Author’s note: A few weeks ago, I detailed why approval voting would be a risky venture for Seattle. Then, I argued that approval voting would start no sooner than ranked choice voting in Seattle, if voters chose it in November. This time: a voting-rights law firm examines the two options’ legality.
In November, Seattle voters will consider whether to stick with their status quo electoral method or switch to approval voting (AV) or ranked choice voting (RCV) for their primary elections. They may wonder what the odds are that either option will fall prey to court challenges.
Sightline retained a Seattle-based law firm with a wealth of experience in campaign finance, election, and voting law to examine the legality of the two systems and offer an opinion, based on an extensive reading of federal and state law and court decisions.
Download the full legal memo prepared by Barnard Iglitzin & Lavitt LLP.
In short, the analysis concludes that
- RCV is safe from legal challenges;
- A state court could reject AV before it took effect under a provision in Washington’s election law;
- After it took effect, AV could be vulnerable to court action under the federal or Washington state voting rights act; and
- A court decision against AV does not seem likely but is possible.
The legal opinion, prepared by attorneys Jacob Harksen and Dmitri Iglitzin at the law firm of Barnard, Iglitzin, and Lavitt, reads, in part:
RCV has repeatedly withstood challenges brought against it based on state and federal law. Relative to approval voting, it is a known quantity, the legality of which is not seriously in doubt. Approval voting, in contrast, has been adopted by only two U.S. cities, and has not been subject to legal challenge in either. This means there are no cases evaluating approval voting’s legality. . . .
Approval voting applied to Seattle’s top-two primaries, as proposed by the November measure, might run afoul of the federal or Washington state voting rights act because it could
. . . allow a majority voting bloc to elect all of its preferred candidates, leaving non-majority voters no meaningful opportunity to elect their candidates of choice. Currently, a majority voting bloc can only advance one of its top two candidates in each primary to the general election, because voters only have one vote each. Under approval voting, the same majority bloc could choose both of the top two candidates in each race, if the majority was organized sufficiently, because each voter can vote for as many candidates as they choose.
Also possible, according to the firm’s analysis, would be a court rejection of AV even before it was implemented. Harksen and Iglitzin wrote:
. . . a different provision of Washington law may pose a more immediate threat to approval voting: a single sentence in the state’s law on primary elections prohibits voters from casting “more than one vote for candidates for a given office.” . . . [T]his law poses at least a potential risk to approval voting, but not RCV.
The analysis does not suggest that either type of court loss for AV would be likely. To the contrary, voting rights cases are hard to mount and win, and the sentence in state law that seems to ban AV has never been interpreted or called upon by a court.
Still, AV faces at least a modicum of legal risk, while RCV faces none.
The full 14-page legal memorandum is available for download here.
Clay Shentrup
There’s simply no evidence for your claims.
Here is the complete text of the voting rights act (VRA). It says nothing of the sort. It’s very short. I suggest everyone take five minutes to read it.
https://www.rangevoting.org/VRAtext.html
This claim by your lawyers is also deeply misguided:
> “there are no cases evaluating approval voting’s legality”
Apparently they weren’t aware that “at-large” elections are already used across the country, in which some voters vote for several candidates, while others only vote for 1 or 2. That is, different voters casting different numbers of votes. This is not categorically different from approval voting, insofar as it relates to voting power. If anything, approval voting is *less* vulnerable to claims of voter inequality, because the unlimited number of voters means that any voter can cast an equal-but-opposite vote to exactly counterbalance the effect of any other voter.
At-large districts have been questioned for their potential to give 100% of seats to 51% of voters (which isn’t an issue in these single-winner elections in Seattle), but no one questions that it’s perfectly legal for different voters to vote for different numbers of candidates.
Ironically, we DO have evidence that Ranked Choice Voting suffers from this problem, because it counts some voters’ X-vs-Y opinion, while ignoring others’ X-vs-Y opinion. For example, in the recent Alaska House election, Peltola won even tho voters preferred Begich to Peltola by a significant 5% majority. Begich was removed before most of those preferences had even been looked at however—so it was impossible to see that nearly all Palin voters preferred Begich to Peltola. This is due to the later-no-harm flaw in RCV, which is described in detail by a Princeton math PhD and voting methods expert here.
https://www.rangevoting.org/IrvIgnoreExample.html
It gets even worse.
> [approval voting could] allow a majority voting bloc to elect all of its preferred candidates, leaving non-majority voters no meaningful opportunity to elect their candidates of choice.
