Enterprise Dynamics and Finance: Distinguishing Mechanism Design from Exogenously Incomplete Markets Models ∗ Alexander Karaivanov Simon Fraser University Robert M. Townsend University of Chicago February, 2008 PRELIMINARY AND INCOMPLETE Abstract We formulate and solve a range of dynamic models of information-constrained credit markets that allow for moral hazard and unobservable investment. We compare them to the exogenously incomplete markets environments of autarky, saving only, and borrowing and lending in a single asset. We develop computational methods based on mechanism design theory, linear program- ming, and maximum likelihood techniques to structurally estimate, compare and statistically distinguish among the competing theoretical models of credit market imperfections. Our meth- ods can be applied with both cross-sectional and panel data and allow for measurement error and unobserved heterogeneity in initial conditions. The models match major stylized facts from the empirical literature on firm dynamics as listed by Cooley and Quadrini (2001). Empirically, we find that using consumption, cashflow and investment data jointly or using dynamic data improves the researcher’s ability to distinguish across the various model regimes relative to us- ing consumption or investment only data, especially in the presence of high measurement error. We also estimate our models using data on Thai households running small businesses. We find that the borrowing and saving only frameworks provide the best fit when using joint data on consumption, cashflow and investment. Keywords: financial constraints, mechanism design, structural estimation and testing JEL Classifications: C61, D82 ∗ We thank Lars Hansen, Ali Hortacsu, Pascal Lavergne, Derek Neal, Sam Schulhofer-Wohl, participants at the Society of Economic Dynamics conference, and seminar audiences at the University of Chicago, Stanford and Berkeley for their insightful comments and suggestions. All remaining errors are ours. 1