Many environmental issues can be attributed to misaligned distribution of the costs of conservation and the benefits of conservation. For instance, biodiversity represents value for the global community, but biodiversity protection imposes various costs on local communities in forested areas of developing countries. Correcting this misalignment requires presenting these local communities with appropriate incentives. Conservation agreements – negotiated transactions in which conservation investors finance direct social benefits in return for conservation actions by communities – are one tool for doing so. The results of this approach depend crucially on effective monitoring of both ecological and socio‐economic impacts to verify that environmental and development objectives are met in a socially equitable, economically efficient, and financially sustainable way. Monitoring also is needed to verify that parties to the agreements are in compliance with their commitments. This paper will present the conservation agreement model and demonstrate the central role of robust monitoring frameworks, using the example of agreements between Conservation International and communities in the Colombian Amazon. These agreements are designed to protect forest areas and two endangered species of fish that are important to local livelihoods and have a high commercial value in neighboring countries. A key feature of this project is that the agreements both depend on and strengthen social and resource governance within the partner communities, thereby promoting self‐determination while enhancing the overall context for socio‐economic development. At the same time, lessons generated by this project inform emerging frameworks for scaling up the approach to advance conservation and development at the national level, requiring integration with national policies. The paper will conclude by identifying the strengths and limitations of the conservation agreement approach, emphasizing that effective monitoring is essential for success and exploring the implications of scaling‐up for design of monitoring frameworks.
