Journal of Business Venturing 31 (2016) 449–467

Contents lists available at ScienceDirect

Journal of Business Venturing

Human capital in social and commercial entrepreneurship☆

Saul Estrin a,⁎, Tomasz Mickiewicz b, Ute Stephan b
a
London School of Economics, CEPR, and IZA, Houghton Street, London WC2A 2AE, UK
b
Aston University and Global Entrepreneurship Monitor UK, Aston Triangle, Birmingham, B4 7ET, UK

a r t i c l e i n f o a b s t r a c t

Article history: We advance research on human capital and entrepreneurial entry and posit that, in order to
Received 3 April 2015 generate value, social entrepreneurship requires different configurations of human capital
Received in revised form 13 May 2016 than commercial entrepreneurship. We develop a multilevel framework to analyse the com-
Accepted 18 May 2016
monalities and differences between social and commercial entrepreneurship, including the im-
Available online 29 May 2016
pact of general and specific human capital, of national context and its moderating effect on the
human capital-entrepreneurship relationship. We find that specific entrepreneurial human cap-
Keywords: ital is relatively more important in commercial entrepreneurship, and general human capital in
Human capital
social entrepreneurship, and that the effects of human capital depend on the rule of law.
Education
© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND
Commercial entrepreneurship
Social entrepreneurship license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Institutions
Rule of law
Property rights
Global entrepreneurship monitor
Multi-level modelling

1. Executive summary

Social entrepreneurs create value, but not in the same way as commercial entrepreneurs; they provide goods and services that
are neither supplied on the market nor addressed by the government. In so doing, they create social welfare while the financial
viability of their venture is their key constraint. In contrast, commercial entrepreneurs maximise private welfare and capture the
residual for themselves. Some of the value that social entrepreneurs create is therefore not incorporated in market transactions
directly: instead their objective is to generate positive external effects for the community from their activities. Because social
and commercial entrepreneurs differ in their goals and in the way that value is created, some of the necessary skills are common
but other skills and abilities may need to be different to succeed in social and commercial entrepreneurship respectively. Hence
the two types of entrepreneur may not be drawn from exactly the same pool of talent, which has implications for the relationship
between education, human capital and the choice of social as against commercial entrepreneurship as well as the context in

☆ Data for this study were provided by the Global Entrepreneurship Monitor (GEM), which is a consortium of national research teams. Names of the members of these
teams, the global coordination team, and the financial sponsors are published in the annual Global Entrepreneurship Monitor reports, which can be downloaded at
www.gemconsortium.org.We thank all the researchers and their financial supporters who made this project possible. The authors acknowledge valuable comments
from Simon Parker, Paul Reynolds, participants of the Aston Social Enterprise Research Day 2015, and from anonymous reviewers of both the Academy of Management
Annual Meeting and, last but not least, of this journal. Ute Stephan gratefully acknowledges financial support by the European Commission, Socioeconomic Sciences and
Humanities grant agreement 613500 (Seforїs project). Any errors are our own. The authors are listed in alphabetical order and contributed equally to the manuscript.
⁎ Corresponding author.
E-mail addresses: [email protected] (S. Estrin), [email protected] (T. Mickiewicz), [email protected] (U. Stephan).

http://dx.doi.org/10.1016/j.jbusvent.2016.05.003
0883-9026/© 2016 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
450 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

which the new venture creation occurs. These issues have not been systematically tackled in the literature and it is our intention
to address this gap.
We start with distinguishing between general human capital, which can be employed across a variety of occupations and in-
dustries, and specific human capital, for which the return to the individual is specific to an occupational context; say a job, sector,
or occupation. We argue that, in evaluating the decision to become either a social or a commercial entrepreneur, it is important to
consider how specific and general human capitals are combined. Both types of entrepreneurs will rely on entrepreneurship-
specific skills and knowledge. However, because the goals of the organisation are different, and more complex, additional abilities
may be needed for social entrepreneurship. In particular, social entrepreneurs must identify and exploit opportunities that are
both financially viable and can generate positive external effects. They therefore need to attend simultaneously to potentially con-
flicting social and economic logics in their enterprise in addition to developing capabilities for relating to a wide set of stake-
holders. Furthermore, their actions may need to be strongly embedded in local communities, in order to mobilise resources
and to stimulate the wider social impact. Hence, we argue that general human capital, which is associated with a more diverse
cognitive perspective, may have relatively greater significance for social entrepreneurs.
Moreover, one needs to consider the motivational aspects of education which reach beyond the rational calculations of human
capital theory. Precisely those human capital characteristics that enhance individual capacities to realise positive externalities may
also be those that motivate people to search for those opportunities in the first place. This is an example of the socialising effects
of education; values will influence an individual's cognitive alertness, and in turn, those values will be affected by what we are
able to notice and understand. As a result, we hypothesise about the relationship between specific versus general human capital
and entry into commercial as against social entrepreneurship, based on both the ability-enhancing (economic approaches) and
the motivation-shaping effects (psychological approaches) of human capital.
The balance of the effects of these factors will be moderated by the nature of the institutional environment in which entrepre-
neurs operate. Institutional economists highlight the rule of law as the fundamental aspect of institutions. Institutional structures
influence the returns to be obtained from human capital across different occupational choices. For example, a weak rule of law
increases the risk of expropriation of entrepreneurial returns, more so than of income from employment, shifting the balance
of incentives in favour of employment. Hence, both commercial and social entrepreneurs are more common under strong
constitutional-level institutions. Furthermore, the returns to different types of human capital in commercial as against social en-
trepreneurship may be sensitive to institutional contexts. Following this observation, we explore whether the rule of law has a
moderating effect on how different types of human capital affect the occupational choice of entrepreneurship comparing its social
and commercial forms.
Our research questions therefore span personal characteristics and national contexts, which suggests multi-level modelling as
the appropriate methodology. We test our hypotheses utilising data from the Global Entrepreneurship Monitor, a cross-national
individual data set, which we combine with indicators of the rule of law and other macro-economic data.
This study contributes to research on both human capital and on entrepreneurship, especially social entrepreneurship. It
broadens our understanding of the role of human capital in entrepreneurship by newly considering socialisation effects of general
human capital alongside the well-established ability effects of human capital. Analysing human capital in relation to social entre-
preneurship highlights a broader insight, namely the need to acknowledge differences among entrepreneurs, their preferences,
and the varying returns they seek. Moreover, we broaden the understanding of the role of human capital by considering context:
combining human capital theory and institutional theory.
Our research helps to answer whether social and commercial entrepreneurship compete for the same entrepreneurial talent;
we find that they do so only to a limited extent. In fact, we further corroborate a ‘crowding in’ effect whereby social entrepreneur-
ship attracts somewhat different talents into the entrepreneurial process. Thus, social entrepreneurship has value not only as a
way of addressing social problems, but also as a route into entrepreneurship for those who otherwise would remain in employ-
ment. Our work has implications for policy makers, but also for practitioners, who may find it valuable to consider entrepreneurial
teams in terms of their relative contributions of entrepreneurial talent and other skills.

2. Introduction

Many aspects of the venture creation process, such as the necessity to innovate, to take risks, and to coordinate resources
(Schumpeter, 1934), will be common to social and commercial entrepreneurs. In this respect, the two occupations may draw
on a similar entrepreneurial talent pool. However, while social and commercial entrepreneurship both create value, they differ
in the primary objective of the activity. Social entrepreneurs identify opportunities arising from “neglected problems in society
involving positive externalities”, which are neither incorporated into the market nor addressed by the government (Santos,
2012, pp.342). By realising those opportunities, social entrepreneurs create “social welfare” (Mair and Marti, 2006; Zahra et al.,
2009) while taking the financial viability of their venture as a constraint. In contrast, commercial entrepreneurs maximise “private
welfare” by creating value while capturing the residual for themselves (Santos, 2012). Because the goals and the way in which
value is created differ for social and commercial entrepreneurs, they may need to rely upon different skills and abilities; implying
that the two types of entrepreneur may not be drawn from exactly the same pool of talent.
Human capital is important for all entrepreneurs in making occupational choices in the labour market between paid employ-
ment and venture creation (Parker, 2009 for a review). It is useful to follow Becker (1964) in distinguishing between general
human capital, which can be employed across a variety of occupations and industries, and specific human capital, for which
the derived value is specific to a context, say a job, sector or occupation (Acemoglu and Pischke, 1998). We argue that, in
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 451

evaluating the decision to become an entrepreneur, it is important to consider how specific and general human capital are com-
bined. We explore the proposition that their relative weights will differ for social as against commercial entrepreneurship because
social entrepreneurs do not pursue the same objectives and this leads to differences in their activities.
Both types of entrepreneurs will rely on entrepreneurship-specific skills and knowledge. However, different and additional
abilities may be needed for social entrepreneurship, in particular to identify and exploit opportunities that can generate positive
external effects. Hence, we argue that general human capital, which is associated with a more diverse cognitive perspective, may
have relatively greater significance for social entrepreneurs. This is because the scope of their objectives is broader and their ac-
tivities are more complex; while employing market-based strategies, they also seek to create value that is not captured within
direct market transactions. Thus they need to attend simultaneously to potentially conflicting social and economic logics in
their enterprise (Battilana and Lee, 2014; Mair et al., 2015) as well as to develop capabilities for relating to a wide set of stake-
holders. Furthermore, their actions may need to be strongly embedded in local communities, to mobilise resources and stimulate
a wider social impact (Austin et al., 2006; Stephan et al., 2016).
To explain engagement in social entrepreneurship fully, we argue that one needs to consider both motivational aspects and
human capital theory. The existing social entrepreneurship literature focusses on the importance of other-regarding values and
prosocial-motivation of individuals as determinants of social entrepreneurship (e.g., Miller et al., 2012; Renko, 2013). The specific
skills and broader abilities needed to act upon those values have been rarely considered (exceptions include: Bacq et al., 2014;
Parker, 2008). Yet an occupational choice of social entrepreneurship is not necessarily purely driven by values: an individual
characterised by other-regarding values may follow a commercial project and realise his/her values outside that project, say by
philanthropic giving (Mickiewicz et al., 2016). The choice to engage in social entrepreneurship will be conditioned on the
individual's human capital that enables the identification and formation of a relatively complex venture which is characterised
by the joint supply of commercial output and of positive external effects. Moreover, the human capital characteristics that en-
hance individual capacities to realise positive externalities may also be those that motivate people to search for those opportuni-
ties; an example of the socialising effects of education proposed in value theory (Schwartz, 2006). Values will influence an
individual's cognitive alertness, and in turn, those values will be affected by perceptions: what we are able to notice and under-
stand.1 This leads us to hypothesise about the relationship between both specific and general human capital and entry into com-
mercial as against social entrepreneurship, considering the ability-enhancing (as typically discussed in economic approaches) and
the motivation-shaping effects (as alluded to in psychological approaches) of human capital.
The impact of these factors will be moderated by aspects of the institutional context. In considering differences in the propen-
sity to enter social and commercial entrepreneurship across nations, the balance of returns from human capital from different oc-
cupational choices is contingent on country-specific institutional characteristics. North (1990) stresses the importance of market
supporting institutions for economic performance, and since then, institutional economists have highlighted the rule of law as the
fundamental aspect of institutions (Acemoglu and Johnson, 2005). A weak rule of law increases the risk of expropriation of entre-
preneurial returns, more so than of income from employment, shifting the balance of incentives to the latter (Estrin et al., 2013a,
b). Hence, both commercial and social entrepreneurs are more common in societies with strong constitutional-level institutions
(Estrin et al., 2013a, b). Furthermore, returns to different forms of entrepreneurship and to different types of human capital
may be sensitive to institutional contexts, and this applies to social as well as to commercial entrepreneurship. The literature
has not previously attempted to unpack these complex cross-country differences by investigating the role that national institu-
tions may play in moderating the effects of individual-level human capital characteristics on different forms of entrepreneurship.
In particular, we explore whether the rule of law has a moderating effect on how different types of human capital support entre-
preneurship comparing its social and commercial forms.
Our research questions therefore span personal characteristics and national contexts, which means that multi-level modelling
is the appropriate methodology. We theorise about why the rule of law moderates the impact of general and specific human cap-
ital on individual choices to become a social entrepreneur, in ways that are different for commercial entrepreneurs. We test our
hypotheses on a large cross-national data set, consisting of population-representative surveys combined with independent indica-
tors of the rule of law.
This study contributes to research on both human capital and entrepreneurship, especially social entrepreneurship. It broadens
our understanding of the role of human capital in entrepreneurship by newly considering socialisation effects of general human
capital alongside the well-established ability effects of human capital. Analysing human capital in relation to social entrepreneur-
ship highlights a broader insight, namely the need to acknowledge heterogeneity among entrepreneurs, their preferences, and the
varying returns they seek — when investigating the effects of human capital. Moreover, we offer a more contextualised under-
standing of the role of human capital by combining human capital theory and institutional theory. Our multi-level framework le-
verages institutional theory to make predictions about the heterogeneous effects of different types of human capital in a variety of
country contexts. This responds to calls for greater consideration of context in entrepreneurship research (Zahra and Wright,
2011; Welter, 2011). In sum, while the link between human capital and entrepreneurial performance is well documented, how
human capital influences entrepreneurial entry is less well researched and the findings are conflicting to date. Our research sug-
gests that theorising about the heterogeneous effects of human capital, among different types of entrepreneurs as well as across
different institutional settings, will lead to sharper results.

