THE OECD JOBS STUDY: EVIDENCE AND EXPLANATIONS

New evidence for the manufacturing sector in the U.S. indicates that the rate of employment gain through establishment openings is no higher than that of Canada and of a sample of Nordic countries. Turnover largely reflects movements within industries rather than across industries. More basic factors, such as the technology base of an industry are the principal determinants of turnover [Baldwin, Dunne and Haltiwanger (1993)]. Openness to trade has an impact on turnover in some countries.
There is also evidence that structural change is concentrated in cyclical downturns; however, data limitations do not allow this hypothesis to be fully tested, although turnover did increase in several OECD countries during the most recent recession. There may be some potential to smooth the cyclical swings of job loss and avoid contemporaneous structural and cyclical change [Caballero and Hammour (1991)]. This largely falls to macro-economic policy, which by smoothing cycles may reduce the cost of closure of older facilities and new investments during upswings which is increased by the unpredictability of demand.
There may be asymmetry in the process of structural change. More extensive allocative shocks, which shift production and employment from one industry to another will produce immediate increases in job losses, perhaps concentrated in cyclical downturns and only gradual increases in job creation through entry of new establishments, perhaps in other sectors [Davis and Haltiwanger (1990)]. The evidence presented in this report is not inconsistent with this view; however, more evidence is needed. This asymmetry in the process of structural change could have implications for unemployment. Jobs lost due to structural change during downturns may not be replaced during a cyclical recovery which would, strictly speaking, restore jobs lost for cyclical reasons. New jobs would eventually emerge through the gradual opening of new establishments, but this could result in a considerable unemployment gap, especially given the high failure rate of new establishments. From a policy perspective, it is important to distinguish this phenomenon from that whereby cyclically unemployed individuals may become structurally.