Ranked Choice Voting and Its Discontents

Ranked choice, or instant-runoff voting, is probably the best-known alternative method to the plurality voting system.  It is supported by reform groups like Fair Vote, Independents, third party advocates – Greens in particular –, and even independently-minded Democrats and Republicans.  Ranked choice has already been implemented in a number of cities across the country, from Maine to Minnesota to California.  San Francisco is set to hold its first mayoral election under the new system.  At Divided We Stand, the Dividist takes a critical view of the process.  Excerpt:
As a direct consequence of our new public financing rules for the mayoral race in San Francisco, we now have a cavalry charge of 16 candidates running for mayor . . . As Ron Popeil might say - "But that's not all!" At no extra charge we will now throw all sixteen candidates into the mix-master of our first ranked voting / instant runoff election for mayor. On November 8th, all San Francisco voters will cast three votes for mayor in rank order of preference. "Rank" being the operative word in that sentence. . . .


Who knows what kind of a gawd-awful mess will come out of this election? Recall that in Oakland's 2010 ranked voting mayoral election, candidate Jean Quan had 10 points fewer first choice votes than Perata in the first vote count. I am talking about - Her Honor Jean Quan, the current mayor of Oakland, who had 10 percentage points fewer first choice votes than the loser Don Perata.

The simple fact is that most SF citizens have no friggin' clue about the ramifications of ranked voting/instant runoffs. While the voters may not understand it, be assured the candidates who would otherwise have zero chance of winning a plurality in the election or a majority in a real runoff know how exactly how the voting system can be gamed.

Net net - As a voter it is more important to decide who to exclude from any of your three votes for mayor than it is to pick who you would prefer to see win as your first choice. In fact you may be better off ranking your favorite as your second or third choice. This is Game Theory Gone Wild. We might as well be drawing lots to pick the next mayor . . . 
Fortunately, for those who are not fans of the instant runoff, ranked choice is not the only alternative to plurality: there are alternatives to the alternative, so to speak, such as approval voting and range voting

11 comments:

Dale Sheldon-Hess said...

Rangevoting.org has a page on this election, at http://rangevoting.org/Oakland2010Mayor.html.

You know I'm a big approval/range supporter, and strongly disfavor IRV. But, despite its shortcomings, IRV probably got the right answer in this election; any internally-consistent interpretation of the ballots as Borda, approval, or range votes would have still yielded Quan as a winner. (Condorcet gives either Quan or Perata, depending on how you interpret non-ranked candidates, and obviously Perata wins by plurality.)

The problems I see with IRV in this election are two fold. One, as pointed out in the link, the election had a "participation paradox": if there were more anti-Perata/pro-Kaplan (the 3rd-place candidate) voters, then Perata would have won the election by IRV. Something like that isn't possible under approval or range (or Borda, for that matter.)

Second, despite FairVote's claims of always picking a majority winner, Quan won with only 45% of the ballots, versus Perata's 44%. The final result is often reported as 51%/49%, but that's because those reports ignore the over 10% of ballots that had been exhausted by the final Quan vs. Perata round. And yet, FairVote still pushes their majority claim.

TiradeFaction said...

The article in question also seems to be a critique of public financing of elections (or at least SF's system in particular). Which is fine, but that seemed another important angle of the article you might have left out. Just saying.

d.eris said...

Thanks for the analysis Dale.

d.eris said...

Indeed, TF, the article has no sympathy for San Francisco's public financing system. I wasn't even aware that they had instituted public financing and so decided to let it be since I'm not familiar with the situation. Would be interested in a different take though. In the main, however, I just wanted to take the opportunity to plug AV/RV
;-)

Now that I'm thinking about it a bit more though, I wonder how the ranked choice system in San Francisco is squared with the top two primary system instituted state wide. Anybody have any info in that regard?

TiradeFaction said...

"Indeed, TF, the article has no sympathy for San Francisco's public financing system. I wasn't even aware that they had instituted public financing and so decided to let it be since I'm not familiar with the situation. Would be interested in a different take though"

I actually generally support public financing of elections, but, I too wasn't even aware SF has embarked on it until I read the article, and I can't really comment on it until I read more about it.

"I wonder how the ranked choice system in San Francisco is squared with the top two primary system instituted state wide. Anybody have any info in that regard?"

I'm 98% sure our Top Two law has zero effect on local races. City and County races should be exempt. I could be wrong though.

DavidPSummers said...

The author has a number of complains about IRV. Howerver, not only do I find them "reaching" but it isn't clear why the proposed alternatives would be any better...

One is the number of candidates (16). However, I note that the previous election had 12 candidates. So I'm not sure what is so much worse. Nor is clear why this would be different with approval/range voting (or any other system that means that more than two candidates have a change, which is the point).

The second seems to be that the person with the most number of first place votes didn't win. But that is the point both IRV and also of the systems the author prefers. You have to win because you have the most support of everyone, not because your opponents split the vote of those who don't like you.

Confusion of voters simply means that voters need to become used to the system. Nor is range/approval voting fundamentally easier to understand.

Another complaint is that voters will game the system. This might be more persuasive if it wasn't so vague. The examples are a candidate who can't win a plurality, but moving away from plurality voting is the point, and candidate who supposedly couldn't win a "real" runoff, but IRV _is_ a real runoff (or more properly a series of runoffs). You just have voters vote in all the runoffs at the same time.

Now I find the alternatives the author proposes to be fine choices, but then so is IRV. I find the article an exercise in tearing down a decent solution that you don't want to make room for the one you do want, at the risk of having no solution at all. That is the kind of thing that the partisan system pushes that many of use want to stop. IRV is a lot better than plurality voting and it would be a shame if we ended up doing nothing because everyone was unwilling to compromise on their own preference.

DavidPSummers said...

"
The problems I see with IRV in this election are two fold. One, as pointed out in the link, the election had a "participation paradox": if there were more anti-Perata/pro-Kaplan (the 3rd-place candidate) voters, then Perata would have won the election by IRV. Something like that isn't possible under approval or range (or Borda, for that matter.)"

Is this because voters were restricted to only three choices? If so, then that would seem to be the problem, not the IRV system itself.

Dale Sheldon-Hess said...

Dale: "The problems I see with IRV in this election are two fold. One, as pointed out in the link, the election had a 'participation paradox'"

David: "Is this because voters were restricted to only three choices? If so, then that would seem to be the problem, not the IRV system itself."

No, it is not because of the restriction. Participation paradoxes are possible (say that three times fast!) in all forms of IRV, whether they force truncation to 3 candidates or not, as are nonomonotonicity paradoxes and a host of others, which approval voting and range voting do not suffer from.

If you're interested in all the possible paradoxical results of IRV, this constructed example shows most of them, and this page calculates the probabilities of them occurring (although only in 3-candidate elections) for various voter-models.

d.eris said...

TF, I think you're right, i.e. that top two doesn't apply to county, city and more local offices. It might only apply to statewide offices, state legislature and US Congress?

Rob Richie said...

I posted this message below, after my signof here, at the other blog and thought folks might want to see it.

Damon - - it's baffling that you seem to like runoffs, but don't like instant runoff voting (even editing John Anderson's support for IRV out of your blog on his Christian Science Monitor piece). That doesn't make any sense for someone who professes to want to encourage multi-party elections.

As to approval/range votng, you should read the analysis of why approval voting is SO much more prone to pre-election day gaming in a way that disqualfies as a reasonable alternative in any meaningfully contested election. See: approvalvoting.blogspot.com

Approval voting/range voting folks have absolutely no positive reform impact -- only as defenders of the status quo. That's because their reform has no legs for reasons that become obvious as soon as you try to convince a legislature to pass it. If they don't like IRV or runoffs, they should focus on proportional voting for legislative elections, but they don't lift a finger to help on that either. A real shame, as I know there are at least some well-intentioned folks among them.

Here was the post at the other site...
Rob

########

I'd say it's not so bad for San Francisco voters to have choices - better than 2-choice system.

As to Oakland in 2010, Jean Quan would have defeated Don Perata one-on-one last November. We know that because more voters ranked her ahead of Perata than the other way around, even as she won the most votes by an Oakland mayoral candidate in almost 2 decades.

That race shows why we should have runoffs or instant runoff voting in races with candidates - to find out which of the top candidates does best when matched against the other top candidate.

As to Damon Eris' questions:

* This is not San Francisco's first RCV elections. It has held them every November since 2004. It's the first open seat mayoral election with RCV. For more on Bay Area election history with RCV.
http://www.fairvote.org/ranked-choice-voting-in-bay-area-elections

* Top two doesn't apply in San Francisco. If it did, this field of candidates would have been reduced last June to two candidates in a low turnout election, with the general election only with 2 choices.

d.eris said...

Rob, thanks for your input and the info and error correction on ranked choice in SF.

As for your (implicit) question about my support for runoffs but apparent opposition to ranked choice, I would have to say that I'm not opposed to IRV, ranked choice, etc. I think we just need to realize that it is not the only alternative to plurality. imo, as a country, we need to explore and experiment with as many potential reforms as possible. This is one of the great strengths of the federal system, the old idea that the states are the laboratories of democracy.

What I am opposed to, or, what I take issue with, generally, is the popular inclination to simply cut and paste reforms from one state into another, rather than actually experiment with different reforms in different states.

I would much rather see IRV implemented in one state/city/district etc., approval voting experimented with in another, while others try out forms of proportional representation elsewhere, rather than see everyone kind of jump on to the same reform bandwagon, or do nothing at all.

 
http://www.wikio.com