1 Introduction

American democracy faces a wide array of challenges. The aspiration of majority rule — central to the goal of representative democracy — is itself in question. Of the five presidential elections that resulted in an Electoral College victory for a popular vote loser, two have occurred in the twenty-first century. The challenges to representative outcomes extend well beyond presidential elections. The Supreme Court’s 2013 Shelby County decision weakened protection of voting rights. The single-member district regime used to elect the U.S. House of Representatives has enabled partisan gerrymandering that, combined with greater partisan sorting, skews the relationship between seats and votes (Drutman, 2020). Amid these challenges, instant runoff voting (IRV) has emerged as a tool to provide more representative outcomes and is gaining significant traction among policymakers and voters. Two American states and over 50 local jurisdictions have adopted or are piloting IRV, with adoptions accelerating since 2016 (FairVote, 2022a). This essay provides an overview of IRV and its benefits, describes IRV’s advantages over other single-winner voting methods, addresses counterarguments, and anticipates future developments in IRV advocacy.

2 Overview of Instant Runoff Voting and why it has gained Momentum

Commonly known as ranked choice voting in the United States and preferential voting by Australia and many NGOs, IRV provides voters with the option to rank multiple candidates in order of preference. That simple change allows voters to express the full scope of their political opinions, rewards candidates for engaging with more voters, and ensures winners do not win over the objections of the majority. Tallies simulate a series of runoff elections, with voters’ first choices counted as one vote each in the first round. If no candidate wins outright with a majority of valid votes cast, then the candidate with the fewest votes is eliminated and their votes are added to the totals of the candidates ranked next on those ballots. This process of eliminating last-place candidates and re-allocating their votes to the next backup choice continues until the field is reduced to the final two candidates: the one with a majority wins the election.

IRV has a long history in the United States and abroad (FairVote, 2022a). Australia and the Republic of Ireland have both used IRV for national elections for more than a century, while Scotland and New Zealand use it for some local and city elections. As of September 2022, a total of 43 U.S. jurisdictions have used IRV in their most recent elections, with at least eight more scheduled to implement it in their upcoming elections. In 2020, Maine extended its use of IRV for all congressional and state primary elections (adopted in 2016 and confirmed in 2018) to presidential elections. Alaska started using IRV for all federal and state elections in 2022. Maine parties each nominate candidates for state and congressional elections with IRV, followed by a general election in which congressional and presidential elections use IRV (adding IRV to general elections for state offices depends on a state constitutional change). In contrast, Alaska has a “top four primary” for state and congressional primaries, where candidates participate in a single primary election from which the top four vote-getters advance to a general election run with IRV.

IRV has gained traction in the U.S. in many contexts. Democratic parties in four states replaced their caucus systems with IRV for presidential primaries (FairVote, 2022a). Voters approved IRV in all 13 municipal ballot measures in 2019–2021, including in New York City by 73–27% (Ballotpedia, n.d.), with nine more voting on it in November 2022 (FairVote Action, 2022a). In Utah, 23 city councils in 2021 voted to use IRV for mayoral and council races. In 2021-22, six states passed pro-IRV laws (FairVote Action, 2022b). More than 90 American colleges use IRV for student elections (FairVote, 2022a). The Boston Globe, Economist, New York Times and Washington Post have endorsed IRV in editorials (FairVote, 2022b).

For years, barriers in election administration hindered IRV’s growth. It was only in 1997 that Cambridge (MA) became the first U.S. jurisdiction to use modern ballot-scanning voting machines to count IRV ballots (FairVote, 1997). Before machine tallies were available, adoption of IRV required hand tallies, which raised concerns about the time and expense it would take to produce a winner. Significant opportunities to win IRV in states were hampered by such election administration concerns and costs, including in New Mexico (where the state senate backed IRV for congressional elections in 1999), Alaska (where a 2002 statewide ballot measure lost amidst concerts about costs), and Vermont (where the governor in 2008 vetoed a bill to use IRV for congressional elections). Today, however, all new voting equipment can run IRV, requiring at most a straightforward software upgrade (Ranked Choice Voting Resource Center, 2021).

Several developments have helped advocates overcome these barriers. First, advancements in election technology have removed barriers to using IRV. Second, the Ranked Choice Voting Resource Center has formed to address election officials’ questions. Third, new rules for absentee and early voting have made runoff elections costlier and harder to administer, which makes the option of IRV all the more attractive.

