Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

Ranked-choice voting gets a hearing in Maine

August 13th, 2020, 8:07am by Sam Wang

Today in Maine, ranked-choice voting will be considered in federal court for the third time. And our team at Princeton has weighed in.

In 2018, ranked-choice voting was used for a federal election for the first time anywhere in the United States, in the 2nd Congressional District of Maine. And it made a difference. Minor-party voters were able to express their first preference, without fear of throwing their vote away. They could then give their support on later rounds to either Jared Golden (D) or Representative Bruce Poliquin (R). In the end, their support was enough to elect Golden to the House seat. Over 95% of voters’ preferences were used in all rounds of vote-counting, including thousands of minor-candidate voters.

Minor-party voters were highly strategic in thsir use of the ballot, most often ranking all the candidates. In contrast, Golden and Poliquin supporters’ most frequent choice was to stop after making one choice.

However, plaintiffs have introduced a novel definition of participation. They argue that supporters who made only one choice didn’t fully participate. Therefore, they claim, ranked-choice voting impairs democracy. Their brief is here. My team here at Princeton took a look at the brief. We found the arguments to be deficient, and we shared some of our thoughts with the court as an amicus brief.

Today Judge Lance Walker, who has turned down several previous challenges, holds an evidentiary hearing. It may provide a first look at how the case will unfold.

Update: Basically, the plaintiffs’ attorneys claimed that their voting rights were impaired by ranked-choice ballot. Notably, all four plaintiffs were Poliquin voters – and therefore all of them had their choices registered in the final tally, no matter how they marked their ballot.

At the end, Judge Walker seemed to say that for there to be a justiciable controversy, voters must be burdened in a way that is specific to the Ranked-Choice Voting Act (as opposed to a general hazard of any kind of voting). He said that in plurality elections, difficulties are not enough to make a constitutional harm. Then he wanted to know: is there some threshold fraction of voters who were confused to constitute a harm that requires intervention? Counsel for plaintiffs, Obermeier, said that the threshold could be very low on account of how important the right to vote is. Overall, Judge Walker was skeptical that this case was any different from a 2018 case, Baber, in which he held that RCV did not pose a constitutional harm.

Tags: 2018 Election · House

9 Comments so far ↓

  • 538_Refugee

    BS has to be absurd enough to be amusing before I can read it. Want to clue me in as to whether some of the partial choice ballots were akin to ‘none of the above’ votes? I’d think a lot of people voting for minor candidates knew exactly what they were doing.

    • Raysgirl

      I voted for Tiffany Bond first, Golden second and Hoar third. I did VERY DELIBERATELY not rank Poliquin, even last. I knew it would never become a vote, but it was still a ranking. I would have left the race blank had he been the sole candidate. That was my choice and my right. I believe you are absolutely correct about those who voted for only the two minor candidates. The campaign among the major party candidates was beyond insulting. I ranked Golden in case he was in the position he was at the end of the day. I clearly prefered him to Poliquin and a blank or non ranking was one vote Poliquin would not have to overcome. I was happy with the outcome. I did informational sessions for the seniors in my area so they could see sample ballots before they hit the polls, and don’t count that segment out. We had a great discussion and they knew what RCV was and how to mark a ballot.

    • Sam Wang

      Raysgirl, you are exactly the kind of person the Ranked-Choice Voting Act was designed for.

      As long as you list a major-party candidate (in this case Jared Golden) anywhere among your choices, then your vote is likely to register in the final round.

      In your case, by listing all three other candidates (Golden, Hoar, Bond), you indeed did implicitly rank Poliquin last without having to name him.

    • 538_Refugee

      I was able to stomach a little more reading.

      Let’s see. If you vote for Nader on a once choice only ballot isn’t your vote “exhausted” anyhow?

      I’ll just have to admit to being a low information participant here. I just can’t read any more of that.

      Sounds like Raysgirl could file her own amicus brief.

    • Sam Wang

      Yes, exactly. Compared with plurality ballots, by their very nature RCV can only reduce the risk of exhaustion.

    • raysgirl

      @Sam Wang, you are correct that implicitly I ranked Poliquin last. However, without naming him, if ballots were examined, e.g. a court case, he could not be counted as any sort of choice.

