The case against "rank choice" voting

As regular readers here know, I’ve been stewing a lot these last couple of weeks about “rank choice” voting, which the City of Portland has adopted for municipal races, and which Oregon voters will be accepting or rejecting in the coming weeks for statewide races. I need to move on to other things in life. And so I’ve saved for this post all of the problems I have with it. I’m hoping for catharsis.

I’ve never liked the idea of “rank choice” voting. To me “one person, one vote” has worked pretty darn well, and there’s nothing there that needs fixing. But now that I’ve gotten deep into the weeds of how votes actually get counted under “rank choice,” I absolutely hate it.

Here is why. I’ll start with my problems with the relatively straightforward version of “rank choice” voting, which the city is about to use in the mayoral race and which Oregon voters are considering (in Measure 117) for many state races (but curiously, not in the races for the state legislature, hmmm). Then I’ll get into the zany version of “rank choice” that’s going to be used in the Portland City Council races. It’s even worse.

* * * * * * * * *

Why is “rank choice” voting bad news? Here’s what I think after learning all about how it actually functions and writing about it (among other places) here and here.

1. A large segment of the electorate doesn’t understand how it works. The government and the media have tried hard to explain it, with childish examples and colorful charts. But you and I could write a simple quiz about it that more than half the voters would flunk. The biggest truth that a lot of people are going to miss is that if you rank somebody the last of your six choices, you’re not voting against them, you’re helping them win. Too many voters will get that wrong this time around, and it will be years before everybody’s got the message. In the meantime, we may get results that are contrary to the voters’ intent. That’s the opposite of democracy, folks.

2. A large segment of the electorate doesn’t trust it. And can you blame them? The only way the election officials can tabulate the votes in a “rank choice” contest is by computer, which will be using software that I bet will not be released to the public for inspection. Hand recounts will be an unbearably tedious and expensive prospect, and no one will be willing to pay for them. Voters will just have to trust the tech bros and county bureaucrats to get it right. And that trust just isn’t there. Which brings me to:

3. It’s the worst possible time in modern U.S. history to be fiddling with the rules for vote-counting. Donald Trump and his hideous entourage have done a great job, on behalf of Vladimir Putin, of weakening confidence in American elections. One answer to the baseless challenges has been to say, “We’ve been doing this a long time. We know from long experience what we’re doing.” But with the switch to “rank choice,” that’s not true, and everyone will be questioning anew. And rightly so.

4. Compared to a runoff, some voters are disenfranchised. They call the “simple” version of rank choice “instant runoff,” but as I showed in this post, it’s not like a real runoff. In a real runoff, every voter gets a say between the two finalists. In “rank choice” voting, many voters’ ballots are “exhausted,” and play no role, by the time the two last candidates standing compete in the final “round” of tabulation. As just noted in point 1, you shouldn’t rank somebody you don’t want elected. And the fewer the candidates you rank, the greater the chance your ballot will be “exhausted” before the final showdown.

5. In a single-winner contest like Portland mayor, a candidate can win with less than 50 percent of the total number of ballots counted, even including second- and later-choice votes. As my post showed, in the final round, the winner need only get a majority of votes counted in that round. And that target will be less than the majority of all the ballots received with a vote for mayor on them.

6. Not everyone’s second choice (or later) is counted. The way they count second and later rankings is to count them only on the ballots of the first-choice supporters of candidates who have been eliminated. Candidates get eliminated from the bottom up. The fans of the lowest-scoring candidate in the first round get their second choices counted before anyone else’s. Once the loser candidates’ fans’ second or later rankings push another candidate over the top, the bureaucrats stop counting and go home. At that point, no one else’s second or later choices matter. In that sense, the fans of the fringe candidates take priority over the fans of the more mainstream candidates.

7. It’s spendy to administer. In addition to the setup costs, which are substantial, jurisdictions running “rank choice” will doubtlessly encounter higher costs than the old primary-and-runoff deal. In Portland, for example, Clackamas and Washington Counties will have to hand-deliver the paper ballots from their small batches of Portland residents to Multnomah County for scanning and tabulation. And if a hand recount is needed, the “rank choice” rules are going to make the process extremely slow and expensive.

