Allocation of attention to portfolio companies and the performance of venture capital firms.
Prior research has identified that venture capitalists commonly participate actively in the monitoring and management of their portfolio companies (Sahlman, 1990). According to Gorman and Sahlman (1989), venture capitalists typically spend 60% of their time managing their investments, while the balance is spent on screening new investments and on administration. Active management of investments is widely regarded as a defining characteristic of venture capitalists (Ehrlich, Denoble, Moore, & Weaver, 1994; Fried & Hisrich, 1995; Hellmann & Puff, 2002; Lerner, 1995; Sapienza, Manigart, & Vermier, 1996), although the level of involvement varies according to the chosen investment strategy, stage of the venture, role of the venture capitalist in a syndicate, level of innovation of the venture, and experience of the venture capitalist (Elango, Fried, Hisrich, & Polonchek, 1995; Macmillan, Kulow, & Khoylian, 1989; Sapienza, Amason, & Manigart, 1994).
The motivation of venture capitalists is to maximize the performance of their investments by providing value-added support and by controlling and monitoring the development of the companies (Gorman & Sahlman, 1989; Lerner, 1995; Macmillan et al., 1989; Rosenstein, Bruno, Bygrave, & Taylor, 1993; Sahlman, 1990; Sapienza, 1992; Sapienza & Gupta, 1994; Sapienza et al., 1996). While the earlier research has been able to validate the value added by a venture capitalists' involvement in a venture (Busenitz, Moesel, & Fiet, 1997; Higashide & Birley, 2000; Macmillan et al.; Rosenstein et al., 1993; Sapienza, 1992; Sapienza et al.), the limitations that they face in these tasks have received less attention. Although the scarcity of their time has been widely acknowledged in the earlier research on venture capitalist involvement (e.g., Gifford, 1997), the empirical analysis concerning this issue is scarce (Lerner, 1995) and is mainly based on anecdotal evidence (Gorman & Sahlman, 1989; Macmillan et al.; Sapienza et al.). The challenge is that the performance of an individual portfolio firm does not directly translate to the whole portfolio's performance. A venture capitalist seeks to allocate the optimal amount of attention to an individual venture in order to maximize the performance over all portfolio companies. Thus, although there is evidence for value added by venture capitalists, the link between their level of involvement and the performance of the portfolio has not been established.
We set out to examine how venture capitalist involvement in portfolio firms is related to the performance of the venture capital (VC) firm. We explore the limitations of venture capitalists' involvement considering the nature of informational and interpersonal aspects of their assistance and governance. The key proposition of the article is that the way a VC firm allocates attention to portfolio companies affects the performance of those companies and thus, the overall performance of the VC firm. We divide our argumentation in two parts. The first part addresses the effects of attention allocated to individual portfolio companies, and the second part discusses the effects of that attention on the portfolio level. Furthermore, we argue that cooperation through syndicated coinvestments with other venture capitalists and through sharing the workload effectively alleviates the constraints on providing attention. Our analysis of the data set on the 94 leading VC firms in the United States provides support for our arguments.
Our findings have both theoretical and practical implications. The results support the recent theoretical research on the optimal portfolio size of venture capitalists (Gifford, 1997; Kanniainen & Keuschnigg, 2003) and expand the theory by introducing the moderating role of syndication on optimal portfolio size. The findings also show that the venture capitalist involvement enhances the outcomes of the ventures. This validates the results of earlier research on the value added by venture capitalist involvement and shows that these efforts are most productive when focused on a suitable number of portfolio companies. Furthermore, our findings provide guidance for venture capitalists on successful strategies and resource allocation, underlining the effects of involvement and the use of time in ventures.
The remainder of the article is organized as follows. Section 2 builds testable hypotheses. Section 3 presents the data and the methods used. Section 4 presents the empirical results. Finally, conclusions and implications are discussed in section 5.
Theory and Hypotheses
Motivations for Involvement
The earlier literature has suggested two motivations for venture capitalists' involvement in their portfolio companies: monitoring and supporting the companies. First, the venture capitalist's role in a venture is primarily that of an owner. A VC fund typically holds a significant proportion of the shares in the ventures, received in exchange for their financial input (Sahlman, 1990). Venture capitalists are only rarely directly involved with the daily operations of the ventures and thus, this separation of ownership and control leaves them reliant on the chief executive officer (CEO) of the firm to manage it in a way that maximizes the value of the venture capitalist's investment (Fama & Jensen, 1983). However, the interests of venture capitalists and CEOs may not be perfectly aligned, potentially leading to opportunistic behavior and to the pursuit of private interests by the CEO (Admati & Pfleiderer, 1994; Bergemann & Hege, 1998; Gompers, 1995; Jensen & Meckling, 1976). In addition, the entrepreneur's potentially limited ability to manage the venture introduces an agency problem requiring venture capitalist involvement (Sapienza & Gupta, 1994). Agency risk gives rise to the venture capitalist's governance with a need to monitor the activities of the ventures to ensure that the management's conduct is aligned with the interests of the venture capitalist (Fama & Jensen, 1983; Sapienza et al., 1996). In addition, monitoring the management and the venture in general transfers information about the venture's quality to the venture capitalist, reducing the amount of asymmetric information between the two entities (Akerlof, 1970; Chan, 1983; Gompers, 1995).
The second motivation for allocating attention to an investment is based on the value of the venture capitalist's assistance. While the governance of ventures concentrates on the value of reduced risks and on the prevention of undesirable outcomes, the assistance perspective considers the venture capitalist's involvement as a valuable resource for the focal ventures. Venture capitalists contribute resources that serve the development of the company as an additional input. The literature has focused on the level of involvement and roles assumed by the venture capitalists (Ehrlich et al., 1994; Elango et al., 1995; Gorman & Sahlman, 1989; Macmillan et al., 1989; Rosenstein et al., 1993; Sapienza et al., 1994) and on the value added by the venture capitalist's assistance (Busenitz et al., 1997; Higashide & Birley, 2000; Macmillan et al.; Rosenstein et al; Sapienza, 1992; Sapienza et al., 1996). Prior research has identified multiple forms of value-added assistance provided by venture capitalists. In addition to the main contribution--arrangement of financing and related activities such as financial monitoring and interfacing with investors--the venture capitalists contribute to the strategy process, operational issues, and recruitment of key personnel (Barney, Busenitz, Fiet, & Moesel, 1996; Gorman & Sahlman, 1989; Hellmann & Purl, 2002; Macmillan et al.; Rosenstein et al.; Sapienza, 1992; Sapienza et al.).
Attention Allocated to an Individual Venture
The assistance provided by venture capitalists to their portfolio companies is largely based on the experience and information offered by the venture capitalists (Barney et al., 1996; Bygrave & Timmons, 1992; Fried & Hisrich, 1995; Norton & Tenenbaum, 1993; Rosenstein et al., 1993; Sahlman, 1990; Sapienza, 1992; Sapienza et al., 1996; Steier & Greenwood, 1995). Operating as a sounding board, supplying business advice and general business knowledge, and participating in strategic and operational planning draw on the venture capitalist's earlier experience and existing information.
While part of the information is explicit, such as contacts and market information, the experience of venture capitalists is the result of the knowledge cumulated throughout their career and is thus largely tacit know-how (Kogut & Zander, 1992; Polanyi, 1967). Know-how is expressed through application (Grant, 1996) and thus, the transfer of such knowledge requires interaction (Kogut & Zander, 1992). Due to the low transferability of the knowledge based on experience, the venture capitalists may choose to address a venture-specific problem directly instead of trying to transfer the specific knowledge to the entrepreneur. Thus, in order to provide assistance based on know-how, the venture capitalist can provide either information or information-processing capacity (Sapienza et al., 1996). In both cases, the attention allocated is specific to an individual venture, and is thus unavailable to other ventures.
