The static firm size distributions that we observe in practice are the cumulated result of underlying firm dynamics involving entry of new firms and growth, decline, and exits of incumbent firms. Our study gives an overview of firm size distributions that result as steady states from models differing in the way these firm dynamics are modelled. Some noteworthy results are:

We arrive at new functional forms to describe firm size distributions, for example the Waring distribution (a generalisation of the well-known Yule distribution) and a particular variation of a generalised hypergeometric distribution (yet unnamed).
Insight is given in the interrelationships between the distributions in terms of underlying firm dynamics. For example, it is shown how the Yule distribution proposed by Ijiri and Simon is related to the geometric distribution proposed by Sutton as the least unequal firm size distribution that we are likely to encounter in practice.
Of all reviewed distributions, possible interpretations of their parameters are given in terms of the underlying firm dynamics, e.g. the entry rate, the extent in which Gibrat's law is violated, et cetera.
It appears that all major features of distributions encountered in practice - distributions with a Pareto shaped right tail with a parameter near 1, lognormally shaped distributions, distributions that are concave to the origin - can be explained satisfactorily by the steady-state approach.

The paper gives insight in the precise relationship between firm dynamics and the resulting firm size distribution. Hence, differences in firm size distributions across industries and over time that we encounter in practice can be explained by differences in the underlying firm dynamics. Furthermore, determinants of firm dynamics can be translated into determinant of firm size distributions and vice versa.

Download a PDF version of this report