National trends have also had a major impact. Attention and funding has grown in tandem with concerns about polarization in government, where the defects in plurality voting that exacerbate that polarization are garnering attention. Crowded presidential primaries in contests without an incumbent can advantage outsiders who, when able to win states with low pluralities, can drum up support within a narrow base. Partisan polarization has resulted in more one-party states and districts, which has generated interest in IRV in two ways. First, the rising number of “safe” seats in Congress puts all the more focus on the primary of a district’s majority, resulting in crowded fields and winners who can effectively take office by winning a primary with less than a fifth of the vote.Footnote 1 Second, outcomes in the decreasing number of districts and states with more balanced partisanship can be flipped by the arrival of a third-party candidate; with higher stakes than ever, campaign operatives are turning to recruit or promote third-party candidates for the express purpose of siphoning votes away from the opposition party.Footnote 2 With IRV, third-party candidates need not worry about becoming spoilers and can instead share their ideas in a more receptive forum and contribute meaningfully to the discourse of a campaign.

3 Benefits of IRV

IRV offers its own benefits as well. By ensuring that each voter is able to express the full scope of their political opinions, IRV provides more representative outcomes in elections. Under plurality rules, candidates can and often do win without majority support from their constituency. Two candidates with similar ideologies running against a third with opposite ideology may split the vote of their base between them, leading to a victory for the third candidate when a majority of the electorate opposes their policies. Such outcomes subvert the purpose of democracy and weaken the efficacy of government.

IRV’s tallying algorithm greatly mitigates vote-splitting. As a result, voters may indicate support for multiple candidates and trust that their vote will further whichever of their preferences is most viable. This has been shown to level the playing field for female candidates and candidates of color, who are often told to stay out of single-choice races should another woman or person of color happen to be running (RepresentWomen, 2020; Otis & Dell, 2021). As of September 2021, women held nearly half of all local election seats elected by IRV, and in November 2021, more than doubled their seats on the New York City council to a majority of seats after many crowded primaries held with IRV (RepresentWomen, 2020; FairVote Staff, 2021).

IRV shifts priorities for candidates in a positive way, incentivizing greater engagement between candidates and voters and greater cooperation among like-minded candidates. Because IRV elections often proceed through multiple rounds of counting and require some combination of first choices and backup rankings to win, it is advantageous for candidates to reach out to communicate with and attempt to meet the needs of as many voters within an electorate as they can. The resulting jump in voter engagement under IRV was documented in a 2013 survey, which found that voters in seven IRV cities were almost 10% more likely to report being contacted by a candidate or campaign than those in comparable non-IRV cities and that voters in IRV cities were more satisfied with the conduct of local campaigns (Smith, 2016).

Winners of IRV elections have been shown to enjoy a wider breadth of support. FairVote (2022d) ascertained the “consensus value” of winning candidates under IRV by calculating the percentage of ballots in their elections that ranked them first, second, or third — a particularly meaningful measure when rankings are voluntary. The report found that more than half of IRV winners earned a consensus value over 75% (FairVote, 2022d). These winners not only better reflect the will of their electorate; they also have a stronger mandate to govern.

Because IRV allows voters to register their opinions in greater detail and gain greater influence on election outcomes, voters therefore have greater incentives to learn about candidates. Smith (2016) finds that voters in IRV cities were more likely to report discussing candidates or policies with family members, friends, coworkers, or other citizens than voters in non-IRV cities. These numbers show IRV encouraging political discussion and debate. IRV also has specific merits in different contexts that will help expand its use, such as in nominating presidential candidates (Richie et al., 2021) and for absentee voters, with five states already having their overseas voters cast IRV ballots in congressional runoff elections (Ranked Choice Voting Resource Center, 2020).

4 How IRV compares to Alternatives

IRV compares favorably with other proposed reforms of plurality voting, including two-round runoff systems, approval voting, score voting, and Condorcet methods. A critical general point is that, due to IRV’s relative immunity to efforts to game the system, voters will reflect their preferences genuinely when voting with IRV ballots. That is less true of other alternatives, as we will discuss here.

Compared to two-round runoffs, IRV generally produces higher, more equitable turnout, involves lower costs and demands on voters, and is more likely to elect the Condorcet candidate (the one able to defeat all others head-to-head). Runoffs operate by conducting a second election between the top two finishers in the first round, typically if no candidate wins a majority. Used widely around the world for presidential elections, two-round runoffs are the most widely used alternative to plurality elections in the United States (Lewis, Richie, & Santucci, 2006). Georgia, Mississippi and Louisiana require runoff elections for major statewide general elections when no candidate receives a majority of the vote (Ballotpedia, n.d.) while nine states have primary runoffs, usually when no candidate wins a majority of the votes cast in their party’s primary.

