The recent intelligence blunders committed by U.S. agencies would be laughable if the results weren’t so tragic: FBI field agents’ unheeded warnings; intercepted phone calls stalled in processing; precious tips lost in “chatter.”
Four professors from the University of Iowa’s Tippie College of Business believe these sorts of mistakes can be cured with capitalist theory and software. Their newly formed company, called Martek Technology Systems, sells software that employs the basic trading infrastructure of Wall Street to create information markets that trade shares in unusual commodities such as political candidates or box office receipts.
While Martek’s main line of business is aimed at customers who want to use these online markets to forecast business performance and open up supply lines, some experts, particularly those in the U.S. defense community, believe this approach may be useful for predicting the next terrorist attack.
For years, mathematicians have known that large groups of individuals are fairly adept at presaging outcomes, from company performance in the stock market to sports betting pools. In 1988, Tippie researchers created the Iowa Electronic Markets (IEM), a system by which people could buy and sell shares in U.S. presidential candidates, producing a more precise result than any pre-election polling data, demonstrating that this method can likely be applied to a wide range of industries. Subsequently, the IEM has set up markets that have predicted box office revenues, the performance of high-tech stocks, and the Federal Reserve bank rate-all with admirable accuracy.
Now this concept of trading on the future has made its way into the defense arena, catching the eye of the Defense Advanced Research Projects Agency, Martek’s first client. The agency is interested in using information markets to predict whether defense contractors will deliver projects on budget and on time. In other words, says Martek co-founder Joyce Berg, a contractor’s employees could go on the Web and delineate the status of current projects, enabling other employees to more accurately bid on whether a project will meet the objectives set out for it. “There are political ramifications inside a firm that keep information from flowing,” says Berg. “People in an organization have information-like the faulty O-rings for the Space Shuttle Challenger-that never makes it to the top. A market is, in a sense, more democratic. You vote with the strength of your belief in your trade.”
Cornering the Market on Terror
But DARPA aims to use information markets to do more than simply keep contractors in line. According to Michael Foster, program manager for DARPA’s Information Awareness Office in Arlington, VA, the next step is to determine whether or not they can predict political unrest in different regions of the world.
“It’s quite likely there are people in parts of the world, like embassy employees, academics, etc., who have information about that region that’s not held in any small group of security experts,” says Foster, whose office recently sponsored a two-day workshop on using markets for decision support. “We want to tap into those broader sources of knowledge.”
Information markets can collect such knowledge more quickly and easily than traditional methods, says Ely Dahan, assistant professor of marketing at UCLA’s Anderson School of Management. Dahan has used information markets to conduct product research. In one market, students bought and sold shares in different sport utility vehicles using a Web-based trading interface that rated each model on features like gas mileage, horsepower and size. Dahan used the prices of each SUV’s shares to determine the product features traders valued most. He says similar techniques could be applied to help prevent terrorist attacks.
“Imagine if FBI agent Coleen Rowley had access to such a market, and the minute she noticed the Moussaoui information she created a security called the Al Queda Threat,” muses Dahan. “Let’s say the payoff is $1 if the threat becomes true, and zero if it doesn’t. The value of the security is the expectation the threat will become true. If the security is trading at 10 cents, the market predicts there’s a 10 percent chance it will happen.”
People with access to inside information about Moussaoui-such as intelligence personnel and others well versed in terrorist activity-begin buying and selling shares in the Al Queda Threat. If the share price suddenly climbs to 40 or 50 cents, it becomes clear that traders in this market think something is about to happen. At the very least, says Dahan, “it would trigger a signal to say let’s pay attention to this one.’”
Such a scheme would have to solve huge ethical and legal problems first, says Foster. For example, since research shows markets are more accurate when traders are given an incentive, how would they be compensated? An even thornier question: how to keep a terrorist from rigging the market by fulfilling the prediction? Then there’s the secret nature of the information itself. “Do we want the public-which includes Al Queda-to know the probability of the Moussaoui threat?” asks Dahan.
Researchers are still studying whether such markets are feasible, says Tom Rietz, another Martek principal. “It’s difficult to say what this approach might work for, what questions it might answer. We have no real predictions’ at this point.”
But for intelligence agencies that lose faxes and ignore phone calls, anything could help.