Quality versus quantity: Government institutions and their effects on social entrepreneurship

(1)

Quality versus quantity: Government institutions and their

effects on social entrepreneurship

Abstract: This cross-national research empirically investigates the influence of government institutions on the prevalence of social entrepreneurship by combining institutional void theory and institutional support theory. Based on a nuanced understanding of these theories, I propose that government size has a negative relationship with prevalence rate of social entrepreneurship, whereas overall government quality and some underlying quality aspects are hypothesized to have a positive relationship. Subsequently, I propose an ideal government configuration where social entrepreneurship thrives most. Using data from the Global Entrepreneurship Monitor (GEM) 2015, I perform a cross-sectional analysis of 58 countries using multiple regression. I find that country prevalence of social entrepreneurship benefits from higher-quality governments but is not significantly influenced by government size. Additionally, I did not find evidence for an ideal government configuration. My research contributes to institutional theory by providing a more nuanced view on the forces of institutional voids and institutional support, and paves the way for future research on formal institutions as a determinant of social entrepreneurial activity.

Keywords: Social entrepreneurship ∙ Institutional void theory ∙ Institutional support theory ∙ Global Entrepreneurship Monitor

(2)

1

Introduction

Social entrepreneurship is a phenomenon that is receiving wide interest due to its potential to provide creative solutions to alleviate modern social issues (Yunus, 1998; Zahra et al., 2009). Social entrepreneurs are particularly focused on social wealth creation instead of pure economic wealth creation (Dees, 1998; Bosma et al, 2016) and effectively combine the realms of entrepreneurship, social movements and not-for-profit management in their business (Mair and Martí, 2006; Dacin et al., 2010). Although the term has only emerged in recent years, the practice itself has a long heritage around the globe (Dees, 1998; Mair and Martí, 2006).

Encouraging nascent entrepreneurs to pursue social objectives may result in higher income equality, social inclusion and green initiatives, along with lower poverty (Jenson, 2015). These potential outcomes have incentivized researchers and policymakers to investigate the drivers of social entrepreneurship in, among others, formal and informal institutional spheres (Harding, 2004; Hechavarría, 2015; Estrin et al., 2013). Investigations into formal institutions in the context of social entrepreneurship have led to the development of the institutional void theory and institutional support theory. Institutional void theory argues that demand for social enterprises increases when formal institutions are lacking in social welfare provision (Mair and Martí, 2009), while institutional support theory suggests that formal institutions can increase social entrepreneurial activity through extensive support and a naturally occurring partnership between the social entrepreneur and the government (Young, 2008).

Among the most prominent examples of social entrepreneurship is Muhammad Yunus’ Grameen Bank, which thrived due to the existence of an institutional void in Bangladesh (Mair and Martí, 2009; Mulgan, 2006; Yunus, 1998). Yunus’ ‘Bank for the poor’ provided microfinancing, small long-term loans without collateral, as a way to reduce rampant poverty in his country (Yunus, 1998). Through his efforts against poverty as a social entrepreneur he was awarded the Nobel Peace Prize for 2006 (Nobel Media AB, 2006).

(3)

2

on a nuanced understanding of institutional voids and institutional support: I analyze the proposed beneficial effects of both forces in the distinct dimensions of government size and quality, and investigate how governments can most effectively be configured in order to maximize their combined beneficial effects with regards to social entrepreneurship. Examining these two institutional dimensions separately contributes to an enriched understanding institutional theory and provides implications for policymakers aiming to facilitate social entrepreneurship. Additionally, to provide a richer understanding of drivers of social entrepreneurship related to government quality, I will analyze how different sub-indicators of government quality influence the prevalence rate of social entrepreneurship.

Despite fast developments, research on social entrepreneurship remains defined by a plethora of theory and case studies (Mair and Martí, 2006) and a lack of empirical research in e.g. country comparisons (Terjesen et al., 2016). In this empirical research, based on cross-sectional data from the Global Entrepreneurship Monitor (GEM) 2015, I will analyze the relationship between two distinct institutional dimensions and the prevalence rate of social entrepreneurship across 58 countries. Although I do not find direct support of the negative relationship of government size with prevalence rate of social entrepreneurship, this finding may be explained by the effects of institutional voids and institutional support offsetting each other. I did find that higher government quality is beneficial for social entrepreneurship. In particular, higher levels of control of corruption, regulatory quality, rule of law and voice and accountability have a positive influence. I find strong support for institutional support theory in the quality dimension, but I did not find evidence for the proposed ideal institutional configuration for social entrepreneurs, consisting of a relatively smaller government of relatively higher quality.

The remainder of this paper is organized as follows. Firstly, the literature review provides a brief overview of the core theory, I develop a working definition of social entrepreneurship and relate it to institutional void and support theory, forming the foundation for the development of my hypotheses. Secondly, I will outline the empirical design and methodology of my research. Thirdly, the results are presented along with various robustness checks. Lastly, I will discuss my findings, present implications for researchers, policymakers and managers, and report limitations and future research directions.

Literature review

(4)

3

on the effects of two distinct dimensions of government institutions on prevalence of social entrepreneurship. Lastly, I visualize all hypothesized relationships in my conceptual model. Social entrepreneurship

Social entrepreneurship is a rapidly growing field that has attracted attention from the media, public officials and a plethora of other sectors (The Economist, 2010). The growing popularity is mainly attributed to its potential to provide creative solutions to alleviate modern social issues (Yunus, 1998; Zahra et al., 2009). Furthermore, individuals see something inherently fascinating about entrepreneurship in general, and the social aspect only adds to that (Martin and Osberg, 2007). Despite the growing popularity of social entrepreneurship, the term has been lacking a commonly used definition that is accepted by most researchers and practitioners alike (Zahra et al., 2009). In the following paragraphs, I will describe key developments on the definition of social entrepreneurship over the years, systematically examining the most widely accepted definitions in the field, eventually resulting in a synthesized working definition for this research.

The earliest descriptions of social entrepreneurship stem from the classical view on entrepreneurship, as it can be considered a form of Schumpeterian entrepreneurship (Sud et al., 2008). Schumpeter’s writings set the groundwork for the current understanding of entrepreneurship (Castaño et al., 2015) and described entrepreneurship as the engine of economic progress through creative destruction, and entrepreneurs as agents of change (Schumpeter, 1943). Although occasionally overlooked, his writings on entrepreneurship have a clear social backdrop, stating that “Not only with regard to the economy, but also socially the entrepreneur must be on top of the pyramid of society” (Schumpeter, 2002, p. 414). In light of this, some consider his writings on entrepreneurship to contain one of the first definitions of social entrepreneurship (Shockley and Frank, 2011). Although he did not explicitly define social entrepreneurship, he laid the foundation for later definitions of social entrepreneurship.

(5)

4

Dees’ definition is, in many ways, similar to a more recent widely used definition developed by Mair and Martí (2006), who broadly define social entrepreneurship as a process of combining resources in an innovative manner in order to seek opportunities and fulfill social objectives. Additionally, they suggest that social entrepreneurship is not exclusively a process happening in new organizations and can also occur in established organizations.

Furthermore, in their highly influential 2007 paper, Martin and Osberg argued that earlier definitions of social entrepreneurship were too inclusive and subsequently developed a more rigorous definition, which is as follows.

