Self-employment and well-being across institutional contexts

Journal of Business Venturing 34 (2019) 105946

Contents lists available at ScienceDirect

Journal of Business Venturing journal homepage: www.elsevier.com/locate/jbusvent

Self-employment and well-being across institutional contexts☆,☆☆ Michael Fritscha, Alina Sorgnerb,c,d, Michael Wyrwiche,

⁎

T

a

Friedrich Schiller University Jena, Faculty of Economics and Business Administration, Carl-Zeiss Str. 3, 07743 Jena, Germany John Cabot University Rome, Via della Lungara 233, 00165 Rome, Italy c Research Fellow at the Kiel Institute for the World Economy (IfW Kiel), Germany d Research Affiliate at the Institute of Labor Economics (IZA Bonn), Germany e University of Groningen, Faculty of Economics and Business, Nettelbosje 2, 9747 AE Groningen, the Netherlands b

ARTICLE INFO

ABSTRACT

JEL classification: L26 I31 D01 D91 P51

This paper investigates whether the relationship between a person's occupational status and wellbeing differs across countries with varying institutional contexts. We find that the relationship between job and life satisfaction of self-employed people as well as of paid employees varies considerably across countries. Our results indicate that entrepreneurship-friendly institutions in a country are conducive to the well-being of those who are self-employed. Remarkably, the quality of entrepreneurial institutions also increases the levels of well-being of paid employees, but the effect is more pronounced for the self-employed.

Keywords: Entrepreneurship Institutions Well-being Life satisfaction Job satisfaction

Executive summary In many countries, creating institutional framework conditions that are more conducive to self-employment are a well-established objective of the policy agenda. Apart from manifold growth-oriented motivations for such policy initiatives trumpeting in favor of a more entrepreneurial society, the ultimate goal of such policies should be the well-being of individuals. While there seems to be general agreement that promotion of self-employment has a positive effect on the overall economic welfare of a society, much less is known about the potential impact of such policies on the well-being of individuals. Another important concern in this respect is whether more entrepreneurship-friendly institutions would enhance the well-being of self-employed individuals at the expense of those who prefer to remain in paid employment. This study investigates whether the relationship between a person's occupational status and his or her subjective well-being differs across countries with varying institutional contexts. The empirical analysis is conducted for 31 European countries that are

The responsibility for all conclusions drawn from the data lies entirely with the authors. Financial support for this paper's research was provided by the European Commission under the Horizon 2020 project “Financial and Institutional Reforms for the Entrepreneurial Society (FIRES), Grant Agreement Number 649378. We are indebted to Niels Bosma, Selin Dilli, Kyriakos Drivas, And-rea Herrmann, Catarina Matos and Mark Sanders for helpful comments on earlier versions of this study. Comments of three anonymous referees and the editor were of great help in improving the paper. ⁎ Corresponding author. E-mail addresses: [email protected] (M. Fritsch), [email protected] (A. Sorgner), [email protected] (M. Wyrwich). ☆

☆☆

https://doi.org/10.1016/j.jbusvent.2019.105946 Received 14 August 2018; Received in revised form 3 July 2019; Accepted 3 July 2019 Available online 23 July 2019 0883-9026/ © 2019 Elsevier Inc. All rights reserved.

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

characterized by highly heterogeneous entrepreneurial institutions. We use nationally representative data from the EU Statistics on Income and Living Conditions (EU-SILC) that are comparable across countries, as our main source of individual-level data. Well-being is measured by people's subjective assessment of their level of overall job and life satisfaction. We use the Global Entrepreneurship Index (GEI) as a measure of the entrepreneurship-facilitating quality of a country's institutional context. The GEI accounts for factors such as education, cultural support of entrepreneurship, availability of risk capital, innovation, and internationalization of the economy. It is primarily focused on crucial conditions for newly founded businesses, rather than for incumbent firms. The empirical analysis accounts for various other factors that may affect well-being at the individual level (e.g., age, qualification level, income, marital status), at the country level (e.g., the national wealth level, the unemployment rate, and the generosity of public welfare spending), as well as industry- and occupation-specific effects. Our findings show considerable cross-country variation in the level of subjective well-being, as measured by an individual's overall satisfaction with job and life. The more favorable institutions in a country are for entrepreneurship, the higher the job and life satisfaction of the self-employed. Remarkably, a higher quality of entrepreneurial institutions also increases the well-being of paid employees, but this effect is less pronounced than for the self-employed. Confirming the results of previous studies, our findings show that employees in small entrepreneurial firms tend to experience, on average, higher levels of job and life satisfaction when compared to employees in larger firms. We also find that the average job and life satisfaction of self-employed persons is lower than that of the paid employees in countries where the entrepreneurship-facilitating quality of the institutional framework is rather poor. Hence, the higher levels of well-being of the self-employed as compared to paid employees found in previous research cannot be regarded as a stylized fact. The reason behind our observation may be a high share of necessity entrepreneurship in these countries. The main policy lesson that can be drawn from our study is that promoting a more entrepreneurial society is not necessarily contrary to the interest of paid employees. Enhanced welfare in an entrepreneurial society does not seem to come at the expense of paid employees. The extent to which such policies are also beneficial for paid employees may, however, depend on the type of institutions. Further research should, therefore, seek to identify those types of entrepreneurial institutions that improve the wellbeing of both groups, the self-employed and paid employees. 1. Introduction Creating a more entrepreneurial society with entrepreneurship-friendly institutional framework conditions is now an important topic of discussion in the political dialogue (Audretsch and Thurik, 2001; Audretsch, 2007).1 It is, for example, an explicit goal on the policy agenda of the European Union (e.g., European Commission, 2010, 2013, 2016). A main motivation behind the attempts of creating a more entrepreneurial society and entrepreneurship-friendly institutions is the recognition that entrepreneurship can be an important driver of economic growth. In particular, entrepreneurship strengthens a country's or region's innovative capacity, may trigger growth processes, and can be of key importance for coping with the challenges of structural change (Schumpeter, 1934; Wennekers and Thurik, 1999; Fritsch, 2013). While there seems to be general agreement that promotion of self-employment has a positive effect on the overall economic welfare, much less is known about the potential impact of efforts to build a more entrepreneurial society and to create entrepreneurshipfacilitating institutions on the well-being of individuals which is another important policy objective. This paper wants to fill this research gap by exploring how entrepreneurship-facilitating institutional framework conditions and being either a paid employee or self-employed, as the main form of entrepreneurship, interact in determining individual well-being. The lack of previous work on this interaction between the country and individual level is puzzling from a conceptual point of view. On the one hand, there is extant evidence that entrepreneurship is positively affecting well-being (for an overview, see Croson and Minniti, 2012; Shir, 2016). On the other hand, institutional conditions can play a key role for the attractiveness of self-employment, and, hence, for the allocation of entrepreneurial talent that determines the supply of people in self-employment (e.g., Baumol, 1990; Sobel, 2008). Thus, if institutions determine the attractiveness of entrepreneurship then the satisfaction with being an entrepreneur should be lower in countries with entrepreneurship-inhibiting ruling institutions, as compared to countries with an entrepreneurship-facilitating institutional framework. In contrast, it is an open question whether entrepreneurship-friendly institutions benefit also regular paid employees but it is important to understand whether creating an entrepreneurial society comes at the expense of paid employees (Wiklund et al., 2019). Our main analysis is based on the Global Entrepreneurship Index (GEI) that indicates the degree to which institutions of a country are either entrepreneurship-facilitating or inhibiting (for details, see Acs et al., 2017). Well-being is measured by people's subjective assessment of their level of overall job and life satisfaction, which are important indicators of subjective well-being, the improvement of which is an ultimate objective of many policy initiatives (Diener and Tov, 2012). We find that the well-being of entrepreneurs is higher in countries with a high GEI score. This means that entrepreneurship-friendly institutions have a positive impact on the wellbeing of entrepreneurs. Another important key result of our study is that the entrepreneurship-facilitating quality of a country's institutions also increases the levels of well-being for paid employees. This effect is, however, less pronounced than for the selfemployed. To account for a possible effect of a person's income on well-being, we also investigate these relationships across different income quartiles. Based on these results we draw conclusions about the relationship between different types of entrepreneurship and individual well-being across institutional contexts. 1 Institutions can be generally defined as “the rules of the game” that govern the interaction of people in a society (e.g., North, 1994). It is common to distinguish between formal institutions that are understood as the set of codified rules, such as laws and constitutions, and informal institutions that comprise non-codified norms, conventions, codes of behavior, and the conduct of a society.

2

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

We find that the ratio of job and life satisfaction of the self-employed over paid employees shows pronounced variation that is clearly related to the entrepreneurship-facilitating character of the respective institutional framework. Hence, higher levels of wellbeing of the self-employed that have been found in several earlier studies (McGrath and MacMillan, 1992; Benz and Frey, 2008a, 2008b; Blanchflower, 2004; Croson and Minniti, 2012; Shir, 2016) cannot be regarded as a stylized fact. Rather, country-specific conditions play an important role in determining the actual level of well-being. The remainder of the paper is organized as follows: Section 2 discusses the link between institutional framework conditions and well-being in entrepreneurship in some more detail. The data and the empirical approach are introduced in Section 3, and Section 4 presents the results of the empirical analysis. Section 5 summarizes the main results, discusses implications for theory and policy, and identifies avenues for further research. The final section (Section 6) concludes. 2. Well-being in self-employment and paid employment and the role of institutions Several empirical studies show that the income of the self-employed is not necessarily or systematically higher than that of paid employees (e.g., Hamilton, 2000; Moskovitz, 2002; Sorgner et al., 2017). This result suggests that the choice of self-employment is not solely driven by income prospects but that non-pecuniary motivations, such as achieving higher procedural utility from more freedom of decision making, flexibility, and autonomy, can play an important role and even compensate for lower economic security and income (for an overview, see Croson and Minniti, 2012; Shir, 2016). Accordingly, self-employed people have often been found to experience higher levels of well-being in terms of work and life satisfaction due to higher degrees of self-determination and selfenhancement (e.g., Frey et al., 2004; Benz and Frey, 2008a, 2008b; Binder and Coad, 2013). To put it in the words of Schumpeter, this sense of well-being is achieved by trying to realize the “dream and the will to found a private kingdom” (Schumpeter, 1942, 93). In terms of values, this orientation can be understood as putting emphasis on autonomy or as a desire of individuals to independently pursue “their own ideas and intellectual directions” and aiming at” affectively positive experience” (Licht et al., 2007, 662). A central reason for this pattern discussed in the literature is based on the concept of procedural utility (Frey et al., 2004). Applied to the generation of economic value, this concept argues that people may draw utility not simply from the outcome of their work process but from the actual work process itself (Shir et al., 2018). Thus, pursuing one's own goals through self-employment can stimulate a feeling of self-determination, and may be a way of experiencing a high level of self-efficacy that is positively related to job and life satisfaction. A possible reason for the occurrence of higher levels of self-efficacy among self-employed people is that selfemployment may require more frequent goal setting and decision making than a position in paid employment.2 Quite a number of empirical studies have confirmed that self-employed people experience significantly higher levels of job satisfaction than paid employees.3 Other studies report slightly higher levels of life satisfaction for the self-employed than for paid employees, with the difference being less pronounced than for job satisfaction (Binder and Coad, 2013, 2016; Hessels et al., 2018). Benz and Frey (2008a, 2008b) in their empirical analyses identify several reasons for higher work satisfaction among the self-employed: a higher level of autonomy (“being one's own boss”) that makes work more interesting, and being able to use one's own initiative. Furthermore, they find that paid employees who work in small firms have higher levels of job satisfaction than those working in larger firms. The notion is that the more pronounced hierarchical structures of larger firms can impede personal autonomy. Prevailing institutions in a country might differently affect the well-being of the self-employed and paid-employed people due to a number of reasons. Assuming that a person's occupational choice (i.e., the decision of being self-employed or work as a paid employee) is governed by his or her subjective utility (e.g., Lucas, 1978; Holmes and Schmitz, 1990; Kihlstrom and Laffont, 1979), the relationship between institutions and well-being in different types of occupations may be approached by examining the effect of institutions on monetary and non-monetary returns (Elert et al., 2017). Many of the empirical studies that deal with the impact of formal institutions on entrepreneurship focus on the effects of entry barriers for transitions into self-employment.4 However, although entry barriers (such as the effort of registration, etc.) may have some effect on the start-up rate, particularly on entry of marginal entrepreneurs, they may not necessarily affect the well-being of those who already are in self-employment. The reason is that most of the entry barriers are only relevant at the time a business is first established. Hence, barriers to entry may cause effort and frustration during the start-up process but should not be pivotal for wellbeing after the business is set-up. The main areas of formal institutions affecting how business is conducted after entry include: the rule of law, protection of property rights, bankruptcy law, regulation of goods and service markets, taxation of profits and labor income, availability of finance, labor market regulation, as well as the organization of the social insurance system (Elert et al., 2017). Strict regulation of business conduct, like the regulation of goods and service markets, may have a permanent negative effect on well-being because it reduces the degree of economic freedom and the autonomy that self-employment can provide. A high transparency of rule of law and strong protection of property rights, in turn, may have positive effects on well-being because it reduces uncertainty in business conduct 2 Some psychologists argue that the fact that people pursue personal goals alone is positively related to well-being (Emmons, 1996). See Shir (2016) for a more detailed exposition. 3 See for example Benz and Frey (2008a, 2008b), Binder and Coad (2013), Blanchflower (2000, 2004), Block and Koellinger (2009), Coad and Binder (2014), Millán et al. (2013), van Praag et al. (2003). Hanglberger and Merz (2015) found that the difference in job satisfaction of the selfemployed and paid employees is the highest in the first years after self-employment, and it diminishes in the following years due to adaptation effects. For a comprehensive overview see Shir (2016). 4 E.g., Djankov et al. (2002); Fonseca et al. (2001, 2007); Klapper et al. (2006); Braunerhjelm and Eklund (2014).

