key: cord-1005862-yw2k9pte
authors: Stephan, Ute; Tavares, Susana M.; Carvalho, Helena; Ramalho, Joaquim J.S.; Santos, Susana C.; van Veldhoven, Marc
title: Self-employment and eudaimonic well-being: Energized by meaning, enabled by societal legitimacy
date: 2020-09-09
journal: J Bus Ventur
DOI: 10.1016/j.jbusvent.2020.106047
sha: e3027cc3826443db8d73494132bd05775e09b06a
doc_id: 1005862
cord_uid: yw2k9pte

This study investigates why and where self-employment is related to higher levels of eudaimonic well-being. We focus on meaningfulness as an important eudaimonic process and subjective vitality as a eudaimonic well-being outcome that is central to entrepreneurs' proactivity. Building on self-determination theory, we posit that self-employment, relative to wage-employment, is a more self-determined and volitional career choice, which enhances the experience of meaningfulness at work and perceptions of work autonomy. In a multi-level study of 22,002 individuals and 16 European countries, meaningfulness at work mediates the relationship between self-employment and subjective vitality and explains this relationship better than work autonomy. We identify moderating effects of context: the societal legitimacy of entrepreneurship in a country affects the choice set of alternative career options that individuals can consider and thus shapes the experience of meaningfulness at work and work autonomy, and thereby indirectly subjective vitality. These findings expand our understanding of eudaimonic well-being, entrepreneurs' work, and the role of context in entrepreneurship and well-being research. They complement existing research on hedonic well-being of entrepreneurs and extend the scarce literature on their eudaimonic well-being.

The share of self-employment in the workforce is growing and currently stands at 15.54% in the European Union (OECD, 2019 ). Yet, the economic returns for the individual self-employed or entrepreneur 3 are often limited. The self-employed may not earn more than in wage employment (van Praag and Versloot, 2007) , but they may find their work more fulfilling and may be happier (Nikolaev et al., 2020) . This has increased researchers' interest in well-being as a benefit that individuals may derive from self-employment (Stephan, 2018; Wiklund et al., 2019) . Indeed, entrepreneurs often regard their well-being as a marker of their success (Wach et al., 2016) .

Two types of well-being can be differentiated. Eudaimonic well-being is associated with activation (energy and vitality) and focuses on self-realization and meaning (Ryan and Deci, 2001) , whereas hedonic well-being 4 is more passive and focuses on attaining pleasure and avoiding pain (Ryan and Deci, 2001) . For instance, life and job satisfaction, which reflect a sense of positive contentment with one's life and job, are frequently used measures of general and work-related hedonic well-being in entrepreneurship research (Stephan, 2018; Wiklund et al., 2019) . Such research finds that entrepreneurs tend to have higher life and job satisfaction than the wage-employed Coad, 2013, 2016) including in cross-country cross-sectional (Blanchflower, 2000; Nikolaev et al., 2020) and longitudinal studies (Hamilton, 2000) . These findings are typically explained by the higher work autonomy that entrepreneurs enjoy (Stephan, 2018) . In other words, their ability to independently decide how, when and with whom to work (Parker, 2014) . Economists suggest that this independent way of working provides 'procedural utility' -utility from the process of work rather than its outcome -which may help explain the higher job satisfaction of the self-employed Frey, 2004, 2008) .

The near neglect of eudaimonic well-being in entrepreneurship research is surprising and limiting (Ryff, 2019; Stephan, 2018; Wiklund et al., 2019) . It is surprising because there is much about being self-employed that should specifically increase eudaimonic well-being. From the self-determined choice to pursue such a career in the first place, to the opportunity to express one's identity through work, shape work in line with one's values, skills and needs, and the personal commitment to and legal responsibility for all work-related decisions and actions (Baron, 2010) . Hence, self-employment, more than wage employment, offers unique possibilities only considered traditional wage-employment may be more likely to also consider becoming self-employed. Thus, following selfdetermination theory we explore if and how the societal legitimacy of entrepreneurship shapes (moderates) the relationship of selfvs. wage-employment with eudaimonia.

We explore our predictions in a multilevel study of 22,002 individuals in 16 European countries controlling for alternative explanations and conducting robustness checks for self-selection. We find that the self-employed experience their work as more meaningful, as having more work autonomy, and they exhibit higher eudaimonic well-being (subjective vitality) than wage employees. The experienced meaningfulness at work, rather than work autonomy, is the main mediator of the positive effect of selfemployment on subjective vitality. We also identify moderating effects of country context: the legitimacy of entrepreneurs among a country's population leads the wage-and self-employed to experience their work in different ways and through this impacts their vitality.

Our study contributes to research on entrepreneurship and well-being. First, it complements existing research on entrepreneurs' well-being which has focused on hedonic well-being, has established differences in hedonic well-being between entrepreneurs and wage employees, and has identified work autonomy as a key driver of these differences (see Stephan, 2018) . It responds to calls to consider entrepreneurs' eudaimonic well-being (Ryff, 2019; Wiklund et al., 2019) by building on and extending the scarce existing research on eudaimonia and entrepreneurship. Previous research has examined context-free general eudaimonic processes and related them to measures of mainly hedonic well-being outcomes (Nikolaev et al., 2020; Shir et al., 2019) . By contrast, our study provides a theoretical framework and evidence that helps to understand why and how work-related eudaimonia differs for entrepreneurs and wage employees. Thereby, it highlights meaningfulness at work as a hitherto overlooked benefit of entrepreneurs' work, which we find is more consequential for eudaimonic well-being (and possibly other outcomes) than work autonomy.

Second, our study advances a context-sensitive understanding of entrepreneurs' well-being. Country differences have been documented for engagement in entrepreneurship and for entrepreneurs' well-being (Amoròs and Bosma, 2014; Benz and Frey, 2008) . However, research on how particular aspects of context may impact well-being is lacking (Stephan, 2018) . We start to develop such theory by integrating insights from self-determination theory (SDT) about contexts enabling self-determined (career) choices (Deci et al., 2017; Katz et al., 2018) with research on the relevance of cultural norms and, in particular, the societal legitimacy of entrepreneurship (Kibler et al., 2014; Stephan and Uhlaner, 2010) . Doing so advances the understanding about normative influences on volitional career choice through impacting the choice-set of both, the wage-and self-employed, as well as about the significance of volitional career choice for the subsequent experience of work (meaningfulness and autonomy). It also expands SDT by newly introducing social normative pressure (societal legitimation) as an element of context that shapes self-determined choice.

Our study also offers implications for the literature on meaningfulness at work. It identifies the employment setting as a predictor of meaningfulness at work, and self-employment as a type of work with a high potential for meaningfulness, which has not yet been considered in this literature. This is an important contribution in light of emerging new forms of work that share features of selfemployment (e.g. in the gig economy). Moreover, we offer new insights on how meaningful work can be shaped by societal contexts that past reviews on meaningful work have called for (e.g., Bailey et al., 2019) . Drawing on self-determination theory, our findings highlight the relevance of the societal legitimation of alternative career options (including less commonly chosen careers such as selfemployment) to enable everyone to make truly volitional choices about their career. This complements existing explanations of how context affects meaningfulness at work, which emphasize congruence with societal values and norms (Florian et al., 2019; Lepisto and Pratt, 2017) rather than enhancing the diversity of choice sets.

Entrepreneurship is an occupational choice to work for one's own account and risk (Hébert and Link, 1982) . This definition of entrepreneurship as self-employment includes solo-workers, business owners or working for oneself (e.g., OECD, 2019). It is common in research on entrepreneurs' well-being (Stephan, 2018) and consistent with our interest in contrasting the self-versus wage employed . Additionally, we explore differences between necessity and opportunity self-employed in one robustness check.

Different theoretical models of eudaimonic well-being offer varying and rich sets of constructs to characterize the fully functioning human being (Ryan and Deci, 2001; Ryff, 1989; Waterman et al., 2010) including fulfilment of psychological needs of autonomy, competence and relatedness, purpose in life, self-acceptance, personal growth, environmental mastery, autonomy, positive relationships with others, self-discovery, etc. Some constructs have been argued to represent sources of well-being rather than well-being itself (Diener et al., 2018) . Hence, as outlined in Section 2, we focus on subjective vitality as a key indicator of eudaimonic well-being and the outcome of eudaimonic processes (Ryan and Deci, 2001; Ryan and Deci, 2008) .

Subjective vitality is "one's conscious experience of possessing energy and aliveness" and having positive energy for "regulatory control of one's self" (Ryan and Frederick, 1997, p.530) . It entails "the activation of physical and mental resources" (Schmitt et al., 2017, p. 444 ) and arises from engaging in self-determined choices and actions that involve a sense of personal agency Ryan and Frederick, 1997) . Subjective vitality is also referred to as vigor (Shirom, 2011) , energy (Fritz et al., 2011) , calm energy (Thayer, 1996) or energetic activation (Quinn et al., 2012) informed by Ryan and Deci (2008, p.703 ) who summarize vitality as "positively toned, energized state", that "one can harness or regulate for purposive actions".

