Documenting the Debate About Ranked Voting in Oregon

16 March witnessed a hearing on Oregon’s potential adoption of ranked-choice voting (RCV) which imparted a sense of momentum for the STAR voting movement in Oregon. The hearing itself is viewable at this link. Almost from the beginning, it was clear that the debate was not about the value of RCV over the current plurality voting system, but instead, its value compared to STAR Voting.

RCV takes a ranking from voters and returns a (potentially constructed) majority (> 50%) winner. Oregon Senate Bill 791 would establish RCV as the voting method for selecting the winner of nomination for and election to non-partisan state offices and county and city offices (with exceptions), as well as establishing RCV as the method for selecting the winner of nomination for major political parties for federal and state partisan offices. Senate Bill 343 permits counties to adopt RCV as their voting system and include provisions for funding and machines.

What’s the Difference?

RCV executes a standard one-person-one-vote election except that voters submit a rank order of candidates. Their ballots are only cast for the candidate currently ranked highest on their ballot. Voters supply rankings so that if no candidate achieves a greater than 50% majority of the votes cast, the candidate receiving the least number of votes can be automatically removed from all ballots. Ballots are then recast for their (new) highest-ranked candidate. The process continues until the system creates a winner with over 50% of the votes.

STAR Voting is a voting method designed in Oregon in 2014 as a compromise between score voting and RCV. Using the score voting method, all voters would indicate their level of desire for all candidates using a set scoring range (i.e., 0 – 5), and the most desired candidate would be declared the winner. Under the STAR Voting compromise, voters still indicate their level of desire for all candidates, but the two most desired candidates then go on to an automatic runoff to identify which of the two candidates was preferred the most in head-to-head comparisons.

The STAR Voting preserves RCV’s requirement that the winning candidate receives over a 50% majority of the vote (constructed). STAR Voting thus neither guarantees the voters most desired candidate will win nor a candidate who the top-choice of more than half the voters would win.

STAR in Oregon

Senator Jeff Goldman, a co-sponsor of SB 791 and the sponsor of 343, brought STAR Voting in as a topic stating that “I may be filing an amendment to [SB] 343 that allows counties to adopt, not just RCV but the preference voting system of their choice,” as he noted STAR Voting is, “a system that I think
deserves some consideration.”

Representative Wlnsvey Campos also took up this line of focus on STAR Voting as a proponent for RCV, noting her concern that STAR Voting in races with more than three candidates “it is possible for a candidate to be the top choice of more than half of the voters and yet still lose by not getting enough scores to make it into the runoff. That means this candidate would lose even though more than half of the voters showed on their ballot that this was their top choice. Outside of the current electoral college system . . . there are no other examples of an election where that could happen, and for good reason. It violates the Democratic ideal of majority rule.”

Representative Zach Hudson, the lone dissenting voice among the five elected leaders who testified against the two senate bills focused his testimony almost exclusively on the reasons why STAR Voting outperformed RCV.

“I was a proponent of ranked-choice voting until I found out that not only are there better ways to show voter preference but also that RCV, because of how it’s designed, can lead to some strange results that aren’t immediately apparent” Representative Hudson began. He stated his belief that RCV did not work well when an election race had three competitive candidates. He used as an example the 2009 mayoral race in Burlington, Vermont, which used RCV, where the third candidate to be eliminated was Andy Montroll.

“The problem is that voters actually preferred Montroll,” he said, “against either of the other two candidates in a head-to-head match.” He concluded this example by saying, “incidentally, Burlington voters repealed RCV shortly after this election.” He also demonstrated to the Senate Rules Committee how, in an RCV election, voters could create a worse outcome for themselves by giving their favorite candidate the highest ranking. He stated his desire to establish a voting methods task force (House Bill 3241) to consider the various options.

Testimony on STAR and RCV Voting

The following testimony appears as given by public witnesses, and claims on either side were not validated for truth and accuracy by the author. Eight public witnesses testified in support of RCV. Most of these witnesses highlighted how many other groups had adopted RCV, decried the failings of plurality voting and noted RCV’s ability to offer “more voices and choices” in elections.

