Stanford GSB: Stanford Business February 2001

Stanford Business

February 2001, Volume 69, Number 2

The Auction Architects:
Part 1
The Auction Architects:
Part 2

Going Once, Going Twice...SOLD on Auctions

Business School architects are erecting new frameworks for an age-old economic institution. As auction fever spreads throughout the economy, people are wondering if fixed price tags are going the way of the horse and buggy.

by KATHLEEN O'TOOLE

I ONCE STUMBLED UPON a knot of people in a park who obviously weren't feeding the ducks. Upon closer inspection, I found they were bidding on iris rhizomes, which proud breeders from all over the state had brought to an auction organized by a horticulture club. As a cheapskate and sometime gardener, I asked for a bidding card, but then was stunned into silence as bulbs named Tequila Sunrise and Brazen Beauty went quickly from $2 to $12 and then $20. I could feel my heart racing as I raised my card fast to bid on Hayride and wound up with a shriveled lump for $6. Within an hour, I spent $32 on irises but was so embarrassed I told my husband it was $15.

That was in the pre-Internet dark ages when some of us considered auction-goers the kin of gamblers. We thought it more civilized to get deals by cutting coupons from newspapers or collecting frequent flyer miles. We didn't know there were "decision scientists" at Stanford's Business School writing equations that showed auctions to be of great personal and social value. Had I been aware, I might have been less surprised when almost everybody started bragging about the deals they had made for Beanie Babies or Woodstock memorabilia on eBay or for "virtually new" steam cleaners and binoculars on retailers' auction sites.

Jeff Skoll, MBA '95 and the first president of the leading auction site eBay, didn't know about the decision scientists either. "We began, ran, and developed our site (to this day) without the benefit of academic research. It was all more 'wing and a prayer' really," he says.

Thanks to the Internet and an international tidal wave of governments seeking to privatize state-owned or regulated industries, auction fever has spread faster than a flu virus. The GSB researchers are game theorists who once studied auctions not because of their popularity but because they were manageable models of more complex business activities.

In demand as consultants and business partners, they have become auction architects and engineers, designing economic structures for real-world use, especially by governments. Business School faculty and their former doctoral students have formulated the rules for the auction sale of everything from Treasury bills to electromagnetic spectrum licenses, electricity, gas, railroads, and universal communication service contracts. Some advise private market-makers for Web-based business auctions and exchanges. Some advise sellers on how to prevent collusion among bidders.

At the apex of the activity is Robert Wilson, the Adams Distinguished Professor of Management, whom other game theorists credit with bringing their specialty to the forefront of microeconomics, a field focused on how industries and firms work rather than on national economies.

In the seventies and eighties, "Bob Wilson fruitfully analyzed business situations and published in economic theory journals, which convinced other economists of the value of auction theory," says John McMillan, author of Games, Strategies, and Managers, a book that Fortune described once as "the most user-friendly guide for business people" to game theory, including auctions. "But the proof is in the pudding, and Bob also took these mathematical techniques the next step by applying them to designing real auctions."

It was not a slam-dunk, however. Old ways of allocating scarce resources had to break down before Wilson and his disciples got a chance to demonstrate the practicality of their mathematics. A doorway opened after the U.S. government decided to allocate airwaves to newly invented personal communication devices.

In the eighties, Congress voted to speed up the process with lotteries. The prospect of windfalls attracted 400,000 applications for cellular licenses, says McMillan, who was hired later by the Federal Communications Commission to oversee auctions for other spectrum rights. "In one not atypical case," McMillan wrote, "an obscure group called the RACDG partnership was chosen by lottery in 1989 to run cellular telephones on Cape Cod; the partners then sold their license to Southwestern Bell for $41 million. The total value of cellular licenses the government gave away during the 1980s, according to a Commerce Department estimate, was $46 billion."

Several countries decided to auction spectrum rights, but the auction rules used in New Zealand and Australia led to public relations disasters and unhappy communications companies. With that in mind, American firms looked around for auction expertise, and soon Stanford's Wilson, Paul Milgrom, and Jeremy Bulow were in meetings with McMillan and others picking apart the problems.

The process was cooperative, instead of adversarial, almost like an academic seminar, participants recall. "The companies who hired the theorists gave them a pretty free hand," says McMillan, today professor of international management and economics at the Business School. "They probably had no idea what their stake was in any particular design and just wanted to make sure the guy advising them during the bidding understood it."

Having bidders understand is also in the sellers' interest, the theorists say, unless the design is so faulty that bidders can easily game it. "There are theorems in economics that say you can't design mechanisms that will work perfectly, that aren't subject to gaming, but subtle differences in design can sometimes be very important," says Bulow, the Richard A. Stepp Professor of Economics, who is currently on leave as chief economist of the Bureau of Economics of the Federal Trade Commission. He and former student Paul Klemperer, now a professor at Oxford, have coauthored papers that apply auction theory broadly to such phenomena as corporate take-over contests. They show that a shareholder tendering an offer for a company has an advantage because he is both a bidder and seller and that auctioning generally leads to higher prices than negotiating.

The initial U.S. spectrum auction designs in 1994 and 1995 were tested on students in the Caltech laboratory of Charles Plott. The results boosted the government's confidence in the design eventually chosen, called a simultaneous, ascending-price, multiple-round auction, which was invented by Milgrom, Wilson, and Preston McAfee of the University of Texas. The actual auction was widely regarded as a success, raising far more revenue than initially estimated and leading to what economists regarded as an overall efficient allocation.

Since then Milgrom, PhD '79 and a professor in Stanford's economics department and at the Business School, has worked on a more complex design for packages of licenses involving combinatorial bidding. He wants to better take into account the fact that licenses have differing values because of their geographical and spectral proximities to each other. After several delays and a joint Stanford-FCC conference on combinatorial bidding last May, the auction for 700 megahertz of third-generation wireless licenses is slated for April. In the meantime, countries around the world have been conducting auctions at a record pace as they try to accommodate the next generation of handheld devices, ranging from high-speed Internet telephones to wireless electronic credit cards.

Auction theory garnered big headlines in Europe last summer after a third-generation spectrum auction codesigned by Klemperer, MBA '82 and PhD '86, raised more than 23 billion pounds for the United Kingdom Treasury—20 billion over the initial estimate. Complimentary headlines soon gave way to industry complaints and strong lobbying in other European countries. Telecom companies claimed the prices were unjustified and would be reflected in higher prices for consumers of new products.

"There is no real justification for that argument in economic theory or in the real world," says Bulow, who participated in early planning for the British auction. "As everybody who has taken Econ 200 knows, whatever costs are sunk are gone. Companies price in a way to maximize their profits regardless of what they had to pay initially for the license."

"Pricing is driven by the competitive realities of the marketplace," agrees McMillan, who adds that is one reason most governments' primary goal in auction design is to maximize competition rather than revenue. If they were just interested in revenue, they would award one company a monopoly and collect royalties, Bulow says.

"We think auctions are a good way for governments to sell public goods like spectrum because it leads to the person who values the spectrum the most, the person with the best business plan, getting it," Bulow says. "It is better to let the best plan be determined by people who put their money where their mouth is than some government bureaucrat saying this plan looks better than that one."

Continue to "The Auction Architects: Part2"