Subjective vitality has been researched in organizational behavior due to its performance benefits. Subjective vitality provides resources needed for motivational processes (Baumeister et al., 2007) , for one's engagement in action and in approach-oriented behaviors (Shirom, 2011) such as creativity at work (Kark and Carmeli, 2009 ) and other discretionary activities (Quinn et al., 2012) as well as performance (Shirom, 2011) . Subjective vitality is also associated with positive long-term health outcomes, including longevity (Chida and Steptoe, 2008) . People high in subjective vitality can mobilize their resources better to deal with disease and fatigue (Muraven et al., 2008) , are more resilient, and appraise their personal problems less negatively (Thayer, 1996) .

While subjective vitality within organizations is well understood, research on entrepreneurs' subjective vitality is in its infancy. As outlined above, the initial career choice as well as the nature of work, including constraints and challenges, are different for the selfand wage employed, not allowing us to generalize findings on vitality from the wage-to the self-employed. The lack of research on entrepreneurs' subjective vitality is surprising as vitality is arguably crucial for engaging in self-starting proactive and innovative behaviors (Gorgievski et al., 2014) that are key to the performance of entrepreneurs' businesses (Campos et al., 2017; Rooks et al., 2016) . Subjective vitality should also equip entrepreneurs with the energy to persist and overcome barriers and to assemble resources to explore the environment for new opportunities (Hahn et al., 2012) . In line with these arguments, as part of the broader concept of work engagement, vigor has been found to be higher among the self-than the wage-employed in one cross-sectional study in the Netherlands (Gorgievski et al., 2010) . Moreover, a two-year study of 122 German business owners found that those who scored higher on vitality at the beginning of the study (measured as vigor) showed more task-and relation-oriented proactive behaviors at the end of the study (Hahn et al., 2012) . In the same study, life satisfaction, an indicator of hedonic well-being, had no effect on proactive behavior. This illustrates the importance of distinguishing eudaimonic from hedonic well-being and the need to complement existing research with investigations of eudaimonic well-being (Stephan, 2018; Wiklund et al., 2019) .

Having established why subjective vitality is an important aspect of well-being consequential for entrepreneurs' health and performance, we now turn to our research model and hypotheses, presented in Fig. 1 . These seek to understand how self-employment relates to subjective vitality.

According to self-determination theory (SDT, Deci and Ryan, 2008) , self-regulated activity is the result of an individual's volitional choice (Moller et al., 2006) . It stems from the self, reflects the self and feels authentic (Ryan, 1992) . Higher levels of volitional self-regulation entail a sense of personal agency that is vitalizing Ryan and Frederick, 1997) . Research finds positive associations between volitional self-regulation of one's actions and subjective vitality (Kasser and Ryan, 1999; Nix et al., 1999) . In contrast, non-self-determined actions (termed externally controlled action), are experienced as demands to feel, think or behave in specific ways (e.g., prescribed by social norms) which can deplete individuals' energy and vitality (e.g. Nix et al., 1999; Ryan and Frederick, 1997) .

Self-employment relative to wage employment, is a more self-determined volitional career choice. Although self-employment is on the rise in most developed economies, including in Europe, it is by no means the 'norm'. Rather, wage employment is the default option and few people choose to become self-employed (Lechmann and Schnabel, 2014 ). In fact, the self-employed remain a minority in the workforce (typically 10 to 20%, OECD, 2019). 7 Following SDT and considering that the career decision for self-employment 7 While we emphasize the positive choice for self-employment, we acknowledge that some individuals also pursue self-employment out of necessity to escape unemployment. We control for unemployment and offer robustness checks for opportunity-necessity entrepreneurship in the U. Stephan, et al. Journal of Business Venturing 35 (2020) 106047 typically entails greater volitional choice than the career decision for wage employment, we propose:

Hypothesis 1 (H1). Self-employment is positively related to subjective vitality.

Meaningfulness at work (MW) is the experience of work as personally enriching and useful. It is a positive experience where work is viewed as contributing to one's personal growth (Steger et al., 2012) , as significant (i.e., important to the individual), and giving the individual a sense of purpose (Pratt and Ashforth, 2003) . Lepisto and Pratt (2017) propose to differentiate two perspectives on meaningful work, which they label "realization" and "justification". The realization perspective frames meaningfulness as the result of the fulfilment of needs, motivations, and desires at work, relating to self-actualization. It points to working conditions as constraints on the ability of the self to "being or becoming fully expressed and realized in one's work" (Lepisto and Pratt, 2017, p.106) . Thus, MW mediates the impact of work characteristics on well-being (Hackman and Oldham, 1975) . By contrast, the justification perspective relates MW more fundamentally to the worthiness of work. It represents meaningfulness as the worth and value one attributes to one's work, answering the question "why is my work worth doing?" (Lepisto and Pratt, 2017, p. 108) . This perspective reflects the meaning-making processes individuals engage in to justify the worthiness of their work, which goes beyond the specific characteristics of a job.

We suggest that these two perspectives can help to understand differences in MW between employees and the self-employed. The realization perspective (Rosso et al., 2010 , for a review) may be especially applicable to understand how wage-employees derive meaning from their work. Wage-employees must align with the goals and constraints determined by the employing organization and often have little opportunity to craft their job in line with their values, needs and skills (especially compared to the self-employed). Therefore, for wage-employees, MW might depend mainly on their job characteristics. The justification perspective is especially helpful to understand the meaning-making processes involved in self-employment. For the self-employed, the process of attributing meaning to their work precedes the design of their job. Rather it starts with the choices they make about the job (a job they consider worth doing) and the work setting that they are going to develop for themselves. For the self-employed, MW may also derive from the dayto-day characteristics of their work (realization perspective) but it is even more deeply rooted in their initial volitional career choice.

The volitional choice to be self-employed gives the self-employed uniquely the opportunity to shape and design their work in line with their values, needs, and skills into work they deem worth doing and which results in high congruence (or 'fit') between work and the self (Baron, 2010 , also for job crafting, Wrzesniewski and Dutton, 2001) . This means that work gives the self-employed the possibility to express their identity and themselves authentically (Baron, 2010) , which are prerequisites for experiencing work as meaningful (Lips-Wiersma and Morris, 2009; Rosso et al., 2010) . In other words, work as a way to express the self and a source of intrinsic motivation leads to a greater sense of purpose and significance, i.e. meaningfulness (Allan et al., 2016; Duffy et al., 2015) . Thus, work for the self-employed compared to wage employees is richer in meaning (Cardon et al., 2009) .

Self-employment is positively related to meaningfulness at work.

Evidence from studies of employees corroborates that the experience of meaningfulness at work (MW) can be an important psychological resource (Niessen et al., 2012; Soane et al., 2013; Spreitzer et al., 2005) , which can fuel one's well-being (Allan et al., 2019; Steger et al., 2012; Tavares, 2016) and work engagement (Lips-Wiersma and Wright, 2012; May et al., 2004; Soane et al., 2013) . However, few studies have related MW to the energetic resources of the individual such as vitality and, to our knowledge, no studies explored this relationship in the self-employed. Niessen et al. (2012) found that when employees experience MW in the morning they tended to feel higher levels of vitality by the end of the workday. Lam et al. (2016) observed that employees who experience a sense of MW at the end of the workday were more likely to find their job energizing and to report higher levels of vigor at the end of the day. Likewise, reflecting on the meaning of one's work was related to feeling energized at work (Fritz et al., 2011) . Thus, the experience of MW likely enhances subjective vitality. To the extent that the self-employed experience more MW (H2a), MW may mediate the relationship between self-employment and vitality.

Meaningfulness at work is positively related to subjective vitality.

Meaningfulness at work mediates the relationship between self-employment and subjective vitality.

As outlined in the Introduction, research on entrepreneurship and well-being has highlighted work autonomy as an important resource for entrepreneurs' well-being. Indeed, it was the most frequently studied work resource in Stephan's (2018) review of this literature. Self-employment is not only a more self-determined career choice, it also offers individuals day-to-day work autonomy. In other words, the chance to make ongoing independent decisions about how, with whom and when they work (Parker, 2014) . Compared to wage employees, the self-employed, as own account workers or business-owner managers, have fewer constraints on their decision-making freedom and on how to organize their work.

(footnote continued) results. The majority of self-employed report opportunity motives including in the recent recession . Our setting are European countries where unemployment benefits are widely available and thus even necessity entrepreneurs' make a choice to become self-employed instead of collecting unemployment benefits.

