterns
of time dependence is available. Since we do not have such a body of research to fall
back on for the rate of IPOs and the rate of acquisitions. we begin with a piecewise
exponential model. This model allows us to treat organizational age non-
parametrically by estimating a constant for each of a set of time periods. So if the
effect of age on the rate is irregular or nonmonotonic. we can evaluate the causal
factors of interest without fear that observed effects will be artifacts of the time
dependence. The piecewise exponential model performs this estimation with a
structure similar to the use of dummy variables in OLS regression.
The time span over which organizations age is divided into a set of separate
periods
tp = 'f. +'f 2 +'f3 + ... +'f L;
and the rate from origin state 0 to destination state d is
In this expression A(od) is a row vector of covariates and a.(otl) is a row vector of
coefficients and 1/ is an index of time periods tp. After exploring the nature of the
duration dependency. we then estimate the simpler exponential model which
substitutes a single constant for the constants that are specific to each time period in
the model above.
The regressors present in the hypotheses are presented in three groupings:

venture capital firm centrality measures, measures that describe the venture itself,
and control variables that describe economic circumstances. Some of these variables
change over the time period under study. In order to allow for such changing
regressors, we employ the common tactic of spell splitting (Tuma and Hannan 1984)
in which the time period over which a venture is observed is divided into quarterly
subspells. Each subspell is treated as censored on the right except for the subspells
ending with a liquidity event.
Venture capital centrality measures describe client-service provider contacts:
CjNV investment firm degree centrality
C_VC venture capital firm centrality
C_NVC centrality of non-VC investors.
These centralities are the number of entrepreneurial ventures in which the focal
investors have put capital at the study's end (1995). To understand how this relates
to the usual network degree centrality measure (see Knoke and Burt (1983) for a
review) we note, first, that venture capital firms never invest in each other nor do
entrepreneurial ventures invest in each other (or at least this is very uncommon). Our
data, described in detail below, were drawn from the ranks of venture capital-backed
ventures. We use the term 'investment firm' here because venture capital funds are
managed in a variety of ways. Some are large firms as described above. Others are
venture investing arms of corporations. Still others are small investors with funds
contributed only by a few wealthy investors. They may call themselves venture
capitalists, or they may simply refer to themselves as private investors. We begin
this analysis assuming that the 'impatience' argument applies to all investors and that
the shortage of time is a function of network centrality. Then we expand the analysis
to see if VCs classically defined and other investors differ in conformity with the
impatience hypothesis.
Venture characteristics-business strategies and locations. Because we are not
interested in the effects of these variables in their own right, and the theories
prompting their inclusion in the model are obvious, we simply list them below with
their expected effects.
Log Age: Any of the three events can occur at any organizational age, but since
all are stochastic, and we follow them over time until one of the events occurs, the
probability that each will occur cumulates. So we would expect each to be a positive
function of the organization's age (here the natural logarithm of age);
CAPITAL: The data tell us how much capital was raised in each round but not
how much each VC invested; this variable is the sum of the amounts previously
raised as of any date.
Some of the ventures in our data are not young, but small businesses acquired
by entrepreneurs when they seek venture capital to support a strategic change. We
should take this difference into account and do so with a dummy variable,
ENTRY: Assumes a value of 1 if the venture started before January 1, 1981; 0
otherwise.
Financial performance is notoriously difficult to measure for entrepreneurial
ventures. They do not have to disclose the data that one would use for such analyses
and their structures and operating procedures are so volatile that such data would be
hard to interpret anyway. We use stage of operation instead. When ventures start
out, they are usually developing their product or service. Then they make prototypes
and put them in 'beta test' mode, distributing them to potential customers for tryouts.
If such tests are successful, the product or service goes into distribution. Finally, if
the market responds well and production can be expanded to efficient levels, the
venture becomes profitable. For this analysis, we compare those that are not yet
shipping product, those in development or beta test mode, with those that are
shipping or are profitable. Performance should be positively correlated with the
probability of an IPO, and negatively correlated with the probability of a failure. It is
difficult to predict the effect on the probability of an acquisition since acquisitions
sometimes are fire sales, in which the investors are simply recouping whatever they
can from their investments. Sometimes, however, the company is acquired at a
premium for its technology and other high potential assets.
BETA: Assumes a value of I if the venture has a product in beta test, 0
otherwise.
SHPG: Assumes a value of I if the venture is shipping a product, 0 otherwise;
PROF: Assumes a value of 1 if the firm reports that it is profitable, 0 otherwise.
In addition, we want to control for the following fixed regressors:
SV Silicon Valley Location within region running from Menlo Park to San Jose,
California. Most observers would expect a high success rate for Silicon Valley
ventures. So the probability of an IPO should be higher, and the probability that the
venture will FAIL should be lower when the venture is in Silicon Valley.
Industry: The events we are studying are affected by a wide variety of industry-
specific factors such as competitive conditions. We control for industry with the
following dummy variables:
CDPR Data processing
C CSO Computer software
CCOM Communications
1_CEL Consumer electronics
I_IEQ Industrial equipment
CMED Healthcare and biotechnology
C TST Test and analytical instrumentation
CCMP Components
CSEM Semiconductors
CRET Retailing
CENV Environmental
CCHM Advanced specialty materials and chemicals.
See Table 1 for the proportions of observations in each industry.
National level economic conditions
R_IPO Number of IPOs occurring each month
R_BFAIL Number of business failures each month
R_BINCORP Number of incorporations each month
A V_TBILL Average treasury bill yield
A V_PRIME Average prime interest rate.
DATA
The data used to test these hypotheses were provided by the San Francisco company,
VentureOne, which has tracked venture capital activity in the United States since
1987. VentureOne provides data to venture capital firms that the ves use to make
investment decisions. The ves provide information to VentureOne about the
ventures in which they invest. VentureOne also monitors the business press looking
for information on new ventures. They then send questionnaires to the ventures
seeking information on the time of founding, markets in which the ventures are
active (or markets in which their management intends to be active), sources of
funding, location, et cetera. It is important to note that both the ves and the ventures
have an interest in reporting as VentureOne's data is often used by investors to
identify companies that might be seeking funding in the future. VentureOne follows
these ventures over time, sending questionnaires to update their files yearly. This
process stops when one of the liquidity events described above occurs.
Data were provided in the form of tables describing the ventures themselves, the
rounds of financing they have conducted, and the investors in those rounds. There
are 4,073 records on entrepreneurial ventures and 22,068 records on investors in
which each record represents an investor funding a venture in one or more rounds of
financing. The data tell us how much was raised in each round of financing, and
which investors participated, but not how much each investor put into the round.
These data were examined to remove name misspellings and obviously erroneous
records. For example, investors named 'Individuals' or 'blankco' were treated as
missing values. A record was deemed incomplete when there was no start date or
city of location. When an investor could not be matched with a venture, the record
on the investor was deleted. This yielded an effective sample size of 4,064 ventures,
along with statistics describing the investors and service providers with which they
are associated.
Table 1 presents descriptive statistics on this sample of ventures. The variable
AGE is reported in quarters. It ranges from January of 1907 to October of 1995. The
last date of observation is October of 1995. There were 83 ventures that were born in
the month before the data were provided by VentureOne. Of course, these are all
right-censored (still in existence at the time of the study). Approximately two-thirds
of the cases are either shipping product or are profitable as of their last observation.
RESULTS
We begin with a simple version of a piecewise exponential model that allows for
changing regressors. Our objective here is to see if the time dependence is
monotonic. The time periods are defined from 0 to 7 quarters, 8 to 15, 16 to 23, 24
to 31, 32 to 39, 40 to 47, 48 to 55,56 to 63 and 64 and above. The effects of these
period constants are negative, increasing toward zero the older the venture is. In the
last period or two, it turns negative again. Further analysis shows this is because of a
small number of left-truncated cases. When a dummy variable representing ventures
begun prior to 1981 is introduced, the effects of duration appear monotonic.
Table 1. Descriptive statistics

Variable N* Mean Std Dev Minimum Maximum
ACQ 4064 0.16 0.37 0.00 1.00
FAIL 4064 0.09 0.28 0 .00 1.00
IPO 4064 0.24 0.42 0.00 1.00
Starting Time 145133 30.02 36.39 0.00 354.00
Ending Time 145133 31.01 36.39 0.01 355.00
AGE (in quarters) 145133 31.02 36.39 0.01 355.00
CENS 145133 0.99 0.12 0.00 0.01
SV (Silicon Valley) 145133 0.12 0.33 0.00 0.01
CAPITAL 145133 6.51 16.71 0.00 438.00
BETA 145133 0.03 0.16 0.00 0.01
PROF 145133 0.35 0.48 0.00 0.01
SHPG 145133 0.35 0.48 0.00 0.01
C_INV 145133 302.63 330.79 0.01 2.14
C_VC 145133 240.07 260.65 0.00 1.60
C_NVC 145133 62.56 117.74 0.00 775.00
CDPR 145133 0.21 0.40 0.00 0.01
CCSO 145133 0.14 0.34 0.00 0.01
CCOM 145133 0.10 0.31 0.00 0.01
CCEL 145133 0.01 0.11 0.00 0.01
UEQ 145133 0.06 0.24 0.00 0.01
CMED 145133 0.07 0.25 0.00 0.01
CTST 145133 0.02 0.15 0.00 0 .01
CCMP 145133 0.04 0.18 0.00 0.01
CSEM 145133 0.03 0.16 0.00 0.01
CRET 145133 0.07 0.26 0.00 0.01
CENV 145133 0.02 0.12 0.00 0.01
CCHM 145133 0.01 0 .12 0.00 0.01
COTHER 145133 0.23 0.42 0.00 0.01
R_IPO 138903 110.91 65.82 0.00 283.00
R_BFAIL 138903 5148.00 2043.00 458.00 9143.00
R_BINCORP 138903 55434.00 7505.00 21034.00 65691.00
AV_TBILL 138903 8.47 1.67 0.06 01.4
AV_PRlME 138903 8.96 2.64 0.05 02.0
EARLY 145133 0.30 0.46 0.00 0.01
. . ..
* For the first three vaflables thIS IS number of orgamzattons; for the rest It IS the number
of firm-year combinations.
In Table 2, and those that follow it, we report various probability values
associated with the estimated effects. The critical value in use is .05, and this is what
is meant when the effects are treated as 'statistically significant.' We can see that the
effect of being in Silicon Valley does not have the expected statistically significant
positive effect on the rate of IPOs. Surprisingly, its positive effect on the rate of
Table 2. Piecewise exponential model with competing risks

ACQ FAIL !PO
Coef. (s.e.) Coef. (s.e.) Coef. (s.e.)
Period I -9.0947*** (0.5049) -8.3369*** (0.4156) -7.4850*** (0.2224)
Period 2 -6.8071 *** (0.1837) -6.5876*** (0.1993) -6.3934*** (0.1431)
Period 3 -5.8340*** (0.1355) -5.6463*** (0.1502) -5.4136*** (0.1043)
Period 4 -5.2968*** (0.1226) -5.4457*** (0.1527) -5.1257*** (0.1009)
Period 5 -5.1188 *** (0.1266) -5.2196*** (0.1547) -4.8286*** (0.1004)
Period 6 -4.4314*** (0.1120) -5.1199*** (0.1700) -4.6288*** . (0.1054)
Period 7 -4.1078*** (0.1177) -4.7285*** (0.1763) -4.1445*** (0.1073)
Period 8 -4.3035*** (0.1617) -4.3019··* (0.1932) -4.4892*·· (0.1558)
Period 9 -4.9382*·* (0.1350) -5.4591 *** (0.2107) -5.1686*** (0.1240)
CAPITAL 0.0023 (0.0018) 0.0019 (0.0027) 0.0124*·· (0.0003)
SV 0.3562··· (0.1067) 0.5179··* (0.1347) 0.1058 (0.0960)
BETA 0.3006 (0.2882) 0.1588 (0.3146) 0.8915**· (0.1882)
SHPG 0.4256·** (0.0937) 0.0274 (0.1132) -0.0990 (0.0948)
PROF -0.4928*** (0.1119) -2.6390·*· (0.2895) 0.4596··· (0.0830)
Events 639 343 955
.** p < .001, X2 = 2452.2 df = 42.
acquisition and failure is statistically significant. Even in the most complex model,
reported in Table 5, the positive effect of SV on FAIL remains positive and
statistically significant. Apparently, the many famous success stories of the area, the
'home runs' sought by so many VCs, mask a lower batting average. The three
performance dummies, BETA, SHPG and PROF, produce results that make more
sense when the model is fully specified, so we will discuss them below.
Table 3 presents a simpler, exponential model, employing the dummy variable
for left-truncation, EARLY, and dropping the piecewise specification, and adding
the logarithm of AGE. The main effects of interest are the centrality measures. VC
centrality has a positive effect on the rate of IPOs, but it also has a positive effect on
the rate of acquisition, consistent with the impatience hypothesis. The effect on the
rate of failure is positive but not statistically significant. These results stand up to
controls as we complicate the model (below).
The effects of centrality of other investors are not statistically significant for
ACQ and FAIL, but are negative for the rate ofIPOs (the latter effect does not stand
up to the addition of more detailed controls, however). So the impatience hypothesis
is supported when centrality of VCs is analyzed, but not when centrality of other
investors is analyzed. An interesting question is whether these effects reflect the
superior ability of central VC firms to choose winners when they invest, or whether
they nurture their portfolio companies better. We return to this issue in the next
section of this chapter, but at this point we simply note that if these empirical
patterns were generated by higher perspicacity of centrally positioned venture
capital firms, rates of acquisition and failure would be lower for their portfolio
companies.
Table 3. Exponential model with competing risks estimating effects of investor centrality
ACQ FAIL IPO
Const -11.2846*** 0.2914 -10.3697 *** 0.3696 -9.8033*** 0.2270
L_AGE 1.9051 *** 0.0808 1.6125*** 0.1051 1.5170*** 0.0637
CAPITAL -0.0082* 0.0031 -0.0057 0.0042 0.0117*** 0.0003
SV 0.2627* 0.1089 0.4984 *** 0.1375 -0.0020 0.0983
BETA 0.1388 0.2885 0.0585 0.3154 0.6867*** 0.1879
SHPG 0.3173*** 0.0941 -0.0736 0.1138 -0.2187* 0.0953
PROF -0.5601 *** 0.1136 -2.6900*** 0.2905 0.3914*** 0.0834
C_VC 0.0006*** 0.0002 0.0003 0.0002 0.0004*** 0.0001
C_NVC -0.0002 0.0003 -0.0008 0.0005 -0.0011 *** 0.0003
EARLY -2.7616*** 0.1582 -2.7187*** 0.2204 -2.5509*** 0.1298
Events 639 343 955
* p < .05; ** P < .01; *** p < .001, X2 = 3159.6 df= 30.
The analyses reported in Table 4 add controls for the industry and economic
conditions. The effects of the VC centrality measures are the same, positive for both
IPO and ACQ, not statistically significant for FAIL. All three effects are not
statistically significant for the non-VC investors.
There are a number of interesting changes among the other variables comparing
these estimates with the previously presented tables. First, CAPITAL now has
statistically significant effects for each of the three events. Its effects are negative for
both acquisitions and failures, but they are positive for IPOs. This makes sense if
investors can forbear from throwing good money after bad. Location in Silicon
Valley continues to show no statistically significant effect on the rate of IPOs, and
the effect on failure is positive. Now, however, there is no statistically significant
effect on ACQ.
The three performance measures have effects that make sense. Looking first at
the rate of acquisitions, the only statistically significant effect is profitability, which
is negative. When we shift to the next column, that for the rate of failure, both
SHPG and PROF have statistically significant negative effects. So if the venture is
shipping product or turning a profit, it is less likely to fail. Being in beta test has a
negative effect, but it is not statistically significant. For IPOs, the pattern
strengthens. This time profitability has a positive effect on the rate of IPOs as it
should. So ventures that are profitable are more likely to experience an IPO and less
likely to fail or be acquired.
The economic macro variables and industry identifiers are present for control
purposes. Some of the effects are curious, such as the positive effect of AV_ TBILL
yields and negative effects of AV_PRIME. It is not obvious why the number of
IPOs in the previous quarter should have negative effects on all three rates.
The industry dummies show some interesting effects. Being in data processing
(CDPR) exposes a venture to a very high failure rate, a high probability of being
acquired, and a large, negative effect on the probability of an IPO. A similar, though
less pronounced pattern can be seen in the effects of CCSO (computer software) and
Table 4. Exponential model with competing risks: economic conditions, industry

and early founding controlled
ACQ FAll.. IPO
Const -72.8680*" 3.3974 -84.3657**· 5.2624 -59.3934*** 2.2090
L_AGE 1.5062*·* 0.0829 1.1354*** 0.1012 1.0999*** 0.0637
CAPITAL -0.0208**· 0.0038 -0.0187*** 0.0053 0.0110·** 0.0004
SV 0.2133 0.1100 0.4433·* 0.1405 0.0311 0.1012
BETA -0.3939 0.2930 -0.4135 0.3215 -0.0048 0.1906
SHPG 0.0883 0.0961 -0.2968* 0.1168 -0.2867** 0.0947
PROF -0.6276·*· 0.1154 -2.7948··· 0.2924 0.3919*** 0.0860
C_VC 0.0009·*· 0.0002 0.0005 0.0002 0.0003*· 0.0001
C_NVC 0.0006 0.0003 0.0001 0.0005 -0.0002 0.0003
R_IPO -0.0109*** 0.0012 -0.0101*** 0.0016 -0.0149*** 0.0013
R_BFAll.. 0.0010·*· 0.0001 0.0012*** 0.0001 -0.0001 0.0001
R_BINCORP 0.0008*** 0.0000 0.0009*·* 0.0001 0.0008*** 0.0000
AV_TBILL -0.4502*** 0.0704 -0.1832 0.0999 -0.0169 0.0604
AV]RIME 1.4468**· 0.0664 1.5717*** 0.0989 0.4233**· 0.0588
_DPR 0.4567*** 0.1318 0.9743*** 0.1826 -0.3974*** 0.1191
_CSO 0.5865*** 0.1229 0.5910** 0.1954 -0.4908*** 0.1177
_COM 0.4412** 0.1405 0.6593** 0.2076 -0.0968 0.1105
_CEL 0.0896 0.3898 0.3207 0.5941 -0.0291 0.2835
_IEQ 0.3420 0.1827 0.9518*** 0.2428 -0.6015** 0.1894
_MED 0.1821 0.1771 0.3748 0.2431 0.2734* 0.1214
_TST 0.9586*** 0.2122 0.6053 0.4023 -0.3966 0.2840
_CMP 0.2247 0.2399 1.1727*** 0.2663 0.0093 0.1953
- SEM -0.4640 0.3184 0.6257* 0.3119 -0.2219 0.1748
_RET -1.0667*** 0.2812 0.3070 0.2444 0.3084 0.1637
- ENV 0.4634 0.3284 0.5387 0.5183 -0.0357 0.3220
_CHM -1.1977 0.7151 0.1585 0.7198 -0.4266 0.3609
EARLY -1.3541*** 0.1452 -1.6581*** 0.2221 -1.4462·** 0.1277
Events 639 343 955
* p < .05;·· p < .01 ; •• * P < .001. X2 = 8044.3 df= 81
CCOM (communications). Since many of the more famous recent entrepreneurial

successes come from these industries (e.g., Cisco Systems, Oracle), it is interesting
to note that success is not the rule.
A PORTFOLIO MANAGEMENT PERSPECTIVE

Suppose the impatience hypothesis is true and works primarily through the superior
deal flow of the more central VCs. Perhaps all that is happening is that more central
VCs, searching for the extraordinary performing investments, behave like options
traders. They seek highly risky investments, knowing that risk and expected return
are inversely related. To economize on time, such investors might make many high
risk investments, walking away from the losers. Of course such behavior might very
well undermine the deal flow advantages that centrality provides. It would also
damage the VC's maintenance of its network position and, consequently, the VCs
stock of social capital. In addition, Of course, for this to be true the market would
have to be characterized by long term imperfections. Otherwise, the rate of return
would exactly compensate for the heightened risk and a strategy specializing in high
risklhigh return investments would have no greater expected return than a random
investment strategy. Nevertheless, such a view should be explored because it renders
the impatience hypothesis less interesting. One would expect to see a positive effect
of centrality on the rates of acquisition and failure simply because the central actors
choose to invest in higher risk ventures.
To explore this possibility, we change unit of observation to the investment
spell, rather than studying the venture's duration. As noted above, ventures often
raise capital repeatedly in 'rounds' of investment. If an investor is pursuing a high
risk and abandonment strategy, the standard deviation of the duration of its
investment spells should be large. Very rapidly growing young ventures will either
experience an IPO quickly, or raise successive rounds of financing in quick
succession. Very poorly performing investments lead to quick exits. On the other
hand, an investor pursuing a more cautious approach will tend to choose ventures
that display steady but slower growth. Rounds come farther apart. The probability of
abandonment is lower. If it is true that centrally connected investors choose higher
risk investments, then we should observe that the standard deviation of the time
between investments in a venture, or between an investment and a liquidity event
should be greater. The mean time may not be different, but the standard deviation
should be higher.
Shifting the unit of observation in this way poses numerous challenges. In
particular, we can anticipate problems with autocorrelated disturbances, reflecting a
lack of independence between observations. For instance, if there are multiple
investors with funds in a venture, and that venture experiences a liquidity event, then
all investment spells terminate at the same time. One is not free to end
independently of another. Clearly, this problem is sufficiently different that a
satisfying analysis will require a new modeling effort. For present purposes, we
simply calculate the product-moment correlation between investor centrality and the
standard deviation of investment duration.
Since we are interested in manifestations of impatience, and a single investor
may participate in repeated rounds, we should look for the correlation in question
controlling for the patience signified by repeated investment. So we measure the
time duration for each investor putting funds into each venture, and count whether
this is the first, second, third, etc. round in which that investor has participated.
The results do not support this portfolio interpretation. That is, the correlation
between investor centrality and the standard deviation of investment duration is
negative, not positive as one would expect from this portfolio view of the problem.
It is worth noting that this is, at best, a suggestive result. It does not solve the
autocorrelation problem. Such a task, while worth the effort, goes well beyond the
purposes of this chapter or the space available in this volume.
Table S. Investor centrality and duration of investment

Zero-order r* Round 1 Round 2 Round 5 Round4 Round 5
Centrality and SD of Duration -0.139 -0.177 -0.159 -0.173 -0.194
Centrality and Mean of Duration -0.195 -0.202 -0.191 -0.204 -0.203
Mean of Centrality 118 131 145 158 167
Number of Observations 12546 6677 3157 1206 354
* P < .0001 for all coefficients
Investors participating in multiple rounds are likely to be more central than

those participating in only one or two rounds. Both the mean duration and the
standard deviation of investment duration are inversely correlated with investor
centrality. One might be tempted to think that more central investors have access to
higher quality investment opportunities or that they simply pick winners more often,
but we should recall that the effects of centrality reported in Table 2 through 4 do
not suggest such a simple explanation.
DISCUSSION
This research shows that the connections ventures have with investors do seem to
matter for the liquidity events under study. Receiving support from centrally
connected venture capitalists has the expected positive effect on the probability of an
IPO. This effect is consistent with intuition and also fits the arguments advanced
here that the superior network connections of prominent VCs help move the young
ventures to become public companies. The high quality social structure the VC has
built, thus yields social capital for the starting company. Of course, the superior deal
flow that these more prominent VCs see allows them to pick the most promising
ventures. In fact, the ability to choose well is one of the qualities that leads investors
to put capital into their funds. On the other hand, venture capital firm centrality also
increases the probability of acquisition. This suggests that the advantages of working
with a well-connected group of investors are tempered by their likely higher
impatience. This is consistent with the common view presented in the business press
that venture capital works for the entrepreneur only so long as that entrepreneur's
interests are aligned with those of the Vc. After that, the ties with the VC, that
brought the entrepreneur social capital in the past, may start to produce social
liability to the entrepreneur. This alignment depends on rapid growth. Sadly, firms
sometimes grow faster than their managers can learn.
The centrality effects of non-VC investors are generally not statistically
significant. This further supports the arguments advanced earlier in the chapter,
particularly the impatience hypothesis. If the effects of VC centrality were artifacts
of some variable such as size of fund, we would expect to see similar effects among
the funds run by universities, corporations and state governments. An important
difference is that such organizations are not driven by the same liquidity demands as
are the classically defined VCs. This is because their capital is not organized in
funds with fixed time horizons.
It is obvious that this analysis has employed only the simplest of network
techniques. There are many other centrality measures. In addition, we have not yet
explored the similarities in network connections among the investors. The effects of
industrial specialization have not been explored here, for example. One particularly
intriguing subject for further study is the tendency for some VCs to invest in early
rounds rather than later rounds of financing. There is much discussion in the
business press and the scholarly literature to the effect that such tendencies affect the
ability and Willingness of the VC to serve as a guide through the community, to
provide consulting services as well as money. In addition, some firms act as the 'lead
investor' and this affects their willingness to invest in subsequent rounds, and the
ability of others to invest in the ventures for which they have taken leadership.
Finally, the principal unit of analysis in this study is the focal venture. The
community itself may be an equally promising subject for study. In some social
systems, the mechanisms for allocating opportunity depends heavily on kinship and
ethnicity. In some circumstances, technical training is important while in others
work experience is the key to admission (Fiet 1991). The efficiency of this system
may explain why entrepreneurship happens more often, and perhaps with greater
probability of success, in some communities rather than others.
If we ask why we have entrepreneurs at all, part of the answer is bound up in the
economy of time that drives the working activities of venture capitalists. A highly
efficient opportunity structure can allocate capital, create social connections, provide
expertise, and link a new venture with suppliers and customers very quickly,
probably faster than the same functions can be performed in a large corporation.
Social structure then efficiently breeds social capital. Viewed in this way,
entrepreneurship is a set of activities that links individuals, organizations and social
context. The entrepreneurial community, an important part of this social context,
produces social capital and social liability in the form of barriers and constraints and
operates through a set of intensely personal connections. Money talks. But so do
trust, reputation, access and legitimacy. All of these factors add value, leveraging the
hard work, intelligence and good luck that lead to entrepreneurial success.
I received help from several sources in conducting the research reported here and in preparing this report.
The San Francisco company VentureOne provided the data and allowed me to use it for research
purposes. Lindy Archambeau contributed research assistance. Jerry Engel and Mario Rosati made
suggestions about the institutional realities of venture capital. Bill Barnett, Glenn Carroll and Mike
Hannan read an earlier draft. Thanks are also owed to the members of Stanford's joint
organizations/strategy workshop. Their comments helped me make numerous improvements. Of course,
my ability to generate ignorant substantive assertions and stupid technical errors remains undiminished.
•
Final Issues
csc: An Agenda for the Future
•
Shaul M. Gabbay
INTRODUCTION
Corporate social capital refers to the set of resources, tangible or virtual, that accrue
to a corporate player through the player's social relationships, facilitating the
attainment of goals. In the opening chapter to this volume, we suggested that social
structure and social capital are related, but distinct, entities. Both reside at various
levels of analysis-in particular, at the levels of the individual and the firm-and
interact with each other at these various levels of analysis.
The overarching agenda of the chapters in this volume thus pertains to the
relations between social structure and goal attainment of corporate players. Social
networks in this framework are discussed in the context of their functional role-
their positive or negative effect. We are only at the initial stages of the development
of an encompassing theory of corporate social capital; this chapter is directed
towards future steps. What the effect of corporate networks is, how to manage them
and how to avoid social liability and create social capital instead, are all questions
grounded in the study of corporate social capital. In the current chapter we will
discuss the research and practical applications for further developments. We will
highlight critical questions that, in our view, should be resolved and, in so doing, we
will open more new questions for future discussions. At its initial stages, CSC is an
emerging research agenda. This, of course, presents a wide open window of
opportunity for future and further contributions.
AN AGENDA
The study of corporate social capital is an exciting area. Using the 'lens' of corporate
social capital, we have an explicit way in which to study effective and ineffective
organization. Because we are only at the outset of an encompassing theory of the
interrelationships of social structure and corporate social capital and liability, many
fundamental questions are still unanswered. If CSC theory is to develop into a
fruitful way of studying organizations, we are challenged to address several topics
first. In this section, we will address four of these challenges. Many other issues are
also of interest, but these four are among the most critical at this stage. These four
challenges concern 1) the measurement of corporate social capital, 2) the conditions
under which social structure conveys social capital or liability, 3) the effect of time
on social structure and its outcomes, and 4), from a management-oriented
perspective, the issue of how social capital can be created and maintained. We will
briefly discuss each challenge below, posing questions rather than answering them.
The challenge of 'Measurement'

Discussing social capital is one thing; measuring it is quite another. In empirical
studies, social capital is often equated with some aspect of the social structure-for
instance, the number of ties an actor has, or the centrality I of his or her network
position. Alternatively, social capital is often measured by the existence of structural
holes in a player's network structure. 2 This is a helpful approach, because data on
network ties is relatively easy to collect and the researcher has a large body of
network measures and methodology at his disposal.
This approach, however, is not consistent with the view of social capital we
have laid out in this volume, especially in our introductory chapter. We do agree that
social capital and social structure are (strongly) associated. However, because we
explicitly regard social capital as inherent in social structure, as an outcome and
generator of social structure, it is inappropriate to equate the two. Consequently, an
alternative approach is needed to measure social capital.
This is easier said than done. If social capital resides in social structure and can
take on many different shapes, how would one measure it? Or, more appropriately,
can social capital be measured at all? Could one justifiably claim that 'I have twice
as much social capital as you'? It is not possible to provide a satisfactory answer to
these questions at this point. Finding ways to measure social capital and social
liability is one of the major challenges of esc research.
In some situations, measurement is relatively straightforward. For instance,
some organizations reward their agents based on their agents' sales volume, which,
in turn, is an indicator of the productive value of the agents' social ties. In this
fashion, firms are in fact rewarding (and measuring) their agents' social capital. 3
In most cases, it is difficult to measure social capital. In this volume, two papers
explicitly address this measurement issue. Han and Breiger concern themselves with
the question of in what aspects of social structure social capital is located. If social
capital is inherent in social structure, the question is whether there is one specific
structural characteristic that yields social capital, or whether we need to look for a
mixture of characteristics. Rather than equating structure and capital, they search for
the structural sources of social capital. Han and Breiger approach this question by
studying models that produce estimates for the number of ties between pairs of
actors, using three indicators of social capital: status, average number of ties sent
and received, and the strength of the relationship. In this approach, models and
measures are duals to each other. Han and Breiger show that their approach not only
CSC: An Agenda for the Future - 485
makes it possible to locate the sources of social capital in social structures, but that
dimensions of this structure can be separated out as well, enabling the researcher to
study their individual effects.
Doreian (this volume) argues that organizations that are well regarded by other
organizations have higher social capital. Moreover, organizations that are well
regarded by other well regarded organizations have even higher social capital. This
argument allows Doreian to use input-output methods for generating measures of
firm-level social capital. His approach allows one to assess from which part of its
network a firm derives its social capital.
The studies by Han and Breiger and Doreian provide CSC theorists with ways
of measuring aspects of social capital. But there are still many other measurement
issues to be addressed in the near future. For instance, social structure can provide a
firm with the social capital of increased learning ability, timely access to strategic
information, access to physical resources, referrals from third parties, or the
opportunity to obtain loans at a lower cost-different measures may be needed for
different outcomes. Also, social relationships are often multiplex, leading to overlap
of structures of advantage and disadvantage, which should be taken into account in
future measures of social capital. In addition, measures are required that take into
account that the same social structure can be beneficial for the attainment of one
goal, but detrimental to the achievement of another goal. Further, social structure at
the firm level and/or the individual level affects the social capital at both levels.
Measuring this interplay of levels of analysis is perhaps the most difficult challenge.
The challenge of 'Contingency'

One of the distinctive characteristics of social capital is its contingent nature. The
same social structure provides some actors with social capital, while bringing social
liability to others. The search for contingency factors can be guided by the following
question: Which social structure is benejiciaVdetrimental for whom. for which goals.
where. and when? Although the elements in this question are interrelated, we will
briefly address each of them in turn, without attempting to fully separate them.
Which social structure

