Because the entrepreneur’s private benefits are relatively large and the VC’s strategic benefits are relatively small, the entrepreneur values more the SCR for private benefits and the VC places more value on the RCR for pecuniary benefits. As the entrepreneur’s human capital becomes more exclusive, he would gain stronger bargaining power and request more SCR. To compromise and maintain the VC’s financial capital and other inputs, the entrepreneur would be willing to transfer some RCR to the VC with pecuniary benefits. Thus, we have the following hypothesis:

In this situation, the VC places more value on the SCR that can bring more strategic benefits and the entrepreneur values more on RCR for pecuniary benefits. As the exclusiveness of the entrepreneur’s human capital increases, the entrepreneur gains more bargaining power over the VC. With an emphasis on pecuniary benefits, the entrepreneur demands more RCR from the VC and is willing to transfer some SCR to the VC to maintain the collaborative relationship, compensating the VC with strategic benefits. Thus, we have the following hypothesis:

Residual control rights: According to contemporary corporate governance theories, RCR is generally deemed equivalent to the control of the enterprise’s board of directors. Entrepreneurs and management exercise RCR through board control (Wang 2014; Wang & Dang 2008), and VCs also exercise RCR through board of directors and supervisors to monitor and control the behaviours of entrepreneurs. Following Aghion and Tirole (1997), Tirole (2001), Wang (2014) and Wang and Dang (2008), this study uses the arithmetic average of the percentages of board members, executive directors and supervising committee members that are occupied by VCs as the measure of VCs’ RCR.

Specific control rights: Lerner and Merges (1998) propose 10 measures on corporate control rights from three dimensions including key management decisions, enterprise scale decisions and corporate governance decisions. Among these proposed measures, management recruitment, top management control and board seat distribution are considered as board power in RCR (Wang 2014). The survey for this study collect data on seven measures regarding the SCR of VCs and entrepreneurs, including daily operations, mid-level manager recruiting and promotion, patent licensing, R&D project decisions, R&D investment amount decisions, new product development and capital project decisions. Following the calculation method of relative control in Liu and Ma (2004) and Wang (2014), we calculate the percentage of the VC’s SCR in the total SCR of the VC and the entrepreneur as the measure of the VC’s SCR in an entrepreneurial firm.

Residual claim: Residual claim is the claim on the revenue that remains after subtracting contractual payments. In Grossman and Hart (1986) and Hart and Moore (1990), measures on the residual claim include shares, deferred shares and options. Following Wang (2014) and Wang and Dang (2008), this study measures residual claims in entrepreneurial firms using the share ownership percentage of the VC and the entrepreneur.

Specificity of entrepreneur’s human capital: The human capital uniqueness measures developed by Lepak and Snell (2002) have eight items, including ‘not widely available in the labour market’, ‘would be very difficult to replace’, ‘not available to our competitors’, ‘widely considered the best in our industry’, ‘unique to our organisation’, ‘difficult for our competitors to imitate or duplicate’, ‘customised to our particular needs’ and ‘distinguish us from our competition’. Medina, Cabrales and Cabrera (2011) use these uniqueness measures in their study to evaluate innovative performance of human capital and social capital. This study adopts the eight measures as above on the specificity of entrepreneurs’ human capital.

Exclusiveness of entrepreneur’s human capital: The owner of exclusive human capital can exert certain control over the owners of non-human capital and has strong bargaining power in negotiations regarding enterprise ownerships. The monopoly power of human capital owners can undermine the completeness of contracts, thus granting certain control to human capital owners (Yang & Zhou 1997).

Following Yang and Yang (2001), this study integrates the measures on the irreplaceability of human capital in Hatch and Dyer (2004), Jardon and Martos (2012) and Ruzzier et al. (2007) and adopts the following four considerations when measuring exclusiveness of human capital: (1) the ability to identify and achieve profitable market opportunities, (2) possession of some key technology with perspective of large commercial profit such as the ownership of important patents, (3) possession of a large amount of financial capital in an environment where such capital is scarce and (4) possession of special social capital that can produce significant commercial opportunities.

