Full opinion text
Reversed and remanded by published opinion. Judge MOTZ wrote the opinion for the court, in which Judge WYNN and Judge FLOYD joined except as to Part V.B. Judge WYNN wrote the opinion for the court as to Part V.B., in which Judge FLOYD joined. Judge MOTZ wrote a separate dissenting opinion as to Part V.B. DIANA GRIBBON MOTZ, Circuit Judge, writing for the court except as to Part V.B.: These consolidated cases challenge provisions of a recently enacted North Carolina election law. The district court rejected contentions that the challenged provisions violate the Voting Rights Act and the Fourteenth, Fifteenth, and Twenty-Sixth Amendments of the Constitution. In evaluating the massive record in this case, the court issued extensive factual findings. We appreciate and commend the court on its thoroughness. The record evidence provides substantial support for many of its findings; indeed, many rest on uncontested facts. But, for some of its findings, we must conclude that the district court fundamentally erred. In holding that the legislature did not enact the challenged provisions with discriminatory intent, the court seems to have missed the forest in carefully surveying the many trees. This failure of perspective led the court to ignore critical facts bearing on legislative intent, including the inextricable link between race and politics • in North Carolina. Voting in many areas of North Carolina is racially polarized. That is, “the race of voters correlates with the selection of a certain candidate or candidates.” Thornburg v. Gingles, 478 U.S. 30, 62, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986) (discussing North Carolina). In Gingles and other cases brought under the Voting Rights Act, the Supreme Court has explained that polarization renders minority voters uniquely vulnerable to the inevitable tendency of elected officials to entrench themselves by targeting groups unlikely to vote for them. In North Carolina, restriction of voting mechanisms and procedures that most heavily affect African Americans will predictably redound to the benefit of one political party and to the disadvantage of the other. As the evidence in the record makes clear, that is what happened here. After years of preclearance and expansion of voting access, by 2013 African American registration and turnout rates had finally reached near-parity with white registration and turnout rates. African Americans were poised to act as a major electoral force. But, on the day after the Supreme Court issued Shelby County v. Holder, — U.S. -, 133 S.Ct. 2612, 186 L.Ed.2d 651 (2013), eliminating preclearance obligations, a leader of the party that newly dominated the legislature (and the party that rarely enjoyed African American support) announced an intention to enact what he characterized as an “omnibus” election law. Before enacting that law, the legislature requested data on the use, by race, of a number of voting practices. Upon receipt of the race data, the General Assembly enacted legislation that restricted voting and registration in five different ways, all of which disproportionately affected African Americans. In response to claims that intentional racial discrimination animated its action, the State offered only meager justifications. Although the new provisions target African Americans with almost surgical precision, they constitute inapt remedies for the problems assertedly justifying them and, in fact, impose cures for problems that did not exist. Thus the asserted justifications cannot and do not conceal the State’s true motivation. “In essence,” as in League of United Latin American Citizens v. Perry (LULAC), 548 U.S. 399, 440, 126 S.Ct. 2594, 165 L.Ed.2d 609 (2006), “the State took away [minority voters’] opportunity because [they] were about to exercise it.” As in LULAC, “[t]his bears the mark of intentional discrimination.” Id. Faced with this record, we can only conclude that the North Carolina General Assembly enacted the challenged provisions of the law with discriminatory intent. Accordingly, we reverse the judgment of the district court to the contrary and remand with instructions to enjoin the challenged provisions of the law. I. “The Voting Rights Act of 1965 employed extraordinary measures to address an extraordinary problem.” Shelby Cty., 133 S.Ct. at 2618. Although the Fourteenth and Fifteenth Amendments to the United States Constitution prohibit racial discrimination in the regulation of elections, state legislatures have too often found facially race-neutral ways to deny African Americans access to the franchise. See id. at 2619; Johnson v. De Grandy, 512 U.S. 997, 1018, 114 S.Ct. 2647, 129 L.Ed.2d 775 ' (1994) (noting “the demonstrated ingenuity of state and local governments in hobbling minority voting power” as “jurisdictions have substantially moved from direct, over[t] impediments to the right to vote to more sophisticated devices” (alteration in original) (internal quotation marks omitted)). To remedy this problem, Congress enacted the Voting Rights Act. In its current form, § 2 of the Act provides: No voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color.... 52 U.S.C. § 10301(a) (2012) (formerly 42 U.S.C. § 1973(a)). In addition to this general statutory prohibition on racial discrimination, Congress identified particular jurisdictions “covered” by § 5 of the Voting Rights Act. Shelby Cty., 133 S.Ct. at 2619. Covered jurisdictions were those that, as of 1972, had maintained suspect prerequisites to voting, like literacy tests, and had less than 50% voter registration or turnout. Id. at 2619-20. Forty North Carolina jurisdictions were covered under the Act. 28 C.F.R. pt. 51 app. (2016). As a result, whenever the North Carolina legislature sought to change the procedures or qualifications for voting statewide or in those jurisdictions, it first had to seek “preclearance” with the United States Department of Justice. In doing so, the State had to demonstrate that a change had neither the purpose nor effect of “diminishing the ability of any citizens” to vote “on account of race or color.” 52 U.S.C. § 10304 (2012) (formerly 42 U.S.C. § 1973c). During the period in which North Carolina jurisdictions were covered by § 5, African American electoral participation dramatically improved. In particular, between 2000 and 2012, when the law provided for the voting mechanisms at issue here and did not require photo ID, African American voter registration swelled by 51.1%. J.A. 804 (compared to an increasé of 15.8% for white voters). African American turnout similarly surged, from 41.9% in 2000 to 71.5% in 2008 and 68.5% in 2012. J.A. 1196-97. Not coincidentally, during this period North Carolina emerged as a swing state in national elections. Then, in late June 2013, the Supreme Court issued its opinion in Shelby County. In it, the Court invalidated the preclearance coverage formula, finding it based on outdated data. Shelby Cty., 133 S.Ct. at 2631. Consequently, as of that date, North Carolina no longer needed to preclear changes in its election laws. As the district court found, the day after the Supreme Court issued Shelby County, the “Republican Chairman of the [Senate] Rules Committee[ ] publicly stated, T think we’ll have an omnibus bill coming out’ and ... that the Senate would move ahead with the ‘full bill.’ ” N.C. State Conf. of the NAACP v. McCrory, - F.Supp.3d -, -, 2016 WL 1650774, at *9 (M.D.N.C. Apr. 25, 2016). The legislature then swiftly expanded an essentially single-issue bill into omnibus legislation, enacting it as Session Law (“SL”) 2013-381. In this one statute, the North Carolina legislature imposed a number of voting restrictions. The law required in-person voters to show certain photo IDs, beginning in 2016, which African Americans disproportionately lacked, and eliminated or reduced registration and voting access tools that African Americans disproportionately used. Id, at ---, -, -, -, 2016 WL 1650774, at *9-10, *37, *123, *127, *131. Moreover, as the district court found, prior to enactment of SL 2013-381, the legislature requested and received racial data as to usage of the practices changed by the proposed law. Id. at ---, 2016 WL 1650774, at *136-38. This data showed that African Americans disproportionately lacked the most common kind of photo ID, those issued by the Department of Motor Vehicles (DMV). Id. The pre-Shelby County version of SL 2013-381 provided that all government-issued IDs, even many that had been expired, would satisfy the requirement as an alternative to DMV-issued photo IDs. J.A. 2114-15. After Shelby County, with race data in hand, the legislature amended the bill to exclude many of the alternative photo IDs used by African Americans. Id. at -, 2016 WL 1650774, at *142; J.A. 2291-92. As amended, the bill retained only the kinds of IDs that white North Carolinians were more likely to possess. Id.; J.A. 3653, 2115, 2292. The district court found that, prior to enactment of SL 2013-381, legislators also requested data as to the racial breakdown of early voting usage. Id. at ---, 2016 WL 1650774, at *136-37. Early voting allows any registered voter to complete an absentee application and ballot at the same time, in person, in advance of Election Day. Id, at ---, 2016 WL 1650774, at *4-5. Early voting thus increases opportunities to vote for those who have difficulty getting to their polling place on Election Day. The racial data provided to the legislators revealed that African Americans disproportionately used early voting in both 2008 and 2012. Id. at ---, 2016 WL 1650774, at *136-38; see also id. at - n. 74, 2016 WL 1650774, at *48 n. 74 (trial evidence showing that 60.36% and 64.01% of African Americans voted early in 2008 and 2012, respectively, compared to 44.47% and 49.39% of whites). In particular, African Americans disproportionately used the first seven days of early voting. Id, After receipt of this racial data, the General Assembly amended the bill to eliminate the first week of early voting, shortening the total early voting period from seventeen to ten days. Id. at -, -, 2016 WL 1650774, at *15, *136. As a result, SL 2013-381 also eliminated one of two “souls-to-the-polls” Sundays in which African American churches provided transportation to voters. Id. at -, 2016 WL 1650774, at *55. The district court found that legislators similarly requested data as to the racial makeup of same-day registrants. Id. at -, 2016 WL 1650774, at *137. Prior to SL 2013-381, same-day registration allowed eligible North Carolinians to register in person at an early voting site at the same time as casting their ballots. Id. at -, 2016 WL 1650774, at *6. Same-day registration provided opportunities for those as yet unable to register, as well as those who had ended up in the “incomplete registration queue” after previously attempting to register. Id. at -, 2016 WL 1650774, at *65. Same-day registration also provided an easy avenue to re-register for those who moved frequently, and allowed those with low literacy skills or other difficulty completing a registration form to receive personal assistance from poll workers. See id. The legislature’s racial data demonstrated that, as the district court found, “it is indisputable that African American voters disproportionately used [same-day registration] when it was available.” Id. at -, 2016 WL 1650774, at *61. The district court further found that African American registration applications constituted a disproportionate percentage of the incomplete registration queue. Id. at -, 2016 WL 1650774, at *65. And the court found that African Americans “are more likely to move between counties,” and thus “are more likely to need to re-register.” Id. As evidenced by the types of errors that placed many African American applications in the incomplete queue, id. at -, - & n. 26, 2016 WL 1650774, at *65, *123 & n. 26, in-person assistance likely would disproportionately benefit African Americans. SL 2013-381 eliminated same-day registration. Id. at -, 2016 WL 1650774, at *15. Legislators additionally requested a racial breakdown of provisional voting, including out-of-precinct voting. Id. at ---, 2016 WL 1650774, at *136-37. Out-of-precinct voting required the Board of Elections in each county to count the provisional ballot of an Election Day voter who appeared at the wrong precinct, but in the correct county, for all of the ballot items for which the voter was eligible to vote. Id. at ---, 2016 WL 1650774, at *5-6. This provision assisted those who moved frequently, or who mistook a voting site as being in their correct precinct. The district court found that the racial data revealed that African Americans disproportionately voted provisionally. Id. at -, 2016 WL 1650774, at *137. In fact, the General Assembly that had originally enacted the out-of-precinct voting legislation had specifically found that “of those registered voters who happened to vote provisional ballots outside their resident precincts” in 2004, “a disproportionately high percentage were African American.” Id. at -, 2016 WL 1650774, at *138. With SL 2013-381, the General Assembly altogether eliminated out-of-precinct voting. Id. at -, 2016 WL 1650774, at *15. African Americans also disproportionately used preregistration. Id. at -, 2016 WL 1650774, at *69. Preregistration permitted 16- and 17-year-olds, when obtaining driver’s licenses or attending mandatory high school registration drives, to identify themselves and indicate their intent to vote. Id. at -, -, 2016 WL 1650774, at *7, *68. This allowed County Boards of Elections to verify eligibility and automatically register eligible citizens once they reached eighteen. Id. at -, 2016 WL 1650774, at *7. Although preregistration increased turnout among young adult voters, SL 2013-381 eliminated it. Id. at -,-, 2016 WL 1650774, at*15, *69. The district court found that not only did SL 2013-381 eliminate or restrict these voting mechanisms used disproportionately by Afincan Americans, and require IDs that African Americans disproportionately lacked, but also that African Americans were more likely to “experience socioeconomic factors that may hinder their political participation.” Id. at -, 2016 WL 1650774, at *89. This is so, the district court explained, because in North Carolina, African Americans are “disproportionately likely to move, be poor, less educated, have less access to transportation, and experience poor health.” Id. at -, 2016 WL 1650774, at *89. Nevertheless, over protest by many legislators and members of the public, the General Assembly quickly ratified SL 2013-381 by strict party-line votes. Id. at ---, 2016 WL 1650774, at *9-13. The Governor, who was of the same political party as the party that controlled the General Assembly, promptly signed the bill into law on August 12, 2013. Id. at -, 2016 WL 1650774, at *13. That same day, the League of Women Voters, along with