Shareholders Rights and the Effect of the Origin of Venture Capital Firms on the Underpricing of US IPOs

601 Corporate Governance: An International Review, 2011, 19(6): 601 621 Shareholders Rights and the Effect of the Origin of Venture Capital Firms on the Underpricing of US IPOs Salim Chahine* and Samer Saade ABSTRACT Manuscript Type: Empirical Research Question/Issue: This paper examines whether firm performance at initial public offerings (IPO) is differentially affected by the origin of Venture Capital (VC) firms. Empirical investigations consider both the anti-director index and the anti-self dealing index of the country of origin of a specific VC firm as indicators for its level of legal protection of shareholders rights. Research Findings/Insights: We find that underpricing of US IPOs is negatively related to the rating of the legal protection rights of VC firms within the VC syndicate of an IPO firm, and the effect is more significant in the subsample of IPOs involving foreign VC firms. There is also evidence of a complementary role played by the legal protection rights of foreign VCs and board independence of IPO firms in reducing underpricing. Further robustness tests confirm empirical findings controlling for the selection bias of IPO firms by foreign VCs, and using the effect of the protection rights of the country of origin of the lead, largest, board member, or oldest, i.e., most experienced, VC firm. There is also evidence of a positive but marginally significant effect of the legal protection rights of VC firms on the long-term performance of their portfolio companies. Theoretical/Academic Implications: Our paper suggests that the institutional framework and national legal differences should matter in considering the effect of VC firms on IPO performance. We find strong support for the institutional perspective whereby institutional predictions were largely supported. Our results also expand on prior research on VC syndication and show that in addition to VC monitoring, the shareholders protection rights of the country of origin of foreign VC syndicate members would signal the quality of portfolio companies at IPO. Moreover, the complementary role between the legal protection rights of the country of origin of foreign VCs and board independence indicates that firm performance is the outcome of complex mechanisms involving both firm and country-level settings. Practitioner/Policy Implications: Evidence on the association between IPO performance and the origin of VC firms suggests that policy-makers and practitioners should view legal protection of shareholders rights as a global issue. Since VC investments are valuable for the development of small and medium-size enterprises, our results should contribute to our understanding of cross-border partnering and the quality of partners within the VC industry. Keywords: Corporate Governance, Venture Capital, Initial Public Offerings, Legal Protection Rights, Underpricing, Institutional Theory INTRODUCTION Venture Capital firms (VC, henceforth) in the United States are widely seen as professional investors who actively monitor and support their portfolio companies (Gompers & Lerner, 2001; Jain & Kini, 1999; Lerner, 1995). They also provide an important certification role in the IPO process (Lin, 1996; Megginson & Weiss, 1991; Puri, 1996, *Address for correspondence: Salim Chahine, The Olayan School of Business, American University of Beirut, Bliss Street, P.O.Box: 11-0236, Beirut Lebanon. E-mail: salim.chahine@aub.edu.lb 1999). VCs use various forms of control consents to better shape the strategy/business plan of their ventures, oust the management teams of these firms prior to or following their IPOs (Kaplan & Stromberg, 2004), and replace them with key employees (Jain & Kini, 1999), more professional managers (Gorman & Sahlman, 1989; Hellmann, 1998), and more independent directors (Suchard, 2009) that reduce the principalagency conflicts of interest related to insiders opportunism and increase the value of the portfolio companies (Hellmann, 1998). While these characteristics are usually associated with VCs in the United States, not all of them originally come from the United States. There are many VCs active in the United doi:10.1111/j.1467-8683.2011.00857.x

602 CORPORATE GOVERNANCE States that come from other nations, which provides a strong test for institutional theory. From an institutional perspective, VC s native legal institutional market would help shape the VC s actions in subtle but pervasive ways. In addition to traditional agency framework, VC s monitoring practices and certification role may depend on the level of shareholders legal protection rights of their country of origin. This paper examines the effect of legal protection rights of the country of origin of VCs on the underpricing of IPOs in the US. 1 It further investigates whether country-level legal protection rights and firm-level corporate governance practices play a complementary role which should reduce underpricing of VC-backed portfolio companies. Prior research shows evidence that differences in national institutions and legal systems impact the effectiveness of corporate governance on the firm level (Aguilera and Jackson, 2003; Aguilera, Filatotchev, Gospel, & Jackson, 2008; La Porta, Lopez-de-Silances, Shleifer, & Vishny, 2000, 2002), and the association between VC monitoring and IPO performance (Bruton, Filatotchev, Chahine, & Wright, 2010). Institutional theory would argue that the effectiveness of the monitoring role played by VC firms is affected by formal institutions, law, and regulations and their ability to protect corporate shareholders from expropriation by controlling stakeholders (La Porta, Lopez-de- Silances, Shleifer, & Vishny, 1997), or expropriation of minority shareholders by corporate shareholders (Claessens, Djankov, Fan, & Lang, 2002; La Porta, Lopez-de- Silanes, and Shleifer, 1999; Shleifer & Vishny, 1997; Zingales, 1994). Compared to the United States, foreign VCs come from different backgrounds and are therefore subject to different legal, institutional, and cultural frameworks. These differences could have important consequences on the VC s legal protection rights effect over the firms they