I’ve spent the last week seeking studies on the impact of Rank Choice Voting, which will be on the ballot in Irvine this November. The pro-RCV folks have posted significant data on talking points on why RCV is better than today’s choose one candidate for an elective office and whoever gets the most votes for that race is the winner.
The notion that the winning candidate must have 50% of the vote is a false premise because Irvine elections never result in 100% turnout. Our voting patterns fall between 75% during presidential electionyears and 54% in non-presidential election years and primaries.
Some research highlighted by the pro-RCV crowd struck me as not current; Ryan Dack shared a study from 2015. So I did some digging for newer research and what I found is the empirical research does not support the sweeping promises often used to sell RCV. And I’m doing the digging here so you don’t have to. Remember, only 62 jurisditions in the United States use RCV out of 10,499. That’s close to six-tenths of 1%. And for Mr. Dack to imply not adopting RCV calls into question Irvine’s status as a city of innovation is laughable given that RCV has been around in Europe since 1850. There’s nothing innovativeabout something 175 years old. RCV is best for selecting Oscar winners (that’s where it’s most often used — awards shows).
The strongest evidence via independent study of RCV breaks down this way:
| RCV Promise | What the research actually suggests |
| “RCV guarantees a majority winner.” | Not necessarily. Ballot exhaustion can mean the winner has a majority only of continuing ballots, not of all voters. |
| “RCV increases turnout.” | Mixed. Some evidence suggests an initial increase followed by declining turnout. |
| “RCV makes campaigns more civil.” | Questionable. Research from Maine found increased negative spending. |
| “RCV creates more competitive elections.” | Mixed. Recent research finds no long-term increase in competition in some municipal datasets. |
| “RCV eliminates wasted votes.” | Debatable. Exhausted ballots remain a significant issue, though their impact varies. |
| “RCV eliminates strategic voting.” | No. Strategic incentives remain, and outcomes can depend on rankings and elimination sequences. |
| “RCV always identifies the most broadly supported candidate.” | Not guaranteed. RCV can theoretically—and occasionally empirically—produce outcomes that differ from other measures of broad preference. |
| “RCV is simple and easy to understand.” | Contested. The process is more complex than plurality voting, and recent research specifically studies the additional complexity of polling and tabulation. |
| “RCV is a nonpartisan reform.” | Not necessarily. Research on adoption shows political coalitions can have strategic incentives for supporting electoral-system changes. |
The research itself shows that some of RCV’s most heavily promoted benefits are conditional, disputed, or unsupported in particular empirical settings.
Several cases have different findings or research themes drawn from the same underlying studies, and some are recent working papers. Here are the issues and the related study with the year published.
The sign in the newsroom at the OC Register says “Your Mother says she loves you…check it out.” That’s what I did here.
Theoretical Problems With Instant-Runoff Voting
Social-choice and voting-theory literature — ongoing research
The Graham-Squire and McCune study summarizes several known theoretical weaknesses of RCV, including the possibility of monotonicity violations, failure to elect a Condorcet winner, and the spoiler effect. Their empirical analysis suggests these problems are uncommon in their dataset, but the fact that they remain mathematically possible is relevant to arguments that RCV is inherently superior to plurality voting.
Bayesian Uncertainty Quantification for Ranked Choice Voting Polls
Evan T. R. Rosenman & Jason Liang — June 30, 2026
This very recent research highlights a problem unique to RCV: the outcome depends not merely on how much support a candidate has, but potentially on the order in which other candidates are eliminated. The authors argue that traditional polling concepts such as a simple margin of error are not straightforward to apply to RCV because the outcome can be path-dependent.
Ballot Exhaustion in Multiwinner Single Transferable Vote Elections
David McCune & E. E. Naber — May 12, 2026
This is a very recent study, although it examines multiwinner STV rather than single-winner RCV. Using more than 5.4 million ballots from 1,070 Scottish local elections, the authors found a raw ballot-exhaustion rate of 27.9%. However, they also found that many exhausted ballots had already contributed to electing a candidate or represented a voter’s preferred candidate. The study therefore cuts both ways—but importantly, it demonstrates why simply saying “RCV has no wasted votes” is too simplistic.
Theoretical Problems With Instant-Runoff Voting
Social-choice and voting-theory literature — ongoing research
The Graham-Squire and McCune study summarizes several known theoretical weaknesses of RCV, including the possibility of monotonicity violations, failure to elect a Condorcet winner, and the spoiler effect. Their empirical analysis suggests these problems are uncommon in their dataset, but the fact that they remain mathematically possible is relevant to arguments that RCV is inherently superior to plurality voting.
The “Majority Winner” Problem
Burnett & Kogan’s findings, revisited
The distinction between a majority of continuing ballots and a majority of all ballots originally cast is critical. Under RCV, a candidate can eventually receive more than 50% of the ballots still active while failing to receive 50% of the voters who originally participated. That distinction is sometimes lost in public messaging about RCV.
