Optimal choice of a reserve price under uncertainty

The revenue function for a standard auction is typically asymmetric around the revenue maximizing reserve price. Thus, choosing a reserve price that is smaller than the revenue maximizing reserve price can result in a substantially different loss than choosing one that is larger by the same amount. Therefore, when the revenue function is unknown, it is important to consider uncertainty around the revenue function and its asymmetric structure. For this purpose, I propose a Bayesian decision rule and illustrate its typical revenue gains. I then apply the rule to the bid data from the U.S. timber sales.