Venture capital investments, institutions and ethnic entrepreneurs in the information technology industry : a cross-country analysis

(1)

Venture capital investments, institutions and ethnic entrepreneurs

in the information technology industry: A cross-country analysis

University of Amsterdam

Master thesis

MSc Business Administration – Entrepreneurship and Innovation

Author: Maya Nankoe

Student Number: 11378352

Supervisor: dhr. B. Szatmari

Second Reader: dhr. A. S. Alexiev

Date: 23/06/2017

(2)

Statement of originality

This document is written by Maya Nankoe, 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)

Executive summary

Research on ethnic entrepreneurship explains that ethnic entrepreneurs face challenges in obtaining financial resources due to their ethnicity (Bengtsson & Hsu, 2015). Venture capital (VC) is an important source of external financing for entrepreneurs. Previous studies on ethnic entrepreneurship and venture capital investments have mostly been carried out in the US, therefore the results of these studies are limited in generalizability. Simultaneously, the venture capital market is growing rapidly, influencing the economy on a global scale. Therefore, this thesis explores the effect of ethnicity of the entrepreneur and of the founding team in obtaining finances by venture capitalists (VCs) through a cross-country analysis. The proposed relationships are empirically tested by using a panel data of 397 startups in the information technology industry for the period 2000 – 2010. This data is extracted from the Thomson One database. A total of 13 countries, consisting of Australia, Canada, China, Finland, France, Germany, Israel, Italy, Japan, Spain, Sweden, Switzerland and the United Kingdom, were examined.

Subsequently, to conduct this study GLS random-effects regressions were carried out in Stata. The results of this study show that native and ethnic entrepreneurs do not differ in VC investments received. Besides, there seems to be no difference for the founding team composition, that is if the team consists of native, mostly native or mostly ethnic

entrepreneurs, in VC investments received. As a result of analyzing in different countries, the role of formal and informal institutions were included to establish whether these country-level variables moderate the relationship between ethnicity of the entrepreneur or founding team and VC investments. There was no significant effect for informal (national culture) and formal (governance) institutions in VC investments for ethnic entrepreneurs or founding team.

Key words: Entrepreneurship; Ethnicity; Startups; Venture Capital; Institutions; Governance institutions; Hofstede; National Culture; Investment.

(4)

Introduction

The European Foundation for the Improvement of Living and Working Conditions (Eurofound) stated that “there appears to be a general awareness that ethnic entrepreneurs find it hard – even harder than native entrepreneurs – to obtain loans or other forms of financial capital” (Rath, 2011, p.50). Additionally, this report discusses measures found in practice to reduce the difficulties for ethnic entrepreneurs to obtain financial funding. Some measures that are mentioned include microfinancing, subsidies, guarantor schemes and loans/ grants provided by the city council.

The study of ethnic entrepreneurship has traditionally been framed as the idea that ethnic entrepreneurs become entrepreneurs out of necessity and that successful outcomes are rare (Aldrich & Waldinger, 1990). The ventures were identified as primarily targeted towards the entrepreneurs’ own ethnic group, which limits the opportunity for growth (Aldrich & Waldinger, 1990). However, recent research suggests that this point of view no longer holds credibility. As ethnic entrepreneurs capitalize on their talents in new markets, influencing the economy and society on a broad scale (Saxenian, 2002). Especially in the technology industry success stories of ethnic entrepreneurs are common. These type of startups usually need more financial resources than less knowledge-intensive startups targeted at a specific ethnic group (Aldrich & Waldinger, 1990; Baum & Silverman, 2004). Since startups often bear great risk investors are found to be hesitant to invest, for the reason that there is no history of success, no revenue coming into the venture, the soft assets are difficult to value, changing market conditions play a role and information asymmetry brings uncertainty for the investors (Fried & Hisrich, 1994; Gompers & Lerner, 2001). The access to investments and loans acts as a hurdle for ethnic entrepreneurial success, since external financial resources are usually necessary for startups to start producing, to develop technologies and to accelerate growth (Bengtsson & Hsu, 2015; Calopa et al., 2014). Nevertheless, VC is a form of external finance

(6)

that has grown rapidly and is known to finance young ventures.

In the US VCs were instrumental in financing the growth of firms such as Microsoft, Compaq and Oracle, which rapidly became dominant players in the technology sector (Jeng & Wells, 2000). The contributions of VC to the economy are not explicitly known, although there is a general belief that VC has a big influence in bringing innovations to the market, thereby creating growth, job opportunities and enhancing technological innovation (Jeng & Wells, 2000).

As stated before, ethnic entrepreneurs seem to have more difficulties in obtaining external financial resources, which places ethnic entrepreneurs in a disadvantaged position (Bengtsson & Hsu, 2015; Blanchflower, Levine & Zimmerman, 2003). Most research on this topic has been carried out in the US. The proposition that ethnic entrepreneurs and VC

investments examined in the US gives a good representation of this relationship, assumes that VCs in different countries would respond more or less the same to ethnic entrepreneurs. Most research on this topic is therefore limited in its generalizability and presents a non-exemplary image of VC investments. There is no clear empirical evidence with regard to ethnic

entrepreneurship and VC investments across different countries, thereby indicating a research gap in the literature on ethnic entrepreneurship and venture capital funding. Examining in multiple countries can give a better insight into this relation. Besides, different countries have unique formal and informal institutions in place that could play a role. This leads to the following research question:

How does the ethnicity of the entrepreneur influence venture capital investments and to what

extent do formal and informal institutions play a role in this relationship?

By combining ethnicity of the entrepreneur and VC investments across multiple countries in a database research, this study adds to existing academic research. The country-level variables

(7)

knowledge can be used to increase understanding and to further examine influences on VC funding for native and ethnic entrepreneurs.

The outcome of this study provides practical contributions for entrepreneurs, VCs, governments and other regulatory bodies. More knowledge on the obstacles for ethnic entrepreneurs helps raise awareness and benefits them, as measures could be taken to influence the funding process for ethnic entrepreneurs. Regulatory bodies could create policies aimed at more equal circumstances for ethnic entrepreneurs in external financing options, not limited to VC funding.

The remainder of this thesis first provides the theoretical foundation of this study. Subsequently, the data and methodology used are described, followed by the empirical results. Consequently, the implication of the results, contributions, limitations and future research directions are discussed. Lastly, a conclusion is formed.

(8)

Theoretical framework

Venture capital funding for ethnic entrepreneurs

The term ethnic entrepreneurship suggests that it is different from the regular form of entrepreneurship. Ethnicity linked to a group is defined by commonalities in background or culture, or by the outgroup distinguishing a group because of certain characteristics in origin or culture (Yinger, 1985). Aldrich and Waldinger (1990) emphasize the sociocultural aspect of ethnicity, the social structures which tie members of ethnic groups to each other. As ethnic entrepreneurs are perceived as a group, individuals that are assigned to these groups can be aware of their group membership.

An alternative term for ethnic entrepreneurship is immigrant entrepreneurship, which is limited to immigrants. This definition excludes ethnic groups who have been living in countries for several decades, such as Chinese people in Europe. However, the term ethnic entrepreneurship does not exclude immigrant or ethnic groups (Aldrich & Waldinger, 1990; Yinger, 1985).