3

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

(Elert et al., 2017). Labor market regulation is of key importance because it determines the availability of personnel and the conditions for hiring employees, such as protection against dismissal, maternity leave, etc. (Golpe et al., 2018; Herrmann, 2019). An obvious expectation in this respect is that an entrepreneur's well-being will improve when his or her employment decisions are unencumbered by regulation. Bankruptcy law may also affect the well-being of entrepreneurs. Insolvency regulations that include discharge clauses, the postponement of debt service and repayment, as well as the possibility of restructuring the business, should not only be conducive to the decision to start an own business, but also lessen a self-employed person's fear of suffering from financial hardships in case of bankruptcy (Armour, 2008). This should positively feed back into well-being. Regulations like the taxation of profits and labor income determine the monetary returns of their economic activity which may also affect the perceived well-being of entrepreneurs. The availability of external finance for investing in growth of a venture may particularly affect the well-being of growth-oriented entrepreneurs (Samila and Sorenson, 2011). Apart from the formal institutions represented by regulatory issues, the main types of informal institutions that may affect wellbeing of entrepreneurs are the social legitimacy of entrepreneurship (Kibler et al., 2014), or the presence of an entrepreneurial culture (Freytag and Thurik, 2007; Fritsch and Wyrwich, 2017). This includes the traditions and attitudes of the population towards selfemployment. If entrepreneurs are socially accepted and perceived as admirable role models, these informal institutions can be regarded as being entrepreneurship-facilitating. Moreover, the level of corruption may have negative effects on the well-being of a society as a whole, including self-employed and paid employees (Avnimelech et al., 2014; Chowdhury et al., 2018a, 2018b). Altogether, there are good reasons to assume that institutional framework conditions affect the well-being of self-employed. It is, however, unclear whether entrepreneurship-facilitating institutions are merely beneficial for the self-employed, or whether they also provide advantages for paid employees. It appears plausible to assume that rule of law, protection of property rights, a low level of corruption, a well-functioning financial system, appropriate regulation of markets for goods and services, as well as a sufficiently large and efficient innovation system are beneficial for people in both occupational categories. Conflicts might particularly arise because of labor market regulation where, for example, a lower level of employment protection may increase the well-being of the self-employed but come at the expense of paid employees who face a greater risk of being laid off. Likewise, low tax rates on profits are beneficial for those who are self-employed, but may require higher taxes on wages or value added that put a burden on paid employees. Summarizing these considerations, one may expect that more favorable institutional framework conditions for entrepreneurship may not necessarily benefit paid employees. This, therefore, remains an empirical matter that we aim to address in this study. Likewise, we can only speculate whether institutions are more important for job than for life satisfaction. Job satisfaction may be more affected because the mechanisms we describe deal primarily with work-related issues. Since satisfaction with one's own job partly determines life satisfaction, the latter should at least indirectly be affected – and to a lower degree - by the institutions we are discussing. 3. Data and empirical approach 3.1. Individual-level data and measurement of individual well-being Our source for individual-level data is the EU Statistics on Income and Living Conditions (EU-SILC). These data are the EU reference source for comparative statistics on income distribution and social exclusion at the European level.5 The EU-SILC provides comparable and high-quality cross-sectional data for 31 European countries.6 The reference population of the EU-SILC is all private households and their current members residing in the territory of the countries at the time of data collection. Persons living in collective households and in institutions (e.g., nursing homes) are generally excluded from the target population. Each year EU-SILC includes a different ad-hoc module in its survey program that provides additional information in selected realms. For this study, we use the cross-section from the year 2013, which includes an ad-hoc module on subjective well-being of individuals. The concept of subjective well-being refers to different types of evaluations that people make of their lives. Indicators of subjective well-being are generally considered as an important complement of more established economic indicators of well-being like income, for instance, in assessments of policy interventions (Kahneman and Krueger, 2006; Diener and Tov, 2012). We use two indicators of individual well-being that are available in the EU-SILC, namely, an assessment of an individual's current overall life and job satisfaction. Overall life satisfaction is a respondent's evaluation of his or her life as a whole. It intends to represent a broad, reflective appraisal a person makes of his or her life. It is by far the most frequently used and best validated concept of measuring well-being (Pavot and Diener, 2008). The variable refers to the respondent's feeling about the degree of satisfaction with his or her life in “these 5

For instance, EU-SILC is used in the context of the “Programme of Community Action to Encourage Co-Operation Between Member States to Combat Social Exclusion”, and for producing structural indicators on social cohesion for the annual Spring Report to the European Council. 6 Various regulations elaborated by the European Commission describe, e.g., sampling rules, list of target variables, and the mode of data collection of EU-SILC by national statistical offices of EU Member States and participating non-EU Member States. Eurostat requires all countries to collect a stratified random sample for EU-SILC to ensure its national representativeness. Furthermore, the list of target variables including type of the variable and their description are precisely prescribed. In the EU-SILC legal basis, different modes of data collection are allowed, but priority is given to face-to-face personal interviews. Proxy interviews, telephone interviews or a self-administration of the questionnaire are only recommended under special circumstances, such as illness of the respondent. For variables included in the module on well-being that we use in the empirical analysis only personal interviews are allowed. At the same time, Eurostat allows some flexibility in terms of data sources and sampling design. For instance, existing data sources and national sampling design can be used and procedures to maximize the response rates can be in accordance with each state's “best practices.” 4

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

days” rather than specifying a longer or shorter period.7 The level of life satisfaction is measured on an 11-point Likert scale, with the lowest value of 0 meaning “not at all satisfied” and the highest value of 10 meaning “completely satisfied”.8 The second variable of interest is a person's assessment of his or her level of job satisfaction, which is also measured at an 11-point Likert scale. It refers to the respondent's opinion about the current degree of satisfaction with his or her work for money, not the work someone does in the household of for recreation. If the respondent has several jobs, the answer about the level of job satisfaction refers to the primary job.9 While job satisfaction may be related to the overall life satisfaction, both metrics may reflect quite distinct aspects of individual well-being, which is the reason why we use both measures in the empirical analysis. While job satisfaction pertains to issues that are related to a person's work, life satisfaction is a much broader concept. Since satisfaction with work is a key element of someone's life satisfaction, there should be a positive correlation between the two types of assessment. However, high work satisfaction might also have negative effects on life satisfaction. For example, a rather satisfying job with high emotional engagement and long working hours could crowd out other activities that are important for life satisfaction, such as satisfying social relationships and good health. For this reason, the correlation between the two concepts may be quite low or even negative.10 Self-employed individuals are identified in the EU-SILC based on their self-reported current labor market status. The indicator includes self-employed persons that work full-time or part-time to earn a profit. We construct a binary variable that equals 1 if a person is regarded as self-employed, and it is 0 if a respondent is a paid employee. Paid employees are defined as persons who work for an employer and who receive compensation, for instance, in the form of wages or salaries. Unemployed, non-employed persons, respondents currently in fulltime education, those in compulsory military community or service, and home workers are not considered in our analysis. It has been shown that the levels of job and life satisfaction that someone experiences in a certain type of occupation—selfemployment or paid employment—varies according to her or his individual characteristics (see Shir, 2016, for an overview). The EUSILC includes a set of socio-demographic variables such as age, gender, and marital status that we use as control variables in our analysis. Furthermore, we use the information about the highest level of education (defined according to the ISCED classification),11 occupation (defined at a 2-digits level of ISCO-08),12 industry sector (according to the NACE rev.2),13 the number of hours usually worked per week in the main occupation, and information on change of job in the previous year. We further account for a person's financial situation, because this may significantly affect the level of individual well-being. The EU-SILC contains information on gross monetary income of paid employees and gross monetary income or losses for self-employed persons during a previous 12-month period (such as the previous calendar or tax year) in national currency.14 We construct country-specific income quartiles to make the income measure comparable between countries.15 Since health status is an important determinant of the overall life satisfaction (van Praag et al., 2003; Binder and Coad, 2013), we include self-reported information on a person's current health condition.16 The final sample contains a total of 158,463 observations.17 Table A1 in the Appendix shows the correlations of variables used in the analysis and Table A2 provides descriptive statistics. 3.2. Country-level data and measurement of institutional contexts 3.2.1. Variables representing the entrepreneurship-facilitating quality of institutions To investigate the variation of contexts we use a metric that indicates the entrepreneurship-facilitating context of a country's institutions: the Global Entrepreneurship Index (GEI). The GEI is a comprehensive measure for the quality of a country's entrepreneurship ecosystem, and accounts for factors such as education, cultural support of entrepreneurship, availability of risk capital, innovation, and internationalization (for details, see Acs et al., 2017). The GEI is focused more on the conditions for newly founded businesses than for incumbent firms. The indicators of the GEI are based on national statistics, on individual assessments of 7 Although the measure of life satisfaction is related to happiness, it differs in the sense that responses to the question regarding a person's life satisfaction tend to be considerably more stable over time and less influenced by momentary incidences (Lucas et al., 1996; Diener et al., 2013). 8 The precise formulation in the questionnaire is as follows: “The following question asks how satisfied you feel, on a scale from 0 to 10. Zero means you feel ‘not at all satisfied’ and 10 means you feel ‘completely satisfied’.” The question than is: “Overall, how satisfied are you with life as a whole these days?”; OECD (2013). This type of question is well-established in empirical research on well-being and it has been shown that responses have a high level of validity; see Diener et al., 2013). 9 The question is “How satisfied are you with your job?”; OECD (2013). 10 Our data reveal that the correlation coefficient between job and life satisfaction is 0.48. 11 The International Standard Classification of Education (ISCED) has been developed by the United Nations Educational, Scientific and Cultural Organization (UNESCO) and provides internationally comparable education statistics. We distinguish between primary education, secondary education, and tertiary education in our analysis. 12 The International Standard Classification of Occupations (ISCO) provided by the International Labour Organization is used by Eurostat to provide internationally comparable information on occupational participation. 13 The statistical classification of economic activities (NACE; Nomenclature Statistique des Activités Économiques dans la Communauté Européenne) is employed by Eurostat to provide internationally comparable information on participation in industrial sectors. 14 In Ireland, the survey is continuous, and indication of income refers to the last twelve months. 15 The only available information about wealth is about homeownership of one of the household members (whose occupational status is not identified). Adding the variable “homeownership of one of the household members (yes/no)” to the empirical models leads to a significantly positive coefficient but leaves the basic results unaffected. 16 Health status is measured on a 5-point ordinal scale ranging from 1 (very bad) to 5 (very good). 17 The number of observations is 157,186 in models with life satisfaction as the dependent variable due to missing values on the variable “health status”, which is a covariate in these models.

5

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

representative population samples, and on assessments made by experts. The higher the value of this index, the more entrepreneurship-facilitating the institutions and economic conditions in a country are supposed to be. The GEI score is determined by 14 pillars, which represent country-specific conditions that can be assumed relevant for entrepreneurship. The pillars include measures for opportunity perception, start-up skills, non-fear of failure, networking, cultural support, prevalence of opportunity start-ups, characteristics of the tech sector, quality of human resources, degree of competition, product innovation, process innovation, high growth ambitions, internationalization activities, and availability of risk capital. The scores are an outcome of the underlying institutional framework conditions and of policy measures within this institutional framework. The respective pillars are summarized in Table A3. The GEI scores for most of the countries in the sample refer to the year 2013.18 An alternative indicator of the quality of national institutions regarding the conditions for entrepreneurship is the Doing of Business (DoB) score provided by the World Bank (2013). The DoB score assesses the regulatory performance of 189 countries, including those in our sample. It covers diverse areas that are relevant for self-employment such as the procedures, time and cost of starting a business, dealing with construction permits, registering property, enforcing contracts, resolving insolvency, as well as the total tax rate on profits. In contrast to the GEI, the DoB index19 relies less on subjective assessments of experts and the surveyed population, but is rather more based on hard facts. Moreover, the DoB index is not particularly focused on start-ups, but assesses the general business-friendliness of a country for all types of firms. We use the GEI scores as our main measure of entrepreneurshipfacilitating quality of institutions and employ the DoB score for a robustness check. The two indicators for the quality of entrepreneurship-facilitating institutions are in accordance with the general arguments made in the literature on varieties of capitalism (Dilli and Niklas, 2018; Herrmann, 2019). Dilli and Niklas (2018) distinguish between the following varieties of entrepreneurial capitalism: (i) Liberal market economies (including Anglo-Saxon economies), (ii) Coordinated market economies (including Continental and Northern European economies), (iii) Mediterranean market economies, and (iv) Eastern European market economies. This classification is based on country differences with regard to institutions governing finance, the labor market, education, as well as inter-organizational relationships. Liberal market economies built mainly on market forces as the key coordination mechanism and have relatively low levels of regulation that restrain entrepreneurial initiative. In contrast, coordinated market economies are characterized by a considerably larger role of public intervention and more restrictions of private initiative. Mediterranean market economies provide relatively uneasy conditions for entrepreneurship with comparatively low levels of spending for education and R&D, low levels of adherence to formal rules but high levels of corruption and a non-transparent bureaucracy. Eastern European market economies are particularly suffering from an often still incomplete transformation from a socialist regime and low levels of political support for entrepreneurial activity. Table 1 clearly suggests that the variability of the GEI and the DoB scores is in line with the concept of varieties of entrepreneurial capitalism. This table is restricted to countries that are both part of the EU-SILC data and the analysis by Dilli and Niklas (2018). It is clear from the table that both the GEI and the DoB scores are highest for the liberal market economies and the coordinated market economies. The average scores of both indicators are lowest for the Mediterranean and the Eastern European market economies. 3.2.2. Country-level control variables Since there might be many confounding factors that could affect the well-being of entrepreneurs and paid employees, we also consider a battery of additional country-level control variables that are available for the year 2013. It might be argued, for instance, that the entrepreneurship-facilitating character of a country's institutional framework on individual well-being may be confounded by a relationship between the institutional framework and the welfare level. In particular, one may assume that countries that have highquality entrepreneurial institutions may also enjoy higher levels of economic welfare, and that this effect may drive our results. Thus, we use the measure of gross domestic product (GDP) per capita to account for differences in economic welfare across countries. In addition, the level of entrepreneurship in a country and individual well-being may be influenced by the level of unemployment in that country. For instance, high unemployment rates may lead to higher levels of necessity entrepreneurship by individuals who are pushed into self-employment because of a lack of alternative employment opportunities. In this case, it can be expected that well-being of such entrepreneurs will be lower. Hence, we control for country-specific unemployment rates in our analysis. The data on unemployment rates come from Eurostat for all countries except of Serbia and Switzerland, for which the data from national statistical offices of these countries was used. Similarly, a high degree of income inequality in a country may lead to more entrepreneurship out of necessity, thus, negatively affecting individual well-being. We use the Gini index provided by the World Bank to account for this issue. The Gini index measures the extent to which the distribution of income among individuals or households within an economy deviates from a perfectly equal distribution. A Gini index of 0 represents perfect equality, while an index of 100 implies perfect inequality. Countries also differ substantially with regard to their social welfare systems. One may expect that individuals in countries that provide generous welfare benefits are less likely to pursue self-employment out of economic desperation or lack of alternative employment opportunities. Consequently, people in these countries may be pulled into self-employment primarily by their desire to 18 For several countries the GEI scores were not available for 2013 and, thus, they were taken from an available wave closest to 2013. This was the case for Iceland (2010), Denmark and Austria (2012), Serbia (2009), Bulgaria (2015), and Cyprus (2017). The results of our analysis are robust to the exclusion of countries, for which the GEI scores were not available before 2013. 19 The DoB index measures the distance of each country to the ‘frontier,’ which represents the best performance observed across all countries in the sample since 2005. A country's distance to the frontier is reflected on a scale from 0 to 100, where 0 signifies the lowest performance and 100 represents the frontier. For example, a score of 75 means that a country was 25 percentage points away from the frontier, constructed from the best performances across all countries and across time.

6

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table 1 Varieties of entrepreneurial capitalism, GEI, and DoB. Varieties of entrepreneurial capitalism

Global entrepreneurship index (GEI)

Doing of Business Index (DoB)

Liberal market economies Coordinated market economies

67.43 65.38

84.86 78.49

Mediterranean market economies Eastern European market economies

48.37 41.96

69.19 70.5

Countries available in EU-SILC Ireland, UK Denmark, Finland, Sweden, Norway, Austria, Germany, Netherlands, Switzerland, Belgium France, Italy, Portugal, Spain Czech Republic, Hungary, Poland, Slovakia, Slovenia

Notes: The classification only considers countries that were considered in the analysis by Dilli and Niklas (2018) and are also part of the EU SILC data. Scores are weighted by country-specific population numbers. High DoB scores indicate closeness to the frontier that is defined as the best performance across countries and over time. A score of 100 indicates the frontier.

pursue a promising entrepreneurial opportunity or out of self-realization motives. Thus, it can be expected that they will likely have higher levels of individual well-being. To account for differences in the levels of social protection across countries we use data on social welfare spending provided by Eurostat as part of the European System of Integrated Protection Statistics (ESSPROS). In detail, we use a measure of total social protection expenditure as a percentage of GDP that includes social benefits relating, for instance, to unemployment, health care, old age, disability, and family (Eurostat, 2016). Moreover, we employ the Human Development Index (HDI), which is used, for instance, to measure a country's development by the United Nations Development Program. HDI is a summary measure of three dimensions: life expectancy at birth, education, and standard of living measured by gross national income per capita (Anand and Sen, 1994). 3.3. Empirical strategy To explore the relationship between individual well-being, employment status, and the entrepreneurship-facilitating quality of the institutional context, we estimate ordered logit models that account for the ordinal nature of our dependent variables, the individual levels of job and life satisfaction. Since the dependent variables are defined at the level of individuals across countries, observations within countries might be correlated. Hence, we report standard errors clustered at the country level in all regressions. In order to understand whether the entrepreneurial institutional context differentially affects people who are in self-employment and in paid employment, we interact our measure of the quality of entrepreneurial institutions, the GEI, with an individual's current employment status. We use the estimated coefficients to calculate marginal effects of the measure of the quality of entrepreneurial institutions on the probability to be completely satisfied with one's job and life for both employment states at the different levels of the GEI and the mean values of the control variables that have been introduced in Sections 3.1 and 3.2. It should be noted that reverse causality related to the selection of more satisfied people into self-employment cannot be ruled out in our context. However, it is not a severe issue since we are not primarily interested in the effect of occupational status on well-being but on the differential effect of institutions on well-being of self-employed and paid employees. Institutions are the same for selfemployed and paid-employees within a country which rules out selection into certain institutions. Our interaction approach captures the differential effect of institutions on well-being while taking occupational choice as given.20 The interaction approach accounts for the multilevel nature of our setting. Since we are interested in assessing to what extent country-level entrepreneurial institutions influence individual-level measures of well-being, multilevel modeling might be an appropriate alternative estimation strategy (see, e.g., Kim and Karl, 2016; Shepherd, 2011). Therefore, we run respective robustness checks even though we find hardly an indication that there is a need to use the multilevel approach (for details, see Section 4.3). 4. Results of the empirical analysis 4.1. Descriptive statistics Figs. 1 and 2 show average levels of job and life satisfaction by countries and employment status. In the left-hand part of the figures (the first four segments), we refer to the countries mentioned in Table 1 that are also considered in the assessment of varieties of entrepreneurial capitalism by Dilli and Niklas (2018). In the right-hand part of the figures (the fifth segment), we show other countries that we additionally consider in our empirical analysis. We find considerable variation in the levels of job and life satisfaction across countries representing varieties of entrepreneurial capitalism. People living in the countries belonging to liberal and coordinated market economies tend to experience, on average, higher levels of job and life satisfaction than residents of Mediterranean and many Eastern European states, also when compared to countries from these regions that were not considered in the 20 Reverse causality in our empirical framework is an issue if the interaction effects are driven by a pattern that happier self-employed are more successful in lobbying for more entrepreneurship-friendly institutions as compared to less happy self-employed. Similarly, reverse causality is an issue if the main effect of institutions is driven by happier employees being more successful in lobbying for entrepreneurship-friendly institutions as compared to unhappy employees. We have no firm argument why there should be such systematic patterns.

7

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Fig. 1. Job satisfaction by country and employment status. Countries are sorted within groups by descending order of the GEI level. 95% confidence intervals are reported.

varieties of entrepreneurial capitalism approach by Dilli and Niklas (2018). Moreover, there seems to be pronounced correlation between the level of individual well-being and the quality of entrepreneurship-facilitating institutions, as measured by the GEI, both within and between the groups of countries that constitute varieties of entrepreneurial capitalism (see Figs. B1 and B2 in the online Appendix). The correlation coefficient between GEI scores and individual job satisfaction is 0.17 (0.31 for self-employed and 0.14 for paid employed individuals). The correlation coefficient between GEI and individual life satisfaction is 0.28 (0.32 for self-employed and 0.27 for paid employed individuals). Also remarkable are the differences in the levels of job and life satisfaction between the self-employed and paid employees across countries. Previous studies often provided evidence of self-employed people being on average more satisfied with their jobs as

Fig. 2. Life satisfaction by country and employment status. Countries are sorted within groups by descending order of the GEI level. 95% confidence intervals are reported. 8

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

compared to their paid employed counterparts (McGrath and MacMillan, 1992; Blanchflower, 2004; Benz and Frey, 2008a, 2008b; Croson and Minniti, 2012; Shir, 2016) or as Blanchflower (2004, 62) puts it “self-employed overall are happier.” Our data do, however, suggest that these differences in the level of job satisfaction vary strongly with the institutional context. In particular, selfemployed people are on average at least as likely to be satisfied with their jobs as paid employees in liberal and coordinated market economies (Fig. 1). This is very different for some of the Mediterranean and Eastern European market economies as well as other countries that are not considered by Dilli and Niklas (2018). Within the Baltic States, the level of job satisfaction is much higher for the self-employed than for paid employees only in Estonia while it is slightly lower in Lithuania and Latvia. In turn, self-employed people report on average significantly lower levels of job satisfaction than paid employed people in all Mediterranean countries, with an exception of France, while the difference is highest in Portugal, Cyprus, and Greece. Among Eastern European countries, we observe the strongest differences in the levels of job satisfaction for both employment types in Romania and Serbia. Only self-employed in Czech Republic, Slovakia, and Hungary report slightly higher levels of job satisfaction than their paid employed counterparts. The differences in the level of overall life satisfaction between the self-employed and paid employees also vary between countries, although to a lesser degree than the levels of job satisfaction (Fig. 2). For instance, we do not find any statistically significant differences for liberal and coordinated market economies, as well as for the Baltic States. Self-employed individuals in Mediterranean countries and in some Eastern European countries are, on average, significantly less satisfied with their lives as compared to paid employees. In sum, the descriptive evidence suggests that the differences in job satisfaction between self-employed individuals and paid employees are more pronounced than for the overall life satisfaction. Moreover, there are clear differences across countries and certain institutional contexts, as defined by the varieties of capitalism approach. 4.2. Results of the multivariate analysis The main results of the ordered logit estimations are presented in Table 2. Four separate models are shown for each dependent variable, job satisfaction (models I–IV) and life satisfaction (models V–VIII). Models I and V show the relationship between self-employment status and individual well-being, models II and VI include the GEI variable, while models III and VII account for an interaction between the selfemployment status and the GEI. In addition, models IV and VIII repeat the models with interaction effects but do not include country-level control variables to understand whether and to what extent they might impact the effect of the GEI on individual well-being. Our models show that self-employed persons are not less likely to be satisfied with their jobs and lives than paid employed persons, conditional on a variety of individual- and country-level control variables (models I and V). We also find a statistically significant and positive direct effect of the GEI scores on individual job and life satisfaction (models II and VI). This suggests that a higher quality of entrepreneurship-facilitating institutions is associated with higher individual well-being. To examine whether this is more likely to be the case for self-employed persons than for paid employed persons, we include interaction terms between employment status and the GEI into the model.21 Both the interaction variable and the main effect of the GEI are found to be statistically significant and positive (models III and VII). While the interaction variable captures the differences in the effect size between self-employed and paid employed individuals, the main effect of the GEI indicates the impact of the GEI on job and life satisfaction of paid employees (for details of interpretation of dummycontinuous interactions, see Brambor et al., 2006). Thus, our results indicate that the effect of GEI on job and life satisfaction of paid employed and self-employed persons is positive, but it is significantly higher for self-employed than for paid employed individuals. Figs. 3 and 4 visualize differences between the self-employed and paid employees regarding the predicted probability of being completely satisfied with one's job and life conditional on a country's GEI score and all control variables being kept at their mean values. These estimations are based on Models III and VII in Table 2. The figures depict on the x-axis the range of the GEI scores from the lowest score of 20 (corresponds approximately to the value for Bulgaria) to the highest score of 80 (corresponds approximately to the value for Denmark). We find that the probability of being completely satisfied with one's job is lower for self-employed individuals than for paid employed persons in countries where the GEI score is relatively low, although this difference is not statistically significant. For instance, for a country with a GEI score of 20, the probability for a self-employed person of being completely satisfied with his or her job is 4.3% as compared to 7.9% for a paid employed counterpart. The probability of being completely satisfied with one's job is, however, significantly higher for the self-employed in countries with a GEI score of about 60 points and more, which is the case for all liberal and coordinated market economies (see Fig. 1). For instance, for a country with a GEI score of 80, the probability of a self-employed person being completely satisfied with his or her job is 33.7%, which is almost twice as high as compared to 15.8% for paid employed persons. Thus, job satisfaction of both employment types increases with an increasing level of the quality of entrepreneurial institutions, as measured by the GEI, although the increase is much more pronounced for the self-employed individuals. This also holds true for the probability of being completely satisfied with one's life (Fig. 4). For instance, the probability of being completely satisfied with one's life is 3.8% (20%) for a self-employed person and 4.6% (14.8%) for a paid employed person in a country with the lowest (the highest) GEI score. However, the difference between the degree of life satisfaction for the self-employed and paid employees is not statistically significant. At the same time, these findings suggest that entrepreneurship-friendly institutions matter more for job than for life satisfaction. The latter is certainly affected by other domains than only work whereas job satisfaction is certainly determined primarily by occupational choice. Furthermore, a 21 Note that the constitutive terms of the interactions indicate the effect for the case where the other interacting variable has the value of zero (Brambor et al., 2006). Hence, the coefficients for the constitutive term can hardly be interpreted in isolation in a meaningful way.

