Académique Documents
Professionnel Documents
Culture Documents
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>
Copyright@I999bySpringerSciCIlCl'+RusinessMediaNewYork
Originali)' published by Kluwer Academic Publishers 1999
Softcover reprint of the hardcover 1st edition 1999
•
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
Roger Th.A.J. Leenders
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.
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).
CSC: The Structure of Advantage and Disadvantage - 5
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.
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.
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
CSC: The Structure of Advantage and Disadvantage - 9
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.
10 - Corporate Social Capital and Liability
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
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
12 - Corporate Social Capital and Liability
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
CSC: The Structure of Advantage and Disadvantage - 13
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.
14 - Corporate Social Capital and Liability
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
18 - Corporate Social Capital and Liability
1
JANE DICK
BANKRIGHT SQUAREBILT
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.
knowledge should consult such didactic texts as Knoke and Kuklinski (1982) or
Wasserman and Faust (1994).
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).
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
24 - Corporate Social Capital and Liability
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
26 - Corporate Social Capital and Liability
.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
(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.
Organizational Networks and Corporate Social Capital - 27
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
28 - Corporate Social Capital and Liability
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
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?
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.
32 - Corporate Social Capital and Liability
STRucrURAL
CONDITIONS
TRUST ) ALLIANCE
FORMS
COMMUNICATION
NETWORK
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)
Organizational Networks and Corporate Social Capital - 37
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.
38 - Corporate Social Capital and Liability
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.
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
44 - Corporate Social Capital and Liability
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.
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.
Social Capital of Organization - 47
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).
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).
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.
.........................................
...
Figure 2. Boundary transaction system comprising four individuals among two organizations
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.
Social Capital of Organization - 53
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.
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
Social Capital of Organization - 55
(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
56 - Corporate Social Capital and Liability
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.
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
58 - Corporate Social Capital and Liability
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
Social Capital of Organization - 59
'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
60 - Corporate Social Capital and Liability
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
Social Capital of Organization - 61
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.
62 - Corporate Social Capital and Liability
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
Social Capital of Organization - 63
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.
64 - Corporate Social Capital and Liability
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
66 - Corporate Social Capital and Liability
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
Social Capital of Organization - 67
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.
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
A Relational Resource Perspective on Social Capital - 71
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
72 - Corporate Social Capital and Liability
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
74 - Corporate Social Capital and Liability
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.
•
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)
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.
80 - Corporate Social Capital and Liability
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 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
A Relational Resource Perspective on Social Capital - 83
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
86 - Corporate Social Capital and Liability
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,
A Relational Resource Perspective on Social Capital - 87
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
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
Social Capital by Design - 93
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.
Alter I Alter 1
Ego Ego
Alter 2 Alter 2
Disunion Union
Figure 1. lIlustration of disunion and union strategies
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
96 - Corporate Social Capital and Liability
viewed the potential union as a threat that would reduce their ability to compete for
customers-that is, to operate using the disunion strategy.
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.
98 - Corporate Social Capital and Liability
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
Social Capital by Design - 99
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
100 - Corporate Social Capital and Liability
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.
~
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
-
102 - Corporate Social Capital and Liability
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
Social Capital by Design - 103
Entrepreneurial
Strategy
Disunion
-~----
.- .- -- --
"
/
/ ""
/
I
- - Disunion
SO/50 - - - - Union
-----
Union
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
104 - Corporate Social Capital and Liability
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.
Social Capital by Design - 105
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.
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
108 - Corporate Social Capital and Liability
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)
There are also some significant differences between the three forms:
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.
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.
Corporate Social Capital and Liability- 113
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.
(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.
Corporate Social Capital and Liability- 115
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
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
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)?
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
-~
122 - Corporate Social Capital and Liability
(1)
Dimensions of Corporate Social Capital- 123
Shearson
Lehman
45.618 ) \ 20.665
9.245 5.96~
Goldman, ~ Smith
( sa:hS ~ Ba~ney \
~ 3.860 ~
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
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)
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
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
('}
IV
-..l
-~
128 - Corporate Social Capital and Liability
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
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.
130 - Corporate Social Capital and Liability
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.
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
Dimensions of Corporate Social Capital- 131
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
132 - Corporate Social Capital and Liability
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
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
Corporate Social Capital
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.
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,
actual or virtual, that accrue to an individual or a group by virtue of possessing a
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.
(1)
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:
(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.
138 - Corporate Social Capital and Liability
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
140 - Corporate Social Capital and Liability
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.
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.
144 - Corporate Social Capital and Liability
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
146 - Corporate Social Capital and Liability
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
150 - Corporate Social Capital and Liability
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.
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
152 - Corporate Social Capital and Liability
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).
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.
154 - Corporate Social Capital and Liability
(1)
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):
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)'
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.