finance. Accordingly, higher legal protection rights of VC s country of origin should translate into a better governance effect of VCs over their backed firms and signal their good quality, which reduces underpricing at IPO. In addition, VC investment is a social business network, and therefore the composition of VC syndicates provides valuable insights into the value of portfolio companies. As such, a VC syndicate with foreign VCs coming from countries with low shareholders legal protection rights may send a bad signal on the quality of the IPO firm, which increases the risk premium, i.e., underpricing, required by outside investors at IPO. The contributions of our study are two-fold. On the one hand, this paper expands on prior research that combines institutional theory and agency perspective. As such, it complements prior results in Bruton et al. (2010) showing that multiple agency relationships are affected by different institutional contexts. Specifically, it shows evidence of the effect of differences in national legal frameworks on the effectiveness of the corporate governance role of VC firms at the time of IPO. Using the anti-director index (La Porta et al., 2000) and the anti-self dealing index (Djankov, La Porta, Lopez-de-Silanes, & Shleifer, 2008), it is the first paper to consider the effects of legal protection rights of the country of origin of the VC firm on the underpricing of IPOs in the US market. It also complements prior research that indicates a lack of evidence on the performance impact of foreign versus domestic VCs (Pruthi, Wright, & Lockett, 2003) and the effect of the VC country of origin (Wright, Pruthi, & Lockett, 2005). Consistent with the monitoring and certification roles played by VC firms, we add to the literature on foreign VCs. We find that VC syndicates involving foreign VCs coming from countries with high shareholders protection rights (e.g., UK and Canada) are likely to improve the overall level of governance standards of their investees. These syndicates also send a positive signal of the quality of the IPO firm, thus reducing IPO underpricing. Our findings also confirm the existence of a complementary role between the level of shareholders protection rights and board independence in affecting IPO underpricing. On the other hand, our study contributes to prior research on VC syndication, and suggests that the weighted average of shareholders protection rights of VC investors may reflect the quality of cross-border partnering and VC network. Although investors in the US are protected by stringent regulations, such as Sarbanes-Oxley legislation, and are able to sue managers who do not act in their best interest, VCs still have the flexibility to adopt or decline specific provisions that affect the level of shareholders legal protection rights (Klapper & Love, 2004). Therefore, the choice of foreign VC partners and the governance standards of their countries affect the risk premium required by investors at the time of IPO. This paper is organized as follows. First, we present the literature review and the hypotheses. Second, we describe the data and the methodology used in this research. Third, we discuss the empirical results and their implications, and finally, we draw some relevant conclusions. REVIEW OF LITERATURE AND HYPOTHESIS Underpricing and the Legal Protection Rights of the Country of Origin of VC firms An extensive body of research relies on the signaling theory to explain IPO underpricing. According to this theory, issuers send positive signals to the market about the quality of their public offerings. Similarly, VC firms might play a certification and monitoring role that helps reduce underpricing. While prior research concentrates on the monitoring/certification role from a signaling perspective, there is limited research that addresses this issue from a legal protection rights perspective, particularly the effect of the country of origin of VCs on the performance of their portfolio companies. Prior to funding a firm, VCs undergo a screening process to identify high growth, highly innovative, new technology, or entrepreneurial projects, such as investment opportunities that could provide a high return if they succeed (Gompers & Lerner, 2001; Gompers, 1995). VCs design financial contracts to monitor and add value through support activities (Kaplan & Stromberg, 2004). As such, they set up corporate governance control mechanisms that ensure the success of the IPO and protect shareholders rights, which in turn improve firm value and reduce IPO underpricing (Barry et al., 1990; Chahine, Filatotchev, & Wright, Volume 19 Number 6 November 2011

SHAREHOLDERS RIGHTS AND THE ORIGIN OF VENTURE CAPITAL FIRMS 603 2007; Megginson & Weiss, 1991). VCs also play an important certification role for their investees, whereby the presence of VCs in US IPOs is associated with superior certification (Megginson & Weiss, 1991). Thus, VC s certification role helps alleviate informational uncertainty about the IPO at the time of the issue (Hamao, Packer, & Ritter, 2000:556), and this is stronger in firms with higher informational asymmetry between insiders and outsiders (Li & Masulis, 2004). However, institutions may influence the goals and beliefs of individuals, groups, and organizations (North, 1990; Scott, 1995), and may have a significant effect on the goal formation and processes of venture capital firms (Wright, Thompson, & Robbie, 1992). The effectiveness of the monitoring role played by VC firms and their ability to design investments and financial contracts depend on the institutional environment (Kaplan & Stromberg, 2004). Institutional features, such as codes of corporate governance and ownership structure, are key variables affecting the governance and the investment behavior of VC firms (Jeng & Wells, 2000). The legislation of a country defines the specific rights and investor protection in its relations with firms. Investors behavior is thus the outcome of different forms of agency problems that are solved through different combinations of corporate governance mechanisms. For example, common law countries, including the United States, the United Kingdom, and other Anglo-Saxon countries, protect investors to a large