Turnout Effects Are Not Consistently Positive
Endersby — 2025
The Endersby analysis is particularly important here. It found an initial turnout increase following RCV adoption but a decline later. That complicates the frequent advocacy claim that RCV will bring more people into the electoral process. The research suggests the effect may be temporary rather than a permanent structural increase in participation.
Candidate Competition Does Not Necessarily Increase
Endersby — 2025
Another common RCV claim is that candidates will have to compete harder for voters because they are seeking second- and third-choice support. Endersby’s research, however, found that candidate competition did not increase over the long term in the cities examined. This directly complicates one of the strongest arguments made for RCV.
Can Ranked-Choice Voting Elect the Least Popular Candidate?
David McCune & Jennifer Wilson — February 25, 2026
This study examines whether IRV can select a candidate who is “weakest” under several different definitions, including a candidate with the most last-place votes or the lowest social utility. The authors conclude that such outcomes are generally rare, but they do occur. This is useful when addressing the broader theoretical claim that RCV necessarily identifies the candidate with the greatest overall public support.
Evaluation of RCV Elections
James W. Endersby — January 13, 2025
This is among the most comprehensive recent empirical analyses. Endersby examined election results involving more than 500 public officials in 21 cities between 2001 and 2020. The research found that turnout tends to rise when RCV is first introduced but declines later. It also found that candidate competition does not increase over the long term. The analysis of majoritarian outcomes found that elections without a first-round majority only infrequently produce a majority winner among all voters.
An Examination of Ranked Choice Voting in the United States, 2004–2022
Adam Graham-Squire & David McCune — January 28, 2023
This study examined 182 American RCV elections. The authors found that several theoretical problems with RCV—such as the possibility of failing to elect a Condorcet winner—were uncommon in their dataset. However, they found that ballot exhaustion frequently produced what they call “majoritarian failures.” This is valuable because it is not an anti-RCV polemic; it acknowledges RCV’s strengths while identifying a significant weakness.
An Empirical Analysis of the Effect of Ballot Truncation on Ranked-Choice Electoral Outcomes
Mallory Dickerson, Erin Martin & David McCune — June 9, 2023
Using a database of 1,171 real-world RCV elections, the authors studied what happens when voters cannot rank every candidate. They found that restricting voters to three or more rankings rarely changed the winner in their dataset, but the study nevertheless demonstrates that ballot design and the number of rankings permitted can affect outcomes. It is useful for examining the assumption that every voter can simply rank candidates honestly and fully without consequences.
The Effect of Ranked-Choice Voting in Maine; Campaigns are Not Always Nice
Maine / MIT Election Lab analysis — 2020
This analysis is particularly useful for challenging the claim that RCV makes campaigns more positive. The research found that negative spending increased significantly in Maine after RCV was implemented. An analysis of Facebook political advertising also found that Maine’s 2018 congressional campaign was more negative than comparable districts elsewhere. The study also found evidence that third-party candidates gained vote share under RCV—an effect that could be viewed as a benefit—but the civility claim was not supported.
RCV Does Not Automatically Make Campaigns More Civil
MIT Election Lab / Maine analysis — 2020
The Maine evidence is particularly powerful because it tests a real-world implementation rather than relying solely on theory. The finding that negative spending increased after RCV’s implementation is difficult to reconcile with sweeping claims that candidates have an incentive to avoid negative campaigning because they need to win second-choice votes.
Use it to challenge: “RCV makes politicians nicer to one another.”
RCV Does Not Eliminate Strategic Voting
Graham-Squire & McCune; Dickerson, Martin & McCune
RCV advocates often contrast ranked voting with plurality elections by suggesting voters can simply express their sincere preferences. The research literature is more complicated. RCV can still create strategic incentives, and the outcome can depend on the order of elimination and on how many candidates voters rank. The existence of ballot truncation and strategic considerations means that the idea of “just rank everyone honestly” does not eliminate every strategic problem.
Maine Ranked-Choice Voting as a Case of Electoral-System Change
John Santucci — 2018
Santucci’s analysis of Maine’s adoption of RCV offers an interesting challenge to the idea that RCV emerges solely from a neutral desire for better democracy. The study argues that RCV adoption can be driven by strategic coalitions and partisan incentives, particularly when political groups believe the new system will benefit their side. It describes RCV as a system that can “manufacture an electoral majority” in a fragmented field.
Harvard Cooperative Election Study publication page
Ballot (and Voter) “Exhaustion” Under Instant Runoff Voting Calls Into Question RCV Guaranteeing Majority Support
Craig M. Burnett & Vladimir Kogan — March 2015
This is probably the single most useful study for challenging the claim that RCV guarantees a majority winner. Examining more than 600,000 ballots from four RCV elections, the authors found ballot exhaustion rates ranging from 9.6% to 27.1%. In all four elections, the eventual winner received less than a majority of the originally cast ballots. The authors specifically conclude that ballot exhaustion challenges a central argument made by RCV proponents.
Leave a Reply