Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Foreign direct investment in Iran wikipedia , lookup
History of investment banking in the United States wikipedia , lookup
History of private equity and venture capital wikipedia , lookup
Private equity in the 2000s wikipedia , lookup
Investment management wikipedia , lookup
Venture capital wikipedia , lookup
Private equity in the 1980s wikipedia , lookup
Corporate venture capital wikipedia , lookup
Cross-Cultural Liaisons: Syndication Partner Choice by Foreign Venture Capital Firms in China March 2013 Xiao Huang School of Economics and Management Southeast University Nanjing, China & Martin Kenney Professor & Donald Patton Research Associate Department of Human and Community Development University of California, Davis [email protected] ABSTRACT Cross-border venture capital investing has increased dramatically over the last two decades. Previous research has shown that foreign venture capitalists tend to partner with local venture capitalists as a strategy for offsetting information asymmetries and liabilities of foreignness. This paper examines the choice of co-investors by foreign venture capitalists investing in the Chinese market. We find that foreign investors are more likely to choose Chinese investors in later investment rounds and in more mature portfolio firms. More experienced foreign venture capital firms are less likely to co-invest with Chinese venture capitalists. Also, having a Chinese office made foreign venture capitalists less likely to co-invest. In seed-stage investments, when risks are the greatest, foreign firms are least likely to co-invest with Chinese venture capitalists, while at the later stage, when risks are presumably the lowest is when they are most likely to co-invest. During this period, the propensity to co-invest with Chinese venture capitalists increased significantly in every stage except the seed stage. 1 1. Introduction The emergence of venture capital (VC) firms with global investments and offices creates a remarkable experiment to measure the degree to which foreign investors co-invest with local firms. China is a particularly interesting laboratory for studying the relationship between foreign and domestic VC investors, as its legal, financial, and social systems are significantly different from that of nearly any other major economy. China invariably scores low on Western indicators for transparency, investor protection, rule of law, etc., all of which should increase uncertainty. And yet, during the last fifteen years China has attracted an enormous number of non-Chinese VC investors. Venture capital investors are invariably faced with pervasive uncertainty regarding markets, competition, management teams, and often even the firm’s technology. Responses to these conditions include investment staging, intensive portfolio firm monitoring, and deal syndication (on staging, see Gompers, 1995; on monitoring, see Lerner, 1995; Kaplan and Stromberg, 2003; on syndication, see Brander et al., 2002; Lerner, 1994). Proximity can ease the costs of monitoring and there is a distinct local bias in venture investing (Cumming and Dai, 2010; Devigne et al., 2013; Florida and Kenney, 1988; Sorenson and Stuart, 2001). Despite the proclivity to invest locally, over the past two decades the VC industry has globalized its investment practice (Guler and Guillen, 2010; Mäkelä and Maula, 2005; Wright et al., 2005). Given the recognized importance of proximity, there is significant evidence that foreign investors are likely to co-invest with local investors, and one might expect this to be particularly true in the uncertain Chinese environment. The reasons for co-investment include 2 knowledge of local legal requirements that local VC firms possess (Mäkelä and Maula, 2006), and their experience in the domestic market (Wright et al., 2005). Not surprisingly, previous research on cross-border co-investment found that foreign firms do co-invest with local investors as this reduces risk and assists in firm monitoring (Mäkelä and Maula, 2006, 2008; Meuleman and Wright, 2011). Syndicates composed of domestic and foreign firms perform better than either those made up of purely international or purely local venture capitalists (Devigne et al., 2013; Chemmanur et al., 2013). More recently, Liu and Maula (2014) found that cross-border co-investment differs by the type of uncertainty. If there is portfolio firm-specific uncertainty, then co-investing with local venture capitalists increases; however market uncertainty deters co-investment with local firms. Because VC investment is a purchase of equity, its economics resemble those of partnerships in two respects. First, the investors become partners with the entrepreneur. Second, the co-investors are partners as they pay the same valuation for the firm. With respect to the co-investors, the portfolio firm and the round’s lead investor usually choose the co-investors. This choice can be seen as an optimization problem balancing two competing objectives. Choosing a similar VC firm will, most likely, bring co-investors that share similar values and goals, making them easier to work with, but they will likely have redundant knowledge, resources, and networks. In contrast, bringing dissimilar venture capitalists into the co-investment syndicate would increase the diversity of knowledge, resources, and networks, but there are more likely to be disagreements regarding strategy or free-rider problems. Generally, organizational theory suggests that in conditions of high uncertainty, where mutual trust and understanding ease stress, firms with similar backgrounds and 3 characteristics will be selected. In contrast, in conditions of greater certainty, co-investor selection could be less homogeneous because dissimilar co-investors would be supplemental in terms of the assets they contribute. Building upon Liu and Maula’s (2014) findings and using similar data, we test the micro-level determinants of co-investment choices by foreign firms investing in China. Put differently, our interest is in understanding the conditions under which foreign venture capitalists are willing to co-invest with Chinese venture capitalists. 