On the perils of 'Instant runoff' voting
A blindfold and a dartboard
Berkeley resident David Reid decided to take a poll this fall on his daily walks, with poodles Lucy and Dora, in the city's Rose Garden neighborhood. How many of his neighbors in this upscale and well-educated swath of Berkeley, he wondered, understood how the new ranked-choice voting system worked?
"I asked city officials, academics, writers ... none of them understood it," said Reid, editor and an author of the 1994 book "Sex, Death and God in L.A."
"Don Perata certainly didn't," Reid added.
Perata's defeat by Jean Quan in the Oakland mayoral race was merely the highest-profile oddity in the ranked-choice results from the Nov. 2 elections in Alameda County and San Francisco. Perata was the first choice of 33.7 percent of Oakland voters, far ahead of Quan's 24.4 percent.
But once the computer started culling the field, dropping the last-place finisher and transferring second- and third-place votes to the remaining candidates, Quan emerged as a narrow winner. Quan appeared to have benefited from the "anybody but Don" strategy employed by her and fellow Councilwoman Rebecca Kaplan.
Advocates of ranked-choice voting, also known as instant-runoff voting, like to claim that the system encourages more positive, substantive campaigns than a traditional runoff election between the top two finishers. In this case, however, the winning strategy was simplistic and decidedly negative.
Results in Oakland, the San Leandro mayor's race and two races for San Francisco supervisor proved every bit as "topsy-turvy" as Anthony Gottlieb warned in a scathing critique of ranked-choice voting in the July 26 New Yorker.
"They are what mathematicians call 'non-monotonic,' which means that something can go up when it should go down, or vice versa," Gottlieb wrote.
Perata, San Leandro Mayor Tony Santos and San Francisco supervisor candidates Janet Reilly (District 2) and Tony Kelly (District 10) all had the most first-place votes - and ended up losing.
Some of them might have lost in a traditional runoff, but it would have been through a democratic process - with the candidates forced to build coalitions with the also-rans, and, most important, all voters having the opportunity to decide between the two finalists.
Under ranked-choice voting, the resolution can become the equivalent of a blindfolded shot at a dartboard.
In San Francisco District 10, centered in the southeast Bayview district, ultimate winner Malia Cohen received just 11.8 percent of first-place votes and would not have qualified for a top-two runoff. Nineteen rounds later, after 19 candidates were eliminated, Cohen scored a narrow victory over Kelly - the candidate with the most first-place votes.
FairVote, a Takoma Park, Md., group that is promoting ranked-choice voting, issued an analysis of the District 10 race that bravely but unconvincingly attempted to characterize the system as a success and declaring Cohen the "strongest candidate" because of her ability to collect second- and third-choice votes. FairVote's analysis predicted that a December runoff (San Francisco's old system) would have produced "mudslinging and hack attack campaigning."
Perhaps. It is at least as likely that a runoff between the top two finishers would have drawn out meaningful differences between the candidates. In 2006, voters in San Francisco's Sunset District might have learned that Ed Jew (first-place choice of 26.2 percent) did not even live in the district. He prevailed in the ranked-choice runoff - and received jail time for his residency deception.
Voter confusion is a serious problem with ranked-choice voting. Forty percent of voters in San Francisco's District 10 did not list three choices for supervisor. Mayor Gavin Newsom acknowledged that he voted for the same candidate for supervisor three times. His mistake was spotted by a poll worker. Voters who make such an error on mail-in ballots do not get a chance to correct them.
The fatal flaw with ranked-choice voting is that voters whose top three choices do not make the final cut are effectively disenfranchised. This injustice was highlighted in District 10 - the ballots of the majority of people who cast a vote in that race had been disqualified when it came to the final choice between Cohen and Kelly. Cohen's winning "majority" represented less than a quarter of all votes cast in the race.
A system that confuses and disenfranchises voters is fundamentally undemocratic. Voters here should do what those in Burlington, Vt., did after two mayoral elections: Scrap this experiment in ranked-choice voting.
Ranked-choice voting: How it works
-- Voters rank their top three choices in order of preference.
-- If a candidate wins a majority, he or she is declared the winner.
-- If no candidate wins a majority, the candidate who received the fewest first-place votes is eliminated - and votes for the second choice of the eliminated candidate's voters (at least for those who named a second choice) are counted in the next round.
-- The process continues, eliminating the lowest-scoring candidate each round, until a candidate reaches a majority. The "instant runoff" for San Francisco's District 10 supervisor's race, which had 21 candidates, went 19 rounds.
John Diaz is The Chronicle's editorial page editor. E-mail: jdiaz@sfchronicle.com.
This article appeared on page E - 2 of the San Francisco Chronicle
Visit this link to sign up for email updates:http://feedburner.google.com/fb/a/mailverify?uri=ProtectUsElections-StopInstantRunoffVoting&loc=en_US
