Testing the Effect of Monitoring in Production Contracts: Evidence from Madagascar

This paper explores the effects of monitoring in contracts between an exporting firm and small agricultural producers in Madagascar. Building on a theoretical framework that incorporates both adverse selection and moral hazard, I test for the effect of monitoring of the agents by the principal on productivity, all the while including an estimate of agent-specific technical inefficiency in the regression used to conduct my test, which obviates the identification problem usually encountered when using cross-sectional data in applied contract theory. Empirical results show that the inclusion of such a technical inefficiency estimate has important implications both for applied contract theory and for policy. JEL Classification Codes: D86, L24, 012, 013, Q12. ∗Assistant Professor of Public Policy and Economics, Terry Sanford Institute of Public Policy, Duke University, 201 Science Drive, Box 90239, Durham, NC, 27708-0239, [email protected]. I am grateful to Bart Minten for allowing me to use these data, and I thank Liz Ananat, Chris Barrett, Pierre Dubois, and David Just as well as participants in the Development Microeconomics Research Seminar at Cornell University for useful comments and suggestions.