‘Do you want to test whether a people is given to industry and commerce? Do not sound its ports, or examine the wood from its forest or the produce of its soil. The spirit of trade will get all those things and, without it, they are useless’.
Alexis De Tocqueville (Reference De Tocqueville1958 [1835]: 116)1. Introduction
De Tocqueville (Reference De Tocqueville1958 [1835]: 116) identified ‘courage to seek prosperity’ and ‘a sense and habit to find it’, combined with ‘the assurance of reaping the benefit’, as the key elements of the ‘spirit of trade’.Footnote 1 Thus, he argued that it is not so much endowment in the material factors of production that results in entrepreneurship, but rather cultural values. This approach is consistent with the recent argument of McCloskey (Reference McCloskey2006, Reference McCloskey2010, Reference McCloskey2016), who posits that innovation and entrepreneurship has been driven by cultural values.Footnote 2 Yet, what exactly are these entrepreneurial values?
To answer this question we rely on the key insight proposed by John and Storr (Reference John and Storr2018): different aspects of entrepreneurship are supported by different cultural elements. In particular, entrepreneurial process can be conceptualised as including opportunity identification and opportunity exploitation as two indispensable, core elements (Ibid.). We will argue that these two elements of entrepreneurship are correspondingly supported by cultural elements of creativity and security that reflect what Tocqueville highlighted as ‘courage and habit to seek’, and ‘the assurance of reaping the benefits’.
What we argue is that the cultural recipe for entrepreneurship is complex and contains different elements. Intriguingly, we find ourselves in agreement with the intuitions that De Tocqueville (Reference De Tocqueville1958 [1835]) presented almost 200 years ago: for entrepreneurship we need courage and willingness to explore, combined with a sense of confidence that the benefits were there to be reaped.
Although we will focus on culture we will not deny the role of institutions. In the long historical perspective both can be seen as interdependent (Greif and Tabellini, Reference Greif and Tabellini2017). However, as we will discuss, we feel that the role of institutions supporting entrepreneurship is by now well understood (Estrin et al., Reference Estrin, Mickiewicz and Stephan2016). In contrast, the role of cultural values remains disputed (Storr and John, Reference Storr and John2020).
We start this paper by outlining the concept of cultural values. Next comes the core of the paper, where we propose the values that are conducive to entrepreneurship, relating them to the two elements of the entrepreneurial process we highlighted above (John and Storr, Reference John and Storr2018), with supporting arguments. In the empirical section, we introduce our data and methods. We will explain how we constructed a multilevel data set to test our hypotheses, combining Global Entrepreneurship Monitor (GEM) data with information extracted from the World Values Survey (WVS). Next, we will discuss the results, including extensions for more specific forms of entrepreneurship, before presenting our conclusions, pointing to limitations, and suggesting avenues for further work in this area.
2. Theoretical overview
2.1. Cultural values
The term ‘culture’ can be understood in different ways, depending on the context, and whether the framework is taken from the fields of sociology, anthropology, linguistics or humanities. In this paper, the notion of culture derives from sociology and is defined as a pattern of shared values, beliefs and behaviour of a group of people – these elements are repeated in various definitions, see for example, Hall (Reference Hall, Dirk and Ortner1980), Chanchani and Theivanathampillai (Reference Chanchani and Theivanathampillai2002) and Taras et al. (Reference Taras, Rowney and Steel2009). We tweak the definitions of Hofstede (Reference Hofstede1980) and Schwartz (Reference Schwartz, van de Vijver, Chasiotis and Breugelmans2011) to posit that culture is something that provides mental patterns, ways for people to arrange their minds. Culture relates to those aspects of personal values, beliefs and norms of behaviour that are shared within a given social environment. Although we will focus on values, these interact with beliefs systematically, and both elements can be seen as part of decision-making heuristics; utilising the latter, people save on the cost of obtaining and analysing information (Alesina and Giuliano, Reference Alesina and Giuliano2015). This decision-making heuristics shapes behaviour. McCloskey (Reference McCloskey2006), in her work on the cultural traits that matter for business and entrepreneurship, reintroduces the classic concept of virtue, within which values and related behavioural predispositions are seen to be coherent cultural traits. However, although we will draw upon McCloskey (Reference McCloskey2006), we will not use that term since recent literature gives it a narrower connotation (Newstead et al., Reference Newstead, Macklin, Dawkins and Martin2018).
There are three channels along which culture may affect economic behaviour, including entrepreneurship. First, culture has direct motivational effect through internalised norms (Pitlik and Rode, Reference Pitlik and Rode2017). For example, trivially, in a country characterised by individualistic values, there will be more people holding these values as their own.