Alvidrez, Otis, and Bartola (2020) find that “turnout declined between the primary and the runoff in 240 of the 248 regularly scheduled primary runoffs in the U.S House and U.S. Senate from 1994 to 2020” and that “the average decline in turnout was 38% and the median decline was 37%” (p. 3). The composition of turnout decline in runoffs can also raise equity concerns. Alvidrez et al. (2020) find that, from primary to runoff, “turnout among people of color decreased by an average of 43.5% in 2020, compared to a 38.3% decrease in white turnout” (p. 5). While some ballots will inevitably become inactive in an IRV contest due to some voters abstaining in the choice between the finalists, this form of “dropoff” is generally much greater in runoff systems. Comparisons show that 41% of IRV races between 2004 and 2020 resulted in a winner with a final vote total that was greater than 50% of the first round vote, while only 13% of federal primary runoffs between 1994 and 2020 did the same (FairVote, 2020).

Runoffs also do not reliably advance the strongest candidates to the second round. In crowded fields, candidates can advance with a low plurality of votes. Internationally, a famous example was the French presidential election in 2002, when extremist Jean-Marie Le Pen earned a runoff spot over the prime minister Lionel Jospin with only 16.9% of the vote (McNicoll, 2017). In top-two primary elections, the minority party has sometimes advanced both its candidates to the general elections, such as in California’s 31st congressional district in 2012 (Ballotpedia, n.d.) and Washington’s state treasurer race in 2016 (FairVote, 2022c).

In contrast, IRV can elect a strong candidate who, due to split votes, is in third place in the first round. Two such results occurred in San Francisco Board of Supervisors races, both electing women of color: Malia Cohen in 2010 (San Francisco Department of Elections, 2010), and Myrna Melgar in 2020 (Ballotpedia, n.d.). Cohen and Melgar were both Condorcet candidates (Otis & Dell, 2021), yet if those elections had been two-round runoff elections and voters’ preferences remained consistent, they would not have advanced to the runoff.

Beyond preserving higher turnout and advancing the strongest candidates to the final round, IRV addresses a number of practical problems associated with runoffs. General election runoffs can lead to members of Congress assuming office after the start of a new session, creating a disadvantage for that state in terms of seniority and generating uncertainty, as with Georgia’s U.S. Senators elected in runoffs in January 2021. A related approach to runoffs is the “top-two primary,” as used for major elections in California and Washington (Ballotpedia, n.d.). This approach produces a decisive result on the general election ballot, but at the expense of allowing lower-turnout primaries to keep important voices off of that higher turnout November ballot. With IRV, there can be more than two candidates in the general election.

Runoff elections and IRV do share an important property: supporting a compromise choice will not hurt one’s favorite candidate. That is not true of cardinal voting systems where voters’ preferences count for more than one candidate at a time. One such proposal is approval voting, which allows voters to vote for or “approve” as many candidates as they choose. This method was recently adopted for one-round elections in Fargo, ND and in the first round of a two-round runoff system in St. Louis, MO. Approval voting is inferior to IRV across three key values: promoting majority rule, requiring a minimum level of core support, and, of the greatest practical importance, rewarding sincere voters.

In a three-candidate race with approval voting, a candidate with zero first choice support could defeat a candidate backed by 51% of voters as a first choice. To illustrate how approval voting violates majority rule, consider what theoretically could happen in a simple election with only two candidates and with 100 voters who like both candidates, but with 99 of whom preferring the first candidate to the second. If those 99 voters all honestly approve of both candidates, the 100th voter can elect the second candidate by tactically withholding approval for the first candidate. As a result, the second candidate would win by 100 votes to 99, even though the first candidate would have won by 99 to 1 in a plurality election.

While exaggerated, this example shows how voting sincerely with approval voting can count against a voter’s first choice. Nagel (2007) refers to the resulting dynamic as “a game of chicken.” In practice, voters using approval voting will tend to “bullet vote,” (cast only one vote), thereby reducing the system to an approximation of plurality voting that particularly disadvantages voters who do not understand how following the ballot instructions can hurt them.

