Wednesday, October 11, 2017

Supreme Court 2016 Statutory Term in Review: Perry v. MSPB, Punctuating Adverbial and Adjectival Phrases, and the Beach Boys

I've been promising people a post on Justice Gorsuch's noted dissent in Perry v. MSPB.  Much, however, as I might appear to relish taking Justice Gorsuch's opinions to task, I (a) don't, and (b) don't think there's much of real importance or methodological interest to say about the Perry dissent. However, a promise is a promise, I do have a few things to say that will probably sound familiar, and those of you who have been looking for a definitive answer to whether Gorsuch was right in Perry will get it.  So, here goes.  

As any reader of this post will no doubt recall, Justice Gorsuch began his dissent with the arresting claim that "Anthony Perry," the petitioner who would prevail 7-2, "asks us to tweak a congressional statute—just a little—so that it might (he says) work a bit more efficiently," and concluded it with the arresting suggestion that "[r]espectfully," the Court ought to "follow" the "perfectly good law" Congress wrote.  In the middle, he wrote there was "a constitutionally prescribed way" to address Perry's concerns and "it's called legislation," denounced "judicial tinkering with legislation," explained that the difficulties of "bicameralism and presentment" are "the point" of the Constitution requiring them ("the better to preserve liberty"), and said many other things in this genre.  (All of which I quite agree with, apart from its application to this statute.)

This predictably got a lot of unreflective attention.  No one, though, at least in anything published that I've seen, actually spent any time addressing whether Gorsuch was right.  The assumption seems to have been, as it usually is in such cases, that such a confident statement (especially one joined by Justice Thomas) that the Court had gone off the textual rails must be true or at least amply arguable; the problem at issue in Perry was too difficult, insignificant and dull to merit any commentary.

In arguing that the Court adopted a "seriously atextual" "tweak" of Congress's perfectly good law, Justice Gorsuch placed much of the blame for the Court's wayward interpretation on the much-pilloried Federal Circuit.  The Federal Circuit, he claimed, in a 1983 en banc, "adopted a fix [all emphases mine] much like what Mr. Perry now proposes: allowing civil service claims to tag along to district court with discrimination claims because, in its judgment, '[f]rom the standpoint of judicial economy, consideration of all issues by a single tribunal is clearly preferable.'"  What Gorsuch said the statute required was bifurcation; in an agency appeal where a federal employee raised both civil-service claims and employment-discrimination claims before the Merit Systems Protection Board, the MSPB's decision on the former claims would be reviewed in the Federal Circuit, while its decision on the latter claims would be reviewed in district court, simultaneously.

To read Justice Gorsuch's dissent, one would think the reason the Federal Circuit allowed civil-service claims to tag along to district court with discrimination claims, rather than taking up the civil-service halves of these sorts of (in the parlance of the field) "mixed" MSPB decisions itself, was "because" the Federal Circuit thought it pragmatically preferable, "in its judgment," "from the standpoint of judicial economy."  Justice Gorsuch doesn't quite make it clear whether the Federal Circuit even attempted to textually justify its "fix," or whether it copped to making a "fix" on the ground of its policy "judgment."  But I think it's at least fair to say that one would never guess from Justice Gorsuch's characterization of the Federal Circuit's decision that the Federal Circuit deemed itself bound by the statute's plain meaning to come out at the "seriously atextual" place Gorsuch says it did, or that pragmatic arguments from judicial economy played only a marginal role in the Federal Circuit's decision.

Well, one would never, as I say, guess it from what Justice Gorsuch wrote, but it's true all the same.  The fact of the matter is that the textual case for the interpretation that Gorsuch described as an atextual tweak is, if perhaps not quite unambiguously correct, compelling enough that the Federal Circuit argued with considerable force in an 8-3 decision that "[t]aking the language [of the statute] literally, one would conclude" that the matters Gorsuch argued were clearly within the Federal Circuit's jurisdiction, and could only be put outside it on atextual and dubious policy grounds, were "clearly beyond our jurisdiction."  

Influenced, however, by their sister circuits, which hadn't exactly found the statute ambiguous, but had in some instances found it difficult enough to write a few pages about it, they did pause to judiciously consider Gorsuch's preferred non-literal alternative; rejected it at great length as textually foreclosed, as had the Fifth, Tenth, First, and D.C. Circuits in interpreting a materially similar version of the statute for much the same textual reasons; and finally added in a brief policy-themed coda (Part IV of a five-part opinion, counting the conclusion) what Gorsuch represents and quotes as the sole driver of the Federal Circuit's decision.  A delightful and elegant concurring opinion by Judge Nichols argued that the case for the majority's position was "even stronger than the opinion spells out," that bifurcation was clearly foreclosed, and that it was "as unreasonable as it is contrary to the statutory language."

Finally, a thoughtful, though to my mind ultimately unpersuasive dissent "disagree[d] with the majority's assertion that [the statute], read literally, 'clearly' indicates" that appeals of this kind were beyond the Federal Circuit's jurisdiction, found a degree of ambiguity in the statute, argued that Congress never considered or addressed whether MSPB appeals would be severed or not (unlike Gorsuch, who claims Congress plainly mandated severance), and ultimately relied rather heavily on statements in committee reports on the general purpose of the Federal Circuit's jurisdictional grant in deciding that severance was the best way to handle this open question.  Moreover, after the Federal Circuit's decision, the Ninth Circuit, Eleventh Circuit, and Third Circuit all followed along for the same reasons.  (So did the Fourth, Sixth, twice, and Eighth, but all in rather cursory fashion, though that only underscores how easy they found the problem.)

No court ever disagreed with these eleven circuits; indeed, the question Gorsuch wanted to decide wasn't the question the Court granted cert to address, the parties both argued he was wrong, the government doing so against its interests, and the Court itself unanimously ratified the lower-court consensus five years ago in another case where the parties both took it as a given.  So to recap, the Federal Circuit majority thought itself textually compelled to reject Gorsuch's position, like four circuits before it, six after it, and the Solicitor General's office; the dissent, acknowledging the majority's determination of unambiguity, thought itself textually free to adopt Gorsuch's position, claimed the statute didn't resolve the question, and only ended up where Gorsuch did on the ground of a freewheeling purposivist frolic through the legislative history. 

Of course, in theory Gorsuch could have seen something in these difficult statutes that the Federal Circuit en banc and ten other circuits didn't see.  That, however, is emphatically not the case; at least if he did he's not telling us.  While there's a lot of fun rhetoric in his dissent, there isn't a word in it disputing the textual reasons that the Federal Circuit and nearly every regional circuit in the country thought that what Gorsuch called an atextual tweak was plainly the law; most of the relevant language isn't even quoted.  After reading the dissent a half-dozen times, I still don't know why he even thinks the statute ambiguous, much less why he thinks it means what it means.

The problem Gorsuch wanted the Court to take up in Perry isn't nearly as complicated as commentary on the case made out to be, though it admittedly isn't very interesting either.  First, the MSPB hears appeals from certain federal employment decisions under 5 U.S.C. 7701, including terminations or reductions in pay.  Under that grant of administrative appellate jurisdiction, it decides questions of civil-service law.  The MSPB's decisions in 7701 appeals are reviewable by petition in the Federal Circuit under 5 U.S.C. 7703(b)(1)(A). That much is undisputed.

Next, under 5 U.S.C. 7702, the MSPB also decides appeals in "case[s]," see 7702(a)(1), (b)(4), and (f), in which an employee "has been affected by an action," otherwise appealable to the MSPB under 7701, and "alleges that a basis for the action" was a violation of certain employment-discrimination statutes.  In a 7702 appeal, the MSPB is required to "decide both the issue of discrimination and the appealable action" as it would in a 7701 appeal.  (All those quotes are from 7702(a)(1).)

