Payments for Environmental Services (PES) are increasingly used in tropical countries as incentives to improve forest conservation outcomes, notably in forests managed or owned by rural communities. However, little is known about the performance and the suitable conditions under which collective PES can be effective. We assess the impact of a collective PES on reducing deforestation in the Cardamom Mountains, Cambodia. We use a quasi-experimental approach combining covariate matching with difference-in-difference estimation as a procedure to assess program effectiveness on reducing deforestation. We find that the PES program contributed additionally saving on average about 0.17 percentage point of the enrolled PES area per year during 2005–12. We also highlight that characteristics such as slope, proximity to roads, and number of households are associated with heterogeneous PES impact. We finally call for improving both data availability and impact assessment methods in order to test better theories of change that take into account not only the ability of payments to reduce opportunity costs of conservation but also to consider how PES are turned into improved community collective action likely to improve the local enforcement of conservation.