Vous êtes sur la page 1sur 11

Economic Modelling 63 (2017) 215225

Contents lists available at ScienceDirect

Economic Modelling
journal homepage: www.elsevier.com/locate/econmod

Determinants of control structure choice between entrepreneurs MARK

and investors in venture capital-backed startups

Lei Wanga, Fangzhao Zhoua, Yunbi Anb,
School of Business, Jiangnan University, Wuxi, Jiangsu 214122, China
Odette School of Business, University of Windsor, Windsor, Ontario, Canada N9B 3P4


JEL: This study examines two kinds of control structures in venture capital-backed startups. Based on incomplete
G32 contracting theory, we analyze the inuence of various factors on these control structures in venture capital
D23 models, from the perspectives of investors, entrepreneurs, and startups. In particular, we show how factors such
G24 as bargaining power, monitoring costs, private benets, and risk aversion impact the allocation of control rights.
Keywords: Using the survey data on a heterogeneous group of venture capital-backed startups in China, we empirically
Control rights allocation structure examine the impacts of various factors on the control structures of these enterprises. Based on the full sample of
Startups enterprises, we nd that the stronger the venture capitalists bargaining power and the higher the monitoring
costs, the more likely investors and entrepreneurs are to prefer joint control. Further, the greater the
Venture capitalists
entrepreneurs nancing need and private benets, the more likely investors and entrepreneurs are to choose
Incomplete contracting theory
joint control. High-tech startups are more likely to choose a joint control model than those in traditional
industries. This is especially true for high-tech startups at an early stage of development. In addition, for high-
tech startups, the probability of choosing joint control shows a negative relationship with investors strategic
benets and a positive relationship with investors risk aversion. Regarding startups in traditional industries,
investors strategic benets and risk aversion have an insignicant impact on their control structures.

1. Introduction performance, but also serve as the mechanism and channel through
which entrepreneurs and venture capitalists manage total risk. The
Startups at an early stage of development need signicant invest- selection of control structures in startups is the result of the negotiation
ment support for R & D and innovation. Typically, startups face greater process between venture capitalists and entrepreneurs, and reects the
diculties and higher costs in accessing nancing through traditional purpose of control rights of the two participant parties. It is particularly
means and have to turn to venture capital (VC), which has been important for us to understand the role of various factors in determin-
growing since the 1950s. VC contributes signicantly to the develop- ing the control structures in VC-backed startups. Previous research
ment of high-tech startups at an early stage of development (Carvell results see the incentives and benets of venture capitalists and
et al., 2013). For example, in the US, venture capital invests in startups entrepreneurs as important factors aecting control structures. There
over 15 billion dollars every year (Chen et al., 2011). However, is a signicant discrepancy between startups risks and the benets of
startups high risk, asymmetric information, and commissioning issues control rights. While both entrepreneurs and venture capitalists play a
imply that venture capitalists and entrepreneurs must determine the critical role in startups, their contributions to these enterprises are
structure of corporate control despite incomplete contracting condi- dicult to measure. Given the diculties of data collection, current
tions so as to protect their future. The allocation of control rights has empirical research concentrates on the description of board seats and
emerged as the key in startup governance with the intervention of VC, votes, etc., under the listed startup VC contract, while there has been
and is a key factor in the success of startups. little empirical research on the factors aecting the selection of control
Previous studies focus primarily on control rights allocation in VC- structures. In addition, previous empirical work in this area focuses
backed startups, and pay little attention to the resulting control primarily on European and US enterprises rather than enterprises in
structures in these enterprises. Control structures, as an important emerging markets.
tool of corporate governance, not only aect startups growth and This paper intends to provide a comprehensive analysis of how

Corresponding author.
E-mail addresses: wanglei0663@163.com (L. Wang), traveloguezfz@qq.com (F. Zhou), yunbi@uwindsor.ca (Y. An).

Received 22 July 2016; Received in revised form 23 January 2017; Accepted 15 February 2017
0264-9993/ 2017 Elsevier B.V. All rights reserved.
L. Wang et al. Economic Modelling 63 (2017) 215225

various factors aect the selection of control structures in VC-backed Cestone (2014) proposes a shared contingent governance structure of
startups in China, the largest emerging market in the world. More control rights, which combines the concept of contingent control and
specically, based on the control right types and structures in startup joint control by investors and entrepreneurs. de Bettignies (2008) sums
contracts, we rst classify control structures into joint control and up extant unilateral control, contingent control, as well as shared
contingent control, and describe their characteristics. Second, based on contingent control, and designs three dependent control right cong-
incomplete contracting theory, we analyze the inuence of various uration structures based on an incomplete contracting model, i.e.,
factors on such control structures in venture capital models, from the entrepreneur control, investor control, and joint control. de Bettignies
perspectives of investors, entrepreneurs, and startups. In particular, we (2008) studies the application of nancing tools such as common stock,
distinguish between two kinds of benets of control rights, and analyze preferred stock, convertible preferred stock, secured loans, and un-
how the control structure choice is related to control right benets as secured loans in control rights conguration contracts. Xiong (2010)
well as the bargaining power and control rights demand of various puts forth a third type of control mechanism that combines unilateral
parties. control and bilateral control to oset the negative impact of adopting
Third, using the survey data on 193 venture capital-backed startups only one method, to achieve better control performance. In addition,
in China, we empirically examine the factors inuencing the selection Fluck (2010) studies the relationship between control rights, incentive
of control structures in Chinese industries, and reveal how various motivation, and entrepreneurship, and examines two control struc-
factors impact the control structure choice in this group of hetero- turescontingent control and unconditional control. Unconditional
geneous enterprises. It is noteworthy that the risks to venture control is basically consistent with joint control. Andrieu (2012)
capitalists in high-tech enterprises and in traditional industries are focuses on striking an optimum balance between a venture capitalists
dierent. In addition, the private benets from control rights and the control rights and an entrepreneurs control rights. After comparing the
strategic benets from VC investment in high-tech enterprises and factors inuencing control structures under an independent venture
other startups in traditional industries are dierent. Thus, we divide capital model and bank-/public-funded venture capital model, Hirsch
our sample rms into high-tech enterprises and traditional enterprises, and Walz (2013) proclaim that the dierences between them under VC
and further analyze whether various factors have a dierent impact on models are less than under other intervention mechanisms. Ma et al.
control structures of startups in high-tech industries and in traditional (2013) analyze the application and performance of three types of
industries. binary right structures under a joint investment model, and propose
We nd that the probability of high-tech startups selecting joint that the rights and statuses of venture capitalists and entrepreneurs
control is greater. Entrepreneurs private benets, venture capitalists should be matched in such models.
bargaining power, investment amount, degree of risk aversion, and Based on the literature, we understand that startup control
monitoring cost exhibit a signicantly positive relationship with the structures can be classied as unilateral control, contingent control,
probability of startups choice of joint control. Further, high-tech and joint control. Both contingent control and unilateral control can be
startups at an early stage of development tend to choose joint control dened as a discrete variable that takes the value of 0 or 1. Unilateral
while those at later stages of development prefer contingent control control refers to the control rights of either the venture capitalist or
structures. Our results indicate that the eects of various factors on the entrepreneur, while contingent control depends on whether a random
choice of control structures may dier between startups in high-tech variable (signal) can be realized in the future. Either the startup
industries and those in traditional industries. Our analysis comple- entrepreneur or the venture capitalist rst owns the control rights
ments the empirical evidence as to various factors aecting control and then reassigns it according to observable corporate signals (Fluck,
structures in VC rms in developed markets, and helps us understand 2010; Masulis and Nahata, 2009). Unilateral control by entrepreneurs
the heterogeneous control structures observed in VC-backed startups. or venture capitalists is a special case of contingent control. Therefore,
Our ndings also provide insight into the optimal allocation of control unilateral control in this paper is classied as a part of contingent
rights in dierent types of startups, and have policy implications for control, and the startup corporate control structure in a VC model is
promoting the development and sound corporate governance of these mainly divided into contingent control and joint control. In the
enterprises. following sections, this paper will combine the theoretical analysis of
The remainder of this paper proceeds as follows. Section 2 provides startup joint control and contingent control to discuss the main factors
a literature review. Section 3 develops the hypotheses based on theories that aect the selection of control structure.
and literature. Section 4 describes the research design and the major
variables in the regressions. Section 5 discusses the empirical results, 2.2. Factors inuencing startup control structure
while Section 6 concludes.
Most studies examining the factors that inuence corporate control
2. Literature review structures have concentrated on venture capitalists, entrepreneurs, and
startups (Chen et al., 2014). As far as venture capitalists are concerned,
2.1. Types of startup control structures the chief factor is the venture capitalists risk attitude. Cumming and
Johan (2007) nd that the suggestions proposed by venture capitalists
According to incomplete contracting theory, control structures refer for hedging risks can signicantly facilitate the eorts of entrepreneurs.
to the allocation of startup control rights between venture capitalists This is necessary to endow venture capitalists with more cash ow and
and entrepreneurs (Wang et al., 2009). Hellmann (1998) is the rst to proportionate control rights to encourage them to provide valuable
propose two types of control structuresentrepreneur control and advice and suggestions. Tan et al. (2008) measure venture capitalists
investor control. He also analyzes the mechanism through which these monitoring cost and risk aversion cost by the time they spend on
control rights could be switched. Hellmann (1998) treats control rights inspecting entrepreneurs. They nd that venture capitalists risk
as a discrete variable that takes the value 0 or 1, based on whether attitudes aect their selection of control structures. This then leads
unilateral control is vested in the venture capitalist or the entrepreneur. to the second factor aecting their choice, which is the cost venture
Adapting Hellmanns (1998) theory, Jukka (2002) constructs a con- capitalists pay to hedge risks. Focusing on venture capitalists risk
tingent control structure for startups under continuous changes, and perception, Antonczyk and Salzmann (2012) nd that venture capital-
shows that a contingent control structure is better than other types. ists risk attitudes and risk aversion aect their risk perception, and by
Schmidt (2003) explores contingent control structures in practice, and controlling other factors during empirical research, they also nd that
describes the method of implementing contingent control via conver- risk perception aects venture capitalists appeal for corporate rights.
tible securities. By combining unilateral control and contingent control, Venture capitalists risk attitudes and the actions they take to mitigate