The social network literature is replete with studies that discuss the structural
characteristics that create advantages and disadvantages for corporate players. In the
social networks literature, a number of structural characteristics have been
associated with success and failure.
Undoubtedly the most commonly studied structural feature in the literature is
the number of ties that reach or reach out from an actor. A large volume of ties has
been argued to provide information and control benefits. It is also seen as an
indicator of power and status. Although this view is certainly useful in a number of
cases, it is built on the implicit assumption that ties and social capital can be
equated. As argued in various places in this volume, that assumption often can not
be maintained. However, it is important to study in which situations the volume of
ties does have an effect on social capital, and when not (or to a lesser extent).
Initiated by the work of White, Boorman, and Breiger (1976) and Burt (1992) is
the idea that the study of absence of ties provides vital information about networks
that goes beyond the classical study of ties present. In particular, Burt's concept of
structural holes has inspired social capital researchers in recent years. Here also
further research is necessary. For instance, do structural holes always provide social
capital? Burt has provided many examples of situations in which structural holes can
be associated with success in a competitive arena. However, Gabbay (1995, 1997)
found that, although bridging actors may have a strategic edge, this does not mean
that they necessarily do any 'better.' Gabbay's findings suggest that only when actors
actively and consciously create structural holes-by adding ties and shaping their
network-do they take advantage of structural holes. In the context of job mobility
of R&D unit managers in Fortune 500 firms, Gabbay and Zuckerman (1998) found
that, depending on contextual factors-the degree to which the unit is oriented
toward individual rather than to collective incentives, the density of the overall
pattern of interaction in the unit, and the extent to which brokerage strategies are
legitimate-there is a varying effect of structural holes on managerial mobility.
Other structural characteristics have been studied, such as density, segregation,
and level of multiplexity-and many potential others come to mind. At present, we
do not yet have a full picture of which features of social networks really do provide
benefits or barriers. It is likely that we should broaden our focus beyond the purely
structural quality of networks and include characteristics of the nodes as well. For
example, for a medical representative, there is a clear difference in opportunity
provided by ties to general practitioners as compared to family ties. As another
example, politicians surely need strong supportive ties with their grassroots support,
but can not achieve their goals without productive ties with those in power.4
In addressing 'which social structure,' we have to search for the differential
characteristics of social structure-including the attributes of the actors comprising
the social structure-that affect the attainment of goals of corporate players. The
findings reported above open the search for alternative characteristics.
For whom
In CSC, an explicit distinction is made between structure and capital/liability at the
level of the firm and at the level of the individual member of the firm. Particular
structures may be beneficial for players at one level, but detrimental to players at
another. Therefore, in addressing issues such as 'which structure' and 'for which
goals,' one should be explicit in addressing 'for whom.'
Most research studying the effects of social structure on effective organization
does not address the issue of these interconnected levels of analysis. Rather, most
studies focus on only one level at a time. Perhaps this is all we can expect at this
point. We should be able to walk before we run. But eventually we will need
theories that explicitly address the interplay between these levels (and even refine
these levels). Theories that do so may prove instrumental in furthering our
understanding of many facets of organizations.
For which goals

Corporate players have various goals. Firm goals can include high sales, low price
of raw materials, timely access to market information, increasing its knowledge
base, the identification of potential partners or employees, creating influence on
csc: An Agenda for the Future - 487
governmental decision-making, increasing productivity, identifying market

opportUnIties, and so forth. Most of these goals require different types of
relationships, both within the firm and between firms.
Goals of firm members can include good career opportunities, job satisfaction,
high individual sales, opportunities to travel, high income, job security, status, and
so forth. Again, most of these goals require different types of relations, both within
the firm and with employees of other firms.
An example of this is provided in a study of R&D team performance (Kratzer,
van Engelen, and Leenders, 1998). R&D teams are confronted with the need to
resolve various issues. Kratzer et al. argue that, in the ideation phase of the R&D
process, a segmented network structure is advantageous to the team's ability to solve
administrative and coordination problems efficiently. However, segmentation
negatively affects the team's problem-solving capacity. The same structural
characteristic thus has a different effect on the accomplishment of two different
goals.
Additionally, corporate players often attempt to advance multiple goals
simultaneously. For the attainment of multiple goals, still other structures may be
necessary.
One of the most fundamental questions in this area is the condition under which
the goals of firm and their members coincide. Since the goals of the corporation and
the goals of its members are often different, the pursuit of the attainment of
individual goals by firm members may not further firm level goals, it may even
impede their achievement. s Just as firm level social structure is not simply an
aggregate of individual level social structure, and firm level social capital is not
simply an aggregate of individual level social capital, so too are firm level goals not
simply an aggregate of individual level goals. For example, Flap and Boxman (this
volume: 215) bring to bear the issue of 'managers who act opportunistically and
misuse corporate social structure to advance their own goals to the detriment of the
goals of the company they work for'. We thus need to search for conditions under
which individual level goals and firm level goals concur or diverge--rnore to the
point, we need to search for conditions under which social structure at the firm and
individual levels lead to the advancement of overlapping goals, or to the obtrsuction of
goals at one of these levels.
This contingency simply indicates that research on corporate social capital
should be explicit in the kind of goals sought after by the corporate players. One
cannot expect a single theory of social capital to expiain structural effects on the
furthering of every goal conceivable. We therefore need theories that are explicit in
the goals addressed, and, preferably, can hint at their value for the attainment of
alternative goals. 6
Where
The fourth contingency factor addressed here is 'where.' Most of the popular
organization research written in the English language concerns corporate players in
Western societies. However, cultural differences make it difficult to generalize
findings. In engaging in business interactions, Arabs rely strongly on the trust and
tacit understandings inherent in their relationships with their exchange partners.
Israelis, on the other hand, rely more on formal contracts and procedures. 7 In
Taiwan, where inheritance is the main organizing factor, the lion's share of
businesses have fewer than twenty employees. Japan is well known for its highly
networked keiretsu; Korea for its vertically structured chaebol. Consistent with
historically developed social structures and practices, businesses organize differently
in different cultures.
But even within the same country social structures may operate differently. Foss
(1998) found that the same structural characteristics of networks of fish farming
entrepreneurs in Norway had different effects in different areas of the country. For a
large part of this century, networks in the eastern part of The Netherlands were
relatively small, because the farming population relied on few, but strong, ties. The
business population in the Western part of the country needed large networks,
mainly to politicians and other businessmen.
It is conceivable that different organizational cultures make for different effects
of the contingency factors. This is especially important because of the strong and
propeling drive toward globalization. Ties between actors in different countries may
carry a different baggage. For example, Harland (this volume: 416) notes that
'parties to UK-based relationships expected more distance than those in Spanish
relationships studied where friendship between supplier and customer was expected.'
With the presence of cultural differences with respect to the interpretation and
expectation of social ties, one should look even deeper for the mechanisms that
translate social structure into social capital and liability.
It will be most interesting to see the direction that will be taken with future
theory that compares how various network structures affect the advancement of
goals of corporate players in different organizational cultures.
When
An anecdote tells about a consultant who did a good job in explaining to a group of
managers the merit and importance of corporate culture. One of the managers
present enthusiastically reacted with 'Sounds great, I would like that in my
corporation, can I have it by Monday morning?'
This anecdote could have been about corporate social capital. Corporate social
capital is not built overnight. In fact, it only exists by virtue of time. Trust needs
time to grow. Relationships need time to mature. Payoff often follows the social
structure after some time has passed. Benefits drawn are often not concentrated in
one point in time, but accrue to a player over a longer time span.
There are two main reasons why a time dimension needs to be incorporated into
the area of CSC research. First of all, social structure changes over time and so does
its effects on corporate players. See, for instance, the chapter by Gargiulo and
Benassi (this volume). Second, social structure both affects and is affected by the
actions of corporate players. In order to determine causality, one needs a
longitudinal approach. We will address the former issue here, and the latter in the
section 'time and causality.'
As exchanges develop, become stronger and more long-standing, trust between
the exchange partners is likely to increase. An employee who has been with a firm
for thirty years may be less likely to search for a position in another firm than an
untenured colleague. Firms that have been involved in multiple collaborations in the
past may be more likely to seek each other out for future collaboration. A shared
history makes for increased levels of trust and creates higher levels of similarity in
culture and narratives. Consequently, transaction costs are decreased. The downside
to longstanding collaboration is that it may hamper the firm's ability to learn and its
access to information on the market.
At different points in time, different structures come into play. In the initial
stages of a firm-to-firm relationship, a third party may be needed in the role of
hostage keeper, assuring both parties that vital information will not be leaked to
outsiders. 8 At later stages, the presence of a third party may no longer be required. In
their chapter in this volume, Higgins and Nohria discuss the effect of mentoring
relationships on a protege's ability to create social capital over the course of his or
her career. Their findings suggest that having a mentor early on in one's career can
be obstructive to the protege's development of social capital. However, once the
protege has developed a professional identity and reputation of his or her own,
having a mentor is helpful in facilitating the formation of new ties. In accordance
with our 'contingency question,9 above, this leads Higgins and Nohria to say that 'we
need to consider not just knowing how these relationships work, but also who and
when affect a protege's ability to progress in his or her career.'
An alternative example of the time contingency is given by Kratzer et al. (1998),
who argue that different structures are necessary for R&D team performance at
different stages in the innovation cycle. For example, controlling for other factors,
they find that team leader centrality is beneficial for the team's performance at the
development stage, but is negatively related to performance at the initial stages of
the R&D process.
The issue of when is an important contingency factor that needs to be addressed
in theories of CSc. Only then can we hope to begin to fully understand the interplay
between social structure and the attainment of goals of corporate players.
The challenge of 'Time and Causality'

The third issue on the agenda of research in the field of corporate social capital
addresses the need for dynamic specifications of social structure and social capital:
is social structure a generator of, or generated by social capital, or both? Below, we
briefly address these questions in turn.
Social capital as an outcome of social structure

Theories of social capital usually treat social structure as a generator of social
capital. Actors are viewed as enmeshed in a web of relationships. Social capital then
refers to the resources inherent in a dyadic relationship or in a set of relationships.
This premise is theoretically limiting because social structure is constantly changing
and actors can alter social structure to suit their needs.1O Firms engage in
collaboration; buyers and sellers enter into exchange with the focal firm.
Relationships are created and dissolved. Individuals move into and out of
organizations, or move between teams or departments and climb the corporate
ladder.11
When the Dutch brewery Heineken set out to expand its business abroad, Pieter
Feith, member of the board of directors, accidentally ran into John Fraser and David
Neave, two representatives of the Singapore and Straits Aerated Water Company.
With their help, Heineken established Malayan Breweries Limited and built the
Singapore Tiger Brewery. Tiger soon became one of the best selling beers in
Singapore. The joint venture was a rich source of corporate social capital for
Heineken and resulted in large sales all over Asia. Years later, Heineken set out to
promote its own brand in Singapore. Unfortunately, Fraser and Neave had created
strong ties to many crucial regional business contacts in Singapore and were more
interested in promoting Tiger beer than in promoting Heineken. This social structure
resulted in poor availability of Heineken beer in Singapore. These unwanted changes
in Heineken's social structure left the market open for European competitors
(particularly for Carlsberg which soon sold 17 times as much beer in Singapore as
did Heineken under its own brand name).12
If social structure is considered to convey social capital and social liability, the
kind and amount of social capital and liability evolves along with the social
structure. A static view of social capital merely presents a snapshot, a single frame
from a movie. For a full picture of the history and evolution of corporate social
capital, a dynamic point of view is required.
Social capital as a source of social structure

Success is attractive. Successful accountants, lawyers, consultants, and sales
representatives carry a large collection of social capital with them. When firms are
able to hire them, their fruitful relationships also (to some extent) become part of the
firm's network; in other words, the social ties of the individual now yield social
capital at the firm level.
Companies central in the social structure have timely access to resources held
by others. 13 This then draws other firms to collaborate with the central player, in
order to gain potential access to these resources (and status). The firm, in turn,
becomes even more central in the social structure. 14
Whereas success is attractive, failure surely is not. Firms that are strongly
embedded in rigid business relationships are prone to suffer from decreasing
learning ability. Such firms are unlikely candidates for generating new ties and
engaging in new ventures. The social liability inherent in the focal firm's network
both prevents it from entering into new relationships, and may even cause its current
partners to disinvest in the mutual tie.
Social structure evolves with social capitaVIiability. There are only few studies
that deal with this causal direction. For the understanding of how social structure is
related to performance of firms and its members, studies in this area are imperative.
Co-evolution of social capital and social structure

If social structure is considered to convey social capital and social liability, the kind
and amount of social capital and liability co-evolves with the social structure.
The social capital (liability) of a focal player may draw other players to initiate
(dissolve) relationships. By creating an additional tie in the social structure, a player
hopes to gain access to resources held by the focal firm that may otherwise be
unavailable. The tie thus conveys social capital. The change in network structure
follows from purposeful actions of players in their quest for social capitaI-in order
to attain their individual goals.
The character of relationships also changes over time. First, relationships
require time to be built and go through various stages. IS The social capital and social
liability inherent in the relationship permutes over these stages. Second, the
multiplex character of most social relationships draws attention to how the contents
of relationships change over time. Two partner firms may move from co-
development of a new product to joint marketing of this product, while collaborating
in the development of a second product. These changes in content, often difficult to
observe for an outside researcher, make for various kinds and amounts of social
capital derived from the collaboration. In order to capture the dynamics of corporate
social capital, a longitudinal point of view is instrumental.
From a methodological point of view, there is an additional reason to collect and
study longitudinal data on social structure and social capital. If social structure and
social capital co-evolve, one researcher can always empirically claim that social
capital is a source of social structure, whereas another researcher can empirically
claim, with the exact same data, that social capital was the generator of social
structure. 16 Co-evolution of social structure and social capital make causal
statements, based on static data, irrelevant and untenable. Only with the help of a
longitudinal research design can one disentangle l7 the interrelationship of structure
and capital.
Members of the firm engage in relationships with outside constituencies that
may, over time, translate into social capital at the firm level. These ties may, over
time, become institutionalized at the firm level. Other individuals within the firm
may then draw from these relationships and extract social capital at the individual
level. Change over time of social structure and social capital cuts across the
analytical levels we distinguish in this volume -the firm and its members. Grasping
this longitudinal crossover is perhaps the biggest challenge in CSC theory
development and research.
The challenge of 'Social Capital Management'

Once we have developed a notion of the contingencies affecting how social structure
and social capital are related, and have put into place the theory and methodology
for dynamic representations, the next part of the agenda refers to the practical
application of the approach. Slightly rephrasing the words of Kanter and Eccles
(1992: 527): it is from an action perspective that managers derive their interest in
corporate social capital. The kind of knowledge that can be made available through
studying the interplay of network structure and social capital and liability at the firm
and individual level needs to be supplemented with knowledge about how to build
and use networks in order to create (maximum) social capital.
'Corporate social capital management' refers to the purposeful alteration of
social structure to fit players' goals. The corresponding questions are: how can
social structure be changed to suit the attainment of goals, and what strategies do
actors use to invest, reinvest, or disinvest in their social ties?
Although Burt (1992) empirically follows the 'social capital as an outcome of

social structure' approach, structural holes theory suggests that actors can change
their networks to make them more efficient and effective. This is manifested in the
theory's explicit prescription of investing in non-redundant ties and disinvesting in
relationships that are redundant. Burt implicitly calls for strategically alternating
social networks to maximize efficiency.
To reiterate, management of corporate social capital can be discussed at the
different levels of analysis central in this volume.
Structure at the firm level

Over the past few decades, industry after industry has become globalized. The
emerging global marketplace poses the challenge to firms of spreading their value
activities worldwide, in pursuit of simultaneously realizing the benefits of low cost
and differentiation. In this pursuit firms increasingly enter into strategic alliances, as
it is often faster and more efficient to work through external alliances than to try to
change the organizational mandates of subsidiaries. 18 This is but one example of the
host of interorganizational ties into which firms enter. 19 Firms engage in
interorganizational relationships in order to secure resources, gain access to new
markets, lower costs, and stay abreast of market conditions. 2O Careful selection is of
utmost importance and will determine whether and how much social capital can be
drawn from alliance relationships.21
Firms can also actively engage in inter-firm relationships with the goal of
increasing the social capital of their employees. For instance, firms sometimes
engage in international collaboration which, at least as a side-goal, gives the firm's
employees the possibility to travel and be involved in international assignments. In
the computer industry, in order to be able to provide their members with various
kinds of employee benefits, many businesses nowadays engage in relationships with
car manufacturers, insurance companies, mortgage companies, and so forth.
Potential emplo~ees then partly base their choice of employers on the benefits
offered to them. 2
Structure at the individual level

Individual members of firms can also engineer the social structure so as to increase
their social capital. For example, creating ties with mentors can be beneficial to
one's career. Creation and maintainance of social ties may be instrumental to
acquiring information about job opportunities. 23 Gabbay (1995, 1997) showed that
the individuals who are most successful in network marketing are exactly those who
initially were at an extreme social structural disadvantage and engineered their
social structure to suit their goals.
Individuals may also alter their social ties so as to further firm-level goals. For
example, studying dependence relations, political alliances, and confidential
discussion networks among decision makers in a cooperative Agri-business,
Gargiulo (1993) shows that leaders build ties of interpersonal obligation with people
directly affecting their performance in the organization. When policies' divergence
or personal friction makes these ties untenable, leaders build strong cooperative
relationships with people who may constrain the performance of the party on whom
they depend.24
Baker and Obstfeld (this volume) devote their chapter to the description of two
strategies an entrepreneur can employ to access and mobilize social capital from
social structure. In the disunion strategy, the social entrepreneur bridges between
two disconnected alters. In the union strategy, on the other hand, ego closes the gap
between two alters who are disconnected or in conflict. Baker and Obstfeld are
convincing in their discussion of the conditions under which these strategies are
beneficial to ego-their chapter is valuable for the study of social capital
management. Baker and Obstfeld's chapter focuses on strategies ego can employ
with respect to two alters, but we also need to look for the opportunities ego can
create in larger systems. By both studying the opportunities ego can create and
maintain in a subset of a network, and those within larger systems, we can start to
understand how corporate social capital can be managed, both for firms and
individuals.
The agenda on corporate social capital management takes the academic work on
CSC to a practical level. The agenda embodies the search for how social capital can
be created and maintained, and which factors playa role in doing so.
To the fundamental question we posed earlier: What social structure is
beneficiaVobstructive for whom, for what goals, where and when?-we now add:
'how do we get and stay there?'
CONCLUSION
The 45 scholars who contributed to this volume have presented the reader with a
thorough overview of the area of Corporate Social Capital. The wave of recent
papers in various literatures employing a social capital framework highlights the
accumulation of critical mass for the study and implications of social capital in
general and corporate social capital in particular. CSC is only at its infancy stage. It
will take additional research to bring it to adolescence.
The prominence of the authors of the chapters in this volume speaks for the
merit CSC has in the study of organizations. We hope that this volume has inspired
the reader, and that the study of Corporate Social Capital will prove fruitful in our
understanding of organizations.
NOTES
I. The number of ties is of course also a measure of centrality. We mention them here separately
because the number of ties is often used in research without reference to centrality.
2. Burt (1992).
3. At first sight, this may seem helpful, but still presents a lot of difficulty. For instance, besides the
social capital the individual brings to the firm, reward systems often also depend on other characteristics.
Some of these are idiosyncratic to the employee; many of them are determined by organizational culture
and history, and the lack of flexibility in readjusting reward systems to new conditions. For instance, see
Van Veen (1997).
4. Leenders (1991).
5. O' Aveni(199I).
6. A related issue, beyond the scope of this chapter, is the issue of how the attainment of these goals
can be measured. This issue in itself is the focus of a large body of research.
7. Gabbay, Semyonov, and Mishal (1999).

S. See Nooteboom's chapter in this volume.
9. What social structure is beneficiallobstructive for whom, for what goals, where, and when?
10. Amomg others, Gabbay and Sato (1996), Gabbay (1995, 1997), Leenders (1995ab, 1996).
11. Han (1996) shows that most contacts between firm members occur within divisions, rather than
between them. Contacts across divisions are almost exclusively concentrated in the higher echelons.
Individuals with higher formal positions tend to have denser personal networks. In addition, Han shows
that hierarchical position affects the extent to which an individual initiates interactions, rather than
receiving them. These findings indicate that an individual's social network changes as she advances in her
career. In the 'social capital as an outcome of social structure' approach, Han's findings suggest that an
individual's social capital changes with the individual's professional advancement.
12. Smit (1996).
13. Cf. Stuart (this volume).
14. Also see Freeman (this volume).
15. See Zajac and Olson (1993) and Omta and Van Rossum (this volume).
16. Leenders (1997).
17. Leenders (1995a, 1995b) calls this the 'untying of the knot: He extensively discusses models and
approaches for the longitudinal study of networks.
IS. Yoshino and Rangan (1995).
19. Also, see Knoke (this volume).
20. In the context of states as actors, Gabbay and Stein (1999) argue for the creation of Regional Social
Capital in the context of creating technological interdependencies which will foster social relationships in
the context of peace in the Middle East. They show how it is in the interest of the third party-the USA-
to foster regional social capital by forming technological infrastructurai dependencies.
21. Also seeOmta and Van Rossum (this volume) on this topic.
22. One of the major Dutch automation companies recently held its job interviews in showrooms of car
dealers. Job candidates, who were informed of the outcome immediately after the interview, could pick
out a car to their liking on the spot and drive home in it. In this fashion, the company tried to lure able
new employees to fill its vacancies.
23. See, for instance, the chapter by Flap and Boxman.
24. Gargiulo (1993:1).
References
•
Abel, R. L.
1989 American Lawyers. New York: Oxford University Press.
Abernathy, W.
1978 The Productivity Dilemma. Baltimore: John Hopkins University Press.
Abolafia, M.
1996 Making Markets: Opportunism and Restraint on Wall Street. Cambridge: Harvard University
Press.
Abrahamson, E.
1991 'Managerial Fads and Fashions.' Academy of Management Review 163: 586-612.
Acevado, M., E. M. Basualdo, and M. Khavisse
1990 Quien as Quien?: Los Duenos del Poder Eeconomico (Argentina 1973-1987). Buenos Aires:
Editora.
Adams,J.S.
1976 The Structure and Dynamics of Behavior in Organizational Boundary Roles.' In M. D. Dunette
(ed.), Handbook of Industrial and Organizational Psychology, 1175-1199. Chicago: Rand
McNally.
Adler, P. S.
1993 'The 'Learning Bureaucracy': New United Motor Manufacturing, Inc.' In L. L. Cummings and B.
M. Staw, Research in Organizational Behavior 15: 111-194.
Agresti,A.
1990 Categorical Data Analysis. New York: Wiley.
Ahuja,G.
1998 'Collaboration Networks, Structural Holes, and Innovation: A Longitudinal Study.' Best Paper
Proceedings, National Meetings of the Academy of Management, San Diego.
Albert, M. B., D. Avery, F. Narin, and P. McAUister
1991 'Direct Validation of Citation Counts as Indicators of Industrially Important Patents.' Research
Policy 20: 251-259.
Aldrich, H. E.
1979 Organizations and Environments. Englewood Cliffs: Prentice-Hall.
Aldrich, H. E., and C. Zimmer
1986 'Entrepreneurship Through Social Networks.' In D. L. Sexton and R. W. Smilor (eds.), The Art and
Science of Entrepreneurship, 3-23. Cambridge: Ballinger.
Alford,R.
1975 Health Care Politics. London: University of Chicago Press.
Allison, G. T.
1971 Essence of Decision. Boston: Little Brown.
Allport, F. H.
1933 Institutional behavior. Chapel Hill: University of North Carolina Press.
Amason, A. C., and D. M. Schweiger
1994 'Resolving the Paradox of Conflict, Strategic Decision Making, and Organizational Performance.'
International Journal of Conflict Management 5: 239-253.
Ammer,D.S.
1968 Materials Management. Homewood: Irwin.
Amsden, P. H.
1989 Asia's Next Giant: South Korea and Late Industrialization. New York: Oxford University Press.
Anderson, P.
1995 'The Microcomputer lndustry.' In G. R. Carroll and M. T. Hannan (eds.), Organizations in
Industry: Strategy. Structure. and Selection. 37-58. New York: Oxford University Press.
Andrews, K. R
1971 The Concept of Corporate Strategy. Homewood: Irwin.
Anheier, H. K., J. Gerhards, and F. P. Romo
1995 'Forms of Capital and Social Structure in Cultural Fields: Examining Bourdieu's Social
Topography.' American Journal of Sociology 100: 859-903.
Ansoff, H. I.
1965 Corporate Strategy. New York: McGraw-Hill.
Aold M., B. Gustafsson and O. E. WilUamson (eds.)
1990 The Firm as a Nexus of Treaties. UK: Sage.
Arnold, B. L., and F. M. Kay
1995 'Social Capital, Violations of Trust and the Vulnerability of Isolates: The Social Organization of
Law Practice and Professional Self-Regulation. ' International Journal of the Sociology of Law 23:
321-46.
Ashkenas, R, D. Ulrich, T. Jick, and S. Kerr
1995 The Boundaryless Organization: Breaking the Chains of Organizational Structure. San Francisco:
Jossey-Bass.
Axelrod, R M.
1980 'Effective Choice in the Prisoner's Dilemma.' Journal of Conflict Resolution 24: 3-25.
Axelsson, B., and G. Easton (eds.)
1992 Industrial Networks: A New View of Reality. London: Routledge.
Baba, v.
1990 'Methodological Issues in Modeling Absence: A Comparison of Least Squares and Tobit
Analyses.' Journal of Applied Psychology 75: 428-432.
Bailey, W..J., R Masson and R Raeside
1996 'Choosing Successful Technology Development Partners: A Best Practice Model.' Proceedings of
the 6th International Forum on Technology Management: 271-83.
Baker, G., M. Gibbs, and B. Holmstrom
1994 The lnternal Economics of the Firm: Evidence from Personnel Data.' The Quarterly Journal of
Economics 109: 881- 919.
Baker, W.E.
1984 The Social Structure of a National Securities Market.' American Journal of Sociology 89: 775-
811.
1990 'Market Networks and Corporate Behavior.' American Journal of Sociology 96: 589-625.
1992a The Network Organization in Theory and Practice.' In N. Nohria and R. G. Eccles (eds.),
Networks and Organizations: Structure, Form. and Action, 397-429. Boston: Harvard Business
School Press.
1992b 'Competition, Cooperation, and Collusion: Three Faces of Economic Behavior.' Paper presented
at the joint annual meetings of the Public Choice Society and Economic Science Association, New
Orleans.
1994a Networking Smart: How To Build Relationships for Personal and Organizational Success. New
York: McGraw-Hili.
1994b The Paradox of Empowerment' Chief Executive 93: 62-65.
Baker, W. E., and R R Faulkner
1991 'Strategies for Managing suppliers of Professional Services.' California Management Review 33:
33-45.
1993 The Social-Organization of Conspiracy-megal Networks in the Heavy Electrical-Equipment
lndustry.' American Sociological Review 58: 837-860.
Baker, W. E ., R R Faulkner, and G. A. F'lSher
1998 'Hazards of the Market: The Continuity and Dissolution of lnterorganizational Market
Relationships.' American Sociological Review 63: 147-177.
Bandura,A.
1977 'Self-Efficacy: Towards a Unifying Theory of Behavioral Change.' Psychology Review 84: 191-
215.
References - 497
Bantel, K. A., and S. E. Jackson

1989 'Top Management and Innovations in Banking: Does the Composition of the Top Team make a
Difference?' Academy of Management Journal 10: 107-124.
Barley, S., J. Freeman, and R. Hybels
1992 'Strategic Alliances in Commercial Biotechnology.' In N. Nohria, and R. G. Eccles (eds.),
Networks and Organiwtions: Structure, Form, and Action, 311-347. Boston: Harvard Business
School Press,
Barnett, W. P.
1997 'The Dynamics of Competitive Intensity.' Administrative Science Quarterly 42(1): 128-160.
Barney, J. B.
1991 'FllTII Resources and Sustained Competitive Advantage.' Journal of Management 17(1): 99-120
Barney, J. B., and M. H. Hansen
1994 'Trustworthiness As a Source of Competitive Advantage.' Strategic Management Joumall5: 175-
190.
Baron,J. N.
1984 'Organizational Perspectives on Stratification. ' Annual Review of Sociology 10: 37-69.
Baron, J. N., A. Davis-Blake, and W. T. Bielby
1986 'The Structure of Opportunity: How Promotion Ladders Vary within and Among Organizations.'
Administrative Science Quarterly 31: 248-273.
Baron, J. N., F. Dobbin, and P. D. Jennings
1986 'War and Peace: The Evolution of Modem Personnel Administration in U.S. Industry.' American
Journal of Sociology 92: 350-383.
Baron, J. N., P. D. Jennings, and F. Dobbin
1988 'Mission Control? The Development of Personnel Systems in U.S. Industry.' American
Sociological Review 53: 497-514.
Barron, D. N.
1992 'The Analysis of Count Data: Overdispersion and Autocorrelation.' Sociological Methodology 22:
179-220.
Barry,C.B.
1994 'New Directions in Research on Venture Capital Finance. ' Financial Management 23: 3-15.
Bartlett, C. A., and S. Gboshal
1989 Managing Across Borders: The Transnational Solution. Boston: Harvard Business School Press.
Bates, R.H.
1988 'Contra Contractarianism-Some Reflections on the New Institutionalism.' Politics and Society 16:
387-401.
Batson, C. D.
1990 'How Social an Animal: The Human Capacity for Caring.' American Psychologist 45: 336-346.
Bauer, M., and B. Bertin-Mourot
1991 Les 200' en France et en Allemagne. Deux modeles de ditection-selectionjonnation de dirigeants de
grands entreprises. Paris: CNRS and Hedrick and Struggles.
Bawn, J. A. C., and C. Oliver
1991 'Institutional linkages and organizational mortality.' Administrative Science Quarterly 36: 187-
218.
1992 'Institutional Embeddedness and the Dynamics of Organizational Populations'. American
Baumol, W.J.
1967 Business Behavior: Value and Growth. New York: MacMillan.
Bearman,P.
1997 'Generalized Exchange.' American Journal of Sociology 102: 1383-1415.
Becker, G. S.
1964 Human Capital. New York: Columbia University Press.
Belliveau, M. A., C. A. O'Reilly, m, and J. B. Wade
1996 'Social Capital at the Top: Effects of Social Similarity and Status on CEO Compensation.' Academy of
Management Journal 6: 1568-1593.
Belons, R. S.
1989 The Contingent Economy: The Growth of the Temporary, Part-Time, and Subcontracted
Workforce. Washington: National Planning Association.
Benassi,M.
1993 Dalla Gerarchia alia Rete: Modelli ed Esperienze Organizzative. Milano: Stas.
Ben-Porath, Y.
1980 'The F-Connection: Families, Friends, and Fmns and the Organization of Exchange.' Population
and Development Review 6: 1-30.
Berger, J., N. G. Noorderhaven and B. Nooteboom
1995 'Detenninants of Supplier Dependence.' In J. Groenewegen, C. Pitelis, and S. E. Sjostrand (eds.), On
Economic Institutions. Aldershot: Edward Elgar.
Berger, P., and Th. Luckman
1967 The Social Construction of Reality. New York: Doubleday.
Berg1ov,E.
1994 'Ownership of Equity and C01J>Orate Governance-The Case of Sweden.' In T. Baums, R. M.
Buxbaum, and K. J. Hopt (eds.), Institutional Investors and Corporate Gove17Ulllce. Berlin: Walter
de Gruyter.
Berk,R.A.
1983 'An Introduction to Sample Selection Bias in Sociological Data.' American Sociological Review
48: 386-398.
Berry, L. L., and A. Parasuraman
1991 Marketing Services: Competing Through Quality. New York: Free Press.
Berscheid, E., and E. H. Walster
1969 Interpersonal Attraction. Reading: Addison-Wesley.
Bhagwati, J.
1982 'Directly, Unproductive Profit Seeking (DUP) Activities.' Journal of Political Economy 90: 988-
1002.
Bian, Y.
1997 'Getting a Job Through a Network of Guanxi in China' In B. Wellman (ed.), Networks in the
Global Village . Boulder: Westview.
Bianchi,M.
1994 'Hayek's Spontaneous Order: The 'Correct' versus the 'Corrigible' Society.' In 1. Bimer and R. Van
Zijp (eds.), Hayek, Coordination and Evolution, 232-251. London: Routledge.
Biemans, W. G.
1992 Managing Innovation within Networks. London: Routledge.
Bills,D.B.
1988 'Educational Credentials and Promotions: Does Schooling Do More than Get You in the Door?'
Sociology of Education 61: 52-60.
Bioscan
1988-1996 Phoenix: Oryx Press.
Birley,S.
1985 'The Role of Networks in the Entrepreneurial Process.' Journal of Business Venturing 1: 107- 117.
Bishop, Y. M. M., S. E. Fienberg, and P. W. Holland
1975 Discrete Multivariate Analysis: Theory and Practice. Cambridge: Harvard University Press
Bitner, M. J., J. D. Nyquist, and B. H. Booms
1985 'The Critical Incident as a Technique for Analysing the Service Encounter.' In T. M. Block, G. D.
Upah, and V. A. Zeitharnl (eds.), Services Marketing in a Changing Environment, American
Marketing Association.
Bitner, M. J., B. M. Booms, and M. S. Tetreault
1990 'The Service Encounter: Diagnosing Favorable and Unfavorable Incidents.' Journal of Marketing
54: 71-84.
Blackwell, N., and A. Santomero
1982 'Bank credit rationing and the customer relation.' Journal of Monetary Economics 9: 121-129.
Blau, J. R., and R. D. Alba
1982 'Empowering Nets of Participation.' Administrative Science Quarterly 27: 363-379.
Blau,P.M.
1955 The Dynamics of Bureaucracy. Chicago: University of Chicago Press.
1964 Exchange and Power in the Social Life. New York: Wiley.
1977 Inequality and Heterogeneity: A Primitive Theory of Social Structure. New York: Free Press.
References - 499
Bleakley, F. R.
1995 'Strange Bedfellows: Some Companies Let Suppliers Work on Site and Even Place Orders.' Wall
Street Journal (January 13). New York: Dow Jones & Company. Inc.
Bleeke, J., and D. Ernst
1993 Collaborate to Compete: Using Strategic Alliances and Acquisitions in the Global Marketplace.
New York: Wiley.
1995 'Is Your Strategic Alliance Really a Sale?' Harvard Business Review 73: 97-105.
Blois, K.J.
1972 'Vertical Quasi-Integration.' Journal of Industrial Economics 20: 253-272.
Blossfelcl, H., and G. Rohwer
1995 Techniques of Event History Modeling. Mahwah: Lawrence Erlbaum Associates.
Bonacich, E.
1973 'A Theory of Middleman Minorities.' American Sociological Review 38: 583-594.
Bonacich, P.
1972 'Technique for Analyzing Overlapping Memberships.' In H. Costner (ed.), Sociological
Metlwdology 1972,176-185. San Francisco: Jossey-Bass.
1987 'Power and Centrality: A Family of Measures.' American Journal of Sociology 92: 1170-1182.
Borgatti, S. P.
1997 'Structural Holes: Unpacking Burt's Redundancy Measures.' Connections 20: 35-38.
Borgatti, S. P., M. G. Everett, and L. C. Freeman
1992 UCINET N Version 1.00. Columbia: Analytic Technologies.
Boulding W., A. Kalra, R. Staelin, and V. A. Zeithaml
1993 'A Dynamic Process Model of Service Quality: From Expectations to Behavioral Intentions.'
Journal of Marketing Research 30: 7-27.
Bourdieu, P.
1972 Esquisse d'une Theorie de la Pratique. Precedee de Trois Etudes d'Ethnologie Kabyle. Geneve:
Droz.
1980 'Le Capital Social: Notes Provisoires.' Actes de la Recherche en Sciences Sociales 3: 2-3.
1986 'The Forms of Capital.' In I. O. Richardson (ed.), Handbook of Theory and Research for the
Sociology of Education, 241-258. New York: Greenwood Press.
1994 Raisons Practiques. Sur la Theorie de [,Action. Paris: Editions du Seuil.
Bourweu, P., and M. de Saint Martin
1982 'Le patronat.' Actes de la Recherche en Sciences Sociales 32: 3-82.
Bourdieu, P., and L. J. D. Wacquant
1992 An Invitation to Reflexive Sociology. Chicago: University of Chicago Press.
Boxman, E. A.W., P. M. De Graaf, and H. D. Flap
1991 'The hnpact of Social and Human Capital on the Income Attainment of Dutch Managers.' Social
Networks 13: 51-73.
Bradach, J. L., and R. G. Eccles
1989 'Price, Authority, and Trust: From Ideal Types to Plural Forms.' Annual Review of Sociology 15:
97-118.
Bradburn, N. M.
1969 The Structure of Psychological Well-Being. Chicago: Aldine.
Brass, D.J.
1981 'Structural Relationships, Job Characteristics, and Worker Satisfaction and Performance.'
1984 'Being in the Right Place: A Structural Analysis of Individual Influence in an Organization.'
1985a 'Men's and Women's Networks: A Study of Interaction Patterns and Influence in an Organization.'
Academy of Management Journal 28: 327-343.
1985b 'Technology and the Structuring of Jobs: Employee Satisfaction, Performance, and Influence.'
Organiwtional Behavior and Human Decision Processes 35: 216-240.
1992 'Power in Organizations: A Social Network Perspective.' Research in Politics and Society 4: 295-
323.
1995a 'Creativity: It's all in Your social network.' In C. M. Ford and D. A. Gioia (eds.), Creativity Action
in Organiwtions, 94-99. London: Sage.
1995b 'A Social Network Perspective on Human Resources Management.' In O. Ferris (ed.), Research in
Personnel and Human Resources Management, 39-79. Greenwich: JAI Press.
Brass, D. J., and M. E. Burkhardt

