Epic Fail of Ranked-choice Voting in Oakland

By mw | Related entries in Bad Decisions, Elections, Politics

This post may be a bit too local and parochial for the Donk, but many of the worst political ideas are ignited here on the Left Coast before burning a path of destruction across the rest of the country. Consider this an early warning of smoke on the horizon.

In any case, Oakland is likely to be in the national news again over the next couple of days. To the surprise of absolutely no one with a modicum of common sense, the City of Oakland will again be forced to roust the Oakland Occupiers from their encampment, with Occupy San Francisco not far behind. As before, it is likely that protesters will be injured, perhaps severely, when they provoke and resist police carrying out their assigned duties and responsibilities. This action in Oakland was made necessary by the Mayor of Oakland Jean Quan, who invited the Occupiers back into Oakland after ordering them rousted once before.

Some insight into how Jean Quan came to be in this position of authority:

One year ago, Jean Quan was elected mayor of Oakland. She never led in any poll at any time during the campaign. She always trailed front-runner Don Perata in every minute of the campaign from beginning to end.

On election day, 36% percent of Oakland voters said they wanted Don Perata as their mayor. Only 24% of Oakland voters said Jean Quan was their first choice to be mayor. In prior years, this would have triggered a runoff election and voters would have chosen between Perata and Quan in a head to head contest. Not this year. This was Oakland’s first Ranked-choice Voting election for Mayor. The other candidates on the ballot were eliminated and the second and third choice votes on their ballots were added to Quan and Perata’s totals. Jean Quan became mayor. Oakland saved the cost of conducting a runoff election.

Jean Quan ran a smart and innovative campaign. She asked Oakland voters for their second place votes. Why not? She is likable and her campaign employed fun YouTube ads. People like to give out consolation prizes. Why not give her your second place vote? What harm would it do?

Advocates for the ranked-choice voting system will tell you that if Quan and Perata ran in a runoff election, we would have seen the same result. They claim this was just a more efficient and less costly way to arrive at that result.

Matt Gonzalez is an RCV advocate. His op-ed in the SF Chronicle makes the case for RCV here in San Francisco. I’ll have more to say about his piece later, but this is what he says about the Oakland election:

“Ranked-choice voting results should be identical to those of a traditional runoff … Others argue that everybody’s second-favorite candidate gets elected, citing Oakland’s 2010 mayoral election, which Jean Quan won. But this misses the point. Quan won because she received more votes in a runoff than Don Perata did. The only difference was that the essentially three-way contest (there were 10 candidates total) used ranked-choice voting, which eliminated the need to hold another election a month later – in which fewer voters would have voted. In fact, Quan won more votes in Oakland than any other mayoral candidate had in a generation.”

It is Gonzalez that misses the point. The operative word in this quote is “should”: “Ranked-choice voting results should be identical to those of a traditional runoff…” Sure they should. We just don’t know if they are.

Gonzalez claims that Quan’s plurality of 2nd choice votes produced exactly the same result as we would have seen in a runoff vs. Perata. The truth is that he does not know that for a fact. No one does. It is just his opinion. My opinion is that Quan would never have beaten Perata in a one on one runoff. No one will ever know because Oakland never had that runoff election. The voters were denied the opportunity to make their choice clear. That is precisely the point. If no one knows whether Quan or Perata would have won, Quan’s legitimacy as an elected mayor is open to question and confidence in our democratic process is undermined. Yes – she won according to the ranked-choice rules, but no one knows if that truly reflected the preference of Oakland voters between Quan and Perata.

Now – all of this would be moot if Quan had proven to be a popular and competent mayor. That didn’t happen. So now Oakland voters are facing the question whether they legitimately elected an incompetent, or if they were denied the opportunity to vote for their preferred candidate for mayor.

Let’s take another look at the Matt Gonzalez case for ranked-choice voting:

“Ranked-choice voting results should be identical to those of a traditional runoff, the only exception being that the winner is decided when turnout is highest and big money hasn’t polarized the race. This is better democracy.”

Two things to note – First, he no sooner finishes claiming that ranked-choice voting yields an identical result to a runoff, when he offers an exception. If you have “big money” and a “polarized” race, well – he admits you might get a different result. In other words, Gonzalez is saying we cannot trust the voters to make a decision under those circumstances. “Big money” and “polarized” are subjective pejoratives. Others may substitute terms like “commitment” to and “support” for the candidate they prefer.

More astonishing is his claim that ranked-choice voting is somehow “better democracy”. Step back and think about what he is really saying here. He is asserting that having a real run-off election, letting the voters make a simple, clear choice between two candidates, vote if they want to, vote for the candidate they prefer, adding up the votes to yield an unambiguous decision where the candidates with the most votes wins, is somehow a less good democracy. It is an absurd claim on its face.

Trusting the voters to make a simple choice between the last candidates standing is not a good enough democracy for Matt Gonzalez. According to Matt, we need this New and Improved Ranked Choice Voting Democracy 2.0! A better democracy! Now in a convenient 16-Pack!

He goes on to argue for the qualities that make ranked-choice voting a “better democracy.”:

“.. the winner is decided when turnout is highest and big money hasn’t polarized the race.. With ranked-choice voting, San Francisco has avoided 15 December runoff elections that typically would have resulted in far lower voter turnout, dramatically increased campaign spending from special interests and cost the taxpayers millions to administer (an estimated $3 million this year alone). Old-fashioned door-to-door politics and coalition-building matters more than with the old system, which gave advantages to money politics.”

None of these “better democracy” arguments are supported by empirical fact. All these “better democracy” claims can be distilled into this: Matt does not trust the voters in a runoff election to make the right decision. He fears voters might make a wrong decision in a polarized election. He is concerned voters might be unduly influenced by big money advertising. Matt wants” door-to-door” and “coalition building” candidates to win. Best not to take the risk that voters will choose the wrong kind of candidate in a real runoff. Net net – Matt believes the kind of candidate he prefers would have a better chance getting elected under RCV. Ranked-choice voting is a way to put his thumb on the electoral scale.

