Here are some early thoughts on today’s historic and unexpected decision in Milligan:
1. The Roberts Court Did This? Before today, the Roberts Court had only chipped away at racial vote dilution claims under Section 2. In LULAC, the Court devised a “cultural compactness” requirement to complement Gingles’s geographic compactness criterion and rejected Section 2 claims that minority influence districts should be created. In Bartlett, the Court further rejected Section 2 claims asking for the creation of crossover districts. And in Cooper, the Court suggested there might not be enough racial polarization in voting in North Carolina (of all places) to satisfy Gingles’s second and third prongs. Until now, the Roberts Court had never issued a Section 2 decision clearly requiring the creation of an additional minority-opportunity district (not even in LULAC). In fact, this is the first such decision ever by the Court (since Gingles involved a challenge to multimember, not single-member, districts).
For those of us (myself included) who have long viewed Chief Justice Roberts as an implacable foe of the VRA, Milligan requires at least a partial reassessment. After imposing limit after limit on Section 2—and after authoring Shelby County!—how did Roberts end up here, resoundingly affirming the Gingles framework and holding that Section 2, construed this way, is constitutional? One answer is Roberts’s commitment to stare decisis and aversion to sweeping legal change (at least some of the time, at least at this stage of his career). His opinion is full of references to stare decisis—for example, his statement that “statutory stare decisis counsels our staying the course.” His dislike for the upheaval that Alabama’s reading of Section 2 would produce is also clear. That reading would “remake our § 2 jurisprudence anew.” Its “inescapable consequence” would be that “Gingles must be overruled.”
Another answer is that Roberts was reassured by evidence about how Section 2 is currently operating, under its existing constraints. The amicus brief I filed with Jowei Chen, Chris Elmendorf, and Chris Warshaw presented this evidence. Most Section 2 racial vote dilution claims fail because plaintiffs are unable to satisfy one or more of the Gingles preconditions. As a result, minority voters are disproportionately underrepresented just about everywhere. Section 2 has not yielded proportional representation by race, as the provision’s advocates once hoped and its opponents (including a young John Roberts) once feared. In his opinion in Milligan, Roberts quotes from our brief extensively. He relies on it to conclude that “proportional representation of minority voters is absent from nearly every corner of this country despite § 2 being in effect for over 40 years”—and thus to reject the dissent’s argument that courts have been “methodically carving the country into racially designated electoral districts.”
2. Disabling of Defenses: Milligan is also notable in how it refutes two conventional defenses that Alabama offered for its districts. (Of course, Alabama also advanced the far more radical proposition that Section 2 liability should be assessed by comparison to a “race-neutral benchmark.”) First, Alabama asserted that it wanted to keep Mobile County whole but the plaintiffs’ demonstration maps failed to do so. The Court responded that this county preservation criterion was only weakly supported by the evidence, and that in any event, the plaintiffs’ maps don’t have to win a “beauty contest” with the enacted plan. This reasoning will knock the legs from under many future defendants’ objections to plaintiffs’ demonstration districts. It means that future defendants won’t be able to make up post hoc rationales for their districts, which plaintiffs’ districts conveniently fail to satisfy. It also means that future defendants won’t be able to complain that plaintiffs’ districts score somewhat worse on traditional criteria. After all, that would be a beauty contest, which is exactly what plaintiffs’ districts don’t have to win.
Second, Alabama claimed that it prioritized “core retention”—the preservation of the basic shapes of its 2010s districts—while the plaintiffs’ demonstration maps didn’t. The Court eviscerated this argument. “A State [can’t] immunize from challenge a new racially discriminatory redistricting plan simply by claiming that it resembled an old racially discriminatory plan. That is not the law.” With this pithy analysis, the Court dismantled one of the main defenses that states have been offering in racial vote dilution and racial gerrymandering cases. States will no longer be able to rebut allegations of discriminatory effect or discriminatory intent by invoking continuity with past plans. If past plans diluted minority electoral influence or racially gerrymandered, then so do their functionally similar successors.
