The IEM and the wisdom of crowds

I was going to post something about the Iowa Electronic Markets (IEM) and the presidential primaries, but this post on the Blog of the Long Now is much better than you would have gotten from me! Quoting from the blog post's introductory explanation:

The Iowa Electronic Markets (IEM) trades real money in bets on future cultural events. Prediction markets sell contracts for an easily decidable future event (someone wins an election, a commodity hits a certain price, a movie sells X number of tickets). If that contract comes true, the different between what the price of the contract was when bought and its full price will be pocketed by the winner. So if $1 contracts for Bill Clinton winning the election were selling for 35 cents when you bought them, you would have profited 75 cents [65 cents?!] after the election.

Outside of the IEM most US predictions markets trade with token money to escape uncertain gambling and investor laws, which may prohibit real money trades. But in all prediction markets the price of a contract is decided by the collective demand, or in other words, by the collective mind. Outcomes with low expectations earn a low price. If you are a contrarian you can buy low value, low-expectation contracts, and if conventional wisdom is wrong (at that time), you’ll gain.

But the odd thing is that in these markets, the conventional wisdom of the crowd is usually right.

It’s probably worth noting that the Presidential Election graph shown in the post is the “Vote Share” market (which currently quotes prices of .527 DEM and .488 REP). However, the “Winner Takes All” market has held at about 60-40 (DEM-REP) since early April 2007, though it has tightened a bit recently, to 58-42.