In my previous comment, I defended Justice Kennedy's decision in the Texas redistricting case, LULAC v. Perry, to refrain from attempting to identify when (if ever) the effect of a partisan gerrymander might suffice to establish a constitutionally cognizable injury. In light of the Court's historical difficulties with this especially complicated and politically sensitive question, and because the plaintiffs in LULAC attacked only the mid-decade timing of the Texas redistricting, Justice Kennedy acted responsibly in leaving this larger question unresolved.
Here, as promised, I wish to defend the other major portion of Justice Kennedy's opinion in LULAC , the part where, writing for a five-member majority of the Court, he invalidated the Texas map's District 23 as a violation of the Voting Rights Act.
The problem with District 23, as the majority saw it, was that it split in half the Laredo-based Latino community, which had been wholly within the district as drawn in the previous map. In the old District 23, moreover, Latinos had comprised a majority of the citizen voting-age population (57.5%) and had begun to exercise effective political control over that district, whereas in the new District 23 Latinos were reduced to a minority of the citizen voting-age population (46%). The split of Laredo 's Latinos, therefore, had the effect of diluting Latino voting power, thereby violating section 2 of the Act.
The state defended against this vote dilution claim by arguing that District 25 of the new map contained a majority of Latino voters and, because both the old map and new map contained six majority-Latino districts, there was no overall dilution of Latino voting power in the new map. Kennedy's opinion for the Court, however, rejected this argument on the ground that the loss of voting power for the Latino community in Laredo could not be offset by an increase in voting power of Latinos elsewhere in the state. Only if those other Latinos were equally entitled under the Voting Rights Act to a district of their own, and only if it would be impossible for the map to provide each separate Latino community with its own majority-Latino district, would it be permissible for the mapmakers to trade off the loss to the Laredo Latinos with a gain to Latinos elsewhere. That situation, however, was not present in this case because the Latinos in the new District 25 were geographically dispersed and therefore not entitled to their own VRA-mandated district.
Chief Justice Roberts, in an opinion joined by Justice Alito, vigorously dissented from the Court's reasoning and conclusion on this vote dilution claim. In essence, Chief Justice Roberts believed that the geographic dispersion of the Latino residents in the new District 25 was irrelevant to a proper analysis of the vote dilution claim. As long as the percentage of Latino-controlled districts in a map is roughly proportional to the percentage of Latinos in the overall population, then the Chief Justice would require dismissal of the vote dilution claim. Although he would prefer to conduct this proportionality inquiry on a regional, rather than statewide basis, Chief Justice Roberts argued that either way the inquiry was satisfied on the facts of this case. As he observed, Latinos controlled 19% of the state's congressional delegation and comprised 22% of the state's citizen voting-age population, figures which should satisfy any standard of rough proportionality.
The Chief Justice's dissent is a rhetorical tour de force, as befits a justice who was one of the nation's leading appellate litigators before becoming a judge. But for all the "style points" his dissent deserves (to use a phrase from the dissent itself), it fails in my judgment to refute the Court's two main points: first, that the specific harm to the Laredo Latinos establishes the basis for a vote dilution claim, even if the map overall provides proportionality; and second, that the creation of the new District 25, with its Latino majority, cannot offset the distinctive harm to the Latino voters of Laredo.
To understand why the Court, rather than Chief Justice Roberts, is more persuasive on these points it is necessary to recall an Equal Protection claim that the plaintiffs presented as an alternative to their Voting Rights Act claim. Although it became unnecessary for the Court to adjudicate the merits of this alternative Equal Protection claim once the Court accepts the Voting Rights Act claim, as Justice Kennedy's opinion for the Court explicitly observes, it hovers in the background and undoubtedly influences the reasoning of the Court's opinion on the Voting Rights Act issues. This influence is hardly inappropriate, as it is standard operating procedure that statutes should be interpreted so as to avoid constitutional questions. Therefore, to the extent that section 2 of the Voting Rights Act is open to alternative interpretations-as it surely is, given its lofty but general, quasi-constitutional language-Justice Kennedy was entirely justified in suffusing the Court's analysis of the section 2 vote dilution claim with Equal Protection overtones.
