Page v. Virginia State Board of Elections

MEMORANDUM OPINION DUNCAN, Circuit Judge: In the political landscape prior to the Supreme Court’s June 25, 2013, decision in Shelby County v. Holder, — U.S. —, 133 S.Ct. 2612, 186 L.Ed.2d 651 (2013), the Virginia legislature undertook the task of crafting United States congressional districts with the overarching goal of compliance with the Voting Rights Act of 1965 (‘VRA”) as it was then interpreted. In describing the methodology used in drawing the abstract lines currently under consideration, Delegate William Janis, the architect of that legislation, explained it thus: I focused on the [Third] Congressional District and ensuring, based on recommendations that I received from Congressman Scott[,the representative from the Third Congressional District,] and from all 11 members of the congressional delegation, Republican and Democrat — one of the paramount concerns and considerations that was not permissive and nonnegotiable under federal law and under constitutional precedent is that the [Third] Congressional District not retrogress in minority voter influence. And that’s how the lines were drawn.... [T]he primary focus of how the lines in [the redistricting legislation] were drawn was to ensure that there be , no retrogression in the [Third] Congressional District. Because if that occurred, the plan would be unlikely to survive a challenge either through the Justice Department or the courts because it would not comply with the constitutionally mandated requirement that there be no retrogression in the minority voting influence in the [Third] Congressional District. Pl.’s Trial Ex. 43, at 25.1 Delegate Janis’s efforts were successful. His proposed legislation was approved by the United States Department of Justice (“DOJ”), which found that it did not effect any retrogression in the ability of minorities to elect their candidates of choice.2 As we explain below, however, the Supreme Court’s Shelby County decision significantly altered the status quo. Before turning to a description of the history of the litigation and an analysis of the issues it presents, we wish to emphasize at the outset what we hope will be clear throughout. We imply no criticism of Delegate Janis or Defendants, and do not question that all attempted to act appropriately under the circumstances as they understood them to be at the time. We must nevertheless determine whether the Virginia legislation passes constitutional muster, particularly in the wake of Shelby County. I. THE LITIGATION Plaintiffs Dawn Curry Page, Gloria Per-sonhuballah, and James Farkas3 (“Plain — 537 — tiffs”) bring this action against Defendants Charlie Judd, Kimberly Bowers, and Don Palmer — in their respective official capacities of Chairman, Vice-Chair, and Secretary of the Virginia State Board of Elections4 — and Intervenor-Defendants Eric Cantor, Robert J. Wittman, Bob Goodlatte, Frank Wolf, Randy J. Forbes, Morgan Griffith, Scott Rigell, and Robert Hurt — all Congressmen in the Commonwealth of Virginia — (collectively, “Defendants”)5 challenging the constitutionality of Virginia’s Third Congressional District as a racial gerrymander in violation of the Equal Protection Clause of the Fourteenth Amendment of the United States Constitution. In light of the evidence, and as State Board of Elections Defendants have acknowledged, we conclude that compliance with Section 5 of the VRA (“Section 5”), and accordingly, race, “was the [legislature’s] predominant purpose ... underlying [the Third Congressional District’s] racial composition in 2012.” (Int.-Def.’s Mem. Supp. Mot. Summ. J. 15, ECF No. 39). Moreover, the redistricting plan cannot survive the strict scrutiny required of race-conscious districting because it is not narrowly tailored.6 Equitable considerations preclude remediation prior to Virginia’s November 2014 elections. Because, however, the constitutional infirmities of the Third Congressional District cannot be remedied in isolation, Virginia should act within the next legislative session to draw electoral districts based on permissible criteria. Resolution of the issues before us involves an analysis of the interplay between the VRA and Virginia law governing voting rights and the redistricting process. We therefore begin by laying out the framework that will guide that analysis. We then set out the factual background and procedural history of this litigation, before proceeding to the issues at hand. A. Voting Rights Act Background A brief description of the history and purpose of the VRA, and its impact on Virginia, is a useful predicate for the discussion that follows. The VRA, passed in 1965, “was originally perceived as a remedial provision directed specifically at eradicating discriminatory practices that restricted blacks’ ability to register and vote in the segregated South.” Holder v. Hall, 512 U.S. 874, 893, 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994) (Thomas, J., concurring). The VRA “is a complex scheme of stringent remedies aimed at areas where voting discrimination has been most flagrant.” South Carolina v. Katzenbach, 383 U.S. 301, 315, 86 S.Ct. 803, 15 L.Ed.2d. 769 (1966). Section 4 of the VRA outlines “a formula defining the States and political subdivisions to which [the statute’s] ... remedies apply.” Id. This “coverage formula” includes states or political subdivisions with the following characteristics: 1) as of November 1964, they maintained a test or device as a prerequisite for voting or registration; and 2) 1964 census data indicated that less than 50% of the voting-age popu — 538 — lation was registered to vote. See 42 U.S.C. § 1973b(b)(l)-(2). Section 5 contains specific redistricting requirements for jurisdictions deemed covered under Section 4. See id. § 1973c. In November 1964, Virginia met the criteria to be classified as a “covered jurisdiction” under Section 5. See id. § 1973b-c. As such, Virginia was required to submit any changes to its election or voting laws to the DOJ for federal preapproval, a process called “preclearance.” See id. § 1973c. To obtain preclearance, Virginia had to demonstrate that a proposed change had neither the purpose nor effect “of denying or abridging the right to vote on account of race or color.” Id. § 1973c (a). The legal landscape changed dramatically in 2013, when the Supreme Court ruled that Section 4’s coverage formula, described above, was unconstitutional. Shelby Cnty., 133 S.Ct. at 2620-31. The Court concluded that the formula, although rational in practice and theory when the VRA was passed in 1965, was no longer justified by current voting conditions. Id. As a result of the invalidation of the coverage formula under Section 4, Virginia is no longer obligated to comply with the pre-clearance requirements of Section 5. See id. B. Factual Background We turn now to the Virginia constitutional and statutory scheme. The Virginia Constitution requires the state legislature to reapportion Virginia’s United States congressional districts every ten years based on federal census data. Districts must be “contiguous and compact territory ... constituted as to give, as nearly as practicable, representation in proportion to the population of the district.” Va. Const, art. II, § 6. Virginia’s Third Congressional District was first created as a majority African-American district in 1991. See Va.Code §§ 24.1-17.303 (1991); 24.1-17.303 (1992); 24.2-302 (1993). At that time, the Third Congressional District had an African-American population of 63.98%, and a black voting-age population (“BVAP,” the percentage of persons of voting age who identify as African-American) of 61.17%. Moon v. Meadows, 952 F.Supp. 1141, 1143-44 (E.D.Va.1997). The 2010 federal census showed that Virginia’s population grew 13% between 2000 and 2010. Pl.’s Trial Ex. 1, at 18. Because the growth was unevenly distributed, Virginia had to redraw its congressional districts in order to balance population totals within each district. See id. Pursuant to that goal, Virginia’s Senate Committee on Privileges and Elections adopted Committee Resolution No. 2, establishing goals and criteria concerning applicable legal requirements and policy objectives for redrawing Virginia’s congressional districts. See Pl.’s Trial Ex. 5. The criteria included: 1) population equality among districts; 2) compliance with the laws of the United States and Virginia, including protections against diluting racial minority voting strength and putting minority voters in a worse position than they were before the redistricting change (“retrogression”); 3) contiguous and compact districts; 4) single-member districts; and 5) consideration of communities of interest. Id. at 1-2. The Virginia Senate noted