Empirical studies confirm the higher work autonomy of the self-relative to the wage-employed (Benz and Frey, 2008; Hytti et al., 2013) and its positive impact on hedonic well-being (job satisfaction). In related research, Shir et al. (2019) find the psychological need for autonomy, understood as experiencing few constraints on how to live one's life, mediates the relationship of self-employment with a broad indicator of mostly hedonic well-being. We know of no research linking work autonomy to vitality in the self-employed. Conceptual arguments and research on employees suggest that having control over one's day-to-day work allows individuals to satisfy their need for autonomy and thus increases intrinsic motivation resulting in feeling more vitality and thriving (Deci et al., 2017; Parker, 2014; Ryan and Deci, 2008) . Thus, parallel to H2a to H2c, we propose.

Hypothesis 3a (H3a). Self-employment is positively related to work autonomy.

Hypothesis 3b (H3b). Work autonomy is positively related to subjective vitality.

Hypothesis 3c (H3c). Work autonomy mediates the relationship between self-employment and subjective vitality.

So far, our predictions concerned the individual level. Yet, whether and how strongly employment status is related to MW, work autonomy and ultimately vitality, may depend on the country context in which individuals make the occupational choice for self-vs. wage employment. National contexts can constrain, or enable, the occupational choices individuals make and, thus, may act as a boundary condition for the relationship of employment status with eudaimonia. Past research demonstrates substantial country variation in hedonic well-being and entrepreneurship (Amoròs and Bosma, 2014) . By defining what is expected and acceptable behavior, cultural norms help explain such country differences (e.g., Autio et al., 2013; Steel et al., 2018; Stephan and Uhlaner, 2010) . Regarding well-being, Stephan (2018) suggested that a particularly relevant cultural norm is the societal legitimation of entrepreneurship, i.e., the shared understanding within a country of how desirable and accepted entrepreneurship is (Kibler et al., 2014; Stephan and Uhlaner, 2010) .

Building on the predictions of self-determination theory on the effects of volitional choice (e.g., Moller et al., 2006) , we suggest that the national societal legitimation of entrepreneurship (NSLE) may interact with self-employment status in the prediction of experienced MW. Key to this effect is that a high NSLE can enhance the portfolio of alternative desirable career options for everyone and thereby enables more individuals to make self-determined career choices. Choice is important, as Deci and Ryan (1985) put it: "a behavior is truly chosen only if the person could seriously consider not doing it. (…) Not being able to seriously consider other options suggests that the behavior does not represent true choice, even if it was decided on" (p. 155). Thus, when NSLE is high and entrepreneurship widely considered a desirable career, opting for being a wage-employee is the result of an active choice between two attractive employment options (self-or wage employment). When NSLE is low, wage employees are likely to simply follow the standard employment path which is wage employment (only 10 to 20% of the workforce are self-employed, OECD, 2019). This reasoning is consistent with research in career psychology that builds on SDT. Such research documents that contexts that make more choice options available lead to more self-determined career choices, which, in turn, lead to more engagement with the chosen career option, more satisfaction, less distress, and better performance (Katz et al., 2018) . 8 In line with SDT, greater self-regulation or choosing is the central means by which individuals exercise control (Geers et al., 2013) and reinforce their sense of agency, which leads to higher intrinsic motivation (Deci et al., 2017) and may enhance the perceived meaningfulness of the chosen option (Allan et al., 2016) . Consequently, we expect that wage employees living in countries where entrepreneurship is seen as more legitimate (high NSLE) would perceive their work as more meaningful than their counterparts living in countries with low NSLE, because it results from an active choice for wage employment and against self-employment. Therefore, in high NSLE contexts, the difference in the experienced MW between the self-and wage-employed would be smaller and their level of MW would, on average, be more similar.

There are two countervailing explanations for how NSLE impacts the experienced MW of the self-employed: one based on SDT and the effects of choice (Moller et al., 2006) and one based on the justification perspective of MW. First, when NSLE is low, being selfemployed means going against the social norm regarding the dominant and socially accepted career pattern (wage employment). Therefore, it is more likely to involve a self-regulated decision that results from an active choice process by the individual (Deci and Ryan, 1985) and more likely an expression of his/her values and identity (Cardon et al., 2009) . Hence, the self-employed would experience their work as more meaningful in low, compared to high, NLSE contexts. Second, work on the MW justification perspective suggests a different mechanism that predicts the opposite. It suggests that the experience of MW occurs in alignment with prevailing social norms and values based on "an increased experience of social validation and support" (Lepisto and Pratt, 2017, p. 115; also Florian et al., 2019) . Applied to our context, high NSLE should reinforce the experience of MW for the self-employed, because the evaluation of the worthiness of one's work is argued to be shaped by the social, cultural and institutional context (Lepisto and Pratt, 2017) and high NSLE would offer social validation and legitimation of their career choice. Hence, low NSLE contexts would diminish the experienced MW for the self-employed. 8 SDT research shows that eudaimonic processes (such as MW) may be shaped by context, but that the way these processes lead to eudaimonic well-being outcomes is not impacted by cultural contexts (Chen et al., 2015; Church et al., 2012) . This is in line with the notion that eudaimonic processes have universal qualities (Deci et al., 2017) and the reason why we do not hypothesize context moderation effects on the MW-vitality relationship.

U. Stephan, et al. Journal of Business Venturing 35 (2020) 106047 Which of the two mechanisms impacts the MW of the self-employed more strongly? The first explanation highlights the relevance of self-determined choice in career decision-making for MW and suggests that there is stronger association between self-employment status and MW in low NSLE contexts (than in high NSLE). The second explanation focuses more on the experience of MW based on shared cultural expectations regarding what is a legitimate and worthy career (social normative influence). We expect the first explanation to be more influential because it is consistent with self-employment as a self-determined choice through which individuals create and design a job for themselves that they view as worthy rather than rely on external social validation of worth. In terms of SDT, the second explanation is consistent with external (rather than self-) determination of one's actions and should be less applicable to the self-employed. In sum, our arguments suggest that low NSLE widens the gap in the experienced MW between the self-and wage employed, whereas high NSLE narrows it.

Hypothesis 4a (H4a). The national societal legitimation of entrepreneurship moderates the positive relationship between selfemployment and meaningfulness at work, such that the relationship is weaker in countries where entrepreneurship is more (compared to less) legitimate.

Following from H4a and on our reasoning that MW mediates the relationship between self-employment and vitality (H2c), we expect a moderated mediation effect such that, Hypothesis 4b (H4b). The national societal legitimation of entrepreneurship moderates the indirect positive effect of selfemployment on individual's subjective vitality via meaningfulness at work, such that the indirect positive effect is weaker in countries where entrepreneurship is more (vs. less) legitimate. NSLE might also shape the relationship of self-employment status with work autonomy. Although our argument here is more tentative and explorative. The self-employed might have more actual work autonomy in countries where entrepreneurship is viewed as a positive career option. This is because their stakeholders (e.g., customers, suppliers) are more likely to see the self-employed as legitimate and, thus, are more likely to buy and trade with them (Stephan and Pathak, 2016, p. 509; Zimmerman and Zeitz, 2002) . This would enhance their access to resources and thereby remove constraints to decision making. With regard to the wage employed, we argue that, in societies with high NSLE, their choice for wage-employment is more self-determined since the pool of alternative career options is larger as opposed to contexts where the legitimacy of entrepreneurship is low and wage employment remains the main or normative option. The individuals who become wage employees in this setting, made a more active choice against taking up self-employment. Thus, they actively decided against pursuing what is widely regarded as a more autonomous job (Kolvereid, 1996; Van Gelderen and Jansen, 2006 ) -self-employment. In countries with high NSLE, wage-employees may use the self-employed as a comparator and, thus, come to view their own daily work autonomy as more limited. In countries with low NSLE, this comparator is less salient and thus less relevant to how the wage employed perceive their work autonomy. Hence, Hypothesis 5a (H5a). The national societal legitimation of entrepreneurship moderates the positive relationship between selfemployment and work autonomy, such that the relationship is stronger in countries where entrepreneurship is more (vs. less) legitimate.

Following from H5a and on our reasoning that work autonomy mediates the relationship between self-employment and vitality (H3c), we expect a moderated mediation effect such that:

Hypothesis 5b (H5b). The national societal legitimation of entrepreneurship moderates the indirect positive effect of selfemployment on individual's subjective vitality via work autonomy, such that the indirect positive effect is stronger in countries where entrepreneurship is more (vs. less) legitimate.

