The Cyclical Behavior of Short-term and Long-term Job Flows

Using a band pass filter, this paper estimates plant-level job flows at different frequencies and examines the characteristics of the high frequency (transitory) and low frequency (permanent) component flows. Because high frequency employment movements, which likely result in changes in the utilization of plant assets, and low frequency movements, which likely coincide with the restructuring of plant assets, result in different costs to the economy, understanding their separate behavior is important. High frequency plant-level employment fluctuations account for the majority of cyclical movements in aggregate manufacturing employment, but the temporal separation between job destruction and job creation is more pronounced for low frequency job flows, suggesting that permanent job flows reflect a more protracted employment adjustment process. To facilitate the evaluation of job flow models, many of which describe either transitory or permanent job flows, time series of job flows at different frequencies are presented in the appendix. Email: [email protected]. Research for this paper was conducted at the Census Bureau’s Center for Economic Studies and supported by a research fellowship from the Brookings Institution. The views in this paper are those of the author and not necessarily those of the U.S. Census Bureau, the Federal Reserve Board, or its staff. I would like to thank John Shea, Judy Hellerstein, John Chao, Plutarchos Sakellaris, Bruce Fallick, Darrell Cohen, Bill Wascher and Stephanie Aaronson. I would especially like to thank John Haltiwanger for supplying me with data, encouragement and much useful advice. -1Much of how one looks at aggregate employment and unemployment fluctuations is determined by whether micro-level (e.g. plant, worker) employment or unemployment flows are long-lasting or not. While there has been substantial research on short-term and long-term worker flows, as of yet there has been little work trying to distinguish short-term from long-term job flows. This is unfortunate because theoretically the two types of job flows are quite distinct, and models have been constructed to describe each type. But without data on the separate flows, it is difficult to evaluate the relevance of these models. This paper presents time series of shortterm and long-term job flows for the manufacturing sector, which should aid the evaluation of job flow models, while, at the same time, complementing previous studies on short-term and long-term worker flows. The different time series reveal that fluctuations in aggregate employment are driven predominantly by high frequency plant-level job flows (those with a period of between one and four years), but that the temporal separation of job flows is more pronounced for low frequency job flows (those with a period greater than 8 years). This suggests that high frequency plant-level job flows are more important for understanding cyclical fluctuations in employment, but that low frequency job flows have more lasting effects and reflect a more protracted adjustment process. Studies on short-term and long-term worker flows have looked at flows into and out of employment and flows into and out of unemployment. Using data on temporary layoffs from a variety of sources, Feldstein (1975) emphasized the importance of temporary layoffs to fluctuations in aggregate employment in an attempt to focus policy measures on the adverse incentives of the Unemployment Insurance system. Examining CPS data on worker flows into and out of unemployment, Clark and Summers (1979) emphasized the importance of the longterm unemployed in an attempt to debunk claims that unemployment was primarily frictional and not unduly costly. Murphy and Topel (1987) also used CPS data to point to the importance of long-term unemployment in accounting for the rise in unemployment from the early 1970s to the late 1980s. In contrast to the large empirical literature on permanent and temporary worker flows, empirical work on permanent and temporary job flows has been scant. Permanent and temporary job flows are distinct from permanent and temporary worker flows because permanent worker separations do not imply permanent job loss. Instead, a plant may lay-off a