An internal poll showed Frey with a 19-point lead in the Minneapolis mayoral race. But in an RCV election, he could still lose.

On Monday, All of Mpls, a political committee spending money to try to, among other things, re-elect Mayor Jacob Frey and to defeat the police-reform ballot question, released an internal poll with the finding that Frey was leading challengers with 44 percent of first-choice votes. Sheila Nezhad came in second, with 25 percent of first-choice votes, and Kate Knuth came in third with 10 percent.

One should take any poll released by a campaign with a grain of salt, or whole shaker: campaigns typically only release polls when they reflect favorably on their preferred candidates. Also, All of Mpls did not release important details about how the poll was done or any crosstabs.

But regardless of whether the poll is good or not, was the release right to proclaim Frey was in a “strong position” to win re-election? While in a traditional election, a 19-point lead would have any candidate feeling comfortable, the ranked-choice voting system used in Minneapolis means Frey’s supposed lead could be less formidable than it seems — if his opposition is united.

Ranked-choice review

In a normal election — like the ones Minnesotans vote in for president, Congress and State Legislature — voters pick one candidate for each race on the ballot. A candidate wins the election by winning at least one more vote than the next closest candidate.

In ranked-choice voting, now used by Minneapolis (and St. Paul, and an increasing number of municipalities in Minnesota that have adopted variations of the system), voters can rank candidates — up to three in Minneapolis — in order of preference for each office.

When votes are tabulated, if no candidate receives more than 50 percent of first-choice votes, the candidates with no mathematical chance of winning are dropped as a group. Ballots with those candidates as their first choice are reallocated to remaining candidates according to their second or third choices. The process then continues, with the candidate with the least votes dropped after each round, until one candidate reaches at least 50 percent of the votes plus one vote, not including ballots that were exhausted because none of the choices they listed were still in the running.

In 2017, for example, Frey received 25 percent of first-choice votes. Tom Hoch got 19 percent, former mayor Betsy Hodges got 18 percent, Ray Dehn got 17 percent and Nekima Levy Armstrong got 15 percent. It wasn’t until Levy Armstrong, Hoch and Hodges were eliminated after rounds of RCV tabulation that the race came down to Frey and Dehn, with Frey eventually winning.

Such an outcome is unsurprising in the world of ranked-choice voting. It’s almost always the case that the candidate who gets the most first-round votes also ultimately wins.

Underdog wins

There are exceptions to the rule, though: one, the Oakland, California mayor’s race in 2010, said Jason McDaniel, associate professor of political science at San Francisco State University.

In that open-seat race, former State Sen. Don Perata was ahead in the polls going into Election Day. When first-choice votes were tabulated, Perata led City Council member Jean Quan by more than 11,000 votes. But as lower-performing candidates were eliminated in the RCV process, more of their votes were allocated to Quan than Perata. In what was considered an upset, Quan won by about 2,000 votes.

A particular facet of the campaign made it possible for Quan to upset the first-round frontrunner, McDaniel said: two viable candidates — Quan and City Council member Rebecca Kaplan — were somewhat allied in opposing Perata, who was seen as the establishment candidate.

In a story published after her win, the New York Times described Quan’s strategy as one of aligning with the candidates in the race who were not Perata, focusing on being voters’ second choice.

“She singled out Mr. Perata, a conservative Democrat who had outspent everyone, and aligned herself with the other nine candidates, particularly the other major challenger, Rebecca Kaplan. She came to be seen as the leader of the “anybody but Don” coalition, appealing to voters who were wary of Mr. Perata,” the Times wrote.

‘Don’t Rank Frey’

That strategy of challengers allying to defeat a perceived frontrunner in an RCV election might sound familiar to Minneapolis voters in 2021. In Minneapolis, a “Don’t Rank Frey,” effort encourages voters to rank anybody but the incumbent mayor, denying him late-round votes as other candidates are eliminated.

Nezhad and Knuth have also formed an alliance, appearing together and announcing the other as their second choice. Their partnership has won the endorsement of Rep. Ilhan Omar, progressive political group TakeAction Minnesota and others.

This strategy may be Frey opponents’ best bet to beat the incumbent mayor — but it only works under the right conditions, McDaniel said.

In a scenario where Frey is ahead in first-round votes, Nezhad and Knuth would each have to get 20 percent to 25 percent of first-choice votes, McDaniel said, and they would need about 70 percent of their supporters to rank the other.

The underdog alliance move doesn’t always work: In San Francisco’s 2018 mayoral election, Supervisor Jane Kim and former State Senator Mark Leno campaigned together to try to blunt the edge of Supervisor London Breed, who was perceived as the frontrunner. Kim and Leno each received 24 percent of first-round votes to Breed’s 37 percent, but Breed stayed on top in subsequent rounds of tabulation as lower-performing candidates were eliminated.

No Minnesota candidate who was behind after the first round of ranked-choice tabulation had ever come back to win the election until 2017, when it happened in two Minneapolis’ city council races where the frontrunner was a polarizing candidate — a condition that can also lead to an underdog win. In the race to represent the 3rd ward, which includes Northeast, Marcy-Holmes and part of downtown Minneapolis, Socialist Alternative candidate Ginger Jentzen was ahead to begin with, with 3,297 first-choice votes. DFLer Steve Fletcher got 2,709. By the end of ranked-choice tabulation, though, their roles were reversed. Fletcher had 4,861 votes and Jentzen had 3,844.

In North Minneapolis’ Ward 4, incumbent Barb Johnson led in first-choice votes, with 2,258, while Phillipe Cunningham was in second, with 2,140. By the time the RCV process was finished, though, Cunningham was elected with 2,605 votes to Johnson’s 2,430.

Assuming Frey actually ends up ahead in first-choice votes, could another candidate come from behind to win in this year’s mayor’s race? Sure. But there is one caveat: McDaniel said he’s not aware of it ever happening to an incumbent.

Still, McDaniel said, “that 19-point lead, even if it’s optimistic for Frey, which it probably is from an internal poll, if he doesn’t have a clear double-digit lead in first-place votes, it’ll probably be a close race.”