Vickrey Auctions--Wronging Rothkopf?

Thirteen Reasons Why the Vickrey-Clarke-Groves Process Is Not Practical by Michael H. Rothkopf is a compelling paper which outlines some of the pitfalls of the "second price sealed bid" auction. It also compels me to wonder why Mr. Rothkopf is handing such a drubbing to Drs. Vickrey, Clarke and Groves.

Before we proceed, we should take a moment to reflect upon the very recent passing of Professor Rothkopf, a Penn State professor who contributed much to mechanism design. He died earlier this year at the age of 68 exercising in the university natatorium.

I became interested in the second price sealed bid after considering how prone the conventional English auction is to the "Winner's Curse." The Vickrey auction delves into the motivation for overbidding and almost forces the bidder to value the auction item at it's true fair worth. In the defense contracting business, where large companies compete for both profit margins and market share, it is essential that fair valuations of contracts be calculated to avoid cutting into profits or losing too many bids (and driving up overhead rates.)

So why wouldn't the auctioneers such as the federal government adopt the Vickery auction or some derivation like the VCG process? Especially if, theoretically, the VCG process promises perfect efficiency?

Well, heck if there aren't thirteen reasons why they shouldn't. Sparing you the entire 7-page read filled with scads of game-theory-speak, I'll put some of the more compelling reasons in the context of a defense aquisition scenario:

• Bid Preparation and Bid Communication Costs: on cost-plus-fixed-fee contracts, the goverment does not necessarily pay for the contractor to prepare his own bids; sometimes the contractor can tack on bid preparation through some R&D mini-contract, but I digress. The more the contractor spends to estimate her value, the worse her profits are.
• Information Revelation: winning the contract award at the second price reveals information to the winner of the best loser's valuation. In two-bidder contests, proprietary cost information would be revealed to the winner and could be used against the loser in future contests.
• Cheating (four kinds!): Conspiracy or collusion is easily achieved between old rivals such as Boeing and Lockheed Martin. I say old rivals because the books have to be cooked in this new age of Sarbanes-Oxley accounting. In the old days Bill Boeing could tell Allan and Malcolm Loughhead, "Hey, I'll give you a quick $2 million dollars if you bid $13 million on this contract I value at $10 million. Don't bother putting much effort into the proposal."
  Other cheating measures include false-name bids and communicating to each other via actual bid amounts, as in an FCC bandwith auction: "I'll take package 22 and you can have bandwith package 19, therefore I'll bid $20,022,019" Hey, it's happened.

The most important reason? Rothkopf asserts that every case is different, but he actually discusses the possible repercussions in a "government sale of extremely valuable assets." How would Congress respond to a contract award which offers significantly higher profits on a percentage basis then would originally have awarded on, say, a cost-plus fixed fee contract.

P.S. In case you didn't know, the Vickrey auction is alive and well on Ebay through the "Name Your Maximum Bid" option. You win when your opponent bidder hits her maximium bid and you pay the bid increment (epsilon,) no more. You can still bid incrementally, but notice how you carefully decide what the item is worth to you when setting the inflated price...because you might just have to pay it!