n.Newfeed

Politics · 7 min read

Ranked-Choice Voting: What Municipal Elections Show About Turnout and Ballot Exhaustion

A bounded review of turnout estimates, ballot exhaustion, majority denominators, disparities, and implementation limits.

Ranked-Choice Voting: What Municipal Elections Show About Turnout and Ballot Exhaustion

One ballot can contain several contingent choices

In the single-winner form of ranked-choice voting used in many U.S. municipal elections, voters rank candidates. If one candidate has the required majority of continuing ballots, that candidate wins. Otherwise the lowest candidate is eliminated and ballots assigned to that candidate transfer to the next ranked continuing choice. The process repeats. This method is also called instant-runoff voting, although "ranked-choice voting" can refer to other systems, including multiwinner single transferable vote.

The counting rule is established; its political effects are empirical questions. Advocates argue that ranking gives voters more expression, reduces the need for a separate runoff, and rewards broader coalitions. Critics point to complexity, delayed tabulation, exhausted ballots, and unequal error rates. Evidence from municipal adoption is useful but difficult to generalize because cities change electoral rules amid other differences in campaign timing, competitiveness, demographics, and administration.

Two outcomes receive especially strong claims: turnout and majority support. The existing record supports modest, conditional conclusions. Ranked-choice voting does not reliably create a large turnout increase by itself. It often transfers many ballots successfully, but some become inactive before the final round. Whether that is a serious defect depends partly on the comparison—plurality elections, two-round runoffs, or another ranked method.

Turnout estimates depend on the comparison

A before-and-after comparison can mislead. A city may adopt ranked choice while moving election dates, changing candidate fields, or combining primaries and general elections. Presidential-year turnout differs from off-cycle municipal turnout regardless of ballot design. Strong studies therefore compare adopting cities with similar non-adopters, include election and demographic controls, or analyze a longer series.

Research has produced mixed estimates. Studies by political scientists including Caroline Tolbert, Todd Donovan, and their collaborators have associated ranked-choice adoption with higher participation in some local contexts, particularly compared with low-turnout separate runoffs. Other work finds little change or a temporary increase that fades. James Endersby and Michael Towle's 2025 analysis of elections for more than 500 public officials in 21 cities from 2001 through 2020 reports that turnout tended to rise when ranked choice was introduced and decline later.

Those results do not establish that the counting rule caused each movement. Adopting jurisdictions are not random. Reform campaigns may coincide with unusually competitive races and voter education. The strongest practical inference is comparative: replacing a later runoff with one ranked general election can prevent the sharp participation drop often observed between rounds. That does not mean more people participate than they would in a high-salience plurality general election.

Turnout should also be separated from participation in a particular contest. A voter may cast a ballot in the election but skip the ranked race, producing an undervote. Portland's first use of ranked voting in 2024, with a 19-candidate mayoral field and multiwinner council contests, prompted debate about race-level participation. One complicated implementation should not stand for all ranked elections, but it shows that candidate count and ballot design belong in the analysis.

Ballots exhaust when listed preferences run out

A ballot becomes exhausted when every candidate ranked on it has been eliminated and no further preference can transfer. Exhaustion is not necessarily an error. A voter may intentionally rank only acceptable candidates. It can also result from limits on how many rankings the ballot permits, confusion, or failure to anticipate which candidates will remain.

Craig Burnett and Vladimir Kogan examined four U.S. local instant-runoff elections in a 2015 Electoral Studies article. They found enough exhaustion that the eventual winner in each case lacked a majority of the ballots cast in the first round, even while winning a majority of ballots still active in the final round. The distinction is arithmetic but politically important. "Majority winner" can mean a majority of initial voters, continuing ballots, or voters expressing a preference between the finalists.

Broader datasets find that exhaustion varies with the number of candidates, rounds, ranking limits, and voter behavior. Endersby and Towle's multi-city work reports that elections without a first-round majority infrequently produce a winner supported by a majority of all initial voters after accounting for nontransferable ballots. Computational analysis of U.S. ranked elections likewise finds ballot exhaustion more common than exotic theoretical failures such as a monotonicity violation.

The appropriate comparison remains contested. In a plurality election, votes for losing candidates also do not count for the winner, but they remain in the denominator. In a delayed runoff, voters whose first candidate lost must return and choose again; many do not. Ranked choice records contingent choices on one ballot, preserving participation for those who rank a continuing candidate. Exhaustion reveals the voters for whom that contingency ended.

Error and burden are not evenly distributed

An overvote occurs when a voter marks more than one candidate at the same ranking in a way the rules cannot interpret. A skipped rank may or may not invalidate later rankings, depending on jurisdiction. Ballot design and tabulation rules therefore affect measured error. Clear instructions and opportunities to correct an in-person ballot can reduce invalid marks.

Several studies have found differences in ranking depth or ballot errors across age, education, race, and language groups in particular elections. These associations require care. They may reflect candidate preferences, differing numbers of acceptable choices, ballot familiarity, outreach, or design—not a fixed capacity of a demographic group. A disparity is a reason to investigate administration and access, not proof that one population cannot use the method.

Surveys often find that most voters report understanding ranked ballots after using them, while a minority finds them difficult. Self-reported understanding is useful but not identical to correct completion. Administrative records show marks, not voter intention. Combining ballots, surveys, and demographic analysis gives a fuller picture, though privacy protections constrain linkage.

Claims that ranked choice moderates campaigns or reduces polarization are less established than claims about tabulation. Candidate surveys and content analyses in some cities report less negative campaigning, but small samples, strategic context, and reform selection make causal inference difficult. The rule permits cross-endorsement incentives; it does not force candidates or voters to use them.

Majority language needs a denominator

Single-winner ranked choice guarantees a final comparison under its stopping rules, not universal support. If ballots exhaust, the winner may have a majority of continuing ballots but less than half of initial valid ballots. If a jurisdiction allows limited rankings, the ballot itself constrains expression. If a voter ranks every candidate except the finalists, abstention between those finalists may accurately represent the voter's preferences.

This does not make the outcome illegitimate by definition. Electoral systems convert incomplete and sometimes indifferent preferences into officeholders through specified rules. Plurality can elect a candidate with a small share in a crowded field. A separate runoff can produce a majority among a smaller returning electorate. Approval, Condorcet, and score methods satisfy different criteria and create other strategic incentives. The relevant question is which tradeoffs a jurisdiction wants and can administer transparently.

Reporting should therefore state the denominator: first-round ballots, continuing ballots, registered voters, or turnout in a hypothetical runoff. It should also publish round-by-round data so observers can reproduce transfers and quantify exhaustion. Transparency cannot eliminate disagreement about the rule, but it can prevent the same word from carrying incompatible claims.

The evidence supports bounded expectations

Municipal studies are not randomized trials, and U.S. cities are a selective sample. Local nonpartisan races differ from national party contests. Adoption often bundles reforms, and learning may change effects over several cycles. New implementations with many candidates, such as Portland's multiwinner system, should not be pooled casually with single-winner mayoral elections.

What is established is that ranked ballots transfer many voters' preferences without a second election and that some ballots exhaust before the final count. Existing studies suggest turnout effects are small, context-dependent, and sometimes temporary; replacing a separate runoff is the clearest route to retaining participation. The frequency and meaning of exhaustion depend on ballot rules and voter choices. Claims about civility, moderation, and representation remain less secure.

The policy decision cannot be read from one metric. Jurisdictions must weigh voter expression, administrative capacity, result timing, education costs, error disparities, and the alternative system actually being replaced. The research narrows plausible promises: ranked choice is a counting method with measurable tradeoffs, not a general cure for low participation or political conflict.

Sources and review note

4 sources, none of them with a link, so treat them as reading, not proof.

  1. Craig M. Burnett and Vladimir Kogan, "Ballot (and voter) exhaustion under Instant Runoff Voting," Electoral Studies 37 (2015), 41-49.
  2. James W. Endersby and Michael J. Towle, Ranked Choice Voting: Implementation and Effects on Democracy (Oxford University Press, 2025), analysis of 21 U.S. cities.
  3. Benjamin Reilly and related comparative electoral-system syntheses on instant runoff and alternative vote systems.
  4. Jurisdictional cast-vote records and official round reports, including San Francisco, Minneapolis, New York City, and Portland, used by the cited empirical literature.

The title is limited to municipal evidence and two outcomes. The draft distinguishes majority of continuing ballots from majority of initial ballots and avoids treating exhaustion automatically as error.

Want the reader to remember this spot? Sign in — it is free.

Next How Citizens' Assemblies Turn Random Selection into Policy Advice

The app

Keep going, or take it with you.

Free accounts read twenty pieces a day. Pro is US$1.99 a month or US$11.99 a year for unlimited reading.