Ranked-choice voting is gaining popularity as a voting system across the US. I wrote this post to explain how ranked-choice voting works and to discuss the claimed pros and cons of ranked-choice voting, including recent empirical studies.
What is Ranked-Choice Voting?
When votes are cast in US elections, voters are usually asked to pick their singular, top choice candidate. This is referred to as “choose-one” voting. In general, the candidate who receives the most top-choice votes, or a plurality, is declared the winner. This traditional form of voting is referred to as “first-past-the-post” voting.
An alternative to first-past-the-post voting is ranked-choice voting. Under ranked-choice voting, instead of casting a single vote, voters can rank candidates from their first choice to their last choice. If a candidate does not win a majority in the first round of voting, voters who chose the last-place candidate as their first choice will have their votes transferred to their second-choice candidate, and totals are tallied again. This process of “instant runoffs” continues until a candidate receives a majority, or there is one candidate remaining.
How Prevalent is Ranked-Choice Voting in the US?
At the time of this post, ranked-choice voting is used by 23 jurisdictions outside of Utah, and 19 cities in Utah. Seven additional jurisdictions plan to use ranked-choice voting in upcoming elections. There are additionally six states which use ranked-choice voting in some form of presidential primary elections, and ranked-choice voting is used in 10 party elections/conventions.
Updated information regarding the use of ranked-choice voting can be found here.
Assume a race has four candidates: Candidates A and B are the major party candidates, while Candidates C and D are independent candidates. Further assume there are a total of 10 voters.
Suppose the first round of voting yields the following results (winning votes are highlighted green):
The two major party candidates, Candidates A and B, together claim 70% of the vote: Candidate A receives four first-choice votes, winning 40% of the vote, while Candidate B receives three first-choice votes, winning 30% of the vote. The two independent candidates, Candidates C and D, together claim 30% of the vote: Candidate C receives two first-choice votes, winning 20% of the vote, while Candidate D receives one first-choice vote, winning 10% of the vote.
Even though Candidate A has a plurality with 40% of the vote, and would therefore win in a traditional election, because no candidate can claim a majority after the first round of voting, an instant runoff will ensue, where Voter 10’s first-choice vote will be transferred from Candidate D to their second-choice candidate, Candidate B:
After the second pass, there is still no victor, as Candidate C trails both Candidates A and B, who now each receive 40% of the vote. As such, those who cast first-choice votes for Candidate C (Voters 8 and 9) will have their votes transferred to their second-choice pick (Candidate B for both voters), resulting in the following third pass results:
Candidate B can now claim a majority with 60% of the vote (three first-choice votes and three second-choice votes), and thus the election is effectively over.
There are two main points I wish to illustrate in this example, beyond the mechanics of ranked-choice voting. First, the candidate who won a plurality in the first round of voting (Candidate A) is not the ultimate winner, as Candidate B had a stronger second-choice performance than Candidate A. In essence, Voters 7, 8, 9, and 10 “blocked” Candidate A from winning by giving Candidate A lower ranks. Second, while the two independent candidates had a strong first-choice performance (together claiming 30% of the vote in the first round, much larger than we usually see in actual elections), only the two major party candidates remained in the final round of voting.
Claimed Advantages and Disadvantages of Ranked-Choice Voting
The primary argument in support of ranked-choice voting is that ranked-choice voting reduces the “spoiler effect” and therefore increases third party share. The spoiler effect occurs when a third party candidate causes a major party to lose, thereby paving the path to victory for the other major party. For example, Green Party and Libertarian Party candidates have been known to siphon off votes from major party candidates, thereby “spoiling” results. Due to fear of a spoiler effect, it is difficult for third parties to gain momentum in the US, as those who sincerely prefer a third party candidate may vote for the major party candidate with the most closely aligned platform, only to deter the other major party candidate from winning. This form of voting is referred to as “strategic voting.” Under ranked-choice voting, voters do not have to choose only one candidate, and thus voters who sincerely prefer a third party candidate can place their favorite candidate in their first-choice ranking and their least favorite candidate in their last-place ranking. This is referred to as “sincere voting” and stands in stark contrast to strategic voting.