The central testimony in support of SB 791 and SB 343 included:

• RCV has been adopted by many organizations and communities across the globe representing over 10 million voters, making it “proven, tried, and tested.”
• RCV elections have resulted in the election of more women (25% under plurality, 49% under RCV) and minorities (38% under plurality 62% under RCV) in four Bay city areas.
• Research on voting methods (The American Academy of Arts and Sciences (2020) on better practices in Democratic Citizenship “Our Common Purpose”) identified RCV as one of the most important (listed third and fourth) recommended government improvements over the current government practices.
• RCV is easy as “picking one, two, three.”
• RCV ensures the election of a numerical majority candidate (> 50% of votes), unlike plurality voting.
• RCV has a large support network to help implement it.
• RCV ensures a majority winner but doesn’t require voters to vote for only a “D” or an “R.”
• RCV avoids using strategic voting (having to choose the lesser of two evils).
• RCV elections end expensive live runoff elections.
• RCV curtails divisiveness as candidates must reach beyond their base to win.
• RCV forces deliberation on more issues.
• (against STAR Voting) STAR Voting has not been used in actual elections.
• (against STAR Voting) STAR Voting favors voters who feel more passionate about candidates and who are more confident. This leads to a white person’s vote counting for more than a person of color (based on the idea that two voters who feel the same about a candidate but score them differently have their different scores counted differently).
• How RCV works is understood by voters comparably to plurality.
• RCV allows voters to express themselves through their preferences.
• RCV increases voter turnout between primary and general elections compared to plurality.
• RCV exhausted votes were mainly caused by voters not putting in all the allowed rankings (three), and almost 88% of voters’ votes counted. (FairVote approximates 5% of ballots are not counted in the final vote determination).
• RCV eliminates the spoiler effect.

Four public witnesses noted their desire to not change the current system, citing the level of confusion and mistrust it could create.

• RCV is hard to understand.
• RCV is strategic in nature.
• RCV is not transparent, and voters would strongly question the results.
• RCV, as it was presented [in SB 791 and 343], was not acceptable and needed to allow ranking two or more candidates at the same level and remove pair-wise losing candidates (candidates that would lose a one-on-one match with every other candidate).

Nine public witnesses came out against SB 791 and 343, with most of them noting that though they fully agreed with the problems of plurality voting and wanted the same result as the RCV advocates, they testified RCV did not live up to its hype. Many noted that they too had once supported RCV, though they now saw it not as “not delivering on its promises” and “it offers more than it actually is.” They also noted their strong support of STAR Voting as an alternative to RCV or at least the need to form an independent committee to review the voting systems (as proposed in HB 3241).

The central testimony against SB 791 and SB 343 included (testimony previously listed against not repeated here):

• RCV fails to allow voters to vote for their most desired candidate.
• RCV’s algorithm for tallying votes is complex and can’t be done precinct by precinct.
• RCV does not work in races with more than two competitive candidates.
• RCV fails to eliminate the spoiler effect (spoilers happen in 15% or more of RCV events / up to 1 in 5 elections when there are more candidates). Spoiled elections disproportional affect historically marginalized communities.
• RCV does not decrease spoiled ballots (incorrectly completed ballots) over plurality.
• RCV does not allow voters to rank two candidates at the same level.
• RCV elections show similar levels of racial and financial disparity as plurality elections.
• Concerns over RCV caused four counties in the last 20 years to repeal its use.
• RCV fails to count all voters’ ballots (ballot exhaustion) in the final count, with exhausted votes ranging from 9.6% to 27.1%.
• RCV has been oversold.
• There are much better voting methods (i.e., STAR Voting) that do everything RCV promises but fails to deliver to include ending the spoiler effect, avoiding the need to vote strategically (ranking a less desired candidate higher to help the candidate you want to win), and ensuring all votes count in the election.

Proponents for RCV most strongly professed that RCV demonstrated superiority over STAR Voting due to its wider acceptance and use than the six-year old new-comer system. However, no group felt that the same logic should result in defining plurality voting as the superior system despite its use by
hundreds of millions of people for hundreds of years.

Proponents for STAR Voting most strongly professed that STAR Voting did all the things RCV promised to do but has been “tested and proven” not to do. However, until communities and governments use STAR Voting more widely, its promises of a better future will remain primarily based on logic and mathematical modeling.

Future voting reform debate in Oregon will focus on what STAR Voting can do and what RCV fails to do. The rest of the nation watches to see if a more modern alternative to the failed plurality voting method has now arrived.