1992 'Centrality and Power in Organizations.' In N. Nohria and R. G. Eccles (eds.J, Networks And
Organiwtions: Structure, Form, and Action. Boston: Harvard Business School Press.
1993 'Potential Power and Power Use: An Investigation of Structure and Behavior.' Academy of
Brass, D. J., K. D. Butterfield, and B. C. Skaggs
1998 'Relationships And Unethical Behavior: A Social Network: Perspective.' Academy of Management
Review 23: 14-31.
Breiger, R. L.
1990 'Social Control and Social Network:s: A Model from Georg Simmel.' In C. Calhoun, M. W. Meyer,
and W. R. Scott (eds.), Structures of Power and Constraint: Papers in Honor of Peter M. Blau,
453-476. New York:: Cambridge University Press.
1995 'Social Structure and the Phenomenology of Attainment.' Annual Review of Sociology 21: 115-
136.
Breiger, R. L., and J. G. Ennis
1997 'Generalized Exchange in Social Network:s: Statistics and Structure.' VAnnie sociologique 47: 73-
88.
Breiger, R. L., and J. M. Roberts, Jr.
1998 'Solidarity and Social Network:s.' In P. Doreian and T. 1. Fararo (eds.), The Problem of Solidarity:
Theories and Models. New York:: Gordon and Breach.
Breton, A., and R. Wintrobe
1982 The Logic of Bureaucratic Conduct. An Economic Analysis of Competition, Exchange, and
Efficiency in Private and Public Organiwtions. New York:: Cambridge University Press.
Bridges, W. P., and W. J. Villemez
1991 'Employment Relations and the Labor Mark:et: Integrating Institutional and Mark:et Perspectives.'
American Sociological Review 56: 748-764.
Brockhaus, R. H.
1982 The Psychology of the Entrepreneur.' In C. A. Kent, D. L. Sexton, and K. H. Vesper (eds.),
Encyclopedia of Entrepreneurship, 39-56. Englewood Cliffs: Prentice·HaIl.
Brockhaus, R. H., and P. S. Hurwitz
1986 'The Psychology of the Entrepreneur.' In D. L. Sexton and R. W. Smilor (eds.), The Art and Science
of Entrepreneurship. Cambridge: Ballinger.
Brogowicz A. A., L. M. Delene, and D. M. Lyth
1990 'A Synthesised Service Model with Managerial hnplications.' International Journal of Service
Industry Management 1: 27-46.
Brown, J. S., and J. P. Duguid
1991 'Organizationall.eaming and Communities of Practice. ' Organiwtion Science 2. (reprinted in M. D.
Cohen & L. S. Sproull (eds.) 1996, Organiwtional learning, 58-82. London: Sage.)
1997 Organizing knowledge. Paper presented at the First Berkeley Forum, UC Berk:eley.
Brown, S. W., and T. A. Swartz
1989 'A Gap Analysis of Professional Service Quality.' Journal of Marketing 53: 92-98.
Bruce M., F. Leverick, D. Littler, and D. Wilson
1995 'Success Factors for Collaborative Product Development: A Study of Information and
Communication Technology. ' R&D Management 25: 33-45.
Briiderl, J., and P. Preisendiirfer
1997 'Network: support and the success of newly founded business.' Journal of Small Business Economics
9: 1-13.
Bruno, A. V., and A. C. Cooper
1982 'Patterns of Development and Acquisitions for Silicon Valley Start-ups.' Technovation 1: 15-37.
Brusco,S.
1982 The Emilian Model: Productive Decentralization and Social Integration.' Cambridge Journal of
Economics 6: 167-184.
Bueno de Mesquita, B., and F. N. Stokman
1994 European Community Decision Making . New Haven: Yale University Press.
Bulder, B., F. Leeuw, and H. F1ap
1996 'Networks and Evaluating Public-Sector Reforms.' Evaluation 2: 261-276.
Burbidge, J. L.
1961 The New Approach to Production.' Production Engineer 40: 769-784.
References - 501
Burchell, B., and F. Wilkinson

1997 'Trust, Business Relationships and the Contractual Environment.' Cambridge Journal of
Economics 21: 217-237.
Burgelman, R. A.
1983a 'Corporate Entrepreneurship and Strategic Management: Insights from a Process Study.'
Management Science 1349-1364.
1983b 'A Process Model of Internal Corporate Venturing in the Diversified Major Firm.' Administrative
Science Quanerly 28: 223-244.
1991 'lntraorganizational Ecology of Strategy Making and Organizational Adaptation: Theory and Field
Research.' Organization Science 2: 239-262.
Burkhardt, M. E., and D. J. Brass
1990 'Changing Patterns or Patterns of Change: The Effect of a Change in Technology on Social
Network Structure and Power.' Administrative Science Quanerly 35: 104-127.
Burt, R. S.
1982 Toward a Structural Theory of Action. New York: Academic Press.
1992 Structural Holes: The Social Structure of Competition. Cambridge: Harvard University Press.
1996 'Entrepreneurs, Trust and Third Parties: Paper presented at the conference Shared Cognition in
Organizations, Northwestern University, November.
1997 'The Contingent Value of Social Capital.' Administrative Science Quanerly 42: 339-365.
Burt, R. S., S. M. Gabbay, G. Holt, and P. Moran
1994 'Contingent Organization as a Network Theory: The Culture-Performance Contingency Function:
Acta Sociologica 37: 345-370.
Burt, R. S., and M. Knez
1995 'Kinds of Third-Party Effects on Trust.' Rationality and Society 7: 255-292.
Burt, R. S., and D. Ronchi
1990 'Contested Control in a Large Manufacturing Plant.' In J. Weesie and H. D. Flap (eds.), Social
Networks through Time, 121-157. Utrecht: Isor.
Burt, R. S., and I. Talmud
1993 'Market Niche'. Social Networks 15: 133-149.
Business Week
1986 'And Now, The Post-Industrial Corporation: March 3: 64-71.
Bygrave, W.D.
1988 'The Structure of Investment Networks of Venture Capital Firms.' Journal of Business Venturing
3: 137-157.
Bygrave, W. D., and J. A. Timmons
1992 Venture Capital at the Crossroads. Cambridge: Harvard Business School Press.
Cabo, P. G., W. van Rossum, V. Mangematin, D. Vinck, and P. Saviotti
1996 Academidlndustrial Research Collaboration in the Biotechnology Programme of the European
Community (1992-1994J. Report for the European Community DO XII.
Cadotte E. R., R. B. Woodruff, and R. L. Jenkins
1987 'Expectations and Norms in Models of Consumer Satisfaction: Journal of Marketing Research 24:
305-314.
CampbellA.
1996 'The Agile Enterprise: Assessing the Technology Management Issues.' Proceedings of the 6th
Internationa"r Forum on Technology Management. 197-211.
Campbell, K. E., P. V. Marsden, and J. S. Hurlbert
1986 'Social Resources and Socioeconomic Status.' Social Networks 8: 97-117.
Cannon, W. B.
1932 The Wisdom of the Body. New York: Norton.
Cappelli, P., and W. F. Cascio
1991 'Why Some Jobs Command Wage Premiums: A Test of Career Tournament and Internal Labor
Market Hypotheses.' Academy of Management Journal 34: 848-868.
Cappelli, P., L. Bassi, H. Katz, D. Knoke, P. Osterman, and M. Useem
1997 Change at Work. New York: Oxford University Press.
Carlisle J. A., and Parker R. C.
1989 Beyond Negotiation: Customer Supplier Relationships. UK: Wiley.
Carroll, G. R., and A. C. Teo
1996 'On the Social Networks of Managers.' Academy of Management Journal 39: 421-440.
Caruso, R. E.
1992 Mentoring and the Business Environment: Asset or Liability? Vermont: Dartmouth.
CASSIS
1987-1996 U.S. Patent and Trademark Office. Washington: U.S. Department of Commerce.
CasteUs,M.
1996 The Rise of the Network Society. Oxford: Basil Blackwell.
Chamberlain, N. W.
1968 Enterprise and Environment. The Firm in Time and Place. New York: McGraw-HilI.
Chandler, A.
1977 The Visible Hand. Cambridge: Harvard University Press.
Chatterji, D.
1996 'Accessing External Sources of Technology.' Research Technology Management 39: 48-56.
Chesbrough, H. W., and D. J. Teece
1996 'When is Virtual Virtuous? Organizing for Innovation.' Harvard Business Review 74: 65-73.
Chiesa, V., and R. Manzini
1996 'Profiting from the Virtual Organization of Technological Innovation: Suggestions from an
Empirical Study. ' Proceedings of the 6th International Forum on Technology Management, 211-
27.
Cho, D. S., S. Nam, and L. Tung
1998 Korean Chaebols. New York: Oxford University Press.
Christopher, M. G.
1992 Logistics and Supply Chain Management. London: Pitman Publishing.
Chung,S.
1996 'Performance Effects of Cooperative Strategies Among Investment Banking Firms: A Loglinear
Analysis of Organizational Exchange Networks.' Social Networks 18: 121 -148.
Chung, S., H. Singh, and K. Lee
1995 Comparative Capabilities and Reputational Capital as Drivers of Strategic Alliance Formation.
Nov. 1995. Snider Entrepreneurial Center, The Wharton School, University of Pennsylvania.
Cicourel, A.
1973 Cognitive Sociology, Harmonsworth: Penguin Press.
Clark.D.
1996 'Software Firm Fights to Remake Business After lll-Fated Merger.' Wall Street Journal, January
12: AI.
Clark. K. B., and T. Fujimoto
1991 Product Development Performance: Strategy, Organization and Management in the World Auto
Industry. Boston: Harvard Business School Press.
Clawson,J.
1980 'Mentoring in Managerial Careers.' In C. B. Derr (ed.), Work, Family and the Career, 144-165.
New York: Praeger.
Clegg,S. R.
1990 Modem Organizations: Organization Studies in the Postmodem World. London: Sage.
C1ogg, C. C., and E. S. Shihadeh
1994 Statistical Modelsfor Ordinal Variables. Thousand Oaks: Sage.
Clutterbuck. D.
1985 Everyone Needs a Mentor. London: Institute of Personnel Management.
Cm3807
1997 The New NHS: Modem, Dependable. London: HMSO.
Coase,R.H.
1937 'The nature of the firm.' Economica 4: 386-405.
Cohen, M. D., J. G. March, and J. P. Olsen
1972 'A Garbage Can Model of Organizational Choice.' Administrative Science Quarterly 17: 1-25.
Cohen, Y., and J. Pfeffer
1986 'Organizational Hiring Standards.' Administrative Science Quarterly 31: 1-24.
Coleman, J. S.
1986 'Norms as Social Capital.' In G. Radnitzky and P. Bernholz (eds.), Economic Imperialism: The
Economic Approach Applied Outside the Field of Economics, 133-155. New York: Paragon
House.
References - 503
1988 'Social Capital in the Creation of Human Capital.' American Journal of Sociology 94 (special
supplement): 95-120.
1990 Fouruiations of Social Theory. Cambridge: The Belknap Press of Harvard University Press.
1994 'A Rational Choice Perspective on Economic Sociology.' In N. Smelser and R. Swedberg (eds.).
The Handbook of Economic Sociology, 166-180. Princeton: Princeton University Press.
Coleman, J. S., E. Katz, and H. Menzel
1966 Medical innovation. New York: Bobbs-Merrill.
Collln,A.
1979 'Notes on Some Typologies of Managerial Development and the Role of Mentor in the Process of
Adaptation of the Individual to the Organization.' Personnel Review 8: 10-14.
Contractor, F. J., and P. Lorange (eds.)
1988 Cooperative Strategies in Internationol Business: Joint Ventures and Technology Partnerships
Between Firms. Lexington: Lexington Books.
Cook, K. S.
1977 'Exchange and Power in Networks of Interorganizational Relations.' Sociological Quarterly 43 :
721-739.
Cook, K. S., and R. M . Emerson
1978 'Power, Equity and Commitment in Exchange Networks.' American Sociological Review 43: 721-
739.
Cook, K . S., R. M. Emerson, M. GUlmore, and T. Yamaglshl
1983 'The Distribution of Power in Exchange Networks: Theory and Experimental Results.' American
Journal of Sociology 89: 275-305.
Corfman, K. P., and D. R. Lehmann
1987 'Models of Cooperative Group Decision-Making and Relative Influence: An Experimental
Investigation of Family Purchase Decisions.' Journal of Consumer Research 14: 1-13.
Cox,A.
1997 Business Success. UK: Earlsgate Press.
Cronin, J. J., and S. A. Taylor
1992 'Measuring Service Quality: A Reexamination and Extension.' Journal of Marketing 56: 55-68.
Crosby, L. A., and N. Stephens
1987 'Effects of Relationship Marketing on Satisfaction. Retention, and Prices in the Life Insurance
Industry.' Journal of Marketing Research 24: 404-411.
Crosby, L. A., K. R. Evans, and D. Cowles
1990 'Relationship Quality in Services Selling: An Interpersonal Influence Perspective.' Journal of
Marketing 54: 68-81.
Crozier,M.
1960 The Bureaucratic Phenomena. Glencoe: Free Press.
Crozier, M., and E. Friedberg
1977 L'Acteur et Ie Systeme. Paris: Seuil.
Cyert, R. M., and J. G. March
1963 A Behavioral Theory of the Finn. Englewood Cliffs: Prentice-Hall.
Daft, R. L.
1995 Organization Theory and Design . (Fifth Edition) St. Paul: West.
Daloz,L.A.
1986 Effective Teaching and Mentoring. San Francisco: Jossey-Bass.
Dasgupta, P., and P. David
1987 'lnfonnation Disclosure and the Economics of Science and Technology.' In G. R. Feiwel (ed.),
Arrow and the Ascent of Modem Economic Theory, 519-542. New York: New York University
Press.
1994 'Toward a New Economics of Science.' Research Policy 23: 487-521.
D'Avenl, R. A.
1990 'Top Managerial Prestige and Organizational Bankruptcy.' Organization Science, I: 121-142.
Davidow, W. H., and M. S. Malone
1992 The Virtual Cooperation: Structuring and Revitalizing the Cooperation for the 21st Century. New
York: HarperBusiness.
Davis,G. F.
1991 'Agents Without Principles? The Spread of the Poison Pill through the Intercorporate Network.'
Davis, G. F., K. A. Diekmann, and C. H. Tinsley

1994 The Decline and Fall of the Conglomerate Finn in the 1980s: The De-Institutionalization of an
Organizational Fonn.' American Sociological Review 59: 547-570.
Davis, H. L., S. J. Hoch, and E. K. Easton Ragsdale
1986 'An Anchoring and Adjustment Model of Spousal Predictions.' Journal of Consumer Research 13:
25-37.
Davis, G. F., and M. Mizruchi
1997 'The Money Center Cannot Hold: Conunercial Banks in the US System of Governance.' Working
paper.
Davis, G. F., C. H. Tinsley, and K. A. Diekmann
1997 'Acquisitions by Large Industrials in the late 1980s: The Twilight of Unrelated Diversification?'
Working paper, Columbia University.
Davis,J.A.
1966 'Structural Balance, Mechanical Solidarity, and Interpersonal Relations.' In J. Berger, M. Zelditch,
and B. Anderson (eds.), Sociological Theories in Progress I : 74- \OJ. Boston: Houghton Mifflin.
Davis, K., and J. W. Newstrom
1985 Human Behavior at Work: OrganiZlltional Behavior (7th edition). New York: McGraw-Hill.
Dawes,RM.
1980 'Social Dilemmas.' Annual Review of Psychology 31: 169-193.
De Graaf, N. D., and IL D. Flap
1988 With a little Help from My Friends: Social Resources as an Explanation of Occupational Status and
Income in West Gennany, The Netherlands, and The United States.' Social Forces 67: 452-472.
Dearborn, D. C., and H. A. Simon
1958 'Selective Perception: A Note on the Departmental Identifications of Executives.' Sociometry 21:
140-144.
Dess, G. G., and R. L. Priem
1995 'Consensue-Perfonnance Relationship Research: Theoretical and Empirical Extensions.' Journal of
Management Studies 32.
Deutsch,M.
1973 The Resolution of Conflict: Constructive and Destructive Processes. New Haven: Yale University
Press.
Devine, T. J., and N. M. Kiefer
1991 Empirical Labour Economics: The Search Approach. New Yorlc: Oxford University Press.
Diamond, D. W.
1984 'Financiallntennediation and Delegated Monitoring.' Review of Economic Studies 51: 393-414.
Dickson, P. H. And M. K. Weaver
1997 'Environmental Determinants and Individual-Level Moderators of Alliance Use.' Academy of
Diener, E., And R A. Emmons
1985 'The Independence of Positive and Negative Affect.' Journal of Personality and Social Psychology
47: 1758-1768.
Dimaggio, P. J .
1994 'Culture and Economy.' In N. Smelser and R. Swedberg (eds.), The Handbook of Economic
Sociology, 27-57. Princeton: Princeton University Press.
Dimaggio, P. J., and W. W. Powell
1983 'The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational
Fields.' American Sociological Review 48: 147-160.
DiPrete, T. A.
1987 'Horizontal and Vertical Mobility in Organizations.' Administrative Science Quarterly 32: 422-
444.
1989 The Bureaucratic Labor Market: The Case of the Federal Civil Service. New Yorlc: Plenum Press.
Ditcher, S. F.
1991 'The Organization of the '90s.' The McKinsey Quarterly I: 145-155.
Dobbin, F., L. Edelman, J. W. Meyer, W. R Scott, and A. Swidler
1988 'The Expansion of Due Process in Organizations.' In L. Zucker (ed.), Institutional Patterns and
OrganiZlltions: Culture and Environment, 71-98. Cambridge: Ballinger.
Dodgson,M.
1993 Technological Collaboration in Industry. London: Routledge.
References - 505
Doherty,J.
1997 'The New Drexel.' Barron's (September 8): 17-21.
Domhoff, G. W.
1975 'Social Clubs, Policy-Planning Groups, and Corporations: A Network Study of Ruling-Class
Cohesiveness.' Insurgent Sociologist 5: 173-84.
Dore,R.
1983 'Goodwill and the Spirit of Market Capitalism.' British Journal of Sociology 34: 459-482.
Doreian,P.
1985 'A Measure of Standing of Journals in in Stratified Networks.' Scientometrics 8: 341-363.
1987 'A Revised Measure of Standing of Journals in Stratified Networks.' Scientometrics II: 71-80.
Doreian, P., and K. L. Woodard
1992 'Fixed List versus Snowball Selection of Social Networks.' Social Science Research 21: 216-233.
Dreher, G. F., and R. A. Ash
1990 'A Comparative Study of Mentoring Among Men and Women in Managerial, Professional, and
Technological Positions.' Journal of Applied Psychology 75: 539-546.
Dubois,A.
1994 Organising Industrial Activities: An Analytical Framework. Dept of Ind. Marketing, Chalmers
University of Technology, Sweden.
Dunleavy, P., and C. Hood
1994 'From Old Public Administration to the New Public Management.' Public Money and Management
July/Sept: 9-16.
Durkheim, E.
1893 De la Division du Travail Social. Paris: Presses Universitaires de France.
189711951 Suicide: A Study in Sociology. Translated by J. A. Spaulding and O. Simpson. New York: Free
Press.
Dyer,J. H.
1996a 'Specialized Supplier Networks as a Source of Competitive Advantage: Evidence from the Auto
Industry.' Strategic Management Journal 17: 271-291.
1996b 'How Chrysler Created an American Keiretsu.' Harvard Business Review 74: 42-56.
1997 'Improving Performance by Transforming Arm's-Length Relationships to Supplier Partnerships:
the Chrysler Case.' OrganilJltion Science (forthcoming).
Easton, G., and L. Araujo
1993 'Language, Metaphors and Networks.' In D. Sharma (ed.), Advances in International Marketing:
Industrial Networks 5: 67-85. Greenwich: JAI Press.
1996 'Characterising Resources as an Organisational Competence: an Industrial Networks Approach.' In
R. Sanchez. A. Heene, and H. Thomas (eds.), Dynamics of Competence Based Competition:
Theory and Practice in the New Strategic Management, 183-207. New York: Pergamon Press.
Easton, G., and M. Quayle
1990 'Single and Multiple Network Sourcing-Network Implications.' Proc. of 6th IMP Conference,
Milan.
Ebeling, H., and L. Doorley
1983 'A Strategic Approach to Acquisitions.' Journal of Business Strategy 3: 44-55.
Eccles, R. G.
1981 'Bureaucratic versus Craft Administration: The Relationship of Market Structure to the
Construction Firm.' Administrative Science Quarterly 26: 449-469.
Eccles, R. G., and D. B. Crane
1988 Doing Deals: Investment Banks at Work. Boston: Harvard Business School Press.
Eccles, R. G., and H. C. White
1988 'Price and Authority in Inter-Profit Center Transactions.' American Journal of Sociology 94
(special supplement): 17-51.
Economist, The
1993 'Within the Wirldwind. A Survey of the Computer Industry.' The Economist February 27.
Edstrom, A., and J. Galbraith
1977 'Transfer of Managers as a Coordination and Control Strategy in Multinational Organizations.'
Edwards, R.
1979 Contested Terrain: The Transjonnation of the Workplace in the Twentieth Century. New York:
Basic Books.
Eisenberg, E. M., P. R. Monge, and K. I. Miller

1984 'Involvement in Communication Networks as a Predictor of Organizational Commitment.' Human
Communication Research 10: 179-201.
Eisenhardt, K. M., and C. B. Schoonhoven
1996 'Resource-Based View of Strategic Alliance Fonnation: Strategic and Social Effects in
Entrepreneurial Firms.' Organiwtion Science 7: 136--150.
Ekeh,P.P.
1974 Social Exchange Theory: The Two Traditions . Cambridge: Harvard University Press.
EI-Ansary, A. I., and L. W. Stem
1972 'Power Management in the Distribution Channel.' Journal of Marketing Research 9: 47-52.
ElIram,L.
1991 'Supply Chain Management: The Industrial Organisation Perspective.' International Journal of
Physical Distribution and Logistics Management 21 : 13-22.
Emanuel, P. M.
1995 'Using Engineering Expertise to Develop a Successful World Wide Patent Portfolio.' In M. A.
Lechter et aI. (eds.). Successful Patenting for Scientists and Engineers. New York: IEEE Press.
Emerson, R. M.
1962 'Power-Dependence relations.' American Sociological Review 27: 31-40.
Emirbayer, M., and J. Goodwin
1994 'Network Analysis. Culture. and the Problem of Agency.' American Journal of Sociology 99:
1411-1454.
Erens, F., R. Stoifelen, F. van de Ven, and L. Wildeman
1996 'Alliance and Networks: The Next Generation.' Proceedings of the 6th International Forum on
Technalogy Management 77- 91.
Erickson, B. H.
1988 The Relational Basis of Attitudes.' In B. Wellman and S. D. Berkowitz (eds.). Social Structures: A
Network Approach. 99-121 . Cambridge: Cambridge University Press.
Etgar,M.
1976 'Channel Domination and Countervailing Power in Distributive Channels.' Journal of Marketing
Research 13: 254-262.
1979 'Sources and Types of Intrachannel Conflict.' Journal of Retailing 55: 61-78.
Etzkowitz, H. and A. Webster
1995 'Science as Intellectual Property.' In S. Jasanoff. O. Mankle. J. Peterson. and T. Pinch (eds.).
Handbook of Science and Technalogy Studies. Thousand Oaks: Sage.
Evans,P.
1996 'Govemment Action. Social Capital and Development-Reviewing the Evidence on Synergy.'
World Development 24: 1119-1132.
Evron, Y.
1992 The Military Industry in Israel. Tel Aviv, Ministry of Defense (in Hebrew).
Faase,L.
1980 Tussenschakels op de Arbeidsmarkt.' [Links on the Labor Market, in Dutch]. Econamisch Statistische
Berichten 65: 1047-1049.
Fagenson, E. A.
1988 The Power of a Mentor: Proteg~ and Nonproteg~' Perceptions of Their Own Power in
Organizations.' Group and Organiwtion Studies 13: 182-192.
1989 The Mentor Advantage: Perceived Career/Job Experiences of Proteg~ Versus Non-Proteg~.'
Journal ofOrganiwtional Behavior 10: 309-320.
1994 'Perceptions of Proteges' vs. Nonproteges' Relationships with Their Peers. Superiors. and
Departments.' Journal of Vocational Behavior 45: 55-78.
Farmer, D. H., and R. Ploos von Amstel
1991 Effective Pipeline Management. UK: Gower.
Farr, C. M., and W. A. Fischer
1992 'Managing International High Technology Cooperative Projects.' R&D Management 22: 55-67.
Feld,S.L.
1981 'The Focused Organization of Social Ties.' American Journal of Sociology 86: 10 15-1 035.
Fenton, E.
1996 The Internal Network Organisation: A Review. University of Warwick: Centre for Corporate
Strategy and Change.
References - 507
Ferlie, E. B., and A. M. Pettigrew

1996 'Managing Through Networks-Some Issues and Implications for the NHS.' British Journal of
Management 7: 81-99.
Ferlie, E. B., L. Ashbumer, L. Fitzgerald, and A. M. Pettigrew
1996 The New Public Management in Action. Oxford: Oxford University Press.
Fernandez, R. M., and N. Weinberg
1997 'Sifting and Sorting: Personal Contacts and Hiring in a Retail Bank.' American Sociological
Review 62: 883-902.
Ferris, G. R., and T. A. Judge
1991 'PersonnellHuman Resources Management: A Political Influence Perspective.' Journal of
Festinger, L.
1954 'A Theory of Social Comparison Processes.' Human Relotions 7: 117-140.
Festinger, L., S. Schachter, and K. Back
1950 Social Pressures in Informal Groups: A Study of Human Factors in Housing. Palo Alto: Stanford
University Press.
Fevre,R.
1989 'Informal Practices, Aexible Firms, and Private Labour Maricets.' Sociology 23: 91-109.
Fie!, J.
1991 'Network Reliance by Venture Capital Firms and Business Angels: An Empirical and Theoretical
Test.' In Frontiers of Entrepreneurial Research, 445-455. Wellesley: Babson College.
Finkelstein, S., and D. C. Hambrick
1990 Top Management Team Tenure and Organizational Outcomes: The Moderating Role of
Managerial Discretion.' Administrative Science Quarterly 35: 484-503.
Flscher,C.
1982 To Dwell Among Friends. Personal Networks in Town and City. Chicago: University of Chicago
Press.
Flannery, T. P., D. A. Hofrichter, and P. E. Platten
1996 People, Performance, and Pay: Dynamic Compensation for Changing Organiwtions. New York:
Free Press.
Flap,H.D.
1988 Conflict, Loyalty and Violence. The Effects of Social Networks on Behaviour. Bern: Peter Lang
Verlag.
1990 'Patronage: An Institution in its Own Right.' In M. Hechter, K. D. Opp and R. Wippler (cds.),
Social Institutions: Their Emergence, Maintenance and Effects. New York: A1dine de Gruyter.
1991 'Social Capital in the Reproduction of inequality: A Review.' Comparative Sociology of Family,
Health and Education 20: 6179-6202.
Flap, H. D., and E. Boxman
1996 'Getting Started. The Influence of Social Capital on the Start of the Occupational Career.'
Communication ala Journie-dibat avec Harrison White, 3 Mai 1996, CNRS, Paris.
Flap, H. D., B. Bulder, and B. VOlker
1998 'Intra-Organizational Networks and Perfonnance. A Review.' Computational and Mathemotical
Organiwtion Theory 4: 109-147.
Flap, H. D., T. Snijders, and F. Van Wlnden
1996 'Creation and Returns of Social Capital. Social Networks in Education and Labour Maricets.'
Application for a research grant from the Dutch Science Foundation.
Flap, H. D., and N. D. De Graat
1986 'Social Capital and Attained Occupational Status.' Nether/ands Journal of Sociology 22: 145-161.
Fligstein. N.
1985 The Spread of the Multidivisional Form Among Large Firms, 1919-1979.' American Sociological
Review 50: 377-391.
1990 The Transformation of Corporate Control. Cambridge: Harvard University Press.
1991 The Structural Transformation of American Industry: An Institutional Account of the Causes of
Diversification in the Largest Firms, 1919-1979.' In W. W. Powell and P. J. DiMaggio (eds.), The
New Institutionalism in Organiwtional Analysis, 311-336. Chicago: University of Chicago Press.
1996 'Maricets As Politics: A Political-Cultural Approach to Maricet Institutions.' American Sociological
Review 61 : 656-673.
Florida, R., and M. Kenney

