However, due to some peculiarities of the state constitution, the state Supreme Court issued an advisory opinion that IRV is unconstitutional for 3 of the 7 kinds of state elections: for Governor, for State House, and for State Senate. I believe that this decision was poorly reasoned, but that doesn’t matter, because I’m not on the Maine Supreme Court.
In response, the state Senate has now passed a bill that would repeal the initiative in its entirety. The state House, on the other hand, has passed a amended version of that bill that would repeal the unconstitutional portions, but implement the portions with no constitutional issues, including for primary elections and for electing US Senators and Representatives. The Senate refused to consider the amended bill or to form a conference committee to create a compromise. So the session ends tomorrow, and the bill will die, leaving IRV in limbo: still a law on the books, but with some portions declared unconstitutional and no funds yet appropriated to begin implementation of the remaining portions.
It’s not clear to me when the Maine Legislature could begin this implementation. It appears that last year from July to December, they only met for two special one-day sessions, not beginning regular work until January. That would leave a tight schedule for passing a bill and implementing IRV before the 2018 primaries in June. I believe they should resolve this issue before then.
The people of Maine have spoken. They chose a better voting method, and insofar as there are no constitutional issues, they deserve to get it. The House chose the correct way forward. If you live in Maine, please contact your state Senator and Representative and urge them to support a special session to re-pass the House version of the bill and begin implementation.
What’s the problem with our voting method, and how do improved methods like IRV solve it?
Almost everywhere in the US, we use the worst voting method of any democracy: plurality voting, also known as First Past the Post (FPTP).
(Note: the term “voting method” refers to mathematical rules for what choices voters make on ballots and how those choices are combined to find a winner. “Electoral system” is a broader term which encompasses both voting method and also other rules such as voter and candidate eligibility, counting and recounting procedures, technology and voting machines, campaign finance rules, etc. US electoral systems have many problems besides the voting method; but this article focuses on just the voting method, where our system is not just bad, but actually one of the worst in the world.)
What’s so bad about FPTP? Because you’re only allowed to vote for one candidate, spending that vote on any candidate besides one of the top two is strategically disfavored. That problem leads to many others. If too many people choose such a “third option”, you can get a spoiled election, where a candidate wins even though they’re hated by a majority. On the other hand, if people are too afraid of spoiled elections, then just being one of the two perceived frontrunners becomes a self-fulfilling prophecy; other voices are shut out of the debate, and incumbents can become complacent and even corrupt. Finally, a two-way race is a zero-sum race, so negative, mudslinging campaigns become common, and turnout falls — which is usually bad for Democrats, because older and whiter citizens are often among the most reliable voters.
These problems are very real in the US in general and in Maine in particular. The issues with narrowed debate, negative campaigning, and low turnout are familiar to everyone here. I’ll also quickly remind you of a few high-profile cases of spoiled elections.
In the US: There is good reason to believe that both Bush 43 and Trump 45 only won the presidency because of spoiled elections in key states; that is, that there was some other candidate who could have easily beaten them in a one-on-one race. I am NOT interested in relitigating the 2016 Democratic or Republican primaries here, but I can say both that without FPTP I don’t think Trump would have been the Republican nominee, and also that I’ve done an in-depth statistical analysis of a high-quality pre-election poll that suggests that if we’d had a general election method able to accommodate more than one Democratic nominee, Trump would have lost badly.
In Maine: Governor Lepage has won twice in probable spoiled elections. In 2010, he won with just 38.1%; in 2014, he got 48.2%. LePage is, of course, the guy who claimed that drug dealers “are guys with the name D-Money, Smoothie, Shifty; these types of guys, they come from Connecticut and New York, they come up here, they sell their heroin, they go back home. Incidentally, half the time they impregnate a young, white girl before they leave, which is a real sad thing because then we have another issue we have to deal with down the road.”
There are a number of possible ways to fix these problems. If you’re electing a whole legislature at once, there’s Proportional Representation (PR) methods like those used in a majority of other democracies — methods such as STV with multimember districts (as proposed in a bill by Rep Breyer), or New Zealand-style MMP, or even newer methods like GOLD voting. For single-winner elections, there are methods like:
- Approval voting: each voter can vote for (“approve of”) as many candidates as they want, and the candidate approved by the most voters wins. Approving of more candidates does not give you extra voting power; in a 10 candidate race, whether I approve one candidate or nine of them, you can counterbalance my ballot exactly with one ballot by approving the ones I didn’t.
- IRV: Each voter ranks the candidates in order of preference. The first-choice votes are tallied. If no candidate has over 50% of the remaining valid votes, the one with the lowest tally is eliminated, and their ballots are transferred to the next preference (if any). These eliminations continue until one candidate has over 50% of the remaining ballots — which must happen by the time there are just 2 candidates left.
- 3-2-1 voting: Each voter rates each candidate “Good”, “OK”, or “Bad”¹. The semifinalists are the 3 candidates with the most “Good” ratings²; the finalists are the two semifinalists with the fewest “Bad” ratings; and the winner is the finalist who is rated higher on more ballots.
Of these three methods, IRV, the method passed by Question 5 in Maine, is clearly the weakest. It is better than FPTP, because at least it deals well with minor spoilers who get under 25% of the vote. However, paradoxically, if an extremist candidate gets more than that, they can still act as a spoiler by causing the centrist to be eliminated prematurely. That actually happened in Burlington, VT in 2009, and IRV was repealed soon after.
