The proponents of prediction markets claim that that society would derive valuable information about the future by looking at the trading rates of these markets. But would this information be of any use in coordinating long term policy?

Currently, Ideosphere has an Immortality by 2050 contract trading at 15%. This is absolutely ridiculous; I don't think anyone seriously believes that those of us who manage to eke out another 45 years will be immortal.

So why don't I short the contract? Because I'd make 15% in 45 years. Not a terribly good expected return.

Generally, if we were all rational agents who believed the immortality-by-2050 proposition is false, and if we also believed - quite rationally - that we could get about 10% per year on stocks over the next 45 years, no one would short the contract and the information derived from the trading rate would be completely useless for the next, oh, 43 years.

The point is: in the face of a 10% return on stocks and a 5% return on bonds, idea markets are pretty useless for predicting trends that take more than 4-5 years to materialize.

Other howlers on Ideosphere: a 25% chance that the Democrats will not recapture the house before 2047 and a 40% chance that the US will have proportional representation by 2012.

Update: OK, after some browing on the Ideosphere web page, I found out that the example here is rather silly: Ideosphere does not trade with real money - the idea is that every trader gets an equal amount of fake money to begin with so that traders who get things consistently right dominate the market.

The point stands for markets that use real money though.