2.1 Unemployment and self-employment
Entrepreneurship occurs primarily at the local/regional level, since actors interact with local conditions that form the regional entrepreneurial environment, which can be diverse from the entrepreneurial environment of other regions (e.g., see Barreneche García
2014; Bosma and Schutjens
2011; Feldman
2001; Fritsch and Storey
2014; Georgellis and Wall
2000; Konon et al.
2018; Saridakis et al.
2020). For instance, and as Georgellis and Wall (
2000) explain, regions differ in their suitability for entrepreneurship (relative to paid employment) which is attributed to factors like industry composition. Other factors, like the local human capital or the market potential, are also responsible for differences among regions (see also Wyrwich
2014). Moreover, notice that different regions exhibit different levels of self-employment persistence, as for example has been reported by Fotopoulos and Storey (
2017) for the case of England and Wales, and by Fritsch and Wyrwich (
2014) for the case of Germany, shifting the analysis from the national level to the regional one.
Among other factors, such as the stock of human capital, the regional wealth, and industry composition, an important determinant of regional entrepreneurship is the level of regional unemployment (see Müller
2016). Research on the relationship between unemployment and entrepreneurship, however, has always been a complex task to handle, since there are strong theoretical and empirical arguments supporting both the ‘recession-push’ and the ‘prosperity-pull’ effects (see Audretsch et al.
2015; Storey
1991). The ‘recession-push’ effect relies mostly on poor aggregate conditions and increased unemployment spells that impede individuals from obtaining paid employment, thus forcing them to consider self-employment instead. Other factors, such as the access to cheap labor (see Konon et al.
2018) or equipment (see Parker
2018, p. 267) during a recession, can also encourage self-employment. On the other hand, the ‘prosperity-pull’ effect argues for the expected returns from entrepreneurship. Although during times of economic growth employment prospects are better and higher wages increase the opportunity cost of self-employment, it can also be expected that in a booming economy individuals can be pulled by opportunities due to higher product demand and thus consider self-employment. On the contrary, under an economic regime of falling incomes and rising unemployment rates, transitions to self-employment are less probable to occur, since, generally, during economic downturns, firms, regardless of their size, tend to face lower demand for their products. Thus, increased unemployment can signal less favorable market conditions and therefore discourage entrepreneurship (see Fotopoulos
2014).
In terms of the effect of regional unemployment on regional self-employment, evidence is mixed. For instance, in an earlier study for UK regions, Georgellis and Wall (
2000) have found both push and pull factors that were subject to the level of unemployment in each region. As such, the authors argue for nonlinear effects. Similarly, and by examining business formation, Hamilton (
1989) also argues for the nonlinear effect of unemployment (see also Storey
1991). Robson (
1998), on the other hand, did not provide evidence in favor of the ‘recession-push’ effect in UK regions. Cueto et al. (
2015), using Spanish data, found that self-employment decreases in a region when the unemployment rate increases, but increases when the unemployment rate in neighboring regions increases. Golpe and van Stel (
2008) found a positive effect of unemployment on self-employment for higher-income regions in Spain, while no similar effect was found in lower-income regions.
In a different context, Bosma and Schutjens (
2011) examined entrepreneurial activity and entrepreneurial attitude in 127 regions of 17 European countries during 2001–2006 and found that increased unemployment negatively affects perceived business opportunities, and, as their regression results suggest, to some extent established business ownership (firms operating more than 3.5 years) as well. Finally, an important contribution to the relevant literature, while at the country level, is that by Parker and Robson (
2004). The authors examined a sample of 12 OECD countries between 1972 and 1996 and found no evidence for a significant interaction term between unemployment and the replacement rate (unemployment benefits). Their contribution, however, lies on the underlying potential mechanism implied by the unemployment rate and an institutional variable, the latter thought to be able to mediate the effect of unemployment on self-employment.
