In approval voting, voters may vote for as many candidates as they choose. The candidate with the highest number of votes, or “approvals,” wins.
While easy to execute, approval voting has rarely been used and is not feasible in the many states that invalidate ballots with “overvotes” (those where voters voted for more candidates than winners) because it is impossible to identify and correct voter error when voters may vote for as many candidates as they choose.
St. Louis tried a “compromise” form of approval voting that included a traditional runoff after a first round of voting which used approval. But the St. Louis system -- like other two-round systems -- relies on a small and unrepresentative primary electorate to control the choices in the higher-turnout general election.
Votes for a back-up choice can harm your first choice in approval voting.
Approval voting can be challenging for voters with strong preferences. A vote for a second choice counts exactly as much as a vote for a first choice, creating incentives to “bullet vote,” or choose only one candidate, even when voters have second- or third-choice preferences. Because voters can’t back compromise candidates without weakening their first choice, the use of strategic voting increases —especially in contested elections. In these races, few voters will vote for more candidates than winners, reverting the system back to plurality-like dynamics.
St. Louis also used approval voting in the first round of seven city council elections in 2021. In those contests, voters “approved” an average of between 1.1 and 1.4 candidates. This dismal rate may reflect voters’ unwillingness to support second-choice candidates when it can undermine their first-choice.
In RCV, voters’ ranking for a backup choice only counts if their first choice is defeated, eliminating the incentive to bury support for a second-choice candidate. Therefore, strategic voting is less likely under RCV. In fact, 71% of voters in RCV elections opt to rank multiple choices.
Approval voting rewards strategic voters and campaigns.
Savvy voters with robust understanding of approval voting can ensure their vote is highly impactful, while other voters may miss this opportunity. Voters who know which candidate poses the greatest threat to their favorite candidate will understand that they can bury support for competitors to boost their favorite candidates’ chances. Voters’ inability to distinguish between strong and weak support increases incentives to vote strategically rather than sincerely, putting voters on uneven footing.
In fact, a poll of Fargo, ND voters who used Approval Voting in 2020 found that nearly a third of voters who voted for one candidate did so strategically.
Approval voting could also incentivize hostile and polarizing gameplay between campaigns. Campaigns could encourage bullet voting to increase their chances of winning. For example, if a left, right, and center candidate (relative to the jurisdiction) ran in an approval voting race, and all voters voted honestly, the center candidate would likely win, having gained approvals from both sides. Sensing this, the left and right candidates could demonize the center candidate, so their respective bases don’t give approvals to that candidate.
In Fargo, multi-winner approval voting was first used in 2020 to elect two city councilors. For the upcoming 2022 election, candidates and voters are exploring the strategic incentives in approval voting to try to "game the system." FairVote's CEO Rob Richie examined the way approval voting is impacting the campaign of Fargo mayoral candidate Shannon Roers Jones, demonstrating the negative impacts on campaigns and voters from a system with such blatant opportunity for strategic exploitation.
RCV, on the other hand, does not redistribute a voter’s ballot until his/her initial preference(s)
have been eliminated. This means a vote for one campaign does not equate to a vote against
another. In fact, candidates will rely on the re-distribution of ballots (the alternate preferences of
other campaigns’ supporters) to win in later round(s). Thus, candidates are incentivized to
coalesce with other candidates.
Approval voting does not always elect candidates with broad support.
Even if voters are willing to “approve” of multiple candidates, already a stretch to imagine, candidates in a large field can win with low overall support from the electorate.
This is because approval voting has no majority criterion, and therefore might not elect a a majority-proffered candidate. Because voters can’t distinguish between strong and weak support on their ballots, it’s possible that a candidate who is the top choice of well over half of voters could lose to a candidate with little-to-no first-choice support.
Approval voting could also result in victory for a candidate with no supporters who would select them as their top choice over a candidate whom more than half of voters support as their first choice.
Dartmouth College in New Hampshire is one of few places that has tried approval voting, but students ditched it after a string of student presidents were elected with support from less than 40% of voters. The University of Colorado saw the same phenomenon, where typically more than 90% of voters vote for only one candidate.
In RCV, by contrast, winning candidates must have a strong base of first-choice support and also broad appeal within the electorate. RCV is a majority system.
RCV improves representation for women and people of color.
No sustained evidence finds that approval voting increases representation among women and people of color in public office.
RCV, however, has a demonstrated track record of doing so. See research on RCV and representation.
RCV is on solid legal ground
Approval voting is so rare that it has never come before a judge, so its ability to stand up to legal challenges is untested. Voters in approval voting elections have varying levels of electoral power based on how they interpret and assign their “approvals”, which could lead to legal challenges.
RCV, in contrast, has been upheld against every federal constitutional challenge brought to date.
Approval Voting is Less Expressive than RCV
With approving voting, voters are unable to express stronger and weaker preferences, and must decide to what degree liking or agreeing with a candidate equates to an “approval.” Voters would have to fit a diverse group of candidates into two boxes: “yes” or “no.” RCV, on the other hand, allows voters to express preferences between all candidates on the ballot.