2. Foreign venture capital firms’ choice for syndicate partner in foreign markets Venture capitalists invest in relatively risky non-publicly traded firms whose future performance is uncertain. Prior to investing, a VC firm must consider the characteristics of the recipient firm and possible investment partners. Because investing in an offshore firm adds an element of risk, most previous research has shown that foreign VC firms tend to coinvest with venture capitalists local to the recipient firm (Mäkelä and Maula, 2006, 2008; Meuleman and Wright, 2011; Devigne et al., 2013; Chemmanur et al., 2013). The primary explanation of the preference of foreign firms co-investing with local investors when entering an overseas markets is that this will mitigate the uncertainties due to information asymmetry and foreignness (Meuleman and Wright, 2011; Dai et al., 2012; Devigne et al., 2013; Chemmanur et al., 2013). There is a sizeable literature finding that social status and organizational similarity affect the choice of co-investors (Sorensen and Stuart, 2001, 2008; Hochberg et al., 2007; Hochberg et al., 2011; Wright and Lockett 2003). The studies of co-investors suggest that when uncertainty is higher, venture capitalists prefer to co-invest with firms that share similar 4 experiences, organizational structures, and goals, i.e. homophily. This tendency is mitigated by the benefits more diverse co-investors might bring in terms of different skills and networks that could assist the portfolio firm. Homophily should be greatest in earlier-stage investments, when uncertainty is highest. In choosing a co-investor, proximity, both geographic and social, plays a role. One of the most robust findings in the literature is that venture capitalists not located in the region of their portfolio firm are more likely to co-invest with firms located in the portfolio firm’s region (Sorensen and Stuart, 2001). Having a co-investor closer to the portfolio firm mitigates the difficulties of long-distance monitoring and advising (Jääskeläinen and Maula, 2013). In general, previous research strongly suggests that foreign venture capitalists should choose to co-invest with domestic venture capitalists. The motivation should be reduced, if the foreign VC firm has a local subsidiary (Zhang, 2011). Foreign VC firms may have co-investment relationships with other venture capitalists that they established in their home markets. Chinese firms would be expected to have less previous experience co-investing with other foreign VC firms. If Chinese investors have less co-investment experience, then foreign investors might be less inclined to include them (Meuleman and Wright, 2011). But there are also differences in practice. Domestic Chinese venture capitalists have been shown to emphasize financial monitoring, while foreign investors actively monitor management and encourage greater use of stock options (Tan et al., 2008). These different practices and experiences suggest that foreigners could be less likely to co-invest with Chinese venture capitalists particularly in situations of high uncertainty such as the earliest stages of the portfolio firm’s life where significant investment 5 of time and energy will be required. The Chinese Setting By 2010, the Chinese VC market was the second largest in the world trailing only the U.S. The government has implemented substantial regulatory reforms over the years, but the Chinese entrepreneurial economy still remains remarkably different from that of Western European nations (see, for example, Ahlstrom et al. 2007; Bruton et al. 2009). Given its size, the significance of foreign venture capitalists, and the rapidity of its growth, the Chinese VC industry has attracted significant scholarly attention from scholars interested in the globalization of the VC industry. In China, portfolio firms receiving investment from a foreign venture capitalist are more likely to list abroad and be affiliated with prestigious law firms, bankers and accountants (Humphery-Jenner and Suchard 2013a; 2013b). Earlier work found that social networks (guanxi) were vital for entrepreneurs seeking investment capital (Batjargal and Liu 2004). In addition, in their study of co-investment decisions among Chinese venture capitalists, Gu and Lu (2013) found that the relationship between a firm’s reputation and its likelihood of coinvestment was curvilinear. That is to say, firms with both high and low reputations are less likely to form co-investment syndicates, but this also depends upon institutional development, which they proxied as a dummy variable for the year 2004. Our study builds upon this previous work while examining the role of uncertainty in determining coinvestment choices. 6 3. Hypothesis Development The variables of interest for this study are the characteristics of the foreign lead VC firm and the portfolio firm. Because our dependent variable is whether or not the foreign firm coinvests, we do not examine the characteristics of the chosen Chinese co-investor. Foreign Venture Capital Firm Characteristics Co-investment preference may depend on the lead venture capitalist’s own characteristics. The characteristics of foreign VC firms that might influence their choice of co-investors are comprised of their experience in operating in China, their overall experience as measured by age, and whether they have an office in China. We separate these into the following characteristics and hypotheses. Foreign Venture Capital Firm’s Chinese Investment Experience Organizational learning theory suggests that prior experience influences later behavior (Sorenson and Stuart, 2001). Foreign firms with greater experience investing in China would be expected to be familiar with and adapted to the local institutional environment (Meuleman and Wright, 2011). Over time, the experience that foreign firms accrue could alter the preference of co-investors in two respects. First, with increased understanding of foreign firms of the Chinese context and actors, prior experience should decrease uncertainty. For example, Sorenson and Stuart (2008) found that, when in known settings, actors were more willingly experiment with dissimilar co-investors. The reduced foreignness of the Chinese market could induce foreign firms to more willingly co-invest with local firms. Such an effect might be especially powerful in China where the importance of networks and connections (also known as guanxi) is well documented (Batjargal and Liu, 2004). For this reason, foreign 7 firms with greater Chinese experience might be more willing to select domestic co-investors. There is another possibility. According to previous findings about investing abroad, foreign firms with little or no Chinese investment experience might be more likely to coinvest with Chinese firms to draw upon their local knowledge. For these reasons, the lead venture capitalist’s previous experience could affect the probability of selecting Chinese coinvestors in either direction, but we phrase the hypothesis as a positive relationship. Hypothesis 1: Foreign venture capital firms with previous investment experience in China are more likely to select Chinese venture capitalists as co-investors. Foreign Venture Capital Firm Age It has been shown repeatedly that there are decision-making differences between inexperienced and more experienced VC firms (Sorenson and Stuart, 2001; Butler and Goktan, 2007; Cumming and Dai, 2010). The most often used measure of experience and success of a firm is its age (Sorenson and Stuart, 2001). From this basic observation, it might be concluded that more experienced foreign firms would have greater confidence in their ability to judge a portfolio firm’s promise and might not be as adverse to co-investing with less prestigious or experienced firms – a result that could be extended to Chinese firms. For this reason, previous results could lead to the expectation that more seasoned foreign firms will be more willing to co-invest