The growing market for political futures
Intrade, one of the best-known prediction markets
TEXT OF STORY
Kai Ryssdal: We won't know who won the presidential election until tomorrow night sometime. That is, barring something unusual happening. But the free market can give us a pretty good idea of how things might wind up. People willing to put a few bucks on the line give us a sense of the odds in what're called the prediction markets.
Justin Wolfers teaches at the Wharton School at the University of Pennsylvania. He's also an occasional commentator for us. And David Leonhardt is a columnist for The New York Times. Hello to you both.
Justin Wolfers: Hey, Kai.
David Leonhardt: Hello, Kai.
Ryssdal: Justin, I'm going to start with you. Tell us your experience from election night 2006? You've got a little story to tell.
Wolfers: Well, like everyone else, I was watching major networks. The only thing I was doing a little bit differently was also tracking the prediction markets on InTrade.com. It actually turned out to be pretty important, because 'round about midnight, most of the major networks were telling us it was a boring election. The Democrats weren't going to win the Senate, and my guests were about to get up and go home. At about that point, the prediction markets moved very, very sharply, suggesting something was up and the Democrats might well take the Senate. My guests decided to stay, have another drink, and it turned out we were drinking to a new Senate majority.
Ryssdal: David, let me ask you this, though -- Is that really to suggest that this offshore gambling company in Ireland knows better than all the collective wisdom of the pundits in the United States of America?
Leonhardt: I think to some extent, it does know better, and there are a couple of reasons for that. One, it's not just this one offshore betting company. There are a number of these markets and usually the odds move together. The second thing is the pundits aren't actually being honest with viewers. They know a lot more than they're saying. Pundits don't like to talk in percentage terms. They don't like to say at this point there's a 70 percent chance that the Democrats are going to take over the Senate and there's a 30 percent chance they won't. And a lot of us are comfortable with odds and like to hear those percent chances. And so, in some ways we can't get that from the TV. It's not that the people on TV don't have the information that these traders do.
Ryssdal: Justin, you are a professional economist. You study these things for a living. How are these futures markets different than polls, which we've all seen ad infinitum in the past couple of months?
Wolfers: One important difference is that these ask you to put your money where your mouth is. It's not too hard to tell a pollster that you're cranky with a current set of politicians, but you may end up doing something very different in the privacy of the polling booth. I think the more important thing, though, is the question they ask. Polls ask you, "Who is it you intend to vote for?" Here, we're asking, instead, "Who is it you think will win? And it may well be a better way of forecasting elections.
Leonhardt: Kai, I think the really interesting thing this year is going to be to compare the forecasting value of these prediction markets with the forecasting value of a Web site called FiveThirtyEight.com, where a guy named Nate Silver is doing some really sophisticated analysis of polls. And the question is are the traders any better at analyzing these polls than FiveThirtyEight is at doing data crunching from these polls. And so, in some ways, 2008 is going to be a really interesting test of these prediction markets in a way that 2006 and 2004 was not.
Ryssdal: David, we've talked about InTrade; we've talked about FiveThirtyEight.com. Where else would you suggest people look?
Leonhardt: Well, (laughs) . . .
Ryssdal: Go ahead, plug The New York Times. Go ahead.
Leonhardt: Well, certainly, I would encourage them to look at the New York Times. I would encourage them to look at all sorts of media sources. I think one of the keys here is that we are in an environment with massive amounts of information out there. What I like about these prediction markets is they're actually an attempt to synthesize all those sources out there. They're not perfect, but they're quite good.
Ryssdal: Justin, last question to you: Polls close at 7 p.m. eastern time tomorrow. Is that when we should start looking for some of this data?
Wolfers: You may even want to look a little bit earlier than that. Sometimes, these markets give early signals; sometimes as we learned in 2004 with the funny business with the exit polls, they give false signals. But I think throughout the day as you see things like weather, turnout in various states and then later on as the exit polls start to hit the market, we'll see news all through the day.
Ryssdal: Justin Wolfers at the University of Pennsylvania. David Leonhardt with The New York Times. Thanks guys.
Leonhardt: Thanks, Kai.