numerous other organizations and individuals, filed suit. Id. at -, 2016 WL 1650774, at *16. These Plaintiffs alleged that the restrictions on early voting and elimination of same-day registration and out-of-precinct voting were motivated by discriminatory intent in violation of § 2 of the Voting Rights Act and the Fourteenth and Fifteenth Amendments; that these provisions had a discriminatory result in violation of § 2 of the Voting Rights Act; and that these provisions burdened the right to vote generally, in contravention of the Fourteenth Amendment. See id. Also that same day, the North Carolina State Conference of the NAACP, in conjunction with several other organizations and individuals, filed a separate action. Id. They alleged that the photo ID requirement and the provisions challenged by the League of Women Voters produced dis: criminatory results under § 2 and demonstrated intentional discrimination in violation of the Fourteenth and Fifteenth Amendments. Id. Soon thereafter, the United States also filed suit, challenging the same provisions as discriminatory in both purpose and result in violation of § 2 of the Voting Rights Act. Id. Finally, a group of “young voters” intervened, alleging that these same provisions violated their rights under the Fourteenth and Twenty-Sixth Amendments. Id. The district court consolidated the cases. Id. Ahead of the 2014 midterm general election, Plaintiffs moved for a preliminary injunction of several provisions of the law. See N.C. State Conf. of the NAACP v. McCrory, 997 F.Supp.2d 322, 339 (M.D.N.C. 2014). The district court denied the motion. Id. at 383. On appeal, we reversed in part, remanding the ease with instructions to issue an order staying the elimination of same-day registration and out-of-precinct voting. League of Women Voters of N.C. v. North Carolina (LWV), 769 F.3d 224, 248-49 (4th Cir. 2014). Over the dissent of two Justices, the Supreme Court stayed our injunction mandate on October 8, 2014, pending its decision on certiorari. See North Carolina v. League of Women Voters of N.C., - U.S. -, 135 S.Ct. 6, 190 L.Ed.2d 243 (2014) (mem.). On April 6, 2015, the Supreme Court denied certiorari. See. North Carolina v. League of Women Voters of N.C., - U.S. -, 135 S.Ct. 1735, 191 L.Ed.2d 702 (2015) (mem.). This denial automatically reinstituted the preliminary injunction, restoring same-day registration and out-of-precinct voting pending the outcome of trial in this case. North Carolina v. League of Women Voters of N.C., 135 S.Ct. at 6. That consolidated trial was scheduled to begin on July 13, 2015. N.C. State Conf., - F.Supp.3d at -, 2016 WL 1650774, at *18. However, on June 18, 2015, the General Assembly ratified House Bill 836, enacted as Session Law (“SL”) 2015-103. Id. at -, -, 2016 WL 1650774, at *13, *18. -This new law amended the photo ID requirement by permitting a voter without acceptable ID to cast a provisional ballot if he completed a declaration stating that he had a reasonable impediment to acquiring acceptable photo ID (“the reasonable impediment exception”). Id. at -, 2016 WL 1650774, at *13. Given this enactment, the district court bifurcated trial of the case. Id. at -, 2016 WL 1650774, at *18. Beginning in July 2015, the court conducted a trial on the challenges to all of the provisions except the photo ID requirement. Id. In January 2016, the court conducted a separate trial on the photo ID requirement, as modified by the reasonable impediment exception. Id. On April 25, 2016, the district court entered judgment against the Plaintiffs on all of their claims as to all of the challenged provisions. Id. at -, 2016 WL 1650774, at *171. The court found no discriminatory results under § 2, no discriminatory intent under § 2 or the Fourteenth and Fifteenth Amendments, no undue burden on the right to vote generally under the Fourteenth Amendment, and no violation of the Twenty-Sixth Amendment. See id. at -, -, -, -, 2016 WL 1650774, at *133-34, *148, *164, *167. At the same time, acknowledging the imminent June primary election, the court temporarily extended the preliminary injunction of same-day registration and out-of-precinct voting through that election. Id. at -, 2016 WL 1650774, at *167. The photo ID requirement went into effect as scheduled for the first time in the March 2016 primary election, and was again in effect during the June primary election. Id. at -,-, 2016 WL 1650774, at *19, *171. Plaintiffs timely noted this appeal. J.A. 24967, 24970, 24976, 24980. They also requested that we stay the district court’s mandate and extend the preliminary injunction, which we did pending our decision in this case. Order Extending the Existing Stay, No. 16-1468 (Dkt. No. 122). On appeal, Plaintiffs reiterate their attacks on the photo ID requirement, the reduction in days of early voting, and the elimination of same-day registration, out-of-precinct voting, and preregistration, alleging discrimination against African Americans and Hispanics. Because the record evidence is limited regarding Hispan-ies, we confine our analysis to African Americans. We hold that the challenged provisions of SL 2013-381 were enacted with racially discriminatory intent in violation of the Equal Protection Clause of the Fourteenth Amendment and § 2 of the Voting Rights Act. We need not and do not reach Plaintiffs’ remaining claims. II. a: An appellate court can reverse a district court’s factual findings only if clearly erroneous. United States v. U.S. Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 92 L.Ed. 746 (1948). This standard applies to the ultimate factual question of a legislature’s discriminatory motivation. See Pullman-Standard v. Swint, 456 U.S. 273, 287-88, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982); Hunt v. Cromartie (Cromartie I), 526 U.S. 541, 549, 119 S.Ct. 1545, 143 L.Ed.2d 731 (1999). Such a finding is clearly erroneous if review of the entire record leaves the appellate court “with the definite and firm conviction that the [district [c]ourt’s key findings are mistaken.” Easley v. Cromartie (Cromartie II), 532 U.S. 234, 243, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (citation and internal quotation marks omitted). This is especially so when “the key evidence consisted primarily of documents and expert testimony” and “[credibility evaluations played a minor role.” Id. Moreover, if “the record permits only one resolution of the factual issue” of discriminatory purpose, then an appellate court need not remand the case to the district court. Pullman-Standard, at 292, 102 S.Ct. 1781; see Cromartie II, 532 U.S. at 257, 121 S.Ct. 1452 (reversing, without remanding, three-judge court’s factual finding that racial intent predominated in creation of challenged redistricting plan); Hunter v. Underwood, 471 U.S. 222, 229, 105 S.Ct. 1916, 85 L.Ed.2d 222 (1985) (affirming Court of Appeals’ reversal without remand where district court’s finding of no discriminatory purpose was clearly erroneous); Dayton Bd. of Educ. v. Brinkman, 443 U.S. 526, 534, 542, 99 S.Ct. 2971, 61 L.Ed.2d 720 (1979) (affirming Court of Appeals’ reversal of finding of no intentional discrimination with remand only to enter remedy order). In Village of Arlington Heights v. Metropolitan Housing Development Corp., 429 U.S. 252, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977), the Supreme Court addressed a claim that racially discriminatory intent motivated a facially neutral governmental action. The