Monday, November 05, 2012

The Rationality of Voting

Every election year, like clockwork, some people feel the need to remind the rest of us that (contrary to the exhortations of politicians and peers) our votes do not, in fact, count. Not only do they not count in New York, California or Texas, they don't count in Colorado, Ohio, or Florida either. While the likelihood that a single vote will be decisive may be incrementally higher in the latter set of states, it is negligible everywhere. Steve Levitt goes so far as to say that "it’s only the not so smart people who vote because they’re actually going to influence the election." Phil Arena is a bit more charitable, arguing that "people who believe that their vote counts are simply mistaken." Kindred Winecoff concurs.

Here's Arena's version of the argument:
If you've ever said something like "My vote doesn't count, because I live in New York", you're the type of person who makes my head hurt.  We may not know for sure how things will turn out in New Hampshire this coming Tuesday, but that doesn't mean that an individual's vote will "count" for much of anything in that state.   The fact that everyone who knows anything about politics knows how things will go in New York (or California, or Texas) doesn't make any meaningful difference to the question of whether individual votes in those states are likely to determine the outcome...  Don't confuse uncertainty over the final outcome with a significant probability of a single vote determining the outcome.  Those two things are not even remotely the same.
And yet we have people waiting in line for hours to cast ballots in Ohio, and making multiple trips to polling stations in Florida, bearing significant burdens to engage in what Winecoff asserts is simply "cheap talk." Would these voters make similar sacrifices to vote in New York or California? And if not, are they somehow deluded or dumb?

I believe that it is Arena and Levitt who are mistaken about the rationality of voting, and not the voters themselves. The premise of their argument is correct, but not the conclusions they draw from it. The likelihood of a single vote being decisive is negligible in all states, and voters by and large are fully cognizant of this fact. And yet it is perfectly rational for some voters to incur significant costs to vote in New York, and to incur even greater costs to do so in Ohio.

To see why, one needs only to recognize that the elation one feels when a preferred candidate wins depends both on the margin of victory and on whether or not one has cast a ballot. A single voter cannot materially affect the former, but can certainly determine the latter. Furthermore, the margin of victory can be forecast with a fair amount of accuracy: there is little doubt that the margin in Ohio, no matter who wins tomorrow, will be smaller than that in New York. Provided that the joy of celebrating a victory is greater when one has cast a ballot, and especially so when the margin of victory is small, it makes perfect sense to incur greater costs to vote in Ohio than in New York.

Similar arguments apply to the grief that comes with defeat. In this case it is the failure to cast a ballot when the margin is tight that can give rise to great regret. People in this situation are perfectly well aware that their vote alone would not have materially affected the outcome, but they are not much comforted by this thought. The point is that if a relatively small coalition could have jointly generated a different outcome, then one's failure to join such a coalition can be a cause of distress. There is nothing irrational about such preferences, and they clearly lead to greater turnout when and where elections are predicted to be close. This turnout differential is not based on mistaken beliefs about what one alone can accomplish, and is not driven by cognitive limitations either.

This perspective on voting also explains why people often vote strategically, rather than always voting their conscience. Think of Nader supporters contemplating a vote for Gore in 2000. The size of the coalition of such supporters who could have blocked a Bush victory in Florida turned out to be extremely small, and the possibility that Nader could play such a spoiler role in the election was certainly anticipated. Some of those who chose to vote their conscience may well have regretted this choice once the outcome of the election was finally determined, and some who chose to vote for Gore may well have done so to avoid such regret. The fact that no single voter was decisive is entirely irrelevant. Collective responsibility for coalitional choices comes naturally to us, especially when the groups involved are not large. It is not the motives themselves, but the pretense that they do not exist that constitutes the true departure from rationality.

People vote for all kinds of different reasons. Some consider it a civic duty, others enjoy the process, and still others take satisfaction from the exercise of voice. Voting can be a powerful expression of identity; an affirmation, as Noah Millman puts it, of membership in a political tribe. It can be a result of peer pressure or the desire to avoid social sanction. But incurring greater costs to vote in closer elections is also perfectly consistent with a calm, reasoned, and above all intelligent response to the preferences with which we are endowed.