1988 'Venture Capital and High Technology Entrepreneurship.' Journal of Business Venturing 3: 301-
319.
Flynn, R., G. Williams, and S. Pickard
1996 Markets and Networks: Contracting in Community Health Services. Buckingham: Open
University Press.
Foley, M. W., and B. Edwards
1997 'Escape from Politics? Social Theory and the Social Capital Debate.' American Behavioral
Scientist 4(J: 550--561.
Fombrun, C. J.
1996 Reputation: Realizing Valuefrom the Corporate Image . Boston: Harvard Business School Press.
Ford, I. D.
1978 'Stability Factors in Industrial Marketing Channels.' Industrial Marketing Management 7: 410--
422.
Forrest, J. E., and M. J. C. Martin
1992 'Strategic Alliances Between Large and Small Research Intensive Organizations: Experiences in
the Biotechnology Industry.' R&D Management 22: 41-53.
Forrester, J. W.
1961 Industrial Dynamics. MIT Press.
Forse,M.
1997 'Capital Social et Emploi.' L'Annee sociologique 47: 143- 181.
Foss, N.J.
1993 'Theories of the Fmn: Contractual and Competence Perspectives.' Journal of Evolutionary &onomics
3: 12744.
Foss, N. J., and C. Knudsen
1996 Towards a Competence Theory of the Firm. London: Routledge.
Frank,O.
1991 'Statistical Analysis of Change in Networks.' Statistica Neerlandica 3: 283-293.
Frank, 0 ., and D. Strauss
1986 'Markov Graphs.' Journal of the American Statistical Association 81 : 832-842.
Frazier, G. L., R. E. Spekman, and C. R. O'Neal
1988 'lust-in-Time Exchange Relationships in Industrial Markets.' Journal of Marketing 52: 52-67.
Freeman, J. H.
1986 'Entrepreneurs as Organizational Products: Semiconductor Fmns and Venture Capital Fmns.' In G.
Libecap (ed.), Advances in the Study of Entrepreneurship, Innovation and &onomic Growth.
Greenwich: lAI Press.
Freeman, J. H., G. R. Carrou, and M. T. Hannan
1983 'The Liability of Newness: Age Dependence in Organizational Death Rates.' American Sociological
Review 48: 692-710.
Freeman, L. C.
1977 'A Set of Measures of Centrality Based on Betweenness.' Sociometry 4(J: 35-41 .
1979 'Centrality in Social Networks I. Conceptual Clarification.' Social Networks I : 215-239.
Freeman, R. B., and J . L. Medoff
1984 What Do Unions Do? New York: Basic Books.
French, J. R. P.,Jr., and B. Raven
1968 'The Social Bases of Power.' In D. Cartwright and A. Zander (eds.), Group Dynamics (3rd ed.),
259-269. New York: Harper and Row.
Frenzen, J. K., and K. Nakamoto
1993 'Structure, Cooperation, and the Flow of Market Information.' Journal of Consumer Research 20:
360-375.
Fried, V. H., and R. D. Hisrich
1995 'The Venture Capitalist: A Relationship Investor.' California Management Review 37: 101-113.
Friedberg, E.
1993 Le Pouvoir et La Regie. Paris: Seuil.
Friedkin, N. E.
1998 A Structural Theory of Social Influence . New York: Cambridge University Press.
References - 509
Friedman, R. A., and J . Podolny

1992 'Differentiation of Boundary Spanning Roles: Labor Negotiations and hnplications for Role
Conflict.' Administrative Science Quarterly 37: 28-47.
Fruin,M.
1992 The Japanese Enterprise System. Oxford: Clarendon Press.
Fukuyama, F.
1995a 'Social Capital and the Global Economy.' Foreign Affairs 5: 89-103.
1995b Trust: The Social Virtues and the Creation of Prosperity. New York: The Free Press.
Gabbay, S. M.
1997 Social Capital in the Creation of Financial Capital. Ph.D. dissertation, Columbia University ..
1997 Social Capital in the Creation of Financial Capital: The Case of Network Marketing. Champaign:
Stipes Publishing.
Gabbay, S. M, and R. Th. A. J. Leenders
1999'Corporate Social Capital (CSC) and Global Market Competition.' Paper presented at the
International Sociological Association Meetings, Tel Aviv, Israel.
Gabbay, S. M., M. Semyonov, and S. Mishal
1999 'Entrepreneurial Networks in an Era of Peace.' The Chaim Center for Middle East Studies and
Diplomacy. Ben Gurion University.
Gabbay, S. M, and Y. Sato
1996 'Revisiting Social Capital through hnages of Social Structure.' Japanese Mathematical Sociology
19: 8-21.
Gabbay, S. M, and A. J. Stein
1999 'Embedding Social Structure in Technological Infrastructure: Constructing Regional Social Capital
for a Sustainable Peace in the Middle East.' In J. Wright (ed.), The Political Economy of Middle
East Peace. Boston: Routledge.
Gabbay, S. M., and E. Zuckerman
1998 'Social Capital and Opportunity in Corporate R&D: The Contingent Effect of Contact Density on
Mobility Expectations.' Social Science Research 27: 189-217.
Galanter, M., and T. Palay
1991 Tournament of Lawyers: The Transformation of the Big Law Firm. Chicago: The University of
Chicago Press.
Galasldewicz, J.
1979 Exchange networks and community politics. Beverly Hills: Sage.
1985 Social OrganiZlltion of an Urban Grants Economy. Orlando: Academic Press.
Gallini, N. T.
1992 'Patent Policy and Costly Imitation.' RAND Journal of Economics 23: 52-63.
Gambardella, A.
1992 'Competitive Advantages from In-House Scientific Research. The US Pharmaceutical Industry in
the 1980s.' Research Policy 21 : 391-407.
Gambetta, D.
1988a 'Can We Trust Trust?' In D. Gambetta (ed.), Trust: Making and Breaking Cooperative Relations,
213-238. Oxford: Blackwell.
1988b 'Mafia: The Price of Distrust.' In D. Gambetta (ed.), Trust: Making and Breaking Cooperative
Relations, 158-175. Oxford: Blackwell.
Gambetta, D. (ed.)
1988c Trust: Making and Breaking Cooperative Relations. New York: Blackwell.
Gargiulo, M.
1993 Two-Step Leverage: Managing Constraint in Organizational Politics.' Administrative Science
Quarterly 39: 1-19.
Gartner, W. B.
1985 'A Conceptual Framework for Describing the Phenomenon of New Venture Creation.' Academy of
Management Review 10: 696-706.
1989 "Who Is an Entrepreneur?' Is the Wrong Question.' Entrepreneurship Theory and Practice 13: 47-
68.
1990 'What Are We Talking About When We Talk About Entrepreneurship?' Journal of Business
Venturing 5: 15-28.
Gasse, Y.
1982 'Elaborations on the Psychology of the Entrepreneur.' In C. A. Kent, D. L. Sexton, and K. H.
Vesper (eds.), Encyclopedia of Entrepreneurship, 57-71. Englewood Cliffs: Prentice-Hall.
Gattoms, J.
1978 'Channel of Distribution Conceptualisations: a State of the Art Review.' European Journal of
Gelauff, G. M. M., and C. den Broeder
1996 'Governance of Stakeholder Relationships, The Gennan and Dutch Experience.' Research
Memorandum 127. The Hague: Central Planning Bureau.
Gerlach, M. L.
1987 'Business Alliances and the Strategy of the Japanese Fmn.' In G. Carroll and C. Vogel (eds.),
OrganilJJtional Approaches to Strategy. Berkeley: University of Califiornia Press.
1992a Alliance Capitalism: The Social OrganilJJtion of Japanese Business. Berkeley: University of
California Press.
1992b 'Twillight of the Keiretzu: A Critical Assessment'. Journal of Japanese Studies 18: 79-118.
Ghoshal S., and P. Moran
1996 'Bad for Practice: A Critique of Transaction Cost Theory.' Academy of Management Review 21:
13-47.
Ghoshal, S., and N, Nohria
1989 'Internal Differentiation within Multinational Corporations.' Strategic Management Journal 10:
323-337.
Ghoshal, S., H. Korine, and G. Szulanski
1994 'Interunit Conununication in Multinational Corporations.' Management Science 40: 96-110.
Giddens, A.
1976 New Rules of Sociological Method. London: Hutchinson.
1979 Central Problems in Social Theory. London: Macmillan.
Gilad, N., S. Kalsh, and J. Ronen
1989 'Infonnation, Search, and Entrepreneurship: A Pilot Study.' The Journal of Behavioral Economics
18: 217-231.
Gilroy, B. M,
1993 Networking in Multinational Enterprises: The Importance of Strategic Alliances. Columbia:
University of South Carolina Press.
Gilson, R. J" and R. H. Mnookin
1985 'Sharing Among Human Capitalists: An Economic Inquiry into the Corporate Law Fmn and How
Partners Split Profits.' Stanford lAw Review 37: 313-392.
Glmeno, J., and C, Woo
1996 'Economic Multiplexity: The Structural Embeddedness of Cooperation in Multiple Relations of
Interdependence.' In J. A. C. Baum and J. E. Dutton (eds.), Advances in Strategic Management:
The Embeddedness of Strategy, 323-361. Greenwich: JAI Press.
Gilman, L. J.
1979 Principles of Managerial Finance. New York: Harper and Row.
Glain,S,
1997 'How Beijing Officials Outnegotiated AT&T on Marine Cable Plans.' Wall Street Journal (July
16). New York: Dow Jones & Company, Inc.
Go~asseres, B,
1996 The Alliance Revolution: The New Shape of Business Rivalry. Cambridge: Harvard University
Press.
Goodman L. A.
1984 The Analysis of Cross-Classified Data Having Ordered Categories. Cambridge: Harvard
University Press.
1985 'The Analysis of Cross-Classified Data Having Ordered andlor Unordered Categories: Association
Models, Correlation Models, and Asymmetry Models for Contingency Tables with or without
Missing Entries.' Annals of Statistics 13: 10-69.
Gouldner, A, W.
1960 'The Nonn of Reciprocity.' American Sociological Review 25: 161-178.
References - 511
Grabber,G.
1993 'The Weakness of Strong Ties: the Lock-in of Regional Development in the Rube Area.' In G.
Grabher (ed.), The Embedded Firm: On the Socioeconomics of Industrial Networks, 255-277.
New York: Routledge.
Granovetter, M. S.
1973 'The Strength of Weak Ties.' American Journal of Sociology 78: 13®-1380.
1982 The Strength of Weak Ties: A Network Theory Revisited.' In P. V. Marsden and N. Lin (eds.),
Social Structure and Network Analysis, 105-130. Beverly Hills: Sage.
1985 'Economic Action and Social Structure: The Problem of Embeddedness.' American Journal of
Sociology 91: 481-510.
1988 'The Sociological and Economic Approaches to Labour Market Analysis, a Social Structural View.' In
G. Farkas and P. Fngland (eds.), Industries, Firms and Jobs: Sociological and Economic Approaches,
187-217. New York: Plenum Press.
1993 The Nature of Economic Relationships.' In R. Swedberg (ed.), Explorations in Economic
Sociology, ~1. New York: Russell Sage.
1994 'Business Groups.' In N. Smelser and R. Swedberg (eds.). The Handbook of Economic Sociology.
453-475. Princeton: Princeton University Press.
1995 Getting A Job: A Study of Contacts and Careers. Chicago: University of Chicago Press (first
edition published in 1974).
Greeley,A.
1997 'Coleman Revisited: Religious Structures as a Source of Social Capital.' American Behavioral
Scientist 40: 587-594.
Gregory, D. A., C. W. Saber, and J. D. Grossman
1994 Introduction to Intellectual Property Law. Washington: The Bureau of National Affairs.
Greif,A.
1994 'Cultural Beliefs and the Organization of Society-A Historical and Theoretical Reflection on
Collectivist and Individualist Societies.' Journal of Political Economy 102: 912-950.
Grieco,M.
1987 Keeping it in the Family. Oxford: Oxford University Press.
Grillcbes, Z.
1990 'Patent Statistics as Economic Indicators: A Survey.' Journal of Economic literature 28: 1661-
1707.
Grimm, C. M., and K. G. Smltb
1991 'Management and Organizational Change: A Note on the Railroad Industry.' Strategic
AlanagementJournaI12:557-562.
Grindley, P. C., and D. J. Teece
1997 'Managing Intellectual Capital: Licensing and Cross-Licensing in Semiconductors and
Electronics.' California Alanagement Review 39: 8-4 I.
Groenewegen, J.
1997 'Institutions of Capitalism: American. European and Japanese Systems Compared.' Journal of
Economic Issues 31 : 333-347.
Gronroos, C.
1990 Service Alanagement and Alarlceting. Lexington: Lexington Books.
Guillen, M. F.
1997 'Business Groups in Economic Development.' Best papers Proceedings. 170-174. Academy of
Management. Annual Meeting. Boston.
Gulatl,R.
1995a 'Social Structure and Alliance Formation Pattern: A Longitudinal Analysis.' Administrative
Science Quanerly40: 619-652.
1995b 'Does Familiarity Breed Trust? The bnplications of Repeated Ties for Contractual Choices in
Alliances.' Academy of Alanagement Journal 38: 85-112.
Gummesson, E.
1987 'Lip Service: A Neglected Area in Services Marketing.' Journal of Services Alarlceting 1: 19-23.
Guo,G.
1993 'Event-History Analysis for Left-Truncated Data.' Sociological Alethodology 23: 217-243.
Gupta, A. K., and V. Govlndrajan
1991 'Knowledge Flows and the Structure of Control within Multinational Corporations.' Academy of
Alanagement Review 16: 768-792.
Hackman, J. R., and R. Wageman

1995 'Total Quality Management: Empirical. Conceptual, and Practical Issues.' Administrative Science
Quanerly 40: 309-342.
Hagedoorn, J.
1993a 'Strategic Technology Alliances and Modes of Cooperation in High-Technology Industries.' In G.
Grabher (ed.), The Embedded Firm: On the Socioeconomics of Industrial Networks, 116-137.
New York: Routledge.
1993b 'Understanding the Rational of Strategic Technology Partnering: International Modes of
Cooperation and Sectoral Differences.' Strategic Management 10umal14: 371-85.
Hagedoorn, J., and J. Schakenraad
1994 'The Effect of Strategic Technology Alliances on Company Performance.' Strategic Management
loumal 15: 291-309.
Hakansson, H. (ed.)
1982 International Marketing and Purchasing of Industrial Goods: an Interaction Approach.
Chichester: Wiley.
1987 Industrial Technological Development: A Network Approach. London: Croom Helm
Hakansson, H., and J. Johanson
1993 The Network as a Governance Structure: Interfirm Cooperation Beyond Markets and Hierarchies.'
In F. Grabber (ed.), The Embedded Firm. London: Routledge.
Hakansson, H., and I. Snehota (eds.)
1995 Developing Relationships in Business Networks . London: Routledge.
Ham,C.
1981 Policy Making in the National Health Service. London: Macmillan.
Hambrick, D. C.
1981 'Environment, Strategy, and Power Within Top Management Teams.' Administrative Science
Quanerly 26: 253-275.
Hambrick, D. C., and G. D. Fukutomi
1991 'The Seasons of a CEO's Tenure.' Academy o/Management Review 16: 719-742.
Hambrick, D. C., and P. A. Mason
1984 'Upper Echelons: The Organizations as a Reflection of its Top Managers.' Academy 0/
Hamel, G., Y. L. Doz, and C. K. Prahalad
1989 'Collaborate with Your Competitors-And Win.' Harvard Business Review 67: 133--139.
Hampson, K., and T. Kwok
1996 'Analysing Strategic Alliances in Building Construction.' In M. Sobolewski and M. Fox (eds.),
Advances in Concurrent Engineering, 172-180. Toronto: Technomic Publishing.
Han,S.-K.
1993 'Churning Firms in Stable Markets.' Social Science Research 21: 406-418.
1994 'Mimetic Isomorphism and its Effect on the Audit Services Market.' Social Forces 73: 637-663.
1996 'Structuring Relations in On-the-Job Networks.' Social Networks 18: 47...fJ7.
Han, S.-K., and R. L. Breiger
1996 'Switching and Stacking: Network Multiplexity in the Supreme Court.' Paper presented at the 16th
Sunbelt Intemational Social Network Conference, Charleston, SC.
Hannan, M. T., and G. R. Carron
1992 Dynamics of OrganiZlltional Populations: Density, Competition and Legitimation. New Yorle Oxford
University Press.
HalllUlIl, M. T., and J. Freeman
1984 'Structural Inertia and Organizational Change.' American Sociological Review 49: 149...fJ4.
1989 OrganiZlltional Ecology. Cambridge: Harvard University Press.
Hannerz, U.
1969 Soulside: Inquiries into Ghetto Culture and Community. New York: Columbia University Press.
Hannifan, L,J.
1920 The Community Center. Boston: Silver Burdette.
Hansen, M. N.
1996 'Earnings in Elite Groups: the Impact of Social Class Origin.' Acta Sociologica 39: 385-408.
Hansen, M. T.
1997 'Integrating Knowledge through the Intra-Company Network: Implications for Product
Development Time.' Working Paper, Harvard Business School.
References - 513
Hardin,R.
1993 'The Street-Level Epistemology of Trust.' Politics and Society 21: 505-529.
1996 'Trustworthiness.' Ethics 107: 26-42.
Harland, C. M.
1995 'The Dynamics of Customer Dissatisfaction in Supply Chains.' Production Planning and Control.
Supply Chain Management 6 (special issue): 209-217.
1996a 'Supply Chain Management: Relationships, Chains and Networks.' British Journal of Management
7: 63-80.
1996b 'Supply Network Strategies: the Case of Health Supplies.' European Journal of Purchasing and
Supply Management 2: 183-192.
1997a 'Empirical Research in Characterising Supply Networks.' Proc. of lst International Conference on
Responsive ManUfacturing, 27-37. September, Univ. of Nottingham.
1997b 'Supply Chain Operational Performance Roles.' Integrated Manufacturing Systems 8: 6-14.
Harland, C. M., and J. R. Wensley
1996 'Supply Chain Management as a Response to the Strategy Leaf Kickers.' Paper Presented at BAM
Conference, September, Aston University.
Harrigan, K. R.
1983 Strategies for Venical Integration. Lexington: Lexington Books.
Harris, R. C., R. C. Insinga, J. Morone, and M. J. Werle
1996 'The Virtual R&D Laboratory.' Research Technology Management 39: 32-36.
Haspeslagb, P. C.
1982 'Portfolio Planning: Uses and Limits.' Harvard Business Review 60: 58-73.
Haspeslagh, P. C., and D. B. Jemison
1991 Managing Acquisitions: Creating Value through Corporate Renewal. New York: Free Press.
Haunsehild, P. R.
1993 'lnterorganizational hnitation: The Impact of Interlocks on Corporate Acquisition Activity.'
Administrative Science Quanerly 38: 564-592.
1994 'How Much Is That Company Worth?: Interorganizational Relationships, Uncertainty, and
Acquisition Premiums.' Administrative Science Quarterly 39: 391-411.
Hauser, R. M.
1980 'On 'Stratification in a Dual Economy' .' American Sociological Review 45: 702-12.
Hayek, F. A.
1978 New Studies in Philosophy, Politics, Economics and the History of Ideas. Chicago: The University
of Chicago Press.
Hayes, R. H., and W. J. Abernathy
1980 'Managing Our Way to Economic Decline.' Harvard Business Review 58: 67-77.
Hayes, R. H., and S. C. Wheelwrigbt
1984 Restoring our Competitive Edge: Competing Through Manufacturing . New York: Wiley.
Hayes, S. L., m
1971 'Investment Banking Power Structure in Flux.' Harvard Business Review 49: 136-152.
1979 'The Transformation of Investment Banking.' Harvard Business Review 57: 153-170.
Haywood-Farmer, J., and J. Nollet
1991 Services Plus: Effective Service Management. Boucherville, Quebec: Morin.
Hecbter,M.
1987 Principles of Group Solidarity. Berkeley: University of California Press.
Heinz, J. P., E. O. Laumann, C. L. Cappell, T. C. Halliday, and M. Scbaalman
1982 Chicago Lawyers: The Social Structure of the Bar. New York: Russell Sage Foundation and
American Bar Foundation.
Heinz, J. P., E. O. Laumann, R. L. Nelson and P. S. Schnorr
1997 'The Constituencies of Elite Urban Lawyers.' Law and Society Review 31: 441-72.
Helper,S.
1990 'Comparative Supplier Relations in the US and Japanese Auto Industries: An Exit-Voice
Approach.' Business Economic History 19: 153-162.
1991 'Strategy and Irreversibility In Supplier Relations--The Case of the United-States Automobile-
Industry.' Business History Review 65: 781-824.
Herrigel, G.
1996 Industrial Constructions: The Sources of German Industrial Power. New York: Cambridge
University Press.
Heskett, J. L., T . O. Jones, G. W. Loveman, W. Earl Sasser, Jr., and L. A. Scblesinger

1994 'Putting the Service-Profit Chain to Work.' Harvard Business Review 72: 164-174.
Heydebrand, W. V.
1989 'New Organizational Fonns.' Work and Occupations 16: 323-57.
Hibbard, J ., and D. Iacobucci
1997 'Toward an Encompassing Theory of Relationships in Business and Consumer Marketing: A
Systematic Empirical Generalization Analysis and Future Directions.' Working Paper.
Hkks,J.
1942 The Social Framework: An Introduction to Economics. New York: Clarendon Press
Higgins, M. C.
1997a 'When is Helping Helpful? Problems of Individual Learning in Organizations.' Harvard Business
School Working Paper #97-044. Boston: Harvard Business School Division of Research.
1997b 'Early Career Change: Toward a Portfolio Theory of Relations.' Harvard Business School Working
Paper #98-016. Boston: Harvard Business School Division of Research.
Higgins, M. C., and D. A. Thomas
1997 'Mentoring. Mobility. and Organizational Attachment in the Careers of Lawyers: A Longitudinal
Study.' Harvard Business School Working Paper #97-080. Boston: Harvard Business School
Division of Research.
Hill, L., and N. Kamprath
1991 'Beyond the Myth of the Perfect Mentor: Building a Network of Developmental Relationships.'
Harvard Business School Teaching Note #9-491-096. Boston: Harvard Business School Division
of Research.
Hill, T.
1985 Manufacturing Strategy. London: Macmillan.
Hirsch, P. M.
1972 'Processing Fads and Fashions: An Organization-Set Analysis of Cultural Industry Systems.'
American Journal of Sociology 77: 63~59.
Hofstede, G. H.
1980 Culture's Consequences. Beverly Hills: Sage.
1994 Cultures and Organisations. London: Harper and Collins.
Homans, G. C.
1950 The Human Group. New York: Harcourt Brace.
1961 Social Behavior: Its Elementary Forms. New York: Harcourt, Brace. and World.
Hood,C.
1991 'A Public Management for All Seasons?' Public Administration 69: 3-19.
Hope,K.
1982 'Vertical and Non-Vertical Class Mobility in Three Countries.' American Sociological Review 47:
99-113.
Hornby,D.
1992 in 'Partnership Sourcing: Partnership Sourcing Ud. CBI
HoshI, T., A. Kashyap, and D. S. ScharCstein
1990 'Bank Monitoring and Investment: Evidence from the Changing Structure of Japanese Corporate
Banking Relationships .• In R. G. Hubbard (ed.). Asymmetric Information. Corporate Finance. and
Investment. 105-125. Chicago: University of Chicago Press.
HoshI, T., A. Kashyap, and D. S. ScbarCstein
1991 'Corporate Structure. Liquidity and Investment: Evidence from Japanese Industrial Groups.'
Quarterly Journal of Economics 106: 33-60.
Houlihan, J.
1984 'Supply Chain Management.' Proc. of 19th International Technical Conference, BPICS. 101-110.
Hout, M., O. D. Duncan, and M. E. Sobel
1987 'Association and Heterogeneity: Structural Models of Similarities and Differences.' In C. C. Clogg
(ed.), Sociological Methodology 1987. 145-184. Washington: American Sociological Association.
Hsiao,C.
1986 Analysis of Panel Data. New York: Cambridge University Press.
Hubbell, C. H.
1965 'An Input-Output Approach to Clique Detection.' Sociometry 28: 277-299.
Humphrey, W.
1987 Managingfor Innovation: Leading Technical People. New Jersey: Prentice Hall.
References - 515
Hutt, M. D., P. H. Reingen, and J. R. Ronchetto, Jr.

1988 'Tracing Emergent Processes in Marketing Strategy Formation.' Journal of Marketing 52: 4-19.
Hyatt,J.
1989 'The Last Shift: INC., February: 74-80.
Iacobucci, D.
1997 'A Game and Norm Theoretic Metaphor for the Customer-Service Provider Dyadic Interaction:
Conceptual Explanations for Poor Customer Service and Managerial Implications for Its
Improvement: In J. Sherry (ed.), Servicescapes: The Built Environment of the Sales Encounter.
Glenview: NTC Publishing Group (forthcoming).
1998 'Services Marketing and Management: What Do We Know and Where Shall We Go?' In T. A.
Swartz, D. E. Bowen, and S. W. Brown (eds.), Advances in Services Marketing and Management:
Research and Practice 7. Greenwich: JAI Press (forthcoming).
Iacobucci, D., and A. Ostrom
1996 'Commercial and Interpersonal Relationships: Using the Structure of Interpersonal Relationships to
Understand Individual-to-Individual, Individual-to-Firm, and Fmn-to-Fmn Relationships in
Commerce: International Journal of Research in Marketing 13: 53-72.
Iacobucci, D., and N. Hopkins
1992 'Modeling Dyadic Interactions and Networks in Marketing.' Journal of Marketing Research 29: 5-
17.
Iacobucci, D., and S. Wasserman
1987 'Dyadic Social Interactions.' Psychological Bulletin 102: 293-306.
1988 'A General Framework for the Statistical Analysis of Sequential Dyadic Interaction Data:
Psychological Bulletin 103: 379-390.
Ibarra, H.
1992 'Homophily and Differential Returns: Sex Differences in Network Structure and Access in an
Advertising Fmn: Administrative Science Quarterly 37: 422-447.
1993a Untangling the Web of Interconnections: An Exploration of Competing Explanations for Gender
Differences in Managerial Networks. Cambridge: Harvard University Graduate School of Business
Administration mimeo.
1993b 'Network Centra1ity, Power and Innovation Involvement: Determinants of Technical and
Administrative Roles.' Academy of Management Journal 36: 471-501.
1993c 'Personal Networks of Women and Minorities in Management: A Conceptual Framework:
Academy of Management Review 18: 56-87.
1997 'Paving an Alternative Route: Gender Differences in Managerial Networks: Social Psychology
Quarterly 60: 91-102.
Ibarra, H., and S. B. Andrews
1993 'Power, Social Influence and Sense Making: Effects of Network Centrality and Proximity on
Employee Perceptions.' Administrative Science Quarterly 38: 277-303.
Imal, K.I., I. Nonaka, and H. Takeuchi
1985 'Managing the New Product Development Process: How Japanese Companies Learn and Unlearn.'
In K. Clark, R. Hayes, and C. Lorenz (eds.), The Uneasy Alliance: Managing the Productivity-
Technology Dilemma. Cambridge: Harvard Business School Press.
Inkpen,A.
1995 The Management of International Joint Ventures: An Organiltltional Learning Perspective.
London: Routledge.
Itami, H.
1987 Mobilizing Invisible Assets. Cambridge: Harvard University Press.
braeH, D. N., and I. Talmud
1997 'Getting Aboard: Mode of Recruitment and Gender Composition: The Case of Women Directors in
Israel' International Review of Women and Leadership 3: 26-45.
braeH, D. N., and I. Talmud
1998 The Gender of Social Capital and Recruitment to the Corporate Board. Paper addressed to a
Showcase Symposium on Women in the Corporate Board, Academy of Management Conference,
San Diego, CA.
Jablin, F. M., and K. J. Krone
1987 'Organizational Assimilation.' In C. Berger and S. H. Chaffee (eds.), Handbook of Communication
Science, 711-746. Newbury Park: Sage.
Jackson, M. 0., and A. Wolinsky

1996 'A Strategic Model of Social and Economic Networks.' Journal of Economic Theory 71: 44-74.
Jacobs,J.
1965 The Death and Life of Great American Cities. London: Penguin.
1970 The Economy of Cities. New York: Random House.
Jacoby,S.
1985 Employing Bureaucracy: Managers. Unions. and the Transformation of Work in American
Industry. 1900-1945. New York: Columbia University Press.
Jaffe, A. B.
1986 Technological Opportunity and Spillovers of R&D: Evidence from r1ClIJS' Patents, Profits, and
Market Value.' American Economic Review 76: 984-1002.
Jaffe, A. B., M. Trajtenberg, and R. Henderson
1993 'Geographic Localization of Knowledge Spillovers. as Evidenced by Patent Citations.' Quarterly
Journal of Economics: 577-598.
Jarillo, J. C.
1988 'On Strategic Networks.' Strategic Management Journal 9: 31-41 .
1993 Strategic Networks: Creating the Borderless Organisation. Oxford: Butterworth Heinemann.
Jarillo, J, C., and H. H. Stevenson
1991 'Co-operative Strategies: The Payoffs and the Pitfalls.' Long Range Planning 24.
Jenkins, R., A. Bryman, J. Ford, T. KeD, and A. Beardsworth
1983 'Information in the Labor Market: The Impact of Recession.' Sociology 17: 260-267.
Jensen, M. C.
1984 Takeovers: Folklore and science.' Harvard Business Review 62: 109-121.
Jensen, M. C., and W. H. Meckling
1976 Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure.' Journal of
Financial Economics 3: 305-360.
Johnston, R., and P. R. Lawrence
1988 'Beyond Vertical Integration: the Rise of Value Adding Partnerships.' Harvard Business Review
66: 94-101.
Johnston, R., S. Chambers, C. M. Harland, A. Harrison, and N. Slack
1997 Cases in Operations Management (2nd. edition). London: Pitman.
Jonash, R. S.
1996 'Strategic Technology Leveraging: Making Outsourcing Work for You.' Research Technology
Jones, C., W., Hesterly, and S. Borgattl
1997 'A General Theory of Network Governance: Exchange Conditions and Social Mechanisms.'
Academy of Management Review 22: 911-945.
Judge,G.
1985 The Theory and Practice of Econometrics (2nd edition). New York: Wiley.
Kahneman, D., and A. Tversky
1982 'The Psychology of Preferences.' Scientific American 246: 161-173.
Kahneman, D., and A. Tversky
1984 'Choices, Values, and Frames.' American Psychologists 39: 341-350.
Kalleberg, A. L., D. Knoke, and P. V. Marsden
1995 'lnterorganizational Networks and the Changing Employment Contract.' Connections 18: 32-49.
Kalleberg, A. L., D. Knoke, P. V. Marsden, and J. L. Spaeth
1996 Organizations in America: Analyzjng Their Structures and Human Resource Practices. Newbury
Park: Sage.
Kalleberg, A. L., and J. Lincoln
1988 'The Structure of Earnings Inequality in the United States and Japan.' American Journal of
Sociology 94 (special supplement): 121-153.
Kanter, R. M.
1977 Men and Women of the Corporation. New York: Basic Books.
1983a The Change Masters: Innovation and Entrepreneurship in the American Corporation . New York:
Simon & Schuster.
1983b'Variations in Managerial Career Structures in High-Technology Firms: The Impact of
Organizational Characteristics on Internal Labor Market Patterns.' In P. Osterman (ed.), Internal
Labor Markets. 109-131. Cambridge: MIT Press.
References - 517
1989 When Giants Learn to Dance. London: Unwin.

