Mechanisms that support reduced emissions from deforestation and forest degradation (REDD/REDD+) have potential to counteract a large share of global greenhouse gas emissions if implemented effectively across the tropics. In 2007 the conference of parties to the United Nations Framework Convention on Climate Change called upon parties and international organizations to promote REDD through investments in capacity building and demonstration activities. This prompted many new actors to become involved in REDD activities at a variety of locations and scales. A global survey of REDD activities was undertaken in 2009 to enable better understanding of the intensity and geographic distribution of these activities. Existing compilations, literature review, web-based sources, face-to-face and telephone interviews, and e-mail questionnaires were used to compile data for the inventory. Inter alia, data were collected on the location of activities and official and unofficial factors influencing location choices. Inventory data were combined with secondary data to estimate a statistical count model (Poisson) of factors affecting the number of REDD activities undertaken in the 64 developing countries that experienced significant emissions from deforestation. The results show that there were at least 79 REDD readiness activities and 100 REDD demonstration activities as of October 2009. Of these, the largest shares of REDD readiness and demonstration activities were implemented in Indonesia (7 and 15 respectively) and Brazil (4 and 13 respectively), countries widely agreed to have the greatest potential for reducing forest-based emissions. The statistical results found no national characteristic to have a statistically-significant effect on the number of REDD readiness activities, but five national characteristics to have significant effects on the number of REDD demonstration projects. Baseline CO2 emissions, forest carbon stock, number of threatened species, quality of governance, and region all had significant effects. The results reveal the importance of biodiversity and good governance, and the relative unimportance of human need and opportunity cost of land. The results also reveal a bias against Africa and toward Latin America. Unless this pattern is countered, REDD and REDD+ may have geographic biases that undermine its broad political support.