extent, but remain indifferent to other stakeholders. This encourages dispersed shareholding structures that rely on outside directors to supervise managers. In contrast to the Anglo-Saxon model, civil law countries, including the Continental European legal system, offer weak legal protection to investors, which encourages the predominance of concentrated ownership to defend its interests, whereby large investors monitor and control managers, but might have conflicts of interest with minority shareholders (La Porta et al., 1997). Furthermore, differences in national institutions and legal systems may affect the effectiveness of corporate governance on the firm level (Aguilera and Jackson, 2003; Aguilera et al., 2008; La Porta et al., 2000, 2002). It may also affect the association between IPO performance and the role played by VC firms (Bruton et al., 2010). On the one hand, the monitoring role played by VC firms may be influenced by the legal protection rights of their countries of origin. VCs coming from countries with high corporate governance standards will, for example, include contractual provisions or extra protective mechanisms that may be lacking in their investees countries (La Porta et al., 1997, 1998, 2002; Kaplan, Martel, & Stromberg, 2007; Lerner & Schoar, 2005). The choice of control rights by VCs affects the corporate governance of their portfolio companies, and thus their performance and the likelihood of a successful exit (Cumming, 2008). Bruton et al. (2010) find that VCs in the UK, a common law country with strong legal protection of minority shareholders, add more value to their portfolio companies at IPO than VCs in France, a civil law country with a weak investor protection system. They conclude that performance outcomes of retained ownership of private equity investors are different depending on the legal system and institutional characteristics of the private equity industry in a specific country (Bruton et al., 2010:493). Additionally, Engelen and Van Essen (2010) posit that IPOs operating in countries with less developed legal frameworks will exhibit higher underpricing levels than IPOs operating in countries characterized by a well-developed legal framework. This suggests that the legal and institutional frameworks are a good proxy for corporate governance standards which, in turn, has a positive effect on firm value (Black, Jang, & Kim, 2006; Klapper & Love, 2004; among others). On the other hand, shareholders protection rights of the country of origin of a VC firm affect its certification role. Despite increasing cross-border VC investments (Mäkelä & Maula, 2008), the inclination of VC firms to invest in their own countries indicates the possible existence of liability of foreignness, which relates to additional costs incurred by a foreign VC firm compared to domestic firms (Lu & Hwang, 2010; Zaheer, 1995). Indeed, cross-border VC firms face various legislative, linguistic, cultural, and geographic barriers that make it challenging for ventures to attract foreign VCs (Mäkelä & Maula, 2005). Foreign VCs, which invest in unfamiliar cultures, bear a higher risk and uncertainty than local VCs (Hazarika, Nahata, & Tandon, 2009). In order to mitigate risk, foreign VCs engage in more rigorous screening and evaluation of their portfolio companies (Hazarika et al., 2009:27) than their domestic counterparts. Crossborder VC investments are thus facilitated by networks that confer social status in an intensely social business (Freeman, 2005:153) whereby VCs often syndicate their investments with others. While syndication allows VCs to share information and access ventures, resources, expertise, and risk (Guler & Guillen, 2010), it sends a signal on the quality of the venture. This is consistent with Podolny (2005) who shows how examples on social status may be used to infer quality in different trading contexts (banking, shipping, etc.), where social status reflects the standing, worth, esteem, or prestige of an actor in a given setting (Burt, 1992). This suggests that VC syndicates reflect the level of social status and the ability of a specific VC firm to attract VCs coming from countries with high governance standards. As such, the level of shareholders legal protection rights of foreign VCs within the VC syndicate may send a signal on the quality of the IPO firm. The presence of a foreign VC with high corporate governance standards may send a strong positive signal to IPO investors, thus reducing underpricing. Alternatively, investors might consider that whenever a US VC accepts to syndicate with VCs from countries with lower corporate governance standards as a bad signal. Outside investors might thus ask for a higher risk premium at the time of a public offering of issuing firms that involve VC firms with weak shareholders legal protection rights. This suggests that IPOs backed by both US and non-us VC firms might benefit from the higher level of protection rights provided by foreign VCs coming from countries with a better legal framework. They are thus likely to have a lower underpricing than those backed by VC firms coming from a weak legal framework. Therefore, we hypothesize the following: Hypothesis 1a. From an institutional perspective, IPO underpricing is negatively related to shareholders legal protection rights of the country of origin of VC firms. Volume 19 Number 6 November 2011

