Incentive-based programs to reduce deforestation are expected to play an increasingly important role in global efforts to protect ecosystems and sequester carbon but their environmental effectiveness is not clear. We investigate program effectiveness and slippage in the context of Mexico's national payments for hydrological services program, which pays private and communal landowners to maintain forest cover on enrolled lands. To measure program impacts, we use matched controls drawn from the program applicant pool to establish counterfactual deforestation rates in the absence of payments. We find statistically significant but small to moderate avoided deforestation impacts. To examine slippage of deforestation to nonenrolled lands, we develop a model of household land allocation to agriculture or forest in which some households are credit-constrained. We illustrate that payments for forest conservation may result in slippage due to substitution, resulting in increased deforestation on other parcels belonging to program recipients, or due to output price effects, resulting in increased deforestation within markets where there are high levels of program participation. Our data show evidence consistent with slippage through both mechanisms. This suggests that incentive-based mechanisms can work to prevent deforestation but that avoided deforestation should be accounted for at a regional or national level in international schemes to reduce carbon emissions from deforestation and forest degradation.