Industry character: Boxall and Purcell (2003) and Datta, Guthrie and Wright (2005) believe that reliance on human capital is much stronger in high-tech enterprises than in traditional enterprises. In this study, we employ a dummy variable for this separation, with industry character equal to one for high-tech enterprises and zero for traditional enterprises.

Non-pecuniary private benefits of entrepreneurs: Non-pecuniary private benefits of control refer to the social status and reputation that are associated with the control of an enterprise. The composition of private benefits arising from control is complex, including pecuniary income or non-pecuniary benefits via tunnelling and reputation gain, which is important to enterprise founders and family businesses (Helwege & Packer 2009). Schmidt (2003) points out that private benefits of control can be significant even when compared to pecuniary benefits. For example, managers with academic backgrounds may pay more attention to the technological value of a project rather than the commercial value. Environmentally conscious managers or advocates of local economic developments may not always maximise enterprise value when making decisions. Based on these analyses, this study adopts four measures on entrepreneurs’ non-pecuniary private benefits, including social status, on-the-job consumption, professional reputation and peer recognition (Wang 2014).

Non-pecuniary strategic benefits of venture capital companies: According to a survey on global corporate venture capital projects by Ernst and Young (2002), 67% of corporations invest in entrepreneurial firms for strategic objectives. Ranked from highest to lowest, the strategic objectives include windows on technology developments, leveraging internal technological developments, importing or enhancing innovation with existing business units, corporate diversification and tapping into foreign markets. The strategic objectives of entrepreneurial investments should be closely connected to commercial activities of entrepreneurial firms, core competence, market strategy, development potential, competitive advantage, resource utilisation and industry functions (Gaba & Bhattachary 2012). With reference to the measures on VCs’ strategic benefits as developed in Wang (2014) and Weber and Weber (2005), this study adopts seven measures on VCs’ non-pecuniary strategic benefits, including windows on technology developments, utilisation efficiency of technological platforms, support in the industrial chain, new market development, corporate diversification, economies of scale and economies of scope. Table 1 summarises the measures on the specificity and exclusiveness of entrepreneurs’ human capital, non-pecuniary private benefits and strategic benefits.

In this study, we control several variables at the enterprise level to separate their influence from the effect of the entrepreneur’s human capital. Kaplan and Stromberg (2004) document that investment stage, firm size, industry and investment amount all affect the control right distribution. Zhang, Jia and Wan (2009) study control right sharing and operation efficiency in public–private joint enterprises in China and find that firm size and the nature of firms negatively affect VCs’ control power, while the investment stage, industry and investment amount all positively affect VCs’ control power. Following the literature, this study includes four control variables at the enterprise level as discussed subsequently.

Firm size: Firms are categorised into three groups based on the number of employees: small firms (less than 200 employees), medium firms (200–1000) and large firms (more than 1000). Large firms tend to have more resources, greater ability, as well as sufficient size for internally generated investment. Firm size is often used as a proxy for an enterprise’s resources and ability. This study uses an ordinal variable size to measure firm size. Size is equal to 1 for small firms, 2 for medium and 3 for large firms.

Industry: This study adopts the 12 categorisations as in Listed Companies Industry Classification Guidelines of China and assigns the variable Industry a value from 1 to 12 for each industry.

Stage: Because it is difficult to collect time series data on the same entrepreneurial firms, this study collects a cross-section sample of entrepreneurial firms and uses the ordinal variable Stage to represent their stages: 1 for start-up, 2 for growth, 3 for maturity and 4 for others.

Investment: The variable Investment is measured by the actual amount invested by the VC (in RMB10 000).

Because there is no public database on entrepreneurial firms in China, we collect the required data through field studies and surveys. In a preliminary round, a small number of field studies and interviews are conducted to gain insight into the current status of control rights in Chinese entrepreneurial firms. Then, a large-sample survey is conducted with mid- to high-level managers in VC companies to collect detailed information for the empirical analyses.