1
Schwartz et al. (2000) provide arguments and general evidence that values influence information processing and attention focus.
452 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

We advance the understanding of social entrepreneurship by drawing attention to the important role of ability and human
capital, where the current discourse is dominated by a focus on motivation. In so doing we respond to a call by Parker (2008)
to provide a simple but theoretically grounded typology to understand who becomes a social entrepreneur, and in particular
how differences in human capital drive different choices in entrepreneurship. This helps to answer whether social and commercial
entrepreneurship compete for the same entrepreneurial talent — they do so only to a limited extent. In fact, our findings further
corroborate a ‘crowding in’ effect, whereby social entrepreneurship attracts new talent into the entrepreneurial process (Estrin
et al., 2013a, b).

3. Theoretical framework

3.1. Entrepreneurship

Entrepreneurship – “new entry” through the efforts towards the creation of a viable business (Gartner, 1989; Reynolds et al.,
2005) – results from an individual's occupational choice to work on his/her own account (e.g., Hébert and Link, 1988). Commer-
cial entrepreneurship implies entry into business activities that rely on market exchange structures with the entrepreneur's objec-
tive being to maximise profits. Thus, commercial entrepreneurs capture privately the residual value created within their
enterprise. In contrast, social entrepreneurship definitions commonly stress that the objective of the organisation is to address so-
cial rather than commercial needs, by social wealth creation (Mair and Marti, 2006; Zahra et al., 2009). Santos (2012) moves this
literature forward by observing that the typical activities by social entrepreneurs can be conceptualised as generating positive ex-
ternalities — value that is created by the enterprise primarily accrues to wider society and will not be contained within the market
exchange in which the enterprise is involved. Of course, commercial entrepreneurs may also generate positive externalities
(e.g., through generating employment), as well as negative ones (e.g. pollution). However, the generation of positive externalities
can be viewed as the objective of social entrepreneurship rather than a potential side effect as in commercial entrepreneurship.
Past research has highlighted how the social goals and positive externalities pursued by social entrepreneurs add greater com-
plexity to their activities compared to commercial entrepreneurs. In contrast to either advocacy- or government organizations, so-
cial enterprises mobilise for bottom-up social change through empowerment processes (Santos, 2012; Stephan et al., 2016). This
extends the scope of activities beyond the boundaries of the enterprise. It typically entails that the enterprise is open to diverse
stakeholder influences, is embedded in local communities, and is ‘relational’ in its approach by shaping networks across sectors
(commercial, non-profit, and government) to stimulate social change as well as to leverage resources (Stephan et al., 2016).
This compares with transactions in commercial enterprises that are more focussed and address a narrower set of stakeholders (fi-
nanciers, suppliers, employees) who are largely aligned with the single goal of residual value capture within the organisation.
An increasing number of studies document how the different logics of action associated with social and commercial goals, and
the need to consider both simultaneously, can lead social entrepreneurs into trade-offs, and increase the complexity of decision-
making (e.g. Mair et al., 2015; Battilana and Lee, 2014, for a review). For instance, to gain legitimacy and resources from funders,
social entrepreneurs need to appeal to a commercial logic, demonstrating their management capability and perhaps offering div-
idends, aligning themselves with respective industry organisations. Yet to be credible to their beneficiaries, they may be expected
to closely engage with local stakeholders, to measure social impacts, to reinvest profits in social impact creating activities, and to
embed themselves in cross-sector networks and partnerships (e.g., Austin et al., 2006; Di Domenico et al., 2010; Pache and Santos,
2013).

3.2. Human capital and entrepreneurship

The economic analysis of an individual's choice to become an entrepreneur focusses on alternative occupations in the
labour market; paid employment as against entrepreneurship (Lucas, 1978). Rational utility maximising individuals choose
to become entrepreneurs if the expected utility they gain from it is higher from that obtained from paid employment. Greater
expected utility from entrepreneurship may be caused, for example, by differences in ability generated by variation in human
capital (van Praag, 2005); in attitudes to risk (Khilstrom and Laffont, 1979); or in utility functions, for example placing greater
emphasis on non-pecuniary rewards such as independence and job satisfaction (van Praag & Versloot, 2008). There have been
as yet few applications of this framework to the phenomenon of social entrepreneurship, though Parker (2008) develops a life
cycle model in which differences in time preferences lead individuals to become social entrepreneurs at different points in
their lives.
We seek to extend economic modelling of the occupational choice of entrepreneurship to include social entrepreneurship by
focussing on the differential effects of human capital. Our approach draws on the distinction between general and specific human
capital (Becker, 1964), with general skills being typically acquired through formal education, and specific skills via experience
(Unger et al., 2011). In line with extant literature; we consider the determinants of entrepreneurial occupational choice to be
the returns from general and specific human capital, compared to their opportunity costs (Le, 1999; Davidsson and Honig,
2003; Parker, 2011; Unger et al., 2011). However, relative to commercial entrepreneurship social entrepreneurship entails broader
objectives and activities, which may alter the balance of advantage of general versus specific human capital. We argue that the
main difference will be through the effect of general human capital; the broader scope of the goals and activities of social entre-
preneurs will draw relatively more intensively on general compared to specific human capital.
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 453

The goals of social and commercial entrepreneurs are sufficiently differentiated that one might expect each activity to attract
individuals characterised by quite different motivations, or value sets (preferences).2 Yet motivation cannot be the sole
differentiator of social and commercial entrepreneurship; for example other-regarding values may also be realised both by
those in salaried employment and in purely commercial entrepreneurship, outside their occupation, for example by charitable ac-
tivity and giving. Thus, it is also the ability to realise benefits from the joint supply of a commercial product/service and positive
external effects, which induces an individual to create a social venture. Both of these elements – the abilities set and the values set
– are defined by his/her human capital characteristics. Hence, we need to understand the features of human capital which drive
the choice of commercial and social entrepreneurship, and which will be associated both with values and with the relative rates of
return.

3.2.1. General human capital and entrepreneurship

Investment in general human capital has positive effects on both entrepreneurial entry and performance because it leads to a
broad knowledge base which enables individuals to integrate new knowledge and adapt to new situations more easily (Lazear,
2005). In particular, it enhances the individual's ability to discover and exploit opportunities (e.g., Davidsson and Honig, 2003;
Unger et al., 2011). However, empirical research linking education to commercial entrepreneurship entry yields a mixed pattern
(also Le, 1999). Some studies report that education is positively associated with the likelihood to engage in commercial entrepre-
neurship (e.g., Arenius and Minniti, 2005; Block et al., 2013; Minniti et al., 2005b; Parker, 2009, 2011). Other studies find no re-
lationship (Van der Sluis et al., 2005, 2008). Finally, some research suggests that it is important to consider the type of education:
beyond secondary education, higher education may not have an additional positive effect on entry into commercial entrepreneur-
ship (e.g., Parker and Belghitar, 2006). This is possibly due to rising opportunity costs, because more highly educated individuals
are likely to be offered managerial jobs in wage employment that like entrepreneurship entail considerable decision latitude and
variable incentives — yet entail less risk bearing.3
Interestingly, there is as yet only limited analysis of the effects of education on social entrepreneurship. Nonetheless, findings
available through the Global Entrepreneurship Monitor executive and special topic reports suggest a positive relationship (Bosma
and Levie, 2010; Terjesen et al., 2012). However studies that explore the education-social entrepreneurship relationship control-
ling for alternative explanations and using inference statistics are rare (for an exception and confirming a positive relationship see
Van Ryzin et al., 2009). Given its broader scope, higher levels of education may be particularly important to identify and exploit
opportunities for social entrepreneurship. While investment in education is likely to increase the returns to commercial entrepre-
neurship relative to alternative occupations, these returns may be even higher in social entrepreneurship. Note that to conform to
the objective of social welfare maximisation for social entrepreneurs, returns must be defined broadly to incorporate the overall
value generated by the enterprise, whether the residual is captured privately or not.4
Psychological approaches stress that education, and especially higher education, has a two-fold socialising effect. It enhances
flexibility, openness and independent thinking (Kohn and Schooler, 1983; Schwartz, 2006) — as also emphasised by economic ap-
proaches. In addition, higher education has been found to enhance other-regarding values and engagement in self-initiated, pro-
social actions such as volunteering and political activism (Abramson & Inglehart, 1994; Schofer and Fourcade-Gorinchas, 2001;
Schwartz, 2006, 2010). Thus, higher education instills preferences and motivations consistent with the core aspiration of social
entrepreneurs to contribute to the welfare of others and to create societal wealth (Stephan et al., 2015).5 These motivations
are less likely to sit comfortably with commercial entrepreneurship (Lukes and Stephan, 2012; Noseleit, 2010). Thus, we argue
that higher levels of education may have a more pronounced effect on social as against commercial entrepreneurs. This is because
socialisation element of education might favour both a better understanding of the more complex nature of social entrepreneurial
opportunities and the objective of social welfare maximisation rather than profit maximisation, introducing a ‘pro-social bias’ (Nga
and Shamuganathan, 2010).
Taken together, the socialising effects of education and human capital theory lead us to expect relatively stronger effects from
education (general human capital) on social than commercial entrepreneurship: (i) adopting other-regarding values is a necessary
condition of social entrepreneurship; and (ii) the latter also requires a broader set of skills conducive to identifying opportunities
in producing positive external effects.

Hypothesis 1a. Individuals who have completed tertiary (higher) education have a greater likelihood to choose social, compared
to commercial, entrepreneurial entry.

The same reasoning does not apply to specific (entrepreneurship related) human capital as we elaborate next.