One very important virtue of Justice Kennedy's approach, rather than relying directly on an interpretation of the Equal Protection Clause itself, is that, if Congress does not like the Court's theory of a successful vote dilution claim as articulated in LULAC, it can always amend the Voting Rights Act to make its position clear. A congressional directive of this kind would give the Court the reason to take "a sober second look" before proceeding directly to interpreting the Constitution to entail a theory of vote dilution liability that Congress considers unwise.
The alternative Equal Protection claim is that Texas legislature engaged in intentional race discrimination when it split apart the Latino community in Laredo , removing almost 100,000 Latinos from the old District 23. Some commentators (myself included on occasion) have loosely described this Equal Protection claim as asserting a race-based gerrymander in violation of the Equal Protection Clause according to the doctrine first articulated in Shaw v. Reno. But, in truth, this intentional vote dilution claim regarding the splitting of District 23 is not a Shaw-derivative racial gerrymandering claim, but rather a different kind of Equal Protection claim, as Justice Scalia's opinion in LULAC correctly observes. Shaw -based claims require no showing of vote dilution; rather, they originated in response to a kind of "affirmative action" undertaken to increase the voting power of racial minorities. There was, indeed, a separate Shaw-based claim in LULAC, but it involved the addition of Latino voters to District 25, not the subtraction of Latino voters from District 23, as Justice Scalia also correctly recognized. (That separate Shaw-based claim also was unnecessary to address, as Justice Kennedy properly noted for the Court, once District 23 was found invalid under the Voting Rights Act, thereby requiring the case to be remanded to make a new map.)
With respect to thinking about the Equal Protection challenge to the removal of Latino voters from District 23, on the ground that it was intentional vote dilution rather than Shaw -based gerrymandering, it appears that members of the Court may have become confused about the proof of intentional discrimination necessary to show an Equal Protection violation. Acknowledging as was inevitable that the Texas legislature knew full well that it was removing almost 100,000 Latino voters from District 23 and that it was doing so on purpose, the State's attorney attempted to defend against a finding of intentional vote dilution on the ground that the reason the legislature was consciously treating Latino voters this way was to protect the incumbency of Representative Henry Bonilla-a reason that the state asserted to be a political rather than racial motive. Justice Scalia bought this argument, while other Justices (including Justice Kennedy and his colleagues in the majority) were evidently troubled by it.
But the State's argument on this point seems to me to confuse the intent to disadvantage a racial minority, which triggers strict scrutiny under conventional Equal Protection analysis, with the reason (or motive) for the government's intent to discriminate, which is the justification that the government asserts in an attempt to satisfy strict scrutiny. If the harm to Latinos in Laredo did not involve a reduction of their voting power caused by the redrawing of district lines, but instead another form of impediment to their exercise of electoral power, I think it would be easier for the Justices to recognize this distinction. For example, suppose that the Texas legislature required Latinos in Laredo to provide an additional form of voter identification not required of Anglo voters within District 23. That requirement undoubtedly would qualify as intentional race-based discrimination triggering strict scrutiny under the Equal Protection Clause. If the State attempted to defend this additional identification requirement on the ground that it was desiring to protect Representative Bonilla's incumbency (because diminishing Latino turnout would increase his chances of winning), that justification would be subjected to strict scrutiny-and presumably would fail-but it would never succeed as an argument that the burden on Latino voters was not form of intentional race-based discrimination.
In the context of an intentional vote dilution claim of the kind involving the conscious and purposeful removal of Latino voters from District 23, the argument that this burden on the exercise of Latino voting power was undertaken with an incumbency-protection motive should meet the same fate as it would if the burden on voting rights was the imposition of a discriminatory identification requirement. This situation is different than a Shaw-based claim, where without diluting minority voting power the legislature decides to include a group of minority voters in one district rather than another because drawing the district lines this way is more favorable to an incumbent. In that non-dilution context, it is plausible to say, as the Court has in post-Shaw cases like Easley v. Cromartie, that this incumbency-motivated line drawing is not intentional race discrimination in violation of the Equal Protection Clause even though the mapmakers are obviously aware of racial demographics when they draw the lines.
But the treatment of the issue of intent for Shaw-based claims should have no bearing when the Equal Protection claim is one of intentional vote dilution. Purposefully manipulating district lines to undercut minority voting power is, regrettably (indeed, outrageously) a longstanding practice in U.S. history. The Supreme Court, most famously, confronted an especially egregious example of this intentional vote dilution in the 1960 case Gomillion v. Lightfoot. There, Alabama had manipulated Tuskegee 's boundary to prevent blacks from exercising political power in the city, and the Court had no difficulty determining that this mapmaking exercise was intentional race discrimination requiring a compelling justification from the state. (None was forthcoming in the case.) Although perhaps not as egregious as the dilution of black voting power in Gomillion, the dilution of Latino voting power in LULAC by splitting Laredo in half is the exact same kind of consciously imposed harm on a local minority community as occurred in Gomillion and should be subject to the same strict constitutional scrutiny. In other words, when a legislature willfully redraws district lines to undermine a local minority community's political power, that deliberate burden should count as intentional race discrimination for Equal Protection purposes, whether the legislature's motive for imposing this burden is a belief in the inferiority of the minority group or a desire to protect an incumbency whose reelection is opposed by most members of the minority group.
In any event, whatever is the appropriate resolution of this Equal Protection issue, an inclination towards this position appropriately informed Justice Kennedy's consideration, on behalf of the Court, of the vote dilution claim under the Voting Rights Act. In a crucial passage in his opinion, where he explains why the proportionality inquiry ultimately cannot defeat the section 2 claim involving District 23, he observes that the removal of Latinos from the district "bears the mark of intentional discrimination that could give rise to an equal protection violation." (Slip op. at 34.) This fact distinguishes the section 2 claim here from an ordinary vote dilution claim under VRA's section 2, which requires no showing that the dilution was intentional. Even if a State's constitution required a computer to draw district lines based solely on maximizing district compactness, without any consideration of racial demographics whatsoever (the computer selecting randomly among maps with equally compact districts), the resulting map could be challenged under section 2 if it failed to produce minority-controlled districts in circumstances where such districts could have been drawn. Demonstrating that the map has the effect of diluting minority voting power is enough to invalidate the map under section 2.
In an ordinary section 2 case, where only the effect of dilution is shown, the proportionality inquiry makes sense. Don't invalidate the computer's compactness-maximizing map just because it failed to produce a minority-controlled district that might have been drawn in one part of the state if it turns out that the percentage of minority-controlled districts in the map overall is roughly proportional to the percentage of minority citizen voting-age population overall. In this situation, the state's minority voters have not suffered a discriminatory harm.
The circumstances are entirely different, however, where, as in LULAC, minority voters in a specific locality are deliberately targeted by the legislature to undermine their political power. In Justice Kennedy's own words, the basic fact of the case that "the State divided the cohesive Latino community in Laredo]" because they "threatened Bonilla's incumbency" meant that "[i]n essence the State took away the Latinos' [political power] because Latinos were about to exercise it." (Slip op. at 33. 34.) Where this kind of deliberate undermining of minority political power occurs, the proportionality inquiry of an ordinary section 2 claim is beside the point. Statewide proportionality, as Justice Kennedy explained, could not "overcome" the harm of vote dilution specifically inflicted upon the Latinos of Laredo.
This conclusion is entirely sound. If Latinos in Laredo were targeted with another kind of discriminatory burden-for example, a reduction of their share of scholarships to Texas state universities, so that Anglos in Rep. Bonilla's district could receive a disproportionately large share of these scholarships-it would not matter that statewide the percentage of scholarships received by Latinos were proportional to the Latino percentage of the statewide population eligible to receive these scholarships. In the special case of a section 2 vote dilution claim predicated on the deliberate targeting of a local minority community, the proportionality inquiry similarly should have no place.
This conclusion also goes a long way toward explaining why the addition of Latino-controlled District 25 does not negate the vote dilution claim regarding District 23. In essence, the State's reliance on District 25 to offset District 23 is just another version of the proportionality defense, writ small. But giving extra scholarships to Latino students in Austin could not justify depriving Latino students in Laredo of scholarships in order to bestow extra scholarships on Anglo students within Rep. Bonilla's constituency. Similarly, giving greater voting power to Latinos in District 25 cannot compensate for diminishing the voting power of Latinos in Laredo in order to protect the political interests of Rep. Bonilla and the predominantly Anglo constituents who support him.
Chief Justice Roberts attempted to deflect this basic point by arguing that the Latinos of Laredo were no more entitled in the first place to a majority-Latino district than were the Latinos of District 25. He observed that the old District 23 was no more compact (or otherwise geographically compelling) than the new District 23 and, therefore, the State originally could have chosen equally between the contours of the old District 23 and the new District 25. But, contrary to the Chief Justice's argument, it does not follow from this observation that, when replacing the predecessor map, the State now is permitted to substitute the new District 25 for the old District 23. Even if Latino students in Laredo were no more entitled to receive scholarships than Latino students in Austin before any of these scholarships were awarded, once they are awarded the Latino students of Laredo cannot be deprived of them on the ground that Rep. Bonilla wants his Anglo constituents to receive more of the scholarships allocated to students in his constituency.
In his opinion for the Court, Justice Kennedy makes much of the fact that the Latino community of Laredo had been exercising its political power within District 23 before that power was taken away to benefit Rep. Bonilla and his Anglo supporters. He was correct to do so. Even if those Laredo Latinos had not been entitled to that political power in the first place, once they have it, the State should not be permitted to take it away from them in order to increase the political power of Anglos whose political interests, as demonstrated by past voting patterns, are opposed to theirs.
To be sure, by deciding that it violates VRA's section 2 for a state to deprive a local minority community of preexisting political power, the Court in LULAC appears to import the non-retrogression principle of VRA's section 5 into a vote dilution claim under section 2. But I would suggest that this apparent importation of the non-retrogression principle is limited to the special circumstance in which a vote dilution claim under section 2 is predicated on the deliberate targeting of minority voting power. In other words, it is limited to only claims of intentional vote dilution under section 2 that resemble an Equal Protection claim. It would not apply to purely effects-based vote dilution claims.
Ultimately, then, it is the deliberate nature of the injury to Latino voters of Laredo that justifies the Court's decision to sustain their Voting Rights Act claim and provides the basis for thinking that Justice Kennedy, rather than Chief Justice Roberts, got the better of the argument. Indeed, the Chief Justice neglects this distinctive feature of the vote dilution claim in LULAC. He treats the removal of Latino voters from District 23 as nothing more than a conventional, purely effects-based claim under section 2. He does not consider whether the purposeful targeting of Laredo Latinos for a loss of voting power to protect Rep. Bonilla's incumbency presents a special case of intentional vote dilution under section 2.
Thus, for all its rhetorical power and flair, Chief Justice Roberts' opinion fails to grapple with what is most important about the Voting Rights Act claim regarding the removal of Latino voters from District 23. In essence, then, he misses (or ignores) the Court's main point.
It would be most unfortunate, however, if academic analysis and public commentary on LULAC also misses this main point, as easily could occur if the focus becomes whether in terms of craft Kennedy or Roberts was the more effective advocate of his position. Roberts may well be the better wordsmith, but Kennedy was ultimately stronger on the substance of the law.
Kennedy kept focused on the key fact that Latino voters had been deprived of voting power to increase the voting power of Anglos. In the end, nothing else mattered-nor should matter, as surely it is a proper interpretation of the Voting Rights Act to find this kind of deprivation unlawful.
CORRECTION: The original version of this column contained a paragraph asserting that Chief Justice Roberts never addressed the plaintiffs' alternative Equal Protection claim. Chief Justice Roberts, however, joined Part III of Justice Scalia's opinion, which did address this claim. Therefore, that paragraph has now been deleted. Furthermore, Justice Scalia's rejection of the Equal Protection claim on the ground that the removal of Latinos from District 23 was politically rather than racially motivated, and thus lacked the necessary element of intentional discrimination for Equal Protection purposes--although mistaken in my judgment (for reasons detailed in the column)--explains why Chief Justice Roberts did not consider whether analysis of the Voting Rights Act claim should be influenced by the alternative Equal Protection claim. Nonetheless, in responding to Justice Kennedy's understanding of the Voting Rights Act claim, Chief Justice Roberts might have acknowledged that, where intentional diminishment of minority voting power is evident, then the analysis under the Voting Rights Act should be different than if only the effect of minority vote dilution is shown.