PLANT‐LEVEL ANALYSIS USING THE ARD: ANOTHER LOOK AT GIBRAT'S LAW

This study uses the Annual Respondents Database to test whether Gibrat's law of proportionate effect holds for the UK manufacturing industry during the period 1973–1998. For a sample of manufacturing industries, four different panel unit root tests were carried out to test the relationship between growth and size for three measures of size: gross output, employment and gross value added. The tests were applied to unbalanced plant‐level panel data. The results show that there is strong evidence to reject Gibrat's law that firm growth is a random process in favour of the alternative proposition of mean reversion.