This is literally the exact opposite of the truth. It’s this proposed variant of ranked choice voting (technically “bottoms up”), which has never been used in the USA, which has this very problem! Because it generally pits the majority choice against the “opposition” choice, which will produce lop-sided victories in which the majority-preferred candidate will generally trounce the minority candidate in a meaningless uncompetitive general election.
Prominent RCV advocacy organization FairVote even says:
“One unique risk of bottoms-up RCV compared to other forms of RCV is that a minority of voters may be able to control a majority of seats.”
Of course, they’ve taken that page down recently, probably because of this very argument.
Whereas approval voting will pit the two most broadly appealing *consensus* (not “majority preferred”) candidates head-to-head, leading to a competitive election in which every vote truly counts. This enfranchises minorities. This is exactly what we saw in the 2020 St Louis mayoral race, where two progressive woman—who clearly aligned with the median St Louis voter rather than with the median of the majority faction of the electorate—advanced to the general, and had a competitive showdown, giving every faction of the electorate the chance to count and potentially push their preferred winner over the line.
By your own stated goals, approval voting is far and away the superior option to this never-before-used-in-the-USA “bottoms-up RCV” process.
Alan Durning
Thanks for reading and commenting, Clay.
Regarding the existence of at-large systems in which voters can choose more than one candidate: are you referring to what we at Sightline call bloc voting or plurality at-large voting? In such contests, all candidates run in one big pool, and voters can vote for as many candidates as there are available seats. It is somewhat like approval voting, I agree.
The legal memo does discuss such systems, and it points out that in US v. City of Eastpointe, that city’s use of at-large, bloc-voting elections was found to violate Section 2 of the VRA. The remedy was RCV.
Clay Shentrup
Alan,
Eastepointe had a completely different issue, which I specifically preemptively addressed:
“At-large districts have been questioned for their potential to give 100% of seats to 51% of voters (which isn’t an issue in these single-winner elections in Seattle), but no one questions that it’s perfectly legal for different voters to vote for different numbers of candidates.”
I repeat: Eastpoint has multi-winner elections, whereas Seattle has single-winner elections. Apples and oranges. Indeed, some cities faced with Eastpointe’s predicament have moved to districts!
In addition, the remedy in Eastpointe was a substantively different method: proportional multi-winner Single Transferable Vote (STV). Whereas the “ranked choice voting” proposed in Seattle is the substantively different “bottoms up” RCV, which has never been used in the US, and is extremely unsuitable for use in single-winner elections, as I explained here.
https://clayshentrup.medium.com/ranked-choice-voting-wrong-for-seattle-e7e08349717e
Again, even FairVote says: “One unique risk of bottoms-up RCV compared to other forms of RCV is that a minority of voters may be able to control a majority of seats.”
Alan Durning
Clay,
I’m puzzled by your argument. You’re saying that AV is like at-large bloc voting; at-large bloc voting exists in the United States; and courts have allowed it. From these things, you conclude that approval voting is safe from voting rights act challenges in the courts.
But courts threw out Eastpointe’s system of at-large bloc voting under the voting rights act. So you draw a distinction concerning the grounds for its rejection: that it allowed the majority to elect all the representatives, not that it allowed voters to cast multiple votes. That’s a distinction without a difference. The method was to allow voters to cast multiple votes, and the outcome was that the majority could elect all the representatives, a voting-rights-act violation.
If at-large bloc voting is like approval voting, as you say, then court rejection of at-large bloc voting anywhere is a sign of legal jeopardy for approval voting. It’s not, as you suggest, a proof that AV is immune from legal jeopardy.
Clay Shentrup
Alan,
You continue to confuse two completely separate properties of at-large elections:
1. Some voters vote for a different number of candidates than other voters. This is perfectly legal, thus approval voting must be legal.
2. A mere majority of voters can pick 100% of the winners. This only applies to at-large elections, and is a known voting rights issue that affected Eastpointe, MI. This is completely irrelevant to Seattle, because Seattle uses single-winner districts.
> If at-large bloc voting is like approval voting, as you say, then court rejection of at-large bloc voting anywhere is a sign of legal jeopardy for approval voting.
Equally valid logic: Cocaine is white. Cocaine is illegal. Therefore sheep are illegal.
Alan Durning
Clay,
Thanks for your reply.
The attorneys we hired to analyze approval voting could not find any evidence of a court ever ruling that, as you say, “some voters vot(ing) for a different number of candidates than other voters” is perfectly legal. They found no rulings on that question. What they found was that at least one court threw out a system of voting that included precisely that kind of voting. You’re right: that voting system had other characteristics, including that it was at-large. But the court didn’t say, “the bloc voting part of this is totally fine, it’s the at-large part that violates the voting rights act.” The court did not make that distinction; you did. The court just threw out the Eastpointe system overall. The Eastpointe decision in itself neither damns nor exonerates AV. It simply leaves AV under a legal cloud.
