Regardless of one’s position on the role that race should play in modern politics, the racial polarization of American voters is undeniable: in 2016, black and Latino voters preferred the Democratic presidential candidate by eighty and thirty-six points respectively, while white voters preferred the Republican presidential candidate by twenty-one points.1 Perhaps it is no surprise, then, that both party and race come into play when sorting voters to achieve either partisan or racial advantage. One method that states use to achieve such advantage is redistricting — the redrawing of district boundaries after each census, intended to ensure that Congress and state legislatures remain representative.2 Redistricting plans can be gerrymandered to dilute racial or political minorities’ votes by “packing” minority voters into a few districts or “cracking” minority groups across many districts.3 Although gerrymandering is often discussed as a partisan issue,4 the Court has dealt with it only as a matter of equal protection for racial minorities, such that racial gerrymandering is unconstitutional,5 whereas partisan gerrymandering is not.6
Last Term, in Cooper v. Harris,7 the Court affirmed that two North Carolina congressional districts were unconstitutional racial gerrymanders.8 The fate of one of the two districts, District 12, rested on whether it was race or politics that predominated the legislature’s decision to sort voters into the district.9 Although gerrymandering jurisprudence attempts to draw a clear line between race and party, that line is difficult to maintain, especially when the Court in Harris admitted that the underlying facts of the case could plausibly point in either direction.10 Given the independent harms of partisan gerrymandering and the unique entanglement of race and party, it is both unnecessary and futile for courts to tackle the impossible challenge of distinguishing between racial and partisan gerrymandering. Rather, the fact that Harris could easily have gone the other way suggests the need for an additional standard that addresses partisan gerrymandering directly and thus allows courts to police more accurately the affront to democracy that both racial and partisan gerrymandering represent.
After the 2010 census, North Carolina redrew its congressional districts to reflect population changes, in compliance with the U.S. Constitution.11 Through the Equal Protection Clause, the same Constitution prevents a state, “without sufficient justification, from ‘separat[ing] its citizens into different voting districts on the basis of race.’”12 Race-based lines, therefore, are unconstitutional where (1) “race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without a particular district,”13 and (2) the district’s design cannot withstand strict scrutiny.14 To pass strict scrutiny, the state must prove that its race-based redistricting scheme is “narrowly tailored” to meet a “compelling interest.”15 Compliance with the Voting Rights Act of 196516 (VRA) — which prohibits voting practices that racially discriminate and, of particular relevance to redistricting, prohibits “vote dilution” that diminishes the strength or effectiveness of a racial minority group’s vote — has been assumed by courts to be a compelling interest.17 Nevertheless, it is not a panacea for Equal Protection challenges; narrow tailoring requires that the legislature had “good reasons” to conclude that the VRA compelled its action.18
Two of the congressional districts that the North Carolina state legislature redrew after the 2010 census — District 1 and District 12 — were at issue in Harris.19 Both districts had been designed as majority-black districts after the 1990 census20 and had been challenged by white North Carolina residents as unconstitutional racial gerrymanders, to varying degrees of success.21 Following the 2000 census, Districts 1 and 12 were drawn with black voting-age populations (BVAP) of less than fifty percent,22 but both districts continued to elect the candidates preferred by black voters in the subsequent five general elections.23 Nevertheless, after the 2010 census, the Republican-controlled legislature24 designed a new map that again redrew Districts 1 and 12 as majority-black districts and thus prompted the present lawsuit.25
In Harris, the plaintiffs were black residents who alleged that the state had packed black voters into these two districts, “thereby diminishing their influence in surrounding districts.”26 They, like their white predecessors in the 1990s, argued that racial considerations predominated in the creation of the two districts and that the race-based lines could not withstand strict scrutiny.27 A three-judge district court agreed.28 For District 1, all three judges found that racial considerations predominated and held that the VRA did not compel a majority-black district.29 For District 12, two judges found that race predominated even over partisanship and held that, without any asserted compelling interest, its boundaries were impermissible.30
The Supreme Court affirmed. Writing for the Court, Justice Kagan31 held that race impermissibly predominated in the redesigning of Districts 1 and 12.32 The Court began by dismissing North Carolina’s claim that the plaintiffs were precluded from bringing suit because they are members of the same organizations that brought an earlier state court lawsuit, in which North Carolina prevailed.33 The state also claimed that the Court should not review the district court’s decision for clear error, given that the North Carolina Supreme Court reached a different outcome.34 The Court concluded that, while relevant, the existence of the state court decision could not alter the clear error standard of review.35 In fact, clear error review is based on the premise that there is often more than one view of the evidence, implying that both the state court’s and the district court’s views could be permissible.36