604 CORPORATE GOVERNANCE Since all VC-backed IPOs do not necessarily involve foreign VC firms, Hypothesis 1a compares the effect on performance of US shareholders protection rights in IPOs that involve US VCs only to a weighted average of protection rights in other IPOs with VCs coming from different institutional frameworks. As such, Hypothesis 1a relates to differences in the choice of institutional framework, being exposed to purely US or international legal contexts. Focusing on IPOs involving foreign VC firms would reflect the impact of the within-country differences by VCs coming from different institutional contexts on IPO performance. Hence, as Hypothesis 1a stipulates, we expect a lower underpricing in firms backed by VCs with higher legal protection rights in IPOs involving foreign VCs (H1b). Hypothesis 1b. From a within-country perspective, IPO underpricing is negatively related to shareholders legal protection rights in IPOs involving Foreign VCs. Underpricing and Complementarities between VC and IPO Firm Governance Standards In addition to the role played by the institutional framework in affecting the signaling and monitoring roles played by VC firms, IPO firms might signal their quality through good governance practices. Corporate governance practices are perceived as a means to reduce the information asymmetry and its associated agency costs that result in high underpricing (Certo, Daily, & Dalton, 2001; Filatotchev & Bishop, 2002). Several research studies indicate that corporate governance practices, such as board structures and characteristics (Certo et al., 2001; Higgins & Gulati, 2003) as well as retained ownership (McBain & Krause, 1989; Mikkelson, Partch, & Shah, 1997), could enhance the firm value, thus reducing underpricing. These practices also induce a better monitoring, thus decreasing agency costs (Fama & Jensen, 1983). A survey conducted by McKinsey and Company in 2002 with 201 institutional investors worldwide reveals that the overwhelming majority of these investors are ready to pay a premium for companies with high corporate governance standards (McKinsey & Company, 2002). This premium varies from 12 per cent in North America to 30 per cent in Eastern Europe and Africa. Accordingly, investors seem to value well-governed firms and IPO firms need to improve their corporate governance standards in order to attract investments. This is consistent with Hochberg (2008) who states that VCs choose to back portfolio firms that are better governed. Accordingly, the issue is to understand the combined effect on underpricing of the level of the protection rights of the country of origin of VC firms and the corporate governance practices of portfolio firms at the time of IPO. Cable and Shane (1997) use the prisoner s dilemma approach, pointing out that VCs and portfolio firms may cooperate or enter into defection, and this could have important repercussions on the success of the start-up. They conclude that the best strategy that should be pursued and that might lead to a successful VC-backed start-up is the cooperation between VCs and their portfolio firms. As such, VCs with high level of protection rights are expected to complement corporate governance practices within their portfolio companies. Prior research indicates that VCs may influence board composition and control over the portfolio companies, which affect the performance and success of these firms (Suchard, 2009). This consists of electing more independent board members and hiring more professional directors (Filatotchev & Bishop, 2002). Board independence is thus a key element that VCs could use to perform their monitoring and strategic functions, which mitigates the agency costs associated with their involvement in IPO firms (Chahine & Filatotchev, 2008). Additionally, as highlighted in Cumming, Schmidt, and Walz (2010), better laws facilitate board representation of investors. This suggests that the effectiveness of corporate governance mechanisms is different from one country to another and is moderated by institutional characteristics of a particular economic system (Dharwadkar, George, & Brandes, 2000; Douma, George, & Kabir, 2006; Hoskisson Johnson, Tihanyi, & White, 2005). Moreover, Klapper and Love (2004) show that firm-level governance is strongly and positively related to countrylevel measures of investor protection rights. In other words, average governance standards are higher in countries with stronger legal protection rights. Since the contractual agreements related to shareholders protection rights are negotiated between VCs and the various parties involved (i.e., insiders) prior to public offering, VCs with high protection rights are likely to invest in IPO firms with good governance practices, such as board independence. Thus, we expect the existence of a complementary role between portfolio firms governance practices and VC s country of origin legal protection rights at the time of the IPO, whereby firm-level practices complement VC-level standards and negatively affect IPO underpricing. Hence, we expand previous research and we hypothesize the following: Hypothesis 2a. IPO underpricing is negatively related to the level of shareholders legal protection rights of the country of origin of VC firms in IPOs with more independent boards of directors. In line with Hypotheses 1a and 1b, Hypothesis 2a relates to the differences in the effect of pure US institutional versus mixed institutional contexts. Hypothesis 2b focuses on IPOs involving foreign VC firms, and examines the impact of the within-country differences by VCs coming from different institutional contexts on IPO performance. Hence, parallel to Hypothesis 1b, we expect a lower underpricing in firms with more independent boards and which are backed by VCs with higher legal protection rights in IPOs involving foreign VCs. Hypothesis 2b. IPO underpricing is negatively related to shareholders legal protection rights of the country of origin of VC firms in IPOs involving foreign VCs and having more independent boards of directors. DATA AND METHODOLOGY The database includes 410 randomly selected US VC-backed IPOs from 1997 to 2007. To collect our studied firms, we have first identified the list of all US IPOs in the US markets from the Securities Data Company (SDC) database. We have then excluded all REITs, ADRs, closed-end funds, foreign IPOs, unit offerings, financial IPOs, and those with an offer price Volume 19 Number 6 November 2011

