Consequences for Policy and Politics

The boldest claim that RCV advocates make is that it can reduce partisan polarization and help to make American democracy more compromise-oriented. This is the most downstream consequence, and thus the hardest to measure in the United States, given the extremely limited use of RCV as of this writing. Certainly, there are good theoretical reasons to expect RCV should have a moderating effect, and the experiences of other democracies that use RCV (most notably Australia) offers strong support for the moderating, compromise-inducing influence of RCV.¹

But so far, in the United States, we have very little data to back up any claims that RCV makes government work better, but also none that indicates it makes matters worse.

Conclusion: Unclear, hard to assess

Here, we review the small but growing U.S.-based literature examining the impacts of RCV on different types of polarization: interparty and affective polarization and racially polarized voting. In some ways, the research hints at RCV’s potential for mitigating interparty divides, but otherwise are more suggestive of the limitations of voting reform inside a two-party system.

Cross-partisan and third party voting has become increasingly rare in America due to high levels of partisan polarization. Lindsay Nielson sought to test (among other things) whether RCV might elicit more cross-party voting or reveal a more nuanced set of preferences which voters are not prompted to access or express under plurality rules. In the RCV treatment group, 20-25 percent of partisans ranked at least one candidate from the opposite political party among their top preferences, and 5 percent of partisan respondents exclusively ranked candidates of other parties. In plurality elections, Nielsen notes, about 90 percent of Democrats and Republicans vote for a candidate within their own party. Results indicate that people will take advantage of the opportunity to express more complex preferences and even their ambivalence toward the two parties.²

Relatedly, as cited earlier, the article "Ranked-Choice Voting, Runoff, and Democracy: Insights from Maine and Other U.S. States" reported that in competitive congressional races in 2020, the most ideologically moderate candidate was a Republican, Susan Collins, from Maine, who happened to have run under RCV but most likely would have run as moderate regardless. The report also found third party and independent candidates fared better (but still poorly) under RCV than under runoff or plurality. Again, these findings only gesture at RCV’s potential to break us out of the binary, us-versus-them thinking that single-mark voting tends to encourage.

Regarding affective polarization, or interparty animosity, there are good reasons to suspect that RCV and other preferential voting systems would lead to less hostility between the parties compared to plurality systems. While winner-take-all encourages a zero-sum, obstructionist approach to politics, RCV and other preferential systems should create a “positive-sum” electoral environment that incentivizes cooperation and bargaining among rivals—behaviors which theoretically carry over from campaigning into governing.

Observational data from the international literature provides modest evidence of a causal link between electoral systems and affective polarization, but typically those links are mediated by effects, such as RCV’s civility effect, which has inspired many scholars to recommend RCV for divided societies.³ The only U.S.-based study to date that looks at the direct relationship of electoral rules and affective polarization is "Electoral Systems and Political Attitudes: Experimental Evidence." Researchers Sean Fischer, Amber Hye-Yon Lee, and Yphtach Lelkes ran a modified dictator game designed to measure the level of interparty animosity under plurality, RCV, and proportional representation, as well as with different numbers of parties. They found, surprisingly, there was more in-group bias (their proxy for interparty animosity) in RCV and proportional systems than plurality systems. In other words, plurality systems yielded the lowest level of interparty animosity. However, critically, increasing the number of parties (especially to four parties) decreased in-group bias for RCV and proportional but not plurality. They concluded, accordingly, that “electoral reform that does not lead to a change in the number of parties in a system may make interparty animosity worse.”⁴

Racially polarized voting is a different type of polarization that has been studied primarily in the Bay Area. In theory, RCV should reduce voting strictly along racial and ethnic lines, because a candidate in a multi-racial, multi-ethnic jurisdiction must appeal not only to voters from her own racial or ethnic group but also campaign for the backup preferences of other groups. Benjamin Reilly’s 2018 article "Centripetalism and Electoral Moderation in Established Democracies," lends empirical support to this racial moderation hypothesis. Reilly’s historical examination of Australian elections under RCV found that as the country has become more socially diverse and preference voting more widely used, ethnic voting has declined.⁵ In contrast, the same candidate running under the traditional plurality, single-mark system is more likely to be able to win on the support of members of her own group alone.

But the evidence from the United States is less clear. Several articles, including "Does More Choice Lead to Reduced Racially Polarized Voting? Assessing the Impact of Ranked-Choice Voting in Mayoral Elections," by Jason McDaniel, have advanced a competing theory, the racial competition hypothesis, which posits that RCV voters will not only continue to be guided by racial group identity cues when choosing to express expanded preferences, but that based on the increased complexity of RCV ballots, people will default to racial group identity cues to inform their vote choice.⁶ McDaniel’s analysis compared racial group candidate vote-share in mayoral elections in Oakland and San Francisco before and after those cities adopted RCV.⁷ The former analysis suggested that racially polarized voting in San Francisco declined significantly after the implementation of RCV. But his analysis also revealed that the level of racially polarized voting decreased by even more in the non-RCV cities in the same time period. Ultimately, he concluded that racially polarized voting did not decrease significantly as a result of the implementation of RCV, and actually led to statistically significant increases in the racial vote division between whites and Asians. However, it is unclear how much of this was driven by candidate entry, as opposed to RCV.

