| Pre-election special |
|
navigational aids: News ticker:
|
28 October 04. First, I'm not gonna tell you that you have to go and vote; that's pretty much up to you. But I am going to tell you how to vote if you do choose to. The rolloff effect It's a fact: people vote less in local elections than in the big-ticket presidential elections. However, in terms of what will make your life better, it's the local elections that really matter. You will notice whether the garbage is collected and the police are around before you notice U.S. policy toward Indonesia. Also, your vote is more likely to matter in a smaller election. Don't be another statistic! Vote in your local elections. Of course, this means you need to know who's running and what they promise; I'd give you links, but you can type ``(name of state) electoral endorsements'' ``(name of state) voter's guide'' into a search engine as quickly as I can. Your vote is not a signal Please don't take your vote to be a signal about any one issue (unless it's a referendum). The easiest way to send a signal about a single issue is to write a letter to your representative or senator clearly expressing your opinion. That would be a much less ambiguous move than casting an anonymous ballot. Please don't take your vote to be a signal that you are dissatisfied with the two-party system. The two-party system is a direct result of the first-past-the-post electoral system---this is known as Duverger's law among the political theorists, and is pretty darn consistent. If we want three parties, more media coverage and tweaking the federal laws won't do it: we'd need a constitutional rewrite. If you wanna push for that, feel free, but casting a ballot vote for a third party is not going to send a signal to anybody that you feel the constitution doesn't specify the right method for selecting the executive branch. Please don't take your vote to be a signal to the people who do win that they should be listening to your subgroup. Bush won on a minority---did that make him a more moderate president? One could argue that in the primaries the party is feeling out the opinions of the constituents, and writes the platform for November accordingly, so voting for Kucinich made some sense as a signal back then, but this is not the case in the final election. Vote as if you're pivotal For the sake of exposition, say that there are seven voters, including yourself. There are a number of possibilities for vote tallies among the other six voters; here are a few:
Let's say that in states A, B, C, E, F, and G, you're going to vote for candidate (oh, I need a random letter) N, but in state D, you'll vote for Kerry. But you can only cast one vote. One option would be to guess at what the state will be, and then vote based on your guess, but this ain't great: if you guess that the world is not in state D but it really is, then you'll be kicking yourself forever; but if you guess that the world is in state D but it actually isn't, then you'll just be a little annoyed. The downside is huge for this strategy and doesn't offset the modest upside. The only equilibrium strategy is to vote as if you're in state D and you are the pivotal voter. Vote to minimize regret There are a lot of little stories about why people vote at all, and my personal favorite is `regret minimization.' It'd suck if the candidate you don't like is elected, and it'd suck if you took time to vote and your vote doesn't matter, but the absolute worst case by far would be if the candidate you don't like is elected and it turns out you were actually pivotal and could have made a difference. I can't tell you how to write your own utility function, but perhaps this is a good way to think about it---and it leads to the same result as above: vote as if you are pivotal. End pre-election special; now back to our usual interesting tidbits from economic theory. The above vote-like-you-matter logic has an interesting application to juries. Here's how the others members of a jury of twelve (besides you) could vote:
That is, your vote only matters if everybody else on the jury believes the guy did it. You may take that into account: if you think the guy is innocent but there's a unanimous opinion that the guy is guilty, then you may start to question your private beliefs/information and switch to convicting. The equilibrium strategy is to vote as if you're in the near-unanimous state, so if juries are good game theorists, the system is biased toward conviction. It may even be the case that a simple majority rule would lead to a lower chance of conviction. Half a decade ago, there was a flurry among the political theorists about the conditions under which this is true; depending on your model, it can be a broad range or a hairline case. Unanimity is best only in the case where jurors are resolute in ignoring all of the opinions and hints from fellow jurors.
[link] [4 comments] Replies: 4 comments
|