The role of corporate governance systems and investment stage on corporate venture capital performance

(1)

1

The role of Corporate Governance Systems and

Investment Stage on Corporate Venture Capital

Performance.

University of Amsterdam Business School

Date: January 27, 2017. Name: Roy Asiedu

Student number: 11116781

Msc of Business Administration: International Management Master Thesis

Word Count: 17.162

Supervisor: Dr. Vittoria Scalera

(2)

2

Statement of Originality

This document is written by Student Roy Asiedu who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

3

Abstract

Research on corporate governance has proven its importance and influence on a firm's way of doing business. However, few studies have focused on the influence of corporate governance on corporate venture capitals, meanwhile these studies show that corporate venture capitals are to some extend subject to corporate governance. Simultaneously, the corporate venture capital market is growing rapidly, influencing both our society and economy. Therefore, in this study I further combine the corporate governance literature with corporate venture capitals, and examine how two dichotomous corporate governance systems, Anglo-Saxon and European Continental, influence corporate venture capitals with regard to their performance and investment stage decision. In turn, I examine the role of investment stage with regard to corporate venture capital performance to establish whether investment stage mediates the relationship between corporate governance systems and corporate venture capital performance. The data consists of corporate venture capital investments between 2002 – 2011 and is extracted from VentureXpert (Thomson One). Subsequently, to conduct this study I use multiple binary logistic regressions. The results of this study show that corporate venture capitals from the Anglo-Saxon system perform better than corporate venture capitals from the European Continental system when performance is measured by the ratio of successful exits of their portfolio companies. Furthermore, later stage investments partly mediate the relationship between corporate governance systems and corporate venture capital performance. This study contributes to the literature by further showing that corporate governance does have an influence on corporate venture capitals. In addition, it contributes by showing the effect of, and importance of acknowledging, the dichotomous corporate governance systems as a whole in relation to corporate venture capital performance and the role of investment stages.

Key words: Corporate Venture Capital; Performance; Corporate Governance; Anglo-Saxon system; European Continental System; Investment Stage.

(4)

4

Introduction

In June 2016, Deutsche Börse, Europe’s largest exchange operator, launched a new Venture Capital (VC) fund targeting fintech in capital markets. Under the name DB1 Ventures, the fund joined the VC market that, according to a report from KPMG international examined by CB insights, financed $14.4bn - almost double to previous year - to only fintech companies (Stafford, 2016). In the same month, Kellogg company announced it was expecting to launch its own $100M Corporate Venture Capital (CVC) fund, investing in innovative companies pioneering in new ingredients, foods and packaging (Finsmes, 2016). Similar CVCs, focusing especially on the IT sector and increasingly investing billions of dollars, have already been established over the past decades by mainly US firms such as Microsoft, Google and Intel (CB Insights, 2015). Due to current low interest rates forcing investors to seek for return (Satariano, Bloomberg, 2016) and rising trends such as crowdfunding, the VC market gains more and more popularity worldwide (Oster & Chen, Bloomberg, 2016) with both societal and economic

benefits1_{(Anderson, 2006).}

CVCs are distinctive from Independent Venture Capital (IVC) since they do not only pursue financial returns, but mainly focus on acquiring technology and innovative insights (Dushnitsky & Shaver, 2009). Several studies have focused on the determinants of CVC activity, such as (legal) institutions and culture (Cumming, 2006; Li & Zahra, 2012; Nahata, Haziraka & Tandon, 2014). Furthermore, International Business (IB) literature shows that Corporate Governance (CG) is an important determinant for a firm’s way of doing business. Examples of these are studies that have examined what the effects are of CG factors on a firm’s performance, strategy, voluntary disclosure, fraud, and other facets (Argawal & Chadha, 2005;

1_{This is a study commissioned by the National Venture Capital Association by Stuart Anderson (2006), Executive}

Director from the National Foundation for American Policy. The societal and economic benefits include innovative products that shape society and growing Portfolio companies that foster job creation.

(7)

7 Baysinger & Butler, 1985; Eng & Mak, 2003). In most CG comparison studies, two main systems are identified; the Anglo-Saxon system and European Continental System. Subsequently, using countries as a proxy for CG systems, studies have examined the CG systems and their effects on a firm’s way of doing business, and more specifically the effects on (C)VCs (Franks & Mayer, 1997; Frick & Lehmann, 2005; Goodijk, 2000; Kaplan, 1997). For example, Black and Gilson (1998) compared the Anglo-Saxon system with those from Germany and Japan and its effect on CVCs. Significant differences and effects on CVCs were found.

CVCs can invest in Portfolio Companies (PC) at different moments, ranging from the moment the PC is founded until the phase just before the exit. Several studies have examined CVC’s investment stage decision and its relation to performance (Elango, Fried, Hisrich & Polonchek, 1995; Ruhnka & Young, 1987; Sahlman, 1990; Tyebjee & Bruno, 1984). However, most of these studies used questionnaires and identified CVCs’ expectations of risk of loss as a proxy for performance, rather than the ‘actual’ performance by looking at the value created or the amount of successful PC exits. Subsequently, although the investment stage is an important decision within the CVC’s strategy, no study yet examined the role of CG systems on investment stage. Previous studies only looked at specific (firm-related) CG characteristics or non-CG characteristics that could influence the investment stage (Anokhin, Peck & Wincent, 2016; Sahlman, 1990). Furthermore, some studies compared the US and Europe with regard to CVCs and their performance (Hege, Palomino & Schwienbacher, 2003, 2009), but selected countries or continents with different characteristics rather than looking specifically at the two dichotomous systems.

Thus, I investigates the effect of the two dichotomous CG Systems on CVC performance. I measure CVC performance by looking at the ratio of their PCs that successfully

(8)

8 exit through either IPO or acquisition, compared to the unsuccessful ones that are written off (Gompers, 1995). Thereafter, I investigate how the investment stage plays a role in the relationship between CG systems and CVC performance. Rather than using separate countries as a proxy for CG systems (e.g. US vs Germany), I specifically combine the countries from

both the Anglo-Saxon system and the European Continental system2. All data for this study is

extracted from VentureXpert (Thomson One). Subsequently, I use multiple binary logistic regressions to examine the relationships. For a robustness check, I also run the regression with robust standard errors. Dealing with some of the limitations of SPSS, I use the Baron and Kenny steps (Baron & Kenny, 1986) for the mediation analysis to examine the role of the investment stage in the relationship between CG systems and CVC performance.

The results of this study show that CG systems influence CVC performance in a way that CVCs from the Anglo-Saxon system perform better than CVCs from the European Continental system with regards to the ratio of successful exits of their PCs. Furthermore, the results show that the investment stage in later stage investments partly mediates the relationship between CG systems and CVC performance, and thus plays a role. Herein, CVCs from the Anglo-Saxon system tend to invest more in later stages, whereas later stage investments have a better performance.