Notice however that the above-mentioned theoretical and empirical considerations offer a general explanation of the possible effect of unemployment during boom-and-bust periods, whereas it does not account for their magnitude, which could be extremely useful when examining severe recessionary periods. That essentially means that our understanding of how unemployment might push into self-employment depends heavily on the period under examination (e.g., see Congregado et al.
2012).
Fritsch et al. (
2015) identify such a mechanism when reporting for Germany (during 1996–2010) that although the relationship between unemployment and entrepreneurial entry rates is counter-cyclical, no significant (push) effect is found under unemployment above the trend. Konon et al. (
2018) provide similar evidence for German regions for the period 1996–2008, but they also notice that it would be important to consider the effect of more severe economic downturns, such as the one following 2008. Therefore, even if both studies examine the case of Germany which performed relatively well during the Great Recession, they also refer directly to diverse implications under profound economic hardships.
Those implications could be more evident in labor markets that were vigorously hit during the Great Recession. For instance, Novejarque Civera et al. (
2020) used data for Spanish regions and found a consistent negative effect of unemployment on firm creation rate, before and after 2008, as well as after controlling for the crisis at the national level. Another example is USA, where there has also been considerable effort to understand how the Great Recession affected entrepreneurial activity (e.g., Fairlie
2013; Fossen
2020). Generally, and as regards the trend in the self-employment rate during 2008–2015, it can be argued that there was a decrease in self-employed with entrepreneurial incentives (see Fondeville et al.
2015; Eurofound
2017), but it must also be noticed that there were heterogeneous trends in the EU member-states. Namely, there were countries that experienced an increase in self-employment (e.g., the Netherlands, UK), others a decrease (e.g., Portugal), while some others only marginal changes (e.g., Austria, Denmark) (see Eurofound
2017). Of course, those heterogeneous trends are also evident within each country, since there are regions diverging from the general patterns pertaining to the country they belong to.
While the inconclusiveness of the relevant literature is obvious, which however depends on factors like the micro- or macro approach, or the labor market under examination, recent studies agree on the unique role played by the magnitude of the Great Recession. Moreover, and as explained by Pissarides (
2013), the Great Recession was not a ‘normal’ recession, especially for the European South, which was depicted in the severe rise of unemployment and the sharp fall in aggregate demand.
Thus, considering both the impact of the recent recession on the European labor markets and the suggestions on the role of a severe recession on entrepreneurship (Fritsch et al.
2015; Konon et al.
2018), it seems rational to assume that in Europe while self-employment could have acted as a last resort, the implied lower product demand following the extensive job destruction in that period could not have made entrepreneurship attractive. Moreover, and although the individual incidence of unemployment can be an important determinant pushing into self-employment, the degree of aggregate unemployment of the local labor market cannot be neglected since it can negatively affect self-employment, as the evidence of Berglann et al. (
2011) suggests. In addition, the evidence provided by Georgellis and Wall (
2000) suggests that in regions with already increased unemployment, additional increases can discourage self-employment and lead to prosperity-pull implications. The above-mentioned considerations indicate that poor aggregate conditions lead to an unfavorable environment for self-employment, especially in regions hit most during the Great Recession (see Novejarque Civera et al.
2020), given the rise of unemployment and the subsequent fall in aggregate demand (Pissarides
2013).
2.2 Employment protection, unemployment, and self-employment
Apart from the effect of factors such as the regional unemployment rate, or other endogenous factors, scholars often argue that there is a role for the institutional context as well when conducting a regional analysis. For instance, Agostino et al. (
2020) explain that there is a trade-off between safe-salaried employment and higher and more volatile returns from entrepreneurial activity, which is affected by the institutional quality. The authors further explain that in a non-crisis context higher institutional quality decreases the risk associated with entrepreneurship, while in times of crisis the already risky choice of entrepreneurship might not be an option even under good institutions. Moreover, and as argued by Bosma et al. (
2018), institutions can determine under what conditions entrepreneurs can utilize inputs such as labor, finance, and knowledge to generate output (see also Agostino et al.
2020).
