Follow-Up: Critiques of Instant Runoff Voting

Since posting my last entry endorsing instant runoff voting, I have received a great deal of feedback, and I’m grateful for all of it. While some responses have been in agreement, most of the comments I have gotten have been arguments against IRV. It seems that most people agree that some form of change from our plurality-based system is desirable, there seems to be a great deal of very well-reasoned opposition to IRV specifically. I’ve learned a lot since my last post, have read about the problems and benefits of IRV and its rivals, and I want to address a few points.

First off, and this is a small point, many people expressed concern that IRV would be more prone to mistakes in counting, litigation, and other such hitches. I grant that IRV would be a bit more complicated than our current system, but in its implementation I think one would have to assume we have the apparatus in place that is reliable enough to ensure an accurate count, be that electronic or manual. Obviously, if we can’t handle the logistics, then neither IRV or any other electoral alternatives would be viable. In my arguments, I assume the ability to do so.

Here’s a criticism from Abou Ben Adhem from DailyKos;

Consider this scenario: There’s a three-way race between the Democrats, the Republicans, and the Greens. Virtually all the Republican voters rank the Dems as their second choice, as do nearly all the Greens. But the Democratic voters are split 60/40 between those who pick the Republicans second, and those who pick the Greens.

As long as the Greens are the first party to be eliminated, the system works as you’ve described. But as soon as the Greens pass the Dems and the Dems are eliminated, everything falls apart. 60% of the Democratic votes are re-assigned to the Republicans, and the Republicans win with 55% of the vote — even though 2/3 of the voters would prefer the Democrats to the Republicans. The race gets thrown to the Republicans as a direct result of voters shifting to the left (to the Greens).

And this isn’t some contrived scenario; if you think about it, something like this is very likely to occur in a three-way race.

I got quite a few comments along these lines, and I agree, if such a scenario were to occur, it would defeat the purpose we’re looking for with IRV. One problem is that it assumes the unlikely event that Republican voters would choose the Democrat as their second choice. Of course, I don’t think the Republicans would choose a Green as a second choice, but rather I would assume that with IRV implemented, there would almost certainly be a fourth, more conservative (or moderate) candidate that the voters would be able to choose. That being the case, even though this person’s horror scenario of the Democrats being eliminated still holds true, the result would not be that the candidate with the most support was eliminated. The fourth candidate would probably be second choice to most Republicans, and vice versa. The Republican would still win, which I can understand upsets anyone at Kos, as it would me, but at least the outcome is more “satisfying” to the voters at large. I can’t say for certain that such a fourth candidate would not emerge, but I have a hard time imagining that one would not with such a system in place.

Here’s a very good example from a DailyKos diary by DenmocraticLuntz;

The Spoiler Effect in Instant Runoff Voting

The one thing people who support instant runoff voting as a way to elect members of the Green party (in a race between the Green party, the Democratic party, and the Republican party) is that people choosing the Democratic candidate as a first choice will NOT always choose the Green party candidate as their second choice

This does not in and of itself speak ill of IRV, but of Democrats with questionable tastes in candidates. Anyway, that’s not the important part. He goes on.

A relatively significant minority of people who tend to vote Democratic are moderate in their political standings, and many of these people would probably choose the Republican as their second choice.

I think that’s an important point to remember, actually, and again, while it may not make progressives happy, we have to learn to accept that it is very often the case. The onus is on progressives to work, as we have been, to change those minds. Back to my critic.

This could lead to the following scenario (this is a plausible but of made-up scenario). In this scenario, most of the people voting Green in the first round would be liberals who currently vote Democratic. They vote Green under Instant Runoff Voting because the Green is closer to their views and they believe the Green won’t be a spoiler under IRV.

      Round 1Green:    45,000Democrat:   35,000Republican:   70,000

Notice that the sum of the votes for the Green and the Democratic candidates is 80,000. It’s likely that nearly all those who voted Green as their first choice in this scenario would’ve voted for the Democrat if it was a plurality election.

Again, a big assumption here is that if there was IRV in place, people would vote for the Greens or some such party in droves. I don’t quite believe that, but I’m sure it’s possible in more liberal areas of the country. As a whole, however, I think Democrats vote for Democrats and Greens for Greens, and IRV would not change that drastically, though probably somewhat significantly.

