IRV battle heats up in Burlington, Vermont

Instant Runoff Voting (IRV) is a voting reform measure endorsed by groups such as Fair Vote. IRV is often cited as helping third parties have a better chance to win. In Burlington, Vermont, the election for Mayor is done through IRV. Now, the community is having an argument about the pros and cons of the system. An anti-IRV group is gathering petitions, and the League of Women Voters is defending IRV to the public.

The following three pieces explore the dialogue surrounding IRV in Burlington:

Opponents of IRV gather and work on anti-IRV petition drive:

Burlington Free Press
IRV opponents to meet today
/ April 29, 2009

A pro-IRV opinion piece by Keri Toksu of Burlington. Toksu is the director of the Champlain Valley League of Women Voters:

Burlington Free Press
My Turn opinion piece
/ May 4, 2009

An interesting piece in the Examiner claims that IRV in Burlington will help ensure that the left – either the Progressives of the Democrats – will always win. And, that the candidate does not matter between those two parties. So, that IRV is unfair to the right and to the Republicans in Burlington:

Examiner.com
Instant runoff voting protects Progressive dynasty

42 thoughts on “IRV battle heats up in Burlington, Vermont

  1. profstampede

    IRV is all well and good to help mitigate the spoiler effect and to reduce the cost of runoff elections, but if Australia is any indication, it doesn’t really help third parties much in the long run. It’s a good thing to implement in mayoral races like they’re doing in Vermont. But proportional representation would do much more for ensuring fair representation of third parties in legislative races.

  2. Fair Voter

    Ranked-choice electoral systems like Instant Runoff Voting help educate voters about multi-party friendly electoral systems like proportional representation using the Single Transferable Vote method (although proportional representation doesn’t require rankings, American voters may have a philosophical problem with “party list” PR given the history of voting for actual candidates).

  3. Erik Geib

    My problem with proportional representation is party lists.

    I’d rather see IRV with individual candidates in individual races (assuming, of course, the many other electoral issues are tackled: ballot access, gerrymandering, debate access, etc.).

  4. Dale Sheldon

    To Profstampede: “mitigating” the spoiler effect is useless (and meaningless). As long as there are spoilers, third parties will suffer from a huge systemic disadvantage. It doesn’t matter whether that disadvantage kicks in starting at 0% (as it does under plurality) or at 25% (as it does under IRV). When it kicks, voters will remember, and desert the party in droves.

    There ARE simple voting methods that DO completely eliminate spoilers: both approval voting and score (AKA range) voting.

    http://rangevoting.org

    Latching onto IRV is a suicide-trap for third parties. If you want to have an actual chance to win elections, you want score voting.

  5. Dale Sheldon

    To Erik Geib: There are proportional representation (PR) methods that don’t involve party list. The PR method on which IRV was based, Single Transferable Vote (STV), is such a method; and STV is much more effective at its stated goals than its bastard step-daughter IRV is.

    (There is also a range-voting-like PR method called Re-weighted Range Voting (RRV), but not much research has been done on it. My gut instinct is that RRV is better than STV, but not as glaringly so as RV is better than IRV.)

  6. profstampede

    Dale Sheldon: I agree for the most part, and think that we see the problems with IRV where it has been implemented. Range voting sounds better but I worry about how to balance candidates who hardly anybody knows enough about to make an informed decision on. There are probably good metrics for this that you know about.

    Fair Voter and Erik Geib: I wonder where the resistance to party lists comes from. You are still voting on people (groups of people, sure) and the result is significant representation for third parties (see the Netherlands or Finland, or if you’d rather have a mixed system, Germany or Scotland).

  7. Dale Sheldon

    To Fair Voter: The “educational” angle is a red herring. No constituency has ever gone from instituting IRV to instituting PR: historically, it has only ever proceeded in the opposite direction (or both have been enacted simultaneously), with PR (via STV) arriving first, and later STV being extended to single-winner elections as IRV (which makes sense since, algorithmically, IRV is identical to STV with only one winner.)

    I don’t want to use a crappy single-winner voting system in the (baseless) hope that someday I’ll see proportional representation; I want a good single-winner voting system.

