The Nature of Entrepreneurship and its Determinants: Opportunity or Necessity?

Within the institutional theory of North (1990, 2005), the objective of this study is to analyse the impact of economic and institutional factors, formal and informal, in the entrepreneurial activity of nations, particularly in Total Entrepreneurial Activity (TEA). In order to evaluate the simultaneous influence of economic and institutional factors on the entrepreneurial activity, a multiple regression approach is used with cross-country data sets. The results show that TEA is negatively related to infrastructural capacity and political stability of a country, and positively related to government spending and freedom of expression and corporate associations (Voice & Accountability) at a country level. It is also tested the relationship between TEA and GDP per capita. Our results confirm a convex relationship between the two variables giving evidence that the entrepreneurial activity is mostly necessity driven rather than motivated by opportunity.


INTRODUCTION
Emerging and increasingly actual, the concept of entrepreneurship by its transversally, heterogeneity and subjectivity (Davidson, 2006), is far from congregating academic consensus (Berglann, Moen, Røed, & Skogstrøm 2011, Martin, Picazo, & Navarro, 2010. Indeed, due to the growing interest in the subject of entrepreneurship, such interdisciplinary should maintain its dynamic influence in the near future (Davis, 2008).
In fact, it is perceptible the contribution of various scientific fields in the construction of the concept of entrepreneurship, such as, psychology (Shaver andScott, 1991, Fishbein andAjzen, 1975), economics (Schumpeter, 1934, Cantillon, 2001, Marshall, 1961, Knight, 1921, Schumpeter, 1942, sociology (Reynolds, 1991, Thornton, 1999 and management (Stevenson and Jarillo, 1990, Sahlman and Stevenson, 1991, Timmons and Spinelli, 2004, Stevenson, 2000. The interdisciplinary nature helps us to decode the different levels of analysis of entrepreneurship: individual-level, corporate-level and country-level. Within psychology, the degree of analysis of entrepreneurship focuses on the individual-level (Shaver and Scott, 1991), in management the level of analysis focuses on corporate-level, and in economics the interest of analysis of entrepreneurship is concentrated at a country level (Nandram and Samsom, 2008). Despite different levels of analysis, it seems obvious the connection between them, to the extent that the individual perception of the entrepreneurship towards the environment that surrounds it is a key factor for the success of the company (Bruno and Tyebjee, 1982), and inherently entrepreneurship (at individual-level and/or corporate-level) enhances the economic development of a country (Lari & Ahmadian, 2012). Indeed, due to the different degrees of analysis and its multidisciplinary nature, as well as the heterogeneity of its determinants, that Verheul et al. (2002) suggest the eclectic theory of entrepreneurship to reach a more comprehensive concept. Therefore, it is important to understand the wide range of determinants that helps to explain, at the country-level, the greater or lesser propensity to entrepreneurship and detect trends that are related to entrepreneurship by necessity or opportunity. Knowing that the study of entrepreneurship at the country-level is not so well developed as it is at the individual or at corporate-level, we focus on the country-level in order to decode a set of determinants and to measure their contribution in explaining entrepreneurial activities. The aforementioned heterogeneity led us to study a wide range of determinants and to understand which ones have more relevance to explain entrepreneurship at the country-level. The value added of our study lies on the variety of the determinants to understand the nature of Entrepreneurship that is, whether it is necessity or opportunity driven. Additionally we test the convex hypothesis between entrepreneurship and per capita income and determine the threshold level that drives the shape of this relationship. This paper is organized in four main sections: literature review, methodology, estimation results and conclusions. In the literature review, we revisited the concept of entrepreneurship, discussing the two types of entrepreneurship motivated by necessity or by opportunity, explaining also the determinants of these types of entrepreneurship. The methodology section describes the nature of the data, the estimation techniques and analyses the obtained results. The final section presents the main conclusions emphasizing the most important aspects of this research.

