The Diffusion of Wal-Mart and Economies of Density


  • Thomas J. Holmes

    1. Dept. of Economics, University of Minnesota, 4-101 Hansen Hall, 1925 Fourth Street South, Minneapolis, MN 55455, U.S.A., Federal Reserve Bank of Minneapolis, and NBER;;
    Search for more papers by this author
    • I have benefited from the comments of many seminar participants. In particular, I thank Glenn Ellison, Gautam Gowrisankaran, and Avi Goldfarb for their comments as discussants, Pat Bajari and Kyoo-il Kim for suggestions, and Ariel Pakes for advice on how to think about this problem. I thank the referees and the editor for comments that substantially improved the paper. I thank Junichi Suzuki, Julia Thornton Snider, David Molitor, and Ernest Berkas for research assistance. I thank Emek Basker for sharing data. I am grateful to the National Science Foundation under Grant 0551062 for support of this research. The views expressed herein are those of the author and not necessarily those of the Federal Reserve Bank of Minneapolis or the Federal Reserve System.


The rollout of Wal-Mart store openings followed a pattern that radiated from the center outward, with Wal-Mart maintaining high store density and a contiguous store network all along the way. This paper estimates the benefits of such a strategy to Wal-Mart, focusing on the savings in distribution costs afforded by a dense network of stores. The paper takes a revealed preference approach, inferring the magnitude of density economies from how much sales cannibalization of closely packed stores Wal-Mart is willing to suffer to achieve density economies. The model is dynamic with rich geographic detail on the locations of stores and distribution centers. Given the enormous number of possible combinations of store-opening sequences, it is difficult to directly solve Wal-Mart's problem, making conventional approaches infeasible. The moment inequality approach is used instead and works well. The estimates show the benefits to Wal-Mart of high store density are substantial and likely extend significantly beyond savings in trucking costs.