The Court then turned to District 1’s redesign. The record showed that the state intentionally established a racial target that “had a direct and significant impact” on the district’s boundaries,37 at the expense of traditional districting principles like compactness, contiguousness, geographical boundaries, and political subdivisions.38 Due to irrefutable evidence of race as a predominant factor in the redesign, the Court next inquired whether the racial gerrymander withstood strict scrutiny.39 The state argued that it had “good reasons” to believe that drawing a majority-minority district was necessary to avoid liability for vote dilution under section 2 of the VRA.40 Nevertheless, the Court concluded that the state could not have good reasons to believe that its district was compelled by section 2 if it did not also have good reasons to believe that, without a majority-minority district, a potential section 2 plaintiff would be able to prove a vote-dilution claim.41
The Supreme Court spoke directly to the latter point in Thornburg v. Gingles,42 in which it identified three threshold conditions for proving a vote dilution claim: (1) a “minority group” must be “sufficiently large and geographically compact to constitute a majority” in a legislative district that is designed sensibly; (2) that minority group must be “politically cohesive”; and (3) the district’s white majority must “vote[] sufficiently as a bloc” such that it typically “defeat[s] the minority’s preferred candidate.”43 Because District 1 consistently elected the black candidate of choice, even with a minority BVAP, the Court concluded that it did not meet the third Gingles condition.44 Thus, the Court affirmed the district court’s holding that racial predominance in the designing of District 1 did not withstand strict scrutiny.45
Next, the Court addressed District 12. In contrast to its District 1 analysis, the Court’s analysis of District 12 rested on the district court’s factual finding of motive. The plaintiffs contended that race predominated in the district’s redesign, whereas the state argued that the district was a “strictly” political gerrymander made without regard to race.46 During the trial for this case, the district court weighed a variety of evidence, including public statements by legislators, the mapmaker’s confirmation of racial intent, the state’s preclearance submission to the U.S. Department of Justice, live testimony by Congressman Mel Watt, and expert reports.47 The Court held that the district court’s finding that race predominated met the plausibility requirement of clear error review, while at the same time acknowledging that the Court may or may not have evaluated the testimony differently if it had had the opportunity to preside over the trial itself.48
Finally, the Court addressed North Carolina’s argument that the plaintiffs had failed to introduce an alternative map of District 12 as evidence.49 The state argued that in a racial-versus-political intent dispute, plaintiffs must show on a map that the state could have achieved its partisan goals without upsetting racial balance.50 The Court concluded that while an alternative map can be useful in such cases, it is not mandatory; the counterfactual evidence is essential only when plaintiffs lack sufficient direct evidence.51 Here, the plaintiffs’ direct evidence was enough to overcome the state’s partisanship defense.52
Justice Thomas concurred. He wrote separately to explain additional grounds on which he would affirm the district court’s decision: any intentional creation of a majority-minority district triggers strict scrutiny, and section 2 of the VRA does not apply to redistricting and thus cannot serve as a compelling interest for a racial gerrymander.53 Finally, he commended the Court’s deferential clear error review.54
Justice Alito concurred in the judgment in part and dissented in part.55 Agreeing with the majority on District 1, he took issue with the finding that race predominated in District 12.56 He explained the risks associated with mistaking a political gerrymander for a racial one, noting the potential usurpation of state authority and the “final, often-unstated danger where race and politics correlate: that the federal courts will be transformed into weapons of political warfare.”57
Justice Alito’s warning has salience, especially considering the strong historical correlation between race and politics. However, rather than steering clear of political warfare altogether, the Court should consider addressing it directly given the nontrivial harms of political gerrymandering. The sharp distinction between the constitutionality of racial versus partisan gerrymandering causes slightly different interpretations of evidence in trial courts to create drastically different outcomes for district lines. In Harris, the lower court’s finding that race predominated over politics in the drawing of District 12 allowed the Court to strike it down as an unconstitutional gerrymander. Nevertheless, the Court recognized that a different trial court could have upheld the district. The unique entanglement of race and politics, and the impractical task of discerning between the two, evinces the need for an additional standard that confronts partisan gerrymandering.
