Prediction markets accurate in elections

One person predicting the future is prophecy, but thousands together is economics.

As Super Tuesday, when over 20 states will vote in this year’s presidential primaries, approaches, the concept of using prediction markets instead of polls to forecast electoral outcomes is gaining credence among economists at Yale and other universities.

“So far, prediction markets have beaten polls in every large-scale comparison” and may one day replace traditional polls, Justin Wolfers, an assistant professor at the Wharton School at the University of Pennsylvania, said in an e-mail.

Prediction markets apply the wisdom of crowds to elections, business, the economy and even humanitarian aid to predict the future by creating monetary incentives for participants to collect relevant information and apply it to speculate about an outcome.

Prediction markets have skyrocketed in popularity since the beginning of the campaign season. CNN recently opened its own political-prediction market, which allows participants to place bets on questions ranging from “Who will be the Republican nominee for president in 2008?” to “Who will win the Georgia Republican primary?”

But experts are skeptical of CNN’s market because it does not allow participants to put money behind their wagers, said Yale Economics professor Ray Fair, who has written about the usefulness of markets in predicting elections.

Other Web sites do allow users to bet money on outcomes. Intrade, which currently runs many of the most popular markets, usually offers $50 contracts, promising to pay out $100 if the prediction attached to the contract turns out correctly. The contracts are then traded on an open market among the users of the Web site.

The price of the security is also the probability assigned to a positive outcome. For example, McCain’s Florida security was trading at $55 dollars when polls opened, while Romney’s was at $45; the market predicted that McCain had a 55-percent chance of winning. By 6 p.m. on the day of the primary, two full hours before polls closed in Florida, McCain had an 80-percent chance of winning.

Although some pollsters had deemed the Florida race too close to call, Intrade’s market accurately predicted a McCain victory.

Prediction markets may often beat polls, but they are not perfect. Like pollsters and pundits, they predicted that Senator Hillary Clinton LAW ’73 had a one-in-10 chance of winning the Jan. 9 New Hampshire primary, in which she eventually emerged victorious.

Once a candidate is sworn in as president, prediction markets may be able to help him or her make policy decisions. In his recent book “Information Markets: A New Way of Making Decisions,” Paul Tetlock, a visiting professor at the Yale School of Management, argued these markets can be used to select the most effective ways to improve health care.

“We don’t know how effective a particular program, such as a vaccination program for children, will be before we implement it,” Tetlock wrote in an e-mail. “Prediction markets, based on the outcomes of policy programs … can help policy makers direct their resources to the programs with the best chances of success.”

As of Sunday evening, Intrade’s markets gave Clinton a 60-percent chance of receiving the Democratic nomination for president and assigned McCain a 90-percent chance of becoming the Republican nominee.

Comments

  • A. Michael Bussek

    My prediction is online since 15th of January and I and my readers made quite some money with it until now - this will going on and on and on since it is a looong way down to 4th November: http://preeedict.blogspot.com/

    A good prediction is pure ca$h!