Small Business Trends
By Scott Shane
There’s an industry of consultants offering advice to city, county and state policy makers on how to generate more entrepreneurs in their locales. Much of this advice is based on research that looks at what’s different about places that have more entrepreneurs from places that have fewer of them.
Many governments are following this advice without looking deeply enough at the research behind it. A brief glance shows four problems with using much of the analysis to develop public policy.
1. The evidence doesn’t always support the theory. Consider, for example, the argument made by economic development guru and consultant, Richard Florida, that increasing demographic diversity will increase entrepreneurial activity. Florida’s own research fails to show that places with more diversity have more entrepreneurs. His diversity index, which measures the share of the population that is gay, and his melting pot index, which captures the foreign born share of the population, have no statistical relationship to the rate of new business formation when other factors are also considered. The theory that increased demographic diversity will lead to more entrepreneurship isn’t supported by the evidence.
2. The causality is often backwards. Again consider Professor Florida’s research. It shows that places with more amenities tend to have more entrepreneurship. But the direction of causality of this relationship is unclear. Does having more amenities lead to more entrepreneurship or does more entrepreneurship result in the creation of more amenities? Because the former interpretation suggests a policy action – build amenities – while the latter does not, policy makers have assumed that more amenities leads to more entrepreneurship and have built amenities in the hope of attracting more entrepreneurs.
3. The studies often focus on only a few types of new businesses. Consider, for example, a new report by Edward Glaeser and William Kerr with the catchy title “What Makes a City Entrepreneurial?”
In their study, the authors defined “entrepreneurship” as the formation of new independent manufacturing establishments with employees. While that definition sounds reasonable, these businesses account for only one percent of all new companies.
Drawing conclusions from a one percent sample wouldn’t be a problem if the places that have a lot of new manufacturing employers also have lots of other types of new businesses. But they don’t. Glaeser and Kerr write, “Our entry metric has a 0.36 … correlation with the 2000 self-employment rates at the city … level.” That is, cities with the highest self-employment rates aren’t the ones that have lots of new independent manufacturing establishments with employees.
Recent research conducted by Larry Plummer for the Office of Advocacy of the U.S. Small Business Administration provides further evidence for why drawing conclusions about entrepreneurship on the basis of new business formation in manufacturing is problematic. Looking at the rate of new independent establishment formation from 1990 to 2006, Plummer found that the rate in manufacturing correlated only 0.33 with the overall rate. Moreover, it correlated only 0.16 with the rate in retail trade, 0.13 with the rate in high tech, and 0.06 with the rate in business services.
Moreover, the top places for new manufacturing business creation are not the same as the superstars for new firm formation in other economic sectors. Of the 20 counties with the highest rate of creation of manufacturing establishments from 1990 to 2006, only one, San Juan, Colorado, was in the top 20 counties for overall rate of new business formation.
4. The factors that stimulate one type of entrepreneurship often dampen others. Consider, for instance, the factors that enhance entrepreneurship in manufacturing and high tech. Plummer’s study showed that places with more college graduates had fewer manufacturing start-ups, but more new high tech businesses. Many would argue that we should increase the number of college graduates in a region, even if that comes at the expense of number of manufacturing businesses started, especially if the rise in the number of degree holders leads to a boost in high tech start-ups.
In short, academic research doesn’t provide strong evidence of specific policies that increase rates of entrepreneurship in a locale. Maybe government officials should take the money that they are using to implement the academics’ recommendations and give it to the entrepreneurs they are trying so hard to create.