This mean effect however does not imply uniformity in values at the individual level. For example, a person may be inclined to risk taking despite being surrounded by a generally risk-averse society, or maybe even because of it, seeing potential gains from boldness to take risks in a risk-averse society. Yet in such ‘incompatible’ environments, a person's entrepreneurial activity may be hampered by the fact that he/she will find difficult to secure business cooperation from those who do not share his/her values. It is how personal enforcement associated with societal cultural values works (Greif and Tabellini, Reference Greif and Tabellini2017; see also Schwartz, Reference Schwartz2006). This relates also to business behaviour; as Langlois (Reference Langlois2016: 3) argues: ‘in any culture that fails to validate commercial activity, people will want to better their conditions but culture will strongly discourage or even prevent them from doing so’. In contrast, and applying Davidsson's (Reference Davidsson2016) terminology, we may say that culture becomes an ‘external enabler’ of entrepreneurship if the corresponding values are conducive to it, ensuring cooperation of others that is essential for the successful implementation of the entrepreneurial projects. As observed by Krueger et al. (Reference Krueger, Liñán and Nabi2013: 705): ‘Two identical people (same experiences, perceptions and preferences) are likely to behave differently if they are under different cultural pressures’. Similarly, Wennberg et al. (Reference Wennberg, Pathak and Autio2013) notice: ‘individuals consider not only their own ability to succeed and the possibility of failure, but also how this action is consistent with prevailing cultural norms and practices’ (Ibid.: 758). More specifically, Chamlee-Wright (Reference Chamlee-Wright1997) argues that persuasion is critical to entrepreneurship. ‘For an entrepreneurial venture to get off the ground, the entrepreneur must persuade investors that his or her ideas are profitable ones, and consumers that the new or product or service is a desirable one’ (Ibid.: 47). Yet, that implies shared meaning and values; that is, culture. If entrepreneurial action is not culturally valued, and therefore not at the centre of cultural meaning, the entrepreneur will find it more difficult to persuade others to secure cooperation she/he needs. This is the second, social channel, along which culture affects entrepreneurship.
Third, culture may influence economic behaviour through the intermediation of formal institutions (Pitlik and Rode, Reference Pitlik and Rode2017). In the long run it is likely that the institutional set-up representing impersonal enforcement affects cultural norms (Greif and Tabellini, Reference Greif and Tabellini2017). However, recent empirical study finds that culture has more significant impact on formal institutions compared to vice versa. In particular, Gorodnichenko and Roland (Reference Gorodnichenko and Roland2017) document this in the context of the individualist-collectivist cultural traits as pitched against institutions. This perspective is consistent with the theory framework that stresses culture as more fundamental to formal institutions (Williamson, Reference Williamson2000). However, that does not imply that institutions have no independent effect. Both culture and institutions, although not static, are persistent (Greif and Tabellini, Reference Greif and Tabellini2017), and this implies that in a short run we focus upon, they may both have an independent effect on entrepreneurship.
Culture is more than values, and complete understanding of culture is impossible without ethnographic and anthropological studies. As argued by Lavoie (Reference Lavoie and Berger1991: 34): ‘Culture is to be understood broadly as the complex of meanings that allows us to comprehend human action’. By focusing on values, we abstract and ignore a lot that is significant (Heine, Reference Heine2020). What we gain, however, is a possibility of quantifiable broad cross-cultural comparisons. For that purpose, cultural values are typically measured by aggregating up survey instruments to the level of countries (or in some cases: regions), with WVS being the most popular source of empirical material recently (Alesina and Giuliano, Reference Alesina and Giuliano2015). The literature proposes different sets of cultural dimensions, such as those of Parsons and Shils (Reference Parsons and Shils1951), Kluckhohn and Strodtbeck (Reference Kluckhohn and Strodtbeck1961), Hofstede (Reference Hofstede1980), Schwartz (Reference Schwartz, Kim, Triandis, Kagitçibasi, Choi and Yoon1994), Trompenaars and Hampden-Turner (Reference Trompenaars and Hampden-Turner1997), Inglehart and Baker (Reference Inglehart and Baker2000) and House et al. (Reference House, Javidan, Hanges and Dorfman2002). There is already rich empirical literature that connects culture and entrepreneurship (De Clercq et al., Reference De Clercq, Lim and Oh2014; Liñán and Fernández-Serrano, Reference Liñán and Fernández-Serrano2014; Minola et al., Reference Minola, Criaco and Obschonka2016; Stephan and Pathak, Reference Stephan and Pathak2016; Stephan and Uhlaner, Reference Stephan and Uhlaner2010; Taylor and Wilson, Reference Taylor and Wilson2012; Thai and Turkina, Reference Thai and Turkina2014; Valdez and Richardson, Reference Valdez and Richardson2013; Wennberg et al., Reference Wennberg, Pathak and Autio2013). However, this literature typically uses composite cultural measures derived empirically from factor analysis. We argue that these techniques combine elements that have conflicting impacts on entrepreneurship into aggregate dimensions. This is consistent with the more general criticism expressed by Hayton and Cacciotti (Reference Hayton and Cacciotti2013: 715) who observe ‘that such clustering loses theoretical meaning and empirical information’. Thus, our approach differs. To borrow an expression from Heine (Reference Heine2020), our objective will be to try to unpack the composite; in our case, proposing the elements that we expect to matter most for entrepreneurial activities.
2.2. Which values matter for entrepreneurship?
We argue that the key societal cultural values that matter for entrepreneurship are security and creativity. In so doing, we primarily build on McCloskey's (Reference McCloskey2006, Reference McCloskey2010, Reference McCloskey2016) work on cultureFootnote 3 that is also consistent with the dimensions highlighted by De Tocqueville (Reference De Tocqueville1958 [1835]: 116). However, we follow Langlois (Reference Langlois2016) arguing that although McCloskey focused her intuitions on values related to innovation, we should shift emphasis towards entrepreneurship. Innovation is critical, yet especially in its most radical forms it is often embedded in new ventures as contrasted to its incremental forms characterising old and large corporations (Baumol, Reference Baumol2010; Martin and Mitchell, Reference Martin and Mitchell1998).Footnote 4
2.2.1 ‘The sense and habit to find’: thinking up new ideas, creativity
The major theoretical perspective on entrepreneurship is that of Schumpeter (Reference Schumpeter2008 [1911]). It emphasises the element of novelty: entrepreneurs create something new, they innovate. What we label after John and Storr (Reference John and Storr2018) as opportunity identification, may be further analytically split into opportunity discovery and opportunity creation (Alvarez and Barney, Reference Alvarez and Barney2007). Both may be seen as the first initial phase of the entrepreneurial process, but need not be. Entrepreneurial opportunities may be created in ‘iterative, inductive, incremental decision making’, continuously along the process of business venturing and development of the new firm (Ibid.: 17).
Here, the link to values is established by McCloskey (Reference McCloskey2010, Reference McCloskey2016); the role of cultural support for innovativeness is probably argued nowhere as strongly as in her work (Mokyr, Reference Mokyr2017). Although at present, acceptance of creativity, innovation and change is ubiquitous, we can also argue that societies still differ in the extent to which this value is embedded. Resistance to creativity stems from the fact that it may come at a cost. Innovations lead to development, but this development sometimes happens through ‘creative destruction’ (Schumpeter, Reference Schumpeter2008 [1911]) even if this is not always or even not mostly the case. According to McCloskey (Reference McCloskey2010), who follows Kirzner (Reference Kirzner1973) here: ‘A big or small entrepreneur, encouraged by dignity and enabled by liberty, alertly notices an opportunity, and takes it’ (McCloskey, Reference McCloskey2010: 20). The third element – noticing opportunity – is creativity, or ‘the sense and habit to find’ to use De Tocqueville's (Reference De Tocqueville1958 [1835]: 116) expression.
Researchers working in the Austrian tradition recognise the link between entrepreneurship, innovation and change (Foss and Klein, Reference Foss and Klein2012, Reference Foss and Klein2015). Within this tradition, Lavoie (Reference Lavoie and Berger1991: 44) sees entrepreneurial creativity as leading to new meaning and understanding: ‘alertness in itself is insufficient to cover all the aspects of entrepreneurship … creativity and judgment are also involved in entrepreneurship … a crucial component of judgement and creativity is interpretation’.
McCloskey emphasises yet something else; she radically departs from the economics approach, in a sense that she sees innovation as separated from pursuit of material gains (see also Storr, Reference Storr2013). In her words:
‘[T]he crux here (I repeat to my economist colleagues) is that a Max U model … cannot work to explain real innovation. That after all is the central point here – that routine maximisations … do not explain the modern world. What explains it, as the Austrian economists would put it is discovery. /…/ There is no U to max and no constraint to obey if real discovery is at issue, as against routine exploration for, say, oil’ (McCloskey, Reference McCloskey2010: 410).
Thus, creativity renders entrepreneurship. Yet, as observed already by Lavoie (Reference Lavoie and Berger1991), creativity is conditioned by culture:
‘Entrepreneurship is not only a matter of opening one's eyes, of switching on one's attentiveness; it requires directing one's gaze … And this raises the question of what gives a predirectedness to the entrepreneur's vision, of why he is apt to read some things and not others. I submit that the answer to this question is culture’ (Lavoie, Reference Lavoie and Berger1991: 46).
Creativity as a societal value may be interpreted as an aspect of individualism because it is based on affirmation of intellectual autonomy (Schwartz, Reference Schwartz, van de Vijver, Chasiotis and Breugelmans2011), which Pitlik and Rode (Reference Pitlik and Rode2017) interpret as ‘self-determination’. Similarly, Bennett and Nikolaev (Reference Bennett and Nikolaev2020) highlight the link between individualistic culture and innovation. The high values placed on creativity and new solutions increase the probability of new venture creation (Lee and Peterson, Reference Lee and Peterson2000). It needs ‘the courage to overcome the fear of change, to bear defeat unto bankruptcy, to be courteous to new ideas, to wake up next morning and face fresh work with cheer, resisting the despairing pessimism’ (McCloskey, Reference McCloskey2006: 508). Here, we see a further link between creativity and confidence that has been recognised in empirical work as the trait that is characteristically associated with starting new ventures (Kahneman, Reference Kahneman2011). High levels of confidence and ‘an optimistic temperament … encourages persistence in the face of obstacles’ (Kahneman, Reference Kahneman2011: 257). Optimism, innovation, and creativity go hand in hand.
Hence, we posit:
Hypothesis 1: A culture characterised by a high value being attached to creativity implies a higher likelihood of an individual starting a new venture.
2.2.2 ‘Assurance to reap the benefits’: security
We now turn to the opportunity exploitation. Security of property rights is a recurring theme in the entrepreneurship research, and empirical results confirm the importance of strong property rights (seen as a component of the rule of law) for entrepreneurship (Estrin et al., Reference Estrin, Mickiewicz and Stephan2016; Levie and Autio, Reference Levie and Autio2011).Footnote 5 This is because entrepreneurship, especially in the Austrian Economics tradition, is seen as a form of investment, where the human and financial resources of the owner-manager become tied to the new venture (Foss and Klein, Reference Foss and Klein2012, Reference Foss and Klein2015; Harper, Reference Harper2003; Mises, Reference Mises1966). The value of this investment depends on a future flow of returns that may materialise as a slow trickle and only after a considerable delay. For this reason, confidence in the medium to long-term security of property rights is an important external enabler of entrepreneurship.
In the extant research, security is considered in the context of institutions rather than that of culture (Estrin et al., Reference Estrin, Mickiewicz and Stephan2016; Levie and Autio, Reference Levie and Autio2011). Security of property rights is viewed as the most business-relevant component of the rule of law, albeit other aspects of the rule of law also matter for entrepreneurship and some may be even more fundamental (such as, e.g. personal security as defined by the Habeas Corpus Act). Generally, entrepreneurs need to experience security; this, within the realm of institutions, is delivered by effective constraints imposed on the executive branch of the government, particularly by the independent judiciary (Acemoglu and Johnson, Reference Acemoglu and Johnson2005; Greif and Mokyr, Reference Greif and Mokyr2016; North and Weingast, Reference North and Weingast1989).
Yet, formal institutions remain effective only as long as there is cultural support for the values and behaviours that sustain them. As Chamlee-Wright (Reference Chamlee-Wright1997) observes:
‘Police and court systems would be incapable of enforcing property rights and contracts, for instance, if most members of a society did not accept the legitimacy of the institutional rules. Such acceptance comes not from social contracts devised in the abstract, but through an evolutionary process within the culture of a community’ (Ibid.: 24).
Thus, the security of entrepreneurial investments depends not only on formal institutions, but also on social values embedded in culture. Furthermore, the latter may play a role independently of formal institutions. This is easy to see in countries, where much entrepreneurship simply happens outside the formal sector. In this case, security is provided by nothing more than informal norms and culture (De Soto, Reference De Soto1989).
Thus, regardless of the shape of the institutions, culture matters. One may say that property rights are as secure as the people in the neighbourhood accept them to be. The same relates to personal security and other aspects of human rights. Before we encounter government, we encounter our neighbours (Acemoglu and Robinson, Reference Acemoglu and Robinson2019), and it matters what their values are. As McCloskey (Reference McCloskey2006: 479) observes: ‘economies grow slowly – or not at all – when stealing … becomes a simpler way to wealth than working and selling’. This is akin to what Storr (Reference Storr2004) identified in his economic ethnography as the contrast between the ‘spirit of piracy’ and the ‘spirit of enterprise’.
Furthermore, pirates when settled become aristocrats, and McCloskey (Reference McCloskey2006) distinguishes two types of respect for property: one inherent to the aristocratic set of values, and one that is part of the domain of bourgeois virtues. For the aristocracy, the stress is on inheritance, while in the bourgeois set of values ‘it is bound up … with a vision of freely chosen work … the bourgeois thinks of himself as entrepreneurial’ (McCloskey, Reference McCloskey2006). In the latter case, security relates to the ‘finding and keeping’ component, which is crucial to entrepreneurship (Harper, Reference Harper2003; Kirzner, Reference Kirzner1989).
What we see here is security that is reciprocally offered within a community, out of mutual respect, and includes both personal security (e.g. freedom from assault) and the securing of assets. Both create space for entrepreneurship, and what follows is that respect for the security of others is a cultural value that has a deeply social character. That is, respect for property is always associated with societal bounds, participation and identification with a group. Mutually granted security is related to the emphasis on the reciprocal respect.
Let us now discuss a counter-argument that would suggest that cultural value of security may have a negative impact on entrepreneurship. Cultures strong in emphasis on feeling secure (collectively granted), might also emphasise the need to do things in the way they are ‘normally’ done. This latter aspect links with security but means that there will be less social support for starting a new business (Mueller and Thomas, Reference Mueller and Thomas2000; Russell, Reference Russell2004).
To solve this difficulty, we apply McCloskey's (Reference McCloskey2006) terminology with respect to security. Where the aspect of tradition prevails, we are in the realm of aristocratic values. In contrast, entrepreneurship-friendly bourgeois values will stress security granted to those who create something new by their own efforts. In that case, as has been pointed out by Kaasa and Vadi (Reference Kaasa and Vadi2010), while people tend to value security, this bourgeois value form part of a strong system of norms that offers support and security to potential and active entrepreneurs. Hence, we formulate our second hypothesis.
Hypothesis 2: A culture characterised by a high value being attached to security implies a higher likelihood of an individual starting a new venture.
To summarise, we follow McCloskey (Reference McCloskey2006) in arguing that an entrepreneurship-friendly culture is a complex phenomenon, where different, even apparently divergent values, need to be present. Cultural affirmation of creativity supports opportunity discovery and creation. Security, both personal and of property resulting from entrepreneurial action, supports opportunity exploitation. Both therefore create a cultural recipe for entrepreneurship.
3. Data and methods
3.1 Data, variables
To test our hypotheses, we use a large dataset constructed by combining information from the GEM, with country-level data added from the WVS, World Bank and the Polity project. The GEM project utilises harmonised cluster sampling, normally of at least 2,000 working age individuals per country, to identify new entrepreneurs. It measures different aspects of entrepreneurship and the related characteristics of individuals (Reynolds et al., Reference Reynolds, Bosma, Autio, Hunt, De Bono, Servais, Lopez-Garcia and Chin2005). GEM provides an internationally comparable dataset that is widely used as a source of data for multilevel entrepreneurship research (e.g. Estrin et al., Reference Estrin, Mickiewicz and Stephan2016; Lim et al., Reference Lim, Oh and De Clercq2016; Schillo et al., Reference Schillo, Persaud and Jin2016). The sampling procedure is reviewed by Reynolds et al. (Reference Reynolds, Bosma, Autio, Hunt, De Bono, Servais, Lopez-Garcia and Chin2005). As argued by Davidsson (Reference Davidsson2016: 207), today the GEM project ‘provides the knowledgeable user with a unique and high quality data set’ enabling a cross-country comparison of entrepreneurship.
3.1.1 Dependent variable
We use the GEM measure of the nascent entrepreneur as our dependent. It is an indicator variable that takes the value 1 if the GEM questionnaire-defined criteria are met. These criteria focus on individuals who moved from having a new business idea to taking concrete steps to start a new venture: the business has not been operational for more than 3 months, that is, it has not yet transferred from the nascent start-up phase to the new business phase. For further details on how this variable is constructed see Reynolds et al. (Reference Reynolds, Bosma, Autio, Hunt, De Bono, Servais, Lopez-Garcia and Chin2005). Nascent entrepreneurship is a popular dependent variable in the empirical literature because it suffers less from either endogeneity vis-à-vis some of the characteristics of individuals or from selection bias.
We drew our sample from every country surveyed in GEM that had data available at the time of writing. Therefore, our starting point is an unbalanced multilevel cross-country panel data of individuals for the period 2001–2015. However, the set of observations we could use was limited with the availability the country level data on values based on the WVS, as we will explain below. As a result, in our regressions we utilise 2005–2015 data.
3.1.2 Core independent variables and culture-related controls
Our hypotheses relate to the impact of societal values on entrepreneurship. The measures of various values are taken from the WVS (2015), which is the most popular source of data on cultural values (Alesina and Giuliano, Reference Alesina and Giuliano2015). We picked questions from the reduced Schwartz Value Inventory included only in waves 5 (2005–2009) and 6 (2010–2014) of this multi-country survey with seven waves since 1981. First, in order to address creativity (hypothesis 1), we use the question about the importance of thinking up new ideas and being creative. Next, for (bourgeois) security we included the question about the importance of living in secure surroundings. We take it as being the most closely correlated to our arguments in Section 2.2.2 and we will use it to test hypothesis 2. However, to alleviate concerns about omitted variable bias, we included importance of always behaving properly, and importance of tradition (both related to the aristocratic dimension of security), accepting risks, striving for wealth and success, importance of being rich, importance of being very successful, importance of having good time and finally importance attached to the environment.
As the aim was to analyse the impact of the cultural environment on entrepreneurship, the cultural variables were included at the country level. Hence individual-level variables from WVS were aggregated by calculating the country means, using the weights provided by the WVS, which is a standard approach in constructing cultural values (Alesina and Giuliano, Reference Alesina and Giuliano2015; Pitlik and Rode, Reference Pitlik and Rode2017). When merging data from WVS with GEM data, we matched the GEM individual data with WVS indicators from the same year or up to 2 years previously, where this was available (depending on the year the particular WVS wave was implemented in a country). If the gap between the GEM observations and the available WVS data was 3 or more years we did not make a match. Exact mapping is available on request.
3.1.3 Other controls
Alongside culture, constitutional level institutions especially are characterised by considerable inertia over time, and to stack odds against our tests we control for effective constraints on the executive power of government (Acemoglu and Johnson, Reference Acemoglu and Johnson2005; Greif and Mokyr, Reference Greif and Mokyr2016; North and Weingast, Reference North and Weingast1989). In particular, higher levels of constraints include the power of a strong and independent judiciary system to put on hold any abuse of power by the executive. Recent examples of the use of this indicator in empirical research on entrepreneurship and institutions include Estrin et al. (Reference Estrin, Mickiewicz and Stephan2013b; Reference Estrin, Mickiewicz and Stephan2016), Estrin et al. (Reference Estrin, Korosteleva and Mickiewicz2013a), Estrin et al. (Reference Estrin, Mickiewicz, Stephan, Wright, Grosse and Meyer2019), and, running parallel, Besley and Mueller (Reference Besley and Mueller2018) for Foreign Direct Investment. The executive constraints are formal institutions' counterpart to the cultural value of security (hypothesis 2), as it controls for constitutional conditions protecting property rights. It is taken from the Polity project (Marshall et al., Reference Marshall, Jaggers and Gurr2010).Footnote 6
At the regulatory level of institutions we include an index that is most directly related to start-up activities, that is Business Freedom, taken from the Economic Freedom dataset offered by the Heritage Foundation and Wall Street Journal project. It weights equally 13 indicators from World Bank ‘Doing Business’ set: four source indicators relate to starting a business, three to obtaining licence, three to closing a business (which may affect entry decisions via value of real options) and three to getting electricity connected. All these are integrated into one index (The Heritage Foundation, 2020).
The entrepreneurial entry decision is not only influenced by institutions but also by other environmental (country) level features, and by the characteristics of individuals. We include the level of development (and its square term, to allow for nonlinearity), which may affect attractiveness of entrepreneurship (Estrin et al., Reference Estrin, Mickiewicz, Stephan, Wright, Grosse and Meyer2019). It is proxied by GDP per person, purchasing power parity, constant prices (from World Bank). In addition, we include unemployment, which may have either a push (necessity) or discouraging (lower opportunities) effect on entrepreneurship (Parker, Reference Parker2018), also from World Bank.
We use indicators from the GEM survey to control for individual characteristics. We include an individual indicator variable that represents the respondent's involvement in another established business (for serial entrepreneurship); the indicator variables that represent a business discontinued in the last 12 months; experience of acting as a business angel in the past 3 years; age and age squared; gender and educational achievement categories. This set of control variables is consistent with other studies on institutions and entrepreneurship (e.g. Estrin et al., Reference Estrin, Mickiewicz and Stephan2016). In GEM data we also have two variables that correspond to fear of failure, and to the level of confidence in one's own entrepreneurial skills. These are useful individual-level controls when placed alongside societal level cultural indicators as described above. Both raise the threshold for confirming hypothesis 1, as they control for the impact of personal traits related to self-direction (Pitlik and Rode, Reference Pitlik and Rode2017) that were internalised.
Table 1 presents a list of the variables used in this study and their description and sources, and contains descriptive statistics. A correlation table is available on request.
Definitions of variables and descriptive statistics
3.2. Estimation strategy
We use multilevel modelling to address the hierarchical structure of our data, with observations on individuals at the lower level and country-years representing the higher level. This structure implies that the individual observations are not independent. Failure to account for this would lead to biased results (Hox et al., Reference Hox, Moerbeek and van de Schoot2017). Regressing individual outcomes on country level values also alleviates concerns about reverse causality (Pitlik and Rode, Reference Pitlik and Rode2017). We follow Cameron and Miller (Reference Cameron and Miller2015) recommendation for modelling standard errors for multilevel discrete choice models based on the exchangeable covariance matrix. This accounts for within cluster correlations between observations; in our case clusters are country-years subsamples.
Following the standard routine, we first estimate the intercept only (empty) model. Next, we add our explanatory variables, as discussed above, and employ the full random intercept model. After that we consider random coefficient (slope) models, which allow for the coefficient(s) of some variables to vary randomly across country-year groups. More specifically, we tested whether the impact of entrepreneurs' age, education, and gender on the likelihood of being engaged in new business creation is the same across country-year groups by introducing random coefficients for these variables. We focus on these characteristics of individuals as they are seen to be core determinants of entrepreneurship (Parker, Reference Parker2018), and we also consider them to be related to how the individuals are affected by culture. By experimenting with these alternative models, we develop an understanding of how robust the results are and how sensitive they are to alternative methods of estimation. Our random intercept model, which utilises all the explanatory variables discussed earlier, is specified as follows:
where Startupij is our GEM-based indicator of being involved in new venture creation; variables with subscripts ij represent individual-level direct effects; variables with subscript j represent country-year effects; the random intercept in the equation is represented by uj, with ɛij representing individual-level residuals. In turn, the random coefficients models (not spelled out for brevity) will have random slopes of the variables we discussed in the random part of the model.
Given that our dependent variable is binary, we apply a logit estimator. Values reported in Table 2 are odds ratios (obtained by exponentiating both sides of the equation) instead of raw logit coefficients, thus facilitating interpretation and enabling immediate evaluation of the size of the effects by the reader. In that form, the model is interpreted as multiplicative. For a unit change in any independent variable, the odds ratio changes by a factor reported for that variable in Table 2. Where an independent variable is a dummy, the reported value shows how much larger being in that category makes the odds ratio of nascent entrepreneurship compared with the benchmark category. For all types of variables, if the value reported is one, the given variable has no impact on the odds ratio of nascent entrepreneurship involvement. A value below 1 implies a negative impact; and above 1, a positive impact (e.g. Long and Freese, Reference Long and Freese2014). The values in brackets are exponentiated standard errors.
Estimation results
Notes: Odds ratios reported instead of coefficients.
Exponentiated standard errors reported in brackets.
***p < 0.001, **p < 0.01, *p < 0.05, +p < 0.10.
4. Results and discussion
4.1. Core results
Table 2 presents the results.
Hypothesis 1 appears to be strongly supported by the results: if thinking up new ideas and being creative is valued in a society, the odds of starting a business are high. We can evaluate the magnitude of this effect by multiplying the odds ratio by the exponential function of standard deviation (Long and Freese, Reference Long and Freese2014) to obtain change in odds for one standard deviation change in the country creativity score. It comes as tangible 2.2 (based on model 1). Similarly, with respect to hypothesis 2, high value placed on security in the society shows a strong relationship with the odds of people being engaged in a start-up, consistent across the models. Interestingly, again evaluating the magnitude of this effect using the same method as before we get the change triggered by move by one standard deviation to be almost the same as in the previous case that is about 2.2 (as here while the odds ratio is lower, this is counterbalanced by higher standard deviation).
Thus, the hypotheses are supported.
At the same time the importance given to tradition, and to behaving properly in a society, appear to have no significant association with the odds of engaging in a start-up. This provides support for the distinction between the aristocratic values focused on tradition and entrepreneurship-friendly bourgeois values emphasising granting security for those putting their own effort for creating something new, as discussed in section 2.2.2. The latter correlates with entrepreneurship, but not the former.
Being able and willing to cope with risks and uncertainty has been seen as a key entrepreneurial trait since the argument was presented by Knight (Reference Knight2009 [1921]). Although the willingness to accept uncertainty may be represented as courage, risks can be insured against. Previously, Koudstaal et al. (Reference Koudstaal, Sloof and Van Praag2016) have found entrepreneurs having similar risk aversion levels compared to managers. The evidence we obtained regarding the societal value of taking risks is mixed at most. The statistical significance is confirmed in just one model and not in others. Here, we may also face the problem that the question combines two somewhat different aspects of values: adventure and taking risks. Adventure may pertain to a hedonistic aspect of culture.
The importance of being very successful in a society turned out to be not significant for the probability of starting a new business. Emphasis on wealth (and more widely: performance) in a society could mean that the achievements related to entrepreneurial activity are highly respected. On the other hand, however, appreciation of wealth may not be the value that differentiates those who are engaged in new business venturing from those who take some well-rewarded employee jobs in business. It seems that our results are more consistent with the latter view. In the case of success and wealth motivation (performance orientation) entrepreneurship may not be perceived as the best way of achieving it. Climbing up the ladder of managerial positions within a large organisation may be a better strategy in this case (Hofstede et al., Reference Hofstede, Hofstede and Minkov2010).
If we follow Teague et al. (Reference Teague, Storr and Fike2020: 2) and define materialism as ‘desires to acquire material possessions for oneself above most other things’, then the ‘importance of being rich’ is a good measure of materialism. Although, our results cannot refute the proposition that it remains central for economic activity, what we see is that it is not specifically associated with entrepreneurship, again arguably because there are ways to get rich within extant organisations, and alternative paths to wealth. We may also notice the results on materialism that Teague et al. (Reference Teague, Storr and Fike2020) obtained. The evidence they provided suggests that countries that embrace markets therefore creating conditions conducive to entrepreneurship are less not more materialistic. Their stronger result is consistent with McCloskey's (Reference McCloskey2010) argument we discussed above: the creative element of entrepreneurship cannot be explained by greed alone.
Developing further the analytical distinctions, we argue that courage and confidence to innovate can be seen as process-oriented values, while wealth and success are values that can be viewed as more result-oriented. It might well be that when starting a new business, success is a goal that is too remote to have much initial impact. In contrast, (bourgeois) courage and confidence to innovate are needed right from the very beginning of the process. This reinforces the emphasis that the literature places on entrepreneurship being seen as a process (Davidsson, Reference Davidsson2016; McMullen, Reference McMullen2015).
4.2. Extensions: innovative entrepreneurship, high aspirations, business angels and young businesses
In the results reported above, our dependent variable is the classic (if narrow) measure of entrepreneurship used in cross-country studies: the likelihood of being involved in a nascent start-up project (Davidsson, Reference Davidsson2016). However, we conducted several additional experiments utilising alternative entrepreneurship-related measures in order to (i) understand more fully the role of cultural values, (ii) compare our results with earlier work and (iii) test the robustness of the results. For reasons of space, we briefly discuss these additional results instead of presenting the tables. The tables, data and codes that generated the results are available at the authors' ResearchGate web pages.
These additional tests are as follows. The first derives from a broad stream of research that explores the link between culture and innovation. Shane (Reference Shane1992, Reference Shane1993) made early contributions to this, and recent empirical study includes Kaasa and Vadi (Reference Kaasa and Vadi2010), Taylor and Wilson (Reference Taylor and Wilson2012) and meta-analysis by Sarooghi et al. (Reference Sarooghi, Libaers and Burkemper2015). In the GEM data, there are questions that rank the degree of product and process innovation of nascent entrepreneurs. Using a combination of these items, we constructed a scale ordering start-ups according to their degree of innovativeness, and used it as a dependent variable in an ordered logit model (with the same set of independent variables and random effects as in Table 2). Interestingly, the results were entirely consistent with Table 2: the cultural factors that correlate with the standard measure of start-up engagement also correlate with entrepreneurship ordered by its innovativeness.
Second, in the entrepreneurship literature, there is an argument that the ambitious, growth-oriented forms of entrepreneurship are particularly important. Cross-country studies based on GEM operationalise this with the level of (employment) growth aspirations of entrepreneurs (e.g. Autio and Acs, Reference Autio and Acs2010; Estrin et al., Reference Estrin, Korosteleva and Mickiewicz2013a). Again, following the same specifications as before, we ran four multilevel count Poisson models, with the GEM measure of expected number of employees in five-years' time taken as the dependent variable, as is used in the extant literature. We found some interesting differences between these and our core results. Secure surroundings as a cultural value turned out to be highly significant, just as it did in our core results, but this time creativity did not. This calls for more exploration, as at the individual level there is evidence on the link between innovation and growth-oriented strategies (Estrin et al., Reference Estrin, Korosteleva and Mickiewicz2020). Next, interestingly, a willingness to take risks was negatively correlated with ambitious entrepreneurship; it is possible that in order to ensure growth, risks have to be rather avoided. In addition, for ambitious forms of entrepreneurship, we now have a significant positive impact of the importance of being very successful. The latter result is intuitive in that the cultural support for business success (performance orientation) provides fertile ground for higher levels of ambition. Thus, although performance-orientation as a societal value does not matter for entrepreneurship overall, it does matter for high growth aspirations ventures.
Third, we explored a different form of entrepreneurship: engagement as an informal investor. This exercise is parallel to study by Li and Zahra (Reference Li and Zahra2012) on cultural support for entrepreneurial finance, yet in their case it is represented by venture capital. The GEM question we use as a dependent variable about being an informal investor is backward-looking, and this time we therefore dropped the variables representing individual fear of failure and entrepreneurial skills since we would risk simultaneity (endogeneity). The results were somewhat weaker than for our core models, yet in two out of four models we obtained significant results both on creativity and on security. At the same time, the coefficients on cultural support for willingness to take risks turned out to be insignificant, again consistent with our argument on deemphasising the role of risk acceptance. However, here, it may also be that, unlike nascent entrepreneurs, business angels have more opportunity to spread their risks, and these risks are also perceived as being lower thanks to their typically higher wealth position. As a result, risk taking aspect matters less.
Fourth and finally, we replaced the nascent start-up dependent variable with being the owner-manager of a young business (up to 42 months old). As this is, by construction, backward-looking we avoided simultaneity by omitting the same independent variables that we dropped when testing the business angel dependent variable. This time, a very strong positive result on the cultural value of security dominates the other results. But creativity comes out as insignificant, and support for risk-taking is mostly insignificant as well. This echoes our results on growth aspirations. Again, it appears that creativity is critical in the initial phase of acting on entrepreneurial opportunities (nascent entrepreneurship) but loses impact later in the process. For fully operational young business we also have a clear-cut negative and significant result of tradition-supporting values.