This study contributes to the literature by further showing that CG does have an effect on CVCs and that future research should not neglect the combination of CG literature with the aspects of CVCs. Furthermore, this study extends the existing literature by comparing a combination of all countries from both systems, rather than separately comparing countries, strengthening the importance of acknowledging the dichotomous CG systems as a whole with regards to CVC performance and the investment stage. In turn, this study shows that the general

2_{The countries United States, Canada and United Kingdom are used as a proxy for the Anglo-Saxon system, and}

(9)

9 acknowledged rule of ‘high risk - high return’ does not always yield since later stage investments, depicted as lower risk, have a better performance than early and balanced stage investments. This difference might be explained by the fact that different studies use different proxies of return to express CVC performance, suggesting future researchers to carefully consider what measurement(s) of CVC performance they use. Lastly, in congruence with former studies by Sahlman (1990) and Ruhnka and Young (1987, 1991), this study sheds light on the importance of the investment stage. The results extend the literature by showing that this factor partially mediates the relationship between CVC’s CG systems and performance, and that future studies comparing the dichotomous CG systems with regards to CVC performance should take this factor into account.

In addition, this study provides practical contributions for managers, investors, governments and other policy makers. First, it shows the importance for CVC managers and investors, whilst drafting their strategies including decisions on finding syndication partners, to acknowledge that CVCs from the Anglo-Saxon system compared to their European Continental peers tend to be more finance orientated, invest more in later stages and have a better performance. Furthermore, governments and other policy makers from specifically the European Continental system should revise their CVC industry and examine possible actions such as deregulation of PC IPOs to increase the attractiveness of their CVC market.

First, I discuss the theoretical foundation of this study by framing the literature to identify a clear research gap and concise research question. In the second section, I describe the development of the hypotheses. Subsequently, I describe the research design and methodology that are to conduct this study which is followed by the description of the results. Lastly, I discuss the implications, managerial implications, limitations and future research recommendations in the discussion and conclusion section.

(10)

10

Literature Review

This literature review starts by exploring (C)VC and CG. Subsequently, I combine both aspects and also examine the investment stage. Thereafter, I examine and clarify the term ‘performance’ since it can be interpreted and measured in many ways. Lastly, the research gap summarizes the findings of the literature review and points out the gap for this research, resulting in the research question.

(Corporate) Venture Capital

VC is defined as equity or equity-linked investments in young, privately held companies, where the investor is a financial intermediary who is often active as a director, advisor or even manager of the firm (Kortum & Lerner, 2000). After a parsimonious $100-200 annual contributions to VC funds during the 1970s, pension fund commitments to VC rose dramatically to in excess of $4 billion by the end of 1980 (Gompers, 1994); a new way of

financing was born3.

The most bespoken difference within the VC literature is the distinction between CVC and IVC. Maula (2001) defines CVC as “equity or equity-linked investments in young, privately held companies, where the investor is a financial intermediary of a non-financial corporation” (p. 9). Also used by Maula (2001), Hellmann & Peri (2002) define IVC as “professionally managed equity investments of young, growth-oriented private companies”, thus leaving out the part of the investor being a non-financial corporation. There are three main differences identified between IVC and CVC. First, their goals and objectives differ. Independent VCs invest in risk-oriented entrepreneurial businesses and seek capital appreciation through lucrative exits, such as initial public offering (IPO) or acquisition

3_{Remarkable is that this solely applied for the VC industry in the US (Gompers, 1994). The European VC industry}

(11)

11 (Sapienza, 1992). Some CVCs also pursue financial returns, however most CVCs focus on acquiring technology and innovative insights (Dushnitsky & Shaver, 2009). Secondly, they differ in their capabilities and value adding activities with regard to their PCs. IVCs provide assistance with strategy formulation, administrative support, personnel recruitment, and networking entrepreneurs with investors and potential acquirers to obtain financing (Maula & Murray, 2001; Sapienza, 1992). On the other hand, CVCs tend to have stronger and unique services that capitalize on corporate resources. This includes providing access to corporate laboratories, customer, partner and supplier networks, distribution channels, and supporting their technological development (Teece, 1986). Lastly, the two differ in the sense that IVCs are solely in the business of financing new ventures, whereas CVCs are part of a corporation and thus might be sensitive to the ventures activity and way of doing business by the corporation (Dushnitsky & Shaver, 2009).

Several studies show CVCs are affected by their institutional environment, whereas institutions are, both formal and informal, defined as humanly devised constraints that structure political, economic and social interaction (North, 1991). Williamson (2000) gives structure to this institutional environment by identifying four levels of institutions. Level one is the embeddedness and is linked to cultural aspects and informal institutions. Level two is the regulatory level and is linked to formal institutions. Furthermore, level three and four are the 'play of the game', incorporating corporate governance and minor rules. Li and Zahra (2012) find that formal institutions have a positive effect on the level of VC activity. However, the effect is weaker in more uncertainty-avoiding and collectivist societies, and thus moderated by cultural aspects. Another study by Nahata, Haziraka and Tandon (2014) supports the first finding of Li and Zahra (2012) by showing that for both developed and emerging economies the superior legal rights (and enforcement), which can be seen as part of the formal institutional

(12)

12 environment, enhance VC performance. However, even in countries and continents with a somehow equal ‘strong’ institutional environment, such as the US and Europe, their CG and its corresponding facets considerably vary. The influence of CG on business is shown by several studies (Argawal & Chadha, 2005; Baysinger & Butler, 1985; Eng & Mak, 2003). Therefore, in the following paragraph I continue to examine CG and later on its relation to CVC.

Corporate Governance

CG may be broadly defined as the study of power and influence over decision making within the corporation (Aguilera & Jackson, 2010). It involves the firm's organization and accountability to various stakeholders, shareholders, employees, local communities, boards and directors. In most studies comparing CG systems there is a distinction between the Anglo-Saxon

system and the Continental European system - Rhineland Model4_{(Aguilera & Jackson, 2010).}

The first is mainly present in the United Kingdom (UK), Canada (CA) and United States (US), whereas the latter is mainly present in Germany (GE), the Netherlands (NL), France (FR), Italy (IT) and Belgium (BE) (Aiginger & Guger, 2006; Cernat, 2004; Koen & Mason, 20055). According to la Porta et al. (1998), deficiencies in the CG systems are linked to the legal tradition of a country. Therefore, the Anglo-Saxon system finds it heritage from the common law, a legal system associated with its origin and development in England. It is based on judicial precedent, whereas the European Continental system operates in a civil law environment, a legal system based on statutes and comprehensive codes (Dainow, 1966). The Anglo-Saxon system is known for dispersed ownership and strong shareholder rights, whereas the

4_{Within CG literature the Continental European system is sometimes called the Rhineland Model and vice versa}

(Aiginger & Guger, 2006).