8. It’s being foisted on the public primarily because the proponents are unhappy about the results of past elections. The rhetoric in favor of “rank choice” voting is mostly about who has won and lost historically. The supposed evil is that candidates favored by people with money tend to win. Candidates favored by only small segments of the population lose. The people pushing “rank choice” want to change that. And they’re willing to sacrifice public confidence in elections in order to Fight The Man™ and push back against majority rule and capitalism. I’m sorry, but I don’t think that’s a valid reason to adopt a new way of choosing among candidates for public office. Find better candidates.

* * * * * * * * *

The “rank choice” system that’s being used for Portland City Council elections is way kookier than what’s being used for the mayor’s race (and future city auditor races, if they’re contested). The list of flaws with “multi-winner" rank choice includes everything just discussed, plus these additional faults:

1. The tabulation is extraordinarily complex. The “rank choice” system for mayor is complicated and suspicion-generating enough, but the multi-winner system is of truly mind-numbing complexity, and the software code is even more of a black box. The pushers of “rank choice” keep pointing out that the only math functions involved are addition, multiplication, and division. That’s true, but the sheer number of different calculations is what boggles the mind. Here is a sample of the results report that will be generated for each Portland City Council district, multiple times, from Election Night until the final count. It will take your breath away. Can the average person ever hope to understand it? No. Will the average person trust it? No.

2. It’s resistant to recounts. Any “rank choice” hand recount is going to be tedious, but in the Portland City Council races, with three winners to be declared in each district, it will drag on forever and really test the skills and attention spans of the elections folks. And who will pay for the recount? Most candidates won’t have the kind of money it will take to finance weeks of recounting.

3. Three seats but only one vote. In the City Council races, three commissioners will be elected in each district. But voters get only one ballot. And many voters’ second and later choices won’t be counted because their favorite candidate will make it to the last round of voting. If you voted as your first choice either the winner or the loser in the final round, only your first choice counts. You have nothing to say about any of the other candidates. This is truly insidious, because other candidates get eliminated without any input from you. For example, say the lowest first-choice vote-getter, Candidate X, was the second choice of every voter who didn’t mark her as their first choice. And assume that no candidate had enough first-choice votes to win in the first round. Then Candidate X is eliminated, and none of the second-choice votes for her count at all. It gets even weirder when winning candidates’ fans’ later choices get counted, only as a fraction, and sometimes as a fraction of a fraction. But leaving that aside, the elimination of Candidate X in the first round when every second-choice ranking is in her favor is ridiculous. And so is giving no weight to the second and later choices of the two finalists for the last of the three seats.

4. Supporters of stronger candidates are disadvantaged. As with the mayoral tally, the second and later choices of the fans of the weakest candidates get counted as full votes – sometimes all six of their choices count. The stronger candidates' fans' second and later choices are counted only as a fraction, or not at all. It’s like Robin Hood, only with votes.

5. In a Portland City Council contest, a candidate can win a seat with less than 25 percent of votes cast, even counting second- and later-choice votes. As this post showed, in the final round, the winner need only get more votes than the other finalist. And that can be less than 25 percent of all ballots cast.

* * * * * * * * *

There’s probably more to gripe about, but that’s about all I have the brains and energy to say at this point. I think “rank choice” is a flawed idea, and its execution is bad for democracy. 

If unlike me, you’re really so dead set against runoffs and need a system that gives people ranking power, there are other setups you could try. How about a system where all six choices of all voters count, with the first choice weighted as 6 points, second choice weighted as 5 points, third choice weighted as 4 points, etc., and the candidate (or three) with the most points wins? Or perhaps give each voter 21 points and let them spread them among anywhere between 1 and 6 choices? It wouldn't be any daffier than what we've got in Portland now. And it might be fairer and more democratic.

Ah, me. This being Portlandia, we’ll probably have “rank choice” with us for a while. But like Measure 110, it will doubtlessly lead to some disaster or other, and then the talk of repeal will begin in earnest. Meanwhile, I sure hope the voters statewide are smart enough to say no to Measure 117.