with Chinese venture capitalists. Even if the foreign firm has less co-investment experience with Chinese firms, the foreign firm may be willing to enter a new co-investment relationship due its overall investment experience. This may be due to the more experienced firms being widely known, thereby increasing the likelihood that other VC firms will offer them good investment opportunities with them (Sorenson and 8 Stuart, 2001). Conversely, domestic firms may be motivated to approach these experienced foreign firms because of their extensive international connections (Mäkelä and Maula, 2006). This would suggest that the older VC firms might co-invest with local firms because they offer access to better opportunities. Alternatively, pre-existing relationships should create trust in their previous investment partner’s abilities and resources and hence they would be more likely to co-invest with them (Sorenson and Stuart, 2008). Alternatively, experienced firms, with their deep knowledge and resources, might have a greater ability to employ risk reduction strategies (Petkova et al., 2013). Because of the great number of foreign firms investing in China, experienced firms might co-invest with other VC firms that they already know well. For this reason, experienced venture capitalists might have less need for local co-investors than less experienced foreign firms that are less knowledgeable about the local environment. Also, during the period of this study the Chinese VC market was attractive and many new foreign firms were formed to invest in China. These novice firms had neither the contacts nor the “brand” to entice other elite foreign venture capitalists to co-invest with them, so they might have been more willing to co-invest with a Chinese firm, because the foreign firm could draw upon the domestic firm’s familiarity with the local environment (Wright et al., 2005). For these reasons, the lead venture capitalist overall experience, as measured by firm age, could affect their probability of selecting a Chinese partner either way. Regarding less experienced firms we hypothesize the following. Hypothesis 2: Younger foreign venture capital firms are more likely to co-invest with Chinese venture capital firms. Chinese Offices of Foreign Firms 9 Foreign VC firms can increase proximity to and knowledge of the local market by establishing a local office (Meuleman and Wright, 2011). This was the case of Israel, as many foreign venture capitalists have established local offices to manage their investments (Avnimelech and Teubal, 2006). During the last decade, many foreign VC firms established local offices in China (Zhang, 2011). To illustrate, in our database 71.3% of foreign firms with four or more investment deals in China had a Chinese office by the end of 2012. Because a local office should increase access to local knowledge (Meuleman and Wright, 2011) and improved monitoring of local portfolio firms, it could influence a VC firm’s decision to co-invest with local VC firms. The office could depress the need to co-invest with local firms in the earliest and most uncertain periods, where homophily would be expected to be of greatest importance, while assisting in the recruitment of local investors during the later and less uncertain periods, when the now larger portfolio firm might benefit from local networks. Therefore, we expect that: Hypothesis 3: Having a Chinese office decreases the probability that the foreign venture capital firm selects a Chinese partner. Portfolio Firm Characteristics The characteristics of the portfolio firm could affect co-investor choice (Hochberg et al., 2011; Liu and Maula, 2014). These characteristics might include a portfolio firm’s maturity and its geographic location. Portfolio Firm Age One of the primary data for investment uncertainty is a portfolio firm’s operating history. 10 Since the financial statistics for portfolio firms are unavailable, age is used as a proxy (Sorenson and Stuart, 2001). For investors, portfolio firms with longer operating histories normally have reduced uncertainty (Sorenson and Stuart, 2001). This may encourage the lead venture capitalist to recruit dissimilar co-investors with complementary resources. Therefore, we expect that: Hypothesis 4: The Portfolio Firm’s Age positively affects the likelihood that a foreign venture capital firm chooses Chinese co-investor. Portfolio Firm Location Venture capital firms often have geographic preferences (Sorenson and Stuart, 2001; Christensen, 2007; Hochberg et al., 2011). It is well-known that foreign investments are highly concentrated in Beijing and Shanghai, while domestic firms are more widely dispersed nationally (Zhang, 2011). In these other regions, not only is distance a factor, but also foreign VC firms are not likely to be embedded locally. Further, regional entrepreneurs may be more resistant to the sale of ownership to “outsiders” (Tan et al., 2008). As a result, foreign VC firms face more risk and uncertainty outside the VC centers. For these reasons, for portfolio firms located outside of VC centers, foreign investors will be more likely to co-invest with Chinese VC firms Hypothesis 5: Foreign venture capitalists are likely to choose a Chinese co-investor, if the portfolio firm is located outside a venture capital center. Investment Characteristics In addition to the characteristics of the foreign VC firms and the portfolio firms, characteristics of the specific investment include investment stage and investment year. 11 Investment Stage Investment stage is the single most important predictor of venture uncertainty. Thus we expect it to have a significant impact upon co-investment. This is particularly the case in China where researchers have found that foreign venture capitalists invest in earlier stages than do Chinese firms (Tan et al. 2008). For this reason, after running our models with dummy variables for the four stages identified by VenturExpert, we ran the same variables in separate regressions for each stage to explore differences. This was necessary in our data, because the preponderance of the portfolio firms in our database did not go through the classical sequence of seed, early, expansion, and later stage. Many received their first round of investment even at the latest stages (See Table 1). For this reason, we expect that: Hypothesis 6: The later the investment stage, the more likely there is to be a Chinese coinvestor. [Table 1 around here] Investment Year The Chinese VC market has matured rapidly, both in terms of changed government policies for encouraging VC and the levels of experience of Chinese VC firms. In the late 1990s when foreign venture capitalists began investing in Chinese firms, there were few domestic Chinese firms and they were inexperienced in nurturing young firms. Moreover, it has been reported that they were risk adverse. Since then there have been many changes in Chinese financial and legal policies (Xu, 2002). For this reason we expect that investment year will have an impact on co-investing. Therefore, we propose Hypothesis 7: Investment Year will be positively correlated with the presence of a Chinese 12 co-investor. 