vote_lede_template.jpgRanked choice voting, otherwise known as instant run-off, has been used in San Francisco elections since 2004. This past November’s elections marked the first time it was used in Oakland, Berkeley and San Leandro.

With RCV, voters rank candidates by 1st, 2nd, and 3rd choice. If no candidate receives more than 50% of the 1st choice votes, then the process of eliminating and redistributing votes begins. Details of how this works are available on the SFElections website. The new system produced some surprising outcomes, and in the wake of the results there has been a flurry of criticism and praise of how RCV plays out for voters.

In San Francisco, District 10’s new supervisor, Malia Cohen, initially in 3rd place after the first round of votes, won her seat in the 19th round of the computerized instant runoff. D10 is of particular interest because it was already a complicated ballot, with 21 candidates running for a single seat. California Watch analyzed the voting data from this district, which includes both Bayview/Hunters Point Housing Projects and middle class homes in Potrero, and came to the conclusion that RCV causes voter fatigue and confusion in especially low-income neighborhoods.

More than 2% of D10 voters (489 people) spoiled their ballots by marking more candidates than are allowed, making their ballots void, according to California Watch. Their analysis shows that this happened more often in the lower income neighborhoods of Bayview/Hunters Point and Visitacion Valley.

CA Watch also points out that Board of Supervisors ballot came after the one for governor, US Senate and Congress, other state races and initiatives, and by the time voters got to the end of the ballot 11 percent of voters didn’t bother with supervisor’s ballot at all.

Tony Santos, San Leandro’s former mayor who lost in his attempt to be re-elected in November complained to California Watch that “RCV is misunderstood by many voters, and it discriminates against minorities and individuals who have a problem with language.”

Santos was once a fan of RCV, but after losing to Steven Cassidy by only 232 votes in the final round, he has changed his mind. However, San Leandro did not experience nearly as many invalid ballots as in SF’s District 10 Race. Only 55 out of 22,484 voters in San Leandro’s mayoral race cast an invalid ballot.

In Oakland, where Jean Quan won by garnering the most 2nd and 3rd choice votes, more than 5 percent of voters made mistakes that discounted their votes. (This commenter pointed out a mistake I made in interpreting the data California Watch presented. It was not that 5 percent of the Oakland ballots were discounted, it’s that 5% of Oakland voters just chose not to cast second and third choices, but instead of leaving the spot blank they wrote in their first choice three times. That just means that for 5% of Oakland voters, those 2nd and 3rd votes weren’t counted, but their first choice was.) Despite this problem, the benefits to RCV can be highlighted in the context of the Oakland election. Beneficial, that is, to those that believe tactics like coalition building and community outreach should outweigh the traditional campaign that relies on endorsements and funding advantages.

Quan went out and told Rebecca Kaplan’s voters to put her as 2nd choice, and told her own followers to do the same for Kaplan. She received 3 times more runoff rankings from Kaplan supporters than her opponent, Don Perata, resulting in her victory.

In SF’s District 10, Malia Cohen won in a similar, yet more drawn out fashion. Each time one of the 21 candidates was eliminated, she gained a little bit more by racking up 2nd choice votes. Like Quan, she endorsed others for second choice and they returned the favor. Another candidate, Eric Smith and she even had joint literature. In addition, the SF Democratic party recommended her for 2nd choice after DeWitt Lacy.

In an op-ed in the SFBG, Steven Hill makes a point that represents a common view of proponents of RCV: “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.”

He also points out that before RCV, a December run-off would have been held and voter turnout would have plummeted. Now, everyone has a say in the run-off. With that said, the complications and confusion around the new system can’t be ignored.

San Francisco has spent $1 million dollars on outreach and education, but as California Watch has pointed out, it does not always appear to be enough. They suggest that there are limits to what a political system can ask of its citizens and maybe using an RCV ballot is too much to ask.

Want more news, sent to your inbox every day? Then how about subscribing to our email newsletter? Here’s why we think you should. Come on, give it a try.

Please make sure your comment adheres to our comment policy. If it doesn't, it may be deleted. Repeat violations may cause us to revoke your commenting privileges. No one wants that!
  • Patrick

    I would be interested to know how many ballots (percentage-wise) are typically voided through voter-error in non-RCV elections.

  • Patrick

    I would be interested to know how many ballots (percentage-wise) are typically voided through voter-error in non-RCV elections.