In addition to increasing sincere voting, ranked-choice voting reformers also believe that ranked-choice voting decreases negative campaigning, as candidates are no longer battling against a single candidate, but instead are vying for second (and possibly third) choice. Further, it has been argued that ranked-choice voting will allow for a more democratic system, as a greater percentage of voters will be afforded the opportunity to vote for the ultimate victor.
Those who oppose ranked-choice voting believe that ranked-choice voting is more expensive and unnecessarily confusing, which may lead to apathy among voters. There is also an argument that “exhausted ballots,” ballots which are not counted in the final round of runoff voting because not all voting options were chosen, will lead to inaccurate or misleading results and voter disenfranchisement. However, this final complaint is often rebuked because, as experts at the Princeton Electoral Innovation Lab have explained, an exhausted ballot is no different from a traditional “choose-one” ballot in which the voter did not vote for the winning candidate.
What Does the Data Show?
I recently had the opportunity to interview Jesse Clark of the Princeton Electoral Innovation Lab. Jesse is an expert in ranked-choice voting and has completed two studies in the field. In this first study, Jesse focuses on the impact of ranked-choice voting on voter confidence, voter satisfaction, and ease of use. In his second study, Jesse primarily examines the impact of ranked-choice voting on campaign civility.
Jesse’s first study consists of an experiment followed by a survey of Mainers who participated in Maine’s ranked-choice voting. Consistent with expectations of rank-choice voter reformers, Jesse’s experiment finds that voting for non-major party candidates increased by five points when participants were presented with a ranked-choice voting ballot, as opposed to a traditional ballot. Jesse’s experiment also reveals many perceived flaws in ranked-choice voting, relative to traditional voting. Specifically, Jesse’s experiment suggests that ranked-choice voting not only results in significantly lower levels of voter confidence, voter satisfaction and ease of use, but also increases the time to vote by nearly 12 seconds per candidate.
Jesse’s survey results support the findings in his experiment regarding a drop in voter confidence when ranked-choice voting is used, relative to traditional voting. In regards to ease of use, Jesse finds that fewer surveyed Mainers found ranked-choice voting ballots difficult to fill out, relative to experiment participants. While this survey result may seem contrary to the ease of use result in his experiment, Jesse believes this discrepancy may be attributable to social desirability bias, as survey participants might be more likely to claim they did not find the ranked-choice voting ballot confusing, but instead would admit that “other people” might find it more confusing.
Jesse’s second study finds that, contrary to the belief of rank-choice voting reformers, not only did negative spending increase in Maine after the adoption of ranked-choice voting, but also the 2018 Maine congressional campaign advertisements were more negative, relative to pair districts across the country. However, this result should be interpreted with caution, as this increase in negativity was not statistically significant compared to the national average. Additionally, consistent with findings in his first study, Jesse’s second study finds that voting for non-major party candidates increased six percent under ranked-choice voting, relative to traditional voting.
The Future of Ranked-Choice Voting
The future of ranked-choice voting is uncertain. There is a lot of momentum to reform the current voting system in the US, and ranked-choice voting is lauded by many proponents of election reform. However, possibly due to the newness of ranked-choice voting, there are only a few empirical studies which examine the impacts of ranked-choice voting, and the studies which have been conducted yield mixed results.
The primary benefit of ranked-choice voting, which is supported by both of Jesse’s studies, is that ranked-choice voting increases sincere voting, related to traditional voting. However, as illustrated in my numeric example, even with a five or six point increase in voting for non-major party candidates, it is still a challenge for a third party candidate to be the last candidate standing under ranked-choice voting. If the goal of ranked-choice voting reformers is to give third party candidates a larger share, ranked-choice voting may help, but it might be only one step in a larger process towards achieving a truly representative voting system.