4. Data and Methodology Data The data for this study was drawn from the Thomson VentureXpert database, which attempts to record all VC firm investments globally and has been used by many researchers (e.g., Sorenson and Stuart, 2001; Liu and Maula 2014). From the VentureXpert database we created a comprehensive dataset of foreign VC investments in China from January 1, 1992 to December 31, 2012. There were several steps in producing our dataset. “foreign venture capital firms” based on the VentureXpert database. First, we defined The VentureXpert database includes a large numbers of VC firms “headquartered” in China that are not domestic Chinese firms, but rather are subsidiaries of foreign VC firms. There are also VC firms that were established abroad by Chinese citizens or ethnic Chinese that were foreigners (e.g., Ceyuan Ventures Management), or were established in China by Westerners (e.g., TDR Capital). To verify our data set we checked each firm against reports by Zero2IPO, which is a Chinese source of information on VC investing. For all those that still could not be categorized, we consulted the firm website or other public websites. In our database all VC firms headquartered outside China, and VC firms headquartered in China that were established by non-Chinese, have been classified as foreign VC firms. Subsidiaries of the same foreign VC firm, but having different names, are combined together as one firm (e.g., SAIF Partners, SOFTBANK China Venture Capital and Softbank Corp). This correction is important because these different subsidiaries may co-invest, but these are 13 not arm’s-length transactions and cannot be considered a normal co-investment. The following investments were excluded from our database as they are not classic VC investments: 1) portfolio firms that were stated-owned enterprises, township/village enterprises, banks, spinoffs, subsidiaries and joint ventures; 2) those mature portfolio firms receiving their first investment when they were more than 15 years old; 3) portfolio firms receiving their first VC investment only in their latest stage and within one year of an IPO; and 4) portfolio firms that were in manufacturing about which there was no further information available in either VenturExpert or in an online search. In addition, the VenturExpert database includes misclassified or multiple listings of the same firm and missing geographic and industrial information. VenturExpert also has a large numbers of “undisclosed” investors or “unknown” locations that were excluded. After this data preparation, there were 1,095 portfolio firms remaining that received 3,365 foreign investments and 696 Chinese investments (an investment is defined as one portfolio firm receiving an investment from one venture capitalist in one investment round). In its China data, VentureXpert categorizes all investments as belonging to one of four mutually exclusive stages. The Seed Stage refers to “portfolio companies that have not yet fully established commercial operations, and may also involve continued research and product development.” The Early Stage refers to portfolio companies after the Seed Stage/Startup and the funds are used for product development, initial marketing, manufacturing and sales activities. The Expansion Stage is investment into portfolio companies that have products and services that are currently available, and require additional capital to expand production to increase revenue. Later Stage investments are those in an 14 already established portfolio firm that has products or services already generating revenue, but may not be making a profit. This is normally the last round of investments before an exit in the form of an IPO or acquisition by a strategic partner. Normally, in later stages firm valuations are higher allowing previous investors to capture some of the value that their investments have created. This is possible because, as a portfolio firm progresses through these stages, the level of uncertainty regarding its management team, market, and technology decreases. If more than one VC firm invests in a particular portfolio firm in a particular round, this is defined as a syndicated co-investment round. The same co-investors usually invest in a target portfolio firm over several rounds. We followed Sorenson and Stuart (2008), but modified their strategy1 and included all co- investment rounds where one or more new firms joined the syndicate. These are defined as the set of “foreign syndicate co-investment rounds”. VenturExpert database provides the financing round number of all the investments. Table One shows the count of co-investments by Investment Round and portfolio firms’ Stage. Sorenson and Stuart (2008) used the count of the financing round to measure the risk and uncertainty of the investment environment. However, from Table One we can see that there are too many first rounds in late stages. If we define seed stage as 1, early stage as 2, expansion as 3, later stage as 4, the Pearson Correlation Coefficient between these stages and investment rounds is only 0.287, which is quite low suggesting that Investment Stage, which is defined clearly, is a better measurement of uncertainty reduction than Investment Round. 1 Sorenson and Stuart (2008) defined syndicate round as those financing rounds in which more than two new-to-thecompany venture capital firms invested in the target company. 15 Therefore, we did not use Investment Round as a variable. In choosing a syndicate partner, the lead VC firm is assumed to decide whether to accept a co-investor. Our goal is to estimate the probability that the lead foreign firm chooses either a Chinese or a foreign co-investor. We identified the lead investor using the following criteria:1) if a foreign firm only has Chinese partners or undisclosed partners in a particular syndicate round, the foreign firm is the lead; 2) if the equity amount of the foreign firm in the syndicate round is the largest, it is the lead VC firm; 3) if more than one foreign firm invested in a syndicate round of a particular portfolio firm with equal equity, we define the firm entering the portfolio firm in an earlier round as the lead; 4) if more than one foreign firm invested in a syndicate round of a particular portfolio firm with equal equity, and the first round they entered the portfolio firms is also the same, we define the VC firm investing the most times in the portfolio firm as the lead.2 Second, we identified those co-investment rounds in which there was more than one lead venture capitalist. To be more specific, if two or more foreign firms made an equal investment in the syndicated round we defined them as co-leads. For these co-lead rounds, we not only count the relationships between each co-lead and its other non-lead investors, but also include the interactional relationships among co-leads. For example, two co-lead foreign firms i and j syndicate invest with a non-lead partner k in a syndicate round for a portfolio firm, we count four relationships: i→j, j→i, i→k, j→k. The dataset includes 555 unique portfolio firms, 745 syndicate rounds and 2,136 relationships by lead foreign firms. Among these samples, lead foreign firms respectively had 542 relationships with Chinese partners 2 In 15 cases foreign VC firms joined a syndicate that Chinese VC firms had already established in an earlier round. As these foreign VC firms were unlikely to have been lead investors they were eliminated from consideration. 