رابطه بین ساختار حکومت و رویکردهای مدیریت ریسک در شرکت های سرمایه گذاری ژاپنی

This paper attempts to understand what drives Japanese venture capital (JVC) fund managers to select either active managerial monitoring or portfolio diversification to manage their firms' investment risks [J. Bus. Venturing 4 (1989) 231]. Unlike U.S. venture capitalists that use active managerial monitoring to gain private information in order to maximize returns [J. Finance 50 (1995) 301], JVCs have traditionally used portfolio diversification to attenuate investment risks [Hamada, Y., 2001. Nihon no Bencha Kyapitaru no Genkyo (Current State of Japanese Venture Capital), Nihon Bencha Gakkai VC Seminar, May 7]. We found that performance pay is positively related to active monitoring and that management ownership is positively related to active monitoring and negatively related to portfolio diversification. The managerial implication of our study is that venture capitalists should be as concerned about the structure of their incentive systems for their fund managers as they are for their investee-firm entrepreneurs. Agency theory says that contingent compensation is a self-governing mechanism for individual effort that is difficult to measure and verify. When properly applied, equity ownership and performance-based pay can have powerful influencing effects on the strategic choices of managers.

What is the impact of the structure and governance mechanisms of a VC firm on its risk management approaches? To answer this question, we build an agency model of the relationship between the JVC firm and the managers who make these choices Robbie et al., 1997, Gompers and Lerner, 1999 and Baker and Gompers, 1999. JVCs, unlike their U.S. counterparts that use active monitoring to attenuate portfolio risks (Lerner, 1995), have traditionally used a passive approach—portfolio diversification (Hamada, 2001). However, the recent increased use of active monitoring by JVCs has allowed us to compare the factors that drive the choice of these risk management approaches Gorman and Sahlman, 1989 and Gompers and Lerner, 2001a. Briefly, our model relates the corporate affiliation of a JVC, compensation structure for its investment managers, and VC firm managerial stock ownership to the choice of active (hands-on monitoring) or passive (portfolio diversification) risk management approach. To examine this model, the paper proceeds as follows. First, we review the JVC industry, followed by the theoretical model. Then, we discuss the methods and results, concluding with some observations on the implications of our findings and directions for future research.

This paper examined two risk management approaches used by JVC firms—portfolio diversification and direct managerial monitoring—and the factors that determined how they were chosen. From agency theory, we hypothesized that corporate affiliation, compensation schemes for investment managers, and management equity ownership determined the choice of risk management approach. The data show that both performance-based pay and managerial ownership are positively related to direct managerial monitoring while managerial ownership is negatively related to portfolio diversification. Although the final effects are small, due in part to the sample size and the constrained model, our results provide some support to the agency theory predictions. While our hypotheses on the corporate affiliation were not supported, other important managerial incentive-based hypotheses were supported, which provide another piece of empirical support to the popular framework used in many U.S.-based studies on VC. Thus, our findings provide some support for the use of standard agency theoretic formulations for non-U.S.-based VC studies. We theorized that an affiliated JVC would more likely adopt a passive risk management approach since this reflects the strategic objectives of its larger parent, which in the Japanese corporate context is focused on stabilizing rather than maximizing portfolio cash flows. However, affiliation was not statistically related to risk management approaches and a closer examination of the correlation table shows why. In fact, the correlation table (Table 2) and full model (Fig. 2) report that JVCs affiliated to financial institutions or other corporate firms had ownership and control structures typically associated with the portfolio risk reduction, rather than value maximization, objectives of their parents. However, it was also highly correlated with the more specific dimensions of ownership and control, implying that it may be measuring the same thing. In hindsight, we should have expected this because it is unlikely that affiliated JVCs could be structured as entrepreneurial units, with the implications for radically different ownership and compensation systems. Basic organization theory suggests that doing so in a culturally monolithic context (i.e., the Japanese corporation) would create internal conflicts in the reward system, which could be severely disruptive to morale and operational effectiveness. Thus, the only type of JVC that could adopt shareholder wealth maximizing approaches would be the independent ones. In short, there was no variance in the data between affiliation status and ownership and control, which accounts for the lack of findings. From a methodological standpoint, this study contributed to the entrepreneurship literature by illustrating a method, structural equation modeling with PLS, that is well tested and utilized in disciplines of management but is still relatively novel in entrepreneurship research. PLS is particularly well suited to entrepreneurship research since it allows for structural equation modeling in circumstances in which the data set is small and/or the assumption of normally distributed data is unadvisable. Such is the case with much of entrepreneurship research, particularly when the domain is constrained to specific industries or life stages but where meaning statements of causality can be made. Here, the researcher is confronted with either small data sets or variables that seriously violate the assumptions of normality, in either of these cases it is likely that the results have been, at best, considered with great suspicion. Through the use of PLS nonnormal data or small data sets can now be used and the results can be looked on with the same or greater confidence as studies based on larger data sets. In terms of future research opportunities, extant studies in VC often suffer from the lack of nuanced data, which can lead to gross generalizations. In this study, we attempted to obtain finer grained data by conducting field interviews with venture capitalists but were still forced to use binary measures for variables (e.g., compensation) that in other studies would have contained more information. Our informants were simply not willing to be more specific although the study (indeed most studies like it) could have benefited from higher quality measures such as data on compensation contracts, portfolio structure, and even management styles. The VC community around the world is known to be tight-lipped with an ethos that puts a premium on access to private information as a competitive advantage in deal making. This study points to the criticality of researcher credibility and personal networks. It suggests that advances in VC research would mostly likely come from teams of researchers (e.g., Manigart et al., 2002) who can combine their personal networks to obtain field-based panel data rather than those that rely solely on secondary data. The managerial implication of our study is this: that venture capitalists should be as concerned about the structure of their incentive systems for their fund managers as they are for their investee-firm entrepreneurs. Agency theory says that contingent compensation is a self-governing mechanism for individual effort that is difficult to measure and verify. However, such schemes must be sensitive to the organizational context in which they are applied. When organizational culture puts a premium on distributive justice, as it is the case with many large Japanese corporations, some forms of contingent compensation such as equity ownership may be difficult to implement. This may be particularly salient for JVCs that are affiliated with other organizations. Nonetheless, our findings suggest that managerial incentives through performance-based pay or ownership can affect a JVC's choice of risk management approach. In conclusion, our contribution to the research literature is a closer examination of the VC firm and VC manager relationship. Whereas it has generally been assumed that an agency gap exists between the VC firm and venture-backed entrepreneur, there appears to be an implicit assumption in the literature that the interests of the VC firm and VC manager converge. We built a model to test this and have shown that variations in compensation and ownership can explain the variations in strategic choices of the managers. In conclusion, our study suggests that general formulations of agency theory appear to operate just as well in the Japanese relational exchange context as it does in the Anglo-American transactional exchange context. Such approaches are particularly useful for modeling governance problems resulting from information asymmetry, unequal risk bearing, and uncertain cash flows.