The social entrepreneur should be understood as someone who targets an unfortunate but stable equilibrium that causes the neglect, marginalization, or suffering of a segment of humanity; who brings to bear on this situation his or her inspiration, direct action, creativity, courage, and fortitude; and who aims for and ultimately affects the establishment of a new stable equilibrium that secures permanent benefit for the targeted group and society at large. (p. 39)

Although it is clearly more thorough and detailed, I believe this definition is in line with the idea Dees (1998) and Mair and Martí (2006) portrayed about social entrepreneurship.

Finally, the Global Entrepreneurship Monitor (GEM), which is broadly recognized as the world’s best source of comparative (social) entrepreneurship data (Lepoutre et al., 2013), also came up with a broad and narrow definition of social entrepreneurship, which have been utilized in their yearly global study of entrepreneurs. For the broad definition, GEM defines a social entrepreneur as someone who is ‘’starting or currently leading any kind of activity, organization or initiative that has a particularly social, environmental or community objective’’ (Bosma et al., 2016, p. 5). Their narrow definition is more exclusive by imposing stricter measures: Social entrepreneurs should prioritize social and environmental objectives over financial objectives, and they have to operate in markets through producing goods and services (Bosma et al., 2016). Although clearly less rigorous than Martin and Osberg’s (2007) definition, I believe that the GEM’s broad definition is in line with, and builds on, the aforementioned definitions of social entrepreneurship. This definition is utilized in the GEM 2015 adult population survey, which is the core dataset for testing my hypotheses, and thus I will work with this definition.

(6)

5

focus is on formal institutions, which include e.g. constitutions and laws that are enforced by a government (North, 1991). Institutional void theory argues that demand for social enterprises increases when formal institutions are lacking in social welfare provision (Mair and Martí, 2009), while institutional support theory suggests that formal institutions can increase social entrepreneurial activity through extensive support and a naturally occurring partnership between the social entrepreneur and the government (Young, 2008).

Institutional void theory

Institutional voids are situations in which government institutions are small and/or of low quality, and noticeably lacking in providing social welfare in their respective countries. (Mair and Marti, 2009). Institutional void theory suggests that the demand for social entrepreneurs is higher when social welfare provision by formal institutions is lacking (Dacin et al., 2010; Mair & Marti, 2009). Such voids can originate from numerous sources and are not necessarily unintentionally created. An intentionally small government may leave a demand for social welfare provision, but a dysfunctional, low-quality government will also likely create voids left to be filled (Estrin et al., 2013). A common pattern in the creation of institutional voids is the dysfunctional interaction between legacy institutions, existing power structures and nascent institutional practices (Mair and Marti, 2009). Institutional voids can both present and limit opportunities for entrepreneurs (Baker et al., 2005), but general entrepreneurial start-up efforts are found to be more likely when a government is relatively smaller (Estrin et al., 2013). Additionally, earlier research has found that the absolute number of social entrepreneurs in a country declines when government spending increases (Ferri & Urbano, 2011; Estrin et al., 2013). Muhammad Yunus’ Grameen bank is a remarkable case of a social enterprise that thrived due to the existence of an institutional void (Yunus, 1998).

While institutional void theory examines what government institutions are lacking, institutional support theory looks at what they have to offer to aid social entrepreneurs in their operations through both general well-functioning and targeted support for social entrepreneurs (Hoogendoorn, 2016). Although the theories postulate alternative views on the effects of government institutions, their suggested beneficial effects on prevalence rate of social entrepreneurship are not mutually exclusive.

Institutional support theory

(7)

6

social entrepreneurship in countries with well-functioning governments. Institutional support in this context may encompass generally efficient, high-quality institutions but also more specific aspects such as grants and subsidies for social entrepreneurs in order to improve social welfare in their country (Sud et al., 2009; Zahra & Wright, 2011). Essentially, this development can be seen as governments partially outsourcing welfare-related tasks because they realize they are more effectively solved by market players, with the two parties becoming natural partners in the process. Several European countries, which are generally qualified as welfare states, are attempting to embrace this approach in order to achieve a more efficient and effective welfare system (Hoogendoorn, 2016; Young, 2000). Furthermore, the European Commission is actively supporting social entrepreneurship by providing extensive access to finance and building an enabling ecosystem for social enterprises (European Commission, 2019). Finally, Stephan et al. (2015) found that government activism has a significant positive effect on individuals’ decision-making process of whether to engage in social entrepreneurship.

Hypothesis development

I argue that the size and quality of government institutions should be investigated as distinct dimensions in regard to institutional void theory, institutional support theory and prevalence of social entrepreneurship. This allows for a more specific examination of where the suggested effects of institutional void theory and institutional support theory are strongest and may provide new insights in the most beneficial configuration of government institutions in order to maximize prevalence of social entrepreneurship. Although the two theories hypothesize on opposite sides of the spectrum of formal institutions, the suggested positive effects from either theory are not necessarily mutually exclusive and can even be complementary.

Government size and social entrepreneurship

Government size is the first of two formal institutional dimensions I will investigate, and it can be a major determinant for the prevalence of social entrepreneurship in a country. I define government size as to what extent a government’s institutions provide services such as social welfare provision, measured by government expenditure as a % of GDP (Aidis et al., 2012; Castles and Dowrick, 1990; Estrin et al., 2013; Stephan et al, 2015).

(8)

7

development of an enabling ecosystem for social enterprises (European Commission, 2019). These opposing forces can vary in strength across countries and will partially cancel each other out with regards to the overall effect of government size on social entrepreneurship. This is evident in earlier research. While recently Estrin et al. (2013) and Ferri and Urbano (2011) found a significant negative relationship between government size and prevalence of social entrepreneurship, Stephan et al. (2015) found a significant positive relationship. Although both forces may play an important role in increasing social entrepreneurial activity, I argue that the force of institutional voids is generally more impactful due to its widespread general impact, as opposed to the targeted impact of support in the form of grants and subsidies.

Hypothesis 1: Government size has a negative relationship with country prevalence rate of social entrepreneurship.

Government quality and social entrepreneurship

Government quality is the second formal institutional dimension of this study and relates to arguments of both institutional void theory and institutional void theory. I define government quality as the level of “traditions and institutions by which authority in a country is exercised” (Kaufmann et al., 2010, p. 4). Government quality is composed of six indicators: control of corruption, government effectiveness, political stability, regulatory quality, rule of law and voice and accountability (Kaufmann et al., 2010).

Entrepreneurship is general is severely harmed by the absence of high-quality government institutions (Chambers and Munemo, 2017), and social entrepreneurship may feel an even stronger negative influence, although it also presents opportunities for them. Similarly to small governments, low-quality governments can be the origin of institutional voids, creating demand for social welfare provision (Dacin et al., 2010; Mair & Marti, 2009). However, I argue that social entrepreneurs have more difficulties fulfilling this demand in countries with low-quality governments because their activities are hindered by multiple characteristics of dysfunctional formal institutions. Drawing on earlier research I propose that the existence of corruption (Anokhin and Schulze, 2009; Aidis et al., 2012) and a weak rule of law (Puumalainen et al., 2015; Hoogendoorn, 2016; Chambers and Munemo, 2017) are the most inhibiting characteristics of low-quality government institutions with regards to social entrepreneurial activity, while high regulatory quality can be a very beneficial aspect of high-quality governments (Puumalainen et al., 2015; Kaufmann et al., 2010).