16 and 1,594 relationships with foreign co-investors partners. Variables Dependent Variable The dependent variable in our analysis is binary and indicates that a lead foreign VC firm undertakes a co-investment with a domestic Chinese VC firm, or a foreign VC firm, in a portfolio firm in a given round. A value of 1 is obtained if a lead foreign firm has a Chinese co-investor, and it is equal to 0 if a lead foreign firm has a foreign co-investor. [Figure 1 around here] Independent Variables VC Experience in China- This variable is used to examine whether previous Chinese experience influences co-investor choice. This variable is the number of previous investment rounds in China a given lead foreign firm has been involved in prior to the date of the given syndicate round. The logarithm of this measure is used. VC Age- To measure changes in co-investor choice by more mature venture capitalists, we employ the lead foreign firm’s age (Sorenson and Stuart, 2001; Wang and Wang, 2011). The age of the lead foreign venture capitalist is calculated by subtracting its founding date (in its own country) from the year of the co-investment. The logarithm is used. Chinese Office- We use a dummy variable to test whether a Chinese office influences the decision to co-invest with a Chinese firm. If the lead foreign firm has a Chinese office at the date of the investment round, the value of this dummy variable is 1 and otherwise it is 0. Portfolio Firm Age - This may influence the decision about choosing a Chinese co-investor. 17 It is measured by the number of months from the portfolio firm’s establishment date to the date of the investment. The logarithm is used. Portfolio Firm Geography- Foreign VC investment in China is concentrated in Beijing and Shanghai. For portfolio firms located in other cities having a local venture capitalist may improve portfolio firm monitoring. For this reason, we created a dummy variable Venture Capital Center for Beijing and Shanghai. The value of this dummy variable is 1 if the portfolio firm is located in a center, otherwise it is 0. Investment Stage- The stage variable provides information on the maturity of the firm (Sorenson and Stuart, 2001). Therefore, we include four dummy variables: Seed Stage, Early Stage, Expansion Stage, and Later Stage. The value of these dummy variables is equal to 1 if the invested portfolio firm is in a particular stage and otherwise 0. Investment Year- There have been major changes in the environment and maturity of Chinese firms. This may have impacted the propensity to choose a Chinese VC partner. This variable is entered by the calendar year of the investment round. Control Variables Prior Investor- A dummy variable is used to indicate whether the portfolio firm has received investment from a co-investment syndicate that included a Chinese firm. The dummy is equal to 1, if the portfolio firm has received such an investment and 0 otherwise. Syndicate Size- The value of this variable is measured by the number of investors in each foreign co-investment round. The logarithm of this measure is used. Methodology 18 Our first regressions are on the entire population and we include stages as an independent variable. The second reported regressions are when we separate our population by stages to examine whether the determinants of co-investing with Chinese firms differs in the stages. Here, we are posing the question of whether within the stages there are differences in the effect of the independent variables on the dependent variables. 5. Results The results of bivariate correlations for our population are reported in Table Two. The correlations are not high and there is little evidence of multicollinearity. Due to the large number of hypotheses and our decision to run regressions on the entire population and then on each stage separately, we report all the results for each hypothesis separately and then summarize the overall findings. [Table 2 around here] The results of our models that examine investment stages with dummy variables are reported in Table 3. The results in our most basic regression Model 1 shows the probability of choosing a Chinese co-investor using the Venture Capital Experience, Venture Capital Age, Portfolio Firm Age, and Investment Year variables and the control variables of Prior Investor and Syndicate Size. There was strong support for Hypothesis 2 that younger foreign firms would be more likely to invest with Chinese firms. There are a number of possible explanations for this. For example, our population contains a number of firms that were formed, quite recently, with the express purpose of investing in China. There is also support for Hypothesis 4 and Hypothesis 7. The coefficient for Venture Capital Experience was not 19 significant meaning that Hypothesis 1 was not supported. [Table 3 around here] The positive effects of portfolio firms’ age indicate that for older portfolio firms where there is less uncertainty so foreign venture capitalists are more likely to take on Chinese coinvestors. Of course, we are not certain regarding the causality here, but our interviews with investors in China suggests that Chinese firms are risk adverse and thus prefer later rounds. Obviously, if a Chinese venture capitalist has been a co-investor in an earlier round we would expect it or other Chinese venture capitalists to be included in later rounds. During the period of our analysis, the Chinese industry was changing due to pro-venture capital government policy changes and increasing experience among local venture capitalists. For these reasons, Investment Year was highly significant suggesting that over time foreign firms were more willing to co-invest with local firms. To conclude, Model 1 demonstrates that younger foreign firms more likely to co-invest with a Chinese partner and foreign firms were more likely to co-invest with Chinese firms if the target portfolio firm was older. In the following models, Model 1 is used as the baseline. Model 2 tested whether having a Chinese Office affected the propensity to co-invest with a Chinese partner, but it was not significant. This indicates that establishing a local office did not change the preference for a Chinese partner, and thus Hypothesis 3 is not supported in this model. Model 3 introduced the dummy variable Venture Capital Center measured whether the location of a portfolio firm in Shanghai or Beijing affected partner preference. The significant negative coefficient suggests that foreign firms are more likely to co-invest with Chinese partner outside these centers supporting Hypothesis 5. 