2
Values and motivations differ in their level of abstraction, although the terms are often used interchangeably. Values refer to general life goals and are more abstract
than motivations which typically are focussed on specific objects or events. In the economics literature, both concepts are typically referred to as preferences. Shalom
Schwartz introduced a general theory of values which is now corroborated by evidence from over 80 countries. It differentiates two broad value dimensions including
self- and other-regarding values and openness to change vs. conservation values, as well as 10 more specific value types (see Schwartz, 2012, for an introduction).
3
The findings on human capital may also be confounded with results on financial capital (Le, 1999); in our study we will proxy for the latter.
4
Bacq et al. (2014) find that in relative terms, the weighting of entrepreneurial skills in human capital is lower for social than for commercial entrepreneurs, while the
former may be superior in general skills; those skills enable them to identify the nature of positive external effects.
5
There may also be a self-selection effect such that those with pro-social values self-select into higher education.
454 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

3.2.2. Specific human capital and entrepreneurship

With respect to specific entrepreneurial human capital, social and commercial entrepreneurial activities have much in com-
mon. Many aspects of the venture creation process are the same for both, notably the necessity for the entrepreneur to identify
business opportunities, to bear the risks involved in new venture creation, and to identify and organise the resources necessary for
success (e.g., Dacin et al., 2010; Meyskens et al., 2010; Reynolds, 2010). All these activities require entrepreneurial skills, i.e. the
specific know-how related to the starting and running of a business, often acquired through previous start-up experience (Unger
et al., 2011). Such skills involve being sensitive to opportunities and crafting business models to exploit them as well as being
proactive and finding solutions to the various obstacles in the way of creating a new business. These skills are not necessarily val-
ued in wage employment where the division of labour gives rise to organisational hierarchies and processes that typically require
some degree of compliance as well as deeper subject matter expertise (Lazear, 2005). This would suggest that some individuals
with specific (entrepreneurial) skills would choose to set up social and commercial enterprises in preference to accepting paid
employment. In economic models of occupational choice of entrepreneurship, individuals seek careers that will maximise their
benefits from their human capital. Individuals with entrepreneurial skills choose entrepreneurship over paid employment when
the returns from the former exceed the market employment wage. In this respect, unlike general human capital, specific entre-
preneurial capital may entail lower opportunity costs vis-a-vis paid employment.
Yet there may also be differences in the relative productivity of the entrepreneurial skills when deployed in social as against
commercial entrepreneurship. This implies that we need to consider not only the differences in terms of opportunity cost of both
types of entrepreneurship against salaried work, but also the opportunity cost of one type of entrepreneurship versus another.
This may lead to differences in the relative likelihood of choosing commercial versus social entrepreneurship for those endowed
with entrepreneurial skills.
In particular, entrepreneurial competence may weigh less heavily in the set of social entrepreneurs' skills; this is likely to be
only one important ingredient alongside other skills necessary in creating positive externalities (and which are captured by gen-
eral human capital). Social entrepreneurship will require additional distinct skills especially with regard to the generation of social
impact (Stephan et al., 2016). Considering the emphasis on social goals, it is likely that, purely entrepreneurial skills will weigh
relatively less heavily in individuals' considerations to start a social venture. These individuals will also be concerned about
their social impact skills and indeed may self-select out of venture creation efforts if they perceive to lack those skills; regardless
of their entrepreneurial skills. The more diverse skill basis appropriate for social entrepreneurship contrasts with commercial en-
trepreneurship, where specific entrepreneurial skills generally yield a stronger impact than general human capital (i.e. education,
Unger et al., 2011). That in turn implies higher opportunity cost of the former for commercial entrepreneurship.
Thus while specific human capital is important for social as well as commercial entrepreneurship, the relative weight of spe-
cific human capital is higher for the latter.

Hypothesis 1b. Individuals who have specific (entrepreneurial) human capital have a greater likelihood to choose commercial,
compared to social, entrepreneurial entry.

3.3. Rule of law, human capital and entrepreneurship

The relationship between human capital, both general and specific, and an individual's occupational choice will also be sensi-
tive to the institutional context in which those choices are made. When considering social as against commercial entrepreneur-
ship, the key contextual element is the institutional structure because, as stressed by North (1990) and Baumol (1990),
institutions shape private incentives by defining individual returns and opportunity costs. Therefore the national institutional
framework affects individual choices about engagement in entrepreneurship, including entry into social as against commercial en-
trepreneurship (Baker et al., 2005; Estrin et al., 2013a, b).
Institutional arrangements in a particular country affect the balance of returns from different occupations. Baumol (1990)
analysed this phenomenon by hypothesising that institutions create incentives which make it more or less attractive for individ-
uals to pursue different forms of entrepreneurship: productive, unproductive and destructive. In many countries, the main threat
to entrepreneurial success is expropriation or graft and a strong rule of law limits or prevents that (Acemoglu and Johnson, 2005).
This implies that the relationship between both general and specific human capital and entrepreneurial activity may vary signif-
icantly, depending on the strength of the rule of law, which in turn leads to security of property and other economic rights
(Epstein, 2011).
In particular, the rule of law influences the extent to which the potential returns from human capital investments can be cap-
tured by the individual and his/her enterprise. For example in the former Soviet Union, where entrepreneurial activities were
largely illegal, the impact of human capital on incomes through the choice of an entrepreneurial career was much lower than
in countries such as the US, where the rewards to entrepreneurship are more protected (Aidis et al., 2008).
In situations where the rule of law is weaker, so that the threat of expropriation is greater, the appropriable returns from
human capital in entrepreneurship compared to alternative occupations are skewed against entrepreneurship. Those with access
to state power or means of violence (officials or criminal gangs) can seize the rewards of successful entrepreneurs because the
latter are relatively few in number, easily identified and perhaps not strongly supported socially (Aidis et al., 2008). It is harder
to seize a share of an individual's income from employment because their numbers are much greater and because social cohesion
supporting that group will be stronger. Variations in the rule of law may therefore help to explain the cross-national differences in
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 455

the returns to human capital observed in previous research on entrepreneurship (Arenius and Minniti, 2005; also Van der Sluis
et al., 2005, 2008 for reviews). A weaker rule of law is likely to constrain relatively more the potential of entrepreneurs who
are characterised by valuable human capital, because their upside gains are threatened to a greater extent. Thus, we hypothesise
a positive moderating effect of the rule of law on the propensity of individuals with more valuable human capital, both general
and specific, to engage in entrepreneurship.
These arguments can be refined to distinguish between the effects of the rule of law on social as against commercial entrepre-
neurship. Where the threat of expropriation is higher, the expected return to entrepreneurship that can be captured privately by
the new organisation is reduced. Hence the rule of law is crucial for commercial entrepreneurs in allowing them to capture the
returns from their own innovations (e.g. Estrin et al., 2013a, b). The implication of the rule of law may differ for social entrepre-
neurship. For commercial entrepreneurship returns are relatively easily identified and subject to expropriation when the rule of
law is weak. In contrast social entrepreneurship produces positive external effects, but the financial gains to the enterprise may
be modest or non-existent. In consequence, social entrepreneurs may be less prone to expropriation under a weak rule of law;
a surplus is not easy to confiscate if shared and spread thinly across thousands of micro credit recipients, for example. Thus
the deterrence effect of a weak rule of law may be less binding on social entrepreneurs. Our argument here is consistent with
the view that social entrepreneurship is more likely to be prevalent where institutional voids are pronounced (Mair and Marti,
2009).6
Therefore we posit that the positive moderating effect of the rule of law on the relationship between human capital and en-
trepreneurial activity may be stronger for commercial entrepreneurs. While a weak rule of law reduces both financial and social
returns the threat of expropriation is stronger for successful commercial entrepreneurs. Financial returns to social entrepreneurs
may be smaller and social returns more difficult to identify and expropriate than the commercial returns captured within a new
organisation. These arguments suggest that weak rule of law may have a greater negative effect on commercial venture creation
by those with more valuable human capital than on social entrepreneurship.
So far we did not distinguish between the general and specific form of human capital, yet the mechanism that links the mod-
erating effect of the rule of law may differ for each of these, even if the expected direction of the effect remains the same. In par-
ticular, for general human capital, in totalitarian societies, where law remains subordinate to political power, such as the former
Soviet Union, education itself may be biased in such a way as to hinder the development of entrepreneurial skills such as oppor-
tunity recognition (Aidis et al., 2008).
In turn, for specific human capital, with a strong rule of law, entrepreneurial success is to a greater extent due to entrepre-
neurs' individual experience, skill and effort rather than to external circumstances (Aidis et al., 2008; Estrin et al., 2013a, b).
There is also more opportunity to acquire entrepreneurial experience. And finally, as already understood by Tocqueville, under
conditions of political freedom, civic skills are developed and those in turn enhance the quality of entrepreneurial skills (De
Tocqueville, 2003 [1835]). For all these reasons, specific entrepreneurial human capital may become more valuable when the
rule of law is stronger.
Thus, while our conclusions for both types of human capital are similar, the reasoning may slightly differ. We hypothesise:

Hypothesis 2a. Where the rule of law is stronger, the likelihood of commercial entrepreneurial entry of those with general human
capital is greater relative to social entrepreneurship.

Hypothesis 2b. Where the rule of law is stronger, the likelihood of commercial entrepreneurial entry of those with specific human
capital is greater relative to social entrepreneurship.

Fig. 1 summarizes our hypotheses.

4. Methods

4.1. Sample, measures and modelling strategy

We merge Global Entrepreneurship Monitor (GEM) data in 2009 with a large group of independent country-level institutional
indicators and macroeconomic controls from different sources.7 Social entrepreneurship is not part of the regular GEM survey, but
was included in 2009 as a special topic, and this provides the basis for our individual level dataset. We exclude some countries
based on quality issues, following a recommendation in the GEM 2009 report (Terjesen et al., 2012), and opt for a rich model,
with low omitted variable bias, yielding a usable set containing 68,885 observations from 37 countries. Our hypotheses focus
on the differential effect of general and specific human capital on the likelihood that an individual starts a social or a commercial
enterprise. Hence, in our core set of models, we constrain our sample to those respondents currently in the start-up process. This

6
The meaning of institutional voids differs in the literature that investigates context and social entrepreneurship sometimes referring to voids related to weak con-
stitutional level institutions (i.e. weak rule of law) and at other times to weak government provision of services (Estrin et al., 2013a, b; Hoogendoorn and Hartog, 2011;
Mair et al., 2012; Stephan et al., 2015). These differences are not our focus here. Low scores on our measure of constraints on the executive (rule of law) are similar to
other conceptualisations of constitutional level institutional voids. In addition, we include a set of country-level control variables that proxy for institutional voids gen-
erated by lack of government provision of services.
7
With very few exceptions, the data consist of representative samples of at least 2000 individuals in each country. The samples are drawn from the working age pop-
ulation which avoids the potential selectivity bias that could affect studies which focus on existing entrepreneurs. National datasets are harmonised across all countries
included in the survey (Reynolds et al., 2005).
456 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

Fig. 1. Multi-level research framework and hypotheses.

leaves us with a sample of 6901. In robustness checks we also use a design where individual decisions to engage in the two types
of entrepreneurial entry are contrasted with those of people not engaged in start-ups at all, that is based on the full 68,885 ob-
servations. Table 1 provides an overview of definitions and sources for all variables in this study. Table 2 lists the 37 countries and
descriptive statistics for the variables central for this study for each country.

