Stochastic models of the growth of firms
Date
1993
Authors
Mills, Peter Gordon
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The purpose of this thesis is investigate the growth rates of large firms; specifically to test whether a stochastic model is able to adequately explain firm growth.
The objectives of the thesis are to specify five functional forms of the stochastic model, suitable for testing, and test whether those models can explain the growth rates observed in recent years within a large sample of Canadian firms.
Two static models are specified in which, at equilibrium, the distribution of individual firm sizes around the mean firm size in the sample would be Pareto and log-normal distributions, resulting from a stochastic growth process. A transitional model, which predicts the growth rates for firms from one period to the next before the stochastic process of growth reaches equilibrium, is also specified. The transitional model is subsequently modified to allow for and test any effects on the growth process due to either initial size or
persistent growth.
Data was assembled on 735 large Canadian firms, covering the period 1986 to 1988, inclusive. The sample was divided into eight sub-samples, each consisting of those firms that were engaged in the same broadly defined area of commercial activity.
None of the five forms specified were able to fully explain the firm sizes or rates of growth in the sample data. Tests of the residuals indicated the presence of heteroskedasticity when the models were tested by Ordi nary Least Squares regression, suggesting that a stochastic model may be a mis-specification of the growth process.
Description
Keywords
UN SDG 9: Industry, Innovation, and Infrastructure