ranked choice voting doesn’t solve the spoiler effect
but approval voting does
the alaska house election this november 2022 will use ranked choice voting (RCV), and will feature three candidates.
moderate republican nick begich is almost certainly preferred head-to-head against peltola to his left, and palin to his right. but based on these numbers, he’ll be eliminated by the RCV process, and then it’ll be peltola vs palin.
if palin wins, this will be a classic “center squeeze” problem. so if you’re a democrat worried that peltola will lose head-to-head against palin, your best strategy is to insincerely/tactically rank begich in 1st place as a “lesser evil”.
approval voting virtually eliminates this spoiler effect, based on two criteria:
approval voting satisfies the sincere favorite criterion, meaning it’s always safe to vote for your favorite. so if those democratic voters were to decide to stategically vote for begich, they’d still want to include a sincere vote for peltola, because that couldn’t possibly hurt them.
as for independence of irrelevant alternatives, this means that adding or removing an “irrelevant” candidate who doesn’t win shouldn’t change the outcome. e.g. the green party shouldn’t be able to change the winner from democrat to republican if voter preferences between the two major parties are the same either way. this is a bit tricker. a voting method has two components: the ballot casting and the tabulation. with approval voting, the tabulation component satisfies independence of irrelevant alternatives. take a stack of cast ballots and remove any losing candidate, and observe that the winner cannot change. whereas with ranked choice voting, removing a losing candidate can absolutely change the outcome, leaving all other rankings unaltered.
as for the ballot casting component of approval voting, it doesn’t satisfy independence of irrelevant alternatives given that voters can be strategic. that is, virtually no one would ever approve or disapprove both candidates in a two-candidate race.
approval voting nearly eliminates any notion of spoilers, and certainly does so more thoroughly than RCV.
here’s a video about this phenomenon from andy jennings, who did his math phd thesis on voting methods and co-founded the center for election science.