RCV lessons for the SF mayor's race

In our winner-take-all society, the incentives of ranked-choice voting find common ground and build coalitions


OPINION Elections using ranked choice voting (RCV) in both San Francisco and Oakland contain important lessons for the upcoming SF mayoral election. Rather than rely on traditional endorsements and funding advantages, winning candidates need to get out in the community, meet people, and build coalitions.

Jean Quan became the first Asian American woman elected mayor of a major city by coming from behind to beat the favorite, former state Senate president and powerbroker Don Perata. Perata outspent her five to one, but Quan countered by attending far more community meetings, forums, and house parties. She would knock on the door of a voter with an opponent's yard sign and say, "I know I'm not your first choice, but please make me your second or third choice."

She also reached out to her progressive opponents, especially Rebecca Kaplan, saying, "In case I don't win, I think Rebecca should be your second choice." As a result, Quan received three times more runoff rankings from the supporters of Kaplan, who finished third, than Perata did. That propelled Quan to victory.

Perata, meanwhile, used the traditional front-runner strategy of spending more money. His campaign never figured out that he needed to seek the second and third rankings from the supporters of other candidates by finding common ground.

A similar story also played out in SF's supervisorial Districts 2 and 10. In those races, victors also won by coming from behind and picking up more second and third rankings from other candidates' supporters.

In D10, some people seem to think that winner Malia Cohen wasn't a strong candidate because she wasn't one of the top-two finishers in first rankings. But this reflects a misunderstanding of this race's dynamics. In the final results, Cohen finished third in first rankings (not fourth, as the early results showed), yet she was only five votes behind Tony Kelly for second place and only 53 votes behind Lynette Sweet in first place.

So Cohen was as much a front-runner as either Kelly or Sweet in an extremely close race with 22 candidates. She prevailed by picking up more second and third rankings from other candidates' supporters, resulting in an African American candidate winning this traditionally black district.

Note that if D10 had used San Francisco's old December runoff, the voter turnout would have plummeted from the high of a November gubernatorial race, and the winner would have won with a handful of votes. The RCV system worked to pick the candidate preferred by the most voters in a single November election.

In D2, fiscal conservative Mark Farrell beat the progressive's choice, Janet Reilly. But this district is not a progressive one, and that's supposed to be one of the benefits of district elections (which was a progressive reform), i.e. each district is able to elect its own representative who conforms to the majority of its district instead of what Big Money interests want. Unfortunately, that also means a progressive candidate probably won't win a nonprogressive district. Farrell built an effort that attracted more second and third rankings from other candidates' supporters, allowing him to come from a point behind to win a close race.

That's the way you win with RCV. With no clear frontrunner, the candidate who can draw significant numbers of second and third rankings is most likely to win. In our overly adversarial, winner-take-all society, the incentives of RCV to find common ground and build coalitions with ranked ballots is a relief for most voters. Mayoral candidates should take note. 

Steven Hill is author of 10 Steps to Repair American Democracy (www.10Steps.net), Europe's Promise (www.EuropesPromise.org) and other books, opeds, and articles. Visit his website at www.Steven-Hill.com.


IRV in D2, D10 and Oakland for that matter elected little more than the candidate everyone hates the least.

D10 was pure random noise, scattershot. The case for IRV in D10 versus a December runoff is tenuous at best, with, what a quarter or so of voters in the first round having their votes survive to elect the winner. Exhausted ballots won in D10.

This disenfranchisement of minority voters would raise voting rights implications if it took place in a jurisdiction with a historical record of disenfranchisement which liberal San Francisco would never, ever approach.

We're ten years into the IRV experiment in San Francisco. It has not had the desired outcomes of moving partisan elections towards IRV to open the door to other political parties. If anything, the partisan system is tighter than its ever been with the imposition of the open primary.

To my mind, the phenomenon of the runoff paid fantastic dividends to progressives because it afforded organizing opportunities. In 1999, 2000 and 2003, mayoral and supervisorial runoffs set the stage for greater gains in the next years. But that door is closed, and we've got a system that elects the person everyone hates the least assuming that person is not named "Exhausted ballot."


Posted by marcos on Dec. 29, 2010 @ 7:47 am

So much of the criticism of ranked choice voting that I have heard from progressives boils down to sour grapes that RCV didn't deliver us shining progressive outcomes on a silver platter.

But the far more important consideration is to imagine what we would have gotten -without- ranked choice. In D-10 this would have been the unacceptable developer puppet Lynette Sweet, and in Oakland, the even more unacceptable corrupt power Democrat Don Perata. In D-6, the running of two strong progressive candidates delivered a trouncing of Downtown backed candidate Sparks; who in a winner take all system would have had a much better shot at winning.

I call all of that RCV success, not failure.

If progressives want better outcomes through ranked choice, they need to get off of their asses and use the process better. For example, in D-8 we made the huge mistake of running only one candidate, when instead, both Tom Radulovich and Rafael Mandelman should have run together, pulling in a more disparate spectrum of progressive voters and giving voters legitimate second choices.

Conservatives in D-2 and D-8 did an excellent job utilizing RCV by running more than one strong candidate, and it worked.

We need to do the same in the upcoming mayoral race. If we do so effectively, we could be looking at the first progressive mayor in San Francisco in decades.

Posted by Eric Brooks on Dec. 29, 2010 @ 1:56 pm

Marc raises the legitimate point that, in D10, there were more exhausted ballots at the end of the count than received by the winner. (In fact, there were more exhausted ballots at the end than received by the remaining two candidates combined.) The best way to reduce the number of exhausted ballots, or at least the involuntary ones, is to get equipment that allows voters to easily rank more choices. For example, the equipment used by Cambridge, Massachusetts, allows voters to rank up to 30 candidates, should there be that many.

(Cambridge uses the proportional-representation cousin of IRV to elect its 9-member city council every two years. They typically have 20 or more candidates for the 9 seats, and voters are allowed to rank all of them, even though only about 10% of the voters rank more than 9; the average voter ranks 5 or 6 candidates. Statistics from their 1999 City Council contest with 24 candidates on the ballot plus five qualified write-in candidates; one voter ranked all 29.)

The San Francisco Voting Systems Task Force is supposed to make recommendations concerning "next generation" equipment, among its other duties. Even so, I was surprised to discover that the latest draft of their report (Draft 9, available at http://www.sfgov2.org/Modules/ShowDocument.aspx?documentid=155 ) makes no mention of allowing voters to rank more candidates. Readers of this post might want to send their own comments to the Task Force.

Yes. Ballots allowing more than three choices are crucial.

And it is also incumbent upon us as campaigners and candidates to strongly educate, and push for voters to use a full ranked choice strategy when they go into the booths. This also will reduce exhausted ballots, which, for example, were far lower in Oakland where the progressive camp aggressively campaigned around RCV.

Posted by Eric Brooks on Dec. 30, 2010 @ 5:39 pm

Steven Hill should disclose the financial windfall he is receiving for rank choice voting, whose only truthful aspect is that it's RANK. You morons used this illegal, violates-my-right-to-one-vote farce to vote in the biggest psycho in the East Bay, Rebecca Quan - that's no typo, the "winner" was a hybrid of two losers.

Even worse is the supposed ENLIGHTENED aspect of rank choiceless ripoff - the joke that it makes "coalition building" - no it makes COLLUSION BUILDING. I can't wait for the next election, I'm getting together with my wacko Tea Party neighbor & making for an election you freaks will never forget. We are putting up 5 candidates, all friends, all of us campaigning for a "vote for me but vote for John Doe as #2". So a useless loser, John Doe, will be the winner because there's no way 5 candidates can't round up enough 2nd place votes to stick him into office.

In the meantime, when Oakland starts to burn I hope everyone will remember the traitors who forced this cr#p on us - Steven Hill will be the first person to be burned in effigy. But of course, he'll have taken his millions and left town by then. I can't wait for you to be exposed for the liar you are, Hill, you are disgusting. YOU DESERVE REBECCA QUAN. The chant is on: GTO. GTO. GTO. GTO. GTO.

Posted by Guest on Jan. 08, 2011 @ 8:09 pm

Guest, if John Doe is everyone's second choice but no one's first choice, then John Doe cannot win, as (with no votes) he will be eliminated in the first round.

Instant runoff voting has been challenged in the courts numerous times, and has always been held to not violate the one-person-one-vote rule.

And what does "GTO" mean? Please explain.

Posted by Steve Chessin on Jan. 19, 2011 @ 7:00 pm

First of all, in what bizzaro universe would it possibly take place that a candidate would get "everyone's" second choice and not anyone's first choice? That is as close to impossible as a vote in real life can get.

And in any case, it is the candidates who receive the smallest number of votes over all who are eliminated first -from the bottom- of the tally in each round, so what you are claiming, is also just completely wrong.

Posted by Eric Brooks on Jan. 20, 2011 @ 12:03 am

"Yes, ballots allowing more than three choices are crucial" NO - ballots allowing me to vote for whom I wish, and the majority of the people choose, is crucial. Take your commie cr#p somewhere else.

Posted by Guest on Jan. 08, 2011 @ 8:11 pm

'Guest' Your preferred voting method is called 'majority rule'.

Majority rule by its very nature, is not democracy.

In ranked choice, though the final winner's totals can be smaller, in the overall election, far more voters are given a chance at -meaningful- participation in the outcome of the vote, because they don't have to fear that their first choices will 'spoil' the election. They can properly reflect their political views at the polls.

Seeing how these voters prefer to vote when they are not fettered by fear, smart candidates will then work to shift their policies enough to please many of those voters in the next election and gain their added support.

A system much closer to true democracy than majority rule.

Posted by Eric Brooks on Jan. 08, 2011 @ 10:25 pm

Double rr'd when I should have doubled nn'd.

It's: tyranny

Posted by Eric Brooks on Jan. 08, 2011 @ 10:50 pm

Progressives have been opposed to mixing it up in elections when they may lose. As in top two and forming a commission on redistricting state wide.

It isn't about getting people involved, its about getting over.

Posted by matlock on Jan. 08, 2011 @ 10:55 pm

PR STV with obviously a complete restructuring of the electoral system, political-economy and re-districting *might* resolve some of the issues.

Top 3? seriously? what century drafted that?

Posted by Sean on Jan. 08, 2011 @ 11:22 pm

Also from this author

  • Guest opinion: RCV is good for progressives

  • Save SF's campaign finance program

    Dipping into the public financing fund for any reason sets a terrible precedent

  • Election security that works

    California should become its own vendor