خود اشتغالی، تحصیلات و قانون حفاظت از اشتغال

We analyse how educational attainment and employment protection influence an individual's decision to become self-employed. By altering expected income from dependent employment, employment protection is likely to affect an individual's choice of occupation, although such a link has not been established in the literature so far. We argue that an interaction between an individual's educational attainment and the institution of employment protection exists when it comes to the decision regarding whether to become self-employed. Based on survey data from OECD countries, we find evidence for a negative interaction, and conclude that only after taking this interaction into account can the effect of employment protection and educational attainment on self-employment rates be assessed.

In a large number of countries, self-employment has become an increasingly important policy instrument for promoting employment (see, for example, the strategy Europe 2020 announced by the European Commission, 2010a and Blanchflower, 2004). Accordingly, various measures for promoting self-employment can be observed, such as the programme of “Self-Employment Assistance” in the US 1 or the start-up grant (“Gründungszuschuss”) funded by the German Federal Agency of Labour 2 to encourage transitions from unemployment into self-employment. 3 Moreover, the economic literature on self-employment that has grown steadily over the last few decades has particularly flourished in recent years. 4 On the one hand, a large number of studies have modelled or estimated microeconomic factors influencing occupational choice, i.e., an individual's choice with regard to self-employment, such as age, gender, marital status, and educational attainment (see Parker, 2009, chapter 4 and the references cited therein). On the other hand, the effects of national institutions, particularly labour market institutions, have been scrutinised with respect to their influence on self-employment rates across countries (see Parker, 2009, chapter 17, or Ardagna and Lusardi, 2010). Employment protection legislation is, among other examples, a natural candidate for this latter type of analysis. Economic wisdom suggests that these regulations, while not necessarily affecting unemployment rates in a significant manner, should result in reduced labour market flexibility. This is reflected in less job creation and longer expected unemployment spells, but more secure jobs for those already employed (see, among many others, Bentolila and Bertola, 1990, Bertola, 1999 and Mortensen and Pissarides, 1999). With self-employment constituting an alternative to unemployment (see, for example, von Greiff, 2009), reduced exit rates from unemployment might therefore increase the share of self-employed individuals in an economy. At the same time, in the presence of stringent dismissal regulations, firms might try to circumvent these rules by outsourcing tasks to people who are (maybe only formally) self-employed (see, e.g., Müller and Arum, 2004). In contrast, if self-employed individuals become employers themselves, employment protection might also negatively affect their expected income by limiting discretion regarding employment adjustments. Correlations between the self-employment rates and measures of the stringency of employment protection for different countries presented in the OECD Employment Outlook ( OECD, 1999) are suggestive of a positive link between the two variables. However, more rigorous empirical investigations based on cross-country analyses mostly reject the validity of a positive relationship between self-employment rates and the stringency of employment protection legislation ( Robson, 2003, Torrini, 2005 and Kanniainen and Vesala, 2005). These findings constitute the starting point for our investigation. We argue that occupational choice is affected by the income difference between available employment opportunities. Both expected income from self-employment and expected income earned through dependent employment differ with respect to the level of education an individual has attained. In particular, the effect of employment protection on the difference in income levels is strongly affected by an individual's level of educational attainment. As a consequence, an interaction between individual educational attainment and the institution of employment protection should be taken into account when analysing the occupational choice decision. Furthermore, this may help to shed more light on the relationship between employment protection and occupational choice per se. With regard to the relationship between educational attainment and expected income, van Praag et al. (2009), among others, find that returns to education may differ for workers and entrepreneurs.5 For OECD countries, Blanchflower (2000) reports a U-shaped relationship between education and the probability of being self-employed, with the least educated individuals showing the highest likelihood. This has also been found in a recent paper by Poschke (2008), whereas Kim (2007), for instance, finds a positive relationship between educational attainment and the probability of choosing self-employment. In addition, Thomas (2009) reports a negative association between the likelihood that young labour market entrants aspire to become self-employed and academic success. With respect to employment protection, low-skilled individuals might be more prone than highly skilled workers to lose in terms of expected income (OECD, 2004 and Burda, 2004). This may especially be the case if productivity investments are complementary to skill and employment protection increases firm-specific investment due to a longer expected job tenure (Belot et al., 2007 and Wasmer, 2006). In a different setting, such an interaction between human capital, investment incentives and employment protection legislation is emphasised by Autor (2003) in his investigation of temporary help agencies in the US. Taken together, these findings might indicate that there is an interdependence between educational attainment and the institution of employment protection when analysing an individual's occupational choice, and it is this which we investigate in this paper.6 In the first part of the paper, we develop a theoretical model that outlines the underlying economic mechanisms operating in a framework of occupational choice. A more pronounced negative effect of employment protection on the probability of finding a job for less educated workers and higher returns from increased investment for highly educated workers indicate that employment protection may have a diverging effect on the respective incomes of less educated and highly educated workers in the labour market. Consequently, employment protection is less likely to promote self-employment for high levels of educational attainment. In the second part of the paper, we use data from the European Values Study and the World Values Survey combined with macro data from the OECD Labour Force survey and the IMF World Economic Outlook. Using linear probability and logit models, and accounting for the clustering of individuals in countries, we confirm the main hypothesis derived from the theoretical model. Indeed, the effect of educational attainment on the probability of being self-employed varies for different levels of employment protection legislation. Analysing this interaction effect, we show that stricter employment protection significantly reduces the probability of choosing self-employment for highly educated individuals in relation to the change in the probability for less educated individuals. Furthermore, controlling for the interaction between employment protection and educational attainment, we find for several specifications a significant positive effect of high educational attainment on the probability of an individual being self-employed. The effect of employment protection per se is less robust and most likely influenced by non-linearities. The paper is structured as follows. The model from which we derive our hypotheses is outlined in Section 2. Section 3 presents the data and the descriptive statistics, and is followed by the empirical results in Section 4. Finally, Section 5 concludes.

In this paper, we investigate occupational choice on the part of individuals, namely the decision to become self-employed or to enter the labour market as a worker. We place our focus on the effect that employment protection regulations exert on this decision, emphasising possible interactions with an individual's level of educational attainment. Theoretical considerations alone are unable to generate a clear-cut prediction regarding whether employment protection per se results in lower or higher rates of self-employment. However, disaggregating the working population according to educational attainment reveals that the effects of dismissal regulations may vary according to individuals' levels of attained skills. Our model predicts that employment protection, by reducing job opportunities for individuals while increasing firm-specific investments, might reduce the incentive to become self-employed for individuals with high levels of educational attainment compared to that for individuals with low levels of educational attainment. In the empirical part of the paper, we use data from the European Values Study and the World Values Survey together with additional macroeconomic control variables to test our hypotheses. We use a pooled linear probability model with cluster-robust standard errors on the country level to account for the multilevel character of our data. With regard to the interaction of employment protection and educational attainment, significant effects are shown to exist. Stricter EPL has a relatively more negative effect on the probability of being self-employed for individuals with high educational attainment compared to individuals with low educational attainment. This relation is robust with respect to various specifications and the inclusion of additional control variables. Whether educational attainment increases or lowers the probability of an individual being self-employed is found to be dependent on the stringency of employment protection regulations in the country considered. Our contribution adds to the literature on labour market institutions and self-employment. In particular, it can be concluded that in order to assess the effects of employment protection legislation and education on self-employment rates, the interaction between these variables has to be taken into account.