L. Wang et al. Economic Modelling 63 (2017) 215225

risks (e.g., monitoring, value-added service, etc.) aect the invested 3. Research hypothesis
startups control structures (Terry and Frye, 2009; Bengtsson and
Sensoy, 2011). On the other hand, Drees et al. (2013) nd that venture 3.1. Venture capitalists and startup control structures
capitalists monitoring of startups aects the control rights they acquire
when analyzing the relationship between corporate ownership struc- With regard to risk attitude, venture capitalists risk preference is
ture and corporate values. Hopp and Lukas (2014) reveal that venture negatively related to the demand for control rights. Kaplan and
capitalists supervision strength/capability and forecast about potential Stromberg (2004) nd that both internal and external risks signi-
risks impact their selection of control structures when analyzing the cantly increase venture capitalists control rights in the board of
appraisal activities of startups investment stages. directors. The participants risk attitudes are a main factor in determin-
The third factor is venture capitalists venture targets. After ing control rights arrangements. Given system and environment
disclosing the black box relationship between venture capitalists conditions, venture capitalists negotiating power and risk attitudes
and entrepreneurs, Turcan (2008) nds that venture capitalists are basic elements in determining the control rights of non-human
investment targets (i.e., the strategic benets of investment) determine capital. As far as venture capitalists with dierent degrees of risk
corporate control structures. Based on the research of initial public aversion are concerned, there is discrepancy in residual claims
oering (IPO) enterprises supported by company venture capital, awarding eects and appraisal of control right values. Hence, startup
Masulis and Nahata (2009) indicate that the strategic targets of entrepreneurs must share control rights with venture capitalists to
company VC as strategic investors aect IPO enterprises allocation encourage VC investment. Empirical data from many countries show
of control rights and equity pricing, and that the complementary or that risk aversion and uncertainty are negatively related to venture
competitive relationship between investors and entrepreneurs is an- capitalists activities (the enthusiasm to participate) (Antonczyk and
other important factor determining IPO corporate control structures. Salzmann, 2012). Venture capitalists try to improve their eorts and
Wang et al. (2012) conduct an empirical analysis of VC-supported help entrepreneurs avoid risk, which increases their cost of risk
enterprises value from the perspective of strategic alliance, and aversion. Entrepreneurs must share control rights with venture capi-
conclude that venture capitalists strategic targets and the value of talists for returns (Cumming and Johan, 2007).
the startups in which they invest determine corporate control struc- Startup entrepreneurs capabilities and behaviors are greatly inu-
tures. The fourth factor is venture capitalists bargaining power. enced by asymmetry in information. Consequently, venture capitalists
Zacharakis et al. (2010) nd that the control structure is an important often begin monitoring startups after they acquire a proportion of the
contracting arrangement to negotiate conicts between venture capi- control rights. The more private information venture capitalists have of
talists and entrepreneurs, and the two parties bargaining power startup entrepreneurs, the stronger their ability to monitor entrepre-
inuences the control rights arrangement. In summary, venture neurs and the less control rights venture capitalists would have.
capitalists risk attitudes (degree of risk aversion), monitoring cost of Therefore, venture capitalists concern about startups control rights
startups, bargaining power, and the realization of strategic targets would inevitably decline. Control rights are a signicant tool of
(strategic benets) are key factors aecting startup control structures. corporate governance. The higher the venture capitalists monitoring
When it comes to entrepreneurs, the most important factor is the costs, the greater the proportion of control rights they would seek, and
amount of nancing that entrepreneurs acquire. According to Terry and the higher is the probability that venture capitalists and entrepreneurs
Frye (2009) and Krishnan et al. (2011), this nancing from venture will adopt joint control (Wang et al., 2010). An examination of the role
capitalists aects their corporate control rights. The second factor is of venture capitalists in determining the board of directors and
entrepreneurs private benets. Lim and Cu (2012) analyze the relation- corporate structures shows that the monitoring strength of venture
ship between venture capitalists and entrepreneurs based on social capitalists aects the structure of the board as well as the control rights
network and contracting characteristics, and nd that entrepreneurs (Sheu and Lin, 2007). The value-added services VC rms provide to
private benets aect cooperation between the two parties as well as startups inuence the allocation of enterprise control rights (decision
control structures. Drees et al. (2013) reveal that corporate value is an rights). When venture capitalists and entrepreneurs have a good long-
important signal aecting corporate control structures. Meanwhile, new term relationship, it greatly reduces venture capitalists monitoring
stockholders private benets have also been found to have an important costs. Venture capitalists may allocate the major decision-making
impact on control rights. Correspondingly, entrepreneurial nancing power to entrepreneurs (Eldridge, 2007), and this may even lead to a
constraints and private benets obtained from operating enterprises are situation where the enterprise is controlled solely by the entrepreneur.
key factors aecting startup control structures. Empirical research indicates that with a xed investment amount,
With regard to enterprises, startups value is a key factor determin- experienced VC rms will rely on stricter monitoring of startups than
ing control structures. Baldenius and Meng (2010) analyze the impact those that are less experienced. This is especially true for large
of startups value on control structures from the viewpoint of venture investment and high-risk industries (Krishnan et al., 2011). Based on
capitalists signal eect. Heughebaert and Manigart (2012) indicate the analysis of corporate governance structure after IPO, it has been
that investors bargaining power aects the appraisal of corporate value found that the monitoring of a VC-backed company is stronger than
and conguration of control rights. Byun et al. (2013) believe that a that of one not backed by VC, and the monitoring costs and eciency of
control rights-centric corporate ownership structure is an eective way capable venture capitalists are higher than those that are less capable.
for investors to implement monitoring via the board of directors; the Under such circumstances, venture capitalists often hold more rights
control structure and the monitoring strength of the board of directors (equity) in corporate governance structure (Terry and Frye, 2009).
are interlocked with corporate value. Meanwhile, the control rights and cash ow governance structure
Although the above literature review covers a discussion of the inuence venture capitalists monitoring eciency and corporate
factors aecting startup corporate control structures from dierent values (Byun et al., 2013). If venture capitalists cost to control startups
angles and helps understand how the startup corporate control and monitor entrepreneurs is high, both parties will prefer joint
structure evolves, it does not oer a systematic analysis and empirical control.
research of the factors that inuence it. Under a VC model, the Apart from providing investment, a venture capitalists most
characteristics of venture capitalists, entrepreneurs, and startups are signicant contribution to a startup is value added and risk manage-
key to determining the control structure in a startup. In this paper, we ment and control. Their non-monetary strategic value is as important
put forward hypotheses on the factors that inuence startup control as the monetary value they bring to the table (Wang et al., 2012). The
structures from the perspectives of venture capitalists, entrepreneurs, resource complementarity between venture capitalists and entrepre-
and enterprises, and test them empirically. neurs promotes social interaction and also reduces entrepreneurs need