No one disputes, not even Gorsuch, really, that a federal employee can raise both civil-service issues and employment-discrimination issues in a single 7702 appeal, not just employment-discrimination. The three Federal Circuit dissenters whose position he rescued from early-80s obscurity wrote that "[i]t is uncontested that Congress envisioned that the MSPB render a single decision in a 'mixed' case."  And, indeed 7702 describes the product of such double-barreled appeals, in 7702(a)(3), as a singular "decision" and "judicially reviewable action."   

Justice Gorsuch, for his part, while not explicitly admitting the existence of these non-bifurcated MSPB proceedings, could at most only suggest that not every 7702 appeal or "mixed case" involves a mix of discrimination and civil-service issues, because some may only involve appealable employment actions, e.g., terminations, that solely raised employment-discrimination issues.  Entirely true as a matter of what 7702 says and means in practice, but equally irrelevant.  The question Gorsuch wanted to decide is whether the unitary 7702 appeals that do present a mix of civil-service and discrimination issues bifurcate on judicial review into two separate cases that land in two different courts.

Finally, section 7703(b)(1)(A) generally grants the Federal Circuit jurisdiction to review the MSPB's "final order[s] or final decision[s]," "[e]xcept as provided in paragraph (2) of this subsection," i.e., 7703(b)(2).  And what 7703(b)(2) says in relevant part is this:  "Cases of discrimination subject to the provisions of section 7702 of this title shall be filed under [enumerated provisions of employment-discrimination law], as applicable."

Now, what are these "cases of discrimination subject to the provisions of section 7702"?  Well, in the view of the Federal Circuit, ten other circuits, and the majority in Perry, they are quite simply the "cases" that the MSPB hears under section 7702.  Section 7702 talks about the MSPB deciding certain "cases," and describes those cases as ones involving appealable employment actions, an alleged basis for which is discrimination.  Again, not necessarily the only basis, but a basis, and the MSPB is required in such a "case" to "decide both the issue of discrimination and the appealable action" and render a unitary "decision" that "shall be a judicially reviewable action."  

So, the Federal Circuit argued rather fulsomely, a 7702 "case," whether containing only employment-discrimination issues or employment-discrimination issues and other issues, is a unitary case before the MSPB, and remains a unitary "case" and "judicially reviewable action" in district court, to which such "cases" are routed.  Perhaps it's a little odd to talk about a "case of discrimination" when one means a case presenting, among other things, discrimination claims, but that's just what 7702 says the "cases" that are "subject to the provisions of 7702" are.

Whatever made Justice Gorsuch think that this rather literal, even somewhat crabbed interpretation of 7702 and 7703 is not only wrong, but a "seriously atextual," willful "judicial tinkering" with the statute?  And what could have led such a self-styled textualist to believe that "cases of discrimination subject to the provisions of 7702" actually only means what 7702 calls "the issue of discrimination" that the MSPB decides in a 7702 "case," leaving the other half of that 7702 "case" or "judicially reviewable action" for review in the Federal Circuit?  

I have no idea.  Again, Gorsuch never says a word about the lower courts' plain-language interpretation of 7703, their harping on 7702's distinction between a discrimination "issue" and a larger "case" concluding in a "decision"/"reviewable action" on all issues before the MSPB, their argument that 7702 defines 7703(b)(2)'s reference to "cases of discrimination subject to . . . 7702," or any of the other textual points in the lower courts' favor.  He quotes the relevant language in 7703(b)(2), not pausing for a second over what it means; misleadingly paraphrases 7702 in three quotation-free sentences as a statute about pure "cases of discrimination," glossing completely over all the awkward language in 7702 about unitary cases, decisions, and judicially reviewable actions with embedded discrimination issues; and finally announces "the statutory scheme is plain": civil service disputes go to the Federal Circuit, and "discrimination cases go to district court."  That just doesn't cut it.

As far as I can tell, the one textual argument Gorsuch's got to support all his textualist Sturm und Drang is the fair enough point that 7703(b)(2) says that "cases of discrimination" are supposed to be "filed under" employment-discrimination statutes.  (Most of Gorsuch's arguments, ironically, are pragmatic or purposivist ones about the difficulties, obscurities and costs of district-court review of civil-service claims, which it's fair to say Congress failed to fully address.)  How do you file an appeal of the MSPB's decision on a civil-service issue under an employment-discrimination statute?  You can't, it would seem, which suggests, perhaps a little paradoxically, that 7703(b)(2) "cases of discrimination subject to . . . 7702" only involve the discrimination "issues" embedded within 7702's "cases."  

The idea that 7702 discrimination "issues" embedded in mixed 7702 "cases" blossom under 7703(b)(2) into "cases of discrimination subject to the provisions of 7702," while the other half of the MSPB's "decision" in the mixed 7702 "case" blossoms under 7703(b)(1)(A) into a freestanding "final order or final decision" of its own reviewable in the Federal Circuit; that the 7702 "case" and MSPB "decision" themselves undergo a process of 7703-induced fission; and that all this is made unarguably clear by 7703(b)(2)'s directive that "cases of discrimination subject to . . . 7702" are to be filed under discrimination statutes, is certainly inventive, if entirely unexpressed in Justice Gorsuch's opinion, which again fails to quote 7702 once or acknowledge the reality of multi-issue 7702 cases.  (A remarkable omission, given that 7703(b)(2) refers you to 7702 to see what the "cases of discrimination" it's talking about are.)  And to be fair, it's not a completely unreasonable view; the Federal Circuit dissenters ably argued the statute could be read that way.  But as far as textual points about the meaning of "cases of discrimination subject to . . . 7702" go, Justice Gorsuch has one decent one to the Federal Circuit's half-dozen, plus its rather sensible ancillary concern about bifurcation impairing judicial economy that he lampoons.  

Besides, even if the consensus reading of 7703(b)(2) didn't have vastly more text on its side than Justice Gorsuch has on his, Justice Gorsuch's one decent textual point is hardly compelling upon careful reflection.  One could quite reasonably, for example, write a statute providing that "cases of discrimination shall be filed under" certain federal discrimination laws in district court, even if another statute allowed plaintiffs to bring pendent state-law claims with their discrimination claims.  One wouldn't suppose that a statute like that meant that pendent state-law claims weren't part of the "case of discrimination" that belonged in district court.  Likewise, the fact that 7703(b)(2) calls for filing under discrimination statutes can't be fairly read to mandate—certainly not unambiguously sothat the entirety of the "case of discrimination subject to . . . 7702" must be filed under those statutes.

At oral argument, the government and Christopher Landau, Perry's counsel, both stood up to Justice Gorsuch's novel reading of 7703(b)(2).  There, Gorsuch was forced to confront the government's well-worn point that 7702 defines a case of discrimination subject to 7702 "as a mixed case that includes both discrimination and other components."  To this, he had a curious response.  

This argument, he said, "raises the question what 'subject to' means, right?"  That's true enough; the basis for saying that 7703(b)(2) "cases of discrimination" are the "cases" described in 7702 is the phrase "subject to the provisions of 7702."  But how many ways are there to read "subject to" in that sentence?  Cases of discrimination subject to the provisions of section 7702 just means, can only mean, cases of discrimination governed by section 7702, which is to say the cases of discrimination described by 7702.  The point of a reference to 7702 is to distinguish 7702 "cases of discrimination," the ones 7702 governs, from all the other discrimination cases in the world.  Justice Gorsuch, however, saw a different possible reading, and at that point things got confusing.  I'll just let Justice Gorsuch and the assistant to the Solicitor General take it from here:
JUSTICE GORSUCH: .  . . And -- and you're equating “subject to” with “meeting the test of.”  But “subject to” can also mean subject to.  It can be tested under.  Not that it meets the test, but it can be tested under, right?  That's often how Congress uses that phrase. 