9

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table 2 Individual well-being, self-employment, and the quality of entrepreneurial institutions. Job satisfaction I Paid employed (reference) Self-employed GEI Self-employed × GEI Country-specific controls Industry fixed effects Occupation fixed effects Number of observations Pseudo R-squared Log likelihood

Life satisfaction II

– 0.136 (0.095) – –

– 0.163* (0.093) 0.017** (0.007) –

Yes*** Yes*** Yes*** 158,463 0.0194 −310,920.032

Yes*** Yes*** Yes*** 158,463 0.0202 −310,642.79

III

IV

V

– −1.201*** (0.2578) 0.0133* (0.0068) 0.0278*** (0.0043) Yes*** Yes*** Yes*** 158,463 0.0216 −310,225.42

– −1.322*** (0.270) 0.016*** (0.004) 0.029*** (0.004) No Yes*** Yes*** 158,463 0.0197 −310,827.41

VI – 0.051 (0.056) – –

– 0.087* (0.049) 0.024** (0.009) –

Yes*** Yes*** Yes*** 157,185 0.0558 −285,128.82

Yes*** Yes*** Yes*** 157,185 0.0575 −284,624.38

VII

VIII

– −0.383** (0.173) 0.022** (0.010) 0.0096*** (0.003) Yes*** Yes*** Yes*** 157,185 0.0576 −284,574.95

– −0.451** (0.181) 0.034*** (0.006) 0.010*** (0.003) No Yes*** Yes*** 157,185 0.0558 −285,123.64

0

.1

Predicted probability .2 .3

.4

Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job satisfaction (models I-IV) and life satisfaction (models VVIII). Standard errors clustered on the country level in parentheses. ***: statistically significant at the 1% level; ** statistically significant at the 5% level; * statistically significant at the 10% level (two sided tests). Control variables are included. Effects for control variables are reported in Table A4 in the Appendix. (The results for control variables indicate that older people and males report lower levels of well-being, while being married has a positive effect. The number of working hours per week and a change of occupation in the previous year are negatively related to overall life satisfaction, but this relationship is not statistically significant for job satisfaction. Both, job satisfaction and overall life satisfaction increase with a person's position in the income distribution. People with higher education levels tend to report higher levels of life satisfaction, while the relationship between educational level and job satisfaction comes out to be negative, which is in line with previous studies (e.g., Clark and Oswald, 1996; Millán et al., 2013). In an attempt to explain this latter result, Clark and Oswald (1996) speculate that higher education induces higher aspirations for characterizing one's situation as “good” or “excellent” that are then not fulfilled in reality. Millán et al. (2013, 665) suggest “that employees with university studies have more demanding jobs and have to meet higher expectations, and thus keeping one's job is more challenging.”)

20

30

40

50 GEI score

Paid employed

60

70

80

Self-employed

Fig. 3. Predicted probability of being completely satisfied with one's job by employment status and the GEI score. 95% confidence intervals are reported.

demanding job may negatively affect work-life-balance, thus, reducing the overall effect of institutions on life satisfaction. Our main estimations also reveal interesting details about the effects of country-level control variables on individual well-being (see Table A4 in the Appendix). We find a significantly negative association between higher social protection expenditures (in % of GDP) and individual job satisfaction. This negative relationship may be explained by the fact that generous social benefits could decrease the additional utility that individuals gain when they decide to work instead of being, for instance, unemployed. In addition, generous social benefits may go along with a higher tax burden placed upon the working population.22 Also, we find that high unemployment rates and strong income inequality, as measured by the Gini index, decrease individual well-being, although these effects are not statistically significant.

22 At the same time, better social protection helps individuals to maintain a high living standard in the event of losing a job or facing other difficulties, thus, contributing to a higher overall life satisfaction. In accordance with this speculation we find a positive relationship between the national level of social protection expenditure and individual level satisfaction that is, however, not statistically significant.

10

Journal of Business Venturing 34 (2019) 105946

0

Predicted probability .1 .2

.3

M. Fritsch, et al.

20

30

40

50 GEI score

Paid employed

60

70

80

Self-employed

Fig. 4. Predicted probability of being completely satisfied with one's life by employment status and the GEI score. 95% confidence intervals are reported.

Although including country-level control variables in the model is indispensable if one wants to ensure that the effect of entrepreneurship-facilitating institutional environment is not confounded by other regional factors, this may introduce a multicollinearity issue.23 Hence, we re-estimate the models with interaction terms and exclude country-level control variables. We find, however, that the effect of the GEI remains almost unchanged in these model specifications, although it is now more significant (Models IV and VIII in Table 2). To further investigate a potential multicollinearity issue, we analyze how the results of our main model estimation with interaction terms between the GEI and employment status change when we introduce the following modifications by including: i) no country-level control variables; ii) each of the country-level control variables separately; iii) all country-level control variables simultaneously. The estimations of these models for job satisfaction are reported in Table B1 and for life satisfaction in Table B2 in the online Appendix. An important finding is that the effects of variables of interest remain virtually unaffected by the above-mentioned variations introduced to our main model. This indicates a robust positive effect of GEI on individual satisfaction with job and life and its stronger relevance for self-employed individuals than for persons in paid employment. At the same time, these estimations also show that social protection expenditure is another important factor influencing individual well-being, which is the only variable to remain statistically significant in different model specifications. Thus, it deserves a more detailed analysis. Social protection expenditure is arguably a potential confounding factor in our analysis of individual wellbeing in the context of different entrepreneurship-facilitating qualities of the institutional environment. Generous social welfare benefits may largely prevent necessity-driven entrepreneurship, thus, increasing the well-being of both paid employed persons and self-employed persons who are likely to set up opportunity-driven businesses. Indeed, a correlation coefficient between GEI and social protection expenditure is about 56%, indicating that countries with more entrepreneurship-friendly institutions are more likely to have higher social protection expenditures. To test if this confounds our main result, we perform a “placebo” test by replacing the GEI variable in our main models with interaction effects by the variable measuring social protection expenditures. Should the interaction between the level of social welfare spending and self-employment have a positive effect on well-being, this would be a strong indicator that our main result is confounded by other country-level factors. The results of this test are reported in Table 3, columns I and III, and show no significant effect of social protection expenditure and of the interaction between social protection expenditure and employment status. Models II and V additionally control for GEI. However, the interaction effect between self-employment status and social protection expenditures remains insignificant. At the same time, these estimations show that social protection expenditures have a significant and negative effect on job satisfaction and no significant effect on life satisfaction of paid employees, which is contrary to a positive and significant effect of GEI in our baseline model. In sum, these results suggest that our results for GEI are not driven by a positive correlation between GEI and social protection expenditure. Moreover, even if a country provides generous social benefits, thus, potentially reducing the level of necessity entrepreneurship, individual well-being of self-employed persons may still strongly depend on the quality of entrepreneurship-facilitating institutions. In turn, a country with low social protection expenditures may still enjoy relatively favorable entrepreneurial institutions, thus, not inducing an additional burden on people who become self-employed because of the lack of alternative opportunities in paid employment. To investigate how entrepreneurship-facilitating institutions influence individual well-being in countries with different levels of social welfare spending, we construct a dummy variable indicating a high level of social protection expenditure in a country, which equals 1 if a country in our sample belongs to the upper quartile of the distribution of social protection expenditures, and it equals 0, otherwise. It is worth mentioning that the group of countries with high social protection expenditures is highly heterogeneous

23

Indeed, there are non-zero correlation coefficients between GEI and country-level control variables (see Table A1 in the Appendix). 11

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table 3 Individual well-being, social protection expenditures, and GEI. Job satisfaction I Paid employed (reference) Self-employed Social protection expenditures, % GDP High social protection expenditures (yes = 1, no = 0) GEI Self-employed × social protection expenditures Self-employed × high social protection expenditures (yes = 1, no = 0) Self-employed × GEI

Life satisfaction

II

III

– −0.372 (0.4679) 0.00748 (0.0102) –

– −0.414 (0.4083) −0.0273⁎⁎ (0.0114) –

–

0.0253⁎⁎⁎ (0.0048) 0.0223 (0.0163)

0.0167 (0.0196) – –

–

High social protection expenditures × GEI

–

–

Self-employed × GEI × high social protection expenditures (yes = 1, no = 0) Country-specific controls Industry fixed effects Occupation fixed effects Number of observations Pseudo R-squared Log likelihood

–

–

Yes*** Yes*** Yes*** 158,463 0.0129 −312,984.6

Yes*** Yes*** Yes*** 158,463 0.0193 −310,943.5

IV

– −1.472⁎⁎⁎ (0.3365) – −0.129 (0.3515) 0.0161⁎⁎⁎ (0.0062) – 0.624 (0.3706) 0.0320⁎⁎⁎ (0.0058) 0.00170 (0.0072) −0.0104 (0.0066) Yes*** Yes*** Yes*** 158,463 0.0197 −310,809.5 ⁎

V

VI

– 0.201 (0.3683) 0.0367⁎⁎ (0.0155) –

– 0.153 (0.2529) −0.0159 (0.0188) –

– −0.0108 (0.0158) –

0.0389⁎⁎⁎ (0.0070) −0.00263 (0.0100) –

–

–

–

–

–

–

Yes*** Yes*** Yes*** 157,185 0.0406 −289,709.7

Yes*** Yes*** Yes*** 157,185 0.0560 −285,052.7

– −0.487⁎⁎ (0.2143) – 0.335 (0.4562) 0.0342⁎⁎⁎ (0.0076) – −0.0330 (0.2845) 0.0110⁎⁎⁎ (0.0038) −0.00403 (0.0086) 0.000154 (0.0054) Yes*** Yes*** Yes*** 157,185 0.0560 −285,063.8

Notes: Results of an ordered logit regression. Dependent variable: 11-point scale measuring job satisfaction (models I-III) and life satisfaction (models IV-VI). Standard errors clustered on the country level in parentheses. ***: statistically significant at the 1% level; ** statistically significant at the 5% level; * statistically significant at the 10% level (two sided tests). Control variables are included. The variable “high social protection expenditures” is a dummy variable that equals 1 if a country is in the above quartile, and equals 0, otherwise. Estimates of control variables are available from the authors upon request.

concerning the varieties of capitalism and the GEI scores.24 Thus, the differentiated analysis of social protection spending, entrepreneurial institutions and individual well-being is meaningful. The results of the model estimations with three-way interaction terms between employment status, GEI, and a dummy variable indicating the level of social protection expenditures are reported in columns III and VI of Table 3.25 We find again a positive and significant effect of the interaction between self-employment and GEI in both models for job and life satisfaction. In addition, the results of these estimations are visualized in Fig. 5 showing the predicted probability of being completely satisfied with one's job at different levels of GEI and social protection expenditures. This analysis supports our main result that a better quality of entrepreneurial institutions is associated with higher levels of individual well-being, and particularly so for self-employed people.26 Remarkably, this seems to be largely independent of the level of social protection expenditures. 4.3. Robustness checks One of our main results is that an entrepreneurship-friendly institutional environment is associated with a higher individual well-being not only among the self-employed, but also for paid employed individuals. However, better entrepreneurial institutions might be differently relevant for paid employees in small and large firms. While small entrepreneurial firms might flourish in an entrepreneurship-friendly environment, it might be of low relevance for large and established companies. In addition, employees in small firms tend have a more entrepreneurial attitude and enjoy higher levels of autonomy (Benz and Frey, 2008b). Entrepreneurship-friendly institutions would improve their prospects of realizing an entrepreneurial venture, thereby improving the well-being of paid employed individuals in smaller firms. If true, these considerations should be reflected in a stronger positive effect of GEI on employees in small firms than on employees in large firms. We investigate this issue by analyzing the relationship between individual well-being, employment in firms of different size, and the GEI. The results in Table 4, columns I and IV, suggest that job and life satisfaction of individuals in paid employment decreases as the size of a firm increases. The highest levels of job and life satisfaction are observed among the self-employed and paid employees working in firms with less than 10 employees. Columns II and V of Table 4 show the differences in this relationship depending on the quality of entrepreneurial institutions. To facilitate the interpretation of the interaction terms, the probabilities of reporting the 24

Countries with high social protection expenditures are Italy, Belgium, Sweden, Netherlands, Finland, Denmark, and France. We include a dummy variable instead of the continuous variable to ensure the interpretation and visualization of the results. 26 A corresponding figure for life satisfaction is very similar, although the differences between employment states are less pronounced, as in our main model estimation. This figure is available from the authors upon request. 25

12

Journal of Business Venturing 34 (2019) 105946

0

.1

Predicted probability .2 .3 .4

.5

M. Fritsch, et al.

20

30

40

50 GEI score

Paid employed, low SPE Self-employed, low SPE

60

70

80

Paid employed, high SPE Self-employed, high SPE

Fig. 5. Predicted probability of being completely satisfied with one's job by employment status, GEI score, and the level of social protection expenditures (SPE) (in % of GDP). 95% confidence intervals are reported.

