The following is a guest post from Kerrel Murray:
Just in time for this week’s reargument of Louisiana v. Callais, I’ve posted to SSRN a draft of my latest Article, False Conflict: Colorblindness and Section Two of the Voting Rights Act, forthcoming next year in the Fordham Law Review. I relied on the Article for an amicus brief I filed in Callais, with the help of a great team at WilmerHale. I’m grateful to Professor Hasen and the Election Law Blog for letting me draw on the Article and the brief to blog about three points that I think are key to the case.
Disparate Treatment, Not Disparate Impact
This blog’s readers know that §2’s prohibition of electoral rules with certain results reflects its amendment in 1982 to repudiate the City of Mobile v. Bolden plurality’s view that the original §2 required proof of discriminatory purpose. That origin sometimes leads commentators to describe the amended §2 as adopting a disparate-impact standard. Yet, as scholars have observed, it would be infelicitous for §2 to be a “pure” disparate-impact provision. Tying liability to electoral proportions alone would risk running afoul of key tenets of colorblind constitutionalism, such as its abhorrence of “racial balancing” and evergreen, indefinite justifications for injecting race into politics. While a carefully calibrated disparate-impact regime can deal with those problems, §2 need not court them. A closer look at its history and context suggests a preferable label: disparate treatment.
The key is §2’s core value: the focus on “racial politics” found not just in the seminal legislative history, but in contemporary coverage of the amendment’s design. That coverage aligned “racial politics” with circumstances in which the effects of past or present intentional discrimination so distort a political system that a genuinely neutral electoral device may treat voters disparately on account of race. Though they do not use the phrase, the key pre-Bolden cases (which §2’s text explicitly invokes) are best read to pursue this quarry. In brief, §2’s design attacks the present-day incorporation of the effects of discrimination into the public machinery for allocating political power. It is unconcerned with the effects themselves unless and until some government action gives them present-day electoral force.
The distinction between disparate treatment and disparate impact may seem subtle. Smoking out present intentional discrimination or remedying past intentional discrimination are among plausible justificationsfor disparate-impact rules, too. But disparate impact might, though it need not, sweep more broadly—for example, by treating group-level disproportionality as a constituent part of the wrongdoing it targets. The disparate-treatment label thus distinguishes §2’s design from the colorblindness-linked risks associated with the disparate-impact label.
“Disparate treatment” also points toward a precise, and important, causal account. As readers likely know, the Supreme Court has held that only intentional discrimination violates the provisions of the Fourteenth and Fifteenth Amendments that §2 most plausibly enforces. How, then, does banning certain electoral devices irrespective of their enactors’ intent enforce those provisions?
As to that problem, the necessary link inheres in §2’s design. The pre-Bolden ethos that §2 incorporated looked to particular kinds of facts about racial politics to support particular inferences about discrimination’s distortive effects. The disparate-treatment label and the evidence that supports it reveal §2’s design as embodying (1) a calibrated assertion thatcausal links between the effects of discrimination and present-day racialized politics may exist, paired with (2) an account of how present-day facts may justify inferring such a connection. That design rejects racial assumptions. And it necessarily assumes that such an inference may be rebutted, even if previously justified. It thus moves §2 away from remedying generalized “societal discrimination” to a focus on particularized present-day facts prima facie best explained by particularized links to past discrimination.
Gingles, Self-Liquidation, and Temporal Limitations
Thus far, I have been speaking of what one might call §2’s original understanding. Subsequent developments only bolster this account.
Thornburg v. Gingles articulated three now-famous preconditions to §2 liability. But we have underappreciated the degree to which Gingles and its progeny implemented the disparate-treatment design that I assert. Every precondition, it turns out, looks to changeable facts about the world that plausibly reflect past or present intentional discrimination. Gingles itself appreciated this aspect of §2 when it approvingly noted the Senate Report’s claim that the results test was necessary to eradicate “the effects of past purposeful discrimination.” Importantly, too, these are the sorts of facts that can and do dissipate (and I identify in the Article recent empirical trends showing that dissipation is not just theory, but fact).