(9)

8

licenses and tax documents (Chambers and Munemo, 2017), uncertainty and ambiguity are higher for (social) entrepreneurs (Anokhin and Schulze, 2009), and unproductive activities such as rent seeking (e.g. lobbying) are rewarded under low-quality institutions, crowding out more productive activities such as entrepreneurship (Chambers and Munemo, 2017).

Secondly, a weak rule of law is suggested to be detrimental to social entrepreneurship. Estrin et al. (2013) found that a stronger rule of law is related to higher prevalence of social entrepreneurship due to the benefits of predictability and a level playing field. Furthermore, Hoogendoorn (2016) found that social entrepreneurship benefits more from a strong rule of law than commercial entrepreneurs do. Arbitrary governments, which are a symptom of a weak rule of law, tend to use their power to pursue private interests on a large scale through lobbying and corruption, particularly crippling social initiatives such as social entrepreneurship (Estrin et al., 2013).

Thirdly, regulatory quality could be considered the most relevant dimension with respect to the institutional support theory. It is defined as governments’ ability to develop and carry out policies and regulations that are beneficial for the private sector (Kaufmann et al., 2010; Puumalainen et al., 2015), which can greatly benefit social entrepreneurs.

Institutional voids may originate from low-quality governments, creating a demand for social welfare provision. However, the benefits of a high-quality government, or rather the disadvantages of a low-quality government, far outweigh the benefits of the social entrepreneurial opportunities created by institutional voids. Thus, largely in favor of the institutional support theory, I propose that overall government quality has a positive influence on the prevalence of social entrepreneurial activity in a country. Furthermore, due to the aforementioned underlying mechanisms of government quality, I hypothesize on the separate effects of control of corruption (a proxy measure used for existence of corruption), rule of law and regulatory quality.

Hypothesis 2a: Government quality has a positive relationship with country prevalence rate of social entrepreneurship.

Hypothesis 2b: Control of corruption has a positive relationship with country prevalence rate of social entrepreneurship.

Hypothesis 2c: Rule of law has a positive relationship with country prevalence rate of social entrepreneurship.

(10)

9

Government configurations and social entrepreneurship

Based on my previous hypotheses, I argue that there should be a most beneficial configuration of the aforementioned government institution dimensions in order to maximize the prevalence rate of social entrepreneurship. A smaller government is suggested to be beneficial, mostly due to the effects of institutional voids, while a higher quality government is largely suggested to be beneficial through the effects of institutional support. Thus, I propose that a relatively smaller government size and relatively higher government quality is the most beneficial formal institutional configuration for social entrepreneurship because it maximizes the combined benefits of institutional voids and institutional support.

Hypothesis 3: There is a moderating relationship between government size and government quality as such that a relatively smaller government size and relatively higher government quality is the most beneficial configuration for the prevalence rate of social entrepreneurship.

Figure 1 presents the conceptual model, including the hypotheses and variables involved in the main analyses.

Figure 1: Conceptual model

Note: Government quality indicators in parentheses have no individual hypotheses but will be tested in the

(11)

10

Empirical design and methodology

Data and sample

In order to test my hypotheses, a dataset with information on prevalence rates of social entrepreneurship that spans over a large number of countries is needed. This data will be sourced from the Global Entrepreneurship Monitor (GEM) adult population study, which is widely recognized as the world’s best source of comparative entrepreneurship data (Shorrock, 2018). Their 2015 dataset includes a special report on social entrepreneurship and is a good fit for addressing my proposed hypotheses. It contains data on social entrepreneurial activity for 167.793 adults in 58 economies, with at least 2.000 respondents per economy (Bosma, 2016). It is a more recent version of the GEM consortium’s first special report on social entrepreneurship, which is based on 2009 data. As described in this paper’s literature review, GEM’s questions on and definitions of social entrepreneurship are rooted in social entrepreneurship literature and further developed through a number of single-country pilot studies (Harding and Cowling, 2004; Levie and Hart, 2011).

Data on governments size is sourced from the Heritage Foundation 2014 index of Economic Freedom (Heritage Foundation, 2014). The database includes data on 183 countries of which 57 are also covered by the GEM dataset. Data on government quality is sourced from the World Bank’s World Governance Indicators (World Bank, 2015b). This data is reliable, suitable for making cross-country comparisons (Charron, 2010), and contains relevant information on all 58 economies covered by the GEM dataset. Data on GPD per capita (ppp) and GDP growth will be sourced from the World Bank’s national accounts data (World Bank, 2015a). Lastly, data on social provision is sourced from the OECD (OECD, 2015).

All predictor variables and control variables are lagged by one year to reduce potential endogeneity.

Dependent variables

(12)

11

environmental value over financial value. Secondly, the activity should operate in a market through production of goods and services (Bosma et al., 2016).

For this research, I will use a combined measure of nascent (start-up phase) and operational (post start-up phase) social entrepreneurship. Using only nascent activity perhaps better reflects the effects of the institutional size and quality per country in recent years, as opposed to further back in history. However, government institutions are relatively static (Olsen, 2009), and combining nascent and operational social entrepreneurial activity greatly increases the sample pools of within-country data. Hence, I will use the combined measure of nascent and operational social entrepreneurial activity.

The broad definition of social entrepreneurial activity will be used in this research, as I deem it the most appropriate regarding the established definitions of social entrepreneurship. Additionally, the narrow definition will be used in a robustness check.

Independent variables

Government size: The size of government’s welfare institutions will be measured by a country’s general government expenditure as a percentage of the country’s total GDP using Heritage Foundation data (Heritage Foundation, 2014). The measure covers multiple aspects of institutional welfare expanse, such as education, health services and unemployment insurance.

Earlier research regarding social entrepreneurship and formal institutions has often looked at general government expenditure to measure the size of a government and relate it to social entrepreneurial activity (Reynolds, 2011; Estrin et al., 2013; Fogel et al., 2005; Aidis et al., 2012). General government expenditure is not a direct measure of government welfare expanse and does not exclude e.g. military spending, which is highly variable across countries (Roser and Nagdy, 2019) while not being relevant in the context of institutional voids and social entrepreneurship. However, to my knowledge, it is still the best measure that provides acceptable coverage of countries.

(13)

12

There are a number of sources providing measures for general government spending data, including General government spending by the OECD (OECD, 2014), Government Spending by the Heritage Foundation (2014). Government expense and General government final consumption expenditure by the World Bank (World Bank, 2016).

My government size variable is based on Heritage Foundation data on government spending because it explains the most variance of the OECD’s social provision expenditure measure out of all listed sources. It has a correlation of 0.86 and explains 74.4% of the variance of social provision expenditure (N=35, p<0.001) across countries. Estrin et al. (2013) additionally noted significant correlations between the Heritage Foundation government spending variable and government expenditure on health and education as a percentage of GDP. Therefore, I consider it the most suitable measure for my research.

In conclusion, following Reynolds (2011); Estrin et al. (2013); Fogel et al. (2005) and Aidis et al. (2012), government size will be measured by a country’s general government expenditure as a percentage of GDP.

