(“When I arrived that evening, I felt like I might’ve been one of the only “undecideds” in attendance.” This writer has written about his or her feeling of confusion regarding the Burlington Ballot item “repeal of IRV” slated for a vote March 2nd, 2010. – promoted by Maggie Gundersen)
Originally posted at http://asrblog.com/2010/02/19/…
For those readers who neither live in Burlington, Vermont, nor are political junkies, IRV stands for Instant Runoff Voting, and it has been the center of some local controversy recently. A few years ago, Burlington voted in a referendum to change its voting system by which the mayor is elected from requiring a plurality of at least 40% to IRV, and the most recent mayoral election was the first second in which the mayor was elected by the new system. That election has been at the center of the aforementioned controversy, and there is now an item on the upcoming ballot proposing IRV’s repeal.
Before I get into the controversy, however, we’ll first run over the whys and hows of Instant Runoff Voting.
The traditional voting system with which we’re all familiar usually gives each person one vote, and if the candidate with the most votes fails to reach the required threshold (usually 50%, in the case of Burlington, 40%), a run-off election is held between the top two vote-getters a few weeks later. However, there’s a glaring problem with that voting system; namely, the “spoiler effect”. Instead of allowing people to vote their consciences, they are incentivized to vote for the least personally objectionable of the two most popular or visible candidates; hence in the 2000 election, people could claim that, by voting for Ralph Nader, people on the left were helping the right by “spoiling” the election for George Bush. For folks who aren’t happy with having a two-choice political system, the spoiler effect is deeply problematic.
IRV was designed to eliminate the spoiler effect by allowing people to express their opinions about the full slate of candidates by ranking them. To use a hypothetical example to illustrate this, let’s assume that the Presidential election had been conducted by IRV, with John McCain, Barack Obama, Ron Paul, and Ralph Nader on the ballot. In a traditional election, if most people supported Barack Obama and John McCain, voting for Paul or Nader could be considered helping the other side. However, under IRV, a hypothetical conservative might vote:
1. Paul (First choice, most in line with personal beliefs)
2. McCain (Not such a fan, but would prefer him to Obama or Nader)
3. Obama (Dislikes, but not as much as that commie Nader)
4. Nader (Bottom of the heap)
If Paul were eliminated in the first round, all first-round Paul voters’ votes would then be transferred to their second choices (in the case of our hypothetical voter, McCain), thus, in theory, eliminating the bystander effect and allowing each voter to have a more nuanced influence on the election which reflects their personal beliefs rather than being intensively molded by the logic of lesser-evilism.
I’d long thought this was a great idea, and was quite cynical when Burlington’s Republican mayoral candidate Kurt Wright decried the system after losing the last election. However, the fact that 2,000 signatures were collected in order to get its repeal on the ballot, combined with some people (whose opinions I deeply respect) prodding me about the issue, forced me to take a second look. Essentially, the anti-IRV camp argues that, in the last election, Bob Kiss was elected by an “artificial majority”, as he wasn’t the recipient of the most first-choice votes. Instead, its asserted, 71% of Burlingtonians voted against Kiss in the four-way race (“voting against” meaning he wasn’t their first choice). A lot of numbers were being thrown about and I was having trouble wrapping my head around the issue, so I was delighted to hear that there was a debate on the issue scheduled at City Hall for the 18th of February. I resolved to attend with an open mind, and to determine my position on the issue after listening to each side’s arguments as objectively as possible.
When I arrived that evening, I felt like I might’ve been one of the only “undecideds” in attendance. IRV supporters sported blue stickers with “40%” crossed out, while anti-IRV partisans sported bright yellow paper fliers attached to a button with a rubber band. The debate had about forty people in the audience, and CCTV (the local government access station) and Seven Days (a local newspaper) were broadcasting it live. The questions addressed to the two teams were posed alternatively by representatives of local media outlets (Seven Days, The Burlington Free Press, WCAX), the audience at City Hall, and TV/Internet viewers sending questions in via email.
