Prediction Markets
A political prediction market is a bit like the stock market, except that you are buying shares whose value depends on the success of a political candidate, rather than the profits earned by a corporation.
Experimental prediction markets were established at the University of Iowa in 1988, and they have since repeatedly outperformed polls. (See “Results from a Dozen Years of Election Futures Markets Research.”) Economic historians have also documented the impressive forecasting record of prediction markets in the period before scientific polling was adopted. (See “Historical Presidential Betting Markets.”)
In the 2004 primaries, prediction markets pointed to the disintegration of Howard Dean’s candidacy in advance of the fateful Iowa caucuses. In the 2004 presidential election, the market favorite won the Electoral College in all 50 states; in 2006 the markets also picked every Senate race.
Prediction markets have long suggested a strong showing for Hillary Clinton, even as popular commentators had earlier dismissed her as unelectable, much as they did prior to her successful New York senate race in 2000.
Prediction markets tend to be forward-looking, while polls are often backward-looking. For instance, Fred Thompson continues to do well in national polls largely due to name recognition, while prediction markets have discounted this advantage, understanding that candidates like Mike Huckabee will become better known through the campaign. Indeed the markets currently believe that Mr. Thompson is less likely to win the Republican nomination than fringe candidate Ron Paul.
The markets predicted Mike Huckabee’s surge a few weeks before the polls.
On the Democratic side, national polls suggest a landslide for Ms. Clinton, while the markets suggest a one-in-three chance that Obama or Edwards will ultimately win the nomination.
“Best Bet for Next President: Prediction Markets,” by Justin Wolfers