Party of Five (or More): Ranking Limit Recommendations for Ranked Choice Voting Elections

Introduction

In January 2024, FairVote updated its model ranked choice voting (RCV) legislation. The update includes a new recommendation – that voters should be able to rank at least as many candidates as there are seats being elected, plus an additional four candidates. This report explains why “N+4” is our recommendation for policymakers, as well as the general implications of ranking limits. 

N+4 allows voters to express a sufficient number of preferences while keeping ballots short and easy to navigate. When N+4 is “tested” on real-world multi-winner RCV ballots from Cambridge, MA, nearly all voters are able to help elect a candidate of their choice. 

Naturally, this recommendation applies to races in which at least N+4 candidates appear on the ballot. Races with fewer candidates will not encounter the question of ranking limits.

Figure 1

Number of winners (N)	1	2	3	4	5	6	7
Suggested minimal ranking limit (N + 4)	5	6	7	8	9	10	11

Ranking limits require trade-offs for policymakers

Choosing a ranking limit means finding a balance between expression and simplicity. Voters should be allowed sufficient rankings to express a broad range of preferences and avoid their ballot becoming inactive due to the ranking limit. At the same time, imposing some limit on the number of rankings can keep ballots shorter. make voting simpler for voters and election administrators alike.

Research shows that more ballots become inactive in contests where voters are unable to rank all of the candidates. The number of inactive ballots drops significantly as more rankings are added. For example, there are significantly more inactive ballots in contests limited to three rankings than in contests limited to five. Ideally, a voter’s ballot should only become inactive if they choose not to rank additional candidates. The only way to guarantee this is to give voters as many rankings as there are candidates.

However, there may be diminishing returns to increasing the number of rankings. A ballot with dozens of candidates and unlimited rankings could appear long and cluttered. According to the Center for Civic Design, when jurisdictions limit ballots to five rankings, two grid-style contests can fit side-by-side on a ballot. In a single-winner race, voters who are able to rank five candidates will have room to rank a frontrunner, and have their vote count in the final round even if that frontrunner isn’t the voter’s first, or even fourth, choice.
For this reason, FairVote previously recommended that jurisdictions allow at least five rankings. However, the rise in popularity of proportional RCV creates the need to update that recommendation for multi-winner races.

Considerations for ranking limits in multi-winner RCV races

Multi-winner contests tend to invite more candidates than single-winner contests, so a ranking limit of five may not allow voters to express a wide enough range of preferences in a crowded field. 

Figure 2 shows that the more winners there are per race, the more candidates tend to run. The analysis is based on cast vote records data from recent contests in Cambridge (MA), Eastpointe (MI), and Minneapolis (MN). 

Figure 2

A ranking limit of two or three times the number of seats would likely allow voters to rank almost all of the candidates. However, in places with a large number of seats, ballots could become very long. For example, Cambridge, MA elects nine candidates at-large to its city council, leading to a ballot with 18-27 rankings. Therefore, a blanket “two or three times the number of seats” ranking limit doesn’t work for all contests. 

Figure 3 shows the full distribution of ranking depths (i.e. how many candidates each voter ranked) for a sample of recent multi-winner RCV races. The boxes represent the middle 50% of the distribution, meaning the number of candidates the “middle half” of voters typically ranked.

Figure 3

Cambridge City Council elections elect more candidates and allow more rankings than any other RCV races in the U.S. Therefore, these races show us how voters rank when they are relatively uninhibited. In both the 2019 and 2021 Cambridge City Council races that elected nine winners, 75% of voters used nine rankings or fewer. In the 2023 Cambridge City Council race, 75% of voters used 10 rankings or fewer – well within FairVote’s N+4 recommendation (which in this case would be 13 rankings allowed).

In the 2019, 2021, and 2023 Cambridge School Board races for six seats, 75% of voters used six rankings or fewer, also well within the N+4 recommendation (which would be ten rankings allowed). 

Testing the N+4 recommendation on real-world ballots

As Figure 3 shows, some voters will rank more than the number of seats to be filled when given the opportunity. However, few voters use more than four rankings beyond the number of seats (N+4). Our analysis finds that even those who use more than N+4 rankings do not need those extra rankings in order for their ballot to count in a meaningful way. In a multi-winner RCV contest, as long as a voter ranked a winning candidate somewhere on the ballot, they helped elect a candidate of their choice. 

To evaluate our recommendation, we asked: If the ranking limit was N+4, would fewer voters have helped elect a candidate of their choice? For example, in a Cambridge City Council race with nine winners, a ranking limit of N+4 would have limited voters to 13 choices, so a voter would have had to rank at least one winning candidate 1st through 13th to have helped elect a candidate. A voter would have been impacted by the N+4 policy if they ranked a winning candidate, but not until their 14th or 15th ranking. (Remember, if a voter ranked a winner 13th or better, they would not have been meaningfully impacted by the N+4 ranking limit.)

Figure 4 suggests that very few, if any, voters would have been significantly impacted in any recent Cambridge election. An N+4 ranking limit would continue to allow virtually every voter in Cambridge to elect a candidate of their choice, demonstrating a suitable standard for other jurisdictions.

Figure 4

Conclusion

Nationally, an N+4 ranking limit can keep ballots short and clean, while still fulfilling the promise of RCV. FairVote will continue to analyze ballot data in the growing number of jurisdictions that use RCV to refine our understanding of how voters engage with the ranked ballot, and issue updated recommendations when appropriate.