Race and politics have long been intertwined, such that both plaintiffs and defendants in redistricting cases often have linked racial and partisan goals. Even in late-nineteenth-century conflicts over North Carolina’s election laws, “racial politics affected the composition of the political parties, and parties engaged in racial politics in part for partisan ends.”58 In the 1990s, the first challenge to Districts 1 and 12 was made by Republicans alleging a partisan gerrymander.59 After that claim failed, these same opponents of the redistricting scheme challenged it successfully as an unconstitutional racial gerrymander.60 The dual nature of gerrymandering claims continues today. The Harris plaintiffs were represented by Democratic Party groups and allies, something that North Carolina’s counsel was quick to point out during oral argument.61 North Carolina urged the Court to force the plaintiffs to prove that they were “bring[ing] race claims and not dressed-up partisan claims” by presenting a feasible alternative map.62 However, this view ignores the harm that partisan claims aim to address, regardless of whether they are “dressed up” or not.
Judge Cogburn discussed these harms in his concurring opinion to the district court decision in this case. While stopping short of advocating for a partisan gerrymandering standard, Judge Cogburn wrote separately to share his concerns about “unfettered” political gerrymandering as an “affront to democracy.”63 He lamented the fact that partisan gerrymandering has become so routine that “the fundamental principle of the voters choosing their representative has nearly vanished” and “[i]nstead, representatives choose their voters.”64 The harm that Judge Cogburn and others — including the Court65 — have articulated is worth remedying on its own terms. This is especially true when meeting the “racial predominance” bar is almost a matter of chance: courts deliver opposite outcomes to similar claims — with similarly tangled partisan and racial justice objectives — depending on their reading of legislative intent and the map-drawing process.66
The Court in Harris recognized the difficulties in distinguishing race from politics. Justice Kagan noted that “political and racial reasons are capable of yielding similar oddities in a district’s boundaries” because of the high correlation between racial identification and party affiliation.67 Justice Alito remarked on the same in his dissent: race and voting approaches perfect correlation in District 12, meaning there is “almost complete overlap” between black and Democratic voters.68 Correlation aside, layered motivations and justifications also lead to confusion. For example, the district court majority treated the redistricting chairs’ statement attempting to counter charges of political gerrymandering as evidence discrediting their strictly partisan account.69 The district court dissent, however, treated statements about racial considerations simply as evidence that multiple factors were at play.70 The Court went even further, admitting that “legislators [may] use race as their predominant districting criterion with the end goal of advancing their partisan interests — perhaps thinking that a proposed district is more ‘sellable’ as a race-based VRA compliance measure than as a political gerrymander and will accomplish much the same thing.”71 At oral argument, Justice Kagan even proposed a hypothetical in which a state wants to segregate by race but uses a political justification because it “sounds better.”72
All of this goes to show that “race or party” is a false dichotomy because both are usually present, at least in effect and oftentimes in motive, justification, and methodology. In two footnotes, Harris acknowledged as much by stating that using race as a proxy for party is just as inexcusable as using race for its own ends.73 And yet, even while plainly stating the jumbled way that race and party interact in redistricting, the Court still required “evidence reveal[ing] that a legislature elevated race to the predominant criterion” for a plaintiff to succeed in showing an unconstitutional gerrymander.74 A predominance requirement suggests that the two factors can be separated, when truly, in cases like District 12’s, race and party are two sides of the same coin.