SHAREHOLDERS RIGHTS AND THE ORIGIN OF VENTURE CAPITAL FIRMS 605 TABLE 1 Data Representativeness Panel A Number and percentage of VC-backed IPOs per year Studied Sample (N = 410) Entire Population (N = 1,343) Year No. % No. % 1997 88 21.46 172 12.81 1998 25 6.10 97 7.22 1999 87 21.22 305 22.71 2000 108 26.34 265 19.73 2001 10 2.44 48 3.57 2002 10 2.44 42 3.13 2003 10 2.44 42 3.13 2004 21 5.12 123 9.16 2005 22 5.37 77 5.73 2006 12 2.93 78 5.81 2007 17 4.15 94 7 410 100 1,343 100 Panel B Number and percentage of VC-backed IPOs per industry Industry Classification No. % No. % Consumer Products and Services 36 8.78 126 9.38 Consumer Staples 6 1.46 22 1.64 Energy and Power 11 2.68 54 4.02 Healthcare 98 23.90 281 20.92 Software & IT Consulting Services 166 40.49 536 39.91 Industrials 22 5.37 50 3.72 Materials 9 2.20 29 2.16 Media and Entertainment 12 2.93 50 3.72 Retail 10 2.44 80 5.96 Telecommunications 39 9.51 113 8.41 Transportation 1.24 2.15 Total 410 100 1,343 100 Hi-Tech IPOs (%) 174 42.44 569 42.37 below five dollars. All IPOs without VC-backing are also excluded, and this resulted in a total list of 1,343 VC-backed IPOs. Our randomly selected subsample represents 30.5 per cent of all VC-backed IPOs over the period of study. VC characteristics are obtained from the VentureXpert database, whereas details on board composition were hand-collected from the IPO prospectuses available from the Securities and Exchange Commission s (SEC s) Electronic Data Gathering, Analysis, and Retrieval system (EDGAR). Table 1 compares the studied sample to the entire population of VC-backed IPOs over the studied period. Panels A and B show the distribution of IPOs across time and industries, respectively. Both panels indicate a very similar distribution to that of the entire population of VC-backed IPOs. Table 1 also indicates a similar percentage of hi-tech IPOs, which confirms the representativeness of the sample. Hence, our random sample is representative of the entire population of VC-backed IPOs. In order to examine the effect of the legal protection rights of VC firms on underpricing, we consider alternatively the effect of the weighted average legal protection rights of VC firms in the entire sample of studied IPOs, WA VCs PR Index (All IPOs), or in the sub-sample of IPOs involving foreign VC firms, WA VCs PR Index (IPOs with Foreign VCs). Since VC-backed IPOs usually involve a syndicate of several VC firms, the effect of legal protection rights may be influenced by the diversity of VC firms. VCs within the VC syndicate will set contractual agreements Volume 19 Number 6 November 2011

606 CORPORATE GOVERNANCE that reflect the VC s country of origin standards in terms of shareholders legal protection rights. Thus, the overall level of these rights should reflect the power game within the VC syndicate, whereby each VC firm will affect the portfolio firms legal protection rights according to its power within the syndicate. WA VCs PR Index is equal to the weighted average rating of the legal protection rights of the country of origin of all venture capital firms (both US and non-us) using their post-ipo ownership as weights as given in IPO prospectus. The rating of the country of origin of VC firms is calculated using both the anti-director rights and the anti-selfdealing rights. On the one hand, the anti-director index defined in La Porta et al. (2000) on the corporate investors legal protection of shareholders rights outlines six measures of shareholders protection rights. The index is a zero to six scale constructed by adding one when: 1) the country allows shareholders to mail their proxy vote to the firm; 2) shareholders are not required to deposit their shares prior to the General Shareholders Meeting (GSM); 3) cumulative voting or proportional representation of minorities in the board of directors is allowed; 4) an oppressed minorities mechanism is in place; 5) the minimum percentage of share capital that entitles a shareholder to call for an Extraordinary Shareholders Meeting (ESM) is less than or equal to 10 per cent (the sample median); and 6) shareholders have preemptive rights that can only be waived by a shareholders vote. On the other hand, the anti-self dealing index defined in Djankov et al. (2008) on minority shareholders legal protection of shareholders rights is the average of ex-ante and ex-post private control of self-dealing, and does not require a shareholder approval for related party transactions to protect minority shareholders against self-dealing. The index is a zero to one scale constructed by computing the average private control of self dealing variable each varying on a scale of zero to one also: (1) approval by disinterested shareholders; (2) disclosures by Buyer; (3) disclosures by Mr James; (4) independent review; (5) each of the elements in the index of disclosure in periodic filings; (6) standing to sue; (7) rescission; (8) ease of holding Mr James liable; (9) ease of holding the approving body liable; and (10) access to evidence (for further information, please refer to the table provided in Djankov et al., 2008). 2 Table 2 indicates that the database includes a total number of 670 VC firms with 162 foreign VC firms, such as ING Barings & Co. from the UK, S-E Banken from Sweden, or Toshiba Corporation from Japan. Regardless of their origin, VCs are general partners of a limited partnership or employees of financial institutions who are specialized in managing investments in new ventures for their respective groups (Bruton et al., 2010). While the US has the highest, i.e., the best, anti-director shareholders protection rights index (a rank of 5), it follows several countries, such as Singapore, the United Kingdom, Australia, and Israel in terms self-dealing director index. This suggests that the latter countries have better selfdealing shareholders protection rights than those in the US. In line with prior research, we calculate underpricing as the first day return (Chahine & Goergen, 2011a). Board independence is the fraction of independent board members, i.e., non-executives and non-related to any of the shareholders. TABLE 2 Distribution of VCs per Country of Origin and Legal Protection Rights Country Number Anti-director Right Index* Self-dealing Right Index* Australia 2 4.79 Belgium 1 0.54 Canada 23 5.65 Denmark 3 2.47 France 11 3.38 Germany 12 1.28 India 1 2.55 Israel 5 3.71 Italy 1 1.39 Japan 16 4.48 Netherlands 12 2.21 Singapore 5 4 1.00 South Korea 4 2.46 Sweden 4 3.34 Switzerland 11 2.27 Taiwan 13 2.56 United Kingdom 38 5.93 United States 508 5.65 Total Number of VC firms 670 *The anti-director rights index and the self-dealing rights index are those defined in La Porta et al. (2000) and Djankov et al. (2008) respectively. In further interactive tests, we have used a Board Independence dummy that is equal to one for IPOs with board