Types of Institutions and Well-Being of Self-Employed and Paid Employees in Europe

This paper analyzes the role of different types of institutions, such as entrepreneurship-facilitating entry conditions, labor market regulations, quality of government, and perception of corruption for individual well-being among self-employed and paid employed individuals. Well-being is operationalized by job and life satisfaction of individuals in 32 European countries measured by data from EU Statistics on Income and Living Conditions (EU-SILC). We find that institutions never affected both occupational groups in opposite ways. Our findings indicate that labor market institutions do not play an important role well-being. The results suggest that fostering an entrepreneurial society in Europe is a welfare enhancing strategy that benefits both, the self-employed and paid employees.


Institutions, entrepreneurship, and well-being
Institutions play a critical role in determining individual behavior and economic performance (North 1994;Acemoglu, Johnson, and Robinson 2005;Boettke and Coyne 2009;Dixit 2009). This is also true in the emergence of new businesses and the role they play in economic development. In many countries, including the European Union, creating institutional framework conditions that are more conducive to self-employment are well-established on the policy agenda (e.g., European Commission 2010;2016). 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 focus on the well-being of individuals. This paper investigates the relationship between different types of institutions and the well-being of self-employed and dependently employed people. The primary purpose of this analysis is to identify those institutions that are particularly important for self-employment, and to explore the differences in these relationships based on employment status, i.e., between selfemployed and paid employees. The paper offers the following contributions to the extant body of literature. First, while there are a number of studies that focus on the role of institutions for new business formation, 1 there is hardly any evidence about the well-being of the self-employed and paid employees in different institutional environments. 2 Second, we cull out discrete aspects of a country's institutional framework and relate these to an individual's subjective well-being. The results may be regarded as an indication of the importance of the different types of institutions for the welfare of society.
Third, by comparing the effect of different types of institutions on the well-being of self-employed and paid employees, we are able to make statements about whether or not certain institutions affect these two groups differently. This is important, because if an institutional reform would favor people in self-employment, but has negative effects for the well-being of paid employees one cannot be sure that this reform enhances the welfare of society as a whole. If, however, a certain reform is beneficial for both self-employed and paid employees, there will be considerably less resistance as compared to a scenario where the 'losers' of a regulatory modification can be clearly identified. Moreover, differences in the effects of institutions on well-being between self-employed and paid employees may create important incentives or disincentives of being self-employed.
Our empirical analysis uses EU Statistics on Income and Living Conditions (EU-SILC) that provides representative data for households in 32 European countries. We find considerable, and somewhat surprising, differences regarding the impact of diverse institutions on individual well-being. There is, however, no indication that any specific set of institutions affects the well-being of self-employed individuals and paid employees in opposite directions.
This implies that any attempt to make the institutional framework more conducive to entrepreneurship will probably not reduce the well-being of paid employees. Our findings do, however, indicate that an attempt to regulate the market in favor of paid employees, for instance, by introducing stricter regulations of employment contracts, is likely to substantially decrease the well-being of the self-employed without having a notable effect on paid employed individuals.
The remainder of the paper is organized as follows. Section 2 discusses the link between specific institutions and well-being of those involved in entrepreneurship in 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.