    If you want to enact STV, go right ahead; I’ll help you get it done. If you want IRV though, expect a fight from me 😉

  8. Dale Sheldon

    Profstampede: Rangevoting makes an allowance for lack of knowledge. Voters are permitted to vote “No Opinion” on candidates, essentially saying “I don’t know enough to make an informed decision, and so leave the choice to those who are better informed.” In practice, most people will give a score of zero to a candidate they know nothing about; so getting your name and your views publicized is still necessary if you actually want to win.

    The hope is that, since under RV all candidates (not just two of them) have a real chance at winning, that the press will cover all candidates (not just two of them).

    To prove whether this is the case though, we’ll need to get RV used in some modern elections for a number of election cycles. That’s the focus of my current efforts.

  9. Morgan Brykein

    I prefer actual runoff voting, where if nobody gets a certain percentage, the top two candidates go to a second round, and there’s actual voting both times.

  10. Erik Geib

    profstampede:
    My resistance to party lists is two-fold. Part of it is my belief in individualism, and part of it relies upon my hesitation towards this being an “improvement.” Party lists make parties even more of a bureaucratic nightmare than they already are, and have too much potential for favoritism.

    Dale: I fear that STV would confuse people too much to ever stand a chance of approval. Personally I don’t think it’s confusing, but I can already hear the populist chorus denouncing it as “confusing to the ill-educated,” while pundits and agitators of the status quo twist it in every manner possible. I tend to favor IRV because I think it stands a better chance of approval within this populist nightmare of an electoral system. I also think range voting would get the ‘confusing’ rap from resistors as well. Too many special interests are invested in the current system to ever allow something like range voting to pass.

  11. Dale Sheldon

    Erik Geib: I’ve heard that argument before, but I have trouble wrapping my head around it. From the only marginally objective measure (“how many lines of computer code does it take to write it”) range voting is simpler than IRV.

    Subjectively: “Give each candidate a score from 0 to 9; highest average score wins.” I challenge anyone to give an overview of IRV in fewer words.

    I do agree with you about special interest though: IRV would continue the two-party system and currently over 99% of elected officials are members of those two parties. It’s going to take significant popular support to get the beneficiaries of a two-party system to put an end to it. And I suspect that those elected officials will gladly accept the help of all the pro-IRV individuals who will allow them to keep their seats while they give the appearance of effecting meaningful change. It will be an uphill battle.

  12. Erik Geib

    Dale,

    Thank you for responding cordially (at times a rare feat on this site- lol).

    The confusion to which I speak is exactly how that computer code determines the winner. I think people would hesitate to support something if they didn’t understand the formula used by the computer to determine the winner. Yes, they’ll understand their personal ballot, but they may have trouble determining the winner. Given this, it makes it too easy for the status quo supporters to deny such a systematic change by merely playing up the “don’t trust what you don’t understand” card. Sadly, these same people constantly encourage the public *to* trust things they don’t understand (see: defense budget). Of course, this likely isn’t news to you.

    I also personally don’t like it because I think it invites too much strategic voting (perhaps more so than other systems, but this is obviously debatable).

    I don’t think there’s such thing as a perfect system, but for now I lean towards IRV. I’m more than open to being convinced otherwise (and look forward to your response) and would certainly take anything over this first-person-past-the-post plurality non-sense we currently have.

  13. Dale Sheldon

    Thank you, and it’s been a pleasure discussing with you as well.

    The most convincing evidence for me was professor Warren D. Smith’s study examining Bayesian regret, as I first encountered it in William Poundstone’s 2008 book, “Gaming the Vote”.

    http://rangevoting.org/BayRegsFig.html

    That graphic illuminates several points: one, honest votes are always better (for society, not themselves) than strategic votes. And two, even if you allow IRV voters to be 100% honest, they still may not do as well as 100% strategic range voters; the results are that conclusive. (That 100% strategic IRV is identical to 100% strategic plurality is another interesting result pointing to IRV’s inability to effect change.)

    If this peeks your interest, I recommend getting a copy of “Gaming the Vote”, it goes into more detail than I possibly could in these comments; or browsing the pages at rangevoting.org; or checking out my blog, The Least of All Evils, at leastevil.blogspot.com 🙂

  14. Ross Levin

    Wouldn’t a 100% strategic vote be identical for all voting systems? With range voting, you could just give someone a 10 out of 10 and everyone else a 0 out of 10 or not vote for them.