that, although “[a]ll of the foregoing criteria [would] be considered in the dis-tricting proeess[,] ... population equality among districts and compliance with federal and state constitutional requirements and [the VRA] [would] be given priority in the event of conflict among the criteria.” Id. at 2 (emphasis added). — 539 — Delegate Janis used the 2010 census data to draw a new plan for Virginia’s United States congressional districts. Delegate Janis presented his plan, House Bill 5004, to the House of Delegates on April 6, 2011; the House adopted it six days later. Pl.’s Trial Ex. 8, at 7. The Virginia Senate, however, rejected Delegate Janis’s plan and replaced it with a plan sponsored by State Senator Mamie Locke. Id. The House and Senate were unable to reconcile the competing plans and the redistricting effort stalled. Id. at 8. The November 2011 elections changed the composition of the Virginia Senate, and, in January 2012, the newly seated House and Senate adopted Delegate Janis’s plan without any changes.7 See id. Governor Bob McDonnell signed the plan into law on January 25, 2012. Id. at 9. The congressional districting plan (“2012 Plan”) is codified at Va.Code Ann. § 24.2-302.2. The 2012 Plan divides Virginia into eleven congressional districts. Plaintiffs describe the boundaries of the Third Congressional District as follows: The northwest corner of the district includes parts of Richmond and the north shore of the James River. It then crosses the James River for the first time and juts west to capture parts of Petersburg. The district again crosses to the north shore of the James River to include parts of Newport News, though this portion of the district is not contiguous with any other part of the district. The district then hops over part of Congressional District 2 to include part of Hampton and crosses the James River and Chesapeake Bay to capture part of Norfolk, which is not contiguous with any other part of [the district]. (Compl. ¶ 34, ECF No. 1). A majority of the voting age population in the 2012 Plan’s Third Congressional District is African-American. Whereas the BVAP of the previous iteration of the Third Congressional District (“Benchmark Plan”), formed after the 2000 census, was 53.1%, the BVAP of the 2012 Plan’s Third Congressional District is 56.3%. PL’s Trial Ex. 27, at 14. There is no indication that this increase of more than three percentage points was needed to ensure nonretrogression, however, because the 2012 Plan was not informed by a racial bloc voting or other, similar type of analysis. See Trial Tr. 342:11-23, 354:18-355:2. A racial bloc voting analysis, which legislatures frequently use in redistricting, studies the electoral behavior of minority voters and ascertains how many African-American voters are needed in a congressional district to avoid diminishing minority voters’ ability to elect their candidates of choice. Trial Tr. 62:22-63:7, 98:16-99:6, 198:5-8; PL’s Trial Ex. 43, at 15. Virginia submitted the 2012 Plan to the DOJ for Section 5 preclearance. As we have noted, the DOJ precleared the plan on March 14, 2012, finding that it did not effect any retrogression in the ability of minorities to elect their candidates of choice. (Def.’s Mem. Supp. Mot. Summ. J. 7, ECF No. 37). On June 25, 2013, the Supreme Court issued its decision in Shelby County. As a result, as we have explained, Section 5’s requirements of review and preclearance for covered areas no longer apply to Virginia with respect to future changes to its voting and election laws. See Shelby Cnty., 133 S.Ct. at 2627-31. — 540 — C. Procedural History Plaintiffs8 brought this action on October 2, 2013, alleging that Virginia used the Section 5 preclearance requirements as a pretext to pack African-American voters into Virginia’s Third Congressional District and reduce these voters’ influence in other districts. (Compl. ¶¶ 3, 40, ECF No. 1). Plaintiffs seek a declaratory judgment that Virginia’s Third Congressional District, as drawn in the 2012 Plan, is a racial gerrymander in violation of the Equal Protection Clause of the Fourteenth Amendment. Id. at 10. Plaintiffs also seek to permanently enjoin Defendants from giving effect to the boundaries of the Third Congressional District, including barring Defendants from conducting elections for the United States House of Representatives based on the current Third Congressional District. Id. Any action under Section 5 must “be heard and determined by a court of three judges in accordance with the provisions of section 2284 of Title 28.” 42 U.S.C. § 1973c; see also Allen v. State Bd. of Elections, 393 U.S. 544, 560-63, 89 S.Ct. 817, 22 L.Ed.2d 1 (1969). Because Plaintiffs’ action “challeng[es] the constitutionality of the apportionment of congressional districts” in Virginia, the Chief Judge of the United States Court of Appeals for the Fourth Circuit granted Plaintiffs’ request for a hearing by a three-judge court pursuant to 28 U.S.C. 2284(a) on October 21, 2013. (ECF No. 10). Virginia Congressmen Eric Cantor, Robert J. Wittman, Bob Goodlatte, Frank Wolf, Randy J. Forbes, Morgan Griffith, Scott Rigell, and Robert Hurt moved to intervene as Defendants in the case on November 25, 2013. (ECF No. 14). On December 20, 2013, all Defendants moved for summary judgment. (ECF Nos. 35, 38). We denied the motions on January 27, 2014. (ECF No. 50). A two-day bench trial began on May 21, 2014. (ECF Nos. 100, 101). We then ordered the parties to file post-trial briefs. (ECF No. 99). After reviewing those briefs, we determined on June 30, 2014, that further oral argument would not assist in the resolution of the issues before the Court. (ECF No. 108). Therefore, this case is now ripe for disposition. We have jurisdiction pursuant to 28 U.S.C. § 1331, 28 U.S.C. § 1343(a)(3), and 28 U.S.C. § 1357. II. ANALYSIS To successfully challenge the constitutionality of the Third Congressional District under the Equal Protection Clause, Plaintiffs first bear the burden of proving that the legislature’s predominant consideration in drawing its electoral boundaries was race. If they make this showing, the assignment of voters according to race triggers the court’s “strictest scrutiny.” Miller v. Johnson, 515 U.S. 900, 915, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995). Then, the burden of production shifts to Defendants to demonstrate that the redistricting plan' was narrowly tailored to advance a compelling state interest. See Shaw II, 517 U.S. at 908, 116 S.Ct. 1894. For the reasons that follow, we find that Plaintiffs have shown race predominated. We find that the Third Congressional District cannot survive review under the exacting standard of strict scrutiny. While compliance with Section 5 was a compelling interest when the legislature acted, the redistricting plan was not narrowly tailored to farther that interest. Accord — 541 — ingly, we are compelled to hold that the challenged Third Congressional District violates the Equal Protection Clause of the Fourteenth Amendment. A. Race As the Predominant Consideration in Redistricting As with any law that distinguishes among individuals on the basis of race, “equal protection principles govern a State’s drawing of congressional districts.” Miller, 515 U.S. at 905, 115 S.Ct. 2475. “Racial classifications with respect to voting carry particular dangers. Racial gerrymandering, even for remedial purposes, may balkanize us into competing racial factions; it threatens to carry us further from the goal of a political system in which race no longer matters.... ” Shaw v. Reno (Shaw I), 509 U.S. 630, 657, 113 S.Ct. 2816, 125 L.Ed.2d 511 (1993). As such, “race-based districting by our state legislatures demands close judicial scrutiny.” Id. To trigger strict scrutiny, Plaintiffs first bear the burden of proving that race was not only one of several factors that the legislature considered in drawing the Third Congressional District, but that race “predominated.” Bush v. Vera, 517 U.S. 952, 963, 116 S.Ct. 1941, 135 L.Ed.2d 248 (1996). The Supreme Court has emphasized that this burden “is a ‘demanding one,’ ” Easley v. Cromartie (Cromartie II), 532 U.S. 234, 241, 121 S.Ct. 1452, 149 L.Ed.2d 430 (2001) (quoting Miller, 515 U.S. at 928, 115 S.Ct. 2475): The plaintiffs burden is to show, either through circumstantial evidence of a district’s shape and demographics or more direct evidence going to legislative purpose, that race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without a particular district. To make this showing, a plaintiff must prove that the legislature subordinated traditional race-neutral districting principles, including but not limited to compactness, contiguity, and respect for political subdivisions or communities defined