Political Futures Markets in the New Hampshire Primary:
Daniel Gross has a short item on this topic at Slate. He concludes that "the price movement tends to respond to conventional wisdom and polling data; it doesn't lead them." Thanks to Matt Bodie for the link.
Isn't this obvious? The market doesn't define the price. The price is emergent from the beliefs of the participants--who individually cannot 'set the price' but who collectively do.
It had value as a measure of people's feelings, which shouldn't be discounted.
But prediction? Pshaw. That's kaput.
I've long been skeptical about the "Magic" of the market. Like any other kind of magic, it's either illusuory or contrived.
I'm throwing this out there, but I would be interested in a rebuttal.
Certain markets, like the stock market, tend to draw the attention of fewer, better-educated people, and can be expected to be more accurate. When you're thinking whether to invest your client's millions in a corporation, you had better know all about the corporation before you buy.
The broader the market, the less informed its members, the less accurate the market. Ditto how little one has riding on it. Few of us really have much invested in who the Democratic or Republican nominee is.
All this seems obvious to me, but people talk about "markets" like it's a magic word or something. Economics describes human behavior, it doesn't improve or recommend it.
The freaky thing is, Intrade was lagging the exit polls.
I saw on CNN how Clinton had a 39/36 edge over Obama, for well over 20 minutes. Intrade still had Obama at over 80%.
I haven't been using Intrade long enough to trust my money to it, but it seemed like Intrade was ignoring the fact that the winner was going to be a toss-up. At least selling Obama down to 60% or so would've been reasonable, especially with the networks declaring the content "too close to call." ... and that market movement did happen, but over the next hour or so, surprisingly slowly.
In other words, the markets were completely blindsided by Hillary Clinton's surprise victory. I follow them because they're by far the best way to turn the conventional wisdom into a single number, but they don't seem to have any special insight into the future.
Also, in a sense markets are smarter than the people in them. It takes very few well-informed people to allow markets to track available information closely. The average trader on intrade could be an idiot and it still could work well. There was a recent article on futures markets that claimed they were most efficient when there was a mix of idiots and well-informed people. (Without the idiots, there's no incentive for the well informed to play.)
There's nothing magical about futures markets, but they can potentially digest available information very well.
If Krugman had the mental discipline to translate his "insights" into specific predictions for, say, the Treasury futures market, people would see how vacuous and self-contradictory most of what he has to say is. Don't hold your breath.
You've obviously never watched Jim Cramer's TV show.
I agree that there are ridiculous claims out there about the power of markets, but I don't see how the markets-are-better-than-polls hypothesis was just tested in the way you suggest. As I understand it, the polls were pointing to an Obama win too. Even if the polls had favored Clinton and intrade predicted Obama, that would be very weak evidence against the hypothesis.
Also prediction markets advocate and expert, Robin Hanson, responds.
The futures prices can be more or less interpreted as probabilities. If you can buy an Obama future at $0.75 that would be worth $1.00 if he wins, you'd want to buy if you thought there was more than a 75% chance of Obama's winning. (This is only more-or-less true; it would be true if participants were risk neutral, i.e. maximizing expected returns.) Of course, a market participant is only guessing what the probability is, but that's the point. If enough of the guesses are well informed, then the market's a good predictor.
There is no poll that tells you the likelihood of a 6-point swing, but the market participants (or at least the serious ones) try to gauge the probability as best they can with the available info.
I don't know who bets on intrade, but I'm sure a lot of them go with conventional wisdom and/or their own wishful thinking. Here's the thing, though. Anyone with better information than the market as a whole can make money on intrade. Let's say a new poll comes out solidly in favor of candidate X. The talking heads take it at face value, and conventional wisdom goes with them. You, however, being particularly well informed, notice that the poll relied on people with land lines who were classified as likely to vote based on an old formula that shouldn't work well in such a hotly contested election. Your well-informed guess is that the poll was biased toward candidate X. You notice that intrade prices have spiked, along with conventional wisdom, for candidate X. You can make some money. You, and other well-informed participants, sell short on X, and the price of X falls. The intrade price of X is now less than what conventional wisdom would predict because a few well informed participants keep a check on the market in their own self interest. Again, the vast majority of traders can be going solely on hunches or conventional wisdom. It only takes a few participants who are well informed.
Of course exit polls affected the intrade market. Assuming the intrade market is working well as a predictor, it's going to update with all new info. Exit polls are extraordinarily convincing new information. It's true that conventional wisdom moved in the same direction as intrade in response to the new information, but unless you're taking on the straw man with the crystal-ball theory of market prediction, this should be no surprise.
I shouldn't have labeled the crystal-ball theory as a straw man. You're right that a lot of market enthusiasts actually have some vague notion that market predictions are magical. It's just that I'm not saying they are.