A recent study by Yuki Atsusaka and Theodore Landsman attempted to extend the literature by using all available information on individual ranked ballots, rather than rely on aggregated election data or voters’ first choices alone, as previous research had done.⁸ Ultimately they concluded that switching from first-past-the-post to ranked-choice voting neither mitigated nor increased the degree of racial polarization among Bay Area voters. They also found circumstantial evidence of “plumping” (when voters only rank one or two candidates in an attempt to concentrate their voting power to their preferred candidates), something future researchers may want to investigate.

As usual, the fact that the data available is limited to nonpartisan, low-information local elections strains generalizability. Racially polarized voting might look completely different in the context of a statewide or national election where voters are more likely to be informed about the candidates and use partisan or policy cues. It might also look different if we combine RCV ballots with increases in district magnitude or the size of the legislature or council.⁹

Finally, while the themes of moderation and sincere voting run through much of the RCV literature, there’s little examination of the ways in which the reform may alter our relationship to ideologically extreme candidates—do we see them as more or less electable, for instance? Our highly polarized political climate has for some time rewarded candidates who are more extreme than their own voter bases, and it’s not unreasonable to assume that extreme candidates would continue to emerge even if we ditched plurality for RCV. In a 2021 experimental study based on social comparison theory, Melissa Baker conducted an experimental study of voters’ responses to ideologically extreme candidates under RCV and plurality rules.¹⁰ Baker observed that candidates, regardless of ideology or ideological extremity, were seen as more representative of the electorate in RCV elections compared to plurality. However, extreme conservative candidates were seen as slightly more electable in elections that used plurality voting compared to RCV. Liberal respondents viewed extreme liberal candidates as less electable in both voting systems, but especially in plurality elections. This result suggests that extreme liberal candidates may be slightly less likely to be punished for ideological extremity by liberals in RCV elections. Future research on this topic could extend this kind of analysis to real-world elections to see if these electability perceptions, and ideological asymmetries, still apply. If future work reinforces the finding that voters perceive candidates as more representative (or electable) in elections that use RCV, scholars might consider how that might impact the odds of women and minority women candidates winning election, in addition to or perhaps instead of changes in campaign environment.

Conclusion: Not supported, very hard to assess

Could RCV ultimately change policy outcomes? Evidence from other countries using preferential systems suggest some positive effects on governing, but overall the effects are still difficult to measure, given the wide range of factors that determine policy.¹¹ In the United States, the limited literature on electoral systems and policy representation contains no real evidence that different electoral rules change responsiveness to public opinion in local government.¹²

A recent study by Arjun Vishwanath was the first direct analysis of whether RCV affects substantive representation in the United States.¹³ With a sample of nine RCV cities, he evaluated changes in key fiscal and ideological variables following the switch to RCV.¹⁴ His empirical findings suggest that there is no significant effect of the reform. Of course, there are other variables to examine, and this study, like so many, was limited to a set of fairly progressive cities. As additional jurisdictions, including states, adopt RCV, more research could help evaluate the reform’s impacts on policy and governance outcomes.

For example, one could collect and analyze interest group endorsements for RCV winners and compare them against the winners’ voting records. One could also conduct surveys and interviews of officials who were elected under both plurality and RCV to understand if/how their approaches to governing or incentives to prioritize (or avoid) certain policies may have changed. Did they feel they could be more productive, or more policy-oriented? Did a councilor or mayor elected under RCV consult more with interest groups that weren’t previously part of their coalition, but perhaps gave her their second- or third-choice votes, or pledged them for the next election? Another possibility is to extend the Vishwanath study to matched plurality cities over the same period of time, for comparison. As polarization has deepened in the United States and politics has become more nationalized, it’s not unreasonable to expect that substantive representation in non-RCV jurisdictions may have worsened. Reflecting on the 2016 study of women candidature in the Bay Area discussed earlier, Terrell et al. noted that RCV cities were less affected than their non-RCV counterparts by a regional downward trend in women running for local office, suggesting that RCV may, in certain respects, be less a force for good than a buffer against broader undesirable trends in the political culture.¹⁵

Additionally, a recent survey experiment conducted in San Francisco may shed some light on the lack of observed effect in Vishwanath’s study. In "Ranked-Choice Voting and Political Expression: How Voting Aids Narrow the Gap between Informed and Uninformed Citizens," Cheryl Boudreau, Jonathan Colner, and Scott MacKenzie evaluated whether citizens’ rankings of candidates in low-profile local elections faithfully reflect their own policy views.¹⁶ They found that overall, respondent rankings are modestly related to their policy views (weaker for respondents with low political knowledge), but that there were no clear differences in agreement between candidates ranked first and those ranked second and third. The results suggest, among other things, that uninformed citizens in particular struggle to distinguish between the candidates’ policy views in local RCV elections (where most RCV elections now take place). This lowers their propensity to use their rankings, but also, to the point of policy representation, deprives those who are elected under this system of the insight into what average citizens care about, a benefit that RCV should theoretically provide.

Certainly, meaningful policy changes are somewhat downstream from electoral changes, and are always hard to prove, given the long causal chain. The theory is that changing electoral rules will lead to different types of candidates running, with different political incentives, and different bargaining structures, which ultimately translate into policy changes. At this point, it is almost certainly more productive to examine these micro-level changes in the causal chain, since there are fewer confounding variables in doing so, and the limited use of RCV thus far makes inferences about policy impacts difficult.