N. Gregory Mankiw, a former adviser to President Bush, on the prediction markets:
Everybody has information from their own little corner of the universe, and they’d like to know the information from every other corner of the universe. What these markets do is provide a vehicle that reflects all that information.
Nice article, I like the closing paragraph best:
Yes. I didn’t excerpt that, I will tell you why. With all due respect to Justin Wolfers (who is quoted on the topic), I wary of the big margin error, two years away from the 08 elections.
Pssttt… And there’s an MP3 file of JW. I will listen to it today, and link to it directly.
Chris:
Your concern about the accuracy of markets two years out seems entirely well placed. But again – and I think it is worth saying this over and over – the relevant question is whether the markets do a better job than alternative information aggregation devices. My own analysis (with Andrew Leigh) of Australian polling data suggests that any polling done more than a year before an election is essentially useless. That is, you are better off simply guessing that the previous election results will repeat themselves, than you are using early polls.
And I should add one more thing: The markets provide a useful measure of our ignorance. For instance, right now, the markets suggest that the Dems are only slightly more likely to win the election than the Republicans. Perhaps this is a forecast that it will be a close election, but perhaps more usefully, it is a forecast that: If nothing else changes the Dems will win, but it seems likely that a lot will happen that we just can’t forecast. As such, prices near 50 are simply a recognition (and quantification) of the uncertainty we face two years prior to a poll.