REDD+ is usually presented as an incentive-based mechanism that can provide payments to compensate for the costs induced by conservation restrictions. Yet in Madagascar REDD+ is implemented through a command-and-control approach with almost no or insufficient compensation. This paper challenges the financial feasibility of an individual cash or in-kind compensation scheme as part of a REDD+ project and assesses the cost of implementing a hypothetical individual compensation scheme for local populations living on the boundary of an ongoing REDD+ pilot project in southeastern Madagascar. In order to estimate a plausible level of compensation, we measured households’ perceived economic losses arising from the project. We carried out this economic evaluation based on households’ declarative statements about their agricultural production (before and after project implementation) and their perceptions of the causes of such changes. We then estimated the start-up and running costs of implementing conditional transfers to compensate for reported losses using first-hand project cost data from different conservation projects in Madagascar, including the one analysed in this paper. Comparing our estimated total cost to the current budget of the REDD+ project, we concluded that compensating households would cost seven times more than the budget initially devoted to field activities during the first phase of the project. Yet we discuss that individual compensation may increase the long-term environmental and social additionality (through greater legitimacy) of the REDD+ project, as it may play a role of safety nets and help farmers, especially the most vulnerable ones, to implement new agricultural techniques to adapt to land use restrictions.