Condorcet rules?

by John Q on March 1, 2004

The comments thread on my last post led me to this site (hat-tip: novalis), advocating Condorcet voting and presenting a critique of the instant runoff/single transferable vote , the core of which is

IRV has serious problems. It allows a sufficiently small minority of voters to safely register “protest” votes for minor-party candidates–but only as long as their candidate is sure to lose. As soon as their candidate threatens to actually win, they risk hurting their own cause by ranking their favorite first, just as they do under our current plurality system. IRV is therefore unlikely to be any more successful than plurality at solving the classic “lesser of two evils” problem.

It’s straightforward to show, however, that this problem can only arise if your preferred candidate would be the loser in a Condorcet system. Hence, voting strategically yields the preferred Condorcet outcome.

To see how the argument works consider three candidates A,B,C and suppose that A’s supporters rank ABC, C’s supporters rank CBA, and that no candidate has an absolute majority. Then (regardless of the preferences of B’s supporters), B is the Condorcet winner. If B has the most or second-best supporters, then a (weakly) dominant strategy is for everyone to vote in line with their preferences, leading to B’s election. Suppose however that B has the smallest number of first-preference supporters and that A would win over C in a pairwise contest. Then, as stated in the critique above, the optimal choice for C’s supporters is to vote strategically for B, so that B finishes the first round ahead of C. The distribution of preferences then ensures that B is the winner. So, in this case, strategic voting does not produce the “lesser evil” as far as the majority of the electorate is concerned.

Things can tricker when there are n (greater than 3) candidates with a serious chance of winning. But the problems for Condorcet are even worse, since the method requires n(n-1) pairwise comparisons.

Taking this a bit further it seems likely that, whatever rule is chosen for resolving cycles, an implementable Condorcet system would be vulnerable to exploitation by strategic choices to run (or not run) particular candidates whose function would be to tip the balance in favor of some other candidate.

{ 8 comments }

1

Conrad barwa 03.01.04 at 11:48 pm

Taking this a bit further it seems likely that, whatever rule is chosen for resolving cycles, an implementable Condorcet system would be vulnerable to exploitation by strategic choices to run (or not run) particular candidates whose function would be to tip the balance in favor of some other candidate.

Possible solution in this case might be to impose stringent conditions for entry into the electoral race –either via some sort of financial yardstick (from individual donors) or some other indicator of a critical level of popular support not restricted to a narrow constituency; to weed out such candidates. Though this has problems in itself.

2

novalis 03.01.04 at 11:59 pm

Then, as stated in the critique above, the optimal choice for C’s supporters is to vote strategically for B, so that B finishes the first round ahead of C.

But, because polling is inexact at best, it’s impossible to determine when to switch from CBA to BCA.

Personally, I’m a fan of approval voting, because of its simplicity. I understand that it (like all other systems) has its problems.

I recently participated in an election (for a small science fiction association). We did sequential (non-instant) run-offs, and the importance of non-primary choices became quite clear. In straw polls, candidates A and B had wide first-tier support. I asked for second choices, and candidate C swept the 2nd votes. Once everyone saw this, C ended up being elected.

3

Dave 03.02.04 at 5:52 am

The main problem with Condorcet is that it’s hard for the “average joe” to understand. I know these people have an argument about why it’s not, but any algorithm that requires college math isn’t going to appeal to a national populace.

I also tend to be wary of arguments like “Condorcet is the best because it is the only method which satisfies the Condorcet Criterion and four other criteria based on it.” The fact that they have an implementation in *Python* of all things scares me too.

I prefer approval voting to any of the other methods. My second preference is plurality, since it is simplest, but plurality wouldn’t work well in a country with a broader political spectrum than the U.S.

4

John Quiggin 03.02.04 at 5:55 am

As noted in my previous post, Australians and Irish (not famously intellectual groups of voters) have no trouble with preferential voting systems, of which Condorcet is an example. But I agree with your scepticism about Condorcet.

5

Dave 03.02.04 at 7:03 am

I have no problem with preferential voting systems, though I think they introduce strategies for voters which are to some extent counter-intuitive.

It does bother me if a candidate wins because his or her supporters understand voting strategy better, rather than that candidate having the most popular support.

6

novalis 03.02.04 at 5:52 pm

Dave, what’s wrong with Python?

(ooh, let’s have a programming language flame war to go with our voting system war ;)

I’m not set on Condorcet (I prefer the simplicity of approval voting). But I don’t think it’s hard to understand — I learned about it in 9th grade, IIRC.

7

Brian Weatherson 03.02.04 at 10:24 pm

I’d be more impressed with the criticism that STV rewards tactical voting if there were more actual examples adduced of tactical voting being rewarded. There have been tens of thousands of elections run under STV in Australia (just counting Federal, State and local elections) and there are no clear cases of this kind of tactical voting making a difference, and only a handful of cases where it is even plausible to claim tactical voting might have mattered.

Maybe it’s just indoctrination into the STV system, but when I see a case like the following, I think it’s natural that B should win.

48 votes for A then C then B
47 votes for B then C then A
5 votes for C then B then A

8

shooting_star 03.03.04 at 12:11 am

I was hoping Condorcet would come up when I posted my comments on the previous STV page about voting systems.

John Quiggen seems to prefer Proportional Representation to centrism. Well, the whole point of a Condorcet ballot is that a candidate has to appeal to multiple groups to elevate his preferences.

Borda can be rigged toward extremism by pushing the main opposition candidate toward the bottom. But there is no advantage to doing this in a Condorcet system.

Condorcet may seem complex, but it is exactly what pollsters do — they match up all candidates with each other in head-to-head contests. I don’t think this is too hard for voters to understand.

Approval is a reasonable alternative to Condorcet and also tends to choose the more centrist candidate. And it also can be polled for easily, so voters will have a better sense of the likely outcome if they want to change their votes.

Proportional Representation has its place, but it isn’t in a single winner election. You really want PR in an assembly.

The known problems with Condorcet occur when you get cyclic rankings. So you need a rule to decide which victories you want to include and which you don’t. Ranked Pairs / Maximize Affirmed Majorities is a reasonably understandable way to handle this. It isn’t hard to do. Pick the largest victory (by total number of votes), include the next largest victory, and leave out any victories that cause a cycle. You’re left with an unambiguous ranking.

Brian Weatherson’s example is interesting. The Condorcet victor would be C (beats B 53:47, beats A 52:48, B beats A 52:48). C is the second choice of a nearly divided set of 95 voters who really dislike the majority opposition candidate (A or B).

In STV, you get 47 happy B voters and 5 semi-satisfied C voters, but 48 very dissatisfied A voters.

If you set the last choice candidate to non-approved, the Approval winner would also be C, with 95 votes C to 52 B to 48 A.

So why should a candidate with lower approval win via STV? It’s only natural if you think using the logic of caucuses, where small losing blocks can sway the plurality when they switch allegiance.

Comments on this entry are closed.