Many people have figured out that the plurality or “first-past-the-post” voting system used in much of the U.S. causes political dysfunction and polarization. It limits voter’s choices and incentivizes politicians to obstruct their opponents, rather than compromise, collaborate, and get things done. Thus there has been a growing election reform movement aimed at making politicians more accountable to the voters.
However, these election reform efforts have been divided. And I think one of the main obstacles to success has been that many reform advocates spend a lot energy fighting among themselves over which specific reform is best. Some prefer ranked choice voting, others approval voting, and still others advocate for “score” voting. Taking power away from entrenched party elites is hard enough. Doing so with a divided and infighting movement will be impossible. The path to success requires a more unified and collaborative election reform movement.
For example, the Equal Vote Coalition recently posted on social media against ranked choice voting, writing that “We do not believe that Ranked Choice Voting is better than the status quo. RCV has undermined the real voting reform movement for too long. It’s time for a new era. Let’s do better.” They and the Center for Range Voting advocate for “score voting.” The Center for Election Science advocates for “approval voting” and specifically not for ranked choice voting. When alerted to critiques of ranked voting, FairVote tends to defensively double-down on instant runoff voting rather than engage seriously with the valid critiques raised.
Fortunately, the election reform movement can readily become a more unified coalition. The advocates already agree that the plurality voting system we use today is the root cause of political dysfunction. We just need to find a compromise voting system that everyone can live with (similar to how they all want a voting system that elects compromise candidates that everyone can live with). And some very clever individuals have pointed to the answer: bottom two runoff IRV (or BTR-IRV).
Before explaining how BTR-IRV works, let me say that I am not saying that everyone needs to always advocate for only this system. But my hope that is election reform advocates should all be open to the BTR-IRV system, and keep it as an option on the table. Having this voting system in the advocate’s toolkit during reform efforts could help achieve more success by increasing collaboration.
How BTR-IRV works
For those familiar with instant runoff voting (IRV), the traditional ranked choice voting, the good news is that BTR-IRV is just a very small tweak to that method. From the voter’s perspective, nothing changes at all. A ballot has a list of candidates, and the voter ranks the candidates in order of their preference. BTR-IRV is still ranked choice voting. The only thing BTR-IRV tweaks compared to traditional IRV is how the ballots are counted.
In BTR-IRV, the winner is determined by ordering the candidates by the number of 1st choice votes each received. The bottom two candidates then go to a head-to-head contest based on all of the ballots cast in the election. Whichever of the two candidates is less preferred across all ballots is eliminated. And then the process continues for the two lowest remaining candidates until one winner remains.
If a candidate receives over 50% of the first-choice votes, that candidate will win the election, and there is no need to do all the elimination rounds. That candidate will clearly be able to win against every other candidate in a head-to-head contest.
There is another variant of BTR-IRV which involves transferring votes, just like IRV. In IRV, in each round the candidate with the fewest votes is eliminated, and their votes are transferred to the next choice indicated on the ballot. BTR-IRV can also be conducted using transferred votes in the same way, and when a candidate is eliminated, their votes get transferred based on the next choices indicated on the ballots.
A clear way to see an advantage of BTR-IRV over traditional IRV is by the mechanism of who gets knocked out each round. In traditional IRV, advocates say that just because someone receives the most 1st choices does not mean they should win the election. But then the method has an asymmetry, because if someone receives the fewest 1st choices that does mean they should be eliminated. What this causes is that a very good choice with broad support (i.e. lots of second choices) can be mistakenly eliminated in the first round. Indeed, as we have analyzed already on this blog, that is exactly what happened in Alaska’s 2022 special election for the House.
Either version of BTR-IRV is great. I prefer the version of BTR-IRV that does not involve transferring votes, because it is simpler. However, a virtue of BTR-IRV that involves transferring votes is that it is only one small tweak to the traditional IRV method. Rather than eliminate the candidate with the fewest first-choice votes like in IRV, BTR-IRV eliminates the candidate who loses in a head-to-head contest between the bottom two fewest first-choice getters.
A clear and concise presentation of BTR-IRV, and how it compares to traditional IRV, is available on the site of Will Klieber.
Why not IRV, scoring, or approval?
Almost any alternative voting system is better than the currently plurality voting we use. And no system is perfect: every voting system must make some trade-offs between desirable and undesirable properties. While BTR-IRV makes very appealing trade-offs, but I am not against other methods. I would just rank them lower on my ballot, and here is just a brief word on why.
Traditional IRV has one very undesirable property, which is that by voting your true preferences, you can induce an outcome that makes you worse off than if you had not participated in the election at all. And this is not just a theoretical possibility: this is exactly what happened in Alaska’s 2022 special election for the House of Representatives. As we discussed previously, in that election Republican Nick Begich was eliminated in the first round because he had the fewest first-choices. After his votes were redistributed, Democrat Mary Peltola defeated Sarah Palin to win the election. But in that election, there were 34,117 voters who ranked Palin > Begich > Peltola. If 5,200 of those voters had stayed home rather than cast a ballot, Begich would have won the election and those voters would have gotten a better outcome for themselves (Begich instead of Peltola). This is because if they had stayed home, Palin would have been eliminated first, and once her votes were redistributed, Begich would have defeated Peltola in the final round. Thus, Palin > Begich > Peltola voters would have gotten a better outcome for themselves if they had stayed home rather than voted!
It is simply untrue that under IRV, you cannot hurt yourself by voting your true preferences. But the good news is that just making a small tweak and going with the BTR-IRV ranked choice system will make statements like this true!
Scoring and approval voting likewise have some undesirable features. For me, the big one is that voting is necessarily “strategic,” meaning that you need to know how others will vote in order to cast your vote optimally. Imagine a simple case of approval voting, where you love option A, hate option B, and are okay with option C. Clearly, the voter should approve A and not vote approve on B. But should option C be given approval or not? The recommendation really requires knowing how others are going to vote. Mote details on this point regarding strategy-proofness are available in a paper by Partha Dasgupta and Eric Maskin called Elections and Strategic Voting: Condorcet and Borda.
Conclusion
If BTR-IRV could be what we mean by “ranked choice voting,” that would eliminate a lot of the resistance to ranked choice voting that is coming from election reform advocates themselves. We would be able to focus on the resistance coming from the people who prefer the status quo – which is an easier task because the status quo is so clearly broken.
My goal is not to say that organizations should give up their research and advocacy for the methods they most prefer. But only that organizations that advocate for ranked voting (FairVote, Rank the Vote, Forward Party, and the many wonderful local reform organizations) should be open to BTR-IRV as a way to implement “ranked choice voting.” And the groups critical of ranked choice voting should hopefully consider that the BTR-IRV method of ranked voting accounts for at least some the valid criticisms they bring. Ultimately, I do believe that debate and discussion across methods can make the movement stronger when done with an open mind, and that BTR-IRV can help bridge some of the divides that currently exist.