Giving power over school management and spending decisions to communities has been a favored strategy to increase school quality, but its effectiveness may depend on local capacity. Grants are one form of such a transfer of power. Short-term responses of a grant to school committees in Niger show that parents increased participation and responsibility, but these efforts did not improve quality on average. Enrollment at the lowest grades increased and school resources improved, but teacher absenteeism increased, and there was no measured impact on test scores. An analysis of heterogeneous impacts and spending decisions provides additional insight into these dynamics. Overall, the findings suggest that programs based on parent participation should take levels of community capacity into account: even when communities are willing to work to improve their schools, they may not be able to do so. The short-term nature of the experiment reduces the extent to which the results can be generalized.