There is a vast literature linking a country's endowment of ‘entrepreneurship’ with economic prosperity. Environments where entrepreneurs can emerge easily are propitious to the creation of firms, to their growth and their success and, with the latter, to the success of the economy as a whole. These ideas, dating back to Richard Cantillon (1755), who first developed the conception of the entrepreneur as a risk-bearer in a well-defined spatial economy, were subsequently developed by Adam Smith (1776), Alfred Marshall (1890) and Joseph Schumpeter (1911). For example, the latter sees the entrepreneur as the carrier of innovation and hence as the true engine of growth. But if entrepreneurship is so central to economic development, what drives it? Why are there so many entrepreneurs in some areas, such as Silicon Valley, and so few in others? Why do we find these clusters in some countries rather than others, and in particular areas within countries, such as in the Italian industrial districts or the German Ruhr? These important questions have often been at the forefront of policy debate and government intervention.
There are two broad classes of explanations for cross-location differences in entrepreneurial density. One possibility is that there are heterogeneous costs of entry, and the locations with lower costs of setting up a firm end up with more entrepreneurs and more firms because even relatively less talented individuals will find it profitable to start a business there. This is the approach implicitly followed by the large literature that focuses on factors – particularly financial – that keep the would-be entrepreneur from actually creating a new firm. The second possibility is that certain areas might have a larger supply of entrepreneurial talents than others. Guiso and Schivardi (2011) use a modified version of the Lucas (1978) occupational choice model with heterogeneous entrepreneurial ability to show that these two explanations have opposite implications for the correlation between entrepreneurial density and average entrepreneurial quality (measured, for example, by total factor productivity or profits) – if heterogeneity is due to entry costs, areas with more entrepreneurs should also have on average less capable entrepreneurs, as the ability of the marginal entrepreneur decreases with entry costs. On the other hand, if heterogeneity is due to the fact that certain areas draw entrepreneurs from a better talent pool, then we should see that entrepreneurial density and average ability are positively correlated.
Using data from Italy, a country with substantial geographical heterogeneity in entrepreneurial density, Guiso and Schivardi (2011) find clear support for the second mechanism – areas with more entrepreneurs are also characterised by a higher average entrepreneurial quality. Lowering entry costs, therefore, might have little effect on entrepreneurship in areas where entrepreneurial skills are scarce. Rather, one should focus on the accumulation of entrepreneurial skills.
The key question is then what can explain differences in the supply of entrepreneurial skills across locations. In a recent paper, we investigate whether selection into entrepreneurship and entrepreneurs' success are affected by learning opportunities (Guiso et al. 2015). While individuals can learn how to become an entrepreneur and how to be a successful one in a variety of ways (e.g. from parents, friends, schools, etc.) and at different stages of their life cycle, we look at one specific but potentially very important channel – learning from one's environment during formative years (adolescence). Arguably, for a young individual growing up in Silicon Valley it should be easier than elsewhere to learn how to set up and run a firm because the high concentration of entrepreneurial activities in the area provides many direct and indirect learning opportunities. We study whether these intuitive predictions receive empirical support. In particular, we test whether firm density in the location where individuals grow up affects the choice of becoming an entrepreneur and their subsequent performance as entrepreneurs.
We study these questions using a sample of Italian entrepreneurs and employees, for which we know both their current location (province) as well as the location in which they were living at 18. Using census data, we construct the measure of entrepreneurial density (number of firms over resident population) both today and at learning age and relate them both to the probability of becoming entrepreneur and to entrepreneurial performance, measured in terms of TFP and profits. Consistent with the learning model, we find that individuals who grew up in a location with a higher entrepreneurial density are indeed more likely to become entrepreneurs. Our finding holds while controlling for the firm density in the current location (reflecting thick-market externalities), for measures of current access to external finance in the local market where the firm is located and in the location at learning age, and for having parents who are entrepreneurs themselves. The effect is sizable – one standard deviation increase in entrepreneurial density at learning age increases the likelihood of becoming an entrepreneur by 1.5 percentage points, around 8% of the sample mean. When we look at variation in performance among entrepreneurs, we find that those who faced a higher firm density at learning age earn a higher income from their business. This finding is consistent with the idea that entrepreneurial abilities can be acquired and that acquisition occurs through exposure to a richer entrepreneurial environment. A one standard deviation increase in firm density at learning age results in an 8% higher income.
In our dataset there are two reasons why individuals living in a given province were exposed at learning age to different learning opportunities. First, they are of different age. Second, they grew up in different places. Hence, identification relies on two sources of variation:
-Differences over time in firm density for people of different ages currently living in the same province where they grew up (stayers); and
-Cross-province differences in firm density for people of the same age who grew up in a different province from the one in which they currently live (movers).
Figure 1 illustrates this source of variability for the largest Italian provinces. We show that focusing on the sample of movers addresses a series of endogeneity issues that can arise from the serial correlation in current entrepreneurial density and entrepreneurial density at learning age for entrepreneurs that did not move. In fact, our results are even stronger when focusing on the sample of movers.
Figure 1. Evolution of entrepreneurial density for the largest Italian provinces
The final question we address is which aspects of entrepreneurship are more prone to be learned. Modern literature on entrepreneurship argues that being an entrepreneur requires a variety of skills/traits (Lazear 2005). Classical theories of entrepreneurship stress the role of personal traits, in terms of the ability to innovate (Schumpeter 1911) and to bear uncertainty and risk (Cantillon 1755, Knight 1921, Kihlstrom and Laffont 1979). These features of entrepreneurship probably have an important innate component and it is still unclear to what extent they can be learned. On the other hand, managerial capabilities are more likely to be learned. We therefore test whether entrepreneurs who grew up in high firm-density provinces adopt better managerial practices, measured using the methodology pioneered by Bloom and Van Reenen (2010b), and develop traits that are traditionally associated with entrepreneurship. We find some evidence that entrepreneurs who grew up in high firm-density locations adopt better managerial practices, although the effect is not precisely estimated. On the other hand, we find no evidence that exposure to firms at learning age affects the traits that have been traditionally associated with entrepreneurship, such as risk aversion, aversion to ambiguity, self-confidence, and optimism. These traits are either learned early in life, possibly within the family, or are innate. Finally, we find no evidence that entrepreneurial density at learning age affects innovation capacity, suggesting that this key aspect of entrepreneurship is also less prone to learning.
Bloom, N and J Van Reenen (2010b), “Why do management practices differ across firms and countries?” Journal of Economic Perspectives, 24(1):203–24.
Cantillon, R (1755)  An essay on economic theory, Ludwig von Mises Institute, Auburn, Alabama.
Guiso, L and F Schivardi (2011), “What determines entrepreneurial clusters?” Journal of the European Economic Association, 9: 61–86.
Guiso L, L Pistaferri and F Schivardi (2015), “Learning entrepreneurship from other entrepreneurs?” CEPR Discussion Paper, DP10997.
Kihlstrom, R and J-J Laffont (1979), “A general equilibrium theory of firm formation based on risk aversion”, Journal of Political Economy, 87: 719–48.
Knight, F H (1921), Risk, uncertainty and profit, Houghton Miffin, New York.
Lazear, E P (2005), “Entrepreneurship”, Journal of Labor Economics, 23: 649–680.
Lucas, Jr., R E (1978), “On the size distribution of business firms”, Bell Journal of Economics, 2: 508–523.
Marshall, A (1890), Principles of economics, Macmillan, London.
Schumpeter, J (1911), The theory of economic development, Harvard University Press, Cambridge, MA.
Smith, A (1776), An inquiry into the nature and causes of the wealth of nations, W Strahan and T Cadell, London.