Conceptual framework
The institutional framework of a country and its entrepreneurship-facilitating or entrepreneurship-inhibiting character can have strong effects on the incentives to become and to remain self-employed (e.g., Baumol 1990Baumol , 1993Elert, Henrekson, and Stenkula 2017). Since entrepreneurship can be an important driver of economic growth (Fritsch 2013), more entrepreneurship-facilitating institutions may lead to higher levels of economic welfare and the general well-being of a society's members.
There seems to be a wide consensus that high degrees of economic freedom (e.g., low barriers to entry and exit, open markets, low taxes on profits), the opportunity of gaining private property on the means of production, reliable legal framework conditions (e.g., enforceability of contracts, low levels of corruption), availability of necessary resources (e.g., finance, qualified labor), and a good quality of government are conducive for entrepreneurship (see for example Boettke and Coyne 2009;Elert, Henrekson, and Stenkula 2017;Parker 2018). The most prominent institutional frameworks that have been investigated empirically with regard to their importance for entrepreneurship are the regulation of entry and exit, 3 the quality of legal institutions (e.g., protection of property rights), the regulation of employment protection, and the institutional framework of credit markets.
While there are a number of studies focusing on the role of institutions on entry and welfare at the country level, almost nothing is known about the role of institutions for the well-being of entrepreneurs as compared to paid employees. It is also unclear whether institutional reform in favor of entrepreneurship comes at the expense of the well-being of paid employees. Conflicts between the self-employed and paid employees may, for example, arise if la-bor market regulations offer a lower level of employment protection increasing the well-being of the self-employed, at the expense of paid employees who face a greater risk of being laid off.
Our attempt to overcome these shortcomings is twofold. First, we use an individual's subjective well-being that we operationalize by his or her level of job and life satisfaction as an outcome for the effect of institutions. In addition, we distinguish between self-employed and paid employed individuals to assess whether institutions affect persons in these two types of occupation states differently. Second, we distinguish between several categories of institutions to compare their impact on individual well-being and to identify those types of institutions that have the most impact on the two occupational groups.
Many studies find that self-employed people enjoy higher levels of job and life satisfaction than paid employees. 4 A main reason for this result discussed in the literature is higher procedural utility that self-employed people draw from the actual work process itself (Frey, Benz, and Stutzer 2004). This includes higher levels of autonomy and flexibility, as well as a stronger feeling of pursuing one's own goals through self-employment that stimulates a feeling of self-determination and self-efficacy (for a detailed exposition, see Shir 2016). Higher levels of well-being could explain why people opt for selfemployment despite less economic security and often lower incomes than available in paid employment (Benz and Frey 2008). Fritsch, Sorgner, and Wyrwich (2018), in an analysis based on the EU-SILC (which is also used for the present study) discover that self-employed individuals tend to enjoy higher levels of job and life satisfaction only in those countries where the entrepreneurship-facilitating quality of the institutional environment has a certain minimum level. In countries where the quality of institutional conditions for entrepreneurship is below this critical level, paid employees reported, on average, higher levels of well-being. This result clearly indicates the important role of institutions for the attractiveness of entrepreneurship. The study by Fritsch, Sorgner, and Wyrwich (2018) did not, however, investigate which types of institutions are most important in this respect.

Expectations
The 'Varieties of Capitalism' (VoC) approach (Hall and Soskice 1981) is a good starting point for discussing how institutions might affect the well-being of self-employed and paid employees. This approach accounts for complementarities between different categories of institutions and distinguishes several types of institutional frameworks such as the 'liberal market economy' and the 'managed market economy'. Dilli, Elert, and Herrmann (2018) classify countries according to the VoC approach taking into account variations in financial institutions, labor market institutions, institutions related to education, and institutions governing inter-firm relations. They then explore how entrepreneurship-related outcomes vary across these groups of countries. This approach does not, however, allow for the identification of the relative strength of the relationships between certain types of institutions and entrepreneurship. In addition, it does not take into account that the quality of institutions may considerably vary within the country groups. Nevertheless, these authors conclude that labor market regulation (especially employment protection) and regulation of financial markets are particularly important for entrepreneurship outcomes.
Labor market regulations are of key importance for entrepreneurship.
These regulations determine the availability of personnel and employment conditions, such as rules for hiring and dismissing employees, as well as employee benefits such as maternity leave, etc. (Herrmann 2019). An obvious expectation in this respect is that the more freedom an entrepreneur has in his employment decisions, the greater his or her well-being will be. At the same time, greater flexibility comes at the expense of paid employees who might face a higher risk of being laid off, or lower levels of compensation.
Therefore, while one can expect that more flexible labor market agreements will have a positive effect on the well-being of self-employed, they may have a detrimental effect for paid employees. Paid employees may also be affected by high levels of corruption, low quality of government, and weak contract enforcement. However, it can be assumed that they are less directly affected by institutions designed to facilitate business activities than self-employed individuals. Hence, the relationship between these types of institutions and well-being of paid employees should be less pronounced. This can be especially expected for those institutions that are related to starting a business, dealing with construction permits, registering property, getting credit, protecting minority investors, trading across borders, and resolving insolvency.
We expect that the relationship between the level of taxation and wellbeing is more pronounced for individuals with higher income who pay higher taxes, than for low income groups. Although studies show that self-employed individuals do not generally earn more than paid employees (Sorgner, Fritsch, and Kritikos 2017), the effect of taxation on these two occupational groups is undetermined.

