Trade and Agglomeration: Theory and Evidence from France

Abstract

Trade openness leads to aggregate welfare gains, but the local effects of trade vary across space. This paper shows that the welfare gains from trade are lower in smaller cities, due to weaker export-specific agglomeration. Using rich micro data from France, I show that firms’ export-to-sales ratio increases with city size, both within and across industries. I develop an open economy economic geography model with heterogeneous firms to rationalize these novel facts: firms jointly choose their location and export behavior in the presence of sectoral differences in factor intensity and external economies of scale in export costs. Within industries, more productive firms sort into larger cities and into exporting, endogenously benefitting from lower export costs. Across industries, more capital-intensive sectors are endogenously more export intensive and overrepresented in larger cities. To quantify the role of export-specific agglomeration forces, I structurally estimate the model: they can account for 1/3 of the differences in export intensity across locations. As a result, counterfactual trade liberalization induces 17% lower welfare gains in bottom size- compared to top size-quartile locations. These results shed new light on the distributional effects of trade openness and help explain the urban-rural divide over protectionist policies.