1991 The Future of Bureaucracy and Hierarchy in Organizational Theory: A Report from the Field.' In
P. Bourdieu and 1. S. Coleman (eds.), Social Theory for a Changing Society, 63-87. Boulder:
Westview Press.
Kanter, R. M., and R. G. Eccles
1992 'Making Network Research Relevant to Practice.' In N. Nohria and R. G. Eccles (eds.), Networks and
Organiwtions: Structure, Form, and Action. Boston: Harvard Business School Press.
Kanter, R. M., and P. S. Myers
1991 'lnterorganizational Bonds and Intraorganizational Behavior: How Alliances and Partnerships
Change the Organizations Forming Them.' In A. Etzioni and P. R. Lawrence (eds.),
Socioeconomics: Toward a New Synthesis, 329-344. Armonk: M.E. Sharpe.
Kanter, R. M., B. Stein, and T. Jick (eds.)
1992 The Challenge of Organiwtional Change: How Companies Experience It and Leaders Guide It.
New York: Free Press.
Kaplan,J.
1995 Stanup! A Silicon Valley Adventure. Boston: Houghton Mifflin.
Katz,L.
1953 'A New Status Index derived from Sociometric Analysis.' Psychometrika 18: 249-56.
Katz,R.
1982 The Effects of Group Longevity on Project Communication and Performance.' Administrative
Science Quanerly 27: 81-104.
Katzenbach, J. R., and D. K. Smith
1993 The Wisdom of Teams: Creating the High-Performance Organiwtion. Boston: Harvard Business
School Press.
Keele,R.
1986 'Mentoring or Networking? Strong and Weak Ties in Career Development.' In L. L. Moore (ed.),
Not as Far as You Think: The Realities of Working Women, 53-68. Lexington: Lexington Books.
Kegan, R.
1982 The Evolving Self: Problem and Process in Human Development. Cambridge: Harvard University
Press.
Kenworthy, L.
1997 'Civic Engagement, Social Capital, and Economic Cooperation.' American Behavioral Scientist 40:
645-656.
Keohane, R. O.
1986 'Reciprocity in International-Relations.' International Organiwtion 40: 1-27.
Kickert, W. J. M., E. H. Klijn, and J. F. M. Koppenjan
1997 Managing Complex Networks. London: Sage.
Kilduff, M.
1990 'The Interpersonal Structure of Decision Making: A social Comparison Approach to
Organizational Choice.' Organiwtional Behavior and Human Decision Processes 47: 27~288.
Kilduff, M., and D. Krackbardt
1993 'Schemas at Work: Making Sense of Organizational Relationships.' Unpublished manuscript.
1994 'Bringing the Individual back in: A Structural Analysis of the Internal Market for Reputation in
Organizations.' Academy of Management Journal 37: 87-108.
Kim, E. U.
1997 Big Business, Strong State: Collusion and Conflict in South Korean Development, 1960-1990.
Albany: SUNY Press.
Kim, Y. H.
1989 Kwangwon Kyongje Tukhye Kyongje [Stateled Economy, Preferential Treatment Economy, in
Korean]. Seoul: Ch'ung Am.
Kimberly, J., and M. Evanisko
1981 'Organizational Innovation: The Influence of the Individual, Organizational, and Contextual
Factors on Hospital Adoption of Technological and Administrative Innovations.' Academy of
Kinch,N.
1992 'Entering a Tightly Structured Network-Strategic Visions or Network Realities.' In M. Forgren
and 1. Johanson (eds.), Managing Networks in International Business, 194-214. Philadelphia:
Gordon & Breach.
Kinston, W.
1992 'Is the United States Right About 'First-to-Invent'?' European IntellectlUJl Property Review 14:
223-226.
Kieimann, E.
1992 Israeli Military Export in the Global Market, Tel Aviv: Am Oved (in Hebrew).
Klein WooltbuIs, R.
1996 'Entrepreneurial Activity through Inter-Organisational Relationships.' Faculty of Technology and
Management, University of Twente, paper RENT X. Research in Entrepreneurship and Small
Business, Brussels, 20--23 November.
Knee, D., and D. Walters
1985 Strategy in Retailing: Theory and Applications. London: Philip Allen.
Knigbt, F. H.
1921 Risk. Uncenainty. and Profit. Hart, Schaftner. and Marx.
Knoke,D.
1990 Organizing for Collective Action: The Political Economies of Associations. New York: Aldine de
Gruyter.
Knoke, D., and R. S. Burt
1983 'Prominence.' In R. S. Burt and M. 1. Minor (eds.). Applied Network Analysis, 195-222. Beverly
Hills: Sage.
Knoke, D., and J. H. Kuklinski
1982 Network Analysis. Beverly Hills: Sage.
Knoke, D., and E. O. Laumann
1982 'The Social Organization of National Policy Domains: An Exploration of Some Structural
Hypotheses.' In P. V. Marsden and N. Un (eds.). Social Structure and Network Analysis, 255-270.
Beverly Hills: Sage.
Knoke, D., F. U. PappI, J. Broadbent, and Y. Tsujinaka
1996 Comparing Policy Networks: Labor Politics in the U.S.. Germany and Japan. New York:
Cambridge University Press.
Kocban, T. A., H. C. Katz, and R. B. McKersie
1994 The Transformation of American Industrial Relations. Ithaca: ILR Press.
Koelemeijer, K.
1995 'The Retail Service Encounter: Identifying Critical Service Experiences.' In P. Kunst and J.
Lemmink (eds.), Managing Service Quality. 2~3 . London: Chapman.
KoelewiJn, J.
1997 Een Nationale Waan [A National Delusion, in Dutch]. Amsterdam: De Bezige Bij.
Kogut,B.
1984 'Normative Observations on the International Value Added Chain and Strategic Groups.' Journal
of International Business Studies Fall: 151-167.
1985 'Designing Global Strategies: Comparative and Competitive Value Added Chains.' Sloan
Management Review Summer: 15-28.
1988 'A Study of the Ufe Cycle of Joint Ventures.' In F. J. Contractor and P. Lorange (eds.),
Cooperative Strategies in International Business: Joint Ventures and Technology Partnerships
Between Firms. 169-185. Lexington: Lexington Books.
Kogut, B., and U. Zander
1996 'What Firms Do? Coordination, Identity and Learning.' Organization Science 7: 502-518.
1997 'Did Socialism Fail to Innovate? A Natural Experiment of the Two Zeiss Companies.' Working
paper. The Wharton School, University of Pennsylvania.
Kogut, B., W. Shan, and G. Walker
1992 'The Make-or-Cooperate Decision in the Context of an Industry Network.' In N. Nohria and R. G.
Eccles (eds.). Networks and Organizations: Structure. Form. and Action, 348-365. Boston:
Harvard Business School Press.
Kollock,P.
1993 'An Eye for an Eye Leaves Everyone Blind: Cooperation and Accounting Systems.' American
Koput, K. W., W. W. Powell, and L. Smith-Doerr
1998 'Interorganizational Relations and Elite Sponsorship: Mobilizing Resources in Biotechnology.'
Paper presented at the annual meetings of the American Sociological Association, San Fransisco.
References - 519
Koput, K. W., L. Smith-Doerr, and W. W. Powell

1997 'Strategies of Learning and Industry Structure: The Evolution of Networks in Biotechnology.' In J.
P. Walsh and A. S. Huff (eds.), Organizational Learning and Strategic Management. Advances in
Strategic Management Research, Vol. 14: 229-254. Greenwich: JAI Press.
Korenman, S., and D. Neumark
1991 'Does Mamage Really make Men more Productive?' Journal of Human Resources 26: 282-307.
Kotter, J. P.
1982 The General Managers. New York: Free Press.
Kotter, J. P., and J. L. Haskett
1992 Corporate Culture and Perfonnance. New York: Free Press.
Krackhardt, D.
1990 'Assessing the Political Landscape: Structure, Cognition and Power in Organizations.'
1992 The Strength of Strong Ties: The hnportance of Philos in Organisations.' In N. Nohria and R. G.
Eccles (eds.), Networks and Organisations: Structure, Form and Action, 216-240. Boston.:
1995 'Entrepreneurial Opportunities in an Entrepreneurial Fmn: A Structural Approach.'
Entrepreneurship: Theory and Practice 19: 53-69.
Krackhardt, D., and D. J. Brass
1994 'Intraorganizational Networks: The Micro Side.' In S. Wassennan and J. Galaskiewicz (eds.),
Advances in Social Network Analysis, 207-229. Thousand Oaks: Sage.
Krackhardt, D., and J. R. Hanson
1993 'Infonnal Networks: The Company Behind the Chart.' Harvard Business Review 71 : 104-111.
Krackhardt, D., and L. W. Porter
1985 'When Friends Leave: A Structural Analysis of the Relationship Between Turnover and Stayers'
Attitudes.' Administrative Science Quarterly 30: 242-261.
1986 The Snowball Effect: Turnover Embedded in Communication Networks.' Journal of Applied
Psychology 71: 5(}-'55.
Krackhardt, D., and R. N. Stem
1988 'Infonnal Networks and Organizational Crisis: An Experimental Simulation.' Social Psychology
Kram,K.E.
1985 Mentoring at Work: Developmental Relationships in Organizational life. Glenview: Scott,
Foresman and Company.
Kramer, R., M. Brewer, and B. Hanna
1996 'ColJective Trust and Collective Action: The Decision to Trust as a Social Decision.' In R. Kramer
and T. Tylor (eds.), Trust in Organizations: Frontiers of Theory and Research, 357-389. London,
Sage.
Kratzer, J., J. M. L. van Engelen, R. Th. A. J. Leenders
1998 'Structuring the communication infrastructure of innovation teams for success.' Journal of Product
Innovation Management. Conference Proceedings Series 9: 43·56.
Kruska1, J. B.
1964 'Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis.' Psychometrika
29: 1-27.
Laat, P. B. de
19% 'Dangerous Liaisons: Sharing Knowledge within R&D Alliances. ' Paper EMOT workshop (European
Management and Organization in Transition, of the European Science Foundation), September,
Modena; Faculty of Philosophy, University of Groningen, the Netherlands.
Lablanca, G., and D. J. Brass
1997 'Exploring the Social Ledger: The Role of Negative Affective Relationships in Social Networks.'
Proceedings of the Academy of Management Meetings, Boston.
Lablanca, G., D. J. Brass, and B. Gray
1998 'Social Networks and Perceptions of Intergroup Conflict: The Role of Negative Relationships and
Third Parties.' Academy of Management Journal 41 : 55-67.
Lamming,R.
1989 The Causes and Effects of Structural Change in the European Automotive Components Industry.
lMVP. MIT. Cambridge. Mass.
1993 Beyond Partnership: Strategiesfor Innovation and Lean Supply. Englewood Cliffs: Prentice Hall.
Langendijk, H. P.
1990 Wisseling van Accountantskantoor: een Fata Morgana in de Toekomst? MaandbUuJ voor
Accountancy en Bedrijfseconomie 64: 264-270.
Lanjouw,J. o.
1993 'Patent Protection: of what value and for how long?' NBER Working Paper #4475.
Lanjouw, J. 0., A. Pakes, and J. Putnam
1996 'How to Count Patents and Value Intellectual Property: Uses of Patent Renewal and Application
Data.' NBER Working Paper #5741.
Larson,A.
1992 'Network Dyads in Entrepreneurial Settings: A Study of the Governance of Exchange
Relationships.' Administrative Science Quanerly 37: 76-104.
Lascelles, D. M., and B. G. Dale
1990 'Examining the Barriers to Supplier Development.' International Journal of Quality and Resource
Management.
Laumann, E. O.
1973 Bonds of Pluralism: The Fonn and Substance of Urban Social Networks. New York: Wiley.
Laumann, E. 0., and D. Knoke
1987 The Organizational State: Social Choice in National Policy Domains. Madison: University of
Wisconsin Press.
Laumann, E. 0., and F. U. Pappi
1976 Networks of Collective Action: A Perspective on Community Influence . New York: Academic
Press.
Laumann, E. 0., P. V. Marsden, and D. Prensky
1983 'The Boundary Specification Problem in Network Analysis.' In R. S. Burt and M. 1. Minor (eds.),
Applied Network Analysis: A Methodological Introduction, 18--34. Beverly Hills: Sage.
Lawler, E. E., m
1992 The Ultimate Advantage: Creating the High-Involvement Organization. San Francisco: Jossey-
Bass.
Lawrence, P. R., and J. W. Lorsch
1986 Organization and Environment. Boston: Harvard Business School Press.
Lawrence, R. Z.
1991 'Efficient or Exclusionist? The Import Behavior of Japanese Corporate Groups.' Brookings Papers
on Economic Activity 1: 311-314.
Lazega,E.
1992a 'Analyse de R~seaux d'une Organisation Collegiale: les Avocats d'Affaires.' Revue Franraise de
Sociologie 33: 559-589.
1992b Bureaucratie, Collegialite et Professionnalisme: Structure et Fonctionnement I"fonnels des
Grandes Finnes d'Avocats d'Affaires en Nouvelle Angleterre. Report to the Swiss National
Foundation for Scientific Research, Geneva, Switzerland.
1993 'Bureaucratie et Collegialite dans les Firmes Americaines d'Avocats d'Affaires.' Droit et Societe,
23/24: 15-40.
1994 'Les Conflits d'lnterets dans les Cabinets Americains d'Avocats d'Affaires: Concurrence et Auto-
regulation.' Sociologie du Travail 35: 315-336.
1995a 'Concurrence, Cooperation et Flux de Conseils dans un Cabinet Americain d'Avocats d'Affaires:
Les Echanges d'Idees entre Collegues.' Revue Suisse de Sociologie 21: 61-84.
1995b 'Protecting the Common Good Among Equals: A Lateral Control Regime of Partners in a Law
Finn.' LAS MAS Preprint N° 96/1.
1998a Riseaux Sociaux et Structures Relationnelles. Paris: Presses Universitaires de France.
1998b The Collegial Phenomenon: Combining Neo-Weberian and Structural Approaches to Collective
Action Among Peers'. Unpublished manuscript.
1999 'Teaming Up and Out? Cooperation and Solidarity in a Collegial Organization.' In W.Raub and
1.Weesie (eds.), The Management of Durable Relations. Amsterdam: Thesis Publishers
(forthcoming).
Lazega, E., and D. Krackhardt
1998 'Spreading and Shifting Costs of Lateral Control in a Law Partnership: Structural Analysis at the
Individual Level.' Structural Analysis 1. Forthcoming.
Lazega, E., and M. O. Lebeaux
1995 'Capital Social et Contrainte Laterale.' Revue Franraise de Sociologie 36: 759-777.
References - 521
Lazega, E., and M. Van Duijn

1997 'Position in Formal Structure, Personal Characteristics and Choices of Advisors in a Law Firm: A
Logistic Regression Model for Dyadic Network Data.' Social Networks 19: 375-397.
Lazega, E., and P. E. Pattison
1998 'Generalized and Multiplex Exchange in Organizational Cooperation.' Forthcoming.
Lazega, E., and S. Vari
1992 'Acteurs. Cibles et Leviers: Analyse Factorielle des Relations de Controle Indirect dans une Finne
Americaine d'Avocats d'Affaires.' Bulletin de Methodologie Sociologique 37: 41-51.
Lazerson, M. H.
1988 'Organizational Growth of Small Firms.' American Sociological Review 53: 330-342.
1995 'A New Phoenix? Modem Putting-Out In the Modena Knitwear Industry.' Administrative Scienc(!
Lee, L., and D. Dobler
1965 Purchasing and Materials Management. New York: McGraw-Hili.
Leenders, R. Th. A. J.
1991 'Prediction of Committees in Collective Decision Making Processes.' NWO Grant Proposal
(subsidized). Department of Sociology. University of Groningen.
1995a 'Models for Network Dynamics: A Markovian Framework.' Journal of Mathematical Sociology
20: 1-21.
1995b Structure and Influence. Amsterdam: Thesis Publishers.
1996 'Evolution of Friendship and Best Friendship Choices.' Journal of Mathematical Sociology 21:
133-148.
1997 'Longitudinal Behavior of Network Structure and Actor Attributes: Modeling Interdependence of
Contagion and Selection.' In P. Doreian and F.N. Stokman (eds.). Evolution of Social Networks.
Langhorne: Gordon and Breach.
1999 'Modeling social influence through network autocorrelation: on constructing the weight matrix.'
Social Networks Forthcoming.
Le Grand, C., R. Szulkin, and M. Tihlin
1994 'Organizational Structures and Job Rewards in Sweden: Acta Sociologica 37: 231-251.
Leifer, E. M.
1988 'Interaction Preludes to Role Setting: Exploratory Local Action.' American Sociological Review
53: 865-878.
1990 'Market and Authority in League Sports: American Journal of Sociology 96: 655-683.
Leonard-Barton, D.
1995 WeI/springs of Knowledge. Building and Sustaining the Sources of Innovation. Boston: Harvard
Business School Press.
Leontlef, W. W.
1951 The Structure of the American Economy, 1919-1939 (2nd. edition). New York: Oxford University
Press.
Lerner,J.
1995 'Patenting in the Shadow of Competitors: Journal of Law and Economics 38: 463-496.
Levaclc, R.
1987 Economic Policy Making. Brighton: Wheatsheaf.
Levl,M.
1996 'Social and Unsocial Capital-a Review-Essay of Putnam. Robert Making Democracy Work:
Politics and Society 24: 45-55.
Levine, D. I.
1990 'Participation, Productivity, and the Firm's Environment.' California Management Review 32: 86-
100.
Levine,J. H.
1972 'The Sphere of Influence.' American Sociological Review 37: 14-27.
Levinson, D. J., C. N. Darrow, E. B. Klein, M. A. Levinson, and B. McKee
1978 Seasons ofa Man's Ufe. New York: Knopf.
Uvi-Strauss, C.
1949 Les Formes Etementaires de la Parente. Paris: PIon.
Lewis,J.D.
1990 Partnerships for Profit: Structuring and Managing Strategic Alliances. New York: Free Press.
Lewis, J. D., and A. Weigert

1985 'Trust as a Social Reality.' Social Forces 63: 967-985.
Likert, R.
1961 The Human OrganiZiltion. New York: Wiley.
Limerick, D., and B. Cunnington
1993 Managing the New OrganiZiltion: A Blueprint for Networks and Strategic Alliances. San
Francisco: Jossey-Bass.
Lln,N.
1982 'Social Resources and Instrumental Action.' In P. V. Marsden and N. Un (eds.), Social Structure
and Network Analysis. Beverly Hills: Sage.
1990 'Social Resources and Social Mobility: A Structural Theory of Status Attainment.' In R. L. Breiger
(ed.), Social Mobility and Social Structure. New York: Cambridge University Press.
1995a 'Les Resources Sociales: une lMorie du Capital Social.' Revue Fran~aise de Sociologie 36: 685-
704.
1995b 'Institutions and Networks: The Infrastructure of Society. ' Paper presented to International Social
Network Conference, London.
Lin, N., W. M. Ensel, and J. C. Vaughn
1981 'Social Resources and Strength of Ties: Structural Factors in Occupational Status Attainment.'
Lin, N., and M. Durnil
1986 'Access to Occupations through Social Ties.' Social Networks 8: 365-385.
Lincoln, J. R., Ahmadjian, and E. Mason
1998 'Organizational Learning and Purchase-Supply Relations in Japan: Hitachi, Matsushita and Toyota
Compared.' California Management Review (forthcoming).
Lincoln, J . R., M. L. Gerlach, and P. Takahashi
1992 'Keiretsu Networks in the Japanese Economy: A Dyad Analysis of Intercorporate Ties.' American
Lindenberg, S. M .
1990 'Homo Socio-Economicus: the Emergence of a General Model of Man in the Social Sciences.'
Journal of Institutional and Theoretical &onomics 146: 727-748.
1993 'Club Hierarchy, Social Metering and Context Instruction: Governance Structures in Response to
Varying Self-Command Capital.' In S. M. Undenberg and H. Schreuder (eds.), Interdisciplinary
Perspectives on OrganiZiltion Studies. Oxford: Pergamon Press.
1996 'Multiple-Tie Networks, Structural Dependence, and Path-Dependency: Another Look at Hybrid
Forms of Governance.' Journal of Institutional and Theoretical Economics 152: 188-196.
Llpnack, J., and J. Stamps
1994 The Age of the Network. Essex Junction: Omneo.
Lippman, S. A., and R. P. Romelt
1982 'Uncertain lmitability: An Analysis of Interfirm Differences in Efficiency under Competition.' Bell
Journal of Economics 13: 418-438.
Long,J.s.
1997 Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks: Sage.
Lorange, P., and J. Roos
1992 Strategic Alliances: Fonnation, Implementation, and Evolution . Cambridge: Blackwell Business.
Lorange, P., J. Roos, and P. G. Bronn
1992 'Building Successful Alliances.' Long Range Planning 25: 10-17.
Lorenz, E. H.
1993 'Flexible Production Systems and the Social Construction of Trust.' Politics and Society, 21: 307-
324.
Lorenzonl, G., and C. Baden-FuUer
1995 'Creating a Strategic Center to Manage a Web of Partners.' California Management Review 37:
146-163.
Lorenzoni, G., and O. Ornati
1988 'Constellations of Firms and New Ventures.' Journal of Business Venturing 3: 41-57.
Lorenzonl, G., and A. Llparinl
1997 'Leveraging Internal and External Competencies in Boundary Shifting Strategies. A Longitudinal
Approach.' Working Paper, Department of Management, Economics faculty, University of
Bologna.
References - 523
Louls,M. R.
1990 'Acculturation in the Workplace: Newcomers as Lay Ethnographers.' In I. L. Goldstein (ed.),
Frontiers of Industrial and OrganilJltionai Psyclwlogy, 85-129. San Francisco: Jossey-Bass.
Loury,G.
1977 'A Dynamic Theory of Racial Income Differences.' In P. A. Wallace and A. LeMund (eds.)
Women, Minorities, and Employment Discrimination . Lexington: Lexington Books.
1987 'Why Should We Care about Group Inequality?' Social Philosophy and Policy 5: 249-271.
Low, M. B., and I. C. MacMillan
1988 'Entrepreneurship: Past Research and Future Challenges.' Journal of Management 14: 139-161.
Luhmann,N.
1979 Trust and Power. Chichester: Wiley.
Lusch,R. F.
1976 'Sources of Power: Their Impact on Intrachannel Conflict.' Journal of Marketing Research 13:
382-390.
Lusch, R. F., and J. R. Brown
1982 'A Modified Model of Power in the Marketing Channel.' Journal of Marketing Research 19: 312-
323.
Macaulay, S.
1963 Non-Contractual Relations in Business: A Preliminary Study. American Sociological Review 28:
55-66.
Macbeth, D. K., and N. Ferguson
1994 Partnership Sourcing: An Integrated Supply Chain Approach. London: Pitman.
MacKie-Mason, J. K.
1990 'Do Firms Care who Provides Their Financing?' In R. G. Hubbard (ed.), Asymmetric Information,
Corporate Finance, and Investment, 63-103. Chicago: University of Chicago Press.
MacLachlan, A.
1995 'Trusting Outsiders to do Your Research: How does Industry Learn to do it?' Research Technology
Macneil, I. R.
1986 'Exchange Revisited, Individual Utility and Social Solidarity.' Ethics 96: 567-593.
Macy, M. W., and A. Flache
1995 'Beyond Rationality in Models of Choice.' Annual Review of Sociology 21 : 73-91.
Maddala, G. S.
1983 Limited-Dependent and Qualitative Variables in Econometrics. Cambridge: Cambridge University
Press.
Malster, D. H.
1993 Managing the Professional Service Firm. New York: Free Press.
Mansfield, E.
1984 'R&D and Innovation: Some Empirical Findings.' In Z. Griliches (ed.), R&D Patents and
Productivity. Chicago: University of Chicago Press.
March,J.G.
1988 Decisions and OrganilJltions. New York: Blackwell.
Markovsky, B., and M. V. Chaffee
1995 'Social Identification and Solidarity: A Revised Formulation.' In B. Markovsky, 1. O'Brien, and K.
Heimer (eds.), Advances in Group Processes 12,249-270. Greenwich: JAI Press.
Markovsky, B., and E. J. Lawler
1994 'A New Theory of Group Solidarity.' In B. Markovsky, J. O'Brien, and K. Heimer (eds.), Advances
in Group Processes II, 113-137. Greenwich: JAI Press.
Markowsky, B., D. Willer, and T. Patton
1988 'Power Relations in Exchange Networks.' American Sociological Review 53 : 220-236.
Marris, R.
1964 The &onomic Theory of Managerial Capitalism. London: MacMillan.
Marsden, P. V.
1982 'Brokerage Behavior in Restricted Exchange Networks.' In P. V. Marsden and N. Lin (eds.), Social
Structure and Network Analysis. Beverly Hills: Sage.
1983 'Restricted Access in Networks and Models of Power.' American Journal of Sociology 88: 686-
717.
1987 'Core Discussion Networks of Americans.' American Sociological Review 52: 122-131.
1995 'The Staffing Process. Recruitment and Selection Methods.' In A. L. Kalleberg, D. Knoke, P. V.
Marsden and J. L Spaeth, OrganiYltions in America. Analyzjng Their Structures and Human
Resources Practices, 133-156. Thousand Oaks: Sage.
Marsden, P. V., and K. E. Campbell
1990 'Recruitment and Selection Processes: The Organizational Side of Job Searches.' In R. L. Breiger
(ed.), Social Mobility and Social Structure, 59-79. New York: Cambridge University Press.
Marsden, P. V., and J. S. Hurlbert
1988 'Social Resources and Mobility Outcomes: A Replication and Extension.' Social Forces 66: 1038-
1059.
Marsh, D., and R. A. W. Rhodes
1992 Policy Networks in British Government. Oxford: Clarendon Press.
Marshal,A.
1890 Principles of &onomics (First Edition). London: Oxford University Press.
1923 Industry and Trade . London: Macrnillam.
Martinelli, A.
1994 'Entrepreneurship and Management.' In N. Smelser and R. Swedberg (eds.), The Handbook of
Economic Sociology, 476-503. Princeton: Princeton University Press.
Mathworks, Inc., The
1994 MATLAB version 4.2c. Natick: The Mathworks, Inc.
Mayer, R. C., J. H. Davis, and F. D. Schoorman
1995 'An Integrative Model of Organizational Trust.' Academy of Management Review 20: 709-734.
McAllister, D. J.
1995 'Affect· and Cognitive-Based Trust as Foundations for Interpersonal Cooperation in
Organizations.' Academy of Management Journal 38: 24-59.
McClelland, D. C.
1967 The Achieving Society. New York: Free Press.
McKenna, C. J.
1985 Uncertainty and the Labour Market: Recent Developments in the Job-Search Theory. Brighton:
Wheatsheaf Books.
McPherson, J. M., P. A. Popielarz, and S. Drobnic
1992 'Social Networks and Organizational Dynamics.' American Sociological Review 57: 153-170.
Mehra, A., M. Kilduff, and D. J. Brass
1998 'At the Margins: A Distinctiveness Approach to the Social Identity and Social Networks of
Underrepresented Groups.' Academy of Management Journal 41: 441-452.
Menon, G., B. Bickart, S. Sudman, and J. Blair
1995 'How Well Do You Know Your Partner? Strategies for Formulating Proxy-Reports and Their
Effects on Convergence to Self-Reports.' Journal of Marketing Research 32: 75-84.
Mergers and Acquisitions
1986-1993 The Journal of Corporate Venture. Washington: Mergers and Acquisitions.
Merges, R. P., and R. R. Nelson
1990 'On the Complex Economics of Patent Scope.' Columbia Law Review 90.
Merton, R. K.
1964 'The Mathew Effect in Science.' Science 159: 56-63.
Meurer, M. J.
1989 'The Settlement of Patent litigation.' RAND Journal of Economics 20: 77-91.
Meyer, J. W., and B. Rowan
1991 'Institutionalized Organizations: Formal Structure as Myth and Ceremony.' In W. W. Powell and P.
J. DiMaggio (eds.), The New Institutionalism in OrganiYltional Analysis, 41-62. Chicago:
University of Chicago Press.
Meyer,M. W.
1994 'Measuring Performance in Economic Organizations.' In N. Smelser and R. Swedberg (eds.),
Handbook of &onomic Sociology. Princeton: Russell Sage.
Meyerson, E. M.
1992 The Impact of Ownership Structure and Executive Team Composition on Firm Peiformance.
Stockholm: Almqvist & Wickselllnternational.
1994 'Human Capital, Social Capital and Compensation: the Relative Contribution of Social Contacts to
Managers' Incomes.' Acta Socioiogica 37: 383-399.
References - 525
Miles, R. E., and C. C. Snow

1986 'Organizations: New Concepts for New Fonns.' California Managment Review 28: 62-73.
1992 'Causes of Failures in Network Organizations.' California Management Review 34: 53-72.
1995 'The New Network Finn: A Spherical Structure Built on a Human Investment Philosophy.'
Organizational Dynamics 23: 5-18.
Milgrom, P., and J. Roberts
1992 Economics, Organization and Management. Englewood Cliffs: Prentice Hall.
1996 'Bargaining Costs, Influence Costs. and the Organization.' In L. Puttennan and R. S. Kroszner
(eds.). The Economic Nature of the Firm: A Reader (2nd. edition). Chicago: University of Chicago
Press.
Miller, J. A.
1976 Exploring Some Alternative Measure of Consumer Satisfaction in Marketing: 1776-1976 and
Beyond. Chicago.
Miller, D.
1991 'Stale in the Saddle: CEO Tenure and the Match between Organization and Environment.'
Management Science 37: 34-52.
Miller, G. J.
1992 Managerial Dilemmas: The Political Economy of Hierarchy. Cambridge: Cambridge University
Press.
MilIer,P.
1994 'Accounting and Objectivity: The Invention of Calculating Selves and Calculable Spaces.' In A.
Megill (ed.). Rethinking Objectivity. Durham: Duke University Press.
Millson, M. R., S. P. Raj, and D. Wilemon
1996 'Strategic Partnering for Developing New Products.' Research Technology Management 39: 41-
49.
Miner, A. S., and S. E. Estler
1985 'Accrual Mobility: Job Mobility in Higher Education through Responsibility Accrual.' Journal of
Higher Education 56: 121-143.
Mintz, B., and M. Scbwartz
1985 The Power Structure of American Business. Chicago: University of Chicago Press.
Mintzberg, H.
1973 The Nature ofManagerial Work. New York: Harper and Row.
1983 Structure in Fives: Designing Effective Organizatians. Englewood Cliffs: Prentice-Hall.
Misblra,A.
1996 'Organizational Response to Crisis: The Centrality of Trust.' I. R. Kramer and T. Tylor (eds.).
Trust in Organizations: Frontiers of Theory and Research. 261-287. London: Sage.
Misurni,K.
1989 'Communication Structure. Trust, and the Free Rider Problem.' Journal of Mathematical Sociology
14: 273-282.
Mitsui, I
1990 'EC-Japan Subcontracting Opportunities.' Proc. of Palais des Congres Conference. Jun 21-22. EC
DGXXIII.
Mizrucbl, M. S.
1992 The Structure of Corporate Political Action. Cambridge: Harvard University Press.
1994 'Social Network Analysis: Recent Achievements and Current Controversies.' Acta Sociologica 37:
329-343.
Mizrucbi, M. S., and L. B. Stearns
1994 'A Longitudinal Study of Borrowing by Large American Corporations.' Administrative Science
Mizrucbi, M. S., P. Mariolis, M. Scbwartz, and B. Mintz
1986 'Techniques for Disaggregating Centrality Scores in Social Networks.' In N. B. Tuma (ed.).
Sociological Methodology 1986.26-48. Washington: American Sociological Association.
Moerbeek, H., H. Flap, and W. U1tee
1997 'How Dutch Men and Women find a Job and What it Does to Their Career.' Paper presented at the
conference on Social Stratification in Modern Welfare States. Research Committee 28. Stockholm.
May 29th-June 2nd.
Monge, P. R., and E. M. Eisenberg