20 If as the literature suggests investment uncertainty decreases in later stages, then we would expect that the later the stage, the higher the probability of including Chinese coinvestors. Because of the relatively high and expected correlation between Portfolio Firm Age and Investment Stage (see Table Two) we excluded Portfolio Firm Age in this model. Therefore, Model Four is a further test to examine Hypothesis 6 using the dummy variables of Investment Stage excluding the age of portfolio firms. Later stage is the reference category. It reveals that the probability of choosing a Chinese partner is lower for investments in the early or expansion stage, but with no statistically significant impact on the seed stage, where we expected the effect to be strongest. This may be due to the relatively small N in the seed stage. In the full Model Five, there were no changes except that the stage significance levels decreased. These results give us some confidence that decreasing uncertainty leads to the inclusion of Chinese co-investors. Our results show the increasing willingness of foreigners to co-invest with Chinese VC firms during the period. And yet, more experienced foreign firms continued to be less likely to co-invest with Chinese firms. Also, if the portfolio firm was located in Beijing or Shanghai co-investment with Chinese was less likely. As expected having a previous Chinese coinvestor increased the likelihood of future co-investment. Finally, the older the portfolio firm was, the greater the likelihood of Chinese co-investment. These results suggest that decreasing uncertainty encourages foreign firms to include Chinese co-investors. However, to better understand how uncertainty affects co-investment, it is possible to run regressions on the population in each stage and these results are reported in the next section. 21 Regressions by Stage If our hypothesis that co-investment decisions should differ by the level of uncertainty is correct, then the variable significance should change by investment stage. In Table Four the same variables as in the full model are used for separate regressions. Each regression can be thought of as proceeding from one uncertainty regime to the next, with each stage being less uncertain. The most noteworthy change is in the seed stage, which provides clear evidence for the proposition that uncertainty affects partner decision making. The most remarkable change is that in the Seed Stage, Investment Year is no longer significant. Moreover, as the stages progress it becomes gradually more significant. This result should be interpreted carefully as the number of co-investment rounds in the seed stage were quite low (68), and because in this stage most investments (90) were solo. It does, though, confirm previous research such as Stuart and Sorenson (2008) that homophily is most powerful in the most uncertain environments. [Table 4 around here] The Seed Stage has further surprises in that having a Chinese office is negatively related to having a Chinese partner and then is positively related at the Later Stage this suggests that the Chinese office may have two functions. The first function is to prospect for early-stage deals that will be syndicated with other foreign venture capitalists, but in the later stages, when the firm is growing to recruit local investors that are likely to have strong connections with all-important government officials. In the general models, Venture Capital Experience in China was not significant. However, in the seed stage it was strongly positively related to a willingness to select a Chinese co-investor suggesting that increased experience in China, 22 which would decrease general uncertainty, may offset the homophily argument. Venture Capital Age, which was negative and significant throughout the general regression remained negative, but was no longer significant outside the Seed Stage. Apparently, the older foreign venture capitalists remained reluctant to accept Chinese co-investors. In the Later Stage, syndicate size was significant in the decision to include Chinese. This may be due to fact that there is relatively low uncertainty, and the portfolio firm is often raising growth capital or may need connections with policy-makers or other actors prior to a liquidity event. Sensitivity Analysis Given the findings in previous research on cross-border co-investment decisions (Du and Vertinsky, 2008; Chemmanur et al., 2013; Devigne et al., 2013; Hoskisson et al., 2000; Meuleman and Wright, 2011), we tested for whether nationality had an impact on the investment behavior. It was not significant in any of the models and therefore it was dropped. We also tested whether the organizational types (including private limited partnerships, corporate venture capitalists, financial venture capital firms and others) of foreign VC firms affected the willingness to co-invest with a Chinese firm and found no significance so we omitted this variable. Previous research has found that industry characteristics may also influence co-investor choice (Du and Vertinsky, 2008; Devigne et al., 2013; Hochberg et al., 2011; Meuleman and Wright, 2011; Sorenson and Stuart, 2001, 2008). For this reason we ran the models with dummy variables for both the information and communication technology and the Internet industries, both separately and combined, but found no significance. When these dummies 23 were dropped from the models there were no substantive changes. 6. Discussion Existing theories of co-investor choice have highlighted the power of homophily (Sorenson and Stuart, 2008), while studies of VC investing in foreign nations have highlighted the importance of local linkages. Learning theories have suggested that prior experience should reduce institutional barriers to co-investment with domestic venture capitalists (Meuleman and Wright, 2011). In general models there was no evidence for this effect, and there was also no evidence that having a Chinese office affected choice of a Chinese partner. However, in the regressions by stage, the results changed. In the seed stage, experience in China now had a positive effect, while having a Chinese office had a powerful negative effect. This can be interpreted as suggesting that, at most uncertain stage, the Seed Stage, a Chinese Office substituted for the local monitoring capability of a Chinese VC firm. The role of the Chinese office was reversed in the Later Stage as it now had a significant positive impact upon co-investment with a Chinese firm. This result suggests that when uncertainty is greatest, when the foreign venture capitalists use their office to undertake monitoring, they opt for homophilous co-investors. If local monitoring is not possible, then a local co-investor is preferred, this qualifies Makela and Maula (2008) finding about choosing a local partner. In the general models, foreign VC firms were more likely to co-invest with a Chinese VC firm when the portfolio firm was located outside of Shanghai or Beijing. Because the foreigners have concentrated their offices and investing in Beijing and Shanghai, as Sorenson 24 and Stuart (2001) found in the case of the U.S., co-investment outside of their home regions was more likely to be with local venture capitalists. While expected, the explanation may not be simply monitoring and may include a greater ability for locals to better interact with the firm and local business networks and government. The Investment Year variable, which is normally a control, behaved as expected given the changes in the Chinese economy during this time period in that it was powerfully positive. However, when we ran the model by stage, in the seed stage, it had no significance. In other words, there was no discernable time effect on the likelihood of co-investment in the seed stage. For every other stage, the passage of time increased the likelihood of having Chinese co-investors. This result suggests that despite the striking changes in the Chinese environment, in the most uncertain early stage, foreigners co-invest with their own kind – dramatic evidence for the uncertainty-homophily linkage. 