Update: Andrew Gelman and Steve Waldman are also worth reading on these issues, although their perspectives differ somewhat from mine. Andrew argues that in swing states the probabilities of being decisive are not effectively negligible, and therefore does not accept the basic premise of Arena's argument. Steve maintains that the argument is "right but wrong-headed" and shows that norms of political participation sustained by sanctions can be stable. Voting is clearly rational in the presence of such norms, as is resistance to the kinds of arguments that Arena makes.

The first action I took as an American citizen was to register to vote; I did this within minutes of receiving my naturalization certificate. I plan to cast a ballot tomorrow in the great and resilient city of New York, even though there isn't a competitive race in sight. Doing so won't affect the outcome, but it will certainly affect the experience of watching the returns come in, no matter who the winner may be. 


  1. I agree with you that voting is rational. But I get there a little differently. To me, the key thing is that one does not, in general, vote as an individual. One votes as a member of a political party, a geographic community, a profession, an ethnic or religious group, a labor union, etc. These kinds of collective actors are unambiguously are able to affect election outcomes, so there is nothing irrational about voting -- provided the "I" that votes is "I the Democrat," "I the New Yorker," "I the teacher," etc. Voting on the basis of a purely personal preference would indeed be irrational, from where I'm sitting. The only *political* decision we make as individuals, I would argue, is what collective(s) to identify ourselves with.

    I don't think there's anything strange or mystical about thinking of political actors as collective in some cases. After all, the whole language of choices, interests, reason, etc. doesn't refer to anything in material reality. There is no sense in which some billions of neurons in a single human body "really do" constitute a moral agent but some trillions distributed across various human bodies do not. The one-to-one correspondence between biological humans and rational agents is just the way we conventionally line up two incompatible, or at least independent, ways of viewing the world. But it's basically arbitrary and in some cases, like voting, clearly inappropriate.

  2. Josh, I was trying to capture some of this when I said that "collective responsibility for coalitional choices comes naturally to us." But I also think that there are individually rational reasons to vote, provided that one is willing to allow for interaction effects between choices and outcomes, even in cases when the choices don't affect the outcomes.

  3. Yes, I guess my comment is just trying to develop the thought in that one line of yours.

  4. Rajiv, on a lighter note -

  5. Thanks... I enjoyed that. Tullock's view is the mainstream one, but I think he makes the common mistake of assuming that if a choice has no impact on an outcome, then it also has no impact on the welfare from that outcome. That assumption is what I'm trying to challenge here.

  6. I have it on good authority from primary sources that while Gordon Tullock regularly taught the argument of the pointlessness of voting, he always did so himself and even bugged those around him if they were not doing so.

    Also, in 1991 there was a House of Delegates race in VA that was decided by one vote. Does not happen often, but it does happen.

  7. But did you see the video Barkley? He says flatly that he doesn't vote. Wonder why he would do that....

  8. I know Gordon, Rajiv, and he is a famous trickster regarding such matters, constantly saying provocative things just to get people thinking.

  9. Rajiv, Levitt goes on to make similar postulates on why people vote ('it is fun to vote'...). But then as you mention, the conclusion does seem different.
    On another note, in the recently concluded college elections here, we noticed the 'elation' people feel when their candidate won (despite the fact that most people expected that candidate to win). However, I wonder if the elation still remains if you support a candidate at heart, but don't bother to go vote. Sort of like, betting on which team wins a match, without the money. The joy of being proven right still remains (as does the 'tragedy' of not predicting) and you lose no money! So if we were to look at this money as the cost one incurs, it does seem possible that rational voters still won't turn up, yet experience what you describe in the post by supporting/not supporting the candidate in the privacy of their heads?

  10. Ah you explain this through the 'margin of victory', which the individual voter can influence.

  11. Deepak, the elation is there regardless of whether or not one votes, my argument requires only that the degree of elation depends on whether one has participated (as a member of a coalition) in giving rise to the victory, and on the size of this coalition. I do not assume that individuals can affect the margin of victory by voting.