Rank-Choice Voting for President in Maine
Staring in 2020 the winner in Maine will be the majority preference.

Starting in 2020, voters in Maine will be allowed to rank presidential candidates when they cast their ballots. The NYT reports: Maine Voters Will Rank Their Top Presidential Candidates in 2020.
Gov. Janet Mills, a Democrat, announced on Friday that she would allow a bill recently passed by the Maine Legislature to become law without her signature. The first vote conducted under the new law will be the general election in November 2020.
Under the new system, voters will be able to rank as many candidates as they like in order of preference. The initial count will look only at their first choices, and if one candidate receives a majority, that candidate would win.
If no one receives a majority, however, the candidate with the fewest first-choice votes will be eliminated, and his or her votes will be redistributed to those voters’ second-choice candidates. This process will repeat until one candidate breaks 50 percent.
This could also be called Instant Run-off Voting (IRV). At a minimum, it means that the candidate who receives the the most overall support statewide, and in each of Maine’s congressional districts, will win its electoral votes (given the way the state allocates its EV’s).
While not a major shift given the size of Maine (and its district allocation system), it means that one state will allocate two electoral votes to the candidate with actual majority support within its borders (and the rest to the candidate who wins absolute majority support in congressional districts).
It a system, if used nationally, that might have had an effect on the outcome of the 2016 election given Trump’s narrow margins winning pluralities of the vote in MI, PA, and WI. Indeed it could have mattered in several other states wherein either Trump or Clinton won by only a plurality, including Maine.
Maine has had RCV for congressional elections since 2016.
Whoa. If this were adopted nationwide it could actually allow third parties to be viable.
In 2016 Hillary Clinton won by a majority of the votes in ten states: California, Illinois, New York, New Jersey, Vermont, Massachusetts, Connecticut, Rhode Island, Delaware, and Maryland. Trump won by a majority in Alaska, Texas, Idaho, Montana, Wyoming, North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Iowa, Missouri, Arkansas, Louisiana, Indiana, Kentucky, Mississippi, Alabama, Ohio, West Virginia, South Carolina, and Georgia—most of the states he carried albeit in general much less populous states.
@Dave Schuler: Sure, but I am not sure of your point, at least not in context..
FWIW: Trump won FL, PA, MI, WI, UT and NE-2 by plurality. Clinton won ME, NH, CO, NM and NV by plurality.
Many of those outcomes might have been different with RCV. Especially considering FL, PA, MI, WI, and NH were each ~a point difference.
Not much of a point, really. Just supplying a little context.
@Dave Schuler: Gotcha.
@Dave Schuler: @Steven L. Taylor: Trump did win ME-02 by an absolute majority.
@Kylopod: He did. But he won NE-2 by a plurality.
I am very much on board with RCV and wish we’d use it everywhere, especially in primaries. It would make third parties viable, decrease intra-party and inter-party sniping (since you want to be people’s second choice) and possibly make extreme candidates less tenable. If we’d had RCV in 2016, a good chance Trump isn’t the nominee.
@Teve:
More like the third party wouldn’t tend to siphon votes from the Democratic candidate. I don’t see how it actually changes anything. Might even help a candidate like Trump more.
@Richard L DeMent: It depends on what you mean by “like Trump”. If you mean a populist candidate would have more of a chance with RCV, that is true, and that would be an adequate description of Trump. If you mean that a candidate who casts themselves as a Messiah would have more of a chance, I’d disagree, because all-or-nothing plays less well when you can submit multiple choices.
@Richard L DeMent: @David S.: Really all it does is guarantee that the person who wins had to have absolute majority support. It does not help any particular type of candidate.
@Steven L. Taylor:
I’m not sure whether that’s true. As I’ve mentioned before, a poll in Feb. 2016 found that when Republican primary voters were asked to choose between Rubio and Trump, or between Cruz and Trump, either Rubio or Cruz beat Trump by double digits. That suggests that Trump’s support was extremely polarized: either you supported him or you didn’t care for him at all, whereas voters who liked Rubio or Cruz weren’t necessarily committed to voting for them. That’s why I agree with @Hal_10000 that RCV likely would have prevented Trump from being nominated. A candidate who’s a cult of personality is likely to benefit more from a plurality-voting system, whereas a candidate who’s broadly liked even among those who don’t vote for him or her will do better under RCV.
@Kylopod: I agree that in a rank-ordered situation that Trump likely would not have won (because I don’t think he ultimately had majority support once second, third, fourth, etc. preferences were taken into account).
That comports with what I said in the comment.
I could be a bit more precise: RCV should produce a winner who has majority support once ranked-order preferences are taken into account.
That depends entirely on how big the cult of personality is.
@Steven L. Taylor:
I don’t think it does. Such a candidate may have majority support, and therefore could win under RCV–but non-cult candidates are likelier to invite thoughts like “I like this candidate, but he’s not my first choice,” and therefore have a bigger advantage under RCV.
@Kylopod:
If a candidate has 50% +1 of support, they win.
So if that is how big the cult of personality is, that candidate wins under RCV.
What am I misunderstanding about your point?
In the primaries, it might help lower-tier candidates, not so sure it makes a difference in the national vote for president other than guarantee a winner with 50% + of the vote.
And I think Steven T. is spot on about the size of the cult of personality.
But I also think that people are trivializing how confusing this would be to a certain non-insignificant number of people like the elderly. People have a huge problem with “pick three” of the following number of candidates. Heck, even candidate order on the ballot has a noticeable effect to the point where some states have experimented with random ballot order.