5_{To which CG system some specifically countries belong is still in debate. For example, some researchers such}

as Aiginger and Guger (2006) also distinguish between the Scandinavian or Mediterranean model. However, I chose to follow the majority of studies that mainly distinguishes between the 2 major systems (Koen & Mason, 2005).

(13)

13 Continental European system is known for concentrated block-holder ownership and weak shareholder rights (Becht & Roell, 1999; La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 1998; Shleifer & Vishny, 1997). Subsequently, the role of banks in the Anglo-Saxon system is minimum, resulting in it being characterized by short-term equity finance and active markets for capital control, in contrast to the European Continental system, where banks play an important role in both long-term debt finance and control (Cernat, 2004). Also the formation of the management board is different, whereas the Anglo-Saxon system has one-tiered boards, and the European Continental system has two-tiered boards; executive and supervisory responsibilities separated (Cernat, 2004). The Anglo-Saxon system is also known as the shareholder-model or market-oriented model examined by Moerland (1995), stating that the cons of this system are short-termism, costs related to defense against hostile takeovers due to capital control and empire-building by corporate management. On the other hand, an advantage is the fact that managers are challenged to do their best due to the existence of an external threat to be replaced. In contrast, the European Continental system is known as the stakeholder-model or network-oriented model. It appears to have more intimate and long term relationships, lower bankruptcy costs due to more support by the network and less agency conflicts between share- and debt-holders since the banks are internalized. However, a disadvantage is the entrenched position by the management which might result in too much protection.

Looking at all the above distinctive CG factors of both systems, one might argue that these differences affect the firm’s way of doing business. For example, Farber (2005) find that fraud firms have fewer numbers and percentages of outside board members, fewer audit committee meetings and CEOs who are also chairmen of the board of directors. Furthermore, La Rocca (2007) shows that the relation between capital structure and a firm’s value is moderated and/or mediated by a firm’s CG. Other researchers address the influence of CG

(14)

14 (factors) on accounting scandals (Argawal & Chadha, 2005), voluntary disclosure (Eng & Mak, 2003), and performance (Baysinger & Butler, 1985). Although the above studies focus on different research topics, all have in common showing that CG is affecting different types of operations and facets of a firm.

According to Kaplan (1997), there is no clear difference between US and German CG systems in disciplining poor managerial performance. However, a remarkable difference is that US systems are more effective in discouraging successful companies from overinvesting. This is due the fact that managers hold larger equity positions in contrast to Germany, and are therefore more vigilant. Franks and Mayer (1997) make a distinction between the UK on the one hand, and France and Germany on the other hand, and found that the European mainland countries, characterized by concentrated ownership, encourage longer-term relationships between the firm and its investors. Both of the above studies show characteristics of the firm that are affected by the CG systems. In general, consistent with its characteristics, all of the stakeholders in the European Continental system are more or less able to monitor and influence a firm’s strategy and performance. This is also appointed by Frick and Lehmann (2005) in their book saying that “CG of German firms is mostly shaped by factors that are internal to the firm” (p. 123). In conjunction with this is the finding of a study by Goodijk (2000), saying that mainly in Europe there are employee councils that discuss strategic topics and decision making with the board of directors.

Thus, altogether the aforementioned studies show that besides differences in the effect of independent CG factors on several facets of the firm, also a distinction is made between CG systems and its effect on the firm’s way of doing business. CG can thus be approached by measuring specific organizational factors or characteristics, such as board composition, or by looking from a broader view and examine the characteristics of the system wherein the

(15)

15 organization operates. This study uses the second approach by looking on a more system level. It is notable that all studies have used countries such as the US vs. Germany as their proxy to distinguish between the two systems. Acknowledging the dichotomous systems, I continue to examine the CG systems and its effect on CVC.

Corporate Governance and CVC

As mentioned before, the Anglo-Saxon system characterizes itself by having a strong market for capital control, including a well-developed stock market (Cernat, 2004). Black and Gilson (1998) find that due to these Anglo-Saxon system characteristics, in contrast to countries such as Germany and Japan known for a strong influence of the banks, VC firms relatively faster exit their PCs. Thereafter, VC has a greater vitality in stock market-centered systems.

The latter is supported by Milhaupt (19966_{). He extends the study of Black and Gilson (1998)}

by not only looking at liquidity as a moderator for VC, but also addressing other CG factors

that differ between US and Japan that affect the VC market7. Although both studies focus on

the VC market as a whole instead of specifically on CVCs, they show that differences in CG affect the (C)VC environment and activity.

On the other hand, Hege et al. (2009) find no evidence that the performance gap of CVCs can be attributed to the legal origin between Common Law and Civil Law countries, which is part of the heritage of the CG systems (Dainow, 1966). Although they only used

European countries to measure this difference8_{, it does arise uncertainty whether CG systems}

indeed have no effect on CVC performance. Furthermore, they find evidence that PCs exiting

6_{The study of Black and Gilson was published in 1998, but already known to Milhaupt in 1996.}

7_{Some of the factors in the study of Milhaupt (1996) are mentioned as ‘factors of the institutional environment’.}

However, CG is known to be rooted in the institutional environment. In congruence with this study, I mention them CG factors.

8_{The Common Law countries used by Hege et al. (2009) are the United Kingdom and Ireland. The Civil Law}

(16)

16 through IPOs have a higher performance (based on Internal Rate of Return - IRR) than PCs exiting through acquisitions. However, they find no evidence that PCs with an IPO exit perform better in US than Europe. Broadly spoken and taking the countries as a proxy for Anglo-Saxon vs European Continental, there is thus no difference in performance per se when examining the exit types between the Anglo-Saxon and European Continental systems. Nevertheless, the Anglo-Saxon system does offer a more stock-market centered system with more liquidity, making exits through IPO or Acquisitions easier (Black & Gilson, 1998). In a prior research by Hege et al. (2003), they find that US VCs perform better than European VCs by measuring performance in 2 ways; type of exit of PC and the Internal Rate of Return (IRR) of the project. However, no clear explanation is given for the difference in performance between US and Europe, leaving behind a conclusion based on assumptions and directions for further research to further clarify and examine the cause of this difference. Another recent study on a firm level (parent corporation) analysis was conducted by Anokhin et al. (2016), examining the role of specific governance factors including board, CEO and institutional ownership characteristics on a firm’s CVC activity. They find evidence that these factors indeed do influence a firm’s CVC activity, and emphasize the fact that the role of governance is important and that subsequent research should not ignore this group of factors.