Our dataset contains individual-level data obtained from the 2010's data wave of the European Working Conditions Survey (EWCS) merged with country-level data on NSLE obtained from Global Entrepreneurship Monitor (GEM; Kelley, Bosma & Amoros, 2011) and control variables from Eurostat, referring also to 2010. The EWCS is an academically driven survey of representative samples focusing on the workplace experience, attitudes, behaviors, and beliefs. It is conducted bi-annually in Europe. The GEM database is the largest international database on entrepreneurship and its societal perception, based on representative samples of the adult population (Reynolds et al., 2005) . Eurostat is the official statistical office of the European Union, which provides harmonized data across countries. Our dataset was confined to the countries which were present simultaneously in EWCS and GEM datasets in the year of 2010. We analyzed data from 16 European countries and 22,002 individuals. Table 1 shows the countries included in our dataset, displays descriptive statistics for all country-and individual-level variables.

To assess employment status, we created a dummy variable from the 2010 EWCS dataset. The underlying question asked respondents whether they were self-employed or an employee. Self-employment status was coded as 0 = wage-employed and 1 = self-U. Stephan, et al employed. Individuals who responded 'other' or 'don't know' were excluded from the sample. Self-declared self-employment status is commonly used in entrepreneurship research on well-being that utilizes representative surveys (Nikolova, 2019; Stephan and Roesler, 2010) .

Feelings of subjective vitality were measured using four items from the 2010's EWCS (α = 0.85), adapted from WHO-5 well-being index (World Health Organization, 1998) . Participants are asked to report how frequently they felt a certain way over the preceding two weeks: "I have felt active and vigorous", "I woke up feeling fresh and rested", "I have felt cheerful and in good spirit", and "I have felt calm and relaxed". These items were answered in a 6-point scale ranging from 1 = all the time to 6 = at no time and were reversed so that higher levels of the variable would correspond to higher levels of subjective vitality.

MW was assessed using two items from the 2010's EWCS dataset (Spearman-Brown coefficient, r SB = 0.73). Individuals rated how often "You have the feeling of doing useful work" and "Your job gives you the feeling of work well done". These items were preceded by "Select the response which best describes your work situation", and responses were given on a 5-point Likert scale from 1 = always to 5 = never. The items were reversed so that higher levels of the variable would correspond to higher levels of MW.

To assess the construct validity of our MW measure, we compared it to previous validated scales of meaningfulness at work by conducting a validation study using Mturk. In addition to the two items from EWCS, we included the following validated scales: Steger et al.'s (2012) Work as Meaning Inventory (10 items), Spreitzer's (1995) subscale of Meaning (3 items) from her Psychological Empowerment in the Workplace Scale, and Hackman and Oldham's (1975) two items measuring Experienced Meaningfulness of Work in the Job Diagnostic Survey. We evaluated correlations with our MW measure for a sample of wage-employees (N wage employees = 159) and a sample of self-employed (N self-employed = 154) obtained with Mturk. To ensure data quality, we controlled for insufficient effort responding bias, as recommended by Cheung et al. (2014) and Huang et al. (2014) , which yielded a wage-employee sample of N = 114 and a self-employed sample of N = 103. We analyzed the Pearson correlations between our 2-item MW measure and the validated meaningfulness scales for both samples. Our 2-item MW measure correlated substantially and positive, as expected, with Steger et al.'s (2012) Work as Meaning Inventory (r wage-employees = 0.83; r self-employed = 0.78), Spreitzer's (1995) subscale of Meaning (r wage-employees = 0.75; r self-employed = 0.71), and Hackman and Oldham's (1975) two items measuring Experienced Meaningfulness of Work (r wage-employees = 0.82; r self-employed = 0.76). Therefore, although the EWCS used an operationalization of MW that is somewhat different from other measures in the literature, our additional study provides evidence for the construct validity of the 2-item measure of MW used in this study and specifically that it captures peoples' perceptions that their work is meaningful.

Work autonomy was measured using three items from EWCS dataset coded as a dummy variable (0 = no; 1 = yes): "Are you able to choose or change your order of tasks?", "Are you able to choose or change your methods of work?"; "Are you able to choose or change your speed or rate of work?" (α = 0.78). A composite variable was computed with a minimum = 0 and a maximum = 3.

NSLE was measured as a country-level variable, using one item from GEM 2010's survey, which referred to the percentage of the adult population (aged between 18 and 64 years old), in each country, that answered yes to "In my country people consider starting a business as a good career choice". 9 The year-to-year stability (re-test reliability) of NSLE was r = 1.00, p < .001, N = 16 for 2010-2011, r = 0.92, p < .001, N = 14 for 2010-2012.

We controlled for variables at the individual and the country level that are known to correlate with well-being. At the individual level, we controlled for gender, age, wage, and education level available on the 2010's EWCS dataset. Gender was coded as a dummy variable (0 = female; 1 = male). Wage was defined as the log of the value in Euros reported by the participants in reply to the question "How much are your net monthly earnings from your main paid job?" Because many participants did not reply to this question, but answered the question "What letter best matches your total net earnings from your main job?", where each of the 21 letters correspond to a given interval of salary values, we used the midpoint of the selected interval to fill-in the missing values, from a minimum of 625 euros to a maximum of 60,000 euros. Education was operationalized with the question "What is the highest level of education or training that you have successfully completed?", answered in a 7-point scale ranging from 1 = No education to 7 = Advanced level of tertiary education.

At the country level, we controlled for the unemployment rate as a possible alternative explanation for the effects of societal legitimation and to account for possible effects of necessity entrepreneurship (see also Robustness Checks Section 5.2.2). A high unemployment rate means that there are few other employment opportunities and people are more likely to enter entrepreneurship out of necessity to find work (Dvouletý, 2018; Estrin et al., 2016) . The unemployment rates taken from Eurostat (2010) are the percentage of the labor force of each country (aged from 15 to 74 years) that was without work during the referenced week, who was available to start working within the next two weeks or who had been actively seeking work in the past four weeks.

In additional robustness checks (available upon request), we also controlled for national wealth (measured as GDP and obtained from Eurostat) and confirmed all effects. We do not include GDP in the main results, because its effect was not significant. Moreover, it correlates at 0.60 with the national unemployment rate, which likely creates multicollinearity and thus unstable regression results. In a second robustness check, we explored whether the effects of societal legitimation of entrepreneurship might be better explained by unemployment (i.e. a push vs. pull into entrepreneurship). Substituting unemployment for societal legitimation in the interaction with self-employment status yielded a non-significant interaction effect.

Our data are nested (individuals within countries). We therefore estimate linear two-level mixed-effects models with random intercepts at the country level to test our hypotheses. To test the moderated mediation effects, we use a two-step approach (Bauer et al., 2006) . The first step involved estimating the total effects of self-employment on work autonomy, MW and subjective vitality (Table 3 : H1, H2a, H3a) and then testing simple mediation models (Tables 4 and 5 : H2b, H2c, H3b, H3c) . The second step (Table 6) involved testing the proposed cross-level moderation effect (H4a, H5a) and the overall moderated mediation model (H4b, H5b) including estimating the direct, indirect and total effects. To test the validity of the random effect assumption required by all models, we applied the Hausman test. 10 To test the magnitude and significance of the hypothesized indirect and total effects in the mediation models, and to obtain 95% bias-corrected confidence intervals (BCs CIs) for them, we used bootstrapping procedures based on 5000 samples. All calculations were performed in Stata.

Means, standard deviations, correlations, and internal reliabilities are shown in Table 2 .

Table 3 presents the total effects of self-employment status on meaningfulness at work (MW), work autonomy, and subjective vitality. After accounting for age, gender, education, and wage (Level 1 control variables) and the country unemployment rate (Level 2 control variable), self-employment status was positively related to subjective vitality (B = 0.08, p < .001), to MW (B = 0.28, p < .001) and to work autonomy (B = 0.75, p < .001), supporting Hypotheses 1, 2a and 3a, respectively. Table 4 reports the first set of tests of the hypothesized mediation models. The results indicate that MW was positively associated with individual's subjective vitality (B = 0.34, p < .001), even after controlling for age, gender, education, wage, countries' unemployment rate, and for work autonomy (Models 3 and 5, Table 4 ). This supports H2b that MW relates positively to vitality.

In Table 4 , Model 4 shows that work autonomy was positively associated with individual's subjective vitality (B = 0.04, p < .001) when controlling for age, gender, education, wage and countries' unemployment rate supporting H3b. However, once MW was added to the model (Table 4 , Model 5), this effect became non-significant (B = −0.001, p > .05). Therefore, H3b is rejected. Consequently, H3c supposing that work autonomy would mediate the relationship between self-employment and subjective vitality is not supported.

