Citizen Blogs are a public space for news and commentary from the community. The views expressed here are those of the individual bloggers, not The Bay Citizen. We aim for this to become a robust forum for discussion of a wide range of issues. If you are interested in becoming a Citizen Blogger on the site, please send an email to info@baycitizen.org and let us know who you are and what you'd like to write about. We look forward to working with you.
San Francisco is in the final days of hotly contested elections for mayor, district attorney and sheriff. These offices will be elected using ranked choice voting (RCV), also known as instant runoff voting, which allows voters to rank a first, second and third choice. San Francisco has used RCV since 2004 in three dozen races to elect the mayor, Board of Supervisors and other citywide offices.
So far, RCV’s accomplishments have been inspiring for those who believe that American democracy needs reform. San Francisco's 11-member Board of Supervisors has become far more diverse, with the number of racial minority supervisors doubling to eight, including four Asian-Americans. The gay community is well-represented, as are progressives, moderates ...even a San Francisco conservative.
Using RCV, San Francisco has avoided fifteen separate runoff elections held in December (which was the method previously used). Those December elections usually were marred by very low voter turnout – as low as 12% of eligible voters for the city attorney runoff in 2001 -- and a quadrupling of Citizens United-type independent expenditures from special interests (according to a study by the San Francisco Ethics Commission). They degenerated into mudslinging slugfests between the two final candidates in which voters usually learned nothing new and everything bad about the finalists. They also were expensive for taxpayers, costing millions of dollars to administer. San Francisco will save $3 million this year alone by avoiding a December runoff.
But under RCV, old-fashioned door-to-door politics and coalition-building have given grassroots candidates a better chance against big money. Voters aren't stuck anymore with a single shot vote for the lesser of two evils, instead they are liberated to rank their three favorite candidates. With RCV, voter choice is king.
Despite its impressive track record, RCV has its critics. Some critics argue that this year RCV has encouraged too many mayoral candidates -- sixteen -- which has made it harder for voters to tell them apart. But in 1999 and 2003 -- before RCV was used -- the mayoral elections drew 18 and 9 candidates respectively; in Board of Supervisors races, some had a dozen or more candidates, with one race having 15. The same complaint was heard before RCV, i.e. hard to differentiate the candidates.
Some critics say that returning to December runoff elections would allow voters to have a "second look" at the top two finishers in a crowded field. But when San Francisco used that system, voter turnout usually plummeted in the second election. In ten of the city's 14 December runoffs between 2000 and 2003, voter turnout declined by more than a third, with most runoff winners having fewer votes than the first-place candidate had in November. Clearly, most voters did not take a "second look" at the candidates
But under RCV, candidates are winning with far more votes than they would have received in a low turnout December runoff or June primary. When Supervisor Sean Elsbernd won his District 7 race with RCV in 2004, he had nearly 50% more votes than his predecessor elected in a December runoff. In Oakland, Jean Quan won more votes in her mayoral election in November 2010 than any other candidate for mayor in a generation, with a 43% increase in turnout over the 2006 June primary election that elected Ron Dellums as mayor. That's been true in virtually every RCV race, and it’s good for democracy.
Let's imagine what this year's race for mayor would be like if San Francisco were using a separate runoff election in December. There would still be a large field of candidates, and according to the latest polls all of them except for front runner Ed Lee would be bunched together with less than 8% support. So it STILL would have been a challenge for voters to discern one candidate from another. But even worse, as candidates vied to face off against Lee, those candidates with the most in common would be trashing each other in order to beat all the others and finish second.
The five Asian mayoral candidates would be knocking each other in order to be the sole candidate winning the Asian vote, instead of appealing for second or third rankings from voters beyond their base, as they are doing now with RCV. The same for progressive candidates Avalos, Adachi and Ting, and moderates like Herrera, Alioto-Pier, Hall and Rees. All of these candidates and their consultants would be engaging in complex strategies and targeted mudslinging to figure out how to knock off their opponents one by one, as if in a shooting gallery. Independent expenditures would have soared.
And then, having attacked each other in the first round, the surviving candidate would face the challenge of unifying the supporters of the candidates they had just finished attacking. And quickly raise a lot of money to defend against the expected fourfold increase in independent expenditures. Good luck.
Those who are pining for the "good ol’ days" of December runoff elections don't remember what those elections were actually like. They were brutal and expensive. That's why voters decided in 2002 to switch to ranked choice voting.
