Ranked-Choice Voting in America: A Map of the Evidence and the Argument

Two Republicans Got 60% of the Vote. A Democrat Won Anyway.

In Alaska's 2022 special election for its lone U.S. House seat, two Republican candidates together took about 60% of first-choice votes. When the counting finished, Democrat Mary Peltola held the seat ^[11][17][26]. Both of those facts are true at once, and that collision is the whole ranked-choice voting fight in miniature.

Ranked-choice voting, or RCV, lets voters rank candidates instead of picking just one. If nobody wins a majority of first choices, the last-place finisher gets eliminated. Ballots cast for that candidate move to whichever candidate each voter ranked next. The process repeats, round by round, until someone has a majority of the ballots still in play ^[6]. That is the whole mechanism — and it is also exactly why the Alaska result happened. One of the two Republicans got eliminated first, and enough of his voters either ranked no one else or ranked Peltola over the remaining Republican for her to win ^[11][17].

To supporters, that is the system working exactly as designed: the winner had to earn support beyond a narrow base. To critics, it is proof the method can bury what a plurality of voters actually wanted. Both readings use the same numbers. That is what makes this fight harder to referee than most.

An Idea From the 1850s, Revived in 2004

RCV is not new. It grew out of single-transferable-vote proposals floated in 1850s Europe, tied to thinkers like Thomas Hare and John Stuart Mill ^[1][2]. The first American use was in Ashtabula, Ohio, for a 1915 city council race. It then spread to roughly two dozen Progressive-era cities before mostly disappearing by mid-century ^[1][2]. San Francisco revived it in 2004, and the current wave dates from there ^[1][2].

Today Maine and Alaska use it for federal and statewide elections. Dozens of cities use it locally, and Washington, D.C. used it for the first time in its June 2026 primary ^[1][3][4]. Maine's path shows how contested this can get even within one state. Voters approved RCV in 2016 with about 52% support. A court then limited it to federal races and primaries. A 2018 "people's veto" restored the full law with about 54% ^[9][10].

That same year, RCV decided its first U.S. House race. Republican Bruce Poliquin led the first round in Maine's 2nd District by about 2,171 votes. After ballots transferred, Democrat Jared Golden won ^[9][10]. Alaska followed a different path in 2020, pairing RCV with a top-four open primary. In 2022 that system elected a Democrat, Peltola, to the House while also electing Republicans Mike Dunleavy and Lisa Murkowski statewide ^[7][8]. A 2024 ballot measure to repeal it failed by roughly 700 votes ^[7][8].

The backlash has been just as fast. Between 2024 and 2026, a wave of states banned RCV statewide — Kentucky, Alabama, Mississippi, Louisiana, Oklahoma and Missouri in 2024; Arkansas, Kansas, North Dakota, West Virginia, Wyoming and Iowa in 2025; Indiana and Ohio in 2026 ^[3][5]. That brings the total to roughly 17 to 19 states, and Nevada voters reversed their own 2022 approval in 2024 ^[3][5][6]. In North Carolina, Republican legislators have repeatedly folded an RCV ban into a larger elections bill, dropping and reviving the provision through 2025 and 2026 ^[24][25].

The party lines are stark. The Republican National Committee unanimously passed a resolution opposing RCV in 2023. Across recent legislative sessions, bills to adopt it are sponsored almost entirely by Democrats, and bills to ban it almost entirely by Republicans ^[11][12].

What the Ballots Actually Show

Strip away the arguments and a few hard numbers remain. Roughly 4.8% of voters — about 1 in 20 — mismark an RCV ballot in some way, whether by skipping a rank or ranking two candidates in the same slot ^[16][17]. Those errors cluster in precincts with more Black, Latino, elderly, foreign-born, and lower-income voters, though that pattern also shows up in ordinary, non-ranked elections ^[16][17].

Then there is "ballot exhaustion" — when a voter's ranked choices all get eliminated before the final round, leaving that ballot with no one left to count. Across U.S. RCV elections, exhaustion averages around 10% to 11%, but it swings wildly, from under 10% to more than half of ballots in some long, multi-round races ^[18][19]. That range matters because it is the number both sides use to make opposite points. Critics call it proof that RCV manufactures a "faux majority" by discarding votes that never got fully counted ^[11][26]. Researchers who have studied it closely find something different: most of those exhausted ballots reflect voters who simply chose not to rank every candidate, not voters confused by the system ^[18][19].

Support and opposition also track something outside the ballot design itself. Across the 2024 ballot measures on RCV, the higher an area's support for Donald Trump, the lower its support for ranked-choice voting ^[6][20]. That correlation does not prove RCV favors Democrats. It shows that RCV has become a partisan signal, which is a different — and, as later sections show, contested — claim.

Five Ways to Read the Same Election

Reform advocates, led by groups like FairVote and RepresentUs, make the most straightforward case. A voter can rank a long-shot favorite first without "wasting" that vote, because a backup choice still counts if the favorite loses ^[4][6]. That neutralizes the classic spoiler problem, where two similar candidates split a majority and hand the win to someone most voters opposed. The eventual winner also holds a majority of the ballots still active in the final round, not just a plurality ^[4]. Maine and Alaska both produced majority winners without needing a separate runoff election, and some studies find RCV nudges candidates toward less negative campaigning ^[4][14]. The gap in this case: careful reviews find the actual effect on who wins is often small, mostly because RCV usually replaces a traditional runoff system that already guaranteed a majority winner ^[18][19].

Conservative opposition, from the Heritage Foundation, the RNC, and state groups like the Foundation for Government Accountability, argues RCV is confusing and erodes trust in results ^[11]. Their sharpest claim is that exhausted ballots let a candidate win with a "faux majority" — a majority only of the ballots still standing, not of everyone who voted ^[11][26]. They point to Alaska's 2022 race, where the two Republicans' combined 60% first-choice share did not translate into a Republican win ^[11][17]. The gap here: the claim that RCV systematically helps Democrats is largely untested, and Alaska's own system has also elected Republicans Dunleavy, Murkowski, and later Nick Begich ^[8][20].

A third camp sits on the left, not the right. Writers at Democracy Journal and The Progressive argue RCV rewards candidates who are broadly tolerable over ones who are intensely wanted by a committed base ^[21][22]. They point to New York City's 2021 mayoral primary, where centrists Eric Adams and Kathryn Garcia had the widest appeal while progressive Maya Wiley finished third, and to Maine's 2018 primary, where a divided left lost to moderate Janet Mills ^[21][22]. Their gap: the same centering effect they call a flaw is precisely what reformers and conservative supporters count as the benefit — and even that effect looks modest and dependent on local conditions ^[14][18].

That leads to a fourth, less expected group: conservative and libertarian supporters. The Cato Institute, the American Enterprise Institute, and Republicans like former Utah GOP chair Stan Lockhart argue RCV lets conservative voters back their true favorite without fear of splitting the vote, and can help Republicans win in Democratic-leaning cities ^[13]. It also removes the cost of a separate runoff election ^[13]. Their gap: this view now runs against a party consensus that has hardened fast, limiting how far it can spread ^[11][20].

Finally, a centrist, empirical camp — MIT's Election Lab, New America, the American Bar Association — treats RCV's effects as real but modest, and highly dependent on local context, since it usually just replaces an existing runoff system ^[18][19][20]. In this reading, the current backlash looks more like partisan branding than a response to hard evidence, and no side has proven a clean partisan advantage either way ^[18][20]. Their own caveat: small effects on average can still hide big swings in individual close races, and there simply is not enough long-run data yet to settle the partisan question ^[18][20].

Why the Fight Is Bigger Than the Ballot

Some of what drives this debate has little to do with vote-counting math. RCV is often paired with open or top-four primaries, as in Alaska, which loosen party control over who even appears on the general-election ballot ^[8][11]. That threat to party gatekeeping is a structural reason both major parties' establishments treat it warily, separate from any argument about fairness ^[8][11].

Adoption and repeal also travel through different doors. Most RCV wins have come through citizen ballot initiatives, while most bans have come through Republican-controlled state legislatures ^[3][5]. That means the same reform can advance in one venue and retreat in another, often in the same state, without either side needing to persuade the other.

And because most American RCV elections replace two-round runoffs — which already produce majority winners — the baseline comparison mutes how big RCV's measured effects can look, no matter which side turns out to be right about the direction ^[18]. Add in the real costs of new vote-counting equipment and voter education, and adoption often turns as much on administrative readiness as on the policy argument itself ^[19].

What Would Actually Settle This

The core partisan question — does RCV help one party over the other — remains largely untested, and the best current evidence points to effects that are small and depend on circumstances, not a uniform tilt toward Democrats ^[18][20]. Nobody has yet run enough large-scale, long-term elections under RCV to answer it with confidence.

The size of any "moderating" effect on candidates is also disputed, and it seems to depend on which counting method is used — instant-runoff versus other approaches like Condorcet counting — which makes it hard to pin blame or credit on RCV itself ^[14][15]. Similarly, how much ballot exhaustion reflects genuine voter choice versus confusion is still argued over, with estimates ranging widely study to study ^[18][19].

Turnout evidence cuts both ways: it rises when RCV merges what would have been separate primary and general elections into one contest, but some single-office mayoral races have shown small declines ^[18]. Even the state ban count keeps moving — the roughly 17-to-19-state figure could shift again depending on how North Carolina's stalled elections bill resolves ^[3][5][24]. And a lesson often cited by advocates — that preferential voting has worked for over a century in Australia — comes with a catch rarely mentioned in the same breath: Australia also has compulsory voting, and how much of its stability comes from mandatory turnout rather than the ballot design itself is genuinely unclear ^[28][29].

Underneath all of it sits a question no dataset can answer. When a system nudges winners toward candidates who are broadly acceptable rather than fiercely preferred, is that democracy functioning as intended, or democracy losing something it should keep? Advocacy groups on both sides answer that question first and cite evidence second; the more hedged, academic sources — MIT's Election Lab, Ballotpedia — tend to lead with the limits of what the data can say ^{[4][10][11][3]}. That gap in framing may be the most durable finding in the whole debate.