NEW INSIGHTS ON THE SIZE DISTRIBUTION OF ITALIAN FIRMS BY GEOGRAPHICAL AREA

Abstract Using a large industrial panel, we show that the size distribution of Italian manufacturing firms can be well approximated by a two-parameter generalized Pareto distribution (GPD); the fitting is particularly good for medium-large companies. This evidence seems to hold using different proxies of size: number of employees, revenues, fixed assets, total costs and debts. The choice of different measures of size is justified by the desire of presenting robust results, but also by the idea that one single optimal measure of size probably does not exist, and it is thus better to combine different points of view. We believe that such an approach represents a clear novelty of our work. The presence of a GPD in the size distribution of Italian firms suggests interesting implications of industrial economics, as discussed in the paper. In more details, we find that, for the entire time window considered, the parameter of the GDP is persistently in the interval [0.5, 1], indicating fat tails - consistently with the so-called Pareto law of industrial dynamics, but also pointing out a certain stability in the shape of the size distribution of Italian firms. For what concerns growth rates, the Laplace distribution provides a good fit, ruling out Gibrat’s law of proportionate effects from the patterns of growth of Italian companies. In the paper we discuss the economic implications of such a finding.