Ezra writes:

I’m not much of a betting man, but as some smarter folks pointed out to me over e-mail, Hillary Clinton’s futures are tremendously overvalued. I know, I know, the crowd is infallible, James Surowiecki told you so, but after the week she’s had, and the way the polls look, I really wouldn’t put her “stock” at $61, while Obama’s only at $33. If someone wants to make some money, they should place a bet. And if Petey’s really so sure about John Edwards winning the whole thing, well, there’s a ton of cash to be made there.

I get on Ezra’s case every now and then when he occasionally expresses market antipathy, and to a certain extent I could do that here, too. For one thing, 61 percent isn’t all that impressive, given her perceived status as clear front-runner. For another, that price has come down in the past few weeks. A month ago, Hillary contracts enjoyed a 60 point lead over Obama contracts, while the lead is now less than 30 points. Finally, a lot of the negative news in recent weeks is the kind of stuff that might resonate with Washington types or maybe even with early primary voters, but it doesn’t necessarily trickle down into the consciousness of voters at large.

On the other hand, Ezra may have a point. Back in 2006, lots of people ridiculed prediction markets for failing to predict the Democratic takeover of the Senate, despite accurately predicting the outcomes of all the Senate races. Defenders of the markets, including myself, pointed out that since no information about control of the Senate was available that wasn’t already taken into account in the individual contracts, bettors could only look at the combined probabilities of the individual races to get the probability of Democratic control. Those races all being close, that probability was not sufficient to cross the 50 percent threshold.

Something vaguely similar may be happening here. Clinton is running close in the early primaries but has commanding leads in many of the later ones. If there’s little information out there about the probability of nomination other than the outcomes of individual primaries, then a bettor might look at the predicted outcomes of all the primary contests and conclude that Hillary has it in the bag.

The problem with this is that trading volume is not very high in later primaries. As a result, bid-ask spreads are huge and prices could shift rapidly. So the best information in those later primary contracts gives a big edge to Hillary, but that’s not telling us all that much. Given that Hillary’s leads are substantially smaller, or nonexistent, where the action currently is, it’s probable that the nomination contracts do overstate her odds.

What should be clear, however, is that political pundits and wonks could benefit from more accurate market prices, and they could contribute to more accurate prices by betting when they believe the prices are wrong. If Ezra bets against Hillary and is right, then his action will move prices in the right direction. If he bets against her and is wrong, then by pushing the price in the wrong direction, he provides profit opportunities for those who know more than him, and who should then come in to push the price back in the right direction.

I’ve argued that economists ought to subsidize prediction markets by trading, whether or not they know much about what they’re doing, because good market prices are such valuable tools for academics. The same is true for political markets. So get in there, Ezra. It’s for science, after all.