There is one and only one good rationalization for Ranked-choice Voting – cost. RCV saves the cost of a runoff election. That is certainly and unarguably true. But it is also unarguable that ranked-choice voting is less good democracy than simply trusting voters in a real runoff. It is utter nonsense to claim that there is a “better democracy” than giving voters a choice between two candidates, let them vote between the two candidates, and declare the one with the most votes the winner.

By utilizing ranked-choice voting, Oakland saved the cost of a runoff election in 2010. They are paying the price of incompetent leadership managing the Occupy Oakland protest now. Oakland will be paying for the additional cost of a recall election in 2012. One occupier paid with his life for Mayor Quan’s indecision. For Oakland, the cost savings of ranked-choice voting are illusory.

We just completed our first ranked-choice mayoral, sheriff, and district attorney race in San Francisco, along with our first public funding of the mayoral candidates. We have yet to learn the real cost of this electoral experiment here.

My suggestion for my comrades a here in San Francisco – If we want better democracy, there is a better way. Trust the voters. and scrap ranked-choice voting in San Francisco before it costs us like it cost Oakland.

X-posted from “Divided We Stand United We Fall


This entry was posted on Monday, November 14th, 2011 and is filed under Bad Decisions, Elections, Politics. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

27 Responses to “Epic Fail of Ranked-choice Voting in Oakland”

  1. Russ G. Says:

    It should provide the same results as a run-off, IF the voters understand what giving somebody their second-place votes means. If they are not familiar with RCV, they could well think that it is just being nice, not expressing the opinion that if there were a runoff, this is whom they would prefer next.

  2. cranky critter Says:

    C’mon, really? That death is the mayor’s fault?

    At the end of the math, ranking is a fully sound approach, unless you presume that people are incapable of ranking things to properly represent their views.

    The other guys’ arguments are flawed and in some cases specious. But I don’t think you’ve made a very compelling argument in favor of a lengthier and more expensive process. Your argument boils down to claiming that the people REALLY NEEDED that final binary choice for their choice to be valid. What, are people too stupid to know whether they prefer coke to pepsi when there are more than 2 beverage choices. Not a very strong argument.

    Gonzalez claims that Quan’s plurality of 2nd choice votes produced exactly the same result as we would have seen in a runoff vs. Perata. The truth is that he does not know that for a fact. No one does. It is just his opinion.

    Well it IS his opinion, but it’s not JUST his opinion. Far and away it’s the most parsimonious expectation. Occam’s razor. We don’t know for certain, of course. But what is the reason to think that someone who ranked Quan 2nd and Perata 3rd realy didn’t prefer Quan? That they were just being nice? Come one. If you ranked Perata 3rd or worse, you didn’t like him very much. That conclusion is unavoidable.

    My opinion is that Quan would never have beaten Perata in a one on one runoff. No one will ever know because Oakland never had that runoff election. The voters were denied the opportunity to make their choice clear.

    Well, in the absence on any compelling argument, THAT really IS just your opinion. I don’t see that voters failed to make their choice clear. Again, your claim is that people are unable to decide whether they like x or y better when there are also other choices. What’s your basis for this.

    The only part that is undeniably true is the possibility that the outcome could have changed had their been more time and more campaigning. But this is true of every election. When a losing candidate says “I think if we had 3 more days I would have won?” That is known as sour grapes.

    I think the best argument that can be made is that voters deserve that final binary choice, so they can be certain, without any alleged distractions. I don’t buy it, and think the cost is un justified, but it’s the best argument. Personally, I thunk it’s insulting to voters to tell them, say, you don’t know whether you prefer Barack Obama to Mitt Romney as long as John Huntsman and Newet Gingrich are also options.

    Again, at the end of the math, ranking is a fully sound approach, unless you presume that people are incapable of ranking things to properly represent their views.

  3. Tully Says:

    Again, at the end of the math, ranking is a fully sound approach, unless you presume that people are incapable of ranking things to properly represent their views.

    Um, no. First, to “presume” is to take your assumptions for granted, which is always a bad idea in a model, and you’re presuming quite a few things there. The claim presumes that [a] ALL other underlying assumptions of the model are correct and precise, [b] that people actually understand the system (and to large extent they don’t), [c] that their ranking preferences of candidates would not change if their top choice(s) were eliminated before a second election, and that [d] other post-vote revelations and events will not change their initial ranking preferences before a subsequent run-off election. The ranking system incorporates such field change assumptions, and those assumptions are questionable, so “presuming” them is not sound application. The math is only as “sound” as the underlying assumptions.

    Voting may be numbers, but people aren’t. They WILL exhibit changing responses to changed conditions. Just as candidates will change positions and tactics post-primary. Is it more “democratic” to limit the voters to one shot at their vote, or to allow them to reconsider their rankings in light of field changes and post-ballot events? Heh.

    MW’s point is at least partially that our brand of democracy greatly prefers a majority vote to win. And you never get a real majority win vote in a ranked system unless you get it in the first count. After that you’re simply presuming the run-off election results with woulda-coulda-shoulda add-on scores derived from the preliminary ranking. At best, you’re meta-aggregating factions together to pick a candidate, when it’s the candidate’s job to do their own post-primary faction aggregation. That itself serves a useful purpose in democracy.

  4. Robert Says:

    This entire scene reminds me very much of a report I recently heard on the radio. It was about when the WWI vets occupied the mall in DC during the depression.
    These vets were promised a bonus check for joining the military during the war and the Federal government chose to ignore them.
    In the end, the leaders attacked them and forced them to move.
    These guys sacrificed everything and just wanted their pay.
    Not that the occupiers have done any such thing, but there are certainly some things to consider here.

  5. kranky kritter Says:

    Tully, my point in saying that is this: it’s not the ranking that is causing the alleged problems. The tool is sound. If people screw up when using the tool, then the bug is with the user. And I think I already pre-acknowledged your objections here.

    You’re right. People will change their preferences over time. BUT that is pretty much always true. To some extent all we’re taking about is a 2nd at-bat.

    There’s a claim here of insufficient consideration and insufficient confidence in the result. How true would that be as folks became familiar with this process?

    I also agree that the process MW favors could be argued to be serving a useful purpose, as you’ve put it. You’re very pro a lengthier and more narrowly focused final chapter that gives the elected candidate the clearest imprimateur possible. I get that. But the rank process achieves a very similar end more efficiently and at less cost.