Furthermore, the results of the regression analyses presented in Table 4 revealed that the positive relationship between selfemployment and subjective vitality was no longer significant when MW was introduced in the regression model (direct effect B = −0.02, p > .05, Model 5). This provides first support for H2c: MW mediates the relationship between self-employment status and subjective vitality. Table 5 offers a formal test of the indirect (mediating) effect. It summarizes the direct, indirect, and total effects of self-employment status on subjective vitality. It shows that the indirect effect of self-employment status on subjective vitality through MW, controlling for age, gender, education, wage, countries' unemployment rate and work autonomy, had a point estimate of 0.07 (95% BCa CI = 0.06, 0.08) and was statistically significant (p < .001). This supported H2c: self-employment status is positively associated with individual's subjective vitality via the experienced MW.

The results in Table 5 further suggest that, if MW would not be considered, self-employment status would have a significant effect on subjective vitality that would be partially mediated by work autonomy (indirect effect with a point estimate of 0.03, p < .001, and a 95% BCa CI = 0.02, 0.04). However, when MW was introduced in the regression model, this indirect effect via work autonomy 10 In three out of the ten models assessed, the Hausman test produced a negative statistic. Theoretically, the Hausman statistic should be nonnegative. Although this is a well-known issue of the Hausman test in empirical work and most researchers interpret such result as evidence favouring the null hypothesis of random effects, we repeated all the analysis in the paper using a linear model with country effects. The results obtained, which are available upon request, were very similar to those produced by the mixed-effects model. The ICC estimates for country-level effects were 0.02 for autonomy, 0.02 for meaningfulness and 0.04 for vitality. Our interest was not in explaining country-level variation in the intercepts of these variables, but rather to examine a cross-level moderation effect, which can exist even in the absence of substantial intercept variation (i.e. when ICCs are low). U. Stephan, et al. Journal of Business Venturing 35 (2020) 106047 was no longer significant −0.00, n.s. and a 95% BCa CI = -0.01; 0.01. This indicates that the experience of MW is the more important mediator. In short, meaningfulness at work 'trumps' work autonomy in its effect on subjective vitality. Table 6 presents the results of the hypothesized cross-level interaction effect between self-employment status (level 1) and national societal legitimation of entrepreneurship (level 2) in the prediction of MW (level 1) -which corresponds to H4a -and in the Table 2 Means, standard deviations, and correlations. prediction of work autonomy (level 1) -which corresponds to H5a. Table 6 also reports the conditional indirect effect (multi-level moderated mediation) via MW (H4b) and work autonomy (H5b). After controlling for age, gender, education, wage and work autonomy at the individual level and unemployment rate at country To illustrate the interaction effect on MW, we plotted the simple slopes for respondents from countries where entrepreneurship has lower societal legitimacy (one standard deviation below the mean) and for respondents from countries with higher legitimacy (one standard deviation above the mean). Fig. 2 shows the moderating effects of the national societal legitimation of entrepreneurship on the relationship between self-employment status and MW and the 95% confidence intervals. Fig. 2 shows that there are no significant differences in the MW reported by the self-employed in countries with different levels of national societal legitimation of entrepreneurship (the slope is not significantly different from zero γ = 0.000, p > .05; 95% CI = −0.006, 0.007). Thus, for the self-employed the experienced MW was always higher than for wage-employees, regardless of the legitimacy of entrepreneurship in their national context. However, for wage-employees, the higher the legitimacy of entrepreneurship, the more employees experienced MW (γ = 0.008, p < .001; 95% CI = 0.003, 0.013). In countries that were more favorable to entrepreneurship (1 SD above the mean), the difference in the MW reported by wage-employees and the self-employed was significantly smaller than in countries that were less favorable to entrepreneurship (1 SD below the mean). Thus, H4a is supported, NSLE negatively moderates the positive relationship between self-employment and MW. The difference between self-and wage-employed individuals in MW is higher in countries where entrepreneurship is seen as less legitimate, when compared with countries where entrepreneurship is seen as more legitimate.

The results in Table 6 (last three rows) also revealed that the conditional indirect effect of self-employment on individual's subjective vitality via MW was stronger when society views entrepreneurship less rather than more positively. This is due to the Table 6 Regression results for the moderating effect of national societal legitimation of entrepreneurship on the relation between self-employment and meaningfulness at work and work autonomy, and conditional indirect effects on subjective vitality. Note: Calculations for indirect effects use the results of the last model in Table 4 . Inference for indirect effects based on 5000 bootstrap samples and use bias-corrected confidence intervals. NSLE = National societal legitimation of entrepreneurship. ΔPseudo R 2 relative to models shown in Table 4 . a 0 = female, 1 = male. b Wage = log of the value in Euros. c 0 = employed, 1 = self-employed. d LR test for one-level ordinary linear regression. e Hausman test for random-effects. ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. diminished difference in MW between self-employed and wage employees in countries with higher levels of NSLE. The estimates were all positive and significant at 0.09 (95% BCa CI = 0.08, 0.11), 0.07 (95% BCa CI = 0.06, 0.08) and 0.05 (95% BCa CI = 0.04, 0.06) when NSLE was low, medium and high, respectively (Table 6 ). Therefore, H4b is supported. Regarding the possible cross-level interaction between self-employment status and NLSE in the prediction of work autonomy (H5a), the results revealed that, after controlling for country's unemployment and individuals' age, gender, education, wage and MW, the cross-level interaction was significant and positive (γ = 0.01, p < .01, Table 6 ), supporting H5a. Fig. 3 shows the moderating effects of societal legitimation on the relationship between self-employment status and work autonomy and the 95% confidence intervals for respondents from countries with low (one standard deviation below the mean), medium (at the mean) and high legitimacy of entrepreneurship (one standard deviation above the mean). Fig. 3 indicates that there are no significant differences in the work autonomy reported by the self-employed in countries with different levels of national societal legitimation of entrepreneurship (NSLE, γ = 0.005, p > .05; 95% CI = −0.005, 0.016). For these individuals, the perceived work autonomy is always higher than for wage-employees, regardless of the national context. Similarly, the perceived work autonomy of wage-employees did not change significantly when we compared participants from countries with different levels of NSLE (γ = −0.004, p > .05; 95% CI = −0.014, 0.005). However, in countries where entrepreneurship was seen as a more legitimate career choice by the population (NSLE 1 SD above the mean), the difference in perceived work autonomy between the self-employed and wage-employees is higher (γ = 0.744, p < .001; 95% CI = 0.680, 0.809) than in countries with lower levels of NSLE (1 SD below the mean) (γ = 0.584, p < .001; 95% CI = 0.509, 0.660). This supports H5a, which expected that Meaningfulness at work as a function of self-employment status and national societal legitimation of entrepreneurship (95% confidence intervals); 'low' and 'high' correspond to 1SD below and above the mean of NSLE, medium is the mean. Fig. 3 . Autonomy at work as a function of self-employment status and national societal legitimation of entrepreneurship (95% confidence intervals); 'low' and 'high' correspond to 1SD below and above the mean of NSLE, medium is the mean. NSLE moderates the relationship between self-employment and work autonomy such that the difference between wage-employees and the self-employed in work autonomy is higher in countries where entrepreneurship is seen as more legitimate compared with countries where NSLE is low.

In line with our previous results, which showed that work autonomy was not a significant mediator of the relationship between self-employment status and subjective vitality (H3c was not supported, see Tables 4 and 5) , we found no significant conditional indirect effect of self-employment on individual's subjective vitality via autonomy at work at any level of NSLE (see Table 6 last three rows). Therefore, we found no support for H5b.

As a robustness test for the effects of self-employment status on MW, work autonomy and subjective vitality, we used propensity score (PS) methods with self-employed workers being the treatment group and wage-employees the control group. We used a logit model to estimate the PS conditional on the following variables, which are assumed to influence the probability of being selfemployed: gender, age, education, and unemployment rate at country level. 11 Then, three alternative PS methods were considered: matching, inverse probability weighting (IPW) and use of PS as an additional covariate in all models estimated in the previous section. Using the first method, we matched each self-employed individual with the untreated individual (or individuals -wage employees) who displayed the closest predicted propensity score(s), and vice-versa. Next, we estimated the response differences for each match. Finally, we estimated the average treatment effect (ATE) by averaging those differences for the whole sample. The second method, IPW, reweights individuals in such a way that the weights of those who received unexpected exposures are increased in the response regression model used to estimate the treatment effects. We used a double robust IPW approach and included in the regression model all control variables considered in our main analysis. Finally, the third PS method simply adds the estimated PS as an additional covariate in the models considered in our main analysis. The results reported in Table 7 are consistent with those reported in Table 3 , providing further confidence in our findings.