L. Wang et al. Economic Modelling 63 (2017) 215225

for self-protection. Entrepreneurs will be willing to share corporate 2008). The two conditions under which joint control is strictly better
control rights with venture capitalists that can promote social interac- than contingent control are as follows. The rst one is that the residual
tion and enhance strategic and nancial benets. The stronger the share from renegotiations accruing to venture capitalists under joint
complementarities between investors and entrepreneurs, the more control is greater than under contingent control. The other is that the
corporate rights the venture capitalists possess and the greater the total investment of venture capitalists (i.e., startup entrepreneurs
degree of social interaction. Thus, startups stand to gain more nancing amount) and venture capitalists cost to monitor and manage
intellectual strategic benets from venture capitalists. One important startups is greater than the expected returns from the initial control
benet of a venture capitalist is the realization of strategic targets such rights of venture capitalists under contingent control (Wang et al.,
as portfolio investment, risk diversication, technology window, and 2009). For these two conditions to be met, entrepreneurs should have a
scale economies from the investment (Weber and Weber, 2005). To high amount of nancing, and venture capitalists costs of monitoring
achieve strategic benets, investors tend to form a strategic alliance and managing the enterprises should also be high. If there is a big
with entrepreneurs to take joint control of startups (Wang et al., 2012). nancing gap for entrepreneurs, namely, their nancing constraints are
VC rms strategic targets have an inuence on control right claims and high, it would be necessary for entrepreneurs to transfer a larger
equity pricing (Masulis and Nahata, 2009). Venture capitalists in a proportion of the control rights to investors, increasing the probability
strategic alliance have a value-added function, and resources and of joint control by venture capitalists and entrepreneurs. On the other
information are important factors aecting the alliances power dis- hand, if startups nancing constraints are higher, namely, entrepre-
tribution (Lindsey, 2008). When the strategic benets venture capital- neurs need a larger amount of nancing from venture capitalists,
ists may gain from an investment far outweigh its possible monetary startups are better o selecting joint control than contingent control.
benets, or when they attach more importance to the strategic benets We are also interested in whether entrepreneurs private benets
of investment, their concerns about control rights will be weaker. On are an important factor aecting the mode of startup control and
the contrary, when the strategic benets that may accrue to them from governance (Helwege and Packer, 2007). With nancial constraints,
an investment are far less than the possible monetary benets, their entrepreneurs have to nd a balance between equity and control rights.
concerns over control rights will be stronger and both parties are more When the private benets they derive from control rights are lower and
likely to select joint control (Turcan, 2008). the venture capitalists negotiation power is stronger, they are more
With regard to partnerships, the bargaining power-based sharing of likely to acquire control rights and there will be a more frequent
control rights and design of incentive contracts are lower than the substitution of startup operators (Hellman, 1998; Wang et al., 2010).
bargaining power that calls for more eorts from the partners. That is, Entrepreneurs private benets are an important factor aecting the
constraints on bargaining power can be more easily satised. The eciency of investment. High private benets from control rights will
greater the bargaining power held by venture capitalists, the more encourage entrepreneurs to take risks while venture capitalists will
likely they are to hold control rights, and the more likely it is that the demand more corporate control and even change the management to
management will be altered. Under joint control, the expected return to avoid and even stop entrepreneurs from taking risks (Lim and Cu,
venture capitalists grows with an increase in the liquidation value of 2012). A VC rm must congure control rights and income rights
the enterprise and in bargaining power. Therefore, only when venture between investors and entrepreneurs to balance the two sides nego-
capitalists have strong bargaining power in renegotiations, namely, tiating power so there is an incentive to invest. When entrepreneurs are
when they acquire a greater share from renegotiations, is selecting joint less likely to derive private benets, unilateral control by the party with
control benecial to both parties (Wang et al., 2009). When venture the larger share of benets or contingent control oers optimal power
capitalists monitoring costs, risk aversion costs, and enterprise liqui- allocation; when entrepreneurs are very likely to derive private
dation value are high, and when cash ow added value and venture benets, the optimal selection is to share control rights. Based on the
capitalists strategic benets and the probability of the projects success above analysis, we have the following hypothesis:
is small, they will have a small residual share in renegotiations under
contingent control. Hypothesis 2:. The tighter a startup entrepreneurs nancial
To sum up, when venture capitalists continuous monitoring and constraints (i.e., the greater the nancing the entrepreneur seeks
management costs are high, it is more dicult for them to realize their from venture capitalists), the greater the private benets of control
strategic targets, their risk aversion cost is high, and it is dicult to rights, the more likely both parties are to choose joint control.
persuade them to intervene. Under such circumstances, the venture
capitalists or entrepreneurs unilateral control is infeasible. The only 3.3. Venture Capitalbacked Enterprise Value and control structure
eective way is for them to enter into joint control. Thus, we have the
following hypothesis: When a startups market liquidation value is high, selecting joint
control of startups is feasible for venture capitalists and entrepreneurs.
Hypothesis 1:. The greater the venture capitalists risk aversion, the Enterprise value is the sum of enterprise performance, venture
higher the supervision and control cost, the stronger is their capitalists strategic benets, and entrepreneurs private benets
bargaining power, the less likely it is that strategic targets are (Wang et al., 2009), and is also an important signal of ability that
realized, the more likely both parties are to choose joint control. entrepreneurs are sending out to venture capitalists. Therefore, with an
increase in enterprise value, venture capitalists corresponding mone-
3.2. Entrepreneur and startup control structure tary and strategic benets will also increase. Venture capitalists as
prot-seeking participants will inevitably require more control rights to
According to Tirole (2001), startup entrepreneurs must transfer enhance their benets. At the same time, entrepreneurs monetary
partial control rights to venture capitalists to ensure the success of benets will grow with an increase in enterprise value, and their private
innovative project nancing. If entrepreneurs own more private wealth benets may also increase greatly. Venture capitalists will need to
and the amount of nancing they demand is less, then VC investment is increase their control rights to oversee and prevent entrepreneurs from
less urgent and important for entrepreneurs. The control rights that harming their own interests. Thus, it is more likely that joint control
entrepreneurs are willing to transfer to VCs are fewer. On the contrary, will be chosen. On the other hand, an entrepreneurs value is an
when entrepreneurs own less wealth, they demand more investment important signal aecting the cooperation between venture capitalists
from venture capitalists. This means that VC investment is more urgent and entrepreneurs (Baldenius and Meng, 2010). The stronger a venture
and important for entrepreneurs. Accordingly, the control rights that capitalists bargaining power, the lower is the appraisal of startups,
entrepreneurs are willing to transfer to investors are greater (Tan et al., which will lead to entrepreneurs being in a disadvantageous position

L. Wang et al. Economic Modelling 63 (2017) 215225

for negotiations on control rights. Entrepreneurs would choose to share Table 1

control rights with venture capitalists and establish a joint control Classifications of the Sample Enterprises.
structure (Heughebaert and Manigart, 2012). Meanwhile, if one partys
Classication Number of Percentage of total
bargaining power is too strong, it will bring conicts during coopera- firms observations
tion between venture capitalists and entrepreneurs. In this case, both
sides choosing joint control of enterprises will be the best way to Industry IT industry 30 15.544
sector Communications and 32 16.580
negotiate conicts (Zacharakis et al., 2012). Based on the above
Intelligent services
analysis, we now have the following hypothesis: Materials, Energy and 35 18.135
Environmental protection
Hypothesis 3:. The higher the enterprise value, the more likely it is Biomedical health care 24 12.435
that venture capitalists and entrepreneurs will choose joint control. General manufacturing 26 13.472
Financial services 21 10.881
Other industries 25 12.953
4. Empirical research design
Size Small-size enterprises 85 44.041
Medium-size enterprises 63 32.642
4.1. Questionnaire design and sample selection Large-size enterprises 45 23.316
Type Non-state-owned 147 76.166
In China, there is no public database of startup information, enterprises
State-owned enterprises 46 23.834
making it dicult to collect data on control structures. Following Region Regional enterprises 132 68.393
Kaplan and Stromberg (2004), we use survey methods to gather data Cross-regional investments 61 31.606
from professional VC rms on invested startups.

