THE OECD JOBS STUDY

This paper reviews recent literature on job gains and job losses. Economies exhibit high rates of gross job reallocation - both high levels of job gains and job losses. For the OECD nations for which data are available, total turnover averaged more than twenty per cent during the 1980s. This is a result of differing behaviour of establishments (firms) in the face of similar general economic conditions.
Two streams have developed in the literature in attempting to characterize the influence of structural change on job turnover. The first sees structural turnover as continuous, and emphasizes the importance of establishment openings as the primary means through which more significant changes in an economy occur. An alternative view emphasizes the concentration of job losses stemming from structural change in cyclical downturns. The timing and pace of structural change can have an important influence on labour market policies.
The behaviour of the four components of job turnover over the cycle gives some indications about how cyclical and structural change are reflected in turnover data. Though limited to four countries, there is evidence that the rate of job gains stemming from the opening of establishments is related to the trend in employment growth, while closure of establishments is correlated with neither the trend nor the cycle. Expansion and contraction of existing establishments is the dominant element in the cyclical pattern of employment change.
The relative stability of openings in several countries during the late 1980s, combined with the expansion of existing establishments (and their reduced contraction) indicates the importance of cyclical conditions in explaining the increase in employment growth over this period. While further evidence is needed, the implication of these findings is that these employment gains were then more subject to loss when cyclical conditions changed than if the growth in employment had taken place through increased entry. These results also point to potential difficulties in trying to increase opening rates, given that they were apparently not heavily influenced by the robust cyclical conditions that prevailed during this period.