highest value on the job satisfaction scale are visualized in Fig. B3 in the online Appendix. The estimations clearly suggest that the quality of entrepreneurial institutions increases the odds of being completely satisfied with one's job for employees of all firm sizes. The magnitude of this effect is most pronounced for employees in firms with less than 10 employees, and it becomes smaller with increasing firm size. The lowest levels of job satisfaction, as well as the smallest increase of this level with more entrepreneurshipfacilitating institutional framework conditions, are observed for people working in firms with 50 and more employees. Estimations in columns III and VI of Table 4 show that the results are robust when excluding the country-level control variables. The results for life satisfaction are similar to the results for job satisfaction, although the magnitude of the effects is substantially lower. Moreover, entrepreneurship-friendly institutions might differently affect the individual well-being of people depending on their financial standing, since they may have different motives for setting up a business. Individuals who have a difficult time earning a living may find it helpful to have a supportive institutional environment to set up a business, even if it is out of necessity. One would not necessarily expect their job satisfaction to improve, though, given that they would prefer finding an appropriate job in paid employment instead of becoming selfemployed. In turn, individuals who are doing well in financial terms and who decide to pursue an entrepreneurial opportunity might also benefit from supporting institutional environment. Their well-being might increase because they were able to realize their dream. Thus, we estimate our main model for different country-specific income quartiles to account for differences in motivations for self-employment. This analysis also accounts for the issue that selection into self-employment in countries with poor institutions or entrepreneurship-inhibiting framework conditions might be driven by necessity, while in countries with high-quality institutional framework conditions there is a higher prevalence of opportunity entrepreneurs. This heterogeneity of self-employment may explain differences in well-being, if necessity entrepreneurs have on average a lower well-being score than opportunity entrepreneurs. Thus, the interaction effect between employment status and the GEI measure might reflect that in some countries certain types of entrepreneurs are more prevalent. This issue can be assuaged by considering income quartiles under the assumption that relative income is informative about whether a self-employed person is a necessity or an opportunity entrepreneur. If our results on the interaction effect between employment status and the GEI measure hold across income groups then selection of certain types of entrepreneurs in countries with certain institutions is unlikely to be a driver of our findings. Table 5 shows that higher GEI increases the well-being of self-employed individuals across all income quartiles and at an approximately similar rate. This pattern is revealed by the positive interaction effect between being self-employed and the GEI on job and life satisfaction. Thus, even for people in entrepreneurship who have a low income and are therefore likely to be a necessity entrepreneur, entrepreneurship-facilitating institutions still have a positive effect on well-being.27 Similarly, having a high income in selfemployment, which is likely indicative of an opportunity entrepreneur, is more positively related to well-being if institutions are conducive for entrepreneurship. Remarkably, higher GEI is conducive to job satisfaction of paid employees only in the first two income quartiles, which is indicated by the coefficient for the GEI measure in the models of Table 5. High-earners among paid employed individuals do not seem to benefit from better entrepreneurial institutions. A possible explanation of this result is that high earners among paid employed people are more likely to work in large firms, for which the effect of the GEI is almost negligible, as our previous robustness check has revealed. In turn, GEI is positively associated with life satisfaction of paid employees across all income quartiles. In our main analysis we used the GEI scores as a measure of the quality of entrepreneurial institutions in a country. In a further robustness check, we use an alternative measure, namely the Doing of Business score (see Section 3.2.1). Despite the differences in the definition of the GEI and the DoB score, both indicators show closely corresponding assessments; the correlation between the two scores

27 There is one exception. Life satisfaction among the poorest self-employed of a country is also positively affected by the GEI but to the same degree as for paid employees.

13

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table 4 Individual well-being and firm size. Job satisfaction

Self-employed (reference) Paid employee, less than 10 employees Paid employee, 11–19 employees Paid employee, 20–49 employees Paid employee, 50 and more employees GEI 2013 Interaction terms: Paid employee, less than 10 employees × GEI

I

II

III

IV

V

VI

– −0.0526 (0.0825) −0.108 (0.0887) −0.173⁎ (0.0930) −0.267⁎⁎⁎ (0.1003) 0.018⁎⁎⁎ (0.0066)

– 0.821⁎⁎⁎ (0.2250) 0.929⁎⁎⁎ (0.2514) 1.245⁎⁎⁎ (0.2820) 1.427⁎⁎⁎ (0.2612) 0.041⁎⁎⁎ (0.0071)

– 0.882⁎⁎⁎ (0.2150) 1.041⁎⁎⁎ (0.2529) 1.385⁎⁎⁎ (0.3005) 1.565⁎⁎⁎ (0.2873) 0.046⁎⁎⁎ (0.0054)

– −0.0397 (0.045) −0.0591 (0.049) −0.0670 (0.054) −0.132⁎⁎ (0.052) 0.024⁎⁎ (0.009)

– 0.157 (0.1506) 0.265 (0.1869) 0.403⁎⁎ (0.2039) 0.552⁎⁎⁎ (0.1614) 0.032⁎⁎⁎ (0.0088)

– 0.200 (0.1534) 0.325⁎ (0.1967) 0.479⁎⁎ (0.2112) 0.630⁎⁎⁎ (0.1686) 0.0447⁎⁎⁎ (0.0066)

–

−0.018⁎⁎⁎ (0.0037) −0.021⁎⁎⁎ (0.0043) −0.029⁎⁎⁎ (0.0047) −0.034⁎⁎⁎ (0.0044) Yes*** Yes*** Yes*** 153,337 0.0221 −299,613.8

−0.0187⁎⁎⁎ (0.0035) −0.0225⁎⁎⁎ (0.0043) −0.0302⁎⁎⁎ (0.0049) −0.0353⁎⁎⁎ (0.0047) No Yes*** Yes*** 153,337 0.0201 −300,210.6

–

−0.00417 (0.0027) −0.00661⁎ (0.0034) −0.0095⁎⁎⁎ (0.0036) −0.0134⁎⁎⁎ (0.0031) Yes*** Yes*** Yes*** 152,080 0.0575 −274,634.6

−0.00480⁎ (0.0027) −0.00740⁎⁎ (0.0034) −0.0104⁎⁎⁎ (0.0037) −0.0144⁎⁎⁎ (0.0031) No Yes*** Yes*** 152,080 0.0557 −275,160.8

Paid employee, 11–19 employees × GEI

–

Paid employee, 20–49 employees × GEI

–

Paid employee, 50 and more employees × GEI

–

Country-specific controls Industry fixed effects Occupation fixed effects Number of observations Pseudo R2 Log likelihood

Life satisfaction

Yes*** Yes*** Yes*** 153,337 0.0202 −300,176.1

– – – Yes*** Yes*** Yes*** 152,080 0.0571 −274,736.6

Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job satisfaction (models I–III) and life satisfaction (models IV–VI). Standard errors clustered on the country level in parentheses. ***: statistically significant at the 1% level; ** statistically significant at the 5% level; * statistically significant at the 10% level (two sided tests). All control variables are included. Effects of control variables are reported in Table B3 in Online Appendix.

Table 5 Individual well-being and income. Job satisfaction 1st income quartile Paid employed (reference) Self-employed (if GEI = 0) GEI Self-employed × GEI Country-specific controls Industry fixed effects Occupation fixed effects Number of observations Pseudo R-squared Log likelihood

– −0.846⁎⁎⁎ (0.2597) 0.0233⁎⁎⁎ (0.0065) 0.0176⁎⁎⁎ (0.0045) Yes*** Yes*** Yes*** 36,769 0.0307 −76,288.5

Life satisfaction 2nd income quartile – −0.855⁎⁎⁎ (0.2789) 0.0163⁎⁎ (0.0081) 0.0215⁎⁎⁎ (0.0051) Yes*** Yes*** Yes*** 39,724 0.0201 −79,094.7

3rd income quartile – −1.366⁎⁎⁎ (0.2633) 0.0113 (0.0078) 0.0315⁎⁎⁎ (0.0044) Yes*** Yes*** Yes*** 40,730 0.0133 −77,849.8

4th income quartile – −0.936⁎⁎⁎ (0.2677) 0.00602 (0.0062) 0.0237⁎⁎⁎ (0.0046) Yes*** Yes*** Yes*** 41,240 0.0106 −75,116.2

1st income quartile – −0.0881 (0.2687) 0.0265⁎⁎⁎ (0.0089) 0.00274 (0.0050) Yes*** Yes*** Yes*** 36,472 0.0620 −69,995.3

2nd income quartile – −0.335⁎ (0.1765) 0.0239⁎⁎ (0.0100) 0.0092⁎⁎⁎ (0.0033) Yes*** Yes*** Yes*** 39,392 0.0546 −72,782.6

3rd income quartile – −0.512⁎⁎⁎ (0.1897) 0.0212⁎⁎ (0.0106) 0.0120⁎⁎⁎ (0.0033) Yes*** Yes*** Yes*** 40,437 0.0504 −72,064.2

4th income quartile – −0.366 (0.2299) 0.0197⁎⁎ (0.0095) 0.00960⁎⁎⁎ (0.0036) Yes*** Yes*** Yes*** 40,884 0.0459 −68,841.6

Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job satisfaction (columns I–IV) and life satisfaction (models V–VIII). Standard errors clustered on the country level in parentheses. ***: statistically significant at the 1% level; ** statistically significant at the 5% level; * statistically significant at the 10% level (two sided tests). Control variables are included. Effects for control variables are reported in Table B4 in Online Appendix.

14

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

in our sample is 0.82 (see Table A1 in the Appendix).28 In Table B5 in the online appendix, we repeat our main analysis by using the DoB score as an alternative measure of the entrepreneurship-facilitating quality of a country's institutional framework instead of GEI scores. This analysis yields rather similar results concerning the effect of the DoB on the well-being of the self-employed, namely, a positive and significant interaction effect between DoB and self-employment status on job and life satisfaction (columns III and VII of Table B5). Another robustness check concerns a methodological issue. Since we are interested in assessing to what extent country-level entrepreneurial institutions influence individual-level measures of well-being, multilevel modeling might be an appropriate estimation strategy (see, e.g., Kim and Karl, 2016; Shepherd, 2011). The multilevel estimation approach allows us to model country-level effects instead of just controlling for them. In more detail, the multilevel approach allows us to assess the impact of country-level predictors on the coefficients and the intercept from the individual-level regressions. Table B6 in the online Appendix contains multilevel estimations for our baseline model, which are comparable to our baseline model estimations by means of ordered logit regressions with standard errors clustered at the level of countries (Table 2). The intra-class correlation (ICC), which is a key parameter that measures the extent to which unobserved factors within each country are shared by individuals, is rather low: it is less than 0.03 in models for job satisfaction and it does not exceed 0.054 in models for life satisfaction. This suggests that there is no indication of a need to use the multilevel approach in place of ordered logistic regression with clustered standard errors. In addition, the multilevel modeling approach relies on a rather problematic assumption that individual random effects and country random effects are uncorrelated with the explanatory variables and with each other (Bryan and Jenkins, 2016). In our main analysis we estimated predicted probabilities of individuals of being completely satisfied with their jobs and lives depending on their employment status and the level of GEI, that is, a probability of reporting the highest possible value on an 11-point Likert scale. In a robustness check, we predict probabilities to report 7 and more points on an 11-point job- and life satisfaction Likert scale based on an estimation of a logit model with clustered standard errors at the level of countries with the binary dependent variable that equals 1, if a reported level of job satisfaction is at least 7 points on an 11-point Likert scale, and equals 0 otherwise. The results for job satisfaction and life satisfaction are presented in Figs. B4 and B5 in the online Appendix, correspondingly. We obtain a similar result, namely, that higher levels of GEI are associated with a significantly higher probability to be satisfied with job and life. However, the differences between self-employed and paid employed individuals disappear. Moreover, the relationship between the GEI and the probability of high job- and life satisfaction becomes concave. This means that promoting entrepreneurship-friendly institutions in countries that have relatively low GEI scores can lead to a greater increase in well-being as compared to similar measures in countries that already have high scores on this index. Finally, we also considered lagged values for the country-level variables we employed in our analysis. By this additional assessment we want to demonstrate that our main results are robust when assuming that there is a delayed influence of country-level variables on well-being. There is an almost perfect correlation of the macro indicators which are all around 0.99 which makes it unlikely that there are meaningful differences in coefficient estimates for lagged macro variables.29 This is confirmed in a series of robustness checks where we used the variables from t-1, t-3, and t-5. The results are reported in the appendix (Table B7 to B9). 5. Discussion and conclusions 5.1. Research contributions The most important result of our empirical analysis is that the considerable cross-country variation in the job and life satisfaction of selfemployed people as compared to paid employees can be explained by the entrepreneurship-facilitating quality of the institutional contexts. While in most countries self-employed individuals experience higher levels of well-being as compared to the paid employed, there are some countries where the opposite holds true. Lower levels of well-being among the self-employed as compared to paid employees are particularly found in Mediterranean countries and in some of the former socialist countries of Eastern Europe, especially in Bulgaria, Romania, and Serbia. Hence, the higher levels of well-being of the self-employed as compared to paid employees found in previous research cannot be regarded as a stylized fact. It is apparent that lack of institutional support for entrepreneurship has consequences not only for the number of people but also the type of people who choose self-employment as their occupation, as well as for the level of well-being that they report. Another important finding is that not only the self-employed, but also paid employees report higher levels of well-being in countries with entrepreneurship-facilitating institutions as compared to countries where the institutions are less favorable. This indicates that promoting an entrepreneurship-friendly framework and the resulting well-being of self-employed persons do not come at the cost of well-being among paid employees. On the contrary, the well-being of people in both occupational categories seems to be positively related to the quality of entrepreneurial institutions. These findings are robust even when we include a country's general level of wealth as measured by GDP per capita. Also, controlling for the generosity of public welfare spending has hardly any statistically significant effect on the basic relationships. 5.2. Implications for theory development and for policy A main implication of our results for theory is that the literature on well-being in different types of labor market statuses needs to 28 It should be mentioned that the GEI uses some data drawn from the Index of Economic Freedom, as prepared by the Heritage Foundation (https://www.heritage.org/index/) that is partly based on the DoB index. This information overlap makes, however, only a very minor part of the GEI and cannot explain the pronounced correlation between the two metrics. 29 The correlation coefficient for the GEI between 2012 and 2013 is about 0.9947. For GDP this value is 0.999. It is 0.988 for social protection expenditures and 0.974 for the unemployment rate.