156 - Corporate Social Capital and Liability
• Retaliation
•
Chump
• Competitive • Competitive
• Cooperative • Cooperative
p p
• Random
• Rand-Ccmp
• • Random
Rand-Comp
• Retaliation
.Chu np
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
Customer Service Dyads - 157
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 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
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.
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.
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
The Sidekick Effect - 165
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
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
The Sidekick Effect - 167
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
168 - Corporate Social Capital and Liability
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.
The Sidekick Effect - 169
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.
170 - Corporate Social Capital and Liability
9 lD 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-.07
-.11 -.lD
-.09 -.08 -.14
-.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*
172 - Corporate Social Capital and Liability
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.
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.
174 - Corporate Social Capital and Liability
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.
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.
178 - Corporate Social Capital and Liability
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.
The Sidekick Effect - 179
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
data base for our analysis, the 1991 National Organizations Study (NOS), we present
findings, and conclude by discussing their implications for organizations and
employees.
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?
188 - Corporate Social Capital and Liability
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
Social Capital in Internal Staffing Practices - 189
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
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.
192 - Corporate Social Capital and Liability
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
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
Social Capital in Internal Staffing Practices - 193
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
194 - Corporate Social Capital and Liability
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
196 - Corporate Social Capital and Liability
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.
8. Statements on statistical significance refer to the .05 level.
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
198 - Corporate Social Capital and Liability
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
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.
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.
Getting a Job as a Manager - 203
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.
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
Getting a Job as a Manager - 205
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.
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%
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)
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
208 - Corporate Social Capital and Liability
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 4. Group-specific differences in employment of social contact in finding a job among Dutch
managers oflarger companies
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
Getting a Job as a Manager - 211
Table 5. Proportion of Dutch managers finding a job through social contacts by two indicators of
social resources, 1987 (n=1359)
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).
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
Getting a Job as a Manager - 213
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.
214 - Corporate Social Capital and Liability
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
Getting a Job as a Manager - 215
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.
216 - Corporate Social Capital and Liability
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
218 - Corporate Social Capital and Liability
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).
220 - Corporate Social Capital and Liability
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.
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
The Changing Value of Social Capital in an Expanding Social System - 223
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).
226 - Corporate Social Capital and Liability
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
The Changing Value of Social Capital in an Expanding Social System - 227
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
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
The Changing Value of Social Capital in an Expanding Social System - 229
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
230 - Corporate Social Capital and Liability
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
The Changing Value of Social Capital in an Expanding Social System - 231
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
232 - Corporate Social Capital and Liability
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
The Changing Value of Social Capital in an Expanding Social System - 233
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
•
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
238 - Corporate Social Capital and Liability
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
240 - Corporate Social Capital and Liability
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
Generalized Exchange and Economic Performance - 241
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
242 - Corporate Social Capital and Liability
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)
244 - Corporate Social Capital and Liability
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.
Generalized Exchange and Economic Performance - 245
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)
246 - Corporate Social Capital and Liability
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).
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
248 - Corporate Social Capital and Liability
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:
Generalized Exchange and Economic Performance - 249
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.
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
Generalized Exchange and Economic Performance - 251
[
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.
Generalized Exchange and Economic Performance - 253
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.
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
256 - Corporate Social Capital and Liability
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.
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
Generalized Exchange and Economic Performance - 259
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
260 - Corporate Social Capital and Liability
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.
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
262 - Corporate Social Capital and Liability
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:
Generalized Exchange and Economic Performance - 263
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
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.
264 - Corporate Social Capital and Liability
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.
Generalized Exchange and Economic Performance - 265
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
268 - Corporate Social Capital and Liability
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
CEO Demographics and Acquisitions - 269
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
270 - Corporate Social Capital and Liability
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.
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.
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
CEO Demographics and Acquisitions - 273
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,
274 - Corporate Social Capital and Liability
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
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
CEO Demographics and Acquisitions - 277
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
278 - Corporate Social Capital and Liability
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
CEO Demographics and Acquisitions - 279
DISCUSSION
Overall. the results provide support for the idea that the networks that come with
educational and functional backgrounds affect acquisitions. These background
280 - Corporate Social Capital and Liability
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
CEO Demographics and Acquisitions - 281
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
CEO Demographics and Acquisitions - 283
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.
3. Statements on statistical significance refer to the .05 level.
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
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.
286 - Corporate Social Capital and Liability
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.
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.
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).
Public Service Organizations: Social Networks and Social Capital - 289
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.
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,
294 - Corporate Social Capital and Liability
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?
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.
296 - Corporate Social Capital and Liability
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.
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.
•
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,
actual or virtual, that accrue to an individual or a group by virtue of possessing a
The Dark Side of Social Capital - 299
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
302 - Corporate Social Capital and Liability
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
The Dark Side of Social Capital - 303
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
304 - Corporate Social Capital and Liability
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.