MR. FLETCHER: I -- I understand.  I -- I may not be grasping in that sense.  I understand that you might use it to say this is true, subject to some other provision that might qualify it. 


MR. FLETCHER: I don't think that's how Congress used it here.

JUSTICE GORSUCH: It may or may not qualify, not that it does qualify. 

MR. FLETCHER: I guess -- 

JUSTICE GORSUCH: Often, it's used in that sense, right? 

MR. FLETCHER: I -- I'm not sure that it -- I -- I can agree with that. 

JUSTICE GORSUCH: Really? Why not? You just gave me a good example.

MR. FLETCHER: Then I think maybe I'm misunderstanding the question.  I -- I agree sometimes Congress says the rule is you go to district court, subject to, in this case, you can go to some other tribunal.  So it's describing an exception.

I think that that's not what it's doing here. Here, it's using that as a description --

JUSTICE GORSUCH: Do you have any authority for that proposition?

MR. FLETCHER: Kloeckner [a unanimous 2012 opinion of the Court rejecting Gorsuch's interpretation in either holding or dictum, though the Perry majority thought it holding], which I think says, you know, this is a--

JUSTICE GORSUCH: Besides Kloeckner. Anything else?

JUSTICE KAGAN [the author of Kloeckner]: Well, Kloeckner certainly says it, but beyond Kloeckner, I mean, merits cases that -- that have been -- have been going to district court for years prior to Kloeckner; isn't that right?
As somebody who once almost wrote a whole article on whether the word "under" in a statute meant "subject to," "by reason of the authority of," or "under the heading of" (the subject, in one instance, of a classic circuit split between then-Judge Ginsburg and Judge Reinhardt), I feel I should be able to understand what Justice Gorsuch is talking about here.  But neither I, nor seemingly Assistant Fletcher, nor anyone I know, can quite figure out what alternative sense of "subject to" he's talking about or how it could fit in this statute.  If I understand him, and it's a big if, "subject to the provisions of section 7702" isn't descriptive; rather, it means something like "if 7702 allows," so that the whole first sentence of 7703(b)(2) should be parsed, "Cases of discrimination, if allowed by 7702, shall be filed under the following discrimination statutes."  That seems to me to be a non-starter for a few reasons.  

First, and it's a rather nitpicky but both intuitive and grammatically correct objection, 7703(b)(2) doesn't contain commas before and after the "subject to" clause; it reads "Cases of discrimination subject to the provisions of section 7702 of this title shall be filed," not "Cases of discrimination, subject to . . . 7702 of this title, shall be filed . . . ."  Now, if "subject to" were an exception to "shall be filed," as Gorsuch seemed to suggest, it would technically be an adverbial phrase modifying "shall be filed."   If it means what I think it means, and what all the courts to interpret it have thought it meant, it's an adjectival phrase that modifies "cases of discrimination."  

If "subject to" is an adjectival phrase modifying "cases of discrimination," as I claim, the lack of commas makes sense, like saying "the bushes under the tree died" instead of "the bushes, under the tree, died."  Adjectival phrases following the nouns or noun-phrases they modify aren't set off by commas unless they're non-restrictive.  Adverbial phrases immediately following nouns or noun-phrases, though, are set off by commas, or they'd be mistaken for restrictive adjectival phrases.  

For example, "Asher, with some exceptions, is usually readable" needs commas because "with some exceptions" is actually modifying "is" in that sentence, not "Asher."  Consider a closer example: "Eastern European nations subject to Soviet control trade with the United States" means that Eastern European nations that are subject to Soviet control do such trade, but "Eastern European nations, subject to Soviet control, trade with the United States" means "Eastern European nations trade with the United States, subject to what's allowed by Soviet control."  So if 7703(b)(2) meant what Gorsuch suggests, it would have read "Cases of discrimination, subject to . . . 7702, shall be filed . . . ."  To claim it means what he says is to claim that 7703(b)(2) as written is a scrivener's error. 

Second, it's actually really hard to find a provision of 7702 that would carve out some exception to 7703(b)(2)'s mandate that "cases of discrimination . . . shall be filed" under discrimination statutes in district court.  Gorsuch, recall, seems to read 7703(b)(2) to mean that cases of discrimination, subject to 7702 allowing it, shall be filed in district court under discrimination statutes.  But 7702 is all about what the MSPB does; it does nothing to stop anyone from filing a "case of discrimination" in district court.  

The one relevant thing in 7702 to Gorsuch's proposed interpretation are several timing provisions about when the MSPB decision in a mixed case is deemed a "judicially reviewable action," but the problem with reading "subject to the provisions of section 7702" as a carve-out for untimeliness given 7702's rules about reviewability accrual is that 7703(b)(2)—you guessed itspecifically invokes those rules elsewhere.  To wit, 7703(b)(2)'s last sentence reads, "Notwithstanding any other provision of law, any such case filed under any such section must be filed within 30 days after the date the individual filing the case received notice of the judicially reviewable action under such section 7702."  So that's not what "subject to . . . 7702" can be all about.

Third, even if Gorsuch were right, we'd still have the following problem.  The Federal Circuit is granted jurisdiction over MSPB decisions in 7703(b)(1), less the carve-out for district-court jurisdiction in 7703(b)(2); that carve-out covers "cases of discrimination."  Bracketing "subject to 7702," cases of discrimination" still aren't any old "case of discrimination" under the enumerated federal laws, but some subset of MSPB orders.  Now, if one wanted to figure out what a case of discrimination that also involved review of an MSPB decision looked like, one would naturally be led to 7702, and there one would find that the MSPB renders unitary decisions, in "cases" presenting (exclusively or non-exclusively) discrimination "issues," that are treated as unitary "judicially reviewable actions," a phrase 7703(b)(2) picks up.  How, then, could one conclude that the 7702 "decision"/"judicially reviewable action" on all issues turns into a judicially reviewable action on discrimination issues only in 7703?  It's a real stretch.  

In sum, you can argue the statute unambiguously means what everyone's always said it clearly means except for Justice Gorsuch and the three Federal Circuit dissenters; you can possibly say it's ambiguous and opt for Gorsuch's reading for purposivist reasons of unitary Federal Circuit doctrine on civil-services law, as the Federal Circuit dissenters did.  Justice Gorsuch's position that the statute is plain his way, however, is simply unarguable, and if it can be argued, it's yet to be done because Justice Gorsuch has never attempted to argue it.  Instead, the comments I quoted above at oral argument were his only attempt to deal with the text that supports the consensus view, and an implausible attempt at that; his dissent either omits that text (7702) completely, or in the case of "subject to . . . 7702," doesn't attempt to explain how it's consistent with his interpretation.

That much is vexing enough.  But what I find really frustrating is that even Justice Gorsuch patently saw at oral argument that the statute is ambiguous at best.  Then, he said that "subject to" could bear multiple meanings, and that one permissible meaning of "subject to," "meeting the test of" (though I would proffer "governed by" or "described in"), supports the overwhelming consensus view.  That being so, I can't understand why he would go to such lengths to deride this confessedly permissible reading as a faithless judicial "tweak," and why he didn't at least explain, in his dissent, why he thought it wrong.   

