► We develop a conceptual framework of determinants of the impact of CCT and PES programs. ► Compliance with conditions in the absence of payments is a good predictor of efficiency. ► CCT and PES efficiency is highly eroded by selection bias in participation. ► The framework can guide policy choice through ex-ante estimates of program impacts. We develop a conceptual framework elucidating the main determinants of the impact of Conditional Cash Transfer (CCT) and Payments for Environmental Services (PES) programs. Using a simple multi-agent model and evaluations of existing programs, we show that (1) the share of the population who would meet the program’s conditions in the absence of payments is a powerful predictor of program efficiency, and that (2) program efficiency is eroded by selection bias (people who already meet conditions self-select into the programs at higher rates than others). We then discuss possibilities for increasing efficiency and criteria for evaluating and choosing between CCTs/PES or other policy instruments.