15

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

integrate the environmental context, i.e., institutional framework conditions, into a coherent framework. Such a framework could facilitate a better understanding of the ways in which different institutions affect job and life satisfaction of both the self-employed and paid employees. Another benefit of such a coherent framework is that it could provide a suitable basis for deriving policy implications, particularly for developing appropriate designs of institutions. The main policy lesson that can be drawn from our study is that promoting entrepreneurial institutions increases the levels of well-being of the self-employed as well as that of paid employees. Hence, policies aiming at promoting a more entrepreneurial society are not necessarily contrary to the interest of paid employees, but can be regarded as being a Pareto-improvement for these groups. Enhanced welfare in an entrepreneurial society does not come at the expense of paid employees. Our results particularly show that the positive effect of self-employment on well-being can be promoted by creating entrepreneurship-facilitating institutions. The extent to which such policies are also beneficial for paid employees may, however, depend on the type of institutions. Further research should, therefore, seek to identify those types of entrepreneurial institutions that improve the well-being of both groups, the self-employed and paid employees. Against this background, it should also be explored inasmuch policy objectives in some countries (especially, coordinated market economies) where social welfare programs are designed to reduce social and economic risk for citizens conflict with encouraging people to become self-employed. Our analysis reveals that the relationship between the GEI (our measure of entrepreneurship-friendly institutions) and well-being is largely independent of social welfare expenditures. Nevertheless, we acknowledge the need for more research on this important issue. 5.3. Limitations A limitation of our study is that the EU-SILC data that we used in our analyses does not provide us with much information about the finer details of the structure and organization of entrepreneurial businesses. For instance, subjective well-being may differ for the selfemployed with and without employees (Sevä et al., 2016). There may also be a difference between necessity and opportunity entrepreneurs (Block and Koellinger, 2009). Although our data do not provide us with a direct measure of motives for self-employment, we were able to at least partly capture the differences between necessity and opportunity motivated entrepreneurs by comparing the wellbeing of the self-employed and paid-employees along an income distribution. The respective analyses (Section 4.4) clearly indicate that our main results do not appear to be driven by cross-country differences in the quality of self-employment. Another shortcoming is that we do not know how long self-employed people have been in business. This may be important because institutions could be designed in a way that protect the incumbent self-employed, while alienating de novo entrants, or the other way around. Our analysis is also limited by the fact that we only have information on well-being from one particular survey year in 2013. The robustness checks show that the estimates are not sensitive to the choice of country-level variables from another year. Nevertheless, a panel data analysis would be superior to rule out the influence of unobserved heterogeneity. Longitudinal data at that scale are not available and need to be addressed by future research. Moreover, the measures of individual's subjective well-being that are available in our data are well-established indicators, although they are not without limitations. One such limitation regards the ability of individuals to adapt to their circumstances, such that different life circumstances such as becoming an entrepreneur or facing institutional change may result in the same level of overall subjective well-being (Diener and Tov, 2012). This adaptive ability hampers measuring the effect of life circumstances on well-being and might explain our less-pronounced results for overall life satisfaction. Future research should develop and use more complex indicators of subjective well-being. Although we have shown the relevance of institutional framework conditions for the well-being of self-employed individuals, as well as of dependently employed persons, there are a number of issues that need to be tackled by future research. One of these issues is measuring entrepreneurial institutions. Differences between groups of countries may reveal some of the heterogeneity across the groups, but do not tell us much about what types of institutions are most important for the well-being of the self-employed and of paid employees. The DoB, and particularly the GEI scores, provide a considerably better description of the entrepreneurship ecosystem of a country, but could, of course, be improved. Better availability of data on various aspects of entrepreneurship-fostering institutions may lead to more differentiated results. Finally, in the current study, we include several macro-level controls to control for the possibility that common omitted variables drive the relationship between the GEI and DoB scores and our measures of satisfaction. Future research might tackle this contextual challenge by considering a wider range of potentially omitted variables. Finally, our analysis is restricted to EU countries. This calls for replication studies for other parts of the world to check for robustness of our results. 5.4. Avenues for future research Generally, the relationship between the institutional framework of a society and the well-being of people is not well understood. Given our current state of knowledge, we can only speculate about the nature of the relevant links. This pertains particularly to the positive relationship that we have found between the entrepreneurship-facilitating character of a country's institutions and the wellbeing of paid employees. One possible explanation for the positive effect of more entrepreneurship-facilitating institutions on job and life satisfaction of paid employees is that such institutions foster entrepreneurial behavior within companies (corporate entrepreneurship) that, in turn, tends to make people happier. There might be also an indirect effect to the extent that the well-being of entrepreneurs positively feeds back to paid employee's motivation and the working atmosphere in their firm. A further explanation would be that entrepreneurship-friendly institutions lead to an increasing competition that challenges established firms, thus, forcing them to improve working conditions for their employees. It might, of course, also be the case that the quality of the institutional framework is correlated with other factors that influence the well-being of paid-employees. Disentangling the specific channels requires a more in-depth analysis that is beyond the scope of the 16

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

present paper. To amplify our understanding, research will have to investigate these issues considering the micro-level of individuals, too. One important avenue for future research is to examine the effect of single elements of the institutional framework on the wellbeing of self-employed persons, as well as of paid employees. Is it, for example, the type of labor market regulation, or is it factors such as quality of government and the level of corruption that play a stronger role here? Such investigations could be of great help in identifying those parts of the framework that are most critical for well-being. Moreover, it is important to know more about the effect of single institutions on different employment categories. This applies to not only necessity-driven self-employment and ambitious opportunity-driven self-employment, but also to the wide variety of paid employees. A related and no less important field for future research concerns the effects of institutions and well-being on individual behavior. Our result that in some countries the self-employed may realize lower levels of job and life satisfaction than paid employees raises the question of why the respective persons remain in self-employment. The most plausible explanation for this phenomenon is that a large part of the self-employed in these countries engage in entrepreneurial activities because they do not have a better employment opportunity (necessity entrepreneurship). Another explanation could be a feeling of “over-confidence” among the self-employed (Koellinger et al., 2007) who hope that their income as a self-employed individual will increase over time. This expectation may also explain why job and life satisfaction among low income self-employed was more positive than among low income paid-employees. Clearly, further research on the effect of institutions on individual behavior is necessary and desirable to throw more light on such phenomena. 6. Final remark This study finds a positive link between the entrepreneurship-facilitating character of a country's institutions and the job and life satisfaction of both the self-employed and paid employees. This implies that shifting to a more entrepreneurial society with more entrepreneurship-friendly institutions is a Pareto-improvement that is beneficial for self-employed and paid employees. Appendix A

Fig. A1. Scatterplot of country-specific mean values of job satisfaction and the GEI.

Fig. A2. Scatterplot of country-specific mean values of life satisfaction and the GEI.

17

Journal of Business Venturing 34 (2019) 105946

0

Predicted probability .1 .2 .3

.4

M. Fritsch, et al.

20

30

40

50 GEI score

60

70

80

Self-employed Paid employed, less than 10 employees Paid employed, 11-19 employees Paid employed, 20-49 employees Paid employed, 50 and more employees

.2

Predicted probability .4 .6 .8

1

Fig. A3. Predicted probability of being completely satisfied with one's job by employment status, firm size, and the GEI score. 95% confidence intervals are reported.

20

30

40

50

GEI score

Paid employed

60

70

80

Self-employed

.4

Predicted probability .6 .8

1

Fig. A4. Predicted probability of reporting 7 and more points on an 11-point scale measuring satisfaction with one's job, by employment status and the GEI scores.

20

30

40

50

GEI score

Paid employed

60

70

80

Self-employed

Fig. A5. Predicted probability of reporting 7 and more points on an 11-point scale measuring satisfaction with life, by employment status and the GEI scores. 18

Male

Married

Primary degree

Secondary degree

Tertiary degree

5

6

7

8

9

2nd income quartile

3rd income quartile

4th income quartile

Health status

Global Entrepreneurship

13

14

15

16

17

19

Gini index

Unemployment rate

23

-0.1562

-0.1144

0.141

0.1495

0.0555

0.1489

0.1718

0.1729

0.1269

0.0341

-0.048

-0.118

-0.0211

0.004

0.1007

-0.0775

-0.0541

0.0261

-0.018

-0.0014

-0.04

0.4846

1

1

-0.1929

-0.1593

0.2265

0.2712

0.1383

0.2202

0.2809

0.2932

0.1246

0.0286

-0.0466

-0.1112

-0.0235

-0.0226

0.1553

-0.1155

-0.094

0.0738

0.0018

-0.0832

-0.0394

1

2

0.0985

0.0514

-0.0596

-0.0597

0.0045

-0.092

-0.0922

-0.0177

-0.0325

-0.0848

-0.0443

0.1671

-0.0334

0.2071

-0.0491

0.0207

0.0705

0.0577

0.1112

0.0998

1

3

-0.0129

-0.0074

0.0049

0.0197

0.0228

0.0193

0.0095

-0.2923

0.126

0.0157

-0.0472

-0.0988

-0.1293

-0.0052

-0.0414

0.0023

0.0983

0.2994

0.0032

1

4

0.0239

0.0088

0.0093

-0.0053

0.0095

-0.0292

-0.022

0.0452

0.1898

0.0585

-0.0946

-0.1607

0.002

0.2861

-0.072

0.063

0.0191

0.0368

1

5

0.0602

0.047

-0.0635

-0.0703

-0.0448

-0.0674

-0.0824

-0.0589

0.0909

0.011

-0.0408

-0.0639

-0.064

0.0108

-0.004

-0.0138

0.0457

1

6

0.1533

0.1377

-0.0095

-0.0598

0.0368

-0.0753

-0.063

-0.0896

-0.0834

-0.0315

0.0246

0.0941

-0.0025

-0.0003

-0.1447

-0.2472

1

7

-0.0367

-0.0165

-0.0998

-0.1083

-0.0948

-0.0949

-0.1151

-0.1046

-0.2747

0.0039

0.1378

0.14

0.0074

-0.0012

-0.923

1

8

1

9

-0.0234

-0.0378

0.1056

0.1343

0.0822

0.1269

0.1425

0.1425

0.3137

0.0086

-0.1505

-0.1804

-0.0066

0.0014

Notes: N = 158,463. Information on health status is only available for 157,185 respondents.

GDP

22

Index

Human Development

21

20

Expenditure, % of GDP

Doing of Business Score

Social Protection

18

19

Index, GEI)

1st income quartile

12

year

Job change since last

Age

4

11

Self-employed

3

Working hours per week

Life satisfaction

2

10

Job satisfaction

1

Correlation matrix.

Table A1

0.0731

0.0117

-0.1096

-0.1378

-0.1099

-0.1044

-0.1351

0.0265

0.1924

0.0774

-0.0125

-0.2672

-0.0282

1

10

-0.0118

0.0201

-0.0046

0.0151

-0.0067

0.0055

0.0154

0.0133

-0.0696

-0.0562

-0.0115

0.1423

1

11

0.0065

-0.004

-0.005

-0.0022

-0.005

-0.0055

-0.0053

-0.0737

-0.326

-0.3233

-0.3179

1

12

1

-0.005

0.0014

0.001

-0.0012

-0.0014

0.0029

0.003

-0.0259

-0.3431

-0.3402

13

1

-0.0033

0.0013

0.003

0.0028

0.0048

0.0015

0.0037

0.0187

-0.3489

14

1

0.0021

0.0012

0.0009

0.0005

0.0014

0.001

-0.0015

0.0778

15

1

0.005

-0.0149

0.1066

0.1038

0.0966

0.0264

0.0795

16

1

-0.5802

-0.4356

0.7405

0.8903

0.5595

0.8209

17

1

-0.5295

-0.3456

0.5079

0.7132

0.3008

18

1

-0.0889

-0.2533

0.5406

0.6704

19

1

-0.4906

-0.4308

0.8068

20

1 -0.4764

-0.3282

21

1 0.4454

22

M. Fritsch, et al.

Journal of Business Venturing 34 (2019) 105946

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table A2

Descriptive statistics. Number of observations

Mean

Job satisfaction

158,463

7.256

Life satisfaction

158,463

7.290

Standard deviation

Median

Minimum

Maximum

2.031

8

0

10

1.868

8

0

10

Dependent variables

Individual-level variables Self-employed

158,463

0.130

0.337

0

0

1

Age

158,463

44.031

10.903

45

18

65

Male

158,463

0.497

0.500

0

0

1

Married

158,463

0.604

0.489

1

0

1

Primary degree

158,463

0.037

0.190

0

0

1

Secondary degree

158,463

0.612

0.487

1

0

1

Tertiary degree

158,463

0.351

0.477

0

0

1

Working hours per week

158,463

39.048

10.039

40

1

99

Job change since last year

158,463

0.058

0.234

0

0

1

Health status

157,185

4.029

0.757

4

1

5

1st income quartile

158,463

0.232

0.422

0

0

1

2nd income quartile

158,463

0.251

0.433

0

0

1

3rd income quartile

158,463

0.257

0.437

0

0

1

4th income quartile

158,463

0.260

0.439

0

0

1

Global Entrepreneurship Index, GEI)

158,463

50.677

14.289

44.978

22.694

77.128

Doing of Business score, DoB)

158,463

72.770

6.388

71.61

60.46

85.63

Social Protection Expenditures, % of GDP)

158,463

24.723

5.543

25.8

14.6

34.2

Human Development Index

158,463

0.875

0.042

0.876

0.771

0.946

GDP

158,463

26,886

18,012

22,000

4,800

85,300

Gini index

158,463

32.116

3.424

32.5

25.4

37

Unemployment rate

158,463

11.164

6.222

10.2

3.15

27.5

Country-level variables

Table A3

Pillars of the Global Entrepreneurship Index. Pillar name Opportunity Perception

Description This pillar refers to the entrepreneurial opportunity perception potential of the population weighted with the size and the level of agglomeration of that country reflecting the potential size of the market.

