Mismatch and the Consequence of Job Loss

This paper asks whether mismatch—or a misalignment in the places where new jobs are created and old jobs are lost—can explain the heterogeneity in earnings losses following job displacement. I offer an empirical answer to this question by building on previous research that uses mass layoff notifications as a worker-level shock; longitudinal administrative data to compare the labor market outcomes of displaced workers with non-displaced workers; and three new data sets on vacant jobs. I first document that the level and persistence of earnings losses is strongly associated with workers’ displacement location: the earnings loss at the 90th percentile is about twice as large as in the 10th percentile of commuting zones. I then assess the extent to which the dispersion in earnings losses depend on the availability of relevant job openings. Using vacancy-unemployment ratios defined by local labor markets, I find that skill and geographic mismatch explain 10-30 percent of the average earnings loss, and more than half of the dispersion in earnings losses across commuting zones. To help interpret the results, I estimate how employment depends on mismatch. I find that full-time employment is five percentage points lower among mismatched workers. By comparison, I find that part-time employment decreases by a smaller amount. These findings suggest that mismatch reduces earnings due to lower job-finding rates, and through fewer hours worked. Moreover, the results imply that the unequal labor market outcomes could be reduced by policies that move jobs or workers across space.