The unique evidentiary challenges in redistricting cases make it particularly difficult to determine racial predominance, which can lead to different outcomes even with the same facts. Although legislative intent is at issue in many areas of the law, here, there was no statute to be interpreted. Instead, courts assessing redistricting plans must evaluate legislatures’ map-drawing processes. As Justice Kagan’s hypothetical shows, this process does not always correspond to ultimate intent, and yet the exercise is expected to establish predominance of one factor over another. Moreover, disputes over legislative privilege and discovery procedures often make for trials that do not present the complete picture,75 even as lower courts undertake a “careful consideration of all evidence presented.”76 At the district court trial in Harris, two key witnesses — a redistricting committee chair and a state representative — did not testify, and the state instead cited their testimony in the earlier state court trial.77 Without an opportunity to see the witnesses on the stand, the district court was “somewhat crippled in its ability to assess either [witness’s] credibility.”78 Trials never provide perfect information, but the disparity among the conclusions reached by the various judges looking at this case indicates just how uncertain this outcome was.79
The Court did not ignore this uncertainty, acknowledging that it may well have reached a different conclusion than the trial court’s.80 During oral argument, Chief Justice Roberts drove home the role chance had played, pointing out the “fortuity that we have the Federal case before us and not the State case. And if it were the State case, we’d be reviewing their factual findings on the same question for clear error.”81 There is no question that District 12’s boundaries could easily have remained intact; a slight difference in the evidence presented at trial or a reordering of the state and federal cases could have led to the opposite outcome. Alone, this is not enough to necessitate a new standard; after all, trials and appeals lead to different outcomes all the time. However, in conjunction with the inextricability of race and party that the Court itself put forward and the independent harms of partisan gerrymandering, the stark difference in outcome is unjust and begs a separate standard that captures partisan gerrymandering.
Importantly, simply tightening the racial standard to remove the “predominance” requirement is not enough. There are a number of states in which gerrymanders have never been litigated, even though partisan gerrymandering is rampant, because those states do not have a sizable minority population. It seems inherently unfair that some states do not have the option to challenge gerrymandering at all, while others rely on racial gerrymander claims to correct for partisanship.82
Conversely, abandoning the racial standard altogether and focusing only on partisanship will also allow some problematic redistricting schemes to slip through the cracks.83 For example, a recent district court decision discussed the dueling goals of protecting a white Democratic incumbent and creating a Latino-majority district in which Latinos could elect the candidate of their choice.84 While the Latino-majority district would likely elect a Democrat, there were still competing partisan and racial interests.85 Furthermore, in a state like Texas that has more than one sizable minority group, the creation of majority-minority districts is further complicated — even when a particular district is likely to elect a Democrat in a general election, a district’s boundaries may dictate whether a black, Latino, or Asian candidate would be viable in a primary race.86 Primaries therefore provide one example of elections in which ignoring race would leave no remedy.
The judicial branch has yet to develop a coherent approach to partisan gerrymandering. In 2004, a plurality of Justices in Vieth v. Jubelirer87 held that partisan gerrymandering claims are nonjusticiable because “no judicially discernible and manageable standards for adjudicating political gerrymandering claims have emerged.”88 Concurring in the judgment only, Justice Kennedy challenged lower courts to keep searching for a standard that would meet the plurality’s bar.89 Even as plausible standards emerge,90 skeptics who do not want courts to interfere in the political realm remain.91 But Harris reveals that courts are already caught up in “political warfare,”92 and as the Court readies itself to hear arguments on a partisan gerrymandering case during the 2017 Term,93 it should establish a clear standard to govern the battlefield rather than allow itself to be used as an unpredictable “weapon[].”94