This Futures Exchange Isn't About Soybeans

IEM trades in election outcomes--with striking accuracy

Al Gore with a 58% chance of beating George W. Bush? Hillary Clinton with a 70% chance of beating Rick Lazio? Conventional polls, to be sure, show the Presidential and New York senatorial elections too close to call. But a fascinating online futures market, followed by political pundits, Wall Street forecasters, and even pollsters, provides a different picture.

Instead of trading soybean or stock index futures, Iowa Electronic Markets has traders bet on election outcomes--winners or their share of votes. Started 12 years ago as a teaching tool by the University of Iowa's Tippie College of Business, IEM has become a serious forecaster of U.S. and foreign elections. It predicted candidates' vote shares in the 1988 and 1992 Presidential elections to within two-tenths of a percentage point, far more accurately than opinion polls. In 1996, it chose the winner but tied with polls on vote totals.

WINNER TAKE ALL. It's also compiling quite a track record abroad. IEM stuck with Boris Yeltsin during his 1996 Rus-sian presidential run and has predicted outcomes in France, Germany, and Turkey--which hasn't escaped Wall Street's notice. "It's a good antidote to the media and many worthless foreign polls," says Martin S. Fridson, chief high-yield-bond strategist at Merrill Lynch & Co.

How does it work? You log on to the Internet site at and open an account--much like an online brokerage account, but you're limited to a deposit of $500. There are no commissions, but there's a $5 signup fee. The most popular markets are winner-take-all: Each share in a winning candidate pays $1, but losing shares are worthless. To get started, you first must buy all the candidates in a race for $1. Then you sell the candidates you expect to lose for whatever price you can get. The price--50 cents, say, for each $1 share--rises or falls depending on how other people are betting. So the price represents the odds of victory.

More than 7,000 traders have poured some $175,000 into IEM so far this year. As of Sept. 27, Gore shares sold for 58 cents and Bush for around 42 cents. In New York's senatorial market, the only congressional race drawing national attention, Clinton is at 70 cents, while Lazio has slid to 30 cents. Traders have bid up Democrats' chances of taking the House to 50 cents, while the cost of a Democrats-take-the-Senate share is just a dime. A share of Republicans controlling both houses runs a tad over 40 cents.

IEM's success has provoked copycats, such as Operating in Washington, it uses fake money and awards prizes. The Iowa exchange "is real people putting their money at risk," says Andrew A. Laperriere, political economist at the International Strategy & Investment Group.

More important is the difference between IEM and polls. Polls typically try to take the electorate's pulse by asking who a participant would vote for if the election were held today. Traders on IEM are betting on an election's outcome. "Most pollsters deny they're in the prediction business. But our market is precisely about that," says Robert Forsythe, senior associate dean at Iowa's business school and one of IEM's architects.

Moreover, players of IEM futures--all political junkies, whether they be grad students or money managers--may know more than most voters. "You probably get the best aggregated wisdom about an election from this," admits Gallup Poll Senior Editor David W. Moore.

Of course, IEM isn't always right. It failed to pick the winner in Mexico's recent presidential election. The exchange did poorly, Forsythe says, because only 12 people wagered, too few to create a viable market. But when there are plenty of players, Wall Street takes notice. "What the pundits will say later shows up in the market graphs now," says Robert A. Rintel, manager of Valgro Mutual Fund and an IEM trader for the past three U.S. Presidential elections.

What about manipulation? Despite the $500 limit, a large group of traders could conceivably gang up on the market to skew results. But Forsythe says prices change little when new money arrives. "The only complaints we've had from politicians," he says, "is that their contracts aren't being traded."

So should Dubya start panicking? Not just yet. Historically, the Iowa predictions are most accurate one week before the election. So he has time to catch up.

Before it's here, it's on the Bloomberg Terminal.