Dr. Emily Campbell has published an analysis of Alaska’s implementation of ranked choice voting in the 2022 House Special Election. That article, published in the American Journal of Cultural Sociology, uses interviews to document the history and experience of ranked voting, as well as the backlash that followed. However, the article does not explore what can be done to mitigate the backlash to election reforms. This is an important issue for advocates to contemplate, and fortunately, a small tweak can make the system more robust to valid criticisms.
Alaska implemented a form of ranked voting known as Instant Runoff Voting or IRV. This common implementation of ranked voting eliminates the lowest first-vote earner in each round, and transfers those votes to the next-preferred option on the ballot. The bad news is that this method has some undesirable theoretical properties and those problems did arise in the 2022 Alaska House Special Election. The good news is that one small tweak fixes the biggest downsides of IRV.
To illustrate the problem with instant runoff, we can look directly at the vote tabulations from the 2022 Alaska House election shown in the table below. Mark Begich had the fewest first-choice votes at 54,009 compared to 58,939 for Sarah Palin and 75,803 for Mary Peltola. So, Begich is eliminated in the first round, and the ballots listing him as the first choice are redistributed to the second choices indicated on the ballot. In the second round, Palin now has 86,197 votes (58,939 + 27,258) and Peltola has Peltola has 91,375 votes (75,803 + 15,572). Thus, Mary Peltola won the election in the second round.
But consider those 34,117 Palin>Begich>Peltola ballots. These are voters who most preferred Sarah Palin among the candidates, and their second choice was Mark Begich. If some of these voters had stayed home rather than vote in the election, Begich would have won the election. Because those voters preferred Begich over Peltola, those voters would have been better off by staying home rather than participate in the election.
Suppose, for example, that just 5,200 of the Palin>Begich>Peltola voters had stayed home. The table below shows the counter-factual tabulation for this hypothetical scenario, where everything else is exactly as before. The only difference is that rather than 34,117 Palin>Begich>Peltola ballots, we now have only 28,917 such ballots, since we are supposing that some of those voters stayed home rather than participate in the election.
In this scenario, Palin now has the fewest first-choice ballots, and is eliminated in the first round (Palin has 53,739 votes compared to the 54,009 for Begich and the 75,803 for Peltola). The second round then has Begich and Peltola remaining. In the second round, Begich has 82,926 votes (54,009 first choice votes + 28,917 transferred from Palin) while Peltola has 79,486 votes (75,803 first choice votes + 3,683 transferred from Palin). Begich would have won the election in this case. What this means is that by staying home, these Palin>Begich>Peltola voters would have been better off.
This is an extremely important point. Many advocates of instant runoff voting claim that an advantage of IRV is that “you can vote your conscious” without concern about strategic voting or making yourself worse off by voting your true preference. But this is untrue, both in theory and in real-world experience.
Mathematician Julie Clelland makes a closely related point in her 2024 article “Ranked Choice Voting And Condorcet Failure in the Alaska 2022 Special Election: How Might Other Voting Systems Compare?” Dr. Clelland shows that the election exhibited “Condorcet failure.” Based on the ballot rankings, Begich would have won a head-to-head contest against Palin, and he would have won a head-to-head contest against Peltola. However, Begich did not win the IRV election.
“Better IRV”: A small tweak to make ranked voting more robust
There is good news though. One small tweak to the classic instant runoff system yields a ranked voting system that fixes this key vote-splitting issue. And it is just a tweak to the way candidates are eliminated. Ranked ballots are still designed and submitted by voters in the same way as instant runoff voting. The tweaked system is called Bottom-Two Runoff IRV or BTR-IRV.
The only difference in BTR-IRV compared to instant runoff is that rather than eliminate the candidate with the fewest first-choice votes each round like in IRV, each round takes the two candidates with the fewest votes and puts them in a head-to-head contest. The candidate that is eliminated each round is the one with the least support among the two bottom candidates.
This one small tweak has important benefits for the outcome of ranked vote elections. Most importantly, it avoids the possibility of eliminating a broad-support consensus candidate in the first round. And in fact that is exactly what happened with instant runoff voting in the 2022 Alaska House Special Election. Begich had a huge amount of second-choice support from both Palin and Peltola voters. Yet because he had fewer first-choice ballots, he was eliminated in the first round, before all of his second-choice support could be counted in his favor.
BTR-IRV fixes this, and in doing so remedies a logical inconsistency of instant runoff voting. Proponents of instant runoff say that it is better than the plurality voting system because a candidate who receives the most votes (a plurality) does not necessary also have the support of the majority of the electorate. And that is true. But then instant runoff voting does apply plurality logic to the candidates it eliminates each round. If the candidate with the most first-choice ballots should not necessarily win, which does make sense, but then why should the candidates with the fewest first-choice ballots necessarily lose?
By eliminating the loser of a head-to-head match-up between the bottom candidates in each round, BTR-IRV prevents a consensus candidate with broad support from being eliminated “too early.” That one small change then drastically mitigates the opportunities for vote-splitting and for voters to need to consider strategically voting. It is a system where the voters can truly just “vote your conscious” and not have to worry about undermining their own preferences in the widest range of circumstances.
This is a really important consideration for minimizing the backlash to election reform. After the 2022 Special Election, some Republicans were upset because although broad support for Begich had been expressed on the ballots, the Democrat won. And the thing is, that is a valid criticism! Some criticisms about election reform are not valid or coming from a good-faith effort to improve our politics. But minimizing backlash requires taking valid criticisms seriously, and not dismissing them. There was a movement to repeal the election reforms on the ballot in Alaska in 2024, and that repeal effort just barely failed. Sustaining election reform requires implementing systems that are robust. The good news is that one small tweak to instant runoff voting yields a BTR-IRV system that has much more desirable properties and is less vulnerable to valid criticism.
Further reading
For anyone looking for a deeper dive into the theoretical properties of voting systems, the 2020 article “Elections and Strategic Voting” by Harvard economist and Nobel laureate Eric Maskin and Cambridge (UK) economist Partha Dasgupta includes the basis for the theoretical claims made in this post. I would summarize the main takeaways from that article at a high level as arguing that (1) while no voting method is perfect, (2) runoff voting is vulnerable to vote splitting, and (3) majority rule, defined as where the winner is the candidate that beats all other candidates head-to-head, has the most desirable properties in the widest range of circumstances, including no vote splitting and strategy-proofness.