4.1.1. Questionnaire design

We prepare and pre-test our rst questionnaire by referring to 193 valid enterprise samples in two phases, 121 of which were high-
existing research and deleting some questions. Our formal questionnaire tech and 72 were from traditional industries. Table 1 reports the
come after a second round of modications based on issues in the pre- detailed description of sample enterprises.
test. The questionnaire adopts the Likert scale and the numbers 15
denote the following: 1completely insignicant (or completely unim-
4.2. Variable selection
portant), 2quite insignicant (or very unimportant), 3general, 4
signicant (or important), and 5very signicant (or very important).
4.2.1. Dependent variable
Since this paper is to examine factors aecting startups control
4.1.2. Large-sample questionnaire
structures, the dependent variable is a measure of the control structure
In this paper, the industries covered by our large-sample ques-
type. While there are many taxonomic studies on startup control
tionnaire include emerging industries such as general high-tech and
structures, this paper adopts a dichotomy method that is widely used
the Internet of Things, the Internet, and a part of the general
in empirical research, and classies control structures as contingent
processing and manufacturing industries as well as traditional indus-
control and joint control. Therefore, this paper utilizes a binary variable
tries such as restaurant chains, education, and training. Samples were
JC to represent the type of a control structure. If joint control is
collected from four cities in Eastern China (Jinan in Shandong
involved in the cooperation between venture capitalists and entrepre-
Province, Shenzhen in Guangdong Province, and Nanjing and Wuxi
neurs, JC=1; otherwise, JC=0.
in Jiangsu Province), Central Chinas Hefei in Anhui Province, and also
from Western Chinas Xian in Shaanxi Province and Wuzhou in
Guangxi Province. We contacted 20 dierent types of VC rms via 4.2.2. Explanatory variables
VC associations in Jinan, Shenzhen, Nanjing, and Xian and the Science Based on our research hypothesis, we empirically examine seven
and Technology Administration Departments in Wuxi, Hefei, and factors, including entrepreneurs non-monetary benets, venture capi-
Wuzhou. A preliminary interview was conducted of about 30 high- talists bargaining power, non-monetary (strategic) benets, degree of
level managers, VC managers, and VC project leaders of these 20 VC risk aversion, monitoring cost, investment amount, as well as enter-
rms, before a part of the questionnaire was modied. This modied prise value.
version was sent out by hard copy and e-mailed to senior management
sta of 20 VC rms in seven cities, inviting them to complete the Entrepreneurs private non-monetary benets. According to
questionnaires and provide feedback about the startups they had Jensen and Meckling (1976), managers can enjoy benets such as
invested in. perquisite consumption and reputation by their corporate control
The survey is divided into two phases. The rst phase was rights. However, Mueller (1989) points out that non-monetary
completed between July and September 2009. During Phase 1, 450 benets and monetary benets are equally important in private
questionnaires were sent out and 205 were collected. Among the benets derived from control rights, which include social status and
collected questionnaires, 52 were eliminated owing to non-cooperation, reputation. Reputational benet is especially important for enterprise
startup liquidation, and other reasons, 31 enterprises fell short of the founders or family businesses. Considering the complexity of private
startup denition or job positions, and 15 were found to be seriously benets derived from control rights, Ehrhardt and Nowak (2015)
lacking in information. That left 107 valid questionnaires that met the believes these benets should contain reputational benets in
basic requirements of large-sample data analysis. The collection rate addition to monetary and non-monetary benets through the
was 45.56%, and the valid collection rate, 23.78%. The second phase tunneling behavior. This paper measures entrepreneurs non-
was completed between July and December 2011. During this phase, a monetary benets (the variable PB) from four aspects: social status,
return visit after two years was paid to 107 valid sample enterprises to perquisite consumption, occupational reputation, and extent of peer
collect corresponding data for the second time. Our follow-up of 107 recognition.
enterprises showed that VC rms withdrew from 12 enterprises, nine
were declared bankrupt, and 15 were in their second or third rounds of
investment. That left 71 enterprises from which VCs had not with- Venture capitalists bargaining power. Venture capitalists
drawn. The nal valid sample had 86 enterprises. There were a total of bargaining power (Zhao et al., 2012) mainly represents their ability to

L. Wang et al. Economic Modelling 63 (2017) 215225

control through cooperation and the benets of such cooperation Therefore, this paper sets corporate factors as the control variables,
(Heughebaert and Manigart, 2012). In this paper, it refers to venture which help improve the prediction stability of our empirical model. In
capitalists ability to claim a share from renegotiations. It (the variable line with the analysis of enterprise attributes in the literature, this
Bargaining) is measured by the residual claims acquired by venture paper controls enterprise size (the variable Size), industry (the variable
capitalists renegotiations (equity), and is calculated by the dierence Industry), regional dierence (the variable Regionality), and the
in equity before and after negotiations. development stage (the variable Stage). The purpose is to control the
eects of these factors at the enterprise level, and reveal how these
factors impact the control structures of startups. Given the diculties Venture capitalists non-monetary (strategic) of collecting startup data by time sequence, we use startup samples at
benets. Monetary benet is an important goal of VC rms. In dierent stages instead of dynamic sample data. The meaning and
addition, research reveals that VC rms make investments based on measurement of our control variables is presented in Table 3.
investment portfolio, risk diversication, opportunities for technology
development and utilization, patent pool eect, upstream and
downstream industry support, development of new markets,
4.3. Validity and reliability analysis
diversied development, economies of scale, and economies of scope
(Weber and Weber, 2005). The strategic targets of VC rms should be
Table 4 reports the Cronbach coecients of the factors involved,
linked to specic startups business activities, core competence, market
which are all greater than 0.7. This implies that the indicators adopted
strategy, growth potential, competitive advantages, resource
in this paper have good reliability. Our validity analysis is divided into
utilization, and industrial functions (Wang et al., 2012). Thus, this
three steps. First, we conduct an exploratory factor analysis during
paper proposes to measure VC rms non-monetary benets (the
index simplication and single dimension check, which ensures no
variable SB) from seven aspects: technology development, eciency
cross-loading of the indices. Second, the factor loads used to calculate
of technology platform utilization, upstream and downstream industry
all indices reported in Table 4 are signicant (P-value < 0.01). Each
support, developing new markets, diversied development, economies
factor load is greater than 0.6, representing factors with good con-
of scale, and economies of scope.
vergent validity. Third, the accumulated variance contribution rate for
each factor reported in Table 4 is greater than 60%, representing
factors with good convergent validity. Venture capitalists risk aversion. Li and Guo (2009) nd that
risk attitudes can be measured by the degree of risk aversion. Chen
(2009) analyzes the factors aecting risk attitudes and their impacts on
startup behaviors. With reference to the measurement of risk attitudes, 5. Empirical analysis
this paper proposes two measureswillingness to invest and the
strength of preference for benets from risk as an indication of risk 5.1. Econometric model
aversion (the variable RA). This method employs the dened standards
of investor risk attitudes in Chinas VC industry using the China The key issue in our empirical research is to examine the relative
Venture Capital Research Institute questionnaire, and also refers to probability of factors aecting startups control structure. As control is
Geert Hofstede and K.E. Warners uncertainty-risk aversion classied as joint control and contingent control, our dependent
measurement from ve cultural dimensions. variable JC is a binary variable that takes the value 0 or 1. The
dependent variable can only take the value 0 or 1 to indicate that an
To ensure the reliability and validity of measurement variables, the event occurs or does not occur. Hence, the relationship between
measurement of VC rms non-monetary benets, strategic benets, independent variables and the dependent variable is nonlinear. In this
and degree of risk aversion in this paper mainly refers to the measures case, if a linear probability model (for example, OLS and WLS) of
used in the literature, with modications as needed. Table 2 presents regression analysis is employed, it may lead to a false conclusion.
the measures of these variables. The measurement of other explanatory Therefore, this paper uses a logistic regression model.
variables is based on appropriate statistical indicators, and their Logistic regression is applied when the dependent variable is a
specic meaning and measurement methods are shown in Table 3. binary variable, and uses maximum likelihood to estimate a nonlinear
model. In this paper, the dependent variable JC takes a value 0 or 1,
representing the probability of selecting joint control. Compared with
4.2.3. Control variable contingent control, the probability of startups selecting joint control is:
Previous studies show that control structures are at the core of
corporate governance, the selection of which also depends on enter- 1
p= ,
prises own resources and capabilities. Thus, the eect of corporate (1 + exp(JC )) (1)
factors (especially corporate resources and capabilities) must be
considered in empirical research. where

Table 2
Explanatory variables.