As one could see, CG literature has tried to circumvent and examine the aspects of

(C)VCs. Indeed, the above studies show that several CG factors are affecting CVCs. However, not all scholars see these factors as completely different amongst different firms. DiMaggio and Powell (1983) argue that CG characteristics undergo the phenomenon of convergence due to isomorphism. These are normative, coercive and mimetic pressures that 'push' firms to adopt similar structures and strategies in the pursuit of legitimacy. On the other hand, Goergen, Manjon and Renneboog (2008) find no clear signs of convergence of the European Continental

(17)

17 system towards the Anglo-Saxon system. There are certain governance mechanisms that incorporated goals/aims associated with the Anglo-Saxon system, but the European Continental system have remained largely unaltered. In addition, IB literature does clearly show severe differences between both systems, making them dichotomous.

Subject to the corporation’s objective and goal, which are usually for CVCs from both CG systems to achieve financial positive returns and new insight and technologies, the CVCs make different strategic choices. One major strategic choice is the choice when to invest in a PC, ranging from the moment it started until the phase just before the exit of the PC. In the following section I continue to examine the investment stage.

Investment Stage

CVCs invest in PCs whereby the moment of investing is done at different moments of the PC’s lifetime. According to Plummer, which is used and referred to in the studies of Sahlman (1990) and Ruhnka and Young (1987), the investment in a PC can be distinguished in the following eight ascending stages; seed investments, startup, first-stage (early development), second stage (expansion), third stage (cash injection), fourth stage (growth and toward liquidity point), bridge stage (mezzanine investment) and eventually liquidity stage (cash-out or exit). A VC can also have a fund in the formative stage, financing seed and early stage, or a fund in the balanced stage, investing in all stages.

The choice for each stage comes with different motives. According to Elango et al. (1995), early stage investors search for higher potential returns than later stage investors. They claim that “Consistent with prior research, this study found that earlier stage investors are interested in unique, proprietary products with high growth potential, whereas late-stage investors demand a market-proven product” (Elango et al., 1995, p. 167). This is in line with Sahlman’s (1990) statement, saying that ‘VC organizations invest in early-stage business that

(18)

18 offer high potential but also high risk’. Ruhnka and Young (1987) find that the perceived risk of loss is very high for seed and start-up investments (66.2% and 53.0%), decreasing to 33.7%

for second stages and around 20.0% for third and exit stages9. Their findings are thus in line

with the aforementioned statement that the earlier an investment stage, the higher the risk. Although early stage investments might be riskier due to more factors that are uncertain, later stage investments also bring their own risks which are not directly present for early stage investments. An example is one of the hypotheses by Ruhnka and Young (1991) saying that stock-market liquidity, which can impair the ability of a successful IPO exit, can be seen as a major determinant of the perceived risk for later stage investments. Different circumstances and aspects influence the uncertainty and thus risk for each investment.

All stages thus have their own characteristics and levels of risk. Nevertheless, CVCs tend to specialize in certain investment stages rather than divide their investments, and thus risk, over several investment stages (Norton & Tenenbaum, 1993). Subsequently, CVCs investing in earlier stages (which are thus having higher risk than later stage investments) are less diversified across industries and firms, and thus not compensating for this relatively higher risk. So this would mean that the role of networks, specialization and information sharing (Bygrave, 1987) tends to have an important role in the reason to specialize in a certain amount of investment stages rather than to diversify the investments (and thus risk) over several stages. However, it is not clear yet how the investments in different stages contributes to the CVC’s performance. According to the model by Tyebjee and Bruno (1984) describing the VC investment decision process, the investment decision is based on the expected return and perceived risk as a result of the evaluation process. Herein, the investment stage can be

9_{The study by Ruhnka and Young (1987) is based on the response of 76 CEO’s or managing partners of US}

Venture Capital firms. Furthermore, one should take in consideration that the respondents were allowed to use their own interpretation of the different stages.

(19)

19 addressed to both expected return and risk. Namely, as described before, different stages come with different levels of risk, which in turn would be expected to have a difference in return (performance).

Corporate Governance and Investment Stage

CVCs investing decisions in different stages is thus subject to the level of risk accepted by the CVC. However, the investment stage is also to some extend subject to CG. For example, VC firms with higher levels of board equity (i.e. more board equity ownership compared to institutional equity ownership which is assumed to have a more risk-averse policy) have more support for risky investments and CVC activity (Anokhin et al., 2016). Herein, more risky investment are related to seed- or early-stage investments. Furthermore, the investment decision is also subject to its environment. Elango et al. (1995) find that VCs from Silicon Valley, known to be more technology-oriented, tend to invest more in early stages, whereas VCs from New York, known to be more finance-oriented, tend to invest more in later stages. In their research they tried to explain differences between regions and found that this could not solely be explained by investment stage. Even though they only measured within the US, and not going deeper into their explanation about why regions differ, their result of the relationship between later stage investors to be more finance-oriented shows that the role (and knowledge) of capital and finance is more important in later stage investments than earlier stage investments. This makes sense, since these stages are namely closer to the date of exit, where finance and capital plays a more important role to successfully exit the PC.

As mentioned before, a positive IPO climate and stock market liquidity, which are more present in the Anglo-Saxon system, are important determinants for PC exits (Black & Gilson, 1998). Acknowledging that these matters are more relevant for later stage investments, these GG factors do influence the investment decision. On the other hand, Allen and Song (2002)

(20)

20 state that “the allocation of investment across different stages of finance and different industries depends more on macroeconomic factors than on CG variables” (p. 3). However, they used only internal CG factors such as investor protection and law enforcement as a proxy for effectiveness of governance. Furthermore, they do acknowledge that fundamental economic conditions and the development of capital markets affect the stage allocation. This is in line with the aforementioned study of Black and Gilson (1998) and their approach of CG, where the development of capital markets including openness and degree of liquidity is also part of the CG systems.

As described before, VC performance is subject to the weigh off between risk and return. Different investment stages have different levels of risk and thus are assumed to have difference in performance. However, acknowledging the characteristics of the different CG systems, this might also have an interaction effect between investment stage and performance. Norton and Tenenbaum (1993) find no evidence for their hypothesis that VCs will diversify across different financing stages when they have the concern of the inability to time “cold” and “hot” IPO Markets. However, whether these CVCs want to diversify or specialize, it would make sense that if they have already chosen to invest in later stages, they are more willing to do this in an environment which is more “hot”. In other words, regardless of their motivation to invest in later stages, it is more likely that if they have chosen to do so, this would be in an environment which is more liquid and thus is related to the Anglo-Saxon system. Furthermore, Mayer, Schoors and Yafeh (2005) find that VC funds in Germany and the UK provide funding to PCs in all stages, with a slight bias towards later stages in the UK, which is in turn part of the Anglo-Saxon system. However, Japan, which is more related to the European system, is mainly investing in later stages. This could be seen as contradictory to the former mentioned view of the importance of a market-system to facilitate successful PC exits which is relatively

(21)

21 not severe present in Japan (Milhaupt, 1996). On the other hand, Japan is known for a large influence of institutions and banks. Their finding of Japanese VCs investing in more later stages is therefore in congruence with the study of Hellman, Lindsey and Puri (2008), showing that banks are subject to regulations and operate more risk averse, avoiding seed and early investments.