Inspired by Aaron Nielson's inclusion of lesser-known Tom Petty songs in his last edition of D.C. Circuit Review, I've included this chronological playlist of fifty-one lesser-known '60s Beach Boys songs in this final edition of OT 2016 Statutory Term in Review for my dear readers.  I've completely skipped their most famous album, Pet Sounds (which I heretically don't like very much anyway), most of the more recognizable hits, and included a bunch of unreleased songs and alternate versions, though even people who don't know the Beach Boys very well will probably recognize a few songs here.  (I've also included what's been described as a proto-shoegaze live recording of "Surfer Girl.")  It's a long list, but I particularly recommend, in no particular order, "Breakaway," "Busy Doin' Nothin,'" "Passing By," "Your Summer Dream," "Little Pad," all the "Can't Wait Too Long" recordings, especially the first two, "Cabin Essence," the pre-Brian-Wilson-nervous-breakdown recording of "Wind Chimes" and the distinctly post-Brian-Wilson-nervous-breakdown recording of "Wind Chimes," "You're So Good to Me," "Please Let Me Wonder," "Don't Hurt My Little Sister," "Why Do Fools Fall in Love," "The Little Girl I Once Knew," and "Let Him Run Wild."

Sunday, October 8, 2017

Partisan Gerrymanders Aren't Intentional Viewpoint Discrimination (or Any Other Type of Discrimination)

As I mentioned in my last post, there's an increasingly popular argument going around that it's unconstitutional for state legislators to draw districts for the purpose of advantaging their party and disadvantaging the opposition party—period.  (Several new articles making this argument are collected here.)  The argument, or a key part of it, goes like this.  

To begin with, everyone agrees it would be unconstitutional to impose a tax on the basis of partisan affiliation in order to put members of one party in a worse financial spot than another; it would also be, indeed has been held, unconstitutional to give out civil-service jobs on the basis of partisan affiliation in order to empower members of one party and disadvantage members of another.  In fact, it's unconstitutional to do just about anything in order to favor members of one party and disfavor members of another; whatever the state did in that vein would be viewpoint discrimination. 

Partisan gerrymandering, the argument goes, isn't any different or exceptional.  Partisan gerrymandering is just another way of disfavoring or, in the language of this article, subordinating members of one party and favoring members of another.  How so?  Well, a "Republican" by definition is just someone who supports Republican candidates, and a "Democrat" by definition is just someone who supports Democratic ones.  When a state draws districts with the purpose of making it hard for Democratic or Republican candidates to get elected, its intention is necessarily to frustrate Democrats' or Republicans' electoral preferences, which are the very preferences that define them as Democrats or Republicans.  Thus, far from partisan gerrymandering being the one sort of law motivated by partisan favor or disfavor that's okay, it's the quintessential case of invidious partisan favoritism. 

I don't think this argument works.  To see why, consider these analogous arguments, or at least arguments that I think are analogous.  It would be unconstitutional to impose a tax on voters who support affirmative action, to discriminate in public employment against voters who support affirmative action, or indeed to do just about anything to disfavor voters who support affirmative action.  That would be impermissible viewpoint discrimination.  Therefore, "rigg[ing] the game," as one court put it, against voters who support affirmative action by prohibiting it in the state constitution, thereby making it much tougher for voters who support affirmative action to get what they want than voters who oppose affirmative action or support other kinds of admissions preferences, must also be unconstitutional. 

Of course, that's just what the Sixth Circuit held in Schuette v. Coalition to Defend Affirmative Action, though even the Sixth Circuit thought its anti-game-rigging theory was limited to policies that benefit racial minorities and didn't protect other groups.  We know what happened next; even that narrow formulation of anti-game-rigging theory was reversed by the Supreme Court 6-2.  It turns out that it's quite constitutional to "rig the game" against supporters of affirmative action when the "game" being played isn't taxes or public employment, but a political fight over affirmative action itself.

This isn't because (or at least not only because) the Fourteenth Amendment "does not require what it barely permits," as Judge O'Scannlain, Judge Sutton, and Justice Scalia all argued.  The Constitution more than "barely permits" marijuana legalization; it absolutely permits it, is completely indifferent to it.  However, the following argument doesn't work:  it would be unconstitutional to discriminate against voters who support marijuana legalization in taxation, public employment, public university admissions, or certainly in the realm of speech—that's all viewpoint discrimination.  Therefore, it must also be unconstitutional to "rig the game" against marijuana-legalization supporters by amending a state constitution to remove drug policy from the ballot-initiative process, or amending the state constitution or state legislative rules to require a supermajority for a ballot initiative on drug legalization, or for passage of a bill legalizing marijuana.  And further still, if it's unconstitutional to merely make it harder to legalize marijuana, do you know what's a really unconstitutional subordination of people who want to legalize marijuana?  Banning marijuana.

Obviously none of that follows.  What does follow from this illustration is that it seems it is unconstitutional to rig the employment, taxation, university admissions, or speech "games" against supporters of marijuana legalization, but not the marijuana legalization game.  More generally, what we learn from this and innumerable hypotheticals that could be spun out just like it is that it is unconstitutional to discriminate against a group of people defined by some political preference except with respect to their achieving that preference.   Or more precisely, "discriminating" against a preference group by thwarting their preferences isn't discrimination against that group at all.

Of course, someone might make the following objection.  When a state bans marijuana or makes it harder to legalize marijuana, it's not doing it to disfavor the people who support marijuana; it's doing it for the purpose of banning marijuana.  That's why those things are constitutional and discriminating against marijuana supporters in taxation/employment/admissions/speech is not—not because it's impermissible to discriminate against people with a particular preference in any way except making it hard for them to achieve their preferences, but because the reason for making it harder to legalize marijuana isn't to disfavor marijuana supporters, while the motive for doing all those other bad things is.  On the other hand, the partisan-intent theorist will claim, the purpose of a partisan-advantage-driven gerrymander is to subordinate members of one party and favor another, not to block the disadvantaged party's political agenda.

I think this gets to the nub of whether partisan gerrymandering is unconstitutional, at least on a theory of invidious motive.  To see precisely when it would be accurate to say that a partisan gerrymander had an invidious motive, consider a claim of an anti-Hasidic-Jewish gerrymander; Hasidic Jews in New York City actually made such a claim in UJO v. Carey.  I don't use the example to suggest that party is different from religion or race; indeed, I'll assume that discriminating against the one is just as bad as discriminating against the other.  Rather, I want to use the example to get clear on the distinction between a gerrymander that's genuinely intended to subordinate some group of voters and a gerrymander that targets preferences, as opposed to the groups that have them.

A state legislature might have a variety of reasons for drawing district lines that reduced the political power of a geographically concentrated Hasidic Jewish community by splitting it up between multiple districts.  One reason might be that Hasidic Jews don't tend to vote for the candidates that black voters support; in order to comply with the Voting Rights Act and draw a sufficient number of districts where black voters are able to elect their preferred candidates, it becomes necessary to crack the Hasidic community.  That's what happened in UJO, and the Court didn't think that was an anti-Hasidic gerrymander, though its purpose was to help candidates that black voters favored and that Hasidic voters happened to disfavor.  The reason we wouldn't call it an anti-Hasidic gerrymander is that it wasn't done for the sake of harming Hasidics or their political power.  Hasidics just happened to be an obstacle to advancing black political power, so their political power was intentionally reduced.

