A large literature examines performance pay for managers in the private sector, but little is known about performance pay for managers in public sector bureaucracies. In this paper, we study performance incentives rewarding school administrators for reducing anemia among their students. Randomly assigning 170 schools to three performance incentive levels and two orthogonal sizes of unconditional grants, we analyze performance pay and its complementarity with discretionary resources. We find that both large incentives and larger unconditional grants reduced anemia substantially, but incentives were more cost-effective. Performance incentives led administrators to innovate by working with parents, mitigating potentially offsetting compensatory behavior among households. Strikingly, we also find that larger unconditional grants completely crowded-out the effect of incentives. Our findings suggest that performance incentives can be effective in bureaucratic environments – but also that discretionary resources can fully crowd-out their effect.