However, and apart from the role of institutions on entrepreneurship in general, we must also examine whether national institutions exert any kind of influence on the regional entrepreneurial activity (see Bosma and Schutjens
2011; Wyrwich
2014). For instance, Wyrwich, (
2014) explains that the institutional shock related to the transformation of post-communist regions into market economies had a positive impact on start-up activity. Audretsch et al. (
2019), and although focusing on the country and city levels, argue that there are indeed (business) regulations governing all entities, which of course shape entrepreneurship, which occurs locally and comes with certain local constraints. Similarly, Di Vita (
2018), and in a different context, argues that while regions have their own economic and institutional individuality, we cannot ignore that they also have to comply with the national laws (again in a different context, see Geppert and Stephan
2008 for accounting for national effects in their regional analysis). Elhorst and Zeilstra (
2007) also criticize the fact that some studies focusing on regional-level variables essentially treat regions as independent entities.
Therefore, the institutional implications brought about by national regulations may affect differently regions even within the same country and thus more attention must be paid. In particular, and within the context of the present study we turn our attention on employment protection. Apart from the effect of unemployment, considerable attention has been paid on the role of rigid labor markets, especially given that employment protection is an important determinant of self-employment (Hipp et al.
2015). Of course, the effect of employment protection may be diverse subject to the level of strictness of the relevant legislation, as shown by Román et al. (
2013), who examine the probability of entering self-employment from unemployment under the occurrence of additional increases in employment protection (both in flexible and rigid labor markets). In addition, and as with the unemployment rate, evidence on the effect of employment protection legislation is, again, mixed.
Some authors argue that increased employment protection could negatively affect self-employment (e.g., Kanniainen and Vesala
2005). In this sense, stricter regulations in hiring and firing are often considered to be an obstacle for employers in deciding how much labor is needed for firms to operate, and thus discourage entrepreneurship. Moreover, strict hiring and firing regulations can also provide disincentives to potential self-employed who know that, in the case of business failure, it can be hard to return to paid employment (Golpe et al.
2008). Van Stel et al. (
2007) explain that, under flexible markets, job insecurity can push employees into entrepreneurship, while employers enjoy this labor market flexibility in order to run their businesses. Thus, and according to the authors, increased regulatory rigidity decreases employer flexibility and increases the costs associated with potential dismissals. Notice that this decreased flexibility, and the increased labor costs associated with it, can lead to less employment given the reluctance of employers to create new jobs (Kahn
2007).
On the other hand, the direct effect of a rigid labor legislation that induces the slanginess of the labor market and impedes job creation, could be also expected to lead to push effects (e.g., see Román et al.
2013). A positive relationship can thus emerge in tighter labor markets where alternative forms of employment (i.e., dependent self-employment) act as a counterweight to strict employment protection legislation (Parker
2010; Román et al.
2011b,
2013; Ulceluse and Kahanec
2018). Moreover, self-employment could act as a last resort. The rationale is that employers cannot easily dismiss the redundant labor, while simultaneously limiting themselves from hiring more, thus pushing individuals who cannot obtain paid employment into self-employment.
Nevertheless, it is also important to consider the existing economic conditions when investigating the effect of labor market flexibility, since a recession and strict employment protection may favor dependent self-employment, or self-employment as a last resort (i.e., employers do not hire labor both due to the recession and the unfavorable legislation). In fact, Román et al. (
2011a) argue that the coexistence of a recession and strict employment protection may lead to dependent self-employment, or self-employment as a last resort. In view of the above, and taking into account the magnitude of the Great Recession and the insight provided by Román et al. (
2011a), we proceed to the next hypothesis:
Notice, however, that much less effort has been made by the relevant literature to identify connections between unemployment and employment protection that could interact and jointly explain entrepreneurship, despite the existence of studies that somehow incorporate, or at least mention, both factors in their analyses. For instance, Bosma and Schutjens (
2011) include both regional unemployment and employment protection legislation in their analysis; however, as these are probably outside their research scope, no joint effects are examined.