    So I don’t see the sturm and drang and the push to kill this in the crib. Epic fail? Hardly. Promising alpha test, more likely.

  6. Tully Says:

    You’re talking about the Bonus Army. The BA people in 1932 were demanding immediate payment of their WW1 service bonuses which Congress awarded them in 1924, which bonuses mostly weren’t due until 1945.

    It’s popular myth that the government reneged on those bonuses. It didn’t. The BA people wanted their payments NOW because of the hard times. Also note that these were NOT enlistment or service bonuses promised during the war, but awards voted out years later. Kind of an early version of the 1944 GI Bill, which sought to preemptively avoid the problems of the WW1 veteran’s bonuses.

  7. Tully Says:

    Sorry, KK, but wrong. You say:

    The tool is sound. If people screw up when using the tool, then the bug is with the user.

    Well, no. The tool is not “sound” other than theoretically and then only IF the underlying assumptions are correct in correlation to reality. In conflicts between theory and reality, reality trumps theory every time. Beautiful theories, ugly realities, etc.

    When designing a tool, the prime criteria is “What is it supposed to do?” In this case the purpose of the tool is nothing more nor less than avoiding holding a second election. It has nothing at all to do with facilitating better or more valid electoral results or better democracy. Like MW, I fail to see how avoiding subsequent votes betters democracy.

    Your response reminds me of the multitudes of Marxists I’ve met who say that communism really does work, it just hasn’t been tried by the right people yet. Not Marx’s fault that the people are screwing it up, the tool is perfectly sound … it’s those pesky people that are the problem!

  8. David P. Summers Says:

    In the end, the author is complaining because someone was elected that he didn’t like. I don’t blame him for not liking Quan, her handling of the situation has been abysmal. However, he is wrong to blame instant-runoff voting.

    Quan’s election in spite of his claims, the voter’s choice. Would he have been any happier if Quan had been able to use the primary system to win against weak candidate in a winner-take-all system?

    He claims that voters didn’t want him, but he admits that a lot of voters were comfortable with Quan as their second choice. Conversely, Perata wasn’t able to garner much support beyond the minority that liked him. Clearly Quan had broader support.

    He complains that IRV isn’t like a real run-off election. And what does he want to scrap IRV voting for? If it is for a series of actual run-off elections, I can understand that, though I don’t think the public will be willing to do that and, in the end, people will be elected just because they had people willing to show up for a number of elections. HOwever, if he wants to scrap it for the old winner-take-all system, he is just returning to a system that produces even more bad results.

  9. cranky critter Says:

    The tool is sound if it works exactly as designed and intended. And it did.

    You haven’t demonstrated a bit of harm that isn’t itself theoretical. Which makes all your lovely theory-reality schtick kind of silly to apply here, even though it’s true stuff.

    At no point did I say it was better. I said it was efficient and saved money. (It saves time too). I asked you to demonstrate the “worse” you’ve claimed.I don’t see it.

    There’s plenty of reason to think people can and will get used to it. And little or no reason to think anything valuable will be lost. Unless you collect paychecks working for political campaigns, of course.

  10. Tully Says:

    LOL. IOW, you don’t care if it’s LESS democratic and produces worse results, the theory is sound! (Repeat observation on Marxism and right people.) Sorry, the rationalization still doesn’t wash, however much you love it in theory. Review my repeated posts on the Law of the Instrument, and apply it in this context. You really want to drive that screw with a hammer, don’t you? And that hammer is still soundly designed for the job by your standard, even though the end product results in mangled screw heads and dented wood and framing that can’t be easily remodeled without tearing the frame members to unusable bits. But that hammer’s still the wrong tool for the intended job, and it’s up to you as a booster to show it does an equal or better job while causing no harm. And you can’t. So you try to redefine the job to fit the tool!

    MW goes straight to the nut of it here:

    Gonzalez claims that Quan’s plurality of 2nd choice votes produced exactly the same result as we would have seen in a runoff vs. Perata. The truth is that he does not know that for a fact. No one does. It is just his opinion. My opinion is that Quan would never have beaten Perata in a one on one runoff. No one will ever know because Oakland never had that runoff election. The voters were denied the opportunity to make their choice clear. That is precisely the point. If no one knows whether Quan or Perata would have won, Quan’s legitimacy as an elected mayor is open to question and confidence in our democratic process is undermined.

    The undermining of confidence in democracy and elected leaders is indeed a harm. And it certainly has undermined HIS confidence in both his local democracy and the current cheif exec of Oakland. So there is your demonstrated harm. Unless you want to call MW a liar.

    Reminder: MW has noted why RCV can indeed be an active harm, and that the claim that it should produce identical results is specious and unproven. You’ve already conceded that RCV can produce different results than the traditional primary/runoff system it replaced. The burden of proof is on the claimant, and the claim being disputed is that RCV will produce identical results to the traditional system while costing less, the first part of which you’ve already conceded ain’t so.

    I can think of another election scheme that would produce even lower costs than RCV, so it’s not the ideal tool for THAT purpose either. Castro holds such elections. They meet your criteria for a “soundly designed” tool for the purpose you claim to desire even better than RCV does. :-)

    I can think of at least one regular multi-candidate national election that would likely have come out differentlyon multiple occasions if RCV had been used. Though not a primary, the principle problems MW elucidates apply to ANY multi-candidate election to which RCV is applied. I’m thinking of the Presidential elections of ’92 and ’00, where eliminating the Perot vote and reassigning it would likely have elected Bush 1 to a second term, and eliminating Nader would have swung the election to Gore.

    The only place I think MW’s arguemnt is weak is his contention that the ill effects of Quan’s incompetence would have been avoided had Oakland had a traditional election and Perata been elected instead. Perata would probably have done better, but as MW himself points out, we’ll never know. However his argument doesn’t stand on that contention, so it’s a side issue. As is the issue of what our own national experience would have been under Bush 1 Term 2, or under a Gore admin. The core issue is that RCV purportedly* trades off lower cost against clear democratic choice.