Variable Measures Sources of Reference

Enterprises non-monetary benets Social status, reputation, perquisite consumption, degree of peer approval Jensen and Meckling (1976),
(PB) Helwege and Packer (2009)
Venture capitalists strategic Technology development window, utilizing internal technology platform, promoting internal Ernst and Young (2002), Weber and
benets (SB) technology innovation, commercialization of high-tech achievements, business diversification, Weber (2005)
developing new market, economy of scope
Venture capitalists degree of risk Risk attitude, investment choice willingness, risk-return preference Jaeger et al. (2010), Chen (2009), Li
aversion (RA) and Guo (2009)

L. Wang et al. Economic Modelling 63 (2017) 215225

Table 3
Meaning and measurement method of variables.

Variable description Variable meaning Measurement method

Dependent variable JC Probability of choosing joint control Select joint control=1; otherwise, =0
Independent variables Bargaining The residual claims that a venture capitalist obtains from re- Calculated by the difference in equity value before and after a venture
negotiations (equity) capitalist possesses equity
Monitoring The costs that a venture capitalist pays to monitor startups Calculated by the number of times a venture capitalist checks enterprise
finance reports every year
RA The cost to venture capitalists for evading risk Likert Scale
SB Non-monetary benefits accruing to venture capitalists from a Likert Scale
PB Non-monetary benefits accruing to entrepreneurs from a Likert Scale
Investment Venture capitalists investment amount in a startup In tens of thousands of RMB
Value Preliminary evaluation of a startups market value In tens of thousands of RMB
Control variables Size Measured by the number of employees in a startup 1: < 200 employees; 2: 2001000 employees; 3: above 1000 employees
Industry The industry that a startup belongs to 1: high-tech industries; 0: traditional industries
Regionality The regions in which a VC firm and a startup are located 1: both parties located in the same region; 0: parties located in different
Stage The development stage of a startup 1: early stage; 2: growth stage; 3: maturity stage; 4: others

Table 4
Questionnaires reliability and convergent validity.

Scale coefficient Factor load Percentage of variance explained

Entrepreneurs non-monetary benets (PB) 0.799 64.51

1. High social status of startups 0.796
2. Excellent perquisite consumption that entrepreneurs enjoy 0.819
3. Scope for improving entrepreneurs occupational reputation 0.802
4. Entrepreneurs peer approval 0.750
Venture capitalists non-monetary benets (SB) 0.808 72.69
1. Technology development opportunities for VCs other invested enterprises 0.845
2. Improving the technology platform utilization eciency for VCs other invested enterprises 0.882
3. Providing upstream and downstream industry support for VCs other invested enterprises 0.786
4. Being able to develop new markets for VC 0.668
5. Being able to realize startups diversied development strategy for VC 0.851
6. Being able to realize scale economies for VC 0.767
7. Being able to realize scope economies for VC 0.810
Venture capitalists degree of risk aversion (RA) 0.719 61.58
1. Which risk category do you belong to? 0.727
2. Which of the following best describes your private investment behavior 0.716
3. If you bear higher risks, you will earn higher returns 0.802

Table 5
p Descriptive statistical analysis of variables.
JC = ln 1 p = 0 + 1 Bargainingit + 2 Monitoringit + 3 RAit + 4 SBit

Variable N Minimum Maximum Mean Std. Deviation
+ 5 PBit
JC 193 0.000 1.000 0.581 0.502
+ 6 Investmentit + 7 Valueit + 8 Sizeit + 9 Industryit + 10 Regionalityit Bargaining 193 0.000 0.150 0.039 0.046
+ 11 Stageit + it Monitoring 193 1.000 5.000 3.268 1.082
RA 193 1.000 5.000 2.785 1.021
(2) SB 193 1.000 5.000 3.103 1.068
PB 193 1.000 5.000 3.558 1.062
The denitions of the variables in Eq. (2) are shown in Table 3. Investment 193 150.000 10000.000 1035.161 1878.304
Value 193 800.000 100000.000 16264.516 19669.970
Size 193 1.000 3.000 1.710 0.643
Industry 193 1.000 2.000 1.419 0.502
Regionality 193 0.000 1.000 0.548 0.506
Stage 193 1.000 4.000 2.161 1.068
5.2. Descriptive statistics and correlation analysis
Note: Variable definitions are in Table 3.
After completing index purication and reliability/validity analysis,
we measure the mean value of dierent indices of a factor to reect the
measured factor. We conduct a descriptive statistical analysis of the LogitP = JC = 0 + 1 Bargainingit + 2 Monitoringit + 3 RAit + 4 SBit + 5 PBit
variables of logistic regression analysis model to nd the extreme value
and make the conversion. The mean value and range of each variable + 6 lg Investmentit + 7 lg Valueit + 8 Sizeit + 9 Industryit + 10 Regionalityit
are reported in Table 5. + 11 Stageit + it ,
According to Table 5, the distribution of nancing and startup value
variables is quite asymmetric, and the average value of each variable is
far greater than its mean value. Therefore, we make a logarithmic where lg Investmentit = log(1 + Investmentit ) and lg Valueit = log(1 + Valueit ).
conversion to eliminate the impact of extreme values. The logistic The correlation coecient and AVE (average variance extracted)
regression model after logarithmic conversion is adjusted as below: values of variables, after log conversion and model modication, are

L. Wang et al. Economic Modelling 63 (2017) 215225

Table 6
Correlation coefficient and AVE values.

JC Bargaining Monitoring RA SB PB Investment Value Size Industry Regionality Stage

JC 0.69
Bargaining 0.509** 0.71
Monitoring 0.674** 0.335 0.68
RA 0.735** 0.326 0.497** 0.75
SB -0.709** -0.363* -0.496** -0.621** 0.76
PB 0.736** 0.242 0.603** 0.496** -0.458** 0.70
Investment -0.183 -0.069 -0.115 -0.319 0.088 -0.426* 0.68
Value -0.470** -0.106 -0.383* -0.290 0.253 -0.626** 0.494** 0.72
Size 0.017 0.094 -0.274 0.037 -0.049 -0.136 0.096 -0.017 0.75
Industry -0.367* -0.065 -0.083 -0.130 0.023 -0.317 0.313 0.367* -0.231 0.72
Regionality 0.537** 0.442* 0.662** 0.476** -0.564** 0.362* 0.164 -0.168 -0.190 -0.020 0.74
Stage 0.394* 0.305 0.627** 0.382* -0.495** 0.231 0.095 -0.106 -0.346 0.088 0.808** 0.71

Note: Variable definitions are in Table 3. The numbers in parentheses are the AVE (average variance extracted) values of each variable. *, **, and *** indicate signicance at 10%, 5%,
and 1%, respectively.

reported in Table 6. The results show that a correlation between factors particularly interested in investing in high-tech enterprises, they also
and conrm that the discrimination of factor design is good. The data invest greatly in traditional industries, such as general manufacturing,
distribution reected in average values and standard deviations is in nancial services, restaurant chains, and educational training. Thus, we
line with the normal distribution, which provides a good basis for the analyze high-tech enterprises and traditional enterprises separately to
next step of data analysis. see whether and how the selection of control structures diers in high-
It is noteworthy that the correlation coecients reect the inter- tech industries and in traditional industries.
relationship of variables. However, the correlation coecient often
reects the comprehensive function of two variables in multiple ways. 5.3.1. Full-sample empirical analysis
Thus, the positive/negative value and signicance level of correlation The full-sample-sized logistic regression results of 193 startups
coecients can only be a reference for nal analysis, without too much control structures are reported in Table 7. The regression coecient of
mandatory meaning. The correlation coecients of some variables are joint control in the model is relative to the default contingent control.
greater than 0.6, so we further test the discriminate validity of Model M1 considers only the impacts of the control variables (size,
variables. The most widely used method to test discriminate validity industry, regional dierences, and development stage) on startup
is to examine whether the AVE values of all variables are greater than control structures. The results indicate that enterprises industry
the squared value of the correlation coecient between variables control variable has a signicantly positive coecient, indicating that
(Shook et al., 2004). From Table 6, we note that the AVE values of high-tech startups are more likely to choose joint control. With regard
the 12 variables in this study are between 0.68 and 0.76, with a to model M2, the impacts of the four variables of bargaining power,
minimum of 0.68, while the correlation coecient among variables is monitoring cost, degree of risk aversion, and investment amount are in
between 0.017 and 0.808, with a maximum of 0.808. Therefore, the line with our expectations. All the regression coecients are positive
maximum squared value of correlation coecient among variables is and signicant at the 1% or 5% level. The regression coecient of
0.6529 (0.8080.808), smaller than the minimum AVE value of 0.68. venture capitalists strategic benet variable is negative (the estimated
Thus, the AVE values of all variables are greater than the squared value coecient is -0.092), which is consistent with expectations, but fails
of the correlation coecient. Consequently, all the variables have good the signicance test. In model M3, the regression coecient of the
discriminate validity. venture capitalists strategic benet variable is still negative but
The logistic model should be adopted for specic regression
analysis so as to further reveal the impact of factors on startup control Table 7
structures. This paper adopts the logistic regression model in SPSS16.0 Full-sample multivariate logistic regression analysis of control structures.
software, and uses a stepwise regression analysis to test the relation-
ship between startup control structures and their explanatory variables. M1 M2 M3

Coecient Wald Coefficient Wald Coefficient Wald

5.3. Model estimation and statistical analysis
Size 0.535 0.425 0.632 0.572 0.487 0.368
Industry 2.078** 5.074 1.363* 3.882 1.505** 4.129
To test our hypotheses, we rst run the regression with the full Regionality -0.362 0.154 -0.208 0.238 -0.417 0.316
sample. Then, we divide the full sample into two groups: high-tech Stage -0.618 1.947 -0.582 1.791 -0.733* 3.867
enterprises and enterprises in traditional industries, and run the Bargaining 0.948*** 7.817 1.606*** 11.318
Monitoring 0.198** 4.234 0.276** 5.585
regressions with these two subsamples separately. We do this for three
RA 0.196** 4.471 0.166** 3.907
reasons. First, the risk to a venture capitalist investing in high-tech SB -0.092 1.831 -0.121 1.964
enterprises is much higher than that in traditional industries. These lg_Investment 0.413** 60.250 0.635*** 9.170
risks aect their monitoring costs. Venture capitalists risk attitudes PB 0.727*** 10.737
and monitoring costs will aect their demand for the protection of lg_Value -0.186 0.132
Constant 0.399 5.681 2.976 10.533 3.138 10.219
control rights to avoid risk (Wang et al., 2009). Second, the private -2 Log 32.963 20.577 19.339
benets from control rights and the strategic benets from VC likelihood
investment in high-tech enterprises and other startups in traditional Cox & Snell R 0.257 0.502 0.521
industries are dierent. These dierent benets of control rights have a Square
Nagelkerke R 0.346 0.675 0.701
varied impact on the mechanism and extent of risk protection. This
dierence will aect various participants demand for control rights
and bargaining power, thus aecting the two parties selection of Note: Variable definitions are in Table 3. *, **, and *** indicate signicance at 10%, 5%,
control structures. Third, in China, while venture capitalists are and 1%, respectively.

L. Wang et al. Economic Modelling 63 (2017) 215225

Table 8 Table 9
High-tech industry samples multivariate logistic regression analysis of control Traditional industry samples multivariate logistic regression analysis of control
structures. structures.

M1(a) M2(a) M3(a) M1(b) M2(b) M3(b)

Coecient Wald Coefficient Wald Coefficient Wald Coecient Wald Coefficient Wald Coefficient Wald

Size 0.726 0.698 0.952 0.827 0.738 0.813 Size 0.437 0.351 0.514 0.46- 0.329 0.279
Regionality -0.643 0.357 -0.576 0.459 -0.817 0.535 Regionality -0.278 0.168 -0.325 0.375 -0.313 0.326
Stage -0.889 2.086 -0.722** 3.890 -0.969** 4.181 Stage -0.539 1.114 -0.483 1.230 -0.644 2.160
Bargaining 1.585*** 12.157 1.866*** 13.396 Bargaining 0.719** 5.233 1.022** 6.459
Monitoring 0.252** 5.129 0.214** 4.973 Monitoring 0.134 2.771 0.159** 3.867
RA 0.349*** 7.630 0.283*** 7.112 RA 0.145 2.316 0.126 3.229
SB -0.377** 5.278 -0.591** 6.362 SB -0.072 0.089 -0.097 0.091
lg_Investment 0.628*** 5.651 0.802*** 7.570 lg_Investment -0.299** 4.353 -0.431** 5.979
PB 0.973*** 11.23 PB 0.525*** 7.638
lg_Value -1.125*** 10.861 lg_Value -0.618** 4.073
Constant 0.409 5.908 2.997 10.825 3.356 10.762 Constant 0.291 4.415 2.176 9.933 2.989 10.01
-2 Log 33.193 20.282 19.116 -2 Log likelihood 33.579 20.798 19.686
likelihood Cox & Snell R 0.221 0.432 0.464
Cox & Snell R 0.262 0.537 0.551 Square
Square Nagelkerke R 0.317 0.494 0.518
Nagelkerke R 0.383 0.709 0.727 Square
Note: Variable definitions are in Table 3. *, **, and *** indicate signicance at 10%, 5%,
Note: Variable definitions are in Table 3. *, **, and *** indicate signicance at 10%, 5%, and 1%, respectively.
and 1%, respectively.
eect on the probability of high-tech venture enterprises selecting joint

