From Casetext: Smarter Legal Research

Covington v. North Carolina

UNITED STATES DISTRICT COURT FOR THE MIDDLE DISTRICT OF NORTH CAROLINA
Jan 19, 2018
1:15CV399 (M.D.N.C. Jan. 19, 2018)

Opinion

1:15CV399

01-19-2018

SANDRA LITTLE COVINGTON, et al., Plaintiffs, v. THE STATE OF NORTH CAROLINA, et al., Defendants.


Before WYNN, Circuit Judge, and SCHROEDER, Chief District Judge, and EAGLES, District Judge. MEMORANDUM OPINION AND ORDER PER CURIAM:

On August 11, 2016, this Court held that the North Carolina General Assembly unjustifiably relied on race to draw dozens of state Senate and House of Representatives district lines, in violation of the Equal Protection Clause of the Fourteenth Amendment. Covington v. North Carolina (Covington I), 316 F.R.D. 117 (M.D.N.C. 2016). The Supreme Court summarily affirmed, without dissent, that determination. North Carolina v. Covington, 137 S. Ct. 2211 (2017) (mem.).

On August 31, 2017, the North Carolina General Assembly enacted Senate and House redistricting plans (the "2017 Plans") intended to remedy the constitutional violations. Plaintiffs, thirty-one North Carolina voters, lodged objections to 12 of the 116 proposed remedial districts, arguing that those districts failed to remedy the identified racial gerrymanders or were otherwise legally unacceptable. Finding 9 of Plaintiffs' 12 objections potentially had merit, this Court appointed Dr. Nathaniel Persily of Stanford University as Special Master (the "Special Master") to assist the Court in redrawing those 9 district configurations (the "Subject Districts"). Thereafter, the Special Master filed draft reconfigurations of the 9 districts for the parties' consideration, invited and considered comments and objections from the parties, and revised his draft plan in light of those comments and objections.

On December 1, 2017, the Special Master submitted to the Court recommended remedial plans (the "Recommended Plans") for the Subject Districts, as well as a report explaining his process for drawing the Recommended Plans and why the Recommended Plans remedy the identified legal problems with the Subject Districts. As further explained below, after careful consideration of the 2017 Plans, the Special Master's report, and the parties' evidence, briefing, and oral arguments, we sustain Plaintiffs' objections to the Subject Districts, approve the Special Master's Recommended Plans for reconfiguring those districts, reject Plaintiffs' challenge to one Senate district, and decline to consider Plaintiffs' remaining objections.

Plaintiffs do not lodge any objections to the remaining 104 districts redrawn in the 2017 Plans, and therefore, we have nothing before us that indicates the districts do not comply with our order.

I.

In early 2011, the North Carolina General Assembly set out to redraw state Senate and House districts to account for changes in population and demographic data revealed in the most recent decennial census. See N.C. Const. art. II, §§ 3, 5. As the appointed chairs of the redistricting committees in their respective chambers, Senator Robert Rucho and Representative David Lewis (collectively, the "Chairs"), both Republicans, led efforts to draw and enact legislative districting maps for use in state elections in North Carolina (the "2011 Plans"). Covington I, 316 F.R.D. at 126. To that end, Representative Lewis and Senator Rucho engaged the assistance of an outside expert, Dr. Thomas Hofeller, to draw the new Senate and House district maps. Id.

Senator Rucho and Representative Lewis instructed Dr. Hofeller to follow three "primary" criteria in drawing the new districting plans, all of which "centered around the creation of what the Chairs called 'VRA districts'"—geographically compact minority population centers for which there was some evidence of a history of racially polarized voting. Id. at 130. The first criterion required that Dr. Hofeller "draw all purported VRA districts to reach a 50%-plus-one [Black Voting Age Population ("BVAP")] threshold." Id. This instruction stemmed from Senator Rucho's and Representative Lewis's belief that the Supreme Court's plurality opinion in Bartlett v. Strickland, 556 U.S. 1 (2009), required that any district drawn to comply with the Voting Rights Act be majority-minority. Id.

Second, Senator Rucho and Representative Lewis directed Dr. Hofeller to draw the so-called "VRA districts" first. Id. at 131. This instruction derived from the North Carolina Supreme Court's opinions in Stephenson v. Bartlett (Stephenson I), 562 S.E.2d 377 (N.C. 2002) and Stephenson v. Bartlett (Stephenson II), 582 S.E.2d 247 (N.C. 2003), both of which sought to harmonize federal election law with the North Carolina Constitution's so-called "Whole County Provision," N.C. Const. art. II, §§ 3(3), 5(3), which requires that, where possible, legislative district lines adhere to county lines, Covington, 316 F.R.D. at 131-32. According to the Chairs, the Stephenson decisions required Dr. Hofeller to identify and draw any VRA districts first. Id.

Third, Senator Rucho and Representative Lewis instructed Dr. Hofeller to draw VRA districts "everywhere there was a minority population large enough to do so and, if possible, in rough proportion to their population in the state." Id. at 130. This instruction again derived from the Chairs' incorrect understanding of governing law. In particular, Senator Rucho and Representative Lewis errantly believed that the Supreme Court's decision in Johnson v. De Grandy, 512 U.S. 999 (1994), held that in order to comply with Section 2 of the Voting Rights Act, the number of majority-minority districts in a state must be proportional to minority voters' share of the state's overall voting population. Covington, 316 F.R.D. at 133. Although the Chairs did not expressly instruct Dr. Hofeller to maximize the number of VRA districts, "the proportionality target functionally operated as a goal to maximize the number of majority-black districts." Id. at 134.

Senator Rucho and Representative Lewis further instructed Dr. Hofeller that any districting proposal had to comply with these three "primary" criteria, two of which—the 50%-plus-one target and the proportionality goal—amounted to "'mechanical racial targets.'" Id. at 135 (quoting Ala. Legislative Black Caucus v. Alabama, 135 S. Ct. 1257, 1267 (2015)). In accordance with Senator Rucho's and Representative Lewis's instructions, Dr. Hofeller first "drew VRA 'exemplar districts,' which were 'racially defined' in that they embodied nothing more than 'concentrations of minority voters' capable of constituting a district that could satisfy the 50%-plus-one BVAP threshold." Id. at 135 (quoting Trial Tr. vol. IV, 228:5-12 (Hofeller); Trial Tr. vol. V, 104:4-105:1 (Hofeller)). By drawing, where feasible, district lines around the black population centers identified in the "exemplar districts," Dr. Hofeller then constructed as many majority-black districts as possible. Id. at 136-37.

Because the Chairs had instructed Dr. Hofeller that the three "primary" criteria could not be compromised, in drawing the districting plans Dr. Hofeller subordinated other race-neutral districting principles such as preserving political subdivisions and communities of interest, compactness, and complying with state districting laws such as the Whole County Provision. Id. at 137-39. As a result of the decision to adhere to the Chairs' mechanical racial targets over traditional race-neutral districting principles, the number of majority-black districts in Dr. Hofeller's proposed state House map increased from nine to thirty-two. Id. at 126, 134, 137. Similarly, the number of majority-black districts in the proposed state Senate map increased from zero to nine. Id. at 126. The state Senate and House considered and adopted, with minor modifications, the 2011 Plans on July 27 and 28, 2011, respectively. Id.

Soon after the General Assembly approved the 2011 Plans, North Carolina voters filed actions in state court alleging that the lines of numerous legislative districts enacted by the General Assembly amounted to unconstitutional racial gerrymanders, in violation of the North Carolina and United States Constitutions. See Dickson v. Rucho, 766 S.E.2d 238 (N.C. 2014), vacated, 135 S. Ct. 1843 (2015) (mem.). A divided Supreme Court of North Carolina held that both the Senate and House districting plans satisfied all "state and federal constitutional and statutory requirements." Dickson, 766 S.E.2d at 260. In April 2015, the Supreme Court of the United States unanimously vacated the state court's ruling without opinion and remanded the case for reconsideration of the federal constitutional and statutory questions presented in light of the Supreme Court's recent decision in Alabama Legislative Black Caucus. Dickson, 135 S. Ct. 1843. On remand, the Supreme Court of North Carolina again concluded that the 2011 Plans complied with federal law. Dickson v. Rucho, 781 S.E.2d 404 (N.C. 2015), vacated, 137 S. Ct. 2186 (2017).