Government quality: Government quality will be measured by a comprehensive variable composed of six indicators from the World Governance Indicators (World Bank, 2015). The World Governance Indicators are based on 31 different sources of data from 25 organizations, representing a balanced and diverse view from a large array of stakeholders (Anokhin and Schulze, 2009). My comprehensive government quality measure is calculated by a simple average of the six indicators, which Chambers and Munemo (2017) established to be an appropriate way to measure overall institutional quality. In addition to the comprehensive government quality measure, the six indicators will also be tested separately on their influence on the prevalence of social entrepreneurship. Although my additional hypotheses only cover control of corruption, regulatory quality and rule of law, all indicators will be tested in the analysis.

The indicators are scaled in a standard normal distribution and range from -2.5 to 2.5. A higher value on an indicator always suggests a more positive outcome on the respective institutional aspect (Kaufmann et al., 2010). Kaufmann et al. (2010) describe the indicators as follows.

Control of corruption captures the perceived extent of public power being used to pursue private interests on a large scale through e.g. lobbying and corruption.

(14)

13

Political stability and absence of violence/terrorism captures perceptions of the probability of political instability and violence motivated by politics.

Regulatory quality measures perceptions of how well governments are able to develop and carry out policies and regulations that are beneficial for the private sector, including entrepreneurship.

Rule of law measures perceptions of the confidence stakeholders have in the rule of law, and the perceived limits of arbitrary power that could be exercised by the government.

Voice and accountability captures to what extent a country’s citizens perceive to have freedoms related to media, expression and association, along with being able to select their government.

Control variables

GDP: GDP per capita (ppp) will be used as a control variable. Including national wealth is an accepted best practice in research on (social) entrepreneurship in order to reduce the effects of potential omitted variable bias (Aidis et al., 2012; Estrin et al., 2013; Stephan et al., 2015). Additionally, Lepoutre et al. (2013) found that the prevalence of entrepreneurship and social entrepreneurship varies with levels of development, which they measured by national wealth. Data will be sourced from the World Bank and transformed into a logarithmic variant due to the distributional nature of the variable.

GDP growth: Changes in national wealth can also have an impact on social entrepreneurship (Stephan et al., 2015), so I will incorporate World Bank data on GDP growth (World Bank, 2019) as a control variable.

An overview of variable definitions and their respective data sources is presented in table 10 which can be found in the appendix.

Data analysis

I employ multiple regression models using ordinary least squares estimation on my cross-country, cross-sectional data to test my hypotheses. This approach is consistent with other country-level analyses in academic entrepreneurship research (Storey & Greene, 2010). Data from all sources was aggregated and cleaned in Stata and regression models were estimated in the same application (StataCorpLLC, 2019), while additionally using the asdoc plugin for Stata to send publication-ready tables to Microsoft Word directly (Shah, 2018).

(15)

14

suggests that multicollinearity is indeed absent among my main independent variables. A VIF of 10 or higher means there is a potential problem. With a mean VIF of 2.493 and the highest VIF value being 3.404 for government quality, there is no cause for concern. Adding the separate indicators that compose government quality will result in severe multicollinearity problems. However, these will only be used separately in analyses, which ameliorates this issue.

Table 1: Multicollinearity statistics

Social entrepreneurial activity VIF1 _1/VIF1

Government size 1.935 .517

Government quality 3.404 .294

GDP per capita (log) 3.307 .302

GDP growth 1.324 .755

Mean VIF 2.493 .

1_{VIF = Variance Inflation Factor}

Results

The results section for this research will feature three parts. In the first part I will describe the sample and analyze the descriptive statistics and correlations of all variables. The second part focuses on the core results of the analyzed hypotheses. The third part contains robustness checks. Descriptive statistics and correlations

Table 2 presents descriptive statistics for all variables used in the main analysis. Overall, the sample includes 58 countries with varying socio-economic backgrounds. Puerto Rico is missing in analyses involving government size due to missing data. Although I consider the sample to give a quite global picture as it includes countries from all major global regions, the distribution of factor-driven, efficiency-driven and innovation-driven economies is overrepresented by innovation-driven economies. The GEM uses economic development classifications from the Global Competitiveness Index (Bosma, 2016, p. 11), and while the Global Competitiveness Index classifies about 27% of economies as innovation-driven (World Economic Forum, 2014, p. 11), the GEM 2015 country sample has a composition containing almost 40% countries with innovation-driven economies. Table 11, which can be found in the appendix, displays the 58 countries and their respective types of economies (Lepoutre et al, 2013).

(16)

15

relationship between the prevalence of overall entrepreneurship and GDP per capita (Wennekers et al., 2005; Wennekers et al., 2010). This observation might indicate that there is a u-shaped relationship at play between the prevalence of social entrepreneurship and GDP, which could be explained by necessity motivations in countries such as Senegal and philanthropy and opportunity motivations due to wealth and extensive institutional support in countries such as Luxembourg. After visually plotting and analyzing these variables, I did not find evidence for a potential u-shaped relationship, thus I will not take this possibility into account for my research. The comprehensive independent variable of government quality has additionally been split up in its six separate indicators in the descriptive statistics and correlation statistics sections in order to further analyze which indicators have the strongest influence on the prevalence of social entrepreneurship. Table 3 can be found on the next page and provides an overview of correlations between the variables. As expected, government quality shows very high correlations with all variables it is composed of, and it presents surprisingly little correlation with government size. Furthermore, government size shows a slightly negative correlation with prevalence of social entrepreneurship. GDP per capita (log) sees especially high correlations with government quality indicators, moderately low correlation with government size, moderately high correlation with prevalence rate of social entrepreneurship.

Table 2: Descriptive statistics of the dependent, independent and control variables

Variables N Mean Std. Dev. Min Max Dependent variables

Social entrepreneurship 58 5.657 3.882 .999 18.079

Narrow SE (robustness) 58 3.521 2.801

Government size 572 _35.779 _10.630 _14.63 _55.100

Social provision (robustness) 26 20.420 5.927 7.568 30.191

Government quality 58 .455 .815 -1.003 1.838 Control of corruption 58 .409 .979 -1.164 2.227 Government effectiveness 58 .594 .791 -.823 2.112 Political stability 58 .178 .804 -1.698 1.403 Regulatory quality 58 .576 .854 -1.458 1.884 Rule of law 58 .507 .949 -1.058 2.100

Voice and accountability 58 .469 .861 -1.617 1.684

Control variables

GDP per capita 58 26788.770 18262.740 3067.034 101297.58

GDP per capita (log) 58 9.962 .733 8.028 11.526

GDP growth 58 3.010 2.170 -2.513 8.557

(17)

16

Table 3: Correlations among the dependent, independent and control variables

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 1 Social entrepreneurship 1.000 2 Narrow SE (robustness) 0.917 1.000 3 Government size -0.103 -0.076 1.000 4 Social provision (robustness) -0.090 -0.050 0.880 1.000 5 Government quality 0.299 0.327 0.189 0.289 1.000 6 Control of corruption 0.338 0.353 0.171 0.262 0.982 1.000 7 Government effectiveness 0.216 0.229 0.137 0.192 0.950 0.951 1.000 8 Political stability 0.173 0.227 0.248 0.409 0.759 0.652 0.595 1.000 9 Regulatory quality 0.364 0.385 -0.024 0.019 0.912 0.908 0.885 0.550 1.000 10 Rule of law 0.311 0.351 0.202 0.268 0.980 0.989 0.959 0.641 0.910 1.000 11 Voice and accountability 0.218 0.227 0.329 0.474 0.951 0.921 0.861 0.799 0.796 0.909 1.000

