This paper measures the impact of year one of the Millennium Challenge Corporation’s Rural
Business Development program for small rice farming households on the Pacific Coast of
Nicaragua. The program was rolled out in the 2009-2010 agricultural year, which was the driest
year on record in the region, likely due to an El Niño event. Estimated impacts show that the
program at best had no effect, and at worst led to a 10 percent reduction in yields. These impacts
are estimated using an econometric model which uses selection on observables as its identifying
assumption, and robustness checks suggest that this is a reasonable approach in this case.
Inference accounts for spatial correlation across households of the unobserved determinants of
agricultural outcomes. The program appears to have been almost exclusively focused on
increasing yields through better and greater application of chemical fertilizers, and minimization
of losses in the post-harvest stages of production. If the pessimistic estimates of program effects
are true, then the program could have been improved by incorporating risk management
strategies into extension advice. On the other hand, farmers may be well insured against climatic
risk, in which case they may have selected into the program knowing that they would be trading
greater risk for higher expected returns. Survey data offers some evidence that the latter is indeed
the case.