While litigation in state court continued, Plaintiffs initiated this action in May 2015. Covington I, 316 F.R.D. at 128. As in the ongoing state court action, Plaintiffs alleged that districts in the 2011 Plans constituted racial gerrymanders and thus violated the Fourteenth Amendment of the U.S. Constitution. First Am. Compl. at 2, July 24, 2015, ECF No. 11. To remedy the alleged constitutional violation, Plaintiffs sought an injunction barring further use of the challenged districts in the 2011 Plans and requiring the General Assembly to adopt constitutionally compliant plans for use in any future elections. Id. at 92-93. Plaintiffs named as Defendants: (1) the State of North Carolina; (2) Senator Rucho, Representative Lewis, President Pro Tempore of the North Carolina Senate Philip E. Berger, and Speaker of the North Carolina House of Representatives Timothy K. Moore (collectively, the "Legislative Defendants"); and (3) the North Carolina State Board of Elections, as well as each of the five members of that body (collectively, the "Board Defendants").

On August 11, 2016, this Court unanimously concluded that Defendants unjustifiably, and therefore unconstitutionally, predominantly relied on race in drawing the lines of twenty-eight majority-minority districts in the 2011 Plans. Covington I, 316 F.R.D. at 176. In particular, this Court concluded that Defendants lacked a "strong basis in evidence" for their belief that race-based districting was necessary to comply with Section 2 of the Voting Rights Act because Defendants never analyzed whether, for each challenged district, the presence of "racial bloc voting . . . would enable the majority usually to defeat the minority group's candidate of choice." Id. at 167 (citing Thornburg v. Gingles, 478 U.S. 30, 51 (1986)). On June 5, 2017, the Supreme Court summarily affirmed, without dissent, this Court's judgment that the Senate and House districting plans violated Plaintiffs' rights under the Fourteenth Amendment. Covington, 137 S. Ct. 2211. Notwithstanding that this Court had found the district lines violated the Constitution in August 2016 and that the Supreme Court affirmed that conclusion in early June 2017, the General Assembly made no effort to begin drawing remedial districting plans until late July 2017.

After obtaining jurisdiction from the Supreme Court, this Court received evidence, briefing, and argument regarding the appropriate remedy for the constitutional violations. In an order entered on July 31, 2017, this Court gave the General Assembly until September 1, 2017, "to enact new House and Senate districting plans remedying the constitutional deficiencies" with the districts found unconstitutional in this Court's August 2016 opinion and order. Covington v. North Carolina (Covington III), --- F. Supp. 3d. ---, 2017 WL 3254098, at *3 (M.D.N.C. 2017). This Court advised that it would extend this deadline until September 15, 2017, if the General Assembly made certain showings regarding the public nature of its redistricting process. Id. The order further explained that the Court selected the September deadlines to ensure that it would have adequate time "(1) to review the General Assembly's enacted remedial district plans, and (2) if the enacted plans prove constitutionally deficient, to draw and impose its own remedial plan." Id. In the same order, and as further explained in a subsequent opinion, this Court denied Plaintiffs' request for a special election. Id. at *2; see also Covington v. North Carolina (Covington IV), --- F. Supp. 3d. ---, 2017 WL 4162335 (M.D.N.C. 2017).

Electing not to make the public showings necessary to obtain an extension of the deadline, the General Assembly's Senate Redistricting Committee and House Select Committee on Redistricting (collectively, the "Joint Committee") put in place a streamlined process designed to ensure enactment of remedial plans in advance of the September 1, 2017 deadline. Representative Lewis and Senator Ralph Hise, who had replaced Senator Rucho as chair of the Senate Redistricting Committee, again engaged Dr. Hofeller to assist the Joint Committee's Republican supermajority in drawing the remedial maps.

The Joint Committee met on August 10, 2017, during which Representative Lewis and Senator Hise proposed the following criteria to govern the drawing of the remedial district plans:

Equal Population. The Committees shall use the 2010 federal decennial data as the sole basis of population for drawing legislative districts in the 2017 House and Senate plans. The number of persons in each legislative
district shall comply with the +/- 5 percent population deviation standard established [Stephenson I].

Contiguity. Legislative districts shall be comprised of contiguous territory. Contiguity by water is sufficient.

County Groupings and Traversals. The Committees shall draw legislative districts within county groupings as required by [Stephenson I, Stephenson II, Dickson I, and Dickson II]. With county groupings, county lines shall not be traversed except as authorized by Stephenson I, Stephenson II, Dickson I, and Dickson II.

Compactness. The Committees shall make reasonable efforts to draw legislative districts in the 2017 House and Senate plans that improve the compactness of the current districts. In doing so, the Committees may use as a guide the minimum Reock ("dispersion") and Polsby-Popper ("perimeter") scores identified by Richard H. Pildes and Richard G. Neimi in Expressive Harms, "Bizarre Districts," and Voting Rights: Evaluating Election-District Appearances After Shaw v. Reno, 92 Mich. L. Rev. 483 (1993).

Fewer Split Precincts. The Committees shall make reasonable efforts to draw legislative districts in the 2017 House and Senate plans that split fewer precincts than the current legislative redistricting plans.

Municipal Boundaries. The Committees may consider municipal boundaries when drawing legislative districts in the 2017 House and Senate plans.

Incumbency Protection. Reasonable efforts and political considerations may be used to avoid pairing incumbent members of the House or Senate with another incumbent in legislative districts drawn in the 2017 House and Senate plans. The Committees may make reasonable efforts to ensure voters have a reasonable opportunity to elect non-paired incumbents of either party to a district in the 2017 House and Senate plans.

Election Data. Political considerations and election results data may be used in the drawing of legislative districts in the 2017 House and Senate plans.

No Consideration of Racial Data. Data identifying the race of individuals or voters shall not be used in the drawing of legislative districts in the 2017 House and Senate plans.
Adopted Criteria for House and Senate Plans, Sept. 7, 2017, ECF No. 184-37.

During the hearing, Democratic members of the Joint Committee objected to the Incumbency Protection criterion as likely to perpetuate the effects of the racial gerrymander by protecting incumbents elected under the racially gerrymandered plans. See, e.g., Joint Select Comm. On Redistricting Meeting Tr. 120:9-121:9, Aug. 10, 2017, ECF No. 184-9 ("[I]t seems just ridiculous to me that [the Republican majority] would get to now say we get to protect the members that we were able to elect using unconstitutional maps."). Likewise, Democratic Joint Committee members expressed concern with the "Election Data" criterion on grounds that the purpose of using such data was unclear and that such data would be used to preserve the partisan makeup of the two chambers achieved under the unconstitutional districting plans. See, e.g., id. at 134:13-139:2. In the course of the discussion on the use of Election Data, Representative Lewis represented that the Joint Committee's Republican leadership did not "have a goal of maintaining the current partisan advantage in the House and the Senate." Id. at 138:15-21. And Democratic Joint Committee members objected to the criterion barring consideration of "racial data" on grounds that it was necessary to consider such data to determine whether remedial plans remedied the racial gerrymander. See, e.g., id. at 151:6-11 ("[I]f the districts were declared unconstitutional because of race, if you don't use race to correct it, how are you going to show the Court that they still are not unconstitutional.").

The Joint Committee unanimously adopted the Equal Population and County Groupings and Traversal criteria. Leg. Defs.' Resp. to Pls.' Objs. ("Leg. Defs.' Objs. Resp.") 8-10, Sept. 22, 2017, ECF No. 192. The remaining seven criteria were adopted by party-line votes. Id. Representative Lewis and Senator Hise directed Dr. Hofeller to follow the adopted criteria in drawing the remedial maps, but the Committee provided Dr. Hofeller with no formal guidance as to the relative precedence of the various criteria. House Select Comm. On Redistricting Meeting Tr. 62:4-6, Aug. 25, 2017, ECF No. 184-18. Legislative Defendants did not introduce any evidence regarding what additional instructions, if any, Representative Lewis or Senator Hise provided to Dr. Hofeller about the proper use and weighting of the various criteria. Nor did they offer any evidence as to how Dr. Hofeller weighted or ordered the criteria in drawing the proposed remedial maps, either in general or as to any particular district.

The General Assembly released Dr. Hofeller's proposed Senate and House Plans on August 19 and 20, 2017, respectively. The General Assembly provided block assignment files and statistical information regarding the 2017 Plans on August 21, 2017. The 2017 Plans altered a total of 116 of the 170 state House and Senate districts. On August 22, 2017, the Joint Committee held a public hearing on the proposed plans in Raleigh, allowing attendees at six satellite locations to participate via teleconference. The Committees also received thousands of public comments through the General Assembly's website.