Approval voting has shown these flaws in practice, including when used by St. Louis in 2021 for the first round of a runoff election (FairVote, 2022c). In the preliminary election for St. Louis mayor, voters indicated approval for an average of only 1.6 candidates to determine the two winners who would advance, low for a two-winner election. In city council elections, voters approved an average of only 1.2 candidates. As a reflection of erratic outcomes, three of the seven runoff winners had trailed in the approval voting primary; in one ward, the first round leader lost despite leading the approval vote by 69–46%.

FairVote’s (2022c) analysis also shows that experiments with approval voting in private organizations have gone poorly. Dartmouth University used and repealed approval voting for both Alumni Association trustee elections (repealed by alumni in 2009 by 82–18%) and for student elections (repealed in 2017) after patterns of non-majority winners and strategic bullet-voting. The largest association to try approval voting, the Institute of Electrical and Electronics Engineers (IEEE), repealed approval voting and reported that more than 80% of voters had been voting for only one candidate.

With another proposed cardinal system called range voting (also known as score voting), voters rate each candidate on a scale, and the candidate with the most points wins. Range voting has almost never been used, because a tactical fringe can overrule a vast majority of sincere voters. Voters may also score candidates at different ranges based on subjective interpretations of those ranges. On a scale of 0 to 5, voters with the same views of a mediocre candidate may give them a ‘1’ or a ‘3.’ A variation of range voting called “STAR Voting” uses such a point system both to advance candidates and then to compare two finalists head-to-head; it, too, can result in a voter’s support for a compromise choice defeating their first choice and can deny victory to a candidate rated highest on more than half of ballots (FairVote, 2022c).

A worthier alternative is Condorcet voting, or “round-robin,” in which voters rank candidates in order of choice, and each candidate is compared with every other in terms of how many voters rank one ahead of the other. If there is a candidate who beats all others in these comparisons, he or she is the winner. If not, a contingent way of selecting a winner is used. Condorcet voting has rarely been tried. One challenge to its viability is the possibility of a candidate winning without being a single voter’s first choice. By putting such heavy emphasis on breadth of support, Condorcet systems may encourage candidates to aspire to merely be the “least offensive.” IRV shares many of the virtues of Condorcet methods while addressing the challenges that inhibit Condorcet’s political viability. In practical terms, IRV also nearly always elects the Condorcet winner. Of the 375 single-winner IRV elections in the United States between 2004 and June 2021 in which FairVote had sufficient ballot data to assess whether the Condorcet winner won, 374 elections were won by the Condorcet winner (FairVote, n.d.).Footnote 3 The exception was in the mayoral election in Burlington (VT) in 2009. In that election, IRV prevented election of the Condorcet loser who led in first choices, produced the same outcome as likely would have occurred under a runoff system, and avoided election of a candidate seen by many voters as weak. Unlike Condorcet voting, candidates under IRV must both have broad appeal and enough first choice support to make it into additional rounds of counting.

While IRV is preferable to many alternative methods, it is a natural complement to the single transferable vote (STV) for legislative elections. STV allows for a candidate-driven form of proportional representation while building on the American tradition of voting for candidates directly, rather than a party list. With STV, legislative districts would have multiple seats (perhaps three to five), voters would rank candidates in order of preference just like with IRV, and ballots would be transferred until all seats have been filled in proportion to candidates’ level of support. While the STV tabulation is slightly more complex, the experience for voters is as straightforward as with IRV: rank candidates in order of preference. Notably, IRV and STV are the featured reforms in the Fair Representation Act (H.R.3863) introduced in Congress in 2021.

STV is often attractive to both scholars and reformers (Drutman, 2020; Bright Line Watch, 2021) for several reasons. First, it could tackle the U.S.’s governance and polarization crises by creating space for more than two intensely polarized factions. The two major parties could be “broad tent parties,” with rural Democrats and urban Republicans winning many more seats in the legislature than they do currently, thereby creating new incentives for increased cooperation across party-lines. Independent and third party candidates would also have a chance to compete under fair rules without being unfairly maligned as “spoilers.” Second, STV would effectively end partisan gerrymandering; when candidates are elected proportionally within multi-member districts, it is much harder to pack or crack a party’s voters into a district to the extent that it denies them fair representation. The flipside of this effect is a third major benefit: all voters would receive fairer representation than under the current winner-take-all system. STV would enable representation for the political left, center, and right in proportion to their numbers, and racial and ethnic groups would have the power to elect candidates of choice even if they are not the majority in a given district.