Altogether, different CG factors such as the role of banks, market liquidity and board equity ownership (instead of institutional equity ownership) thus influence the investment stage. Furthermore, these investments are done in different stages and thus have different levels of risk and expectations of performance. In the following paragraph I briefly elaborate on the terminology of ‘performance’.

Performance

As mentioned before, CVCs are mainly setup for 2 goals: to achieve financial positive returns and gain new insights and technology (Dushnitsky & Shaver, 2009). Due to these different objectives and goals, there are different ways to measure performance of CVCs. For example, Dushnitsky and Lenox (2005) measure patent data as a proxy for innovation which is related to a CVC’s goal of obtaining new insights and technology. However, in order to ensure continuity of the fund, it needs to be profitable, i.e. achieve positive financial returns. Ideally, I would measure the fund's performance directly by having access to data showing the returns of the CVCs. However, returns for individual funds are not widely available to researchers. According to Hochberg, Ljungqvist and Lu (2007), some researchers had access to this type of data but in aggregate form. Furthermore, they imply that the fund’s IRR is mainly based on the on the success of their PCs. Therefore, as the majority of studies do, I measure fund performance indirectly to look at the performance of their PCs. In congruence with Gompers (1995), Gompers and Lerner (2000), Hochberg et al. (2007) and other researchers, I examine “exit

(22)

22 rates” as a proxy of VC performance. This is the amount of PCs that, respectively, successfully exit through IPO, is sold by Merger and Acquisition (M&A), or is written off.

Some studies have extended this proxy of VC performance. For example on PC level, Hochberg et al. (2007) measure also whether the PC survives to obtain an additional round of funding as a proxy for performance. However, measuring the effect of different stages on performance already includes different rounds of funding and thus does not fit as a proxy for performance in my study. Furthermore, Hege et al. (2003) criticized the fact that valuation data has not been used as a proxy to measure VC performance. Therefore, they also measure VC performance by measuring the IRR of the project between the self-reported valuation date 0 and the last self-reported valuation of the project. I support their view that only measuring ‘exit rate’ as a proxy for VC performance might be scarce, and that additional valuation data gives a more comprehensive and intriguing data source to examine performance. However, this is only possible for PCs of which the valuation of the PC is known before the initial investment. Unfortunately, this type of data is not widely available for researchers.

Research Gap

As described throughout the literature review, several studies address the importance of CG and its effect on different aspects of a firm and more specific CVCs. For example, CVCs from the Anglo-Saxon system are more willing to exit their PC by IPO rather than Mergers and Acquisitions due to their open capital markets (Black & Gilson, 1998). Also, these - Anglo-Saxon known - open capital markets with high market liquidity increase the chance of a successful PC exit (Jeng & Wells, 2000). CVC’s strategic choices might thus be affected by, or subject to, specific environments formed by their CG factors and heritage. This, in turn, might affect the performance of CVCs. One of the strategic choices for a CVC to make is the investment stage: to determine in what stage to invest based on risk and expectations of

(23)

23 loss/profit. Several studies examine the difference between investment stages, how CVCs tend to make decisions for and within these stages and how the stages are related to CVC performance (Elango et al., 1995; Ruhnka & Young, 1987; Tyebjee & Bruno, 1984). Surprisingly, former research lacks the examination of what influence CG systems might have on a CVC’s investment stage. Furthermore, the aforementioned studies used questionnaires to identify decisions on the investment stage by examining the CVC’s expectations of risk of loss as a measurement for performance, rather than measuring the ‘actual’ performance by looking at IRR or success ratio of PC exits. Some studies have shown that US CVCs, as the largest representative of the Anglo-Saxon system, perform better than some European countries (Hege et al., 2003, 2009). However, rather than randomly comparing two countries or continents, or looking only independently at firm specific CG factors, this study approaches CG on a system level and specifically combines all countries from both the Anglo-Saxon system (US, CA &

UK) and the European Continental system (GE, NL, BE, FR, IT) 10.

Thus, this study measures the effect of CG systems on CVC performance. Furthermore, I examine the effect of CG systems on investment stage, and how this in turn has an effect on CVC performance. I use an indirect measure for CVC performance by looking at the success ratio of their PC exits. Altogether, I aim to investigate the effect of CG systems on CVC performance, and the role of investment stage in this relationship. The following research question is formed:

How do Corporate Governance systems influence Corporate Venture Capital performance? And how do investment stages play a role in this relationship?

10_{According to the majority of CG studies, the countries used in this study belong to the most salient ones from}

their systems (Koen & Mason, 2005). Therefore, I stated that this study combines ‘all’ countries from both systems, meaning all countries from the Anglo-Saxon system and the most salient ones of the European Continental system, that also actively participate in the CVC industry.

(24)

24

Hypothesis Development

In the following chapter, I build upon the literature review by critically reviewing and working towards clear directions of relationships, resulting in the hypotheses. The framework in figure 1 illustrates the relationships between CG systems, CVC performance and Investment Stage.

Figure 1. Conceptual Model

‘The presence of a strong strategic focus is critical to the success of venture capital funds’ (Gompers & Lerner, 2000, p. 46). However, the question remains to what extend this strong focus explains the CVC’s performance. Several studies examining CVC performance find positive effects of experience, portfolio diversity or syndication on CVC performance (Hochberg et al., 2007; Yang, Narayanan & Zahra, 2009). Herein, some factors have a stronger effect on the innovation part of the CVC’s performance rather than the financial returns. Besides examining the antecedents of CVC performance, several studies compare the differences between countries and CVCs in terms of performance. For example, Hege et al. (2003) find that CVCs from the US have a better performance than CVCs from Europe, where performance is both measured by the success ratio of their PCs and the IRR. According to Hege et al. (2003), these findings suggest either that US CVCs are more sophisticated, or are more successful due to more experience, knowledge and a network based environment. In a subsequent study, they address more firm specific characteristics from both countries that might

(25)

25 explain the difference in performance (Hege et al., 2009). However, the phenomenon of a difference in performance of CVCs can be attributed to more than the different characteristics of the CVC, the firm or the country.