On August 23, 2017, Plaintiffs sent a letter to the House Select and Senate Committees on Redistricting and Defendants' counsel raising the following objections to the 2017 Plans: (1) several of the proposed districts failed to remedy the racial gerrymander; (2) the plans, when analyzed as a whole, amounted to "grossly unconstitutional partisan gerrymanders" in violation of the Equal Protection Clause; (3) the House plan's reconfiguration of certain districts in Mecklenburg and Wake County untainted by the racial gerrymander violated the North Carolina Constitution's prohibition on mid-decade redistricting; and (4) proposed district configurations in Cabarrus and Greene Counties violated the North Carolina Constitution's requirement that, where possible, state legislative districts respect county lines. Letter to Counsel, Sept. 15, 2017, ECF No. 187-1. Plaintiffs also provided the Committees with alternative maps that addressed Plaintiffs' objections, and Democratic representatives offered those maps as amendments during the legislative process.

The Committees did not revise the proposed remedial plans to address Plaintiffs' objections and rejected Plaintiffs' alternative redistricting plans. By party-line vote, the Senate Redistricting Committee approved Dr. Hofeller's proposed Senate plan on August 24, 2017. The House Redistricting Committee approved Dr. Hofeller's proposed House plan on August 25, 2017, also by a party-line vote. The General Assembly adopted, with minor modifications, both 2017 Plans on August 31, 2017.

One week later, Legislative Defendants filed with this Court the 2017 Plans and supporting data and materials required by the Court's July 31 order, including the complete legislative record. On September 15, 2017, Plaintiffs filed objections to 12 of the 116 redrawn districts, alleging essentially the same violations that they had identified in their August 23, 2017 letter to Defendants and the Committees. Objs. ("Pls.' Objs."), Sept. 15, 2017, ECF No. 187. Along with their objections, Plaintiffs filed several supporting records, affidavits, and expert analyses. One week later, Legislative Defendants responded to Plaintiffs' objections, asserting that this Court was without jurisdiction to consider the objections and that the objections otherwise were without merit. See generally Leg. Defs.' Objs. Resp. The State of North Carolina and Board Defendants (collectively, the "State Defendants") took no position on Plaintiffs' objections.

On October 12, 2017, this Court held a hearing on Plaintiffs' objections. This Court gave Legislative Defendants the opportunity to introduce evidence—in addition to the legislative record, data, and other materials submitted in accordance with the Court's July 31, 2017 order—and present witnesses to establish that the General Assembly's proposed remedial plans cured the identified constitutional violations and were not otherwise legally unacceptable. Legislative Defendants elected not to offer any such evidence, either in written submissions or at the hearing.

That same day, the Court issued an order directing the parties to confer and, if possible, jointly submit a list of three persons qualified to serve as a special master under Federal Rule of Civil Procedure 53 to assist the Court in its remedial efforts. Order, Oct. 12, 2017, ECF No. 200. The order further stated that if the parties failed to reach an agreement as to a list of candidates, the Court would select a special master. Id. The parties subsequently informed the Court that they had conferred but failed to reach an agreement as to the requested list of special master candidates. Notice, Oct. 18, 2017, ECF No. 201.

On October 26, 2017, the Court informed the parties that, after carefully considering Plaintiffs' objections, it was concerned that nine district configurations in the 2017 Plans either failed to remedy the identified constitutional violations or were otherwise legally unacceptable. Order, Oct. 26, 2017, ECF No. 202. The Court further informed the parties that in light of its concerns, it intended to appoint Dr. Nathaniel Persily of Stanford University as Special Master to assist the Court by drawing alternative remedial districting plans. Id. The Court gave the parties an opportunity to object to the appointment of Dr. Persily. Id. Pursuant to the Court's invitation, Legislative Defendants objected to the appointment of a special master and Dr. Persily, in particular, but they did not identify any alternative candidate to serve as special master. Obj., Oct. 30, 2017, ECF No. 204.

In a November 1, 2017 order, the Court overruled Legislative Defendants' objections and appointed Dr. Persily as Special Master. Order ("Appointment Order"), Nov. 1, 2017, ECF No. 206. The Appointment Order described the Court's concerns with the Subject Districts and set forth the scope of the Special Master's responsibilities. Id. The Appointment Order also directed the Special Master to adhere to the following guidelines in redrawing Subject Districts:

a. Redraw district lines for [2011 Enacted Senate Districts 21 and 28 and House Districts 21, 33, 38, 57, 99, 102, and 107] and any other districts within the applicable 2017 county grouping necessary to cure the unconstitutional racial gerrymanders. As to House District 57, the redrawn lines shall also ensure that the unconstitutional racial gerrymanders in 2011 Enacted House Districts 58 and 60 are cured. As to 2011 Enacted House Districts 33, 38, 99, 102, and 107, no 2011 Enacted House Districts which do not adjoin those districts shall be redrawn unless it is necessary to do so to meet the mandatory requirements set forth in Paragraphs 2(b) through 2(e) of this Order, and if the Special Master concludes that it is necessary to adjust the lines of a non-adjoining district, the Special Master shall include in his report an explanation as to why such adjustment is necessary.
b. Use the 2010 Federal Decennial Census Data.

c. Draw contiguous districts with a population as close as possible to 79,462 persons for the House Districts and 190,710 persons for the Senate Districts, though a variance up to +/- 5% is permitted and authorized if it would not conflict with the primary obligations to ensure that remedial districts remedy the constitutional violations and otherwise comply with state and federal law, would enhance compliance with state policy as set forth in subsection (f) below, and would not require redrawing lines for an additional district.

d. Adhere to the county groupings used by the General Assembly in the 2017 Enacted Senate and House Plans.

e. Subject to any requirements imposed by the United States Constitution or federal law, comply with North Carolina constitutional requirements including, without limitation, the Whole County Provision as interpreted by the North Carolina Supreme Court.

f. Make reasonable efforts to adhere to the following state policy objectives, so long as adherence to those policy objectives does not conflict with the primary obligations of ensuring that remedial districts remedy the constitutional violations and otherwise comply with state and federal law:

i. Split fewer precincts than the 2011 Enacted Districts;

ii. Draw districts that are more compact than the 2011 Enacted Districts, using as a guide the minimum Reock ("dispersion") and Polsby-Popper ("perimeter") scores identified by Richard Pildes & Richard Neimi, Expressive Harms, "Bizarre Districts," and Voting Rights: Evaluating Election-District Appearances After Shaw v. Reno, 92 Mich. L. Rev. 483 (1993); and

iii. Consider municipal boundaries and precinct lines.

g. After redrawing the districts, in view of the policy decision by the General Assembly that efforts to avoid pairing incumbents are in the interest of North Carolina voters, the Special Master may adjust district lines to avoid pairing any incumbents who have not publicly announced their intention not to run in 2018, but only to the extent
that such adjustment of district lines does not interfere with remedying the constitutional violations and otherwise complying with federal and state law. Additionally, the Special Master shall treat preventing the pairing of incumbents as "a distinctly subordinate consideration" to the other traditional redistricting policy objectives followed by the State. Ga. State Conf. of NAACP v. Fayette Cty. Bd. of Comm'rs, 996 F. Supp. 2d 1353, 1363 (N.D. Ga. 2014) (collecting cases).

h. Except as authorized in Paragraph 2(g), the Special Master shall not consider incumbency or election results in drawing the districts. See, e.g., Wise v. Lipscomb, 437 U.S. 535, 541 (1978) (noting that courts lack "political authoritativeness" and must act "in a manner free from any taint of arbitrariness or discrimination" in drawing remedial districts) (quoting Connor v. Finch, 431 U.S. 408, 417 (1977)); Wyche v. Madison Par. Police Jury, 769 F.2d 265, 268 (5th Cir. 1985) ("Many factors, such as the protection of incumbents, that are appropriate in the legislative development of an apportionment plan have no place in a plan formulated by the courts."); Wyche v. Madison Par. Police Jury, 635 F.2d 1151, 1160 (5th Cir. 1981) (noting that "a court is forbidden to take into account the purely political considerations that might be appropriate for legislative bodies"); Favors v. Cuomo, Docket No. 11-cv-5632, 2012 WL 928216, at *18 (E.D.N.Y. Mar. 12, 2012), report and recommendation adopted as modified, No. 11-cv-5632, 2012 WL 928223, at *6 (E.D.N.Y Mar. 19, 2012); Molina v. Cty. of Orange, No. 13CV3018, 2013 WL 3039589, at *8 (S.D.N.Y. June 3, 2013), supplemented, No. 13CV3018, 2013 WL 3039741 (S.D.N.Y. June 13, 2013), report and recommendation adopted, No. 13 CIV. 3018 ER, 2013 WL 3009716 (S.D.N.Y. June 14, 2013); Larios v. Cox, 306 F. Supp. 2d 1214, 1218 (N.D. Ga. 2004); Balderas v. Texas, No. 6:01CV158, 2001 WL 36403750, at *4 (E.D. Tex. Nov. 14, 2001).

i. The Special Master may consider data identifying the race of individuals or voters to the extent necessary to ensure that his plan cures the unconstitutional racial gerrymanders and otherwise complies with federal law.
Id. The Appointment Order further directed the Special Master to submit to the Court by December 1, 2017, a report that included reconfigured districting plans for each of the Subject Districts, an explanation of those plans, and a comparison of those plans with the related districts in the 2017 Plans and districts submitted by Plaintiffs. Id.