by actual shared interests, to racial considerations. Miller, 515 U.S. at 916, 115 S.Ct. 2475. The Supreme Court has cited several specific factors as evidence of racial line drawing: statements by legislators indicating that race was a predominant factor in redistricting, see id., 515 U.S. at 917-18, 115 S.Ct. 2475; evidence that race or percentage of race within a district was the single redistricting criterion that could not be compromised, see Shaw II, 517 U.S. at 906, 116 S.Ct. 1894; creation of non-compact and oddly shaped districts beyond what is strictly necessary to avoid retrogression, see Shaw I, 509 U.S. at 646-48, 113 S.Ct. 2816; use of land bridges in a deliberate attempt to bring African-American population into a district, see Miller, 515 U.S. at 917, 115 S.Ct. 2475; and creation of districts that exhibit disregard for city limits, local election precincts, and voting tabulation districts (“VTDs”), see Bush, 517 U.S. at 974, 116 S.Ct. 1941. As we demonstrate below, all of these factors are present here.9 Moreover, we do not view any — 542 — of these factors in isolation. We consider direct evidence of legislative intent, including statements by the legislation’s sole sponsor, in conjunction with the circumstantial evidence supporting whether the 2012 Plan complies with traditional redistricting principles. 1. Direct Evidence of Legislative Intent When analyzing the legislative intent underlying a redistricting decision, we agree with the dissent that there is a “presumption of good faith that must be accorded legislative enactments.” Miller, 515 U.S. at 916, 115 S.Ct. 2475. This presumption “requires courts to exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race.” Id. Such restraint is particularly warranted given the “complex interplay of forces that enter a legislature’s redistricting calculus,” id. at 915-16, 115, S.Ct. 2475, making 'redistricting possibly “the most difficult task a legislative body ever undertakes,” Smith v. Beasley, 946 F.Supp. 1174, 1207 (D.S.C.1996) (three-judge court). Nevertheless, “the good faith of the legislature does not excuse or cure the constitutional violation of separating voters according to race.” Id. at 1208. Here, “[w]e do not question the good faith of the legislature in adopting [the 2012 Plan]” so long as “[t]he members did what they thought was required by [Section 5] and by the Department of Justice at the time.” Id. At this stage of the analysis, we are concerned only with whether legislative statements indicate that “race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without [the Third Congressional District].” Miller, 515 U.S. at 916, 115 S.Ct. 2475. We find such statements here, drawn from multiple sources. We must also note, however, that it is inappropriate to confuse this presumption of good faith with an obligation to parse legislative intent in search of “proper” versus “improper” motives underlying the use of race as the predominant factor in redistricting, as the dissent does here. The legislative record here is replete with statements indicating that race was the legislature’s paramount concern in enacting the 2012 Plan. Yet the dissent urges us to consider such statements as mere legislative acknowledgments of the supremacy of federal law, specifically the VRA. The dissent argues that subjecting a redistricting plan to strict scrutiny when it separates voters according to race as a means to comply with Section 5 “trap[s] [legislatures] between the competing hazards of [VRA and Constitutional] liability,” Bush, 517 U.S. at 992, 116 S.Ct. 1941 (O’Connor, J., concurring),10 but this is a red herring. While “[a]pplying traditional equal protection principles in the voting-rights context is ‘a most delicate task,’ ” Shaw II, 517 — 543 — U.S. at 905, 116 S.Ct. 1894 (quoting Miller, 515 U.S. at 905, 115 S.Ct. 2475), we must apply strict scrutiny when, as here, there is strong direct and circumstantial evidence that race was the only “nonnegotiable” criterion. a.Defendants’ Statements Defendants concede that avoiding retrogression in the Third Congressional District and ensuring compliance with Section 5 was the legislature’s primary priority in drawing the 2012 Plan. Defendants acknowledge that the legislature’s top two priorities were “compliance with applicable federal and state laws, expressly including the [VRA]” and population equality. (Def.’s Mem. Supp. Mot. Summ. J. 12, ECF No. 37). Moreover, Defendants “concede[ ] that compliance with Section 5 was [the legislature’s] predominant purpose or compelling interest underlying District 3’s racial composition in 2012.” (Int-Def.’s Mem. Supp. Mot. Summ. J. 15, ECF No. 39). Of course, we do not view the language of the Intervenor-Defen-dants’ summary judgment brief as a “binding concession,” as the dissent suggests. Rather, we take it for what it is — a candid acknowledgement of the incontrovertible fact that the shape of the Third Congressional District was motivated by the desire to avoid minority retrogression in voting, b.Racial Threshold As the Means to Achieve Section 5 Compliance Defendants’ expert, John Morgan, also acknowledged that the legislature “adopted the [2012 Plan] with the [Third Congressional District] Black VAP at 56.3%” because legislators were conscious of maintaining a 55% BVAP floor. Int. Def.’s Trial Ex. 13, at 27. In 2011, the legislature enacted “a House of Delegates redistricting plan with a 55% Black VAP as the floor for black-majority districts” with strong bipartisan support. Id. at 26. Given the success of this prior usage of a 55% BVAP floor, the legislature considered a 55% BVAP floor for the 2012 congressional redistricting “appropriate to obtain Section 5 preclearance, even if it meant raising the Black VAP above the [53.1%] level[] in the Benchmark plan.” Id. at 26-27. The legislature therefore “acted in accordance with that view,” id. at 27, when adopting the 2012 Plan, despite the fact that the use of a 55% BVAP floor in this instance was not informed by an analysis of voter patterns. Indeed, when asked on the House floor whether he had “any empirical evidence whatsoever that 55[% BVAP] is different than 51[%] or 50[%],” or whether the 55% floor was “just .a number that has been pulled out of the air,” Delegate Janis, the redistricting bill’s author, characterized the use of a BVAP floor as “weighing a certainty against an uncertainty.” Pl.’s Trial Ex. 45, at 7. c.Statements by the Author of the 2012 Congressional Maps In addition to Defendants’ statements, we credit explanations by Delegate Janis, the legislation’s sole author, stating that he considered race the single “nonnegotiable” redistricting criterion. PL’s Trial Ex. 43, at 25. In disagreeing, the dissent attempts to discount the meaning of these statements by placing great reliance on remarks by legislative opponents characterizing the redistricting legislation as an incumbency protection plan, and by parsing Delegate Janis’s statements regarding compliance with federal law generally from the necessary antecedent of relying on race to do so. In the face of Delegate Janis’s clear words, we do not find these efforts. persuasive.11 — 544 — Delegate Janis emphasized that his “primary focus” in drawing Virginia’s new congressional maps was ensuring that the Third Congressional District maintained at least as large a percentage of African-American voters as had been present in the district under the Benchmark Plan. Pl.’s Trial Ex. 43, at 25; see also Pl.’s Trial Ex. 13, at 8 (“[W]e can have no less [percentage of African-American voters] than percentages that we have under the existing lines.”). For example, at the second floor reading of the redistricting bill in Virginia’s House of Delegates on April 12, 2011, Delegate Janis noted that “one of the paramount concerns in the drafting of the bill was [the VRA mandate] that [the legislature] not retrogress minority voting influence in the 3rd Congressional District.” PL’s Trial Ex. 43, at 10 (emphasis added). He continued to reiterate this sentiment, noting that he was “most especially focused on making sure that the [Third] Congressional District did not retrogress in its minority voting influence.” Id. at 14-15 (emphasis added). Delegate Janis also stated that the avoidance of retrogression in the Third Congressional District took primacy over other redistricting considerations because it was “nonnegotiable”: [0]ne of the paramount concerns and considerations that was not permissive and nonnegotiable ... is that the 3rd Congressional District not retrogress in minority voter influence.... [T]he primary focus of how the lines in House Bill 5004 were drawn was to ensure that there be no retrogression in the 3rd Congressional District. Because if that occurred, the plan would be unlikely to survive a challenge either through the Justice Department or the courts because it would not comply with the constitutionally mandated requirement that there be no retrogression in the minority voting influence in the 3rd Congressional District. Id. at 24-25 (emphasis added). Unlike the dissent, we deem it appropriate to accept the explanation of the legislation’s author as to its purpose. And there is further support. 2. Circumstantial Evidence of the Third Congressional District’s Shape and Characteristics In addition to the evidence of legislative intent, we also consider the extent to which the district boundaries manifest that legislative will.12 Evidence of a “highly — 545 — irregular” reapportionment plan “in which a State concentrated a dispersed minority population in a single district by disregarding traditional districting principles such as compactness, contiguity, and respect for political subdivisions,” indicates that racial considerations predominated during the 2011-12 redistricting cycle. Shaw I, 509 U.S. at 647, 113 S.Ct. 2816. We consider each of these factors below. a. Shape and Compactness As the Supreme Court has recognized, “reapportionment is one area in which appearances do matter,” Shaw I, 509 U.S. at 647, 113 S.Ct. 2816, and the “obvious fact that the district’s shape is highly irregular and geographically non-compact by any objective standard” supports the conclusion that race was the predominant factor in drawing the challenged district. Shaw II, 517 U.S. at 905-06, 116 S.Ct. 1894 (internal quotation marks omitted). Moreover, compactness is one of two redistricting criteria required by the Virginia Constitution. Va. Const, art. II, § 6 (“Every electoral district shall be composed of contiguous and compact territory----”). Because, as he explained to the Senate Committee on Privileges and Elections, Delegate Janis “didn’t examine compactness scores” when drawing the 2012 congressional maps, Pl.’s Trial Ex. 14, at 8, we begin with a visual, rather than mathematical, overview of the Third Congressional District’s shape and compactness. See Karcher v. Daggett, 462 U.S. 725, 762, 103 S.Ct. 2653 (Without applying any mathematical measures of compactness, “[a] glance at the [congressional] map shows district configurations well deserving the kind of descriptive adjectives ... that have traditionally been used to describe acknowledged gerrymanders.”). Plaintiffs contend that the Third Congressional District is the least compact congressional district in Virginia. Trial Tr. 73:10-14. And, indeed, the maps of the district reflect both an odd shape and a composition of a disparate chain of communities, predominantly African-American, loosely connected by the James River. See Trial Tr. 42:13-16; PL’s Trial Ex. 48. Defendants do not disagree. In fact, Defendants’ expert, Mr. Morgan, concedes that the three primary statistical procedures used to measure the degree of compactness of a district all indicate that the Third Congressional District is the least compact congressional district in Virginia. Trial Tr. 375:21-24, 376:9-13. While Defendants acknowledge the irregularity of shape and lack of compactness reflected by the Third Congressional District, they submit that a desire to protect Republican incumbents explains the District’s shape, a contention we discuss later. See infra section II.A.3; see also Trial Tr. 14:24-15:6. b. Non-Contiguousness In addition to requiring compactness, the Virginia Constitution also requires the legislature to consider contiguity when drawing congressional boundaries. See Va. Const, art. II, § 6. The Virginia Supreme Court has concluded that “land masses separated by water may ... satisfy the contiguity requirement in certain circumstances.” Wilkins v. West, 264 Va. 447, 571 S.E.2d 100, 109 (2002). While the Third Congressional District is not contiguous by land, it is legally contiguous because all segments of the district border the James River. Trial Tr. 74:22-24. Therefore, the Third Congressional Dis — 546 — trict is legally contiguous under Virginia Law. See Wilkins, 571 S.E.2d at 109; see also Trial Tr. 221:12-14. Yet contiguity and other traditional districting principles are “important not because they are constitutionally required,” but rather “because they are objective factors” courts may consider in assessing racial gerrymandering claims. Shaw I, 509 U.S. at 647, 113 S.Ct. 2816. To show that race predominated, Plaintiffs need not establish that the legislature disregarded every traditional districting principle. See Miller, 515 U.S. at 917, 115 S.Ct. 2475 (holding that circumstantial evidence such as shape does not need to be sufficient, standing alone, to establish a racial gerrymandering claim). Rather, we consider irregularities in the application of these traditional principles together. Here, the record establishes that, in drawing the boundaries of the Third Congressional District, the legislature used water contiguity as a means to bypass white communities and connect predominantly African-American populations in areas such as Norfolk, Newport News, and Hampton. See Trial Tr. 75:15-76:1. Such circumstantial evidence is one factor that contributes to the overall conclusion that the district’s boundaries were drawn with a focus on race. c. Splits in Political Subdivisions “[R]espect for political subdivisions” is an important traditional districting principle. Shaw I, 509 U.S. at 647, 113 S.Ct. 2816. A county or city is considered split by a congressional district when a district does not entirely contain that county or city within its borders. See Pl.’s Trial Ex. 27, at 8. The Third Congressional District splits more local political boundaries than any other district in Virginia. Trial Tr. 76:18-20. It splits nine counties or cities, the highest number of any congressional district in the 2012 Plan. PL’s Trial Ex. 27, at 9. Moreover, the boundaries of the Third Congressional District contribute to the majority of splits in its neighboring congressional districts. See id. The Third Congressional District also splits more voting tabulation districts, or VTDs, than any of Virginia’s other congressional districts. Trial Tr. 78:17-19; see also PL’s Trial Ex. 27, at 10. A VTD is a Census Bureau term referring to what is commonly thought of as a voting precinct. Trial Tr. 78:5-8. In total, the 2012 Plan splits 20 VTDs; the Third Congressional District contributes to 14 of them. Trial Tr. 78:20-21; see also PL’s Trial Ex. 27, at 10. While some of these are “technical splits” (i.e., a VTD split that does not involve population; for example, a split across water), such technical splits were used strategically here, as they would not have been necessary “if [the legislature was not] trying to bypass [white] communities using water” and bring predominantly African-American communities into the district. Trial Tr. 79-80. The dissent contends that the population swaps involving the Third Congressional District — and resulting locality splits— were necessary to achieve population parity in accordance with the constitutional mandate of the one-person-one-vote rule,13 and can also be explained by the traditional redistricting criterion of “preserving district cores.”14 See post at 28, 36. The — 547 — evidence does not substantiate either of these arguments. It is true that the Virginia legislature needed to add 63,976 people to the Third Congressional District to achieve population parity. See Trial Tr. 87. Yet, though the dissent asserts that “it is extremely unlikely that any combination of ‘whole’ localities in the vicinity of [the Benchmark Plan] could have been added to the [Third Congressional] District to augment the population by exactly 63,976 people,” post at 36, Plaintiffs’ alternative plan maintains a majority-minority district and achieves the population increase needed for parity, while simultaneously minimizing locality splits and the number of people affected by such splits. See Pl.’s Trial Ex. 29, at 1. Although this alternative plan results in only one less locality split than the 2012 Plan, it reduces the number of people affected by the locality splits between the Third Congressional District and Second Congressional District by 240,080.15 See Trial Tr. 112; PL’s Trial Ex. 29, at 5, tbl. 3. The alternative plan also reduces the number of VTD splits involving the Third Congressional District from 14 in the 2012 Plan to 11. Trial Tr. 111. Moreover, Plaintiffs’ alternative plan, unlike the 2012 Plan, keeps the cities of Newport News, Hampton, and Norfolk intact.16 See id. at 112. This is a particularly important accomplishment because it reflects the fulfillment of a strong public sentiment, as expressed during 2010 redistricting forums,17 against splitting localities, and in favor of keeping the integrity of cities like Hampton and Norfolk intact. See Pl.’s Trial Ex. 29, at 5; see also Pl.’s Trial Ex. 11-12. The evidence similarly undercuts the dissent’s contention that the boundaries of the Third Congressional District reflect an allegiance to the traditional redistricting principle of preserving district cores. Far from attempting to retain most of the Benchmark Plan’s residents within the new district borders, the 2012 Plan moved over 180,000 people in and out of the districts surrounding the Third Congressional District to achieve an overall population increase of only 63,976 people. Trial Tr. 87. Tellingly, the populations moved out of the Third Congressional District were predominantly white, while the populations moved into the District were predominantly African-American. Id. at 81-82. Moreover, the predominantly white populations moved out of the Third Congressional District totaled nearly 59,000 residents — a number very close to the total required increase of 63,976 people. See Pl.’s Trial Ex. 27, at 15, tbl. 6; see also Trial Tr. 87. While “[t]he Constitution does not mandate regularity of district shape,” Bush, — 548 — 517 U.S. at 962, 116 S.Ct. 1941, Plaintiffs’ circumstantial evidence of the Third Congressional District’s irregularities and inconsistencies with respect to the traditional districting criteria described above, coupled with clear statements of legislative intent, supports our conclusion that, in this case, “traditional districting criteria [were] subordinated to race.” Id. 3. Predominance of Race over Politics Defendants, as well as the dissent, rely heavily on isolated statements in the legislative record, made by opponents of Delegate Janis’s bill, suggesting that incumbency protection and partisan politics motivated the 2011-12 redistricting efforts. See, e.g., Pl.’s Trial Ex. 43, at 48-49 (opponent of Delegate Janis’s plan stating that Janis “admitted today that one of the criteria that he used in development of the plan was incumbent protection,” and deeming the redistricting effort “one for incumbency protection first, last, alpha, and omega”); id. at 27 (opponent of the 2012 Plan suggesting that Delegate Janis used incumbency protection as a permissive redistricting criteria). The Supreme Court has made it clear, however, that the views of legislative opponents carry little legal weight in characterizing legislation. See, e.g., Shell Oil Co. v. Iowa Dep’t of Revenue, 488 U.S. 19, 29, 109 S.Ct. 278, 102 L.Ed.2d 186 (1988) (“The fears and doubts of the opposition are no authoritative guide to the construction of legislation.”); see also N.L.R.B. v. Fruit & Vegetable Packers, 377 U.S. 58, 66, 84 S.Ct. 1063, 12 L.Ed.2d 129 (1964) (“[W]e have often cautioned against the danger, when interpreting a statute, of reliance upon the views of its legislative opponents. In their zeal to defeat a bill, they understandably tend to overstate its reach.”); Schwegmann Bros, v. Calvert Distillers Corp., 341 U.S. 384, 394-95, 71 S.Ct. 745, 95 L.Ed. 1035 (1951) (“It is the sponsors that we look to when the meaning of the statutory words is in doubt.”). The rationale for this authority is patent: a bill’s opponents have every incentive to place a competing label on a statute they find objectionable. Defendants and the dissent are inarguably correct that partisan political considerations, as well as a desire to protect incumbents, played a role in drawing district lines. It would be remarkable if they did not. However, in a “mixed motive suit” — in which a state’s conceded goal of “producing] majority-minority districts” is accompanied by “other goals, particularly incumbency protection” — race can be a predominant factor in the drawing of a district without the districting revisions being “purely race-based.”18 Bush, 517 U.S. at 959, 116 S.Ct. 1941. Indeed, the Supreme Court has observed that “partisan politicking” may often play a role in a state’s redistricting process, but the fact “[t]hat the legislature addressed these interests [need] not in any way refute the fact that race was the legislature’s predominant consideration.” Shaw II, 517 U.S. at 907, 116 S.Ct. 1894. The dissent’s attempts to analogize this case to Cromartie II are unavailing. Cro-martie II involved a challenged district in which “racial identification correlated] highly with political affiliation,” 532 U.S. at 258, 121 S.Ct. 1452, and the plaintiffs were ultimately unable to show that “the legislature could have achieved its legitimate political objectives in alternative ways that are comjparably consistent with traditional — 549 — districting principles” because the challenged redistricting plan furthered the race-neutral political goal of incumbency protection to the same extent as it increased the proportion of minorities within the district, id. While it may be true, as the dissent observes, that Democratic votes in the Third Congressional District, and presumably many similarly-situated districts, “can generally be predicted simply by taking the BVAP of a VTD and adding about 21 percentage points,”19 post at 26, the evidence of political justification for the redistricting at issue in Cromartie II is quite different than that presented in this case. In Cromartie II, there was overwhelming evidence in the record “articulating] a legitimate political explanation for [the state’s] districting decision,” 532 U.S. at 242, 121 S.Ct. 1452, including unequivocal trial testimony by state legislators. While Defendants have offered post-hoc political justifications for the 2012 Plan in their briefs, neither the legislative history as a whole, nor the circumstantial evidence, support that view to the extent they suggest. For example, Defendants point to a rather ambiguous statement by Delegate Janis that one goal of the 2012 Plan was to “respect ... the will of the Virginia electorate.” (Post-Trial Br. Int.-Def.’s and Def.’s at 11-12, ECF No. 106 (citing Pi’s. Trial Ex. 43, at 19)). Taken- in context, however, it is clear that this goal was “permissive” and subordinate to the mandatory criteria of compliance with the VRA and satisfaction of the one-person-one-vote rule. See Pi’s. Trial Ex. 43, at 18-19. In support of the argument that political concerns trumped racial ones, the dissent points to Delegate Janis’s remarks that incumbent legislators confirmed their satisfaction with the lines of their respective congressional districts. See id. at 5-6. It is undisputed, however, that the incumbents were not shown the entire 2012 Plan when they were solicited for their input, but were instead shown only the proposed changes to the lines of their individual districts. See Int.-Def.’s Trial Ex. 9, at 9. Delegate Janis testified that he had not asked any congressional representatives “if any of them supported the [redistricting] plan in its totality,” or “[spoken] with anyone who plan[ned] to run against those incumbents” regarding the redistricting plan. Id. at 13-14. Delegate Janis stated: “I haven’t looked at the partisan performance. It was not one of the factors that I considered in the drawing of the district.” Id. at 14. Finally, the nature of the population swaps and shifts used to create the Third Congressional District suggests that less was done to further the goal of incumbency protection than to increase the proportion of minorities within the district. “[A]mong the pool of available VTDs that could have been placed within the Third Congressional District that were highly Democratic performing,” those with a higher BVAP were placed within the Third Congressional District, and those VTDs that were largely white and Democratic were left out, and instead shifted into the — 550 — Second Congressional District.20 Trial Tr. 89.' The record before us presents a picture similar to that in Shaw II, in which the Supreme Court found the evidence sufficient to trigger strict scrutiny: First, the District Court had evidence of the district’s shape and demographics. The court observed the obvious fact that the district’s shape is highly irregular and geographically non-compact by any objective standard that can be conceived. In fact, the serpentine district has been dubbed the least geographically compact district in the Nation. The District Court also had direct evidence of the legislature’s objective. The State’s submission for preclearance expressly acknowledged that [the] overriding purpose was to comply with the dictates of [the DOJ] and to create two congressional districts with effective black voting majorities. Shaw II, 517 U.S. at 905-06, 116 S.Ct. 1894 (internal quotation marks omitted). As we noted earlier, we do not find the dissent’s attempts to distinguish Shaw II from the case at hand persuasive. As an initial matter, it is irrelevant that the challenged district in Shaw II was not only the least compact in the state, as is the Third Congressional District, but also the least compact district in the nation. Irregularities in shape need not be so extreme as to make the district an outlier nationwide; courts simply consider a “highly irregular and geographically non-compact” shape evidence of the predominance of race. Id. at 905-06, 116 S.Ct. 1894. As the least compact and most bizarrely shaped district in the 2012 Plan, the Third Congressional District displays such characteristics. And again, we see no reason why it should make a difference whether Defendants’ “explicit and repeated admissions,” post at 42, of the predominance of race were made in the course of hearings on the House of Delegates floor, as here, or in the State’s Section 5 preclearance submission, as in Shaw II. These specific and repeated references, when taken together with the circumstantial evidence of record, compel our conclusion that race was the legislature’s paramount concern. B. Strict Scrutiny Analysis The fact that race predominated when the legislature devised Virginia’s Third Congressional District in 2012, however, does not automatically render the district constitutionally infirm. Rather, if race predominates, strict scrutiny applies, but the districting plan can still pass constitutional muster if narrowly tailored to serve a compelling governmental interest. See Abrams v. Johnson, 521 U.S. 74, 91, 117 S.Ct. 1925, 138 L.Ed.2d 285 (1997); see also Miller, 515 U.S. at 920, 115 S.Ct. 2475. While such scrutiny is not necessarily “strict in theory, but fatal in fact,” Johnson v. California, 543 U.S. 499, 514, 125 S.Ct. 1141, 160 L.Ed.2d 949 (2005) (quoting Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 237, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995)), the state must establish the “most exact connection between justification and classification.” Parents Involved in Cmty. Sch. v. Seattle — 551 — Sch. Dist. No. 1, 551 U.S. 701, 720, 127 S.Ct. 2738, 168 L.Ed.2d 508 (2007) (quoting Gratz v. Bollinger, 539 U.S. 244, 270, 123 S.Ct. 2411, 156 L.Ed.2d 257 (2003)). And because, as we address below, compliance with the VRA is a compelling state interest, the redistricting plan would not fail under the Equal Protection analysis if it had been narrowly tailored to that interest — if it had not gone “beyond what was reasonably necessary to avoid retrogression.” Bush, 517 U.S. at 984, 116 S.Ct. 1941. While the Third Congressional District was drawn in pursuit of the compelling state interest of compliance with Section 5, Defendants have failed to show that the 2012 Plan was narrowly tailored to further that interest. 1. Compelling Interest The fact that Shelby County effectively relieved Virginia of its Section 5 obligations in 2013 does not answer the question of whether Section 5 compliance in 2012 was a compelling state interest. The appropriate inquiry is whether the legislature’s reliance on racial considerations was, at the time of the redistricting decision, justified by a compelling state interest, not whether it can now be justified in hindsight. See Ala. Legislative Black Caucus v. Alabama, 989 F.Supp.2d 1227, 1307 (M.D.Ala.2013) (three-judge court) (“We evaluate the plans in the light of the legal standard that governed the Legislature when it acted, not based on a later decision of the Supreme Court that exempted [the state] from future coverage under section 5 of the [VRA].”). Although the Supreme Court has yet to decide whether VRA compliance is a compelling state interest, it has assumed as much for the purposes of subsequent anal-yses. See, e.g., Shaw II, 517 U.S. at 914, 116 S.Ct. 1894 (“We assume, arguendo, for the purpose of resolving this suit, that compliance with § 2 [of the Voting Rights Act] could be a compelling interest”); Bush, 517 U.S. at 977, 116 S.Ct. 1941 (“[W]e assume without deciding that compliance with the results test [of the VRA] ... can be a compelling state interest^]”). Particularly because the parties do not dispute that compliance with Section 5 was a compelling interest pre-Shelby County,21 we likewise do not. 2. Narrow Tailoring We now consider whether the 2012 Plan was “narrowly tailored to achieve that compelling interest.” Shaw II, 517 U.S. at 908, 116 S.Ct. 1894. The Supreme Court has repeatedly struck down redistricting plans that did more than was necessary to avoid “a retrogression in the position of racial minorities with respect to their effective exercise of the electoral franchise.” Bush, 517 U.S. at 983, 116 S.Ct. 1941 (quoting Miller, 515 U.S. at 926, 115 S.Ct. 2475); see also Shaw II, 517 U.S. at 910-18, 116 S.Ct. 1894 (concluding that districts were not narrowly tailored to comply with the VRA). — 552 — Indeed, “the [VRA] and our case law make clear that a reapportionment plan that satisfies Section 5 still may be enjoined as unconstitutional.” Shaw I, 509 U.S. at 654, 113 S.Ct. 2816. Section 5 does not “give covered jurisdictions, carte blanche to engage in racial gerrymandering in the name of nonretrogression,” and a reapportionment plan is not narrowly tailored to the goal of Section 5 compliance “if the State went beyond what was reasonably necessary to avoid retrogression.” Shaw I, 509 U.S. at 655, 113 S.Ct. 2816; see also Miller, 515 U.S. at 921, 115 S.Ct. 2475 (“[C]ompliance with federal antidiscrimination laws cannot justify race-based districting where the challenged district was not reasonably necessary under a constitutional reading and application of those laws.”). Courts have cited several specific examples of characteristics of a redistricting plan that would suggest that the plan did more than was “reasonably necessary” to avoid retrogression. These include significantly increasing the total number of African-American voters in a historically “safe” majority-minority district, see Bush, 517 U.S. at 983, 116 S.Ct. 1941; using a BVAP threshold for majority-minority districts, see Smith, 946 F.Supp. at 1210; and generally failing to take specific steps to narrowly tailor a district, such as by conducting a racial bloc voting analysis before making redistricting changes, see Moon,, 952 F.Supp. at 1150. As we explain below, all of these factors are present here, a. BVAP Increase in a Safe Majority-Minority District Although the Third Congressional District has been a safe majority-minority district for 20 years, the 2012 Plan increased the total number of its African-American voting age residents by 44,711.22 Pl.’s Trial Ex. 27, at 11, 14; Trial Tr. 52:18-54:5. This change also increased the district’s BVAP from 53.1% to 56.3%. PL’s Trial Ex. 27, at 14. Congressman Bobby Scott, a Democrat supported by the majority of African-American voters in the Third Congressional District, has represented the District since 1991. PL’s Trial Ex. 21, at 33; Trial Tr. 52:18-21. In the six elections between 2002 to 2012, Congressman Scott ran unopposed in three; he ran opposed in the general elections in 2010 and 2012, but was reelected each time. PL’s Trial Ex. 27, at 11; Trial Tr. 53:7-22. In 2010, Congressman Scott won 70% of the vote, while in 2012 — under the redistricting plan at issue here — he won by an even larger margin, receiving 81.3% of the vote. Id. In this respect, the legislature’s decision to increase the BVAP of the Third Congressional District is similar to the redistricting plan invalidated by the Supreme Court in Bush. See 517 U.S. at 983, 116 S.Ct. 1941. In Bush, a plurality of the Supreme Court held that increasing the BVAP from 35.1% to 50.9% was not narrowly tailored because the state’s interest in avoiding retrogression in a district where African-American voters had successfully elected their representatives of choice for two decades did not justify “substantial augmentation” of the BVAP. Id. Such an augmentation could not be narrowly tailored to the goal of complying with Section 5 because there was “no basis for concluding that the increase to a 50.9% African-American population ... was necessary to ensure nonretrogression.” Id. “Nonretrogression is not a license for the State to do whatever it deems necessary to ensure continued electoral success; it merely mandates that the minority’s op — 553 — portunity to elect representatives of its choice not be diminished, directly or indirectly, by the State’s actions.” Id. While the BVAP increase here is smaller than that in Bush, the principle is the same. Defendants show no basis for concluding that an augmentation of the Third Congressional' District’s BVAP to 56.3% was narrowly tailored when the district had been a safe majority-minority district for two decades. b. BVAP Threshold At trial, Defendants’ expert, Mr. Morgan, confirmed that the legislature adopted a floor of 55% BVAP for the Third Congressional District throughout the 2011-12 redistricting cycle. See Int.-Def.’s Trial Ex. 13, at 26-27. This BVAP threshold was viewed as a proxy for the racial composition needed for a majority-minority district to achieve DOJ preclearance. See id. at 26. Thus, the legislature altered the Third Congressional District’s boundaries in order to meet or exceed that threshold. See id. at 26-27 (noting that legislators “viewed the 55% black VAP ... as appropriate to obtain Section 5 preclearance, even if it meant raising the Black VAP above the levels in the benchmark plan”). Because “[njarrow tailoring [in the dis-tricting context] demands ... that the district chosen entails the least race-conscious measure needed to remedy a violation,” Prejean v. Foster, 227 F.3d 504, 518 (5th Cir.2000), similar ad hoc uses of racial thresholds have been held to defeat narrow tailoring by other three-judge courts. For example, one court invalidated a plan implementing a 55% threshold as arbitrary without supporting evidence. See Smith, 946 F.Supp. at 1210 (holding that narrow tailoring requires legislatures to consider the fact that a 55% BVAP will not be needed to elect a candidate of choice in districts where most minority citizens register and vote, and cautioning against “insist[ing] that all majority-minority districts have at least 55% BVAP with no evidence as to registration or voter turnout”). The legislature’s use of a BVAP threshold, as opposed to a more sophisticated analysis of racial voting patterns, suggests that voting patterns in the Third Congressional District were not “considered individually.” Id.23 Considering the foregoing factors, we, conclude that the 2012 Plan was not narrowly tailored to achieve compliance with Section 5, and therefore fails strict scrutiny. III. REMEDY Having found that the 2012 Plan violates the Equal Protection Clause, we must now address the appropriate remedy. Plaintiffs seek to have us enjoin the use of the current congressional plan for the upcoming 2014 election, and to have the court draw an interim plan. Our consideration of this issue is guided by the Supreme Court’s decision in Reynolds v. Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964). In Reynolds, the Court stated: [0]nce a State’s legislative apportionment scheme has been found to be unconstitutional, it would be the unusual case in which a court would be justified in 'not taking appropriate action to [ensure] that no further elections are conducted under the invalid, plan. However, under certain circumstances, such as where an impending election is imminent and a State’s election machinery is already in progress, equitable considerations might justify a court in withhold- , ing the granting of immediately effective — 554 — relief in a legislative apportionment case, even though the existing apportionment scheme was found invalid. 377 U.S. at 585, 84 S.Ct. 1362. We must also be conscious of the powerful concerns for comity involved in interfering with the state’s legislative responsibilities. As the Supreme Court has repeatedly recognized, “redistricting and reapportioning legislative bodies is a legislative task which the federal courts should make every effort not to pre-empt.” Wise v. Lipscomb, 437 U.S. 535, 539-40, 98 S.Ct. 2493, 57 L.Ed.2d 411 (1978). As such, it is “appropriate, whenever practicable, to afford a reasonable opportunity for the legislature to meet constitutional requirements by adopting a substitute measure rather than for the federal court to devise ... its own plan.” Id. It should also be noted that courts have repeatedly allowed elections to proceed under unconstitutional apportionment plans when elections are imminent, or necessity so requires. See, e.g., Upham v. Seamon, 456 U.S. 37, 44, 102 S.Ct. 1518, 71 L.Ed.2d 725 (1982) (“[W]e have authorized District Courts to order or to permit elections to be held pursuant to apportionment plans that do not in all respects measure up to the legal requirements, even constitutional requirements.”); Kilgarlin v. Hill, 386 U.S. 120, 121, 87 S.Ct. 820, 17 L.Ed.2d 771 (1967) (affirming the district court’s decision allowing state legislative elections to proceed even though districting plan was “constitutionally infirm in certain respects”). With these considerations in mind, we determine that general principles of equity dictate that Virginia’s 2014 elections should proceed as scheduled under the challenged districting plans. This case presents precisely the “unusual” case referred to by the Reynolds Court, where competing interests weigh heavily against Plaintiffs’ equal protection rights. Delaying the elections or attempting to configure an interim districting plan would unduly disturb Virginia’s election process. The general election is certainly imminent, roughly two months away. Virginia’s primary elections were held as scheduled in June, candidates have spent significant time and money campaigning, and voters have begun to familiarize themselves with the candidates. Delaying the elections would cause significant and undue confusion. Moreover, Plaintiffs are largely responsible for the proximity of our decision to the November 2014 elections. Although the 2012 Plan was enacted in January 2012, Plaintiffs delayed bringing this action until October 2013, nineteen months later.24 (See Compl. [signature page], ECF No. 1). If we enjoined the general election, we could not limit the intrusion to the unconstitutional Third Congressional District. Any interim remedy would have to apply to all of Virginia’s electoral districts because we could not predict how repairing the Third Congressional District would alter the boundary lines of neighboring districts. On the other hand, we recognize that individuals in the Third Congressional District whose constitutional rights have — 555 — been injured by improper racial gerrymandering have suffered significant harm. Those citizens “are entitled to vote as soon as possible for their representatives under a constitutional apportionment plan.” Cos-ner v. Dalton, 522 F.Supp. 350, 364 (E.D.Va.1981). Therefore, we will require that new districts be drawn during Virginia’s next legislative session to remedy the unconstitutional districts. In accordance with well-established precedent that a state should have the first opportunity to create a constitutional redistricting plan, e.g., Wise, 437 U.S. at 539-40, 98 S.Ct. 2493, we allow the legislature until April 1, 2015, to enact a remedial districting plan. IV. CONCLUSION Because Plaintiffs have shown that race predominated in Virginia’s 2012 Plan, and because Defendants have failed to establish that this race-based redistricting satisfies strict scrutiny, we find that the 2012 Plan is unconstitutional, and will require the Commonwealth to act within the next legislative session to draw a new congressional district plan. It is so ORDERED. Footnotes [1] Because of Delegate Janis's key role as sponsor of the legislation at issue, we cite his views frequently. [2] As we discuss in greater detail below, in distinguishing the case before us from that in Shaw v. Hunt (Shaw II), 517 U.S. 899, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996), the dissent finds it significant that the legislative goal of maintaining minority voting strength in the Third Congressional District was not also articulated in the preclearance submission. With respect, we do not. [3] Named Plaintiff Dawn Curry Page was dismissed from this case via stipulation of dismissal on April 9, 2014. (ECF No. 79). [4] Original Defendants, the Virginia State Board of Elections and Kenneth T. Cuccinelli, II, Attorney General of Virginia, were dismissed from this case via stipulation of dismissal on November 21, 2013. (ECF No. 14). [5] Plaintiffs do not seek different remedies against Defendants and Intervenor-Defen-dants. Since there is no distinction between the interests of Defendants and Intervenor-Defendants, we refer to them collectively. [6] Defendants do not contend otherwise. Defendants make only limited narrow tailoring arguments, but do not assert that any kind of racial voting analysis informed their decisions. [7] Delegate Janis’s bill was renamed House Bill 251 but remained identical to the original House Bill 5004. [8] Named Plaintiffs are all United States citizens who are registered to vote in the Commonwealth of Virginia and reside in the Third Congressional District. (Compl. ¶¶ 7-9, ECF No. 1). [9] In contending that Plaintiffs do not make this ''initial” showing, the dissent notes, among