| OVER THE YEARS, ECONOMISTS HAVE NOTED THAT the orange juice futures market has been a better predictor of Florida weather changes than meteorologists’ forecasts. And the futures markets for oil, wheat, and other commodities have predicted prices far better than any individual analyst.
Those results are not accidental. Markets are not only about goods and services but also about marshaling and using information. As Professor Charles Plott of the California Institute of Technology puts it, “The markets can know more than any individual knows because the market is collecting and integrating information.” In recent years economists have sought to capitalize on the ability of markets to distill disparate strands of information by creating “information markets” or “decision markets” in which buying and selling are facilitated largely to display what the participants think rather than to hedge risks or make a profit.
“What markets have always done as a side effect is aggregate information,” notes Professor Robin Hanson, an economist at George Mason University, “and recently people have said maybe we can do this on purpose—maybe we could make a market in something because we want to know about it.”
Robert Forsythe, who got his Ph.D. at CalTech, helped set up the Iowa Electronic Market (IEM) at the University of Iowa. This year, Americans will see repeated references to the IEM, a futures market in which people buy contracts on the candidate they think will be elected president. As in any futures market, the prices of the contracts rise and fall over time, depending on what purchasers
believe the outcome will be. The IEM, which has been around since 1988, has outperformed most opinion polls as a predictor. It was initially open only to students and faculty and was used mainly as a teaching tool, but now anyone can open a trading account online (www.biz.uiowa.edu/iem), deposit up to $500, and begin to buy and sell contracts. These days, there are some 5,000 participants.
One of the two main presidential markets is the “winner take all” market: Participants buy and sell contracts that pay a dollar each if the candidate invested in is elected president, while contracts on the losing candidates expire worthless. “At any point in time, you can use the price of the contracts as a prediction of the probability that each candidate will be the winner,” Forsythe says. If one
candidate’s contracts are dropping in price, that’s a sign that futures investors are losing confidence in his or her chances of winning the election. There is also the “vote share” market, where the payoff
on each candidate is proportional to the share of the popular vote he or she gets; thus, if George W. Bush shares are trading at 53.4, the market is anticipating he will get 53.4 percent of the popular vote.
According to one analysis of the last four presidential elections, the Iowa market was compared with 600 surveys, and its results were closest to the election results 75 percent of the time. “It’s been demonstrated that the price system in a futures market does a pretty good job of aggregating and disseminating the pieces of information that are in traders’ heads,” Forsythe says.
While the IEM attracts national attention, Forsythe adds, “The markets we’ve been least successful with are ones that you could probably describe as ‘thin markets.’” In a market on a congressional election in Utah, for example, he says, “The 30 students in the class that set it up were the only ones willing to trade.”
Iowa runs markets in other events besides elections. One market covers Federal Reserve monetary policy. Another targets box office returns on newly released motion pictures. This is used in marketing classes to help students learn about forecasting revenues for new products.
Hewlett-Packard and half a dozen other major companies are using internal information markets to assemble forecasts of next-quarter sales and other variables. They, too, are finding that information
markets do a better job of predicting than do the opinions of individual experts or broad-based surveys.
Proponents cite several reasons for the highly prescient nature of information markets. Different people know different things, and markets can compile information, while surveys can’t. Maybe you don’t have a clue how far it is from New York to Duluth, but if several people get involved and one knows how far it is from New York to Chicago while another knows the distance from Chicago to Minneapolis, you end up with a far better estimate than if you just guessed. Markets facilitate
the aggregation process by enabling people to express the information they have through buying and selling.
Moreover, if you’re just responding to a survey, talk is cheap, Hanson says, but “you can be sitting around the dinner table talking, and as soon as someone says, ‘You want to bet a hundred bucks on that?’ people shut up.” Markets get people to put their money where their mouth is, even if their trades only involve $20 or $40. However, the Hollywood Stock Exchange, a market that offers trading in Oscar nominations and movie grosses, and the Foresight Exchange, a market that trades in a wide range of topics, permit people to buy and sell contracts without using real money.
Although there could be bubbles and irrational exuberance in information markets, their finite time periods limit that: Most are focused on events that quickly come to pass—or don’t—and then the market is over.
While the structure of information markets is closest to futures and options markets, the information markets idea also is similar to betting. The flow of money into wagers aggregates into a fairly accurate picture of the likely outcome of an event. The changing odds or point spreads offer a good reflection of the probabilities of various outcomes.
Thus, it’s not surprising that a major information market is run by Tradesport.com, an Irish betting shop. It not only takes wagers on major sporting events, but also on elections in major countries, and it has made markets in such things as “Next Supreme Court justice to leave the bench.” Reflecting its
hybrid nature, the Tradesport Web site offers participants a choice between a betting screen and a trading screen.
Somewhere between the bookies and the academics are two other play money markets: the Foresight Exchange (www.foresightexchange.com), which Hanson helped create while he was a graduate student, and Newsfutures.com. Each lets participants “invest” in contracts on a wide range of news events. And both have good track records in predicting the outcomes of those events.
Given these capabilities, the U.S. Department of Defense sought to bring this methodology to bear on questions of terrorism. Last July, the Pentagon announced plans for a Policy Analysis Market (PAM), an experimental online futures exchange that would have allowed participants to speculate
on political stability indexes. The market was being put together by a San Diego firm called Net Exchange, created by CalTech faculty, with Professor Robin Hanson—a CalTech graduate—as a consultant.
The goal was to improve intelligence forecasting by sharing views on stability, but the prospect of veering off into assassinations and terrorism quickly got PAM under fire as “a futures market in
death.” As soon as senators Byron Dorgan and Ron Wyden held a press conference to denounce the market, the Defense Department scrapped the project. Association with the project cost John Poindexter, controversial retired admiral and former National Security Advisor, his Pentagon job.
Plott says, “The guys at the Pentagon had a really good idea and were pursuing exactly what they should be pursuing in trying to find new ways to solve extremely hard problems. The way they
were treated by the press and the political system was scandalous.” Forsythe, on the other hand, says he had “grave doubts that those markets would actually work.” He adds, “In my mind there was something unseemly about it.” He says PAM could have stuck to more-objective variables,
like oil production, that provide an indicator of political stability.
The debate over the abortive Pentagon market is reflective of continuing efforts to define precisely where and how the information markets idea can be deployed. Ultimately, its proponents say
these markets can be useful anywhere there is a specific outcome—something that people can buy a contract on. It also helps to have a substantial number of people participating in the market.
“It’s only a matter of time before people start using this more broadly,” Forsythe says. Plott believes “it’s going to continue to grow. It’s like any other discovery in science—there’s a long road
from understanding it at a basic level to the ultimate applications.” The Pentagon fiasco is a reminder of that. So the jury is out on how far and how fast this idea will take hold. If there were an information market that was trading information markets futures, we might have a better idea. 
|
