Conservation auctions have been advocated as a way to increase the cost-effectiveness of payment for ecosystem services (PES) programs by reducing the informational rents captured by participating landowners. Most PES programs have continual or periodic (rather than one-time) enrollment. In repeated auctions, it is possible for participants to learn the winning bids from previous auctions and use this information to strategically set their bids, thereby capturing more informational rents. We develop an agent-based model, using data from Costa Rica's Pago de Servicios Ambientales (PSA) program, to examine how strategic behavior, specifically through learning about previous winning bids, affects program participation and cost-effectiveness. When learning and strategic behavior occur among landowner agents in the model, informational rents increase and environmental benefits captured per dollar decrease over time. However, we also show that the distribution of participation can be adjusted by targeting for social as well as environmental benefits.