At a glance
| Argument for | Argument against |
|---|---|
| Winner needs a majority of continuing ballots | That majority can exclude exhausted ballots |
| Reduces vote splitting between similar candidates | Can still eliminate a broadly liked "center" candidate |
| Replaces a separate runoff election | Final results take longer to report |
| Most surveyed voters say it's easy | Ballot errors are more common than on single-choice ballots |
| Some studies find less negative campaigning | Evidence on turnout is mixed |
Majority winners
For
Under plurality, a candidate can win a crowded race with a third of the vote. Instant runoff continues until someone has more than half of the ballots still in play, so the winner beat the last rival head to head.
Against
Ballots that rank no remaining candidate drop out. A study of four local elections found that in every one, the winner got less than a majority of all ballots cast, with exhaustion rates of 9.6% to 27.1% (Burnett & Kogan, Electoral Studies, 2015).
Context: a traditional two-round runoff has a similar issue, because many voters don't return for round two. See exhausted ballots and the real figures on our election pages.
Vote splitting and spoilers
For
Voters can rank a long-shot first without "wasting" their vote, and similar candidates don't split a majority. In our vote-splitting example, plurality elects a candidate that 60% of voters left off their ballots entirely; instant runoff doesn't.
Against
Instant runoff can eliminate the candidate who would beat every other one-on-one (the "center squeeze"), as analysts argue happened in Burlington, Vermont, in 2009. It is also not monotonic: in rare cases, more support can make a candidate lose (Stanford Encyclopedia of Philosophy). Methods compared.
Do voters understand it?
For
In an Edison Research exit poll after New York City's 2021 primary, 95% of respondents said the ballot was simple to complete and 77% wanted to keep using it. The poll was commissioned by groups that support ranked-choice voting (IVN summary).
Against
Opponents argue that ranking many candidates is harder than marking one, that it advantages voters with more time and information, and that results are harder to verify by hand (Foundation for Government Accountability, an opposing group).
Ballot errors
Rejected ballots are rare, but studies find they are more common in ranked races than in single-choice races on the same ballot, and more common in lower-income and less-educated areas. A study of New York City found more overvote errors in such neighborhoods in 2021; a follow-up found errors fell substantially by 2025, though the education gap persisted in weaker form (Cormack, 2024; follow-up). Some researchers dispute these studies' methods (Utah Valley University critique).
Turnout
Mixed. One study found a 17% higher predicted probability of voting in local elections in ranked-choice cities, along with more direct contact from campaigns (Dowling, Tolbert, Micatka and Donovan, Electoral Studies, 2024). Studies by Jason McDaniel found lower turnout and wider gaps by age and education in some cities. Ballotpedia keeps a summary of the academic literature (Ballotpedia).
Campaign tone
For
Candidates want to be voters' second choice too, which may discourage attacks. Voters surveyed in ranked-choice cities were less likely to describe campaigns as negative than voters in comparable plurality cities (Donovan, Tolbert & Gracey, Electoral Studies, 2016).
Against
The effect is measured through voter perceptions in local races; it's less clear in high-profile partisan contests, and cross-endorsements can look like coordination to voters.
Counting time and transparency
For
One election instead of two saves the cost of a separate runoff and the drop in turnout that often comes with it. First-choice results are usually available on election night.
Against
Final results wait for every ballot. Alaska's 2022 special election ran its ranked rounds about two weeks after election day; New York's 2021 primary was certified four weeks later, after a preliminary count wrongly included about 135,000 test ballots (NPR). Counting must be centralized rather than done precinct by precinct.
Who runs and wins
Evidence is mixed here too. A study of Bay Area cities found more candidates of color ran after ranked-choice voting was adopted, but not more women (New America summary). An experiment with hypothetical candidates found voters penalized candidates of color at similar rates under ranked-choice and plurality rules; its title sums it up: "Neither Panacea nor Catastrophe" (Crowder-Meyer, Gadarian & Trounstine, 2023). Advocacy groups on both sides publish stronger claims.
Legal challenges
Federal courts have upheld ranked-choice voting against "one person, one vote" and Voting Rights Act challenges, including the Ninth Circuit in Dudum v. Arntz (2011, San Francisco) and federal courts in Maine in Baber v. Dunlap (2018) and Hagopian v. Dunlap (2020) (Congressional Research Service). State constitutions are a separate question: Maine's highest court advised in 2017 that it can't be used for state general elections there.
A note on sources
Much of the public writing on ranked-choice voting comes from advocacy groups, for it (such as FairVote) and against it (such as the Foundation for Government Accountability). We cite peer-reviewed research and official records first, and label advocacy sources when we use them. If you spot an error, email [email protected].