Governance of portfolio companies faces restrictions similar to those related to providing assistance and information. The monitoring is based on the presumption that, once the venture capitalist acquires information to verify the behavior of the entrepreneur, the behavior of the entrepreneur is more likely to be aligned with the interests of the venture capitalist (Eisenhardt, 1989; Fama & Jensen, 1983). Venture capitalists gather information on the behavior of the management and on the performance of the firm and assess this information against their experience and expectations. Similar to the assistance, the time and attention allocated to the governance and the monitoring of an individual portfolio company consume the attention available to other ventures.
Although the value of the assistance increases with the amount of attention allocated, the knowledge base of a venture capitalist is limited, thus implying a limit to the value of assistance at high levels of involvement. In addition, the benefits of governance are also likely to become exhausted at higher levels of monitoring. Extensive involvement may also be destructive if the venture capitalist attempts to assist the venture beyond its learning capacity, or if the monitoring becomes intrusive and interferes with the venture's development. Thus, extensive involvement can reduce the marginal return on attention to zero or even negative. However, with the generally noted opportunity cost of venture capitalists and their tendency to economize on time, it is likely that they normally operate at such levels of involvement that the return on effort is increasing, although at a decreasing rate.
Portfolio Size and the Effects of the Allocation of Attention
Since a venture capitalist must allocate time among many investments, the amount of attention that is available to each venture is closely related to the number of investments managed by the venture capitalist. The average time available for each company rapidly decreases as the number of companies in the portfolio increase, thus decreasing the potential for value added.
As an increasing number of managed companies reduce the amount of attention available to an individual venture, it also reduces the total time available to the management of these companies due to the portfolio's increasing size and diversity. Each additional venture introduces more overhead in the form of nonproductive tasks, such as travel time to visit the ventures. This accumulating overhead reduces the total time available for the management of the portfolio companies.
A larger number of managed companies also increase the diversity of the ventures. In their article on the diversification of companies, Prahalad and Bettis (1986) suggest that the factor limiting diversification is the ability of management to handle the strategic variety of the business units. The management acts on and interprets new information on the basis of existing experiences. These belief structures and mental representations enhance the processing of information from situations and businesses that are familiar to the management. Thus, strategically similar businesses can be managed using a single dominant logic. However, although efficient in a given context, the dominant logic hinders the ability to manage businesses that do not fit within that sphere. The venture capitalist faces a similar situation to the management of a diversified firm. Each company in the portfolio differs from the other and, although possibly operating in the same industry, may face situations that are strategically very different. The venture capitalist may be fully able to manage an individual investment, but when a portfolio has a number of companies in different strategic situations, development stages and environments, managing the diversity becomes difficult. The larger the number of managed companies, the more diversity there is and the slower it is for the venture capitalist to adapt to each specific situation. Thus, each additional portfolio company reduces the total time available for the management of the portfolio due to its increased diversity.
To summarize, the effect of dividing a venture capitalist's attention is twofold. First, the average time available for each company decreases in relation to the number of managed companies, therefore reducing the effect of venture capitalist's involvement. Second, the increasing number of companies increases the size and the diversity of the portfolio and thus increases the time required to switch the focus of attention from one firm to another. This reduces the total amount of time available to support the companies under management.
On the portfolio level, this implies two opposite effects. On one hand, if we consider a venture capitalist managing a single company, more time is available for the company than might be beneficial for its development. Therefore, a second company could be managed without affecting the attention needed for the first company. As the number of managed companies increases, so do the potential outcomes as the portfolio is larger. On the other hand, if the venture capitalist continues to increase the number of portfolio companies, the time available to each of them decreases, thus reducing the value added from the venture capitalist. The positive effect of portfolio size eventually slows and begins to reverse. That is, economies of scale initially exist in relation to venture capitalist attention, but diseconomies of scale eventually emerge as the portfolio increases beyond a manageable size.
Therefore, we propose a curvilinear (inverted U-shaped) relationship with respect to the extent of the involvement of the venture capitalist in portfolio companies and the performance of the VC firm. In particular, we propose that the relationship between the extent of involvement and the performance of the firm exhibits diminishing marginal returns, and passes the point of diminishing total returns and thus, eventually exhibits negative marginal returns for each additional portfolio company.
Hypothesis 1: There is a curvilinear (inverted U-shaped) relationship between the attention allocated to the investments (i.e., portfolio companies per partner) and the performance of the VC partnership.
Cooperation and Syndication
Syndication Frequency. As the increasing size of a portfolio reduces the attention available to each company, the time available to the venture capitalist forms a constraint on the venture capitalist's attention and available time. This limits the number of portfolio companies the venture capitalist can effectively handle. The syndication of investments offers a mechanism for venture capitalists to reduce the time required to manage an individual venture by sharing the workload with syndicate partners.
VC syndicates are typically recurring among partners, with the participants alternating between the roles of lead and nonlead investor (Bygrave, 1987, 1988; Wright & Lockett, 2003). The role of a lead investor is to coordinate the investor group and to act as the managing investor of the investment. On one hand, the existence of a lead investor solves the potential problem of "free riding" in the investor group, resulting from the fact that the benefits of the efforts contributed to the investment are shared among all participants. On the other hand, the commitment of the lead investor to the investment is frequently secured both by a larger equity stake and by the reputation effects implied by the reciprocal syndication relationships (Wright & Lockett, 2003). According to Gorman and Sahlman (1989), the lead investor in a syndicate uses 10 times the time the other venture capitalists spend in direct contact with the venture. Participating in syndicates as nonlead investors enables venture capitalists to increase the portfolio size with a lower commitment of resources than acting as a sole investor. Thus, by syndicating investments and delegating a part of the lead investor role to the syndication partners, the venture capitalist is able to increase the number of investments with a lesser negative effect on performance, potentially resulting in a higher optimal number of investments.
We hypothesize that VC firms with a higher level of syndication should be able to manage larger portfolios. Accordingly, the point of diminishing total returns on the allocation of attention should be reached at a higher portfolio size for VC firms with high levels of syndication compared with VC firms syndicating less.
Hypothesis 2: Syndication frequency positively moderates the curvilinear relationship between the attention allocated to the investments and the performance of the VC partnership.
Syndication Role. While a coinvestor is able to join a syndicate with a significantly lower commitment of managerial resources, the management of the investment is delegated to the lead investor (Gorman & Sahlman, 1989; Wright & Lockett, 2003). Therefore, although syndication is an effective method to share workload and thereby to increase the optimal number of investments for the coinvestor, the amount of attention required from the lead investor to manage a syndicated investment is comparable with that of a nonsyndicated investment. Furthermore, the workload of the lead investor is increased by the effort invested in the management of the syndicated investment (Wright & Lockett, 2003) Thus, as predicted in hypothesis 2, syndication should have a positive effect, but this effect should be lower for lead syndication compared with nonlead syndication.
Hypothesis 3: The higher the share of investments syndicated as a lead investor, the lower the positive impact of syndication on the curvilinear relationship between the attention allocated for the investments and the performance of the VC partnership.
To test the effect of the allocation of attention on the performance of the venture capitalists, we used the VC firm as the unit of analysis. By focusing on the performance of VC firms as the unit of analysis, our research complements the prior research in which the value of venture capitalist attention has been primarily addressed at the level of an individual venture (Barney et al., 1996; Busenitz, Fiet, & Moesel, 2004; Gorman & Sablman, 1989; Hellmann & Puri, 2002; Macmillan et al., 1989; Rosenstein et al., 1993; Sapienza, 1992; Sapienza et al., 1996). Combining the outcomes from the ventures and funds enables us to address the question of the effect of allocation of attention and syndication on the performance of a portfolio of ventures.
We tested our hypotheses employing longitudinal data on the largest private U.S. VC organizations, defining the size of firms as the number of portfolio companies the firms had invested in cumulatively by the end of the year 2000. This set of VC firms includes only those U.S. VC partnerships that Venture Economics classifies as "Independent private partnerships," thus excluding investment bank affiliates, corporate investors, endowments, individuals, and other private equity investors. The sample of firms consists of 94 VC firms and their investments. Altogether, the data include 25,009 investments in 6,044 portfolio companies during the years 1986-1998, resulting in 997 firm-year observations. The investments made by the sample firms comprise 48% of all investments recorded in the Securities Data Corporation's (SDC) Venture Economics database for the time period.
