It took the Alameda County Registrar of Voters more than a week to officially declare Mayor Jean Quan the winner of the historic November 2010 Oakland mayoral race, the first time the city had used ranked-choice voting, a system in which voters select their top three candidates. On election night, political heavyweight Don Perata had 34 percent of the first choice votes, while Quan lagged 10 points behind. But by November 10, eight days and 10 rounds of ballot counting later, Quan had clinched 50.96 percent of the votes after second and third choices had been tabulated.
Now, two years after Quan took office, some voters still don’t understand why Quan won, so voting officials are working hard to make sure people understand the system better this year, and that ranked choice ballot results are returned more quickly.
Under ranked-choice voting, people mark their choices on the ballot in succession—first, second and third. On election day, if no candidate wins the majority of the vote (50 percent plus one), the candidate with the fewest votes is eliminated, and the ballots are recounted. If a voter’s first choice candidate has been eliminated, then their second choice on the ballot is counted instead. That process of elimination continues until a candidate wins a majority.
But in 2010, said Dave Macdonald, Alameda County’s Registrar of Voters, on election night his office only released the results of the candidate with the most first choice votes, which seemed to put Perata ahead. The ranked-choice tally, which takes into account second and third choice votes after the trailing candidates have been eliminated, wasn’t run until three days after the election. If voting officials had recounted the second and third place votes that night, it would have revealed a much tighter race—with Quan, as well as Councilmember Rebecca Kaplan and Joe Tuman, pulling ahead.
“One of the things that added to the confusion was waiting for results to get back,” Macdonald said. In addition, he pointed out, not all ballots can be tallied on election night. “There are thousands and thousands of vote-by-mail ballots that are not counted until days after the election,” he said. “There are a large number of provisional ballots, and we still had to verify all the signatures.”
This year on November 6, voters will use ranked-choice to select candidates for council districts 1, 3, 5, 7, the at-large seat, the Oakland City Attorney and school board members for districts 1, 3, 5, and 7. Elections for government positions at the county, state and federal level do not use ranked-choice voting, and local and state ballot measures are also exempt. This year, said Macdonald, “We’re taking a look at running the ranked-choice voting algorithm on election night. It’ll be unofficial, but the benefit is it could give you a better snapshot of what the results may be.”
The Registrar of Voters has 29 days to verify results and officially certify the election.
In 2010, several factors added to the confusion over who had won, said Corey Cook, a professor at the University of San Francisco, who teaches classes on American political systems, and who also tracks ranked-choice voting outcomes in San Francisco and Alameda counties. Not only was it Oakland’s first use of ranked-choice voting, but new voting machines were being rolled out, the public was clamoring for the Registrar of Voters to call the race as soon as possible and media pundits were too quick to announce a Perata victory.
“We’re trained that when we see election results on Tuesday night that we have some sense of what those mean,” said Cook. “In Oakland what happened with Don Perata is he had a nine-point lead, and no one had ever come back from that—most of the time the frontrunner wins.”
But things were different in Oakland, Cook said. Historically, he said, even with ranked-choice voting, voters rarely use up all of their choices. That means that typically, whoever is ahead in the first round wins the election. “But in Oakland that was different because we had things we hadn’t seen before, like people using all three choices,” Cook said. “And Perata was uniquely unpopular.”
Nevertheless, Cook said that the media and political pundits were quick to call the election for Perata that night, instead of waiting for the registrar to tally the second and third choice votes. “I was one of them quoted in the Chronicle and on NPR the night of the election saying ‘Anything can happen,’” Cook said. “I was better than some others—some said, ‘There’s no way Quan can win—this race is over.’ I was careful not to say that, but I sure as hell thought that.”
But ultimately, he said, Quan “was able to do something we didn’t expect.” After Tuman fell out in round eight of the ballot count, with 15,46 votes, and Kaplan was eliminated in round nine, with 32,719 votes, Quan pulled ahead in round 10. She defeated Perata with 50.96 percent of the vote, compared to his 49.04. That meant she had about 2,000 more votes than Perata—53,897 compared to 51,872.
Perata’s campaign initially circulated documents claiming that ranked-choice had caused voter confusion, and campaign spokesman Rhys Williams told Oakland North that ranked-choice voting (RCV) “short-changed the public” and made campaigning difficult. “Voters we met door-to-door didn’t understand RCV. We had to dedicate campaign resources to explain a basic right: how to vote,” he told Oakland North in 2010.
But Perata ultimately conceded the election to Quan, telling reporters, “The results are pretty clear. I have no quarrel with the way the election was conducted.”
Support for ranked-choice voting among Oakland’s political leaders has always been mixed. Councilmember Patricia Kernighan, who represents District 2 and was one of the two members of the City Council to get ranked-choice voting on the 2006 ballot as Measure O, said the voting system helps level the political playing field because it encourages candidates to work together and it eliminates the need for a June primary. “For me, the strongest reason to support it was that I think it is very burdensome and expensive for candidates to run two elections—one in June and another six months later,” Kernighan said. “Ranked-choice voting gives the average person to run for office, because they don’t have to raise so much money.”
She said it also encourages politicians to work together, asking people for their second or third place votes. “It doesn’t eliminate negative campaigning entirely, but it does seem to encourage alliances and there is a tendency for candidates to work together,” she said.
This incentive for collaboration, said Macdonald, is often cited as an argument in favor of ranked-choice voting. “It discourages candidates from running negative campaigns,” he said. “They can say, ‘If I don’t get your first choice vote, vote for me second. It encourages candidates to work together, and that’s what we saw two years ago when there were three candidates considered the favorite—Perata, Quan and Kaplan.”
Quan said that asking people for their second or third vote if she wasn’t their first helped her win. “Rebecca and I both had strong grassroots campaigns,” Quan said. “It wasn’t a formal thing, but asking voters to support both of us strengthened our campaign.”
She also cited higher voter turnout, saying that without a June primary, more people turn out in November. In 2010, nearly 123,000 of the city’s 200,000 registered voters showed up at polling places. “If you look at voter turnout in 2010, it’s much more representative of Oakland,” Quan said, “because more minority populations like Asians and blacks vote in the general than the primary election.”
An analysis by The Center for Voting and Democracy, a think tank based in Maryland that lobbies for higher voter turnout and fair election practices, which also runs FairVote.org, showed that over the period from 1992 to 2010, Oakland voter turnout in June primaries averages 76,475 voters, compared to 125,637 in November general elections.
In an email, Jason Overman, spokesman for third place finisher Rebecca Kaplan, echoed Quan, saying that ranked-choice voting encourages positive campaigning, that it’s less expensive than having two elections and that its results are based on higher voter turnout. “One of the really valuable benefits of ranked-choice voting is that it encourages more positive campaigning around issues and coalition-building,” Overman wrote.
But some members of the City Council remain critics of the system. Councilman Ignacio De La Fuente said the system marginalizes some populations. “I believe in one person, one vote,” De La Fuente said as he left a council meeting two weeks ago. “Ranked-choice voting doesn’t allow people who know where their vote goes.”
According to a final report by the Alameda County Registrar of Voters, 99.7 of the ballots in the 2010 mayoral race were valid, but just 78 percent of voters fully used the ranked choice option by marking three separate choices, instead of only choosing one candidate or marking the same candidate three times. (In such a case, only the person’s first choice will be counted.)
“A voter can do whatever they want on their ballot,” said Noe Lucio, of the Registrar of Voters, who gives presentations on how the system works. “But we don’t recommend it, because if you vote for the same candidate three times in a row, or you only vote for one person, then the ballot becomes exhausted and there’s nothing else we can do with it.”
Cook said it’s typical for voters to not use all three choices in ranked-choice elections. But, he said, the fact that nearly 80 percent of voters ranked all three candidates during the 2010 Oakland election is “really high.”
“That’s why the election surprised people like me who thought that Jean Quan couldn’t win,” he said. Normally, “it’s closer to half of voters who use all three choices.”
Cook and Macdonald agree that it’s important this year to clear up any remaining voter confusion, using voter outreach programs and staffers at polling stations to answer questions people might have. In general, Cook said, “Minority language voters and older voters are more likely to make mistakes because there is more confusion about how ranked-choice voting works.” But in Oakland, he said, because voting officials send out workers to the polls to answer questions and because of the public information drives at senior centers and in neighborhoods where there is a high percentage of people who speak English as a second language, error rates in the county are lower than neighboring San Francisco, who has used ranked-choice voting since 2004.
The Alameda County’s Registrar of Voters is also increasing outreach by giving talks at community groups and sending out explainers in the mail. On a recent afternoon, a group of about 20 elderly people gathered to discuss issues important to their senior living community—the Allen Temple Arms Retirement Community in East Oakland. Shelley Burton, the retirement community’s service coordinator, had requested a presentation from the Registrar of Voters to explain how ranked-choice voting works.
“What happens if I vote for the same person all three times?” asked Bernice Randle, 72, who lives at Allen Temple Arms and said she doesn’t understand ranked-choice voting. “What if I don’t want to vote for three people, and I only vote once?”
Lucio, who was there to answer questions, explained that only the first choice would count.
Patricia Shoate, 72, said she remembered ranked-choice voting from voting in the 2010 election. “At first I didn’t understand it, but after someone came in and explained it to me, I got it. And I don’t like it,” she said. “I like the old system, with two people, better.”
Cook and Macdonald agree that there were hiccups in reporting the results of the 2010 election. Cook said the delay in announcing the final results—rather than confusion about how to mark the ballot—had the biggest effect on voter trust. “Where you get into problems is in this delay, like with the Mayor Jean Quan race, is when people started saying, ‘Wait a minute, now. You need to explain to me how you got there. I want to know your math now, because for three days Jean Quan was losing to Don Perata and now you’re telling me something different,’” Cook said.
For his part, Macdonald said he hopes that news outlets and political analysts will be more careful this year about announcing election results right away. “It drives me crazy when the media declares winners on election night,” Macdonald said. “It happens all the time, when—whoops!—a few days later someone else pulls ahead. The results on the first night can be misleading since we haven’t processed all those other ballots.”
Cook agreed that people like himself, whom many turn to for election results analysis, will likely be more careful this year: “They’ll say, ‘Hey, remember Oakland, 2010, when something we didn’t think was possible became possible?”