Freakonomics author says it's stupid to vote

Politics & Elections
81

Are there no economists who are willing to drop the assumption that people are instrumentally rational? Instrumental rationality isn’t the only kind of rationality, so it’s not like dropping that assumption would end all hope of axiomatization, is it?

With the voting example, for example, what about the following principle:

U: Do not refrain from doing something if, were everyone to refrain from doing it, something would happen which you disvalue.

Following U might not be instrumentally rational–because if you do it even when there are defectors about, depending on what’s at stake, you may be in a position to be harmed. But just because it’s not instrumentally rational, isn’t it rational in some other sense? Indeed, in a sense close to “instrumental” since, so long as enough people are following U, the resulting system is likely to be stable and apt to support the existence of things which its members value? (I say “close to instrumental” but not instrumental simpliciter because no one following U thereby causes the stable supportive system to come about. But everyone–or enough people–following U does in some sense constitute such a stable supportive system.)

Are Causal Decision Theory and Evidential Decision Theory things that economists talk about? I ask because in the prisoner’s dilemma with psychologically identical prisoners, evidential decision theory recommends non-defection while causal decision theory recommends defection, and I’ve never understood why people seem to tend to respond to this fact by saying so much the worse for EDT. I’d say, rather, that it’s evidence that EDT is the right decision theory in at least some cases. And since all people are psychologically similar in some sense, EDT might be applicable in a fairly wide range of cases. This may be relevant to the voting case (I’m not sure). Anyway, I don’t even know if CDT and EDT are a thing in economics so I should wait to hear about that first before saying anything more…

82

That sounds super interesting, actually.

83

If California is likely to be safe Democrat, then it doesn’t matter what I vote because it will still be safe Democrat. If California is likely to turn Republican, that means so many other states are safe Republican that it doesn’t matter what California votes.

It’s impossible to concoct a plausible scenario where California matters in this election.

84

This is kind of what I was getting at back in post #5 when I mentioned Kant’s categorical imperative.

I’m not sure exactly how that factors in to models of human behavior. I do think we’ve all been socially conditioned to some degree to feel an obligation to “do our part” when it comes to things like this. And that’s a good thing.

85

To clarify, I’m saying I vote not because I feel societal pressure to vote, specifically, but because I feel like it would be morally wrong for me to not vote for my chosen candidate while hoping that other people do vote for him. And the reason I feel that way is probably because society has conditioned me to believe in something like Kant’s categorical imperative: that it’s wrong to behave one way while wanting everyone else to behave a different way. In other words, I feel like it’s important to do my part (even though my individual contribution is trivial), so as not to be a hypocrite or a freeloader. And probably the reason society encourages these sorts of values is evolutionary, because ultimately societies that cared about everyone doing their part (no matter how small) were more successful than ones that didn’t.

Of course a lot of this is speculation, and I don’t know enough about the history of different societies and their values to know how well supported it is. (I’d be very interested to learn.) But it makes sense to me.

86

Yeah, it does. I would totally attempt to read that if it’s not too heavy with technical terminology that I wouldn’t understand. Will it be available online when you’re done with it?

87

Even though instrumental rationality is imperfect, there is nothing better right now to put in its place. All instrumental rationality says is that individuals are goal directed and do things to achieve their goals. It can be translated into mathematics very easily. It is the most bare-bones assumption you can make: it cannot be broken down any further.

Other plausible-sounding possibilities don’t do as well when you try to hang a theory on them. First is altruism. People definitely do altruistic things, so it is tempting to think that humans have some innately altruistic motivations that complicate the picture we like to paint of instrumental rationality. But imputing altruism as a motivation does not explain when people choose not to behave altruistically. On the other hand, it is very easy to use instrumental rationality to motivate altruistic action.

Consider the idea that people “satisfice”, that is, they optimize their decisions just enough to get by. Or perhaps Jon Elster’s argument that people are “reward-sensitive but not reward-maximizing.” This sounds plausible, and satisficing lines up with some experimental results. But no one to my knowledge has been able to reduce these ideas into nice axioms on which to derive some conclusions. At what margin do people satisfice? When do people stop maximizing rewards? It’s easy to show how people satisficed after they have taken some action, but anyone can tell a story ad hoc. These are verbal, qualitative arguments that sound good on the surface. But try to draw up some simple supply & demand curves using the concept of satisficing instead of the classical rationality assumptions about human motivation. You’ll get lost in the woods.