Court recognized that a facially neutral law, like the one at issue here, can be motivated by invidious racial discrimination. Id. at 264-66, 97 S.Ct. 555. If dis-criminatorily motivated, such laws are just as abhorrent, and just as unconstitutional, as laws that expressly discriminate on the basis of race. Id.; Washington v. Davis, 426 U.S. 229, 241, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976). When considering whether discriminatory intent motivates a facially neutral law, a court must undertake a “sensitive inquiry into such circumstantial and direct evidence of intent as may be available.” Arlington Heights, 429 U.S. at 266, 97 S.Ct. 555. Challengers need not show that discriminatory purpose was the “sole[ ]” or even a “primary” motive for the legislation, just that it was “a motivating factor.” Id. at 265-66, 97 S.Ct. 555 (emphasis added). Discriminatory purpose “may often be inferred from the totality of the relevant facts, including the fact, if it is true, that the law bears more heavily on one race than another.” Davis, 426 U.S. at 242, 96 S.Ct. 2040. But the ultimate question remains: did the legislature enact a law “because of,” and not “in spite of,” its discriminatory effect. Pers. Adm’r of Mass. v. Feeney, 442 U.S. 256, 279, 99 S.Ct. 2282, 60 L.Ed.2d 870 (1979). In Arlington Heights, the Court set forth a nonexhaustive list of factors to consider in making this sensitive inquiry. These include: “[t]he historical background of the [challenged] decision”; “[t]he specific sequence of events leading up to the challenged decision”; “[departures from normal procedural sequence”; the legislative history of the decision; and of course, the disproportionate “impact of the official action-whether it bears more bears more heavily on one race than another.” Arlington Heights, 429 U.S. at 266-67, 97 S.Ct. 555 (internal quotation marks omitted). • In instructing courts to consider the broader context surrounding the passage of legislation, the Court has recognized that “[o]utright admissions of impermissible racial motivation are infrequent and plaintiffs often must rely upon other evidence.” Cromartie I, 526 U.S. at 553, 119 S.Ct. 1545. In a vote denial case such as the one here, where the plaintiffs allege that the legislature imposed barriers to minority voting, this holistic approach is particularly important, for “[discrimination today is more subtle than the visible methods used in 1965.” H.R. Rep. No. 109-478, at 6 (2006), as reprinted in 2006 U.S.C.C.A.N. 618, 620. Even “second-generation barriers” to voting, while facially race neutral, may nonetheless be motivated by impermissible racial discrimination. Shelby Cty., 133 S.Ct. at 2635 (Ginsburg, J., dissenting) (cataloguing ways in which' •facially neutral voting laws continued to discriminate against minorities even after passage of Voting Rights Act). “Once racial discrimination is shown to have been a ‘substantial’ or ‘motivating’ factor behind enactment of the law,the burden shifts to the law’s defenders to demonstrate that the law would have been enacted without this factor.” Hunter, 471 U.S. at 228, 105 S.Ct. 1916. When determining if this burden has been met, courts must be mindful that “racial discrimination is not just another competing consideration.” Arlington Heights, 429 U.S. at 265-66, 97 S.Ct. 555. For this reason, the judicial deference accorded to legislators when “balancing numerous competing considerations” is “no longer justified.” Id. Instead, courts must scrutinize the legislature’s actual non-racial motivations to determine whether they alone can justify the legislature’s choices. See Mt. Healthy City Sch. Dist. Bd. of Educ. v. Doyle, 429 U.S. 274, 287, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977); cf. Miss. Univ. for Women v. Hogan, 458 U.S. 718, 728, 102 S.Ct. 3331, 73 L.Ed.2d 1090 (1982) (describing “inquiry into the actual purposes underlying a statutory scheme” that classified based on gender (emphasis added) (internal quotation marks omitted)). If a court finds that a statute is unconstitutional, it can enjoin the law. See, e.g., Hunter, 471 U.S. at 231, 105 S.Ct. 1916; Anderson v. Martin, 375 U.S. 399, 404, 84 S.Ct. 454, 11 L.Ed.2d 430 (1964). B. In the context of a § 2 discriminatory intent analysis, one of the critical background facts of which a court must take notice is whether voting is racially polarized. Indeed, to prevail in a case alleging discriminatory dilution of minority voting strength under § 2, a plaintiff must prove this fact as a threshold showing. See Gingles, 478 U.S. at 51, 56, 62, 106 S.Ct. 2752. Racial polarization “refers to the situation where different races ... vote in blocs for different candidates.” Id. at 62, 106 S.Ct. 2752. This legal concept “incorporates neither causation nor intent” regarding voter preferences, for “[i]t is the difference between the choices made by blacks and whites-not the reasons for that difference-that results” in the opportunity for discriminatory laws to have their in- tended political effect. Id. at 62-63, 106 S.Ct. 2752. While the Supreme Courthas expressed hope that “racially polarized voting is wan- ing,” it has at the same time recognized that “racial discrimination and racially polarized voting are not ancient history.” Bartlett v. Strickland, 556 U.S. 1, 25, 129 S.Ct. 1231, 173 L.Ed.2d 173 (2009). In fact, recent scholarship suggests that, in the years following President Obama’s election following President Obama’s election in 2008, areas of the country formerly subject to § 5 preclearance have seen an increase in racially polarized voting. See Stephen Ansolabehere, Nathaniel Persily & Charles Stewart III, Regional Differences in Racial Polarization in the 2012 Presidential Election: Implications for the Constitutionality of Section 5 of the Voting Rights Act, 126 Harv. L. Rev. F. 205, 206 (2013). Further, “[t]his gap is not the result of mere partisanship, for even when controlling for partisan identification, race is a statistically significant predictor of vote choice, especially in the covered jurisdictions.” Id. Racially polarized voting is not, in and of itself, evidence of racial discrimination. But it does provide an incentive for intentional discrimination in the regulation of elections. In reauthorizing the Voting Rights Act in 2006, Congress recognized that “[t]he potential for discrimination in environments characterized by racially polarized voting is great.” H.R. Rep. No. 109-478, at 35. This discrimination can take many forms. One common way it has surfaced is in challenges centered on vote dilution, where “manipulation of district lines can dilute the voting strength of politically cohesive minority group members.” De Grandy, 512 U.S. at 1007, 114 S.Ct. 2647 (emphasis added); see also Voinovich v. Quilter, 507 U.S. 146, 153-54, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993). It is the political cohesiveness of the minority groups that provides the political payoff for legislators who seek to dilute or limit the minority vote. The Supreme Court squarely confronted this connection in LULAC. There, the record evidence revealed racially polarized voting, such that 92% of Latinos voted against an incumbent of a particular party, whereas 88% of non-Latinos voted for him. 548 U.S. at 427, 126 S.Ct. 2594. The Court explained how this racial polarization provided the impetus for the discriminatory vote dilution legislation at issue in that case: “In old