independence that is higher than the median value of 42.86 per cent, zero otherwise. Empirical investigations also add a list of control variables that are usually used in IPO literature. First, we consider some VC-related variables that include VC syndicate, VC Syndicate Retained Ownership, VC Syndicate Age, and VC Syndicate Participation. VC Syndicate is the total number of VC firms within the IPO syndicate. De Clercq and Dimov (2008) argue that VC syndicates share knowledge and collaborate with other firms, which create value to their portfolio companies. VC Syndicate Retained Ownership is the post-ipo ownership of VC firms as listed in the IPO prospectus. Prior research indicates that VCs may make decisions that are in their own interest and act to the detriment of the other shareholders. Gompers (1996) argues that VCs may grandstand and take their portfolio companies public prematurely to enhance their own reputation. Investors are thus likely to face greater conflicts of interest and may require a higher risk premium, which increases underpricing. This suggests that VCs may first start by grandstanding, and over time, become more reputable and credible to certify the quality of their backed IPOs (Chahine, Filatotchev, & Zahra, 2011). Volume 19 Number 6 November 2011

SHAREHOLDERS RIGHTS AND THE ORIGIN OF VENTURE CAPITAL FIRMS 607 Underpricing is thus negatively related to VC age and VC reputation. Since we do not have access to the reputation of VCs in an international context, we have referred to their age as a proxy for experience (Lee & Wahal, 2004). VC Syndicate Age is the cumulative age of VC firms within the syndicate. VC Syndicate Participation is equal to the number of shares sold by the VCs in the IPO as a fraction of the total number of shares offered in the IPO. In terms of firm characteristics, we consider variables related to firm size and industry membership. Pre-IPO Asset is the total assets for the last year prior to the IPO date. Empirical tests use a log-normalized asset using the logarithm of the Asset, Log (Asset). Firm Age is the age of the IPO firm, in years, since inception. Hi-tech dummy is equal to one if the IPO firm is a hi-tech firm, and zero otherwise. 3 We further control for the effect of the IPO method on underpricing. According to Jenkinson and Jones (2004) and Ljungqvist, Jenkinson, and Wilhelm (2003), book building procedure characterized by firm commitment contracts minimizes the risk of IPO failure. We control for the type of underwriting contract by using a Firm Commitment dummy variable, taking the value 1 if the underwriting contract is of firm commitment type and zero otherwise. We have also added Aftermarket std-dev which measures ex-ante uncertainty of the IPO firm. Beatty and Ritter (1986) indicate that underpricing could be explained by the ex-ante uncertainty surrounding the intrinsic value of an issue. Aftermarket std-dev is the standard deviation of returns over the 20 post-ipo trading days, excluding the first trading day return. Field and Hanka (2001) indicate that VCs may be subject to lock-up agreements that are used to support the aftermarket performance of IPO shares. We control for this effect using Lockup period that is equal to the difference, in days, between the IPO date and the lock-up date (Chahine & Goergen, 2011b). We also control for the impact of the quality of regulation of the exchanges listing requirements and standards using a Nasdaq dummy variable taking a value of one if the portfolio firm is listed on the NASDAQ and zero otherwise. The quality of regulation, especially in terms of corporate governance rules, is higher for the NYSE or AMEX than that of the NASDAQ (Affleck-Graves, Hedge, Miller, & Reilly, 1993). Moreover, firms listed on the NASDAQ are usually associated with higher underpricing levels as these firms are surrounded by a higher uncertainty. We have also added Underwriter Reputation which is calculated based on underwriter ranking in Loughran & Ritter (2004) (0 to 9, least to most prestigious underwriter). To control for market conditions, we have used Market Return, which is equal to the compounded daily return of a value weighted index over the 20 trading days prior to the IPO date. We have finally controlled for the bubble period using a Bubble Period dummy which is equal to one if the IPO occurs in 1999 2000, and zero otherwise. Using the WA VCs PR Index, we first examine the simple effect of the weighted average protection rights index of VC firms (Equation 1). Later, we verify the complementary versus substitution role played by both board independence and the weighted average protection rights of VC firms (Equation 2). Our two main testable equations are as follows: Underpricing = α + β 1 WA VCs PR Index ( All IPOs or IPOs with ForeignVCs) + β VC Syndicate + β VC Syndicate Retained Ownership + β 2 VC 3 4 Syndicate Age + β5 VC Syndicate Participation + β6 Board Independence + β7 Log( Asset) + β8 Firm Age + β9 Hi- tech dummy + β10 Firm Commitment dummy + β11 Aftermarket std- dev + β12 Lockup period + β13 Nasdaq dummy + β Underwriter Reputation + β Market Return + β 14 15 16 Bubble dummy + ε 1 (1) Underpricing = α + β1 Independent Board dummy + β2 WA VCs PR Index ( All IPOs or IPOs with Foreign VCs) + β 3 Independent Board dummy WAVCsPRIndex( AllIPOsorIPOswithForeignVCs) + β + β + β 4 VC Syndicate 5 VC Syndicate Retained Ownership 6 VC Syndicate Age + β7 VC Syndicate Participation + β8 Log( Asset) + β9 Firm Age + β Hi- tech dummy + β Firm Commitment dummy + 10 11 β12 Aftermarket std- dev + β13 Lockup period + β14 Nasdaq dummy + β15 Underwriter Reputation + β16 Market Return + β Bubble dummy + ε 17 2 (2) EMPIRICAL RESULTS Descriptive Statistics Table 3 presents the descriptive statistics for the entire sample of 410 IPOs as well as for both subsamples of IPOs with and without foreign VC firms within the VC syndicate. This shows that more than one-third (36.6 per cent) of the studied sample involves foreign VC firms. Panel A describes the firm and market conditions at the time of IPOs whereas Panel B exposes the characteristics of VC firms. Panel A of Table 3 indicates an average underpricing of 40.1 per cent, which is highly skewed when compared to the median value of 