Start-up skills

This pillar captures the perception of start-up skills in the population and weights this aspect with the quality of human resources available for entrepreneurial processes in the country

Nonfear of failure

Nonfear of failure captures the inhibiting effect of fear of failure of the population on entrepreneurial action combined with a measure of the country's business risk

Networking

This pillar combines two aspects of networking, 1) a proxy of the ability of potential and active entrepreneurs to access and mobilize opportunities and resources and, 2) the possible use of the internet.

Cultural Support

This pillar combines how positively a given country's inhabitants view entrepreneurs in terms of status and career choice and how the level of corruption in that country affects this view.

Opportunity Start-up

The opportunity start-up pillar captures the prevalence of individuals who pursue potentially better quality opportunity-driven start-ups as opposed to necessity-driven start-ups and weights this against regulatory constraints.

Tech Sector

This pillar reflects the technology-intensity of a country's start-up activity combined with a country's capacity for firm-level technology absorption.

Quality of Human Reso-

This pillar captures the quality of entrepreneurs by weighting the percentage of start-ups founded by individuals with higher than secondary education with a qualitative measure of the

urces Competition Product Innovation

propensity of firms in a given country to train their staff This pillar measures the level of the product or market uniqueness of start-ups combined with the market power of existing businesses and business groups This pillar captures the tendency of entrepreneurial firms to create new products. The pillar was created by weighting the percentage of firms that offer products that are new to at least some of their customers with a complex measure of innovation.

Process Innovation

This pillar captures the use of new technologies by start-ups combined with the Gross Domestic Expenditure on Research and Development, GERD). GERD serves as measurement of the systematic research activity as opposed to easy to copy technological improvements.

High growth

The high growth pillar is a combined measure of, 19 the percentage of high-growth businesses that intend to employ at least ten employees and plan to grow more than 50 % in 5 years and, 2) business strategy sophistication.

Internationalization

The internationalization pillar captures the degree to which a country's entrepreneurs are internationalized, as measured by businesses' exporting potential weighted by the level of economic globalization of the country.

Risk capital

This pillar combines two measures of finance: informal investment in start-ups and a measure of institutional venture capital. Availability of risk capital is to fulfill growth aspirations.

20

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Table A4

Determinants of job- and life satisfaction. Job satisfaction I Paid employee, reference) Self-employed

II

Life satisfaction III

IV

V

VI

VII

VIII

-

-

-

-

-

-

-

-

0.136

0.163⁎

-1.201⁎⁎⁎

-1.322⁎⁎⁎

0.0507

0.0871⁎

-0.383⁎⁎

-0.451⁎⁎

(0.0951)

(0.0932)

(0.2578)

(0.2698)

(0.0559)

(0.0489)

(0.1729)

(0.1809)

GEI

-

0.0175⁎⁎⁎

0.0133⁎

0.0162⁎⁎⁎

-

0.0239⁎⁎

0.0224⁎⁎

0.0341⁎⁎⁎

(0.0068)

(0.0068)

(0.0045)

Self-employed x GEI

-

-

0.0278⁎⁎⁎

0.0292⁎⁎⁎

(0.0043)

(0.0044)

-0.0187⁎⁎

-0.0194⁎

-0.0184⁎

-

(0.0093)

(0.0100)

(0.0097)

Social Protection Expenditures, % GDP Human Development Index GDP Gini Index Unemployment rate Age Male Married Secondary degree Tertiary degree Working hours per week Job change since last year Total gross yearly working income: 2nd quartile Total gross yearly working income: 3rd quartile Total gross yearly working income: 4th quartile Health status

3.546

-0.985

-0.828

(2.9531)

(3.5870)

(3.4136)

0.000

0.000

0.000

(0.0000)

(0.0000)

(0.0000)

-0.0219

-0.0208

-0.0201

(0.0148)

(0.0147)

(0.0138)

-0.0228⁎⁎

-0.0153

-0.0150

(0.0113)

(0.0110)

(0.0106)

-0.0066⁎⁎⁎

-0.00652⁎⁎⁎

-0.0068⁎⁎⁎

(0.0019)

(0.0018)

(0.0018)

-0.0953⁎⁎⁎

-0.0881⁎⁎⁎

(0.0209) 0.128⁎⁎⁎ (0.0166)

-

(0.0094)

(0.0095)

(0.0060)

-

-

0.00958⁎⁎⁎

0.0105⁎⁎⁎

(0.0031)

(0.0031)

-0.0186

-0.0198

-0.0195

-

(0.0147)

(0.0159)

(0.0158)

10.27⁎⁎

4.151

4.213

(4.1389)

(4.7270)

(4.6757)

-0.000

-0.000

-0.000

(0.0000)

(0.0000)

(0.0000)

-0.0264

-0.0247

-0.0244

(0.0258)

(0.0251)

(0.0249)

-

-0.0225

-0.0124

-0.0123

(0.0204)

(0.0200)

(0.0198)

-

-0.00658⁎⁎⁎

-0.0121⁎⁎⁎

-0.0119⁎⁎⁎

-0.0120⁎⁎⁎

(0.0019)

(0.0025)

(0.0024)

(0.0024)

(0.0024)

-0.0851⁎⁎⁎

-0.107⁎⁎⁎

-0.0878⁎⁎⁎

-0.0783⁎⁎⁎

-0.0770⁎⁎⁎

-0.0872⁎⁎⁎

(0.0209)

(0.0208)

(0.0209)

(0.0216)

(0.0214)

(0.0213)

(0.0249)

0.134⁎⁎⁎

0.134⁎⁎⁎

0.132⁎⁎⁎

0.495⁎⁎⁎

0.504⁎⁎⁎

0.504⁎⁎⁎

0.500⁎⁎⁎

(0.0150)

(0.0146)

(0.0167)

(0.0419)

(0.0406)

(0.0407)

(0.0440)

-0.0116⁎⁎⁎

-0.0740

-0.0548

-0.0764

0.0486

0.120

0.146

0.139

0.275⁎⁎⁎

(0.1241)

(0.1127)

(0.1016)

(0.1120)

(0.0820)

(0.0910)

(0.0889)

(0.1056)

-0.195

-0.200⁎

-0.230⁎⁎

-0.137

0.184⁎

0.179⁎

0.169⁎

0.278⁎⁎⁎

(0.1336)

(0.1186)

(0.1058)

(0.1097)

(0.0944)

(0.0942)

(0.0912)

(0.1065)

-0.0003

-0.0001

-0.0005

0.001

-0.006⁎⁎⁎

-0.006⁎⁎⁎

-0.006⁎⁎⁎

-0.005⁎⁎⁎

(0.0020)

(0.0019)

(0.0019)

(0.0018)

(0.0011)

(0.0011)

(0.0011)

(0.0014)

0.0367

0.0357

0.0344

0.0242

-0.120⁎

-0.121⁎⁎

-0.122⁎⁎

-0.123⁎

(0.0445)

(0.0411)

(0.0420)

(0.0463)

(0.0617)

(0.0572)

(0.0575)

(0.0642)

0.189⁎⁎⁎

0.185⁎⁎⁎

0.180⁎⁎⁎

0.165⁎⁎⁎

0.174⁎⁎⁎

0.168⁎⁎⁎

0.166⁎⁎⁎

0.148⁎⁎⁎

(0.0374)

(0.0372)

(0.0396)

(0.0386)

(0.0298)

(0.0296)

(0.0301)

(0.0331)

0.410⁎⁎⁎

0.406⁎⁎⁎

0.402⁎⁎⁎

0.379⁎⁎⁎

0.340⁎⁎⁎

0.333⁎⁎⁎

0.332⁎⁎⁎

0.304⁎⁎⁎

(0.0513)

(0.0521)

(0.0523)

(0.0525)

(0.0359)

(0.0365)

(0.0362)

(0.0430)

0.675⁎⁎⁎

0.675⁎⁎⁎

0.671⁎⁎⁎

0.644⁎⁎⁎

0.553⁎⁎⁎

0.551⁎⁎⁎

0.549⁎⁎⁎

0.516⁎⁎⁎

(0.0661)

(0.0670)

(0.0671)

(0.0705)

(0.0431)

(0.0441)

(0.0435)

(0.0528)

-

-

-

-

0.665⁎⁎⁎

0.671⁎⁎⁎

0.671⁎⁎⁎

0.661⁎⁎⁎

(0.0492)

(0.0480)

(0.0481)

(0.0514) Yes***

Industry fixed effects

Yes***

Yes***

Yes***

Yes***

Yes***

Yes***

Yes***

Occupation fixed effects

Yes***

Yes***

Yes***

Yes***

Yes***

Yes***

Yes***

Yes***

Number of observations

158,463

158,463

158,463

158,463

157,185

157,185

157,185

157,185

Pseudo R-squared Log likelihood

0.0194

0.0202

0.0216

0.0197

0.0558

0.0575

0.0576

0.0558

-310,920.0

-310,642.8

-310,225.4

-310,827.4

-285,128.8

-284,624.4

-284,575.0

-285,123.6

Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job satisfaction, Models I-IV) and life satisfaction, Models VVIII). Standard errors clustered on the country level in parentheses. ***: statistically significant at the 1% level; ** statistically significant at the 5% level; * statistically significant at the 10% level (two sided tests).

Appendix B. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbusvent.2019.105946.

References Acs, Zoltan J., Szerb, Laszlo, Autio, Erkko, Lloyd, Ainsley, 2017. Global Entrepreneurship Development Index. Washington, D.C.: The Global Entrepreneurship and Development Institute. doi:https://doi.org/10.1007/978-3-319-63844-7_3. Anand, Sudhir, Sen, Amartya K., 1994. Human Development Index: Methodology and Measurement. Human Development Report Office. https://ora.ox.ac.uk/ objects/uuid:98d15918-dca9-4df1-8653-60df6d0289dd/download_file?file_format=application/pdf&safe_filename=HDI_methodology.pdf&type_of_work= Report. Armour, John, Cumming, Douglas, 2008. Bankruptcy law and entrepreneurship. Am. Law Econ. Rev. 10(2), 303–350. doi:https://doi.org/10.1093/aler/ahn008. Audretsch, David B., 2007. The Entrepreneurial Society. Oxford University Press, Oxford. Audretsch, David B., Thurik, Roy A., 2001. What is new about the new economy: sources of growth in the managed and entrepreneurial economies. Ind. Corp. Chang. 10, 267–315. https://doi.org/10.1093/icc/10.1.267.