    [*--likewise an unproven contention. We don't know that it's actually cheaper either, particularly when candidates get piles of public money for running. Once again, theory is fine but it ain't reality until you prove it in the real world.]

  11. cranky critter Says:

    At no point have I seen any actual accounting of the alleged harm to democracy, it’s all 100% allegations from both of you. That one person feels unhappy is not relevant, as data collectors and lovers like you well know. What would matter is the people overall felt the system had been undermined, and if, over time, they continued feel that way instead of grasping the condensed process.

    Where I agree with you 100% is that the people (as opposed to 1 person) are the measuring stick that matters. If a tool works exactly as designed, it’s a sound tool. If people don’t like the tool and don’t want to use it, then it is still a failure as a product. I know you well enough to know that you would NEVER EVER EVER accept a single person’s account as conclusive proof of harm. Shame on you for sporting it here.

    IF the people in general feel unhappy and unsettled by this process, then you’re right. But if they accept it and are fine with it, then you’re wrong.

    There’s one part where you repeat MW’s specious claim, that the people were denied the right to make their choice clear. That’s silly. There is NOTHING unclear about a ranking. It’s crystal clear.

    I will happily concede that such a system could get unwieldy if there were many candidates and no frontrunners. Can you envision a realistic circumstance under which an unknown could emerge? I really can’t. Now that I think about that, I can see how if there were 2 frontrunners, and one groups supporters were more determined to rank the other guy last, odd things could happen.

    How is it unproven that the city would save money by having one election instead of 2? What am I missing?

  12. mw Says:

    @David Summers

    “…what does he want to scrap IRV voting for?… if he wants to scrap it for the old winner-take-all system, he is just returning to a system that produces even more bad results. – DS

    I don’t know what you mean by a “winner take all system”. However, I will point out that you are doing here exactly what you accuse me of doing – not liking the outcome, so wanting to change the election rules.

    I’m guessing that by “winner take all” you are referring to a “first across the line” election where the largest number of votes determines the winner (even if less than a majority of votes cast) – where a candidate with a plurality but not a majority can be elected. While I do think that even that is better than RCV, it would not be my preference.

    My preference would be the use of a single runoff election between the top two candidates. Using an RCV election to arrive at the top two would be fine by me, but I would also be perfectly content with using the candidates with two highest totals of first place votes. However, I do want to see a clear unambiguous expression of the voters preference between those two candidates so there is no doubt about the intention and will of the voters. A runoff unequivocally stamps the electorate’s imprimatur of democratic legitimacy on the elected. Oakland did not get that with RCV in their 2010 election.

    @ CC 11-14-11 10:01

    C’mon, really? That death is the mayor’s fault? – cc

    I guess if I was a lawyer, I would argue that while she was clearly not the direct cause, she was certainly a proximate cause of the protester’s death as there was a foreseeable risk of that eventuality. The encampment was known to harbor a criminal and potentially violent element. The police were instructed to stay out of the encampment unless invited in by the Occupy Oakland “security”. If she had not invited the protesters to reoccupy after the first eviction this death would not have occurred – Chip Johnson:

    “Even Quan couldn’t ignore last week’s shooting death of an Occupier by another Occupier just outside the encampment. That pretty much sealed the camp’s fate.”

    Completely predictable. Hard to see how she does not bear some responsibility for this murder.

    As to the rest – What Tully said.

    @CC 11-15-11 11:05

    “There is NOTHING unclear about a ranking. It’s crystal clear.” – cc

    Not to many San Francisco voters. Many voted for the same candidate three times. Many did not vote for second or third choices. Confidence in the system is undermined and legitimacy of the elected is open to question – particularly if a candidate with a large plurality of first choice votes loses the election – as happened in Oakland.

    Astonishingly, uber-liberal Jon Avalos (2nd place finisher in the SF Mayoral race here) now wants to complicate it even more. He suggests that voters be permitted to cast all three votes for the same candidate, and when there is a large field, expand the ranking – so that voters would need to rank up to nine candidates. Really. I guess we need to keep cranking up the complexity until he gets elected mayor.

  13. cranky critter Says:

    There’s a general point here that I strongly agree with, MW. Too much complexity is a bad thing. And I strongly agree that IF the people really are confused and so untrusting of the method, then it has to go.

    I know a very little bit about game theory, and I’ve had occasion to review other folks thoughts about fairly elaborate alternate methods of voting, which I would call baroque, or even preposterous. As you and Tully both suggest, it’s extremely important that the outcome feel like a pretty clear expression of the public’s preference.

    I don’t feel like it’s beyond the capability of regular folks to be able to grasp and use a ranking system for the sake of efficiency, even if there are initial bugs. It seems eminently possible that folks will be fine with this approach. They may even come to prefer it if it decreases the length of the campaigning portion of each election cycle.

    But I don’t buy the precious claims that it’s more democratic or a better version of democracy. That stuff is all special pleading to me.

  14. David P. Summers Says:

    >mw Says:
    >November 15th, 2011 at 8:25 pm
    >
    >@David Summers
    >
    > “…what does he want to scrap IRV voting for?… if he wants to scrap it >for the old winner-take-all system, he is just returning to a system that >produces even more bad results. – DS
    >
    >I don’t know what you mean by a “winner take all system”.

    I mean the old previous system where you have one election with only one round and the person with the most votes wins.

    > However, I will point out that you are doing here exactly what you >accuse me of doing – not liking the outcome, so wanting to change the >election rules.

    I’m not sure where you get this from. I agree with you that Quan has been a poor a mayor and I don’t have much opinion either way about Lee.

    >I’m guessing that by “winner take all” you are referring to a “first across >the line” election where the largest number of votes determines the >winner (even if less than a majority of votes cast) – where a candidate >with a plurality but not a majority can be elected. While I do think that >even that is better than RCV, it would not be my preference.

    Well, that system is even more subject to manipulation than IRV and has the added feature that the two parties can use it to restrict competition and accountability. If you can look at American politics, with the partisanship, inability to get anything done, and the politics of person destruction and say there is no problem, then sure IRV isn’t for you. But I don’t see how you can.

    >My preference would be the use of a single runoff election between the >top two candidates.

