So here's an application of economic reasoning to voting. (And I caution you that I don't necessarily believe everything I write. This is brain exercise intended to get you think differently about a problem.)
(And is it ok to write things you don't necessarily believe? Well, sure, as long as you don't intend to run for office. Academics exist to try to provoke the rest of the world into thinking hard.)
When should you vote?
Obviously, when the benefit of voting exceeds the cost. The cost is a time cost --- it can be a hassle to get to the local polling place.
The benefit is that you might influence the outcome. When will your vote influence the outcome? Your vote influences the outcome if and only if your vote is pivotal.
Example: Suppose there are nine voters: you and eight others. When does your vote matter? Your vote matters in exactly the case where the other eight voters end up split 4-4. In that case, your vote is pivotal --- it has determined the outcome. If the other voters split 5-3, or 6-2 or 7-1, then your vote isn't pivotal. For example, if the others split 5-3, then your vote will make it 6-3 or 5-4, but the outcome isn't changed. (The TV game show "Survivor" is a really interesting context to use to think about strategic voting...)
So, in trying to figure the benefit of voting, you should try to figure the probability that it's a tie in the absence of your vote. Given the current polls, it's unlikely that any one person's vote will influence this year's gubernatorial race in Utah --- that race looks like a Huntsman landslide. So the benefit of voting is pretty small.
But what if you're voting for president? And what if you live in Ohio or Florida? One lesson people claim to take from from the 2000 Florida recount is that "every vote matters." The margin there was so small that it seems reasonable that one vote really could matter for picking a president.
But I think that's the wrong lesson to draw. Here's why. Let's suppose you live in Ohio. Let's suppose you think it's reasonably likely that the election will end up tied and that your vote will determine the winner. Then maybe we'd conclude that it's really important for you to get out there and vote.
But what did we learn from Florida 2000? We learned that close elections are decided not by the voters, but by the courts deciding what ballots to count. To put this another way, suppose McCain wins Ohio by one vote. And suppose that the electoral college is so close that the Ohio winner is guaranteed to win the election. In that case, did the one Ohio voter's vote matter? Well, in an election that close, the dems will litigate and the courts will decide.
So the probability that a single Ohio voter can determine the outcome isn't almost zero. It's literally zero. The benefit of voting is zero, and the cost is positive. Don't vote.
Posts
5 comments:
Scott--
Interesting idea. What you didn't say is that bviously the cost of voting is outweighed by something else--civic pride, patriotism, guilt. So people still vote despite the fact that their vote almost certainly doesn't matter. How do economists place a value on other motivations besides time or money? Is there a way to measure that?
-rm
EMBA '09
Good points, all... Exactly right that there must be some other benefit that outweighs the hassle cost --- I'm sure my colleagues in Poli Sci here would have a better sense of what those benefits are.
Economists are not blind to the fact that actual human motivation is complex. And we do have techniques that allow us to place values on factors like this. This is a good thing for me to write about sometime... I'll put it on the list.....
Also, actual voting behavior doesn't fit this "rational voting" model at all. Turnout tends to highest in the elections where the probability of being pivotal is lowest.
Sir
Interesting take on voting.
But isn't the viewpoint sort of individualistic ?
A single vote compared to the population size would obviously seem irrelevant & beneficial only in case of exceptional pivotal scenarios. But wouldn't it be statistically better to consider a bigger sample size of voters ?
Also as 'rm' points out, voters might weigh benefits from a different perspective than economists.
-vs
MBA '09
Hello vs - I'm not sure what you mean by "statistically better". When I go to the polls, I increase the total number of votes cast by exactly one. So I might want to compare cost and benefit.
Sample sizes are relevant if we're trying to do statistical inference. But I don't see how the "Should I vote?" question comes down to statistical inference.
Now, if we're trying to do a poll to predict how an election will turn out, that's statistical inference and we'd like to ask a number of people.
Post a Comment