Forget Polls: Look at Prediction Markets on the Election

(Moderators’ Note: Nothing here to be construed as an endorsement of any political action—this analysis is for educational purposes only.)

During election season, pundits and politicians are obsessed with polls. But are they obsessed with the wrong thing? There is a growing academic literature on the role of prediction markets, particularly in political contests, and it has been shown routinely that these markets are excellent (albeit not perfect) predictors of outcomes.

With this we turn to the ever-increasing obsession with November’s Presidential election. I’ve gotten a couple of emails recently comparing Obama and McCain on a number of issues, and to listen to most people in the media, the Presidential election is a two-horse race even though it isn’t. The structure of the system has some interesting implications, though.

First, we know that a single vote will not make a difference to the outcome of the election. A standard (and interesting) discussion to have in economics 101 is the question of whether it is “rational” to vote. The stock answer is that it isn’t, but that is only true if your only goal is to actually influence the outcome of the contest. I vote because it is a cheap way to let everyone know my preferences (Friend: “who are you voting for?” Me: “Candidate X.” This exchange says a lot about the package of ideas I most agree with among those being debated in the public square.

With this in mind, voters shouldn’t let fear that “the wrong guy will win” influence their decision. Voting “against” the major candidate one likes the least squanders an opportunity to cast a vote that means something. It’s the very essence of a wasted vote.

Armed with the knowledge that a single vote will not swing the election, the conscientious voter should refuse to hold his or her nose and pick the lesser of two evils—or whichever combination of electoral clichés you prefer—and should instead find a candidate who best represents his or her ideas. Presidential candidates abound from across the political spectrum, and I would guess that most voters could, with little effort, find a candidate with a platform they find preferable to the platforms offered by Obama and McCain. Extreme liberals would likely prefer Ralph Nader. Extreme conservatives would likely prefer Chuck Baldwin. Hardcore environmentalists would likely prefer Cynthia McKinney. Libertarians would likely prefer Bob Barr. And so on.

In my experience, a common voter complaint is that the major party candidates are usually uninspiring Washington insiders offering more of the same. It doesn’t have to be that way: by supporting candidates from outside the political establishment, people can move the debate away from personalities and posturing and toward substantive issues. The debates between Obama and McCain will be polished affairs in which the candidates offer rehearsed answers memorable more for their cleverness than for anything of substance. A debate featuring the other candidates would be a real debate about ideas. Instead of debating whether to expand the welfare state by X billion dollars or X.1 billion dollars, a debate between (say) Baldwin, Nader, McKinney, and Barr would likely be an animated contest in which the combatants consider bigger issues, like the nature and role of the state.

According to the prediction markets at, there is at least an 85% chance that McCain will win Tennessee (where my wife and I live) and a greater than 90% chance that McCain will win Alabama (where our families live). The market is predicting similar probabilities that Obama will win California, the largest electoral prize in the country. In other words, it is virtually certain that McCain will carry Tennessee and Alabama and Obama will carry California in November. The likelihood that a single vote will determine the outcome of a even a hotly-contested election is practically zero; in cases like these where the outcome is virtually certain, a vote for a candidate with whom the voter does not identify is a vote wasted.

For better or for worse, presidential election season is a time when most people are paying attention to ideas of fundamental importance. We sell ourselves short if we only focus on two candidates. There are a lot of candidates out there with a lot of ideas. I hope the 2008 election is an election that moves the discussion forward.

Art Carden is a Research Fellow at the Independent Institute in Oakland, California, Associate Professor of Economics and Business at Rhodes College.
Full Biography and Recent Publications
Posts by Art Carden | Full Biography and Publications
  • Catalyst