    There is no evidence this would have made a bit of difference. Looking at people’s first choice (I guess, under the old system, people might have voted for someone who wasn’t their first choice, but having to vote for someone you don’t really want is, to me, the big problem), it would have been Quan vs Perata. Those who voted for Kaplan and Tuman generally chose Quan. Now maybe if their had been a run-off, Perata would have been able to sell himself head-to-head. However, what would he do that he didn’t do before, except attack Quan. Is what we need more negative campaigning? We both may not like how Quan has performed, but there is not reason to believe the expense time of a run-off will produce winners who are better leaders.

    Don’t get me wrong, I don’t have any fundamental objection to using actual run-offs (though a singe run-off round isn’t quite as good). But it seems likely the repeal of IRV is going to mean going back to a broken “first across the line”. And, in the end, there is no clear reason why it is better (any different really).

    >Using an RCV election to arrive at the top two would be fine by me, but I >would also be perfectly content with using the candidates with two >highest totals of first place votes. However, I do want to see a clear >unambiguous expression of the voters preference between those two >candidates so there is no doubt about the intention and will of the >voters. A runoff unequivocally stamps the electorate’s imprimatur of >democratic legitimacy on the elected. Oakland did not get that with RCV >in their 2010 election.

    The voters for Tuman broke pretty evenly for Quan and Perata, but the voters for Kaplan chose Quan over Perata by a margin of about 3-1 in the later rounds. That seem pretty unequivocal to me. I’m not sure how making people get in their cars and wait in line would change that.

  15. David P. Summers Says:

    “Many did not vote for second or third choices.”

    Of course many also don’t bother to vote in run-off elections….
    http://www.jstor.org/pss/2131348
    Australia and India seem to be able to conduct IRV elections, so I don’t think it is beyond San Francisco or Oakland voters.

    Though, in the end, even if a voter completely fails to vote for more than one choice, they are no worse off than the old system where they could only vote for one candidate anyway.

  16. Tully Says:

    Damn site ate my comment. I’ll forbear some, but …

    I think we can all agree that transparency is a requirement for functional elections. RCV is enormously less transparent than the traditional primary/runoff structure. In other contexts (non-political) I have supervised and tallied in RCV ballot scenarios. It’s enormously more complex than a primary/runoff structure. People do not really understand the system, and in statutorily-regulated elections I would expect to see much higher rates of ballot spoilage, and contested results could be a legal nightmare that would make Bush v. Gore look downright kindergarten-simple, not to mention brief. I have seen fourth-place entries eventually win after all the “adjustments” were made. Second- and third-place first-rounders eventually winning are not uncommon.

    Applying that to the domestic context, think of President Perot, a great potential had the ’92 election utilized RCV. After all, given known and demonstrated partisan animosities, neither Bush nor Clinton voters were liable to list the “other side” candidate as their SECOND choice. Think about that in light of the fact that the heaviest backers of RCV are also the prime backers of the “National Popular Vote” movement. (Among them is, unsurprisingly, John Anderson. Your own mileage there.)

    There’s also the Aspen example. Aspen used RCV for ONE election, and it resuled in such a FUBAR that the citizenry subsequently voted the method back out of use by a 2-1 margin. To illustrate the problems, the top first-round vote-getter eventually lost, but would have won if 75 of the voters who listed him 1st had instead listed him 2nd. The only people who genuinely liked the system seem to have been the eventual winner and the vote-machine contractor. There were problems with over 12% of the ballots, which is quite high, and 7% of the absentee ballots had overvotes — and mind you, Aspen is a wealthy community with one of the highest education rates anywhere, with 80% or more of the residents holding 4 yr degrees.

    I can’t see that as an improvement in the democratic electoral process over the traditional primary/runoff structure, however nifty CC may think of RCV as a conceptual hammer for driving in electoral screws. The only true utility I can see is avoiding having that subsequent runoff election, something that seems a lot more useful in a one-off process ballot that can’t be repeated than in the public arena. And the runoff itself serves a useful purpose, as noted previously. Aspen residents complained later that a runoff would have been very helpful to them, allowing them better knowledge of the top candidates before making a final decision. Knowledge that was tough to come by in a crowded race with limited “discovery” time available before the vote.

  17. cranky critter Says:

    Well, I have to bow to your experience-based testimony.

    To illustrate the problems, the top first-round vote-getter eventually lost, but would have won if 75 of the voters who listed him 1st had instead listed him 2nd.

    OK, clearly I didn’t take the time to properly understand how the method works (or doesn’t). Because that sounds preposterous. I withdraw my support, and concede that I am a jackass.

  18. David P. Summers Says:

    Tully Says:
    November 16th, 2011 at 5:19 pm

    Damn site ate my comment. I’ll forbear some, but …

    I think we can all agree that transparency is a requirement for functional elections. RCV is enormously less transparent than the traditional primary/runoff structure. In other contexts (non-political) I have supervised and tallied in RCV ballot scenarios. It’s enormously more complex than a primary/runoff structure.

    I don’t what “lack of transparency” exists. In the Oakland and SF races you can see who voted for whom and why it turned out the way it did.

    I’m also not sure why “indicating how you want to vote in any run-offs where you first choice has been eliminated” is so complex. Instant Run-off voting goes back to at least 1984 in Cambridge (they had it when I lived there) and they use it in Australia and India. I hardly think it was warranted to throw up one’s hands and say voters can’t understand it.

    It might be harder to tally (though even then I’m not sure it is harder to tally than n separate election), but who cares. It is not like having the tally people take another day to do it really hurts anything.

    People do not really understand the system, and in statutorily-regulated elections I would expect to see much higher rates of ballot spoilage, and contested results could be a legal nightmare that would make Bush v. Gore look downright kindergarten-simple, not to mention brief. I have seen fourth-place entries eventually win after all the “adjustments” were made. Second- and third-place first-rounders eventually winning are not uncommon.

    Your examples have to be of winning, after “adjustment”, in a very close election. Of course the examples you cited (only voting for one candidate and marking one candidate multiple times) aren’t that hard to handle. You just treat multiple votes for one candidate as one vote for only that candidate (though I see in SF it was even easier, they just pointed out the error and had them vote again). Voting for only one candidate is allowed and if someone does it, he still has had just as much choice as he had under the old system (they got to vote for one guy they like), hardly a reason to take having more choice away from everyone else.