Measuring individual well-being and self-employment
Our data source for job and life satisfaction is the EU Statistics on Income year EU-SILC includes an ad-hoc module in its survey program that provides additional information in a selected realm. For this study, we use the 2013 data that includes an ad-hoc module on individual well-being.
We use two indicators of individual well-being that are available in the EU-SILC, namely, the assessment of current overall job satisfaction and the respondent's satisfaction with his or her life as a whole. Life satisfaction is intended to represent a broad, reflective appraisal a person makes of his or her life. It is the by far most frequently used concept of measuring well-being, and has a high level of validation (Pavot and Diener 2008). The variable refers to the respondent's feeling about the degree of satisfaction with his or her life in "these days" rather than specifying a longer or shorter time period. 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, Diener, and Suh 1996;Diener, Inglehart, and Tay 2013).
The While life satisfaction is a rather broad concept, job satisfaction pertains only to issues that are related to a person's work. 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. This could be the case if a poor work environment that offers little satisfaction leads an individual to report lower levels of life satisfaction. There may, however, also be a negative effect of job satisfaction on life satisfaction. For example, a 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 6 The non-response rate in the EU-SILC is rather low. For example, the share of missing values for the variable measuring job (life) satisfaction is 0.6% (0.4%). An analysis of non-responses showed that older individuals, individuals with lower levels of formal education, and those with lower income were more likely not to report their satisfaction with job and life. To test for the presence of a non-response bias, we run the analysis with imputed responses based on the information about the characteristics of the respondent. The results of this analysis were robust. Given a very low share of missing values, we decided to report the results of analyses based on the original real values.
relationships and good health. For this reason, the correlation between the two concepts may be quite low or even negative.
Self-employed individuals are identified in the EU-SILC based on their self-reported current labor market status. An individual is considered self-employed if he or she works full-time or part-time in self-employment to earn a profit. Paid employees are defined as persons who work for an employer and who receive compensation, for instance, in the form of wages or salaries. We construct a binary variable that equals 1 if a person is self-employed, and 0 if a respondent is a paid employee. While we are well aware that self-employ- In addition, we investigate different categories of self-employed individuals, such as income levels, to account for heterogeneity within this group.
It has been shown that the levels of job and life satisfaction someone experiences in self-employment or paid employment varies based on her or his individual characteristics, as well as job-specific characteristics. Education and income levels, personality, motivation and preferences, and the tasks performed at one's job all come into the equation (see Shir 2016, for an overview). To account for these characteristics, our analysis uses the set of socio-demographic variables included in the EU-SILC as control variables, such as age, gender, and marital status. We also use the information about education levels (defined according to the ISCED classification), 7 occupation (defined at a 2-digits level of ISCO-08), 8 industry sector (according to the NACE rev.2), 9 the number of hours usually worked per week in the main occupation, and information on change of job in the previous year.
We also 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 of self-employed persons during a previous 12-month period (such as the previous calendar or tax year) in national currency. 10 We construct country-specific income quartiles to make the income measure comparable between countries. 11 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 provided by the EU-SILC that is measured on a 5-points ordinal scale ranging from 1 (very bad) to 5 (very good).
The final sample contains 161,127 observations. It does not include unemployed or otherwise economically inactive persons, respondents currently in full-time education, those in compulsory military community or service. We also do not consider home workers in our analysis.

