The Charter Commission of Portland, Maine, has arranged to ask city voters if they wish to amend the charter, and elect a Mayor with a citywide vote using Instant Runoff Voting. The popular vote will be on November 2, 2010. See this story.
Currently, Portland voters do not elect a Mayor. Instead they merely elect a city council, and the council chooses one of its own members to act as Mayor. Thanks to Thomas MacMillan for the link.
Pingback: Ballot Access News » Blog Archive » Portland, Maine Voters to Vote … | Maine Guide
Portland officials were misled into believing that IRV provides a majority win.
San Francisco had to amend their charter to reflect the IRV form of “majority”, which is the majority of votes left after others are sorted, allocated, eliminated, reallocated and ballots are exhausted.
Many IRV elections award winners with less than a majority of votes cast.
http://www.instantrunoffvoting.us/majority.html
Joyce McCloy made this argument and many others to the commission. Her concerns were evaluated and rejected by a special committee that held several meetings. They weren’t “misled” — rather, looking at the same facts they came to a different conclusion, one that happens to be the same conclusion as made by policymakers around the world who have chosen to go to IRV.
You can’t force a majority of people to like a candidate in every election, so if you allow abstentions, you have to accept that no candidate may get a majority of registered voters. But runoff elections are commonly said to produce majority winners even when turnout sharply declines from the first round. And a candidate who wins more than 50% is said to have a majority even if most registered voters abstained from participating. By that logic, instant runoff elections also produce majority winners — meaning a “majority of the willing.” See a blogpost on this point here:
http://irvfactcheck.blogspot.com/2010/06/rebutting-majority-failure-argument.html
If Portland, Maine wants a strong mayor, then a conventional runoff is better. The primary reason that Burlington, Vermont eliminated IRV was because, while Bob Kiss was universally regarded as a nice guy, he turned out to be an ineffectual leader when significant issues arose. Though it possibly sounds like Portland wants a sort of mush, with the city manager presenting the budget, and the mayor commenting on it. But it doesn’t make much sense for the mayor to be paid an executive salary in that case.
A runoff for a contested mayoral election will produce similar turnout to the general election, and will let voters closely examine the two finalists. It is similar to if you were hiring an executive for a company, and called the two finalists back for a 2nd more extensive interview.
IRV = THE method for the election of Hitler and Stalin clones when the Middle is divided.
34 H–M–S
33 S–M–H
16 M–H–S
16 M–S–H
99
Is the U.S.A. more than a bit divided ??? Duh.
App.V. — pending major education about head to head Condorcet math.
Pingback: Portland, Maine Voters to Vote on Mayoral Election Using Instant Runoff Voting – Ballot Access | Portland, ME Blog
As a Portland resident, I am pleased with the introduction of IRV to our municipal voting process. I wish we had included it for all city council members. I wish we had included a stronger elected Mayor, but the increased powers are an improvement of what we had before.
I find Rob’s argument on majority winner persuasive. But I only find it persuasive referring to those who, for whatever reason, don’t rank their entire ballot (good analogy to runoffs). The major flaw with IRV is that even though a voter is expressive, this doesn’t mean that the expression itself is taken in during the tally. A voter’s expression is only tallied when votes transfer. Otherwise, those 2nd, 3rd, and other choices are never heard from. Despite the IRV tally ignoring them, they matter.
It’s largely this ignoring of lower rankings that causes IRV to suffer from such calamities as nonmonotonicity and a better-to-stay-at-home paradox. Indeed, vote splitting may continue to occur under IRV, unfortunately. It’s not to say IRV will always mess things up, but it botches the outcome enough that it makes the cost of its complexity not worthwhile by far.
When we look at computer simulations, we see IRV doing poorly and goofing up. We can even look at the real world. Australia uses IRV for their lower house (get that STV out of your head from the upper house). In the past some hundred years that IRV (preferential voting there) has been implemented in Australia, there’s no alternate party presence in the lower house (the multiple parties there are a split of the same group that do not compete against each other). And there are voter cards for strategic voting.
But what’s our goal here? Is it to prevent vote splitting? Then look to systems in which the entire ballot is automatically counted and votes for candidates can be the same. Two systems do this–approval and range. And that’s why they avoid vote splitting. They’re also very easy to count and understand. They even work on our current voting machines (insecure as they are). Even more, there’s an interpretable expression of support for all candidates in either method.
Why then is it being insisted that we use an inferior and more complicated system–IRV? I’ve heard the expression that you don’t switch horses midstream. But you do get off the horse when that horse is clearly drowning. We have the same goals here. It’s time to switch horses. You have two great ones to choose from–and their names are approval and range voting. (range being the technically superior of the two, approval being simpler)
Longer term — Condorcet head to head math using Number Votes (1, 2, etc.) for ALL offices — and YES/NO (tiebreaker)
ALL combinations of —
Test Winner(s) versus Test Loser — with all others deemed Losers — transfer votes to the highest ranked Test Winner(s) or the Test Loser.
Legislative bodies and 1 or more executive / judicial officers.
Each Test Winner will / will NOT win in ALL of his/her combinations. YES votes for tiebreakers.
Will need computers in any large election.
Legislative body winners to have a voting power equal to the final votes each receives.
@Rob Richie,
IRV only ensures that the winner is preferred by a majority to at least one other candidate. That’s not saying much.
Here’s an example IRV election where huge majorities are overruled.
% of voters, their vote
35% A > C > D > B
17% B > C > D > A
32% C > D > B > A
16% D > B > C > A
Instant Runoff Voting selects candidate B as the winner, beating A in the final round, 65% to 35%. But a huge 67% majority of voters would rather have candidate C than candidate B. And candidate C received nearly twice as many first-place votes as candidate B, 32% to 17%. And an even larger 83% super-majority of voters would rather have candidate D than B (and D got just a little fewer first-place votes than B).
Ordinary delayed “top-two runoffs” (aka TTR) have very different behavior than IRV. For instance Matt Gonzalez, Green Party member and former president of the San Francisco board of supervisors, faced Gavin Newsom in a runoff for mayor in 2003. In the first round, Gonzalez received 19.6% of the total vote to Newsom’s 41.9%. Newsom outspent Gonzales $4.4–4.9 million to $800,000–900,000. Yet Newsom won the runoff by a mere 5.6% margin. I believe that has a lot to do with the time leading up to the runoff. Voters had time to consider and get to know just those two candidates. The media had every reason to cover Gonzalez, even if they thought hey had no chance to win, because he was the only alternative to Newsom at that point. Thus the effect of Newsom’s cash and major party backing was significantly diminished.
How are things going now that San Francisco has had IRV for a few years? Well, consider that SF supervisor Ross Mirkarimi, who co-founded California’s Green Party in 1990, and coordinated Ralph Nader’s 2000 presidential campaign in California, switched from the Green Party to the Democrat Party in March of this year.
Relatign to Matt Gonzalez in San Francisco, he’s a strong backer of instant runoff voting (and proportional representation where that can be won). So he obviously thinks differently than Clay Shentrup does.
Approval voting and range voting are better in theory than in practice, however limited that practice is. Its advocates dismiss the “later no harm” criterion, but when indicating support for a lesser choice counts aginst your favorite choice and your vote counts for more than one candidate at a time, you have problems that all the passionate theoretical talk can’t wish away.
#11 The question is why he thinks that way since it irrational. The same thing happened in Minneapolis where a Green city council candidate won in the general election, after the other candidate actually received a majority in the (Top 2) primary.