Pursuant to the Court's Appointment Order, the Special Master immediately set out to draw new configurations for the Subject Districts. On November 14, 2017, the Special Master disclosed to the parties and filed with the Court draft reconfigurations of the Subject Districts as well as an explanation of his rationale behind those reconfigurations. Special Master's Draft Plan and Order, Nov. 14, 2017, ECF No. 213. In accordance with the Court's Appointment Order, the Special Master's draft plan made no effort to avoid pairing incumbents. Id. at 4. Rather, the Special Master ordered the parties to submit objections and proposed revisions to the draft plan, including suggestions "as to how incumbents shall be unpaired without degrading the underlying features of the [draft] plan." Id. at 19.

Pursuant to the Special Master's order, Plaintiffs submitted comments on the Special Master's draft plan on November 17, 2017, stating, inter alia, that they believed the draft plan remedied the constitutional flaws with the subject districts. Pls.' Resp. & Proposed Modifications to the Special Master's Draft Plan, Nov. 17, 2017, ECF No. 216. Plaintiffs further suggested several approaches the Special Master could take in revising his draft plans to avoid pairing incumbents in some, but not all, of the reconfigured districts. Id.

By contrast, Legislative Defendants elected not to raise any objection to specific aspects of the Special Master's draft plan or offer suggestions as to how the Special Master could improve his draft plan or avoid pairing incumbents, representing that they lacked authority under State law to advise the Special Master on the drawing of remedial districts. Leg. Defs.' Response to Special Master's Draft Rep. ("Leg. Defs.' Draft Rep. Resp.") 5, Nov. 17, 2017, ECF No. 215 (explaining that "the legislative defendants do not themselves speak for the entire General Assembly" and therefore that "[a] few members of the legislature, even if they are leaders, are not authorized to state how the entire legislature would vote on, or amend, draft districts proposed by a law professor"). Rather than offering any substantive comments or suggestions regarding the Special Master's draft plan, Legislative Defendants elected to renew their objections to this Court's jurisdiction and the Special Master's authority to draw remedial districts. See generally id.

In response, Plaintiffs asserted that Legislative Defendants' jurisdictional arguments were without merit. Pls.' Resp. to Leg. Defs.' Nov. 17, 2017 Filing, Nov. 21, 2017, ECF No. 217. The Legislative Defendants then objected to Plaintiffs' suggestions for unpairing incumbents on grounds that the suggestions served to benefit Democratic candidates and recommended that the Special Master advise the Court to adopt the General Assembly's 2017 Plans in full, rather than his proposed remedial plans. Leg. Defs.' Resp. to Pls.' Proposed Modifications to Special Master's Draft Plan, Nov. 21, 2017, ECF No. 218.

On December 1, 2017, after receiving comments and suggestions from the parties, the Special Master filed with this Court his Recommended Plan and Report and numerous supporting materials. Special Master's Rec. Plan & Rep. ("Rec. Plan & Rep."), Dec. 1, 2017, ECF No. 220. In his 69-page report, the Special Master presented his Recommended Plans for the Subject Districts and thoroughly explained how those configurations conformed to the Court's guidelines and advanced traditional redistricting criteria; described how the Recommended Plans addressed the Court's concerns with the Subject Districts and cured the constitutional violations with the related districts in the 2011 Plans; explained why his remedial configurations were superior to those proposed by Plaintiffs; and offered alternative configurations to address several potential concerns with his Recommended Plans. See generally id. Notwithstanding that Legislative Defendants elected not to suggest how incumbents should be unpaired—and categorically objected to Plaintiffs' suggestions for unpairing certain incumbents—the Special Master's Recommended Plans avoids pairing all but two of the incumbents—one Republican and one Democrat—in his reconfigured districts and did not pair any incumbents of the same party. Id. at 30, 37.

On December 8, 2017, Plaintiffs notified the Court that they had no objections to the Special Master's Recommend Plan. Pls.' Pos. on the Special Master's Recommended Plan, Nov. 8, 2017, ECF No. 223. That same day, Legislative Defendants filed with the Court numerous objections to the Special Master's Recommended Plan and Report, Leg. Defs.' Resp. to Special Master's Recommended Plan & Report ("Leg. Defs.' Rec. Plan Resp."), Nov. 8, 2017, ECF No. 224, notwithstanding that Legislative Defendants had previously represented that they lacked authority under state law to comment on or provide suggestions regarding the Special Master's reconfigurations, Leg. Defs.' Draft Rep. Resp. 5. Legislative Defendants maintained that the Recommended Plans "reveal[] the [S]pecial Master's single-minded focus on race" and that the recommended districts, if adopted by the Court, would "impose on the State a racial gerrymander that favors one political party." Leg. Defs.' Rec. Plan Resp. at 2-3. Although Legislative Defendants had offered no substantive suggestions to the Special Master regarding his earlier draft plan, Legislative Defendants raised several district-specific objections to the Recommended Plans and argued that the 2017 Plans were superior to the Recommended Plans. Id. at 8-17. Finally, Legislative Defendants objected to the Special Master's unpairing of Democratic incumbents, but appeared to acquiesce in the Special Master's unpairing of Republican incumbents. Id. at 20 ("The special master agreed to allow plaintiffs' requests and submitted a final plan that un-pairs numerous Democratic incumbents, even where doing so required him to make changes to his draft districts in a way that did not improve the scoring of the districts under traditional redistricting principles.").

On January 5, 2017, the Court held a hearing during which the Special Master presented his Recommended Plans and addressed numerous questions raised by the parties. At the hearing, Legislative Defendants also introduced expert and testimonial evidence pertaining to alleged infirmities with the Recommended Plans. Having carefully reviewed the 2017 Plans; the Special Master's Recommended Plan and Report, and the materials appended thereto; and the parties' evidence, briefing, and oral arguments, we sustain Plaintiffs' objections to the Subject Districts and approve and adopt the Special Master's Recommended Plans for reconfiguring those districts.

II.

Before addressing the merits of Plaintiffs' objections to certain districts in the 2017 Plans, including the Subject Districts, we first must address several threshold arguments made by Legislative Defendants, which seek to circumscribe the scope of this Court's review of the General Assembly's proposed 2017 Plans. In particular, Legislative Defendants argue that: (1) the enactment of the 2017 Plans rendered this action moot; (2) this Court's review of the 2017 Plans extends, at most, to determining whether the plans corrected the racial gerrymander; (3) this Court lacks jurisdiction under the three-judge panel statute to consider any of Plaintiffs' objections other than the racial gerrymandering allegations that initially served as the basis of this panel's jurisdiction; and (4) this Court may not, as a matter of federalism, consider Plaintiffs' state law objections. We address each of these arguments in turn.

A.

Legislative Defendants first contend that the General Assembly's enactment of the new districting plans rendered this case moot. Leg. Defs.' Objs. Resp. 19-21. In particular, Legislative Defendants argue that because the districting plans that served as the basis of Plaintiffs' challenge have been replaced, "[P]laintiffs no longer have a concrete stake in the outcome of the case." Id. at 20. This argument is without merit.

The Supreme Court long has held that when a federal court concludes that a state districting plan violates the Constitution, the appropriate state redistricting body should have the first opportunity to enact a plan remedying the constitutional violation. Reynolds v. Sims, 377 U.S. 585, 586 (1964). But after finding unconstitutional race-based discrimination—as this Court did here—a district court also has a "duty" to ensure that any remedy "so far as possible eliminate[s] the discriminatory effects of the past as well as bar[s] like discrimination in the future." Louisiana v. United States, 380 U.S. 145, 154 (1965); see also, e.g., Lane v. Wilson, 307 U.S. 268, 275 (1939) (holding invalid State's proposed remedy for state constitutional provision that violated the Fifteenth Amendment because it "part[ook] too much of the infirmity" of the original constitutional provision). To that end, if the state fails to enact "a constitutionally acceptable" remedial districting plan, then "the responsibility falls on the District Court." Chapman v. Meier, 420 U.S. 1, 27 (1975); see also Reynolds, 377 U.S. at 586 (holding that a district court "acted in a most proper and commendable manner" by imposing its own remedial districting plan, after the district court concluded that remedial plan adopted by state legislature failed to remedy constitutional violation).

