Title: Regional beveridge curves
Authors: Michael Owyang - Federal Reserve Bank of St Louis (United States) [presenting]
Hannah Shell - Federal Reserve Bank of St Louis (United States)
Abstract: The Beveridge curve - the relationship connecting the unemployment rate with job vacancies - is typically estimated at the national level. Labor markets, however, are largely local level phenomena, with usually only a small fraction of workers migrating from place to place to seek employment. Moreover, local labor markets have been found to exhibit some heterogeneity, depending in part on things like industrial composition and the local minimum wage. We construct MSA-level unemployment and vacancy data using techniques similar to those used in previous papers for the national level. We then estimate MSA-level Beveridge curves for 51 U.S. cities. Previous estimates of local Beveridge curves used time fixed effects as identifying restrictions, imposing that the slope of the Beveridge curve was invariant across locations. Our method identifies the slope of the Beveridge curve by restricting it to be constant for locations with similar labor market characteristics but allows it to vary across labor markets with, say, different industrial compositions. In this sense, we form regions with similarly sloped Beveridge curves. We then study whether the advent of the internet and subsequent adoption of online job postings has reduced the heterogeneity in local labor markets, reducing the number of regions.