insignicant, while the regression coecient of startups private

benets is positive, consistent with expectations, and signicant at 5.3.3. Empirical analysis of subsamples in traditional industries
the 1% level. The regression coecient of startups value variable is Table 9 reports the results of the multivariate logistic regression
negative, contrary to expectations, and fails the signicance test. with 72 enterprises in traditional industries. In M1(b), control vari-
Our analysis of the full sample shows that there is discrepancy in ables do not seem to have a signicant impact on the selection of joint
enterprise control structures. The probability of high-tech startups control. In M2(b), the impacts of the three variables of bargaining
selection of joint control is greater. Entrepreneurs private benets, power, monitoring cost, and degree of risk aversion are in line with
venture capitalists bargaining power, investment amount, degree of hypotheses. All regression coecients are positive, of which venture
risk aversion, and monitoring cost exhibit a signicantly positive capitalists bargaining power is signicant at the 5% level, while the
relationship with the probability of startups choice of joint control. monitoring cost and risk aversion variables fail the signicance test.
Enterprise value and venture capitalists strategic benets exhibit an The estimated coecient of the venture capitalists strategic benet
insignicantly negative relationship with the probability of selecting variable is negative, consistent with hypotheses, but insignicant. The
joint control. coecient of the venture capitalists investment amount variable is
negative, contrary to expectations, and signicant at the 5% level. In
5.3.2. Empirical analysis of subsamples in high-tech industries M3(b), the eect of the entrepreneurs monitoring cost variable
Table 8 reports the empirical results of the multivariate logistic strengthens, with its regression coecient positive and signicant at
regression with 121 startups in high-tech industries. In M1(a), the the 5% level. The regression coecient of venture entrepreneurs
control variables fail to show a signicant impact. In M2(a), the private benet is positive, consistent with expectations, and signicant
regression coecient of the enterprise development stage control at the 1% level. The regression coecient of the enterprise value
variable is negative and signicant at the 5% level. The impacts of variable is negative, consistent with expectation, and signicant at the
the four variables of bargaining power, monitoring cost, degree of risk 5% level.
aversion, and investment amount are in line with expectations. All Our analysis shows that entrepreneurs private benets, and
these estimated coecients are positive, of which venture capitalists venture capitalists bargaining power, and monitoring costs all have a
bargaining power, degree of risk aversion, and investment amount signicantly positive relationship with the probability of startups in
variables are signicant at the 1% level, while the coecient of the traditional industries selecting joint control. Venture capitalists in-
monitoring cost variable is signicant at the 5% level. The venture vestments and enterprise value have a signicantly negative relation-
capitalists strategic benet variable is negative and signicant at the ship with the probability of selection of joint control by startups in
5% level, which is consistent with our expectation. In M3(a), the traditional industries. Venture capitalists degree of risk aversion and
coecient of entrepreneurs private benets is positive, consistent with strategic benets show an insignicant relationship with the prob-
the expectations, and signicant at the 1% level, while that of the ability of startups in traditional industries selecting joint control.
enterprise value variable is negative, contrary to expectations, and An enterprises development stage, degree of risk aversion, and
signicant at the 1% level. strategic benets variables all have a signicant eect on the control
Our analysis shows that the probability of high-tech startups at an structures of high-tech startups, which is consistent with our hypoth-
early development stage selecting joint control is higher, while that of esis. However, for startups in traditional industries, the impacts of
high-tech startups at a later development stage selecting contingent these three factors are not statistically signicant. The degree of risk
control is higher. Entrepreneurs private benets, venture capitalists aversion reects venture capitalists uncertainty of the benets of the
bargaining power, investment amount, degree of risk aversion, and enterprises in which they have invested. The degree of risk of startups
monitoring cost pose a signicantly positive eect on the probability of in high-tech industries is high, while that of startups in traditional
high-tech venture enterprises selecting joint control. Enterprise value industries is relatively low. Therefore, the degree of risk aversion in
and venture capitalists strategic benets pose a signicantly negative high-tech industries has a greater impact on control rights, while it has

L. Wang et al. Economic Modelling 63 (2017) 215225

a lower impact in traditional industries. In addition, the success of traditional industries is not signicant. The heterogeneity of industries
venture capitalists investments in high-tech startups brings greater causes great discrepancy in venture capitalists strategic benets. The
reputation, which means that the strategic benets in high-tech strategic benet targets between corporate venture capital and general
industries are more signicant than those in traditional industries. venture capital are dierent. Corporate venture capital investments are
Furthermore, as far as the development stage of enterprises is to achieve strategic objectives, such as industry support, economies of
concerned, since the returns in high-tech industries is particularly scope, and economies of scale, while the main purpose of general
uncertain, venture capitalists tend to choose joint control of enterprises venture capital investment is to obtain nancial benets. Again,
that are at an early stage of development to eectively control and technological uncertainty and human capitals high heterogeneity, high
supervise startups. The coecient of startups development stage in specicity, and unpledged property in high-tech enterprises cause the
traditional industries is still negative, but not signicant, indicating investment risk in high-tech enterprises to be much higher than that in
that the development stage of startups in traditional industries is not traditional enterprises. Therefore, the degree of risk aversion of venture
an important factor for venture capitalists. The impact of investment capitalists investing in high-tech enterprises to hedge risk is far higher
(enterprises nancing amount) on enterprises in traditional industries than that in traditional enterprises, leading to a more pronounced
is contrary to our hypothesis and signicant. This can be due to the impact of the degree of VCs risk aversion in high-tech enterprise
following two reasons. On one hand, it may be caused by the analysis.
dierences in the scale of investment needed in dierent industries. Inspired by de Bettignies (2008) who divides control structures into
Namely, the overall investment amount in traditional industries is three types and Fluck (2010) and Gestone (2014) who classify control
relatively low. On the other hand, the prospects and expected benets structures into contingent control and unconditional control, in this
of startups in traditional industries are less variable. Either entrepre- paper we classify startup control structures into joint control and
neurs or venture capitalists possess the control rights, and it is easier to contingent control, from the perspective of corporate governance.
reassign control based on the changing dynamics during the growth Using a dichotomy method, our analysis shows that this classication
stages. Therefore, the more venture capitalists invest in traditional is appropriate for the analysis of this type. In contrast with Kaplan and
industries startups, the more likely they are to choose contingent Strombergs (2003, 2004) analysis of the structure and allocation of
control for the sake of exit or unilateral control. control rights in venture capital contracts, our main focus is on the way
in which various factors impact the selection of control structures in
6. Conclusion venture capital-backed startups. Schmidt (2003), Yan and Fei (2007),
and de Bettignies (2008) nd unilateral control or contingent control
This paper examines the inuence of the characteristics of venture has advantages over joint control. However, using the survey data on
capitalists and entrepreneurs on the selection of control structures in heterogeneous enterprises in China, we nd that joint control is
venture capital-backed startups. More specically, we investigate how preferred by high-tech enterprises, particularly those at an early stage
factors such as venture capitalists bargaining power, monitoring cost, of development. The implication of this nding is that joint control
nancing amount, degree of risk aversion, strategic benets, startup could be more eective, depending on the types of startups and the
value, and entrepreneurs private non-monetary benets impact the development stage of startups. In addition, we nd that entrepreneurs
control structures in Chinese startups. The main conclusions are as in the Chinese market are generally risk averse, rather than risk neutral
follows. 1) The probability of high-tech startups choosing joint control as assumed by Schmidt (2003) and de Bettignies (2008) in their
is greater. High-tech startups at an early stage of development tend to theoretical framework. On the other hand, capitalists degree of risk
choose joint control while those at later stages of development prefer aversion and bargaining power have signicant impacts on the selec-
contingent control structures. 2) Overall, venture capitalists and tion of control structures of invested enterprises. Tirole (2001) docu-
entrepreneurs prefer joint control of startups when the former have ments two types of non-pecuniary benets of control, while Masulis
stronger bargaining power and higher monitoring costs; they also tend and Nahata (2009) nd that the benets of control aect the allocation
to choose joint control when entrepreneurs' nancing amount and of control rights in enterprises. This paper conrms the existence of
private benets are greater. In traditional industries, venture capital- these two types of non-pecuniary benets of control, and further shows
ists strategic benets and the degree of risk aversion have an that their size varies in dierent enterprises. Importantly, we nd that
insignicant impact on the selection of startup control structures. In the mechanisms through which these two types of non-pecuniary
high-tech industries, the venture capitalists bargaining power, mon- benets of control impact the selection of control structures are
itoring costs, degree of risk aversion, investment amount, and strategic dierent. Our overall ndings imply that many factors can impact the
benets all have a signicant impact on the probability of startups selection of control structures in startups, and that these impacts can
selecting joint control, which is consistent with our hypotheses. 3) For be dierent in high-tech enterprises and traditional enterprises, as well
startups in traditional industries, only the eects of venture capitalists as in developed markets and emerging markets.
bargaining power and monitoring cost are consistent with our hypoth- Similar to most previous studies on enterprise control structures,
esis, while other factors such as the stage of development of startups, this analysis is conducted from the perspective of incomplete contracts.
venture capitalists degree of risk aversion, and strategic benets have However, research on startup control rights under incomplete con-
an insignicant impact on the probability of startups choosing joint tracts ignores factors such as cooperation, trust, and the reputation of
control. 4) We nd no signicant impact of enterprise value on the the two parties. Research on these issues requires the introduction of
selection of control structures. relational contract theory and human capital theory, which is beyond
Our empirical analysis indicates that factors aecting the choice of the scope of this study. However, it is an important area for future
control structures show signicant dierences between startups in research.
high-tech industries and those in traditional industries. The reasons
are as follows. First, there is a lack of an eective pricing mechanism in References
Chinas high-tech enterprises, both in theory and in practice, making it
dicult to determine a fair value for startups; hence, survey sample Andrieu, G., 2012. Optimal allocation of control rights in venture capital contracts. In:
data fail to reect the real value of startups. Therefore, the value of Proceedings of the International Conference of the French Finance Association
(AFFI). Available at SSRN: http://ssrn.com/abstract=1833417.
enterprises does not show a signicant inuence on the selection of Antonczyk, R.C., Salzmann, A.J., 2012. Venture capital and risk perception. J. Bus. Econ.
control structures in our empirical analysis. Second, while the strategic 82 (4), 389416.
eect of high-tech industries technology window and technology Baldenius, T., Meng, X., 2010. Signaling rm value to active investors. Rev. Account.
Stud. 15 (3), 584619.
platform support is noticeable, the strategic eect of investment on

