Postwar Period Changes in Employment Volatility: New Evidence From State/Industry Panel Data

Abstract

Many recent studies have identified a decline in the volatility of U.S. real output over the last half century.This study examines a less discussed and analyzed trend, but one as significant as the drop in output volatility, namely a substantial decline in employment volatility during the postwar period.Using a new panel data set covering industry employment by state since 1952 we find that a large decline in employment growth volatility began in the early 1950s and largely ended by the mid to late 1960s.This study also illuminates the geographical dimension of the declines, an aspect that has heretofore been unexamined.The data indicate that all states have shared in the volatility decline, although the magnitudes have differed.A pooled cross-section/time-series model indicates that fluctuations in statespecific (state level differences in demographic and industrial composition) and macro variables (e.g., changes in monetary policy regimes) have each played a potentially substantial role in explaining volatility trends.We find that state-specific forces account for between 1 percent and 24 percent of the variations in employment volatility across time and space.Macro variables account for between 30 percent and 76 percent of the movements in employment volatility, a range broadly consistent with the findings of Stock and Watson (2002).An important finding of this study is that "unknown forms of good luck," in the form of smaller shocks to employment, account for between 1 percent and 10 percent of the observed fluctuations.This latter finding suggests a reduced role for unknown forms of good luck in describing the postwar decline in volatility compared to the findings in Stock and Watson's (2002) analysis of the variance of real output growth.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 16 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.