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.