  15. Dale Sheldon

    To Ross Levin: You could, but that wouldn’t be your best strategy under range. And no, the best strategy is not identical under all systems.

    Check out rangevoting.org and my blog leastevil.blogspot.com; both have some entries about strategy and honesty (and based on your comment, I plan to soon write another one) but it needs more detail than I feel can be covered here.

  16. Michael Cavlan

    IRV will be implimented in Minneapolis this election and will be on the ballot in St Paul.

    I agree that it is far from perfect but it is a start.

    Eventually Proportional Representation will be here. Until then, IRV is a good starting point.

  17. Green Ferret

    “Latching onto IRV is a suicide-trap for third parties.”

    I’m not convinced of this. The Progressive party just won the mayoral race in Burlington under IRV, to use just one example.

    The problem with approval and range voting is that people will still vote strategically. In actual IRV elections, voters don’t vote strategically – or they want to, but realize that it’s no use trying. IRV works as well in practice as it does in theory, which is sadly not true for approval.

    In college I took a class on the mathematics of voting systems. Of course, we covered Arrow’s theorem, which tells us that there can be no perfect voting system (yarr!). At the end of the course, we held a debate between approval voting, instant runoff voting, and plurality voting (full disclosure: I was one of the IRV advocates). After considering the merits, our class of some 40 students voted for voting systems, using 5 different systems (approval, IRV, plurality, borda count (a score-based range system), and one other I can’t recall. Instant runoff won every vote.

    Of course, that doesn’t prove the merits of any voting system over any other. What that shows is that a group of people educated in voting systems will choose IRV over approval. Combined with referendum results that show voters choosing IRV over plurality, it’s clear to me that IRV is the voting reform with the best chance of widespread adoption.

  18. Gary

    Proportional representation is the ONLY way to go. Having 4 to 8 parties in a body will break the back of one-party dictatorship in any legislative body.

  19. Nate

    I think IRV is simpler than range voting. Not in terms of how to explain the two, not in terms of how to figure out who won, but in the more important aspect of how to vote.

    It’s incredibly easy to choose which candidate suits you best, then the next best candidate, and so on. What is not so easy is deciding whether you like X only 7/9 as much as Y or 8/9.

    The only positive of range voting is that one can rank two candidates as equal, where in IRV one has to rate one higher than the other. (Or exclude both, but that only works if both are equally the worst choice, no other positioning.)

  20. Nate

    Proportional representation doesn’t work for single seats like mayor. I also think 8 parties is probably too many. 4 or 5 seems like a happy medium.

  21. Melty

    The counting is the more important aspect.

    If Range seems hard for in the voting booth, just boil it down to fewer breakdowns and put it in words, not numbers, like great / good / tolerable / bad. Or else just go for Approval voting… can’t get it any simpler than that.

  22. Melty

    Proportional Representation is spiffy for filling boards of appointees (example: the proportion of the vote your party’s candidate for governor got, that’s the proportion of seats your party gets to fill on the Election Board). I’m not in favor of PR, though, for electing anybody.

  23. DebbieKat

    I personally like approval voting, where you can indicate approval for several candidates. This way, your least-liked candidate doesn’t win just because the more popular candidates have split votes. If 60% of the population is left-leaning and 40% is right-leaning, but the left-leaning votes are split between two different left-leaning candidates, you’re left with the right-leaning candidate winning by 40% and the other two candidates with 30% of the votes each. Approval voting would eliminate this (and the “spoiler” effect).

    My recommendation is to eliminate primaries altogether and use approval voting in the general election instead. Let everyone run, regardless of party. It would be much less expensive in the long-run.

    Hi Ross! 🙂

  24. Dale Sheldon

    Green Ferret: The Vermont Progressive Party has been able to win elections under plurality, too; IRV has nothing to do with their success. Unless you’d like to argue that plurality is /also/ a fine system for third parties, because it elected VPP candidate Bernie Sanders as mayor several times.

    No, the simple fact is, and this has been shown through theorem (covered extensively at rangevoting.org) and through real-world data (Australia for instance), is that IRV, just like plurality, leads to a two-party dominated system. Therefore, third parties should not support it.

  25. Dale Sheldon

    Green Ferret: Oh, also, Arrow’s Theorem is only about rank-order systems. Since range isn’t a rank-order system, Arrow’s Theorem doesn’t apply to it.