    It is also true that the same thing happens with old “first across the line” election. The senate race where Franken was elected was after votes were “adjusted”.

    Applying that to the domestic context, think of President Perot, a great potential had the ’92 election utilized RCV. After all, given known and demonstrated partisan animosities, neither Bush nor Clinton voters were liable to list the “other side” candidate as their SECOND choice. Think about that in light of the fact that the heaviest backers of RCV are also the prime backers of the “National Popular Vote” movement. (Among them is, unsurprisingly, John Anderson. Your own mileage there.)

    Yes, the supporters of the two big parties might be so polarized as to be unwilling to put in a second choice. I don’t see why that is a problem, they had as much choice as they wanted and if they don’t want more, I don’t see why they should take our ability to not vote for “major party approved” candidate away from use.

    It is indeed no surprise that John Anderson supports IRV. He has seen how the two party system systematically excludes those who don’t win major party approval. I don’t like it either…

    There’s also the Aspen example. Aspen used RCV for ONE election, and it resuled in such a FUBAR that the citizenry subsequently voted the method back out of use by a 2-1 margin. To illustrate the problems, the top first-round vote-getter eventually lost, but would have won if 75 of the voters who listed him 1st had instead listed him 2nd.

    Of course in a the old system, you can similarly game the system. For example, if you have a safe primary, you can (like Grey Davis did) put out adds attacking the more dangerous candidate and drive him down to get a weaker opposition.

    Heck in a straight run-off election you can pull the _same_ trick you are citing with IRV (after all, it is the same thing), namely vote for someone you don’t like (but consider a weaker opponent) so they can edge out someone who you actually like better (but presents a weaker foe for your guy). You game any system (and you can game a regular run-off system in the exact same way you can game IRV). So what is better, IRV or the old system. Something that could have been gamed in a really close election (if you don’t mind risking blowing it entirely), or the old system which can also be gamed (in may ways worse, because it doesn’t have to that close to elect someone most people don’t want) but which _also_ routinely excludes choice?

    The only people who genuinely liked the system seem to have been the eventual winner and the vote-machine contractor. There were problems with over 12% of the ballots, which is quite high, and 7% of the absentee ballots had overvotes — and mind you, Aspen is a wealthy community with one of the highest education rates anywhere, with 80% or more of the residents holding 4 yr degrees.

    I can’t see that as an improvement in the democratic electoral process over the traditional primary/runoff structure, however nifty CC may think of RCV as a conceptual hammer for driving in electoral screws.

    You mean the old system that systematically divides the vote of the center (discouraging moderate candidates), the old system which (though gerrymandering and other tools) politicians are becoming increasing adept at “managing” (some would say “rigging”), the old system which marginalizes third parties (another source of options for voters who don’t want the stranglehold of the main parteis), and which is a zero sum game (which means that mean and dirty campaigning works better)? I have to say I think almost anything would be an improvement over that. I’m not sure a random lottery would be any worse.

    The only true utility I can see is avoiding having that subsequent runoff election, something that seems a lot more useful in a one-off process ballot that can’t be repeated than in the public arena. And the runoff itself serves a useful purpose, as noted previously. Aspen residents complained later that a runoff would have been very helpful to them, allowing them better knowledge of the top candidates before making a final decision. Knowledge that was tough to come by in a crowded race with limited “discovery” time available before the vote.

    Well, if your article was about how true run-off elections would be better than IRV, I would still (in fact) disagree with you. But it would be a much smaller disagreement. I’m not sure why another round where candidates are freer to go negative (since it is a smaller field) is worth the expense of running another elections and forcing everyone to drive someplace and stand in line. But it still solves the main problem I see. In fact, if you spend a few years with a traditional run-off system, I think people will start wondering why they just don’t go to an IRV. But lets be clear, the old broken system is regarded as the default, and if you are simply saying “repeal” IRV, you are pushing toward going back to the old system.

  19. David P. Summers Says:

    [I messed up the blockquoting. If anyone has been confused, here is the section that was messed up again, with the error corrected...]

    Tully Says:
    November 16th, 2011 at 5:19 pm
    There’s also the Aspen example. Aspen used RCV for ONE election, and it resuled in such a FUBAR that the citizenry subsequently voted the method back out of use by a 2-1 margin. To illustrate the problems, the top first-round vote-getter eventually lost, but would have won if 75 of the voters who listed him 1st had instead listed him 2nd.

    Of course in a the old system, you can similarly game the system. For example, if you have a safe primary, you can (like Grey Davis did) put out adds attacking the more dangerous candidate and drive him down to get a weaker opposition.

    Heck in a straight run-off election you can pull the _same_ trick you are citing with IRV (after all, it is the same thing), namely vote for someone you don’t like (but consider a weaker opponent) so they can edge out someone who you actually like better (but presents a weaker foe for your guy). You game any system (and you can game a regular run-off system in the exact same way you can game IRV). So what is better, IRV or the old system. Something that could have been gamed in a really close election (if you don’t mind risking blowing it entirely), or the old system which can also be gamed (in may ways worse, because it doesn’t have to that close to elect someone most people don’t want) but which _also_ routinely excludes choice?

    The only people who genuinely liked the system seem to have been the eventual winner and the vote-machine contractor. There were problems with over 12% of the ballots, which is quite high, and 7% of the absentee ballots had overvotes — and mind you, Aspen is a wealthy community with one of the highest education rates anywhere, with 80% or more of the residents holding 4 yr degrees.

    I can’t see that as an improvement in the democratic electoral process over the traditional primary/runoff structure, however nifty CC may think of RCV as a conceptual hammer for driving in electoral screws.

    You mean the old system that systematically divides the vote of the center (discouraging moderate candidates), the old system which (though gerrymandering and other tools) politicians are becoming increasing adept at “managing” (some would say “rigging”), the old system which marginalizes third parties (another source of options for voters who don’t want the stranglehold of the main parteis), and which is a zero sum game (which means that mean and dirty campaigning works better)? I have to say I think almost anything would be an improvement over that. I’m not sure a random lottery would be any worse.