Next, perhaps Hasidic Jews uniformly favor Republicans (I have no idea if this is the case), and New York wants to draw district lines that favor Democrats.  In order to do this, in multiple cases, they have to splinter Hasidic communities between multiple districts, or put them all in one where they can't have any impact on the elections in others.  Again, that would not be deemed an anti-Hasidic gerrymander; though the state would have intentionally harmed Hasidic voting strength, it didn't do it to harm Hasidics, but to benefit Democrats, who Hasidics hypothetically disfavor.  The Court has upheld that kind of politically motivated ethnic sorting of voters (so long as it's part of a larger political sorting of voters and entire districts aren't organized around ethnic stereotypes of political behavior) in cases like Easley v. Cromartie; it doesn't understand it as a racial gerrymander.  Maybe it's an impermissible partisan gerrrymander, but not an anti-Hasidic one.

So what would count as an anti-Hasidic gerrymander?  Suppose Hasidics didn't reliably favor any particular party, but swung from election to election.  Suppose too that Hasidics don't pose a problem for Voting Rights Act compliance.  An anti-Semitic legislature that simply doesn't want Hasidics to have a meaningful say in the outcome of elections splinters compact Hasidic communities up into so many tiny pieces that their votes will rarely have a dispositive effect on any election.  That would be an anti-Hasidic gerrymander, one motivated to politically subordinate Hasidics and make it difficult for them to achieve their preferences whatever they happen to be, not to thwart particular electoral preferences that Hasidic voters happen to have.

Now, are partisan gerrymanders like that sort of gerrymander, or like the gerrymanders I described that target preferences rather than subordinating groups of voters?  I think they're obviously gerrymanders of the second variety.  A Republican state, for example, doesn't diminish Democratic voting strength for the sake of thwarting registered Democratic voters' preferences whatever they happen to be, whether they're for electing Democrats or Republicans.  If a bunch of registered Democrats started voting for Republicans in state-legislative elections, while continuing to vote for Democrats in national elections, the state wouldn't keep its state-legislative gerrymander in place the way an anti-Semitic legislature would keep its anti-Hasidic gerrymander in place regardless of who Hasidics happen to vote for.  Rather, it would rejigger its state-legislative gerrymander while keeping its congressional gerrymander constant.

The reason it would rejigger the one gerrymander and not the other is because the purpose of partisan gerrymanders isn't to disempower some group of voters, but to prevent candidates of a particular party from getting elected.  Partisan gerrymanders target preferences, not people; they kneecap candidates, not voters.  To the extent they intentionally reduce particular voters' political power, they do so only because of the transient preferences those voters happen to have at the moment, not because of an inelastic desire to subordinate those voters.  So it's not true that partisan gerrymanders meaningfully differ from laws that "gerrymander" marijuana-legalization politics; both are just intended to block some political preference, not to harm the people who have it.  

Next, one might object that the motives for partisan gerrymanders are more "tribal" than the motives for rigging marijuana legalization politics against the legalizers.  The politicians who change the procedural rules of legalizing drugs to make marijuana legalization harder have some good-faith policy rationale for doing what they're doing, but partisan gerrymanderers have no more motive than increasing the political power of their "tribe" and diminishing that of the opposing tribe.  Even if partisan gerrymanders' targets are political preferences rather than the voters  who have them, their motive is so base that we shouldn't think of them like other preference gerrymanders.

I think that's wrong, but in the first place I'm not sure why it matters.  Consider the following example.  Nothing is more tribal than intrastate college-football rivalries; there can be no high-minded purpose in supporting one team and opposing another.  And if a state had two state universities with popular football teams, it couldn't discriminate against supporters of the one and favor supporters of the other in employment, taxation, speech, or anything else.  That would be viewpoint discrimination of an exceptionally inane kind.  

However, the state obviously could favor one team over the other in the sphere of football, disappointing the other team's fans in the process.  It could make one of the two teams the state's official football team, decide to pour money into the one team and starve the other, simply disband one of the two teams, or to get closer to gerrymandering, have one set of procedures for appropriating funds for the one team and another much more onerous set of procedures for appropriating funds for the other team.  All those policies would serve no other purpose than to delight fans of the one team and immiserate fans of the other, but like my other examples, that's not viewpoint discrimination; it's picking a side of a funding dispute on which people have views, not discriminating against the people with the rejected view.  As stupid, petty and tribal as the choice to favor one of the teams would be, it still isn't viewpoint discrimination.

That said, I don't think partisan gerrymanders are anything like picking a side in a football rivalry.  The reason politicians gerrymander for their party isn't ultimately loyalty to the party, but their desire to advance the policies for which it stands and hinder the policies for which its opponent stands.  For example, suppose that segregationist Southern Democrats discover in 1964 that their party has just become the party of civil rights, voting rights, and welfare for the non-white poor on the national level.  Will they continue to draw Democratic gerrymanders and do so around the new voters who join the party while diluting the electoral strength of the voters who leave?  Or will they switch over time to another party and draw gerrymanders in order to thwart the policies that Democrats used to oppose?  I think we know the answer to that question.

Or, suppose that President Trump announces tomorrow that the Republican Party is now the party of single-payer healthcare, progressive taxation, and immigration reform.  Will Republican state legislatures still churn out Republican congressional-districting gerrymanders, attempting to draw districts that diminish the power of former Republican voters who left the party and maximize the power of all the new Republican voters who join up?  Or might they start a third party and gerrymander for that party?  I think it's unlikely that loyalty to the name brand would persist for very long.  Instead of Republican gerrymanders, we would soon see gerrymanders by formerly Republican legislators for whatever party opposes single-payer healthcare, progressive taxation, and immigration reform.  Partisan gerrymandering may, as Justin Levitt argues, be motivated by "tribal partisanship," but the "tribes" are held together by some minimal set of policy preferences and attitudes, not, when the rubber meets the road, a party logo. 

One might argue that I'm just wrong and that thwarting a group's preferences is, in at least some contexts, discrimination against that group.  For example, wouldn't a ban of challah discriminate against Jews?  And if so, isn't that only because they like to eat challah in their religious rituals?  Didn't same-sex-marriage bans discriminate against gays and lesbians, given their preference for same-sex marriage over opposite-sex marriage?  

I think the answer to these questions is that, if challah bans or same-sex marriage bans are discriminatory, they are only so because their purpose is to harm the groups that like challah and same-sex marriage, not because banning or impeding a group's preference is ever tantamount to discrimination by itself.  Take challah.  As we know, it's not deemed religious discrimination to ban peyote, even though some religions smoke peyote in their rituals.  That's because the purpose of those bans is to ban peyote for the sake of banning peyote, not to discriminate against religions that smoke peyote.  The same could be true of any number of substances or foods used in religious ceremonies.

In the case of challah, though, there's no conceivable purpose for banning challah other than animus against Jewish people, unless challah were banned as part of a general ban of fattening breads.  Such a law would fall, despite its facial neutrality, under Lukumi Babalu Aye.  Unless it can be shown that the purpose of partisan gerrymanders is really to harm the voters who support the minority party independent of their political preferences, partisan gerrymanders aren't like banning challah.  In reality, of course, partisan gerrymanderers gerrymander to increase their party's chance of winning elections and instantiating its agenda, not to subordinate voters who like the minority party for the sake of subordinating those voters.

As for same-sex marriage, I tend to think it's more correct to explain Obergefell in terms of  burdening the right to marry than in terms of discrimination against gays and lesbians.  (Likewise, if partisan gerrymandering poses a First Amendment problem, it's a problem of indirect burden on people with partisan views, not viewpoint discrimination.  I may address the question of indirect burden in a later post.)  The universal prohibition of polygamy doesn't seem to me to discriminate against people who wish to form polygamous marriages; at most, it unduly restricts their right to marry.  State marriage laws' age requirements don't discriminate against high-school students who, as the Beach Boys sang, "want to get married, but [are] so young [and] can't marry no one"; age requirements may pose a liberty problem, and they may pose an age-discrimination problem, but they do not discriminate against the class of persons who want to get married in their teens.  

