LABOR MARKET POLARIZATION AND THE GREAT URBAN DIVERGENCE

Labor market polarization is one of the most important features in recent decades of advanced country labor markets. Yet key spatial aspects of this phenomenon remain under-explored. Is the polarization present in both small and large cities? Do middle-paid jobs decline most sharply in cities where those jobs were initially most present? How do observations on polarization concord with facts stressed in the literature on the great urban divergence? Are large and small cities losing the same types of middle-paid jobs? We develop four key facts addressing these questions for the case of France 1994-2015. We argue that no leading existing theory accounts well for these facts. We then develop a parsimonious theory that builds on existing models of routinization and offshoring, as well as of heterogeneous labor in

a spatial equilibrium of cities. To these we add a structured approach to how the interaction of skill and technology differs across sectors, owing to individual- and city-level comparative advantage. The resulting theory does account for our four key facts.