Ed Glaeser has a very interesting post and an accompanying working paper on differences in entrepreneurship across cities. His post covers some history of entrepreneurship in economics (he mentions Smith, Marshall, Schumpeter, Knight, and Chinitz, but not Cantillon), how to measure entrepreneurship, and some preliminary results from their working paper:
The big fact about entrepreneurship and cities is that average firm size strongly predicts urban success.
The chart below shows that a 10 percent increase in the number of firms per worker in 1977 is associated with a 9 percent increase in employment growth between 1977 and 2000. An abundance of small, independent firms is, along with January temperature and share of the population with college degrees, one of the best predictors of urban growth. [chart omitted] …
Is this relationship largely spurious, the result of some omitted variable — or set of variables — that increases both the number of firms and city growth? If the relationship is real and an abundance of smaller firms actually causes urban success, then why are some places more entrepreneurial than others?
The analysis looks across industries within cities and finds what seems like largely a life cycle result — smaller firms are associated with faster growth rates in new firms than in established firms. But by doing it across industries and controlling for industry and for firm age, the faster growth rates for new firms are arising from some other process than the traditional industry life cycle. Very interesting.