Speaker: Joseph P. Kaboski is the David F. and Erin M. Seng Foundation Professor of Economics in the Department of Economics at the University of Notre Dame. His research focuses on growth, development and international economics, with an emphasis on structural change, finance and development, schooling and growth, microfinance, explaining international relative price patterns, and the role of inventories in international trade. More »

Abstract: In developing countries, industrial agglomerations are often fostered because, theoretically, local external economies can lead to inefficiently low levels of agglomeration and investment. However, agglomerations can also foster collusion. We explore the relationship between geographic industrial agglomeration and collusion both theoretically and empirically. We develop a model of agglomeration and collusion, in which external economies and collusion have very different impacts. Markups increase profitability but lower market share, while external economies increase profitability but raise market share. Moreover, the model motivates a novel identification strategy for collusion that is analogous to Townsend (1995)’s risk-sharing regression: we solve the planner’s problem of an agglomerative cartel of perfectly colluding firms rather than a village-level syndicate of perfectly risk-sharing households. Outside of a cartel, markups depend on a firm’s market share but not on the total market share of firms in the agglomeration, whereas in a perfect collusion, markups are constant across firms and depend only on the overall market share of the agglomeration. These effects are more easily seen for highly tradable goods, for which the size of the local market is negligible. Empirically, we test for collusion using data from Chinese manufacturing firms (1999-2008), which exhibit industrial agglomeration. We indeed find evidence for collusion both within agglomerations and within certain industries more broadly.

Last Updated: Apr 16, 2015