As mentioned throughout the literature review, CG has proven to be an important determinant in business and more specifically CVC. CG exceeds the countries’ borders, making it possible for different countries to have the same CG characteristics, which leads to the two most known and bespoken dichotomous CG systems; the Anglo-Saxon system and the European Continental system. In contrast to the European Continental system, the Anglo-Saxon system is known to be more capital oriented with ‘short term’ equity finance and having a stock market centered system (Cernat, 2004; Moerland, 1995). This means that firms from the Anglo-Saxon system usually fulfill their financial needs through the stock exchange, rather than

through banks which is the case in the European Continental system11. As emphasized by Koen

and Mason (2005) in their book, this leads to a system based on liquid stock markets. In turn, higher market liquidity knowns to be an important determinant for a successful PC exit (Jeng & Wells, 2000), which in turn measures the CVC performance. Consistent with the findings of Black and Gilson (1998), US CVCs tend to be more successful due to this liquidity in active IPO markets. Regardless of the CVC’s skills, characteristics or origin, they all share the similar goal of (financial) performance: exiting the PC through IPO or acquisition (Dushnitsky & Shaver, 2009). For both options, market liquidity is vital in order to have no obstruction of an ongoing flow of successful exits. CVCs from the Anglo-Saxon system experience this advantage over CVCs from the European Continental system, increasing their probability of successfully exiting their PCs, and thus performance.

Mainly to diversify the risks, syndication tends to be commonplace within the VC

11_{Presumably firms from the Anglo-Saxon system do have huge loans and debt looking at the flat numbers, but it}

(26)

26 industry (Lerner, 1994). This applies for the investors investing in the PC, but also for the buyers who are acquiring the PC when it exits. However, within the European Continental system, firms tend to have long-term investment goals and are known for concentrated block-shareholding (Koen & Mason, 2005). Also, as mentioned before, financial needs are mainly fulfilled through banks. Therefore it might be hard to find parties that are willing to co-invest to acquire a PC. This applies for both finding ‘a few’ acquisition partners, or the many individuals and investors on the stock market which is thus illiquid. On the other hand, within the Anglo-Saxon system, it might be easier to find other parties who are willing to co-invest to acquire a PC. Due to the underlying shareholder theory of the Anglo-Saxon system, firms tend to be more focused on maximizing shareholder value (short-term) and diversification of their portfolios. This is again in line with a higher market liquidity and short-term equity finance, saying there are more parties with the desire and resources to invest, and also willing to diversify by acquiring different (parts of) PCs. For example, Schwienbacher (2005) find that European CVCs face less liquid markets, resulting a longer duration of the exit stage and making syndicated deals more difficult. Furthermore, both Hochberg et al. (2007) and Lerner (1994) find that a general increase in the number of IPOs on the NASDAQ Composite Index is associated with a significant increase in the probability that a PC will exit in the next quarter. Thus, altogether, the presence of higher market liquidity and short-term equity finance in the Anglo-Saxon system makes it easier for CVCs to successfully exit their PCs compared to CVCs from the European Continental system.

H1: CVCs from the Anglo-Saxon system have a better performance than CVCs from the European Continental system.

(27)

27 As described throughout the literature review, CVCs invest in PCs in different stages (Ruhnka & Young, 1987; Sahlman, 1990). For the choice of each stage are different motives, usually to do with the level of risk and profit potential associated with the stage. For example, early stage investments tend to be more risky due to more factors that are uncertain in contrast to later stage investing in PCs that are more ‘market-proven’ (Elango et al., 1995, p. 167). However, in addition to profit potential and risk, it seems to be appropriate that also CG systems have their influence on investment stage decisions, suggesting that CVCs from the Anglo-Saxon system tend to invest more in later stages in contrast to CVCs from the European Continental system.

First, as highlighted before, the aspects and characteristics of the Anglo-Saxon system being more capital oriented with ‘short term’ equity finance and having a stock market centered system with high market liquidity (Cernat, 2004; Moerland, 1995) are relevant for a successful PC exit (Black & Gilson, 1998; Jeng & Wells, 2000). However, this relevance seems more salient when a PC is in the stage before the exit, rather than when the PC is just founded (early stage). Of course, also when investing in early stages there is a desire to eventually successfully exit the PC, but the immediate urge of being in a stock-market centered and market liquid environment might be less than for later stage investments. At the seed or early stage, the PC still has several years to go before actually exiting wherein several developments can take place. On the other hand, when the PC is in the later stage, it is close to the exit and thus plays the probability to be sold through IPO or acquisition a larger role. This is supported by the study of Jeng and Wells (2002), showing that an active IPO market has no effect on early stage investments, but does have an effect on the attractiveness of later stage investments. Moreover, Allen and Song (2002) find that low GDP countries have less developed stock markets, which in turn also discourages later stage investments. Consistent with these findings is the study by

(28)

28 Elango et al. (1995), showing that later stage investing tend to take more place in finance oriented environments. Although the study of Elango et al. (1995) was conducted within the US comparing finance and non-financed oriented states, the Anglo-Saxon system clearly shows to be a more finance related system than the European Continental system. Furthermore, CVCs tend to specialize in certain investment stages rather than divide their investments, and thus risk, over several investment stages (Norton & Tenenbaum, 1993). If this is true, it seems obvious that CVCs in a finance oriented environment such as the Anglo-Saxon system are more willing to specialize themselves in later stage investments. As highlighted before, later stage investments are namely closer to the exit, and thus demand more relevance and expertise with regards to financial markets in order to successfully exit the PC. Lastly, stock market liquidity can be seen as a major determinant of the perceived risk for later stage investments (Ruhnka & Young, 1991). This means that CVCs from the European Continental system, known to be in an illiquid stock-market environment, might be more reluctant to invest in later stages due to a higher risk of not being able to successfully exit their PC. Altogether, CVCs from the Anglo-Saxon system might thus be more likely to invest in later stages than CVCs from the European Continental system.

In addition, investments in seed and early stages are expected to have higher returns when estimated by the CVCs themselves (Elango et al., 1995; Ruhnka & Young, 1987). This makes sense since they tend to be more risky, and thus have higher returns. However, these expectations of performance are measured as the total return of a finance project rather than in terms of success ratio of the project. Since seed and early investments are more risky and have a longer way to go before the exit, the number of PCs not reaching the final exit are likely to be more than later stage investments, which are already closer to the exit and thus are expected to have a higher success ratio. It is thus expected that, when measuring CVC performance as the

(29)

29 ratio of PCs that successfully exit through IPO or acquisition, investments in later stages have a better performance than seed, early and balanced stage investments.

Altogether, I thus expect that CVCs from the Anglo-Saxon system tend to invest more in later stages compared to CVCs from the European Continental system. Subsequently, I expect that investments in later stages have a better performance than seed, early and balanced stages when looking at the success ratio of PC exits. Together, this leads to the expectation that the relationship between CG systems and performance is mediated and thus (partly) explained

by the investment stage in later stage investments.

H2: Later stage investing is mediating the relationship between CVC’s CG systems and CVC performance.

(30)

30

Methodology

This study is one with an explanatory nature, where the emphasis is on studying a situation or a problem in order to explain the relationship between variables (Saunders, Lewis & Thornhill, 2012). It has a deductive approach, starting from a theoretical basis and building on the existing literature. To test the aforementioned hypothesis, this study has an experimental design. In the following section, I elaborate on the data collection, population and sample. Subsequently, I describe and define the different variables. Lastly, I explain the used methods and method of analysis.

Database

The data is collected from the database VentureXpert, which is accessible through Thomson One. VentureXpert contains worldwide VC related investment information, and is

officially endorsed by the National Venture Capital Association and the

PricewaterhouseCoopers MoneyTree Survey (Thomson Financial, 2016). In addition, both associations approve the research methodology used by VentureXpert for obtaining the data. Lastly, other acknowledged authors such as Anokhin et al. (2016), Cumming, Knill and Syvrud (2016), and Dushnitsky and Shaver (2009) have also used Thomson One VentureXpert in their CVC related studies. All data is extracted from this database and transferred to Excel and eventually SPSS for further analysis.

Data Collection and Sample

The population of this study is all CVCs from both CG systems. However, a list of all CVCs from both systems is not available, which makes it hard to conduct probability sampling. Furthermore, it is unlikely that all CVCs from both systems are in the database VentureXpert. Nevertheless, VentureXpert is peer-reviewed and by many acknowledged authors and universities seen as the most comprehensive and reliable database for financial related data

(31)

31 (Thomson Financial, 2016). Therefore, the results can be generalized to all CVCs from both CG systems, also known as external validity (Saunders, Lewis & Thornhill, 2012).

Initially, the data is gathered on investment level with a total of 192.000 investments done by CVCs from both CG systems, investing in PCs worldwide. Usually, venture capitalists want to cash-out their gains five to ten years after the initial investment (Tyebjee & Bruno, 1984). Therefore, the timespan of CVC investments is 10 years, between 2002 - 2011. However, only the investments in PCs that exited are incorporated in the analysis. This means that the exit data runs until 2016 (e.g. A PC can receive an investment in 2010 and exit through an IPO in 2015. This would be incorporated into the analysis and classified as a successful exit). Thus, all CVCs being investigated all have at least one PC in their portfolio that either successfully exited or is written off. Subsequently, within the total list of investments between 2002 - 2011, I eliminated all the duplicate cases to make sure that multiple investments by one CVC fund in the same PC (different rounds) eventually counts for one exit. Herein, several PC exits are thus counted multiple times, since there might be multiple CVC funds investing in the same PC (e.g. two CVC funds from the Anglo-Saxon system that both invested in the same PC between 2002 – 2011 that eventually is written off, is thus depicted as two right offs related to CVCs from the Anglo-Saxon system). This led to a sample of 2682 investments (N = 2682) conducted by a total of 568 different CVC Funds investing in 661 different PCs.

Description of Variables

In this section I comprehensively describe the variables and operationalization of all variables. Table 1 shows an overview of all the variables.

Dependent Variable

(32)

32 by the ratio of their PCs that successfully exit through either IPO or acquisition, compared to the unsuccessful PCs that are written off. As mentioned before, the majority of studies related to CVC performance use this proxy of performance (Gompers, 1995). There is thus no deeper analysis on PC level, as the only measure is whether the PC successfully exits, whereby the PC is subservient to a CVC Fund. Notable is that PCs that are still active, and thus not have exited yet, are not incorporated in the analysis. Unfortunately, the database does not mark all written off PCs as write offs, but leaves a blanc field indicating the data is missing. Previous researchers also experienced this problem. Tian (2011) solved this problem by saying that PCs that not have received an investment for more than 10 years, are considered as written off. However, CVCs invest in PCs with the goal to cash-out their investment within 5-10 years (Tyebjee and Bruno, 1984). Thereafter, according to Dr. Scott Shane, Professor at Case Western Reserve University,

the average time to exit for PCs from their initial investment date is 5.5 years (Shane, 201512).

Therefore, I classify a PC as a write off if it did not receive any investment subsequent investment for more than 5 years. The DV performance can thus be either a PC that is exited by IPO or Acquisition, which is perceived as successful, or is written off. Therefore, I create the dummy variable D1_SUCCESFULL, which is equal to one (1) if the PC of the CVC exited successfully, and 0 otherwise (which is in this case unsuccessful / written off).

Independent Variable

The Independent Variable (IV) is the Corporate Governance Systems, which is the CVC’s country from either the Anglo-Saxon system or the European Continental system. The US, CA and UK represent the Anglo-Saxon system, whereas GE, NL, BE, FR and IT represent the European Continental system. As mentioned before, rather than randomly comparing

12_{Dr. Shane used data from the National Venture Capital Association showing the time to exit in years from 1992}

(33)

33 countries, I select all the countries that, based on the literature, represent both systems. To make a clear distinction between countries from both systems, I define the dummy variable D1_ANGLO-SAXON_FUND which is equal to one (1) if the CVC is from either US, UK or CA, and 0 otherwise. In this case, these are CVCs from the European Continental model (GE, NL, BE, FR and IT).

Mediator

The mediator variable (M) is the investment stage, which is the CVC’s main focus of a specific stage to operate in, and thus mainly invest in PCs which are in that current stage. The investment stage data is thus related to the CVC fund, but takes into account that for example CVCs that have their main focus on seed stage investments, also might do little investments in early or later stage investments. Some studies have chosen to measure investment stage by operationalizing it in different investment rounds on PC level, being able to measure specific characteristics such as average amount invested, the duration or the level of syndication at the different rounds and thus stages (Hege et al., 2003, 2009; Sahlman, 1990). However, with this approach the researcher will still encounter a CVC investing in different stages. Furthermore, this study tries not to explain specific characteristics on PC level with regard to investment stages. Therefore, I use the simplified variable of fund stage, that relates to the CVC’s main focus of a stage to invest in.

Initially, I looked at the eight ascending investment stages used by Sahlman (1990) and Ruhnka and Young (1987; 1991). Due to the amount of perceived risk per stage shown by Ruhnka and Young (1987) and the characteristics of each stage, I combined the stages ending up with the following four stages: 1. seed stage, 2. early stage, 3. later stage, and 4. balanced stage. However, in congruence with the literature and my hypotheses, only later stage investments are expected to have a mediation effect. Therefore, to make a clear distinction

(34)

34 between the effect of the later stage against the early and balanced stage, I create the dummy variable D1_LATERSTAGE. This dummy is equal to one (1) if the CVC is mainly investing in later stages, and 0 otherwise. In this case these are the seed, early and balanced stage investments.

Control Variables

In addition to the IV and DV, the model also needs control variables in order to held other influences constant (Saunders, Lewis & Thornhill, 2012). The first control variable is experience, which is measured by 2 variables. The first variable is FUND_AGE, which is the age of the CVC at the moment when the related PC received its last investment, measured on a continuous scale. This last received investment could be from any fund, and is thus not specifically related to the fund in question. However, this last received investment date is used to measure the age of the fund in question, because the last received investment date gives the best and closest moment of the PC before it exits. More preferably would be to measure the age of the CVC at the moment of exit of its PC. However, as mentioned before, not all write offs in the dataset have an ‘exit date’. Therefore it is not possible to measure CVC age by looking at the exit date. In order to strengthen the control for experience, and due to the fact that I am not able to measure age most preferably by looking at the exit date, the second variable within the control of experience is NR_OF_DEALS_FUND, which is the total number of deals that the fund has been involved in, measured on a continuous scale. Obviously, an increasing number of deals results in an increase of experience. Furthermore, both Gompers (1996) and Barry et al. (1990) find significant effects for CVC age and performance with regards to the valuation of the PC exit IPO price. This is not directly related to the PC successful exit rate, but does show severe relevance of age with regards to CVC.

(35)

35 The second control variable is home country effect, which is the case when a CVC is investing in a PC from its own country. This might have an advantage since the CVC has no costs of doing business abroad, also known as the negative effects of liability of foreignness (LOF) (Zaheer, 1995). The LOF arises due to differences in for example cultural understanding or the knowledge of laws and regulations in the host country. These costs are borne both by the CVC and the PC, which in turn might affect the CVC performance. To control for this, I create the dummy variable HOME_COUNTRY, which takes the value of one (1) when the CVC and the PC are from the same country, and 0 if the countries between the two differ.

The third control variable is fund size, which is the estimated amount of capital invested by the firm who owns the CVC(s) (fund). These estimations are provided by VentureXpert, and appear to be more precise than the known amount of capital invested, since funds and firms hardly disclose all their investment information (Hochberg et al., 2007). More preferably would be to have the estimated equity invested by the CVC (fund) itself, but unfortunately this data is missing. Some other studies have also used the firm (parent corporation) as a proxy for the CVC size (Dushnitsky & Shaver, 2009). Elango et al. (1995) have shown that firm size is significantly affecting investment stage, whereas smaller VC firms are more interested in seed investments, and large VC firms in LBO’s (Later stage). Furthermore, fund size has also a significant effect on VC performance (Hege et al., 2009; Kaplan & Schoar, 2005). The variable FUND_SIZE is measured on a continuous scale.

The fourth control variable is time, which is the year when the PC received its last investment. As described before in the DV section, not all write offs have an ‘exit’ date. Therefore, I use last investment year instead of exit year. Instead of controlling for all years separately, I aggregated the years to control for the global crisis starting in 2008. This is a major phenomenon in the period of my research data (2002 - 2011) that affected markets worldwide.

(36)

36 Therefore, I created the dummy variable TIME, which takes the value of one (1) when the last investment year was during the financial crisis (2008 - 2010), and 0 for the years when there was no financial crisis (2002 - 2007, 2011).

The last control variable is industry, which represents the industry of the PC invested in by the CVC. Throughout IB literature industry is acknowledged as a determinant to have a huge influence. In my study, some industries might be riskier than others, influencing both investment stage (which also represents risk) and performance. Gompers (1996) also used industry as a control variable in his regression measuring the effect on IPO valuation (i.e. Performance). According to the United States Department Of Labor, there are ten major divisions subdivided in 83 major groups representing their own industry (SIC Division Structure, 2016). Like most studies, I have aggregated the industries, resulting in the following groups: 1. Agriculture, forestry, fishing, mining, construction and manufacturing, 2. Transportation, communications, electric, gas and sanitary services, 3. Wholesale and retail trade, 4. Finance, insurance and real estate and 5. Services (includes nine cases of public administration and unspecified). In order to properly run the regression, I cannot leave them categorical (Stockburger, 2016). Therefore, I create four industry dummy variables (e.g. D1_INDUSTRY2), where the dummy is one (1) if it represents the group of that particular

industry, and 0 otherwise. Group 1 herein represents the comparison group13_.

13_{I aggregated the ten major industries of the SIC (2016) into five groups to enhance the feasibility of this study.}

(37)

37

Table 1. Names, descriptions, operationalization and possible outcomes of variables.

Name Description Operationalization Outcome

Dependent Variable

CVC Performance Classifies whether the PC of a CVC successfully exits (IPO / Acquisition) or is written off and therefore unsuccessful.

VentureXpert provides exit type (IPO, Trade Sale or Written Off). All PCs without exit data and without subsequent investment for five years are labelled as written off, and thus unsuccessful.

1 = Successful PC exit.

0 = Unsuccessful PC exit

Control Variables

CVC Fund Age Classifies the age of the CVC at the date the PC received its last investments.

Extracts CVC founded year from year of last received investment corresponding PC.

Continuous

Nr. Of Deals CVC Fund Classifies the number of deals that the CVC fund has been involved in.

VentureXpert provides number of deals of CVC fund.

Continuous

Home Country Classifies whether a CVC invested in a PC from its own (home) country, or in a PC from a foreign country.

VentureXpert provides data on country information for all CVCs and PCs. I manually distinguished between CVCs and PCs from the same country and CVCs investing in PCs from foreign countries.

1 = Home Country PC

0 = Foreign Country PC

CVC Fund Size Classifies the size of the parent firm that owns the CVC fund based on the estimated amount of capital invested by this firm.

VentureXpert provides data on estimated amount of capital invested by firm. I manually listed the firms with the corresponding CVC funds.

Continuous

Time Classifies whether the PC of a CVC received its last investment during the years of crises or in times of non-crises.

VentureXpert provides data on last investment date. I manually converted them to years. I classify the years of crises 2008-2010, and non-crises 2002-2007, 2011.

1 = Crises

0 = Non-crises

Industry Classifies the industry from a PC the CVC invested in.

I manually distinguish 5 different industries based on SIC Code. One reference group and 4 dummies.

1 = Industry Group 1

0 = Other groups

Independent Variable

Corporate Governance Systems

Classifies whether a CVC is from the Anglo-Saxon system or from the European Continental system.

I use countries as a proxy for the CG systems. CVCs from US, UK and CA belong to the Anglo-Saxon system. CVCs from GE, NL, FR, BE and IT to the European Continental system. 1 = Anglo-Saxon system CVC 0 = European Continental system CVC Mediator Variable

Investment stage Classifies the current stage of a PC when a CVC conducts an investment.

VentureXpert provides the CVC’s main focus of investment stage. For my study, I distinguish between later stage investments and early or balanced stage investments.

1 = Later stage

0 = Early or Balanced stage

The role of corporate governance systems and investment stage on corporate venture capital performance