Protecting the world’s remaining forests is a global policy priority. Even though the value of the world’s remaining forests is global in nature, much of the protection has to come from national policies. Here, we combine global, high resolution remote sensing data on forest outcomes (tree-cover loss, forest degradation, net primary production) and two complementary econometric research designs for causal inference to first quantify how much it matters in which country a forest is located, secondly, the role of public policies, and third, under which conditions such pubic policies tend to be most successful. We find considerable border discontinuities in remotely sensed forest outcomes around the world (in a regression discontinuity design) and these are largely explained by countries’ policies (using a differences-in-discontinuities design). We estimate that public policies reduce the risk of tree cover loss by almost 4 percentage points globally, but there is large variation around this. The best explanations we find for these heterogenous treatment effects are a country’s policy enforcement, its policy stringency, its property rights, and its rule of law (in that order). Our results motivate international cooperation to finance and improve (a) countries’ public policies for forest protection and (b) countries’ capacity to implement and enforce them well.

保护世界上剩余的森林是全球政策的优先事项。尽管世界剩余森林的价值本质上是全球性的，但大部分保护必须来自国家政策。在这里，我们结合了关于森林结果的全球高分辨率遥感数据（树木覆盖损失、森林退化、净初级生产）和两个互补的计量经济学研究设计，以进行因果推断，首先量化森林位于哪个国家的重要性其次，公共政策的作用，第三，在什么条件下，此类公共政策往往最成功。我们发现世界各地遥感森林结果存在相当大的边界不连续性（在回归不连续性设计中），这些很大程度上是由国家政策（使用不连续性差异设计）来解释的。我们估计，公共政策可将全球树木覆盖率损失的风险降低近 4 个百分点，但这方面存在很大差异。我们发现，对这些异质性治疗效果的最佳解释是一个国家的政策执行、政策严格性、产权和法治（按顺序）。我们的研究结果推动了国际合作，以资助和改善（a）各国的森林保护公共政策以及（b）各国良好实施和执行这些政策的能力。