Reducing CO2 emissions from forests was slow to get off the ground as a subject of international climate negotiations, but it has picked up considerable momentum since 2005. In particular, agreement has been reached on the urgence to set up a global REDD+ mechanism. The mechanism aims to provide developing countries with incentives to reduce emissions from deforestation and forest degradation and to increase forest carbon stocks through appropriate forestry practices or through planting. Agreement has also been reached that REDD+ incentives should be result-based and ultimately awarded at the national level. Nevertheless, local initiatives are a useful mean of tackling deforestation. However, when carbon incentives depend on national performance, linking them to local initiatives is a technical and financial challenge. Technically, the national accounting framework must be able to track emissions-reduction initiatives at the sub-national level (regional, local or projet level). Financially, investors are likely to be scared away if their reward depends on deforestation occuring outside the area of their investment. Ultimately, the issue of transferring national incentives coming from supranational agreements to the local level can be reduced to a political decision on risk sharing between the State and private stakeholders. Industrialized countries have already faced this issue during the first commitment period under the Kyoto protocol and they have often found it difficult to develop satisfactory solutions. Two notable exceptions are New Zealand, which included its forest sector in its emissions trading scheme, and Australia, which is developing a „Carbon Farming Initiative‟ for forestry and agricultural offsets. This study draws lessons from a comparison of the treatment of the Land Use, Land Use Change and Forestry (LULUCF) sector in industrialized countries during the first Kyoto protocol commitment period and the current discussions and initiatives on the architecture of a future REDD+ mechanism. Two opposite alternatives are identified: a State guarantee that project developers are rewarded based on the success of their project no matter the national performance, as in New Zealand; and a carbon incentive for project developpers which is scaled down in proportion of the national performance, as it has been discussed in France. Beyond the political decision concerning the risk sharing, three technical keys to a successful transfer of carbon incentives to local actors are presented through the analysis of a number of case studies: (1) a consistent set of baselines; (2) a monitoring system which includes both emissions reductions and the causes behind them; and (3) a clear regulatory framework. The variety of approaches to including project reference baselines in the national reference level is analyzed based on experiences in the European Union, Guyana and Peru. Regarding monitoring, building on national GHG inventories from Annex 1 countries and Brazil‟s experience in monitoring deforestation could help leapfrogg the technical challenges posed by Monitoring, Reporting and Verification (MRV) in REDD+. Lastly, the study explores the progress made by Colombia in setting up a national regulatory framework and by the state of Acre in Brazil in building an institutional framework for REDD+ at a regional level.