On the Contribution of Agglomeration Economies to the Spatial Concentration

1Many thanks to Shannon Mail and Jim DiSalvo for timely research assistance and to Albert Saiz and Jordan Rappaport for several thoughtful comments. Comments from seminar participants at FRB Richmond and Kansas City, University of Texas, and ITAM are also gratefully acknowledged. The views expressed in this paper are those of the author and do not necessarily represent those of the Federal Reserve Bank of Philadelphia or of the Federal Reserve System. Why does the level of economic activity vary so much across space? One reason given is “agglomeration economies, ” meaning that a firm’s or household’s production costs (of market and home goods, respectively) are lower when production is car-ried out in close proximity to other firms and households. In this paper I explore, via a quantitative spatial macroeconomic model, the contribution of agglomeration economies to the observed spatial concentration of US employment. The approach is analogous to “business-cycle accounting ” or “growth accounting. ” The results of the “spatial accounting ” performed in this study depend on the details of the model used. The critical detail pertains to how the model rationalizes the stability of low-density localities. If it is rationalized via an appeal to restrictions on labor mobility, the accounting implies that the bulk of spatial concentration results from an unequal distribution of natural advantages. In contrast, if it is rationalized via an agglom-eration threshold (an employment level below which local increasing returns do not operate), the accounting implies that the bulk of the spatial concentration results from increasing returns. 1 I