District 23 the increase in Latino voter registration and overall population, the concomitant rise in Latino voting power in each successive election, the near-victory of the Latino candidate of choice in 2002, and the resulting threat to the” incumbent representative motivated the controlling party to dilute the minority vote. Id. at 428, 126 S.Ct. 2594 (citation omitted). Although the Court grounded its holding on the § 2 results test, which does not require proof of intentional discrimination, the Court noted that the challenged legislation bore “the mark of intentional discrimination.” Id. at 440, 126 S.Ct. 2594. The LULAC Court addressed a claim of vote dilution, but its recognition that racially polarized voting may motivate politicians to entrench themselves through discriminatory election laws applies with equal force in the vote denial context. Indeed, it applies perhaps even more powerfully in cases like that at hand, where the State has restricted access to the franchise. This is so because, unlike in redistricting, where states may consider race and partisanship to a certain extent, see, e.g., Miller v. Johnson, 515 U.S. 900, 920, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995), legislatures cannot restrict voting access on the basis of race. (Nor, we note, can legislatures restrict access to the franchise based on the desire to benefit a certain political party. See Anderson v. Celebrezze, 460 U.S. 780, 792-93, 103 S.Ct. 1564, 75 L.Ed.2d 547 (1983).) Using race as a proxy for party may be an effective way to win an election. But intentionally targeting a particular race’s access to the franchise because its members vote for a particular party, in a predictable manner, constitutes discriminatory purpose. This is so even absent any evidence of race-based hatred and despite the obvious political dynamics. A state legislature acting on such a motivation engages in intentional racial discrimination in violation of the Fourteenth Amendment and the Voting Rights Act. III. With these principles in mind, we turn to their application in the case at hand. A. Arlington Heights directs us to consider “[t]he historical background of the decision” challenged as racially discriminatory. 429 U.S. at 267, 97 S.Ct. 555. Examination of North Carolina’s history of race discrimination and recent patterns of official discrimination, combined with the racial polarization of politics in the state, seems particularly relevant in this inquiry. The district court erred in ignoring or minimizing these facts. Unquestionably, North Carolina has a long history of race discrimination generally and race-based vote suppression in particular. Although we recognize its limited weight, see Shelby Cty., 133 S.Ct. at 2628-29, North Carolina’s pre-1965 history of pernicious discrimination informs our inquiry. For “[i]t was in the South that slavery was upheld by law until uprooted by the Civil War, that the reign of Jim Crow denied African-Americans the most basic freedoms, and that state and local governments worked tirelessly 'to disenfranchise citizens on the basis of race.” Id. at 2628. While it is of course true that “history did not end in 1965,” id. it is equally true that SL 2013-381 imposes the first meaningful restrictions on voting access since that date-and a comprehensive set of re- strictions at that. Due to this fact, and because the legislation came into being literally within days of North Carolina’s release from the preclearance require- ments of the Voting Rights Act, that long- ago history bears more heavily here than it might otherwise. Failure to so recognize would risk allowing that troubled history to “piek[ ] up where it left off in 1965” to the detriment of African American voters in North Carolina. LWV, 769 F.3d at 242. In consideringPlaintiffs’ discriminatory results claim under § 2, the district court expressly and properly recognized the State’s “shameful” history of “past dis- crimination.” N.C. State Conf., - F.Supp.3d at -, 2016 WL 1650774, at *83-86. But the court inexpli- cably failed to grapple with that history in its analysis of Plaintiffs’ discriminatory in- tent claim. Rather, when assessing the in- tent claim, the court’s analysis on the point consisted solely of the finding that “there is little evidence of official discrimination since the 1980s,” accompanied by a foot- note dismissing examples of more recent official discrimination. See id. at -, 2016 WL 1650774, at *143. That findingis clearly erroneous. The record is replete with evidence of instances since the 1980s in which the North Carolina legislature has attempted to suppress and dilute the voting rights of African Americans. In some of these instances, the Department of Justice or federal courts have determined that the North Carolina General Assembly acted with discriminato- ry intent, “reveal[ing] a series of official actions taken for invidious purposes.” Ar- lington Heights, 429 U.S. at 267, 97 S.Ct. 555. In others, the Department of Justice or courts have found that the General As- sembly’s action produced discriminatory results. The latter evidence, of course, proves less about discriminatory intent than the former, but it is informative. A historical pattern of laws producing dis- criminatory results provides important context 'for determining whether the same the same decisionmaking body has also enacted a law with discriminatory purpose. See, e.g., Veasey v. Abbott, No. 14-41127, 830 F.3d 216, 2016 WL 3923868 (5th Cir. July 20, 2016) (en banc) (considering as relevant, in intentional discrimination analysis of voter ID law, DOJ letters and previous court cases about results and intent). The record reveals that, within the time period that the district court found free of “official discrimination” (1980 to 2013), the Department of Justice issued over fifty objection letters to proposed election law changes in North Carolina — including several since 2000 — because the State had failed to prove the proposed changes would have no discriminatory purpose or effect. See U.S. Dep’t of Justice, Civil Rights Div., Voting Determination Letters for North Carolina (DOJ Letters) (Aug. 7, 2015), https://www.justice.gov/crVvoting-determination-letters-north-carolina; see also Regents of the Univ. of California v. Bakke, 438 U.S. 265, 305, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978) (referring to objections of the Department of Justice under § 5 as “administrative finding[s] of discrimination”). Twenty-seven of those letters objected to laws that either originated in the General Assembly or originated with local officials and were approved by the General Assembly. See DOJ Letters. During the same period, private plaintiffs brought fifty-five successful cases under § 2 of the Voting Rights Act. J.A. 1260; Anita S. Earls et al., Voting Rights in North Carolina: 1982-2006, 17 S. Cal. Rev. L. & Soc. Just. 577 (2008). Ten cases ended in judicial decisions finding that electoral schemes in counties and municipalities across the state had the effect of discriminating against minority voters. See, e.g., Ward v. Columbus Cty., 782 F.Supp. 1097 (E.D.N.C. 1991); Johnson v. Halifax Cty., 594 F.Supp. 161 (E.D.N.C. 1984) (granting preliminary injunction). Forty-five cases were settled favorably for plaintiffs out of court or through consent degrees that altered the challenged voting laws. See, e.g., Daniels v. Martin Cty. Bd. of Comm’rs., No. 4:89-cv-00137 (E.D.N.C. 1992); Hall