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
306 - Corporate Social Capital and Liability
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
The Dark Side of Social Capital - 307
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
308 - Corporate Social Capital and Liability
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.
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
310 - Corporate Social Capital and Liability
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):
The Dark Side of Social Capital - 311
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
312 - Corporate Social Capital and Liability
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
The Dark Side of Social Capital - 313
(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.
The Dark Side of Social Capital - 315
Table 2. Structural hole effects on the number of individual coordination failures (Standardized
OLS coefficients)
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
316 - Corporate Social Capital and Liability
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.
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
The Dark Side of Social Capital - 319
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
320 - Corporate Social Capital and Liability
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.
322 - Corporate Social Capital and Liability
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
324 - Corporate Social Capital and Liability
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.
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
326 - Corporate Social Capital and Liability
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).
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
Social Capital, Social Liabilities, and Social Resources Management - 327
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
328 - Corporate Social Capital and Liability
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
Social Capital, Social Liabilities, and Social Resources Management - 329
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.
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
330 - Corporate Social Capital and Liability
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
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)
332 - Corporate Social Capital and Liability
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
Social Capital, Social Liabilities, and Social Resources Management - 333
(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,
334 - Corporate Social Capital and Liability
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
Social Capital, Social Liabilities, and Social Resources Management - 335
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
336 - Corporate Social Capital and Liability
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.
Social Capital, Social Liabilities, and Social Resources Management - 337
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
338 - Corporate Social Capital and Liability
•
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.
342 - Corporate Social Capital and Liability
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
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
The Triangle: Roles of the Go-Between - 345
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).
346 - Corporate Social Capital and Liability
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
The Triangle: Roles of the Go-Between - 347
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:
348 - Corporate Social Capital and Liability
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
The Triangle: Roles of the Go-Between - 349
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.
350 - Corporate Social Capital and Liability
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
352 - Corporate Social Capital and Liability
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,
The Triangle: Roles of the Go-Between - 353
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
354 - Corporate Social Capital and Liability
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.
The Triangle: Roles of the Go-Between - 355
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
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
358 - Corporate Social Capital and Liability
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
The Management of Social Capital in R&D Collaboration - 359
developed in cooperation with the partners, which add varying degrees of value
(Lorenzoni and Baden-Fuller 1995).
t f f
f
I Implement corrective actions
I'-
Measure network's
results
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
The Management of Social Capital in R&D Collaboration - 361
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.
362 - Corporate Social Capital and Liability
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.
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.
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
The Management of Social Capital in R&D Collaboration - 363
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.
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
364 - Corporate Social Capital and Liability
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
The Management of Social Capital in R&D Collaboration - 367
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.
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
368 - Corporate Social Capital and Liability
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.
The Management of Social Capital in R&D Collaboration - 369
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.
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.
The Management of Social Capital in R&D Collaboration - 371
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
372 - Corporate Social Capital and Liability
Focus
Objectives
Strategy
Environment
Policy
Technologies
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
The Management of Social Capital in R&D Collaboration - 373
Scenario
Strategy
Location
Partners
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
The Management of Social Capital in R&D Collaboration - 375
NOTES
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
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
Technological Prestige and the Accumulation of Alliance Capital - 379
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.
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:
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
Technological Prestige and the Accumulation of Alliance Capital - 383
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
384 - Corporate Social Capital and Liability
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
Technological Prestige and the Accumulation of Alliance Capital - 385
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
386 - Corporate Social Capital and Liability
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
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
Technological Prestige and the Accumulation of Alliance Capital - 389
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
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.
Collaboration and Patenting in Biotechnology - 393
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 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.
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.
396 - Corporate Social Capital and Liability
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:
Collaboration and Patenting in Biotechnology - 397
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.
398 - Corporate Social Capital and Liability
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.
400 - Corporate Social Capital and Liability
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
Collaboration and Patenting in Biotechnology - 401
Table 1. Panel regression models of the effects of network centrality and patenting activity
Dependent variables at time t+ I
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)
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.
Collaboration and Patenting in Biotechnology - 403
R&D Ties
Centrality
Network Diversity
Figure 1. Visual representation of the relationship between network measures and patent
activity
Table 2. Logistic regression models predicting initial patents, and initial network ties
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)
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
406 - Corporate Social Capital and Liability
APPENDIX
Descriptive Statistics and Correlations ll
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
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.
408 - Corporate Social Capital and Liability
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.
410 - Corporate Social Capital and Liability
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.
Table 1. Supply chain management: the industrial organization perspective (derived from Ellram
1991)
Business management
Business
support
Operations management
Human resource management
Financial management
Marketing
Information systems management
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
416 - Corporate Social Capital and Liability
Supplier's Customer's
perception of perception of
requirements requirements
mismatch 1
mismatch 2
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
418 - Corporate Social Capital and Liability
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.
