Yesterday, the US Supreme Court handed down a historic and somewhat unexpected ruling that upheld both the Voting Rights Act and election science. In Allen v Milligan, Chief Justice John Roberts, joined by Justices Ketanji Brown Jackson, Elena Kagan, Brett Kavanaugh, and Sonia Sotomayor, rejected the state of Alabama’s argument that only “racially blind” methods of redistricting meet constitutional requirements of race neutrality. As a result of this decision, Alabama and several other states (Georgia, Louisiana, possibly Texas) will likely be required to redraw Congressional and/or state legislative districts.
This is a landmark case with a lot to unpack, and I will have more to say in the future (especially about the Court’s continuing struggle to reconcile the principles of proportionality and political equality), but what’s immediately notable is the central role that election science played in the outcome of this case, and the potential for the Court’s interpretation of that science to shape future Voting Rights Act cases. Importantly, the high court reaffirmed the constitutionality of what is known as the Gingles test for identifying and remedying racial discrimination in redistricting, which came as surprise for most of the nation’s election law scholars and voting rights advocates.
A fair application of scientific standards
First, the high court upheld the scientific standards used to analyze racial vote dilution, which led the majority to a fair ruling. Second, the use of a specific method, namely algorithmic-generated mapping simulations, played a central role in all of the opinions offered by the Court, which placed clear limits on their application.
It may seem unusual to point this out, but I am not the only observer who was surprised to see this Supreme Court squarely address the question before it through such a straightforward application of established scientific standards and analysis.
For background: The Gingles test, established in the 1986 racial vote-dilution case Thornburg v Gingles, is a three-prong test that the Court developed to assess whether a districting plan violates section two of the Voting Rights Act:
“First, [the] ‘minority group’ [whose interest the plaintiff represents] must be ‘sufficiently large and geographically compact to constitute a majority’ in some reasonably configured legislative district. Second, the minority group must be ‘politically cohesive.’ And third, a district’s white majority must ‘vote sufficiently as a bloc’ to usually ‘defeat the minority’s preferred candidate.’” [Cooper v. Harris, 581 U. S. 285, 301–302]
The establishment of the Gingles test represented a major voting rights litigation advancement. It offers an evidence-based framework through which racially polarized voting can be demonstrated, and it shows how altering the proportion of like-minded voters within district boundaries can alter electoral outcomes by diluting the voting strength of targeted voters.
The Alabama legislature’s decision to remove one of two African-American opportunity districts from their Congressional delegation by splitting the Black Belt population in the southern part of the state was a fairly clear case of vote dilution. Alabama’s argument in Allen v Milligan was fundamentally a challenge to this test which the Court has relied on for nearly four decades. Alabama had argued that the use of racial demographic and voting data in the Gingles test unconstitutionally made race the predominant factor in determining redistricting outcomes. The Court’s response to the challenge was unambiguous. It ruled that Alabama’s approach would require abandoning this precedent and overruling the interpretation of Section 2 of the Voting Rights Act as set out in nearly a dozen of previous cases. As the majority ruling put it: “We decline to take that step.”
For voting rights advocates, this ruling came as a pleasant surprise. As election scholar Guy-Uriel E. Charles’ noted: “It has been such a long time since I’ve expected a majority of the Court to fairly apply its prior voting rights precedents, to interpret the VRA without malice, and to read the record fairly that I have forgotten what that looks like.”
Chief Justice Roberts was clearly averse to the sort of legal upheaval that Alabama’s re-interpretation of the Voting Rights Act would have produced. Instead, this is the first time the Roberts Court has made a decision that requires the drawing of a new minority-opportunity district.
At least for now, science—and justice—prevailed.
I say “for now” because, while the majority opinion reaffirmed that Section 2 of the Voting Rights Act “demands consideration of race” given that the question of whether additional majority-minority districts can be drawn necessarily involves a “quintessentially race-conscious calculus,” Justice Kavanaugh did not concur with this section of the decision. He claimed in his own opinion that “the authority to conduct race-based redistricting cannot extend indefinitely into the future.” Further, the majority opinion made it clear that this decision, far from expanding voting rights, anticipates a reduction in successful voting rights litigation. The majority wrote: “as residential segregation decreases—as it has ‘sharply’ done since the 1970s—satisfying traditional districting criteria such as the compactness requirement ‘becomes more difficult.’”
Appropriate limits on applying mapping simulations
Chief Justice Roberts is no fan of math. In 2017, during oral arguments in the partisan gerrymandering case Gill v Whitford, the Chief Justice referred to methods for estimating partisan vote dilution (including one proposed by my research team) as “sociological gobbledygook.” Yesterday’s opinion was similarly critical of the use of ensembles of algorithmically-generated districting maps as a means to evaluate valid alternatives.
The opinion correctly notes one of the primary limitations of the approach. Citing amicus briefs from computational scientists who note that it is “effectively impossible to generate a complete enumeration of all potential districting plans,” the Supreme Court dismissed the claim that a large number, or any number, of ensembles of maps could constitute an accurate reflection of what the actual distribution of all possible maps could be.
This is an important clarification, because it is intuitive for a judge (and even some statisticians) to look at a “normal” distribution of simulated maps and see outliers as evidence of a map that is “abnormal” or intentionally biased, for example. But since the 1960s, when mapping algorithms were first developed, it has been known that automated algorithms cannot be counted on to search the landscape of possible legal redistricting plans in an unbiased manner.
Lower courts should take note. Alabama took the erroneous reliance on automated algorithms a preposterous step further, claiming that their so-called “blinded” (i.e. excluding racial data) simulations should be used as a benchmark for maps that are, by their circular definition, racially neutral, on the false premise that racial blindness is the equivalent of racial neutrality in the generation of districting maps.
This is not to say that ensembles are not an important technology. Ensembles are especially useful in revealing the consequences of specific tradeoffs incorporated into the redistricting process. The process must try to balance conflicting criteria: contiguity, population equality, compactness, concurrence with other government boundaries, respect for communities of interest, and, as computational scientist Moon Duchin explained as an expert in the case, the “non-negotiable” criterion of political equality, that is, equally weighted votes, for those protected under the Voting Rights Act.
In this case, at least, science won out, preserving Section 2 of the Voting Rights Act and supporting the federal government’s obligation to protect voters from racial discrimination, as required by the 15th Amendment. Allen is a historic decision with a lot more to unpack, but today science advocates and defenders of democracy across the country have reason to celebrate.