21

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Avnimelech, Gil, Zelekha, Yaron, Sharabi, Eyal, 2014. The effect of corruption on entrepreneurship in developed vs non-developed countries. International Journal of Entrepreneurial Behaviour & Research 20, 237–262. https://doi.org/10.1108/IJEBR-10-2012-0121. Baumol, William J., 1990. Entrepreneurship: productive, unproductive, and destructive. J. Polit. Econ. 98, 893–921. https://doi.org/10.1086/261712. Benz, Matthias, Frey, Bruno S., 2008a. Being independent is a great thing: subjective evaluations of self-employment and hierarchy. Economica 75, 362–383. https:// doi.org/10.1111/j.1468-0335.2007.00594.x. Benz, Matthias, Frey, Bruno S., 2008b. The value of doing what you like: evidence from the self-employed in 23 countries. J. Econ. Behav. Organ. 68, 445–455. https:// doi.org/10.1016/j.jebo.2006.10.014. Binder, M., Coad, Alex, 2013. Life satisfaction and self-employment: a matching approach. Small Bus. Econ. 40 (4), 1009–1033. https://doi.org/10.1007/s11187-0119413-9. Binder, Martin, Coad, Alex, 2016. How satisfied are the self-employed? A life domain view. J. Happiness Stud. 17 (4), 1409–1433. https://doi.org/10.1007/s10902015-9650-8. Blanchflower, David G., 2000. Self-employment in OECD countries. Labour Econ. 7 (5), 471–505. https://doi.org/10.1016/S0927-5371(00)00011-7. Blanchflower, David G., 2004. Self-employment: more may not be better. Swedish Economic Policy Review 11, 15–73. Block, Joern, Koellinger, Philipp, 2009. I can't get no satisfaction–necessity entrepreneurship and procedural utility. Kyklos 62 (2), 191–209. https://doi.org/10.1111/ j.1467-6435.2009.00431.x. Brambor, Thomas, Clark, William Roberts, Golder, Matt, 2006. Understanding interaction models: improving empirical analyses. Polit. Anal. 14, 63–82. https://doi. org/10.1093/pan/mpi014. Braunerhjelm, Pontus, Eklund, Johan E., 2014. Taxes, tax administrative burdens and new firm formation. Kyklos 67, 1–11. https://doi.org/10.1111/kykl.12040. Bryan, Mark L., Jenkins, Stephen P., 2016. Multilevel modelling of country effects: a cautionary tale. Eur. Sociol. Rev. 32 (1), 3–22. Chowdhury, Farzana, Audretsch, David B., Belitzki, Maksim, 2018a. Institutions and entrepreneurship quality. Entrepreneurship in Theory and Practice. https://doi. org/10.1177/1042258718780431. Chowdhury, Farzana, Sameeksha, Desai, Audretsch, David B., 2018b. Corruption, Entrepreneurship, and Social Welfare—A Global Perspective. Springer Nature, Heidelberg. https://doi.org/10.1007/978-3-319-64916-0. Clark, Andrew E., Oswald, Andrew J., 1996. Satisfaction and comparison income. J. Public Econ. 61, 359–381. https://doi.org/10.1016/0047-2727(95)01564-7. Coad, Alex, Binder, Martin, 2014. Causal linkages between work and life satisfaction and their determinants in a structural VAR approach. Econ. Lett. 124 (2), 263–268. https://doi.org/10.1016/j.econlet.2014.05.021. Croson, David C., Minniti, Maria, 2012. Slipping the surly bonds: the value of autonomy in self-employment. J. Econ. Psychol. 33, 355–365. https://doi.org/10.1016/j. joep.2011.05.001. Diener, E., Tov, W., 2012. National accounts of well-being. In: Land, K.C., Michalos, A.C., Sirgy, M.J. (Eds.), Handbook of Social Indicators and Quality of Life Research. Springer, New York, NY, pp. 137–156. Diener, Ed, Inglehart, Ronald, Tay, Louis, 2013. Theory and validity of life satisfaction scales. Soc. Indic. Res. 112, 497–527. https://doi.org/10.1007/s11205-0120076-y. Dilli, Selin, Elert, Niklas, Herrmann, Andrea, 2018. Varieties of entrepreneurship: exploring the institutional foundations of different entrepreneurship types through ‘varieties-of-capitalism’ arguments. Small Bus. Econ. 51, 293–320. doi:https://doi.org/10.1007/s11187-018-0002-z. Djankov, Simeon, La Porta, Rafael, Lopez-de-Silanes, Florencio, Shleifer, Andrei, 2002. The regulation of entry. Q. J. Econ. 118, 1–37. doi:https://doi.org/10.1162/ 003355302753399436. Elert, Niklas, Magnus, Henrekson, Stenkula, Mikael, 2017. Institutional Reform for Innovation and Entrepreneurship—An Agenda for Europe. Springer Nature, Cham. https://doi.org/10.1007/978-3-319-55092-3. Emmons, Robert A., 1996. Striving and feeling: Personal goals and subjective well-being. In: Gollwitzer, Peter M., Bargh, John A. (Eds.), The Psychology of Action: Linking Cognition and Motivation to Behaviour. Guilford Press, New York, NY, US, pp. 313–337. European Commission, 2010. Europe 2020- a Strategy for Smart, Sustainable and Inclusive Growth. (3 March 2010, COM (2010) 2020). European Commission, 2013. Entrepreneurship 2020 Action Plan-Reigniting the Entrepreneurial Spirit in Europe. (9 January 2013, COM(2012) 795 final). European Commission, 2016. DG GROW strategic plan 2016-2020. http://ec.europa.eu/atwork/synthesis/amp/doc/grow_sp_2016-2020_en.pdf. Eurostat, 2016. European system of integrated social protection statistics - ESSPROS. Manual and user guidelines. https://ec.europa.eu/eurostat/documents/ 3859598/7766647/KS-GQ-16-010-EN-N.pdf/3fe2216e-13b0-4ba1-b84f-a7d5b091235f. Fonseca, Raquel, Lopez-Garcia, Paloma, Pissarides, Christopher A., 2001. Entrepreneurship, start-up costs and employment. Eur. Econ. Rev. 45, 692–705. https://doi. org/10.1016/S0014-2921(01)00131-3. Fonseca, Raquel, Michaud, Pierre-Carl, Sopraseuth, Thepthida, 2007. Entrepreneurship, wealth, liquidity constraints, and start-up costs. Comparative Labor Law & Policy Journal 28, 637–673. Frey, Bruno S., Benz, Matthias, Stutzer, Alois, 2004. Introducing procedural utility: not only what, but also how matters. Journal of Institutional and Theoretical Economics, JITE/Zeitschrift für die gesamte Staatswissenschaft 160, 377–401. https://www.jstor.org/stable/40752468. Freytag, Andreas, Thurik, Roy, 2007. Entrepreneurship and its determinants in a cross-country setting. J. Evol. Econ. 17, 117–131. https://doi.org/10.1007/s00191006-0044-2. Fritsch, Michael, 2013. New business formation and regional development—a survey and assessment of the evidence. Foundations and Trends in Entrepreneurship 9, 249–364. https://doi.org/10.1561/0300000043. Fritsch, Michael, Wyrwich, Michael, 2017. The effect of entrepreneurship for economic development—an empirical analysis using regional entrepreneurship culture. J. Econ. Geogr. 17, 157–189. https://doi.org/10.1093/jeg/lbv049. Golpe, Antonio A., Millán, José Maria, Román, Concepción, 2018. Labour market institutions and entrepreneurship. In: Congregado, Emilio (Ed.), Measuring Entrepreneurship: Building a Statistical System. Springer Science, New York, pp. 279–296. https://doi.org/10.1007/978-0-387-72288-7_14. Hamilton, Barton H., 2000. Does entrepreneurship pay? An empirical analysis of returns to self-employment. J. Polit. Econ. 108, 604–632. https://doi.org/10.1086/ 262131. Hanglberger, Dominik, Merz, Joachim, 2015. Does self-employment really raise job satisfaction? Adaptation and anticipation effects on self-employment and general job changes. Journal for Labour Market Research 48 (4), 287–303. https://doi.org/10.1007/s12651-015-0175-8. Herrmann, Andrea M., 2019. A plea for varieties of entrepreneurship. Small Bus. Econ. 52, 331–343. https://doi.org/10.1007/s11187-018-0093-6. Hessels, Jolanda, Arampatzi, Efstratia, van der Zwan, Peter, Burger, Martijn, 2018. Life satisfaction and self-employment in different types of occupations. Applied Economic Letters 25, 734–740. https://doi.org/10.1080/13504851.2017.1361003. Holmes, Thomas J., Schmitz, James A., 1990. A theory of entrepreneurship and its application to the study of business transfers. J. Polit. Econ. 98, 265–294. https:// doi.org/10.1086/261678. Kahneman, D., Krueger, A.B., 2006. Developments in the measurement of subjective well-being. J. Econ. Perspect. 20 (1), 3–24. https://doi.org/10.1257/ 089533006776526030. Kibler, Ewald, Kautonen, Teemu, Fink, Matthias, 2014. Regional social legitimacy of entrepreneurship: implications for entrepreneurial intention and start-up behaviour. Reg. Stud. 48, 995–1015. https://doi.org/10.1080/00343404.2013.851373. Kihlstrom, Richard E., Laffont, Jean-Jaques, 1979. A general equilibrium theory of firm formation based on risk aversion. J. Polit. Econ. 87, 719–748. https://doi.org/ 10.1086/260790. Kim, Phillip, Wennberg, Karl, Croidieu, Grégoire, 2016. Untapped riches of Meso-level applications in multilevel entrepreneurship mechanisms. Acad. Manag. Perspect. 30(3), 273–291. doi:https://doi.org/10.5465/amp.2015.0137. Klapper, Leora, Laeven, Luc, Rajan, Raghuram, 2006. Entry regulation as a barrier to entrepreneurship. J. Financ. Econ. 82, 591–629. https://doi.org/10.1016/j. jfineco.2005.09.006. Koellinger, Philipp, Minniti, Maria, Schade, Christian, 2007. “I think I can, I think I can”: overconfidence and entrepreneurial behavior. J. Econ. Psychol. 28, 502–527. https://doi.org/10.1016/j.joep.2006.11.002. Licht, Amir N., Goldschmidt, Chanan, Schwartz, Shalom H., 2007. Culture rules: the foundations of the rule of law and other norms of governance. J. Comp. Econ. 35, 659–688. https://doi.org/10.1016/j.jce.2007.09.001. Lucas, Robert E., 1978. On the size distribution of business firms. Bell J. Econ. 9, 508–523. https://doi.org/10.2307/3003596.

22

Journal of Business Venturing 34 (2019) 105946

M. Fritsch, et al.

Lucas, Richard E., Diener, Ed, Suh, Eunkook, 1996. Discriminant validity of well-being measures. J. Pers. Soc. Psychol. 71, 616–628. https://doi.org/10.1037/00223514.71.3.616. McGrath, Rita, MacMillan, Ian, 1992. More like each other than anybody else? A cross-cultural study of entrepreneurial perceptions. J. Bus. Ventur. 7, 419–429. https://doi.org/10.1016/0883-9026(92)90017-L. Millán, José Maria, Hessels, Jolanda, Thurik, Roy A., Aguado, Rafael, 2013. Determinants of job satisfaction: a European comparison of self-employed and paid employees. Small Bus. Econ. 40, 651–670. doi:https://doi.org/10.1007/s11187-011-9380-1. Moskovitz, Tobias J., Vissing-Jorgensen, Annette, 2002. The returns to entrepreneurial investment: a private equity premium puzzle? Am. Econ. Rev. 92, 745–778, doi:https://doi.org/10.1257/00028280260344452. North, Douglass C., 1994. Economic performance through time. Am. Econ. Rev. 84, 359–368. OECD, 2013. OECD Guidelines on Measuring Subjective Well-Being. OECD Publishing, Paris. https://doi.org/10.1787/9789264191655-en. Pavot, William, Diener, Ed, 2008. The satisfaction with life scale and the emerging construct of life satisfaction. J. Posit. Psychol. 3, 137–152. https://doi.org/10.1080/ 17439760701756946. Samila, Sampsa, Sorenson, Olav, 2011. Venture capital, entrepreneurship, and economic growth. Rev. Econ. Stat. 93, 338–349. https://doi.org/10.1162/REST_a_ 00066. Schumpeter, Joseph A., 1934. The Theory of Economic Development. Harvard University Press, Cambridge, MA. Schumpeter, Joseph A., 1942. Capitalism, Socialism and Democracy, 3rd edition. George Allen and Unwin, London (1976). Sevä, Ingemar J., Vinberg, Stig, Nordenmark, Mikael, Strandh, Mattias, 2016. Subjective well-being among the self-employed in Europe: macroeconomy, gender and immigrant status. Small Bus. Econ. 46, 239–253. https://doi.org/10.1007/s11187-015-9682-9. Shepherd, Dean A., 2011. Multilevel entrepreneurship research: opportunities for studying entrepreneurial decision making. J. Manag. 37, 412–420. https://doi.org/ 10.1177/0149206310369940. Shir, Nadav, 2016. Entrepreneurial Well-Being—The Payoff Structure of Business Creation. Ph.D. Dissertation Stockholm School of Economics. https://ex.hhs.se/ dissertations/849175-FULLTEXT02.pdf. Shir, Nadav, Nikolaev, Boris N., Wincent, Joakim, 2018. Entrepreneurship and well-being: the role of psychological autonomy, competence, and relatedness. J. Bus. Ventur. https://doi.org/10.1016/j.jbusvent.2018.05.002. Sobel, Russel S., 2008. Testing Baumol: institutional quality and the productivity of entrepreneurship. J. Bus. Ventur. 23, 641–655. https://doi.org/10.1016/j. jbusvent.2008.01.004. Sorgner, Alina, Fritsch, Michael, Kritikos, Alexander, 2017. Do entrepreneurs really earn less? Small Bus. Econ. 49, 251–272. https://doi.org/10.1007/s11187-0179874-6. van Praag, Bernard M.S., Frijters, Paul, Ferrer-i-Carbonel, l Ada, 2003. The anatomy of subjective well-being. J. Econ. Behav. Organ. 51, 29–49. doi:https://doi.org/10. 1016/S0167-2681(02)00140-3. Wennekers, Sander, Thurik, Roy A., 1999. Linking entrepreneurship to economic growth. Small Bus. Econ. 13, 27–55. https://doi.org/10.1023/A:1008063200484. Wiklund, Johan, Nikolaev, Boris, Shir, Nadav, Foo, Maw-Der, Bradley, Steve, 2019. Entrepreneurship and well-being: past, present, and future. J. Bus. Ventur. 34, 579–588. https://doi.org/10.1016/j.jbusvent.2019.01.002. World Bank, 2013. Doing Business 2014—Understanding Regulations for Small and Medium-Size Enterprises. World Bank, Washington, D.C.

23