Network
breadth
Network length
....1 - - - - - - (number of levels or echelons) ------t~~
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.
420 - Corporate Social Capital and Liability
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.
Supply Network Strategy and Social Capital - 421
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.
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
424 - Corporate Social Capital and Liability
Manufacturer Manufacturer
Manufacturer Manufacturer
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
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
432 - Corporate Social Capital and Liability
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
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.
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.)
436 - Corporate Social Capital and Liability
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
Choosing Ties from the Inside of a Prism - 437
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
438 - Corporate Social Capital and Liability
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
Choosing Ties from the Inside of a Prism - 439
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.
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.
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.
Choosing Ties from the Inside of a Prism - 441
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
442 - Corporate Social Capital and Liability
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.
Choosing Ties from the Inside of a Prism - 443
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.
444 - Corporate Social Capital and Liability
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
Choosing Ties from the Inside of a Prism - 445
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
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.
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.
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).
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
452 - Corporate Social Capital and Liability
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
Corporate Social Capital and the Cost of Financial Capital - 453
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.
454 - Corporate Social Capital and Liability
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
Corporate Social Capital and the Cost of Financial Capital - 455
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
Corporate Social Capital and the Cost of Financial Capital - 457
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
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
458 - Corporate Social Capital and Liability
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
Corporate Social Capital and the Cost of Financial Capital - 459
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.
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
462 - Corporate Social Capital and Liability
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.
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
464 - Corporate Social Capital and Liability
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
Venture Capital as an Economy of Time - 465
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.
Venture Capital as an Economy of Time - 467
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
468 - Corporate Social Capital and Liability
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.
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
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.
470 - Corporate Social Capital and Liability
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.
472 - Corporate Social Capital and Liability
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.
Venture Capital as an Economy of Time - 473
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
474 - Corporate Social Capital and Liability
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.
Venture Capital as an Economy of Time - 475
Table 3. Exponential model with competing risks estimating effects of investor centrality
ACQ FAIL IPO
Coef. (s.e.) Coef. (s.e.) Coef. (s.e.)
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
476 - Corporate Social Capital and Liability
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.
478 - Corporate Social Capital and Liability
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
Venture Capital as an Economy of Time - 479
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
•
Roger Th.A.J. Leenders
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
484 - Corporate Social Capital and Liability
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.
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.
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.
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.
488 - Corporate Social Capital and Liability
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
CSC: An Agenda for the Future - 489
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.
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.
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.
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.
494 - Corporate Social Capital and Liability
•
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.
496 - Corporate Social Capital and Liability
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
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.'
Administrative Science Quarterly 26: 331-348.
1984 'Being in the Right Place: A Structural Analysis of Individual Influence in an Organization.'
Administrative Science Quarterly 29: 518-539.
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.
500 - Corporate Social Capital and Liability
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.'
Administrative Science Quarterly 36: 583-613.
504 - Corporate Social Capital and Liability
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.'
Administrative Science Quarterly 22: 248-263.
Edwards, R.
1979 Contested Terrain: The Transjonnation of the Workplace in the Twentieth Century. New York:
Basic Books.
506 - Corporate Social Capital and Liability
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
Marketing 12: 470-512.
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.
512 - Corporate Social Capital and Liability
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.
514 - Corporate Social Capital and Liability
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
Sociological Review 58: 768-786.
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
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
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
Management 38: 48-53.
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.
524 - Corporate Social Capital and Liability
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
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
528 - Corporate Social Capital and Liability
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
Marketing 49: 41-50.
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
Management Journal 13: 29-46.
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:
Cambridge University Press.
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.
530 - Corporate Social Capital and Liability
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
Quarterly 39: 458-483.
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:
Harvard Business School Press.
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
Management Review 30: 67-87.
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:
Harvard Business School Press.
Powell, W. W., and L. Smith-Doerr
1994 'Networks and Economic life.' In N. Smelser and R. Swedberg (eds.), The Handbook of Economic
Sociology, 368-402. Princeton: Princeton University Press.
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.
532 - Corporate Social Capital and Liability
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.
534 - Corporate Social Capital and Liability
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:
Cambridge University Press.
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.
536 - Corporate Social Capital and Liability
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).
538 - Corporate Social Capital and Liability
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
Management Journal 37: 688-702.
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.
540 - Corporate Social Capital and Liability
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
Quarterly 29: 74-92.
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
Management Review 15: 421-458.
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
Sociological Review 56: 60-71.
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
Sociology, 77-107. Princeton: Princeton University Press.
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.
542 - Corporate Social Capital and Liability
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
546 - Corporate Social Capital and Liability
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
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
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
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
558 - Corporate Social Capital and Liability
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.
560 - Corporate Social Capital and Liability
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.
562 - Corporate Social Capital and Liability
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.
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.