September 18, 2004 | Financial TimesPredicting the next president of the United States is an inexact science and in a close election opinion polls may not offer definitive clues about who will win. Among 12 incumbent presidents seeking re-election since 1936, eight of the nine winners were ahead of their opponents just after Labor Day, according to the Gallup organization. Two who went on to lose were lagging after Labor Day, while the third, Jimmy Carter, saw his lead fade. But some find it instructive to follow the money. Political futures markets and online betting sites historically claim to have been better predictors of political races than polls. IOWA ELECTRONIC MARKETS, a real-money futures market, claims its predictions have proved more accurate than opinion polls 76 per cent of the time. Its vote-share market shows a final, election day split of 49.3 percent for Kerry and 50.7 percent for Bush -- much closer than most polls indicate. "It's going to be a close race," says JOYCE BERG, co-director of IEM and an associate professor of accounting at the University of Iowa. "There is still a lot of uncertainly about what the outcome would be." An essential difference between opinion polls and markets is that traders put their money on who they think will win the election, versus who they want to win, Berg says. "Our traders have an incentive to put their money where their mouth is. Traders can also express the strength of their belief where in polls everyone is equally weighted."
Contact: Joyce Berg