3. The Next Dominoes: Milligan indisputably means that Alabama will have to draw a second Black-opportunity congressional district. It also likely has this implication for Georgia and Louisiana: both states where one more compact, reasonably-configured, Black-majority congressional district can easily be drawn. Additionally, Milligan may complicate Republicans’ plans in North Carolina to obliterate the Black-opportunity congressional district in the eastern part of the state. Racial polarization in voting would have to be established first (as it was not in Cooper), and it would have to be possible to draw a reasonable Black-majority district in that region. More generally, Milligan should change the gestalt around racial vote dilution claims, especially for conservative judges. In recent cases, some conservative judges have seemed to believe that they can (even should) always rule for defendants, including by inventing new doctrines (like Section 2 not containing a private right of action), requiring a racial explanation for racially polarized voting, and intimating that Section 2 is unconstitutional. Milligan should bring an end to this behavior. Conservative judges must now live with a controlling precedent authored by a conservative Chief Justice clearly finding liability under many of the same circumstances where they previously ruled for defendants.
4. Redistricting Simulations: As in Rucho, Roberts was highly critical of redistricting simulations in Milligan. Here, his skepticism was warranted. Section 2 already has a benchmark for assessing racial vote dilution: the demonstration map(s) offered by the plaintiffs. So Section 2 has no need for a different baseline, like the mean or median map randomly generated by a computer algorithm. This alternative baseline is especially inappropriate because it’s typically race-blind. A race-blind baseline makes sense if the question is whether a plan or district was motivated by racially discriminatory intent. But that’s not the issue under Section 2, which is primarily or exclusively violated by racially discriminatory effect.
That said, even after Milligan, Section 2 plaintiffs may want to use race-conscious computer algorithms to find maps with more compact, reasonably-configured, majority-minority districts. Having identified such maps, plaintiffs could then use them to satisfy Gingles’s first prong. This usage, critically, wouldn’t entail the introduction of a whole array of computer-generated maps, with which the enacted plan would be compared. Rather, plaintiffs would just offer one or a handful of computer-generated maps, to show that more compact, reasonably-configured, majority-minority districts could be drawn.
In addition, while redistricting simulations may be conceptually unsuited to racial vote dilution claims, they remain relevant to claims of racial gerrymandering. In a racial gerrymandering case—unlike in a racial vote dilution case—intent is indeed the key. Strict scrutiny follows if race was the predominant motivation for the construction of a given district. As I noted above, redistricting simulations aren’t useful for establishing racially discriminatory effect, but they can be helpful circumstantial evidence of predominant racial intent. The idea is to include all of a jurisdiction’s nonracial criteria in an algorithm but to exclude race itself. The racial demographics of the computer-generated districts are then compared to those of the enacted districts. If these racial demographics are significantly different in the computer-generated districts, that suggests that race predominated in the design of the enacted districts. If race hadn’t predominated, the enacted districts would have resembled the computer-generated districts more closely.
5. The Constitutionality of Section 2: Finally, and maybe most importantly, Milligan resolved the question of Section 2’s constitutionality. For decades, the Court had assumed that compliance with Section 2 is a compelling state interest without squarely addressing Section 2’s validity. In Milligan, the Court definitively “reject[ed] Alabama’s argument that § 2 as applied to redistricting is unconstitutional under the Fifteenth Amendment.” Justice Kavanaugh fully joined this portion of the opinion notwithstanding his cryptic suggestion in his concurrence that “the authority to conduct race-based redistricting” might not “extend indefinitely into the future.”
This holding that one of the pillars of the VRA is constitutional is obviously important for its own sake (especially after Shelby County). It also takes off the table one of the common claims of conservative litigants and judges: that a narrow construction of Section 2, or a ruling in favor of defendants, is necessary to avoid constitutional problems. It’s now clear that the reasonable application of the Gingles framework—including when this application yields victories for plaintiffs—is perfectly constitutional. Milligan has implications, too, for other disparate impact statutes like Title VII of the Civil Rights Act and the Fair Housing Act. If Section 2 is valid even though it “outlaw[s] voting practices that are discriminatory in effect,” not intent, the same should be true of other statutes that similarly target racial disparities. If it comes to pass, this dispelling of the constitutional cloud that hangs over disparate impact law would be even more consequential than anything Milligan said about Section 2 itself.