Seventeen people ran for governor of Alaska in the recent Aug. 18 primary. Four of them will be on your ballot in November. Two Democrats and two Republicans.
Supporters of ranked choice voting and the jungle primary told us this system would give voters more choice and clearer results. Let’s look at what it actually gave us.
This system does not produce clarity. It produces arithmetic.
Most Alaskans remember the 2022 special election for U.S. Congress. Mary Peltola beat Sarah Palin after Nick Begich was eliminated in the first round.
Then mathematicians went back and studied the actual ballots. Begich would have beaten both of them.
Jeanne Clelland, a professor of mathematics at the University of Colorado Boulder, ran the numbers from the ballots Alaskans marked. Against Peltola one-on-one, Begich wins with about 52%. Against Palin one-on-one, he wins with about 61%. He beat every opponent head-to-head, yet he was the first of the final three eliminated.
Ranked choice voting was sold to Alaskans with a simple promise. No more winners with less than half the vote. That is not what happened.
The simplest way to say it is this. Among voters who marked a second choice, Begich won an overwhelming share. The system threw those votes out before counting them.
Mathematicians call a candidate who beats every opponent one-on-one the “Condorcet winner.” Clelland found Begich was the Condorcet winner in 2022, and Palin the Condorcet loser. The candidate who could beat either opponent was eliminated first. The one who could beat neither made the final round.
That was not voters filling out ballots wrong. Those ballots did exactly what voters told them to do. The result came from the counting method itself.
Two other researchers, Adam Graham-Squire and David McCune, studied the same ballots and found something stranger. Their analysis showed Peltola could have lost that election by getting more votes. Under certain changes in support, gaining supporters could have cost her the race.
They also found what mathematicians call a “no-show paradox.” Some voters could have gotten a result they preferred by staying home. According to the researchers, Alaska in 2022 appears to be the first American ranked choice election ever documented to produce that outcome.
Read. That. Again. We built an election system where showing up to vote could work against the result a voter wanted.
I have said before that ranked choice voting lets an algorithm decide our elections. That is not just a campaign line. When it takes a peer-reviewed study to explain why the strongest candidate lost, we have made voting harder than it needs to be.
Ranked choice voting was sold to Alaskans with a simple promise. No more winners with less than half the vote. That is not what happened.
The Division of Elections reported Peltola with 91,266 votes and 51.48 percent in the final round. But that percentage was figured against the 177,292 ballots still active at the end, not against every ballot Alaskans cast.
Thousands of Alaskans ranked only one candidate. When that candidate was eliminated, those ballots had nowhere to go. The state counted 11,243 of them as exhausted.
Measure Peltola’s 91,266 votes against the 188,582 ballots active in the first round and she finishes at 48.40 percent. That is Clelland’s calculation, using the state’s own report. Under half. The one thing this system promised to fix is the thing it failed to do.
Now go back to the recent Aug. 18 primary.
Under the old system, each party held its own primary and one candidate came out of it. That primary was a filter. It was not glamorous or fun, but it did a job. It took a crowded field and narrowed it to a simpler choice before November.
The 2020 ballot measure removed that filter and replaced it with an open top four primary. Every candidate runs together and the top four advance no matter their party.
You can dislike the rules and still know how to use them. We lose this race in the third round or we don’t.
That is how 17 candidates go into a governor’s primary and the November ballot comes out two Democrats and two Republicans.
Defenders will say this is rare. Graham-Squire and McCune examined 182 American ranked choice elections held between 2004 and 2022 and found only two that failed to elect the head-to-head winner.
One was Burlington, Vermont, in 2009. The other was Alaska.
But rare is not the same as never, especially when you live in one of the two places it happened. Burlington repealed ranked choice voting after theirs.
Ballot Measure 2 is on the ballot this November. A yes vote repeals ranked choice voting and the open top-four primary that came with it.
I am voting yes, and my reason is simple. An election should make the voter’s job easier and the result easier to understand, not require a mathematics paper afterward to explain what happened. This system does not do that.
ALASKA WATCHMAN DIRECT TO YOUR INBOX
Vote YES on 2. End the mess.
Until that vote is counted, these are still the rules governing every race on that same ballot.
We used to get a binary choice. Now we get a menu. If two of those four names are more acceptable to you than the other two, and for most Republicans they will be, rank them both, because a ballot that stops after one name stops mattering after one round.
Marking only one name does not give that candidate extra support. Ranking someone second does not take a vote away from your first choice.
Eleven thousand Alaskans found that out the hard way in 2022.
You can dislike the rules and still know how to use them. We lose this race in the third round or we don’t.
That part is up to you, not to them.
The views expressed here are those of the author.