To address the concern that we may identify different effects for people that chose to be self-employed out of opportunity versus necessity, we ran additional robustness checks. We checked for differences in the means of MW, work autonomy, and subjective vitality between individuals that immediately before becoming self-employed were employed (a proxy for opportunity entrepreneurship) and individuals that previously to being self-employed were unemployed (a proxy for necessity entrepreneurship, following the approach used by Binder and Coad, 2013; Nikolova, 2019) . The results show that after controlling for gender, age, education and wage, there are no significant differences between opportunity vs. necessity entrepreneurs in the means of MW (the regression coefficient of opportunity vs. necessity on MW was b = 0.050, p = .513), work autonomy (b = −0.011, p = .900), and subjective vitality (b = −0.103, p = .380). Results were similar and non-significant when we controlled additionally for work autonomy and MW respectively.

In a multi-level study of 22,002 individuals from 16 European countries, we found self-employment to be associated with higher eudaimonic well-being (subjective vitality) which was explained by the self-employed experiencing their work as more meaningful than wage employees, in line with our predictions drawing on self-determination theory. The findings could not be better explained by differences in work autonomy or self-selection. Although the self-employed reported higher work autonomy than wage employees, it was meaningfulness at work, and not work autonomy, that explained why the self-employed experienced higher vitality than wage employees. These processes were shaped by context, in line with predictions that in countries in which entrepreneurship is seen as an attractive career, everyone makes more self-determined career choices, which enhances the meaningfulness of the chosen option (especially for wage employees), while magnifying differences in perceived work autonomy. Our study expands the understanding of eudaimonia and self-employment.

Our study responds to calls to consider entrepreneurs' eudaimonic well-being (Ryff, 2019; Stephan, 2018; Wiklund et al., 2019) and advances our understanding of eudaimonic well-being in entrepreneurship. It develops the scarce existing research on entrepreneurs' eudaimonic well-being (Hahn et al., 2012; Nikolaev et al., 2020; Shir et al., 2019) by newly drawing attention to meaningfulness at work as an critical eudaimonic process and to vitality as an important eudaimonic outcome. In doing so, our study provides new insights into the mechanisms of eudaimonia. It also complements existing research on entrepreneurs' well-being which has focused mainly on hedonic well-being, its antecedents and consequences, and has highlighted work autonomy as a key driver of these differences (see Stephan, 2018) .

Our study provides a theoretical framework and evidence that helps to understand why and how work-related eudaimonia differs for entrepreneurs and wage employees, and reveals MW and vitality as overlooked non-pecuniary benefits of self-employment. Our findings are consistent with the notion that the initial volitional career choice to be self-, rather than wage-, employed allows individuals to choose a job that they find worth doing, which drives eudaimonic benefits (MW and subjective vitality).

Our findings suggest that this initial self-determined choice (through creating meaningful work) is more consequential than the ongoing day-to-day decision-freedom (work autonomy) for eudaimonic well-being. The meaning entrepreneurs attribute to their work helps to understand why self-employment is positively related to vitality, even though it can, at times, also contain menial work as the self-employed have to be 'jack of all trades' (e.g., Lazear, 2005) and can even be precarious work (e.g., working in the gig economy or for a sole client, Lewchuk, 2017; Petriglieri et al., 2018) . This finding also fits with emerging research. For instance, on the one hand, research suggests that on a daily basis the self-employed may experience many constraints and limits to their work autonomy (van Gelderen, 2016; Van Gelderen et al., 2019) . On the other hand, research on precarious work highlights that if this work is experienced as meaningful, it can still offer some well-being benefits (Deery et al., 2019) . In sum, while existing work has often hailed positive job characteristics as an advantage of self-employment and a source for well-being, our results suggest an alternative lens. The experienced MW is more central to understanding entrepreneurs' vitality and thriving, and possibly other wellbeing outcomes and performance than job characteristics.

Finally, our research not only enhances our understanding of the nature of entrepreneurs' work experience by drawing attention to MW; it also newly highlights vitality as an essential component of eudaimonic well-being and of particular importance to entrepreneurship. Vitality appears critical for entrepreneurial agency and performance (Hahn et al., 2012 ). Yet, vitality rarely receives theoretical or empirical attention in the entrepreneurship literature and thus remains poorly understood. We hope our study helps to instigate research interest in vitality. In addition to links with performance, vitality can also be important to understand the microdynamics of entrepreneurs' well-being over a day or week, and thus develop new theoretical and practical insight to support entrepreneurs in managing their energy and productivity. Diary studies of employee's energy management can provide inspiration for such research (Fritz et al., 2011; Niessen et al., 2012) .

In sum, we complement emerging research that draws attention to general context-free eudaimonic processes and their association with mainly hedonic well-being (Nikolaev et al., 2020; Shir et al., 2019) . We do so by shining a light on work-related eudaimonic processes which advance our understanding of how and why self-employment can be an energizing work setting.

Our study contributes by theorizing and testing the role of context for eudaimonic processes and establishes the societal legitimacy of entrepreneurship as an important boundary condition for these processes. It advances a context-sensitive understanding of entrepreneurs' well-being that complements past research, which rarely accounts for context effects (Stephan, 2018) . Equally, the emerging research on entrepreneurs eudaimonic well-being is based on single country samples (Hahn et al., 2012; Shir et al., 2019) or controls country influences 'away' instead of theorizing them in multi-level frameworks (Nikolaev et al., 2020) . We hope our multilevel theorizing can inspire more research on eudaimonia in different contexts.

More specifically, our contribution develops theory by integrating, on the one hand, predictions from self-determination theory on the types of contexts that enable individuals to make self-determined (career) choices (Deci et al., 2017; Katz et al., 2018) with cultural research on the relevance of norms and, in particular, the legitimacy of entrepreneurship (Kibler et al., 2014; Stephan and Uhlaner, 2010) . This complements culture-entrepreneurship fit explanations that are commonly evoked to theorize the effect of context on entrepreneurship (Hayton et al., 2002) . Equally, we do not constrain our arguments, as is common, to consider only the effect of cultural norms on the self-employed and the support they can mobilize from others in their culture (Autio et al., 2013; Kibler et al., 2014; Stephan and Uhlaner, 2010) . Instead, we conceptualize and demonstrate that cultural norms enhance volitional career choice for both the self-and wage-employed. Thus, future research on culture should expand its theorizing to consider additionally how culture effects the non-entrepreneur 'others'.

In sum, there have been repeated calls for more context-sensitive theory and research in entrepreneurship (Welter, 2011; Zahra and Wright, 2011) , in organizational behavior (Johns, 2006 (Johns, , 2017 and on entrepreneurs' well-being (Stephan, 2018) . Our study a Effect of self-employment on autonomy at work, meaningfulness at work and subjective vitality controlling for gender, age, education, wage (Level 1) and unemployment (Level 2). Results presented previously in Table 3 . b Robust standard errors, based on the correction by Abadie and Imbens (2016) , are reported in parenthesis. ⁎⁎⁎ p < .001.

responds to such calls. Moreover, it proposes important nuances and guidance 'where to look' for context effects in research on wellbeing. In line with self-determination theory, context is less likely to shape or moderate basic psychological processes such as the link between experienced meaningfulness and vitality (Chen et al., 2015; Church et al., 2012 ). Yet, context can shape antecedent conditions whether meaningfulness at work arises in the first place.

Drawing on self-determination theory and the role of volitional choice, our study identifies the employment setting (self-vs. wage-employed) as a relevant predictor of meaningfulness at work. Specifically, we find self-employment has a high potential for MW that has not yet been recognized in the literature on MW. Considering that volitional choice and autonomy are important antecedents in models of meaningfulness at work (e.g., Allan et al., 2016) , we suggest that this literature would benefit from considering selfemployment more seriously. As Lips-Wierma and Morris (2009, p. 494) suggest, "the mark of true meaningfulness is that it is based on personal discovery and free choice rather than prescription and domination" -a description fitting self-employment particularly well. We suggest that the consideration of self-employment, as an important determinant of MW, may not only help to extend research on meaningfulness at work but also benefit policies of inclusion in the labor market. Self-employment can uniquely offer meaningful work to individuals who are seen to have disadvantages when it comes to fitting into standard wage-employment settings, such as neuro-diverse individuals with ADHD, dyslexic or autistic tendencies (Wiklund et al., 2018) and those with disabilities (Pagán, 2009) .

