Did the Alaska Ranked-Choice Voting Work? It’s Complicated.

Alaska’s first election under the new top-four, ranked-choice voting system has come to a close, and in a surprise, Democrat Mary Peltola emerged victorious with 91,206 votes to Republican Sarah Palin’s 85,987; fellow Republican Nick Begich finished in third, eliminated in the first round. While this outcome wasn’t a total surprise, given Palin’s low favorability numbers and the brutal two-way fight between her and fellow Begich, it might not have been possible without this new system.

It’s worth looking, then, at how ranked-choice voting worked, and whether it met the goals reformers wanted: eschewing extremism for more moderate candidates, and creating winners more appealing to broad segments of the electorate.

Speed of Counting: Much Better than Expected

As I wrote about earlier this week, Alaskan voters had to wait a while to find out the winner. Over two weeks had passed before the RCV count was announced, which isn’t outside the norm for Alaska. Because of the large number of votes from military and disconnected rural areas, elections can take a while before results are known, even under the old FPTP system.

That being said, while the counting took a while to happen, the results were announced right on schedule yesterday. This is a really positive sign that this system actually won’t take substantially longer than before. Further improvements can be made on the speed of tabulating ranked-choice votes, but I was pleasantly surprised to have the results come in instantly last night, even if it was a few weeks past election day.

Extremism: A Mixed Bag

One of the most-touted benefits of ranked-choice voting is that its design discourages extremist candidates by making it harder for them to win. Rather than sneaking through with a plurality of the vote, candidates need to appeal to the broadest segments of the electorate and to the coalitions of other candidates in order to succeed.

In this regard, RCV succeeded in one obvious area: the more moderate candidate, Peltota, defeated Palin, the more extreme candidate. But the problem comes when you look under the hood – is this result functionally any different than what would have happened under a standard FPTP election? Under the old system, Palin wins the Republican primary and Peltota wins the Democratic one. It’s quite possible that, even under the old system, Peltola wins the general, or at least comes close. In either scenario, though, Palin – the more extreme candidate than Begich – is effectively the last Republican standing.

The only difference would be that, in a traditional FPTP system with a partisan primary, Palin would have had a few months to campaign and win over Begich voters. Under the current system, the two were bitter rivals campaigning directly against each other even after the primary. This actually disincentivized either candidate from moderating or appealing to the other side, as both wanted to be ranked above the other. Moreover, it gave Peltolta a mostly clear path for a third lane. While a smart campaign from Palin and Begich would have emphasized ranking each other as a backup, this goes against what most politicians instinctively understand; being ranked second doesn’t help you if you lose.

If anything, the new system might even encourage candidates to appeal to ideological extremes in order to ensure they make it to the next round. Virtually all Alaskan elections for the last decade would have had at least two Republicans making the top four, and some others would have seen two Democrats make it as well. In this scenario, where voters can only rank one candidate first, what’s the more rational play – appealing to moderate voters that might have you as a second or third choice, or appealing to a committed base of partisans willing to rank you first?

Majoritarianism: A Swing and a Miss

Another benefit ranked-choice voting advocates frequently cite is majoritarianism – the candidate that a majority of voters want will win. This argument is mostly sophistic, however. Ranked-choice voting actually doesn’t yield a majority winner every time, unless you pretend that exhausted ballots aren’t actual votes cast. The Alaska special election was no exception:

The candidate with the most votes, Mary Peltola, won, as you would expect in any rational election system. However, Peltola did not win the majority of votes cast. She only won a majority if you disregard the 10,726 exhausted ballots (votes which didn’t rank a candidate).

Like in many elections in Alaska, Peltola won with a plurality of the vote. This isn’t abnormal, and has long been a staple of Alaskan politics under FPTP; independent Bill Walker won the 2014 Gubernatorial race over Republican Sean Parnell by a similar margin (48.1% to 45.9%). However, it’s pretty clear that ranked-choice voting didn’t actually deliver on the claim of majoritarian support. It’s just not a substantially different outcome to previous races under FPTP.

Conclusions

Ranked-choice voting’s first experiment was far from a disaster, but it also didn’t quite meet the lofty goals that advocates have argued. In particular, the more extremist candidate (Palin) was nearly elevated to a position of victory, while the candidate that would have received the highest percentage of votes according to polling (Begich), was eliminated immediately.

I still maintain that this new system has inherent, structural flaws. These flaws can’t quite be addressed through reform. A system like approval voting would have likely yielded a different outcome where a candidate (Begich) would receive a majority of votes cast, while a fringe-oriented candidate like Palin would have been locked in a distant third, unable to gain beyond her initial coalition. One potential reform could be to reduce the number of candidates from four to three; another could be to adopt a Condorcet-style method, where all the individual head-to-head matchups are accounted for.

Ultimately, given that most races in Alaska will, for the foreseeable future, have at least two Republicans on the ballot, the Alaska GOP and its candidates will need to adjust to this. While discussing election reforms or proposing new systems is nice and all, ultimately you have to play the game under the rules that are set, not the rules you want.