12 GDP per capita (log) 0.349 0.294 0.246 0.341 0.780 0.774 0.780 0.573 0.638 0.774 0.781 1.000

(18)

17

Government size, government quality and social entrepreneurship

I find no support for hypothesis 1: Government size has a negative relationship with country prevalence rate of social entrepreneurship. Table 4 shows the results of the analysis of the relationships between government size, government quality and the prevalence of social entrepreneurial activity. Including just control variables in the regression results in an R-squared value of 0.102. As shown in model 1, government size, controlled by GDP per capita (log) and GDP growth, shows a statistically insignificant coefficient of 0.052 in relation to social entrepreneurship. Additionally, controlling for government quality, as shown in model 3, did not yield a significant result.

I find strong support for hypothesis 2a: Government quality has a positive relationship with country prevalence rate of social entrepreneurship. Table 4, model 2 shows the regression results of the relationship between government quality and prevalence rate of social entrepreneurship. The effect of government quality is significant at p<0.05, even when controlling for government size, as shown in table 4, model 3.

Table 4: Regression results explaining social entrepreneurial activity using government size and

quality

Social entrepreneurial activity Controls Model

1 Model 2 Model 3

Government size 0.052 0.003

(0.063) (0.063)

Government quality 2.654** 2.644**

(1.009) (1.088)

GDP per capita (log) 0.751 0.378 -1.721 -1.731

(0.729) (0.864) (1.167) (1.198) GDP growth 0.615** 0.682** 0.562** 0.570** (0.246) (0.271) (0.235) (0.264) Constant -3.677 -2.032 19.903* 19.881* (7.582) (7.922) (11.500) (11.779) Observations 58 57 58 57 R-squared 0.102 0.107 0.204 0.198

Standard errors are in parenthesis *** p<0.01, ** p<0.05, * p<0.1

Government quality indicators and social entrepreneurship

(19)

18

Table 5: Regression results explaining social entrepreneurial activity using government quality indicators

Social entrepreneurial activity Model

4 Model 5 Model 6 Model 7 Model 8 Model 9 Control of corruption 2.107** (0.799) Government effectiveness 1.259 (1.246) Political stability 0.320 (0.914) Regulatory quality 2.445** (0.927) Rule of law 1.969** (0.912)

Voice and accountability 2.582***

(0.814)

Government size 0.004 0.038 0.046 0.036 0.003 -0.016

(0.062) (0.064) (0.066) (0.060) (0.065) (0.062)

GDP per capita (log) -1.525 -0.721 0.182 -1.883 -1.387 -1.107

(1.092) (1.390) (1.036) (1.185) (1.169) (0.925) GDP growth 0.539** 0.622** 0.674** 0.543** 0.537* 0.730*** (0.263) (0.278) (0.275) (0.263) (0.271) (0.251) Constant 18.218* 8.852 0.122 20.078* 16.756 13.845 (10.744) (13.373) (10.090) (11.252) (11.593) (8.867) Observations 57 57 57 57 57 57 R-squared 0.212 0.124 0.109 0.213 0.181 0.252

Government configurations and social entrepreneurship

(20)

19

Table 6: Regression results explaining social entrepreneurial activity through the interaction of

government size and government quality

10

Government quality 2.156

(2.559)

Government size -0.004

(0.072)

Government quality * government size -0.013

(0.064)

GDP per capita (log) -1.725

(1.210) GDP growth 0.560** (0.270) Constant 20.033* (11.910) Observations 57 R-squared 0.199

Robustness checks

Several robustness checks are performed in order to evaluate the sensitivity of the main results when variables are from a different cross-section or when alternative variables are added to the models. For the additionally generated models, the model numbers in parentheses correspond with the hypothesis the robustness check is performed for.

Firstly, in table 7, model 9 and 10 show how the main results change when replacing government size with government social provision expenditure. The coefficients are similar to when using government size, which adds to measurement validity.

Table 7: Robustness test replacing government size with social provision

11 (1) Model 12 (1)

Social provision -0.031 -0.031

(0.131) (0.133)

Government quality 1.016

(1.978)

GDP per capita (log) 0.917** 0.941**

(0.431) (0.441) GDP growth 2.864 1.725 (1.911) (2.949) Constant -25.179 -14.397 (18.884) (28.454) Observations 26 26 R-squared 0.313 0.321

(21)

20

Secondly, I regressed government size and quality on social entrepreneurial activity according to GEM’s narrow definition. Results are similar to the main results in both cases, adding to the robustness of my main results for hypotheses 1 and 2a (table 8, model 13, 14 and 15).

Table 8: Robustness test using narrow definition of social entrepreneurship

Narrow social entrepreneurial activity Model

13 (1) Model 14 (2) 15 (1+2) Model

Government size 0.032 -0.001

(0.046) (0.047)

Government quality 1.766** 1.779**

(0.741) (0.799)

GDP per capita (log) 0.569 -0.846 -0.849

(0.629) (0.858) (0.880) GDP growth 0.435** 0.356** 0.360* (0.198) (0.172) (0.194) Constant -4.615 10.069 10.128 (5.772) (8.450) (8.653) Observations 57 58 57 R-squared 0.094 0.175 0.173

Lastly, I tested hypotheses 1 and 2a using data relevant for 2009 prevalence rates of social entrepreneurial activity to achieve higher compatibility with earlier findings (Estrin et al., 2013; Stephan et al., 2015; Hoogendoorn, 2016). The GEM 2009 country sample differs from the 2015 sample and it consists of 48 usable countries instead of 58. Furthermore, the Heritage Foundation only provides a quadratic variant of their government spending data, which I converted back to a simple ratio. The results for hypothesis 1 are similar, again finding no support for the proposed relationship between government size and country prevalence of social entrepreneurial activity (table 9, model 16 and 18). Outcomes for hypothesis 2 are similar when using 2009 data, although significance is slightly lower (table 9, model 15).

Table 9: Robustness test using GEM 2009 data

Social entrepreneurial activity (2009) Model

16 (1) Model 17 (2) 18 (1+2) Model

Government size (2008) 0.043 0.023

(0.040) (0.041)

Government quality (2008) 1.772* 1.134*

(1.048) (0.636)

GDP per capita (log) (2008) 0.384 -2.315** -0.358

(0.574) (1.149) (0.698) GDP growth (2008) -0.154 -0.342 -0.122 (0.146) (0.231) (0.144) Constant -1.986 26.547** 5.489 (5.736) (11.287) (6.996) Observations 48 48 48 R-squared 0.160 0.130 0.218

(22)

21

Discussion

In this paper I have investigated how government size and government quality influence country prevalence rates of social entrepreneurial activity. In the following parts I will discuss my results and their implications for managers and theory. Furthermore, I will address some important limitations and provide potential future research directions.