    Also, you offer no support for your theory that “IRV voters won’t be strategic, but range voters will”. It’s actually very easy to vote strategically in IRV; and that strategy leads directly to two party domination (since it’s the same strategy as plurality.)

    Nor do you offer support for your theory that approval/range fails in practice. In truth, it’s been successful for many professional organizations, in selecting the UN’s secretary general, and historically in the two longest-lasting democracies on record (Sparta and Venice); hardly a failure.

  26. Ross Levin

    Dale – I think the Progressives have most of their elected officials in Burlington, which uses IRV. So at least some of their success could be attributed to that, but I’m not sure.

  27. Erik Geib

    DebbieKat,

    Your criticisms are more a reflection of plurality voting. Most elecotral reforms I’ve ever heard advocated do away with plurality voting as part of the package.

  28. DebbieKat

    Ross – Things are good! 🙂

    Erik – My first paragraph definitely is, but I have heard some criticism about IRV with respect to third party candidates and that, in the long run, the two major parties still have an advantage.

    Here’s a link with some criticisms of IRV:

    http://www.approvalvoting.org/irv.html

  29. Erik Geib

    DebbieKat,

    I don’t doubt the possibilities of IRV having flaws- I think they all have some flaw or another. As I said to Dale, I favor IRV but am certainly open to other proposals. I do doubt IRV could replicate the problem of plurality, however.

  30. Joyce McCloy

    The League of Women Voters endorsement is meaningless. This is the same organization that nationally endorsed paperless computerized voting. It takes them years to figure out that they’ve pushed ideas that are poisonous to democracy. And admit it – NEVER!

    It took a huge effort by members of LWV to force the organization to back away from its dangerous endorsement of paperless computerized voting.

    IRV is much more insidious than paperless voting, because with IRV, even with paper ballots, transparency is greatly diminished. Heck, voters have trouble understanding how the tallies provide the results that they do.

    The IRV groups in North Carolina are waging war on election transparency – by publicly endorsing the central counting of ballots cast at polling places – making it easier to stuff the ballot boxes, or perhaps fill in a few choices on some of those IRV ballots. Change just a few per polling place and voila – you have the results you want.

    Another nasty thing IRV proponents are pushing is the use of uncertified software. Pierce County Washington got suckered into this, and their precinct optical scanners, with IRV software, can’t be used to count – IRV because they can’t handle the ballot images.

    Scotland was talked into abandoning hand counted paper ballots for STV in 2007, and ended up with machines that didn’t do the job, 100,000 spoiled ballots party due to the machines, mostly due to voter confusion.

    IRV is the trojan horse for election fraud and is insidious in nature – it sounds so good, so well intended, it inspires the emotions of people who can’t put their support behind one candidate – and in an academic setting, its fine. But for real life elections, IRV damages the integrity of the elections and it also weakens political debate.

    see http://www.instantrunoffvoting.us for more

  31. Rob Richie

    Joyce McCloy is wrong about Scotland and most everything else in her post. A few corrections:

    1. The spoiled ballots she cites were in the non-ranked choice voting elections for the Scottish parliament — due to a bad ballot design along the lines of the butterfly ballot. Scotland was quite pleased with the results from its STV elections and has expanding STV for use in more elections.

    2. Ranked systems can be done without sacrificing election integrity. Many election integrity leaders in the US have been working hard for IRV and STV. Here’s a good statement from the leader in Ireland’s fight against toucshcreen machines:
    http://www.fairvote.org/?page=2536

    3. McCloy fought against getting optical scan machines that could do full IRV tallies at the precinct in NC and now remarkably complains that… the machines can’t do full IRV tallies at the precinct. But North Carolina’s board of elections has developed a very sensible proposal for handling IRV counts that does not allow “ballot stuffing.” McCloy knows this, of course, but won’t admit it. The total number of first, second and third rankings would be tallied at the precinct and then compared at the central counting place where the IRV tally would be done — any changes would be noted. (And of course one needs to keep ballots secure for possible recounts anyway.)

    4. It’s funny to have her suggest that IRV “weakens debate” when it gives third party and independent candidates an unfettered chance to put their best case forward. The Australian Greens ran candidates in every seat in 2007 and the MEDIAN number of candidates in races was seven. True, it’s in the PR-STV senate seats that they’re winning ,but Aussie Greens like IRV within the limitations of a winner-take-all system.

