The Performance of a Repeated Discriminatory Price Auction for Ecosystem Services

Government agencies often rely on repeated discriminatory price auctions to procure ecosystem services from private landowners despite limited evidence on this mechanism’s performance. This study presents an agent-based model of a repeated discriminatory price procurement auction in which the auctioneer informs bidders of their respective bid outcomes and the average price of successful bids after each round. The model introduces a new learning algorithm through which bidders adapt to these price signals. Simulations are used to compare the mechanism’s performance to an equivalent uniform (second) price auction, providing several findings. First, the performance of the discriminatory mechanism tends to deteriorate over time relative to the uniform mechanism as bidders learn. Second, minimal changes in bidders’ price expectations have a large influence on the relative performance of the mechanisms. Third, the discriminatory mechanism maintains high levels of efficiency and cost-effectiveness over time if bidders have highly heterogeneous opportunity costs and neutral or moderately low price expectations. Fourth, the system’s price paths have a high degree of stochasticity and path dependency, making it difficult to predict a single realisation’s trajectory. Based on these findings, we provide several suggestions regarding auction design.