Twice this month (with the Franken-Coleman recount raging on), I've had to hear about instant-runoff voting, and how great it is. It is the darling of political science majors who still cosplay at fan conventions or live with their parents. Admittedly, it sounds so much more attractive than the current simple plurality system in the US (in most cases: pick one, ONLY one) that it makes some people think that IRV is the best and only alternative out there.
It's overrated. I don't like it. It's a bad idea.
Not that the current setup is peaches and cream either. I'm gonna push my own preference in a bit, but first, let's talk about why IRV isn't all it's cracked up to be.
Basically, in an IRV election, voters are allowed to rank any number of candidates they like in the order they like them. Voter ranks Joe as number 1, Michelle as number 2, Richard as number 3, and so on. If that voter decides not to rank some of the candidates, fine.
At poll closing, count the votes. A plurality (more votes than all others) is not enough; if no one has an outright majority (50% + 1), the instant runoff takes place. Eliminate the candidate with the fewest votes, and distribute those votes to the candidates of the voters' next preference. So if Joe is eliminated, the vote of the voter above then goes to Michelle, ranked second. If a majority is found, great! If not, repeat the process.
At first glance, it sounds like a wonderful idea. Wouldn't it be nice for a voter to go into a booth and say, "Well, I like both candidates, I just like one guy more than the other, so let me vote for both"?
Hey, it does sound good. I'm not denying that.
But I have two things. First, it does bother me that simple plurality is being scapegoated as the reason the Minnesota recount is going on and on. One example of this is a comment to which I have replied at Political Wire (yes, my name is Roehl).
I'm going to call this the "we need the other thing" syndrome. It's new! It's shiny! It's not what we have! We must have it! And I can't help but think this is what's going on, in this and in many cases I've stumbled across when people complain about the system.
It's not that it's automatically invalid, but once we get the new thing, I doubt very much that all, or even many of the problems associated with voting will go away. IRV wouldn't stop a recount between candidates with very close vote totals (as we have with Franken and Coleman). It wouldn't resolve the problems of decoding voter intent in some of the most insane ballots to ever have come out of Florida in 2000 and Minnesota in 2008. And it wouldn't fix the voting machines that Diebold gave us.
IRV is not a panacea, much in the same way that Prime Minister's Question Time, or snap elections, or any number of instruments others have that we don't have is not going to fix American politics overnight without some actual critical thinking and an analysis of whether or not, as well as why, those instruments would fix our problems.
My second criticism is more abstract, and hasn't been seen in practice as far as I know. But the possibility is just as real.
Monotonicity is a deal-breaker for me. Essentially, in a three-man race, the second-place finisher in the first round can still end up on top. It nearly happened in the Georgia Senate race. Extrapolated further, a multi-candidate elections can see a third-, fourth- or even fifth-placed finisher in the first round accumulate enough votes from eliminated candidates and come out on top in the final round.
I don't know about you, but that seems strange to me.
Some have called the potential phenomenon a nonissue, that it is unlikely to happen in the real world. But "unlikely" is a far cry from "impossible," and got home tonight reading about an 8-vote difference between Franken and Coleman. To me, it matters very little that such an outcome is labeled as "unlikely." That it can happen is something that needs to be seriously considered and negotiated.
I do agree that simple plurality could potentially become more trouble than it's worth. Personally, I am a fan of approval voting. Basically, if you like more than one candidate, why not vote for as many candidates as you want? There is no ranking of candidates, but the trade-off is simplicity in voting; don't ask me to choose between vanilla and coffee ice cream, for example. I can't do it, and I won't do it. It removes the problem of spoilers presented by simple plurality (Nader can get all the votes he wants, just as long as our guy gets those votes too). And unlike IRV, no one can game the system to help a long-shot candidate win in the later rounds when it couldn't be done at the outset.
Whichever is better, I do believe there are advocates of instant-runoff voting who aren't merely impressed by its shiny "otherness." However, I am not one such supporter, and merely think that, though IRV is a step in the right direction, we can do much better.