European Central Bank Operations: Experimental Investigation of Variable Rate Tenders

In an experimental study auction procedures of variable rate tenders, which are prevalent in practice, the discriminatory American auction and the competitive Dutch auction, are investigated on a repeated basis. At the center of this study stands the comparison of the two auction formats with respect to true bidding, efficiency, seller's revenue and bidders' payoffs. We observe bid shading in both auction formats and a higher level of underbidding in the American auctions than in the Dutch auctions. Furthermore, subjects understate their willingness to pay more and more in the course of time in the American auction. As a result seller's revenue decreases and bidders' payoffs increase. However, the experimental results neither favor the American nor the Dutch auction format, in terms of revenue. Moreover, no differences are found with respect to efficiency. The outcomes of both auction formats are inefficient. However, the level of efficiency increases over time in the auctions.