Revenue-maximizing Dutch auctions with discrete bid levels
This paper is concerned with setting a predetermined number of bid levels in a Dutch auction to maximize the auctioneer’s expected revenue. As a departure from the traditional methods used by applied economists and game-theorists, a novel approach is taken in this study to tackle the problem by formulating the auctioning process as a constrained nonlinear program and applying standard optimization techniques to solve it. Aside from proposing respective closed-form formulae for computing the optimal bid levels and the auctioneer’s maximum expected revenue, we also show that the bid decrements should be increasing if there are two or more bidders in the Dutch auction. Additionally, the auctioneer’s maximum expected revenue increases with the number of bidders as well as the number of bid levels. Finally, managerial implications of the key findings as well as limitations of this research work are discussed.