President Obama? Web sites bet it's a done deal

After the votes were tallied on Election Day four years ago, the big winners turned out to be the betting Web sites that predicted George W. Bush's re-election.

U.K.-based Betfair correctly predicted that Bush would stay in office and gave him 2-to-1 odds of beating his Democratic rival, Sen. John Kerry. The odds at Dublin-based Tradesports.com were similarly accurate, giving Bush a 58 percent chance to win and Kerry a 42 percent chance.

This year, the predictions are far more dramatic. Betfair's national election chart puts the probability of a President Barack Obama at 94 percent and the probability of John McCain winning at just 7 percent. Obama's odds are up from below 90 percent last week.

What a change a few months make. In May, Obama and McCain were at rough parity. In mid-September, they were almost at parity--in other words, even odds--after Sarah Palin was nominated as the Republican vice presidential candidate. Then the odds of an Obama presidency broke sharply upward, roughly at the same time the financial crisis hit and the bailout bill was approved.

Similarly, Intrade puts Obama at about 92 percent, and McCain at about 9 percent. Other relevant odds: a 36 percent chance Obama will win 370 or more electoral college votes; an 80.7 chance of a U.S. recession in 2008; and a 79 percent chance of substantial income-tax hikes next year.

In general, betting exchanges tend to be surprisingly accurate in their predictions. One explanation is that it's easy enough to offer predictions when being wrong means only potential embarrassment for a blogger or would-be pundit. But when there's a financial penalty for making a mistake, results tend to line up far more closely with reality.

Well, most of the time. Sometimes betting markets--also called prediction markets--do a better job of reflecting conventional wisdom than offering a glimpse of the future.

During the 2008 New Hampshire primary, Obama attracted more enthusiastic crowds than rival Hillary Clinton. Intrade put his odds of winning at a remarkable 54 percent, while a traditional pollster said Obama would get only 28 percent of the vote.

We know how that turned out: Clinton won the state. (Perhaps more to the point, as recently as winter 2007, Intrade had Obama at something like a 10 percent of winning and Rudy Giuliani at 20 percent.)

While Internet-based betting may be novel, political odds-making enjoys a venerable tradition. Two economics professors, Paul Rhode and Koleman Strumpf, once calculated that betting markets predicted the popular-vote winner in every presidential race but one between 1884 and 1940. (In that 1916 race, gamblers were offering even odds between Woodrow Wilson and Charles Hughes.)

Rhode and Strumpf are working on a paper--a publicly available version (PDF) is dated August 2008--that looks at the international history of betting markets. It says election betting was popular in Italian city-states as far back as 1500. And in the United States during the 1830s, politicians including Martin van Buren either bet on themselves or encouraged others to do so.

"While it is sometimes claimed that political betting markets are a recent invention, they clearly are not," Rhode and Strumpf write. "Rather, it is the absence of such markets during the mid- and late-20th century which is the exception."