L. Wang et al. Economic Modelling 63 (2017) 215225

Bengtsson, O., Sensoy, B.A., 2011. Investor abilities and nancial contracting: evidence empirical analysis of venture capital contracts. Rev. Econ. Stud. 70 (2), 281315.
from venture capital. J. Financ. Intermed. 20 (4), 477502. Kaplan, S.N., Stromberg, P., 2004. Characteristics, contracts, and actions evidence from
Byun, H.S., Ji, H.L., Park, K.S., 2013. Ownership structure, intensive board monitoring, venture capitalist analyses. J. Financ. 59 (5), 21732206.
and rm value: evidence from Korea. Asia-Pac. J. Financ. Stud. 42 (2), 191227. Krishnan, C.N.V., Ivanov, V.I., Masulis, R.W., Singh, A.K., 2011. Venture capital
Carvell, S.A., Kim, J.Y., Ma, Q., Ukhov, A.D., 2013. Economic and capital market reputation, post-IPO performance and corporate governance. J. Financ. Quant. Anal.
antecedents of venture capital commitments (19602010). Int. Entrep. Manag. J. 9 46 (5), 12951333.
(2), 167182. Li, T., Guo, J., 2009. Risk attitude and stock market investment. Econ. Res. J. (2), 5667.
Cestone, G., 2014. Venture capital meets contract theory: risky claims or formal control. Lim, K., Cu, B., 2012. The eects of social networks and contractual characteristics on the
Rev. Financ. 18 (3), 10971137. relationship between venture capitalists and entrepreneurs. Asia Pac. J. Manag. 29
Chen, B., 2009. A case study on the attitude to risks on peasant-workers returning to (3), 573596.
their hometown to start a business. Manag. World (3), 8491. Lindsey, L., 2008. Blurring rm boundaries: the role of venture capital in strategic
Chen, K., Chu, T., Billota, R., 2011. A spatial investigation of venture capital investment alliances. J. Financ. 63 (3), 11371168.
in the US biotechnology industry, 19952008. GeoJournal 76 (3), 267282. Ma, D., Rhee, M., Yang, D.Y., 2013. Power source mismatch and the eectiveness of
Chen, T.Q., Ding, S.H., He, J.M., Li, X.D., 2014. Research on control right transfer and interorganizational relations: the case of venture capital syndication. Acad. Manag. J.
incentive mechanism in risk enterprise nancing. Syst. Eng.-Theory Pract. 34 (5), 56 (3), 711734.
11451152. Masulis, R.W., Nahata, R., 2009. Financial contracting with strategic investors: evidence
Cumming, D., Johan, S.A.B., 2007. Advice and monitoring in venture nance. Financ. from corporate venture capital backed IPOs. J. Financ. Intermed. 18 (4), 599631.
Mark. Portf. Manag. 21 (1), 343. Schmidt, K.M., 2003. Convertible securities and venture capital nance. J. Financ. 58 (3),
de Bettignies, J.E., 2008. Financing the entrepreneurial venture. Manag. Sci. 54 (1), 11391166.
151156. Sheu, D.F., Lin, H.S., 2007. Impact of venture capital on board composition and
Drees, F., Mietzner, M., Schiereck, D., 2013. Eects of corporate equity ownership on ownership structure of companies: an empirical study. Int. J. Manag. 24 (3),
rm value. Rev. Manag. Sci. 7 (3), 277308. 573581.
Ehrhardt, O., Nowak, E., 2015. . Private benets of control in founding-family owned Tan, J., Zhang, W., Xia, J., 2008. Managing risk in a transitional environment: an
rms: an analysis of the dynamics of disproportionate ownership and control in exploratory study of control and incentive mechanisms of venture capital rms in
family rm IPOs (July1, 2015). Available at SSRN: https://ssrn.com/ China. J. Small Bus. Manag. 46 (2), 263285.
abstract=423506 or http://dx.doi.org/10.2139/ssrn.423506. Terry, I.I., Frye, M.B., 2009. Venture capitalist monitoring: evidence from governance
Eldridge, D., 2007. Optimal allocation of decision rights for value-adding in venture structures. Q. Rev. Econ. Financ. 49 (2), 265282.
capital. Manag. Decis. 45 (5), 897909. Tirole, J., 2001. Corporate governance. Econometrica 69 (1), 135.
Ernst & Young, 2002. Corporate venture capital report. Ernst & Youngs Global Turcan, R.V., 2008. Entrepreneurventure capitalist relationships: mitigating post-
Corporate Venture Capital Report for 2002, http//:www.ey.com. investment dyadic tensions. Ventur. Capital. 10 (3), 281304.
Fluck, Z., 2010. Optimal nancial contracting: control rights, incentives, and Wang, H., Wuebker, R.J., Han, S., Ensley, M.D., 2012. Strategic alliances by venture
entrepreneurship. Strat. Change 19 (1-2), 7790. capital backed rms: an empirical examination. Small Bus. Econ. 38 (2), 179196.
Hellman, T., 1998. The allocation of control rights in venture contracts. Rand J. Econ. 29 Wang, L., Dang, X.H., He, L.P., 2009. Research on the selection and the inuencing
(1), 5776. factors of control rights structure in venture capital-backed new high-tech rms
Helwege, J., Packer, F., 2009. Private Matters. J. Financ. Intermed. 18 (3), 362383. based on incomplete contract. Chin. J. Manag. Sci. 17 (5), 166173.
Heughebaert, A., Manigart, S., 2012. Firm valuation in venture capital nancing rounds: Wang, L., Dang, X.H., Wang, X.L., 2010. Research on the inuencing factors of control
the role of investor bargaining power. J. Bus. Financ. Account. 39 (3), 500530. rights allocation in venture capital-backed rms based on the incomplete contract.
Hirsch, J., Walz, U., 2013. Why do contracts dier between venture capital types? Small Sci. Res. Manag. 31 (4), 5966.
Bus. Econ. 40 (3), 511525. Weber, C., Weber, B., 2005. Corporate venture capital organization in Germany. Ventur.
Hopp, C., Lukas, C., 2014. Evaluation frequency and evaluators experience: the case of Capital. 7 (1), 5173.
venture capital investment rms and monitoring intensity in stage nancing. J. Xiong, Y., 2010. Relationships among trust, control mechanism and supplier
Manag. Gov. 18 (2), 126. performance. Manag. Rev. 22 (9), 113120.
Jaeger, D.A., Dohmen, T., Falk, A., Human, D., Sunde, U., Bonin, H., 2010. Direct Yan, Z.X., Fei, F.Y., 2007. Control allocation in rm nancings and entrepreneurs
evidence on risk attitudes and migration. Rev. Econ. Stat. 92 (3), 684689. incentive. Econ. Res. J. (2), 112123.
Jensen, M.C., Meckling, W.H., 1976. Theory of the rm: managerial behavior, agency Zacharakis, A., Erikson, T., George, B., 2010. Conict between the VC and entrepreneur:
costs and ownership structure. J. Financ. Econ. 3 (4), 305360. the entrepreneurs perspective. Ventur. Capital. 12 (2), 109126.
Jukka, V., 2002. An incomplete contracts approach to nancial contracting: a comment. Zhao, D., Wang, Z.J., Zhang, H.H., 2012. Product heterogeneity, cost dierence and
Econ. Bull. 7 (1), 13. technology licensing of enterprise with incomplete bargaining power. J. Manag. Sci.
Kaplan, S.N., Stromberg, P., 2003. Financial contracting theory meets the real world: an China 15 (2), 1527.