Two-round runoff voting uses two elections. The first is a “choose-one” election (often a primary) to narrow the field to two candidates, and the second (often a general, but sometimes a special runoff) between the top two vote-getters. Like RCV, runoffs seek to uphold the principles of majority rule that voters can support their compromise choices without hurting their first choices. But there are downsides.
RCV saves taxpayer money
RCV allows cities and states to condense two elections into one, saving taxpayer money. In 2013, New York City paid $13 million for a delayed runoff, a cost it eliminated when it switched to RCV in 2021. Runoffs in Texas and Louisiana also come with multimillion-dollar price tags.
To compete in second elections, candidates must go back to their donors to quickly refill their campaign coffers.
RCV prevents turnout drop-off
Because two-round runoffs require voters to return to the polls, turnout often declines in the second election (by nearly 40 percent in congressional primary runoffs). What’s more, runoff voters are generally less representative of the voting population as a whole.
When RCV is used as an "instant runoff" to condense two-round elections into one, the same group of voters participate in every round, and more voters take part in the decisive election. Under RCV, voters rank candidates in order of preference. Some may not rank all candidates. These ballots become inactive if all ranked candidates are eliminated during the count. But even when taking into account these inactive ballots, RCV outperforms two-round runoff elections in final round turnout and representation.
This chart shows that two-round runoff winners often earn even fewer votes in the runoff than they earned in the first round, reflecting the small number of voters who participate in runoff elections.
RCV has a stronger mandate
Two-round runoffs are valuable in that they elect a winner with a majority of votes (among the voters who participate in the final round). However, when there are only two candidates, winning 50% of votes does not necessarily equate to 50% support. It could mean being the “least worst” out of two options. In an RCV race between a less-limited general election field, a majority of votes carries a much stronger mandate since the winner had to potentially beat several opponents and win support from multiple voter coalitions.
Two-round runoffs are more susceptible to tactical voting than RCV
A two-round runoff is susceptible to tactical voting because people can change their preferences between rounds (for example, a voter votes for a weak candidate in the first round, so their actual favorite candidate has a weak opponent in the runoff. Then, in the runoff, the voter votes for their actual favorite). With RCV, voters cannot switch their votes between rounds, and they cannot vote tactically (because they do not know who will be eliminated in what order), and are therefore less likely to vote tactically.
RCV prevents vote-splitting
Crowded fields in the first round can result in vote-splitting between similar candidates, potentially leaving large factions without representation on the runoff ballot. For example, in 2021 Texas held a two-round runoff election to fill the vacancy in its 6th congressional district. Two Republicans advanced out of a field of twenty-three candidates. Because Trump only carried the district by 3% in 2020, Democrats hoped to advance one candidate into the runoff, but their votes were split between four of the top-eight finishes, shutting them out of the runoff. Such instances of vote-splitting are generally avoided under RCV, since votes can be redistributed to other candidates between rounds.
This table compares election results from 84 RCV races in 10 cities and one state since 2003 with 257 municipal, congressional, and statewide primary runoffs since 1994, the 22 statewide runoff elections held in 2018, and the 14 runoff races which took place in San Francisco from 2000-2003, before the city started using RCV.
https://e.infogram.com/ae4e1bd3-b855-4a36-8b92-88590daf6513?src=embedRanked Choice Voting vs Runoffs800668no0border:none;allowfullscreen
As intended, both runoffs and RCV resulted in a majority of active votes for every winner in the final round. However, RCV winners’ median vote share in the final round was significantly higher (49% of the first-round vote) than two-round runoff winners’ median share (37% in congressional primary runoffs, 34% in San Francisco runoffs, and 36% in statewide runoffs).
Disclaimer: Demographics and voter turnout statistics were determined from voter data from L2, which uses weighted census data and surname classifications to estimate voters’ racial and ethnic background. Though blunt, this method provides a workable estimate for voter turnout by ethnic group in prior elections.
Note on methodology: L2 separates data by party for primaries but not runoffs. For example, data are available for all participants in Republican primaries, but runoff data don’t identify who participated in which party’s runoff. This is a problem because many runoff states have open primaries (meaning that restricting voters by party in the runoff doesn’t offer an accurate picture).
To get around this issue, both primary and runoff data are filtered for the party in question. For example, if 1,000 Republicans participate in a Democratic primary, they are excluded in the runoff (regardless of whether they turned out for it or not) because of the filter for the Democratic party. To avoid skewing the data, those Republicans must also be excluded from the primary data.
This is especially necessary when looking at changes in turnout by demographics because the different parties have different demographic compositions, so including different parties’ voters in the primary but not in the runoff would significantly alter demographic data. Since few people vote in other parties’ primaries, the exclusion should not have a large impact on the results.
The election below is a real election from San Francisco in 2020. Seven candidates ran for Board of Supervisors in the 7th district to fill an open seat.
Below is a table of first-choice results.
Because these voters voted using ranked ballots, we can simulate hypothetical head-to-head matchups based on ranked ballot data. The chart below shows a table of all head-to-head matchups. The table is best read across rows. For example, the first row with data can be read as “Engardio is preferred over Nguyen by 52% of voters, Engardio is preferred over Melgar by 47% of voters, …”
We can clearly see that Melgar would win a head-to-head matchup against every other candidate, making her a consensus choice and the Condorcet winner.
A two-round runoff election would have advanced Engardio and Nguyen to the runoff election, where Engardio would have prevailed 52%-48%. Two-round runoff elections can fail to elect a consensus candidate if such a candidate is not in the lead in first choices. Such a candidate may be impacted by vote-spitting in choose-one elections, in which some voters from their primary base of support may have divided their support with another similar candidate.
Below are the ranked choice voting results.
Melgar consolidates support as trailing candidates are eliminated, until ultimately facing Engardio in the final round and defeating Engardio 53%-47%. Ranked choice voting elects a consensus winner, or a Condorcet winner, when a two-round runoff election would have failed to do so.