This is actually the exact field I study, believe it or not. A system where everyone cooperates because everyone believes everyone else will cooperate is called an “institution.” The earliest and most inspirational book on the subject is Andrew Schotter’s An Economic Analysis of Social Institutions from 1981. After that my favorite writer on the subject is Randall Calvert.

The simplest example is “rationing by waiting” at movie theatres. People wait on line to get the good seats. The movie theatre doesn’t police lines, individuals do. If I barge in, I can get the good seats for myself. But people typically don’t. To make a long story short, stable institutions usually rely on the credible threat of low cost punishment. If I believe that everyone in the line will verbally abuse me and I dislike verbal abuse sufficiently, I won’t muscle my way to the front of the line. But in any event, I form an expectation of what other people do because I go to movie theatres over and over. So the institution is self-enforcing because it creates these expectations of what other people do and because it enables punishment of defectors. So I am willing to cooperate even in circumstances when I might seriously consider defection.

All of this follows logically from instrumental rationality.

I have never heard of these terms before. I am also a political scientist on the extreme quantitative end. So while there is a lot of overlap between us and economists, this just might be something that I have not heard ever come up but a real economist might be aware of.

88

Love you both. I have no idea if it will be online, actually. There is a lot of uncertainty about putting dissertations online. In some cases it makes it very difficult to publish them afterwards; in other cases the reverse seems to be true. I’m not exactly close to finishing, so I will figure it out in a year or two.

There is quite a bit that is both mathematical and deep in the details of texts in long-dead languages, but aside from that, I am trying to make it as readable as I can. :slight_smile:

89

How do people’s notions of right and wrong factor in, if at all? I mean, I wouldn’t steal even if I thought I could get away with it, because I don’t think it’s the right thing to do. Some people certainly would, so it’s a good thing we have police, but I imagine a lot of people wouldn’t, even if you removed the threat of punishment. (Although maybe we’ve been conditioned at this point by having that threat of punishment for our entire childhoods. I don’t know.)

And likewise with voting. It seems like some people will vote just because they feel it’s their civic duty, or it’s not fair of them to ask it of others without doing it themselves, or whatever. Sure, a lot of people wouldn’t (given that there aren’t any real consequences of an individual not voting), but maybe that’s why voter turnout is always so low.

90

I don’t know the answer to this question at all. I don’t have a good answer for why pedestrians wait to cross the street until a green light when there is no oncoming traffic. There may be social norms, there may not. Maybe there are multiple social norms that operate on different levels. Maybe institutions themselves shape our social norms. Sociologists and sociological institutionalists have some insight into these issues. I definitely do not have any good answers.

91

Man, this thread is frustrating, with people saying things that don’t make sense or that have nothing to do with what was actually said.

So let’s start with the basics: is it true that one is being irrational if they believe that, in isolation, their vote has any non-negligible chance of deciding the election? Of course it’s true! No one here thinks that their vote has a significant chance of deciding the election.

The problem with what Levitt said is that it is so self-evident that absolutely no one votes for that reason. You can see it all over this thread–no one recommends voting because they think their vote alone will matter. The most rational argument is based entirely on the idea that your vote, in aggregate, matters. That is also true. It’s simple math

The problem with saying something self-evident is that people assume you are saying something with meaning, and search for that meaning. That is why this thread is entitled “Freakonomics author thinks it’s stupid to vote” when he clearly allowed for all other reasons for voting besides the one that is self-evidently irrational.

No one is saying that your vote is worthless. It just will not, with fairly high certainty, decide the election. And, if you think it will, you are, in fact, stupid. The problem is that, by saying it at all, Levitt implies there are a lot more people who believe this stupid thing than there are. None of you think it.

But this isn’t all Levitt said. He also said that people vote because it’s fun. This is only true if, by fun, you mean that in the same way it is fun to fulfill any other utility function. But then that’s again self-evident, as that’s how utility functions work–we enjoy fulfilling them, to some extent. Heck, thinking your vote will decide the election would still make voting “fun” in that sense.

Using “fun” to mean more than that is a problem, though, and I believe it is unclear if this is what Levitt is doing. If so, the word fun means that the good feelings you get when performing the action are the only significant worthwhile properties of the action. In other words, voting because your vote matters in aggregate is not “fun,” nor is voting to encourage other people to vote. If he really thinks the only utility in voting is that it makes you feel good, then he is sadly mistaken. He is completely ignoring why doing so makes you feel good. That reason is that you have fulfilled one of your utility functions.