  • John E. Palmer

    Without comparisons regarding ballot spoilage, it’s hard to know whether more education on RCV is needed. Many relatively new systems (such as RCV in SF) require an adjustment period, but it doesn’t mean they are not an improvement over the past. The real problem in District 10 was an incredibly fractured race (a large field of candidates and no obvious frontrunner). Is it possible that we can improve the system in this situation, but still have the benefit of RCV? This should be considered. RCV allows for more “sincere” voting, and the elimination of the “spoiler” factor present in pluralilty elections. It also allows races to be less partisan (party primaries can be eliminated) and quite possibly less negative (vying for those second/third choice votes).

  • John E. Palmer

    Without comparisons regarding ballot spoilage, it’s hard to know whether more education on RCV is needed. Many relatively new systems (such as RCV in SF) require an adjustment period, but it doesn’t mean they are not an improvement over the past. The real problem in District 10 was an incredibly fractured race (a large field of candidates and no obvious frontrunner). Is it possible that we can improve the system in this situation, but still have the benefit of RCV? This should be considered. RCV allows for more “sincere” voting, and the elimination of the “spoiler” factor present in pluralilty elections. It also allows races to be less partisan (party primaries can be eliminated) and quite possibly less negative (vying for those second/third choice votes).

  • Preston Jordan

    It is not true that 5% of voters made mistakes that discounted their ballots in the Oakland Mayor’s race. I hope SFAppeal will issue a correction.

    The official tally at the Alameda County Registrar’s site (http://www.acgov.org/rov/rcv/results/rcvresults_2984.htm) indicates only 355 out of 122,268 ballots were discounted due to be being marked incorrectly (“exhausted by overvotes”). This is only 0.2% of the ballots. All other ballots that were marked counted in the race.

    Regarding the higher spoilage of ballots in lower income areas, that is an interesting analysis. What would make it more interesting though is comparing the number of valid votes in those areas in the most recent election versus the previous December runoffs. For instance within 1 minute on the web I just learned that only 4,542 Bayview-Hunters Point residents cast votes in the December 2000 runoff as opposed to 8,533 in the most recent election. And the 2000 runoff was not a sleeper. The favorite candidate of Bayview-Hunters Point lost to that of Portrero Hill, suggesting strong motivation to get to the polls.

    So why isn’t the media trumpeting RCV as a great civil rights success by increasing turnout in low income communities rather than trying to gin up arguments that RCV somehow failed this community? Set against this, the single percentage spoilage rate should be a footnote rather than the story lead.

  • Preston Jordan

    It is not true that 5% of voters made mistakes that discounted their ballots in the Oakland Mayor’s race. I hope SFAppeal will issue a correction.

    The official tally at the Alameda County Registrar’s site (http://www.acgov.org/rov/rcv/results/rcvresults_2984.htm) indicates only 355 out of 122,268 ballots were discounted due to be being marked incorrectly (“exhausted by overvotes”). This is only 0.2% of the ballots. All other ballots that were marked counted in the race.

    Regarding the higher spoilage of ballots in lower income areas, that is an interesting analysis. What would make it more interesting though is comparing the number of valid votes in those areas in the most recent election versus the previous December runoffs. For instance within 1 minute on the web I just learned that only 4,542 Bayview-Hunters Point residents cast votes in the December 2000 runoff as opposed to 8,533 in the most recent election. And the 2000 runoff was not a sleeper. The favorite candidate of Bayview-Hunters Point lost to that of Portrero Hill, suggesting strong motivation to get to the polls.

    So why isn’t the media trumpeting RCV as a great civil rights success by increasing turnout in low income communities rather than trying to gin up arguments that RCV somehow failed this community? Set against this, the single percentage spoilage rate should be a footnote rather than the story lead.

  • Eve Batey

    Thanks, Preston, we’ve fixed the error in the Oakland passage.

  • Eve Batey

    Thanks, Preston, we’ve fixed the error in the Oakland passage.

  • SFoodie

    It is fair to give a comparison of an error rate in a typical runoff election. According to the SF Department of Elections, runoff elections for D12 State Assembly was 0.02% and for the State Senate D8 it was 0.03%. This was found with the same link Williams used to find the over 3% in BayView Hunter. Ballot spoilage in a real runoff is very low as the report shows.

    What is not fair is for Preston to cherry pick his numbers on turnout to make RCV look better. If so, we could look at Minneapolis’ first RCV election where they had the lowest turnout in nearly 100 years. Last time they had a turnout that low, women didn’t have the right to vote. You can find this easily on the web. There’s what’s called correlation and causation. They fact that it did happen doesn’t mean RCV was the “cause”. Though green party members and RCV apologists would like you to think this was the case. It would be like saying the higher turnout of African American males in the 2008 Presidential election was because a white woman was on the ticket. Correlation and causation.

    It is not unusual for RCV supporters to misdirect attention from the failed experiment called RCV by trying to tie one item with another.

    Turnout for the 2003 December traditional runoff actually increased by 20%. They are not always low turnout elections. Just depends on how vigorous the competition is. Simple as that.

    What happened in D10 with Cohen winning with 24% support should be a real worry for the upcoming SF Mayoral Race. This race will have at least 7 STRONG competitors, and upwards to 15 candidates. Dufty, Lee, Ting, Chiu, Leno, Peskin, Gonzales, Mirkarimi, Herrera, Rees, Adachi and Hall have either filed or have been mentioned as mayoral candidates. Like D10, there may be more exhausted ballots that valid ballots, with the winner not having more than 50% support of San Franciscans. SF needs a Mayor with a true mandate. Funny thing, if you go to the elections website, you will see that “exhausted ballots” won the D10 race!

    RCV is a fun way of electing your Prom King and Queen, or the president of your local green party club. You can say, I voted for your Sara, and you Tina, and you Sasha without taking a stand and hurting any candidates’ feelings. But for Mayor, the system is broken. How about an election in the spring consolidated with the state/federal and a runoff if needed in November like the majority of cities in California? This would save the $ needed for a special runoff election in December, one of the driving forces behind the change to RCV.

    In the end, it is expected there would be vigorous replies from RCV supporters, they have just been told their baby is ugly. No one likes to hear this, but sometimes it’s true.

  • SFoodie

    It is fair to give a comparison of an error rate in a typical runoff election. According to the SF Department of Elections, runoff elections for D12 State Assembly was 0.02% and for the State Senate D8 it was 0.03%. This was found with the same link Williams used to find the over 3% in BayView Hunter. Ballot spoilage in a real runoff is very low as the report shows.

    What is not fair is for Preston to cherry pick his numbers on turnout to make RCV look better. If so, we could look at Minneapolis’ first RCV election where they had the lowest turnout in nearly 100 years. Last time they had a turnout that low, women didn’t have the right to vote. You can find this easily on the web. There’s what’s called correlation and causation. They fact that it did happen doesn’t mean RCV was the “cause”. Though green party members and RCV apologists would like you to think this was the case. It would be like saying the higher turnout of African American males in the 2008 Presidential election was because a white woman was on the ticket. Correlation and causation.

    It is not unusual for RCV supporters to misdirect attention from the failed experiment called RCV by trying to tie one item with another.

    Turnout for the 2003 December traditional runoff actually increased by 20%. They are not always low turnout elections. Just depends on how vigorous the competition is. Simple as that.

    What happened in D10 with Cohen winning with 24% support should be a real worry for the upcoming SF Mayoral Race. This race will have at least 7 STRONG competitors, and upwards to 15 candidates. Dufty, Lee, Ting, Chiu, Leno, Peskin, Gonzales, Mirkarimi, Herrera, Rees, Adachi and Hall have either filed or have been mentioned as mayoral candidates. Like D10, there may be more exhausted ballots that valid ballots, with the winner not having more than 50% support of San Franciscans. SF needs a Mayor with a true mandate. Funny thing, if you go to the elections website, you will see that “exhausted ballots” won the D10 race!

    RCV is a fun way of electing your Prom King and Queen, or the president of your local green party club. You can say, I voted for your Sara, and you Tina, and you Sasha without taking a stand and hurting any candidates’ feelings. But for Mayor, the system is broken. How about an election in the spring consolidated with the state/federal and a runoff if needed in November like the majority of cities in California? This would save the $ needed for a special runoff election in December, one of the driving forces behind the change to RCV.

    In the end, it is expected there would be vigorous replies from RCV supporters, they have just been told their baby is ugly. No one likes to hear this, but sometimes it’s true.

  • DT

    RCV is not easily understood by the majority of the voting populace; they do not have solid Math and English skills.

    Combined with District Elections, RCV has resulted in people voted into office with less than 1% of city-wide votes, thus millions are wasted on extremely narrow pet projects and special interests.

    Perhaps voters might be a little bit more cautious if there was some language in 32 point or larger red ink stating that “ballot will be spoiled if RCV done incorrectly.”

  • DT

    RCV is not easily understood by the majority of the voting populace; they do not have solid Math and English skills.

    Combined with District Elections, RCV has resulted in people voted into office with less than 1% of city-wide votes, thus millions are wasted on extremely narrow pet projects and special interests.

    Perhaps voters might be a little bit more cautious if there was some language in 32 point or larger red ink stating that “ballot will be spoiled if RCV done incorrectly.”

  • DT

    RCV is not easily understood by the majority of the voting populace; they do not have solid Math and English skills.

    Combined with District Elections, RCV has resulted in people voted into office with less than 1% of city-wide votes, thus millions are wasted on extremely narrow pet projects and special interests.

    Perhaps voters might be a little bit more cautious if there was some language in 32 point or larger red ink stating that “ballot will be spoiled if RCV done incorrectly.”

  • DT

    RCV is not easily understood by the majority of the voting populace; they do not have solid Math and English skills.

    Combined with District Elections, RCV has resulted in people voted into office with less than 1% of city-wide votes, thus millions are wasted on extremely narrow pet projects and special interests.

    Perhaps voters might be a little bit more cautious if there was some language in 32 point or larger red ink stating that “ballot will be spoiled if RCV done incorrectly.”

  • Al

    If people don’t read “Vote for One”, already printed at the top of the column, what makes you think they’ll read “ballot will be spoiled if RCV done incorrectly”?

    I just don’t see a 3% failure rate as something to get worked up about.

    BART Board elections, for instance, don’t use RCV, and it’s too bad: James Fang (who is awful) got less than 50% but the opposition was split down the middle. RCV could have changed that, and I’m sorry it wasn’t used.

    The main problem I see is that the ballots could be better-designed, but you’d need new counting machines for that. Computer screen voting could help, but it has its own problems.

  • Al

    If people don’t read “Vote for One”, already printed at the top of the column, what makes you think they’ll read “ballot will be spoiled if RCV done incorrectly”?

    I just don’t see a 3% failure rate as something to get worked up about.

    BART Board elections, for instance, don’t use RCV, and it’s too bad: James Fang (who is awful) got less than 50% but the opposition was split down the middle. RCV could have changed that, and I’m sorry it wasn’t used.

    The main problem I see is that the ballots could be better-designed, but you’d need new counting machines for that. Computer screen voting could help, but it has its own problems.

  • Steve Chessin

    I agree with Al about the need for better machines. Cambridge, Massachusetts, uses a related ranked-choice system to elect its nine-member City Council. All nine seats are up for election every two years and they typically get more than 20 candidates for the nine seats. The equipment they use allows voters to rank up to 30 candidates, so any exhausted ballots are because the voter doesn’t care anymore if all their choices are eliminated.

    The San Francisco Voting Systems Task Force (see http://www.sfgov2.org/index.aspx?page=1862 ) is supposed to make recommendations to the Board of Supervisors on “next generation” voting equipment, but at least as of Draft 9 of their report (available at http://www.sfgov2.org/Modules/ShowDocument.aspx?documentid=155 ) they haven’t mentioned the need to let voters rank more candidates. I suggest you let them know what you think.

    Steve Chessin
    President, Californians for Electoral Reform
    http://www.cfer.org

  • Steve Chessin

    I agree with Al about the need for better machines. Cambridge, Massachusetts, uses a related ranked-choice system to elect its nine-member City Council. All nine seats are up for election every two years and they typically get more than 20 candidates for the nine seats. The equipment they use allows voters to rank up to 30 candidates, so any exhausted ballots are because the voter doesn’t care anymore if all their choices are eliminated.

    The San Francisco Voting Systems Task Force (see http://www.sfgov2.org/index.aspx?page=1862 ) is supposed to make recommendations to the Board of Supervisors on “next generation” voting equipment, but at least as of Draft 9 of their report (available at http://www.sfgov2.org/Modules/ShowDocument.aspx?documentid=155 ) they haven’t mentioned the need to let voters rank more candidates. I suggest you let them know what you think.

    Steve Chessin
    President, Californians for Electoral Reform
    http://www.cfer.org