San Francisco made headlines recently with the election of Chesa Boudin as district attorney. Soon to be sworn in as the city and county’s top prosecutor, Boudin, a public defender, holds radical views on prosecution and law enforcement. Though his victory has been seen as evidence that famously progressive San Francisco has lurched even further left, Boudin was an outlier candidate even by local standards, and he won with a margin of less than 1 percent. His victory was largely due to San Francisco’s use of Instant-Runoff Voting, a poorly understood alternative-voting system used nationally in about two dozen city elections—including in New York City—and one state system.

IRV is a variation of ranked voting. In ranked-voting systems, instead of selecting for one candidate, voters rank the candidates in order of their preference, from first choice to last. If a candidate has a majority of first-choice ranks, he wins the election outright. But if no candidate wins a majority, IRV automatically runs a series of virtual runoff elections by eliminating the candidate with the fewest first-choice votes and redistributing those votes to each voter’s second-choice candidate. The system redistributes the votes through multiple rounds until a winner emerges.

IRV remains rare, but it has become the alternative-voting method of choice in the United States. The ability to vote for a candidate with little chance of winning, then have that vote transferred to a more viable candidate, is said to encourage participation by allowing people to support their first choice without feeling that their vote was wasted. In an IRV system, Al Gore could have won Florida (and with it, the presidential election) in 2000 when Ralph Nader voters might have picked Gore over George W. Bush as their second choice; in 2016, Donald Trump would have won the popular vote (and a few states that he barely lost, like Minnesota) when Gary Johnson and Evan McMullin voters could have chosen him as their second choice over Hillary Clinton. In these hypothetical cases, IRV could have reduced the claims that the election was unfairly decided.

IRV also allows compromise candidates (“everyone’s second choice”) to rise from crowded fields. Imagine an electorate 20 percent far left, 30 percent center left, 30 percent center right, and 20 percent far right. With IRV, centrist voters can rank centrist candidates higher, and as counting rounds continue, presumably a centrist candidate eventually wins. Alternatively, an electorate 30 percent left, 40 percent center, and 30 percent right ends up with an extreme winner if two centrists split the initial vote and only the far-right and far-left candidates proceed to a run-off. With IRV, after one centrist is eliminated, the other presumably gets his votes, leading to a centrist win when one of the more ideological candidates is eliminated next.

So what happened in San Francisco? How did the most extreme candidate end up winning, instead of the centrist, establishment-approved Suzy Loftus or the law-and-order-focused candidates Nancy Tung and Lief Dautch? The answer lies in what are called “exhausted” votes. IRV does not require that voters rank every candidate; in a field of four, they can rank all four or choose to rank only one, two, or three. But if a voter’s first choice is eliminated, and the voter doesn’t mark a second choice, then his vote is not counted in subsequent rounds. As a result, the simplistic examples above (all center-left voters mark the center-right candidate as their second choice) break down in a real-world application.

This is what happened in San Francisco, as the detailed Board of Elections results show. First, not all voters rank their choices in ways that mirror logical examples like those above. Almost 5,000 of those who voted first for Dautch, endorsed by the Deputy Sheriffs Association, voted second for Boudin, though Boudin was strongly opposed by law enforcement. Even more significant are the exhausted votes. Over 6,000 of Dautch’s votes, and almost 16,000 of Tung’s, didn’t count in the final tally. Considering that Boudin won over Loftus by a margin of less than 3,000 votes, a fraction of those 22,000 exhausted votes could have tipped the election to Loftus had voters marked her as their second choice instead of leaving that spot blank.

Several factors could explain why so many voters for Dautch and Tung didn’t rank Loftus as a later choice even though she was closer to their views than the radical Boudin. First, some voters might not understand how the system works, especially newcomers to San Francisco from jurisdictions with traditional voting systems. More significantly, to fill a vacancy, Mayor London Breed controversially appointed Loftus as acting district attorney a few weeks before the election. Breed’s support for Loftus was heavily criticized as a presumptuous power grab by the mayor and may have influenced voters to avoid ranking Loftus.

IRV’s advocates tout it as a system that ensures a clear majority winner and produces elected officials in tune with the broad electorate, rather than with one extreme or the other. This is probably true; IRV has many positives compared with traditional voting. In San Francisco, however, it failed. Due to ballot exhaustion, Boudin—who, even in left-wing San Francisco, held views outside the mainstream—won with less than 45 percent of valid votes cast. San Francisco’s lesson, for jurisdictions considering adopting IRV: no voting system, no matter how ingenious, can take the place of informed and active voters.

Photo by Kena Betancur/Getty Images


City Journal is a publication of the Manhattan Institute for Policy Research (MI), a leading free-market think tank. Are you interested in supporting the magazine? As a 501(c)(3) nonprofit, donations in support of MI and City Journal are fully tax-deductible as provided by law (EIN #13-2912529).

Further Reading

Up Next