10 per cent. High skeweness is related to the large fraction of firms going public during the bubble period within the studied sample (47.8 per cent). Table 3 also shows no significant difference between firms with and those without foreign VC firms. In terms of corporate governance, the average firm has 41.5 per cent independent board members, and this is significantly higher in firms with foreign VCs. Board independence is significantly different (p < 1 per cent) for IPOs backed by US VCs only (39.3 per cent) and those involving foreign VCs (45.2 per cent). One explanation is that foreign VCs rely on independent board members to help monitor their portfolio companies. Although not shown in Table 3, this represents around three members out of an average board size of seven members. Although not shown in the table, further investigations indicate that IPOs with a lead VC firm coming from the UK (one of the highest legal protection rights, an anti-director right index of 5, and an anti-self-dealing right index of.93) exhibit an average underpricing of 3.73 per cent, which is lower than the average firm in the studied sample. They also indicate an average board independence of 51.9 per cent that is higher than the average firm in the studied sample. In terms of firm characteristics, an average IPO firm has total asset size of $176.59 million and was created 8.46 years prior to IPO date. It goes public following a positive market Volume 19 Number 6 November 2011

608 CORPORATE GOVERNANCE TABLE 3 Descriptive Statistics Total Sample (N = 410) IPOs without Foreign VCs (N = 260) IPOs with Foreign VCs (N = 150) Mean Median s.d. Mean s.d. Mean s.d. T-test Diff Panel A Firm Characteristics and Market Conditions at IPO Underpricing.40.10.73.41.77.38.66.65 Board Independence.41.40.21.39.21.45.20.00** Pre-IPO Total Assets 176.58 27.81 683.40 211.71 833.17 115.70 263.53.17 Firm Age 8.45 5.77 9.67 8.60 10.15 8.19 8.79.67 Hi-tech dummy.42.00.49.40.49.45.49.36 Firm Commitment dummy.94 1.00.23.94.23.94.22.85 Aftermarket std-dev.05.05.03.05.03.05.03.52 Lockup period 124.36 180.00 89.57 131.84 87.92 111.41 91.19.02* NASDAQ dummy.92 1.00.27.90.30.96.18.02* Underwriter reputation 8.60 9.00 1.15 8.56 1.32 8.68.78.33 Market Return.01.01.04.01.03.01.04.28 Bubble dummy.47.00.50.45.49.51.50.27 Panel B VCfirmCharacteristics VC Syndicate 3.56 3.00 1.91 3.15 1.77 4.28 1.96.00*** US VC Number 3.06 3.00 1.74 3.15 1.77 2.90 1.68.16 Foreign VC Number.50.00.80 1.37.75 VC Syndicate Retained Ownership.40.41.19.38.19.42.20.09 US VC Retained Ownership.35.35.20.38.19.29.20.00*** Foreign VC Retained Ownership.04.00.09.13.12 VC Syndicate Age (Total) 41.32 33.00 37.09 38.96 37.67 46.52 35.39.06 US VC AGE 37.30 28.00 36.36 38.96 37.67 33.66 33.19.19 Foreign VC AGE 4.02.00 9.54 12.85 13.35 VC Syndicate Participation.04.00.15.04.13.04.17.84 PR Index Using Anti-director Rights (AD) WA VCs PR Index (AD) 4.79 5.00.55 5.00.00 4.43.79.00*** PR Index Using Anti-self dealing Rights (SD) WA VCs PR Index (SD).64.65.06.65.00.62.10.00*** ***, **, *, stand for statistical significance at the.1%, 1%, 5%, and 10% level, respectively. Table 3 presents the descriptive statistics for a sample of 410 US IPO firms that went public from 1997 to 2007 in the US markets. It describes the entire sample as well as both subsamples of firms with and those without foreign VC firms (260 and 150 IPOs, respectively). Panel A presents statistics on firm and market conditions characteristics at IPO date, and Panel B exposes statistics on the characteristics of VC firms within the VC syndicate. Underpricing is the first day return. Board Independence is the fraction of independent board members, i.e., non-executives and non-related to any of the shareholders. Pre-IPO Asset is the total assets for the last year prior to the IPO date. Firm Age is the age of the IPO firm, in years, since inception. Hi-tech dummy is equal to one if the IPO firm is a hi-tech firm, and zero otherwise. Firm Commitment dummy is equal to one if the company goes public using a firm commitment agreement, and zero otherwise. Aftermarket std-dev is the standard deviation of the rate of return of IPO firm during the 20 day-period following the first day of trading. Lockup period is the difference, in days, between the IPO date and the lock-up date. Nasdaq dummy is equal to one if the IPO firm goes public in the NASDAQ, and zero otherwise. Underwriter Reputation is calculated based on underwriter ranking in Loughran and Ritter (2004). Market Return is equal to the compounded daily return of a value weighted index over the 20 trading days prior to the IPO date. Bubble Period dummy is equal to one if the IPO occurs in 1999 2000, and zero otherwise. VC Syndicate is the total number of VC firm within the VC syndicate. US VC Number is the total number of US VCs in the VC Syndicate. Foreign VC Number is the total number of Foreign VCs in the VC Syndicate. VC Syndicate Retained Ownership is the total post-ipo ownership of all venture capital firms as listed in the IPO prospectus. Foreign VC Ownership is the total post-ipo ownership of foreign (non US) venture capital firms as listed in the IPO prospectus. VC Syndicate Age is the cumulative age of VC firms within the syndicate. It is calculated for both US and foreign VCs, US VC Age and Foreign VC Age, respectively. VC Syndicate participation is equal to the number of shares sold by the VCs in the IPO as a fraction of the total number of shares offered in the IPO. WA VCs PR Index stands for the weighted average rating of legal protection rights index of VC firms within the VC syndicate. The rating of legal protection rights refers to both the anti-self dealing rights (SD) and the anti-director rights (AD) (Djankov et al., 2008). Weights are calculated based on post-ipo ownership as listed in the IPO prospectus. Volume 19 Number 6 November 2011