    5. The League of Women Voters states groups that have been endorsing IRV after full-blown studies are largely the ones that led the fight for paper trails against national League policy. Joyce may want to cast aspersions against all state League activists, but that’s foolhardy and unreflective of the reality of the grassroots democratic nature of much state League activity.

    McCloy has a simple tactic. Find URLs that seem to support her point of view, ignore all other facts and then throw lots of mud, hoping some of it will stick. A good example is her site’s recent post on Aspen, where:

    * There was 100% valid ballots in the mayor’s race.
    * The city had its highest turnout ever.
    * Two of three incumbents were ousted, as reflective of the voter’s mood.
    * Big money lost out again, as is becoming a pattern in IRV races, with the biggest spenders in the mayor’s race and council race losing.
    * A post-election audit showed the counting system worked well and the count was done very transparently

    Sadly, McCloy does all this against IRV and proportional voting (which she lambasts as unAmerican and awful, by the way) with no real alternative. She gives token support to alternative voting methods, but is doing nothing to help them in NC as far as I can tell. She has enough friends in the legislature that she could get a bill introduced for something like approval voting — but I’ll believe it when I see it.

  32. icr

    All states but one somehow feel the need for bicameral legislatures. There is no good reason why one house couldn’t be elected by statewide party-list PR. More or less local replications-with maybe a higher threshold of say 5%-of the parliaments of the Netherlands and Israel.

    Nothing could do more to quickly break up the “two party” duopoly.

    PR is the norm outside the US, Canada and the UK. I tend to think that that most of the resistance to it comes from cultural prejudice.

    The evidence seems to be in favor of range voting when it comes to elections for single offices.

    And remember that some states have initiative and referendum. This allows the electorate-barring excessive procedural hurdles-to bypass D and R party hacks. A small state would probably be a more opportune target than a larger one.

  33. Rob Richie

    Several folks have mentioned approval voting and range voting in posts above, but note that they’re largely untested in competitive elections. One of the few examples was the Dartmouth Alumni Association elections for seats on its board of trustees. Today it was announced that 82% of alumni voted to amend the association’s constitution to replace approval voting with a traditional runoff system.

    The system was rejected amidst evidence that approval voting was being gamed. Backers of petition candidates apparently were more likely to “bullet vote” for their preferred candidate and defeat candidates nominated by the board, whose backers were less tactical.

    Approval voting also was abandoned in one of its few other high-profile uses, for elections for the leadership of the the Institute of Electrical and Electronics Engineers.

    Meanwhile, instant runoff voting is used for hundreds of private association and college elections, including many hotly contested ones. The Liberal Party of Canada has also just announced that it will use instant runoff voting when it moves to a national primary to select its leader. See a listing of organizations in the US using IRV here

    http://www.fairvote.org/?page=1964

    The 51 American colleges and universities with it are listed here:
    http://www.fairvote.org/?page=61

  34. Melty

    Voting for just one when you could opt for more, “bullet voting”, can easily be an honest indication of voter preference. Instant Runoff offers more opportunity for tactical voting than Approval Voting does.

    Instant Runoff is very complex in the counting, and needlessly so. I suggest reading “Reality Mars Instant Runoff” by Kathy Dopp on the link @34 bottom.

  35. Melty

    I’m just defining the term “bullet vote” which was mentioned in the entry just before my last one. I’ve not said anything about what voting systems allow or do not allow bullet voting.

    Actually, not every voting system allows for it. Last I checked, some parts of Australia use a version of instant runoff that discards ballots that have not ranked all candidates on the ballot. This garantees a majority outcome. Versions of instant runoff that allow you to fill out the ballot partially do not insure a majority, but it’s better to allow partial voting and bullet voting, than to disallow it. Also, under Condorcet you may be expected to do all pairwise comparisons, though this is not done in order to insure any majority outcome.

  36. Ross Levin

    Discarding ballots like that brings on a whole new set of problems, although I guess it might be different in Australia because they have mandatory voting, I think.

  37. Melty

    My main point here is that elimination & reallocation systems such as Instant Runoff are no friend to budding parties.

Leave a Reply

Your email address will not be published. Required fields are marked *