February 5, 2008 — The quality of your future wireless access to the Internet is being determined right now as the Federal Communications Commission auctions off sections of radio spectrum well-suited to becoming a future superhighway of access to the Internet. Or possibly something else.
What exactly this "beachfront" spectrum is used for will be determined by the auction bidders — rather than decided by political lobbying or even sophisticated economic theory of how it would bring the most revenue — thanks to a new auction design co-developed by Charles A. Holt, the A. Willis Robertson Professor in Political Economy at the University of Virginia, and Jacob Goeree, an economics professor at the California Institute of Technology.
The bidders include heavyweights new to the wireless business, like Google and Cox Cable, alongside entrenched companies like Verizon and AT&T.
The 700-megahertz spectrum being auctioned off is highly prized for several reasons. A 700-MHz frequency signal can penetrate walls for improved indoor reception, and each tower broadcasting in this range can cover at least four times as many square miles as conventional cell-phone towers. This means fewer towers, and thus less expense, to any potential bidder, making the so-called "C block" an express route to creating a new national or regional network, Holt explained.
"This spectrum is like the interstate highway system for the Web marketplace of the 21st century," he said.
Those qualities are why this spectrum is currently being used for TV broadcasts. However, television stations will vacate these airwaves in February 2009 as they move from analog to digital broadcasting.
The C Block of spectrum — the largest single swath of spectrum offered in this auction — comes with open-access requirements requested by Google, such that a network must be "open to any devices and services." Advocates hope that requirement will promote innovation, improve services and enable a third major avenue of broadband Internet access (along with DSL and cable), one that wouldn't be controlled by any one company.
Holt and Goeree's novel auction design, known as "hierarchical package bidding," is being used by the FCC to sell the C block, allowing the block to be offered for sale in two separate configurations: under one national license, or divided into eight regional licenses that can be bid on separately or in combination. If the sum of the eight regional auction bids eclipses the top bid for the national license, the spectrum will be sold in those pieces. Otherwise, it will be sold to the national bidder. Either way, the spectrum is expected to fetch billions of dollars.
These auction rules "let the market discover the optimal allocation, so the market decides how things get packaged, which takes the heat off of the public officials who must respond to lobbying pressures on these issues," said Holt.
Or at least that's the theory. The new auction design has never before been used in an auction involving billions of dollars.
The FCC auction has been under way for about two weeks, and thus far it is operating exactly as Holt and Goeree predicted it would.
The leading bid for the C block has already shifted back and forth at least twice between the cumulative prices for the eight regional licenses and bids for the national license, demonstrating that the market is determining how the spectrum will be most valued. As Holt put it, "the lead has switched about the same number of times as in the final quarter of the Super Bowl."
As reported by the New York Times, bidding on the national license reached $4.71 billion on Jan. 31, then was trumped Feb. 4 by bidding on the eight regional licenses, which totaled $4.73 billion — both totals well above the $4.6 billion reserve price for the C block. After less than two weeks, the overall bidding is at about $19 billion, almost double the total reserve prices of about $10 billion. The auction will continue until the bidding is exhausted.
The rules of and effectiveness of hierarchical package bidding were refined and proven by rigorous testing in lab experiments involving more than 300 students at Caltech, carried out over a six-month span. Groups of five to seven students placed bids in simulated auctions that typically lasted for about two hours.
"Just as you would never consider a new airplane design without wind tunnel tests, it makes sense to run laboratory tests of a new auction design," said Holt.
The use of laboratory experiments to study how people actually behave in economic interactions is a relatively new field of economic research known as experimental economics. Holt has written a textbook on the topic and is "one of the top researchers in experimental economics," noted Bill Shobe, director of business and economics research at U.Va.'s Weldon Cooper Center for Public Service.
"For me as a professor, one satisfying aspect of this laboratory research is its use in teaching," said Holt. "We put students from our classes into the test auction environments and let them discover insights about the public policy and auction design issues for themselves. Economic theory can sometimes be abstract and dry, and the classroom auction simulations provide an active learning experience that brings relevance and excitement to the class."
In experiments similar to those conducted for the FCC auction, U.Va. students have been involved in research by Holt and Shobe on auctions to allocate carbon emissions, commissioned by the Regional Greenhouse Gas Initiative, a consortium of about 10 Northeastern states that plans to begin auctioning rights to emissions in 2009. The students have helped recruit test bidders, run the experiment sessions in a special economics computer lab in Wilson Hall and analyzed the results.
Nearly all of the presidential candidates have called for the creation of a cap-and-trade system for emissions as a major part of their policies to deal with global warming. The observations in Wilson Hall may well influence the workings of future cap-and-trade auctions, said Holt. "These students see the interplay between policy analysis and an ongoing research project in the lab."