Aggregate Job Creation, Job Destruction and Job Turnover in the Irish Manufacturing Sector*

Using an annual employment survey data set we construct aggregate job flow rates for the Ir ish manufacturing sector for the period 1974 to 1994. We report the existence of simul­ taneous job creation and job destruction, inducing job turnover well above that necessary to accommodate net employment changes at the aggregate or even very refined sectoral level, over the entire sample period. This job turnover is caused by a large proportion of the total plant population making mostly persistent adjustments to its employment level. The properties of the aggregate job flows in the Irish manufacturing sector conform well to the stylised facts derived from studies of aggregate job flows in the manufacturing sectors of other developed countries. Moreover, these properties hold regardless of whether we restrict our sample to continuing plants or entering and exiting plants. I I N T R O D U C T I O N T he relat ively recent emergence of studies on job flows has unveiled the existence of a large amount of ongoing reshuff l ing of employment beyond t h a t necessary to accommodate net employment changes at the aggregate or even very refined sectoral level. The heterogeneity of f i r m level labour demand behaviour over the whole business cycle imp l i ed by th i s *Thanks are due to Mark Cassidy; Joseph McLoughlin; Aidan Meyler; and Marius Bruelhart for comments on earlier drafts. We are especially grateful to Forfas for their help. We also wish to thank Gerry Boyle for supervising the refereeing process. f inding, which lies i n contrast to t radi t ional representative firm models found i n s tandard textbooks, has k ind l ed much interest i n job flows among economists. Through the use of establishment level employment data sets, the characteristics and cyclical properties of aggregate job flows have now been documented for numerous countries; Contini and Revil l i (1987) for I ta ly ; Davis and Ha l t iwange r (1992) for the US; Boeri and Cramer (1992) for Germany; Ba ldwin , Dunne and Hal t iwanger (1994) for the US and Canada; Salvanes (1996) for Norway; Konings (1995) for the U K ; and Albsek and S0rensen (1995) for Denmark, to name a few. 1 These studies have confirmed the existence of large magni tudes of s imultaneous, though cycl ical ly asymmetric, aggregate job creation and job destruction internat ional ly . To date, however, there has been no parallel study for I re land . 2 The purpose of th is paper is to document the properties of the job flows of the I r i s h manufac tur ing sector. These are constructed by aggregating the number of jobs created and jobs destroyed from the plant level us ing an annual employment panel survey. Our data set is a par t icular ly attractive one i n tha t i t covers v i r t ua l ly a l l plants i n the manufacturing sector over the period 1974 to 1994 and, hence, allows us to undertake a study of aggregate job flows over a long t ime period while encountering few problems of sample selection bias. Where possible, we compare our results to those derived from studies of other countries. Such a comparison is of part icular interest given t h a t I r e l and has implemented a strong in tervent ionis t i ndus t r i a l policy, ma in ly v ia an extensive provision of grants, since the 1950s i n an effort to a t t ract foreign direct investment and promote indigenous indust ry growth. One migh t thus expect different employment behaviour by plants located i n I re land, and hence different aggregate job flow properties, relative to plants i n other developed countries. We proceed as follows. I n Section I I our data set is described. I n Section I I I we document the properties of the I r i sh aggregate job flows using a variety of indices and measures. The importance of plant turnover i n the aggregate job flows is investigated i n Section IV. We conclude i n Section V.