7. Conclusion Because our data is only for China, there are limitations to the generalizability of our results, but they reinforce the conclusion of Liu and Maula (2014) that uncertainty affects the decision to co-invest with local firms. Also, our data does not allow us to conclusively determine whether the reason for the lack of Chinese co-investors is due to the choice of foreign firms to not co-invest with Chinese venture capitalists, or due to Chinese firms not wanting to invest in seed-stage firms because of the greater risks. Anecdotal evidence suggests that it is likely a combination of both. Another general limitation is that the VentureXpert database has serious data-quality 25 problems requiring substantial time investment to correct. Uncritical use of the database is likely to lead to estimation problems. The most significant of these is that investments attributed to the Chinese subsidiary of a global VC firm are considered domestic. In international business studies, few scholars would consider a subsidiary as equivalent to a domestic firm. Should the same hold for VC subsidiaries in foreign nations? Unfortunately, much current research on overseas VC investing is not explicit concerning how this is addressed. Hypotheses may have been tested on data having serious collection-related bias – a problem that is likely to be particularly true in developing nations with young VC industries and weak reporting standards. The final important limitation of the data is the large number of “undisclosed” investors. It is possible that these are mainland Chinese, overseas Chinese, or Taiwanese who, for tax or possibly political purposes, keep their identity secret. This might mean that foreign-domestic co-investment was greater or smaller than what we capture. We have shown that Chinese investors are most likely to co-invest with foreign firms during the later stages, after the portfolio firm is already well-established. While this decreases investment risk, there is also less potential for learning from experienced foreigners the most difficult skills of the venture capitalist’s craft such as assessing a new opportunity and advising, nurturing and monitoring an entrepreneurial team during the most perilous period. This problem of learning is increased even more, as the foreign firms least likely to work with Chinese co-investors are the ones that are the most experienced, i.e., members of the global elite. Choosing to work with like-minded partners in situations of high uncertainty is likely an outcome of natural social tendencies, it may limit the transfer of skills for identifying and nurturing fledgling firms. 26 REFERENCES Ahlstrom, D., Bruton, G., Yeh, K. (2007) Venture capital in China: Past, present, and future. Asia Pacific Journal of Management, 24, 247–268. Avnimelech, G., & Teubal, M. (2006). Creating venture capital industries that co-evolve with high tech: Insights from an extended industry life cycle perspective of the Israeli experience. Research Policy, 35(10), 1477-1498. Batjargal, B. and Liu, M. (2004) Entrepreneurs’ access to private equity in China: The role of social capital. Organization Science, 15 (2): 159-172. Brander, J., Amit, R., Antweiler, W. (2002) Venture capital syndication: improved venture selection versus the value-added hypothesis. Journal of Economics and Management Strategy, 11: 423-452. Bruton, G. D., Ahlstrom, D., & Puky, T. (2009). Institutional differences and the development of entrepreneurial ventures: A comparison of the venture capital industries in Latin America and Asia. Journal of International Business Studies,40(5), 762-778. Butler, A., Goktan, M. (2007) Does Distance Still Matter? Evidence from the Venture Capital Market. Working Paper, University of Texas. Chemmanur, J., Hull, T., Krishnan, K. (2013) Expertise or Proximity? International Venture Capital Investments and the Development of Venture Capital Markets. AFA 2012 Chicago Meetings Paper. Available at SSRN: http://ssrn.com/abstract=1670319. Christensen, J.L. (2007) The Development of Geographical Specialization of Venture Capital. European Planning Studies, 15: 817-833. Cumming, D., Dai, N. (2010) Local Bias in Venture Capital Investments. Journal of Empirical Finance, 17: 362-380. Dai, N., Jo, H., Kassicieh, S. (2012) Cross-border venture capital investments in Asia: Selection and exit performance. Journal of Business Venturing, 27: 666-684. Devigne, D., Vanacker, T., Manigart, S., Paeleman, I. (2013) The role of domestic and crossborder venture capital investors in the growth of portfolio companies. Small Business Economics, 40: 553-573. Du, Q., Vertinsky, I. (2008) Risk mitigation strategies of foreign venture capitalists in China. University of British Columbia Working Paper. 27 Florida, R., Kenney, M. (1988) Venture Capital, High Technology and Regional Development. Regional Studies, 22: 33-48. Jääskeläinen, M, Maula, M.(2013) Do networks of financial intermediaries help reduce local bias? Evidence from cross-border venture capital exits. Journal of Business Venturing Manuscript Draft. Gompers, P. A. (1995). Optimal investment, monitoring, and the staging of venture capital. The Journal of Finance, 50(5), 1461-1489. Gu, Q., Lu, X. (2013) Unraveling the mechanisms of reputation and alliance formation: A study of venture capital syndication in China. Strategic Management Journal, published online: 21 MAR. Guler, I., Guillen, M.F. (2010) Knowledge, Institutions and Organizational Growth: The Internationalization of U.S. Venture Capital Firms, Journal of International Business Studies, 41: 185-205. Hoskisson, R., Eden, L., Lau, C, Wright, M. (2000) Strategy in emerging economies. Academy of Management Journal, 43:249–267. Hochberg, Y., Ljungqvist, A., Lu, Y. (2007) Whom you know matters: Venture capital networks and investment performance. The Journal of Finance, 1: 251-301. Hochberg, Y., Lindsey, L., Westerfield, M. (2011) Partner selection in co-investment networks: evidence from venture capital. Northwestern University working paper. Humphery-Jenner, M., & Suchard, J. A. (2013a). Foreign venture capitalists and venture success: Evidence from China. Journal of Corporate Finance, 21, 16-35. Humphery-Jenner, M., & Suchard, J. A. (2013b). Foreign venture capitalists and the internationalization of entrepreneurial companies: Evidence from China. Journal of International Business Studies, 44: 607-621. Kaplan, S.N. and Stromberg, P. (2003) Financial Contracting Theory Meets the Real World: An Empirical Analysis of Venture Capital Contracts. Review of Economic Studies, 70: 281315. Lerner, J. (1995) Venture capitalists and the oversight of private firms. Journal of Finance, 50: 301-318. Lerner, J. (1994). The syndication of venture capital investments. Financial management, 1627. 