The VC investment data in this article are obtained from the SDC database. This extensive source has been used in previous VC research (e.g., Bygrave, 1987; Gompers, 1995; Lerner, 1994; Podolny, 2001; Sorenson & Stuart, 2001). Venture Economics has gathered VC investment data since the 1970s using annual reports of VC funds, personal contacts with funds' personnel, initial public offering (IPO) prospectuses, and acquisitions announced in the media. The database contains information on over 210,000 private equity investments (one whole financing round consists of several single investments), and it is widely recognized as a leading source of U.S. VC investment data.
Data on the personnel resources of the sample VC firms are gathered from the back issues of Pratt's Guide to VC Sources. This widely used publication lists the managing partners, the key personnel, and a variety of other parameters of most of the U.S. VC firms each year. The most relevant records for our study are the managing partners of the firms. These data are reported consistently each year with names and positions. Using data from the publication, we tracked the total number of partners and their names in each VC firm each year. To further ensure the validity of the managing partner data, we collected the current resumes of 28% of the partners of our sample firms and backtracked their career years in each venture partnership. We then compared the publication's listings of partners with the subsample of resumes without observing significant inconsistencies between these two sources.
Performance of Venture Capitalist. A VC fund is typically structured as a limited partnership (Sahlman, 1990). While the investors of the fund, i.e., limited partners, contribute nearly 99% of the capital (Sahlman, 1990), the share of general partners, i.e., venture capitalists, over the profits range from 20 to 25%. Thus, the behavior that maximizes the venture capitalist's personal returns also maximizes the return on the funds invested by limited partners. Although the performance of a VC fund can be defined straightforwardly as return on investment, assessing the performance of a VC fund is hard for an outsider due to the secretive nature of information relating to venture capitalists' profits. However, as the largest valuations and returns to venture capitalists are most often realized in IPOs (Bygrave & Timmons, 1992), we can use the number of IPOs (controlling for the number of investments; see Control Variables section) as a proxy for the venture capitalist' s performance. Although only a fraction of VC investments reach the IPO, most of the total value to the investors is created in these exits (Bygrave & Timmons, 1992; Gompers & Lerner, 1999). Thus, we can consider the IPO as the preferred exit vehicle of most VC firms.
We measured the performance of the VC firm as the number of IPOs generated from the investments in new ventures. We identified the investments made by the venture capitalists and tracked down the number of IPOs performed by these companies. We observed the investment activity between the years 1986-1998 and the IPOs from these investments until the end of June 2003.
The number of IPOs from new company investments is biased downward from the actual value toward the end of the time range of our sample. This results from the fact that new company investments require a certain amount of time before the exit is realized. The median time from the first investment of a venture capitalist to the IPO is 4.18 years in the first half of the sample. This indicates that only slightly over half of the investments made in 1998 had reached the potential IPO by June 2003, when we observed the IPOs. Thus, we observed too few IPOs in the last years of our sample for all venture capitalists, and especially for those firms who concentrate on early-stage investments. Given that our unit of analysis is the VC firm, not a single venture, we could use hazard rate models that could be used to measure the likelihood of single ventures reaching an IPO. However, as the bias is downward, the analysis would tend to reject the hypothesis, making the results more significant if the data were complete. In addition, our use of year dummies also partially remedied this problem. As a robustness test, we reran the analyses limiting the time frame to 1993 instead of 1998 and obtained qualitatively identical results.
While the IPOs can be considered as the vehicle for achieving the largest profits, venture capitalists also use many other kinds of exits including trade sales to industrial acquirers, secondary sales to other investors, buybacks, and write-offs (Cumming & Macintosh, 2003). However, of the different exit methods, trade sales are frequently considered as the second best option for the investor bringing in a significant share of returns (Cochrane, 2005). While trade sales are generally smaller in size than IPOs, individual trade sales can be substantial and comparable in size with IPOs. Thus, as a further robustness test, we used the number of all positive exits (identified by Venture Economics as IPOs, mergers, acquisitions, and buyouts) as an alternative measure of the performance of the venture capitalist. We reported the deviations in results for these two performance measures in the following analyses.
Allocation of Attention. The personnel structure of a VC firm is typically an upside-down pyramid, with a high proportion of upper management and a limited lower level staff (Wasserman, 2002). The experience and operations of a VC firm are embodied in its partners and due to the knowledge-intensive nature of the work, tasks are generally nondivisible (Wasserman, 2002). Managing partners are the key contacts between the VC partnership and the portfolio company (Sahlman, 1990). According to Gorman and Sahlman (1989), in an average firm, a partner has 8.8 investments to manage, while on an associate level, this figure is 3.6. We use the number of partners in a VC firm to measure the total amount of attention available for the management of the investments. We classified personnel as partners if their position title included the term "partner," "vice president," or "managing director." The number of partners each year is taken from our combination of the records from Pratt's Guide to VC Sources and the resumes of the partner subsample. We recognized that the records from the publication represent the situation of the previous year, so we used each year's published data for the previous year's entries. We further assumed that the number of partners during a single calendar year remains constant. In rare cases where the partner data are unavailable from both sources for a certain year, we used the data of the previous year.
We measured the size of the portfolio of a venture capitalist as the number of companies the venture capitalist is involved in. We identified the portfolio companies of a venture capitalist using the investment records of Venture Economics, containing information on the dates of the investment as well as whether a company has made an IPO and when. We recorded the entry date to the portfolio as the date of the initial investment by the venture capitalist. The exit date was recorded either as the date of an IPO or as the date one year after the last observed investment round. This corresponds to the median interval between investment rounds (Gompers, 1995).
We measured the allocation of attention as the size of portfolio relative to the number of partners in each VC firm. Thereby, we were able to account for the differences in the sizes of the firms and to make the measure comparable across firms. Furthermore, the number of new investments per partner offered a measure for the capacity of an individual venture capitalist to manage the investments.
Syndication Frequency. We measured the frequency of syndication as the ratio of syndicated investments to the total number of new investments made within a given year. We recorded an investment as a syndicated one if the Venture Economics database records more than one investor for a given investment round. To capture all possible syndication relationships, we first identified all portfolio companies in which the VC firms within our sample have invested during the years under investigation. We then proceeded by identifying all investors of these portfolio companies for each investment round. We coupled the investors that have invested in a company on the same round, marking them as syndication partners and recording the investment as syndicated.
Syndication Role. We further refined the syndication measure by making a distinction between those investments where the investor acts as a lead investor and those where the investor is one of the coinvestors. We identified the lead investor as the one who makes the largest investment in a given round, and the coinvestor as the one who makes a smaller investment in the same round. Both lead and nonlead syndications were measured as frequencies, i.e., the ratio of lead or nonlead investments to the total number of new investments made within a given year.
The measures for the lead and nonlead syndication frequencies contained a potential downward bias. The measure for lead investor was based on the estimated investments made by investors. When the exact amounts are unknown, the Venture Economics database reports equally sized investments by all investors participating in an investment round. Thus, with equally sized investments, we were unable to distinguish the roles within the syndicate and thus, we tailed to recognize some of the investments where the investor acts as a lead or as a nonlead. However, since we were able to identify the lead investor in 66.2% of all investment rounds with more than one investor, the measure should still be powerful enough to capture the effect of lead investing.
Similarly, the measure for the syndication frequency is based on our ability to observe syndicated investments. According to Lerner (1994), investments making up an individual investment round are occasionally recorded on separate dates and thus, using these dates, we were unable to observe all syndications. Although the records of the Venture Economics database are continuously augmented and corrected, we acknowledged this potential bias. However, as both the measures for the syndication frequency and the lead syndication frequency were biased downward, our estimates were more conservative than they would have been with perfect observation.