Now, if he is using fun to just mean that one feels pleasure while doing it, rather than just after it is done, then he may have something. But, still, that’s not really saying anything profound.

92

Certainly not. We have come to think this way because of the shorthand used in politics, where each side is assigned a certain number of electoral votes associated with their “safe” states, and then the plotting begins to try to win enough of the “swing states” to get to 270. But as has been pointed out upthread, those “safe” states are only safe as long as all those individuals expected to vote for their candidate actually do so.

Somewhat surprisingly, California has not delivered its electoral votes to a winning presidential candidate, who would have lost without those EVs, since Nixon in 1968. But by the yardstick I’m using–whether a vote is part of a winning coalition–California was part of Obama’s winning coalition in 2008 and before that for Clinton in 1996. If Obama wins again, it will likely be with California making up a sizable chunk of the electoral votes he needs.

Sure, neither candidate is advertising and campaigning in the state. But politics isn’t just about what happens in one cycle. In a longer view sense, Democrats won this biggest of all electoral prizes away from Republicans (and it’s only since the '90s that it has been a Democratic state, let’s remember). When all those Democratic-voting Californians go out and vote consistently for Democrats, and–let’s not forget–allow Democrats to husband their financial resources and use them elsewhere instead of in that expensive collection of media markets, they do matter. To act as though voters in deep blue (or deep red) states “don’t matter” is a fundamental misunderstanding in my view.

(An individual California voter could make their vote “matter more” by moving to Ohio, I suppose; but if too many of them did this, California might again become a swing state or even a red state.)

93

Is that “your” singular or plural?

94

I have a strict policy as regards Kant. Whenever I find myself agreeing with him, I stop what I am doing and retrace all my logical steps to make sure there’s no error.

95

When things don’t make sense to you, I suggest you ask questions.

This is not completely wrong. But the problem is Levitt’s sleight of hand. He really offers only two alternatives: either people vote to affect the outcome or they vote because it’s “fun”. The former cannot be true in presidential elections (though it certainly can be true in local ones), so anyone who believes it must be an idiot. So the only other alternative is that it is “fun”. In other words, he sees no other way to make voting consistent with rational self-interest. He takes refuge in the tautological property of utility, that people only do things they want to do that achieve some goals. This is an assumption from which we derive action, not a straightforward reason for action. Especially when that action, voting, is strategic.

It’s a jejune explanation because voting is not fucking fun. If you go after work, it can be a real hassle and a drag. It can be costly, and people often don’t enjoy it. But they do it anyway.

Levitt doesn’t say something like, “look, we really don’t know why people vote.” Either the voter must somehow be having fun or he must be stupid. Of course this sounds self-evidently true, but it reveals a pretty big interpretive problem that makes thinking people suspicious of the validity of this way of doing social science.

96

I’m shocked that no one has attributed this finding to the right author. Anthony Downs got to this point in 1957, it’s not something Freakonomics found out last week. That it is true is easy enough; given the assumption you only care about influencing the outcome, the ‘costs’ of actually turning up (or just raising the pencil) are far higher than the expected utility (assuming a non-ridiculous benefit from actually deciding the election) because of the low probability of being decisive.

This is non controversial and thought to thousands of Poli Sci students each year. The fact that many people do vote (and all the discusion above) just point out that the assumption of people voting to ‘be decisive’ is false. People vote for other reasons; for instance a sense of civic duty, it’s easier to just go and don’t think about the use of voting, or just because you want to have someone to cheer for during the election party. Non of this makes the orginal finding any less true.

97

You have completely misinterpreted the median voter theorem. Downs’ simple model is a frictionless pedagogical tool that rarely holds in reality. Under some very rarefied conditions (unidimensional issue space and a single-peaked preference profile) you can get a peculiar result. This result obtains reasonably well for things like school board budget elections, but it does not hold at all for things like presidential elections. The initial conditions are violated, so the result does not obtain.

98

Disagree vehemently. Who or what is “affecting the outcome” if not “people”? Okay, sure: things like weather have a marginal effect. But the preponderence of what affects the outcome of elections (including presidential elections) is the people who vote in them.

99

While I don’t find it fun, believe it or not some people do. I’ve actually met them.

I mean, people like all kinds of weird shit.

Levitt’s got a strange blind spot that comes up from time to time where he simply can’t fathom that people do things for reasons that are different from his reasons.

100

Unfortunately, they probably don’t have the required 2% of the vote needed for federal funding to kick in.