    • Sam Wang

      Having sat through hours of expert testimony and examining Maine’s rules, as it turns out you are wrong! The definition of a complete RCV ballot is to rank n-1 out of n candidates. Happily for your preference, under Maine law your vote counted in the final round as a vote for Jared Golden. You were part of the win over Bruce Poliquin.

  • Mitch

    Choosing NOT to rank more than one candidate on an RCV ballot is the same as not showing up to a subsequent round in a multi-round runoff election.

    Which we also know what happens, with less participation than RCV, and with terrible results.

    The problem with multi-round runoff elections was illustrated quite clearly in the 2017 Lewiston, Maine Mayoral Race.

    The eventual winner had FEWER votes than the Candidate who was leading in the first round.

    In essence, the voters didn’t determine this election, voter participation did.

    In the second round of voting…
    Shane Bouchard won with 3,663 votes to Benjamin Chin’s 3,518 votes
    Out of 7,186 voters from 26,795 eligible voters.
    27% voter turnout.

    However, in the first round of voting:

    Benjamin Chin lead with 4,239 votes compared to Shane Bouchard’s 2,979 votes.
    Out of 10,148 voters from 26,795 eligible voters.
    35% voter turnout

    Ranked Choice Voting results will always show a gain in the number of votes for the winning candidate from the first to the final round.

  • Nathan Rasmussen

    Wait, let me see if I’ve got this right.

    All four plaintiffs are in their 70s and prefer Susan Collins (R) for the Senate. They complain as follows:

    I. One does not understand how to both maximize support for the one candidate he likes, and keep his ballot from exhaustion. This is like saying (in a traditional single-choice election) that you don’t understand how to both maximize support for your top choice, and avoid voting for someone who doesn’t win. You decide whom to support. You do not decide whether the rest of the electorate gives any chance of victory to the candidates (or traditionally, the one candidate) that you do support. And you never will, because democracy.

    II. Two more object to giving Sara Gideon (D) a rank on their ballots at all — something neither RCV nor the ballot instructions force them to do.

    II-A. Leaving one candidate unranked is fully equivalent to ranking that candidate last. In particular, either will guarantee that your ballot can never count in that candidate’s favor. This ought to satisfy the plaintiff who objects to ranking Gideon specifically, but plans to rank her below the also-rans out of necessity.

    II-B. Leaving multiple candidates unranked indicates dispreferring them all, equally, to those that you ranked. It guarantees that your ballot will never count toward any of them, with the corollary that it will be exhausted if (and only if) they are the only ones left in the race. (If such exhaustion would upset you, you do not disprefer them equally after all — but you can prevent it by ranking them accordingly.) This ought to satisfy the plaintiff who claims to object on principle to ranking any candidate except Collins (but who nevertheless also plans to rank all candidates, Gideon last).

    II-C. These objections reduce either to the bogus “avoid exhaustion” one, or directly to “people run that I don’t like.”

    III. The fourth, an engineer and former state representative with multiple college degrees, usually claims to understand “2nd, 3rd, 4th choice candidate” as meaning “to whom my ballot should go in the 2nd, 3rd, 4th round of counting,” even after consulting experts for clarification.

    III-A. Yes, a poor explanation of how votes are counted would burden the right to vote, but that is not the harm he complains of.

    III-B. And yes, the system that he claims to think RCV is would impair effective, purposeful voting for those who see merit in more than one candidate. But he sees no merit in any candidate but Collins, and he intends to vote in a way that will express that view as effectively in real RCV as in the system he professes to think he’s voting in.

    III-C. In fact, plaintiffs don’t really even call the justice of that putative system into question, do they?

    (Please, someone tell me again that those dang liberal millennials want a ruling elite instead of respecting the electorate (see I), whine for ‘safe spaces’ where they can cower from the mere mention of things they don’t like (see II), and ultimately just can’t grasp facts and logic (see III).)

    Ranking just one candidate in RCV is at least as expressive as a traditional single-choice election; the worst case is identical (no additional ballots accrue before your candidate is dropped and your ballot exhausted), and many cases are better (because people with rarer first choices than yours may join you, especially if they know how their candidate is polling). And of course, ranking a second choice potentially offers you expression in circumstances where your first choice could not sway a traditional election at all.

    Voters may choose, as plantiffs did, to express themselves as far as a traditional election would permit, but not further.

    Since RCV offers further opportunities for expression, they are not fully participating voters, says their expert witness. Therefore RCV reduces voter participation.


Leave a Comment