Last month, New York City’s primary voters used ranked-choice voting (RCV). In a November 2019 referendum, the city adopted RCV by an overwhelming margin. The reform was touted as a faster and cheaper alternative to the previous runoff system and as an approach that would help eliminate the so-called spoiler effect, through which an unpopular candidate wins because their opponents split all the other votes.

But reformers often conflate two types of spoiler effects. In one, the candidate with the most votes — the winner — may take the victory with less than 50 percent of the vote. In the other, the winner-takes-all voting system discourages candidates from running for fear of vote-splitting. The way reformers try to “solve” the second effect might actually make the first one more likely.

What exactly is a ‘spoiler’?

When Americans think about spoiled elections, they often have the 2000 presidential election in mind. In that election, whichever candidate won Florida would win the electoral college, and thus the presidency. Three candidates were on the ballot: Democrat Al Gore, who won 48.84 percent of the votes; Republican George W. Bush, who won 48.85 percent; and third-party candidate Ralph Nader, who took 1.64 percent. Some have argued that, had Nader not been on the ballot, most of his voters might have voted for Gore instead — and Gore would have been president instead of Bush. Commonly, Democrats think of Nader as the spoiler here, imagining that without him in the race, Gore would have won.

But when we’re talking about a spoiler effect — in which something in the system is spoiled — we see two possibilities. First, the winner took the electoral college prize without a majority of Florida’s votes, because some went to an unpopular candidate. Second, going in, Nader — or any third-party candidate in the U.S. system — knew he could not win, which in most cases would deter candidates from minor parties from running. That second spoiler effect is speculative, because there’s no way to readily measure whether someone decides not to jump into a race.

Those who advocate for ranked-choice voting often discuss both views. Groups such as FairVote say that ranked-choice voting will eliminate spoilers in which someone wins without a majority of the votes, and that numerous candidates will be free to run without risking vote-splitting.

How does New York City’s system work?

In New York City, voters may rank up to five candidates in order of their preference. If no candidate wins a majority from those first choices, the ballots are retabulated. The last-place candidate is eliminated, and every ballot that ranked that candidate first now gets its second choice counted as its vote. If a ballot does not have a ranking for a candidate still in the race, that ballot is considered “inactive.” The majority threshold is recalculated, so that candidates need a majority only of the still-active ballots, not of all the ballots cast.

If no candidate gets that majority in the next round, this recalculation continues until one remaining candidate wins a majority of still-active ballots.

So did New York City’s first ranked-choice primaries eliminate a spoiler effect?

To find out, we looked at the 46 primary races for Democratic city council nominations. Only 13 of 46 city council elections were decided in the first round, with clear majorities favoring one candidate. For the others, so many ballots became “inactive” as top-ranked candidates were eliminated that most winners did not earn a majority of the votes cast.

Did the buzz around ranked-choice voting encourage too many liberals to jump in?

In our second spoiler effect, candidates simply don’t run because they don’t think they can win. It’s not entirely clear that this actually happens, and some observers argue that it doesn’t. However, groups such as FairVote often encourage people to run, saying that the system will protect against the first spoiler effect, in which an unpopular candidate wins because other candidates split the vote. We cannot know whether that encouragement actually brought more candidates into the New York City primaries — there’s no way to measure who would have run without ranked-choice voting. But in 2017, the average number of candidates in Democratic primary city council races was three; in 2021, the average number was six. RCV supporters claimed the greater numbers of candidates as a victory for the change.

What does this mean elsewhere?

Our point is not that ranked-choice voting necessarily hurts liberals. In Australia, since the 1980s, ranked-choice voting has tended to benefit candidates on the left of the political spectrum. Rather, the change to a new system — and the discussion about what that system might have achieved — could have itself had side effects.

Any electoral system has its limits, and any election reform involves trade-offs. If other jurisdictions want to adopt ranked-choice voting, reformers may wish to think about how candidates might navigate the potential spoiler effects.

Editor’s note: This post has been updated to remove discussion of a particular New York City council race.

Lindsey Cormack (@DCinbox) is assistant professor of political science at Stevens Institute of Technology and author of “Congress and U.S. Veterans: From the GI Bill to the VA Crisis” (Praeger 2018).

Jack Santucci is assistant teaching professor in political science at Drexel University.