This has great entertainment value for political junkies, but do online betting parlors really have predictive value? Actually, yes, and they offer a nice lesson in how futures markets work. Such markets, which allow individuals to bet on the likelihood of future events, are based on the theory of "the wisdom of crowds"--the idea that a large group of individuals can predict outcomes better than an expert. Or as F.A. Hayek put it: The myriad data necessary for predicting or directing economic outcomes are "not given to a single mind." That's why few stock pickers can consistently outperform the market as a whole.If you check out Tradesports today, the GOP would lose four seats in the Senate (Montana, Ohio, Pennsylvania, and Rhode Island) but they would keep the seats in Missouri, Tennessee and Virginia. They would lose 19 to 24 seats in the House.
There are plenty of validations of the reliability of betting markets. A study in the American Economic Review once found that the futures market for orange juice is a better predictor of weather patterns in Florida than weathermen, because they have so much money at stake in coming up with the right answer. The predictions markets select the Oscar winners correctly much more often than supposed insiders.
So it is with political betting markets. "It has been empirically proven," boasts Mr. Delaney [CEO of Tradesports], "that Tradesports beats the pollsters 80% of the time." The day before the 2004 elections, Tradesports correctly picked the Senate winner in 33 of 34 races. Unlike polls, which are snapshots through the rear-view window, election markets occur in real time, with the odds rapidly changing to incorporate new and relevant information--even from "insider traders."
However, I wonder if there is really more insight here or just a reflection of what the polls say. I would be interested in seeing a comparison of races, perhaps from 2004, when the markets predicted something different from the polls. Tradesports is proud of having picked 33 of 34 Senate winners. How many Senate elections the day before the election in 2004 were picked right by the polls?
In the smaller markets, I would also be more wary about the effect of one or two people buying a large number of shares. We've bet on several things at Tradesports and both lost and won. It was, however, a lot of fun and made following stories such as predicting the margin by which Alito would be approved much more engrossing. If you're optimisitic about how the Republicans are going to do this November, there are some real opportunities for you. But do you want to deal with the disappointment of a bad election day coupled with having lost money?