Syndication Frequency (Direct Effects). We hypothesized that syndication has a moderating role on the optimal portfolio size and hence, has an effect on performance. Syndication may have also several direct effects on performance. First, the syndication of investments may increase the quality of the investments, both by increasing the amount of information venture capitalists receive concerning potential investment targets (Bygrave, 1987; Lerner, 1994) and by increasing the geographical reach of the venture capitalist (Lerner, 1995; Sorenson & Stuart, 2001). Having a larger pool of potential investments may increase the quality of the best proposals received, thus resulting in better investments. Second, the shared decision making of a syndicate is likely to further enhance the quality of investments (Brander, Amit, & Antweiler, 2002; Lerner, 1994). If syndicate partners independently review a proposal and decide to invest upon the approval of all partners, their decision is likely to be of better quality than the one made by an individual decision maker (Sah & Stiglitz, 1986; Wilson, 1968), thus resulting in better investments. Third, the syndication of investments increases the number of investors of a venture, thus potentially giving it an access to a larger pool of resources. A syndicate with multiple venture capitalists provides complementary skills and contacts that contribute to the assistance and governance of the venture (Brander et al., 2002; Lockett & Wright, 2001). Furthermore, the venture capitalist may use later stage syndication as a means of "window dressing," seeking association with successful investments (Lerner, 1994). We controlled these potential direct effects by including the ratio of syndicated investments to all investments as a control to the models. In addition, to control for the network effects of syndication, we measured the number of syndication partners of the focal firms (i.e., the number of investors with whom they have syndicated) in a given year and included this measure in the models.
Average Syndicate Size. The size of the syndicate may have an effect on the division of work within the syndicate and on the contribution of the investor group. On average, the larger the investor group, the smaller the share of value-added workload that is left for each investor. Thus, we controlled for the effects of the syndicate size by measuring the average size of the syndicate at the initial investment.
Investments in New Companies. The IPOs are generated from the pool of new companies that the venture capitalists invest in yearly. The number of these new companies forms the upper limit for the number of IPOs. Thus, to control the venture capitalists' yearly potential for IPOs from new investments, we included the number of new portfolio companies in regressions.
Capital under Management. The main resources of a VC company are the capital to be invested, information on potential investment targets, and the partners managing the firm and its investments (Bygrave, 1987). The scarcity or availability of financial resources is likely to have a significant effect on the operations of the firm and thus, they needed to be controlled in the model.
When raising funds, venture capitalists negotiate capital commitments from limited partners and invest them gradually over a few years in promising target companies. The total amount of these commitments in a firm is referred to as "capital under management." We calculated the amount as a sum of the nonexpired VC funds and excluded funds that were raised for investments in buyouts, for example. We included only funds that Venture Economics classifies as "VC" in our sample. We further assumed that a fund would expire 10 years after the raising of the fund was completed, which Sahlman (1990) found to be the case in 72% of the funds in his sample. Thus, the total size of a fund was calculated as part of capital under management for 10 years after the fund was raised.
Age of VC Firm. The age of VC firms greatly affects its operations. The older the firm, the more contacts, experience, and prominence it has. Moreover, the younger the firm, the more it tries to establish a reputation by opportunistically striving toward successful exits. This is a phenomenon called "grandstanding" (Gompers, 1996). We control for the age effects. We calculate the age of each firm in our sample based on the founding dates in the Venture Economics database. We crosschecked the validity of these records using the back issues of Pratt's Guide to VC Sources. In some rare and ambiguous cases, we found that Venture Economics had allocated VC investments to the firm before the reported founding date. In these cases, we set the founding year of a company equal to the year of its first investment.
Investment Stage Mix of Investments. The nature of an investment target and the stage of the development of a venture affect the risk of investment, the expected time to exit, and the nature of involvement required in the investment. To control for these effects, we calculated the percentage of investments made in seed, early, expansion, and later stage companies as well as in acquisitions and buyouts for each firm each year according to Venture Economics classifications. In addition, Venture Economics classifies a proportion of investments to the "other/unknown" category containing unknown and special situations, mostly later stage and public market investments. We also included this category in the analysis. Furthermore, to differentiate between investment strategies with respect to the stage of investment, we calculated the Herfindahl index for the stage shares. With this stage focus index, we were able to control the effects of stage-focused investment strategies.
Average Round of Company at Entry. While the investment stage mix captures the effect of the stage of the investments in the portfolio, we measured the average round the venture capitalists entered their portfolio companies in order to control for the investment structure of the portfolio companies. If the initial entry occurs through syndication on a later round, the role of the investor and the interaction between the investors and the venture are substantially different from those of first-round investors.
Industry Mix of Investments. We also controlled for potential industry effects. We used the classification of the Venture Economics database and percentage-of-investments variables for the following industry sectors: communications, computer hardware, computer software, semiconductors/electronics, internet communications, internet/computer related, medical, biotechnology, and non-high-technology ventures as recorded by Venture Economics. To measure the degree of specialization in specific industries, we also calculate the Herfindahl index for the portfolio with respect to the industry sectors.
Time-dependency. As our sample was a time series of cross sections, it was necessary to control for differences between sample years. In all regressions, we included dummy variables for the sample years.
We tested the hypotheses using longitudinal data from 94 firms over a period of 13 years, with a count variable as the dependent variable. We modeled the dependent variable, the number of IPOs, as a Poisson model, analyzing the regressions using the generalized estimating equations (GEE) methodology (Liang & Zeger, 1986).
Utilizing a Poisson regression is the standard approach to analyze count data models. However, the Poisson regression assumes that the mean count of events equals the variance of the number of events. We tested this assumption in our data, (1) and found the potential effects of overdispersion statistically insignificant. Furthermore, we tested an alternative, negative binomial model that relaxes the assumption of the equality (Cameron & Trivedi, 1986; Greene, 2000). The analysis did not produce any qualitative differences.
To control for the heterogeneity of individuals, we used a population-averaged approach, i.e., averaged partial effects, (Baltagi, 1995; Hsiao, 1986; Woolridge, 2002). We estimated the models using Stata. To control for the expected existence of heteroscedasticity in the data, we estimated the GEE models using robust option, providing White heteroscedasticity consistent estimates.
Table 1 reports the descriptive statistics and correlations for the variables in the data of 997 observations over 94 firms. The statistics are from the whole sample and, as such, they pool observations across firms and years. This pooling contributes to the reported variation of observations.
The indicators of the size of the VC firm--the number of partners and the number of new investments made--reflect both the variation and similarity of firms in the sample. The standard deviation (SD) of the number of partners, 2.65, over the mean of 5.54, demonstrates that the size of VC firms is relatively homogenous. However, as the range shows, the largest firms have up to 22 partners, indicating existence of alternative organizational structures. The number of new investments made yearly shows larger variation with a mean of 18.96 and an SD of 14.76, with the largest number being 95. This reflects both differences in the strategies as well as the variation due to the pooling of firm-year observations. However, as the average number of companies per partner, 6.58, and its SD, 3.70, show, notably large portfolios are a clear minority. Thus, our assumption appears valid; venture capitalists typically operate portfolios of a size that allows them to be involved in their portfolio companies.
The mean syndication frequency is 80%, suggesting that venture capitalists cooperate extensively by syndicating their investments. Higher levels of syndication are more common than lower levels. When decomposed to lead and nonlead syndication, the lead syndication ratio, 15%, is slightly more than a third of the nonlead syndication ratio, 43%. This is in line with the average syndicate size of 4.22 investors. The difference between the overall frequency of syndication and the sum of lead and nonlead syndications results from both the inability to observe roles in all syndicates and the fact that for each lead investor, there is typically more than one coinvestor.
Performance and Allocation of Attention
Table 2 presents the results of the population-averaged GEE regression analysis on the number of portfolio companies that ultimately reached the IPO. The nonstandardized regression coefficients are presented with the corresponding standard errors. Although not presented in the table, all regression analyses include year dummy variables to control for potential differences between years.