4.1.1. Social and commercial entrepreneurship

For our core set of results, we apply a multilevel logit model comparing the individual likelihood of social and commercial en-
trepreneurial entry respectively. Our baseline category is commercial entry (coded “0”); social entry is coded “1”. We conduct ro-
bustness tests focussing only on those who undertake start-up activity while being in paid employment (Table 5 columns 2 and
3). As mentioned above, in further robustness checks we utilise the whole sample and test our hypotheses using multinomial
multilevel modelling with three possible outcomes: no engagement in employment, those engaging in commercial start-up and
those engaged in social start-up. In all these cases our results correspond to those reported below as the core.
In this study, we define social and commercial entrepreneurial entry in terms of start-up or nascent activity. To be classified as
starting-up or nascent entrepreneurs in GEM, respondents answered affirmative that (a) they are alone or with others are currently
trying to start a new business, (b) they have actively taken action to start the new business over the past 12 months, (c) they will
at least part-own this business, and (d) they have not paid wages, salaries, or ‘in kind’ for more than three months. Respondents
were asked a corresponding set of questions about starting and owner-managing “any kind of activity, organisation or initiative
that has a particularly social, environmental or community objective” to be identified as social entrepreneurs. Respondents who
stated that their social entrepreneurial activity was the same as their commercial entrepreneurial activity (declared earlier in
the survey) were treated as social entrepreneurs, as in Hoogendoorn and Hartog (2011).

4.1.2. Individual-level predictors (H1): human capital

As our preferred measure of general human capital, we consider tertiary (higher) Education to be the relevant aspect of human
capital accumulation. This is because it is to a significant extent under the control of the individual and has been closely associated
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 457

Table 1
Definitions of variables and descriptive statistics.Source: GEM 2009 unless specified otherwise (t-1) indicates lagged variables. Institutional variables: values before
mean-centring.

Variable Definition Mean S.D.

Dependent variable (baseline category for logit model: engaged in commercial start-up)
Social enterprise Respondent actively involved in social start-up (1, 0 otherwise) 0.42 0.49

Dependent variable (baseline category for multinomial model: not engaged in start-up)
Social enterprise (SE) Respondent actively involved in social start-up (1, 0 otherwise) .027
start-up
Commercial enterprise (CE) Respondent involved in commercial start-up (1, 0 otherwise) .037
start-up

Individual-level variables
Education: Highest level of education (dummy coded, baseline: no education beyond primary)
Some secondary - Respondent has incomplete secondary education .308
Secondary - Respondent has completed secondary education .348
Tertiary - Respondent has completed tertiary education .344
Entrepreneurship experience Respondent sold, shut down, discontinued or quit a business in the past 12 months that s/he owned and .036 .186
managed, and this business continued its activities after the entrepreneur disengaged
Female Female (1, 0 male) .527 .499
Social entrepreneur (SE) Respondent is currently owner-managing a young or established social enterprise (1, 0 otherwise) .027 .161
engagement
Commercial entrepreneur Respondent is currently owner-managing a young or established commercial enterprise (1, 0 otherwise) .106 .308
(CE) engagement
Business angel in last 3 years “Have you, in the past three years, personally provided funds for a new business started by someone else, .032 .177
excluding any purchases of stocks or mutual funds?” (1-yes, 0-no)
Know an entrepreneur Respondent knows an entrepreneur (1, 0 otherwise): “Do you know someone personally who started a .378 .485
business in the past 2 years?” (1 = yes, 0 no)
Fair of failure Respondent would not start a business out of fear of failure (1, 0 otherwise) .365 .481
Age Age of respondent between 15 and 64 (inclusive) 40.80 13.16
In employment Respondent is currently in full or part time employment (1, 0 not in employment) .575 .494

Country-level variables
Executive constraints (t-1) Polity IV ‘Executive Constraints’; scores from 1 = ”unlimited authority” to 7 = ”executive parity or 6.17 1.59
subordination”; higher value: less arbitrariness
Government spending (t-1) Government spending/GDP (authors' calculations, based on Heritage Foundation data) 36.67 9.98
% Working age population Percentage of working age population in total population (based on World Bank) 66.34 4.28
(t-1)
% Unemployment (t-1) Percentage share of unemployed in economically active population in 2008 7.78 3.66
Survival vs. self-expression Survival vs. self-expression culture, averaged scores from 1999/2000 and 2005/2008 World Values Survey .536 .964
culture
Traditional versus rational Traditional vs. rational culture, averaged scores from 1999/2000 and 2005/2008 World values survey −.031 .803
culture
Logarithm of GDP per capita Gross Domestic Product (GDP) per capita in purchase power parity (in natural logarithm) 9.84 0.78

Robustness checks
Start-up skill Respondent believes has skills for start-up (1-yes, 0-no skill) .527 .499
Household income Head of household's income, dummy coded categorised into three groups of equal number of respondents .33
for each country (baseline: bottom third) each
Rule of law Freedom House Rule of law as measured in the Freedom House data base 10.408 4.436
Rule of law World Bank Rule of law as measured in the World Bank Worldwide Governance Indicators data base .644 .216

with identification of opportunities in entrepreneurship (e.g., Van Praag et al., 2013). Education is measured with four categories
indicating whether the individual's highest completed level of education is tertiary education, secondary education, incomplete
secondary, or less than that. The latter is the reference category in our regressions.
We use Entrepreneurship experience as a proxy for specific human capital, in line with the conceptualisation of specific human
capital as skills obtained through relevant, specific practise (Becker, 1964). We construct this indicator based on two questions in
the GEM survey. We coded respondents as possessing entrepreneurship experience if they answered affirmatively that (a) they
have sold, shut down, discontinued or quit a business in the past 12 months that they owned and managed, and (b) that this
business continued to exist after the respondent departed from it. Thus, our indicator of specific human capital captures recent
start-up experience where the venture had some degree of success.8 All human capital measures come from the GEM survey.

4.1.3. Individual-level control variables

We include a set of individual level control variables taken from the GEM adult population survey. Previous research shows
that men, middle-aged, and people in employment are more likely to start a business (Reynolds et al., 1999; Minniti et al.,

8
We add as a control variable whether a respondent is currently running a commercial or social enterprise as owner-manager. However, we do not use this as our
measure of specific human capital as it confounds effects of skills with opportunity costs when the decision to start another business is taken.
458 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

Table 2
Country descriptive statistics for main variables.

Country Social relative to commercial Entrepre-nerial Some secondary Secondary Tertiary Executive
start-up experience education education education constraints

Algeria 0.45 0.02 0.31 0.35 0.28 5

Argentina 0.35 0.06 0.13 0.29 0.40 6
Belgium 0.51 0.05 0.11 0.41 0.41 7
Brazil 0.12 0.09 0.10 0.40 0.11 6
Chile 0.49 0.04 0.11 0.46 0.28 7
China 0.31 0.04 0.32 0.35 0.21 3
Colombia 0.42 0.06 0.14 0.38 0.29 6
Croatia 0.67 0.04 0.11 0.61 0.26 7
Finland 0.45 0.06 0.08 0.42 0.42 7
Germany 0.32 0.01 0.18 0.33 0.49 7
Guatemala 0.20 0.02 0.16 0.21 0.02 6
Hungary 0.49 0.02 0.27 0.30 0.30 7
Iran 0.48 0.06 0.14 0.32 0.30 2
Israel 0.45 0.01 0.07 0.38 0.51 7
Italy 0.54 0.08 0.26 0.49 0.19 7
Jordan 0.36 0.07 0.34 0.32 0.22 3
Korea 0.43 0.04 0.09 0.32 0.58 6
Latvia 0.36 0.03 0.07 0.56 0.37 7
Malaysia 0.02 0.02 0.17 0.30 0.24 5
Netherlands 0.35 0.01 0.08 0.63 0.29 7
Norway 0.50 0.05 0.00 0.40 0.52 7
Peru 0.35 0.06 0.15 0.41 0.20 7
Romania 0.73 0.02 0.14 0.49 0.21 7
Russia 0.44 0.11 0.07 0.07 0.86 4
Saudi 0.16 0.07 0.15 0.45 0.30 1
Arabia
Slovenia 0.42 0.03 0.00 0.57 0.37 7
South Africa 0.49 0.03 0.38 0.41 0.07 7
Spain 0.31 0.04 0.31 0.16 0.42 7
Switzerland 0.43 0.03 0.06 0.59 0.32 7
Uganda 0.17 0.06 0.22 0.04 0.07 3
UK 0.32 0.02 0.18 0.42 0.39 7
United 0.54 0.05 0.10 0.24 0.63 7
States
Uruguay 0.28 0.04 0.42 0.33 0.17 7

2005a, 2005b). Past research and conceptual arguments also suggest that the gender and age distributions differ between social
and commercial entrepreneurship; relative to commercial entrepreneurship, women, as well as both younger and older individ-
uals, are more likely to enter social entrepreneurship (Estrin et al., 2013a, b; Parker, 2008). We therefore include a dummy var-
iable for gender with female = 1 (Female), introduce the individual's Age (also in a quadratic form) and employment status (In
employment) as control variables to address these possibilities. Furthermore, those who currently run a commercial or social ven-
ture may be likely to continue in the same line of entrepreneurship; thus we also control for an individual's commercial and social
entrepreneurship engagement (CE engagement and SE engagement), defined as currently running a business, measured through cur-
rently being a young or established entrepreneur in the GEM dataset. Young or established entrepreneurs are those whose busi-
ness has paid salaries, wages or in kind for three months or longer and they own and manage that business.
Past research points to the importance of access to capital for potential entrepreneurs who engage in the start-up process, both
commercial (e.g., Ho and Wong, 2007; Korosteleva and Mickiewicz, 2011) and social (Meyskens et al., 2010). It has also been
identified as critical to include alongside human capital variables, as the two forms of capital are correlated, and otherwise an
omitted variable bias could result. We proxy access to capital first through the GEM question whether the respondent has
been a Business Angel in the past 3 years. In our robustness checks, we also control for respondents' household income, measured
as being in the lower, middle or upper third of household incomes in the respondent's country of residence. The inclusion of
household income reduced the available sample size, and more importantly, missingness in income is highly correlated with
age, gender and employment status; therefore we do not have the same level of confidence in these tests. Access to capital
may play a more significant role in commercial compared to social projects.
We also control for knowing an entrepreneur as it influences individual's engagement in business start-up positively, e.g. via
role modelling effects (e.g., Arenius and Minniti, 2005; Wagner and Sternberg, 2004). As the prevalence of commercial entrepre-
neurs is higher than of social entrepreneurs (Terjesen et al., 2012), knowing an entrepreneur may affect the individuals' choice of
commercial over social entrepreneurship. Finally, while recent evidence suggests that risk aversion does not differ between entre-
preneurs and non-entrepreneurs; loss aversion does. In our regressions we address this by controlling for the Fear of failure, which
captures the idea that the potential loss from entering entrepreneurship is weighed more heavily than the potential gains, i.e. loss
aversion. This is the dimension of risk that has been confirmed experimentally as relevant for entrepreneurs (Koudstaal et al.,
2015). Social compared to commercial nascent entrepreneurs face a higher risk of failure (Renko, 2013); however, the impact
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 459

of failure may be more serious for commercial entrepreneurs given the higher value of personal assets typically invested in a com-
mercial project. On balance, we expect that greater fear of failure may lead individuals to enter into social instead of commercial
entrepreneurship.