Some RCV critics have proposed that instead of going back to December runoffs that San Francisco should move to a June primary followed by a November runoff election. But that method was used by Oakland for many years and suffers from similar problems as December runoffs. That's why Oaklanders also voted to change their elections to RCV.
That doesn't mean that San Francisco's RCV elections can't be improved, and the Board of Supervisors should hold hearings about ways to do that. Although voters are handling RCV well (about 99.7% of voters cast valid ballots in most races, and about two-thirds of those voting in competitive races use all three of their rankings), voter education efforts could include more information on how the ballots are counted and not just the mechanics of how to rank candidates.
For those desiring more than three rankings, San Francisco's voting machines could be modified. RCV elections this year in St. Paul (MN) will allow six rankings (using equipment similar to San Francisco's), while Portland (ME) is allowing 15 rankings. San Francisco also could tweak public financing rules to ensure it is fulfilling its worthy goals. Debate organizers could begin limiting the number of participating candidates to no more than the six front runners as Election Day draws closer.
San Francisco has rightly been recognized as a national leader with RCV, with more cities using it every year. Next year, we'll certainly wish we had RCV for presidential elections if more than two candidates run, to prevent another Gore-Nader-type split. The freedom to rank your three favorite candidates is a blessing that we should treasure and make work. Mend it, don't end it.
Steven Hill (www.Steven-Hill.com) is the former director of the political reform program at the New America Foundation and author of "10 Steps to Repair American Democracy" (www.10Steps.net). He is the architect of the ranked choice system in San Francisco and Oakland
David Cary
Ranked Choice Voting has been good for democracy in San Francisco, improving voter participation and the ability for voters to express their preferences.
It is no surprise that California voters in a recent scientific poll by California Forward, said by a 2-to-1 margin that they'd like to follow San Francisco's lead and use RCV in their elections.
The most powerful critics of RCV are those who dislike it because of its success, because it has strengthened democracy, because it has shifted decision making power to voters. They've been working overtime spreading disinformation about RCV, trying to confuse voters about RCV.
It is nice to see the real RCV story being told here.
Chris Jerdonek
I am a San Francisco resident and will be a polling place inspector in a San Francisco election for the tenth time in next Tuesday's municipal election.
I strongly support ranked choice voting (RCV). With all the improvements that the US needs in its elections, it would be a real shame if some well-funded organization were to try to reverse RCV. That would be a huge waste of time and a large step backwards during a time in which we should instead be focused on advancing our country's democracy forward.
John Palmer
I strongly support RCV. I see the "confusion" argument that is largely present in most negatively-slanted articles as a total red herring. Other than picking them, how hard is it, really, to vote your first, second, and third choice?? The political consultants in this town hate RCV because it takes away their jobs. If it's the right method for the electorate, who cares!? There may be ways we can improve this system (more rankings when there are many candidates, ballot format, etc.), and we should move forward to do this, rather than move backward to the old system.
John E. Palmer
Clay Shentrup
John Palmer,
Since getting IRV, SF voters have made SEVEN TIMES as many ballot-invalidating spoilage mistakes. So clearly something is more confusing to them.
http://ScoreVoting.net/SPRates.html
Further there's confusion about how to cast the most effective ballot. A lot of IRV proponents make the utterly false claim that a voter's best strategy is to sincerely rank the candidates. Nope.
http://www.electology.org/irv-plurality
I agree the old system was not good either. IRV and Top-Two Runoff ("TTR") are two of the worst voting systems in existence, according to Bayesian Regret calculations. The best system in terms of simplicity AND representativeness (low Bayesian Regret) is clearly Approval Voting.
http://www.electology.org/approval-voting-vs-irv
SF residents who don't want to see a repeal back to the old system would be well advised to promote this compromise.
Rob Richie
Clay -- Your about strategic voting is entirely theoretical. AFTER the election, one might go back and say "if we had done X or Y, it would have helped." But at this point in the contest, you are absolutely wrong that voters should do anything other than vote sincerely. The only reason to moderate that statement in San Francisco is tied to limiting to three rankings-- e.g, rank candidates sincerely, but try to include a candidate you think will be in the final round. If San Francisco was like the may places that don't limit rankings to three, even that caveat would not be necessary.
With approval voting, even your ally Jameson Quinn admits that it would elected the candidate in Burlington who would lost one-on-one to two other candidates -- the reason would have been tactical voting incentives part of approval voting.