Variables representing the entrepreneurship-facilitating quality of institutions
We use several data sources for measuring the quality of different types of institutions in a country. One of these data sources is the Doing of Business Index provided by the World Bank for the year 2013. The Doing of Business score assesses the regulatory performance of more than 180 countries in terms of general business-friendliness. It covers various areas that are relevant for self-employment such as the ease of starting a business, dealing with construction permits, getting electricity, registering property, getting credit, protecting minority investors, paying taxes, trading across borders, enforcing contracts, and resolving insolvency (see Table 1). One may expect that some of the above-mentioned facets of institutions like getting electricity should not be an issue for entrepreneurship in well-developed, high-income countries. However, they may be relevant in low-income European countries. Measures the paid-in minimum capital requirement, number of procedures, time and cost for a small-to medium-sized limited liability company to start up and formally operate in economy's largest business city. Dealing with construction permits All procedures required for a business in the construction industry to build a warehouse along with the time and cost to complete each procedure. In addition, it measures the building quality control index, evaluating the quality of building regulations, the strength of quality control and safety mechanisms, liability and insurance regimes, and professional certification requirements. Getting electricity All procedures required for a business to obtain a permanent electricity connection and supply for a standardized warehouse. These procedures include applications and contracts with electricity utilities, all necessary inspections and clearances from the distribution utility and other agencies, and the external and final connection works.

Registering property
Procedures necessary for a business to purchase a property from another business so that the buyer can use the property for expanding its business, use the property as collateral in taking new loans or, if necessary, sell the property to another business. It also measures the time and cost to complete each of these procedures.

Getting credit
Indicates the legal rights of borrowers and lenders with respect to secured transactions through one set of indicators and the reporting of credit information through another. The first set of indicators measures whether certain features that facilitate lending exist within the applicable collateral and bankruptcy laws. The second set measures the coverage, scope and accessibility of credit information available through credit reporting service providers such as credit bureaus or credit registries.

Protecting minority investors
Protection of minority investors from conflicts of interest and shareholders' rights in corporate governance.

Paying taxes
Measures taxes and mandatory contributions that a medium-size company must pay in a given year as well as the administrative burden of paying taxes and contributions and complying with postfiling procedures. Taxes and contributions include the profit or corporate income tax, social contributions and labor taxes paid by the employer, property taxes, property transfer taxes, dividend tax, capital gains tax, financial transactions tax, waste collection taxes, vehicle and road taxes, and any other small taxes or fees.

Trading across borders
The time and cost associated with the logistical process of exporting and importing goods. It measures the time and cost (excluding tariffs) associated with three sets of procedures-documentary compliance, border compliance and domestic transport-within the overall process of exporting or importing a shipment of goods. Enforcing contracts Time and cost for resolving a commercial dispute through a local firstinstance court and the quality of judicial processes index, evaluating whether each economy has adopted a series of good practices that promote quality and efficiency in the court system.

Resolving insolvency
Time, cost and outcome of insolvency proceedings involving domestic entities as well as the strength of the legal framework applicable to judicial liquidation and reorganization proceedings. We use two indicators to assess the general quality of government in a country. First, the Corruption Perception Index provided by Transparency International ranks countries based on a score indicating the perception of how corrupt a country's public sector is. The Corruption Perception Index is a 1 The Doing of Business Report for the year 2013 covers 185 countries. None of the European countries in our sample represents the frontier for the overall DoB Index. While the UK is among the countries that represent the frontier for the pillar "getting credit", none of the European countries in our sample reaches a score of 100 with regard to the other pillars of the DoB Index (see Figure A1 in the Appendix).
widely used indicator that draws on data sources from independent institutions specialized in governance and business climate analysis. A higher score of the Corruption Perception Index indicates a lower level of perceived corruption in a country's public sector. 3 The second indicator, the European Quality of Government Index, focuses on both perceptions and experiences with public sector corruption, along with the extent to which citizens believe various public sector services are impartially allocated and of good quality. 4 Table A1 in the Appendix provides descriptive statistics of all variables used in the empirical analysis.

