Estimating the effectiveness of forest protection using regression discontinuity

This article uses satellite data to estimate the effectiveness of government protection on forested land across the globe over 2000–2022. Since deforestation can have significant negative externalities, measuring the effectiveness of protected areas is important for the future of conservation. It uses a regression discontinuity design at the boundaries of protected forest to counter the fact that protection is not randomly assigned. It estimates that protected areas are 30% effective on average, with significant heterogeneity between countries. Many countries with significant forest have extremely ineffective protection, such as Indonesia, the DRC, and Bolivia, suggesting that improvements to the quality of protection are just as important as the quantity of protected areas to conserve biodiversity.