As to voter errors, you are talking only about the modest changes in overvotes. For instance, in Oakland's mayoral race. 0.3% of voters (3 out of every 1000) made that kind of mistake. That's more than an X-type ballot, but pales beside the fact that RCV elected Jean Quan with more votes than any Oakland mayoral cancidate in a generation - and that one clear finding in San Francisoc is that many fewer votes skip the RCV races for Board of Supervisors then they did in the old runoff races.
Joanne McKray
RCV means that a larger number of citizens participate in choosing government leaders who they favor or find acceptable.
RCV means that taxpayers do not have to pay for two elections to choose one set of leaders.
RCV means that the top two candidates do not have to do double fund-raising, thus incuring double obligations from their contributors.
RCV is good for democracy.
Joanne McKray
Clay Shentrup
Joanne McKray,
> RCV means that a larger number of citizens participate in choosing government leaders who they favor or find acceptable.
Simply false. This participation increase is based on the notion that runoff elections tend to have lower turnout. But IRV has nothing to do with that. You can eliminate runoff elections WITHOUT adopting IRV. The alleged benefit of IRV is to still ensure you get a "majority winner" after you do that. But clearly IRV does not guarantee majority winners either.
http://www.electology.org/core-support
Moreover, the very premise that runoffs have lower participation is itself misleading. For instance, in the mayoral runoff between Gavin Newsom and Matt Gonzalez, turnout went UP by around 25% over the first round election. This is often the case with more significant races. If you look at the historical average turnout for a few decades before we got IRV, compared to the time since we got IRV, turnout has actually gone down a bit.
http://www.electology.org/irv-turnout-sf
Again, it would be nice if IRV proponents would demonstrate some basic understanding of the issue, instead of just repeating false pro-IRV talking points from groups like FairVote.
Rob Richie
RCV has nothing to do with the fact that runoffs generally had lower turnout? In Board of Supervisors races in particularly in San Francisco that went to a runoff, the pattern is __overwhelming__. Ten of 14 runoffs had turnout drop by more than a third, and in most of those runoffs, the winner in December had __fewer_ votes than the leading candidate in that race had earned in November.
Joanne McKray
Clay Shentrup,
You can eliminate separate runoff elections without adopting RCV, but you have to settle for a plurity winners. RCV elections do guarentee winners having a majority of those favored or acceptable to voters, whereas plurity elections obviously do not, resulting in minority rule.
Regarding voter participation, only once in 15 separate runoff elections between 2000 and 2005 did the number of voters not significantly decrease compared with the first round.
Joanne McKray
Clay Shentrup
Joanne McKray,
> RCV elections do guarentee winners having a majority of those favored or acceptable to voters
False. In the last IRV mayoral race in Burlington, Vermont, the winner was Progressive Bob Kiss. But a majority of voters ranked Democrat Andy Montroll ahead of kiss. (Incidentally, Burlington repealed IRV after that (their second) IRV election.)
http://ScoreVoting.net/Burlington.html
It's even possible for IRV to elect candidate X even though Y is preferred to X by a majority of voters AND Y got more first-place votes than X.
http://www.electology.org/core-support
> You can eliminate separate runoff elections without adopting RCV, but you have to settle for a plurity winners.
So then the argument you're making is that IRV is a way to get majority winners instead of plurality winners, not that IRV increases participation or decreases costs. You get the alleged cost and participations benefits by eliminating runoffs, which has nothing to do with IRV.
I explained all of this in my previous post, and even cited the same links. But as is typical with the vast majority of IRV proponents I have encountered over the years, you apparently did not read them, and so you just repeated the same false claims.
Clay Shentrup
David Cary,
Actually, turnout has gone DOWN since San Francisco got Instant Runoff Voting. Here's an analysis using the city's official turnout data.
http://www.electology.org/irv-turnout-sf
The alleged turnout increase (as well as cost savings) are based on eliminating runoff elections. But eliminating runoffs can be done WITHOUT adopting IRV. So IRV has nothing to do with these "savings". And IRV actually introduces a lot of new costs. AND a lot of runoff elections (the ones voters feel are most contentious and important) actually have MORE turnout, like the last mayoral runoff between Gavin Newsom and Matt Gonzalez.
You say IRV improves "the ability for voters to express their preferences", but that's a statements which demonstrates lack of understanding of election theory. The ability to express more information on your ballot is only one of three crucial factors governing the representativeness of election outcomes. The other two are:
1) The efficiency with which the votes are tabulated. For instance, there are lots of algorithms for looking at a stack of ranked ballots and determining the winner. For instance the Borda method assigns scores based on the rankings, and the candidate with the most points wins. Condorcet (condor-say) methods attempt to find the candidate who would beat all rivals head-to-head. Some of these algorithms are more optimal than others. IRV is one of the least optimal algorithms. It discards an enormous amount of the information. This is explained here by a Princeton math Ph.D. whose been studying this issue for over a decade.
http://scorevoting.net/IgnoreExec.html
2) The inaccuracy of the information expressed by the voter, due to strategic behavior (and to some extent by voter confusion disparities due to the fact that some ballot formats are easier to understand than others). IRV is extremely susceptible to tactical behavior, particularly a problem called the Naive Exaggeration Strategy, which especially affects ranked voting methods.
http://scorevoting.net/NESD.html
It is only after combining all of these effects simultaneously that we can see the net performance of a voting system. And the only way to do that is to employ computerized election simulations to calculate "Bayesian Regret", a measure of avoidable voter dissatisfaction. The lower the Bayesian Regret, the better the system. And it turns out that Top-Two Runoff and Instant Runoff Voting both have comparable Bayesian Regret, as you can see in the second chart at this page.
http://scorevoting.net/StratHonMix.html
As for other voters wanting to "follow our lead", I believe that's largely due to false and misleading IRV propaganda popularized by groups like FairVote. For instance, myths such as "IRV prevents spoilers" or "IRV elects majority winners" or "IRV saves money" dominate the dialog. Whenever I have asked IRV proponents that I've met in real life to justify their support for IRV, I have ALWAYS heard one or more of these false beliefs. Once voters adopt IRV and see the reality, that's when it tends to be repealed, as it was in Cary, North Carolina; Burlington, Vermont; Pierce County, Washington; and Aspen, CO.
Proponents of election reform who do their homework tend to support better simpler systems like Approval Voting.
http://www.electology.org/approval-voting-vs-irv
Rob Richie
Clay writesy "IRV is extremely susceptible to tactical behavior" without a single example of it. Check out the _thousands__ of RCV races and find examples. Start with Bay Area races and last week's presidential election in Ireland.
Meanwhile, approval voting creates obvious tactical incentives - if your first choice has a chance to win, you will want to only vote for that person if you understand the system unless you feel equally with other candidates or are mostly going out there to oppose someoone you don't like. See the problem explained at www.approvalvoting.blogspot.com
You talk as if approval voting as if it never gets repealed (and exaggerate what's happened with RCV in that respect._) Despite approval being a largely theoretical system with no track record, that limited track record is far from promising -- repeals of it in the NGO elections (Dartmouth alumni, who voted it out 81% to 19%, and IEEE.
I hope you can start trying to make a positive case for your reform. I know this is hard for you.
Clay Shentrup
Rob Richie,
You know perfectly well that Republicans in Burlington could have gotten their second choice instead of their third choice if they had insincerely ranked the Democrat in first place. Voters obviously would have wised up to this if they had kept using IRV. But Burlington repealed IRV, as you know.
> Meanwhile, approval voting creates obvious tactical incentives - if your first choice has a chance to win, you will want to only vote for that person
Obviously. Just like with Plurality Voting, if your favorite candidate has a chance to win, you vote for that candidate. Whereas if your favorite is Ralph Nader, you'll tend to vote for e.g. Gore.
With Approval Voting, you just vote for the same candidate you would with Plurality Voting, PLUS everyone you like better. So a Gore fan would probably want to vote only for Gore, whereas a Nader fan would probably want to vote for Gore and Nader.
Your blog link is full of misinformation. For instance, you talk about the "bullet voting" problem, which is actually a bigger problem for IRV than for Approval Voting.
http://www.electology.org/bullet-voting
Approval Voting has a record going back hundreds of years, giving us a more than adequate picture of how it works in real life. It has recently been used to elect leaders and nominees in the German Pirate Party, which just won 15 seats in the Berlin parliament. These were real, contentious elections, and the results showed that Approval Voting worked exactly as expected.
In fact, in the very elections you mention (at Dartmouth) there was an average of 1.81 approvals per voter, further contradicting your very own argument about "bullet voting". It turns out that Approval Voting was apparently repealed there because it WORKED (the cited reasons for repeal are pretty obviously untrue). See this thorough analysis:
http://scorevoting.net/DartmouthBack.html
Whereas IRV has historically been repealed because of VERY REAL cost and complexity concerns. That is the very problem we're experiencing here in San Francisco, and why IRV is likely to be repealed by voters next June.
Clearly I'm more than capable of making the positive cas for Approval Voting. It is superior to IRV in every meaningful way, from representativeness to simplicity.