The first precondition requires that the plaintiff be a member of a minority group “sufficiently large and geographically compact to constitute a majority in a single-member district.” Here, I think commentators have underemphasized the link between past or present discrimination and this precondition’s implicit requirement of persistent racially identifiable housing. To be sure, this has not been the Court’s explicit explanation for this prong. But history teaches in detail of the links between discrimination and racially identifiable housing patterns. The heuristic is imperfect, but housing sufficiently integrated to thwart a §2 claim plausibly suggests either intentional discrimination’s absence or the eradication of its effects. If anything, this precondition may produce false negatives, preventing §2 from activating where it should. But if so, §2 is to that extent less likely to trigger where it ought not, which cuts against any narrative in which §2 has overflowed the channel Congress cut in 1982.
For their part, the second and third preconditions, taken together, require a bloc-voting racial majority that consistently votes against and defeats the preferred candidates of a bloc-voting racial minority. This “racially polarized voting” is, in brief, outcome-determinative racial agonism. When a government’s vote-aggregation device incorporates that state of affairs into its results, the harm that follows is (as the Court recognized two years ago in Allen v. Milligan) “plausibly on account of race.” Specifically, it is plausibly on account of past or present discrimination’s effects. Racially polarized voting is evidence (not conclusive evidence, but evidence nonetheless) that some exogenous distortion, like discrimination, has left a political system unable to aggregate beyond race. After all, the idea that melanocytes inherently determine voting preferences is at war with colorblind constitutionalism. This phenomenon, too, is the sort that can and does wane.
Finally, Gingles’ progeny require plaintiffs to independently satisfy a totality-of-the-circumstances standard. Doctrine elaborating that standard makes clear that the inferences justified by the satisfaction of the preconditions are rebuttable. Defendants may present evidence on the point, including evidence that complained-of electoral losses are on account of (e.g.) politics, rather than race. This too helps keep §2 limited to instances of present-day disparate treatment flowing from the effects of discrimination.
Scholars often claim that §2 “self-liquidates,” i.e., that its ultimate quiescence is built into its design. By tying §2’s activation to changeable facts indicative of discrimination’s effects, Gingles both implements §2’s disparate-treatment design and ensures its self-liquidation. We thus can and should have confidence in two things. First, when §2 does activate, it does so with good reason. Second, the need for the race-conscious redistricting that it sometimes requires will not, to quote the concern Justice Kavanaugh expressed in Milligan, extend “indefinitely into the future.”
The Zero-Sum Problem’s Inapplicability
Last is a point relating to the zero-sum idea critical to the Chief Justice’s majority opinion in Students for Fair Admissions v. Harvard. SFFA placed substantial weight on the idea that college admissions are zero-sum; if one gives race-based preferences to some but not others, the latter group loses out. Whatever might be said about that reasoning, it is inapt here.
Scholars often recognize the “right” to vote as comprising at least two facets: a participation right (the right to cast a ballot that is counted) and an aggregation right (the right to have one’s ballot combined with one’s copartisans on equal terms with all cast ballots). As to participation, the zero-sum idea is misplaced. My right to cast a ballot and have it counted loses no value when other voters exercise their own right, and my vote stops no one else from voting.
So too for aggregation. Of course, a voter might self-interestedly prefer district lines which ensure that aggregation in her district combines her ballot with enough of her copartisans to outnumber the aggregated ballots of her electoral foes. Likewise might she prefer statewide district lines guaranteeing her preferred political party legislative control. But the aggregation right contemplates a right to have one’s ballot aggregated on equal terms with others, not a right to maximally beneficial aggregation rules. Section 2 does not burden that right when, to cure racial vote dilution, it moves a majority-race voter from one aggregative regime that does not dilute her vote to a different, still nondilutive regime. The playing field §2 alters is thus singularly non-zero sum. Because of its disparate-treatment design, it deprives voters only of that to which they are not legally entitled: electoral advantages flowing from an electoral system distorted by racial politics. It strikes me that any argument to the contrary would invoke something unfamiliar to American law, namely, a freestanding right to a preferred electoral outcome.
***
The Article says much more, including an argument (aligning with one made by Professor Stephanopoulos in his amicus brief) that §2 does not involve “racial classifications” at all. But I have gone on long enough. Hopefully, long enough to encourage you to read the whole thing!