1987 'Emergent Communication Networks.' In F. M. Jablin. L. L. Putman. K. H. Roberts. and L. W.
Porter (eds.). Handbook of Organiwtional Communication: An Interdisciplinary Perspective.
304-342. Newbury Park: Sage.
Monge, P.R., M. Cozzens, and N. Contractor
1991 'Communication and Motivational Predictors of the Dynamics of Organizational Innovation.'
Organiwtion Science 2: 1-25.
Moore,G.
1992 'Gender and lnfonnal Networks in State Government.' Social Science Quanerly 73: 46-61 .
Moran, P., and S. Ghoshal
1996 'Value Creation by Firms.' Academy of Management Best Papers Proceedings: 41-45.
Morek, R., A. Shleifer, and R. W. Vishny
1990 '00 Managerial Objectives drive bad Acquisitions?' Journal of Finance 45: 31-48.
Morgan, I.
1987 The Purchasing Revolution.' McKinsey Quanerly Spring: 49-55.
Morgan, G., and S. Maddock
1997 Markets, Managers and the NHS: An Institutionalist Approach. University of Manchester:
Manchester Business School.
Mowery, D. C. (ed.)
1988 International Collaborative Ventures in US ManUfacturing. Cambridge: Ballinger.
Mueller, R. K.
1986 Corporate Networking. New York: Free Press.
Nahapiet, J ., and S. Ghoshal
1998 'Social Capital. Intellectual Capital, and the Organizational Advantage.' Academy of Management
Review 23: 242-266.
Nakatani, I.
1984 'The Economic Role of Financial Corporate Grouping.' In M. Aoki (ed.). The Economic Analysis
of Japanese Firm. 227-258. New York: North-Holland.
Neckennan, K. M., and R. M. Fernandez
1997 'Keeping a Job: Network Hiring and Turnover in a Retail Bank.' Paper presented at the American
Sociological Association. Toronto. Ontario. Aug. 13. 1997.
Nelson, R. E.
1989 The Strength of Strong Ties: Social Networks and Intergroup Conflict in Organizations.' Academy
of Management Journal 32: 377-401.
Nelson, R. R., and S. G. Winter
1982 An Evolutionary Theory of Economic Change. Cambridge: Harvard University Press.
Nelson, R. L .
1988 Partners With Power: The Social Transformation of the Large Law Firm. Berkeley: University of
California Press.
Neu,D.
1991 Trust, Contracting and the Prospectus Process.' Accounting, Organiwtions and Society 16: 243-
256.
Newcomb,T.
1961 The Acquaintance Process. New York: Holt, Rinehart, and Winston.
Newton,K.
1997 'Social Capital and Democracy.' American Behavioral Scientist 40: 575-586.
Nicholls, J. G.
1978 The Development of the Concepts of Effort and Ability. Perception of Academic Attainment, and
the Understanding that Difficult Tasks Require More Ability.' Child Development 49: 800-814.
Nicholson, N.
1993 'Purgatory or Place of Safety? The Managerial Plateau and Organizational Agegrading.' Human
Relations 46: 1369-1389.
Niemiiller, B., W. H. Van Schuur, and F. N. Stokman
1980 'Stochastic Cumulative Scaling: Mollen Scale. Mollen Test.' STAP User's Manual, Volume 4.
Technisch Centrum FSW. University of Amsterdam.
Nishiguchi, T.
1994 Strategic Industrial Sourcing: The Japanese Advantage. New York: Oxford University Press.
References - 527
Nishiyama, T.
1982 'The Structure of Managerial Control: Who Owns and Controls Japanese Businesses.' Japanese
Economic Studies Fall: 37-77.
Nobeoka, K., and J. H. Dyer
1998 The Influence of Customer Scope on Supplier Learning and Peiformance in the Japanese
Automobile Industry. Working Paper, The Wharton School, University of Pennsylvania,
Philadelphia.
Noe, R.A.
1988a 'An Investigation of the Determinants of Successful Assigned Mentoring Relationships.' Personnel
Psychology 41: 457-479.
1988b 'Women and Mentoring: A Review and Research Agenda.' Academy of Management Review 13:
65-78.
Nohria,N.
1992a 'Is a Network Perspective a Useful Way of Studying Organizations?' In N. Nohria and R. G. Eccles
(eds.), Networks and Organizations: Structure, Form, and Action. Boston: Harvard Business
School Press.
1992b 'Information and Search in the Creation of New Business Ventures: The Case of the 128 Venture
Group.' In N. Nohria and R. G. Eccles (eds.), Networks and Organizations: Structure, Form, and
Action, 240-261 . Boston: Harvard Business School Press.
Nohria, N., and R. G. Eccles
1993 'Face-to-Face: Making Network Organizations Work.' In N. Nohria and R. G. Eccles (eds.),
Networks and Organizations: Structure, Form, and Action, 288-308. Boston: Harvard Business
School Press.
Nohria, N., and R. G. Eccles (eds.)
1992 Networks and Organizations: Structure, Form, and Action. Boston: Harvard Business School
Press.
Nohria, N., and S. Ghoshal
1997 The Differentiated Network: A New Model for Organizing Multinational Corporations. San
Francisco: Jossey Bass.
Nohrla, N., and J. D. Berkley
1995 'From Structure to Structuring: A Pragmatic Perspective on Organizational Design.' Harvard
Business School Working Paper #96-053.
Nonaka, I., and H. Takeuchi
1995 The Knowledge Creating Company. Oxford: Oxford University Press.
Nooteboom, B.
1992 'Towards a Dynamic Theory of Transactions.' Journal of Evolutionary Economics 2: 281-299.
1996 'Trust, Opportunism and Governance: A Process and Control Model.' Organization Studies 17:
985-1010.
1997 'Voice and Exit Based Forms of Corporate Control: Anglo-American, European and Japanese.' Paper
for the EAEPE conference, Athens, 6-9 November.
1998a 'Cost, Quality and Learning Based Governance of Transactions.' In M. Colombo (ed.), The Changing
Boundaries of the Finn. London: Routledge.
1998b 'Trust as a Governance Device: Theory and Empirical Evidence.' Paper for a symposium on cultural
factors in economic growth. Marienrode, Germany, 3-5 April.
1999 'The Dynamic Efficiency of Networks.' In A. Grandori (ed.), Inter·Firm Networks: Organizational
and Industrial Competitiveness. London: Routledge (forthcoming).
Nooteboom, B., J. Berger, and N. G. Noorderhaven
1997 'Effects of Trust and Governance on Relational Risk.' Academy of Management Journal 40: 308-338.
Norburn, D., and S. Birley
1988 'The Top Management Team and Corporate Performance.' Strategic Management Journal 9: 225-
237.
North,D.C.
1990 Institutions, Institutional Change and Economic Peiformance. Cambridge: Cambridge University
Press.
North, D. C., and R. Thomas
1973 The Rise of the New World: ANew Economic History. Cambridge: Cambridge University Press.
Northrup, H. R., R. M. Cowin, L. G. Vanden Plas, and W. E. Fulmer
1978 The Objective Selection of Supervisors. Philadelphia: Industrial Research Unit, The Wharton
School, University of Pennsylvania.
Obstfeld, D.
1997 'Different Fonns of Social Capital: Sorting Out Two Alternative Conceptions of Social
Entrepreneurship.' Paper presented at the Interdisciplinary Consortium for Organizotional Studies
(ICOS). University of Michigan.
Office of Technology Assessment and Forecast, U.s. Department of Conunerce, Patent, and
Trademark Office
1976 Technology Assessment and Forecast. U.S. GPO. Washington.
Obmae,K.
1989 The Global Logic of Strategic Alliances.' Harvard Business Review 67.
Oliver, R. K., and M. D. Webber
1982 Outlook. Booz Allen and Hamilton Inc
Olson,M.
1965 The Logic of Collective Action. Cambridge: Harvard University Press.
Ornta,S. W.F.
1995 Critical Success Factors in Biomedical Research and Pharmaceutical Innavation. Dordrecht:
Kluwer Academic Publishers.
O'Neil, R., and R. E. Quinn
1983 'Applications of the Competing Values Frameworlc:.' Human Resource Management 32: 1-7.
Opotow,S.
1990 'Moral Exclusion and Injustice: An Introduction.' Journal of Social Issues 46: 1-20.
O'ReDly,B.
1994 The New Deal: What Companies and Employees Owe One Another.' Fortune 13: 44-52.
Osborn, R. N., and C. C. Baughn
1990 'Fonns of Interorganizational Governance for Multinational Alliances.' Academy of Management
Journal 33: 503-519.
Ostrom,E.
1994 'Constituting Social Capital and Collective Action.' Journal of Theoretical Politics 6: 527-562.
Packer, K., and A. Webster
1996 'Patenting Culture in Science: Reinventing the Wheel of Scientific Credibility.' Science,
Technology, and Human Values 21: 427-53.
Pakes, A., and M. Schankerman
1984 The Rate of Obsolescence of Patents, Research Gestation Lags, and the Private Rate of Return to
Research Resources.' In Z. Griliches (ed.), R&D, Patents, and Productivity. Chicago: University of
Chicago Press.
Pakes, A., and Z. Griliches
1980 'Patents and R&D at the Firm Level: A First Look.' Economics Letters 5: 377-81.
Palmer, D. A., P. D. Jennings, and X. Zhou
1993 'Late Adoption of the Multidivisional Form by Large U.S. Corporations: Institutional, Political,
and Economic Accounts.' Administrative Science Quarterly 38: 1~131.
Papa,M.J.
1990 'Communication Networlc: Patterns and Employee Performance with a New Technology.'
Communication Research 17: 344-368.
Parasuraman, A., V. A. Zeithaml, and L. L. Berry
1985 'A Conceptual Model of Service Quality and its Implications for Future Research.' Journal of
Parcel, T. L., R. L. Kaufman, and L. Jolly
1991 'Going Up the Ladder: Multiplicity Sampling to Create Unked Macro-to-Micro Organizational
Samples.' In Peter V. Marsden (ed.), Sociological Methodology 1991, 43-79. Oxford: Basil
Blackwell.
Parker, R. E.
1994 Flesh Peddlers and Warm Bodies: The Temporary Help Industry and Its Workers. New
Brunswick: Rutgers University Press.
Parsons, T.
1951 The Social System. Glencoe: Free Press.
1963 'On the Concept of Influence.' Public Opinion Quarterly 27: 37-62.
References - 529
Pennings, J. M.
1980 Interlocking Directorates. San Francisco: Jossey Bass.
Pennings, J. M., and F. Barianto
1992 The Diffusion of Technological Innovation in the Commercial Banking Industry.' Strategic
Pennings, J. M., K. Lee, and A. Van Witteloostuyn
1998 'Human and Capital, Social Capital, and Firm Dissolution.' Academy of Management Journal 41:
425-440.
Penrose,E.
1959 The Theory of the Growth of the Firm. Oxford: Basil Blackwell.
Perrow,C.
1992 'Small Firm Networks.' In N. Nohria and R. G. Eccles (eds.), Networks and Organiwtions:
Structure, Form, and Action, 445-470. Cambridge: Harvard Business School Press.
Petersen, M. A., and R. G. Rajan
1994 The Benefits of Lending Relationships: Evidence from SmaIl Business Data.' Journal of Finance
49: 3-37.
1995 'The Effect of Credit Market Competition on Lending Relationships.' Quarterly Journal of
Economics 110: 407-443.
Pettigrew, A.
1974 The Politics of Organiwtion Decision Making. London: Tavistock
Pettigrew, A. M., E. B. Ferlle, and L. McKee
1992 Shaping Strategic Change. London: Sage.
Pfeffer,J.
1977 'Toward an Examination of Stratification in Organizations.' Administrative Science Quarterly 22:
553-567.
1981 Power in Organiwtions. Cambridge: Ballinger.
1983 Organizational Demography.' In L. L. Cummings, and B. Staw (eds.), Research in Organiwtional
Behavior 5: 299-357. Greenwich: JAI Press.
1987 'A Resource Dependence Perspective on Intercorporate Relations.' In M. Mizruchi and M.
Schwartz (eds.), Intercorporate Relations: The Structural Analysis of Business, 25-55. New York:
1989 'A Political Perspective on Careers: Interests, Networks. and Environments.' In M. B. Arthur, D. T.
Hall, and B. S. Lawrence (eds.), Handbook of Career Theory, 380-396. New York: Cambridge
University Press.
1992 Managing with Power: Politics and Influence in Organiwtions. Boston, MA: Harvard Business
School Press.
Pfeffer, J., and Y. Cohen
1984 'Detenninants of Internal Labor Markets in Organizations.' Administrative Science Quarterly 29:
550-572.
Pfeffer, J., and P. Nowak
1976 'Joint Ventures and Interorganizational Interdependence.' Administrative Science Quarterly 21:
398-418.
Pfeffer, J., and G. R. Salancik
1978 The External Control ofOrganiwtions: A Resource Dependence Perspective. New York: Harper
and Row.
Pfeffer, J., and J. Ross
1982 The Effects of Marriage and a Working Wife on Occupational and Wage Attainment.' Administrative
Science Quarterly 27: 66-80.
Phillips-Jones, L.
1982 Mentors and Proteges. New York: Arbor House.
Piaget,J.
1932 The Moral Judgement of the Child. London: Kegan, Paul, Trench, and Teubner.
Pl1des, R. B.
1996 'The Destruction of Social Capital Through Law.' University of Pennsylvania lAw Review 144:
2055-2077.
Pinchot,G.
1985 Intrapreneuring: Why You Don't Have to Leave The Corporation to Become an Entrepreneur.
New York: Harper and Row.
Pinfield, L. T.
1995 The Operation of Internal Labor Markets: Staffing Practices and Vacancy Chains. New York:
Plenum Press.
Platteau, J. P.
1994a 'Behind the Market Stage Where Real Societies Exist I : The Role of Public and Private Order
Institutions.' Journal of Development Studies 30: 533-577.
1994b 'Behind the Market Stage Where Real Societies Exist 2: The Role of Moral Norms.' Journal of
Development Studies 30: 753-817.
Podolny, J. M.
1993 'A Status-Based Model of Market Competition.' American Journal of Sociology 98: 829- 872.
1994 'Market Uncertainty and the Social Character of Economic Exchange.' Administrative Science
Podolny, J. M., and J. N. Baron
1997 'Resources and Relationships: Social Networks and Mobility in the Workplace.' American
Sociological Review 62: 673-{i93.
Podolny, J. M., and A. Feldman
1997 'Is it Better to Have Status or Know What You Are Doing: Position and Capability in Venture
Capital Markets.' Graduate School of Business, Stanford University: Working Paper.
Podolny, J. M., and K. L. Page
1998 'Network Forms of Organization.' Annual Review of Sociology (forthcoming).
Podolny, J. M., and D. J. Phillips
1996 The Dynamics of Organizational Status.' Industrial and Corporate Change 5: 453-472.
Podolny, J. M., and T. E. Stuart
1995 'A Role-Based Ecology of Technological Change.' American Journal of Sociology 100: 1224-
1260.
Podolny, J. M., T. E. Stuart, and M. T. Hannan
1996 'Networks, Knowledge, and Niches: Competition in the Worldwide Semiconductor Industry,
1984-1991.' American Journal of Sociology 102: 659-{j89.
Podolny, J. M., and J. Baron
1997 'Resources and Relationships: Social Networks and Mobility in the Workplace.' American
Sociological Review 62: 673-{i93.
Polanyi,K.
1957 The Great Transformation. Boston: Beacon Press.
1958 Personal Knowledge, London, Routledge & Kegan Paul.
Pollitt, C.
1990 The New Managerialism and the Public Services: The Anglo American Experience . Oxford: Basil
Blackwell.
Porter, M. E .
1980 Competitive Strategy: Techniques for Analyzing Industries and Competitors. New York: Free
Press.
1985 Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press.
1987 'Managing Value-From Competitive Advantage to Corporate Strategy.' Harvard Business
Review 65: 43-58.
1991 'Toward a Dynamic Theory of Strategy.' Strategic Management Journal 12: 95-117.
Porter, M. E., and K. Fuller
1986 'Coalitions and Corporate Strategy.' In M. Porter (ed.), Competition in Global Industries. Boston:
Portes, A., and J. Sensenbrenner
1993 'Embeddedness and Immigration: Notes on the Social Determinants of Economic Action.'
American Journal of Sociology 98: 1320-1350.
Posner, E. A.
1996 The Regulation of Groups-The Influence of Legal and Nonlegal Sanctions on Collective Action.'
University of Chicago Law Review 63: 133-197.
Powell, W. W.
1982 Getting into Print. The Decision-Maldng Process in Scholarly Publishing. Chicago: University of
Chicago Press.
1987 'Hybrid Organizational Arrangements: New Form or Transitional Development?' California
References - 531
1990 'Neither Market nor Hierarchy: Network Fonns of Organization.' In L. L. Cummings and B. M.
Slaw (eds.), Research in Organiwtional Behavior 12: 295-336. Greenwich: JAI Press.
Powell, W. W., and J. Owen-Smith
1998 'Commercialism in Universities: life Sciences Research and its Linkage with Industry.' Journal of
Policy Analysis and Management 17.
Powell, W. W., and P. Brantley
1992 'Competitive Cooperation in Biotechnology: Learning Through Networks?' In N. Nohria and R. G.
Eccles (eds.) Networks and Organiwtions: Structure, Form, and Action, 366-394. Boston:
Powell, W. W., and L. Smith-Doerr
1994 'Networks and Economic life.' In N. Smelser and R. Swedberg (eds.), The Handbook of Economic
Powell, W. W., K. W. Koput, and L. Smith-Doerr
1996 'Interorganizational Collaboration and the Locus of Innovation: Networks of Learning in
Biotechnology.' Administrative Science Quarterly 41: 116-45.
Prahalad, C. K., and G. Hamel
1990 "The Core Competences of the Corporation.' Harvard Business Review 68: 79-90.
Prelsendijrfer, P., and T. Voss
1988 'Arbeitsmarkt und Soziale Netzwerke: Die Bedeutung Sozialer Konta1cte beim Zugang zu Arbeitsplat-
zen.' Sowle Welt 39: 104-119.
Price, L. L., E. J. Arnould, and P. Tierney
1995 'Ooing to Extremes: Managing Service Encounters and Assessing Provider Performance.' Journal
of Marketing 59: 83-97.
Provan, K. G., and H. B. Milward
1995 'A Preliminary Theory of Interorganizational Network Effectiveness: A Comparative Study of
Four Community Mental Health Systems.' Administrative Science Quarterly 40: 1-33.
Pruitt, D. G., and M. J. Klnunel
1977 'Twenty Years of Experimental Gaming: Critique, Synthesis, and Suggestions for the Future.'
Annual Review of P:rychology 28: 363-392.
Putnam, R. D.
1993a Making Democracy Work: Civic Traditions in Modem Italy. Princeton: Princeton University
Press.
1993b 'The Prosperous Community: Social Capital and Public life.' American Prospect 13: 35-42.
1995a 'Bowling Alone: America's Declining Social Capital.' Journal of DefTWcracy 6: 65-78.
1995b 'Bowling Alone, Revisited.' The Responsive Community Spring: 18-33.
Puto, C., W. Patton, and R. King
1985 'Risk Handling Strategies in Industrial Vendor Selection Decisions.' Journal of Marketing 49: 89-
98.
Rablnow,P.
1996 Making PCR: A Story of Biotechnology. Chicago: University of Chicago Press.
Raftery, A. E.
1986 'Choosing Models for Cross-Classifications.' American Sociological Review 51 : 145-46.
1995 'Bayesian Model Selection in Social Research.' In P. V. Marsden (ed.), Sociological Methodology
1995,111-163. Washington: American Sociological Association.
Raglus, B. R., and D. B. McFarUn
1990 'Perceptions of Mentor Roles in Cross-Gender Mentoring Relationships.' Journal of Vocational
Behavior 37: 321-339.
Raglus, B. R., and T. A. Scandura
1994 'Gender Differences in Expected Outcomes of Mentoring Relationships.' Academy of Management
Journal 37: 957-971.
Ral, A., S. Borah, and A. Ramaprasad
1996 'Critical Success Factors for Strategic Alliances in the Information Technology Industry: An
Empirical Study.' Decision Sciences 27: 141-155.
Ranson, S., and J. Stewart
1994 Managementfor the Public Domain: Enabling the Learning Society. London: Macmillan.
Rapoport, A., and M. Guyer
1966 'A Taxonomy of 2x2 Games.' In General Systems: Yearbook of the Society for General Systems
Research 11: 203-214.
Raub, W., and J. Weesie

1990 'Reputation and Efficiency in Social Interactions: An Example of Network Effects.' American
Jou17IlJ1 of Sociology 96: 626-654.
Ravenscraft, D. J.
1987 'The 19805 Merger Wave: An Industrial Organization Perspective.' In L. E. Browne and E. S.
Rosengren (eds.), The Merger Boom, 17-37. Boston: Federal Reserve Bank.
Rees,A.
1966 'Information Networks in Labor Markets.' American Economic Review 56: 559-566.
Reichheld, F. F., and W. E. Sasser, Jr.
1990 'Zero Defections: Quality Comes to Services.' Harvard Business Review 68: 105-111.
Requena,F.
1991 'Social Resources and Occupational Status Attainment in Spain: a Cross-National Comparison with
the United States and the Netherlands.' International Jou17IlJ1 of Comparative Sociology 32: 233-242.
Reskin, B. F., and H. I. Hartmann
1986 Women's Work, Men's Work: Sex Segregation on the Job. Washington: National Academy Press.
Reve, T.
1990 'The Firm as a Nexus of Internal and External Contracts.' In M. Aoki, B. Gustafson, and O.
Williamson (eds.), The Firm as a Nexus of Treaties, 133-161. London: Sage.
Rhodes, R. A. W.
1997 From Marketisation to Diplomacy: It's the Mix That Matters. University of Newcastle:
Department of Politics.
RIchardson, G. B.
1972 'The Organisation of Industry.' Economic Jou17IlJ182
RIley, S., and D. Wrench
1985 'Mentoring Among Female Lawyers.' Jou17IlJ1 of Applied Social Psychology 15: 374--386.
RIng, P. S., and A. H. Van de Ven
1992 'Structuring Cooperative Relationships Between Organisations.' Strategic Manogement JouT7UJl
13: 483-498.
1994 'Developmental Processes of Cooperative Interorganizational Relationships.' Academy of
Manogement Jou17IlJ119: 90-118.
Roberts, K. H., and C. A. O'Reilly, m
1979 'Some Correlates of Communication Roles in Organizations.' Academy of Manogement Jou17IlJ1
22:42-57.
Roche,G.
1979 'Much Ado About Mentors.' Harvard Business Review 57: 14--28.
Rogers, E. M .
1971 Communication of Innovations. New York: Free Press.
Romo, F. P., and M. Schwartz
1995 The Structural Embeddedness of Business Decisions: The Migration of Manufacturing Plants in
New York State, 1960-1985.' American SOCiological Review 60: 847-907.
Roncek, D. W.
1992 'Learning More from Tobit Coefficients: Extending a Comparative Analysis of Political Protest.'
Rosenbaum, J.
1979a 'Organizational Career Mobility: Promotion Chances in a Corporation during Periods of Growth
and Contraction.' American Jou17IlJ1 of Sociology 85: 21-48.
1979b 'Tournament Mobility: Career Patterns in a Corporation.' Administrative Science Quarterly 24:
22~241.
Rosenburg, L. J., and L. W. Stem
1971 'Conflict Measurement in the Distribution Channel.' Jou17IlJ1 of Marketing Research 8: 437-442.
Roussel, P. A., K. A. Saad, and T. J. Erickson
1991 Third Generation R&D. Boston: Harvard Business School Press.
Rowley, S., and Rowley, R.
1916 and 1960 Rowley on Partnership. New York: Bobbs-Merrill.
Russell, J. A.
1979 'Affective Space is Bipolar.' Journal of Personality and Social Psychology 37: 345-356.
References - 533
Saaty, T.L.
1990 The Analytic Hierarchy Process. Planning. Priority Setting. Resource Allocation (2nd. edition).
Pittsburgh: RWS Publications.
1990 Multicriteria Decision Making. Pittsburgh: RWS Publications.
Sabel,C.F.
1995 'Design, Deliberation and Democracy: On the New Pragmatism of Firms and Public Institutions.'
Paper presented at the liberal Institutions. Economic Constitutional Rights and the Role of
Organiwtions. European University Institute, Florence.
Sabel, C. F., G. Herrigel, R. Kazis, and R. Deeg
1987 'How to Keep Mature Industries Innovative.' Technology Review 90: 26-35.
Sabel, C. F., and J. Zeitlin
1997 'Stories. Strategies, Structures: Rethinking Historical Alternatives to Mass Production.' In C. F.
Sabel and J. Zeitlin (eds.), World of Possibilities. Flexibility and Mass Production in Western
Industrialisation, 1-33. Cambridge: Cambridge University Press.
Sailer, L. D.
1978 'Structural Equivalence: Meaning and Definition, Computation and Application.' Social Networks
1: 73-90.
Sako,M.
1991 Price. Quality and Trust: lnteljinn Relations in Britain and Japan. Cambridge: Cambridge
University Press.
Salancik, G. R.
1986 'An Index of Subgroup Influence in Dependency Networks.' Administrative Science Quarterly 31 :
194-211.
1995 'Wanted: A Good Network Theory of Organization.' Administrative Science Quarterly 40: 345-
349.
Samuel, Y.
1998 'Sociotechnical Realigning to Environmental Change: Defense Industries in Israel.' Unpublished
manuscript, University of Haifa, Israel.
Sandefur, R. L., and E. O. Laumann
1997 'The Changing Value and Character of Social Capital in an Expanding Social System: A
Preliminary Test.' American Bar Foundation Working Paper #9606. Chicago: American Bar
Foundation.
1998a 'Organization and Inequality: Changing Patterns of Income Inequality in the Chicago Bar, 1975
and 1995.' Paper prepared for presentation at the Annual Meetings of the American Sociological
Association. San Francisco, August.
1998b 'A Paradigm for Social Capital.' Rationality and Society (forthcoming).
Sasser, W. E., R. P. Olsen, and D. D. Wyckoff
1978 Management of Service Operations: Text, Cases and Readings. Boston: Allyn and Bacon.
Saxenian, A. L.
1990 'Regional Networks and the Resurgence of Silicon Valley.' California Management Review 33:
89-112.
1991 'Response to Richard Florida and Martin Kenney 'Silicon Valley and Route 128 Won't Save Us."
California Management Review 33: 136-142.
1994 Regional Advantage-Culture and Competition in Silicon Valley and Route 128. Cambridge:
Harvard University Press.
Scandura, T. A.
1992 'Mentorship and Career Mobility: An Empirical Investigation.' Journal of Organiwtional Behavior
13: 169-174.
Scandura, T. A., and B. R. Ragins
1993 'The Effects of Sex and Gender Role Orientation on Mentorship in Male-Dominated Occupations.'
Journal of Vocational Behavior 43: 251-265.
Scandura, T. A., and C. A. Schriesheim
1994 'Leader-Member Exchange and Supervisor Career Mentoring as Complementary Constructs in
Leadership Research.' Academy of Management Journal 37: 1588-1602.
Scharpf, F. W.
1993 'Coordination in Hierarchies and Networks.' In F. W. Scharpf (ed.). Games in Hierarchies and
Networks: Analytical and Empirical Approaches to the Study of Governance Institutions, 125-165.
Frankfurt: Campus Verlag !Westview Press.
Schecter, M. D.
1995 'How to Read a Patent.' In M. A. Lechter, E.C. Clifford, and R.B . Famiglio (eds.), Successful
Patenting for Scientists and Engineers. New Yorlc ffiEE Press.
Schein, E.
1968 'Organizational Socialization and the Profession of Management.' In B. Staw (ed.), Psychological
Foundations ofOrganiwtional Behavior, 210--233. Santa Monica: Goodyear Publishing.
Scher,M.J.
1996 'The Relational Access Paradigm and Japanese Interfinn Netwooo; Why Outsiders Seldom Wm:
Outcomes and Policy Implications.' Paper presented at the EMOT workshop (European Management
and Organization in Transition, of the European Science Foundation) Turin, November.
Schere,J.
1982 'Tolerance of Ambiguity as a Discriminating Variable Between Entrepreneurs and Managers.'
Proceedings 404--408. Academy of Management, New Yon.
Schermerhorn, J. R.
1977 'Information Sharing as an Interorganizational Activity.' Academy of Management Journal 20:
148-153.
Schleifer, A., and R. W. Vlshny
1990 The Takeover Wave of the 1980s.' Science 17: 745--749.
Schlesinger, L. A., and J. L. Heskett
1991 'Breaking the Cycle of Failure in Services.' Sloan Management Review Spring: 17-28.
Schmidt, V. A.
1996 From State to Market?: The Transformation of French Business and Government. New Yon:
Schmookler, J.
1966 Invention and Economic Growth. Cambridge: Harvard University Press.
Schonberger, R. J.
1986 World Class Manufacturing: The Lessons of Simplicity Applied. New Yon: Free Press.
Schumpeter, J. A.
1934 The Theory of Economic Development. Cambridge: Harvard University Press.
Scott, W. R.
1992 Organiwtions: Rational, Natural. and Open Systems. Englewood Cliffs: Prentice Hall.
1995 Institutions and Organiwtions. Thousand Oaks: Sage.
Scott, W. R., and J. W. Meyer
1991 The Organization of Sectors: Propositions and Early Evidence.' In W. W. Powell and P. J.
DiMaggio (eds.), The New Institutionalism in Organiwtional Analysis, 108-140. Chicago:
University of Chicago Press.
Seabright, M. A., D. A. Levinthal, and M. Flchman
1992 'Role of Individual Attachments in the Dissolution of Interorganizational Relationships.' Academy
of Management Journal 35: 122-160.
Seron,C.
1996 The Business of Practicing Law: The Work lives of Solo and Small Firm Attorneys. Philadelphia:
Temple University Press.
Sexton, D. L., and N. Bowman
1985 The Entrepreneur: A Capable Executive and More.' Journal of Business Venturing 1: 129-140.
Shah,P.
1996 'Network Destruction: The Structural Implications of Downsizing.' Paper presented at Academy of
Management Meetings, Cincinnati.
Shapiro, D. L., B. H. Sheppard, and L. Cheraskln
1992 'In Theory: Business on a Handshake.' Negotiation Journal 8: 365--377.
Shapiro,E.
1997 'Judge Rules Against Viacom in Dispute with Seagram Over Cable-TV Conflict.' Wall Street
Journal May 16: B6.
Shaw,M.E.
1964 'Communication Netwooo'. In L. Benowitz (ed.), Advances in experimental social psychology 1:
111-147. New York: Academic Press.
Sheppard, B. H., and M. Tuchlnsky
1996 'Micro-OB and the Netwon Organization.' In R. M. Kramer and T. R. Tyler (eds.), Trust in
Organiwtions. Thousand Oaks: Sage.
References - 535
Sherer, P. D.
1995 'Leveraging Human Assets in Law Finns: Human Capital Structures and Organizational
Capabilities.' Industrial and Labor Relations Review 48: 671-691.
Sherman, J. D., H. L. Smith, and E. R. Mansfield
1986 'The Impact of Emergent Network Structure on Organizational Socialization.' Journal of Applied
Behavioral Science 22: 53-63.
Sheth,J.N.
1973 'A Model of Industrial Buyer Behavior.' Journal of Marketing 37: 50-56.
Shl, Y., and M. J. Gregory
1994 'International Manufacturing Strategy: Structuring World·Wide Manufacturing Network.' Working
Papers in Manufacturing, No II, Cambridge University.
Shrum, W., and R. Wuthnow
1988 'Reputational Status of Organizations in Technical Systems.' American Journal of Sociology 93:
882-912.
Slmmel,G.
1950 The Sociology of Georg Simmel. Translated by Kurt H. Wolff. Glencoe: Free Press.
1955 Conflict and the Web of Group-Affiliations. Translated by K. H. Wolff and R. Bendix. New York:
Free Press.
1959 Sociology of Religion. Translated by C. Rosenthal. New York: Philosophical Library.
Simmons, D. B.
1985 'The Nature of the Organizational Grapevine.' Supervisory Management Nov.: 39-42.
Simon, H. A.
1957 Models of Man: Social and Rational: Mathematical Essays on Rational Human Behavior in Social
Settings. New York: Wiley.
1991 'Organizations and Markets.' Journal of Economic Perspectives 5: 25-44.
Skinner, W.
1969 'Manufacturing-Missing Link in Corporate Strategy.' Harvard Business Review 47: 136-145.
Skowronski, J. J., and D. E. Carlston
1989 'Negativity and Extremity Biases in Impression Formation: A Review of Explanations.'
Psychological Bulletin 105: 131-142.
Slack, N.
1991 The Manufacturing Advantage. London: Mercury Business Books.
Slaughter, S., and L. L. Leslie
1997 Academic Capitalism: Politics. Policies. and the Entrepreneurial University. Baltimore: Johns
Hopkins University Press.
Smart,A.
1993 'Gifts, Bribes, and Guanxi-A Reconsideration of Bourdieu's Social Capital.' Cultural
Anthropology 8: 388-408.
Smelser, N., and R. Swedberg (eds.)
1994 Handbook of Economic Sociology. Princeton: Princeton University Press.
Smigel, E. O.
1969 The Wall Street Lawyer. Professional Organization Man? Bloomington: Indiana University Press.
Smit,B.
1996 Heineken. un Leven in de Brouwerij [Heineken, A life in the Brewery, in Dutch]. Nijmegen: Sun.
Smith,H.C.
1996 Sensitivity to people. New York: McGraw-Hill.
Smith, K. G., S. J. Carroll, and S. J. Ashford
1995 'lntra- and Interorganizational Cooperation: Toward a Research Agenda.' Academy of Management
Journal 38: 7-23.
Smith-Doerr, L.
1997 'Framing Scientific Careers: Biotechnology as Legitimate Science.' Paper presented at the annual
meetings of the Society for Social Studies of Science, Tucson, AZ.
Smitka, M. J.
1991 Competitive Ties: Subcontracting in the Japanese Automotive Industry. New York: Columbia
University Press.
Snijders, T. A. B.
1996 'Stochastic Actor-Oriented Dynamic Analysis.' Journal of Mathematical SOCiology 21: 149-172.
Snow, C. C., and S. A. SneU