The first model in Table 2 is the base model testing the effects of the control variables. The sized related controls--the amount of capital under management, the number of partners and the number of new investments--have an expected positive impact on the number of IPOs. Interestingly, the frequency of syndication does not appear to affect the performance directly. In addition, neither the number of syndication partners nor the average size of the syndicates has a significant effect. As expected, the average round when the venture capitalists enter their investment has a strong positive effect on the number of IPOs from these investments.
In hypothesis 1, we suggested that there is a curvilinear relationship between the number of investments per partner and the performance of the VC firm. Models 2 and 3 of Table 2 present the analyses testing the hypothesis. The first model includes the linear effect of the number of companies per partner, and the second introduces the quadratic effect. The coefficient of the linear term is positive, whereas the coefficient of the quadratic term is negative. Both terms turn out to be highly significant. Thus, the performance of the VC partnership increases as the relative size of the portfolio increases, but the rate of this increase is diminishing. In order to see whether this deceleration turns the total effect negative in the range of the observed portfolio sizes, we differentiate the model with respect to the number of companies per partner. The first-order derivative is zero when the companies per partner ratio is 13 (in model 3 of Table 2). As the centered measure for the range of companies per partner is from -6.33 to 28.41, the model has its optimum within that range. We emphasize that this number is contingent upon the estimated model, the sample, and the period under investigation. Thus, the result is merely a validation of the existence of an optimum rather than a prescriptive objective for the size of the portfolio. The importance of the result is that there exists an optimal portfolio size beyond which the inclusion of additional portfolio companies results in negative marginal returns.
Moderating Effect of Syndication
The previous analysis demonstrates the inverted U-shaped relationship between the number of companies per partner and the performance of the venture capitalist. In hypothesis 2, we suggested that the syndication of VC investments moderates the optimal number of companies per partner in such way that the more syndication there is, the higher the optimal number of companies per partner. In model 4, we include an interaction term of the syndication frequency and the number of portfolio companies per partner. The effect of interaction on the performance is positive and significant.
The effect of the interaction term on the optimal number of portfolio companies per partner is twofold. First, as the syndication frequency increases, the optimum moves to the right on the axis of the number of companies per partner, thus increasing the optimal number of companies per partner. Second, the number of IPOs from investments in the optimum increases as the syndication frequency increases. Thus, the increased frequency of syndication increases both the number of portfolio companies a partner is able manage optimally and the number of IPOs these companies produce.
Although syndication increases both the optimal size and the performance at this optimum, when syndicating as a lead investor, the benefits of work sharing are mitigated by the workload of the lead investor. Thus, in hypothesis 3, we suggested that the amount of investments syndicated as lead investors negatively moderates the optimum, lowering the benefits of syndication. Model 5 tests the hypothesis by including the interaction terms of the companies per partner and the frequency of lead and nonlead syndication.
The interaction between lead syndication frequency and the number of companies per partners does not have a statistically significant coefficient, thus failing to support hypothesis 3. However, the nonlead syndication frequency positively moderates the optimal size of the portfolio on a significant level. The more a venture capitalist syndicates as a nonlead investor, the higher the optimal portfolio size. Thus, it appears that the positive effect of syndication on the portfolio size is mainly due to the benefits of reduced workload when participating in a syndicate in a nonlead role.
To test the robustness of our results, we repeated the previous analyses using the alternative performance measure of all positive exits resulting from the new investments. (2) While our results on optimal portfolio size were directly supported with the alternative variable, the effects of syndication on the all-positive exits were slightly different, although the qualitative results remained the same. (3) With the all-positive exits as the dependent variable, the moderating effect of syndication was not significant (corresponding to model 4 in Table 2). However, when we repeated the analyses dividing the syndication between lead and nonlead syndication, the coefficient of lead syndication moderation became statistically significant, having a negative effect on performance, while nonlead syndication was statistically insignificant. Thus, when using this alternative measure, the increased frequency of syndication in the role of a lead investor lowered the optimal size of the portfolio, suggesting a cost for having the lead investor role.
In addition to the alternative performance measure, we also tested the robustness of the choice of the moderating effects, as the optimum is determined both by the linear and quadratic term of the portfolio size. Analyzing the number of IPOs, we tested the moderating effect of syndication using an interaction term between the syndication frequency and the number of companies per partner squared. In the case of overall syndication effect, the interaction between the syndication frequency and quadratic term was not significant, indicating that the linear interaction term captures the moderating effect. When we divided the syndication between lead and nonlead syndication, the interaction effects with the number of companies per partner squared were identical to the first-order interactions. That is, the nonlead syndication was positive at a significant level and lead syndication did not have a statistically significant effect. However, when we included both the first-order and second-order interaction effects in the same regression, only the first-order nonlead syndication was significant and negative. Thus, our analyses yielded identical results regardless of our choice of interaction terms.
Discussion and Conclusions
In this article, we set out to examine the limitations to the allocation and value of the venture capitalist attention. While earlier research has validated the value of venture capitalists' involvement in an individual venture, the link between the allocation of attention and the performance of the venture capitalist had not been addressed in prior research.
We argued that the increasing size of the portfolio divides the venture capitalist's attention, producing two countering effects. On one hand, the larger the size of the portfolio, the more firms the venture capitalist will have that can potentially reach desirable outcomes. On the other hand, the larger the size of the portfolio, the less time the venture capitalist has for being involved in portfolio companies, and the smaller the value of this involvement. The overheads created by managing the individual ventures, as well as the diversity of the portfolio, reduce the total time available for managing the portfolio. With these arguments, we hypothesized that there exists a curvilinear inverted U-shaped relationship between the performance and the number of companies per partner and that this relationship is positively moderated by the frequency of syndication. The hypotheses were tested utilizing a data set on the 94 leading U.S. VC firms and their investments during 1986-1998.
Our hypotheses are supported by the results of the analysis, suggesting that there exists an optimal portfolio size with respect to the number of companies per partner. As the number of companies per partner increases, the performance shows a corresponding increase, although with diminishing marginal returns. When the number of investments exceeds the optimum, performance starts to deteriorate and leads to negative total returns. Given the shape of the relationship, there exists an optimal number of investments for a venture capitalist to manage. This optimum is moderated by the syndication activity. The more the venture capitalist syndicates its investments, the higher the size of the portfolio it can manage optimally. Additional analyses indicated different benefits from syndication depending on the role of the investor in syndicates. Supporting hypothesis 3, our analyses showed that acting as a nonlead investor is more beneficial compared with acting as a lead investor.
The foremost implication of our results is that the venture capitalist attention is valuable to the portfolio companies. Should the venture capitalist merely pick and choose the best investment targets with no postinvestment contribution, the outcomes of the portfolio companies would be independent. In this case, we would observe potentially diminishing returns on the relative portfolio size, resulting from the decreasing quality of the investments. However, the existence of an optimal portfolio size demonstrates that with a large enough portfolio, the marginal return on each additional portfolio company turns negative. Thus, the success of a portfolio company is not independent of the other ventures in the portfolio. Each additional venture affects the outcomes of existing portfolio companies by reducing the amount of attention available.
As the restrictions to the portfolio size stem from the limited attention of the venture capitalist, sharing the workload of managing the investments relaxes these constraints. According to our results, syndication serves as a mechanism to loosen the limitations and to increase the size of the portfolio. The increased portfolio size enables the venture capitalist to diversify its portfolio more widely than when investing alone. The difference in the effects of acting as a lead versus nonlead investor in syndicates suggests that reciprocity between investors, which is commonly observed (Wright & Lockett, 2003), is important in the syndication of VC investments.