    The only true utility I can see is avoiding having that subsequent runoff election, something that seems a lot more useful in a one-off process ballot that can’t be repeated than in the public arena. And the runoff itself serves a useful purpose, as noted previously. Aspen residents complained later that a runoff would have been very helpful to them, allowing them better knowledge of the top candidates before making a final decision. Knowledge that was tough to come by in a crowded race with limited “discovery” time available before the vote.

    Well, if your article was about how true run-off elections would be better than IRV, I would still (in fact) disagree with you. But it would be a much smaller disagreement. I’m not sure why another round where candidates are freer to go negative (since it is a smaller field) is worth the expense of running another elections and forcing everyone to drive someplace and stand in line. But it still solves the main problem I see. In fact, if you spend a few years with a traditional run-off system, I think people will start wondering why they just don’t go to an IRV. But lets be clear, the old broken system is regarded as the default, and if you are simply saying “repeal” IRV, you are pushing toward going back to the old system.

  20. Tully Says:

    But CC, you’ll deprive me of the chance to pedantically pontificate some more! (Yeah, and like we’re not all jackasses …)

    Yeah, the Aspen experience shows that RCV really can become a major FUBAR in excess of the traditional primary/runoff structure SNAFU. When everything works right, great. No problem. When it doesn’t the available permutations of what can go wrong are multiplied, as are the probabilities.

    DPS, you miss the point with the Aspen fiasco. No one gamed the system in Aspen (as Quan did overtly in Oakland). It worked exactly as designed, if not exactly as advertised. Poorly and in unexpected (but predictable) ways. Encountering the real world showed there were huge potentials for startlingly adverse effects in both execution and results, and in Aspen they materialized. And the Aspen election was a small election by comparison to cities such as Oakland and SF. The rest of your post sounds like facile cheerleading rationalization to me.

    I note we only see RCV elections here in America in (purportedly) non-partisan elections, where the field is open to everyone to enter regardless of party. That seriously kicks the props out from under the “locked in a partisan box” arguments. Those are typically local/municipal elections. The only places where RCV is in place in America seem to be overwhelmingly “owned” by Democrats as guaged by voter registrations, so likewise there. If you and yours wish to run your municipality that way, fine by me. You get to live with the results. But please spare me the party pleadings.

  21. David P. Summers Says:

    Tully Says:
    November 17th, 2011 at 12:27 pm

    But CC, you’ll deprive me of the chance to pedantically pontificate some more! (Yeah, and like we’re not all jackasses …)

    Yeah, the Aspen experience shows that RCV really can become a major FUBAR in excess of the traditional primary/runoff structure SNAFU. When everything works right, great. No problem. When it doesn’t the available permutations of what can go wrong are multiplied, as are the probabilities.

    DPS, you miss the point with the Aspen fiasco. No one gamed the system in Aspen (as Quan did overtly in Oakland). It worked exactly as designed, if not exactly as advertised. Poorly and in unexpected (but predictable) ways. Encountering the real world showed there were huge potentials for startlingly adverse effects in both execution and results, and in Aspen they materialized. And the Aspen election was a small election by comparison to cities such as Oakland and SF.

    Whether or not your objection was that is was deliberately gamed (as was done by Davis under the old system and, debatably, may have been done by Quan) or merely didn’t work as you expect, the fact remains that _same_ situation arises under a traditional run-off elections. This is _not_ something that was introduced by the IRV method.

    If you object to that, then you have to object to holding any type of run-off elections. In the end, you either want a new voting method that allows real choice beyond what the two party system provides you and doesn’t expect candidates to have anything but a plurality of support, or you don’t. The “first past the post” system is simple. However, it also, on a routine basis, allows the parties to manipulate the system, freezes out choices, promotes negative campaigning and partisanship, and is even more prone to artifacts and gaming (manipulation of primaries by the other party, election of minority candidates through split opposition, etc. all occur all the time, not just now and then in a close election).

    The rest of your post sounds like facile cheerleading rationalization to me.

    It was, IMO, quite specific on the points and the fact that you chose to respond to it with a vague characterization that could be used to duck any points you don’t think you can take on is worth noting.

    If you want to talk about the issue, then do so. If you want to make snide remarks about people you are talking with, the partisan blogs love that sort of thing.

    I note we only see RCV elections here in America in (purportedly) non-partisan elections, where the field is open to everyone to enter regardless of party. That seriously kicks the props out from under the “locked in a partisan box” arguments. Those are typically local/municipal elections.

    I see why this conclusion holds. If you are locked in a partisan box then you will try and find ways out of it.

    The only places where RCV is in place in America seem to be overwhelmingly “owned” by Democrats as guaged by voter registrations, so likewise there. If you and yours wish to run your municipality that way, fine by me. You get to live with the results. But please spare me the party pleadings.

    For someone who just tried to attack how I make my arguments, this is an awfully vacuous statement. I could say that if you want to have a locality where entrenched parties tell you what choices you get to vote for, fine by me, but don’t take away my ability to have more choice. The fact is that we all have to live with how elections are run and my view is just as valid as yours.

  22. Tully Says:

    LMAO. Seems I poked someone’s True Believer button. I’ll just let that rant speak for itself, complete to its battalion of straw men.

    But I will note one thing: YOU are the one who raised the “gaming” issue, not me. That’s YOUR straw man you’re trying to torch there. I just pointed out (after YOU tried to make an issue of it) that of the two specific elections discussed here, as an issue it didn’t appear to apply to Aspen and obviously did apply to Oakland. I’m pretty sure that most of us are quite aware that elections of any kind can be and are often gamed.

  23. David P. Summers Says:

    Tully Says:
    November 17th, 2011 at 6:07 pm

    LMAO. Seems I poked someone’s True Believer button. I’ll just let that rant speak for itself, complete to its battalion of straw men.

    But I will note one thing: YOU are the one who raised the “gaming” issue, not me. That’s YOUR straw man you’re trying to torch there. I just pointed out (after YOU tried to make an issue of it) that of the two specific elections discussed here, as an issue it didn’t appear to apply to Aspen and obviously did apply to Oakland. I’m pretty sure that most of us are quite aware that elections of any kind can be and are often gamed.