Overall, the results indicate that the country prevalence rate of social entrepreneurial activity benefits from a higher quality government, while government size exerts no significant influence on it. Furthermore, the government quality sub-indicators control of corruption, regulatory quality, rule of law and voice and accountability have significant positive effects. I did not find a significant relationship for my proposed ideal configuration of government institutions for social entrepreneurship. These findings result in support for hypothesis 2a, 2b, 2c and 2d, but not for hypothesis 1 and 3, and contribute to the development of a more nuanced understanding of institutional void theory and institutional support theory.

The insignificant result for hypothesis 1 is not necessarily unexpected, it can be explained through several plausible causes. The most straightforward explanation is that the forces exerted by institutional support were stronger than expected with respect to social entrepreneurship in the government size dimension. Although I argued that institutional void theory had the upper hand in this dimension, this finding may signal both forces being at play in a more balanced manner, effectively cancelling out each other’s effects with regard to the overall effect of government size on social entrepreneurship.

(23)

22

available, which is GEM 2015. The 2009 dataset features a different set and smaller sample of countries. In this six-year difference formal institutions and the economic situation have changed. The findings based on GEM 2009 data may be affected by the financial crisis of 2007-2008 and recent developments like active support for social entrepreneurs by the European Union may have affected GEM 2015 data (Widuto, 2017). Nevertheless, my robustness test using 2009 data did not yield results significantly different from my main model. To conclude, Estrin et al. (2013) and Stephan et al. (2015) find contradictory support for institutional void theory and institutional support theory respectively, and my result may signal both factors being at play through finding a coefficient that lies between them.

I find support for hypothesis 2a. There is a significant positive relationship between prevalence rate of social entrepreneurship and government quality. This relationship mostly remains significant in the performed robustness checks. Social entrepreneurship thrives in countries where government institutions are of high quality, regardless of the size of these institutions. This finding is in line with institutional support theory but does not necessarily disprove institutional void theory. It does confirm that, even though institutional voids may increase demand for social provision, social entrepreneurs are only able to effectively respond to this demand as much as their government’s level of institutional quality allows them to. This result extends Chambers and Munemo’s (2017) finding that the prevalence of entrepreneurship is harmed by the absence of high-quality institutions by specifically analyzing social entrepreneurship. It also expands on Estrin et al. (2013), who found that a strong rule of law is beneficial for social entrepreneurship, by looking at a broader measure for institutional quality.

(24)

23

explained by country characteristics that are related to some perceived aspects of freedom. For example, Eikenberry et al. (2004) find that freedom of expression is important for the existence of nonprofit organizations, and my finding may extrapolate this earlier finding to social entrepreneurship. Another possibility for the significant relationship could be the high correlation of voice and accountability with other government quality indicators.

Hypothesis 3 is not supported. I did not find evidence for the proposed ideal institutional configuration for social entrepreneurs, consisting of a relatively smaller government of relatively higher quality. This insignificant result is related to the insignificant result of hypothesis 1, where I found a coefficient close to zero for the relationship between government size and social entrepreneurial activity. The forces implicated by institutional void theory are not strong enough to outweigh the countering force implicated by the institutional support theory in the government size dimension.

Theoretical implications

This study’s main contribution is to entrepreneurship research and institutional theory, by presenting a more nuanced view on institutional void theory and institutional support theory and a division between government dimensions in the context of social entrepreneurship. While earlier research would often see these theories as opposing (Stephan et al., 2015; Hoogendoorn, 2016), I proposed that a differentiation should be made within governmental institutional contexts regarding size and quality, simultaneously tolerating support for both institutional theories. Government institution quality is clearly beneficial to the prevalence of social entrepreneurship, which is in accordance with institutional support theory. The effects of government institution size are inconclusive as to which theory it supports regarding social entrepreneurship, potentially due to offsetting beneficial effects of institutional voids and institutional support.

Furthermore, this study adds to an ongoing stream of literature on the contextual determinants of social entrepreneurship, contributing to the exploration of factors that influence the prevalence of social entrepreneurship. Specifically, I have used a novel, comprehensive measurement of government quality and its effect on prevalence of social entrepreneurship, and added to the ongoing debate of the effect of government size by finding an inconclusive result based on data that is significantly more recent than earlier research.

(25)

24 Policy and managerial implications

Overall, my findings are more impactful for policymakers than for managers. Social entrepreneurs thrive in environments with high-quality government institutions. Policymakers aiming to enhance the prevalence of social entrepreneurship in their country should take governmental institutions into account and can achieve the steadiest results by putting their resources in improving the quality of those institutions. Cutting or increasing overall government expenditure as a way to make space for, or support social entrepreneurs respectively is a more doubtful practice for enhancing prevalence of social entrepreneurship.

Furthermore, this study encourages individuals or managers who are, or are considering becoming, social entrepreneurs to be acutely aware of how their country’s institutions may facilitate or hinder their process, and helps them act accordingly. This also applies to social entrepreneurs who are considering scaling across country borders. These individuals are advised to consider general government institutional quality, and also examine levels for control of corruption, regulatory quality and rule of law, while government size should be considered on a case-by-case basis.

Limitations and future research directions

Despite the contributions of this study, it is subject to a number of limitations that have to be considered before drawing definite conclusions.

Firstly, the sample of countries used is limited. Although I consider the sample to give a quite global picture as it includes countries from all major global regions, the distribution of factor-driven, efficiency-driven and innovation-driven economies is overrepresented by innovation-driven economies. This improper representation of the overall population may hurt external validity. Additionally, countries with factor-driven economies display significantly higher heterogeneity within their group compared to countries with efficiency-driven and innovation-driven economies (Bosma, 2016, p. 12, figure 2), resulting in higher uncertainty in the analysis. Future researchers investigating the prevalence of social entrepreneurship should consider weighing the sample in order to represent all economic development classifications proportionally. This advice holds especially with investigations focused on differences in prevalence across countries with different levels of economic development, e.g. when examining opportunity motives and necessity motives of social entrepreneurship.

(26)

25

by one year to reduce potential endogeneity. Future researchers may be able to make use of a future GEM dataset that also includes a special report on social entrepreneurship, and synthesize the 2009, 2015 and future dataset in order to improve explanatory power.

Thirdly, social entrepreneurship is still an ill-defined concept. Apart from discrepancy between academic definitions, the term may also be perceived differently across different people and countries. Potential local contingencies regarding the understanding of the phenomenon may be the cause of a difference in reported social entrepreneurship rates compared to actual rates in countries. Additionally, the broad type of social entrepreneurship is identified by only one question in the GEM study questionnaire, and the individual answers were aggregated to national-level data in the used dataset. These issues slightly hurt measurement validity and make the data prone to type 1 errors respectively. Furthermore, the GEM’s broad definition of social entrepreneurship was used. This definition does not exclusively measure social entrepreneurs who put their social objective above their financial objective, but also includes entrepreneurs who just have a particularly social objective. This may clash with some researchers’ proposed definitions of social entrepreneurship.

(27)

26

References

Aidis, R., Estrin, S., & Mickiewicz, T. M. (2012). Size matters: Entrepreneurial entry and government. Small Business Economics, 39(1), 119–139. https://doi.org/10.1007/s11187-010-9299-y

Anokhin, S., & Schulze, W. S. (2009). Entrepreneurship, innovation, and corruption. Journal of Business Venturing, 24(5), 465–476. https://doi.org/10.1016/j.jbusvent.2008.06.001

Baker, T., Gedajlovic, E., & Lubatkin, M. (2005). A framework for comparing entrepreneurship processes across nations. Journal of International Business Studies, 36(5), 492–504.