Approval is better than IRV; it’s simpler, it doesn’t require centralized counting, and simulations suggest it would have a higher Voter Satisfaction Efficiency. However, it’s still unsatisfying to some voters, because it doesn’t allow you to distinguish your favorite candidate from one you think is an acceptable compromise.
3-2-1 voting is better still. For a large majority of voters, it would let them make all the distinctions they care about the most, while still having a relatively simple ballot and counting procedure that don’t give an unfair advantage to strategic voters. Unlike IRV, it doesn’t prematurely eliminate centrists in a “center-squeeze” scenario; unlike Approval, it provides a safe, honest, and strategically strong way to vote in a “chicken dilemma” where the supporters of two similar candidates have to cooperate in order to beat a third dissimilar candidate (think Hillary and Bernie versus Trump).
So Maine, by choosing IRV, didn’t go for the best method possible. But they were making an improvement. And repealing their reform would definitely be backsliding, as well as being a terrible precedent for voting reform in general.
Constitutional issues
Why did the Maine Supreme Court say that IRV is unconstitutional for electing Governor, State Representatives, and State Senators? Well, when the Maine constitution was originally written, it required a “majority” for these offices. Then, in 1878, no candidate for Governor got a majority, and “Fusionist” (Democratic and Greenback) legislators tried to choose Democrat Alonzo Garcelon, who’d come in third place, as the governor. A standoff resulted, in which Civil War hero Joshua Chamberlain dramatically faced down an armed mob and said:
It is for me to see that the laws of this state are put into effect, without fraud, without force, but with calm thought and sincere purpose. I am here for that, and I shall do it. If anybody wants to kill me for it, here I am. Let him kill!
Nobody did kill Chamberlain there, and soon after, the Maine Supreme Court ruled that the Republican first-place finisher should be seated. To prevent this from happening again, the state constitution was changed in 1880 to require that the candidate “with the highest number of votes” (a plurality) should win for the three state offices mentioned above.³
So, after Question 5 passed, the Maine Senate asked the Supreme Court for an advisory opinion on IRV’s constitutionality for these three offices. The Supreme Court found that it was unconstitutional, because by requiring a winner to have a majority (over 50%) in the last round of counting, it prevents anyone from winning with a plurality (“the highest number of votes”).
In fact, I’d argue that this decision is mistaken, and that it’s actually the fault of FairVote, the organization behind Question 5 in the first place. (Full disclosure: though I’ve been a member of FairVote in the past, I’m currently involved with Electology.Org, a rival voting-reform charity which sees methods like approval voting as superior to IRV). You see, one of the arguments FairVote uses to promote IRV is that it supposedly “guarantees a majority”. But that’s only true if all voters rank all candidates. If some voters only rank one or just a few candidates, their votes can become exhausted, so that over 50% of the non-exhausted votes in the final counting round can be less than 50% of the full original number of votes. So sometimes the IRV winner does just have a plurality, and I think it should pass constitutional muster in Maine. By hiding this possibility — in my view, somewhat dishonestly — FairVote made the Supreme Court’s finding of unconstitutionality more likely, scuttling their own reform.
But that’s not important, because no matter how mistaken their logic, if the Maine Supreme Court says it’s unconstitutional, it is. Still, note that even by the logic of the Supreme Court, approval voting would be constitutional for all races in Maine. (As for 3-2-1, it’s debatable; so approval is probably a better option there.)
Legislative response
So the legislature had to decide what to do. Would they amend the constitution to allow IRV for all races; allocate funds to implement IRV for the races without constitutional issues; or would they just repeal the whole thing?
They tried an amendment first. This passed both houses, but failed to get the necessary ⅔ majority in a final vote.
Then, as noted above, the Senate passed a repeal bill. The House took it up and, after one of the original sponsors heard from voters and changed his mind, amended it to keep IRV for the races without constitutional problems. This amended bill passed, but the Senate refused to take it up again or send it to conference. So as of tomorrow, IRV will be in legal limbo in Maine.
What to do?
This is an unacceptable situation. So if you’re a Maine voter, first look up your state representative and state senator, then contact them using the switchboards at 1-800-423-2900 (Rep) and 1-800-423-6900 (Sen). (You’ll probably get an answering machine, so you need to know the names of your legislators before calling). Ask them to support a special session to allow IRV implementation to begin. If you feel like it, you can also mention that approval voting would be a good alternative to IRV for the three offices with constitutionality issues.
Summary
- FPTP is a horrible system, and reforming it should be an important progressive priority. More democracy and higher turnout will help our side.
- IRV is not the best reform, but it’s a definite step up.
- Maine voters supported reform, and they deserve to see it implemented carefully, not hastily starting in January.
Footnotes
¹ If you just rate one candidate “Good”, you can leave any of the rest of them blank, and your rating for that blank candidate will be filled in by the public pre-declared rating they got from your favorite candidate.
² There are two extra rules about the semifinalists to rule out certain rare problem situations. First, one party cannot sweep all three semifinalist positions; if they would, the third semifinalist slot goes to the strongest candidate from another party. Second, the third semifinalist must have at least half as many “good” ratings as the top semifinalist; if they don’t, the top two go straight to being finalists. This prevents a weak unknown candidate from winning a closely-divided two-way race “by mistake”.
³ Actually, for State Representative, the constitution had already been changed in 1847. As for Chamberlain, he died in 1914 of an infection in his 50-year-old wound from the Civil War; so the shot that killed him was actually fired long before the 1879 standoff.