    Heck, I just answered you post point for point. (Including that your objection to what happen to Aspen has little to do with IRV, which you seem to be interested in ignoring in favor of side issues) I guess talking about the issues, rather than labeling people as “true believers” and stuff, is novel to you. (It sure does help you dismiss arguments without considering them). In the end, I’m very happy to let my post stand on its own and if others have any real responses, I would be happy to hear them.

  24. Tully Says:

    Heh. But others can read for themselves, note the order of the posts, and sort it out if they like.

    My points still stand — RCV is only currently used in the US in basically one-party areas where it is used in “non-partisan” open elections. In application it has been shown to be vulnerable to serious problems in execution, and to at times produce some seriously weird results that do not comport with the claims made for it, nor with the traditional American notions of democracy. That latter may well be the attraction for some — YMMV as to why. MW’s original posting stands pretty well unrebutted.

    Having been an election judge and ballot counter on multiple occasions and contexts, including where RCV was involved, I think it generally sucks in any but a low-level context where actual runoffs are not practical, and think the label of “Instant Runoff Voting” is misleading as RCV can (and DOES) indeed produce different results than those produced by the traditional primary/runoff structure. In addition, it lacks the utility functions of a traditional runoff, where the field is already narrowed and the voters have additional time to focus on and consider the issues and candidates still standing. Some think that’s a feature, because campaigning annoys them. Others may find the idea that others should take that opportunity for additional information and consideration away from them to be patronizing, at best.

  25. David P. Summers Says:

    Heh. But others can read for themselves, note the order of the posts, and sort it out if they like.

    My points still stand — RCV is only currently used in the US in basically one-party areas where it is used in “non-partisan” open elections.

    Well, aside form the fact that I find this to be just a “the wrong people are using it, so it must be bad” kind of argument, breaking out of the two parties is of interest to many on the left and right, who get shut out by the two party system just as much as moderates.

    The fact is that in districts that very liberal or conservative, there is a need for alternatives to the the liberal or conservative party one would be expected top support without thinking, however, there is also need for local politics to “mesh” with the larger political system. Runoff elections help them address these needs.

    In application it has been shown to be vulnerable to serious problems in execution, and to at times produce some seriously weird results that do not comport with the claims made for it, nor with the traditional American notions of democracy. That latter may well be the attraction for some — YMMV as to why. MW’s original posting stands pretty well unrebutted.

    Well, agree with you on one thing. People can read for themselves and see the counter points that multiple people have made. Including the about the point that was suppose to be such and example, Aspen, but which actually doesn’t have anything to do with IRV itself.

    Having been an election judge and ballot counter on multiple occasions and contexts, including where RCV was involved, I think it generally sucks in any but a low-level context where actual runoffs are not practical, and think the label of “Instant Runoff Voting” is misleading as RCV can (and DOES) indeed produce different results than those produced by the traditional primary/runoff structure. In addition, it lacks the utility functions of a traditional runoff, where the field is already narrowed and the voters have additional time to focus on and consider the issues and candidates still standing. Some think that’s a feature, because campaigning annoys them.

    Well, if you don’t think the old system is broken, then you have no interest in IRV. But I do think it is broken, then producing different results something to like in it.

    As to IRV vs a traditional run-off. I don’t see that another round where, because of a narrow field, we can have air waves filled with character assassination, is really worth the cost and making people drive and wait in line. However, I don’t really object to tradiational run-offs. They waste effort but they do achieve the necessary reform. In fact, my guess is that as people get used to them, they will lay the ground work for later moves to IRV to accomplish the same thing at lower cost.

    Others may find the idea that others should take that opportunity for additional information and consideration away from them to be patronizing, at best.

    Well, given your remarks about “true believer” and such, I found your own comments to be more than a little patronizing…

  26. Tully Says:

    I think I can encapsulate the lack of substance in your verbiage and your attempts to evade the point with your own dismissal of the very relevant example of Aspen’s RCV clusterflorp. You don’t like the real-world example, so you attempt to simply dismiss it. Likewise with the factual points on all the other real-world examples. Rather than attempt to come to grips with the points, you simply dismiss and evade. They are indeed a glaring example of facile rationalization. (Also of the layering of logical fallacies.)

    There are still some perfect systems by your overt and implied metrics available, of course, as I also pointed out earlier to CC.

  27. David P. Summers Says:

    I think I can encapsulate the lack of substance in your verbiage and your attempts to evade the point with your own dismissal of the very relevant example of Aspen’s RCV clusterflorp. You don’t like the real-world example, so you attempt to simply dismiss it. Likewise with the factual points on all the other real-world examples. Rather than attempt to come to grips with the points, you simply dismiss and evade. They are indeed a glaring example of facile rationalization. (Also of the layering of logical fallacies.)

    Well, your “encapsulation” avoids the point I made above. You completely failed to address the point that your “example” isn’t what you say it is. It could have occurred the same way in any run-off system, not just IRV. I would say you should look to yourself as the one who is trying to dismiss facts because they don’t want to address them.

    If you can’t even respond the point that changing votes to drop someone out of the first round works the same in any runoff system, you can’t claim to be the one addressing fact. To this, add the irony that you are using claims that I won’t address facts to avoid addressing facts.

Leave a Reply


NOTE TO COMMENTERS:


You must ALWAYS fill in the two word CAPTCHA below to submit a comment. And if this is your first time commenting on Donklephant, it will be held in a moderation queue for approval. Please don't resubmit the same comment a couple times. We'll get around to moderating it soon enough.


Also, sometimes even if you've commented before, it may still get placed in a moderation queue and/or sent to the spam folder. If it's just in moderation queue, it'll be published, but it may be deleted if it lands in the spam folder. My apologies if this happens but there are some keywords that push it into the spam folder.


One last note, we will not tolerate comments that disparage people based on age, sex, handicap, race, color, sexual orientation, national origin or ancestry. We reserve the right to delete these comments and ban the people who make them from ever commenting here again.


Thanks for understanding and have a pleasurable commenting experience.


Related Posts: