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

A Reform That Keeps Coming Back

Ranked-choice voting asks voters to do something small and unfamiliar: instead of picking one candidate, rank them in order of preference. If nobody wins a majority of first choices, the last-place finisher is dropped and those ballots move to whichever candidate each voter ranked next, and the process repeats until someone crosses 50 percent ^[6]. The mechanism itself is not new — it traces back to single transferable vote theories developed in 1850s Europe, associated with Thomas Hare and John Stuart Mill — and it has already had one American life and death, arriving in Ashtabula, Ohio, in 1915, spreading to roughly two dozen Progressive-era cities, and then being largely repealed by mid-century before resurfacing in San Francisco in 2004 ^[1][2]. What makes it contested now is not the arithmetic, which is fairly settled, but what the arithmetic is thought to prove: whether a system that rewards broad acceptability over narrow intensity is democracy working better, or democracy being quietly re-engineered.

That question has become urgent because the reform is no longer a curiosity. Maine and Alaska use it for federal and statewide races, dozens of cities use it locally, and Washington, D.C., adopted it for its June 2026 primary ^[1][3][4]. At the same time, a wave of state legislatures has moved to ban it outright, and the Republican National Committee has taken an official position against it ^{[3][5][11][12]}. A reform that spent decades as a niche good-government idea is now argued over in explicitly partisan terms, which is itself part of the story.

What Isn't Actually in Dispute

Strip away the advocacy on both sides and a fairly stable factual record remains. Adoption has proceeded mostly through citizen ballot initiatives: Maine voters approved RCV in 2016 with about 52 percent, saw it narrowed by a court to federal races and primaries, and restored it through a 2018 people's veto at about 54 percent ^[9][10]. That system produced the first U.S. House race decided by RCV, when Democrat Jared Golden overtook Republican Bruce Poliquin's roughly 2,171-vote first-round lead in Maine's 2nd District in 2018 ^[9]. Alaska adopted a different hybrid — a top-four open primary feeding into an RCV general election — in 2020; in 2022 it elected Democrat Mary Peltola to the U.S. House in the same cycle that Republicans Mike Dunleavy and Lisa Murkowski won statewide, and a 2024 repeal effort failed by roughly 700 votes ^[7][8].

The backlash is equally documented. Since 2024, roughly 17 to 19 states have banned RCV statewide — among them Kentucky, Alabama, Mississippi, Louisiana, Oklahoma, and Missouri in 2024; Arkansas, Kansas, North Dakota, West Virginia, Wyoming, and Iowa in 2025; and Indiana and Ohio in 2026 — and Nevada voters reversed their own 2022 approval in 2024 ^[3][5][6]. The pattern of who is doing the adopting and who is doing the banning is stark: bills to bring in RCV are sponsored almost entirely by Democrats, and bills to ban it almost entirely by Republicans, following the RNC's unanimous 2023 resolution against it ^[11][12]. On the mechanics, independent research finds that about 4.8 percent of voters — roughly 1 in 20 — mismark an RCV ballot in some way, with errors concentrated in precincts with more Black, Latino, elderly, foreign-born, and lower-income voters, a pattern that mirrors error rates in conventional contests ^[16][17]. "Exhausted" ballots — those with no remaining ranked choice by the final round — average around 10 to 11 percent nationally but swing from under 10 percent to over 50 percent in long, crowded races; researchers who have studied this closely conclude most of that exhaustion is voluntary rather than a sign of confusion ^[18][19].

Five Ways to Read the Same Numbers

The pro-reform case, made by groups like FairVote and RepresentUs and echoed by a number of political scientists, treats RCV as a fix for two related problems: vote-splitting and manufactured pluralities. Because a second choice still counts, voters can back a favorite without fear of "wasting" the vote or acting as an accidental spoiler, and the eventual winner holds a majority of the ballots still in play rather than a mere plurality of a divided field; the theory adds that candidates chasing second-choice votes have an incentive to build coalitions and soften negative campaigning ^[4][6]. Advocates point to Maine and Alaska, which delivered majority winners and viable third candidates without a separate runoff election, and to academic work finding a modest moderating tendency under instant-runoff counting ^[4][14]. The caveat critics raise is that the measured real-world effect on who actually wins, and on policy outcomes, tends to be small — largely because RCV usually replaces runoff systems that already produced majority winners, muting the size of any improvement ^[18][19].

Conservative opposition, articulated by the Heritage Foundation, the RNC itself, the Foundation for Government Accountability, and state-level outlets such as Carolina Journal, argues almost the opposite: that RCV is needlessly complex, erodes public trust, and produces a "faux majority" by discarding exhausted ballots rather than counting a true one ^[11][12]. This school leans heavily on Alaska's 2022 special election, where two Republican candidates combined drew about 60 percent of first-choice votes yet a Democrat won after eliminations — offered as proof that the system can manufacture outcomes at odds with a plurality's evident preference ^[11][26]. The counterargument, made by defenders of the reform, is that a systematic partisan tilt remains largely untested in the data, and that Alaska's same voting system has also elected Republicans Dunleavy, Murkowski, and later Begich, which complicates any simple "rigged for Democrats" reading ^[8][20].

A third, cross-cutting view comes from the left rather than the right. Progressive skeptics writing in Democracy Journal and The Progressive argue that RCV's real effect is to reward candidates who are broadly tolerable rather than intensely preferred, which tends to favor centrist figures over movement-left ones — treating the system's celebrated moderation as a bug rather than a feature ^[21][23]. They point to New York City's 2021 mayoral primary, where moderates Eric Adams and Kathryn Garcia drew the widest support while progressive Maya Wiley finished third, and to Maine's 2018 gubernatorial primary, where a divided left flank lost out to moderate Janet Mills ^[21][22]. The blind spot here, as reform advocates and conservative supporters of RCV both note, is that the very centering effect progressive skeptics distrust is precisely what proponents are selling as the system's chief virtue — and the underlying effect itself appears to be modest and highly dependent on context ^[14][18].

A smaller, cross-cutting group of conservatives and libertarians takes the opposite tack from Heritage and the RNC. Voices at the Cato Institute, the American Enterprise Institute, and R Street, along with Republican figures like former Utah GOP chair Stan Lockhart, argue that RCV expands the "marketplace of ideas," lets conservative voters express their true preference without fear of spoiling a race, saves taxpayers the cost of separate runoff elections, and can actually help Republican candidates in Democratic-leaning territory by rewarding broad appeal ^[13]. They cite conservative wins under RCV systems and the long-running stability of preferential voting in center-right-governed Australia ^[13]. This position, however, now runs against an entrenched consensus within its own political coalition, and RCV's partisan branding appears to be hardening faster than this counter-narrative can spread ^[11][20].

Finally, a more empirical, largely academic reading — from the MIT Election Data and Science Lab, New America, and the American Bar Association — holds that because RCV typically replaces runoff elections that already deliver majority winners, its measurable effects on turnout, moderation, and outcomes are real but small and highly conditional on local context, and that the recent wave of bans looks more like partisan signaling than a response to hard evidence ^[18][19][20]. The acknowledged limitation of this view is that small effects on average can still mask large effects in specific close races, and that the large, long-run datasets needed to settle the partisan-advantage question simply do not yet exist ^[18][20].

The Forces Pushing on the Debate

Beneath the arguments sit structural dynamics that would shape this fight regardless of who has the better data. Once the RNC formally opposed RCV and its use clustered in Democratic-leaning cities and states, the reform itself became a partisan symbol; support and opposition now correlate more strongly with an area's Trump vote share than with any actual feature of the ballot ^[6][20]. RCV is also frequently bundled with open or top-four primary structures, as in Alaska, which loosen party control over who gets nominated — a structural threat to gatekeeping that gives both major-party establishments reason for caution, independent of ideology ^[8][11]. The two paths of institutional change run in opposite directions, too: adoption has come almost entirely through citizen ballot initiatives, while bans have come almost entirely through Republican-controlled state legislatures, meaning the same reform advances and retreats through different doors ^[3][5]. Because most American RCV contests replace two-round runoffs that already guarantee a majority winner, that shared baseline caps how large any measured difference can look, whichever side of the debate is correct ^[18]. And underneath all of it are mundane administrative questions — new tabulation equipment, multi-round counting procedures, voter education — that shape whether a jurisdiction can adopt RCV at all and whether the public trusts the results once it does ^[19].

What the Record Doesn't Settle

Several questions remain genuinely open rather than merely contested. Whether RCV systematically benefits one party is largely untested with the data currently available; the best reading available is that any effect is conditional and probably modest rather than uniformly favoring Democrats ^[18][20]. The size and durability of the "moderation" effect is disputed and appears to differ depending on the counting method used — instant-runoff versus Condorcet-style approaches — which makes clean causal attribution difficult ^[14][15]. How much ballot exhaustion reflects genuine voter error versus a deliberate decision not to rank every candidate is also disputed, with estimates of inactive ballots varying widely across studies ^[18][19]. Turnout effects are mixed: RCV appears to raise turnout when it merges what would otherwise be separate primary and general elections into one contest, but some single-office comparisons show modest declines ^[18]. Even the basic tally of banning states is a moving target, with fights like North Carolina's repeatedly revived elections-bill provision still unresolved as of mid-2026 ^{[3][5][24][25]}. And how much can be learned from long-running preferential-voting systems abroad, such as Australia's since 1918, is unclear given that Australia pairs the ballot with compulsory voting — a feature the voluntary-turnout United States does not share and that may be doing much of the work credited to the ballot design itself ^[28][29].

How the Coverage Itself Splits

The sourcing on RCV tracks the same fault lines as the argument. Advocacy outlets like FairVote present the reform's benefits as settled fact, largely dropping the caveats that more careful researchers attach ^[4]. Heritage Foundation and Alaska Policy Forum material leans on charged language — "faux majority," "failed experiment," "disenfranchise" — and leads with a verdict before laying out evidence ^[11][26]. A counter-current within the right, represented by the Cato Institute, argues openly against the RNC's position, an unusual instance of intra-coalition dissent ^[13]. Reference and academic sources — Ballotpedia, the MIT Election Lab, New America, the American Bar Association — tend toward tables, tallies, and hedged language like "mixed" or "conditional," foregrounding the limits of what the data can show rather than pressing a conclusion ^{[3][10][18][19]}. And a distinct left-leaning skeptical strand, in Democracy Journal and The Progressive, breaks from the reform coalition it might be expected to join, framing the same moderating tendency that centrists and conservative reformers celebrate as a structural problem rather than a solution ^[21][22]. Read together, the spread of sources suggests the debate over ranked-choice voting is less a dispute over what happens when ballots are counted than over what kind of winner an election is supposed to produce.