Why the President Is Still a Heavy Favorite on the Prediction Markets
Ohio — which has remained Obama friendly — is one of three key states where Romney needs to gain support in order to swing the election in his favor. Photo by Mandel Ngan/AFP/Getty Images
On the NewsHour tonight we explore three ways to predict the presidential election: polling, economic modeling, and betting in the prediction markets. Economist David Rothschild tracks all three. His was a fascinating interview, filled with nuances about presidential prediction and the betting markets in particular. The most surprising fact of all: they used to be much bigger than they are today. But you’ll have to read the entire exchange below to find out how much and when. (Unless you skip ahead, that is.)
David Rothschild: The uncertainty in election forecasts is not about what would happen if the election were held today. It is about what could happen between now and Election Day that could shake things up. And so when a state is 55 percent, or 60 percent, or 75 percent for a given candidate, what we’re saying is that there is a 45 percent, or 40 percent, or 25 percent chance the other candidate could win if something happens that shifts that balance, not what would happen if the election were held that day.
If you want a quick and meaningful rubric for understanding the likelihood of an election, one way to think about it is take every state and the District of Columbia — the 51 Electoral College districts — and rank them from the most likely for Romney to the most likely for Obama. Rank them all, number them 1 to 51. You have a bunch of states that are bunched up near 100 [percent likelihood] for Obama and a bunch of states bunched up near 100 [percent likelihood] for Romney. Then you have this whole middle ground from leaners [to the Democrat ticket], to tossups, to leaners [toward the Republican ticket]. And what happens is that, barring a major event, any given state is just going to start drifting, over time, towards the candidate they’re most likely [to vote] for. So every day that goes by on which that state does not flip [to the other party], it’s going to slowly drift in the direction of the candidate that it favors in terms of probabilities. And that’s what we’re seeing happen.
Paul Solman: So a state like Ohio has simply become more and more Obama friendly?
DR: No. What I would say is that a state like Ohio has maintained its Obama friendliness but as it gets closer and closer to Election Day, it becomes more and more likely that it will actually vote for Obama come Election Day. So it’s not a shifting of opinion as much as it’s saying, “Romney is running out of time to shift that opinion.” And as that happens, you see very few states cross each other, in essence, and so that ranking holds. We’ve had one state that’s really cut across and that’s North Carolina. Though it went for Obama in 2008, it was a heavy leaner towards Romney in the beginning of the cycle and drifted to the point where it was slightly in Obama’s category and is now slightly in Romney’s category. But for the most part this order has maintained.
So take that order, start adding up electoral votes until you hit 270. You’re going to find a pivot state: a state that if Obama holds he wins and that if Romney can reach he wins, and that state has been Ohio for virtually the entire election cycle. It means that if Romney can get all the states up to that point, he can win this election. But that includes Florida and Virginia on top of Ohio and those are three big hurdles and that means on the flip [side] that all Obama has had to do is hold one of those states. Just one of those main three states, and that’s what’s been giving him that edge this entire time.
The ranking method is not 100 percent. It is possible that Obama could win the state of Ohio and that Romney could capture everything and then somehow lightning strikes and there’s a massive economic issue, or something that somehow flips another state past Ohio that gives him the election.
PS: What’s the likelihood that that could happen?
DR: It’s extremely low at this point because there just aren’t that many states in play. … There aren’t that many swing states left.
PS: As between prediction markets and polls, which do you trust?
DR: I’m going to take prediction markets because prediction markets have all of this polling information available to them as well as additional information. They understand some things that will definitely happen that polls have not picked up yet. And I’m going to give you a good example: job numbers. We know that they’re going to affect the trajectory of the election. They’re going to be involved in the debates, they’re going to be in commercials, they’re going to help change donations. But unlike you and me, most people are not refreshing the Bureau of Labor Statistics website at 8:30 in the morning on the first Friday of every month. It’s the prediction markets who know about these numbers before the polls do, and so that’s what makes a difference.
PS: So prediction markets reflect more information because they are played by people with an incentive to stay on top of the information as opposed to react naturally to it over time.
DR: That’s correct. Putting your money where your mouth is incentivizes you to go out there and gather as much information as possible and more importantly in some ways, to then reveal that information correctly. And so those kinds of things combine to provide a very high level of information in these prediction markets that you don’t necessarily have in a poll that’s a snapshot of today, versus a prediction market which is looking at what’s going to happen on Election Day.
People always ask me, “Well, how would prediction markets do if there weren’t these polls? Are they just regurgitating these polls?” And the answer is, prediction markets do quite well, thank you. And we can tell that because prediction markets existed prior to polling. Standard polling that we know and love started in 1936 with George Gallup. And two great researchers, Koleman Strumpf and Paul Rhode, wrote a paper showing the history of prediction markets from the late 19th century up through that time and how they were incredibly effective in predicting elections. In 1916, around $200 million in today’s dollars were exchanged on prediction markets on that election. That’s a lot of money.
PS: So in 1916 Woodrow Wilson beat Charles Evans Hughes and $200 million in today’s dollars was bet on that election alone?
DR: Yes. Back in the day they were actually some of the most traded stocks on Wall Street, predictions on elections every four years, and they did great.
PS P.S.: If you want to follow the presidential race on the prediction markets yourself, here are three links: Intrade; Betfair, where to translate odds like 1.47 into a percentage, you divide 1.47 into 1 (about 68%); and finally IEM — the Iowa Electronic Markets, where the current odds are in the two boxes on the lower right; Obama, above; Romney, below. Be advised, though, that Intrade and Betfair represent wagering on who becomes President. IEM betting is on who wins the popular vote.
This entry is cross-posted on the Making Sen$e page, where correspondent Paul Solman answers your economic and business questions