Thursday, April 28, 2011

What Can We Learn from Intrade?

Intrade is a website that allows users to place bets on whether certain future events will occur. It is most known for allowing people to make bets on the outcomes of elections. Its tagline reads, "The World's Leading Prediction Market." It's often claimed that by studying how the overall market judges the likelihood of events, one can make accurate predictions. But is it true?

The short answer is "No." The slightly longer answer is "No, not until shortly before the election."

Elephant Watcher has observed Intrade activity over the last few elections. Intrade markets tend to do a poor job of predicting the outcome of elections. Worse, they are inconsistent. Typically, the market will make an incorrect prediction over a long period of time and, shortly before the election, it will shift radically in the proper direction. In other words, Intrade is usually no better at predicting an election than a Gallup poll. This shouldn't be surprising, since public polling data is what most Intrade investors use to make up their own minds.

But Intrade can occasionally serve as a reflection of conventional wisdom. Over the course of the Republican 2012 primary, Elephant Watcher will periodically take a look at what the Intrade market has to say about the primary. Intrade's market for the primary may be found here.

On Intrade, candidates with very low chances of winning tend to be somewhat inflated, because the potential payout is so large: If an event occurs, the price of the "stock" goes to 100, so a candidate with a 1% chance of winning is like a 99:1 bet.

Candidates with a greater chance of victory have more stable "prices," and can be interesting to watch. In a future post, we will take a closer look at who Intrade is giving the best odds, and why they are right or wrong.