1993 'Staffing as Strategy.' In N. Schmitt and W. C. Bonnan (eds.), Personnel Selection in
Organizations, 448-478. San Francisco: Jossey-Bass.
Snow, C. C., R. E. Miles, and H. J. Coleman, Jr.
1992 'Managing 21st Century Network Organizations.' Organizational Dynamics 20: 5-20.
Sobel, M. E., M. Hout, and O. D. Duncan
1985 'Exchange, Structure, and Symmetry in Occupational Mobility.' American Journal of Sociology
91: 359-372.
Solomon, M. R., C. F. Surprenant, J. A. Czepiel, and E. G. Gutman
1985 'A Role Theory Perspective on Dyadic Interactions: The Service Encounter.' Journal of Marketing
49: 99-111.
Song,J. H.
1982 'Diversification Strategies and the Experience of Top Executives of Large FI!IIlS.' Strategic
Sorenson, A.
1996 'The Structural Basis of Social Inequality.' American Journal of Sociology 101: 1333-65.
Spaeth, J. L., and D. P. O'Rourke
1994 'Design and Implementation of a National Sample of Work Establishments.' American Behavioral
Scientist 37: 872-890.
Spector,M.
1973 'Secrecy in Job Seeking Among Government Attorneys: Two Contingencies in the 1beory of
Subcultures.' Urban Life and Culture 2: 211-229.
Starr, J. A., and I. C. MacMiUan
1990 'Resource Cooptation Via Social Contracting: Resource Acquisition Strategies for New Ventures.'
Strategic Management Journal 11 : 79-92.
Steers, R. M., Y. K. Shin, and G. Ungson
1989 The Chaebol: Korea's New Industrial Might. New York: Harper and Row.
Stevens, G, C,
1989 'Integrating the Supply Chain.' International Journal of Physical Distribution and Materials
Management 19: 3-8.
Stewart, J. B.
1991 Den of Thieves. New York: Simon & Schuster.
Stewman,S.
1986 'Demographic Models of Internal Labor Markets.' Administrative Science Quarterly 31: 212-247.
Stewman, S., and S. L. Konda
1983 'Careers and Organizational Labor Markets: Demographic Models of Organizational Behavior.'
American Journal of Sociology 88: 637~85.
Stigler, G. J.
1961 'The Economics of Information.' Journal of Political Economy 69: 213-225.
1962 'Information in the Labour Market.' Journal of Political Economy 70: 94-105.
Stinchcombe, A, L.
1965 'Social Structure and Organizations.' In James G. March (ed.), Handbook of Organizations, 142-193.
Chicago: Rand McNally.
1985 'Contracts as Hierarchical Documents.' In A. Stinchcombe and C. Heimer, Organization Theory
and Project Management, 121-171. Oslo: Norwegian University Press.
1986 'Nonns of Exchange.' In A. L. Stinchcombe, Stratification and Organization, Selected Papers,
231-267. New York: Cambridge University Press.
1990 Information and Organizations. Berkeley: University of California Press.
Stock, J. R., and D. M. Lambert
1987 Strategic Logistics Management (2nd. edition). Homewood: Irwin.
Stoelhorst, J. W.
1997 'In Search of a Dynamic Theory of the FInD.' Doctoral dissertation, Twente University, The
Netherlands.
Stokman, F. N., J . van der Knoop, and F. W. Wasseur
1988 'Interlocks in the Netherlands: Stability and Careers in the Period 1960-1980.' Social Networks 10:
183-208.
Stokman, F. N., R. Ziegler, and J. Scott
1985 Networks of Corporate Power. Cambridge: Polity Press.
References - 537
Stolzenberg, R. M.
1980 'Measurement and Decomposition of Causal Effects in Nonlinear and Nonadditive Models.'
Sociological Methodology 11: 459-488.
Strauss, D., and M. Ikeda
1990 'Pseudolikelihood Estimation for Social Networks.' Journal of the American Statistical Association
85: 204-212.
Streed" W.
1997 'Beneficial Constraints: On The Economic limits of Rational Voluntarism.' In 1. R. Hollingsworth
and R. Boyer (eds.), Contemporary Capitalism. The Embeddedness of Institutions, 197-219.
Cambridge: Cambridge University Press.
Stuart, T. E.
1998a 'Interorganizational Alliances and the Perfonnance of Firms: A Study of Growth and Innovation
Rates in a High-Technology Industry.' Working paper, University of Chicago.
1998b 'Producer Network Positions and Propensities to Collaborate: An Investigation of Strategic
Alliance Formation in a High-Technology Industry.' Administrative Science Quarterly
(forthcoming).
Stuart, T. E., H. Hoang, and R. Hybels
1999 'Interorganizational Endorsements and the Performance of Entrepreneurial Ventures.'
Administrative Science Quarterly (forthcoming).
Suchman, M. C.
1994 'On Advice of Counsel: Law Firms and Venture Capital Funds as Information Intermediaries in the
Structuration of Silicon Valley.' Unpublished doctoral dissertation, Department of Sociology,
Stanford University.
Sundqvlst, S. I.
1990 Agarna och Markten I Sveriges Borsbolag 1986. [Owners and Power in Sweden 1986, in
Swedish] . Stockholm: Dagens Nyheters ForJag.
Surprenant, C. F., and M. R. Solomon
1987 'Predictability and Personalization in the Service Encounter.' Journal of Marketing 51 : 86--96.
Suzuki, S., and R. W. Wright
1985 'Financial Structure and Bankruptcy Risk in Japanese Companies.' Journal of 1ntemational
Business Studies 16: 97-110.
Swan, J. E., and F. Trawick
1980 'Satisfaction Related to Predictive vs. Desired Expectations.' Association of Consumer Research
Swartz, T. A., and S. W. Brown
1989 'Consumer and Provider Expectations and Experiences in Evaluating Professional Service QUality.'
Journal of the Acackmy of Marketing Science 17: 189-195.
Swedberg, R.
1994 'Markets as Social Structures.' In N. J. Smelser and R. Swedberg (eds.), The Handbook of
Economic Sociology, 255-282. Princeton: Princeton University Press.
Tagiuri,R.
1958 'Social Preference and its Perception.' In R. Tagiuri and L. Petrullo (eds.), Person Perception and
Interpersonal Behavior. Stanford: Stanford University Press.
Talmud,I.
1992 'Industry Market Power, Industry Political Power, and State Support: The Case of Israeli Industry.'
In G. Moore and J.A. Whitt (eds.), Research in Politics and Society 4, 35-62. Greenwich: JAI
Press.
1994 'Relations and Profits: Imperfect Competition and Its Outcome.' Social Science Research 23: 109-
135.
Talmud, I., and D. N. Izraell
1998 'The Relationship Between Gender and Performance Issues of Concern To Directors: Correlates
Or Institution?' Journal of Organizational Behavior 19 (forthcoming).
Talmud, I., and G. S. Mesch
1997 'Market Embeddedness and Corporate Instability: The Ecology of Inter-Industrial Networks' Social
Science Research 26: 41~1.
Talmud, I., and Y. Yanovitzky
1998 'The Contradictory Demand Paradox: Social Embeddednesss and the Political Economy of the
Firm' Israeli Sociology 1: 1-43 (forthcoming, in Hebrew).
Tam,S.
1990 'Centrifugal and Centripetal Growth Processes: Contrasting Ideal Types for Conceptualizing the
Developmental Patterns of Chinese and Japanese Finns.' In S. R. Clegg and S.G. Redding (eds.),
Capitalism in Contrasting Cultures, 153-183. New York: De Gruyter.
Tarrow,S.
1996'Making Social-Science Work Across Space and Time-A Critical Reflection on Putnam, Robert
'Making Democracy Work'.' American Political Science Review 90: 389-397.
Taylor, J. B.
1994 'Discovery of New Medicines.' In J. P. Griffin, 1. O'Grady, and F.O. Wells (eds.), The Textbook of
Pharmaceutical Medicine (2nd. edition). Antrim: Greystone Books.
Taylor, S. E.
1991 'Asymmetrical Effects of Positive and Negative Events: The Mobilization-Minimization
Hypothesis.' Psychological Bulletin 110: 67-85.
Taylor, M., and S. Singleton
1993 'The Communal Resource-Transaction Costs and the Solution of Collective Action Problems.'
Politics and Society 21 : 195-214.
Teece,D.J.
1986 'Profiting from Technological Innovation: Implications for Integration, Collaboration, licensing and
Public Policy.' Research Policy 15: 285-305.
1988 'Technological Change and the Nature of the FJCID.' In G. Dosi, C. Freeman, R. Nelson, G. Silverberg,
and L Soete (eds.), Technical Change and Economic Theory, 256-281. London: Pinter.
Tevlin,J.
1997 'Non-Compete Clauses are a Growing Trend.' Minneapolis Star Tribune May 23: 01.
Thackray, J.
1986 'America's Vertical Cutback.' Management Today, June.
Thaler R., and H. Sherrin
1981 'An Economic Theory of Self-Control.' Journal of Political Economy 89: 392-406.
Thibaut, J. W., and H. H. Kelley
1959 The Social Psychology of Groups. New York: Wiley.
Thomas, D. A.
1993 'Racial Dynamics in Cross-Race Developmental Relationships.' Administrative Science Quarterly
38: 169-194.
Thomas, D. A., and M. C. Higgins
1996 'Mentoring and the Boundaryless Career: Lessons from the Minority Experience.' In M. Arthur and
D. Rousseau (eds.), Boundaryless Career: A New Employment Principle for a New Organimtional
Era. New York: Oxford University Press.
Thompson, G.
1982 'The Finn as a Dispersed Social Agency.' Economy and Society 11: 233-250.
Thompson, G., J. Frances, R. Levacic, and J. Mitchell
1991 Markets, Hierarchies and Networks. London: Sage.
Thompson, J. D.
1967 Organimtions in Action: Social Science Bases of Administration. New York: McGraw-Hill.
Thorelli, H. B.
1990 'Networks: Between Markets and Hierarchies.' In D. Ford (ed.), Understanding Business Markets.
London: Academic Press.
Timmons, J. A.
1995 New Venture Creation: Entrepreneurship for the 21" Century. Burr Ridge: Irwin.
Tolbert, P. S., and L. G. Zucker
1983 'Institutional Sources of Change in the Formal Structure of Organizations: The Diffusion of Civil
Service Refonn, 1880-1935.' Administrative Science Quarterly 28: 22-39.
Towill,D.
1991 'Supply Chain Dynamics.' Computer Integrated Manufacturing 4: 197-208.
Trajtenberg, M.
1990 'A Penny for Your Quotes: Patent Citations and the Value of Information.' Bell Journal of
Economics 21 : 172-187.
Travers, J., and S. Milgram
1969 'An Experimental Study of the 'Small World' Problem.' Sociometry 32: 425-443.
References - 539
Triandls, H. C.
1995 Individualism and Collectivism. Boulder: Westview.
Trice, H. M., and D. A. Morand
1989 'Rites of Passage in Work Careers.' In M. B. Arthur, D. T. Hall, and B. S. Lawrence (eds.),
Handbook of Career Theory, 397-416. New York: Cambridge University Press.
Tsui, A. S., and J. L. Fahr
1997 'Where Guanxi Matter. Relational Demography and Guanxi in the Chinese Context.' Work and
Occupation 24: 56-79.
Tsui, A. E., and C. A. O'ReiUy, III
1989 'Beyond Simple Demographic Effects: The Importance of Relational Demography in Superior-
Subordinate Dyads.' Academy of Management Journal 32: 402-423.
Tsui, A. E., T. D. Egan, and C. A. O'Reilly, m
1992 'Being Different: Relational Demography and Organizational Attachment.' Administrative Science
Quanerly 37: 549-579.
Tuma, N. B., and M. T. Hannan
1984 Social Dynamics: Models and Methods. New Yorlc: Academic Press.
Turban, D. B., and T. W. Dougherty
1994 'Role of Protege Personality in Receipt of Mentoring and Career Success.' Academy of
Turnbull, P., and J. P. Valla (eds.)
1986 Strategies for International Industrial Marketing. London: Croom-Helm.
Tushman, M. L.
1977 'Special Boundary Roles in the Innovation Process.' Administrative Science Quanerly 22: 587-
605.
1978 'Task Characteristics and Technical Communication in R&D.' Academy of Management Journal
21: 624-645.
Tushman, M. L., and P. Anderson
1986 'Technological Discontinuities and Organizational Environments.' Administrative Science
Quanerly 31: 439-465.
Upton, D. M., and A. McAfee
1996 'The Real Virtual Factory.' Harvard Business Review 74: 123-133.
Useem,M.
1979 'The Social Organization of the American Business Elite and Participation of Corporate Directors
in the Governance of American Institutions.' American Sociological Review 44: 553-572.
1982 'Classwide Rationality in the Politics of Managers and Directors of Large Corporations in the
United States and Great Britain.' Administrative Science Quanerly 27: 199-226.
1996 Investor Capitalism: How Money Managers Are Changing the Face of Corporate America. New
York: Basic Books.
Useem, M., and J. Karabel
1986 'Pathways to Top Corporate Management.' American Sociological Review 51: 184-200.
Uzzl,B.
1996a 'The Sources and Consequences of Embeddedness for the Economic Performance of
Organizations: The Networlc Effect.' American Sociological Review 61: 674-698.
1996b 'Coase Encounters of a Sociological Kind: A Sociological Perspective on Strategy.' In J. Dutton
and J. A. C. Baum (eds.), The Embeddedness of Strategy. New York: JAI Press.
1997a 'Social Structure and Competition in Interfirm Networks: The Paradox of Embeddedness.'
Administrative Science Quarterly 42: 3~7.
1997b 'A Structural Embeddedness Approach to Deindustriatization and Organizational Decline.' The
International Journal of Sociology and Social Policy 17: 111-155.
Uzzi, B., and Z. Barsness
1998 'Contingent Employment in British Establishments: Organizational Determinants of the Use of
Fixed-Term Hires and Part-Time Workers.' Social Forces 76: 967-1005.
Uzzi, B., and J. Gillespie
1998 'Embeddedness and Market Dynamics: The Case of Banks, Corporate Borrowers and the Cost and
Availability of Financial Capital.' Working paper. Northwestern University.
Van de Mark, D.
1996 'Ovitz Out at Disney.' Cable News Network December 12.
Van de Ven, A. H.
1993 'The Development of an Infrastructure for Entrepreneurship.' Journal of Business Venturing 8:
211-230.
Van de Ven, A. H., D. C. Emmit, and R. Koenig
1975 'Frameworks for Inter-Organisational Analysis.' In A. R. Negandhi (ed.), Inter-Organisational
Theory. Kent State: Kent State University Press.
Van Engelen, J. M. L., P. Terlouw, D. J. Kiewiet, and S. W. F. Ornta
1999 'Towards a Polarity Strategy in Managing Product Development Teams'. Unpublished Manuscript,
School of Management and Organization, University of Groningen.
Van Lange, P. A. M., W. B. G. Liebrand, D. M. Messick, and H. A. M. Wilke
1992 'Social Dilemmas: The State of the Art.' In W. Liebrand, D. M. Messick, and H. Wilke (eds.),
Social Dilemmas: Theoretical Issues and Research Findings. Oxford: Pergamon Press.
Van Maanen, J.
1975 'Police Socialization: A Longitudinal Examination of Job Attitudes in an Urban Police
Department.' Administrative Science Quarterly 20: 207-228.
1983 'Golden Passports: Managerial Socialization and Graduate Education.' The Review of Higher
Education 6: 435-455.
Van Maanen, J., and E. Scbein
1979 'Toward a Theory of Organizational Socialization.' In B. M. Staw (ed.), Research on
Organizational Behavior 1: 209-264. Greenwich: JAI Press.
Viilker, B., and H. Flap
1998 'Getting Ahead in the GDR. Social Capital in the Status-Attainment under Communism.' Acta
Sociologica 41 (forthcoming).
Voss, C. A., C. G. Armistead, R. Johnston, and B. Morris
1985 Operations Management in Service Industries and the Public Sector. Chichester: Wiley.
Wagner, G. W., J. Pfeffer, and C. A. O'ReUly, m
1984 'Organizational Demography and Turnover in Top-Management Groups.' Administrative Science
Waldinger, R.
1995 'The Other Side of Embeddedness-A Case-Study of the Interplay of Economy and Ethnicity.'
Ethnic and Racial Studies 18: 555-580.
Walker, G., B. Kogut, and W. Shan
1997 'Social Capital, Structural Holes and the Formation of an Industry Network.' Organization Science
8: 109-125.
Walker, M. E., and S. Wasserman
1987 TRIADS. A Computer Program for Triodic Analyses. Urbana: University of Dlinois.
Walsh, J. S. Kuker, N. Maloney, and S. M. Gabbay
1999 'Connecting Minds: CMC and Scientific Work.' Journal of the American Society for Information
Science. Forthcoming.
Walsh,J.P.
1988 'Selectivity and Selective Perception: An Investigation of Managers' Belief Structures and
Information Processing.' Academy of Management Journal 31 : 873-896.
Walsh, J. P., and J. K. Seward
1990 'On the Efficiency of Internal and External Corporate Control Mechanisms.' Academy of
Walton,M.
1997 Car: A Drama of the American Workplace. New York: Norton.
Wasserman, S., and D. Iacobucci
1988 'Sequential Social Network Data.' Psychometrika 53: 261-282.
Wasserman, S., and K. Faust
1994 Social Network Analysis: Methods and Applications. New York: Cambridge University Press.
Waters,M.
1989 'Collegiality, Bureaucratization, and Professionalization: A Weberian Analysis.' American Journal
of Sociology 94: 945-72.
Waterson, M.
1990 'The Economics of Product Patents.' American Economic Review 80: 860-869.
Watson, D., and A. Tellegen
1985 'Toward a Consensual Structure of Mood.' Psychological Bulletin 98: 219-235.
References - 541
Webb,E.
1996 'Trust and Crisis.' In R. Kramer and T. Tylor (eds.), Trust in Organizotions: Frontiers of Theory
and Research, 288-301. London: Sage.
Weber,M.
1978 Economy and Society: An Outline of Interpretive Sociology. Translated by E. Fischoff et al., edited
by G. Rothard and C. Wittich. Berkeley: University of California Press. (first edition published in
1922).
Wegener,B.
1991 'Job Mobility and Social Ties: Social Resources, Prior Job, and Status Attainment.' American
Weick, K. E,
1979 The Social Psychology of Organizing. Reading: Addison Wesley.
1995 Sensemaking in Organizotions. Thousand Oaks: Sage.
Weldenbaum, M., and S. Hugbes
1996 The Bamboo Network. New York: The Free Press.
Wells,M.
1995 'Hollywood Super Agent Michael Ovitz Named Disney President.' Detroit News August 15.
Wernerfelt, B.
1984 'A Resource Based View of the Finn.' Strategic Management Journal 5: 171-180.
Weston, J., K. Cbung and S. Hoag
1990 Mergers, Restructuring. and Corporate Control. Englewood Cliffs: Prentice-Hall.
Wetzel, W. E.
1987 'The Infonnal Venture Capital Market: Aspects of Scale and Market Efficiency.' Journal of
Business Venturing 2: 299-313.
White,H.C.
1961 'Management Conflict and Sociometric Structure.' American Journal of Sociology 67: 185-199.
1981 'Where Do Markets Come From?' American Journal of Sociology 87: 517-547.
1988 'Varieties of Markets.' In B. Wellman and S. D. Berkowitz (eds.). Social Structure: A Network
Approach, 226-260. New York: Cambridge University Press.
1992 Identity and Control: A Structural Theory of Social Action. Princeton, NJ: Princeton University
Press.
White, H. C., S. A. Boorman, and R. L. Brelger
1976 'Social Structure from Multiple Networks I: Blockmodels of Roles and Positions.' American
Journal of Sociology 81 : 730-780.
Whitely, W., T . W. Dougherty, and G. F. Dreher
1991 'Relationship of Career Mentoring and Socioeconomic Origin to Managers' and Professionals'
Early Career Progress.' Academy of Management Journal 34: 331-351.
Wiersema, M. F., and K. A. Bantel
1992 'Top Management Team Demography and Corporate Strategic Change.' Academy of Management
Journal 35: 91-121.
Wllklnson, I, F.
1981 'Power, Conflict and Satisfaction in Distribution Channels-An Empirical Study.' International
Journal of Physical Distribution and Materials Management II: 20-30.
Williamson, O. E.
1975 Markets and Hierarchies: Analysis and Antitrust Implications. New York: Free Press.
1981 'The Economics of Organization: The Transaction Cost Approach.' American Journal of Sociology
87:548-577.
1985 The Economic Institutions of Capitalism: Firms, Markets. Relational Contracting. New York: Free
Press.
1993 'Ca1culativeness, Trust. and Economic Organization.' Journal of Law and Economics 36: 453-486.
1994 'Transaction Cost Economics.' In N. J. Smelser and R. Swedberg (eds.), The Handbook of Economic
1996 The Mechanisms of Governance. New Yode Oxford University Press.
Wills, T.A.
1991 'Social Support and Interpersonal Relationships.' In M. S. Clark (ed.), Prosocial Behavior. Review
of Personality and Social Psychology 12: 265-289. Newbury Park: Sage.
Windolf, P., and S. Wood

1988 Recruitment and Selection in the Labour Market. A Comparative Study of Britain and West Germany.
Avebury: Aldershot.
Winship, C., and R. Mare
1984 'Regression Models with Ordinal Variables.' American Sociological Review 49: 512-525.
Wistow,G.
1992 'The Health Service Policy Community.' In D. Marsh and R. A. W. Rhodes (eds.). Policy Networks
in British Government. Oxford: Clarendon Press.
WoUT,M.F.
1994 'Building Trust in Alliances.' Research Technology Management 37: 12-15.
Womack, J. P., D. T. Jones, and D. Roos
1991 The Machine That Changed the World: The Story of Lean Production. New York: Rawson
Associates.
Woodard, K. L.
1995 'Introduction and Acceptance of Inter-Organizational Agreements: The Experience of Seventy-
Five Administrators in One County.' Administration in Social Work 19: 51-83.
Woodard, K. L., and P. Doreian
1994 'Utilizing and Understanding Community Service Provision Networks: A Report of Three Case
Studies Having 583 Participants.' Journal of Social Service Research 18: 1--41.
Woolcock, M.
1997 'Social Capital and Economic Development: Toward a theoretical synthesis and policy framework.'
Theory and Society 27: 151-208.
Worldbank
1997 'Social capital: The Missing linkT Expanding the Measure of Wealth: Indicators of Environmentally
Sustainable Development. Environmentally Sustainable Development Studies and Monograph Series
17. Washington D.C.: The World Bank.
Wright, E. 0., and D. Cho
1992 ibe Relative Permeability of Class Boundaries to Cross-Class Friendships: a Comparative Study of
the United States, Canada, Sweden. and Norway.' American Sociological Review 57: 85-102.
Wrong,D.H.
1961 The Oversocialized Concept of Man in Modem Sociology.' American Sociological Review 26:
183-193.
Yoffie,D.B.
1996 'Competing in the Age of Digital Convergence.' California Management Review 38: 31-53.
Yoshino, M. Y., and T. B. Lirson
1986 The Invisible Link: Japan's Sogo Shosha and the Organization of Trade. Cambridge: MIT Press.
Yoshino, M. Y., and U. S. Rangan
1995 Strategic Alliances: An Entrepreneurial Approach to Globalization. Boston: Harvard Business
School Press.
Zagare, F. C.
1984 Game Theory: Concepts and Applications. Newbury Park: Sage.
Zaheer, A., and N. Venkatraman
1995 'Relational Governance as an Interorganizational Strategy: An Empirical Test of the Role of Trust
in Economic Exchange.' Strategic Management Journal 16: 373-392.
Zajac, E. J., and C. P. Olson
1993 'From Transaction Costs to Transactional Value Analysis: Implications for the Study of
Interorganizational Strategies'. Journal of Management Studies. 30: 131-145.
Zeggelink, E. P. H.
1994 'Dynamics of Structure: An Individual Oriented Approach.' Social Networks 16: 295-333.
Zenger, T. R., and B. S. Lawrence
1989 'Organizational Demography: The Differential Effects of Age and Tenure Distributions on
Technical Communications.' Academy of Management Journal 32: 353-376.
Zey,M.
1984 The Mentor Connection. Homewood: Dow Jones-Irwin.
Zucker, L. G.
1986 'Production of Trust: Institutional Sources of Economic Structure 1840-1920. 'In B. M. Staw & L.
L. Cummings (eds.). Research in Organizational Behavior. 8: 53-III. Greenwich: JAI Press.
References - 543
Zuckerman, E. W.
1999 'The Categorical hnperative: Securities Analysts and the Legitimacy Discount.' American Journal
of Sociology (forthcoming).
Zukin, S., and P. DiMaggio
1990 'Introduction.' In S. Zukin and P. DiMaggio (eds.), Structures of Capital. The Social Organization
of the Economy, 1-36. New York: Cambridge University Press.
Zuscovitch, E.
1994 'Sustainable Differentiation; Economic Dynamism and Social Nonns.' Paper presented at the J. A.
Schumpeter conference, Miinster, August.
Index
•
A allocation 461
absence of ties 92. 103 of resources 446
acceptance 163 altercentric uncertainty 431. 433. 436-439. 443
access 164.389.392.479 Analytic Hierarchy Process (AHP) method 370.
restrictions 377 372.373.375
to infonnation 299. 356. 453 anticipated cooperation 305
accessibility 80. 81 Apple 22. 67. 95. 423
accumulation processes 461 application 394
acquisition strategies 272 appropriable social structure 219
acquisitions 266-283. 462. 468. 474. 478 Araujo. Luis 59. 69. 71. 75. 76. 78. 300. 410. 505
action arm's-Iength ties 446. 452. 453. 458
dilemmas 77 Arrow's paradox 349
set 29 Asian societies 54
activity model 410 assessments 134. 136
actor asset 74. 358. 447. 462
centrality 26 of social capital 318
connectedness 25 specifity 30
dissimilarity 4. 328. 329 assignment of patents 391
prestige 377 associate group 47
similarity 327. 329. 334 associates 247. 250
•s location 444 attitude similarity 328
actors 80. 123. 167. 285. 300. 302. 303. 324. audit firms 59. 60
327.379. 400.409.410.449.451.458.461 authority 24
boundaries 22 structures 85
positions 18 system 243
adaptability 308. 318 automobile industry 100. 102
adhocracies 462 automorphic equivalence 27
administrative costs 185 avoidance of ties 95. 97
advantage viii
adversarial ties 120
adverse selection 448 B
advertisements 205. 207 background effects 281
advice 248. 263 Baker. Wayne E. 37. 80. 93. 95. 96. 97. 98. 99.
advising 24. 245 100. 104. 111. 119. 149. 180. 300. 338. 447.
age 402 449.450.452.453.454.458.493.496
aggregation 46 bank-firm relationship 450. 454. 459
alliance banking relationship 451
capital 376-389 base technologies 368
formation 30. 387. 432 baseline model 249
networks 29 basic advice network 263
partners 433 Bayesian infonnation criterion (BIC) 128
relationships 491 behavioral theories 424
alliances 24. 165 Benassi. Mario 4. 58. 80. 115. 117. 292. 294.
outcomes 36 306.325.488.498
benefit-rich networks 426 Chandlerian principles 424

benefits 219. 220 change 294
Benetton 7. 409. 423. 426 changes in the social system 221
between-finn collaborative 404 changing network
betweenness centrality 26. 41 patterns 32
bilateral governance 341 structures 42
biotechnology 390-408. 432.438.447 channel 431
Blau-tie 253-256. 258 behavior 424
Boeing 3. 66. 369 management 423
boundaries of organizations 43 Chicago lawyers 217. 223. 224. 226. 232
boundary 410. 411 China 73, 82,95
spanner 47. 351. 354 chump pattern 153. 155, 157
system's identity 52 clinical
transaction systems 51. 65 networks 291. 292. 296
Boxman. Ed 6. 19. 60. 87. 181. 200. 211. 212. trials 399
213.214.331.431.487.494.499.507 close ties 453
Brass. Daniel J. 4. 14.40.41.42. 162. 167. 199. closeness centrality 26. 330
232. 294. 324. 325. 327. 328. 329. 330. 333. closer ties 99
334.335.337.499.500.501.519.524 closure 94. 305, 355
Breiger. Ronald L. 2. 92. 104. 119. 120. 122. cluster actors 27
133. 136. 137. 147. 213. 265. 377. 399. 441. co-attendance 23
484,485.500.512,522, 524, 541 coaching 163
bright side of social capital/ties 299. 304 codes of conduct 69
buddy relationships 177 co-evolution of firms and networks 40 1
bull-whip effect 416 cognitive
business corporate capital 107-110, 113. 115, 116
friendship 450. 452 dimension of social capital 90
group 30. 52. 54, 59. 64 embeddedness 70
relationships 358. 490 social capital 106
-risk view of trust 33 cohesion 27
strategy 74 collaboration 358-361. 363. 364. 369. 371. 373.
buyer-seller interactions 148-158 375.390-408
to patents 393
c collaborative
activities 399. 402
calculativeness 83 agreement 371
capital viii, 73, 74 capital 390. 392, 394, 395,406. 414
capital-in-general metaphor 86 experience 400, 403
career 174, 176. 178-180.336 relationships 414
capital 39 social capital 406
-cycle hypothesis 208 ties 120
development 168. 334, 336 collaborator 390
failure 165 collaborators in development 345
help 163. 164 collective
history 169 action 182
progress 162, 163 dilemma 72. 84
cartel 29,49, 52 dilemma 70
Castellucci, Fabrizio 59.110.137.147.419.448. goods 80
465.468 social
causality 489 capital 240
central firm 405 structure 241
centrality 41. 42. 123. 125. 392. 397. 399. 400. collectivism 100
402-405,470.477,478 collegial organizations 238, 239. 241
CEO co-management relations 119
demographics 266-283 commercial chain 416
tenure 272. 278. 279 comrnitment 332, 374
chaeboI30.49,50,54-58,loo communication 326. 374
challenge 163 networks 34. 83, 334
Index -547
communities control lIS

of practice 344 structure 106-117
of trust 91 strategy 359
community structures 464 superiority 113
competence 343-345, 348, 350, 357, 365, 370, technological prestige 378
374,393,394 corrective action 374
trust 346 cost of financial capital 446-459
competencies of motivation and morality 348 costs 420
competition 155 of search 207
competitive counseling 163
advantage 27,182,379 coworker ties 245
corporate capital 107, 111, 113 coworkers 248, 255, 263
corporate structure liS conshr.rints 249
modalities 21 network 240, 251, 256, 263
patterns 157 teams 259
social capital 106 creation of social capital 76, 77, 83
complementarity 84 credibility 290
complex ties 399 cross-firm economy of learning 345
complications 362 cross-licensing 396
composition of the network 304 cross-stockholdings 64
concept of organization 45 esc 1,2,9, 10, 11, 12, 13,483,484,485,486,
conceptualization 43 489,491,493,509
confirmation 163 cultural
conflict 337,424 conditions 99
connectivity 98 context 99
consistency 59 embeddedness 70
conshr.rinment24O organizations 436
conshr.rint 248 customer-employee dyad 151
consultation network 318 customer service dyads 148-157
consumer loyalty 149 customers 148
consumer-to-consumer relations ISO
contingencies 352, 485
contractual arrangements 371 D
contribution to standing 141 dark side
control 219 of cooperation 357
controllability 80, 81 of social capital 298-322
cooperation 155, 369, 374, 424 data interactions 153
cooperative Davis-Blake, Alison 181,497
behavior 152 decentralized organization 242
modalities 21 decision uncertainty 271
patterns 157 dedicated biotechnology firm (DBF) 390-391,
relationships 356 394,398,399,404,405
coordination 326 definitions of social capital 118
failures 309, 311, 313, 315-317 demographic
core group 47, 48 composition 377
corporate variables 328
acquisitions 267-270 demographics 329
actors 89 dense networks 110
alliances 32 density 98
entities 20 depreciation 76
leverage III design 88-105
liability 106-117 destruction of social capital 77
social development of social capital 161-179
capital 1, 10-13, 106-134, 231, 238, differentiated markers vii
240, 291, 312, 320, 446-459, 483- diffused reciprocity 77
493 diffusion of the hr.rining 333
capital management 491, 492 dimension of social capital 103
capital theory 458, 484, 485, 491
direct education
approaches 183 level 276
social ties 181 sector 135, 138, 141, 142, 144, 145
dissatisfaction 415, 416 educational background 266-283
dissimilarity 328 effective ties 115
dissolution 62 effectiveness of social capital 320
distribution 399 egocentric
of exchange 457 network 22
of infonnation 5 uncertainty 431-445
of power 5 ego-network 446, 451453, 458, 459
of strategies 100 size 449, 454, 456
disunion strategy 88, 89, 93-95, 98-100, 102, structure 452
104,493 eigenvector centrality 123
diversity 406 elite ties 223, 228
of network activity 403 elites 217, 220-224, 227, 228, 230, 231
division of labour 70 embedded ties 45, 453, 458
domain-expanding acquisitions 269, 270 embeddedness 36, 46, 58, 59, 64, 65, 68-70, 80,
domination 40 86, Ill, 237, 238, 241, 245, 259-262, 377,
Doreian, Patrick 14, 104, 136, 138, 149, 441, 448, 459,465
485,500,505,521,542 approach 446459
downstream partners 357 embedding of relations 342
dual mode emerging technologies 367
ego-network structures 458 emotional support 199
networks 457 employability 38
duplex ties 255 employee-customer relations 149
durability 76 endowments of social capital 79
duration of relationships 454, 458 end-users 148
Dutch managers 201-203, 205, 207, 210-212 entrepreneurial
dyad 162,240,312,342,399,446 community 462, 465, 468
dyad-specific social capital 77 firms 436, 437, 440
dyadic organizations 303, 320
interaction model 154 research 90
interaction strategies 152 strategies 93, 97, 100
international corporate alliances 32 successes 476
modeling 149 ventures 468
parameter patterns 155 entrepreneurs45,220, 302,459,463,478
relationship 150, 410, 413, 415, 416, 450, entrepreneurship 80, 88, 89, 320, 460, 461
451 equity 274
supply relationships 418 -based contracts 35
ties 251, 253 partnerships 28, 29
dynamic task environments 318 pressures 186
dynamics of social capital accumulation 401 swap 29
equivalence 27
erosion of social capital 17
E establishment of collaborative agreement 360,
Easton, Geoff 59, 69, 71, 74, 75, 76, 78, 285, 363
300,410,419,496,504,505 Europe 71, 369
economic evaluability 78
exchange 68 evaluation of social capital 78, 79
performance 237-265 exchange 449
relations 21 of hostages 348
resources 68, 69, 74, 75, 410 partners 378, 451
sociology 458 relations 377, 432, 450
theory 347, 424 system 241, 242
value 396 exchanges of resources 260
of ties 217 exclusivity 99
economy of time 460-479 expatriates 174, 176
economy-society interaction 68 expectations 414416
Index -549
expert based organizations 284 forms of capital 73-75