This last experiment deserves attention as it creates an opportunity to extend GEM-based work on the stages of entrepreneurship (e.g. Bergmann and Stephan, Reference Bergmann and Stephan2013; Mickiewicz et al., Reference Mickiewicz, Nyakudya, Theodorakopoulos and Hart2017; Wennekers et al., Reference Wennekers, Van Wennekers, Thurik and Reynolds2005) where the role of the cultural environment could be hypothesised to change during the entrepreneurial process. As already highlighted, there may be good reasons why support for creativity matter most for nascent businesses and, in contrast, why the societal value of security is what becomes critical in the more advanced phase. This theme is worth exploring in further research.
5. Conclusions
Although preserving the role of individual human agency in the theory of entrepreneurship, we need to account for social environmental influences (Davidsson, Reference Davidsson2016), and look beyond the role of institutions, and consider culture (Alesina and Giuliano, Reference Alesina and Giuliano2015; Gorodnichenko and Roland, Reference Gorodnichenko and Roland2017). We subscribe to entrepreneurial theory built on Austrian Economics with its emphasis on entrepreneurial judgement leading to entrepreneurial action, a prime example of which is new firms creation, the outcome we analyse in this paper. New firms creation process can be presented as consisting of two key elements, opportunity identification and opportunity exploitation (John and Storr, Reference John and Storr2018). The first element is supported by creativity. At the same time, the aspect of investment related to start-ups emphasised by Foss and Klein (Reference Foss and Klein2015) implies the importance of security, which as we argue is supported not only by institutions but also by cultural values (Chamlee-Wright, Reference Chamlee-Wright1997).
When considering the role of culture in entrepreneurship, we postulate that using highly aggregated constructs of culture may not suit the purpose when it comes to entrepreneurship (Hayton and Cacciotti, Reference Hayton and Cacciotti2013). Instead of matching pre-existing aggregate measures of culture with entrepreneurship, we identify two cultural traits that could be considered to be consistent with entrepreneurial behaviour. This is based on the assumption that ‘culture is decomposable’ (Mokyr, Reference Mokyr2017: 8). In particular our analysis leads us first to stress creativity, an aspect of individualism, akin to Pitlik and Rode (Reference Pitlik and Rode2017) self-direction construct. However at the same time, we supplement it with the second element, security. Thus, we follow McCloskey (Reference McCloskey2006), arguing that a business- and entrepreneurship-supporting culture is a balance of different, sometimes apparently conflicting elements, and the right mixture is not easy to identify. However, the approach proposed by John and Storr (Reference John and Storr2018) is promising. Here we start with the identification of the key components of the entrepreneurial process and next ask which elements of culture support those.
We conclude that our results are consistent with the thesis that creativity and security form a bundle of cultural ingredients that support entrepreneurship. In contrast, the stress on wealth, success and achievement (performance orientation and materialism) that has been emphasised in the classic entrepreneurship literature (such as by Baumol and Strom, Reference Baumol and Strom2007) does not appear to play a role. This seems consistent with results obtained by Teague et al. (Reference Teague, Storr and Fike2020), as discussed above.
We think there are two explanations for it. First, there are alternative ways to achieve wealth, either within pre-existing business organisations, or sometimes not by value-creation, but by plunder (by piracy, to use Storr's (Reference Storr2004) terminology); the latter being consistent with aristocratic not bourgeois set of values in McCloskey's (Reference McCloskey2006) sense, with oligarchic structures as described by Acemoglu and Robinson (Reference Acemoglu and Robinson2019), and with destructive entrepreneurship in Baumol's (Reference Baumol1990) terminology. Second, following McCloskey (Reference McCloskey2010) we should avoid economic reductionism and see drive to achieve material wealth as one of many and not necessary a dominant cultural value.
Similarly, our results on the affirmation of risk-taking are at best mixed. Risk can be insured against, and the uncertainty-handling aspect seen as core to entrepreneurship is arguably more closely related to innovativeness, and therefore creativity. In that sense the two major entrepreneurship theory traditions, that of Knight (Reference Knight2009 [1921]) stressing uncertainty, and that of Schumpeter (Reference Schumpeter2008) stressing innovation, may be closer to each other than it appears.
Taken together, this calls for a reconsideration of the basic assumptions about human motivation upon which the classic entrepreneurship theory is built. As we have argued, pursuit of wealth may be common, but is not a feature specific to entrepreneurship, unlike what some of the classic entrepreneurship theory emphasise (Baumol and Strom, Reference Baumol and Strom2007). This is a fundamental reason why the classic economics' approach (based on wealth maximisation) is too narrow to interpret the entrepreneurship phenomena in full.
To conclude, we think that Tocqueville was surprisingly accurate arguing almost 200 years ago that wealth produced by entrepreneurs is supported by cultural values that combine ‘courage to seek’ and ‘habits to find’ with ‘assurance of reaping the benefits’.
Acknowledgements
This study was supported by the Estonian Research Council Grant PRG380. We also express gratitude to the editor and the reviewers of this journal, Anna Grosman, Johannes Kleinhempel, and participants of seminars at Aston University, Birkbeck College, Syracuse University, Tartu University and AOM conference in Boston, for comments and feedback.