Method
In  Table 2 shows the distribution of scores on both satisfaction scales by employment status. We observe that a higher percentage of self-employed individuals (13.58%) are completely satisfied with their jobs, as compared to paid employed individuals (11.82%). At the same time, in our sample there are more self-employed persons (1.82%) who are completely unsatisfied with their jobs than paid employed persons (0.75%). Our analysis shows that when compared to self-employed individuals, a slightly higher percentage of paid employees are strongly satisfied with their lives in general. On average, the self-employed report a significantly lower satisfaction score on both scales compared to paid employees, and there is a stronger variation in the satisfaction scores among the self-employed than among paid employed individuals.   Notes: Satisfaction scales are 11-point Likert scales ranging from 0 "not satisfied at all" to 10 "completely satisfied". t-test of equal means, as compared to the sample of paid employed persons; ***: statistically significant at the 1% level.  Notes: All correlation coefficients except those marked with an " a " are statistically significant at 10% level of significance. Correlation coefficients above 0.7 are in italic. Notes: All correlation coefficients except those marked with an " a " are statistically significant at 10% level of significance.
closely related (r = 0.943), and there are relatively high correlations between the Quality of Government Index and the following pillars of the Doing of Business Index: "dealing with construction permits", "getting electricity", "paying taxes", "trading across borders", "enforcing contracts", and "resolving insolvency".
Remarkably, the correlations between our measures of individual wellbeing and the indicators of employment protection legislation are rather low.
The strongest relationship that we find in this category of institutions is between individual life satisfaction and regulation of temporary contracts (r = -0.097). It is also noticeable that regulations and practices that directly affect starting a business as measured by the Doing of Business pillar "starting a business" are most strongly and positively correlated with the general quality of government. The statistical relationships between the pillar "starting a business" and job and life satisfaction are, however, rather low (r = 0.069 and r = 0.021, respectively).  status. However, the relationship between these indices and job satisfaction is stronger for the self-employed than for paid employees.
To summarize, the correlations indicate a moderate relationship between the measures for the different types of institutions used in our analysis and individual well-being. These relationships are stronger for job satisfaction than for life satisfaction and for the self-employed than for paid employees.

Results of multivariate analysis
This section presents the results of our multivariate analysis. As a first step,

Individual well-being and the institutional environment
We begin our analysis by estimating the model including the Doing of Business Index and its interaction with an individual's employment status (Table   5). The estimated coefficient for the Doing of Business score represents the for the interaction term indicates the extent to which the relationship between the institutional variable and the well-being of self-employed individuals differs from the relationship for paid employees. 1 We find that the overall ease of doing business is positively related to job and life satisfaction of paid employees, but even more so for self-employed respondents. This can be regarded as an indication of a stronger relevance of entrepreneurship-facilitating institutions for the self-employed than for paid employed individuals. This finding is also in line with our expectations (see Section 2.2).
To facilitate the interpretation of this result, we keep all control variables at their mean values and plot the predicted probabilities of being completely satisfied with one's job and life for both employment states based on the observed scores of the Doing of Business Index (Figures 1 and 2). 2 The probability that an average self-employed person living in a country with a low Doing of Business score (60 out of 100, corresponds to Serbia) to report the highest value on the job satisfaction scale is only 4.7%, while it is 8.1% for a comparable paid employed person. This difference, however, is not statistically significant.
Moreover, there is an almost 41% probability that a self-employed person living in a country with a high Doing of Business score (90 out of 100) will be completely satisfied with his or her own job, while this probability is only about 18.3% for a comparable paid employed person (Figure 1). Similar results are observed for the relationship between the Doing of Business score and the probability of being completely satisfied with one's life in general.
However, the predicted probabilities are lower in this case, and there are no significant differences between employment states (Figure 2). 1 The coefficients for the dummy variable that represents the occupational status can hardly be interpreted in a meaningful way. It measures the relationship for the self-employed in the unrealistic case that the institutional variable has the value of zero. Hence, in further analyses we only report the effect of the institutional variable and its interaction with the employment status. Also, see Brambor, Clark, and Golder (2006) for more details on the interpretation of models with interaction terms.  Table 6 shows the relationship between the Doing of Business Index with job satisfaction and life satisfaction, as explained in detail above (see Table 5).  Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job and life satisfaction. 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. All models include a dummy variable for self-employment status (yes=1: no=0), variables for individual characteristics, industry fixed effects, and occupation fixed effects.
We find statistically significant positive coefficients for many of the sub-indices of the Doing of Business score. Pillars measuring the ease of "dealing with construction permits", "getting electricity", "trading across borders", "enforcing contracts", and "resolving insolvency" are significantly positively related to job satisfaction of the self-employed, while the relationship is not statistically significant for the paid employees. A high value for the subindex "paying taxes", which measures the level of tax contributions and the administrative burden of paying taxes, is positively related to job satisfaction for both groups, but the effect is significantly stronger for self-employed respondents. Quite remarkably, no statistically significant relationship is found for the "starting a business", "getting credit", and "protecting minority investors" sub-indices. The reason for this somewhat surprising result may be that these regulations matter more for nascent entrepreneurs than for paid employees or self-employed persons. We find again similar but weaker relationship between the pillars of the Doing of Business Index and individual overall life satisfaction. 4 Turning to the strictness of labor market regulation, we find (and this is quite surprising) that none of the indicators for labor market regulation affects the job and life satisfaction of paid employees. There is, however, the expected significantly negative relationship between restrictions on the use of temporary contracts with the job and life satisfaction of self-employed individuals. To demonstrate this result, Figure 3 shows the predicted probabilities of being completely satisfied with one's job depending on the level of regulation of temporary contracts. An average self-employed individual facing a weak regulation of temporary contracts has a 21.6% probability of being completely satisfied with his or her job, compared to just a 2.4% likelihood if a very strict regulation of temporary contracts applies. For paid employed persons, the probability of being completely satisfied with a job also decreases with the increasing strictness of this regulation, but this decrease is not statistically significant. The probability of being completely satisfied with one's job is highest for selfemployed individuals (the maximum value is 36.6%) if the Corruption Perception Index is very high (corresponding to low a perceived level of corruption).
The probability of being completely satisfied with one's job also increases for paid employees, but at a considerably lower rate. Lower levels of perceived corruption seem to enhance the job satisfaction of self-employed individuals more strongly than paid employees.
All in all, the results clearly suggest that those types of institutions that prove to be statistically significant for job or life satisfaction work in the same direction for both the self-employed and paid employees. None of the institutional variables has an opposite effect on the two groups. There is no indication that a positive effect for self-employed respondents comes at the expense of a negative impact on the well-being of paid employees, or vice versa. In general, the effect is significantly stronger for the self-employed than for paid employed persons. There is no type of institution for which the effect is significantly weaker for self-employed than for paid employed individuals. This suggests that shaping institutions to be more entrepreneurship-facilitating does not necessarily imply a lower level of well-being for paid employees.