SHAREHOLDERS RIGHTS AND THE ORIGIN OF VENTURE CAPITAL FIRMS 609 return of 1.2 per cent over the 20-day period prior to IPO date; it is followed by an aftermarket standard deviation of 5.7 per cent, and is likely to be managed by a reputable underwriter with an average of 8.609 using the Loughran and Ritter (2004) ranking (0 to 9, least to most prestigious underwriter). Although not shown in Table 3 and 78 per cent of the studied sample was underwritten by a Class 9 prestigious underwriter. The studied sample also includes a significant fraction of hi-tech firms (42.4 per cent), and a large fraction of firms that go public using a firm commitment method (94.4 per cent). IPO firms are likely to go public on the NASDAQ (92 per cent for the total sample), and this is more significant for firms with foreign VCs compared to those involving US VCs only (90 per cent versus 96.6 per cent, p < 5 per cent). They have an average lock-up period of 124 days, whereby IPOs backed by US VCs and those involving foreign VCs have an average lockup period of 131 and 111 days, respectively. This difference is more significant for IPOs involving foreign VCs (p < 5 per cent). Overall, there is a weak difference in IPO firms characteristics between firms with and those without foreign VCs. Panel B shows that an average IPO firm involves an average of 3.566 VC firms, and the number of VCs in IPO firms with foreign VCs is significantly higher than that with US VCs only (4.280 versus 3.154, p <.1 per cent). Panel B indicates that the number of US VC is not significantly different between both sub-samples, and that the average number of US VCs within the sub-sample of firms involving foreign VCs is equal to 2.907 (versus 1.373 foreign VC firms). VC firms retain an average 4.1 per cent, with US VC firms holding an average of 35.3 per cent following the IPO date. IPOs with foreign VCs have a significantly higher VC retained ownership than IPOs without foreign VCs (42.2 per cent versus 38.8 per cent, p < 10 per cent). The cumulative age of VC firms within the syndicate reaches an average of 41.33 years per IPO firm, and this is mainly related to US VC firms with an average cumulative age of 37.30 years for all US VCs within each syndicate. VC-backed IPOs involving foreign VCs have VCs which are slightly older than IPOs without foreign VC involvement (46.5 years versus 38.9 years, p < 10 per cent). Moreover, VC syndicate participation in public offering is equal to 4.4 per cent on average, and this is not significantly different between both subsamples. Using the anti-self dealing rights to calculate our Weighted Average of VCs Protection Rights Index, we find a.65 WA PR Index for IPOs without foreign VCs, and a.63 weighted average protection rights index for those firms that involve foreign VCs. Using the anti-director rights, the weighted average of VCs protection rights index for IPOs without foreign VCs is 5, and 4.44 for IPO firms including foreign VCs. Both methods used for VC firms PR indices indicate that IPO firms without foreign VCs have a higher weighted average protection rights index than firms involving foreign VCs (p <.1 per cent). This suggests that foreign VCs investing in the US come from environments that are characterized by lower shareholders protection rights than US VC firms. Table 4 exhibits the correlation coefficients between studied variables. This shows a negative, but weak, association between underpricing on the one hand, and foreign VC presence and PR indices, on the other hand. Although not reported in Table 4, the Variance Inflation Factor (VIF) test is lower than 1.93 which rejects the likelihood of multicolinearity, and this is consistent with the low correlation coefficients between the independent variables used in Equations 1 and 2. Underpricing and Legal Protection of Shareholders Rights of VC Firms Table 5 includes the regressions run for underpricing on the legal protection rights indices of VC firms using the rating of the country of origin based on both anti-director and antiself dealing rights. Panel A refers to the rating of the country of origin using the anti-director rights whereas Panel B presents results based on the rating of the country of origin using the anti-self dealing rights. Models 1a and 1b examine the effect of the weighted average legal protection of shareholders rights for VC firms in all IPOs using the rating of the country of origin based on anti-director rights and anti-self dealing rights, respectively. Both models indicate a negative association between underpricing and the weighted average protection rights index of VC firms in all IPOs (t =-2.26 versus -2.16, p < 5 per cent), which is consistent with Hypothesis 1a. This suggests that the higher the legal protection of shareholders rights, the lower the first day return. Accordingly, VCs coming from countries with higher legal protection rights than those of the US will play an important screening and certification role of their portfolio companies, which reduce underpricing at the time of IPO. Models 2a and 2b focus on the effect of the weighted average legal protection of shareholders rights of VC firms in the subsample of IPOs involving foreign VCs using the rating of the country of origin based on anti-director rights and anti-self dealing rights, respectively. Both models show consistent results with Models 1a and 1b and indicate a negative and significant association between underpricing and legal protection of shareholders rights (t =-2.68 versus -2.85, p < 1 per cent), and the results are more significant than in Models 1a and 1b, which is consistent with Hypothesis 1b. Although further investigations not shown in Table 5 do not show any significant relationship between Foreign VC dummy and underpricing, our results confirm the positive role played by differences in legal and institutional frameworks rather than US versus non-us origin of VC firms. This also suggests that VC firms with higher legal protection of shareholders rights, especially foreigners, are likely to play a better monitoring and certification role, thus reducing underpricing. In terms of control variables, all models show that underpricing is positively related to the size of VC syndicate (p < 10 per cent), and negatively related to the age of VCs within the syndicate (p < 10 per cent). Moreover, there is a negative association between board independence and underpricing (at the 1 per cent level or better). Underpricing is also negatively related to firm size (p < 10 per cent), and to the length of the lockup period (p < 10 per cent). On the Volume 19 Number 6 November 2011