That said, it could be argued that the only thing that motivated same-sex marriage bans was animus against the people who wanted to obtain gay marriages, just as the only thing that could possibly motivate a challah ban would be animus against the Jewish people who eat challah.  If and (I think) only if that is so, same-sex marriage bans were an instance of anti-gay discrimination.  If not, talk of same-sex marriage bans as discrimination or subordination is just another instance of natural but fallacious conflation between thwarting a preference and discriminating against the people who have it—the same conflation that grounds arguments that partisan gerrymanders are a way of intentionally subordinating minority party members.  


I should note that it's possible that, while partisan gerrymanders don't disfavor or subordinate minority party voters, the purpose to advantage one party and disadvantage another is simply an impermissible purpose, regardless of whether all policies motivated by this purpose discriminate against minority party voters.  For example, while this article by Michael Kang argues that partisan gerrymandering discriminates against minority party voters, it relies at times on the somewhat different claim that official government partisanship is impermissible—that the government couldn't endorse candidates for office, or contribute funds to only one party's candidates' campaigns.  Laws like that wouldn't discriminate against voters (they would discriminate against candidates), but they'd still probably be invalid on the ground of illicit purpose.  So too, one might argue, are partisan gerrymanders for the same reason, even absent viewpoint discrimination against minority-party voters.  

I tend to think that partisan gerrymanders can be defended from this argument too; the government probably couldn't officially endorse one side of a ballot initiative or contribute to one side of a ballot-initiative campaign, but I believe it could, as I said above, adopt a supermajority requirement for ballot initiatives on certain subjects, or remove a subject from the ballot-initiative process for the purpose of preventing an officially disfavored position from succeeding.  But it's a better argument.

Wednesday, October 4, 2017

A Miscellany of Thoughts on Justice Kennedy's Maximal-Partisanship Hypothetical in Gill

1.  Justice Kennedy's hypothetical at oral argument in Gill and the answers he received.

As you probably know if you read this blog, Justice Kennedy asked the following hypothetical at oral argument in Gill v. Whitford Tuesday of Erin Murphy, counsel for the Wisconsin State Senate:
You've probably considered the hypo many times.  Suppose a state constitution or state statute says all districts shall be designed as closely as possible to conform with traditional principles, but the overriding concern is to increase -- have a maximum number of votes for party X or party Y.  What result?
It didn't, however, seem like Murphy had considered the hypo any times—which is odd, for as I'll discuss below, Kennedy had already asked and answered essentially just this hypo himself 13 years ago in his essential concurring opinion on the subject in Vieth v. Jubelireras she began with a tentative "at least you're closer [to something unconstitutional]" answer, drifted into an "I don't think . . . that you've proven a constitutional violation" answer on further questioning from Justice Ginsburg, returned seconds later on still further questioning on the hypo from Justice Alito to allowing that "at least at that point, you know the intent," then confidently told Justice Kagan minutes later that "just finding the intent isn't a problem," after which, when told by Justice Kennedy that "I don't think you ever answered the question," she (a) volunteered that his hypo "could be your instance of a -- a problem that can be actually solved by the Constitution," (b) said that whether his hypo violated the Equal Protection Clause or First Amendment was "a little hard to say at this point because, you know, it really just hasn't been fully explored, this concept of how you would come at all this from a First Amendment perspective," and finally (c) answered about 10 seconds later:
Yes.  It would be an unconstitutional, if it was on the face of it, and I think that that would be better thought of probably as an equal protection violation, but you could think of it just as well, I think as a First Amendment violation in the sense that it is viewpoint discrimination against the individuals who the legislation is saying you have to specifically draw the maps in a way to injure, but, again, I --

By my count, that's four maybes (the two "at least" answers, the "could be your instance" answer, and the "a little hard to say" answer), two nos, one final and definitive yes, and one stunningly quick switch on "how you would come at all this from a First Amendment perspective."  (None of which is to dump on Murphy, who's done a lovely job in her past appearances at the Court.)  Wisconsin's SG, for his part, recognizing the significance of the question, picked it up in rebuttal and said that Kennedy's hypothetical statute would violate the First Amendment because it would stigmatize the minority party, not because it would be evidence of a facially partisan gerrymander.  This is clever but I don't think it will help Wisconsin, as I doubt anyone will agree that the only problem with that hypothetical is stigma.

2.  The answer I would have given to his hypothetical.

Lamenting how argument went with an appellate-lawyer friend who, like me, wants Wisconsin to win (though I want Wisconsin to win for Frankfurterian reasons and he or she has some troglodytic reason or another having to do with the powers of common-law courts at Westminster), my friend asked how I would answer Justice Kennedy's question.  That's easy, I said.  (Sorry, Erin Murphy.)  

As I had always understood the Court's partisan-gerrymandering precedents and Justice Kennedy's concurring opinion in Vieth, or even the Vieth dissents, a gerrymander isn't unconstitutional merely because it has partisan purposes; virtually all redistricting does.  Rather, a gerrymander has to go too far, excessively diminish the electoral power of the minority party or excessively entrench the then-majority power.  And whether even a law that facially requires maximized partisan advantage in districting, subject to traditional districting principles, results in excessive gerrymanders will depend on the districting plans that maximization generates and how much partisan advantage maximization gets you in those plans.  For sometimes a maximally partisan gerrymander may not be all that partisan, given the constraints of traditional districting principles and compliance with federal law, both constitutional (one-person one-vote) and statutory (the Voting Rights Act), that are baked into Justice Kennedy's hypothetical.  We could easily imagine, for example, that a particular districting plan that maximized partisan advantage as much as it could while adhering to traditional districting principles and federal law would only translate 50% of statewide votes to 54% of legislative seats in a given election cycle.  Is that unconstitutional?  It sounds pretty unexceptional.

In order to decide if an admittedly maximally partisan gerrymander did go too far, then, the Court couldn't just decide on the ground of the state's stated aims; it would still have to engage in the line-drawing and measurement problems that have confuted all attempts to decide partisan-gerrymandering claims since the cause of action was recognized.  That is to say, first the Court has to draw a non-arbitrary line for how far is too far; next it has to somehow predict whether a districting map will, in future elections, go too far.  So far, both have proved impossible tasks, though that doesn't mean they always will.  So, I told my friend, a challenge to Justice Kennedy's hypothetical law would be just as nonjusticiable as a challenge to a facially neutral districting scheme, no more no less, and even if Justice Kennedy believes he knows that this scheme was intended to maximize partisan advantage, as do I, whether that maximization is excessively partisan given the constraints on maximization is a nonjusticiable question.  At least, that would be my answer were I in Ms. Murphy's position.

3. . . . which turns out to have been flatly rejected by Justice Kennedy's opinion in Vieth.

The funny thing is, though, is that everything I've just said is completely undermined by what Justice Kennedy's concurring opinion in Vieth actually said.  In Vieth, Kennedy famously couldn't find partisan-gerrymandering claims justiciable yet, but was unwilling to hold they could never be found justiciable on the ground of some standard that had yet to be offered.  People have always thought that the standard he was looking for was an excessiveness standard, but that doesn't seem to be right at all, or at least not in the sense that he's been searching for a quantitative measure of excessiveness.