Beach, W. W., & Kane, T. (2008). Methodology: Measuring the 10 Economic Freedoms. 17.

Bosma, N., Schøtt, T., Terjesen, S. A., & Kew, P. (2016). Global Entrepreneurship Monitor 2015 to 2016: Special Topic Report on Social Entrepreneurship (SSRN Scholarly Paper ID 2786949). Social Science Research Network. https://papers.ssrn.com/abstract=2786949

Castaño, M.-S., Méndez, M.-T., & Galindo, M.-Á. (2015). The effect of social, cultural, and economic factors on entrepreneurship. Journal of Business Research, 68(7), 1496–1500. https://doi.org/10.1016/j.jbusres.2015.01.040

Castles, F. G., & Dowrick, S. (1990). The Impact of Government Spending Levels on Medium-Term Economic Growth in the Oecd, 1960-85. Journal of Theoretical Politics, 2(2), 173–204. https://doi.org/10.1177/0951692890002002003

Chambers, D., & Munemo, J. (2017). The Impact of Regulations and Institutional Quality on Entrepreneurship (SSRN Scholarly Paper ID 3066312). Social Science Research Network. https://papers.ssrn.com/abstract=3066312

Charron, N., Lapuente, V., & Rothstein, B. (2010). Measuring the quality of government and subnational variation. Report for the European Commission Directorate-General Regional Policy Directorate Policy Development. Brussels: European Commission.

Chowdhury, F., Audretsch, D. B., & Belitski, M. (2019). Institutions and Entrepreneurship Quality.

Entrepreneurship Theory and Practice, 43(1), 51–81.

https://doi.org/10.1177/1042258718780431

Dacin, P. A., Dacin, M. T., & Matear, M. (2010). Social Entrepreneurship: Why We Don’t Need a New Theory and How We Move Forward From Here. Academy of Management Perspectives, 24(3), 37–

57. https://doi.org/10.5465/amp.24.3.37

Dees, J. G. (1998). The meaning of social entrepreneurship. Kauffman Center for Entrepreneurial

Leadership.

https://centers.fuqua.duke.edu/case/wp-content/uploads/sites/7/2015/03/Article_Dees_MeaningofSocialEntrepreneurship_2001.pdf

Drucker, P. F. (1985). Innovation and entrepreneurship.

https://www.academia.edu/38623791/Innovation_and_entrepreneurship_-_Peter_F_Drucker Eikenberry, A. M., & Kluver, J. D. (2004). The Marketization of the Nonprofit Sector: Civil Society at

Risk? Public Administration Review, 64(2), 132–140. https://doi.org/10.1111/j.1540-6210.2004.00355.x

Estrin, S., Mickiewicz, T., & Stephan, U. (2013). Entrepreneurship, Social Capital, and Institutions: Social and Commercial Entrepreneurship Across Nations. Entrepreneurship Theory and Practice, 37(3), 479–504. https://doi.org/10.1111/etap.12019

(28)

27

Evans, P. (1996). Government action, social capital and development: Reviewing the evidence on synergy. World Development, 24(6), 1119–1132. https://doi.org/10.1016/0305-750X(96)00021-6

Ferri, E., & Urbano, D. (2011). Social entrepreneurship and environmental factors: A cross-country comparison. Research Work International Doctorate in Entrepreneurship and Business Management Department of Business Economics & Administration, Universitat Autonoma de Barcelona.

Fogel, K., Hawk, A., Morck, R., & Yeung, B. (2008). Institutional Obstacles to Entrepreneurship. The

Oxford Handbook of Entrepreneurship.

https://doi.org/10.1093/oxfordhb/9780199546992.003.0020 Harding, & Cowling. (2004). The GEM UK social entrepreneurship monitor.

Harding, R. (2004). Social Enterprise: The New Economic Engine? Business Strategy Review, 15(4), 39–

43. https://doi.org/10.1111/j.0955-6419.2004.00338.x

Harding, R. (2016). Understanding Social Entrepreneurship. Industry and Higher Education. http://journals-sagepub-com/doi/10.5367/000000007780222723

Hechavarría, D. M. (2016). The impact of culture on national prevalence rates of social and commercial entrepreneurship. International Entrepreneurship and Management Journal, 12(4), 1025–1052. https://doi.org/10.1007/s11365-015-0376-1

Heritage Foundation. (n.d.-a). Download Index of Economic Freedom Data, Maps and Book Chapters. Retrieved 30 December 2019, from //www.heritage.org/index/download

Heritage Foundation. (n.d.-b). Government Spending: Data, Economics, Growth, National Debt, Examples. Retrieved 30 December 2019, from //www.heritage.org/index/government-spending

Hoogendoorn, B. (2016). The Prevalence and Determinants of Social Entrepreneurship at the Macro Level. Journal of Small Business Management, 54, 278–296. https://doi.org/10.1111/jsbm.12301 Jenson, J. (2015). Social Innovation: Redesigning the Welfare Diamond. In A. Nicholls, J. Simon, & M.

Gabriel (Eds.), New Frontiers in Social Innovation Research (pp. 89–106). Palgrave Macmillan UK. https://doi.org/10.1057/9781137506801_5

Kaufmann, D., Kraay, A., & Mastruzzi, M. (2010). The Worldwide Governance Indicators: Methodology and Analytical Issues (SSRN Scholarly Paper ID 1682130). Social Science Research Network. https://papers.ssrn.com/abstract=1682130

Kibler, E., Salmivaara, V., Stenholm, P., & Terjesen, S. (2018). The evaluative legitimacy of social entrepreneurship in capitalist welfare systems. Journal of World Business, 53(6), 944–957. https://doi.org/10.1016/j.jwb.2018.08.002

Lepoutre, J., Justo, R., Terjesen, S., & Bosma, N. (2013). Designing a global standardized methodology for measuring social entrepreneurship activity: The Global Entrepreneurship Monitor social entrepreneurship study. Source: Small Business Economics, 40(3), 693–714. https://doi.org/10.1007/sl

Levie, J., & Hart, M. (2011). Business and social entrepreneurs in the UK: Gender, context and commitment. International Journal of Gender and Entrepreneurship, 3(3), 200–217. https://doi.org/10.1108/17566261111169304

(29)

28

Mair, J., & Marti, I. (2009). Entrepreneurship in and around institutional voids: A case study from Bangladesh. Journal of Business Venturing, 24(5), 419–435. https://doi.org/10.1016/j.jbusvent.2008.04.006

Martin, R. L., & Osberg, S. (2007). Social Entrepreneurship: The Case for Definition. https://ssir.org/articles/entry/social_entrepreneurship_the_case_for_definition

Mulgan, G. (2006). The Process of Social Innovation. Innovations: Technology, Governance, Globalization, 1(2), 145–162.