The theoretical implications contribute to the increasing literature on the optimal size of the portfolio. By confirming the existence of optimal size, it provides both support for earlier theoretical models and an extension to these approaches. By relating the allocation of attention to the outcome of the portfolio, we extend the prevailing theoretical stream that approaches the issue from contracting and incentives (Gifford, 1997; Kanniainen & Keuschnigg, 2003). As our performance measure is the number of IPOs (controlling for the number of investments), we are not able to capture the more subtle effects of contracting on the profits of venture capitalists. However, our results indicate a size of portfolio that maximizes the outcomes for the portfolio companies as a group. Thus, our results are complementary to the existing research and also contribute to the empirical studies on the portfolio size and its determinants and variations (Cumming, 2006; Elango et al., 1995; Murray & Marriott, 1998; Sahlman, 1990). Furthermore, our findings expand the research by introducing the moderating role of syndication on the optimal size of the portfolio.
Our findings also have important practical implications. The results provide valuable guidance for venture capitalists on successful strategies and resource allocation. Venture capitalists should carefully consider how many companies they can manage and still add value. Too few companies imply that the venture capitalist's value-adding potential is not fully used. However, exceeding the optimal number of companies results in deteriorated performance as the venture capitalist's value-adding activities become too fragmented. Syndication can be used to overcome resource constraints. However, given that acting as a lead investor consumes almost as much resources as investing alone, reciprocity should be endorsed and/or taking the lead role should be compensated for in the syndication of VC investments.
Similarly, entrepreneurs seeking financing should consider the limitations that their investors face. While the results provide further evidence that venture capitalists add value to the ventures beyond the capital, the level of involvement from the venture capitalist affects the value added. Thus, observing and assessing the venture capitalist's ability to manage its investments is of importance also for the ventures. The larger the existing portfolio of the venture capitalist, the less value one can expect from the investor. In addition, the results validate the opportunity cost of venture capitalist attention. As suggested by Gifford (1997), with the opportunity cost of attention, the venture capitalists' incentives do not necessarily match those of the entrepreneur. It is not always optimal for venture capitalists to maximize the value of an individual venture, as this may decrease the value of other portfolio firms.
Although our approach enables us to examine the performance of the venture capitalist, it has limitations that are consequently a source of opportunities for further research. First, as we measure the performance on the level of VC firm with the number of IPOs, we are able to capture the optimum for the group of portfolio ventures, but we fail to resolve some of the finer details. Thus, more sophisticated measures of performance are needed to address the question of division of the proceeds from an IPO. The emerging literature on risk and return in private equity (e.g., Cochrane, 2005) can provide new opportunities to examine these aspects in the coming years.
Second, our sample contains the leading U.S. venture capitalists and thus, it may not be representative of the whole VC industry. According to Rosenstein et al. (1993), only the top 20 venture capitalists in their sample seemed to add value to their investments. Although this questions the generalizability of the results regarding the value of venture capitalist involvement, the restrictions that the venture capitalist faces in involvement apply even if the involvement would not be as valuable as that of the best venture capitalists.
Third, while our results provide support for the value of the venture capitalist's attention, it leaves open the question of the relative importance of the forms of involvement, as the involvement is both motivated by assistance and by governance. The dominant source of value added has implications for both the ventures and venture capitalists. Should the value of involvement stem from governance, the entrepreneur may be better off with less involvement, as this would mean less intense monitoring. The opposite would apply if the value stems from assistance, which makes involvement more desirable. Our results do not differentiate between these two alternatives, and further research on the issue is called for.
To conclude, this article is one of the first to empirically link the organization of investment activities to the performance of VC firms. Building on and extending prior theory, we developed and empirically tested hypotheses arguing that there exists a curvilinear inverted U-shaped relationship between the number of companies per partner and the VC firm performance, and that this relationship is positively moderated by the frequency of syndication and the role in syndicates. In so doing, we hope our article stimulates further research on the strategies and success factors of venture capitalists.
The authors would like to thank a number of individuals, who provided valuable advice on previous versions of the article, including Ari Hyytinen and the participants of the 2002 Babson College--Kauffman Foundation Entrepreneurship Research Conference, Boulder, CO; the 22nd Annual International Conference of the Strategic Management Society, Paris, France; and the Entrepreneurship Research Workshop of UNIEI 2004. Mikko Jaaskelainen also acknowledges the Jenny and Antti Wihuri Foundation and the National Technology Agency of Finland (TEKES), for the financial support provided to the research. An earlier version of the article was published in the Frontiers of Entrepreneurship Research 2002.
Admati, A.R. & Pfleiderer, P. (1994). Robust financial contracting and the role of venture capitalists. Journal of Finance, 49(2), 371-402.
Akerlof, G.A. (1970). The market for "lemons": Quality uncertainty and the market mechanism. Quarterly Journal of Economics, 84(3), 488-500.
Baltagi, B.H. (1995). Econometric analysis of panel data. Chichester, UK: John Wiley & Sons Ltd.
Barney, J.B., Busenitz, L.W., Fiet, J.O., & Moesel, D.D. (1996). New venture teams' assessment of learning assistance from venture capital firms. Journal of Business Venturing, 11(4), 257-272.
Bergemann, D. & Hege, U. (1998). Venture capital financing, moral hazard, and learning. Journal of Banking & Finance 22(6-8), 703-735.
Brander, J.A., Amit, R., & Antweiler, W. (2002). Venture capital syndication: Improved venture selection vs. the value-added hypothesis. Journal of Economics & Management Strategy, 11(3), 423-452.
Busenitz, L.W., Fiet, J.O., & Moesel, D.D. (2004). Reconsidering the venture capitalists' "value added" proposition: An interorganizational learning perspective. Journal of Business Venturing, 9(6), 787-807.
Busenitz, L.W., Moesel, D.D., & Fiet, J.O. (1997). The impact of post-funding involvement by venture capitalists on long-term performance outcomes. In P.D. Reynolds, W.D. Bygrave, N.C. Carter, W.B.P. Davidsson, C.M. Gartner, Mason, & P.P. McDougall (Eds), Frontiers of entrepreneurial research (pp. 498-512). Wellesley, MA: Babson College.
Bygrave, W.D. (1987). Syndicated investments by venture capital firms--a networking perspective. Journal of Business Venturing, 2(2), 139-154.
Bygrave, W.D. (1988). The structure of the investment networks of venture capital firms. Journal of Business Venturing, 3(2), 137-157.
Bygrave, W. & Timmons, J.A. (1992). Venture capital at the crossroads. Boston, MA: Harvard Business School Press.
Cameron, A.C. & Trivedi, P.K. (1986). Econometric models based on count data: Comparisons and applications of some estimators and tests. Journal of Applied Econometrics, 1, 29-53.
Chan, Y. (1983). On the positive role of financial intermediation in allocation of venture capital in a market with imperfect information. Journal of Finance, 38(5), 1543-1568.
Cochrane, J.H. (2005). The risk and return of venture capital. Journal of Financial Economics, 75(1), 3-52.
Cumming, D.J. (2006). The determinants of venture capital portfolio size: Empiric evidence. Journal of Business, 79(3).
Cumming, D.J. & MacIntosh, J.G. (2003). A cross-country comparison of full and partial venture capital exits. Journal of Banking & Finance, 27(3), 511-548.
Ehrlich, S.B., Denoble, A.F., Moore, T., & Weaver, R.R. (1994). After the cash arrives--a comparative-study of venture capital and private investor involvement in entrepreneurial firms. Journal of Business Venturing, 9(1), 67-82.
Eisenhardt, K.M. (1989). Agency theory: An assessment and review. Academy of Management Journal, 14(1), 57-74.
Elango, B., Fried, V.H., Hisrich, R.D., & Polonchek, A. (1995). How venture capital firms differ. Journal of Business Venturing, 10(2), 157-179.
Fama, E.F. & Jensen, M.C. (1983). Separation of ownership and control. Journal of Law & Economics, 16, 301-325.
Fried, V.H. & Hisrich, R.D. (1995). The venture capitalist--a relationship investor. California Management Review, 37(2), 101-113.