While ethnic entrepreneurs progressively contribute to innovation and job creation, acquiring financial resources is generally more difficult for them in comparison to their native peers (Blanchflower et al., 2003; Rath, 2011). Ethnic minority immigrants form a large

subgroup of technology and engineering entrepreneurs in the US (Saxenian, 2002). The traditional viewpoint that ethnic entrepreneurs have low knowledge-intensive ventures is being challenged by these numerous high-growth technology startups founded by ethnic entrepreneurs (Saxenian, 2002). The information technology industry is a popular industry for VC investments, as VCsspread over this industry trying to capture fast-growing and

continuously redefining ventures fueled by emerging technologies (Gaba & Meyer, 2008). First, we look at how VCs assess new ventures and then we explore how this relates to ethnic

(9)

entrepreneurs in the information technology industry.

As mentioned before, VC is a popular form of external financing. VCs overcome difficulties in assessing new ventures by developing specific capabilities to measure risk levels of a new venture. The selection process consists of different stages, in which screening takes place (Fried & Hisrich, 1994). Due to standardized decision making processes in these stages it is broadly accepted that venture capitalists are experts in making funding decisions based on objective information, as the business plan and other specified requirements (Fried & Hisrich, 1994). However, as VCs have to deal with an abundance of funding proposals and a costly screening process, the screening involves fast decision making next to high

uncertainty, which results in decision making based on inadequate and incomplete

information (Kirsch, Goldfarb & Gera, 2009). Thus although the VCs are seen as objective, under conditions of high uncertainty, information asymmetry and time constraints subjective reasoning (i.e. heuristics) can influence the decision making process, as a mechanism to evaluate proposals swiftly (Kirsch et al., 2009).

Given many uncertainties in new ventures, VCs look for signs of ventures’ future performance and quality. New ventures must mobilize a diverse range of resources and relationships in order to be successful. Prior research indicates that biotechnology startups require access to social capital, intellectual capital and human capital, which form key signals of startup potential (Baum et al., 2000; Baum & Silverman, 2004). Research in other dynamic industries shows that the likelihood of VC investments is higher for ventures with patents, for entrepreneurs who have successful experiences in startup companies, highly educated

entrepreneurs and entrepreneurs who have a broad social network (Bengtsson & Hsu, 2015; Hsu & Ziedonis, 2007). This indicates that the aforementioned key signals also form

legitimate signals in the information technology industry.

(10)

entrepreneurs has been recognized in multiple studies (Aldrich & Waldinger, 1990; Bengtsson & Hsu, 2015). Social networks and strong ties provide access to resource mobilization and knowledge exploration, which also serve as external signals by suggesting that the startup has a capable network necessary for success. Additionally, alliances (e.g. business partners) signal the quality of a startup as alliances suggests that the startup has earned a positive assessment, especially an experienced alliance signals high quality of the startup (Baum & Silverman, 2004).

Ethnic entrepreneurs rely heavily on their social network. Family members and friends often form an important source of funding (Zimmer & Aldrich, 1987). Moreover, ethnic entrepreneurs create strong networks and partnerships predominantly with co-ethnics (Aldrich & Waldinger, 1990; Zhang, Wong & Ho, 2016). Very recent studies find that co-ethnicity between entrepreneurs and VCs gives more benefits to the entrepreneur in comparison to non-co-ethnic relationships and facilitates the funding process through increased trust. However, it results in worse outcomes in the performance of the startup (Bengtsson & Hsu, 2015; Zhang et al., 2016). Ethnic enclaves in networks and financing limits entrepreneurs to gain access to diversified resources, which are critical for entrepreneurial success. The opposite perspective is that co-ethnic networks form a support structure for entrepreneurs, through solidarity and trust mechanisms. However, reliance on ethnic networks does result in lower social status in comparison to mainstream networks (Zhang et al., 2016).

Secondly, the intellectual capital can help startups to acquire resources, by signaling good knowledge of industry, innovative capabilities, intellectual properties and customer understanding (Baum & Silverman, 2004). A possible disadvantage that ethnic entrepreneurs face, especially first generation ethnic entrepreneurs, is the lack of knowledge in processes and lack of host country language skills, which decreases the probability of finding relevant information during the startup phase and constraints communication with (potential)

(11)

customers and alliances (Clark & Drinkwater, 2000). Furthermore, limited knowledge in the operating market and industry can influence signals of intellectual property (e.g. patents) that are not given priority by the entrepreneur, but do have a strong relation to the startups’

potential and therefore to the likelihood of VC investments (Baum & Silverman, 2004; Zhang et al., 2016).

Lastly, VCs commonly overestimate the startups’ human capital when making

investment decisions (Baum & Silverman, 2004; Vinig & de Haan, 2001). Vinig and de Haan (2001) found that VCs in the Netherlands and the US assign more importance to attributes of the entrepreneur (i.e. track record and leadership), compared to other screening criteria as the idea, product and market. The entrepreneurs’ prior experience, the prominence of the prior employer and diverse skills are signals that VCs look at (Baum & Silverman, 2004). There are different levels in human capital, as there can be highly educated ethnic entrepreneurs with prior business experience, but there can also be low educated ethnic entrepreneurs with limited experience present. It is however clear that who the entrepreneur is plays a role in VCs investments (Kirsch et al., 2009). Therefore, discrimination in the host country can also play a role in VC investments for ethnic entrepreneurs. As the ethnic entrepreneurs can be attributed a lower status than their native peers. Previous research in the US finds significant differences in credit access in certain demographic groups, which can likely be explained by discrimination (Blanchﬂower et al., 2003; Cavalluzzo & Cavalluzzo, 1998). Yet, more recent research in Germany and the UK finds that ethnic entrepreneurs are not more likely to be rejected for credit or loans or to receive smaller loans than requested (Bruder, Neuberger & Räthke-Döppner, 2011; Fraser, 2004). Building upon studies that find a disadvantaged position for ethnic entrepreneurs (Bengtsson & Hsu, 2015; Blanchflower et al., 2003), the proposition that ethnic entrepreneurs receive lower VC investments in comparison to native entrepreneurs is formed.

(12)

Hypothesis 1 (H1): Ethnic entrepreneurs receive lower VC investments in comparison to

native entrepreneurs.

Institutions

Startup and VC activity is related to economic growth. According to North (1991) institutions and limitations of the economy defines the possibilities in economic activity and growth. Institutions are an important determinant of venture capital activity, because

“institutions provide the incentive structure of an economy; as that structure evolves, it shapes the direction of economic change towards growth, stagnation or decline” (North, 1991, p.97). North distinguishes institutions into two categories; formal and informal institutions.

According to North (1991) formal institutions present the political environment of a country, such as legal systems and regulations. Whereas informal institutions present beliefs and assumptions, that is the cultural setting, which complements the formal institutions.

