Misallocation in Firm Production: A Nonparametric Analysis Using Procurement Lotteries
How costly is the misallocation of production that we might expect to result from distortions such as market power, incomplete contracts, taxes, regulations, or corruption? This paper develops new tools for the study of misallocation that place minimal assumptions on firms' underlying technologies and behavior. We show how features of the distribution of marginal products can be identified from exogenous variation in firms' input use, and how these features can be used both to test for misallocation and to quantify the welfare losses that it causes. We then consider an application in which thousands of firms experience demand shocks derived from a lottery-based assignment of public procurement contracts for construction services in Ecuador. Using administrative tax data about these firms, we reject the null of efficiency but estimate that the welfare losses resulting from misallocation are only 1.6% relative to the first-best. Standard parametric assumptions applied to the same setting would suggest losses that are at least an order of magnitude larger.
We thank our team of outstanding, dedicated research assistants, as well as David Baqaee, Richard Blundell, Kirill Borusyak, Arnaud Costinot, Pete Klenow, Jeremy Majerovitz, Matt Masten, Virgiliu Midrigan, Emi Nakamura, Whitney Newey, Michael Peters, Richard Rogerson, Cian Ruane, Jesse Shapiro, John Sturm, Alex Torgovitsky, Ivan Werning, and numerous seminar participants for helpful comments. We are grateful to the Centro de Estudios Fiscales and the Departamento de Control of the Ecuadorian Tax Authority for outstanding collaboration and to Innovations for Poverty Action (IPA), International Growth Centre (IGC) and the Stanford Institute for Innovation in Development Economies (SEED) for generous research support. This paper has also benefited from funding from CEPR and UK Department for International Development (DFID) (under the Private Enterprise Development in Low-Income Countries program, reference MRG004 3834), the European Research Council (grant reference 758984), and the Swiss National Science Foundation (grant 100018 192588). The views expressed are not necessarily those of CEPR, DFID, ERC or SNF. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research.
I, Paul Carrillo, have visited and taught short courses at SRI for nominal compensation. I do not see any conflicts of interest but am happy to disclose in accordance with NBER policy.