Were AV challenged in Seattle or elsewhere, attorneys defending it could make the argument you outline above as a way to differentiate AV from the principles in the Eastpointe precedent. Perhaps the court would accept that argument. Perhaps not.
The point here is that we cannot know, so AV faces legal risk.
I’m not confused about the Eastpointe case, Clay, nor by your attempt to invent a distinction not made by the court. I confess, though, that I am a little confused why you are arguing so strenuously with a piece of legal analysis that concludes the risks facing AV are rather modest. I would have thought that’s a pretty good result, considering how new AV is, with only three uses in public elections so far. Most voting innovations, including not just RCV but also automatic voter registration, vote by mail, and nonpartisan, top-two primaries, have run through legal gauntlets when first adopted. AV will have to, as well.
Clay Shentrup
> The attorneys we hired to analyze approval voting could not find any evidence of a court ever ruling that, as you say, “some voters vot(ing) for a different number of candidates than other voters” is perfectly legal.
No one’s had to rule on it because it’s never been challenged. That’s the whole point. And it has massive historical precedence. We have absolutely no reason to think there’s any legal issue around approval voting.
> the court didn’t say, “the bloc voting part of this is totally fine, it’s the at-large part that violates the voting rights act.”
Of course they did! The whole issue was that the minority was shut out of a multi-winner election, because 51% of voters can control 100% of the winners. This obviously can’t be an issue in a single-winner district election, because there’s only one winner.
This is explained in detail by RCV advocacy organization More Equitable Democracy:
Note the crucial part there: “this occurred because the entire city got to vote in each city council race”. The issue was not that different voters voted for different numbers of candidates.
They even note that one considered remedy was moving to districts, like Seattle already has.
> The court did not make that distinction; you did.
You’re just clearly and obviously wrong on this. The court’s issue was about minority representation in multi-winner elections. If we elect 3 city council members in one big election, a 51% majority of voters can focus all their votes on 3 candidates and guarantee they pick all three winners, for instance. This obviously objectively mathematically has nothing to do with the number of votes each voter gets. It’s about having multi-winner at-large elections.
> The court just threw out the Eastpointe system overall. The Eastpointe decision in itself neither damns nor exonerates AV. It simply leaves AV under a legal cloud.
No, it doesn’t. The issue was with multi-winner elections. They even considered a move to single-winner district elections as a potential remedy, like other cities have done.
> Were AV challenged in Seattle or elsewhere, attorneys defending it could make the argument you outline above as a way to differentiate AV from the principles in the Eastpointe precedent. Perhaps the court would accept that argument. Perhaps not.
Whether the court accepts it or not is a separate issue. The objective reality is that the critical issue with approval voting—that some people vote for more candidates than others—has never been legally challenged in spite of having vast historical precedence. The issue in Eastpointe was about a completely unrelated issue that’s irrelevant to Seattle, because Eastepointe uses multi-winner elections whereas Seattle uses single-winner elections. Full stop.
> The point here is that we cannot know, so AV faces legal risk.
You’ve presented no evidence for this. You’re being intentionally obtuse. You know perfectly well the difference between a single-winner and multi-winner election.
> I am a little confused why you are arguing so strenuously with a piece of legal analysis that concludes the risks facing AV are rather modest.
Because your argument is clearly wrong and, in my view, appears to be politically motivated and disingenuous.
Alan Durning
Clay,
You wrote, “Because your argument is clearly wrong and, in my view, appears to be politically motivated and disingenuous.” I’m afraid that’s how I read your arguments.
But I will rest my case now. Readers can judge. Voters will decide. Perhaps in time, judges will decide about approval voting, as they have decided about RCV.
Clay Shentrup
A clarification:
> The objective reality is that the critical issue with approval voting—that some people vote for more candidates than others—has never been legally challenged in spite of having vast historical precedence.
Note that I meant specifically with regard to at-large plurality voting elections.
There was a 1915 challenge to multiple votes with the Bucklin system (like a hybrid of RCV and approval voting), previously used in about 40 US cities. However,
1. This ruling would have applied to RCV as well, based on the logic supplied by the court.
https://www.rangevoting.org/BrownSmallwood.html
2. Multiple votes in at-large elections has continued for over a century since then, with no legal challenge that your lawyers could find. (The challenge was to minority representation in multi-winner districts, not to how many votes each voter could cast.)
3. The reasoning was incredibly flimsy, and likely to be overturned by a higher court. This is especially obvious when you consider that lots of other states used Bucklin voting and had no such rulings apparently.