Nissan, E., Carrasco, I., & Castaño, M.-S. (2012). Women Entrepreneurship, Innovation, and Internationalization. In M.-A. Galindo & D. Ribeiro (Eds.), Women’s Entrepreneurship and Economics (pp. 125–142). Springer New York. https://doi.org/10.1007/978-1-4614-1293-9_9 Nissan, E., Castaño, M.-S., & Carrasco, I. (2012). Drivers of non-profit activity: A cross-country analysis.

Small Business Economics, 38(3), 303–320. https://doi.org/10.1007/s11187-010-9276-5

Nobel Media AB. (2006). The Nobel Peace Prize 2006. NobelPrize.Org. https://www.nobelprize.org/prizes/peace/2006/press-release/

North, D. C. (1991). Institutions. Journal of Economic Perspectives, 5(1), 97–112. https://doi.org/10.1257/jep.5.1.97

OECD. (2015a). General government—General government spending—OECD Data. TheOECD. http://data.oecd.org/gga/general-government-spending.htm

OECD. (2015b). Social Expenditure Database (SOCX) -. https://www.oecd.org/social/expenditure.htm Olsen, J. P. (2009). Change and continuity: An institutional approach to institutions of democratic

government. European Political Science Review, 1(1), 3–32. https://doi.org/10.1017/S1755773909000022

Puumalainen, K., Sjögrén, H., Syrjä, P., & Barraket, J. (2015). Comparing social entrepreneurship across nations: An exploratory study of institutional effects. Canadian Journal of Administrative Sciences / Revue Canadienne Des Sciences de l’Administration, 32(4), 276–287. https://doi.org/10.1002/cjas.1356

Reynolds, P. D. (2011). New Firm Creation: A Global Assessment of National, Contextual and Individual Factors. Foundations and Trends® in Entrepreneurship, 6(5–6), 315–496. https://doi.org/10.1561/0300000034

Rose-Ackerman, S. (1999). Corruption and Government: Causes, Consequences, and Reform (SSRN Scholarly Paper ID 189828). Social Science Research Network. https://papers.ssrn.com/abstract=189828

Roser, M., & Nagdy, M. (2013). Military Spending. Our World in Data. https://ourworldindata.org/military-spending

Schumpeter, J. (1943). Capitalism, Socialism and Democracy. https://eh.net/book_reviews/capitalism-socialism-and-democracy/

Schumpeter, J. A. (2002). New Translations from Theorie der wirtschaftlichen Entwicklung. American Journal of Economics and Sociology, 61(2), 405–437. https://doi.org/10.1111/1536-7150.00167 Shah, A. (2018, February 23). How to use asdoc: A basic example. StataProfessor.

https://fintechprofessor.com/2018/02/23/use-asdoc-basic-example/

(30)

29

Shorrock, A. (2008). Opening Address. United Nations University-WIDER Conference on Entrepreneurship and Economic Development, Helsinki.

StataCorp LLC. (n.d.). Stata | Why Stata? Retrieved 30 December 2019, from https://www.stata.com/why-use-stata/

Stephan, U., Uhlaner, L. M., & Stride, C. (2015). Institutions and social entrepreneurship: The role of institutional voids, institutional support, and institutional configurations. Journal of International Business Studies, 46, 308–331. https://doi.org/10.1057/jibs.2014.38

Storey, D. J., & Greene, F. J. (2010). Small business and entrepreneurship. Financial Times/Prentice Hall. http://sro.sussex.ac.uk/id/eprint/58723/

Sud, M., VanSandt, C. V., & Baugous, A. M. (2009). Social Entrepreneurship: The Role of Institutions. Journal of Business Ethics, 85(S1), 201–216. https://doi.org/10.1007/s10551-008-9939-1 Terjesen, S., Hessels, J., & Li, D. (2016). Comparative International Entrepreneurship. Journal of

Management, 42(1), 299–344. https://doi.org/10.1177/0149206313486259

The Economist. (2010, August 12). Social Innovation: Let’s hear those ideas. The Economist. https://www.economist.com/node/16789766/print

Wennekers, S., Stel, A., Carree, M., & Thurik, R. (2011). The Relationship between Entrepreneurship and Economic Development: Is it U-Shaped? http://www.nowpublishers.com/article/Details/ENT-023

Wennekers, S., van Wennekers, A., Thurik, R., & Reynolds, P. (2005). Nascent Entrepreneurship and the Level of Economic Development. Small Business Economics, 24(3), 293–309. https://doi.org/10.1007/s11187-005-1994-8

Widuto, A. (2017, March). EU support for social entrepreneurs—Think Tank. http://www.europarl.europa.eu/thinktank/en/document.html?reference=EPRS_BRI(2017)599 346

World Bank. (2015a). GDP per capita, PPP (current international $) | Data. https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD?view=chart

World Bank. (2015b). WGI 2019 Interactive > Documentation. https://datacatalog.worldbank.org/dataset/worldwide-governance-indicators

World Economic Forum. (n.d.). The Global Competitiveness Report 2014-2015. Global Competitiveness Report 2014-2015. Retrieved 12 January 2020, from http://wef.ch/1pxXIWS

Young, D. R. (2000). Alternative Models of Government-Nonprofit Sector Relations: Theoretical and International Perspectives. Nonprofit and Voluntary Sector Quarterly, 29(1), 149–172. https://doi.org/10.1177/0899764000291009

Young, D. R. (2008). A Unified Theory of Social Enterprise [Chapters]. Edward Elgar Publishing. https://econpapers.repec.org/bookchap/elgeechap/12543_5f9.htm

Yunus, M. (1998). Creating a World Without Poverty.

https://ssir.org/books/reviews/entry/review_creating_a_world_without_poverty

Yunus, M., & Weber, K. (2009). Creating a World Without Poverty: Social Business and the Future of Capitalism (Reprint edition). PublicAffairs.

(31)

30

(32)

31

Appendices

Table 10: Description of variables used in the regression models

Variables Description Source

Dependent variables

Total social entrepreneurial activity

Percentage of individuals aged 18-64 who are engaging in starting or owning-managing a business with a social purpose. People are considered social entrepreneurs if they answer yes to the question ‘Are you, alone or with others, currently trying to start or currently owning and managing any kind of activity, organization or initiative that has a particularly social, environmental or community objective?’ (Bosma et al., 2016, p. 5) GEM 2015 Adult Population Survey Narrow social entrepreneurial activity (robustness)

Percentage of individuals aged 18-64 who are engaging in starting or owning-managing a business that prioritizes social goals over financial goals and produces goods or services in the market. (Bosma et al., 2016, p. 5)

GEM 2015 Adult Population Survey

Government size General government consumption expenditure as a percentage of a country’s GDP Heritage Foundation Index of Economic Freedom Social provision (robustness)

Government social provision expenditure, including social benefits in areas such as health, education and unemployment, as a percentage of a country’s GDP

OECD Social Expenditure Database (SOCX)

Government quality Simple average of the six following governance indicators measuring formal institutional quality

World Bank World Governance Indicators (WGI)

Voice and accountability

Captures to what extent a country’s citizens perceive to have freedoms related to media, expression and association, along with being able to select their government

WGI

Political stability and absence of violence

Captures perceptions of the probability of political instability and violence motivated by politics

WGI

Government effectiveness

Reflects perceptions of the quality of a country’s public services and civil services, and how credible the government is in its commitments to these services

WGI

Regulatory quality Measures perceptions of how well governments are able to develop and carry out policies and