Does Trust Matter For Entrepreuneurship : Evidence From A Cross-Section Of Countries

To the extent that trust is necessary to conduct informal sector business activities, its absence could possibly constrain entrepreneurial spirit and overall economic growth. This paper tests the hypothesis that differences in trust levels between countries explain the observed differences in entrepreneurial spirit amongst them. Analyzing a cross-section of 60 countries in 2010, our findings suggest that about half of the variation in entrepreneurial spirit across countries in the world is driven by trust considerations. This result is robust to regional clustering, to outliers and to alternative conditioning variables. The findings of the study suggest that while formal incentives to nurture entrepreneurship must be maintained, policy-makers should also pay attention to the role of trust cultivated through informal networks.


Introduction
Trust has recently received increasing attention in the economic development literature.
Scholars have paid a particular attention to two broad dimensions of trust namely, its causes (see notably, Alesina and La Ferrara, 2002;Bjornskov, 2006;and Smith, 2008) and its impact.
The present study follows the latter broad approach to the subject of trust by investigating its impact on entrepreneurship. Despite the existence of numerous theoretical foundations linking trust to entrepreneurial spirit, no prior empirical study in the literature to date, to the best of our knowledge, has explicitly tested this relationship. According to Harper (2003), trust is crucial to cultivating entrepreneurship. As business transactions rely on trust: where there is trust, businesses generally thrive. Additionally, most prior studies linking trust to entrepreneurship have utilized a microeconomic or management framework. This paper analyzes this linkage from a macroeconomic perspective. Indeed, nothing genuinely to confirm the results found so far can explain the differences at the country level for entrepreneurship. The specific and sectorial contexts of each country but also micro data effectively validate this argument.
From a macroeconomic viewpoint, the absence of trust need not necessarily constrain business activity as long as formal institutions that bridge the trust gap exist. However, the absence of such formal institutions in many countries highlights the crucial role of trust in nurturing entrepreneurial spirit. Fafchamps (2002) emphasizes this thesis for Sub-Saharan African countries. While Berggren and Jordahl (2006) emphasize the link between social capital and economic freedom, Hafer and Jones (2012) instead emphasizes the connection between economic freedom and entrepreneurship. The present paper goes beyond both preceding views by directly linking entrepreneurship to trust.
The research question this paper seeks to answer is whether trust is necessary for entrepreneurial activity to flourish? An empirical answer to this question would offer great insight into why some countries have superior entrepreneurial culture than others. Also, to the extent that the literature (e.g. Holcombe, 1998 ;Caree & Thurik 2003;Audretsch, et al. 2006;Kirzner, 1997;andLazear, 2004 &2005) attributes a great role of entrepreneurship in economic development, understanding what drives entrepreneurship is helpful not only in understanding why some countries have superior entrepreneurship culture but also, why some countries are more developed than others. The paper performs cross-sectional analysis on 60 countries for the year2010. The data for entrepreneurship is from the recently published Global Entrepreneurship Monitor (GEM) by Acs & Szerb (2010). Following the literature, we use the trust variable from the World Values Survey which measures the extent to which people trust each other. These two measures also constitute the novelty of this paper. Indeed, we take a broader view of entrepreneurship than most of the prior studies. Further, our variable has a wider acceptance and is more comparable at the national level.
Furthermore, another major innovation of this paper is the use of the cross-sectional approach at the national level. Despite several studies conducted on the relationship between social capital and small/medium businesses; few have employed this technique as confirmed by Geindre et Dussuc (2012: 12) We employed a 2SLS methodology using suitable instruments to control for endogeneity of the entrepreneurship variable. The findings does suggest that trust has a strong positive impact on entrepreneurship and the result is robust to the presence of outliers, continental/regional clustering and alternative conditioning variables. Quantitatively, our results do suggest that about half of the variability in entrepreneurial spirit across the world is driven by trust considerations. Further empirical analysis suggest that causality runs from trust to entrepreneurship.
The rest of the paper is organized as follows: section two discusses the conceptual framework of the paper ; while section three discusses the data. Section four presents the methodology, while section five presents and discusses the empirical results. Section six concludes.

