Conservation impacts and hidden actions in a randomized controlled trial of a marine pay-to-release program.

Incentive payments could cost-effectively and equitably achieve biodiversity conservation goals but could also trigger unintended countervailing actions. Here, we report on a preregistered, randomized controlled trial of a pay-to-release program among small-scale, Indonesian fishing vessels for the release of two critically endangered marine taxa from fishing gear: hammerhead sharks and wedgefish. A conventional monitoring approach, which quantifies impacts based on conservation-relevant actions (i.e., numbers of live releases), implies that the program was successful: a 71 and 4% reduction in wedgefish and hammerhead shark mortality, respectively. The experimental data, however, imply that the pay-to-release program also induced some vessels to increase their catch, thereby decreasing wedgefish mortality by only 25% [confidence interval (CI): -49 to 10%] and increasing hammerhead mortality by 44% (CI: 8 to 92%). Our results do not imply that pay-to-release programs cannot work but rather demonstrate the complexity of designing incentive-based conservation programs and the importance of piloting them using experimental designs before scaling up.