Robust design of multiscale programs to reduce deforestation

A framework is provided for structuring programs aimed at reducing emissions from deforestation and forest degradation (REDD). Crediting reference levels and the coordination among different implementing entities at multiple geographic scales are discussed. A crediting reference level has an error component if it differs from the business-as-usual (BAU) without REDD. Both the BAU emissions and the impact of REDD actions are uncertain, implying that participating in REDD entails stakeholder risk, the distribution of which depends on REDD program design. To categorize REDD architectures we define <italic>scale-neutrality</italic> whereby, for a given REDD design, crediting relative to the reference level at a given scale is not affected by errors in reference levels at scales below it. Sufficient conditions are derived for scale-neutrality to hold. A Brazilian Amazon example is provided, comparing potential REDD architectures, and highlighting how a cap-and-trade approach may match the environmental outcome obtainable with perfect foresight of the BAU emissions.