LITERATURE REVIEW
The current academic divergence (Agca et al., 2012), which prevails from many decades ago (Cole, 1942), according to Iversen, Jorgensen, & Malchow-Moller (2008), arises from difficulties in the conceptualization and definition of theoretical models to measure entrepreneurship. In this vein, also because the phenomenon of entrepreneurship is being complex, dynamic and with diversified purposes (Bruyat and Julien, 2001), there are different concepts associated with economics, management and psychology perspectives.
The awareness that entrepreneurship is essential for economic growth (Naudé, 2010) assumed as the "main vehicle of economic development" (Anokhin et al., 2008: 117), seems to indicate the prevalence of the economic nature over the management or psychology ones. In fact, extensively studied since its inception by many economists like Knight , Schumpeter, Kirzner, Baumol, Marshall, among many others, the effect of entrepreneurship on economic variables, such as employment, innovation and wellbeing (Acs et al., 2008) can justify its importance in this area.
Despite the current academic divergence some consensus prevails in line with the Schumpeterian doctrine, that entrepreneurship is manifested through the relentless pursuit of business opportunities through innovation and creativity (Bjørnskov & Foss, 2008).
The impact of entrepreneurship on the economy has been studied at the level of the company, sector or region in detriment of the comparative analysis between nations (Stel, Thurik, & Carree, 2005). The level of economic development of a country is an important factor in explaining his entrepreneurial activity (Carree, Stel, Thurik, & Wennekers, 2007;Wennekers, Stel, Thurik, & Reynolds, 2008). However, several authors confirm the inverse relationship between GDP per capita and entrepreneurial activity (Stel et al., 2005). Some authors partially verify this inverse relationship to describe a convex curve between entrepreneurship and gross domestic product per capita (Acs, Audretsch, & Evans, 1994;Wennekers & Thurik, 1999). The entrepreneurship by necessity, opposing the entrepreneurship by opportunity / capacity may explain the inverse relationship between the two variables (Reynolds, Camp, Bygrave, Autio, & Hay, 2001). Therefore, we should differentiate the necessity driven entrepreneurship from the opportunity driven entrepreneurship. The former stems from the belief that the creation of self-employment grants to its promoter bigger utility and it is usually a result of employment that has been lost or from a saturation of the labor market (Block and Wagner, 2010). The latter, opportunity driven entrepreneurship, relates to the identification of an opportunity arising from an innovative idea (Valdez et al., 2011). The value-added generated by entrepreneurship due to necessity is residual and ephemeral to the economy. On the contrary entrepreneurship by opportunity generates higher value-added lasting longer, due to its innovative nature associated with technology-based activities.
In parallel to this relationship between GDP per capita and entrepreneurship, the institutional theory of North (1990North ( , 2005 which confirms the contribution of institutions in economic development in the long term, is the basic reference for the study of entrepreneurship (Díaz-Casero et al., 2012, Bjørnskov and Foss, 2008, Veciana and Urbano, 2008, Álvarez and Urbano, 2011, Salimath and Cullen, 2010. In agreement with this theoretical stream, we must distinguish between the informal role of institutions in the creation of ideas, beliefs, attitudes and personal values and the formal role, which includes a set of political-legal rules, economic rules and contractual procedures. If, on the one hand, the role of informal institutions through its governance has impact on entrepreneurial activity (McMillan & Woodruff, 2002), his formal role as more Several authors have confirmed the existing relationship between some indicators of economic freedom (published by The Heritage Foundation) and entrepreneurship measured by Total Entrepreneurial Activity (Díaz-Casero et al., 2012, Bjørnskov and Foss, 2008, McMullen et al., 2008. According to Acs & Armington (2004), Wennekers, Stel, Thurik, & Reynolds (2005) and Alvarez & Urban (2011) the factors of competitiveness also have significant impact on the entrepreneurial activity in a country.
In the early studies on entrepreneurship, the factors explaining its performance were mainly economic (Grilo & Thurik, 2005). However, given the weak explanation of the economic factors in the process of the entrepreneurship (Freytag & Thurik, 2007), several authors suggested also cultural dimensions to improve the degree of explanation , Wennekers et al., 2007, Hofstede et al., 2004, Osman et al., 2011 such as, education, religion, language, ethnic factors, the role of women in the labor market or Hofstede's cultural index, among others.

METHODOLOGY
This section explains the methodological aspects of the study used to identify the relevant determinants of total entrepreneurial activity of a country (TEA).

The sample
The Global Entrepreneurship Monitor (2011) reports data on TEA for a set of 54 countries. However due to missing values on the explanatory variables the sample is reduced to 36 countries and this set of countries is presented in Table 1.

Estimation technique
Focus will be given to understand the influence of different factors on the rate of entrepreneurship, as measured by Total Entrepreneurial Activity (TEA) and published by the Global Entrepreneurship Monitor (GEM) in 2011. To this end, we use a multiple regression analysis based on cross-country data. The estimation approach will make possible to compare the relative effect of various independent variables on the variable of interest (TEA). The cross-section multi-country model will be estimated initially by OLS and check its relevance through the usual diagnostic tests. Adaptations and corrections will be made to the regression model in case of detection of violation of some basic assumptions.

Variables of control
Entrepreneurship is a multidisciplinary concept, therefore, a vast number of explanatory variables can be used to explain its behavior. For this reason Table 2 reports the main factors that are likely to influence the decision to undertake a business activity. To optimize the chosen model, we follow the backward mode of estimation, starting with the whole set of explanatory variables and eliminating sequentially the variables with no statistical significance after performing an F-test on the joint significance of the population parameters. Doing so we have reached to a parsimonious model that includes the four most relevant explanatory variables explained in Table 3. Captures perceptions of the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including politically-motivated violence and terrorism. Another task of our study is to verify the typology of entrepreneurship whether it is driven by opportunity or by necessity. In doing so we analyze the relationship between GDP pc and TEA through a linear regression to verify the convexity hypothesis

Models to estimate
When all the usual assumptions on the error term and the explanatory variables are fulfilled, the OLS (ordinary least squares) estimation approach can be used to estimate our cross-section model. For comparison, we perform also the GLS estimator  Table 3): Our second goal is to understand the relationship between TEA and income per capita with the purpose of identifying whether the entrepreneurship is motivated by opportunity or by necessity. We propose a log-log model (Model 2), which allows us to obtain the elasticity between the two variables: A negative elasticity of this relationship will provide evidence in favor of entrepreneurship motivated by necessity. The lower the standard of living in a country the higher the necessity for people to create their own jobs.
Finally, in an attempt to verify the convexity hypothesis described by some authors between TEA and GDP pc, we tested the following quadratic function (Model 3): . From this regression we are able to determine the threshold level of income that beyond this point the shape of the relation inverts its initial negative tendency. Table 4 reports some elementary descriptive statistics on the variables used in the estimation approach. Observing the data we can highlight some relevant aspects. The dependent variable Total Entrepreneurial Activity (TEA) represents the percentage of the population able to develop a professional activity, that is, actively involved in the creation of a business, either in the starting phase of business activity or 42 months after the birth of a business unit (Bosma, Wennekers, & Amoros, 2012). On the basis of the observed data, the values of TEA vary between 3.7 (minimum value for Slovenia) and 23.7 (maximum value for Chile).

Descriptive statistics
The variable 'Infrastructure' is an index that varies between 1 and 7 points and describes the quality of the infrastructures in a country in three major areas: transports, energy and communications. On the basis of the observed data, the values of this variable vary between 3.16 (minimum value for Algeria) and 6.27 (maximum value for France). The higher the value, the higher the structural facilities in a country. Republic of Iran) and 1.382 (maximum value for Finland). The higher the value, the higher the political stability and the lower the level of potential violence in a country.
According to the Coefficient of Variation (see Table 4), the higher variability is shown by the variables 'Voice & Accountability' and 'Political Stability' and the lower dispersion is represented by the variable 'Infrastructures'. The full set of the crosssection data is reported in Table I in the Appendix.    Note: ***, **, * indicates that coefficients are statistically significance at 1%, 5% and 10% level respectively.

ESTIMATION RESULTS
Homoskedasticity is not rejected, the GLS estimation results are also reported in column 2 for the sake of comparison. Both estimation approaches reveal similar results.
As it can be seen, the OLS estimation reveals satisfactory results, in particular: all population coefficients of the variables show high statistical significance at 1% level; the goodness of fit is reasonable showing evidence that 67% of the variation in the entrepreneurial activity is explained by the controlled variables; through the conventional White-test, Heteroskedasticity is rejected, therefore estimates are efficient; through the Chow test the model is stable; and finally the Reset-test shows that the model specification is appropriate.
Interpreting the marginal impacts of the covariates we can conclude the following: Every one point increase in the variable 'Infrastructure' (ranging from 1 to 7 points) is responsible for 24.0% decrease in TEA, everything else remaining constant. This evidence seems to be in line with the claim that less infrastructures in transports, energy and telecommunication leaves more space for developing entrepreneurial activities in these sectors.
With respect to the variable 'Voice & Accountability' (ranging from -2.5 to 2.5 points) the evidence shows that 1 point increase in this index is associated with 70.9% increase in TEA, everything else staying unchanged. This is an expected result, since more liberty in expression and more facilities in doing businesses are necessary conditions for developing more entrepreneurial activities.
The impact of the variable 'Government Spending' (ranging from 0 to 100) is estimated to be lower in magnitude. Assuming one point increase in this scale it is predicted that TEA increases by only 1.4%, everything else being unchanged. Having in mind that higher values of this scale indicate less state intervention, more space is left for the private sector to develop business activities.
'Political Stability and Absence of Violence' (ranging from -2.5 to +2.5 points) has a negative impact on TEA. It is estimated that one point increase in this scale is responsible for 44.1% decrease in entrepreneurial activities, everything else being constant. This evidence is interesting and in line with the claim that less political stability accompanied with violence create conditions of auto-defense, promoting therefore self-employment activities driven by necessity.
Model 2 of Table 5 shows the results of the relationship between TEA and income per capita using a log-log specification and a set of 53 countries. As has been explained this simple relation will help us to understand the nature of entrepreneurship whether is motivated by necessity (negative correlation) or by opportunity (positive correlation). In fact looking at the evidence we observe a negative relation between the two variables confirming therefore that entrepreneurship in this group of countries is necessity driven.
Countries with lower per capita income show higher propensity of developing business motivated by necessity and not by opportunity, and vice versa. More specifically, every one percent increase in GDP per capita is associated with 0.26% fall in TEA and this relationship is statistically significant at the 1% level. Table 5 (53 countries), shows that along with income per capita, the variable 'Government Spending' presents also statistical significance. It was not possible to find any statistical significance for the other explanatory variables due to colinearity problems. In this regression it is expected that one point increase in the scale of 'Government Spending' induces only 1% increase in entrepreneurial activities, everything else assumed constant. On the other hand, it is predicted that when income per capita increases by 1%, TEA decreases by 0.135% which is lower than the elasticity

DISCUSSION AND CONCLUSIONS
In this study we estimate cross-section models using two samples of countries: initially a set of 36 countries applied to the full model to identify the most relevant factors explaining entrepreneurial activities; and later a set of 53 countries to test the hypothesis of the convex relationship between entrepreneurial activity and income per capita.
In the regressions we run we found that the variables 'Infrastructure' and 'Political Stability' have a negative impact on entrepreneurial activities (TEA), and in a separate regression we found evidence of an inverse relationship between TEA and income per capita. Combining all these results we can assert the prevalence of entrepreneurial activities motivated by necessity rather than by opportunity.
In particular the negative impact of 'Infrastructures' on TEA is in line with the study of Fontenele (2010)  constitutes one of the main explanation that entrepreneurship is necessity driven.
Reflects the idea that, in developing countries, entrepreneurship is particularly important in light of the pressing needs of the populations and higher unemployment. Our evidence also supports the convex relationship between the two variables collaborating other studies by Acs et al, (1994) and Wennekers & Thurik (1999 Our evidence also reveal a positive impact of the variables 'Voice & Accountability' and 'Government Spending' on TEA collaborating the findings of Lecuna (2014) and Díaz-Casero et al. (2012) . This result suggests that greater freedom of citizens and the predominance of economic liberalism, with less State presence in the economy, induce higher entrepreneurial activity. The underlying idea of less government spending is that less State in the economy will bring more opportunities for the private sector and raise up the entrepreneurial activity (Bjørnskov & Foss, 2008).
The prevalence of entrepreneurship motivated by necessity is associated with selfemployment activities involving less skilled labor and causal work of short duration.
This kind of activities is also characterized by low value-added acting mostly in the nontradable sector. But most importantly this kind of activities are less innovative and lowtech users. On the contrary, entrepreneurship by opportunity is the kind of activity related to skilled labor, driven by innovation and technical progress improvements.