The design of the formal survey follows the issue-oriented principle and is based on the hypotheses to be tested. The questionnaire adopts the popular Likert scale (Dawes 2008; Likert 1932) with the following choices: 1 – completely insignificant (or completely unimportant), 2 – insignificant (or unimportant), 3 – neither significant nor insignificant, 4 – significant (or important) and 5 – very significant (or very important). The degree of conformity with the question increases as the number increases.

This study focuses primarily on high-tech enterprises financed by VC companies, which are also the research subject of most existing literature. In recent years, VC companies in China have supported not only high-tech enterprises such as Internet of Things and new energy sources but also traditional enterprises such as manufacturing and financial businesses, chain restaurants, and education and training institutions. These traditional types of enterprises are also included in our survey so that the results can be compared to those of high-tech enterprises.

The survey sample includes regions with different levels of economic and market development, and such regional balance can ensure the generality of analytical results. This project selects representative cities from three regions: east, central and west China. In the east region, we selected Jinan, Shenzhen, Nanjing and Wuxi; in the central region, we selected Hefei; and in the west, we selected Xi’an and Wuzhou, where entrepreneurial investment has developed fairly quickly in recent years.

Most existing literature on corporate governance targets only top managers in their surveys. However, in our field studies, we realise that besides top-level managers, the mid-level managers, such as investment managers and project managers in VC firms, are also appropriate survey subjects. These managers maintain direct contact with entrepreneurial firms and thus are quite familiar with entrepreneurs and their firms. For this reason, the survey subjects include top- and mid-level managers and investment decision committee members.

The survey in this project was conducted in two stages. The first stage was in the period of July to September 2009. Survey questionnaires were distributed via local networks (such as local entrepreneur associations or municipal government branches) in the target regions. In total, 450 copies of surveys were distributed, and 205 copies were later collected, with an overall response rate of 45.56%. Among the returned surveys, 98 were invalid for various reasons including incomplete information, bankruptcy of entrepreneurial firms or disqualification of the respondents. Therefore, 107 valid completed surveys were received, with an effective response rate of 23.78%. The sample size is sufficient for our statistical analysis. The second stage of the survey was conducted in the period from July to December 2011, which was meant to follow up with the 107 respondents in the valid sample from the previous survey. The second round of the survey revealed that among the 107 respondents, VCs exited successfully from 12 entrepreneurial firms, 9 firms were bankrupt, 15 firms received the second- or third-round investment from VCs and the remaining 71 were still working with the VC companies. The two rounds of survey yielded a valid sample size of 193, including 121 in strategic emerging or high-tech industries and 72 in traditional industries.

Table 5 reports the regression results on the RCR with the full sample in the two situations.

Model M1.1 only considers the effects of control variables. The estimated coefficient of investment is positive and significant, indicating that the higher the investment amount, the more RCR the VC has in the firm. Other variables are insignificant.

M1.2 adds to M1.1 residual claims of the VC and the entrepreneur as well as industry character as explanatory variables, which increase the explanatory power of the model (adjust-R² increases from 0.151 to 0.227). Venture capitalist’s RCR is positively related to the industry character at the 10% level, which indicates that the VC tends to have more RCR in a high-tech enterprise. Venture capitalist’s RCR is negatively related to his residual claim at the 1% level, while the residual claim of the entrepreneur is insignificant.

M1.3 adds to M1.2 the specificity and exclusiveness of entrepreneurs’ human capital, as well as the squared terms of specificity and exclusiveness for potential nonlinear effects. The model’s adjust-R² further increases to 0.433. Among the control variables, the investment stage becomes more significant at the 5% level, while the effects of the other control variables remain unchanged. Industry character is now positive and significant at the 5% level, and the effects of the VC’s and the entrepreneur’s residual claims have not changed. The effect of the specificity of entrepreneur’s human capital is significantly negative (at 10%), yet in a nonlinear fashion, and the effect of exclusiveness is highly positive (at 1%) in a nonlinear fashion. This finding is consistent with our hypothesis that the VC’s RCR is negatively related to the specificity of the entrepreneur’s human capital and is positively related to the exclusiveness.