What types of institutions are particularly important for individual well-being?
Due to the considerable correlation between some of the measures of the different categories of institutions (see Section 4.1), there is a concern that the results of models including each measure separately may result in overestimating their relationship with individual well-being. To account for this Notes: Results of ordered logit regression. Dependent variable: 11-point scale measuring job and life satisfaction. 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. All models include a dummy variable for self-employment status (yes=1: no=0), variables for individual characteristics, industry fixed effects, and occupation fixed effects.
concern and to shed more light on the relative importance of institutional regulations for individual well-being, we estimate models that include all sub-categories of institutions simultaneously. Specifically, we include the pillars of the Doing of Business Index in the model and we exclude the Corruption Perception Index and the Quality of Government Index, since they do not reflect certain types of regulation. 5 Table 7 reports the results of the model estimations. The effects of institutions become less significant in this model specification. In particular, we find the differences between self-employed and paid employed individuals are less significant, as reflected in the effects of the corresponding interaction terms. There is robust evidence for a strong positive association with individual job satisfaction and entrepreneurship-fostering institutions, particularly the sub-indices of the Doing of Business Index, such as "registering property", "enforcing contracts", and "resolving insolvency". The latter two indices are also positively associated with life satisfaction, while there is a significantly negative relationship between the sub-index "protecting minority investors" and individual life satisfaction. In line with the previous analysis, the different types of institutions are more strongly associated with job satisfaction than with life satisfaction for both the self-employed and paid employees.
In contrast with the previous results, the effect of restrictions on the use of temporary contracts, which was significantly and negatively related to the individual well-being of self-employed persons, is no longer statistically significant. Moreover, the effect of paying taxes, which was found to be significantly positive in the previous analysis, is now not statistically significant for paid employees, and its interaction term with self-employment status is significantly negative. This means that the burden of tax regulation decreases the job satisfaction of self-employed but not of paid employees. Figure 5 presents the predicted probabilities of being completely satisfied