Although the concept of culture has existed for thousands of years, recently research is exploring the influence of culture in order to gain insights in the globalized business world. Studies examine differences in cultural dimensions to explain relationships between host country and foreign country direct investments, acquisition behavior, cost transactions and other variables (Bhardwaj, Dietz & Beamish, 2007; Brouthers & Brouthers, 2000)

Li and Zahra (2012) are one of the first to provide empirical evidence for the Hofstede’s cultural dimensions linked to venture capital activity and formal institutions. Li and Zahra (2012) find that both formal as informal institutions are important explanatory variables for VC activity. Countries with a low score on the dimensions collectivism and uncertainty avoidance were positively associated with formal institutions, resulting in a positive relationship for VC activity as these countries had a higher number of VCs and VC investments in the country compared to countries with a high score on the dimensions

(13)

collectivism and uncertainty avoidance. Even though research on the impact of national culture on venture capital investments for ethnic entrepreneurs is rather limited, studies on national culture and VC investments build evidence for this relationship. Therefore, in the remainder of this thesis the informal institutions are limited to national culture.

National culture

National culture presents one of the core elements of differences between countries. Hofstede (1994, p.1) stated that culture is: “the collective programming of the mind which distinguishes the members of one category of people from another”. The category in

Hofstede’s model (1994) is specified to a nation, but culture can also be specified to a region, ethnic group, gender, a social class and so on. The main tie in this category is the system of collectively held values. These values and therefore the culture stay relatively stable over time (Hofstede, 1994). Hofstede’s (2001) original framework distinguishes four cultural

dimensions: uncertainty avoidance, individualism, power distance and masculinity. The role of uncertainty avoidance, individualism and power distance in VC investments for ethnic entrepreneurs are further examined in this study.

Uncertainty avoidance

Uncertainty avoidance reflects to “what extent a culture programs its members to feel either uncomfortable or comfortable in unstructured situations. Unstructured situations are new, unknown and different situations from the usual. Uncertainty avoiding cultures try to minimize the possibility of such situations” (Hofstede, 1994, p.5). Societies vary in their level of uncertainty tolerance, thus in the degree in which a society is ready to accept ambiguous situations and risk. A country with a high score on the uncertainty avoidance index (UAI) is characterized by low tolerance for risk and uncertain situations. As a consequence individuals

(14)

feel the need for hierarchy, rules and structure (Erramilli, 1996).

Several studies show that a low score on uncertainty avoidance is related to high levels of innovation rates within a country (Shane, 1993; Sundqvist, Frank & Puumalainen, 2005). Countries that are uncertainty avoiding often accept innovations through imitation. Imitation of innovation decreases the risk-taking that comes with innovation (Sundqvist et al., 2005). In the VC literature Li and Zahra (2012) provide evidence that countries with a low score on uncertainty avoidance have a higher number of VCs and a higher number of VC investments, through a positive relation with formal institutions. The acceptance of uncertainty forms a necessary condition to create an environment in which innovations and startups can grow, because these activities require a tolerance for risk and change (Li & Zahra, 2012; Shane, 1993). The perceptions of risk are subjective to individuals.

In countries that score high on uncertainty avoidance additional barriers for ethnic entrepreneurs could be present, because of the feeling “what is different, is dangerous” which can be associated with ethnocentrism (Hofstede, 1999). The ethnic entrepreneur is assigned a higher level of risk than their native peers. In high uncertainty avoiding countries people can experience a discomfort in interpersonal relationships with people from other cultures, which can lead to feelings of frustration and hatred towards these people (Mangundjaya & Lecturer, 2011). Thus, putting ethnic entrepreneurs at a disadvantage in their social capital, as creating social networks and building strong ties for resource mobilization is harder if people from the society are not willing to create ties. This in turn influences the ability to signal intellectual capital. Besides, the human capital is already observed as more negatively, due to higher risk perceptions. Therefore the proposition has been formed that ethnic entrepreneurs are more disadvantaged in high uncertainty avoiding countries than in low uncertainty avoiding countries, which influences the amount of VC investments received.

(15)

Hypothesis 2a (H2a): Ethnic entrepreneurs receive lower VC investments in comparison to

native entrepreneurs; this effect is moderated by uncertainty avoidance, the effect is stronger for countries with high uncertainty avoidance scores.

Individualism

Hofstede characterizes individualism as “societies in which the ties between individuals are loose: everyone is expected to look after himself or herself and his or her immediate family” (Hofstede, 1994, p.3). In these societies the priority is given to be independent and autonomous, which includes the desire to be independent from any group relations. On the contrary in collectivistic societies ties are strong and individuals are

integrated into groups. These groups have a high cohesion and can prioritize members of the in-group, as a form of loyalty towards the group (Hofstede, 2001).

Several studies show that a high score on individualism is related to high levels of innovation rates within a country (Shane, 1992; Shane, 1993). Indicating that individualism is important, because of the connection with autonomy, independence, and freedom. However, collectivistic countries are also becoming more innovative (Shane, 1993). Demonstrating that the VC environment can be well-established in countries at either end of the dimension. In countries with high collectivism individuals have the need of saving face and confirm to group agreements to maintain their social position. Co-ethnic ties in collectivistic countries tend to be strong and reciprocity and mutual obligations are highly valued (Bochner & Hesketh, 1994). In contrast to this, in individualistic countries the personal goals and

rewards have a central role and decisions are made autonomously. Loyalty is less emphasized, which gives entrepreneurs more opportunities to build a diverse social network and to gather necessary information from diverse resources (Shane, 1992). Besides, individualistic

countries are more willing to look at the individual themselves. Which can imply that ethnic entrepreneurs do not receive a lower status, as individuals are not strongly related to groups

(16)

(e.g. outgroup) in individualistic countries. Whereas in collectivistic countries the dominant in- and outgroup view places ethnic entrepreneurs at a disadvantage, in high individualistic countries these in- and outgroup views are not given as much importance (Hofstede, 2001). From this point of view, in high individualistic countries an ethnic entrepreneur has a higher probability to receive similar amounts of VC investments as native entrepreneurs.

Hypothesis 2b (H2b): Ethnic entrepreneurs receive lower VC investments in comparison to

native entrepreneurs; this effect is moderated by individualism, the effect is weaker for countries with high individualism scores.

Power distance

Power distance refers to the extent to which members of institutions and organizations accept and expect power inequality in relationships (Hofstede, 1994). This indicates that in countries with a high power distance score there is a strict hierarchy of power, whereby the upper levels rule over the lower levels. The unequal distribution of power is accepted and followed as a matter of procedure. The opposite applies for countries with a low power distance score, where equality plays a bigger role. In these countries there is a big strive for equal distribution of power.

Erramilli (1996) found that countries with a high score on power distance have

decision makers who are more inclined towards centralized management styles. Furthermore, Hofstede (1984) found that decision makers in high power distant countries use power, wealth and prestige to keep the distribution of power in their control. As a consequence social

inequality is created and reinforced. Innovations are not viewed as favorable, because innovations often cause changes that can result in a redistribution of power, which is not desirable for individuals who hold positions of power (Shane, 1993). Additionally, individuals from high power distant countries are high task-oriented and low people-oriented. The

(17)

structure within organizations is to give out tasks, without consulting employees (Bochner & Hesketh, 1994). Decision makers from low power distant countries lean towards decentralized management styles, showing more willingness to share control with others. Low scores on power distance is important for innovations and startups, as this indicates a tolerance for change in the social hierarchy and distribution of power in the innovation process (Shane, 1993).

