Yesterday's NY Times reports that, "Polls Say Kerry. Futures Say Bush". It appears that Bush futures have a slight lead in both of the the major political futures electronic markets, the one at Intrade and the one at the University of Iowa -- and both are contradicted by various election polls.
The article, by Daniel Gross, reports that the electronic markets have the better track record in getting presidential election outcomes right. Gross notes that bets on futures markets are by people who are putting their money where their mouths are. He then stumbles by noting that, "The people who trade securities online look a lot more like the crowd at the '21' Club -- most of them affluent white men -- than the voter registration lists."
Huh? Are PC demographic descriptors relevant on any futures market? If so, there is money to be made.
I used to be confused when political pollsters were identified by their party affiliations and leanings. I thought that they were in the business of discovery. So what does it matter what their politics are? A little reflection suggested that their views do matter because polling is really about focus groups and testing campaign themes. It has much less to do with forecasting. Now that advances in communications and funds transfers have made more and more futures markets possible, the forecasting function of political polling is probably on the way out.