v. Kennedy, No. 3:88-cv-00117 (E.D.N.C. 1989); Montgomery Cty. Branch of the NAACP v. Montgomery Cty. Bd. of Elections, No. 3:90-cv-00027 (M.D.N.C. 1990). On several occasions, the United States intervened in cases or filed suit independently. See, e.g., United States v. Anson Bd. of Educ., No. 3:93-cv-00210 (W.D.N.C. 1994); United States v. Granville Cty. Bd. of Educ., No. 5:87-cv-00353 (E.D.N.C. 1989); United States v. Lenoir Cty., No. 87-105-cv-84 (E.D.N.C. 1987). And, of course, the case in which the Supreme Court announced the standard governing § 2 results claims — Thornburg v. Gingles — was brought by a class of African American citizens in North Carolina challenging a statewide redistricting plan. 478 U.S. at 35, 106 S.Ct. 2752. There the Supreme Court affirmed findings by the district court that each challenged district exhibited “racially polarized voting,” and held that “the legacy of official discrimination in voting matters, education, housing, employment, and health services ... acted in concert with the multimember district-ing scheme to impair the ability” of African American voters to “participate equally in the political process.” Id. at 80, 106 S.Ct. 2752. And only a few months ago (just weeks before the district court issued its opinion in the case at hand), a three-judge court addressed a redistricting plan adopted by the same General Assembly that enacted SL 2013-381. Harris v. McCrory, No. 1:13-CV-949, 159 F.Supp.3d 600, 603-04, 2016 WL 482052, at *1-2 (M.D.N.C. Feb. 5, 2016), prob. juris. noted, — U.S.-, 136 S.Ct. 2512, — L.Ed.2d -, No. 15-1262, 2016 WL 1435913 (June 27, 2016). The court held that race was the predominant motive in drawing two congressional districts, in violation of the Equal Protection Clause. Id at 603-04, 621 & n. 9, 2016 WL 482052, at *1-2, *17 & n. 9. Contrary to the district court’s suggestion, see N.C. State Conf., — F.Supp.3d at - n. 223, 2016 WL 1650774, at *143 n. 223, a holding that a legislature impermissibly relied on race certainly provides relevant evidence as to whether race motivated other election legislation passed by the same legislature. The district court failed to take into account these cases and their important takeaway: that state officials continued in their efforts to restrict or dilute African American voting strength well after 1980 and up to the present day. Only the robust protections of § 5 and suits by private plaintiffs under § 2 of the Voting Rights Act prevented those efforts from succeeding. These cases also highlight the manner in which race and party are inexorably linked in North Carolina. This fact constitutes a critical — perhaps the most critical — piece of historical evidence here. The district court failed to recognize this linkage, leading it to accept “politics as usual” as a justification for many of the changes in SL 2013-381. But that cannot be accepted where politics as usual translates into race-based discrimination. As it did with the history of racial discrimination, the district court again recognized this reality when analyzing whether SL 2013-381 had a discriminatory result, but not when analyzing whether it was motivated by discriminatory intent. In its results analysis, the court noted that racially polarized voting between African Americans and whites remains prevalent in North Carolina. N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *86-87. Indeed, at trial the State admitted as much. Id. at-, 2016 WL 1650774, at *86. As one of the State’s experts conceded, “in North Carolina, African-American race is a better predictor for voting Democratic • than party registration.” J.A. 21400. For example, in North Carolina, 85% of African American voters voted for John Kerry in 2004, and 95% voted for President Obama in 2008. N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *86. In comparison, in those elections, only 27% of white North Carolinians voted for John Kerry, and only 35% for President Obama. Id. Thus, whether the General Assembly knew the exact numbers, it certainly knew that African American voters were highly likely, and that white voters were unlikely, to vote for Democrats. And it knew that, in recent years, African Americans had begun registering and voting in unprecedented numbers. Indeed, much of the recent success of Democratic candidates in North Carolina resulted from African American voters overcoming historical barriers and making their voices heard to a degree unmatched in modern history. Despite this, the district court took no issue with one of the legislature’s stated purposes in enacting SL 2013-381— to “mov[e] the law back to the way it was.” N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *111. Rather, the court apparently regarded this as entirely appropriate. The court noted repeatedly that the voting mechanisms that SL 2013-381 restricts or eliminates were ratified “relatively recently,” “almost entirely along party lines,” when “Democrats controlled” the legislature; and that SL 2013-381 was similarly ratified “along party lines” after “Republicans gained ... control of both houses.” Id. at -, -, 2016 WL 1650774, at *2-7, *12. Thus, the district court apparently considered SL 2013-381 simply an appropriate means for one party to .counter recent success by another party. We recognize that elections have consequences, but winning an election does not empower anyone in any party to engage in purposeful racial discrimination. When a legislature dominated by one party has dismantled barriers to African American access to the franchise, even if done to gain votes, “politics as usual” does not allow a legislature dominated by the other party to re-erect those barriers. The record evidence is clear that this is exactly what was done here. For example, the State argued before the district court that the General Assembly enacted changes to early voting laws to avoid “political gamesmanship” with respect to the hours and locations of early voting centers. J.A. 22348. As “evidence of justifications” for the changes to early voting, the State offered purported inconsistencies in voting hours across counties, including the fact that only some counties had decided to offer Sunday voting. Id. The State then elaborated on its justification, explaining that “[cjounties with Sunday voting in 2014 were disproportionately black” and “disproportionately Democratic.” J.A. 22348-49. In response, SL 2013-381 did away with one of the two days of Sunday voting. See N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *15. Thus, in what comes as close to a smoking gun as we are likely to see in modern times, the State’s very justification for a challenged statute hinges explicitly on race — specifically its concern that African Americans, who had overwhelmingly voted for Democrats, had too much access to the franchise. These contextual facts, which reveal the powerful undercurrents influencing North Carolina politics, must be considered in determining why the General Assembly enacted SL 2013-381. Indeed, the law’s purpose cannot be properly understood without these considerations. The record makes clear that the historical origin of the challenged provisions in this statute is not the innocuous back-and-forth of routine partisan struggle that the State suggests and that the district court accepted. Rather, the General Assembly enacted them in the immediate aftermath of unprecedented African American voter participation in a state with a troubled racial history and racially polarized voting. The district court clearly erred in ignoring or dismissing this historical background evidence, all of which supports a finding of discriminatory intent. B. Arlington Heights also instructs us to consider the “specific sequence of events leading up to the challenged decision.” 