Gifford, S. (1997). Limited attention and the role of the venture capitalist. Journal of Business Venturing, 12(6), 459-482.
Gompers, P.A. (1995). Optimal investment, monitoring, and the staging of venture capital. Journal of Finance, 50(5), 1461-1489.
Gompers, P.A. (1996). Grandstanding in the venture capital industry. Journal of Financial Economics, 42(1), 133-156.
Gompers, P.A. & Lerner, J.A. (1999). Venture capital cycle. Cambridge, MA: MIT Press.
Gorman, M. & Sahlman, W.A. (1989). What do venture capitalists do? Journal of Business Venturing, 4(4), 231-248.
Grant, R.M. (1996). Toward a knowledge-based theory of the firm. Strategic Management Journal, 17(Special Issue: Knowledge and the Firm), 109-122.
Greene, W.H. (2000). Econometric analysis. Upper Saddle River, NJ: Prentice-Hall, Inc.
Hellmann, T. & Puri, M. (2002). Venture capital and the professionalization of start-up firms: Empiric evidence. Journal of Finance, 57(1), 169-197.
Higashide, H. & Birley, S. (2000). Value created through the socially complex relationship between the venture capitalist and the entrepreneurial team. In P.D. Reynolds, E. Autio, C.G. Brush, W.D. Bygrave, S. Manigart, H.J. Sapienza, & K.G. Shaver (Eds), Frontiers of entrepreneurship research (pp. 308-319). Wellesley, MA: Babson College.
Hsiao, C. (1986). Analysis of panel data. Cambridge: Cambridge University Press.
Jensen, M.C. & Meckling, W.H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305-360.
Kanniainen, V. & Keuschnigg, C. (2003). The optimal portfolio of start-up firms in venture capital finance. Journal of Corporate Finance, 9, 521-534.
Kogut, B. & Zander, U. (1992). Knowledge of the firm, combinative capabilities, and the replication of technology. Organization Science, 3(3), 383-397.
Lerner, J. (1994). The syndication of venture capital investments. Financial Management, 23(3), 16-27.
Lerner, J. (1995). venture capitalists and the oversight of private firms. Journal of Finance, 50(1), 301-318.
Liang, K. & Zeger, S. (1986). Longitudinal data analysis using generalized linear models. Biometrika, 73(1), 13-22.
Lockett, A. & Wright, M. (2001). The syndication of venture capital investments. Omega-International Journal of Management Science, 29(5), 375-390.
Macmillan, I.C., Kulow, D.M., & Khoylian, R. (1989). Venture capitalists involvement in their investments--extent and performance. Journal of Business Venturing, 4(1), 27-47.
Murray, G.C. & Marriott, R. (1998). Why has the investment performance of technology-specialist, European venture capital funds been so poor? Research Policy, 27(9), 947-976.
Norton, E. & Tenenbaum, B.H. (1993). Specialization versus diversification as a venture capital-investment strategy. Journal of Business Venturing, 8(5), 431-442.
Podolny, J.M. (2001). Networks as the pipes and prisms of the market. American Journal of Sociology, 107(1), 33-60.
Polanyi, M. (1967). The tacit dimension. Chicago: University of Chicago Press.
Prahalad, C.K. & Bettis, R.A. (1986). The dominant logic: A new linkage between diversity and performance. Strategic Management Journal, 7, 485-501.
Rosenstein, J., Bruno, A.V., Bygrave, W.D., & Taylor, N.T. (1993). The CEO, venture capitalists, and the board. Journal of Business Venturing, 8(2), 99-113.
Sah, R.K. & Stiglitz, J.E. (1986). The architecture of economic system: Hierarchies and polyarchies. American Economic Review, 76(4), 716-727.
Sahlman, W.A. (1990). The structure and governance of venture-capital organizations. Journal of Financial Economics, 27(2), 473-521.
Sapienza, H.J. (1992). When do venture capitalists add value. Journal of Business Venturing, 7(1), 9-27.
Sapienza, H.J., Amason, A.C., & Manigart, S. (1994). The level and nature of venture capitalist involvement in their portfolio companies: A study of three european countries. Managerial Finance, 20(1), 3-17.
Sapienza, H.J. & Gupta, A.K. (1994). Impact of agency risks and task uncertainty on venture capitalist-CEO interaction. Academy of Management Journal, 37(6), 1618-1632.
Sapienza, H.J., Manigart, S., & Vermier, W. (1996). Venture capitalist governance and value added in four countries. Journal of Business Venturing, 11(6), 439-469.
Sorenson, O. & Stuart, T.E. (2001). Syndication networks and the spatial distribution of venture capital investments. American Journal of Sociology, 106(6), 1546-1588.
Steier, L. & Greenwood, R. (1995). Venture capitalist relationships in the deal structuring and post-investment stages of new firm creation. Journal of Management Studies, 32(3), 337-357.
Wasserman, N. (2002). The non-division of labor: Knowledge separability., structure, and the upside-down venture capitalist. Working paper: Harvard Business School.
Wilson, R. (1968). The theory of syndicates. Econometrica, 36(1), 119-132.
Woolridge, J. M. (2002). Econometric analysis of cross section and panel data. Cambridge, MA: MIT Press.
Wright, M. & Lockett, A. (2003). The structure and management of alliances: Syndication in the venture capital industry. Journal of Management Studies, 40(8), 2073-2102.
Mikko Jaaskelainen is a researcher at the Institute of Strategy and International Business at Helsinki University of Technology, Helsinki, Finland.
Markku Maula is a professor of Venture Capital at the Institute of Strategy and International Business at Helsinki University of Technology, Helsinki, Finland.
Tuukka Seppa is a manager with The Boston Consulting Group, Helsinki, Finland.
(1.) To confirm this statistically, we calculated the Lagrange multiplier test for overdispersion. It tests the Poisson assumption of the equality of mean and variance as [H.sub.0] against the negative binomial model (Cameron & Trivedi, 1986; Greene, 2000). For the test, we estimated the base model in Table 2. The [chi square] (1)--distributed test statistic had a value of 3.593 or the [H.sub.0] of [alpha] = 0; thus, we were unable to reject the hypothesis with a p-value of 0.058. Therefore, the overdispersion in the data is not significant, making the Poisson model applicable.
(2.) We analyzed the number of all successful exits using a negative binomial regression with fixed effects. The population-averaged Poisson model was not applicable due to overdispersion of the variable.
(3.) Some potential factors explaining the differences include (a) whereas IPOs are most often considered a successful outcome, mergers and acquisitions include unsuccessful exits also, causing investor losses; and (b) due to the statistical properties of this variable, some of the assumptions used in the main analysis were not valid (the assumption concerning over dispersion), forcing us to use an alternative specification (negative binomial instead of Poisson).
Please send correspondence to: Mikko Jaaiskelainen, e-mail: Mikko.Jaaskelainen@hut.fi at P.O. Box 5500, FIN-02015 HUT, Finland.
