Working Paper: NBER ID: w3229
Authors: Daniel S. Hamermesh
Abstract: This study examines what one can infer from aggregate time-series of employment under the assumption that adjustment at the micro level is discrete because of lumpy adjustment costs. The research uses various sets of quarterly and monthly data for the United States and imposes assumptions about how sectoral dispersion in output shocks affects adjustment through aggregation. I find no consistent evidence of any effect of sectoral shocks on the path of aggregate employment. I generate artificial aggregate time series from microeconomic processes in which firms adjust employment discretely. They produce the same inferences as the actual data. Standard methods of estimating equations describing the time path of aggregate employment yield inferences about differences in the size of adjustment costs that are incorrect and inconsistent with the true differences at the micro level. This simulation suggests that the large literature on employment dynamics based on industry or macro data cannot inform us about the size of adjustment costs, and that such data cannot yield useful information on variations in adjustment costs over time or among countries.
Keywords: No keywords provided
JEL Codes: No JEL codes provided
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| sectoral shocks (F41) | aggregate employment dynamics (J69) |
| methodology used (C80) | incorrect conclusions about micro-level adjustment costs (D21) |
| discrete employment adjustment processes (J63) | similar inferences as actual data (C20) |