Moreover, we offer new insights on how meaningful work can be shaped by societal contexts that past reviews on meaningful work have called for (Bailey et al., 2019, p. 22) . While there is first research that suggests that the changing legitimacy of social issues can impact the meaningfulness of volunteers' work (Florian et al., 2019) , we offer a complementary lens. This lens emphasizes how social legitimacy of more 'niche' career options, which are typically only picked up by a minority of individuals in the workforce, enables volitional choice even for those who do not embrace this career and thereby enhances their meaningfulness of work. 12 Whereas research on the meaningfulness at work emphasizes that meaningfulness derives from the congruence with societal values and norms (Lepisto and Pratt, 2017) , we find evidence consistent with an alternative explanation based on self-determination theory. Our theorizing emphasizes the role of self-determined choice for meaningfulness at work, which has been established at the individual level (Duffy et al., 2015; Tims et al., 2016 ), but has not yet been theorized with regard to (country) context in a multi-level model. Moreover, an important difference between our study and theorizing on meaningfulness at work (Lepisto and Pratt, 2017 ) is the focus on the self-employed. Lepisto and Pratt (2017) and the meaningfulness at work literature focus on dependent wage employment. Societal worth attributed to a job might be more important for wage employees. It may compensate for the relative lack of other intrinsic sources of meaningfulness.

Regarding self-determination theory (SDT), our research expands the view of what aspects of context are relevant for shaping eudaimonic processes by newly highlighting how social pressures and legitimacy can impact choice options. SDT mainly theorizes the role of autonomy supportive contexts and typically considers proximal levels of context, such as the family or supervisors in work organization. Their autonomy supportive behavior is key in supporting volitional choice of children or employees (Deci et al., 2017; Katz et al., 2018) . Our findings introduce the idea that social pressures and especially the social legitimation of alternatives constrains or enhances the set of choice options that can be considered. As there is only volitional choice if there are options to choose from (Deci and Ryan, 1985) , such social pressures and legitimacy can be an important constraint or enabler of self-determination.

While SDT has been successfully employed to understand work, work motivation and well-being (Deci et al., 2017) , its application to career decision and career choices has been more limited (Katz et al., 2018) . We hope our study can inspire future work on SDT and career choices, including choices for less 'standard' careers such as self-employment.

One surprising pattern of results in our study suggests that experienced MW may shape perceptions of having work autonomy. Our data are cross-sectional and this pattern needs to be corroborated by future longitudinal research. Yet, it is intriguing considering that the dominant view in organizational behavior research is that work autonomy and other features of 'good' job design enable experiences of meaningfulness (Hackman and Oldham, 1975; Parker, 2014) in line with the realization perspective of MW (Lepisto and Pratt, 2017) . Our findings imply that experiencing work as significant, purposeful and as an extension of oneself might enable individuals to generate spaces where they feel they can make self-determined decisions. Conversely, if work is experienced as meaningless, useless and without impact, individuals are less likely to be motivated to exercise decision-making autonomy even if it is provided. In other words, our findings suggest that MW generated through the justification perspective facilitates meaningfulness from the realization perspective. It thus offers first insights into Lepisto and Pratt's (2017, p.116) call for research to explore how the two perspectives work together. We hope future research can explore such patterns in more depth as they offer insights beyond entrepreneurship to research on meaningful work and the outcomes of meaningful work.

First, this study is limited in its ability to make causal inferences due to the cross-sectional nature of the research design. All research comes with trade-offs and the use of data from 16 different countries enabled us to test a multi-level model, but it meant that we had to rely on cross-sectional data. We know of no dataset that can provide harmonized longitudinal data for that many countries. At the same time, our findings are consonant with previous literature that uses stronger research designs and which suggest that MW is a predictor of individuals' energetic resources (Blanco-Donoso et al., 2017; Fritz et al., 2011; Lam et al., 2016; Niessen et al., 2012) . Nevertheless, using a longitudinal design in future research would allow to test possible reciprocal relationships between MW and subjective vitality. It is possible that people feeling more energized and cognitively alive would use that appraisal as input information to evaluate the purpose and significance of their work (cf. Affect infusion model, Forgas, 1995) , reporting thus higher levels of MW.

Second, as is common in research on well-being (e.g. Nikolaev et al., 2020; Shir et al., 2019) , we relied on self-report data for MW, work autonomy, and subjective vitality. Since MW and vitality are constructs based on subjective experience, self-reports are appropriate, and they showed good internal reliability in our study.

Third, we used a short general measure of subjective vitality. Future research could employ longer multi-facetted measures of subjective vitality to refine our understanding of the physical, emotional and cognitive resources involved in the concept of vitality (Shirom, 2003) . Such an approach is likely to be particularly suitable for survey or diary studies, rather than for cross-country studies such as ours, which often must economize on scale length. Future research could explore how vitality and MW are relevant for the long-term success of entrepreneurs and survival of their enterprises, and whether prior performance may influence the purpose and meaningfulness that the self-employed attribute to their work.

Fourth, by considering the societal legitimacy of entrepreneurship we investigated a specific aspect of context that is particularly relevant for the setting of self-employment in a relatively small sample of 16 European countries. Yet, we were able to establish that context (societal legitimacy of entrepreneurship) acts as a boundary condition even in this relatively small and homogenous set of countries. Institutional conditions are similar in these 16 countries because they all operate within the legal frameworks of the European Union. This helped to isolate the effects of societal legitimation. It is likely that societal legitimacy and other cultural norms would have even more pronounced effects when more varied countries are considered, and in emerging economies where the career choice set might look very differently (due to lack of wage employment options, e.g., Estrin et al., 2018) . Moreover, our time frame was 2010, it would be interesting to expand our findings to further time periods.

Fifth, there is considerable heterogeneity in self-employment. Our main analyses sought to control for necessity-based entrepreneurship by including the national unemployment rate in our empirical models. Interestingly, additional robustness checks did not find differences between necessity and opportunity entrepreneurship in MW, work autonomy, and subjective vitality. However, past research established systematic differences in the hedonic well-being of opportunity and necessity entrepreneurs (Stephan, 2018 for a review). One reason might be that our measure of opportunity-necessity was too crude (although it was the same as in Coad, 2013 and Nikolova, 2019) . Future research could use direct measures of entrepreneurial motivation as well as more nuanced measures that differentiate challenge-and purpose-driven motivations and relate these to eudaimonic well-being.

Sixth, as we indicated in Section 3.2, different models of eudaimonic well-being exist. Nikolaev et al. (2020), Shir et al. (2019) and our study investigate different aspects of eudaimonia and thus offer complementary insights. Our study only investigates one eudaimonic process and outcome. Future research might try to combine the different models of eudaimonic well-being Ryff, 1989; Waterman et al., 2010) into one study to disentangle diverse eudaimonic processes and outcomes and determine which are most relevant for entrepreneurship. However, such research needs to be mindful that concepts and models of eudaimonic well-being partly overlap, and that context-free measures of eudaimonia of psychological functioning and psychological well-being (Ryff, 1989) show medium to large relationships with the Big 5 personality traits (Anglim et al., 2020) . This challenges researcher to disentangle what eudaimonic well-being benefits originate from the choice to engage in self-employment and the nature of selfemployed work as opposed to the self-employed's personality traits.

Finally, while we depict subjective vitality as a positive energetic resource in line with eudaimonic models of well-being, we also encourage future research to investigate the long-term consequences of high subjective vitality for entrepreneurs. It could be that vitality is a double-edged sword, leading entrepreneurs to engage in more work and 'over-work' themselves without enough time for rest and recovery. This would in the long run enhance entrepreneurs' stress and diminish their well-being.

Our findings also have implications for policymakers. The European Commission (EC) presents entrepreneurship as a political priority considering it is "a powerful driver of economic growth and job creation" (European Commission, 2013: Entrepreneurship 2020 Action Plan, p. 3). Policymakers around Europe have sought to stimulate self-employment to deal with the aftereffects of the 2008 economic crisis to improve employability levels, mitigate unemployment and foster economic growth. An ambition that is likely to be salient again to deal with the aftereffects of the Covid-19 pandemic. The EC in its Entrepreneurship 2020 Action Plan posits that "we must work on ensuring that being an entrepreneur is an attractive prospect for Europeans" (EC, 2013, p. 5), but it also admits that in Europe there is still "a widespread culture that does not recognize or reward entrepreneurial endeavors enough and does not celebrate successful entrepreneurs, as role models who create jobs and income" (p. 4). Our results suggest that the lack of appreciation of entrepreneurship is not only detrimental to entrepreneurial activity, but can also diminish the quality of work life among wage employees. Improving the societal legitimation of entrepreneurship would also change what it means to be a wage employee in that national context.

In this sense, our study provides further re-enforcement for the EC's ambition to celebrate and appreciate self-employment more. Public communication campaigns promoting entrepreneurship and entrepreneurial role models, as well as conveying the benefits, rewards and challenges of this career choice, are certainly a helpful tool (EC, 2013) . We would add that role models should be varied, not only representing high-growth entrepreneurs but also appreciating that entrepreneurship may be small scale self-employment, which has an important economic function (Welter et al., 2017) . One challenge in enhancing the legitimacy of self-employment is that it is a relative 'niche' segment in the workforce (most people are in wage employment). This is because legitimacy is typically conferred to entities that are common, widespread and therefore can be taken for granted (Stephan and Uhlaner, 2010; Zimmerman and Zeitz, 2002) . Hence, aside from showcasing role models, incorporating entrepreneurship in school, vocational and university curricula would ensure that knowledge about, and acceptance of, self-employment becomes more widespread and thus that selfemployment would be increasingly seen as a legitimate career path.

