SECTORAL SHOCKS AND METROPOLITAN EMPLOYMENT GROWTH

Horvath and Verbrugge (1996) argue that when investigating the sources of aggregate fluctuations, it is important to use the highest frequency data available.Using monthly data for the U.S. economy they show that industry-specific shocks are more important in explaining fluctuations in industrial production than are common aggregate shocks.With the exception of Coulson (1999) studies that examine the issue at the sub-national level have used low frequency, spatially aggregated data.We examine the relative importance of national disturbances versus local industry shocks for employment fluctuations using monthly data on five metropolitan statistical areas (MSAs).Input-output tables are used to quantify the strength of interindustry linkages, which are then used to help identify a structural VAR model for each MSA.Within-MSA industry shocks are found to explain considerably more of the forecast-error variance in industry employment growth (87-94 percent) than do common national shocks to productivity and monetary policy, and the manufacturing, services, and government sectors make the largest individual contributions to local employment variance.We also find that the measured importance of national shocks for employment fluctuations increases as the level of spatial aggregation increases.