How to Not Campaign in a Ranked Choice Election

Don’t attack your allies

Background

Alaska’s new electoral system, adopted by voters in a 2020 referendum, has created some interesting results. In the first round (often referred to as a primary), voters have one vote for all candidates of all parties listed on the ballot. The top four vote-getters advance to the second round, the general election. Voters can then rank up to four candidates in an instant runoff system.

For those unfamiliar with Alaskan politics, the dynamics are very weird compared to the lower 48. But that’s a story for another time.

This “final four voting” system replaces the old partisan primary system, where candidates from the same party run against each other in the primary, and the winner of each party’s primary runs against each other in the general election. The incentives are different for each candidate

Supporters argue that this system causes more cooperation and less negative partisanship, claiming that ideologically aligned candidates need to work together to gain each others’ second choices.

I argue this is partially true. The top-four primary mostly increases coordination and decreases negative partisanship, but does not always produce the most agreeable candidate due to Condorcet failures. (Please disregard the headline; Peltola is not that progressive.)

We have two paths: one of coordination between two ideologically aligned candidates, and one with constant backstabbing.

The 2022 Alaska Governor’s Election

Les Gara (D) vs Bill Walker (I) vs Mike Dunleavy (R)

In the 2022 Alaska Governor’s election, Les Gara, a Democrat, and Bill Walker, the Independent former governor, faced off against Mike Dunleavy, the incumbent Republican governor.

The ranked-choice component may increase positive campaigning if both ideologically aligned candidates are willing to attack their opponent. This is evident in Gara’s debate performance.

We’ve seen candidates campaign together and ask each other’s voters to rank the other candidate second. But defending your opponent’s record is a hell of a thing to do in a debate.

In the general election, Gara and Walker ended up splitting the not-Dunleavy vote, and Dunleavy won outright wihout needing to run the RCV procedure. Dunleavy would have won with 59% of the vote in a hypothetical head to head maximum round against both Gara and Walker.

The 2022 Alaska House of Representatives Special Election

Sarah Palin (R) and Nick Begich III (R) vs Mary Peltola (D)

This race has a long backstory (with Santa Claus and lots of backstabbing involved), even before the special election.

Palin, the 2008 Vice Presidential candidate, was endorsed by Trump and represented the MAGA wing of the party, with her base in the Matanuska-Susitna borough, where she was formerly mayor of Wasilla. Nick Begich was a mainstream conservative, against the far-right wing of the Republican party. Peltola ran as a pragmatic candidate on the slogan “Fish, Family, Freedom”, with her base in the Alaskan Bush. Expectations were that a Republican would easily win.

The Alaska Republican Party’s “Rank The Red” initiative was launched to encourage ranking Republican candidatess, but was unsuccessful because Republican candidates did not present a unified message. Palin actually appeared to help Peltola, praising her. This stems from the two’s tenure together in the Alaska statehouse. In fact, Palin spent most of her energy attacking Begich.

From a game theory perspective, Palin and Begich were victims of a prisoners’ dilemma. Both should have attacked Peltola in a coordinated manner, but did not due to strategic and personal reasons. In the special, Palin did not disclose who she would rank second, while Begich cooperated and told his supporters to rank Palin second, yet continued to feud with Palin during the campaign. After both Republicans lost the special election, Palin made an attempt to cooperate and told her supporters to rank Begich second, yet still continued the feud.

Peltola vs Begich

This simplistic scenario is not entirely accurate, but is good enough for a general overview of the election.

2024 and beyond

Democrats and aligned factions are united behind Peltola. There is no sufficiently large left faction to challenge her position. Incumbency is also a powerful advantage in Alaska, because in a low population state, it is easier to build relationships with constituents. We see this with home-state effects being more powerful in small states.

Republicans are split between many factions: the center-right wing and the Trump-aligned wing. Therefore, Republicans will have multiple candidates advancing to the final round of IRV, and Democrats will most likely have one candidate. Overall, Peltola is more favored than the state’s Trump+10 partisan lean suggests.

Ballot Exhaustion in IRV

Palin’s voters were more likely to bullet vote than Begich’s voters.

Ballot exhaustion rates were higher among the Republicans, due to Democrats’ tendencies to coalesce around one candidate more often.

In a traditional runoff (like Georgia), having one candidate of each party makes the race more partisan. Thus, one is more likely to vote for the lesser of two evils. In contrast, with RCV people oftentimes do not rank beyond their first couple choices, leaving candidates they dislike unranked.

We even see ballot exhaustion persist down the ticket, at the state legislative level.

Finally, I leave you with this gem. Many voters simply do not understand the RCV system! The author of this article fails to protest vote effectively in the 2020 Maine Senate election. Ranking the two independents (who endorsed Gideon (D)) first, then ranking Gideon is the correct way. As long as a voter ranks all candidates but the disliked one, ranking the disfavored candidate last and not ranking that candidate have the same effect.

Conclusion

Running multiple candidates of similar ideologies is usually detrimental due to ballot exhaustion and inefficient resource allocation. Coordination between candidates with similar ideologies is key.

And finally, RCV has a marginally good to highly positive effect on enabling more agreeable candidates to win! There is no scenario where the partisan primary system produces more representative winners than RCV.