In accordance with Chapman and Reynolds, the U.S. Court of Appeals for the Fourth Circuit has held that when, as here, a state enacts a redistricting plan in an effort to remedy a constitutional violation, a district court must "consider whether the proffered remedial plan is legally unacceptable because it violates anew constitutional or statutory voting rights—that is, whether it fails to meet the same standards applicable to an original challenge of a legislative plan in place." McGhee v. Granville Cty., N.C., 860 F.2d 110, 115 (4th Cir. 1988). Numerous other courts have reached the same conclusion—federal courts must review a state's proposed remedial districting plan to ensure it completely remedies the identified constitutional violation and is not otherwise legally unacceptable. See, e.g., Large v. Fremont Cty., Wyo., 670 F.3d 1133, 1138, 1148-49 (10th Cir. 2012) (rejecting governmental entity's proposed districting plan to remedy Voting Rights Act violation because it failed to comply with state law); Ketchum v. Byrne, 740 F.2d 1398, 1411-12 (7th Cir. 1984) (rejecting governmental entity's proposed remedial districting plan because it failed to completely remedy Voting Rights Act violation); Williams v. City of Texarkana, Ark., 32 F.3d 1265, 1268 (8th Cir. 1994) ("If an appropriate legislative body offers a remedial plan, the court must defer to the proposed plan unless the plan does not completely remedy the violation or the proposed plan itself constitutes a . . . violation [of the Voting Rights Act]." (emphasis added)); Harris v. McCrory, No. 1:13-cv-949, 2016 WL 3129213, at *2 (M.D.N.C. June 2, 2016) (holding, in racial gerrymandering case, that a district court "must determine whether the legislative remedy enacted at its behest is in fact a lawful substitute for the original unconstitutional plan"); United States v. Osceola Cty., Fla., 474 F. Supp. 2d 1254, 1258 (M.D. Fla. 2006) (rejecting governmental body's remedial districting plan because it was "not a full and adequate remedy" of the identified Voting Rights Act violation).

Additionally, we emphasize that the General Assembly redrew the Subject Districts pursuant to the opportunity provided by this Court's order to "enact new House and Senate districting plans remedying the constitutional deficiencies." Covington III, 2017 WL 3254098, at *3. It is axiomatic this Court has the inherent authority to enforce its own orders. See, e.g., Carlisle v. United States, 517 U.S. 416, 438 (1996) (noting that "[e]xamples of the exercise of the federal courts' inherent powers are abundant in both our civil and our criminal jurisprudence" and collecting cases); see also Degen v. United States, 517 U.S. 820, 827 (1996); Spagnuolo v. Whirlpool Corp., 717 F.2d 114, 122 (4th Cir. 1983). This is especially so here, given that the state constitution prohibited the General Assembly from engaging in mid-decade redistricting absent this Court's order. Thus, this Court has a strong interest in ensuring both that the legislature complied with, but did not exceed, the authority conferred by this Court's order.

Legislative Defendants do not cite any persuasive authority supporting their position that the enactment of the proposed remedial plans rendered this action moot. Nor do Legislative Defendants acknowledge, much less try to distinguish, the voluminous authority contrary to their unsupported position. Accordingly, the General Assembly's enactment of its remedial plans did not moot this action.

B.

Second, Legislative Defendants argue that even if the case is not moot, our review of the proposed remedial districts is limited to determining, at most, whether the General Assembly corrected the racial gerrymanders previously identified by this Court. According to Legislative Defendants, this Court, therefore, may not consider whether the remedial plans otherwise violate federal or state constitutional or statutory law. Leg. Defs.' Objs. Resp. 22-28, 51-52.

In support of their argument that this Court may consider only those challenges to a remedial districting plan that rely on the same legal theory as the original violation, Legislative Defendants principally rely on the Supreme Court's statement in Upham v. Seamon, 456 U.S. 37 (1982), that a court-drawn interim remedial plan may not "'reject[] state policy choices more than . . . necessary to meet the specific constitutional violations.'" Leg. Defs.' Obj. Resp. 23 (quoting Upham, 456 U.S. at 42 (emphasis retained)). According to Legislative Defendants, the Supreme Court's use of the phrase "specific constitutional violations" limits this Court's review to determining whether the remedial plans corrected the racial gerrymanders identified by this Court.

But in Upham, the Supreme Court struck down a court-drawn interim remedial plan because the district court redrew an entire state districting plan, notwithstanding that only two of twenty-seven districts were the subject of an ongoing challenge by the Attorney General. 456 U.S. at 43 ("We have never said that the entry of an objection by the Attorney General to any part of a state plan grants a district court the authority to disregard aspects of the legislative plan not objected to by the Attorney General."). Unlike in Upham, this Court and the Supreme Court have rendered final decisions that the General Assembly's 2011 districting plans violated the Constitution. Also unlike in Upham, this Court has given the legislature the first opportunity to draw new districts. And most significantly, unlike the district court in Upham, which redrew districts unaffected by the alleged violation, this Court did not—indeed, could not—direct the General Assembly to redraw districts unaffected by the constitutional violation. Upham, therefore, does not constrain this Court's authority to ensure that the General Assembly's proposed remedial plan complies with federal and state law.

Legislative Defendants similarly misplace reliance on the Fourth Circuit's decision in McGhee. There, a district court found that a municipal districting plan that elected all five county commissioners in county-wide, at-large districts violated Section 2 of the Voting Rights Act by freezing a sizable minority of African-American citizens (approximately 40 percent of the voting age population) out of any representation on the commission. McGhee, 860 F.2d at 112-13. To remedy the violation, the county adopted a new plan composed of seven single-member districts. Id. at 113. Only two of the seven remedial districts were majority-minority, meaning that, according to the plaintiffs, the preferred candidates of African-Americans would make up, at most, 28 percent of the commission, less than their proportional representation in the county. Id. at 113-14. In order to provide African-American representation on the commission in proportion to the population of African-Americans in the county, the district court rejected the proposed plan and adopted an alternative plan akin to cumulative voting. Id. at 114.

The Fourth Circuit concluded the district court erred in rejecting the county's proposed plan and adopting the cumulative voting plan. Id. The Court emphasized that the plain language of the Voting Rights Act stated that minority groups have no right to "proportional representation." Id. at 119. Because (1) the county's plan provided a "complete remedy" for the Section 2 violation and (2) the proportional representation plan adopted by the court exceeded the relief to which the plaintiffs were entitled under the Voting Rights Act, the district court erred. Id. at 115, 120-21 (internal quotation marks omitted). Unlike the McGhee plaintiffs' request for proportional representation, Plaintiffs do not ask this Court to provide relief exceeding that to which they are entitled under the Constitution or law, nor is this Court ordering any such relief. Rather, Plaintiffs simply ask this Court not to approve a proposed remedy for the racial gerrymandering that "violates anew constitutional or statutory voting rights"—a proposition McGhee expressly supports. Id. at 115.

Contrary to Legislative Defendants' argument that Upham and McGhee foreclose review of violations other than those originally alleged, numerous courts, including three- judge panels in this circuit bound by Upham, have held that their review of a remedial redistricting plan extends beyond the particular legal theory that was the basis for invalidating the original plan. Large, 670 F.3d at 1148 (rejecting municipal redistricting plan imposed to remedy Voting Rights Act violation due to noncompliance with state constitutional provision); Harris, 2016 WL 3129213, at *2 (rejecting Legislative Defendants' argument that the court's review of remedial maps was "limited to whether the new Congressional Districts 1 and 12 pass constitutional muster," and stating that "precedent suggests that we have a responsibility to review the plan as a whole" (citing McGhee, 860 F.2d at 115)); Personhuballah v. Alcorn, 155 F. Supp. 3d 552, 564 (E.D. Va. 2016) ("[T]hough the [legislator intervenors] urge us not to consider the requirements of Section 2, as no Section 2 claim was raised in Page II, we think it appropriate to implement a plan that complies with federal policy disfavoring discrimination against minority voters." (footnote omitted)); Jeffers v. Clinton, 756 F. Supp. 1195, 1199 (E.D. Ark. 1990) (rejecting districts created by remedial plan that failed to comply with Voting Rights Act, notwithstanding that such districts were not subject to original challenge); Sullivan v. Crowell, 444 F. Supp. 606, 611-12 (W.D. Tenn. 1978) (finding that legislative remedial plan enacted to cure one-person, one-vote violations violated state constitution); cf. Burns v. Richardson, 384 U.S. 73, 83 (1966) (holding that court considering remedial apportionment plan "must consider the scheme as a whole"). Again, Legislative Defendants fail to acknowledge, much less distinguish, this contrary authority.