For the first part of the debate, things didn’t look up for the pro-IRV team of State Rep. Mark Larson and Keri Toksu of the League of Women Voters. They seemed to spend more time rehearsing trite slogans than they did explaining the value of IRV, and Larson in particular was almost painfully inarticulate. The anti-IRV side, by contrast, was lucid and forceful in their arguments, and by the time questions rolled around to the community, I was tentatively against IRV.
Then came an incredibly insightful question from the Internet. A large part of the anti-IRV side’s argument hinged on the assertion that IRV voters are more or less “voting in the dark”, whereas in a traditional runoff they have access to enough information to make an informed decision; the questioner essentially inquired into the nature of that information differential. If, on election day, I know that I like A more than B, and that I like B more than C, and rank them as such, how does that change if there ends up being a runoff between B and C? The anti-IRV team’s answer wasn’t satisfactory to me, and I spent the next few minutes wrapping my head around the implications of that question.
I wasn’t able to get in a question asking for more clarification during the debate, but afterwards I was able to talk to Kurt Wright and Sandy Baird, two IRV opponents. Mr. Wright told me that he was the candidate who received the most “bullet votes” in the last election; that is, votes in which the voters only marked one candidate, and the implication was that he considered those people to have been disenfranchised by the IRV system.
That finished the debate for me; essentially, the problem in the last election wasn’t with the system, it was with voters. All voters had the equal right to rank the candidates according to their preferences; I ranked them all, and thus exerted the maximum possible amount of agency allowed by the system. Anyone who didn’t do so, and instead voted for just their favorite candidate, was the equivalent of a voter who turned out for the election but failed to show up to the runoff. The instructions were clearly printed on the ballot; I read them and marked my ballot accordingly. Anyone who didn’t do so was making a choice and is responsible for results their actions. Disenfranchisement happens when an external force removes an individual’s agency; “bullet voters” voluntarily chose to not rank all the candidates and thus were not disenfranchised by the IRV system. Once I grasped this idea, the anti-IRV argument fell apart for me.
No system of elections is perfect; in fact, by the basic laws of the universe, no system can be perfect, period. For me, though, IRV’s elimination of the Spoiler Effect is a step in the right direction, and the repeal of IRV can thus only be justified by successfully arguing that the results generated by the system do not reflect the will of Burlington’s voters. As the bullet voting issue indicates, though, the “disenfranchisement” trumpeted by anti-IRV advocates was really the moral equivalent of failing to vote in the runoff, and the responsibility for the result thus lay with the bullet voters who voluntarily chose to not have a say in future rounds, not the system. I’m always open to rebuttals, but I’m pretty confident with my logic at this point. Unless I encounter a really good argument, then, I plan on slapping on a blue sticker and voting no on 5 this town meeting day.
One clarification:
I believe the 2006 election was the first and the 2009 was the second.
Kurt Wright has a serious blind spot and a gross inability to reflect upon the obvious. IRV is the ONLY reason a fringe entity like Kurt Wright was taken seriously enough to run for mayor.
Instant Runoff Voting is about choice.
There are those, such as the anti-IRV agitators, who will tell you choice is good as long as they get to give you the choices.
And there are those, such as the pro-IRV agitators, who will tell you choice is good, and you deserve to choose from ALL the political offerings.
I can see no downside to having real choices in elections.
Face it. These morons disenfranchised themselves. They were too stupid and/or too arrogant to mark their ballot more than once, like decent folks who give a rat’s ass about democracy do.
Typical GOP mindset – my way or the highway.
The Wright voters bullet vote counted all the way through. There are many that think that their vote did not count all the way through, or that others got to vote twice or more times.
The reality is that the bullet votes got counted multiple times, once in each round, just as others got to vote once in each round whether their first choice candidate was eliminated or not. Which, by the way, is the exact same thing that happens in a delayed run-off.
