Friday, December 14, 2012

Complicated auctions are more proftable

There was a time where auction theory limited itself to studying very simple auction, where the subtleties were whether the first or second price should be paid (or the first less an increment). Now, auction theory looks at much more complex mechanisms, for example where bidders may or may not reveal their bids, or whether they are bidding, or where bidding comes with a fix price. Not all these mechanisms try to obtain a surplus maximizing outcome. Some maximize the profits of the seller, sometimes by confusing or even misleading the seller. The most extreme example are penny auctions, about which I posted before (I, II).

Andrea Gallice discusses a variation of the Dutch auction where the current winning bid price remains hidden but can be observed against a fee. This so-called price reveal auction has an additional twist: paying that fee makes the winning bid fall by a predetermined amount. An auction so complex must be designed to maximize someone's surplus. It is the seller. And his profits are even higher if he manages to keep the number of bidders secret. This is not unlike penny auctions, where the profits come from the fees, not the winning bid.

OK, this maximizes profits, but I do not think this maximizes overall well-being. Obfuscation is not likely to be beneficial, and I am quite surprised the author does not address this. Until convinced of the contrary, I am going to assume that such obfuscation is detrimental for society and should be outlawed. And with rules so complex, it would not surprise me that bidders would have a hard time behaving rationally.

1 comment:

Andrea said...

Hi, here is the author of the above mentioned paper. Let me just point out that I obviously agree with Economic Logician's concerns about the "moral virtues" of price reveal auctions. These mechanisms (I also include lowest unique bid auctions and penny auctions) do not aim at allocating a good efficiently. They rather have been designed with the explicit goal of raising seller's profits, possibly also exploiting agents' enthusiasm, curiosity and naivete. In the paper I focus on a pure game-theoretic analysis of the mechanism and therefore I do not mention these issues because I find them quite obvious. Notice that the study of the theoretical properties of these games remain interesting as pay-per-bid mechanisms could be succesfully used in other (more ethical) contexts such as fundraising activities and charities, i.e., situations in which agents are more consciously and happily willing to "lose some money" but still may enjoy the fun of playing and the chance to win a prize.