In the first place, Justice Kennedy already answered his own hypothetical in Vieth in a way that rules out my answer to his question.  Arguing that partisan-gerrymandering claims at least might be justiciable, he wrote that "[i]f a State passed an enactment that declared 'All future apportionment shall be drawn so as most to burden Party X's rights to fair and effective representation, though still in accord with one-person, one-vote principles,' we would surely conclude the Constitution had been violated."  But why?  It can't be because the mere existence of partisan purposes in districting is unconstitutional; that, Kennedy wrote a page later, "describes no constitutional flaw."  Nor can it logically be because maximizing partisan advantage necessarily works an unconstitutional amount of partisan disadvantage.  

The reason is simply this: political "classifications, though generally permissible," are unconstitutional, or so Kennedy claims at the very beginning of his opinion, if they "were applied in an invidious manner or in a way unrelated to any legitimate legislative objective."  The absolute subordination of other "aims of apportionment" to partisanship just is, Kennedy thinks, unlawful.  So when Kennedy writes, immediately after he poses his hypothetical, of "the possibility . . . that a legislature might attempt to reach the same result [of maximized partisan advantage" without [an] express directive," and of "a standard [that] might emerge that suitably demonstrates" whether a legislature is making that attempt, what he imagines is a standard that "establishes the [partisan] classification is unrelated to the aims of apportionment and thus is used in an impermissible fashion."  And later, when he explains why the plaintiffs' claims in Vieth failed, he says they "[f]ail[ed] to show that the alleged classifications are unrelated to the aims of apportionment."

Kennedy's implicit rejection of some definition of quantitatively excessive partisan advantage as the Holy Grail of partisan-gerrymandering law, in favor of some as-yet unknown test for excessive subordination of apportionment aims to the purpose of creating partisan advantage, gets much more explicit at the end of his opinion.  In what can only be described as one of the more confused and confusing passages in his work, he posits that—while mere partisan classifications are permissible—all "gerrymanders" are bad, no matter how quantitatively excessive:
That said, courts must be cautious about adopting a standard that turns on whether the partisan interests in the redistricting process were excessive. Excessiveness is not easily determined. Consider these apportionment schemes: In one State, Party X controls the apportionment process and draws the lines so it captures every congressional seat. In three other States, Party Y controls the apportionment process. It is not so blatant or egregious, but proceeds by a more subtle effort, capturing less than all the seats in each State. Still, the total effect of Party Y's effort is to capture more new seats than Party X captured. Party X's gerrymander was more egregious. Party Y's gerrymander was more subtle. In my view, however, each is culpable.
Besides that "Party Y's effort" can't sensibly be aggregated across multiple independent states and their independent legislatures that Party Y controls, a point I would expect Justice Kennedy of all people to understand, the conclusion is nonsense; as described, "Party Y" hasn't done anything "culpable" at all, though it's hard to understand from the hypothetical what exactly "Party Y" did.  But read most charitably, which requires reading a great deal into the hypothetical, it seems the idea is that, so long as a legislature's districting "scheme" is a partisan scheme, it doesn't matter how much partisan injury it causes.

4.  Why Justice Kennedy hasn't announced an intent standard already.

We know, then, that thirteen years ago Kennedy believed that the absolute subordination of "legitimate" apportionment aims to partisan advantage was unconstitutional, and I think we probably know from his repeatedly asking Murphy whether she agreed that he still thinks it today.  What mystified me at first on rereading Vieth, then, is why he didn't settle on a simple intent-based standard for adjudicating partisan-gerrymandering claims long ago.  It shouldn't be too hard to find out, as a factual matter, whether the only thing a legislature was interested in when drawing its district maps was partisanship.  

Indeed, Justice Kennedy has demanded much harder and quite similar findings of district courts before; in instructing them on how to decide racial-gerrymandering claims, he required them to decide whether "race was the predominant factor" motivating the shape of a particular district, and whether "the legislature subordinated traditional race-neutral districting principles . . . to racial considerations."  Deciding whether a legislature's only motive was partisanship, aside from compliance with one-person-one-vote, the Voting Rights Act, and a superficial show of district compactness and internal contiguity, should be much easier than deciding whether a legislature's predominant motive was race, a standard which is widely seen as meaningless (in what sense can motives predominate?), especially given that political gerrymandering is easily mistaken for racial gerrymandering.

The reason, I think, that Kennedy didn't write an opinion like that in Vieth is that he doesn't really have an absolute-subordination standard in mind; rather, he's looking for some purpose-based threshold that's impossible to describe, and certainly can't be described in quantitative terms.  An absolute-subordination standard would describe a null set; bracketing compliance with one-person-one-vote and the Voting Rights Act, a districting plan will always attempt to do at least something besides afford the ruling party partisan advantage, whether it's drawing contiguous and compact districts, protecting incumbents, or avoiding splitting at least some municipalities, communities of interest, or political subdivisions.  And even if it didn't describe a null set now, absolute subordination would describe a null set in the future given that states could work around it by pursing some tertiary aim to some minute degree.

Absolute subordination also isn't quite what he says he wants a standard to measure; rather, he wants a standard that can "establish[ a partisan] classification is unrelated to the aims of apportionment," that "the alleged classifications are unrelated to the aims of apportionment," that partisan "classifications . . . were applied . . . in a way unrelated to any legitimate legislative objective."  The question, oddly, on this formulation isn't whether the legislature did or didn't also serve some legitimate aim of apportionment, but whether the legislature's partisan classifications, taken by themselves, are related to legitimate aims of apportionment.  That is, the partisan motive for the partisan classification must be a purely partisan, illegitmate motive, though it can be mixed with other legitimate motives and legitimate non-partisan classifications.  For this reason, an attempt to maximize partisan advantage subject to all sorts of traditional districting criteria is still unconstitutional on Kennedy's view, because the partisan motive is, as far as it goes, pure and illegitimate.

Probably needless to say, this can never be an operable test (but see sec. 6 below), as classifying voters on the basis of partisan affiliation is always (or at least can always be described and viewed as) purely partisan and illegitimate.  Any districting map that districts on the basis of partisanship at all singlemindedly pursues partisan advantage to the extent it singlemindedly pursues partisan advantage; it may do all sorts of other things, may even be dominated by those other things, but to the extent it classifies some of the voters on the map on a partisan basis, it won't have done so for some legitimate non-partisan reason (absent things like incumbency protection).  

So, unless Kennedy thinks that all districting schemes drawn by politicians are invalid, he can't be serious about invalidating any partisan classification that's unrelated to non-partisan aims.  Nor could one explain his holding in Vieth if he were serious; is there any doubt that partisanship had been pursued purely for the sake of partisanship to some extent in Vieth?  It would seem that what's being groped for here is some sort of predominant-motive test, like his test for racial gerrymandering, but even that isn't quite stringent enough for what Kennedy seems to want.

One of the reasons Kennedy was unable to settle on a predominant-illicit-motive test is that, at least in 2004, he seemed deeply unsure of what the legitimate "aims of apportionment" were.  His opinion complains that there isn't "any agreed upon model of fair and effective representation," any "agreed upon substantive principles of fairness in districting," that the parties hadn't shown him "statements of principled, well-accepted rules of fairness that should govern districting," and most oddly, that he had "not been able to discover[] helpful discussions on the principles of fair districting discussed in the annals of parliamentary or legislative bodies."  It was "[b]ecause," he wrote at one point, "there are yet no agreed upon substantive principles of fairness in districting" that the Court had "no basis on which to define . . . standards for measuring the particular burden a given partisan classification imposes on representational rights."  