Additionally, were this Court to accept Legislative Defendants' argument, the General Assembly could draw a map to remedy their racial gerrymander that plainly violated, for example, the Equal Protection Clause's one-person, one-vote requirement. According to Legislative Defendants, this Court nonetheless would be required to approve the map, and wait for Plaintiffs to bring a separate one-person, one-vote claim. Plaintiffs then would be forced to incur the costs of litigating a new action, and the majority party in the legislature would reap the benefits of using an unconstitutional districting plan for another election cycle. Indeed, a legislature could adopt seriatim unconstitutional or unlawful districting plans as remedial plans so long as each new plan violated a different constitutional or statutory provision. To be sure, some challenges to a remedial districting plan—like Plaintiffs' partisan gerrymandering objection—would demand development of significant new evidence and therefore be more appropriately addressed in a separate proceeding. But in the absence of a demonstration that objections to a remedial districting plan require such factual development, this Court declines to create the perverse incentive Legislative Defendants propose.

C.

Third, Legislative Defendants assert that, as a general matter, this Court lacks jurisdiction under the three-judge panel statute, 28 U.S.C. § 2284, to consider any objections other than racial gerrymandering, including objections premised on violations of state law. Leg. Defs.' Objs. Resp. 26. Legislative Defendants are correct that Section 2284 establishes the jurisdiction for a three-judge panel to hear federal constitutional challenges relating to the apportionment of any statewide legislative body. See Kalson v. Paterson, 542 F.3d 281, 287 (2d Cir. 2008) (holding the three-judge requirement under Section 2284 is jurisdictional). But "once convened, 'the jurisdiction of the [three-judge] District Court so constituted . . . extends to every question involved, whether of state or federal law, and enables the court to rest its judgment on the decisions of such of the questions as in its opinion effectively dispose of the case.'" Armour v. Ohio, 775 F. Supp. 1044, 1048 (N.D. Ohio 1991) (quoting Sterling v. Constantin, 287 U.S. 378, 393-94 (1932)); see also Page v. Bartels, 248 F.3d 175, 190 (3d Cir. 2001), as amended (June 25, 2001) (holding that the pendent jurisdiction of a three-judge panel extends to all claims that are "inextricably intertwined" with the claim that served as the basis of the panel's jurisdiction). To that end, a number of three-judge panels have exercised their pendent jurisdiction over state law claims in redistricting cases, particularly when state law claims are "inextricably intertwined" with their federal constitutional claims. See, e.g., Page, 248 F.3d at 190; Armour, 775 F. Supp. at 1048; Sullivan, 444 F. Supp. at 613 (noting that "pendent jurisdiction of a properly convened three-judge court is measured by the same standards applicable to a one-judge district court" and therefore exercising pendent jurisdiction over claim that multimember remedial districts violated state constitution).

Legislative Defendants identify two decisions in which three-judge district courts have declined to exercise their pendent jurisdiction over state law claims or non-redistricting federal claims. But in those cases the courts did not dispute their authority to exercise pendent jurisdiction over related state or federal claims; rather, they declined to exercise such jurisdiction because the state law claims or non-redistricting federal claims were unrelated to the claim giving rise to the panel's jurisdiction. See, e.g., Robertson v. Bartels, 148 F. Supp. 2d 443, 461-62 (D.N.J. 2001) (declining to exercise pendent jurisdiction in racial gerrymandering case over claim that durational residency requirement violated state constitution); Adams v. Clinton, 90 F. Supp. 2d 35, 39 (D.D.C. 2000) (declining to exercise pendent jurisdiction in case challenging denial of apportionment of representative to District of Columbia to various other claims premised on denial of home rule). Accordingly, this Court has authority under Section 2284 to consider Plaintiffs' federal and state law objections to the General Assembly's remedial plan, at least to the extent such objections are "inextricably intertwined" with the claim that serves as the basis of this Court's jurisdiction.

There are no doubt cases when it is appropriate for a three-judge panel to decline to exercise jurisdiction over an allegedly pendent claim, such as when the claim implicates an unsettled question of state law. See Robertson, 148 F. Supp. 2d at 461-62. Indeed, we reach that conclusion with regard to Plaintiffs' arguments that two configurations in the 2017 Plan fail to comply with the North Carolina Constitution's Whole County Provision. See infra Part III.B.2.

But having considered the factors of judicial economy, convenience, fairness to the litigants, and comity, the Court finds that the exercise of pendent jurisdiction over Plaintiffs' objections premised on Legislative Defendants' alleged failure to comply with the North Carolina Constitution's prohibition on mid-decade redistricting is particularly appropriate here. See Sullivan, 444 F. Supp. at 613; Hagans v. Lavine, 415 U.S. 528, 545 (1974) (noting that "[n]eedless decisions of state law should be avoided both as a matter of comity and to promote justice between the parties" (quoting United Mine Workers of Am. v. Gibbs, 383 U.S. 715, 726 (1966)). Indeed, declining to exercise such jurisdiction would cause significant problems. As further explained below, this Court's order invalidating the lines surrounding the twenty-eight districts provided the sole authority for the General Assembly to ignore the North Carolina Constitution's prohibition on mid-decade redistricting. See infra Part III.B.1. Because this Court's order governed the scope of the General Assembly's redistricting authority, this Court is in the best position to determine whether the General Assembly exceeded its authority under that order by redrawing districts allegedly untainted by the identified constitutional violation.

D.

Finally, Legislative Defendants assert that, as a matter of federalism, this Court is barred from considering whether the proposed remedial plans comply with state law. But Legislative Defendants cite no cases holding that, having found that a districting plan violates the Constitution or federal law, a federal court may not consider whether a remedial plan violates state law. On the contrary, several courts have rejected remedial plans as violative of a state constitution or statute. Large, 670 F.3d at 1146 ("When a political subdivision of a State substantively contravenes the laws of that State—at least insofar as that contravention is not sanctioned by higher federal law—it no longer acts as an agent of that sovereign, and therefore is due no federal-court deference."); Sullivan, 444 F. Supp. at 611-12 (finding that legislative remedial plan enacted to cure one-person, one-vote violations violated state constitution). Legislative Defendants make no effort to address, much less distinguish, these cases.

More significantly, as Legislative Defendants concede, in apportionment cases, federal courts tasked with drawing or reviewing remedial maps should not "displac[e] legitimate state policy judgments with the court's own preferences." Perry v. Perez, 565 U.S. 388, 394 (2012). Here, North Carolina citizens have enshrined in their constitution a "policy judgment[]" that the General Assembly should not engage in mid-decade redistricting or disregard county lines unless compelled to do so by federal law. It would be paradoxical to hold, as Legislative Defendants argue, that this Court must defer to the legislature's policy decisions regarding redistricting, but not to the people of North Carolina's sovereign decisions in their constitution regarding the policies the legislature must follow in engaging in such redistricting.

* * * * *

In sum, we reject Legislative Defendants' efforts to circumscribe this Court's review of the remedial plans. Accordingly, in determining whether each of the General Assembly's remedial plans completely remedies the constitutional violation, we must also assess whether the "proffered remedial plan is legally unacceptable because [they] violate[] anew constitutional or statutory voting rights" under federal or state law. McGhee, 860 F.2d at 115.

III.

Having disposed of Legislative Defendants' arguments pertaining to the scope of our review, we now turn to Plaintiffs' specific objections to aspects of the 2017 Plans. In particular, Plaintiffs assert (1) that four of the districts—proposed remedial Senate Districts 21 and 28 and House Districts 21 and 57—fail to remedy the racial gerrymander that served as the basis for invalidating the 2011 version of those districts and (2) that several of the districts and district configurations violate provisions in the North Carolina Constitution. We address each objection in turn.

Plaintiffs' also assert that the plans, when analyzed as a whole, amounted to "grossly unconstitutional partisan gerrymanders" in violation of the Equal Protection Clause. Pls.' Objs. 42-43. Plaintiffs, however, acknowledge that in the absence of discovery, this Court does not have an adequate record to rule on their partisan gerrymandering objection. Id. Accordingly, Plaintiffs do not presently raise any partisan gerrymandering objection, and therefore we do not address whether the 2017 Plans are unconstitutional partisan gerrymanders.

A.

As detailed more fully in this Court's earlier opinion, a state legislature engages in impermissible racial gerrymandering, if, in drawing the district lines, consideration of "race predominated over traditional race-neutral redistricting principles," absent a showing by the State that the "'districting legislation [wa]s narrowly tailored to achieve . . . [a] compelling state interest.'" Covington I, 316 F.R.D. at 129 (quoting Shaw v. Hunt, 517 U.S. 899, 908 (1996)). Predominance may be shown "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." Id. (quoting Alabama, 135 S.Ct. at 1267). "In general, that requires proof that 'the legislature subordinated traditional race-neutral districting principles, including . . . compactness, contiguity, and respect for political subdivisions . . . to racial considerations." Id. (quoting Miller v. Johnson, 515 U.S. 900, 907 (1995)). Relevant circumstantial evidence that the Supreme Court has considered in determining whether racial considerations predominated includes, but is not limited to: "bizarre or non-compact district shape" and "district lines that cut through traditional geographic boundaries or local election precincts." Id.