other things, that Plaintiffs failed to produce an adequate alternative plan showing “that the legislature could have achieved its legitimate political objectives in alternative ways that are comparably consistent with traditional redistricting principles.” Cromartie II, 532 U.S. at 258, 121 S.Ct. 1452. While the dissent acknowledges "that the attacking party is not confined in its form of proof to submitting an alternative plan,” post at 49, it makes much of the fact that the alternative plan proffered by Plaintiffs accomplishes a more favorable result for Democrats than does the Enacted Plan. However, the significance of the discrepancy between these political outcomes is overstated, and relies on an — 542 — assumption that the legislature’s political objective was to create an 8-3 incumbency protection plan. See Trial Tr. 180-81 (noting that the Alternative Plan would only undermine incumbency protection objectives if it was the legislature’s political goal to have an 8-3 split, which is something "we don’t have knowledge of”). This inference is not supported by the record, as we develop more fully below. [10] The dissent relies solely on Justice O'Con-nor’s concurrence in Bush to make this argument. The language quoted by the dissent appears in the context of Justice O’Connor’s assertion that compliance with Section 2 of the VRA is a compelling state interest, see Bush, 517 U.S. at 990-93, 116 S.Ct. 1941 (O’Connor, Jr; concurring), but Justice O'Con-nor's opinion also specifically notes that using race as a proxy for VRA compliance should be subject to strict scrutiny, see id. [11] Perhaps this is also the appropriate juncture at which to address the dissent’s rejection of the credibility of Plaintiffs’ expert, Dr. Michael McDonald, and endorsement of De — 544 — fendants’ expert, Mr. Morgan, which we find somewhat puzzling. We find it no more damning that Dr. McDonald has testified differently in different contexts than that Mr. Morgan has testified consistently on the same side. Nor is the exploration of issues in an academic piece, written before Dr. McDonald was retained by Plaintiffs and before he fully evaluated the evidence here, of particular relevance. We do, however, find significant the following facts: that Mr. Morgan proffers no academic work, does not have an advanced degree, that his undergraduate degree was in history, that he has never taken a course in statistics, that he has not performed a racial bloc voting analysis, that he did not work with or talk to any members of the Virginia legislature, and that he miscoded the entire city of Petersburg’s VTDs. See Trial Tr. 334-35, 338-43, 361-65. [12] At this juncture, we must take issue with the manner in which the dissent considers Plaintiffs' circumstantial evidence. When evaluating evidence of the Third Congressional District’s shape, compactness, contiguity, political subdivision splits, and population swaps, the dissent considers each in isolation, concluding that no factor alone carries Plaintiffs’ burden of showing that race predominated. In addition, the dissent implies that Plaintiffs must, for each of these factors, make a "necessary showing" that these circumstantial irregularities, considered individually, resulted from racial, rather than political, motivations. See post at 34. Precedent counsels, — 545 — however, that courts must consider whether these circumstantial factors "together weigh in favor of the application of strict scrutiny.” Bush, 517 U.S. at 962, 116 S.Ct. 1941 (emphasis added). No one factor need be “independently sufficient” to show race predominated. Id. [13] This principle, contained in art. I, § 2 of the United States Constitution, requires all congressional districts to contain roughly equal populations. See Wesberry v. Sanders, 376 U.S. 1, 17, 84 S.Ct. 526, 11 L.Ed.2d 481 (1964). [14] A new district preserves district cores when it retains most of the previous bench — 547 — mark district’s residents within its boundaries. Trial Tr. 379. [15] The total population affected by the Third Congressional District’s locality splits with the Second Congressional District in the 2012 Plan is 241,096, while the population affected by the splits between these districts in the alternative plan is only 1,016. Trial Tr. 112; Pl.'s Trial Ex. 29, at 5, tbl. 3. [16] The fact that the 2012 Plan splits these cities, despite the demonstrated feasibility of achieving population parity while keeping them whole, further refutes the dissent’s contention that the population swaps were based on “a desire to limit locality splits.” Post at 36. Despite the fact that doing so was unnecessary, the legislature split Newport News and Hampton when it excluded certain low-BVAP VTDs from the Third Congressional District. See Pl.’s Trial Ex. 27, at 17 (showing that VTDs in Newport News and Hampton with BVAPs of 23.1% were excluded from the Third Congressional District). Similarly, the legislature's removal of predominantly white VTDs from the Third Congressional District contributed to otherwise unnecessary splits in Norfolk. See Trial Tr. 436-39. [17] Virginia attached the transcripts of these hearings to its Section 5 submission. See PL’s Trial Ex. 11-12. [18] We do not, as the dissent implies, suggest that a different legal test applies to a “mixed-motive suit.” We simply observe that, when racial considerations predominated in the redistricting process, the mere coexistence of race-neutral redistricting factors does not ' cure the defect. [19] Aside from the clear distinctions between Plaintiffs’ case here and Cromartie II, the dissent's contention that the legislature used BVAP as a predictor for Democratic votes is precisely the sort of race-based consideration the Supreme Court has confirmed triggers strict scrutiny. See Bush, 517 U.S. at 968, 116 S.Ct. 1941 C‘[T]o the extent that race is used as a proxy for political characteristics, a racial stereotype requiring strict scrutiny is in operation.”); Shaw I, 509 U.S. at 653, 113 S.Ct. 2816 ("[W]e unanimously reaffirmed that racial bloc voting and. minority-group political cohesion never can be assumed, but specifically must be proved____”). [20] Defendants' expert, Mr. Miorgan, contends that the majority-white populations excluded from the Third Congressional District during redistricting were predominantly Republican. Int.-Def.'s Trial Ex. 13, at 13-14. The evidence at trial, however, revealed that Mr. Morgan's analysis was based upon several pieces of mistaken data, a critical error. See Trial Tr. 359:1-14, 361:10-365:10 (indicating that Mr. Morgan had miscoded several VTDs as to whether they were part of the Third Congressional District); see also id. at 404:17-25 (Mr. Morgan’s coding mistakes were significant to the outcome of his analysis). [21] Plaintiffs make limited arguments that Section 5 compliance is no longer a compelling state interest. Plaintiffs first contend that Shelby County applies retroactively (See Pl.'s Trial Br. at 21-23, ECF No. 86), relying on Citizens United v. Fed. Election Comm’n, 558 U.S. 310, 130 S.Ct. 876, 175 L.Ed.2d 753 (2010), which held only that the Supreme Court’s decision that a particular interest does not qualify as a compelling state interest may have retroactive effect. The Supreme Court decided no such thing in Shelby County, so this assertion misses the mark. Plaintiffs also argue that compliance with Section 5 cannot be a compelling interest when the legislature conducted no analysis to determine whether an increase in the Third Congressional District’s BVAP was necessary, but this point is relevant only to the narrow tailoring prong of the strict scrutiny analysis. (See Pl.’s Trial Br. at 23-24, ECF No. 86; Pl.’s Post-Trial Br. at 30-31, ECF No. 105). [22] African-American voters accounted for over 90% of the voting age residents added to the Third Congressional District. Pl.’s Trial Ex. 27, at 14. [23] We pause to clarify that, while the legislature did not conduct a racial bloc voting analysis in enacting the 2012 Plan, we do not find that one is always necessary to support a narrow tailoring argument. [24] Plaintiffs have offered the intervening decision in Shelby County as an excuse for their delay. (See Pl.’s Br. Available Remedies 2, ECF No. 30 (stating that Plaintiffs brought their challenge “in the wake of the Supreme Court's recent ruling in Shelby County ”); Pl.'s Opp'n Def.’s Mot. Summ. J. 21-22, ECF No. 42 (arguing that courts are capable of undertaking mid-decennial redistricting when an intervening decision of the Supreme Court, such as Shelby County, establishes that a current plan is no longer valid)). Shelby County, however, dealt with Section 4 of the VRA, not Section 5, and therefore provides no support for their position.