High power distant societies are inclined to manipulatively make use of power and lack of equal opportunities for minorities, like gender and race (Carl, Gupta & Javidan, 2004). The distribution of prestige, social status, wealth and classes are retained in these societies. Hence, the following proposition:

Hypothesis 2c (H2c): Ethnic entrepreneurs receive lower VC investments in comparison to

native entrepreneurs; this effect is moderated by power distance, the effect is stronger for countries with high power distance scores.

Formal institutions

Formal institutions are the rules that facilitate economic interaction, these rules include the legal and regulatory framework (North, 1991). According to North (1991) businesses can grow if government policies adopt proper frameworks for political and economic efficiency. The effect of formal institutions in economic efficiency differs, as formal institutions can create economic certainty, through clear legislation which leads to reasonable expectations for economic actors. On the other hand, formal institutions can create barriers for economic activity, an example of such a barrier is inefficient bureaucracy. Thus, formal institutions can create limits, but can also present opportunities (Globerman & Shapiro, 2003). These broad political and legal institutions that influence the economy of a country, is also referred to as the governance infrastructure (Globerman & Shapiro, 2003). A positive and approving

(18)

governance infrastructure benefits conditions for economic growth. According to Globerman and Shapiro (2003, p.3) “a beneficial governance infrastructure might therefore include: an effective, impartial and transparent legal system that protects property and individual rights; public institutions that are stable, credible and honest; and government policies that favour free and open markets.”

As governance institutions build the framework wherein businesses operate, they also influence the position of startups in the economy. Studies about governance institutions and entrepreneurial activity find that economic policies can influence whether individuals engage in entrepreneurship, which in turn influences the demand for VC investments (Audretsch, 2002; Baumol, 1996; Bruton, Ahlstrom & Puky, 2009). Governance institutions could install strict establishment regulations and exit regulations of young ventures, as a way to monitor the amount and quality of young ventures. If the establishment regulations to start a business are very demanding, time consuming, costly and complex, starting a business becomes less attractive.

Moreover, in countries with weak governance institutions legislative procedures are time consuming and are frequently problematic. Weak governance institutions lead to reliance on relationships with mostly powerful individuals (e.g. elite groups) or reliance on economic exchanges (e.g. bribes) to influence processes (McMullen, Bagby & Palich, 2008).

Entrepreneurs and VCs need to establish personal contacts and a strong network in countries with low governance, as these relationships form the basis to provide access to information and help to deal with regulations (Bruton et al., 2009). Interpersonal relations play a more important role in countries with weak governance institutions, since ties with others are an alternative and often necessary source to overcome obstacles (Fried & Hisrich, 1994; McMullen et al., 2008). As ethnic entrepreneurs often lack diverse ties and usually create a network with their co-ethnics, ties to overcome regulations and restrictions are often weak or

(19)

not present. However, in countries with strong governance institutions processes are well organized, which means less reliance on social networks and ties for ethnic entrepreneurs. Therefore, the final proposition was formed.

Hypothesis 3 (H3): Ethnic entrepreneurs receive lower VC investments in comparison to

native entrepreneurs; this effect is moderated by governance infrastructure, the effect is weaker for countries with a high scores on governance infrastructure.

Conceptual framework

The framework in figure 1 illustrates the relationships described in the theoretical framework, between ethnic entrepreneurs, VC investments, national culture and governance institutions.

Figure 1. Conceptual framework

-

+

-Ethnicity entrepreneur (H1) VC investments

H2a: Uncertainty avoidance H2c: Power distance

H2b: Individualism H3: Governance institutions

(20)

Table 1. Overview of national culture dimensions and governance institutions with the expected effects for ethnic entrepreneurs.

Dimension Low scores on

dimension High scores on dimension Expected outcome for ethnic entrepreneurs for

high scores on the dimensions

Uncertainty avoidance index (UAI)

- Tolerance for risk - Risk aversion Negative outcome H2a Individualism index (IDV) - Collectivism, individuals belong to a group - Existing networks influence decisions - Focus on being independent and autonomous - In decision making distant from groups

Positive outcome H2b

Power distance index (PDI) - Decentralized structures - Collaboration between levels - Centralized structures

- Higher levels exert power over lower levels Negative outcome H2c Governance institutions (WGI) - Weak governance infrastructure; time consuming and frequently problematic processes - Strong governance infrastructure, well organized and standardized processes Positive outcome H3

(21)

Method

Database

For this study, multiple data-sources were used for data collection. The firm- and industry-level data were gathered from the Thomson One database, which includes the VentureXpert database. The names of founding members of startups are commonly not mentioned in Thomson One. Therefore the names of startup founding members were collected from company reports, corporate websites, newspaper sources (LexisNexis Academic) and personal pages (as LinkedIn). The national culture dimensions were taken from Hofstede’s dimensions (2001). The governance institutions came from the World Governance Indicators developed by the World Bank (Kaufmann, Kraay & Mastruzzi, 2009).

Data collection and sample

The sample consists of 397 information technology startups backed by venture capitalists over the period of 2000-2010 in 13 countries (i.e. Australia, Canada, China, Finland, France, Germany, Israel, Italy, Japan, Spain, Sweden, Switzerland and the United Kingdom).

To select the sample, data was gathered from the Thomson One database. The

information technology industry was defined by using the Venture Economics Industry Code (VEIC) 1000. The search was restricted to ‘Venture capital deals’, specified to firms that are “Currently private equity/ venture capital backed” or “Formerly private equity/ venture capital backed” to verify that the sample only includes firms who received at least one private equity/ venture capital investment. To limit the dataset to firms that were in fact startups, the search was restricted to firms whose first funding round was identified as ‘Venture capital seed/ start-up stage’ (Mann & Sager, 2007). Excluding firms whose first funding round was identified as late stage, balanced stage or identified as backed by other forms of funding. The

(22)

round of investment was restricted from the first till the third funding round. This means that the investment could take place in either funding round 1, 2 or 3, taking into account that ventures overall proceed to other stages of funding (i.e. late stage or balanced stage) after 3 or more funding rounds.

The time span of 2000 to 2010 was chosen to have a time recent enough that can be representative of current conditions. However, not too recent as the database can be more incomplete for the last few years. As the study is focused on information technology startups, the first year of data collection was chosen to be 2000. The internet bubble lasted till the late 1990s (Li & Zahra, 2012). This study focuses on post “bubble” investments, as during the bubble internet sectors and information technology sectors grew rapidly. VC has also become an increasingly accepted form of investment in different parts of the world since the late 1990s (Bengtsson & Hsu, 2015; Li & Zahra, 2012).

The selection criteria resulted in a population of 2291 unique firms from a total of 45 countries. The US was dropped from the sample, as other countries had similar scores on the moderating variables and most research on ethnic entrepreneurship has been carried out in the US, which already provides a good perspective on the conditions for ethnic entrepreneurial financing in the US. Additionally, countries with less than 5 firms were excluded from the dataset, resulting in 623 firms. Due to equity amount data unavailability 52 of the 623 cases were dropped. Moreover, after analyzing the data 8 countries were dropped. As the sample of 571 firms in 21 countries, included countries without ethnic entrepreneurs present in the sample (appendix table 1). The moderating variables in this study are country-level variables, therefore countries without ethnic entrepreneurs were excluded. Resulting in a sample of 404 firms in 13 countries. After removing the outliers from the dataset, the final sample contains

(23)

founded by an ethnic entrepreneur (13.6%).