4.1.4. Country-level predictor: rule of law (H2)

We measure rule of law using the Polity IV indicator of efficient constraints on the arbitrary power of the executive branch of
the government, Constraints on the executive. Compared to other indicators of institutional quality, this measure has the advantage
of capturing the arguably key necessary condition of the rule of law (Epstein, 2011), which in the context of entrepreneurial ac-
tivity is related to the risk of expropriation (Acemoglu and Johnson, 2005). As robustness checks we include results which sub-
stitute for this variable with alternative measures of the rule of law, in particular the Rule of Law indicator from the Freedom
House dataset and the Rule of Law indicator drawn from the World Bank Worldwide Governance Indicator database (as compiled
by the comprehensive Quality of Government database: Dahlberg et al., 2016).9 However we consider the Polity IV measure su-
perior to the other two: as argued by political scientists, Polity IV stands out as a highly transparent and robust set of measures
ranked higher than that constructed by Freedom House (Munck and Verkuilen, 2002). The one offered by World Bank is probably
even further away from the rule of law concept, as it merges a number of disjoint dimensions into one factor. In addition, com-
pared with the other two measures, Polity IV also has weaker correlation with GDP per capita, alleviating multicollinearity
concerns.

4.1.5. Other country-level control variables

Our empirical framework also requires controls at the country level to alleviate omitted variables bias. Social and commercial
entrepreneurship are known to vary with a country's level of economic development; we control for this using Per capita GDP at
purchasing power parity (World Development Indicators), in logarithm to allow for the expected nonlinear relationship. We fur-
ther control for the level of government activism, which past research found to impact both social and commercial entrepreneur-
ship (Fogel et al., 2006; Estrin et al., 2013a, b), using a measure of the size of the government based on Wall Street Journal/
Heritage Foundation data.10 We also control for the national level of unemployment to capture the state of the labour market
(and therefore the opportunity cost of entrepreneurship) and likewise for the share of the working age population in the overall
population, to help us to separate the individual age effects from the age structure in the environment. We lag all these variables
by one year to reduce potential endogeneity. Importantly, this set of country-level control variables also allows us to account for
possible confounding effects of the extent of social needs, i.e. ‘opportunities’ for social entrepreneurship.
We also control for the level of existing entrepreneurial activity, both social and commercial, in each country, by including the
national rate of social and commercial young and established enterprise owners respectively. We also include the country level
means of higher education attainment to ensure that the individual effects of general human capital are isolated more finely. Sim-
ilarly we add country level prevalence rate of business angels, in addition to the individual effect.
While our primary interest is in formal institutions, we also control for a potential impact of informal institutions. For that pur-
pose we follow both Reynolds (2010) and Hechavarría (2015) and include two cultural scales based on the World Value Survey:
one spans from “traditional” to “secular-rational” culture, the other from “survival” culture to “self- expression” culture closely re-
lated to postmaterialism values (Stephan et al., 2015).
Definitions of all variables discussed above are reported in Table 1, the correlation matrix of the individual level data is pre-
sented in Table 3, and correlations at the national level are presented in Table 4. There is relatively high correlation between
GDP per capita and both the prevalence of tertiary education and survival values.11 We include these variables in our specifica-
tions however, as we do not use them to test our hypotheses; rather we consider them as important controls.

4.2. Estimation

We follow Autio and Acs (2010) and Estrin and Mickiewicz (2011) among others in using multilevel modelling within the
context of a cross-country, cross-individual entrepreneurship dataset. Multilevel modelling takes account of the fact that our
dataset has a hierarchical structure in which individuals represent level one and countries represent level two. This allows us
to control for unobserved country-level heterogeneity related both to macro factors that are not directly included in the model
and to sub-sample specific measurement errors. At the same time, we address the problem of unit dependencies, where, for ex-
ample, two respondents from the same country in the same year are more likely to exhibit similar patterns in their behaviour. In
this case, the independence assumption does not hold, and a multi-level, random effects model should be employed (Peterson
et al., 2012; Rabe-Hesketh et al., 2005). We test the significance of the country effects. For the null model, where we only include
random country effects, the intra-class correlation (ICC) is 0.137 and highly significant (p b 0.001). This supports the use of multi-
level modelling. Based on Model 1 (with all variables of interest including our predictors) in Table 5, the intra-class correlation

9
The Quality of Government database also contains the rule of law measure from Bertelsmann Stiftung. Unfortunately this cannot be used in our empirical work be-
cause it only covers a subset of countries: developing economies.
10
We follow Reynolds (2010) and transform this back to the simple ratio of government expense to GDP.
11
When we obtained variance inflation factors (VIF) after running a regression model corresponding to Model 1 below, the only VIF above the conventional 10 was for
logarithm of GDP p.c. (apart from age and age square which are correlated by construction).
460 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

Table 3
Correlations among individual-level variables; (N = 68,885 individuals).

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

1 Social relative to
commercial start-up
2 Incomplete secondary −0.03
education
3 Secondary education −0.00 −0.36
4 Tertiary education 0.07 −0.35 −0.55
5 Entrepreneurial −0.04 0.02 0.01 −0.02
experience
6 Female 0.06 −0.01 0.01 −0.02 −0.01
7 Age 0.03 0.04 −0.10 0.01 −0.01 0.00
8 Age squared 0.04 0.04 −0.09 −0.01 −0.01 0.00 0.99
9 Existing social 0.21 −0.04 0.00 0.05 0.01 −0.01 0.02 0.01
business
10 Existing commercial 0.10 0.01 −0.03 −0.01 −0.04 −0.11 0.03 0.02 −0.01
business
11 Household income — 0.01 0.05 0.04 −0.11 0.06 0.03 −0.01 0.00 −0.01 −0.04
middle third
12 Household income — −0.00 −0.12 −0.04 0.23 −0.03 −0.09 −0.05 −0.06 0.03 0.07 −0.64
top third
13 Business angel −0.02 −0.02 −0.01 0.03 0.02 −0.06 −0.03 −0.03 0.04 0.10 −0.04 0.05
14 Knows entrepreneurs −0.02 −0.05 0.01 0.05 0.01 −0.12 −0.16 −0.16 0.07 0.12 −0.03 0.11 0.16
15 Fear of failure 0.05 0.02 −0.02 −0.01 −0.02 0.07 0.01 0.00 −0.02 −0.08 0.02 −0.04 −0.03 −0.03
16 In employment −0.07 −0.11 −0.02 0.18 0.07 −0.20 0.00 −0.05 0.06 0.22 −0.04 0.19 0.05 0.10 −0.03

Note. See Table 1 for variable definitions and reference categories.

(ICC) decreases to 0.033. Yet it is significantly different from zero (p b 0.001): our country level variables still leave some overall
country level variance in dependent variable unexplained.
We estimate a multilevel logit model for social and commercial entrepreneurship, taking commercial start-up activity as a
baseline category and present odds ratios (OR) instead of coefficients for ease of interpretation. Thus positive effects (OR N 1)
mean that a variable has a stronger effect on social as compared to commercial start-up. By comparison, negative effects
(OR b 1) mean that a variable has a stronger effect on commercial as opposed to social start-up. The drawback in using ORs is
that as these represent responses to unit change in independent variables, the values for country prevalence rates will be very
high. Table 5 presents the results per estimation model, where the second column “Model 0” present a baseline model with all
control variables but not predictor variables included.

5. Results

5.1. Education and entry into social and commercial entrepreneurship (H1a, H1b)

Model 0 of Table 5 presents the baseline model containing control variables only. Model 1 adds our predictor variables based
on which we evaluate H1a and H1b. The data support H1a; we see a positive effect of tertiary education (OR = 1.35, p b 0.05)
meaning that it has a stronger positive effect on social compared to commercial entry. In support of H1b, we observe a negative
effect of entrepreneurship experience (OR = 0.63, p b 0.01) meaning that it has a stronger (positive) effect on commercial com-
pared to social entry.

Table 4
Correlations among country-level variables (N = 37 countries).

1 2 3 4 5 6 7 8 9 10

1 Executive constraints
2 Prevalence of social start-up 0.20
3 Prevalence of commercial start-up −0.34 0.04
4 Prevalence of tertiary education 0.35 0.26 −0.46
5 Prevalence of business angels −0.33 0.08 0.58 −0.34
6 GDP per capita (natural logarithm) 0.59 0.01 −0.54 0.69 −0.47
7 Government spending 0.44 0.24 −0.46 0.51 −0.51 0.59
8 Working age population 0.10 0.02 −0.19 0.50 −0.15 0.56 0.08
9 Unemployment −0.02 −0.32 −0.15 −0.38 −0.10 −0.24 −0.04 −0.11
10 Survival values 0.49 0.43 −0.32 0.40 −0.37 0.68 0.53 −0.09 −0.22
11 Traditional values 0.38 0.27 −0.16 0.56 −0.28 0.58 0.52 0.47 −0.28 0.18
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 461

Table 5
Multilevel logistic regressions on the likelihood of social against commercial start-up.

Model 0 Model 1 Model 2 Model 3 Model 4

Constant 0.548 1.070 1.024 1.085 0.436

(1.437) (2.790) (2.675) (2.841) (1.117)

Individual-level control variables

Age 0.966* 0.962* 0.962* 0.963* 0.959*
(0.015) (0.015) (0.015) (0.015) (0.016)
Age squared 1.000* 1.000* 1.000* 1.000* 1.001**
(0.0002) (0.0002) (0.0002) (0.0002) (0.0002)
SE engagement 4.667*** 4.656*** 4.658*** 4.699*** 4.803***
(0.444) (0.444) (0.445) (0.449) (0.504)
CE engagement 2.589*** 2.595*** 2.597*** 2.595*** 2.573***
(0.210) (0.211) (0.211) (0.211) (0.228)
Business angel 0.830* 0.839+ 0.838+ 0.832* 0.871
(0.075) (0.076) (0.076) (0.076) (0.086)
Know entrepreneur 0.931 0.923 0.924 0.925 0.913
(0.054) (0.054) (0.054) (0.054) (0.058)
Fear of failure 1.333*** 1.343*** 1.343*** 1.341*** 1.324***
(0.081) (0.082) (0.082) (0.082) (0.090)
Female 1.271*** 1.265*** 1.267*** 1.268*** 1.303***
(0.071) (0.070) (0.071) (0.071) (0.081)
In employment 0.669*** 0.649*** 0.650*** 0.647*** 0.677***
(0.043) (0.042) (0.042) (0.042) (0.049)

Country-level control variables

Rate young & established SE 2422* 1432+ 1479+ 1562+ 830.9+
(9432) (5529) (5718) (6056) (3152)
Rate young & established CE 0.006*** 0.007** 0.008** 0.008** 0.010**
(0.008) (0.011) (0.012) (0.011) (0.015)
Education secondary or higher 3.867+ 3.617+ 3.526+ 3.561+ 2.241
(2.735) (2.543) (2.485) (2.514) (1.606)
Business angel 1411* 841.3* 889.1* 854.5* 971.4*
(4269) (2529) (2677) (2579) (2891)
Logarithm of GDP p.c. 0.710 0.629 0.631 0.622 0.675
(0.210) (0.186) (0.187) (0.185) (0.195)
Government spending 1.007 1.008 1.009 1.008 1.014
(0.010) (0.010) (0.010) (0.010) (0.010)
Working age population 1.039 1.048+ 1.048+ 1.049+ 1.054*
(0.026) (0.027) (0.027) (0.027) (0.027)
Unemployment rate 1.043+ 1.036 1.036 1.035 1.027
(0.024) (0.024) (0.024) (0.024) (0.023)
Survival vs. self-expression 1.067 1.078 1.078 1.081 1.038
(0.141) (0.142) (0.142) (0.143) (0.134)
Traditional vs. rational culture 1.016 0.978 0.976 0.976 0.940
(0.127) (0.126) (0.126) (0.126) (0.119)
Executive constraints 1.065 1.078 1.080 1.083
(0.056) (0.059) (0.057) (0.058)

