Labour economists have long studied net changes in employment but have only recently turned their attention to gross levels of job creation and destruction, which are also of interest for analysis of the business cycle and industrial organization. In Discussion Paper No. 912, David Blanchflower and Simon Burgess relate recent theoretical models of these processes to UK data on private and public sector manufacturing and services establishments during 1980-90. They find that many common beliefs about job creation and destruction and employment growth rates are false or misleading.

Blanchflower and Burgess find that the dispersion of employment growth rates was greater in manufacturing than in services. Less than half of jobs created (destroyed) arose in plants growing (declining) at more than 20% per annum. About half of all job creation and destruction were concentrated in just 4% of plants. Manufacturing accounted for around one-third of total job creation (roughly its share in total employment). Many manufacturing plants prospered during the 1980 recession, with 25% of survivors growing by 5% or more. Jobs were created and destroyed in groups of roughly equal size. Employment was only marginally more stable in large plants than small ones. Job turnover among plants 25 years old or more was far from negligible, at 9% in 1980 and 15.5% in 1984.

Blanchflower and Burgess maintain that neither conventional labour demand nor simple sector sectoral flows models can properly account for the diversity and concentration of employment growth rates, and selection and passive learning models account for only a small part of turnover. Dropping their common assumption of perfect competition may identify winners and losers, however, and the persistence of their success (or otherwise) will then depend on such models' structure. Models of the introduction and diffusion of innovations may account for the concentrations of employment gains and losses, while the channel through which aggregate macroeconomic shocks also affect the distribution of employment growth may shed light on the modelling of such aggregate labour market phenomena as unemployment.

Job Creation and Job Destruction in Britain: 1980-90
David G Blanchflower and Simon M Burgess

Discussion Paper No. 912, February 1994 (HR)