The Origins of Regional Specialization

Why do US states specialize in different sectors? We document that employment specialization is highly persistent, which suggests that specialization may deviate from natural advantage, and reallocation could increase aggregate output. We develop a quantitative spatial model in which workers move across state-by-sector labor markets in response to exogenous changes in local fundamentals and endogenous agglomeration, as well as mobility frictions and labor market-specific idiosyncratic skills. We quantify the model with historical Census microdata that tracks workers' joint regional and sectoral mobility, which yield novel estimates of regional and sectoral mobility frictions as well as new evidence that workers carry both state- and sector-specific skills. Mobility frictions play a substantial role in the persistence of specialization. Although spatial differences in productivity within a sector are small, the ability of individual workers to reallocate across labor markets is valuable. Migration costs are the main barrier to workers' reallocation; spatial reallocation facilitates better matching not only across states but also across sectors.