Therefore, in a plurality system, the Democratic candidate would have won the seat.Let’s assume that 80% of those who chose the Democrat as their choice chose the Green as their second choice, and that 20% chose the Republican as their second choice (in many cases the Republican would probably get more than 20% of the original vote). In that case, this would be the final result.

        Round 1   Round 2Green:        45,000(+28,000)  73,000Democrat:      35,000(-35,000)  ELIMINATEDRepublican:    70,000(+7,000)  77,000

The Republican candidate ends up winning despite the fact that the majority of the voters would have preferred the Democratic candidate, and despite the fact that they were probably the last choice of over 45% of the electorate.

So yes, if this were to be the actual outcome of a real election, IRV’s purpose is once again defeated. But I have to reassert that I find it highly unlikely, and if it were likely to occur, chances are polling would tell us as much well in advance, and the candidates and parties would react to deal with the situation, probably with bargaining and negotiation, which is a good thing. If we have learned nothing else from the Bush foreign policy, an inclination toward diplomacy is a virtue. We want the candidates and the parties reacting to the wishes of the voters, and certainly Greens don’t want Republicans in office any more than the Constitution Party wants Democrats. It has seemed to me that most nightmare scenarios assume that the parties and campaigns are not running as though IRV is in place.

I received a lot of similar nightmare scenario examples, and while they showed many plausible ways in which a candidate not supported by a majority of voters would end up winning, all of them struck me as being highly unlikely, especially given that our two-party system is extremely entrenched. I have yet to be convinced that the implementation of IRV would, other than in a negligible set of cases, threaten the loss of a majority-supported candidate. But perhaps the loss of any of those candidates is too much.

The critiques I have gotten of IRV have been extremely
well thought out and reasoned. So much so, that some of them were almost too complicated for me to digest on a first reading! A very solid example of an IRV critique I’ve seen is from a blogger by the name of Auros, and I recommend you read his post here. Also, the comment in response to my last entry from Dr. Warren D. Smith of the Center for Range Voting, which sites a specific example in Peru, was very compelling, and I thank him for his comment. Click on the link for his organization to get a good idea of what range voting is all about.

I’d like to ask Dr. Smith (if he’s reading), to help me and others reading to understand, even in the Peru example, don’t we still get a winner that has the support of a majority, even if that support is not all-around first-tier? Also, while I am always happy to ask voters to rise to the occasion, how plausible is it to ask the average voter to give numerical values that accurately reflect their feelings toward candidates, when it is difficult enough to educate them sufficiently for even something like IRV? There is a good point made in the blog Critical Mastiff where Mastiff says

Unfortunately, what often happens in places like Amazon.com is that the overall rating in disproportionally determined by strong partisans for and against a candidate. If a voting system were to allow voters to remove their views from consideration, outcomes would be determined largely by the fringes.

If a candidate has done such a poor job of outreach that many voters have no opinion one way or the other, he should not be rewarded by making his few supporters relatively more influential. There should be no “express no opinion” option.

I sincerely would love for you to respond to my questions and Mastiff’s point, Dr. Smith. I’ll even give you your own entry here!

There seems little love, on the whole, for the current plurality system, and in his post Auros gives several examples of other alternative voting methods, and I won’t go into all of them now, but their goal is the same as IRV’s, to ensure that the candidate winning an election has the support of most voters. The Wikipedia entry on voting systems has a good summary of various systems, including the complicated Condorcet method, the aforementioned range voting, and the intriguing approval voting system, among others. Most of these seem preferable to what we have, but to me, IRV seems the best of the bunch, despite its flaws.

For my own edification I will point out, though it surely is unscientific, that I placed a poll asking about folks’ position on IRV on my Kos diary, and while most commenters were anti-IRV, 69% of the people who voted in the poll (out of a mere 65 votes cast) voted for the implementation of IRV. 24% were against, and to my amusement, 6% “didn’t get it”. I can’t imagine what the results would have been if IRV were implemented in the poll itself.

While I imagine my next post will cover a different topic, this entry does not close the debate at FifteenNineteen for IRV, and I encourage folks to keep talking back. I am always open to changing my mind based on sound argument, as I hope are others. I would like to ask those in favor of IRV to make their voices heard here and post some comments. My slight experience tells me that people are less likely to post a comment if they agree with you, more likely if they don’t. That makes sense, since the whole point is to spur discussion and debate. Consider this a call to those who agree with me to “get my back”.

So, you game, Dr. Smith?

Thanks for reading. As always, I always appreciate a Digg, if you dig.