In finding that race predominated in the drawing of dozens of district lines in the 2011 districting plans—including the previous versions of the four districts subject to Plaintiffs' racial gerrymandering objections—this Court relied on both direct and circumstantial evidence. In particular, Representative Lewis's and Senator Rucho's instructions that Dr. Hofeller draw, where possible, majority-African American "VRA districts"—which Dr. Hofeller implemented by searching for minority population centers and, where feasible, drawing district lines around those population centers—provided direct evidence that the General Assembly predominantly relied on race in drawing the challenged districts. Id. at 130-37. We also relied on circumstantial evidence of the General Assembly's subordination of traditional race-neutral principles, such as the challenged districts' bizarre shapes, lack of compactness, and division of counties, municipalities, precincts, and communities of interest along racial lines. See, e.g., id. at 137-38, 143-51. With this evidence as a backdrop, we now must consider whether each of the four districts "so far as possible eliminate[s] the discriminatory effects" of the racial gerrymander in each of the four districts. Louisiana, 380 U.S. at 154.

Plaintiffs and Legislative Defendants disagree as to the governing burden of proof. According to Plaintiffs, Defendants bear the burden of establishing the 2017 districts completely remedy the constitutional violation. By contrast, Legislative Defendants assert that Plaintiffs bear the burden of proving that the 2017 districts fail to remedy the constitutional violation. Plaintiffs are correct that, outside the context of redistricting, the Supreme Court has held that once a governmental action is found to violate the Equal Protection Clause, the governmental defendant bears the burden of demonstrating that its proposed remedial plan remedies the constitutional violation. See, e.g., United States v. Virginia, 518 U.S. 515, 547-48 (1996) (holding, in sex discrimination case, that "[h]aving violated the Constitution's equal protection requirement, Virginia was obliged to show that its remedial proposal 'directly address[ed] and relate[d] to' the violation" (quoting Milliken v. Bradley, 433 U.S. 267, 280 (1977)); Greene v. Cty. School Bd. of New Kent Cty., Va., 391 U.S. 430, 439 (1968) ("The burden on a school board today is to come forward with a plan that promises realistically to work, and promises realistically to work now."). But the Supreme Court never has addressed where the burden lies in the context of a challenge to a state redistricting plan adopted to remedy a racial gerrymander. We need not decide that unsettled question, however, because we conclude that regardless of whether the burden lies with Defendants or Plaintiffs, Senate Districts 21 and 28 and House Districts 28 and 57 fail to remedy the constitutional violation.

In doing so, we also must keep in mind that we are not confronted with an original racial gerrymandering challenge to the four proposed remedial districts. Rather, we consider these districts after already having found that their preceding versions violated the Constitution. This remedial posture impacts the nature of our review. Generally, state legislative enactments—including districting plans—are presumed valid and entitled to substantial judicial deference. See Upham, 456 U.S. at 43 ("[I]n the absence of a finding that the . . . reapportionment plan offended either the Constitution or the Voting Rights Act, the District Court was not free . . . to disregard the political program of the . . . State Legislature."); Wise v. Lipscomb, 437 U.S. 535, 540 (1978) ("The new legislative plan, if forthcoming, will then be the governing law unless it, too, is challenged and found to violate the Constitution."). "The district court need not defer to a state-proposed remedial plan, however, if the plan does not completely remedy the violation . . . ." Harvell v. Blythe Sch. Dist. No. 5, 126 F.3d 1038, 1040 (8th Cir. 1997) (emphases added); cf. Abrams v. Johnson, 521 U.S. 74, 85 (1997) (holding that legislative "plan is not owed Upham deference to the extent the plan subordinated traditional districting principles to racial considerations"). Accordingly, when, as here, "the districting plan is offered as a replacement for one invalidated by the court[,] . . . the court has an independent duty to assess its constitutionality, and cannot ignore substantial evidence of improper racial motivation." Wilson v. Jones, 130 F. Supp. 2d 1315, 1322 (S.D. Ala. 2000), aff'd sub nom., Wilson v. Minor, 220 F.3d 1297 (11th Cir. 2000).

In the remedial posture, courts must ensure that a proposed remedial districting plan completely corrects—rather than perpetuates—the defects that rendered the original districts unconstitutional or unlawful. See Abrams, 521 U.S. at 86. To that end, a remedial districting plan cannot be based on considerations that "would validate the very maneuvers that were a major cause of the unconstitutional districting." Id.

Of particular relevance here, see infra Parts III.A.1-4, efforts to protect incumbents by seeking to preserve the "cores" of unconstitutional districts or through reliance on political data closely correlated with race—particularly attempts to ensure an incumbent will prevail in his or her new district—have the potential to embed, rather than remedy, the effects of an unconstitutional racial gerrymander in a proposed remedial districting plan. Although the Supreme Court has not squarely addressed whether, and by what means, a state redistricting body tasked with drawing remedial districts may protect incumbents elected in racially gerrymandered districts, four Justices have stated that whether "the goal of protecting incumbents is legitimate, even where, as here, individuals are incumbents by virtue of their election in an unconstitutional racially gerrymandered district . . . . is a questionable proposition." Easley v. Cromartie, 532 U.S. 234, 262 n.3 (2001) (Thomas, J., dissenting) (noting that that question was not presented to the Supreme Court or district court and, therefore, that the Court had not addressed it). Lower courts likewise have expressed concern that remedial districts drawn to protect incumbents elected under an unlawful or unconstitutional plan may serve to perpetuate the identified violation. See, e.g., Ketchum, 740 F.2d at 1408 (expressing skepticism about efforts to protect incumbents in maps drawn to remedy impermissible race-based districting because "many devices employed to preserve incumbencies are necessarily racially discriminatory"); Jeffers, 756 F. Supp. at 1199-1200 (rejecting remedial districts that violated the Voting Rights Act, notwithstanding that governmental defendant asserted the districts were drawn to protect incumbents, because "[t]he desire to protect incumbents, either from running against each other or from a difficult race against a black challenger, cannot prevail if the result is to perpetuate the violations of the equal-opportunity principle contained in the Voting Rights Act").

The potential for efforts to protect incumbents to perpetuate a constitutional violation is greater with some forms of incumbency protection than others. Outside of the remedial context, the Supreme Court has recognized that in drawing district lines a legislature may seek to "avoid[]" pairing incumbents in the same district. See Karcher v. Daggett, 462 U.S. 725, 740-41 (1983). But the Supreme Court has emphasized that, even when a legislature is not seeking to remedy an unconstitutional districting plan, other forms of incumbency protection—most notably, efforts to ensure an incumbent will prevail in his new district—pose greater concerns, particularly when efforts to protect incumbents rely on considerations closely correlated with race.

In League of United Latin American Citizens v. Perry (LULAC), 548 U.S. 399 (2006), the Supreme Court considered a mid-decade redistricting plan that removed Latinos from a district in order to protect an incumbent "from a constituency that was increasingly voting against him." Id. at 440-41. Notwithstanding that the district court concluded that the legislature removed the Latino voters from the district "for political, not racial, reasons," the Supreme Court found the districting plan violated Section 2 of the Voting Rights Act. Id. In reaching this conclusion, the Court stated that "incumbency protection can be a legitimate factor in districting, but experience teaches that incumbency protection can take various forms, not all of them in the interests of the constituents." Id. at 440-41 (citation omitted).

If the justification for incumbency protection is to keep the constituency intact so the officeholder is accountable for promises made or broken, then the protection seems to accord with the concern for the voters. If, on the other hand, incumbency protection means excluding some voters from the district simply because they are likely to vote against the officeholder, the change is to benefit the officeholder, not the voters. By purposely redrawing lines around those who opposed [the incumbent], the state
legislature took the latter course. This policy, whatever its validity in the realm of politics, cannot justify the effect on Latino voters.
Id. Lower courts have reached the same conclusion—drawing districts "on a block-by-block or neighborhood- or town-splitting level to corral voters perceived as sympathetic to incumbents or to exclude opponents of the incumbents" is a "form of incumbent protection [that] is much different" than the form of incumbent protection that the Supreme Court has sanctioned: avoiding the pairing of incumbents. Vera v. Richards, 861 F. Supp. 1304, 1336 (S.D. Tex. 1994), aff'd sub nom., Bush v. Vera, 517 U.S. 952 (1996) (finding unconstitutional decennial redistricting plan that shifted voters among districts based on race in order to protect incumbents). Therefore, "[i]ncumbent protection is a valid state interest only to the extent that it is not a pretext for unconstitutional racial gerrymandering." Id.

