This paper uses systems analysis to examine the factors that influence leakage in Reduced Emissions from Deforestation and Degradation (REDD) projects in tropical forest areas in developing countries. A causal loop diagram is used to identify some key intervention points for stakeholders to reduce the negative impacts of leakage. One of the most important intervention points identified is to provide livelihood alternatives to supplement the resources and the income that neighbouring small-scale landholders and communities must forego for the REDD initiative to be successful. The question remains, however, of how best to organise and distribute these resource allocations and payments to support the REDD project objectives. It is concluded that in many cases, developing countries with tropical forest coverage are unlikely to benefit from REDD in the foreseeable future, because of constraints related to economies of scale, but also because of political instability and lack of institutional capacity in many of these nations. Moreover, the risk remains that many of these developing countries will suffer from cross-border market leakage due to REDD projects conducted elsewhere, and that this will lead to negative local forest conservation and livelihood outcomes.