M1.4 is the baseline model of this situation that includes control variables only, and its result resembles that of M1.1. Adding residual claims and industry character as independent variables to M1.4 forms M1.5, which shows increased explanatory power (adjust-R² increases from 0.126 to 0.211). The investment amount is positively related to the VC’s RCR (significant at 5%), and industry character is positively significant at 10%. The effect of VC’s residual claim is negatively significant at 10%, while the entrepreneur’s residual claim is insignificant.

As in M1.3, M1.6 adds the quadratic effects of the specificity and exclusiveness of the entrepreneur’s human capital. The effects of control variables remain the same. The effect of specificity is significantly positive (at 10%) although nonlinear, and the effect of exclusiveness is just the opposite (negative at 10%). Such results show that the VC’s RCR is positively related to the specificity of the entrepreneur’s human capital and is negatively related to the exclusiveness.

To ensure the reliability of the results in Table 5, we conduct robustness checks to address potential concerns of omitted variables, heteroscedasticity and multicollinearity. The p-value of the omitted variable test is greater than 0.1, of the heteroscedasticity test is greater than 0.05, and the average variable inflation factor is less than five. These tests demonstrate that the models do not suffer from these concerns. The above results provide compelling evidence in favour of Hypotheses 1–3 regarding the impacts of specificity and exclusiveness of the entrepreneur’s human capital on the VC’s RCR.

To test Hypothesis 4, we decompose the sample into two subsamples, one with observations for enterprises in high-tech industries and the other for enterprises in traditional industries. The estimated results using the two subsamples are reported in Table 6.

Based on the results in Table 6, we find that the effects of control variables are similar with slightly different significant levels for both subsamples. The impact of the VC’s residual claim (ownership percentage) is negative for high-tech enterprises but positive for traditional enterprises. This suggests that in traditional industries the VCs’ RCR is largely in line (proportional) with their ownership, but in high-tech enterprises the VCs tend to have less RCR relative to their ownership. This observation implies that in high-tech industries entrepreneurs have more representation in boards and supervising committees, which reflects the bigger role that entrepreneurs play.

The estimated coefficients on Specificity² are significantly negative for both subsamples, indicating that the impact of the specificity of the entrepreneur’s human capital is negative (nonlinear though) for both high-tech and traditional companies. On the other hand, the estimated coefficients on Exclusiveness² are significantly positive for both subsamples. Moreover, in terms of the size and significance of the estimated coefficients, the influences of these two factors are more pronounced for high-tech companies than for traditional companies.

In this scenario, the effects of control variables and residual claims are almost the same as they are in the previous scenario. The effect of the specificity of the entrepreneur’s human capital is positive in both subsamples, while the effect of the exclusiveness is negative in both subsamples. The effects of specificity and exclusiveness are stronger for high-tech enterprises than for traditional enterprises, which are consistent with the predictions in Hypothesis 4. Additionally, the exclusiveness seems to have a stronger general impact (judged by significance of coefficients) than the specificity on the VC’s RCR.

In this case, among the control variables, firm size and industry are insignificant for all samples. Stage is marginally negative (10% level) in two samples, and investment amount is positively significant. The effect of specificity is strongly positive in a convex fashion for all three samples, while the effect of exclusiveness is linear and negative. These effects are generally stronger in the high-tech sample than in the traditional company sample, which supports Hypothesis 4.

In this case, compared to the results reported in the last subsection, the effects of all control variables are largely unchanged except that the investment amount becomes more significant, especially in the sample of high-tech enterprises. The industry character is highly positive, suggesting significant difference in SCR between high-tech and traditional enterprises. The effect of specificity is significantly positive for all three samples, which is inconsistent with the negative relationship as hypothesised. The impact of exclusiveness is strongly positive in a convex fashion in all cases. Similar to the case in which entrepreneur’s private benefits outweigh VC’s strategic benefits, these effects are generally stronger in the high-tech sample than in the traditional company sample. Collectively, the preceding results generally support (in three out of four cases) our hypotheses about the ways in which specificity and exclusiveness of the entrepreneur’s human capital are related to the VC’s SCR.

To ensure the reliability of the results, robustness checks are conducted for the regressions in Table 7. Testing results show no concern for omitted variables, heteroscedasticity or multicollinearity.