The article reviews some of the critical issues for including Reduced Emissions from Deforestation and forest Degradation (REDD) into a new global climate agreement. Four different REDD models (regimes) are discussed based on two dimensions: scale (national vs. project) and funding (market vs. funds). One of the most troublesome issues concerns setting national baselines (= emission quotas). Research provides few definite answers on how to do this, yet it has huge implications for the possible payments to developing countries. The paper argues that the expectations about the magnitude of such transfers are unrealistically high, and may reduce the prospect for reaching an agreement and increase the chances for ‘hot air’ from the South.