Accordingly, regardless of whether it is ever legitimate for a state redistricting body to draw a remedial districting plan to protect incumbents elected to racially gerrymandered districts—a question the Supreme Court has yet to squarely address—a redistricting body's desire to protect such incumbents must give way to its duty to completely remedy the constitutional violation. That is particularly true where, as here, a state redistricting body relies on redistricting criteria closely correlated with race in its pursuit of the far more suspect goal of seeking to ensure that incumbents elected in a racially gerrymandered district prevail in their remedial district.

For example, although state redistricting bodies may use political data for certain purposes when initially drawing district lines, see Gaffney v. Cummings, 412 U.S. 735, 752-53 (1973) (holding that state legislature did not violate Equal Protection Clause by relying on political data "to create a districting plan that would achieve a rough approximation of the statewide political strengths of the Democratic and Republican Parties"), the consideration of political data to ensure incumbents will prevail in their remedial district may serve to carry forward the discriminatory effect of the original violation, see Jeffers, 756 F. Supp. at 1199-1200; c.f. Personhuballah, 155 F. Supp. 3d at 564 ("[A]t some point political concerns must give way when there is a constitutional violation that needs to be remedied."). And whereas a state redistricting body may have a "legitimate" interest in "preserving the cores of prior districts" so as to ensure an incumbent prevails in his new district when initially drawing a redistricting plan, Karcher, 462 U.S. at 740, Legislative Defendants concede that a remedial plan drawn to preserve the "core of [a] racially gerrymandered district" "would perpetuate [the] racial gerrymander," Leg. Defs.' Objs. Resp. 52; Easley, 532 U.S. at 265 n.7 (Thomas, J., dissenting) ("Of course, considering that District 12 has never been constitutionally drawn, Dr. Weber's criticism—that the problem with the district lies not just at its edges, but at its core—is not without force."); cf. Personhuballah, 155 F. Supp. 3d at 561 n.8 ("[M]aintaining district cores is the type of political consideration that must give way to the need to remedy a [racial gerrymandering] violation.").

The Court emphasizes that its holding regarding the propriety of the use of political data and core preservation to protect incumbents is limited to the remedial phase and should not be construed to address the legislature's ability to consider such factors outside the remedial context.

In light of the remedial context—and in view of the compelling evidence presented by Plaintiffs that the General Assembly's efforts to protect incumbents by preserving district cores and through use of political data perpetuated the unconstitutional effects of the four districts that are the subject of Plaintiffs' racial gerrymandering objections, see infra Part III.A.1-4—we reject Legislative Defendants' two principal arguments in response to Plaintiffs' racial gerrymandering objections: (1) that the adopted criterion barring the use of racial data in drawing the 2017 Plan categorically precludes a finding that any of the districts in the plans continues to be a racial gerrymander and (2) that sustaining Plaintiffs' racial gerrymandering objections would be tantamount to holding that a state redistricting body must consider race in drawing a redistricting plan to remedy a racial gerrymander. Leg. Defs.' Objs. Resp. 30 (citing Adopted Criteria for House and Senate Plans, Sept. 7, 2017, ECF No. 184-37); Leg. Defs.' Rec. Plan Resp. 16.

As to the first argument—that the race-blind criterion immunizes the proposed remedial districts from any claim of racial gerrymandering—the Supreme Court long has recognized that a statute enacted by a state legislature to remedy an unconstitutional race-based election law can perpetuate the effects of the constitutional violation, and thereby fail to constitute a legally acceptable remedy, even when the remedial law is facially race-neutral. For example, in Lane v. Wilson, the Court considered a statute enacted by the Oklahoma legislature to remedy a racially discriminatory voter qualification provision in the Oklahoma Constitution that the Court previously had held violated the Fifteenth Amendment. 307 U.S. at 269-71; see also Guinn v. United States, 238 U.S. 347, 367 (1915) (striking down Oklahoma constitutional provision excluding lineal descendants of persons entitled to vote prior to January 1, 1866, from being subject to literacy test as a precondition to voting on grounds that provision "by necessary result re-creates and perpetuates the very conditions which the [Fifteenth] Amendment was intended to destroy"). Notwithstanding that the remedial statute was facially race-neutral, the Court nonetheless struck down the remedial statute as perpetuating the constitutional violation because it "part[ook] too much of the infirmity [of the violative state constitutional provision] to be able to survive." Lane, 307 U.S. at 275; see also Kirksey v. Bd. of Sup'rs of Hinds Cty., Miss., 554 F.2d 139, 146-47 (5th Cir. 1977) ("Where a [redistricting] plan, though itself racially neutral, carries forward intentional and purposeful discriminatory denial of access that is already in effect, it is not constitutional. Its benign nature cannot insulate the redistricting government entity from the existent taint."), superseded by statute on other grounds as recognized in League of United Latin Am. Citizens, Council No. 4434 v. Clements, 999 F.2d 831, 866 (5th Cir. 1993).

Like the remedial election law at issue in Lane, even though the General Assembly here forbid the mapdrawers from considering race, the district configurations that are the subject of Plaintiffs' racial gerrymandering objections "partake too much of the infirmity" of their racially gerrymandered versions and therefore continue to constitute racial gerrymanders. Id. In particular, as explained more fully below, even though the Adopted Criteria barred Representative Lewis, Senator Hise, and Dr. Hofeller from considering race in drawing the remedial plans, several of the challenged districting configurations in the remedial plan preserve the "core of the racially gerrymandered district" configurations—which derived from Dr. Hofeller's 2011 VRA exemplars—thereby "perpetuat[ing] [the] racial gerrymander." Leg. Defs.' Objs. Resp. 52; see also infra Part III.A.1-4. Likewise, even though the mapdrawers could not consider race in drawing the 2017 Plan, the mapdrawers' use of partisan election results—which, Legislative Defendants concede, are correlated with race, Hr'g Tr. 115:8-15—to try to ensure incumbents would prevail in their remedial districts carried forward the effects of the identified racial gerrymanders, see infra Part III.A.1-4.

The fallacy of Legislative Defendants' argument that the race-blind criterion precludes any finding of racial gerrymandering is most evident when one follows the argument to its logical conclusion. Under Legislative Defendants' argument, a state redistricting body tasked with redrawing districts to remedy a racial gerrymander could adopt the exact same districts as those held unconstitutional so long as the redistricting body relied on only the prior district lines, not race, in drawing the purportedly remedial districts. Such a result plainly would not "so far as possible eliminate the discriminatory effects" of the racial gerrymander, as the Constitution demands. Louisiana, 380 U.S. at 154; see also Perez v. Abbott, --- F. Supp. 3d ---, 2017 WL 3495922, at *43 (W.D. Tex. Aug. 15, 2017) (rejecting State's argument that "a Legislature could . . . insulate itself from a Shaw-type challenge simply by re-enacting its plan and claiming that it made no decisions about who to include in the district at the time of re-enactment"). Nor would this result comply with this Court's order that the General Assembly "enact new House and Senate districting plans remedying the constitutional deficiencies." Covington III, 2017 WL 3254098, at *3.

As to Legislative Defendants' contention that sustaining Plaintiffs' racial gerrymandering objections is tantamount to requiring that a state redistricting body consider race in redrawing districts to remedy a racial gerrymander, again we disagree. We do not hold that a legislative body tasked with redrawing districts to remedy a racial gerrymander must consider race. Rather, we hold that when, as here, a legislative body faced with such a task chooses to rely on redistricting considerations that have the potential to carry forward the effects of the constitutional violation—like preserving district cores and relying on political data to draw districts that ensure incumbents will prevail in their new districts—then the legislative body must ensure that its reliance on those considerations did not serve to perpetuate the effects of the racial gerrymander. Accordingly, the General Assembly's obligation to be conscious of the prior racially drawn districts to ensure that the proposed 2017 Plans remedy the racially gerrymander derives not from judicial mandate, but instead from the General Assembly's choice to adopt redistricting criteria that posed a risk of carrying forward the effects of the racial gerrymanders in the 2011 Plans.

* * * * *

In sum, we conclude that this Court has an independent duty to assess whether the remedial plans "completely remedy" the constitutional violation. And we further conclude that in the remedial context, a state redistricting body may not rely on an otherwise legitimate redistricting consideration—such as seeking to ensure incumbents will prevail in their remedial districts—if doing so would prevent it from completely remedying the identified constitutional violation. With these principles in mind, we now analyze the four proposed remedial districts subject to Plaintiffs' racial gerrymandering objections.

1. Senate District 21

The General Assembly's proposed remedial version of Senate District 21 encompasses all of Hoke County and a portion of Cumberland County. