Measuring entrepreneurship (I)

This post is coauthored with Francisco Campos

A bat and a ball cost Rs. 1100 in total. The bat costs Rs. 1000 more than the ball. How much does the ball cost?    A culturally appropriate GRE? No, this question comes from the cognitive portion of a test designed to measure entrepreneurship in Sri Lanka. 

But why would we want to measure entrepreneurship? First, it’s going to lead to better programs.   And second, it’s going to lead to better evaluations.   In this post, and one coming next week, we’ll discuss this issue and some approaches to measuring entrepreneurship and what they might be trying to tell us.  

First, though, it’s worth considering why this might be different in developing countries.   In a recently published paper (ungated version here), De Mel, McKenzie and Woodruff argue that that there are four main differences between entrepreneurs in developing countries vs. developed countries.    First, risk matters more and innovation matters less.   The risk point is obvious.   They equate innovation with cutting edge technology creation and use – but their point here may be a bit weaker if a broader definition of innovation (e.g. creativity) were considered.   Second, developing country entrepreneurs face a bigger agency problem.   Given the state of legal systems and the lack of formal information systems, this makes the institutional environment tougher.   Third, managerial ability matters more in developing countries given the lower level of schooling the workers are likely to have and the agency issues firms face.   Finally, getting help is hard – the market for business services is thinner in developing countries.   

So even if we knew how to measure entrepreneurship really well in developed countries (which we don’t), we probably wouldn’t want to unquestionably use the tools from developed countries to design surveys or even programs in developing countries.  

Indeed, when we think about designing programs to boost entrepreneurship, these issues present a problem for both the suppliers and for the entrepreneurs who are possible clients.    On the client side, it is likely to be difficult for them to find a program in the first place and then to know if it any good (given the information issues).   The thinness in the market for business services is also likely to make entrepreneurs more skeptical of the potential quality of any program.   And finally, if entrepreneurs have to pay for the training, agency issues will constrain their ability to mobilize the cash, if they don’t have it on hand.  

On the supply side – let’s say the government -- these issues also make it hard to design a program – in terms of content, but also in terms of targeting.    One can imagine a government conversation going something like this:  

-          Should we target start-ups? they’re upcoming entrepreneurs with potential!

-          No, 80% of the start-ups die in the first 3 years of operation (we don’t want losers).

-          Should we provide the training to all sectors? Give everyone a chance?

-          No, just the priority sectors because we don’t want retail shops that won’t grow.

But other questions and other objectives could lead to a reverse set of conclusions.   At the end of the day, if the government is after growth, be it straight economic growth or employment growth, they will need some sense of how to measure entrepreneurship foremost to know who to target.  

Now the empirical evidence is pretty clear – we aren’t very good at this. After all, if we were we would be rich – and one doesn’t see a lot of rich empirical microeconomists or a lot of venture capitalists running surveys.   But, as we tour the literature, perhaps an argument can be made that we are making progress.   

One main strand of the measurement literature comes from psychology, where authors have tried to identify the main traits of an entrepreneur. A core start came from the identification of the Big Five broad category of personality traits, which was developed among others by Costa and McCrae (1987) (gated version here). Broadly, the Five-Factor model includes the following measures: (1) Openness to experience (specific traits include being artistic, curious, imaginative, insightful, original, and wide interests), (2) Conscientiousness (efficient, organized, planful, reliable, responsible, thorough), (3) Extraversion (active, assertive, energetic, enthusiastic, outgoing, talkative), (4) Agreeableness (appreciative, forgiving, generous, kind, sympathetic, trusting), (5) and Neuroticism (anxious, self pitying, tense, touchy, unstable, worrying).

These broad measures and from there specific ones have been used to identify the characteristics of entrepreneurs, particularly in comparison with managers or even other individuals. These studies often involve personality questionnaires to identify traits strongly correlated with business creation. For instance, in a meta-analysis, Rauch and Frese (ungated version here) find need for achievement, generalized self-efficacy, innovativeness, stress tolerance, need for autonomy, and proactive personality, as traits significantly correlated to entrepreneurial behavior of business creation and success.

The ways of measuring these traits will be part of our post for next week, where we will provide a few examples, notably from impact evaluation studies. We will also explain how initiatives such as the Entrepreneurial Finance Lab are helping governments, banks and other interested parties in identifying which entrepreneurs to target.