Research on entrepreneurship and well-being is focused on hedonic well-being; with little attention given to eudaimonic processes. Our study widens the scope of entrepreneurship and well-being research by drawing attention to a hitherto overlooked eudaimonic process (meaningfulness at work) and outcome (subjective vitality) and how they are shaped by context (the societal legitimacy of entrepreneurship). Our findings showcase that the meaningfulness at work is more central to explaining eudaimonic well-being (subjective vitality) than work autonomy. They develop our understanding of meaningfulness at work as an important well-being resource for entrepreneurs and complement past research that focuses on work autonomy as the main contributor to entrepreneurs' well-being. We outline contributions to the understanding of eudaimonia in entrepreneurship and the role of context for (eudaimonic) well-being. Our research has also implications for the organizational literature on meaningfulness at work and selfdetermination theory in relation to career choices. 

Matching on the estimated propensity score

Self-determination and meaningful work: exploring socioeconomic constraints

Outcomes of meaningful work: a meta-analysis

Global Entrepreneurship Monitor

Predicting psychological and subjective well-being from personality: a meta-analysis

Consequences of cultural practices for entrepreneurial behaviors

A review of the empirical literature on meaningful work: progress and research agenda. Hum. Resour

Job design and entrepreneurship: why closer connections=mutual gains

Conceptualizing and testing random indirect effects and moderated mediation in multilevel models: new procedures and recommendations

Self-regulation and the executive function: the self as controlling agent

Being independent raises happiness at work

The value of doing what you like: evidence from the self-employed in 23 countries

Social job resources as sources of meaningfulness and its effects on nurses' vigor and emotional exhaustion: A cross-sectional study among Spanish nurses

Life satisfaction and self-employment: a matching approach

How satisfied are the self-employed? A life domain view

Self-employment in OECD countries

Teaching personal initiative beats traditional training in boosting small business in West Africa

The nature and experience of entrepreneurial passion

Basic psychological need satisfaction, need frustration

Amazon Mechanical Turk in organizational psychology: An evaluation and practical recommendations

Positive psychological well-being and mortality: a quantitative review of prospective observational studies

Need Satisfaction and Well-Being: Testing Self-Determination Theory in Eight Cultures

Intrinsic Motivation and Self-Determination in Human Behavior

Self-determination theory: a macrotheory of human motivation, development, and health

Self-determination theory in work organizations: the state of a science

Can dirty work be satisfying? A mixed method study of workers doing dirty jobs

Advances and open questions in the science of subjective well-being

Work volition and job satisfaction: examining the role of work meaning and person-environment fit

How to analyse determinants of entrepreneurship and self-employment at the country level? A methodological contribution

Human capital in social and commercial entrepreneurship

The Oxford Handbook of Management in Emerging Markets

Struggling with meaningfulness when context shifts: volunteer work in a German refugee shelter

Mood and judgment: the affect infusion model (AIM)

Towards a psychology of entrepreneurship: an action theory perspective

It's the little things that matter: an examination of knowledge workers' energy management

Why does choice enhance treatment effectiveness? Using placebo treatments to demonstrate the role of personal control

Advancing the psychology of entrepreneurship: a review of the psychological literature and an introduction

Work engagement and workaholism: comparing the self-employed and salaried employees

Relating work engagement and workaholism to entrepreneurial performance

Development of the job diagnostic survey

Happy and proactive? The role of hedonic and eudaimonic well-being in business owners' personal initiative

Does entrepreneurship pay? An empirical analysis of the returns to self-employment

National culture and entrepreneurship: a review of behavioral research

Insufficient effort responding: Examining an insidious confound in survey data

Determinants of job satisfaction for salaried and self-employed professionals in Finland

The essential impact of context on organizational behavior

Alive and creating: the mediating role of vitality and aliveness in the relationship between psychological safety and creative work involvement

The relation of psychological needs for autonomy and relatedness to vitality, well-being, and mortality in a nursing home1

Parental support for adolescents' autonomy while making a first career decision

Global Entrepreneurship Monitor: Global Report

Regional social legitimacy of entrepreneurship: implications for entrepreneurial intention and start-up behaviour

Organizational employment versus self-employment: reasons for career choice intentions

Going the extra mile and feeling energized: an enrichment perspective of organizational citizenship behaviors

Absence from work of the self-employed: a comparison with paid employees

Meaningful work as realization and justification

Precarious jobs: where are they, and how do they affect well-being? The Economic and

Discriminating between 'meaningful work' and the 'management of meaning

Measuring the meaning of meaningful work

Enacting creative calling when established career structures are not in place: the case of the Dutch video game industry

The psychological conditions of meaningfulness, safety and availability and the engagement of the human spirit at work

Entrepreneurial action and the role of uncertainty in the theory of the entrepreneur

Choice and ego-depletion: the moderating role of autonomy

Helpful self-control: autonomy support, vitality, and depletion

Entrepreneurship and subjective well-being: the mediating role of psychological functioning

Switching to self-employment can be good for your health

Revitalization through self-regulation: the effects of autonomous and controlled motivation on happiness and vitality

Self-Employment Rate

Self-employment among people with disabilities: evidence for Europe

Beyond motivation: job and work design for development, health, ambidexterity, and more

Agony and ecstasy in the gig economy: cultivating holding environments for precarious and personalized work identities

Fostering meaningfulness in working and at work

Building a sustainable model of human energy in organizations: exploring the critical role of resources

Global entrepreneurship monitor: Data collection design and implementation

Unpacking the personal initiative-performance relationship: a multi-group analysis of innovation by Ugandan rural and urban entrepreneurs

On the meaning of work: a theoretical integration and review

Agency and organization: intrinsic motivation, autonomy, and the self in psychological development

On happiness and human potentials: a review of research on hedonic and eudaimonic well-being

From ego depletion to vitality: theory and findings concerning the facilitation of energy available to the self

On energy, personality, and health: subjective vitality as a dynamic reflection of well-being

Happiness is everything, or is it? Explorations on the meaning of psychological well-being

Entrepreneurship and eudaimonic well-being: five venues for new science

Entrepreneurial job characteristics: an examination of their effect on entrepreneurial satisfaction

Feeling vital after a good night's sleep: the interplay of energetic resources and self-efficacy for daily proactivity

Entrepreneurship and well-being: the role of psychological autonomy, competence, and relatedness

Vigor as a positive affect at work: conceptualizing vigor, its relations with related constructs, and its antecedents and consequences

Feeling vigorous at work? The construct of vigor and the study of positive affect in organizations

The association of meaningfulness, well-being, and engagement with absenteeism: a moderated mediation model

A socially embedded model of thriving at work

Psychological, empowerment in the workplace: Dimensions, measurement and validation

The happy culture: a theoretical, meta-analytic, and empirical review of the relationship between culture and wealth and subjective well-being

Measuring meaningful work

Entrepreneurs' mental health and well-being: a review and research agenda

Beyond cultural values? Cultural leadership ideals and entrepreneurship

Understanding motivations for entrepreneurship. Retrieved from. Department for Business

Health of entrepreneurs versus employees in a national representative sample

Performance-based vs socially supportive culture: a cross-national study of descriptive norms and entrepreneurship

How does creativity at work influence employee's positive affect at work?

The Origin of Everyday Moods

Job crafting and its relationships with person-job fit and meaningfulness: a three-wave study

Striving for entrepreneurial autonomy: a comparison of Russia and the Netherlands. Manag

What is the value of entrepreneurship? A review of recent research

More than money: developing an integrative multi-factorial measure of entrepreneurial success

The Questionnaire for eudaimonic Well-Being: Psychometric properties, demographic comparisons, and evidence of validity

Contextualizing entrepreneurship-conceptual challenges and ways forward

Everyday entrepreneurship-a call for entrepreneurship research to embrace entrepreneurial diversity

Mental disorders in the entrepreneurship context: when being different can be an advantage

Entrepreneurship and well-being: past, present, and future

Well-Being Measures In Primary Health Care: The DepCare Project

Crafting a job: revisioning employees as active crafters of their work. Acad

Entrepreneurship's next act

Beyond survival: achieving new venture growth by building legitimacy

We would like to thank Daan van Knippenberg for his helpful comments on earlier drafts of the manuscript. This work was partially supported by the Portuguese Science and Technology Foundation [UID/GES/00315/2019].