Robustness check: the role of the income level
As a final step of our analysis, we investigate whether our main results differ depending on the individual income level. Relative income level can be regarded as an indication of entrepreneurial success and more productive entrepreneurship (Sorgner, Fritsch, and Kritikos 2017). Fostering successful, productive entrepreneurship is therefore crucial for economic growth (Shane 2009  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. All models include a dummy variable for self-employment status (yes=1: no=0), variables for individual characteristics, industry fixed effects, and occupation fixed effects. estimating our baseline model (as in Table 5) separately for individuals in the lowest (1 st ) and the highest (4 th ) quartiles of the country-specific income distribution. The results are shown in Table 8, which only reports the coefficients of models for job satisfaction, since they are more pronounced as compared to the results for life satisfaction. 6 The results suggest that the ease of doing business is important for the self-employed in both income quartiles, but the effects are stronger for entrepreneurs with high incomes. This indicates that entrepreneurship-fostering institutions might be particularly relevant for successful entrepreneurs.
The results are quite different for paid employees, though. We observe that regulations related to the ease of doing business are mainly relevant for the job satisfaction of respondents in the lowest (1 st ) income quartile while they are almost irrelevant for paid employees in the highest (4 th ) income quartile.
The only exception is the ease of registering property, which is only important for paid employed persons with high incomes. When we examine the effects of strictness of employment regulation, the results are robust for both income quartiles. The strictness of temporary contract regulation is significantly and negatively related to the well-being of 6 The results for life satisfaction can be found in Table A2 in the Appendix. The results are in line with the results for job satisfaction. For instance, the same institutions that are relevant for the life satisfaction of paid employees are also relevant for their job satisfaction. However, we find the differences between both employment types to be less significant in the lowest income quartile. Moreover, institutional variables seem to be more relevant for the life satisfaction of paid employees with high levels of income than for their job satisfaction. The findings clearly show that a country's institutions can have a rather significant effect on the well-being of its population. However, they seem to matter more for well-being among self-employed individuals than for paid employees. An important finding of our investigation is that our institutional variables do not have an opposite effect for individuals in either self-employment or paid employment. In other words, there is never a positive relationship between a specific institutional category and the well-being of self-employed persons, but a negative effect on the well-being of paid employees, or vice-versa. This also holds for institutional categories where one could expect to find such opposite effects, like labor market regulations of individual and collective dismissals or regulations of temporary employment. This result suggests that introducing more entrepreneurship-facilitating institutions does not come at the expense of the well-being of paid employees. Hence, promoting an institutional framework for a more entrepreneurial society appears to be a strategy that benefits both the self-employed and paid employees.
The three global indicators for a country's institutional framework that we use in our analyses, the Doing of Business Index, the Corruption Perception Index, and the Quality of Government Index are closely correlated and have a highly significant impact on the self-employed as well as on paid employees. According to the empirical estimates, the most important single types of institutions are "enforcing contracts", "trading across borders", and "dealing with construction permits". While "enforcing contracts" describes the liability of the legal system, "trading across borders" indicates the openness of an economy and exposure to international competition. "Dealing with construction permits" may be regarded as a measure of the general density of regulation and the efficiency of a country's public administration.
An unexpected result of the analysis is that the labor market regulations, particularly restrictions for temporary contracts, do not seem to play a significant role for the job and life satisfaction of paid employees. This result is quite remarkable given that many authors assume that labor market regulation plays a key role for entrepreneurship (see Herrmann 2019). Restrictive labor market regulation shows, however, the expected negative relationship with the well-being of the self-employed. Another surprising finding is that the Doing of Business Index sub-index "ease of starting a business" is not significantly related to the well-being of either the self-employed or paid employees. A possible explanation for this result may be that a large majority of the self-employed respondents own well-established businesses that have been operating for a long period of time. Hence, the effort of starting a business is no longer relevant for their current well-being.
Our analysis is not without limitations. Overcoming these limitations presents promising avenues for further research. First, the analysis is based on a pure cross-section due to data limitation. Hence, we are unable to assess the dynamics of institutional change that could better allow for the iden-