28 Liu, Y. Maula, M. (2014) To Partner Or Not with Local Firms in Emerging Markets? The Effects of Different Types of Uncertainty and Experiential Learning on Partnering. Academy of Management Journal Forthcoming. Mäkelä, M., Maula, M. (2005) Cross-border venture capital and new venture internationalization: An isomorphism perspective. Venture Capital: An International Journal of Entrepreneurial Finance, 7: 227-257. Mäkelä, M., Maula, M. (2006) Interorganizational Commitment in Syndicated Cross-Border Venture Capital Investments. Entrepreneurship Theory and Practice, 30: 273-298. Mäkelä, M., Maula, M. (2008) Attracting cross-border venture capital: the role of a local investor. Entrepreneurship and Regional Development, 20:237-257. Meuleman, M., Wright, M. (2011) Cross-border private equity syndication: institutional context and learning. Journal of Business Venturing, 26: 35-48. Petkova, A., Wadhwa, A., Yao, X., Jain, S. (2013) Reputation and Decision Making under Ambiguity: A study of U.S. Venture Capital Firms’ Investments in the Emerging Clean Energy Sector. Academy of Management Journal. Published online before print May 28, 2013. Sorenson, O., Stuart, T. (2001) Syndication Networks and the Spatial Distribution of Venture Capital Investments. American Journal of Sociology, 106: 1546-1588. Sorenson, O., Stuart, T. (2008) Bringing the Context Back In: Settings and the Search for Syndicate Partners in Venture Capital Investment Networks. Administrative Science Quarterly, 53: 266-294. Tan, J., Zhang, W., Xia, J. (2008) Managing Risk in a Transitional Environment: An Exploratory Study of Control and Incentive Mechanisms of Venture Capital Firms in China. Journal of Small Business Management, 46: 263–285. Wang, L., Wang, S. (2011) Cross-border venture capital performance: Evidence from China. Pacific-Basin Finance Journal, 19: 71-97. Wright, M., Lockett, A. (2003) The Structure and Management of Alliances: Syndication in the Venture Capital Industry. Journal of Management Studies, 40: 2073-2102. Wright, M., Pruthi, S., Lockett, A. (2005) International venture capital research: From crosscountry comparisons to crossing borders. International Journal of Management reviews, 7:135-165. Xu, E. (2002) Venture capital finance in China. Journal of Entrepreneurial Finance, 7:11-24. 29 Zhang, J. (2011) The spatial dynamics of globalizing venture capital in China. Environment and Planning A, 43: 1562-1580. 30 Appendix Syndicate relationship Foreign lead VC firm Invest Co-investors Invest Dependent variable: Chinese or Foreign co-investor Target portfolio firm Figure 1 Visual Explanation of Population Table 1: Co-Investment Ties by Stage and Round Investment Stage/ Round Seed stage Foreign-Foreign Foreign-Chinese Early stage Foreign-Foreign Foreign-Chinese Expansion stage Foreign-Foreign Foreign-Chinese Later stage Foreign-Foreign Foreign-Chinese Total Foreign-Foreign Foreign-Chinese 1 2 3 126 91 35 289 210 79 422 338 84 146 85 61 983 724 259 12 9 3 193 147 46 286 215 71 83 49 34 574 420 154 3 3 47 33 14 185 145 40 78 54 24 313 232 81 4 19 14 5 95 85 10 46 36 10 160 135 25 5 6 7 Total 16 15 1 12 10 2 16 15 1 12 10 2 141 100 41 550 404 146 1044 831 213 401 259 142 2136 1594 542 2 2 28 23 5 48 35 13 78 58 20 31 2 3 1.VC Experience 0.240 ** 0.617* * 2. VC Age 1 0.086* * 3. Chinese Office 4. PF Age 5. Prior Investor 6. VC Center 7. Syndicate Size 8.Seed Stage 9.Early Stage 1 Table 2: Correlation Matrix 4 5 6 7 8 9 10 0.098 0.057* -0.035 -0.028 0.111* 0.058* 0.039 ** * * * 0.086 0.098* 0.043* -0.018 0.062* -0.029 0.017 ** * * 0.092 0.076* 0.141* 0.093* 0.176* -0.008 -0.035 ** * * * * 0.157* 0.131* 0.210* 1 0.105* 0.492* 0.326* * * * * * * 0.150* 0.228* 1 0.109* 0.01 * * 0.043* * 1 -0.022 1 11 0.473* * 0.128* * 0.367* * 0.249* * 0.067* * 0.111* 0.046* -0.002 0.064* * * * 0.178* -0.019 0.075* 0.115* * * 1 0.157* 0.260* 0.073* * * * 1 0.576* -0.002 * 10. Expansion 1 Stage 11. Investment Year Notes: **, * respectively mean that correlation is significant at the 0.01 level and 0.05 level. -0.018 1 32 Table 3: Logistic Regression on whether the Co-Investor is a Chinese Venture Capital Firm N = 2,136 Model 1 Model 2 Model 3 Model 4 Model 5 Constant 6.040*** 5.982*** 5.525*** 5.053*** 5.297*** 0.666 0.697 0.699 0.697 0.785 -0.026 -0.028 -0.024 0.000 0.004 VC Experience 0.040 0.050 0.040 0.040 0.050 † -0.128* -0.115 -0.127* -0.110* -0.120* VC Age 0.056 0.060 0.056 0.055 0.060 0.080 0.036 Chinese Office 0.162 0.166 † 0.121* 0.140* 0.104 0.182* Portfolio Firm Age 0.059 0.060 0.059 0.083 0.277* 0.280* 0.339* 0.308* 0.323* Prior Investor 0.138 0.139 0.141 0.137 0.144 -0.271* -0.235* Portfolio Firm Geography: VC Center 0.124 0.129 -0.036 -0.040 -0.049 -0.099 -0.079 Syndicate Size 0.104 0.105 0.104 0.106 0.110 Portfolio Firm Investment Stage: -0.153 0.267 Seed Stage 0.234 0.325 -0.464** -0.175 Early Stage 0.154 0.200 0.687*** 0.530*** Expansion Stage 0.138 0.151 2.035*** 1.943*** 2.015*** 1.972*** 1.806*** Investment Year 0.254 0.257 0.254 0.247 0.259 2218.482 2165.065 22.13.789 2257.790 2137.050 -2 Log likelihood 0.086 0.085 0.089 0.100 0.105 Nagelkerke R square Notes: Significance levels of the coefficients: † = .1 * = .05 ** = .01 *** = .001 33 Table 4: Logistic Regressions on whether the Syndicate Partner is a Chinese VC Firm by Portfolio Firm Investment Stage Seed stage: # Expansion stage: Later stage: # Early stage: # of of Co-invest # of Co-invest of CoCo-invest= 550 =141 =1,044 invest=401 # of Unique # of Unique # of Unique # of Unique Portfolio Firms Portfolio Portfolio Firms Portfolio Firms = 204 Firms = 65 = 289 = 89 -2.856 -3.469** -6.183*** -14.235*** Constant 2.958 1.207 1.194 2.56 0.643* -0.076 -0.023 0.038 VC Experience 0.267 0.095 0.078 0.117 -0.510* -0.027 -0.089 -0.257† VC Age 0.227 0.116 0.098 0.148 -2.677*** 0.384 -0.281 0.975* Chinese Office 0.808 0.312 0.244 0.448 -0.075 0.283† 0.308* 0.138 Portfolio Firm Age 0.198 0.16 0.153 0.276 -20.917 0.484 0.652** -0.254 Prior Investor 28420.722 0.302 0.207 0.369 -0.397 -0.227 0.401 PF Geography: VC Center 0.721 0.593 0.281 0.203 0.300 0.487 -0.305 -0.204 0.700** Syndicate Size 0.57 0.288 0.171 0.241 1.536 0.865* 1.849*** 4.647*** Investment Year 1.066 0.400 0.392 0.889 127.207 575.693 940.156 410.961 -2 Log likelihood 0.202 0.079 0.087 0.313 Nagelkerke R square Notes: Significance levels of the coefficients: † = .1 * = .05 ** = .01 *** = .001 34 35