Flow of Firm Creation

The failure of the approach based solely on Gibrat’s principle stems, at least in part, from the fact that it attempts to derive the distribution of firm sizes directly from the distribution of the asset value of a single firm. Indeed, many models start with the implicit or explicit assumption that the set of firms under consideration was born at the same origin of time. This approach is equivalent to considering that the economy is made of only one firm. Therefore, the distribution of firm sizes can reach a steady-state if and only if the distribution of the asset value of a single firm reaches a steady state, which seems rather counterfactual. An alternative approach to model a stationary distribution of firm sizes is to account for the fact that firms do not appear at the same time but are born according to a more or less regular flow of newly created firms. For instance, Bonaccorsi Di Patti and Dell’Ariccia (2004) report a yearly average (over all industry branches) rate of birth for the period 1996–1998 equal to 5.6% for Italian firms with a maximum of 32% in some industry branches. Reynolds et al. (1994) give the regional average firm birth rates (annual firm births per 100 firms) of several advanced countries in different time periods: 10.4% (France; 1981–1991), 8.6% (Germany; 1986), 9.3% (Italy; 1987–1991), 14.3% (United Kingdom; 1980–1990), 15.7% (Sweden; 1985–1990), 6.9% (United States; 1986–1988). They also document a large variability from one industrial sector to another.