Subjectivity of Preference in Approval Voting
Voters interpret “approval” in different ways, giving them different amounts of power over election outcomes. One voter might only vote for candidates they actively like. Another might vote for all candidates whom they find palatable. These voters have a different impact on the result because they interpret “approval” differently. Rating a movie either “thumbs up” or “thumbs down” is highly subjective; so too is rating candidates this way. And both yield inconsistent returns.
RCV, on the other hand, is less susceptible to strategic voting. Voters simply rank candidates in order of preference, knowing that a second-choice ranking will not harm their top choice, rather than worrying over exactly how much support to give to a second-favorite candidate.
Approval voting has never been used in a single-winner context for public government elections. In recent years, two U.S. cities — St. Louis and Fargo, North Dakota — have used multi-winner or modified approval voting for multi-winner contests.
St. Louis used a modified version of approval voting in 2021, using approval voting in the first round of a two-round runoff election. Results were mixed. In the mayoral preliminary election, voters “approved” of an average of 1.6 candidates to determine the two winners who would advance — a high rate for a typical single-winner election but low for a two-winner election. This was true even though the leading newspaper endorsed two mayoral candidates, and voters knew they could more safely vote for two candidates because two would advance.
As a result, voters in the general election -- in which turnout was 30% higher -- were presented with only the two choices identified by approval voting from the primary electorate. In other words, voters were presented with fewer choices at the time when turnout was highest. David Kimball, a political science professor at the University of Missouri at St. Louis found that voters in majority-White wards were more likely to cast multiple votes than voters in majority-Black wards. As a result, White voters likely had a larger say in determining the two finalists than Black voters.
Three city council runoffs resulted in different winners than those who led the approval-voting preliminary elections. The approval voting leader in Ward 17, for example, earned approval votes from 69% of voters, while the second-place candidate won approval from only 46% percent. Nonetheless, the second-place candidate beat the first-place candidate in the head-to-head runoff. Approval voting on its own was clearly unable to determine the candidate who would win in a head-to-head matchup.
Fargo used multi-winner approval voting in 2020 to elect two city councilors. Both winners had majority support, and voters used an average of 2.3 “approvals” per voter in a field of seven candidates. This version of approval voting is a winner-take-all voting method, which means a cohesive majority can control all seats, shutting out minority voices.
Private Experiments with Approval Voting
Political parties, universities, and private organizations have used approval voting and later replaced it. In 2016, the Independent Party of Oregon used this method in an online primary but found that over 70% of participants bullet-voted for only one candidate. The party switched to STAR voting in 2020, which also yielded mixed results.
Similarly, Dartmouth College used and repealed approval voting in both alumni and student elections after patterns of non-majority winners and strategic bullet-voting. In 2009, the school’s alumni group voted overwhelmingly to repeal the use of the method in trustee elections (82% - 18%). Dartmouth students also voted to repeal the method in student body elections in 2017.
The Institute of Electrical and Electronics Engineers (IEEE) repealed the use of approval voting in 2002 because the vast majority of voters (more than 80%) voted for only one candidate under the system.
Ultimately, the strong incentive to bullet-vote in approval voting elections often nullifies the promise of increased voter engagement. Approval voting does not make good on its promise to deliver better outcomes than plurality voting.
In an approval voting election, half of faction A voters (30%) vote for the candidates of their own party and the Independent candidate, whom they believe they can live with. The other 30% interpret approval voting differently and vote only for candidates they genuinely like, rather than those they can live with. They only vote for candidates from faction A.
Voters aligned with faction B vote for faction B candidates and the Independent because they understand that backing the Independent offers their best chance to defeat candidates from the majority faction.
|Number of Voters||Preferences||Approval Voting Behavior||RCV Behavior|
Prefer faction A;
oppose faction B
Half vote for A and I
Half vote for A only
Half rank A, I;
Half rank A
Prefer faction B;
oppose faction A
|Vote for B and I||Rank B, I|
neutral between A and B
|Vote for I only||Rank I|
In an approval voting election, votes would be counted as follows:
The Independent candidate wins, despite a strong majority preference for faction A. This is a non-majority outcome.
Minority factions will be very satisfied with this result: The 10% of Independent voters got their top choice, and faction B voters were able to deny a victory to their chief opponents, even though they comprised only 30 percent of the electorate.
The majority of voters (60%) will be disappointed in this result. In such an election, faction A voters have a strong incentive to bury support for any candidates other than their absolute favorite because of approval voting’s inability to differentiate between strong and weak support. In the simulation above, faction A voters could secure a more desirable outcome by denying votes to the Independent candidate. Ultimately, the faction that best understands the strategic incentives behind approval voting has a significant advantage over less-savvy voters.
In RCV, elimination rounds would proceed as follows:
|Candidate||Round 1||Round 2|
Under RCV, candidate A wins — regardless of whether some Independent voters ranked A or B as their second choice. And it would remain even if a second candidate from faction A or faction B joined the race. This stable outcome leads to a majority-preferred candidate — even if voters attempt to vote strategically.