Table 1 Descriptive Statistics Mean SD Min 1. Number of IPOs 2.38 2.42 0 2. Number of partners 5.54 2.65 1 3. Number of new investments 18.96 14.76 1 4. Companies/partner ([dagger]) 6.58 3.70 0.25 5. Syndication frequency ([dagger]) 0.80 0.19 0.00 6. Lead syndication frequency 0.15 0.12 0.00 ([dagger]) 7. Nonlead syndication 0.43 0.21 0.00 frequency ([dagger]) 8. Capital under management 207.48 198.41 3.40 ([double dagger]) 9. Age ([double dagger]) 16.29 9.01 0.50 10. Average investment round at 1.96 1.20 0.00 entry ([double dagger]) 11. Average size of syndicate 4.22 1.75 1.00 12. Number of syndication 44.62 34.75 0.00 partners 13. Stage specialization index 0.33 0.15 0.18 14. Industry specialization index 0.31 0.18 0.13 Max 1 2 1. Number of IPOs 14 -- 2. Number of partners 22 .43 -- 3. Number of new investments 95 .69 .51 4. Companies/partner ([dagger]) 35.00 .33 -.25 5. Syndication frequency ([dagger]) 1.00 .02 -.05 6. Lead syndication frequency 1.00 .l6 .19 ([dagger]) 7. Nonlead syndication 1.00 .11 -.02 frequency ([dagger]) 8. Capital under management 1,929.20 .44 .55 ([double dagger]) 9. Age ([double dagger]) 51.50 .09 .32 10. Average investment round at 11.00 .18 .06 entry ([double dagger]) 11. Average size of syndicate 13.66 .09 -.05 12. Number of syndication 234.00 .59 .37 partners 13. Stage specialization index 1.00 -.30 -.23 14. Industry specialization index 1.00 -.35 -.31 3 4 5 1. Number of IPOs 2. Number of partners 3. Number of new investments -- 4. Companies/partner ([dagger]) .50 -- 5. Syndication frequency ([dagger]) -.01 .01 -- 6. Lead syndication frequency .23 .07 -.01 ([dagger]) 7. Nonlead syndication .11 .18 .54 frequency ([dagger]) 8. Capital under management .58 .17 -.08 ([double dagger]) 9. Age ([double dagger]) .18 .01 .00 10. Average investment round at .15 .14 .08 entry ([double dagger]) 11. Average size of syndicate .03 .12 .66 12. Number of syndication .81 .48 .30 partners 13. Stage specialization index -.42 -.29 .04 14. Industry specialization index -.48 -.34 .01 6 7 8 1. Number of IPOs 2. Number of partners 3. Number of new investments 4. Companies/partner ([dagger]) 5. Syndication frequency ([dagger]) 6. Lead syndication frequency -- ([dagger]) 7. Nonlead syndication -.15 -- frequency ([dagger]) 8. Capital under management .22 -.05 -- ([double dagger]) 9. Age ([double dagger]) -.02 -.12 .27 10. Average investment round at .03 .19 .10 entry ([double dagger]) 11. Average size of syndicate -.11 .66 -.09 12. Number of syndication .16 .42 .41 partners 13. Stage specialization index -.15 -.17 -.32 14. Industry specialization index -.13 -.22 -.37 9 10 11 1. Number of IPOs 2. Number of partners 3. Number of new investments 4. Companies/partner ([dagger]) 5. Syndication frequency ([dagger]) 6. Lead syndication frequency ([dagger]) 7. Nonlead syndication frequency ([dagger]) 8. Capital under management ([double dagger]) 9. Age ([double dagger]) -- 10. Average investment round at -.03 -- entry ([double dagger]) 11. Average size of syndicate -.09 .20 -- 12. Number of syndication .08 .24 .51 partners 13. Stage specialization index -.05 -.30 -.08 14. Industry specialization index -.12 -.32 -.15 12 13 14 1. Number of IPOs 2. Number of partners 3. Number of new investments 4. Companies/partner ([dagger]) 5. Syndication frequency ([dagger]) 6. Lead syndication frequency ([dagger]) 7. Nonlead syndication frequency ([dagger]) 8. Capital under management ([double dagger]) 9. Age ([double dagger]) 10. Average investment round at entry ([double dagger]) 11. Average size of syndicate 12. Number of syndication -- partners 13. Stage specialization index -.40 -- 14. Industry specialization index -.48 .66 -- Note: The table reports descriptive statistics and correlations for the dependent and independent variables (94 firms, 997 firm-year observations). All correlations higher than .06 are significant on the level .05. The table excludes industry, stage, and location controls as well as dummy variables for years. These variables do not correlate with other variables on a level higher than 0.4. Mean, SD, and range are reported for nontransformed observations, while correlations are estimated for' centered and = logarithmic transformations that are used in regressions. SD, standard deviation; IPO, initial public offering. Table 2 The Performance of Venture Capitalists Relative to the Number of Companies per Partner and the Moderating Effect of Syndication Number of IPOs out of new company investments/year 1986-1998 Model 1 Model 2 Companies/ .023 * (.01) partner (Companies/ partner) (2) Syndication frequency x companies/ partner Lead syndication frequency x companies/ partner Nonlead syndication frequency x companies/ partner Syndication -.166 (.23) -.160 (.22) frequency Lead syndication frequency Nonlead syndication frequency Number of new .019 *** (.00) .016 *** (.00) investments Number of VC .021 (.01) .045 ** (.02) partners Capital under .116 ([dagger]) (.06) .111 ([dagger]) (.06) management (log) Avg. .335 *** (.06) .340 *** (.06) investment round at entry Avg. size of -.008 (.03) -.006 (.03) syndicate Number of .003 (.00) .003 (.00) syndication partners Age (log) -.123 ([dagger]) (.06) -.143 * (.06) Stage -1.227 *** (.34) -1.236 *** (.34) specialization index Industry -.850 ** (.30) -.771 ** (.29) specialization index N 997 997 Wald 923.66 992.27 df 37 38 Probability 0.00 0.00 Number of IPOs out of new company investments/year 1986-1998 Model 3 Model 4 Companies/ .052 *** (.01) .048 *** (.01) partner (Companies/ -.002 ** (.00) -.002 * (.00) partner) (2) Syndication .112 *** (.04) frequency x companies/ partner Lead syndication frequency x companies/ partner Nonlead syndication frequency x companies/ partner Syndication -.173 (.22) -.121 (.21) frequency Lead syndication frequency Nonlead syndication frequency Number of new .014 ** (.00) .016 *** (.00) investments Number of VC .061 ** (.02) .059 *** (.02) partners Capital under .101 (.06) .102 ([dagger]) management (.06) (log) Avg. .354 *** (.06) .350 *** (.06) investment round at entry Avg. size of .001 (.03) .004 (.03) syndicate Number of .002 (.00) .001 (.00) syndication partners Age (log) -.158 * (.07) -.147 * (.06) Stage -1.190 *** (.34) -1.167 *** (.341 specialization index Industry -.673 * (.29) -.720 * (.30) specialization index N 997 997 Wald 1.09 1,095.82 df 39 40 Probability 0.00 0.00 Number of IPOs out of new company investments/year 1986-1998 Model 5 Companies/ .048 *** (.01) partner (Companies/ -.002 * (.00) partner) (2) Syndication frequency x companies/ partner Lead .036 (.06) syndication frequency x companies/ partner Nonlead .074 * (.03) syndication frequency x companies/ partner Syndication frequency Lead -.162 (.23) syndication frequency Nonlead -.006 (.16) syndication frequency Number of new .016 *** (.00) investments Number of VC .060 *** (.02) partners Capital under .101 (.06) management (log) Avg. .363 *** (.06) investment round at entry Avg. size of -.011 (.03) syndicate Number of .001 (.00) syndication partners Age (log) -.159 * (.07) Stage -1.261 *** (.34) specialization index Industry -.729 * (.30) specialization index N 997 Wald 1,474.42 df 42 Probability 0.00 * significant at the .05 level. ** significant at the .01 level. *** significant at the .001 level. ([dagger]) significant at the .1 level. We use one-tailed tests for hypothesized relationships and two-tailed tests for controls. Note: The dependent variable is the number of initial public offerings from investments made in new companies each year. The analysis applies a population-averaged Poisson model estimated with generalized estimating equations. Nonstandardized regression coefficients and the corresponding standard errors (in parentheses) are presented; year dummies are included in the analysis but not reported. Table excludes industry, stage, and location controls as well as dummy variables for years. IPO, initial public offering; VC, venture capital; df, degree of freedom.
|Printer friendly Cite/link Email Feedback|
|Author:||Jaaskelainen, Mikko; Maula, Markku; Seppa, Tuukka|
|Publication:||Entrepreneurship: Theory and Practice|
|Date:||Mar 1, 2006|
|Previous Article:||Adverse selection and capital structure: evidence from venture capital.|
|Next Article:||Venture capital investment strategy and portfolio failure rate: a longitudinal study.|