USS Clueless - Instant runoff voting

Stardate 20031014.1725

(Captain's log): In response to yesterday's post on the American electoral system, Stephen writes:

Read your post regarding US political parties and the interactions before and after primaries, etc. etc.

Here is an interesting alternative system (As I understand it, something like this is used in Ireland), and I was curious as to your take on it.

In a presidential election, everybody votes _twice_, selecting a "first choice" and a "second choice".

The first choices are counted. If someone gets more than 50%, game over, they win.

If nobody gets at least 50%, then whoever is in third place is removed from the count, and all the votes that are first choice for the loser are counted as votes for the "second choice" on that vote.

Under this system, Bush would probably have won in 1992, assuming Perot didn't win in a major upset (as more people would have voted for him since they would no longer be "throwing their vote away"). All those Republicans voting for Perot would have listed Bush as second choice. When nobody has 50% and Perot was in third place, the Perot voters' second choices would be counted, and most likely go more for Bush than Clinton.

The question is this: Do you think this would be a better system for running elections? Would it better reflect the will of the people or would it have some hidden potential (I can't see it) to subvert the intent of the system?

The electoral system you're describing is called "Instant runoff voting".

All other things being equal, if it had been in use in 1992 then it probably would have led to a Bush win. But that's rather pointless to speculate about because other things would not have been equal. If the system had been different, the candidates would have used different strategies during the campaign. For instance, the Clinton campaign would have known it could not rely on Perot to be a spoiler and would have worked to make sure it got enough votes in the states it thought it could win so as to make sure it could win even with Perot eliminated in the second round. So it isn't actually clear that Bush would have won.

Would it be a better system? I have my doubts. It would be different, but I'm not really so sure it would be better. What concerns me about it isn't its direct characteristics so much as the secondary consequences if it were in place.

It would eliminate the intrinsic penalty involved in creating splinter parties. It would mean that someone like Perot or Nader could create their own party without fearing that they'd hand the election to the other side. That would mean that there would no longer be a systemic tendency to sustain the two-party system, and over a period of 50 years the political landscape would mutate in unpredictable ways.

I don't claim that the two-party system is essential, but it's something we understand and know how to deal with. It's a component of our political system, though it's not mandated Constitutionally. But with IRV in place, a Perot could say "Vote for Reform first round and vote for the Republicans second round" or a Nader could say "Vote Green first and the Democrats second" and then they could use each successive election cycle to build first-round strength until they could start winning. But that wouldn't necessarily happen the same everywhere; Nader and the Greens would probably be more successful in California and the NE than in Texas, and Perot exactly the opposite. And it isn't the presidential election that concerns me so much as elections for the US House of Representatives.

With only two parties, then disregarding the occasional rare "independent" (who usually ends up aligning with one of the parties anyway) then one party always has a majority.

On the other hand, if the Greens manage to elect three or five representatives out of districts in northern California, and maybe one or two from Vermont or Massachusetts, and if the Reform party sends three from Texas and a couple from Georgia or Kentucky, then there might be no party in majority in the House. That would require coalitions, which is what happens in the European parliamentary system. Of course, it would be a coalition to control the chamber rather than to form a government, but the basic problem is that it would give those splinter groups disproportionate influence just as it does in Europe.

It wouldn't stop with them. The "religious right" could manage to control a few house seats from the Bible Belt. Someone might form an "Agrarian Party" in Iowa and Nebraska, representing farmer interests. How many such parties might form? There's no way to tell, but they'd only have to be strong in a few places to gain control of seats in the House, so that they could play in the big game there.

I think one of the basic strengths of our system compared to those in Europe is what I referred to in this article as a high threshold of noise rejection. The idea is that fringe or extreme political viewpoints cannot significantly influence the system through local dominance in a small area or by having a small number of followers spread around everywhere. It's structural, and what it means is that we are relatively less vulnerable to extreme political opinions. Which means we can

Captured by MemoWeb from on 9/16/2004