The other important clarification is that the spoiler issue is much more prevalent in a 40% threshold (former Burlington situation and the one we would return to if IRV is repealed) or a simple plurality scenario (President, defacto all statewide offices in Vermont).
With 50% as the threshold, spoiler is less of an issue as far as mechanics, because if two people split 48% of the vote…then neither one would have won even if all of the third place candidates voters supported the second place person.
That is why the 50% issue does matter. Kurt’s folks would rather focus on this as a referendum on Bob (which is an easy target…and somewhat justifiable target these days). But the major change is the 40% threshold as that allows a truly minority supported candidate to win.
As David observes, the “bullet” Wright voters weren’t even uncounted at all — their votes counted in every round, and propelled Wright over Montroll to the showdown with Kiss.
Now in theory, if Wright had come in third earlier on, those voters wouldn’t have gotten a say in the last round — and as you say that’s not disenfranchisement, it’s just not voting. A generous take could read their votes at that point as a “no-confidence,” none-of-the-above vote. (But it could also be poor education, laziness, or plain stupidity.)
But if we’re talking in theoretical terms, it’s also important to recognize that Montroll could conceivably have beaten either of the other two had he just had a little more first round support, or that Wright voters may have gotten their least favorite candidate elected (Kiss) by voting for their favorite rather than their second-favorite. So there are logical cases to be made against IRV — so it’s odd that the anti-IRV folks don’t seem to be making them. They don’t fit on a bumper sticker, I suppose.
So I would also vote to uphold IRV in this case, because the pros outweigh the cons here. But I would also switch in an instant to approval voting, if only people were advocating it… what’s not to like about simply voting for all the candidates you think would do well in a position? Why should we have to choose just one most of the time, the lesser of two evils? Is it so hard to understand a race being won with someone taking 60% of the vote, beating the candidates with only 55% and 48%? Wouldn’t it tell us a lot more about the actual mandate of a victor?
… in fact it was seriously flawed.
but returning to the old system is even more flawed. i’m voting “No” on 5.
BTW, IRV had it’s own spoiler, and it was Kurt Wright. Wright was a loser whose presence in the race changed who the winner is. that’s at least “spoiler-lite”.
and it didn’t eliminate the burden of strategic voting, but transferred that burden from the Majority (we didn’t have to make a painful choice between Bob Kiss and Andy Montroll like we would have in the old system) to the “GOP Prog-hater” minority who found out that by marking Wright as their first choice, they actually caused the election of Bob Kiss, their least preferred candidate. so if IRV survives to 2012, what are these folks gonna be thinking when they go to the polls and vote for mayor? could it be?: “In this town full of liberals, I gotta choose between Liberal and More Liberal, because when I vote for the guy I really like, More Liberal gets elected.”
IRV screwed up in 2009, but the screwup was not failing to elect Kurt Wright. in the IRV final round we know that Bob Kiss was preferred to Kurt Wright by 252 votes (4313 to 4061). that’s what our ballots said. 4313 voters marked Kiss higher up on their ballot than Wright while a fewer number of voters (4061) marked a preference of Wright over Kiss.
but what the IRV failed to consider is that Andy Montroll beat Wright by a greater 930 vote margin and even more anomalously, Montroll beat IRV-winner Kiss by 587 votes.
even though the GOP doesn’t want to admit it, Wright was the third most preferred candidate last year, Bob Kiss was second, and Andy Montroll was preferred over every other candidate when the voters were asked to choose between the two. that is what our ballots said.
there is a method named after Condorcet that would elect this candidate who beats all other candidates in a paired mano-a-mano faceoff. it would have elected the True Majority candidate (Montroll), it would not have suffered the “Spoiler-Lite” problem (where Wright’s presence in the race kept Montroll from beating Kiss), and it would not have transferred the burden of strategic voting from the majority to the minority as IRV did.
but like the 2-party system, this binary choice between IRV and 40% Plurality is like choosing between Dumb and Dumber. and IRV is the former, which is why i am voting No on 5.