Variables and measures

Dependent variables

Venture capital investments. The venture capital investments are measured by the amount of

funding a startup received during the funding round (Bengtsson & Hsu, 2015). This is an absolute measure of the funding amount received, not in proportion to the funding request.

Pre-money valuation. The pre-money valuation equals the amount of valuation at the funding

round minus the funding amount received (Gompers, 1995; Hsu, 2004; Miloud, Aspelund & Cabrol, 2012). This is an inverse way to measure the potential of a startup. Startups often get higher evaluations than funding, therefore the outcome of this equation is mostly positive.

Independent and moderator variables

Ethnic entrepreneurs. In the context of this thesis the ethnic entrepreneurs present the

non-native entrepreneurs, as non-native entrepreneurs can label non-non-native entrepreneurs as the outgroup in their country and the ethnic entrepreneurs can be aware of their group membership to the outgroup. Ethnicity of the entrepreneur was measured as a dummy-variable, taking the value 0 if the entrepreneurs, also known as founders, were native and taking the value 1 if at least one member of the founding team was from an ethnic

background.

The ethnicity was determined based on the names of the founding members, following the approaches of Kerr (2008) and Bengtsson and Hsu (2015). The names are identified by ethnicity through the ethnic-name database by Kerr (2008). For countries not covered by the database or in unsure cases company communications and personal pages were used to identify the ethnicity. If the ethnicity is not able to be determined based on previous

(24)

mentioned sources, the founders are coded as native as this is the most common scenario (Bengtsson & Hsu, 2015).

To illustrate this methodology, suppose the country of analysis is China, in this case the Chinese surnames are coded as native (value = 0). Indian surnames would be coded as non-native in this country (value = 1). The limitations that come with this form of analysis are explained in the discussion section.

Ethnicity founding team. As entrepreneurs often start ventures in teams. The team

composition, that is ethnicity of the founding team, was added as an independent variable. The dummy ‘0’ was assigned to teams with no ethnic entrepreneurs. The dummy ‘1’ was assigned to teams with ethnic entrepreneurs, but in which the majority of founders were native. The dummy ‘2’ presents teams of which the majority has an ethnic background (i.e. non-native background).

Moderator variables:

Uncertainty avoidance, Individualism and Power distance. These dimensions were taken

from the Hofstede’s dimensions, measured on a 0-100 point scale. The most recent indices were used, as culture is considered to be relatively stable over time. The cultural dimensions are therefore treated as time-invariant (Li & Zahra, 2012). The indices for the cultural dimensions were retrieved from http://www.geert-hofstede.com/hofstede_dimensions.php1

Formal institutions. The formal institutions in this research present the governance

institutions of a country. The governance institutions were measured by using the Worldwide Governance Indicators (WGI) index (Kaufmann et al., 2009). This index distinguishes governance quality through six dimensions: voice and accountability, political stability and absence of violence/terrorism, government effectiveness, regulatory quality, rule of law and

(25)

lastly control of corruption. The dimensions are measured on a score between -2.5 and 2.5.

Li and Zahra (2012) have pointed out that these six dimensions are seen as a reliable measure for governance institutions, because it does not rely on a few data observations which ensures low error in this measure. Li and Zahra (2012) have used this measure in examining differences in venture capital activity across countries, making it an appropriate measure for the context of this study.

As in previous research (Kaufmann et al., 2009; Li & Zahra, 2012) the six dimensions are averaged to measure one combined value over the period of 2000-2010. A negative value indicates that countries have poor governance institutions in place, whereas a positive score indicates the opposite.

Control variables

As this research focuses on startups and their investments (firm-level), so were most of the control variables included in the research. The firm age at funding was taken as a control, as older firms have more experience and have a form of proven skill and ability, linked to their longer survival (Fried & Hisrich, 1994). The firm age at funding is the age of the venture at the funding round measured in months. Additionally, the firm size was taken as a control variable, since bigger firms can receive benefits due to their size. Firm size is measured through the total assets of the venture (Anderson & Reeb, 2003). As all investment stages have their own risk which are connected to the firm’s development stage, the firm’s

development stage was controlled for. The firm’s development stage was coded as a dummy variable, coded ‘0’ for ‘seed stage’ and ‘1’ for ‘early stage’ investments (Gompers, 1995; Jeng & Wells, 2000). Additionally, the number of investors during the investment stage can influence the VC investment. Furthermore, industry dummies based on the SIC-code were created to control for the industry. The industries were chosen based on the amount of firms belonging to a certain industry, resulting in 4 categories. If the firm’s main business was

(26)

Prepackaged software (SIC = 7372) the industry was coded as dummy 1. Dummy 2 was used for Semiconductors and Related Devices (SIC = 3674). Dummy 3 was used for Information Retrieval Services (SIC = 7375). Dummy 0 was applied to all the remaining SIC codes and firms of which the industry was unknown.

Method of analysis

Given the fact that the dependent variable in this panel dataset is a continuous variable and the independent variables are categorical, a regression model can be used. The specific regression model depends on the assumptions being met.

As the sample forms an unbalanced panel dataset with multiple countries that differ in characteristics a generalized least squares (GLS) regression model was proposed to test the hypothesis. The GLS random-effects model controls for unobservable country characteristics and allows diverse variances among countries (Li & Zahra, 2012). Additionally,

time-invariant characteristics as the national culture dimensions can be included in this model. In this model the control variables were measured first, followed by the independent variable, and finally the interaction effects, leading to the following regression model:

Venture capital investment = ɑ + β1 Firm age + β2 Firm development stage + β3 Firm industry + β4 Number of investors + β5 Ethnicity + β6 (Ethnicity*UAV) + β7

(Ethnicity*IDV) + β8 (Ethnicity*PDI) + β9 (Ethnicity*WGI) + ε

Ethnicity = ‘Ethnicity of the entrepreneur’ or ‘Ethnicity of the founding team’ UAV = Uncertainty avoidance

IDV = Individualism

PDI = Power distance

(27)

Results

First, the data observations are presented. Second, the assumptions regarding GLS regression will be discussed. Followed by the descriptive statistics. Finally, the results regarding the hypotheses will be discussed.

Data observations

After analyzing the panel data sample, the variable pre-money valuation was excluded as a dependent variable as the valuation was only available for 20 firms in the dataset. The control variable total assets was also excluded as this information was present for 18 firms in the dataset. See Appendix 2 for an overview of the variables employed to test the hypotheses (also shown in the correlation matrix).

Assumptions

In this section the assumptions regarding GLS regressions will be discussed. If one of the assumptions is being violated, there will be adjustments made in order to interpret the results in a correct way. The assumptions are: normal distribution, normality of residuals, homoscedasticity and the absence of multicollinearity.

Normal distribution

When checking the dependent variable venture capital investment for normality, it turned out that it did not meet the assumption. Consequently, venture capital investment was log-transformed to improve the normal distribution. However, after the log transformation venture capital investment was heavily skewed, with high levels on the right side of the spectrum (see appendix 3). The dependent variable VC investment (log-transformed) had a substantial negative skewness of 1.189 and a positive kurtosis of 3.913. The dependent

(28)

variable VC investment (before the log transformation) had a positive skewness of 0.139 and a positive kurtosis of 1.857. The variable VC investment had a significantly different

distribution than normal, as shown in the Kolmogorov-Smirnov test. However, the skewness and kurtosis are within the boundaries of -2 and 2 for an acceptable normal distribution to run regressions with and are thus not expected to influence the results severely (George &

Mallery, 2010).

Firm age was also log-transformed to ensure a better normal distribution. However, after log transformation there was a negative skewness of 1.091 and a positive kurtosis of 4.444. Before the log transformation there was a positive skewness of 0.373 and a positive kurtosis of 1.939, falling between the acceptable values to run regressions with according to George and Mallery (2010). Therefore the original values of the two continuous variables VC investment and Firm age were used to test the hypothesis, instead of the log-transformed variables. Lastly, the ordinal scales were centered to improve interpretation.

According to Lumley, Diehr and Chen regressions do not require the assumption of normal distribution to be met in order to interpret the relationship between variables (e.g. analyze differences and trends) in large samples, which is often around 100 data observations. When regressions are used to predict outcomes for individuals, knowing the distribution of the outcome is critical to estimate prediction intervals. As this study examines differences between groups, the normal distribution is not critical. Besides, the continuous variables comply to the boundaries of George and Mallery (2010), consequently testing the other assumptions was continued.

Normality of residuals

To test the normality of the residuals the Shapiro-Wilk W test was conducted. The Shapiro-Wilk W test is significant (p < .001). This means that the null-hypothesis indicating

(29)

not being met. However, since there are more observations than 20, the assumption of

normality of residuals does not have an impact due to the high number of observations (Long & Ervin, 2000). Subsequently, a histogram of the residuals and the P-P plot shows that the residuals approach normality (appendix 4). It is therefore concluded that the assumption of normality of residuals is met.

Homoscedasticity

The assumption of homoscedasticity of the residuals is met when the variance of a residual is not homogeneous. To test this assumption the Breusch-Pagan/ Cook-Weisberg test was conducted, which was not significant (p =.477). The null-hypothesis states that the variance of the residuals is homogeneous. The null-hypothesis is rejected. This means that the assumption for homoscedasticity of the residuals is met (appendix 5).

Absence of multicollinearity

A Spearman's correlation was executed to assess the relationship between the variables (see table 4). The ethnicity of the entrepreneur and the ethnicity of the founding team are highly correlated with each other (r = .99, p < .001). Therefore, to avoid potential

multicollinearity issues, separate regression models for each independent variable were used for the empirical analysis. As the correlation between the other variables were not above r = .80, multicollinearity problems were unlikely to be cause for concern.

Additionally, the variance inflation factor (VIF) scores of each variable was analyzed to make sure that no subtle forms of multicollinearity were present. The rule of

multi-collinearity is that the VIF-values should not be higher than 10 and the tolerance level should not be lower than 0.20 (Field, 2013; O’Brien, 2007). The variables score well below 10 on the VIF-values and the tolerance levels score higher than 0.20, except for ethnicity of the

(30)

variables are not executed in the same regression model. Consequently, multicollinearity is unlikely to form a problem. The assumption of absence of multicollinearity was met (appendix 6).

Hausman test

Lastly, the Hausman Test was performed to determine whether a fixed- or random-effects model needs to be used. A significant result for this test means a fixed-random-effects model should be used, in other cases a random-effects model is more appropriate. The results of this test were not significant for the ethnicity of the entrepreneur (chi-square = 0.710) and

ethnicity of the founding team (chi-square = 0.976), thereby indicating the use of a random-effects model to be more appropriate.

Descriptive statistics and correlations

For this study a panel data of 397 startups in 13 countries was used. Most firms in the sample originate from Canada (19.6%), France (19.1%) and the United Kingdom (19.1%), while Japan (1.5%), Spain (1.3%) and Switzerland (1.5%) formed the least represented countries. In this sample 54 firms (13.6%) of the total amount of 397 firms were founded by an ethnic entrepreneur. Most firms with ethnic entrepreneurs were located in Canada, France, Germany and the United Kingdom, as can be seen in table 2. Furthermore, firms where the majority of the founding team had an ethnic background (4.3%) were mostly located in Canada and the United Kingdom.

Table 3 displays the summary statistics of the variables. As mentioned before, the ordinal scales were mean-centered to improve interpretation.

(31)

Table 2. Overview of firms by country

Countries Firms with at least one ethnic founder(s) Firms with majority ethnic founder(s) Firms with native founder(s) Total number of firms Australia 2 1 15 17 Canada 14 6 64 78 China 1 - 17 18 Finland 1 - 14 15 France 6 1 70 76 Germany 8 2 30 38 Israel 1 - 26 27 Italy 1 - 7 8 Japan 1 1 5 6 Spain 1 - 4 5 Sweden 1 - 26 27 Switzerland 2 2 4 6 United Kingdom 15 4 61 76 Total 54 17 343 397

Table 3. Descriptive statistics (N = 397)

Variables Mean SD Min Max

1. VC investment 192.09 113.26 8 400 2. Ethnic entrepreneur .13 .34 0 1 3. Ethnicity team .18 .48 0 2 4.Uncertainty avoidance (c) .00 20.97 -27.71 35.29 5. Individualism (c) .00 10.99 -28.42 15.58 6. Power distance (c) .00 14.90 -14.66 34.34 7. Governance institutions (c) .00 .51 -1.85 .58

8. Firm age at funding 71.37 44.81 1 162

9. Firm development stage -.40 .49 -1 0

10. Firm industry .70 .92 0 3

11. Number of investors 2.37 1.42 1 10

(32)

Correlations

The correlation matrix in Table 4 shows that several variables correlate with venture capital investment. The control variables correlate with the dependent variable in the expected directions (p < .01). However, the ethnicity of the entrepreneur and ethnicity of the founding team do not significantly correlated to VC investment, contrary to the expectation of H1 (r = -.03, p > .10). Furthermore, there is a significant correlation between the dependent variable and individualism (r = -.15, p = .004) and governance institutions (r = -.27, p = .000). Contrary to the expectations, uncertainty avoidance, power distance and governance

institutions do not correlate significantly with ethnicity of the entrepreneur or ethnicity of the founding team (p > .10). However, individualism shows a marginally significant correlation with ethnicity of the entrepreneur (r = .09, p = .088) and ethnicity of the founding team (r = .09, p = .088). Lastly, governance institutions correlates significantly with uncertainty

avoidance (r = -.38, p = .000), individualism (r = .19, p = .000) and power distance (r = -.68, p = .000).

(33)

Table 4. Correlation statistics

Variables 1 2 3 4 5 6 7 8 9 10 11

1. VC investment -

2. Ethnicity entrepreneur -.03 -

3. Ethnicity founding team -.03 .99*** -

4.Uncertainty avoidance .07 -.04 -.04 -

5. Individualism -.15** .09† .09† -.49*** -

6. Power distance .06 -.06 -.07 .57*** -.22*** -

7.Governance institutions -.27*** .05 .05 -.38*** .19*** -.68*** -

8. Firm age at funding .48*** -.07 -.07 .03 -.02 .01 -.07 -

9. Firm development stage .15** -.03 -.03 -.03 -.00 .00 -.03 .11* -

10. Firm industry -.13** .09† .09* -.09† .07 .06 .04 -.17*** -.07 -

11. Number of investors .13** .03 .03 .16** .01 .17*** -.10† -.09† .10* .10* -

a. *** p < .001,** p < .01, * p < .05, †p < .10

(34)

Results

The results of the GLS random-effects regression models are reported in table 5 for the independent variable ethnicity of the entrepreneur and table 6 for the independent variable ethnicity of the founding team. First the results for the independent variable ethnicity of the entrepreneur are discussed.

As can be seen in Models 1-7, the control variables Firm age and Number of investors have a positive, statistically significant effect on VC investment (p < .001). The Firm

development stage has a marginally significant effect on VC investment (p < .10). Whereas, the control variable Firm industry does not have a significant effect on VC investment. All the models (1-7) are significant as a whole (p < .001).

Model 1 shows the baseline model, including only the control variables. In Model 2, the linear effect of the ethnicity of the entrepreneur on VC investment is added to the baseline model. The interaction terms of Uncertainty avoidance, Individualism, Power distance and Governance institutions are added separately in Models 3 - 6. Finally, Model 7 includes all the variables.

The first hypothesis in this study proposed a negative, direct relationship between the ethnicity of the entrepreneur and VC investments. Model 2 shows a statistically insignificant (p = .841) effect for the ethnicity of the entrepreneur on VC investments. Therefore,

hypothesis 1 is rejected.

The other hypotheses propose moderating effects. To test for these moderating effects, interaction terms were created. Hypothesis 2a predicted that high scores on uncertainty

avoidance positively moderates the relationship between ethnicity of the entrepreneur and VC investments. Hypothesis 2b posits that individualism negatively moderates the relationship between ethnicity of the entrepreneur and VC investments. Hypothesis 2c expected that power distance positively moderates the relationship between ethnicity of the entrepreneur and VC

(35)

investments. However, the interaction terms in model 3 for uncertainty avoidance (p = .809), model 4 for individualism (p = .588) and model 5 for power distance (p = .571) are

statistically not significant. Therefore, hypothesis 2a, 2b and 2c are not supported. The results of the moderators are not significant and as the coefficients are close to zero, they are too low to draw conclusions about the direction of the relationship.

Hypothesis 3 posits that governance institutions negatively moderates the relationship between ethnicity of the entrepreneur and VC investment. The interaction term is statistically not significant (p = .444). The coefficient does indicate a negative relationship between governance institutions and ethnicity of the entrepreneur on VC investment. However, as the interaction is not significant, hypothesis 3 is not accepted.

As can be seen in Model 7, the results of the tested hypotheses remain the same in the full model, i.e. the effects are still statistically insignificant. The hypotheses have been rejected. The possible reasons for the insignificant findings are presented in the discussion section.

Secondly, the results of the GLS random-effects regression models with independent variable ethnicity of the founding team are discussed. The models are run in the same order as the previous analysis. Model 1 in table 6 shows the baseline model, including only the control variables. In Model 2, the linear effect of the ethnicity of the founding team on VC investment is added to the baseline model. The interaction terms of Uncertainty avoidance, Individualism, Power distance and Governance institutions with ethnicity of the founding team are added separately in Models 3 - 6. Finally, Model 7 includes all the variables. All the models (1-7) are significant as a whole (p < .001).

The first hypothesis in this study proposed a negative relationship between the ethnicity of the founding team and VC investments. Model 2 shows a statistically insignificant (p = .778) effect for the ethnicity of the founding team on VC investments.

(36)

Therefore, hypothesis 1 is rejected.

The other hypotheses propose moderating effects, for which interaction terms were created. However, the interaction terms in model 3 for uncertainty avoidance (p = .666), model 4 for individualism (p = .229) and model 5 for power distance (p = .707) are statistically not significant. Therefore, the moderating effects of ethnicity of the founding team with uncertainty avoidance, individualism and power distance are not supported. The coefficient of the interaction term for hypothesis 3 indicates a negative relationship, however the term is statistically not significant (p = .521). As can be seen in Model 7, the results of the tested hypotheses remain insignificant in the full model. The results were not as expected and possible reasons for the insignificant results will be given in the discussion section.

(37)

Table 5. GLS random-effects regression models (Independent variable: Ethnicity entrepreneur).

Dependent variable: VC investments

Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Constant 58.698** (19.889) 58.385** (20.715) 58.115** (21.579) 59.548** (21.496) 57.724** (20.164) 55.998** (14.913) 56.079** (14.883)

Firm age at funding 1.335***

(.102) 1.336*** (.103) 1.337*** (.103) 1.333*** (.103) 1.337*** (.103) 1.337*** (.105) 1.336*** (.106) Firm development stage 16.952†

(9.271) 16.963† (9.276) 16.909† (9.280) 16.833† (9.280) 17.167† (9.297) 17.821† (9.534) 17.428† (9.597) Firm industry -4.774 (5.073) -4.833 (5.081) -4.833 (5.084) -4.964 (5.088) -4.794 (5.090) -4.548 (5.205) -4.654 (5.238) Number of investors 14.023*** (3.283) 14.020*** (3.285) 14.096*** (3.302) 13.951*** (3.289) 14.157*** (3.299) 13.969*** (3.350) 14.156*** (3.377) Ethnicity 2.660 (13.226) 3.079 (13.358) .803 (13.656) 4.615 (13.681) 5.933 (14.015) 3.812 (14.489)

Ethnicity * Uncertainty avoidance .170

(.702) 1.225 (1.536) Ethnicity * Individualism .663 (1.222) 1.903 (2.135)

Ethnicity * Power distance .602

(1.062)

-.674 (2.142)

(38)

a. *** p < .001, ** p < .01, * p < .05,†p < .10. b. Standard errors are reported in parentheses.

Ethnicity * Governance institutions -27.610

(36.102)

-49.273 (59.248)

Wald chi square 199 199 199 199 199 191 190

Number of observations 397

(39)

Table 6. GLS random-effects regression models (Independent variable: Ethnicity founding team).

Dependent variable: VC investments

Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

Constant 58.698** (19.889) 58.268** (20.833) 59.044** (21.944) 60.493** (21.819) 57.913** (21.826) 56.005*** (15.092) 54.491*** (14.883)

Firm age at funding 1.335***

(.102) 1.337*** (.103) 1.336*** (.103) 1.334*** (.102) 1.338*** (.103) 1.338*** (.105) 1.356*** (.109) Firm development stage 16.952†

(9.271) 16.961† (9.273) 16.938† (9.273) 16.508† (9.266) 17.048† (9.280) 17.692† (9.507) 17.891† (9.886) Firm industry -4.774 (5.073) -4.871 (5.084) -4.897 (5.084) -5.129 (5.080) -4.854 (5.085) -4.560 (5.199) -4.998 (5.380) Number of investors 14.023*** (3.283) 14.004*** (3.285) 13.914*** (3.293) 13.965*** (3.281) 14.107*** (3.298) 13.987*** (3.346) 13.511*** (3.406) Ethnicity 2.648 (9.413) 1.969 (9.553) -.258 (9.708) 4.149 (10.211) 5.766 (11.053) 7.045 (11.699)

Ethnicity * Uncertainty avoidance -.229

(.530) .848 (1.342) Ethnicity * Individualism 1.018 (.845) 1.396 (1.637)

Ethnicity * Power distance .324

(.863)

-.1.104 (1.776)

(40)

a. *** p < .001, ** p < .01, * p < .05, †p < .10. b. Standard errors are reported in parentheses.

