In light of today’s Benisek argument, and at the risk of getting banished to another forum (I jest, I think), I offer a modest addition to the Pildes–Stephanopoulos discussion on partisan gerrymandering and the First Amendment.
I agree with much of what both Pildes and Stephonopoulos say. And (of course?), I’ll quibble with each.
First, a caveat. It’s natural to compare the claims about partisanship in redistricting to the claims about race in redistricting, because the latter are more fully developed, and law loves analogies to settled practice. But it has to be noted that injuries based on partisanship and those based on race aren’t equivalent (and I don’t think either Rick or Nick would disagree). Political science shows that partisanship is mostly consistent, but race is (largely) immutable; injuries have been perpetrated based on partisanship, but they pale in comparison to those (still) perpetrated based on race. The different premises means there’s no reason to transplant the doctrine directly.
And yet, because law loves analogies to settled practice, the comparison to race claims is natural. It may even be illuminating. But I think Rick leaves out the most natural analog, and the one that may best characterize both Benisek and Gill. And I think this natural analog also explains the appropriate remedy in ways that diverge from Nick’s approach. (Much, much more, here. And there’s lots in Michael Kang’s recent work along very similar lines.)
Rick mentioned two lines of race claims in redistricting. In fact, there are three.
As he says, one line is about racial vote dilution – call it the White v. Regester line (picked up in much of the Voting Rights Act). Though it depends on a whiff of danger signs of intent, the real work is done by evaluating a tangible effect revealing diluted political power.
Another line follows Shaw v. Reno. The Court has repeatedly said that the non-invidious consideration of race — in the mix with other stuff — is fine. But if (and only if) race represents the predominant reason for sorting voters into one district rather than another, there had better be a really good reason. This is an intent test, yes. But it’s not really the one that fits best.
There’s also a third line of race claims, distinct from both. It focuses on specific state intent to injure based on race. The right redistricting analog here is Gomillion – when Tuskegee was redrawn to exclude (nearly) all the black voters from the city. Gomillion is often understood as a dilution case. It’s not. It didn’t depend on measuring black voters’ municipal voting power (and wouldn’t have been OK if Alabama had drawn a neighboring “Black Tuskegee,” where black voters had unlimited municipal voting power). Instead, Gomillion recognized that the reason that Tuskegee was redrawn was to subjugate the local black citizens – and that’s an illegitimate reason for the state to use its power, whether it’s “predominant” or not. As long as plaintiffs can establish standing, a drop of this illegitimate motive is enough to force Alabama to prove that it would have passed the same plan without the illegitimate motive – or redraw the map.
So now to the (imperfect) partisanship analogs.
In some cases, the Court has shown First Amendment concern along the lines of Regester, where the effect is the thing that really matters. (Most cases about undue burdens on voting fit this pattern.) Some of the partisan gerrymandering theories, like those based on extreme overall skew or permanent-minority status, seem along these lines. This may be consonant with the claim in Gill. It’s not really the claim in Benisek.
I don’t know that there’s any Shaw equivalent under the First Amendment, where it’s fine to consider partisan viewpoint, just not too much. Maybe we’ll see one here.
But the Court doesn’t have to invent Shaw for partisanship in order to find a First Amendment analog for the Benisek claim. There’s plenty of First Amendment precedent that follows Gomillion in declaring off-limits state action premised on a specific intent to subjugate based on protected characteristics. Indeed, most First Amendment doctrine suggests that a specific intent to subjugate based on partisanship doesn’t depend at all on either the extent of the impact or the presence/absence of other considerations. Heffernan, from just last Term, makes clear that if the government takes adverse action against an employee based on their perceived political preference, that’s usually unconstitutional. The size of the impact doesn’t matter. And adverse action is just as present when there’s a clear failure to promote – “we’ll keep those people in their place” – as when there’s a firing.
(This doctrinal thread is not exclusive to partisanship – intent to injure based on religion works the same way. Nor is it exclusive to the First Amendment. In plenty of other arenas of con law, the specific intent to injure a disfavored group is sufficient to declare the action unconstitutional: bills of attainder, the dormant commerce clause, the 8th Amendment, cases like Moreno and Windsor.)
Indeed, it’s even easier than presenting an analog from the First Amendment employment cases. Because this is also the logic of Cox v. Larios, and the lower court cases striking down unequal population based on impermissible partisanship. These cases exist in a zone where the impact of population inequalities is unremarkable on its own. The only thing that drives the outcome is impermissible intent to injure or subjugate based on party.
The logic of Gomillion — and Cox — is applicable to both individual districts and district-wide plans. But it doesn’t demand the impossible: it doesn’t demand districts drawn with no partisan information. What it demands, instead, is that information in every legislator’s head not be demonstrably used to pass legislation designed specifically to injure the partisan opposition.
This is the standard to which every non-redistricting piece of legislation is held. Legislators no doubt consider the preferences of their partisan voters in every bill they pass. But even in these jaded partisan times, I think it would still inspire shock if a bill required Republicans to pay more in taxes because they are Republicans. The amount doesn’t matter. Neither would the presence of thousands of other reasons for the bill.
This is also the question that Justice Kennedy fixed on in the Gill argument. He asked if it would be constitutional for a state to require that all legitimate redistricting factors be used specifically to favor one party and disfavor another. The answer does not depend on the resulting impact of the map. It depends on whether the demonstrated intent is permissible.
The real meat of this test comes in demonstrating specific intent to injure, and this is where I think I most disagree with Nick. Nobody writes the bill that Justice Kennedy asked about, and so litigants have to turn to other tools. Nick suggests they’ll find it wherever one party’s in control … but though it may only be possible to find there, that’s not even close to passing muster under current law. Feeney clearly says that invidious intent has to be proven, not just presumed. And that standard is hard. Really hard. Go ahead and count the number of cases in the last decade won based on findings of invidious intent to injure based on race. (The dearth isn’t because we’re post-racial.) If the Court is prepared to revisit Feeney, that’s a whole different conversation – but I’m not getting that sense. With all of the possible reasons why district lines might fall in X place rather than Y place, actually proving that the reason was a specific intent to injure opposing partisans will be a whole lot more difficult than simply citing the partisan legislative lineup.
Which comes to the last question: is it worth it? I think the logical remedy for a First Amendment theory along the lines of Gomillion, or Cox, is not to require partisan-blind districts, but to require districts drawn for reasons other than using state power to injure or subjugate the partisan opposition. This will not purge the process of partisanship. It won’t even purge the process of partisan intent to injure – as in all litigation, there will always be a gap between what exists in reality and what is provable to a court on the evidence available. But it would limit the extremes of obvious partisan excess. And compared to the alternative, that sure seems like it’d be worthwhile.