Individual-level human capital (H1)

Some secondary education 0.979 0.978 0.979 1.160
(0.125) (0.125) (0.125) (0.161)
Secondary education 1.048 1.041 1.052 1.170
(0.124) (0.124) (0.125) (0.150)
Tertiary education 1.360** 1.350* 1.365** 1.444**
(0.162) (0.162) (0.163) (0.186)
Entrepreneurship experience 0.631** 0.631** 0.559*** 0.541***
(0.095) (0.095) (0.089) (0.092)

Cross-level interaction effects (H2)

Tertiary education * 0.970 0.966
executive constraints (0.035) (0.038)
Entrepreneurship experience* 0.781** 0.816*
executive constraints (0.062) (0.070)
Log variance of random effect 0.116*** 0.113*** 0.113*** 0.114*** 0.100***
(0.042) (0.040) (0.041) (0.041) (0.041)
Log likelihood −4138.051 −4120.338 −4119.974 −4115.622 −3381.151
Intra-class correlation1 0.034 0.033 0.033 0.033 0.029
(0.012) (0.011) (0.011) (0.011) (0.012)

Note. Number of individual-level observations: 6901; number of countries: 37. Relative odds ratios (OR), OR N 1 indicates positive effect, OR b 1 indicates negative
effect. Standard errors in parenthesis; CE — commercial enterprise, SE — social enterprise; ***p b 0.001, **p b 0.01, *p b 0.05, +p b 0.10; the likelihood ratio test
comparing Model 0 (without the hypotheses-related predictors) with Model 1 renders χ2(5) = 26.63, which is significant at p b 0.001. 1The intra-class correlation
is a measure of residual, i.e. unexplained country-level variation.
462 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

5.2. The moderating effect of institutional quality (H2a, H2b)

To evaluate the moderating effect of institutional quality on human capital, we first add the interaction between executive
constraints and higher education in Model 2, and next we introduce the interaction between executive constrains and entrepre-
neurial skills in Model 3. Finally, as the most stringent test of our hypotheses, we include both interactions in Model 4 (all in
Table 5).
The results provide support for H2b, but not for H2a. Institutional quality makes a significant difference to how people use
their specific entrepreneurial skills (OR = 0.82, p b 0.05, Model 4), but affects the use of their general human capital less
(OR = 0.97 n.s., Model 4). This can be seen as fundamentally consistent with Baumol's (1990) perspective: it is the use of entre-
preneurial talent that is predominantly affected by institutions. We plot the significant moderating effects in Fig. 2 displaying the
association of entrepreneurial skills with the likelihood to choose social over commercial entry. We see that the association of en-
trepreneurial skills with commercial entry is stronger in the presence of strong (vs. weak) rule of law (executive constraints).

5.3. Further results

Some additional results are noteworthy. First, while the rule of law (executive constraints) does not have any impact on the
choice between social and commercial entrepreneurship, this should not be interpreted as lack of evidence for a positive net effect
of the rule of law on entry. Additional multilevel multinomial estimations run on the whole sample, showed a positive effect of
the rule of law on entry into both types of entrepreneurship when contrasted with no entrepreneurial activity.
Second, current engagement in a social enterprise (SE engagement) has a consistent significant positive effect on the choice to
start a new project as a social versus commercial enterprise, with the corresponding relative odd ratios remaining remarkably sta-
ble and high (OR = 4.66, Model 1 Table 5). This suggests that individuals, who are first attracted to social entrepreneurship, be-
come serial social entrepreneurs.
Third, current engagement in a commercial enterprise (CE engagement) also has a positive effect on choosing social as opposed
to commercial entry (OR = 2.59, Model 1 Table 5). This is very interesting: Estrin et al. (2013a, b) identified a route leading from
social to commercial projects, but here we also see the evidence for a positive spillover in the opposite direction — entrepreneurs
currently running a commercial business are also more likely to start a new project as a social compared to a commercial enter-
prise. However, high rates of commercial entrepreneurship engagement in a country privilege commercial over social entry
(OR = 0.007, Model 1 Table 5).
Fourth, we also observe a highly significant positive effect of gender on likelihood of choice of social versus commercial entry
(OR = 1.65, Model 1 Table 5): women are relatively more likely to become social than commercial entrepreneurs supporting
Hechavarria et al. (2012) and Terjesen et al. (2012). This suggests that socially oriented projects could be an important channel
for women to enter into entrepreneurship.
Fifth, being in employment has a negative effect on the choice of social over commercial entry (OR = 0.649, Model 1 Table 5).
This suggests that those not in employment are more likely to choose social rather than commercial entrepreneurship, and thus
that social entrepreneurship may also be an important entry channel for those currently detached from the labour market.

5.4. Robustness tests

We conducted a range of robustness checks which support our findings. The full results are available from the authors upon
request.
First, we replicated the results in Table 5 using the indicator of entrepreneurial skills based on self-assessment instead of mea-
sure of experience with earlier projects. In this variant, the interaction effect with rule of law for H2b was only significant at
p b .10.

Fig. 2. Interaction effect specific human capital (entrepreneurial experience) and constraints on the executive (institutional quality) (H2b, Table 5, Model 3).
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 463

Second, we repeated all the estimations restricting the sample only to those in employment, which could be seen as a design
that comes closer to the logic of the occupational choice (Le, 1999). The results for all hypotheses and significance levels are ex-
actly as in Table 5.
Third, we used alternative measures of the rule of law. Applying the Freedom House measure instead of Polity IV makes no
difference to significance levels for our hypotheses. Using the World Bank measure weakens the results for the rule of law.
This may result both from the fact that the conceptual basis of their measure is weaker being more a catch-all factor rather
than anchored in theory (Langbein and Knack, 2010), and it suffers from very high multicollinearity with the level of GDP per
capita.
Fourth, we replicated our specifications running multilevel multinomial logit models with two outcomes: entry into commer-
cial and social entrepreneurship, as contrasted with no entry. Here, to test our hypotheses we relied on post-estimation tests for
differences in coefficients between social and commercial entry (Wald tests). The significance and pattern of results for the hy-
potheses were the same as in Table 5.
Finally, we also ran models controlling in addition for household income. This did not change our results and the income ef-
fects were always insignificant. Therefore we did not use them in our main specifications.

6. Discussion

This multi-level study compared two forms of human capital as drivers of social and commercial entry while simultaneously
considering contextual effects of the rule of law. We found that general human capital is relatively more important for social en-
trepreneurship while specific human capital is relatively more important for commercial entrepreneurship. Furthermore the rule
of law moderated the effects of specific human capital on entrepreneurship.
This study advances research on how and why human capital influences entrepreneurial entry by outlining important contin-
gencies of this relationship related to heterogeneity among entrepreneurs and across national contexts. Past research on human
capital largely focusses on the consequences of human capital for firm performance (e.g., Rauch and Rijsdijk, 2013; Unger et al.,
2011; Van der Sluis et al., 2005, 2008; Van Praag et al., 2013), while the effects of human capital on entrepreneurial entry are less
well understood. Indeed, previous research generating mixed findings for entry (e.g., Block et al., 2013; Le, 1999; Parker, 2009).
Our study contributes to the latter in two ways.
First, through the novel application of a human capital lens to the analysis of social and commercial entrepreneurship, our
study advances and broadens the conceptualisation of human capital effects. In particular we include the ability enhancing effects
of human capital, as stressed in economic theory, and the hitherto overlooked socialising effects of education in terms of (pro-so-
cial) preferences. For human capital theory more broadly, our study highlights the importance of considering heterogeneity in the
effects of general human capital on both individuals' skills and preferences, which helps to explain why different types of human
capital may lead to different entry modes (social versus commercial).
Our analysis of human capital and social entrepreneurship also illustrates the need for a broader view of returns to occupation-
al choices. We find that general human capital is of particular relevance in situations where the external benefits of occupational
choices are greater and add to the purely private returns, as occurs with social entrepreneurship. The implications extend beyond
the analysis of social entrepreneurship. The literature documents a wide variety of possible motives for entrepreneurship that go
beyond the aim of accumulating monetary returns — for example, for open-source, high-tech entrepreneurs, ethnic and immi-
grant entrepreneurship, for family businesses, and for those pursing entrepreneurship to realise opportunities for greater personal
independence. Greater consideration of this heterogeneity in returns to entrepreneurship can help future research to establish an
even deeper understanding of how human capital relates to entrepreneurship and entrepreneurial entry, and likely yields more
consistent findings.
Second, our research also offers a more contextualised understanding of human capital and entrepreneurship by integrating
predictions from institutional theory and human capital theory. We outline how national institutions act as an important contin-
gency influencing the opportunity costs and potential returns from human capital when it comes to occupational choices for en-
trepreneurship. This offers an additional explanation for the mixed findings in past research on human capital and entry
mentioned earlier and helps to unpack the drivers of the repeatedly observed national variation in the effect of human capital
on entrepreneurship (e.g., Arenius and Minniti, 2005; Parker, 2009; van der Sluis et al., 2005, 2008). Specifically, while effective
institutions are important for both commercial and social entrepreneurship (Estrin et al., 2013a, b), our results show that the
moderating impact of the rule of law on the returns to specific human capital is more central for the former than the latter
type of entrepreneurship. When the rule of law is strong, it ensures that commercial entrepreneurs have a better chance of keep-
ing the private returns from their venture. This is consistent with Baumol's (1990) perspective that it is the use of entrepreneurial
human capital which is particularly sensitive to the quality of institutions in the environment. In turn, social entrepreneurship is
focussed on the generation of positive external effects rather than private gains (Santos, 2012); these are more widely dispersed
and thus more difficult to expropriate, and therefore the moderating effect of the rule of law is less important for social
entrepreneurs.
Moreover, the underlying resources and opportunities may differ in the different contexts. For example, entrepreneurial
human capital may be rarer in countries where institutions are weaker (Aidis et al., 2008). Institutional weaknesses may also cre-
ate more opportunities for social entrepreneurship (Mair and Marti, 2009), offsetting to some extent the negative incentive ef-
fects. Our results therefore help to explain why social entrepreneurship may play an important positive role in countries
characterised by dysfunctional institutions and therefore we complement qualitative research in this area (e.g. Mair et al.,
464 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

2012). In addition, the absence of the hypothesised significant moderating relationship between general human capital and the
rule of law merits further careful research, particularly in environments where institutional voids exist.
Finally, for entrepreneurship theory, our findings contribute to a better understanding of the similarities and differences be-
tween social and commercial entrepreneurship, and they highlight the important role of human capital and ability next to the
commonly considered differences in motivation between the two types of entrepreneurship. We show that social and commercial
entrepreneurship attract different types of individuals, consistent with the notion that these two forms of entrepreneurship are
differentiated by both the type of human capital they require, and by the objectives and motivation of the entrepreneurs. For so-
cial entrepreneurship, our research highlights the importance of general human capital. This complements the field's focus on mo-
tivation as the key driver of social entrepreneurship (e.g. Dacin et al., 2010; Miller et al., 2012). This is not trivial. Prosocial
motivation can be realised through channels other than social entrepreneurship, and it is only when this motivation is combined
with the opportunity-recognition ability associated with higher education that individuals engage in the relatively more complex
endeavour of setting up a social enterprise. In so doing, we respond to the call by Parker (2008) to provide a simple but theoret-
ically grounded typology to understand who becomes a social entrepreneur. We also add more generalisable insights, derived
from studying population-representative samples across a variety of countries, to social entrepreneurship research, a field still
dominated by conceptual and case-based analysis (Gras et al., 2014; Dacin et al., 2010; Nicholls, 2010, for reviews).
Our additional findings provide renewed support for the notion that social entrepreneurship attracts different individuals than
commercial entrepreneurship. Women, the highly educated and those who are not currently employed are more likely to become
social than commercial entrepreneurs. Together with past findings that social entrepreneurship is a way into commercial entre-
preneurship (Estrin et al., 2013a, b) 12 the former opens up an important channel for valuable human talent to become entrepre-
neurs and (re-)enter the labour market.