exposure 163, 164 Forrester effect 416, 417
external fosters trust 327, 450
corporate ties 271 France 198, 362
economy of cognitive scope 345 Freeman, John 8, 24, 32, 66, 90, 93, 132, 293,
networks 49, 280 321 , 336, 448,452, 463,494,497,508, 512
pressures 186 frequency of an entrepreneurial strategy 97
relationships 358 friendship 163, 199,245,248,392
resources 374 networks 263, 327
extra-organizational social capital 232 ties 249, 253-256, 259, 335
function of network ties 432
functional background 266-283
F functions of metaphors 71
face-ta-face groups 76
failure 165,462, 468,474
family 84 G
Far East 369 Gabbay, Shaul M. 2, 4, 14, 87, 110, lIS, 117,
favoritism 183, 186 174, 215, 216, 219, 238, 264, 299, 300, 323,
female managers 199 325, 343, 408, 448, 449, 459, 486, 492, 493,
Ferlie, Ewan 14, 149, 288,289, 297,507,529 494, SOl, 509, 540
finance 280 game theory 148, ISO, 155
financial gaming strategies 157
aid 24 Gargiulo, Martin 3, 4, 58, 80, lIS, 117, 292, 294,
capital 44, 53, 436, 442, 446, 459 299, 300,309, 325, 492,493
capitallIlllrlcets 446 gender 162, 177
planning 451 gender-differentiated network 39, 40
finders 243 general management 270
firm age 400 Genoese traders 70
firm-environment interface 50 gentlemen's agreement 243
firm-level geographical associates 70
social capital 237, 256 Germany 198, 362
variables 400 gift 82
firm social capital 240 exchange 73
firm's boundaries 47 Gillespie, James 1. 8, 87,418,447,450,459,539
firm-ta-firm relations 150,488 global network structures 42
fixed-effects specification 40 1 globalization 7
Rap, Henk 6, 13, 19,60,87, 119, 181, 198,200, goal
202, 203, 211, 212, 213, 214, 215, 216, 238, definition 360, 362
239,299, 300, 318, 331 , 431,487,494, 499, specifity 3
500,501,504, 507,525,540 ga-between 341-355, 374, 465
flexibility 352 Gorman, Elizabeth H. 6, 201, 331
flow of information 458 governance 348
focal mechanism 341
firm 433 govemment355
producer 431 grapevine 6
focus 413, 416, 418 grinders 243
Foller 2-3 group
form of social capital 97 cohesiveness 305
formal dynamics 91
differentiation 98 group-level
methods 183-185, 187 social structure 237
organizations 85, 230 social ties 241, 249
peer review system 244 GSS 194,195
qualifications 232 guanxi 82
relationships 9, 392
rules 243
rules for lIIllrlcet order 70 H
selection procedures 181 halfway model 126, 128
Han, Shin-Kap 104, 1l0, 117, 119, 122, 133, immigrant communities 80
136, 137, 147, 203, 399, 433, 441, 484, 485, impatience 468, 477
493,512 IMP 409, 410, 424
Harland Christine M. 7, 410, 411, 415,417,418, implementation 360, 363, 374
420,422,424,488,513,516 income 205, 229, 231
Haunschild, Pamela R. 119,267, 268, 271, 274, indegree centrality 248
281,283,431,513 independent social capital 176
health indirect social ties 181
care 286, 287, 289, 296 individual
management 284 actions 182, 299
sector 135, 138, 141, 143-145 commitment 260
Heinz, John P. 14,220,222,223,224,231,232, economic performance 260,261
233,513 networks 6
heterogeneity 400 social capital 240, 248, 405
Heineken 14,490 social network 240
hierarchical social relationships 237
networks 177 241,251
relations 28, 85 individual-collective distinction 46
status 239 individual-level social capital 240
hierarchy 98, 102, 164,248,285,326 individual-level social structure 237
Higgins, Monica 6, 9, 19, 178, 179, 209, 306, individualism 100
319,336,489,514,538 individuals 91, 163, 167,391
high performers 254, 255 influence 40, 219
high prestige 377 informal
firms 380, 387 channels 183, 209
high quality relationships 458 contacts 198, 205, 208, 211
high status 432 interorganizational networks 288
actors 388 methods 184-186, 191, 193, 197
firms379,434,435,438,441~ procedures 185
organization 376 recrui tment 183
high-technology relationships 9, 176
firms 378 search 202, 212
industries 379 social contact 220
organizations 284 staffmg 183, 184, 187, 189
Hitachi 66, 369, 383 ties 180, 203, 207, 337
Holland 166,200 information
home runs 468 access 161
homosocial reproduction 328, 331 exchange 23, 416
homogeneous RC(l) model 126, 128 network 406
homophily 39, 327, 377, 387 transfer 82, 183, 219
honor 73 - and communication technology (ICT)
hostage keeping 348, 349, 353, 354 344
HRM implications 294 infrastructure of the network 418
human capital 44, 53, 72, 84,164,177,182,203, initial
205,210-212,218,225,300,324,327,334 mentoring 174
investments 20 public offering of stock (IPO) 462, 467-
theory 201 469,473,478
human resources 389 innovations 332
management323,324,332,334, 337 input-output models 136
hybrid interfaces 458 institution building 69
institutional
conditions 214
I context 88-105
Iacobucci, Dawn 14,25,148,150, 154,514,515, institutionalization of ties 9
540 institutionalized trust 52
mM 50, 66, 95, 369, 379, 380, 383, 388, 396 instrumental ends 83
ideal expectations 415 integrity 83
identity-making ties 114 intellectual
Index - 551
capital 390, 391, 394-3%, 403, 405 networks 5-7, 37,41,108,427

competencies 396 performance data 239
output 390, 399 relations 116
property views 396 ties 42, 166
rights 395 inventory control 99
Intel 38, 67, 96 investment 399
intentional banks 119, 120, 128, 131 , 132
sense 348 lOR network (ION) 135, 137, 142
trust 346, 347, 350 Israel110-115
interagency alliances 293 Italy 77, 100, 306, 362
intercorporate
coalitions 377
relations 378, 379 J
interdependencies 317, 319, 374, 451 Japan 71, 166, 342,362, 413, 419, 452
interdependent social systems 19 job 202, 208, 213
interests 119 attainment 324
inter-firm relationships 9, 46, 49, 50, 54, 149, process 197
352, 450,491 control 183
intergrating mechanisms 176 descriptions 195
interlocking of social relations 245 -finding
internal method 204, 211
management process 199, 200, 203, 209,214
structures 290 placement 222
styles 290 posting 183, 184, 186
networks 49 satisfaction 205, 324, 329, 330
rate of return (lRR) 465, 467 security 38
recruitment 182 job-seeker 19, 20
social cohesiveness 221 job-search theory 201
staffing practices 180, 182, 187-193 joint
intemet vii actions 23
interorganizational problem solving 374
alliance network 378 venture 29, 52, 399
boundaries 410 judicial sector 135, 138, 142-145
chains 416 jumping the gun 165, 166
collaboration 7,390, 392,395 just-in-time (JIT) 99, 414
connections 18
links 49,271,327
network 27, 36, 268, 269, 392, 406, 409, K
410, 418,426,427 keiretsu 30, 49-51, 54-56, 58, 100
relationships 357 key technologies 368
strategy 424, 427 Knoke, David 6, 7, 22, 23, 29, 38, 40, 41, 87,
relations (lOR's) 342, 343,377,391 , 392 105, 181 , 187, 195, 196,219, 287, 330, 378,
social capital 282, 391 383,470,494,501,516, 518,520,524
supply networks 409, 427 knowledge 393, 406
ties 42, 285 -based firms 116
trust 35 production 390-408
interpersonal spillover 406
relations 347, 416 Koput, Kenneth W. 7, 8, 50, 99, 132, 285, 293,
skills 185 377, 378, 382, 391, 392, 394, 399, 400, 402,
transactions 152 405, 413,447,453,518,519,531
intersubsidiary contacts 168, 169, 174
intervention of the state 77
intra-group relationships 57 L
intraorganizational Labianca, Giuseppe 4, 162, 232, 294, 324, 325,
boundaries 177 337,519
connections 18 labor
level 237, 238 contracts 237 -265
mobility to careers 181 markets 200, 205, 212
lack of social linkage 115 market

large-scale controls 70 economies 69
Laumann. Edward O. 2. 14.22.219. 220. 222. exchanges 457
223. 224. 228. 231. 232. 233. 287. 513. 518, networks 444
520.533 opportunities 433
lawyers 217-223, 225, 228, 229, 232, 246, 250 relations 28
Lazega. Emmanuel 14,49,77,81,87,214,215, stuff 431. 432
218, 229, 239, 241, 242, 244, 246, 256, 257, ties 458
262,264,265,303,306,334,520,521 uncertainty 433, 443
leadership 424 marketability 113
learning 333, 414, 435 marketing focus 416
by doing 463 markets 285
Lee, Kyungmook 6, 7. 9, 14, 30, 50, 51, 63, 75, Marsden, Peter V. 6, 22, 23, 38, 119, 181, 187,
77,87,132,215,285,347,408,502,529 195, 196,201,207,213,239,299,331,377,
Leenders, Roger Th.A.J. 4, 6, 14, 25, 32, 87, 104, 378, 501, 511, 516, 518, 520, 522, 523, 524,
215, 216, 219, 286. 293. 294, 299, 325. 327. 528,531
328, 329, 333, 343, 389, 408, 449, 452, 459, Mathew Effect 466
487,489,493,494,509,519,521 matrix hoppers 290
legislation 77 maximum likelihood
legitimacy 164, 479 procedures 123
of an entrepreneurial strategy 97 techniques 122
lending relationships 456, 459 measure of standing 140
length of relationship 449 measurement 484
level of social capital 97 measurements of performance 253
levels of analysis 4, 43 measures
liability 220, 240, see also social liability ofeSe 118
license alliance 379, 385 of performance 246
licensing 396. 399 membership 392
link 50 mental health
local cycles 259 agencies 135, 141, 145
location effects vii sector 138,141,142,144
logic of embeddedness 449 mentor 335
long term mentoring relationships 161-179, 256
buyer-supplier relationships 64 metaphor 74, 86
collaborative relationships 414 Microsoft 49, 50, 67
dependencies 358 minders 243
licensing agreement 52 minority equity investment 29
relationships 70,150,415,419,420,450 mixed
satisfaction 415 coupling 459
lottery ticket's value 395 ego-network coupling 453
low perfonners 254, 255 mode of management 287
low-status firms 435, 442, 444 mobility 205
mode of acquisition 220
model
M of independence 126
macro-level sources of trust 34 of quasi-symmetry 123
Maghribi traders 70 models of ese 118
magnitude of ties 25 modular cooperation 358
maintenance of social capital 77, 83, 301, 354 moral
management borderlands 79
of social capital 356-375 resources 19
of supply 411 morality 344
managerial background 267 multi-group membership 47
managers 178-216,299,300,302,307-3\0,317, multilevel
318 embeddedness 238, 260
, networks 199,300,306,310 social capital network 20
manufacturing 399 multinational subsidiary boundaries 161, 162,
marginal workers 47 176
Index - 553
multinationals 167, 168, 178, 306, 362 fonns 25

multi-participant alliances 28 fundamentals 21
multiple generated (corporate) social capital 134,
cycles 259 136
levels 195 hierarchy 311, 313, 315, 316, 318
linear regression (MLR) 169 in health care management 284, 295
multiplex level 446, 451
boundary system 52 of coworkers 249
combinations of dyadic ties 251 of lending relationships 458
exchange system 238 of partial cooperation 345
fonns of organizational boundaries 49 of social relations 461
generalized exchange system 240, 256- opportunities 169
258, 260,262 organization 39, 338, 358
links 49, 57 organizational design 98
networks patterns 32
relations 42, 46 portfolio diversity 400
ties 450, 451, 456 relational contents 23
web of network connections 50 relations 356, 358, 400
multiplexity 450, 454, 458, 486 size 451
multiplicative models 122 structure viii, 5, 248, 304, 318, 352, 366,
multi-site organization 195 427,446,452,484
mutual sustainability 293
competition pattern 152, 155 ties 120, 402,432
cooperation pattern 152, 155 volume 25
cowork tie 253 networking 334
abilities 39
phenomenon 289, 290
N networks 202, 246, 256, 285, 293-295, 298, 364
national subsidiaries 167 of civic engagement 77
natural paradigm 11 new
nature business organizations 463
of organizations 44 ventures 45
of relationships 448 niches 146
of social capital 97 Nissan 30, 420, 423
negative Nohria, Nitin 6, 9, 14, 19, 97, 102, 103, 162,
asymmetry 325, 332, 336 167, 169, 209, 285, 286, 289, 300, 306, 319,
relationships 324-326, 328, 329, 331, 332, 336, 338, 366, 489, 496, 497, 500, 510, 517,
336,337,434 518,519,527,529,531
Netherlands 198,342, 353,361,362 nominalist strategy 22
network 167, 198, 222,253,269, 299,341,352, non-consolidated network 459
365, 374,39~8 , 453 non-equity partnerships 28
activity 400, 405 nonfriendly relationships 416
actors 18 nonprofit sector 195
alters 22 nonredundant contacts 391
based nonsupervisory mentoring relationships 177
management 285, 288, 289, 294, 296, 297 nontradability 61
organizations 284, 293, 295 Nooteboom, Bart 24, 57, 78, 87, 91, 92, 305,
centrality 41,391,394,395, 402 342, 343, 345, 347, 350, 351, 352, 355, 366,
concepts 17 373, 377,432, 451,465,494
configuration 98 nonns 69, 77, 78,100
cons~nt311,313 , 315,316 of generalized reciprocity 77
coupling 449, 451 , 452, 455, 458, 459 NOS (National Organizations Study) 187, 188,
data 237 194, 195
design 103 notion of social capital 72
diversity 403 number
dynamics 17,39 of players in the supply network 422
embeddedness 64 of ties 118, 132, 136,454
formation 18
o patents 390, 391, 393, 395-399, 402, 404, 405

Obstfeld, David 92, 94, 493, 528 patb-dependence 79
occupational categories 195 Pennings, JohaMes M. 6,7,9, 14,30,47,50,51,
OECD countries 284, 286 63,75,77,87,132,215,285,347,408,529
old boy network 52 perceived
Omta,OMo 14,49,99,364,392,494,528,540 constraints 169
open system paradigm 11 similarity 327
openness 100 perception of performance 415-417
opportunities for access 377 performance 238, 239, 318, 319, 334, 415
opportunity 165, 299 assessment 374
costs 185, 325 of corrective actions 360, 364
optimal performance
embeddedness strategy 111 benefits 36
retaliation pattern 153 implications 43, 53
organization measures 241, 242, 246
's network of ties 452 peripheral
stratification 48 backgrounds 271 , 280, 281
-to-organization ties 392 functions 270
organizational person based networks 45
action 299 personal
benefits 391 forms of organizational boundaries 49
boundaries 43, 44, 461 goals 19
career 166, 174, 176 inclusion 48
competencies 344 networks 39
employment policies and practices 180 trust 52
goals 10 personality characteristics 162
level of analysis 46 personalized ties 245
networks 17 personalizing work ties 245
performance 318 personnel department 195
resource 44 pharmaceutical firms 357
role 168,374 philos relationships 81
standing 134-147 physical
organizations 91, 176, 241, 306, 355 capital 72,84, 177
origins of social capital 17 position in the supply network 422
output functions 270 piecewise exponential model 469, 472
outsourcing 367, 369, 413 pipeline 416
overlapping membership 47 pitfalls 362, 394
Owen-Smith, Iason 7, 8, 32, SO, 99, 132, 285, player 422,441,490
293,378,382,395,397,413,531 Podolny, Joel M. 37, 59, 87, 94, 108, 110, 111,
114, 119, 128, 130, 131, 132, 137, 147, 181,
p 185, 215, 299, 300, 305, 318, 319, 321, 377,
378,379, 382,383,389,394,399,419, 432.
pacing technologies 368 433, 434, 435, 444, 447, 448, 449, 458, 465,
panel-regression model 400 466,468,509,530
paradigms 11 political
paradox of embeddedness 351 actor 44. 45
partner selection 360, 363, 369 corporate capital 107, 110-113, 116
partnering embeddedness 71
decisions 369 power 41
matrix 367 relations 40
partners 244, 247 social capital 106
partnership poor customer service 149
agreement 242 popularity 125
management 356,371 position 27,119
sourcing 414 positional advantage 341
patent activity 403 poverty sector 135, 139, 141-145
patenting 390-408
Index - 555
Powell, WaIter W. 7, 8, 28, 32, 36, 49, 50, 57, rational

97,99,100, III, 132,215,285,293,294,295, actors 43, 45
300, 377, 378, 382, 391, 392, 394, 395, 397, paradigm 11
399, 400,402, 405, 406, 413, 447, 448, 453, rationality pressures 186
504,507,518,519,524,530,531,534 RC(2) model 126, 128
power6,73,85,324,330,424 realist approach 22
in the supply network 422 reciprocity 77-79, 83, 87, 100, 125, 290, 302,
resources 40 303,317,327
structures 6 recognition 23
prejudice 183, 186 recruitment procedures 182,214,331
presence of ties 92, 103 redundant
prestige 125,376,387-389 market areas 110
primary contribution 141, 142 ties 47
prism 431-445 referrals 181, 183,201
prismatic perspective 432, 444 regular group 47, 48
problem of order 120 relational
process 359 competence 341, 344
product quality 113 dimension of social capital 90
production capacity 113 inertia 303, 317
professional proximity 131
elites 217, 220, 232 ties 44
identity 176 relations 25
networks 287 relationship
recognition 262 between
reputation 176 strategy and cultural context 99
services firms 54 status and volume 130
profitability 274 duration 449, 456
project teams 307, 313 formation 377
promotion 39, 180-182, 184, 194, 199,324 interfaces 458
practices 180 multiplexity 446, 449
systems 181 to actors 80
property 390 relationships viii, 50, 219, 221, 229, 300, 301,
protection 163 319,351,399,409,414,446,447,450,490
protege 161-166, 168, 176-178 between organizations 465
proximity 118, 122, 125, 126, 128, 131-133 reliability 59
psychological help 163 reputation 73, 164, 479
public reputational status 136
goods 80 research and development 395, 399, 400, 487
sector 195 collaborations 356-375, 391
public service managers 357, 361, 370
organizations 284-297 partnership 51, 52, 57, 362-365
settings 286 relationships 356, 358, 365, 366, 370
pyramidal vehicle networks 423 ties 402, 403, 405
resource 59, 68, 78, 119, 161, 256, 258, 299,
301, 341, 343, 344, 376, 409, 410, 432, 446,
Q 460,465
quality 59, 149,454 allocation decisions 431, 433, 435
of a relationship 446 allocations 444
quasi-symmetry models 122, 125, 126, 128 dependence 31, 32
theory 308, 309, 452
flows 432
interdependencies 31
R value 61
resource-flow network 168
random
restraint 261
data patterns 157
restricted access 377
interactive behaviors 156
retaliation 155,353
random-competitive data 156
dyadicinteraction 157
randomly generated pattern 156
revelation problem 341, 349, 354 metaphor 86

rewards 180 of ties 220
risk sharing 58 theory 202, 358
role closure 200
modeling 163 clubs 392
of networks 431 cohesiveness 88, 89, 92-94,103
of go-between 341 connections 19
of trade 342 constraints 78, 365
contagion 6
s debts 82
domain 68
Sandefur, Rebecca L. 14, 219, 224, 228, 232, embeddedness 237-265
233,533 entrepreneurship 88, 90
satisfaction 415, 416 exchange 416,426
scale advantages 57 system 261
scarcity 61, 78, 79 groups 20
search costs 202 liability 3,5, 10, 11,31,42,59, 150, 186,
secondary contribution 141 291, 298, 318, 323-338, 358, 374, 409,
selection 410, 416,417,421,423, 449, 458, 479,
methods 182,331 483,489
of ties 97 mobility 461
self-entrapment 262 nature 460
semiconductor industry 380 network
seniority 183,239,247,249,250 analysis 459
lockstep system 243 theory 448
services industries 148 networks 163, 185, 194, 197, 198, 200,
sex-specific division 199 208, 211, 215, 238, 249, 284-297, 299,
short-term satisfaction 415 302,391
should expectations 415 organization 222, 223, 230
sidekick effect 161-179, 336 power 40
Silicon Valley 29, 82, 83,99,357,369,463,471 , psychological explanation of trust 33
473,475 relationships 108,209,329,330,334
Simmelian problem of order 120, 122 resources 20, 240
simplex management 323-338
boundary system 52 rules 69
web of network connections 50 service organizations 134, 135
size 195,274,326,400,402,454,456 services 146
in the supply network 422 solidarity 69, 219
small status 198
business entrepreneurs 449 structure 3, 68, 217, 237-239, 249, 301,
businesses 8, 446, 454-456, 458, 459 304,305,447,448,478,485,489,490
companies 358 systems 19,218,221
firm 45, 357 ties 2, 3,6,212,219,229, 241,281
networks 29 socialization 332
group dynamics 65 societal sector 135
towns 19 socioeconomic status 162
Smith-Doerr, Laurel 7, 8, 32, SO, 99, 132, 285, solidarity 69, 87
293, 377, 378, 382, 391, 392, 394, 397, 399, sourcing 417, 419
400,402,405,406,413,447,453,518,519, Spain 198,416
531,535 specific reciprocity 77
social spill-over control 350, 352-354
actors 18,91,93 sponsorship 163, 358
capital I, 17, 20, 36, 44, 72-76, 89, 90, spread ofsocial capital 18
162, 163, 203, 205, 210-212, 217-233, staffing methods 182-184
251,262,284-297,300-302,324,459 status 7, 118, 122, 125, 128, 131, 132, 162,249-
creation 285 251,262,431-445
explanation of problems 364 homophily 131
management 491-493 stimulation 262
Index - 557
stock of social capital 161, 163, 166, 168, 178 system boundaries 22
strategic
action 267
alliance network 377 T
alliances 37, 49, 97, 356, 374, 377 talking shops 291
license alliances 378 Talmud, Dan 107, 108, 109, 110, Ill , 112, 113,
models 424 114, 115, 116, 117, 501 , 515,537
strategies 88-105 task
strong tie 47, 110,248,256,259,318,319,331, environment 299, 306, 318
332 interdependence 309-310, 312, 313
strategy 335 teamwork 100,414
trust relations 35 technological
structural collaboration 358-360, 362
conditions 98 prestige 376-389
dimension of social capital 90, 98 spillover 395
embeddedness 71, 446, 448, 449, 451 technology
equivalence 27 management processes 374
hole theory 304, 317 networks 357
holes 76, 88-92, 98, 102, 103, 110, 119, temporary
305,306,310,317,318,335,391 or marginal category 48
structure workers 47
of advantage 1 tenure 281
of disadvantage 1 tenured partners 246
of institutional context 97 tertius gaudens 91, 93-95, 98, 99,102
of social capital 92, 103 theory
of the network 304, 364 of corporate social capital 483
of ties 119 of knowledge 345
structures 88-105 third parties 97
Stuart, Toby E. 8, 25, 32, 50, 99, 130, 132,378, third-party enforcement 69
379, 382, 383, 389, 399, 408, 432, 433, 444, Third World 79
464,468,494,530, 537 throughput fUnctions 270
subnetworks 177,409,410 ties 50, 128, 136, 146, 150, 155, 161-163, 166,
subordinates 183 167, 169, 176, 179, 182,204, 217, 221, 222,
subsidiary 168, 169, 174,178 225,227, 231 , 238,240, 303
success of firms 280
of a collaboration 359 time 489
of an entrepreneurial strategy 97 span in research procedures 25
supervisors 183, 334 Tobit regression model 456
supervisory mentoring relationships 177 top management team (TMn 268
supplier-customer partnerships 356, 409 total
supplier-manufacturer networks 453 network 22
supply 410, 411 , 413, 416, 418 quality management (TQM) 414
base reduction 421 Toyota 7, 29, 30, 57, 58, 64, 369, 409, 420, 421,
chain 416-418 423,426
management 412 tradeability 80
focus 416 trade
network 424 association 392
breadth 420 expansion 70
strategy 409-427 trading partners 457
structure 418, 419 training 195, 332, 333
supply networks 410, 418, 421 , 423 programs 176
support 24 transaction
surplus consultation 308 cost
Sweden 200, 362 economics (TeE) 342, 343, 345, 347,
symbolic capital 73, 74 348
syndicate relations 432 theory 201,345, 452
syndicates 120 costs 30, 238
synergy 414 interfaces 458
transactional economics 411

transactions 23, 450
w
transfer 180-184,194 weak tie 47, 110,203,208,213,331,335
transfers of resources 260 strategy 336
transfonnation of social capital 17 web 358
triads census 100 will expectations 415
trilateral governance 341, 343, 348, 353, 354 women 199,327
triplex reciprocated ties 254, 255, 260 work
trust 24, 33-37,41, 52, 70, 78, 79, 81, 93, 100, assignments 163
115, 290, 346, 350, 353, 354, 371, 373, 374, experience 164
414,416,465,479 work-related perceptions 41
trust-violation 36
turnover 336
U
U.K. 284,287,288,362,416
uncertainty 274
understandability 85, 290
union
presence 195
strategy 88, 89, 94-97, 99, 100, 102-104,
493
universalism 183
university-industry collaboration 364
upstream partners 357
U.S. 71, 80, 119, 178, 180, 183, 188, 193, 198,
218, 237, 272, 273, 357, 362, 369, 382, 383,
392-394,396,419,432,452
Uzzi, Brian 8, 36, 45,59,65,66,67,71,76, 81,
87,95,96, 107, 108, 109, 111, 116,240,301,
302, 334, 342, 343, 347, 350, 351, 418, 435,
447, 448, 449, 450, 451, 452, 453, 454, 455,
456,459,539
v
valence 220
valuation 78
value 78, 79, 149,217,300,359,379,394,396,
399,409,417,427,450,458
adding process 409
of conduct 344
of elite ties 222, 223, 228
of status 443
Van Rossum, Wouter 49,99,357, 364, 392, 494,
501
venture capital 436, 460-479
finn 438, 440, 442, 461-479
markets 431-445
versatility 83
virtual
corporation 358
organization 358
visibility 163-165
volume 118, 122, 128, 132,458
CONTRIBUTORS
•
Luis Araujo
Senior Lecturer in Marketing, The Management School, University of Lancaster,
Great Britain.
Wayne E. Baker
Associate Professor of Organizational Behavior and Human Resource Management,
Business School, and Faculty Associate, Institute for Social Research, University of
Michigan.
Mario Benassi
Assistant Professor, Department of Computer and Management Sciences, University
of Trento, Italy.
Ed Boxman
Consultant Operational Auditing, Dutch Post Office Organization Postkantoren BV.,
The Hague, The Netherlands.
Daniel J. Brass
Professor of Organizational Behavior, Smeal College of Business Administration,
The Pennsylvania State University.
Ronald L. Breiger
Goldwin Smith Professor of Sociology, Department of Sociology, Cornell
University.
Fabrizio Castellucci
Ph.D. Candidate, Graduate School of Business, Stanford University.
Alison Davis-Blake
Associate Professor of Management, Graduate School of Business, University of
Texas-Austin.
Patrick Doreian
Professor of Sociology, Department of Sociology, University of Pittsburgh.
Geoff Easton
Professor of Marketing, The Management School, University of Lancaster, Great
Britain.
Ewan Ferlie
Professor of Public Services Management, The Management School, Imperial
Management School, Great Britain.
HenkFlap
Associate Professor of Sociology, Department of Sociology and Interuniversity
Center for Social Science Research and Methodology, University of Utrecht, The
Netherlands.
John Freeman
Helzel Professor in the Haas School of Business, University of California, Berkeley
Martin Gargiulo
Assistant Professor of Organizational Behavior, INSEAD, France.
James J. Gillespie
Ph.D. Candidate, Department of Organization Behavior, J.L. Kellogg Graduate
School of Management, Northwestern University.
Elizabeth H. Gorman
Ph.D. Candidate, Department of Sociology, Harvard University.
Shin-Kap Han
Assistant Professor of Sociology, Department of Sociology, Cornell University.
Christine M. Harland
Senior Research Fellow, Centre for Research in Strategic Purchasing and Supply
(CRiSPS), School of Management, University of Bath, Great Britain.
Pamela R. Hauoschild
Associate Professor of Organizational Behavior, Graduate School of Business,
Stanford University.
John P. Heinz
Owen L. Coon Professor of Law, School of Law, Northwestern University, and
Distinguished Research Fellow, American Bar Foundation.
Andrew D. Henderson
Assistant Professor of Management, Graduate School of Business, Columbia
University.
Contributors - 561
Monica C. Higgins
Assistant Professor of Organizational Behavior, Harvard Business School, Harvard
University.
Dawn Iacobucci
Professor of Marketing, J.L. Kellogg Graduate School of Management,
Northwestern University
David Knoke
Professor of Sociology, Department of Sociology, University of Minnesota.
Kenneth W. Koput
Associate Professor, Department of Management and Policy, University of Arizona.
Giuseppe Labianca
Assistant Professor of Organizational Behavior, A.B. Freeman School of Business,
Tulane University.
Edward O. Laumann
George Herbert Mead Distinguished Service Professor of Sociology, The University
of Chicago
Emmanuel Lazega
Associate Professor of Sociology, University of Versailles and LASMAS-CNRS,
France.
Kyungmook Lee
Assistant Professor of Management, College of Business Administration, Seoul
National University, Korea.
Peter V. Marsden
Professor of Sociology, Department of Sociology, Harvard University.
Nitin Nohria
Professor of Organizational Behavior, Harvard Business School, Harvard
University.
Bart Nooteboom
Professor of Industrial Organization, School of Management and Organization,
University of Groningen, The Netherlands.
David Obstfeld
Doctoral Candidate, Organizational Behavior and Human Resource Management,
Business School, University of Michigan.
Onno (S.W.F.) Omta

Associate Professor of Innovation and Business Development, School of
Management and Organization, University Groningen, The Netherlands.
Jason Owen-Smith
Ph.D. Candidate, Department of Sociology, University of Arizona.
Johannes M. Pennings
Professor of Management, The Wharton School, University of Pennsylvania.
Joel M. Podolny
Associate Professor of Strategic Management and Organizational Behavior,
Graduate School of Business, Stanford University.
Walter W. Powell
Professor of Sociology, Department of Sociology, University of Arizona.
Wouter van Rossum

Professor of Commercial Management, Marketing, and Innovation Management,
University of Twente, The Netherlands.
Rebecca L. Sandefur
Ph.D. Candidate, Department of Sociology, University of Chicago.
Laurel Smith-Doerr
Ph.D. Candidate, Department of Sociology, University of Arizona.
Toby E. Stuart
Assistant Professor of Organizations and Strategy, Graduate School of Business,
University of Chicago.
Dan Talmud
Lecturer, Graduate Program in Applied and Organizational Sociology, Department
of Sociology and Anthropology, University of Haifa, Israel.
Brian Uzzi
Associate Professor of Organization Behavior and Sociology, J.L. Kellogg Graduate
School of Management, Northwestern University.
THE EDITORS
Shaul M. Gabbay
Lecturer. Davidson Faculty of Industrial Engineering and
Management. Technion. Israel Institute of Technology.
Israel.

Assistant Professor of Business Development and Business
Research Methods. School of Management and
Organization. University Groningen. The Netherlands.

Gabbay 1999 Corporate Social Capital and Liability

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Gabbay 1999 Corporate Social Capital and Liability

Transféré par

Droits d'auteur :

Formats disponibles

Corporate Social

Capital and Liability

Library of Congress Cataloging-in-Publication Data

Corporate social capital and liability I edited by Roger Th. AJ.

1. Business networks. 2. Social nelworks. 3. Industrial

AII rights reserved. No part of this publicatian may be reproduced, stored in a

Printed an acid-free paper.

SOCIAL CAPITAL AND ORGANIZATIONS

CORPORATE SOCIAL CAPITAL

represented various aircraft manufacturers; their length represented the number of

SOCIAL CAPITAL AND SOCIAL LIABILITY

Organizations and their members continuously engage in social relationships.

Relations Among Individuals

by having access to other (important) people in the organization, and by having

Relations Among Organizations

An organizational risk attached to the personal nature of relationships is that

THE ADDED VALUE OF esc

head Tele-TV, a new video-program joint venture of three regional telephone

SOCIAL CAPITAL EMBEDDED IN SOCIAL RELATIONS

Figure 1. Schematic of a multi-level social capital network

university hiring and promotion procedures typically require letters of

SOME NETWORK FUNDAMENTALS

Identifying Actors and System Boundaries

Total and Ego-Centric Networks

Network Relational Contents

networks, or more narrowly targted on specific recipients, for example, a marketing

types of governance structures designed to monitor and control the partners'

PRODUCTION DEPARTMENT SALES FORCE

Figure 2. Graph and matrix representations of a hypothetical eight-actor network

Hierarchical relations Subsidiaries

Equity partnerships Minority equity investments

Non-equity partnerships Franchising

Multi-participant alliances Data banks and information clearinghouses

Market relations Spot exchanges

franchisers such as McDonald's Corp. exercise centralized coordination to safeguard

Theories of Alliance Formation

Resource dependence principles seemed more helpful than transaction cost

together and that dense co-location in an alliance network enhances mutual

Trust as Corporate Social Capital

Figure 4. Trust is an intervening factor in the alliance formation process

the moral communities within which trustworthiness conveyed great importance in

producer Novell Inc.'s disastrous effort to integrate its subsidiary WordPerfect

Rigid bureaucratic hierarchies yielded to experiments in cross-functional teams that

Networks and Career Capital

other men. Women employees differentiated according to network contents,

Networks as Power Resources

A core theoretical proposition, derived from resource dependence principles, is

CONCEPT OF ORGANIZATION AND ITS SOCIAL CAPITAL

Boundary Spanner or Multiple-Group Membership

participate in that tournament will 'plateau,' become sidetracked or might even be

Marginal: free agents, sub contractors,

Associate: temporary employees,

Core: long term employees, partners,

Figure 1. Stratification of firms based on partial inclusion of their members

Multiplex versus Personal Forms of Organizational Boundaries

contracted for the pre-arranged installation of Microsoft's operating system in what

In short, corporate social capital bifurcates into personalized and depersonalized

managing external dependencies depends critically on the quality of the relationship

Boundary Transaction System

Table 1. Examples of social capital among organizations

Mediated by Individuals (Simplex) Mediated by Systems (Multiplex)

PERFORMANCE IMPLICATIONS OF SOCIAL CAPITAL

In the next two sections, we belabor these implications by reviewing two

Social Capital of Business Groups