RESUMEN
Between 2004 and 2012, multiple policies contributed to one of the great conservation successes of the twenty-first century-an 84% decrease in the rate of Brazilian Amazon deforestation. Among the most prominent of these policies is the Amazon Soy Moratorium (ASM), an agreement by grain traders not to purchase soy grown on recently deforested land. The ASM inspired widespread adoption of similar zero-deforestation commitments, but its impact is poorly understood due to its overlap with other conservation policies. Here, we apply an econometric triple-differences model to remotely sensed deforestation data to isolate the ASM's impact within Brazil's Arc of Deforestation. We show that the ASM reduced deforestation in soy-suitable locations in the Amazon by 0.66 ± 0.32 percentage points relative to a counterfactual control, preventing 18,000 ± 9,000 km2 of deforestation over its first decade (2006-2016). Although these results highlight potential benefits of private conservation policies, the ASM's success was dependent on complementarities with public property registries and deforestation monitoring.