429 U.S. at 267, 97 S.Ct. 555. In doing so, a court must consider “[d]epar-tures from the normal procedural sequence,” which may demonstrate “that improper purposes are playing a role.” Id. The sequential facts found by the district court are undeniably accurate. N.C. State Conf.- F.Supp.3d at -, 2016 WL 1650774, at *8-13. Indeed, they are undisputed. Id. And they are devastating. The record shows that, immediately after Shelby County, the General Assembly vastly expanded an earlier photo ID bill and rushed through the legislative process the most restrictive voting legislation seen in North Carolina since enactment of the Voting Rights Act of 1965. Id. The district court erred in refusing to draw the obvious inference that this sequence of events signals discriminatory intent.. The district court found that prior to Shelby County, SL 2013-381 numbered only sixteen pages and contained none of the challenged provisions, with the exception of a much less restrictive photo ID requirement. Id. at -, ---, 2016 WL 1650774, at *8, *143-44. As the court further found, this pre-Shelby County bill was afforded more than three weeks of debate in public hearings and almost three more weeks of debate in the House. Id. at -, 2016 WL 1650774, at *8. For this version of the bill, there was some bipartisan support: “[f]ive House Democrats joined all present Republicans in voting for the voter-ID bill.” Id. The district court found that SL 2013-381 passed its first read in the Senate on April 25, 2013, where it remained in the Senate Rules Committee. Id. At that time, the Supreme Court had heard argument in Shelby County, but had issued no opinion. Id. “So,” as the district court found, “the bill sat.” Id. For the next two months, no public debates were had, no public amendments made, and no action taken on the bill. Then, on June 25, 2013, the Supreme Court issued its opinion in Shelby County. Id. at-, 2016 WL 1650774, at *9. The very next day, the Chairman of the Senate Rules Committee proclaimed that the legislature “would now move ahead with the full bill,” which he recognized would be “omnibus” legislation. Id. at -, 2016 WL 1650774, at *9. After that announcement, no further public debate or action occurred for almost a month. Id. As the district court explained, “[i]t was not until July 23 ... that an expanded bill, including the election changes challenged in this case, was released.” Id. at -, 2016 WL 1650774, at *144. The new bill-now fifty-seven pages in length-targeted four voting and registration mechanisms, which had previously ex- panded access to the franchise, and provided a much more stringent photo ID provision. See 2013 N.C. Sess. Laws 381. Post-Shelby County, the change in accept- ed photo IDs is of particular note: the new ID provision retained only those types of photo ID disproportionately held by whites and excluded those disproportionately held by African Americans. N.C. State Conf., - F.Supp.3d at -,-, 2016 WL 1650774, at *37, *142. The district court specifically found that “the re- moval of public assistance IDs” in particu- lar was “suspect,” because “a reasonable legislator [would be] aware of the socioeco- nomic disparities endured by African Americans [and] could have surmised that African Americans would be more likely to be more likely to possess this form of ID.” Id. at -, 2016 WL 1650774, at *142. Moreover, after the General Assembly finally revealed the expanded SL 2013-381 to the public, the legislature rushed it through the legislative process. The new SL 2013-381 moved through the General Assembly in three days: one day for a public hearing, two days in the Senate, and two hours in the House. Id. at-, 2016 WL 1650774, at *9-12. The House Democrats who supported the pre-Shelby County bill now opposed it. Id. at ---, 2016 WL 1650774, at *12. The House voted on concurrence in the Senate’s version, rather than sending the bill to a committee. Id. at -, 2016 WL 1650774, at *12. This meant that the House had no opportunity to offer its own amendments before the up-or-down vote on the legislation; that vote proceeded on strict party lines. Id.; see J.A. 1299; N.C. H.R. Rules 43.2, 43.3, 44. The Governor, of the same party as the proponents of the bill, then signed the bill into law. N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *13. This hurried pace, of course, strongly suggests an attempt to avoid in-depth scrutiny. See, e.g., Veasey, 830 F.3d at 237, 2016 WL 3923868, at *12 (noting as suspicious voter ID law’s “three-day passage through the Senate”). Indeed, neither this legislature — nor, as far as we can tell, any other legislature in the Country — has ever done so much, so fast, to restrict access to the franchise. The district court erred in accepting the State’s efforts to cast this suspicious narrative in an innocuous light. To do so, the court focused on certain minor facts instead of acknowledging the whole picture. For example, although the court specifically found the above facts, it dismissed Plaintiffs’ argument that this sequence of events demonstrated unusual legislative speed because the legislature “acted within all [of its] procedural rules.” N.C. State Conf., — F.Supp.3d at -, 2016 WL 1650774, at *145. But, of course, a legislature need not break its own rules to engage in unusual procedures. Even just compared to the process afforded the pre-Shelby County bill, the process for the “full bill” was, to say the very least, abrupt. Similarly, the district court accused Plaintiffs of “ignor[ing] the extensive debate and consideration the initial voter-ID bill received in the spring.” Id. at -, 2016 WL 1650774, at *146. But because the pre-Shelby County bill did not contain any of the provisions challenged here, that debate hardly seems probative. The district court also quoted one senator who opposed the new “full bill” as saying that the legislators had “a good and thorough debate.” Id. at —, -, 2016 WL 1650774, at *12, *145. We note, however, that many more legislators expressed dismay at the rushed process. Id. at -, 2016 WL 1650774, at *145. Indeed, as the court itself noted, “[sjeveral Democratic senators characterized the bill as voter suppression of minorities. Others characterized the bill as partisan.” Id. at -, 2016 WL 1650774, at *12 (citations omitted). Republican senators “strongly denied such claims,” while at the same time linking the bill to partisan goals: that “the bill reversed past practices that Democrats passed to favor themselves.” Id. Finally, the district court dismissed the expanded law’s proximity to the Shelby County decision as above suspicion. The Court found that the General Assembly “would not have been unreasonable” to wait until after Shelby County to consider the “full bill” because it could have concluded that the provisions of the “full bill” were “simply not worth the administrative and financial cost” of preclearance. Id. at -, 2016 WL 1650774, at *144: Although desire to avoid the hassle of the preclearance process could, in another case, justify a decision to await the outcome in Shelby County, that inference is not persuasive in this case. For