Is there misallocation of production in developing countries, as we might expect to occur in the presence of credit constraints, imperfect competition, regulations, political connections, or other distortions? We develop a nonparametric test for allocative efficiency based on the way that firms respond to demand shocks. We then apply this test to a setting where firms experience exogenous demand shocks (equal to about 50% of annual sales for the median firm) due to the randomized allocation of thousands of public procurement lotteries in Ecuador. Using monthly data on firm-to-firm transactions and employer-employee relations, we first document average responses: firms respond to demand shocks by rapidly expanding total scale and input use, and do so with no crowd-out or crowd-in of other activities contemporaneously or in the future. These findings have much to say about the incidence of demand shocks, but have no implications for inferences about misallocation in our nonparametric context. However, the key prediction from a nonparametric efficient economy is that these responses should be common across firms. Consistent with that prediction, we document that firms with different predetermined characteristics respond similarly to lottery-driven demand. Further, a test based on Fisher exact p-values fails to reject the null of allocative efficiency at standard levels.

Written with Paul Carrillo (George Washington University), Dina Pomeranz (University of Zurich) and Monica Singhal (University of California, Davis)