Many people naturally read this passage as a request for a model to measure quantitative deviations from some as-yet undiscovered quantitative measure of fairness; that is certainly a possible reading of the opinion, especially given his later hope that "new technologies may produce new methods of analysis that make more evident the precise nature of the burdens gerrymanders impose on . . . representational rights."  But I think it also suggests an uncertainty about what the "legitimate legislative objectives" and permissible "aims of apportionment" even are, and absent some notion of that, even a purpose-based test is hopeless.

Finally, there is no doubt that Kennedy gave political scientists and election lawyers ample reason in Vieth to think he was searching for some quantitative measure of the "burden" on "representational rights."  For example, at one point he suggested that gerrymanders would violate the First Amendment just so long as "a State did impose burdens and restrictions on groups or persons by reason of their views," but that, "[o]f course, all this depends first on courts' having available a manageable standard by which to measure the effect of the apportionment and so to conclude that the State did impose a burden or restriction on the rights of a party's voters."  

Even this, though, is bizarrely non-quantitative; how, if it's true that the mere use of political classifications is constitutional, can it be that the imposition of "a burden" is a per se First Amendment problem no matter the burden's size?  And what sort of quantitative standard would measure whether "a burden" exists or not?  Yet at least this motivates his answer to his hypothetical; minimizing the political power of one party, to the extent compatible with traditional districting principles and federal law, would certainly count as "a burden" on that party, even if minimization wasn't ultimately that minimal.

5.  Finally, why Justice Kennedy is wrong about the invalidity of "maximizing" partisan advantage.

Of all this, I can only say, and must say, that Kennedy's thoughts on partisanship and districting in Vieth were extraordinarily incoherent, and appear to remain so in their most critical respects.  If Kennedy wants to hold that partisan intent in any degree invalidates a districting plan, that would be a workable, if implausible, standard (but see note below acknowledging recent contrary scholarship).  But the rule or intuition to which he seems powerfully inclined—that though partisan motives are generally permissible, a motive to "maximize" partisan advantage to the extent that can be done while also serving a bunch of other motives, like hewing to traditional districting principles, is invalid—is just nonsense.  

For one thing, "maximizing" partisan advantage subject to the pursuit of other restrictive aims can't sensibly be described as maximization, anymore than one can sensibly talk about living a maximally hedonistic life subject to keeping a reasonable diet, avoiding undue risks, obeying the law, and working fifty hours a week.  If that's living a maximally hedonistic life, everyone lives a maximally hedonistic life except for all the ways in which they don't, and if a state can "maximize" partisan advantage subject to (a) district compactness, (b) district contiguity, (c) respect for political subdivisions, (d) avoiding one-person, one-vote liability (which doesn't mean going right to the edge of it), and (e) avoiding Voting Rights Act liability (ditto), probably every state controlled by a single party maximizes the partisan advantage of the party in power.  For maximizing one aim subject to five others is simply what it means to rationally pursue that aim at all.  

For example, suppose that a state with sixty districts has three partisan options that would comply with traditional districting criteria and federal law: a plan that would give the majority party control of 31 seats, a plan that would give the party control of 32 seats, and a plan that would give the party control of 33 seats.  Picking the one with 33 safe seats would "maximize" partisan advantage subject to traditional districting criteria, and picking the ones with 31 and 32 safe seats would not, but what rational legislature that cared to pursue partisan advantage at all would pick those plans when the 33-seat plan satisfied every other objective the legislature had?  No legislature would do so, anymore than a gourmand on a diet would choose to eat, as between three healthy, appetizing foods, the one that only sounded the second or third most appetizing to that person at that point in time.  Any person, gourmand or otherwise, in that situation would pick the most appetizing healthy option.  You could call that person a "maximal" hedonist, or you could just call them sane. 

Even if "maximizing" partisanship subject to other goals meant anything besides pursuing partisanship in a non-irrational way, it is impossible to explain why such a "maximal" pursuit of partisanship is per se invalid on grounds that don't invalidate non-maximal pursuits of partisanship.  "Maximal" pursuits of partisanship can't be invalid because they're necessarily all that partisan or that maximal; traditional districting criteria may, or may not, prevent them from getting that partisan.  A theory of First Amendment burdens on representational or associational rights doesn't explain anything either; in what sense is a non-maximal burden not a burden?  How, as a matter of First Amendment law, could the Court sensibly hold that only "maximal" burdens on a party's viewpoints were invalid?  Obviously that would never fly in any other area—one is tempted to say any real area—of First Amendment law.  

All that is left, it would seem, is that maximizing partisanship is a bad motive, but if the Court held this tomorrow, why wouldn't it also follow that it would be bad for the state in my hypothetical to pick the plan with 32 safe seats instead of one with 31, 30, or 0?  Why would the theoretical possibility of a more partisan and unlawful 33-safe-seat plan that complied with traditional districting criteria save the state's intentional choice of the most partisan non-maximally partisan plan it could draw?  It shouldn't, at least not on whatever theory would per se invalidate the "maximally" partisan plan, anymore than people who condemn living a maximally hedonistic life have grounds to condone people who attempt to skirt moral objections by always opting to behave in the second-most hedonistic way they can think of.  I can only say that someone whose central intuition about regulating redistricting is this  incoherent has no business regulating redistricting.

6.  A note on the partisan-intent theorists

I am aware that a growing chorus of election-law scholars, whose articles are helpfully collected in the last sentence of this post, argue that at least some sort of partisan motive is  invalidating, at least when that motive is met with some success.  The author of one of these articles, Michael Parsons, has just written an excellent post cheering on Justice Kennedy's interest at argument in invalidating gerrymanders that stated their maximally partisan aims.  I would say just two things about their work in relation to Kennedy's feelings about intentions to maximize partisan advantage.  

The first is that it certainly doesn't explain a rule against "maximally" pursuing partisan advantage (or maximally pursuing partisan advantage). Partisan-intent theorists might think such a rule is better than nothing, as it captures a subset of what they're worried about, but it makes no more sense on their lights than permitting the non-maximal pursuit of racial subordination would. 

The second thing I would say is that I don't see this as a standard that Justice Kennedy is willing to accept.  At least some of the intent theorists' articles argue that not all partisan intent is bad, just partisan intent to subordinate the minority party, or "tribal partisanship."  This is intended both to palliate concerns that their rule would invalidate all partisan districting, and to get around the objection that the Court in general and Justice Kennedy have repeatedly said that some partisanship is alright.  I guess I don't understand, though, how subordinating the minority party (which is just a pejorative way of talking about favoring the majority party) is a practically lesser subset of partisan redistricting purposes or partisan classifications; what other sorts of partisan purposes or classifications are there?  

When some of the intent theorists argue that when Justice Kennedy writes that partisan classifications aren't per se impermissible, all he means is that states may seek to promote proportionality or perhaps incumbency advantage by way of districting for partisan balance, I have a tough time following them.  It's true that Kennedy's repeated citation for the proposition that partisan classifications are okay by themselves is an early case that upheld a bi-partisan gerrymander that sought to achieve a rough measure of proportionality.  But it strikes me that if Kennedy thought that pursuing partisan advantage to any degree and with any success was per se unconstitutional, he would have adopted a very simple intent test a long time ago.  It wouldn't be too difficult for district courts to decide whether states pursued partisan advantage in part in drawing their districting plans; Kennedy has asked district courts to decide the much harder question of whether race predominated over other districting aims.  And I don't understand why Kennedy would keep returning to the theme of maximizing partisan advantage, or the pursuit of partisanship in a fashion that's unrelated to any legitimate aim, if he were actually interested in setting aside districting schemes motivated by any hopes of achieving partisan advantage.