here, the General Assembly did not simply wait to enact changes to its election laws that might require the administrative hassle of, but likely would pass, preclearance. Rather, after Shelby County it moved forward with what it acknowledged was an omnibus bill that restricted voting mechanisms it knew were used disproportionately by African Americans, id. at -, 2016 WL 1650774, at *148, and so likely would not have passed preclearance. And, after Shelby County, the legislature substantially changed the one provision that it had fully debated before. As noted above, the General Assembly completely revised the list of acceptable photo IDs, removing from the list the IDs held disproportionately by African Americans, but retaining those disproportionately held by whites. Id. at-,-, 2016 WL 1650774, at *37, *142. This fact alone undermines the possibility that the post-Shelby County timing was merely to avoid the administrative costs.' Instead, this sequence of events — the General Assembly’s eagerness to, at the historic moment of Shelby County’s issuance, rush through the legislative process the most restrictive voting law North Carolina has seen since the era of Jim Crow-bespeaks a certain purpose. Although this factor, as with the other Arlington Heights factors, is not dispositive on its own, it provides another compelling piece of the puzzle of the General Assembly’s motivation. C. Arlington Heights also recognizes that the legislative history leading to a challenged provision “may be highly relevant, especially where there are contemporaneous statements by members of the decisionmaking body, minutes of its meetings, or reports.” 429 U.S. at 268, 97 S.Ct. 555. Above, we have discussed much of what can be gleaned from the legislative history of SL 2013-381 in the sequence of events leading up to its enactment. No minutes of meetings about SL 2013-381 exist. And, as the Supreme Court has recognized, testimony as to the purpose of challenged legislation “frequently will be barred by [legislative] privilege.” Id. That is the case here. See N.C. State Conf., — F.Supp.3d at - n. 124, 2016 WL 1650774, at *71 n. 124. The district court was correct to note that statements from only a few legislators, or those made by legislators after the fact, are of limited value. See id. at-, 2016 WL 1650774, at *146; Barber v. Thomas, 560 U.S. 474, 485-86, 130 S.Ct. 2499, 177 L.Ed.2d 1 (2010); Hunter, 471 U.S. at 228, 105 S.Ct. 1916. We do find worthy of discussion, however, the General Assembly’s requests for and use of race data in connection with SL 2013-381. As explained in detail above, prior to and during the limited debate on the expanded omnibus bill, members of the General Assembly requested and received a breakdown by race of DMV-issued ID ownership, absentee voting, early voting, same-day registration, and provisional voting (which includes out-of-precinct voting). N.C. State Conf., — F.Supp.3d at- -,-, 2016 WL 1650774, at *136-38, *148; J.A. 1628-29, 1637, 1640-41, 1782-97, 3084-3119. This data revealed that African Americans disproportionately used early voting, same-day registration, and out-of-precinct voting, and disproportionately lacked DMV-issued ID. N.C. State Conf., — F.Supp.3d at-, 2016 WL 1650774, at *148; J.A. 1782-97, 3084-3119. Not only that, it also revealed that African Americans did not disproportionately use absentee voting; whites did. J.A. 1796-97, 3744-47. SL 2013-381 drastically restricted all of these other forms of access to the franchise, but exempted absentee voting from the photo ID requirement. In sum, relying on this racial data, the General Assembly enacted legislation restricting all — and only — practices disproportionately used by African Americans. When juxtaposed against the unpersuasive non-racial explanations the State proffered for the specific choices it made, discussed in more detail below, we cannot ignore the choices the General Assembly made with this data in hand. D. Finally, Arlington Heights instructs that courts also consider the “impact of the official action” — that is, whether “it bears more heavily on one race than another.” 429 U.S. at 266, 97 S.Ct. 555 (internal quotation marks omitted). The district court expressly found that “African Americans disproportionately used” the removed voting mechanisms and disproportionately lacked DMV-issued photo ID. N.C. State Conf., — F.Supp.3d at-, 2016 WL 1650774, at *37, *136. Nevertheless, the court concluded that this “disproportionate[ ] use[ ]” did not “significantly favor a finding of discriminatory purpose.” Id. at-, 2016 WL 1650774, at *143. In doing so, the court clearly erred. Apparently, the district court believed that the disproportionate impact of the new legislation “depends on the options remaining” after enactment of the legislation. Id. at -, 2016 WL 1650774, at *136. Arlington Heights requires nothing of the kind. The Arlington Heights Court recognized that “[t]he impact of [a governmental] decision” not to rezone for low-income housing “bear[s] more heavily on racial minorities.” 429 U.S. at 269, 97 S.Ct. 555. In concluding that the zoning decision had a disproportionate impact, the Court explained that “[mjinorities constitute^] 18% of the Chicago area population, and 40% of the income groups said to be eligible for” the low-income housing. Id. The Court did not require those minority plaintiffs to show that the Chicago area as a whole lacked low-income housing or that the plaintiffs had no other housing options. Instead, it was sufficient that the zoning decision excluded them from a particular area. Id. at 260, 265-66, 269, 97 S.Ct. 555; see also City of Memphis v. Greene, 451 U.S. 100, 110, 126, 101 S.Ct. 1584, 67 L.Ed.2d 769 (1981) (indicating that closing a street used primarily by African Americans had a disproportionate impact, even though “the extent of the inconvenience [was] not great”). Thus, the standard the district court used to measure impact required too much in the context of an intentional discrimination claim. When plaintiffs contend that a law was motivated by discriminatory intent, proof of disproportionate impact is not “the sole touchstone” of the claim. Davis, 426 U.S. at 242, 96 S.Ct. 2040. Rather, plaintiffs asserting such claims must offer other evidence that establishes discriminatory intent in the totality of the circumstances. Id. at 239-42, 96 S.Ct. 2040. Showing disproportionate impact, even if not overwhelming impact, suffices to establish one of the circumstances evidencing discriminatory intent. Accordingly, the district court’s findings that African Americans disproportionately used each of the removed mechanisms, as well as disproportionately lacked the photo ID required by SL 2013-381, if supported by the evidence, establishes sufficient disproportionate impact for an Arlington Heights analysis. As outlined above, the record evidence provides abundant support for that holding. Moreover, the district court also clearly erred in finding that the cumulative impact of the challenged provisions of SL 2013-381 does not bear more heavily on African Americans. See Clingman v. Beaver, 544 U.S. 581, 607-08, 125 S.Ct. 2029, 161 L.Ed.2d 920 (2005) (O’Connor, J., concurring) (“A panoply of regulations, each apparently defensible when considered alone, may nevertheless have the combined effect of severely restricting participation and competition.”). For example, the photo ID requirement inevitably increases the steps required to vot