Abstract.
In this paper we analyze the relationship between the distribution of firm size and stochastic processes of growth. Three main models have been suggested by Gibrat (1931), Kalecki (1945) and Champernowne (1973). The first two lead to lognormal distribution and the last to Pareto distribution. We fitted lognormal and Pareto distribution to two Italian sectors: ICT and mechanical. For ICT we found that lognormal distribution must be rejected and Pareto fits reasonably well to the last 30% of largest companies. For mechanical sector we can not reject lognormal distribution. Furthermore, we perform some experiments to corroborate the theoretical models. By means of transition matrices we found that ICT shows features very close to Gibrat’s and Champernowne’s models, while Kalecki’s model strongly fits to mechanical.
Similar content being viewed by others
Author information
Authors and Affiliations
Corresponding author
Additional information
JEL Classification:
L00, L25, D21
Correspondence to: Luigi Grossi
This research was partially supported by grants from Ministero dell’Istruzione, dell’Universitá e della Ricerca (MIUR). Despite being the results of a joint work, Sects. 1, 4, 8 and 10 should be attributed to Ganugi, Sects. 3, 6, and 7 to Grossi and Sects. 2, 5, and 9 to Crosato.
Rights and permissions
About this article
Cite this article
Ganugi, P., Grossi, L. & Crosato, L. Firm size distributions and stochastic growth models: a comparison between ICT and Mechanical Italian Companies. Statistical Methods & Applications 12, 391–414 (2004). https://doi.org/10.1007/s10260-003-0073-z
Issue Date:
DOI: https://doi.org/10.1007/s10260-003-0073-z