Ethnicity * Governance institutions -21.369

(33.269)

-71.417 (50.594)

Wald chi square 199 199 199 201 199 191 185

Number of observations 397

(41)

Discussion

Implication of results

The purpose of this study was to develop a better understanding and to analyze the moderating role of formal (i.e. governance) and informal (i.e. national culture) institutions on VC investments for ethnic entrepreneurs. Therefore, the research question of this study states:

‘How does the ethnicity of the entrepreneur influence venture capital investments and to what

extent do formal and informal institutions play a role in this relationship?’

The proposed relationships were empirically tested by using a panel data sample of 397 startups from the information technology industry for the period 2000-2010.

This study predicted a negative relationship between the ethnicity of the entrepreneur and VC investments. Previous research found opposing conclusions, whereas in some studies ethnic entrepreneurs were credit constrained, other studies found no such difference between non-ethnic and ethnic entrepreneurs (Blanchﬂower et al., 2003; Bruder et al., 2011; Fraser, 2004). As previous studies were conducted in a specific country, this study presents a broader view as a result of testing in multiple countries. The result of this study shows that there seems to be no difference in VC investments for native and ethnic entrepreneurs. There also seems to be no difference for the founding team composition, that is if the team consists of native, mostly native or mostly ethnic entrepreneurs, for the VC investments received. Furthermore, the results of this study show no moderating effect of uncertainty avoidance on VC investments for ethnic entrepreneurs. In contrast to the proposition of Hofstede (1999) and Mangundjaya and Lecturer (2011) conveying that ethnic entrepreneurs experience a disadvantaged position in high uncertainty avoiding countries, as higher level of risks are assigned to ethnic entrepreneurs due to a different ethnic background. In this study there was no such difference found in VC investments for native and ethnic entrepreneurs, in neither low nor high uncertainty avoiding countries. Likewise, uncertainty avoidance did not

(42)

show a moderating effect on VC investments for ethnicity of the founding team. Additionally, findings of this research are not in line with studies that find

disadvantages in VC investments for ethnic entrepreneurs in low individualistic countries, as there was no support for a moderating effect of individualism on VC investments for ethnic entrepreneurs. The proposition that in collectivistic countries group agreements and loyalty towards the native group (in-group) provides obstacles for ethnic entrepreneurs in mobilizing resources (Bochner & Hesketh, 1994; Hofstede, 2001; Shane, 1992) is thus not supported by this study. There was also no support for a moderating effect of individualism on VC

investments for ethnicity of the founding team, indicating there are no differences in VC investments for a team consisting of only native, mostly native or mostly ethnic entrepreneurs in low or high individualistic countries.

Moreover, the results of this study show no moderating effect of power distance on VC investments for ethnic entrepreneurs. Former research shows that in high power distant countries social inequality is created, reinforced and accepted (Hofstede, 1994; Shane, 1993). There is low tolerance for change and low willingness to share control with others, which can result in misusage of power and placing minorities at a disadvantage (Carl, Gupta & Javidan, 2004). However, this does not seem to be the case for ethnicity of the entrepreneur and ethnicity of the founding team. By linking national culture to the ethnicity of the entrepreneur in this study, the conclusion can be formed that national culture does not influence differences in VC investments for native and ethnic entrepreneurs or for the founding team composition. Lastly, the influence of governance institutions did not show a moderating effect for VC investments for ethnic entrepreneurs or for the founding team composition. The

proposition that ethnic entrepreneurs receive lower VC investments in countries with low governance institutions, in comparison to native entrepreneurs is not supported. These findings are not in congruence with previous research, which state that in low governance

(43)

countries the reliance on ties is strong, especially with powerful groups, and ways to influence processes are necessary. As ethnic entrepreneurs often lack necessary social ties and resources to access or influence the processes in low governance countries, a disadvantage for ethnic entrepreneurs was expected (Bruton et al., 2009; McMullen et al., 2008). However, this study did not provide evidence for this expectation.

Contributions

This study contributes to the existing literature by adding multiple countries in the analyses, where most previous studies are executed in one country. By taking multiple countries in the dataset, the existing literature is extended by showing differences in one model, rather than comparing countries separately. Consequently, country-level moderators were added to VC investments for ethnic entrepreneurs, which showed that there seems to be no differences for native and ethnic entrepreneurs across countries in the VC investments received. Also for the founding team composition, that is if the team consists of native, mostly native or mostly ethnic entrepreneurs, there seems to be no difference in VC investments across different countries.

The results indicate that the hypothesized relationships are not valid, which perhaps indicate that the theoretical arguments for the expected relationships are not accurate. However, it is also possible that some elements of the research design have affected the results. As to my knowledge, this study is the first in its kind, this study has paved the way for future research, concerning the relationship of VC funding for ethnic entrepreneurs across multiple countries and to explore the effect of country-level variables.

The practical recommendation is that entrepreneurs and VCs should be aware that the environment of the country they operate in can have an impact on their operations. Although this study does not provide empirical evidence for the influence of country-level variables, in

(44)

reality different countries do have different circumstances. These circumstances, for example rules and regulations can influence the startup funding process. However, this

recommendation is based on previous research (Li & Zahra, 2012) and what is seen in practice rather than the results of this study.

Limitations of the study

There may be several reasons for the statistically insignificant findings. The method design limits the study in several ways. Firstly, this study is limited by the measurement of ethnicity of the founders through names. Ethnicity is measured by coding the names, which results in imprecision as some entrepreneurs may be misclassified based on their name (Bengtsson & Hsu, 2015). However, these imprecisions are randomly distributed over the dataset, therefore the results should not be biased. Moreover, the database used to code

ethnicity by Kerr (2008) is based on inventors in the US. However, if there was no match with the database, company communications and personal pages of the founders (as LinkedIn) were used to prevent inaccuracy in establishing ethnicity. The distinction between native and non-native entrepreneurs as ethnic entrepreneurs was taken consciously, as a result of testing in multiple countries in which minorities differ. However, most previous research examines ethnic minorities within a specific country, for example the ‘black community’, but also Indian and Chinese minorities in the US (Blanchflower et al., 2003; Saxenian, 2002). Distinguishing non-native entrepreneurs as ethnic entrepreneurs can be problematic, as previous research was limited to specific ethnic groups within a country. Therefore, the theoretical literature used to form the hypotheses presents an important theoretical limitation. Namely the theoretical foundation for the hypotheses exists mostly of research analyzing specific ethnic groups within one country. As a result of testing in multiple countries ethnic

Venture capital investments, institutions and ethnic entrepreneurs in the information technology industry : a cross-country analysis