In addition, the existing literature that employs relevant interactions remains scarce and in a different context. Parker and Robson (
2004), for example, examine the interaction between the unemployment rate and the replacement rate, which however accounts for the benefits of unemployment and only indirectly, and loosely, accounts for its relevant implications on workers. For the record, the authors find no evidence supporting the subsequent push effect of unemployment. In a different context, Fu et al. (
2018) examine whether labor market regulations could affect ex-entrepreneur re-entry into entrepreneurship, while accounting for their current work status (employed or unemployed). The authors find evidence supporting that currently employed ex-entrepreneurs are more likely to re-enter entrepreneurship, than unemployed ones, attributing this finding to the fact that, while employed under strict labor regulations, employed individuals enjoy the legal safety nets and are less financially constrained.
Another shortcoming is that many studies focus on the individual level. Individual-level transitions from unemployment to entrepreneurship, however, can be hindered due to human and financial capital constraints (Parker
2018, p. 268) and thus not directly or solely related to the level of employment protection and/or aggregate conditions. As such, studies like that by Baumann and Brändle (
2012) are more informative since they account for the interaction of employment protection and educational attainment. In addition, evidence shows that while unemployment might be positively related to entrepreneurship at the individual level, the relationship can be negative at its aggregate counterpart (Berglann et al.
2011). That is because the latter is an indication of the aggregate prevailing conditions within a region/country’s labor market, which, in conjunction with the strictness of employment protection, manifests the level of labor market rigidity (Barbieri
2001).
Moreover, it is one thing to interpret the effect of unemployment and employment protection independent to each other in an analysis, but quite another to examine whether the joint effect of these two measures can explain self-employment, especially in the context of a recession. Strikingly enough, this is overlooked by the related empirical research, especially since evidence from the Great Recession shows the potential connections existing between unemployment and employment protection. Evidence, for example, shows that the unemployment rate in Germany during the Great Recession was held at lower levels compared to the rest of European economies, but the overall performance of the German labor market was enhanced by (prior) reforms that aimed to stimulate job search, self-employment and deregulate the job market (Ter Weel
2015; Weber
2015; Pissarides
2013). In this sense, Germany’s prior reforms helped mitigate the adverse unemployment effects of the recession, while countries like Greece and Spain trying to implement fiscal and labor market reforms at the same time were caught in a phase characterized by both a sharp decline in aggregate demand and rigid labor institutions, eventually leading to more unemployment (Pissarides
2013). It is thus evident that labor market institutions (and the reforms related to them), are closely related to the diverse response and impact of unemployment of different economies to the recession.
Given that no direct evidence is as yet available that shows any potential link between increased unemployment and stringent labor legislation that could affect self-employment during a recession, the possibility to build hypotheses based on previous findings and/or theoretical arguments is, at best, limited. Nevertheless, the studies that somehow formally consider both factors, do not provide (strong) evidence in favor of a joint push effect of unemployment and employment protection. As such, we first build on Barbieri (
2001) who examined whether self-employment in Italy can be explained by the unemployment rate and employment protection, both high at the time, thus signaling labor market rigidity. The author, however, found no evidence that this labor market rigidity could explain self-employment movements. Second, we consider that Román et al. (
2013) find evidence that increased national unemployment and more stringent labor legislation in countries with already strict EPL discourage self-employment with employees, but impel own account self-employment. The latter, however, was found to be only weakly statistically significant according to the regression results. In addition, we consider that i) high aggregate regional unemployment signals less promising entrepreneurial prospects (Bosma and Schutjens
2011; Georgellis and Wall
2000; Novejarque Civera et al.
2020), and ii) strict labor regulations lead to increased labor dismissal costs (van Stel et al.
2007), jointly implying an overall labor market rigidity (Barbieri
2001) (i.e., product markets are in a downturn and firing and hiring comes with increased legislative obstacles), and thus, at the end of the day, ‘prosperity-pull’ implications (Audretsch et al.
2015; Storey
1991). We thus proceed to Hypothesis
3: