FairVote's First Take on RCV Elections in Four Bay Area Cities
On November 16, this piece was updated to reflect final results.
Ballots for Tuesday’s ranked choice voting (RCV) elections in four cities in the Bay Area has now largely been counted. It is clear that RCV has again worked well. Because of RCV, voters in Berkeley, Oakland, San Francisco and San Leandro were able to elect leaders in a high turnout presidential election instead of having to rely on either low turnout runoffs in December or low turnout primaries in June. As a result, a far larger and more representative electorate participated in city elections than when these cities used non-RCV voting systems.
oday we focus on three particularly important points emerging from the 2012 elections.
CANDIDATES OF COLOR AGAIN THRIVED IN RCV RACES, EVEN WHEN OUTSPENT
Racial minorities continue to do remarkably well as candidates in ranked choice voting elections in the Bay Area. In San Francisco, for example, there are 18 offices chosen by RCV, and 16 next year will be held by people of color -- including two new winners this year in the District 5 and District 7 races for the Board of Supervisors. Once these new members are seated, the Board of Supervisors will have five members who are Asian American, two who are African American, two who are Latino and two who are white. In 2004, when RCV was first used, the Board had only four people of color, including only one Asian American.
As evidenced by Oakland’s District 3 City Council election, voters of color are using the RCV system to clear effect. Three strong African American candidates ran in District 3, with their backers generally promoting ranking these candidates first, second and third. Although a white candidate led in first choices, Lynette McElhaney prevailed, with particularly strong support from voters who had backed Derrick Mohammed and Neisha de Witt as a first choice. McElhaney won despite being outspent by Sean Sullivan – a common pattern in RCV races in which the winning candidate has been particularly active in direct outreach to voters.
RATE OF VOTER ERROR IS LOW: VOTERS HANDLE RCV BALLOT WELL
The results demonstrate that election officials in both Alameda County and San Francisco continue to do an excellent job in ensuring voters know how to fill out an RCV ballot. Even though Berkeley, Oakland, San Francisco and San Leandro have all held RCV elections before, many voters came out to vote for president this year who had never voted locally in an RCV election.
Nevertheless, very few voters made an error in indicating a first choice in RCV races. In Alameda County RCV races, for example, more than 99.5% of voters in multi-candidate RCV races cast valid ballots, including more than 99.6% of voters in the races with the most candidates in Berkeley, Oakland and San Leandro. In San Francisco’s large-field RCV races in District 5 and District 7, more than 99.4% of voters cast valid ballots for their first choice, and more than three in four voters indicated a valid second choice as well. These rates of invalid ballots are far lower than the invalid ballot rate in the U.S. Senate primary in June 2012 , and lower than the expected rate of error in the San Francisco school board races.
Ballot image reports are not yet available for the Alameda County races, but are available for analysis in the two San Francisco races that require multiple rounds to determine a winner (District 5 and District 7). In both races two-thirds of voters used all three of their rankings and eight in ten voters ranked two – this despite the fact that many endorsing entities like the San Francisco Chronicle endorsed only one or two candidates.
A TRUE “INSTANT RUNOFF" IN 2012 : MOST RCV OUTCOMES CLEAR ON ELECTION NIGHT
Election officials in San Francisco and Alameda County should be applauded for choosing to run the ranked choice voting algorithm on election night. Doing so not only provided faster results, but also clarified that RCV has not been the reason for any past delays in determining outcomes.
Indeed the timing for reporting results should be no different than when reporting non-RCV results. Just as with non-RCV elections, most RCV races will be known as soon as the RCV tally is run. If an RCV election is close, of course, then the winner may not be decided until absentee and provisional ballots have been completely processed. That of course is no different than in non-RCV races, such as the California Attorney General race in 2010 that required weeks to decide, in which the race is close and all absentee and provisional ballots must be processed before the winner can be conclusively determined.
This year, RCV tallies on election night made it instantly clear who was the winner in nearly every race in which there was no majority winner in the first round. In San Francisco’s District 5, for example, London Breed led with less than 30% of first choices, but was a decisive winner over incumbent Christina Olague in the RCV tally – showing strength in securing second and third choices that propelled her to being a certain winner. Lynette McElhaney was second in first choices in Oakland’s District 3, but was a clear winner after running the RCV tally.
By running the tally early, the public and the candidates had more clarity about who is ahead and who is behind, as is normal when election returns are reported. In San Francisco’s District 7, for example, Norman Yee leads in first choices, but is locked in a very tight race in the final round of the RCV tally with F.X Crowley. In San Leandro’s District 2, incumbent Ursula Reed leads in first choices and is in a tight race with challenger Morgan Mack-Rose in the final round of the RCV tally.
The only two RCV races where outcomes were in doubt as remaining ballots were tallied were ones where the election was extremely close. The winner was ultimately the finalist who earned the most votes in the “instant runoff” – in other words, was preferred to the other finalist by more voters. Norman Yee in San Francisco and Ursula Reed in San Leandro ultimately won these clsoe races, but all the votes has to be counted, just as would have been the case in a close non-RCV race that had uncounted ballots.