Conceptual Framework
The role of trust in entrepreneurship is not novel. As we havealready noted, trust is essential to entrepreneurship. A number of authors, notably, Chabaud & Ngijol, (2005), Bhagavatula et al. (2010), Audretsch et. al. (2011 have shown that by creating room for new opportunities, trust enables entrepreneurship. Some other authors have shown that trust facilitates the creation of enterprises, (e.g. Mueller, (2006), Davidsson & Honig, (2003), Rodríguez & Santos, (2007), Clarke &Chandra, (2011) andDeakins et al. (2007)) while others suggest that trust enables access to other resources (e.g. Baron & Markmann, (2003), Runyan et al. (2006), Honig et. al, (2006 and Packalen, (2007)). Trust has also been found to be indispensable in the birth of new enterprises (e.g. Geindre (2009) and Aarstad et. al. (2010)) and also for the growth and development of small and medium-sized enterprises (e.g. Bosma et. al. (2004), Mosek et. al. (2007), Han, (2007) and Coviello& Cox, (2007). While certainly informative and relevant, most of these prior studies linking trust to entrepreneurship have utilized a microeconomic or management framework. There is therefore need for a macroeconomic perspective to the subject, which is the object of the present study.
However, most prior empirical studies have not discussed this relationship. Some studies have empirically investigated the importance of confidence in entrepreneurship, through its effect on innovation (e.g. Akçomak and terWeel, 2009;Doh and Acs, 2010). Because trust affects innovation, we can logically think that trust would be useful in entrepreneurship because it is the foundation of innovation.
Wede fine a production function 2 of entrepreneurship as follows: Where Q represents entrepreneurship, T, trust, Z traditional determinants and , unobservable factors influencing Q. measures the externalities related to social capital. If > 0, we deduce that > 0, social capital thus increases entrepreneurship. If = 0, entrepreneurship is orthogonal to T, and if < 0, distrust reduces the spirit of entrepreneurship.
Simply put, we can settle for0 < < 1, where the average increase in social capital has a positive impact on entrepreneurship.
There are many reasons to consider a relationship between these two variables, such as the work cited above suggests. Akçomak and ter Weel (2009) believe it is easier to finance entrepreneurial activity based simply on the reputation of a company or a person. This in turn can help to create a business or to support a business. Starting a business is not always an individual process: the idea could flourish or die depending on the level of trust and the surrounding business environment. Retention of information is difficult in the presence of a higher confidence, which minimizes the asymmetry of information, further lowering the cost of information and transactions. One can easily get leads for his company, so we can develop entrepreneurial attitudes exchanging with entrepreneurs from various backgrounds. It is also understandable that confidence stimulates the ambition to become an entrepreneur or own these same attitudes.

Data and Descriptive Findings
Following the tradition in the literature, we use the World Values Survey (WVS) trust indicator which surveys the proportion of a population that answers "yes" to the fundamental question: ''in general, do you think that most people can be trusted, or can't you be too careful?'' WVS data for a number of countries has been available since 1981 and is generally accepted as a reliable indicator of trust at the aggregate level. National social trust scores have proved it to be a fairly valid measure of honesty, trust, and trustworthiness. Further, this indicator has been widely utilized in previous works.
Data for the aggregate national entrepreneurship activity is obtained from the Global from multinational organizations such as the UNIDO or OECD. While we tried to find a single institutional variable for each of the individual variables, sometimes it proved to be not executable. Therefore some of these institutional variables are themselves complex "indexes".
Comparing to the previous versions of our index, we avoided the duplication or the multiplication of the same institutional factors in different part of the index." Table 1 describes the sources of different variables included in this study. Idem The empirical analysis of the data follows two stepsthe summary descriptive statistics and then the analysis of partial correlations. Table 2 presents the summary descriptive statistics of the variables used in this study. It follows from the analysis of individual country statistics for the two key variables of interest, namely, entrepreneurial spirit and trust, that Uganda received the lowest score for the entrepreneurship variable, while Denmark received the highest one. The mean score position was earned by Japan and the coefficient of variation of 46.15 suggests great heterogeneity in entrepreneurial spirit amongst the countries included in the study. Regarding the trust variable, Sweden received the highest score, Russia was at the mean score position while Brazil received the lowest one. Again the coefficient of variation of 51.18 suggests great heterogeneity in trust amongst countries.  Figure 1 presents the scatter plot between Entrepreneurship (y-axis) and Trust (x-axis), and sub-indexes of entrepreneurship and trust for the countries included in our sample. The evidence clearly suggests a positive relationship between these two variables. This positive relationship is further confirmed in Table 3  Considering that entrepreneurial spirit (and sub-indexes) is a function of many different factors, these correlation figures must not be taken seriously unless further examination of the partial correlation of these other variables with entrepreneurial spirit on the one hand, and with trust on the other hand, is undertaken. This is the objective of Table 3. As expected, the evidence in Table 3 suggests that entrepreneurship is strongly correlated with many other variables, such as, economic freedoms, human capital and regulatory quality. Hence, the relationship presented in Figure 1 might change or weaken in strength once these other variables are taken into account.

Empirical Model
The question we seek to answer in this study is whether differences in trust levels between countries can explain observed differences in entrepreneurial spirit amongst these countries?
We specify a regression model of the form (following Hafer and Jones (2012): Where for Entrepreneurship, we use GEDI as the main indicator. Subsequently, we will use  (2012) recently show that entrepreneurship spirit is a positive function of the level of IQ. Thus, human capital being an important driver of entrepreneurial activity, we control for this by including the Intelligence Quotient (IQ) 4 and the average of years of schooling. We measure intelligence using the IQ data by Lynn and Meisenberg (2010), which has also been used by Jones and Schneider (2010) (2012), we include a measure of economic freedom to control for the influence of economic freedom on entrepreneurship. Finally, we include regional dummies to take account of the specificities of different regions of the world.
We perform our analysis on the empirical model specified in equation (1) above using essentially ordinary least square (OLS) regression model. To correct for likely heteroskedasticity, we present white-corrected standard errors.
Reverse causality is a concern in this study. Indeed, trust is a variable that is not entirely endogenous. Trust at the national level may be affected by entrepreneurship. Some believe that in extreme cases, entrepreneurship encourages greed, which in turn creates mistrust, through the exploitation that it generates. This is essentially the view of Marxist theories. If so, then variations in trust between citizens at the national level is driven by the spirit of enterpriseresulting in reverse causality. Further, in our case, we can think, for example, of variables such as tax rates, and labor forces participation that have been omitted resulting in omitted variable bias. Two Stage Least Squares (2SLS) is employed to correct for the fact that we cannot control for all the possible sources of endogeneity in the association between trust and entrepreneurship. This technique requires the instruments to be correlated with confidence, but not correlated with entrepreneurship. The instruments use dare those that have been used in Bjørnskov (2010Bjørnskov ( ,2012, namely, monarchy, the pronoun drop and the average temperature. The author justifies theoretically these instruments as follows: "I firstly include a dummy for whether countries are monarchies, which Bjørnskov (2007) Kashima and Kashima (1998) in arguing that cultures in which the language forbids dropping the personal pronoun traditionally have been more respectful of individual rights and have therefore developed stronger trust norms." (Bjørnskov, 2012:6). "These are supplemented by the average temperature in the coldest month of the year, based on the premise, dating back to Aristotle, that trust and social cohesion historically has been relatively more important for survival in regions with cold winters, and that cultures of such regions may have selected high-trust institutions through an evolutionary process." (Bjørnskov, 2010:336) Beyond the use of these dedicated instruments; we use standard statistical approaches to validate them. On the one hand, we use the Sargan and Hausmann tests for over identifying, the results accompany each estimate and secondly we look at the behavior of these instruments in the first stage regressions. These regressions are presented in the following  table  are naturallyinsertedintotheseestimates.Wereproducethecoefficientsofinstrumentalvariables. We observe the significance of all the instruments, except temperature. We therefore proceed with these results to estimate the 2SLSwith these instruments.
To further test the robustness of our results, and consistent with the approach by Bjørnskov (2010); we consider the influence of outliers. The approach is to eliminate outliers using both the Student test and the Iteratively Weighted Least Squares (IWLS) techniques. These two latest techniques therefore permit to verify whether the results found are not driven by the presence of outliers. As further test of robustness, we use regional clusters to account for regional heterogeneity and also use alternative conditioning variables.

Main regression results
The main regression results are presented in Table 5 The results in Table 5 thus confirms the strong explanatory power of trust on entrepreneurship. In particular, that trust explains about 50 percent of the variation in entrepreneurial spirit in the sample of countries considered. Other determinants found to have an important impact on entrepreneurship include, former communist background, economic freedoms, and human capital. While a former communist background was found to negatively affect entrepreneurship, economic freedom and human capital (measured by the average years of schooling) instead has a strong positive impact. The statistical significance of the former communist background variable is however unstable and changes with the introduction of controls for regional specificities. The likely intuition for this could be that former communist countries that fail to undertake institutional reforms to favor entrepreneurship are likely going to continue witnessing the detrimental effects of communism whereas those countries that reform their institutions to make them conducive to entrepreneurship are less likely to suffer the negative effects.
Income inequality measured by the Gini coefficient, has a negative but statistically insignificant effect on entrepreneurship while the level of development of a country, as well as all the regional dummies are statistically insignificant. If anything, the lack of statistical significance of the sub-Saharan African dummy suggests that entrepreneurial weakness is not We test for the robustness of our main results in the next section (section 4.1.2).

1. 2 Robustness Checks
We conduct two forms of robustness checks namely, continental clustering outlier observations and endogeneity (Table 6) and using alternative conditioning variables, outlier observations and endogeneity (Table 7). It makes sense to perform a continental clustering considering the extent of heterogeneity observed in both variablestrust and entrepreneurshipacross countries included in our sample. We would have wished to use an alternative variable for trust in our robustness checks but the non-availability of suitable proxies constrained this option. We were thus left with the sole option of using alternative conditioning variables, which is the approach that has been used in some studies, see notably, Potrafke (2011). There are a number of differences between the conditioning variables in our main results (Table 5) and Table 7. First, instead of Economic freedom used in Table 5, we use regulatory quality in Table 6. Also, instead of Education 1 (average years of schooling in population aged 25 and above) used in Table 5, we use instead Education 2 (average years of schooling in population aged 15 and above) in Table 6. We also use a dummy variable for GDP per capita instead of high income countries. Finally, we use dummies for regional classification of countries instead of continents. Of course, the decision to use alternative proxies for Economic freedom and human capital is justified by the fact that both variables were significant in our main regression. As we have already explained, the ideal robustness check would involve using alternative proxies for the principal explanatory variable (trust) but data constrains limited this option. We were thus left with the option of using alternative proxies for the chief conditioning variables, hoping to minimize bias in our results that would have been brought about by measurement errors in our conditioning variables. which is a dual means of controlling for likely simultaneity and for the presence of outliers.
The results confirm a positive statistically significant impact of trust on entrepreneurship.
Using alternative conditioning variables and controlling for the presence of outliers and endogeneity in Table 7 basically upholds our main result: that trust matters for entrepreneurial spirit. As observed in Tables 6 and 7, both specifications do not fundamentally change our main finding: that trust matters for entrepreneurial spirit.

Regression Results with sub-indexes
Regressions in Table 8 relate to the components of GEDI. We consider the same control variables as in the previous tables. Estimates are made in two stages. Initial estimates using the explanatory variables in Table 5. The explanatory variables in Table 7 are used in the second part of the Table 8. For entrepreneurial attitudes, trust is statistically related to this variable. This relationship is not a correlation. We are in the presence of causality. There is a strong correlation between entrepreneurial activity and trust. However, the single linear relationship observed between the two variables entrepreneurial aspiration and trust in Figure 1 disappears completely. We thus find a similar result by changing certain variables by other measures and estimation techniques.

Conclusion
This paper sought to investigate whether differences in trust levels between countries can explain differences in entrepreneurial spirit amongst them. We employed a 2SLS methodology using suitable instruments to control for endogeneity of the entrepreneurship.
The findings does suggest that trust has a strong positive impact on entrepreneurship and the result is robust to the presence of outliers, continental/regional clustering and alternative conditioning variables. Quantitatively, our results do suggest that about half of the variability in entrepreneurial spirit across the world is driven by trust considerations.
The implications of this findings are that, to spur entrepreneurial spirit, countries that lack formal trust-building institutions, for instance, sub-Saharan African countries, can benefit from creating conditions that favor the expansion of informal networks where trust is cultivated.
Our study has been limited by a number of factor amongst which are sample size, unique time period of study, the lack of a rigorous treatment of endogeneity issues and an exploration of the transmission mechanisms between trust and entrepreneurship. Further studies should consider probing deep into these important issues.