As of September 2022, STV is being used in the United States in six localities: Minneapolis, MN; Cambridge, MA; Eastpointe, MI; Albany, CA; Palm Desert, CA; and Arden, DE. While IRV is beneficial in and of itself, its use can also increase the potential for adoption of STV by normalizing ranked ballots for voters and election administrators.

5 Addressing counter-arguments

No voting method is perfect. Any proposed major change to a well-established system will likely face skepticism, and IRV has been no exception. The most common argument against IRV is that it is too complicated for voters to use. While voters have navigated well-designed IRV ballots with ease in both the U.S. and abroad, it is true that ballots allowing the ranking of multiple candidates can appear complex and election results will look different. Combined with transitional hurdles to fast and efficient results, the difference in ballots and results presentation can make IRV seem exotic. Voters who truly dislike or feel confused by IRV can choose to rank only one candidate as usual. But most voters choose to rank. As of February 2021, the median portion of American voters who ranked multiple candidates in IRV races was 68%, with that share rising when the election received more attention (FairVote, 2022d). In the 2021 New York City Democratic mayoral primary, 95% of poll respondents found their ballot simple to complete (Common Cause NY and Rank the Vote NYC, 2021). Similar trends emerged after the first uses of IRV for elections in Maine.Footnote 4 In general, when offered the chance to register more of their political opinions, most voters have embraced that option and accurately ranked candidates.

Some critics claim that IRV will depress turnout. The argument is that because ranking multiple candidates is harder than marking only one, fewer people will be willing to vote. In practice, the opposite is often true. Turnout is consistently higher in Malta and Ireland, both of which have used IRV ballots for decades, than in the United States (International Institute for Democracy and Election Assistance, 2020). Evidence in the U.S. thus far seems to suggest that IRV has at least a neutral effect on turnout, although the introduction of IRV coincided with a 9.6% point jump in turnout in the Minneapolis-St. Paul area (McGinn, 2020), and fewer voters skip over IRV contests when held down ballot from more prominent elections (FairVote 2021f). The fact that IRV makes each voter’s ballot more powerful by ensuring that it will most likely influence the outcome offers the potential to drive more voters to the polls over time.

IRV has also been shown to be accessible and equitable for voters of various races and ethnicities. Historically, ballot error has tended to be higher among Black American voters than voters of other racial groups, and there is some concern that IRV’s heightened complexity will exacerbate that disparity. However, Neely et al. (2015) find that while rates of overvoting (marking more candidates than is permitted) were higher in precincts with more Black voters, this difference is roughly equal in IRV and non-IRV elections. Furthermore, Neely et al. (2008) found that Black, Latino, and Asian voters in San Francisco were more likely to fully complete their IRV ballots than voters of other racial groups, and Otis and Dell (2021) echoed these findings when analyzing 2020 elections across four IRV cities. High levels of support for and understanding of IRV exist across racial and ethnic groups, and voters of all identities have been able to make use of the IRV ballot and the advantages it offers.

Advancements in election technology have made adopting and administering IRV elections easier than ever for U.S. jurisdictions and allow election officials to run smooth, auditable elections. Risk-limiting audits were used in major IRV elections in 2019 and 2020 (Ranked Choice Voting Resource Center, 2020a). All three of the largest voting equipment vendors (Election Systems & Software, Dominion Voting Systems, and Hart InterCivic) have built capacity for ranked ballots into their systems, which can typically be accessed with a one-time software upgrade (Ranked Choice Voting Resource Center, 2020b). While this upgrade may present a cost, jurisdictions that implement IRV to replace runoff elections will actually save money by eliminating an additional round of voting.

6 Conclusion

The IRV movement has reached an inflection point, with a critical mass of jurisdictions using ranked ballots. U.S. voting machines can handle IRV, and data show that the vast majority of voters are capable of and enjoy ranking their preferences. Looking forward, we anticipate several new milestones: 10 states could incorporate IRV ballots into their 2024 presidential primaries, Congress could incentivize new IRV adoptions and use-cases, and more states and cities could enact IRV. The benefits of IRV—empowering voters to indicate honest preferences, promoting majority winners and voter engagement, and incentivizing civility—are meeting a moment in which U.S. politics is succumbing to polarization, gridlock, and even political violence. IRV will not fix all these problems on its own and—as true of all electoral systems—is imperfect, but it helps move the country toward “a more perfect union” and lays the groundwork for more comprehensive reform.