6.1. Strengths and limitations

Apart from being able to draw on population representative samples across a wide range of countries, a further strength of our
study is the use of multi-level modelling which allows us to test individual-level relationships at the same time as country-effects.
This addresses aggregation and disaggregation biases (Peterson et al., 2012), namely that relationships observed at one level of
analysis (e.g. country-level) may not generalise (and may be different from that equivalent relationship) at a different level of
analysis (e.g. individual-level).
A limitation of the GEM dataset is the cross-sectional nature of the data, which gives rise to concerns about reverse causality.
Our analyses alleviate such concerns somewhat as we investigated the effects of human capital that individuals obtained in the
past (highest degree, past experience in commercial entrepreneurship) and used lagged data for country-level institutions and
GDP. There is a need for future research to address all these questions and findings using longitudinal data (see Renko, 2013,
for a good example).
The use of secondary datasets such as GEM also restricted the measures of human capital. We focussed on general education,
rather than the specific type of subject studied, and on obtaining a degree rather than years of schooling. This choice of measures
was probably adequate for our purposes; reputation effects of higher education (which determine opportunity cost of entry) are
arguably contingent on obtaining a degree rather than years of schooling. The use of highest degree obtained also created greater
comparability across countries. Nevertheless, future research could usefully investigate more fine-grained measures of education
and could also help to unpack the mechanisms, including skills, motivation and confidence, that link higher education to entre-
preneurial entry. In line with past research, we emphasised the socialising function of education, but we could not control for al-
ternative explanations. For instance, education effects in part reflect differences in individual persistence, family background and
ability. However, a recent analysis by Block et al. (2013) suggests that controlling for endogeneity strengthens education effects
on entrepreneurial entry, which in turn implies that those are possibly underestimated rather than overestimated by us.
We also acknowledge that GEM used a specific question for social entrepreneurship as an initiative, activity or organisation
with a particular community, social or environmental objective. The interpretation of such objectives may vary across cultures.
Although to help interpretation GEM gave specific examples for community and social objectives, but not for environmental ob-
jectives. Thus the latter may be underrepresented in the sample of social entrepreneurs. As of yet, GEM is however the only large
scale database on social entrepreneurship.
Apart from the directions we highlighted already, future research could explore the moderating effects of “lower level” (in the
sense of Williamson, 2000) regulatory national institutions and policies, which were captured by a control variable (government
spending) in our analyses. While our focus was on the rule of law, informal institutions including culture and social capital may
also influence the higher education — entrepreneurship link, as suggested by Stephan et al. (2015).

6.2. Policy and practical implications

Policy makers, for example in the European Commission, have adopted as an objective the creation of a favourable environ-
ment for the development of social businesses, because such firms are argued to “contribute to social cohesion, employment
and the reduction of inequalities” (European Commission, 2013). Policy interest in social entrepreneurship stems from doubts

12
We note that our findings on gender are consistent with earlier findings by Hechavarria et al. (2012) and Terjesen et al. (2012) as well as Estrin et al. (2013a, b)
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 465

about how much can be achieved towards social goals from for-profit motivation, and from scepticism about the effectiveness of
bureaucratic and centralised political interventions. In contrast to interventions by the state, social entrepreneurship generates
highly decentralised modes of action, focussed on resolving social problems at a local or relatively small-scale level. Social inter-
ventions are thereby differentiated, drawing from local knowledge, competing and subject to continuous innovative pressure. Our
theoretical framework also emphasises the joint supply characteristics of the social enterprise business model, and the commercial
logic combined with focus on realisation of positive external effects, which can make social entrepreneurship a superior mode of
action compared to the civil society charitable action, at least for some categories of social problems.
Our findings on the relationship between different forms of human capital and entrepreneurial entry provides the basis for
improved targeting of business policies, as well as of policies related to education including higher education. Our findings suggest
that social enterprises are more likely to be started by those with higher education. For educators, especially in higher education,
this suggests that a greater awareness and support of social entrepreneurial activities among students may be warranted
(Lawrence et al., 2012).
We find beneficial effects from specific entrepreneurial human capital for social entrepreneurship. This reinforces the case for
programmes developing entrepreneurial skills among all types of students while emphasising that entrepreneurial skills are useful
for the realisation of multiple objectives including social objectives. This contrasts with our experience in which entrepreneurship
training and courses primarily focus on commercial entrepreneurship and on business and engineering students. Although further
research is necessary, our findings seem to imply that those studying social sciences, medicine or humanities are well positioned
to take up social entrepreneurial projects. In turn, those who build entrepreneurial teams for social objectives should be well
aware of the need for different human capital, as applicable in both commercial and social entrepreneurship.
Moreover, we found that social entrepreneurship attracts new categories of people into entrepreneurial activity; for example
women or the more highly educated (Estrin et al., 2013a, b; Hechavarria et al., 2012 and Terjesen et al., 2012). Policymakers who
seek to reduce gender discrimination in occupational choice may therefore be interested in promoting social entrepreneurship.
This conclusion is reinforced by the phenomenon of “crowding in” (Estrin et al., 2013a, b), i.e. social entrepreneurship helps in
developing entrepreneurial skills with broader applicability. This is an important finding for business support policies, highlighting
that supporting social entrepreneurship may generate positive externalities such as stimulating commercial entrepreneurship.
Last but not least, our results can be interpreted as offering some predictions about the role of social entrepreneurship in dif-
ferent institutional environments. We have found that social entrepreneurship projects have a comparative advantage, relative to
commercial entry, in weaker institutional environments. This has important implications for policy design supporting human de-
velopment across nations. In particular, while commercial projects may be risky where the rule of law is weak, social entrepre-
neurship projects are less so. Thus, there may be deeper reasons why the organisational form adopted for example by the
Grameen Bank (Yunus, 2003) makes a good fit with the local environment, and why social entrepreneurship may be more effi-
cient in developing than developed countries.

6.3. Conclusion

Our study reinforces the importance of human capital – abilities and skills – for entry into entrepreneurship. It extends human
capital theory to social entrepreneurship and across national contexts; in doing so we find that national institutions act as impor-
tant contingencies. Our findings also enrich existing research on social entrepreneurship and on the differences between social
and commercial entrepreneurship. While past research has mainly focussed on differences in motivation, this paper highlights
the importance of taking abilities and skills into account to understand who is likely to become a social entrepreneur.

References

Abramson, P.R., Inglehart, R., 1994. Education, security, and postmaterialism: A comment on Duch and Taylor’s ‘Postmaterialism and the economic condition’. Amer-
ican Journal of Political Science 38 (3), 797–814.
Acemoglu, D., Johnson, S., 2005. Unbundling institutions. J. Polit. Econ. 113, 949–995.
Acemoglu, D., Pischke, S.J., 1998. Why do firms train? Q. J. Econ. 113 (1), 79–119.
Aidis, R., Estrin, S., Mickiewicz, T., 2008. Institutions and entrepreneurship development in Russia: a comparative perspective. J. Bus. Ventur. 23 (6), 656–672.
Arenius, P., Minniti, M., 2005. Perceptual variables and nascent entrepreneurship. Small Bus. Econ. 24 (3), 233–247.
Austin, J., Stevenson, H., Wei-Skillern, J., 2006. Social and commercial entrepreneurship: same, different, or both? Enterp. Theory Pract. 30 (1), 1–22.
Autio, E., Acs, Z., 2010. Intellectual property protection and the formation of entrepreneurial growth aspirations. Strateg. Entrep. J. 4 (3), 234–251.
Bacq, S., Hartog, C., Hoogendoorn, B., 2014. Beyond the moral portrayal of social entrepreneurs: an empirical approach to who they are and what drives them. J. Bus.
Ethics 1–16.
Baker, T., Gedajlovic, E., Lubatkin, M., 2005. A framework for comparing entrepreneurship processes across nations. J. Int. Bus. Stud. 36 (5), 492–504.
Battilana, J., Lee, M., 2014. Advancing research on hybrid organizing: insights from the study of social enterprises. Acad. Manag. Ann. 8 (1), 397–441.
Baumol, W.J., 1990. Entrepreneurship: productive, unproductive, and destructive. J. Polit. Econ. 98 (5), 893–921.
Becker, G., 1964. Human capital: A theoretical and Empirical Analysis with Special Reference to Education. University of Chicago Press, Chicago, IL.
Block, J.H., Hoogerheide, L., Thurik, R., 2013. Education and entrepreneurial choice: an instrumental variables analysis. International Small Business Journal 31 (1),
23–33.
Bosma, N.S., Levie, J., 2010. Global Entrepreneurship Monitor 2009 Executive Report. Global Entrepreneurship Research Association (retrieved from) www.
gemconsortium.org.
Dacin, P.A., Dacin, M.A., Matear, M., 2010. Social entrepreneurship: why we don't need a new theory and how we move forward from here. Acad. Manag. Perspect. 24,
37–57.
Dahlberg, S., Holmberg, S., Rothstein, B., Khomenko, A., Svensson, R., 2016. The Quality of Government Basic Dataset, version Jan16. University of Gothenburg: The
Quality of Government Institutehttp://www.qog.pol.gu.se http://dx.doi.org/10.18157/QoGBasJan16.
Davidsson, P., Honig, B., 2003. The role of social and human capital among nascent entrepreneurs. J. Bus. Ventur. 18 (3), 301–331.
466 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467

De Tocqueville, A., 2003 [1835]. Democracy in America. Penguin, London.

Di Domenico, M., Haugh, H., Tracey, P., 2010. Social bricolage: theorizing social value creation in social enterprises. Enterp. Theory Pract. 34 (4), 681–703.
Epstein, R., 2011. Design for Liberty. Private Property, Public Administration and the Rule of Law. Harvard University Press, Cambridge, MA.
Estrin, S., Mickiewicz, T., 2011. Institutions and female entrepreneurship. Small Bus. Econ. 37, 397–415.
Estrin, S., Mickiewicz, T., Stephan, U., 2013a. Entrepreneurship, social capital, and institutions: social and commercial entrepreneurship across nations. Enterp. Theory
Pract. 37, 479–504.
Estrin, S., Korosteleva, J., Mickiewicz, T., 2013b. Which institutions encourage entrepreneurial growth aspirations? J. Bus. Ventur. 28 (4), 564–580.
European Commission, 2013. The social business initiative of the European commission. http://ec.europa.eu/internal_market/publications/docs/sbi-brochure/sbi-
brochure-print_en.pdf (accessed 23 July 2014).
Fogel, K., Hawk, A., Morck, R., Yeung, B., 2006. Institutional obstacles to entrepreneurship. In: Casson, M., Yeung, B., Basu, A., Wadeson, N. (Eds.), Oxford Handbook of
Entrepreneurship. Oxford University Press, Oxford, pp. 540–579.
Gartner, W.B., 1989. “Who is an entrepreneur?” is the wrong question. Enterp. Theory Pract. 13 (4), 47–68.
Gras, D., Moss, T.W., Lumpkin, G.T., 2014. The use of secondary data in social entrepreneurship research: assessing the field and identifying future opportunities. In:
Short, J. (Ed.), Social Entrepreneurship and Research Methods. Bingley, Emerald, pp. 49–75.
Hébert, R.F., Link, A.N., 1988. The Entrepreneurs: Mainstream Views and Radical Critiques. Praeger, New York.
Hechavarría, D.M., 2015. The impact of culture on national prevalence rates of social and commercial entrepreneurship. Int. Entrep. Manag. J. http://dx.doi.org/10.1007/
s11365-015-0376-1.
Hechavarria, D.M., Ingram, A., Justo, R., Terjesen, S., 2012. Are women more likely to pursue social and environmental entrepreneurship? In: Hughes, K.D., Jennings, J.E.
(Eds.), Global Women's Entrepreneurship Research: Diverse Settings, Questions and Approaches. Edward Elgar, Cheltenham, pp. 135–151
Ho, Y.P., Wong, P.K., 2007. Financing, regulatory costs and entrepreneurial propensity. Small Bus. Econ. 28 (2–3), 187–204.
Hoogendoorn, B., Hartog, C., 2011. Prevalence and determinants of social entrepreneurship at the macro-level. Scales Research Reports H201022. EIM Business and
Policy Research.
Khilstrom, R.E., Laffont, J.F., 1979. A general equilibrium theory of firm formation based on risk aversion. J. Polit. Econ. 87, 719–749.
Kohn, M.L., Schooler, C., 1983. Work and Personality. Ablex, Norwood, NJ.
Korosteleva, J., Mickiewicz, T., 2011. Start-up finance in the age of globalisation. Emerging Markets Finance and Trade 47 (3), 23–49.
Koudstaal, M., Sloof, R., van Praag, M., 2015. Risk, uncertainty and entrepreneurship: Evidence from a lab-in-the-field experiment. Management Science http://dx.doi.
org/10.1287/mnsc.2015.2249.
Langbein, L., Knack, S., 2010. The worldwide governance indicators: six, one, or none? J. Dev. Stud. 46, 350–370.
Lawrence, T., Phillips, N., Tracey, P., 2012. From the guest editors: educating social entrepreneurs and social innovators. Acad. Manag. Learn. Educ. 11 (3), 319–323.
Lazear, E., 2005. Entrepreneurship. J. Labor Econ. 23, 649–680.
Le, A.T., 1999. Empirical studies of self-employment. J. Econ. Surv. 13 (4), 381–416.
Lucas, R., 1978. On the size distribution of business firms. Bell J. Econ. 9 (2), 508–523.
Lukes, M., Stephan, U., 2012. Entrepreneurs and leaders of non-profit organizations: similar people with different motivations. Ceskoslovenska Psychologie 56 (1),
41–55.
Mair, J., Marti, I., 2006. Social entrepreneurship research: a source of explanation, prediction, and delight. J. World Bus. 41, 36–44.
Mair, J., Marti, I., 2009. Entrepreneurship in and around institutional voids: a case study from Bangladesh. J. Bus. Ventur. 24 (5), 419–435.
Mair, J., Martí, I., Ventresca, M.J., 2012. Building inclusive markets in rural Bangladesh: how intermediaries work institutional voids. Acad. Manag. J. 55 (4), 819–850.
Mair, J., Mayer, J., Lutz, E., 2015. Navigating institutional plurality: organizational governance in hybrid organizations. Organ. Stud. 36 (6), 713–739.
Meyskens, M., Robb-Post, C., Stamp, J.A., Carsrud, A.L., Reynolds, P.D., 2010. Social ventures from a resource-based perspective: an exploratory study assessing global
Ashoka fellows. Enterp. Theory Pract. 34 (4), 661–680.
Mickiewicz, T., Sauka, A., Stephan, U., 2016. On the compatibility of benevolence and self-interest: philanthropy and entrepreneurial orientation. International Small
Business Journal 34 (3), 303–328.
Miller, T.L., Grimes, M.G., McMullen, J.S., Vogus, T.J., 2012. Venturing for others with heart and head: how compassion encourages social entrepreneurship. Acad.
Manag. Rev. 37 (4), 616–640.
Minniti, M., Arenius, P., Langowitz, N., 2005a. Global Entrepreneurship Monitor: 2004 Report on women and entrepreneurship. Centre for women's leadership at
Babson College & London Business School.
Minniti, M., Bygrave, W., Autio, E., 2005b. Global Entrepreneurship Monitor: 2005 Executive Report. Babson College, Wellesley, MA.
Munck, G.L., Verkuilen, J., 2002. Conceptualizing and measuring democracy. Evaluating alternative indices. Comparative Political Studies 35 (1), 5–34.
Nga, J.K.H., Shamuganathan, G., 2010. The influence of personality traits and demographic factors on social entrepreneurship start up intentions. J. Bus. Ethics 95 (2),
259–282.
Nicholls, A., 2010. The legitimacy of social entrepreneurship: reflexive isomorphism in a pre-paradigmatic field. Enterp. Theory Pract. 34, 611–633.
North, D.C., 1990. Institutions, Institutional Change and Economic Performance. Cambridge University Press, Cambridge.
Noseleit, F., 2010. The entrepreneurial culture: guiding principles of the self-employed. In: Freytag, A., Thurik, R. (Eds.), Entrepreneurship and Culture. Springer, Berlin,
pp. 41–54.
Pache, A.-C., Santos, F., 2013. Inside the hybrid organization: selective coupling as a response to competing institutional logics. Acad. Manag. J. 56 (4),
972–1001.
Parker, S.C., 2008. Social entrepreneurs: a neoclassical theory. In: Shockley, G.E., Frank, P.M., Stough, R.S. (Eds.), Non-Market Entrepreneurship. Edward Elgar, Chelten-
ham, pp. 206–216.
Parker, S.C., 2009. The Economics of Entrepreneurship. Cambridge University Press, Cambridge.
Parker, S.C., 2011. Intrapreneurship or entrepreneurship? J. Bus. Ventur. 26 (1), 19–34.
Parker, S.C., Belghitar, Y., 2006. What happens to nascent entrepreneurs? An econometric analysis of the PSED. Small Bus. Econ. 27, 81–101.
Peterson, M.F., Arregle, J.L., Martin, X., 2012. Multilevel models in international business research. J. Int. Bus. Stud. 43 (5), 451–457.
Rabe-Hesketh, S., Skrondal, A., Pickles, A., 2005. Maximum likelihood estimation of limited and discrete dependent variable models with nested random. J. Econ. 128
(2), 301–323.
Rauch, A., Rijsdijk, S.A., 2013. The effects of general and specific human capital on long-term growth and failure of newly founded businesses. Enterp. Theory Pract. 37
(4), 923–941.
Renko, M., 2013. Early challenges of nascent social entrepreneurs. Enterp. Theory Pract. 37 (5), 1045–1069.
Reynolds, P.D., 2010. New firm creation: a global assessment of national, contextual, and individual factors. Foundations and Trends in Entrepreneurship 6 (5–6),
315–496.
Reynolds, P.D., Hay, M., Camp, S., 1999. Global Entrepreneurship Monitor: 1999 Executive Report. Babson College, London Business School and Kauffman Foundation.
Reynolds, P.D., Bosma, N., Autio, E., Hunt, S., De Bono, N., Servais, A., Lopez-Garcia, P., Chin, N., 2005. Global entrepreneurship monitor: data collection design and im-
plementation 1998–2003. Small Bus. Econ. 24, 205–231.
Santos, F.M., 2012. A positive theory of social entrepreneurship. J. Bus. Ethics 111 (3), 335–351.
Schofer, E., Fourcade-Gorinchas, M., 2001. The structural contexts of civic engagement: voluntary association membership in comparative perspective. Am. Sociol. Rev.
66, 806–828.
Schumpeter, J.A., 1934. The Theory of Economic Development: An Inquiry Into Profits, Capital, Credit, Interest, And The Business Cycle. Transaction publishers, New
Brunswick, NJ.
Schwartz, S.H., 2006. Les valeurs de base de la personne: Théorie, mesures et applications [Basic Human Values: Theory, Measurement, and Applications]. Revue
Française de Sociologie 47, 929–968.
 S. Estrin et al. / Journal of Business Venturing 31 (2016) 449–467 467

Schwartz, S.H., 2010. Basic values: how they motivate and inhibit prosocial behavior. In: Mikulincer, M., Shaver, P. (Eds.), Prosocial Motives, Emotions, And Behavior:
The Better Angels of Our Nature. American Psychological Association Press, Washington, pp. 221–242.
Schwartz, S.H., 2012. An overview of the Schwartz theory of basic values. Online Readings in Psychology and Culture 2, 1–20.
Schwartz, S.H., Sagiv, L., Boehnke, K., 2000. Worries and values. J. Pers. 68 (2), 309–346.
Stephan, U., Uhlaner, L.M., Stride, C., 2015. Institutions and social entrepreneurship: the role of institutional voids, institutional support, and institutional configura-
tions. J. Int. Bus. Stud. 46, 308–331.
Stephan, U., Patterson, M., Kelly, C., Mair, J., 2016. Organizations driving positive social change: a review and an integrative framework of change processes. J. Manag.
http://dx.doi.org/10.1177/0149206316633268.
Terjesen, S., Lepoutre, J., Justo, R., Bosma, N., 2012. 2009 Report on Social Entrepreneurship. (Retrieved from) www.gemconsortium.org/docs/2519/gem-2009-report-
on-social-entrepreneurship.
Unger, J.M., Rauch, A., Frese, M., Rosenbusch, N., 2011. Human capital and entrepreneurial success: a meta-analytical review. J. Bus. Ventur. 26 (3), 341–358.
Van der Sluis, J., Van Praag, M., Vijverberg, W., 2005. Entrepreneurship selection and performance: a meta-analysis of the impact of education in developing economies.
The World Bank Economic Review 19 (2), 225–261.
Van der Sluis, J., Van Praag, M., Vijverberg, W., 2008. Education and entrepreneurship selection and performance: a review of the empirical literature. J. Econ. Surv. 22
(5), 795–841.
Van Praag, M., 2005. Successful Entrepreneurship: Confronting Economic Theory With Empirical Practice. Edward Elgar, Cheltenham.
Van Praag, M., Versloot, P.H., 2008. What is the value of entrepreneurship? A review of recent research. Small Bus. Econ. 29 (4), 351–382.
Van Praag, M., van Witteloostuijn, A., van der Sluis, J., 2013. The higher returns to formal education for entrepreneurs versus employees. Small Bus. Econ. 40 (2),
375–396.
Van Ryzin, G.G., Grossman, S., DiPadova-Stocks, L., Bergrud, E., 2009. Portrait of the social entrepreneur: statistical evidence from a US panel. Voluntas 20, 129–140.
Wagner, J., Sternberg, R., 2004. Start-up activities, individual characteristics, and the regional milieu: lessons for entrepreneurship support policies from German micro
data. Ann. Reg. Sci. 38, 219–240.
Welter, F., 2011. Contextualizing entrepreneurship—conceptual challenges and ways forward. Enterp. Theory Pract. 35 (1), 165–184.
Williamson, O., 2000. New institutional economics. J. Econ. Lit. 38, 595–613.
Yunus, M., 2003. Banker to the Poor. Aurum Press, London.
Zahra, S.A., Wright, M., 2011. Entrepreneurship's next act. Acad. Manag. Perspect. 25 (4), 67–83.
Zahra, S.A., Gedajlovic, E., Neubaum, D.O., Shulman, J.M., 2009. A typology of social entrepreneurs: motives, search processes and ethical challenges. J. Bus. Ventur. 24
(5), 519–532.