Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.584166
Title: Exploration of the stochastic approach to modelling industrial dynamics and size distribution
Author: Paull, N. M.
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 2007
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
Under the stochastic process known as Gibrat's Law firm growth follows a random walk with no persistence in growth rates and both firm growth rates and the variance of growth rates independent of firm size. Over time, for a fixed population of firms, such a stochastic process gives rise to a lognormal limiting distribution of firm sizes. Utilising an official U.K. regional database of manufacturing plants, the empirical validity of the propositions of Gibrat's Law and the lognormal distribution are examined across three sequential U.K. economic cycles, making use of a range of econometric and statistical methods to assess the sensitivity of the results. Formal statistical tests and non-parametric kernel density estimation methods provide evidence of generally stable aggregate plant size distributions, though frequency exhibiting deviations in the form of positive skewness and leptokurtosis compared to the lognormal distribution. Examination of age and selection effects indicates some right-shift of the plant size distributions with time, though not generally in the form of a simple or consistent evolution towards the lognormal distribution. Plant growth rate distributions are found to often exhibit highly peaked distributions with "fatter tails" than the normal or Laplace distributions. Considerable heterogeneity is observed in both plant size and growth rate distributions across industries. Using a number of econometric model specifications the proposition of equi-proportional plant growth is statistically rejected for aggregate samples with mean reversion observed, a finding robust to corrections for heteroscedasticity and sample selection bias. However, equi-proportional plant growth cannot be rejected for a number of sub-groups. Consistent with previous research, plant growth is found to decline with age. Statistical tests reject the proposition of the variance of plant growth rates being independent of plant size, with a consistent negative relationship identified. Regression approaches suggest the absence of significant plant growth rate persistence effects.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.584166  DOI: Not available
Share: