In the Autumn of 1955, Vernon Smith, a young Economics professor from Purdue University, was having difficulties explaining basic microeconomic theory to his students. So on the first day of the following semester he tried something new. Instead of teaching his students what economic theory had to say about the way markets work, he would run a simulation of the stock market in the classroom itself. Smith split the class in two. One half would play the role of buyers, the other would play the role of sellers. Smith assumed his little classroom experiment would teach students how competitive price theory really works.

There was only one problem: the experiment didn't work.

If economic experiments can be used to predict how buyers and sellers might behave in a given market, why not use experiments to predict other types of events? That's exactly what the folks at the Pentagon's Policy Analysis Market (PAM) had in mind when they created an online market for predicting future events. If analysts already use changes in the price of oil as indicators of future events in the Middle East, the hypothesis went, why shouldn't they create a dedicated online market to assess the likelihood of future political and military events? The market would enable people to make bets on, say, whether the House of Saud will fall in the next 12 months. If the price of the underlying contract suddenly moved upward, that would mean that "the market" had reason to believe that something dramatic is about to happen in Saudi Arabia.

Unfortunately for the people behind PAM, not everyone in the US Senate thought it was such a good idea. "I think this is unbelievably stupid," said US Senator Byron Dorgan. "It combines the worst of all of our instincts... it is a tragic waste of taxpayers' money, it will be offensive to almost everyone." Dorgan wasn't alone. Other US Senators called the program anything from "grotesque" to "an incentive actually to commit acts of terrorism." Aside from the moral implications of operating such a market, the opinion of the Senate was that "trading on corn futures is real different than trading on terrorism and atrocity futures."

But is a market for future terrorist events really that different from any other predictive market?

Consider the following: The Iowa Electronic Markets, which was set up in 1988 to predict local and federal elections in the US has consistently outperformed any other forecasting tool, including general polls, expert polls and historical extrapolations. Or consider the Hollywood Stock Exchange, an online site that lets you bet on how much money a certain movie will make in its first month. Hollywood studio executives frequently use information gained from the Hollywood Stock Exchange to decide whether they should allocate more or less marketing dollars to promote a movie. Shrek 2, for example, is now trading at $197. This means that the market estimates it will bring in $197 million in its first month. If you're in charge of setting the advertising budget for Shrek 2, this information could be incredibly valuable, considering the fact that the movie isn't even due out until May 2004.

If the collective wisdom of a crowd can help a movie studio predict the future success of a cartoon movie, why can't it be used by a pharmaceutical company to predict which drugs in its pipeline will bring in the most money, or by a toy manufacturer to decide how much advertising it will need to invest in order to maximize sales of a new action figure?

This is exactly the idea Mikael Edholm had when he founded Predicom, a new venture that will use predictive markets to help companies improve their forecasting. Edholm, Ericsson's former director of marketing & strategic business development, is no stranger to predicting the future. At Ericsson he spearheaded a multi-year study to develop multiple scenarios of what the future might look like. Edholm and his colleagues now believe that predictive markets can help companies predict a wide array of future events, from next quarter's production bottlenecks to whether releasing a new consumer product will be an instant hit with customers.

Take sales forecasting, for example. Ask your average salesperson how much they forecast they'll sell by the end of next quarter and they'll answer "How much is my quota?". Given the politics and rivalry found in most firms, the likelihood of getting a realistic sales projection is incredibly low. But imagine you had a predictive market set up for your sales team. Each salesperson would log on, anonymously, and bet on what their region's sales figures would look like, with whoever comes closest to the actual figure winning $500 at the end of the quarter.

It turns out that just like traders in many other markets have no incentive other than making the right prediction, chances are that the results produced by your company's sales prediction market will, by far, outstrip any other forecasting mechanism you currently use.

The idea that predictive markets can help firms become better at forecasting is hard to digest at first. That's because we don't normally associate making bets and speculating about future events with better decision making. But what's important to remember about Vernon Smith's experimental economics and the world of predictive markets is that they simply offer us a tool for removing traditional limitations on how information flows within an organization. And as organizations become more and more interconnected, tapping into new forms of distributed intelligence will change the way we work, live and play.

--> See what the Policy Analysis Market would have looked like, had it not been shot down.
--> Check out the Iowa Electronic Markets.
--> Visit the Hollywood Stock Exchange.
--> See the Foresight Exchange Prediction Market, which enables you to bid on a host of future events.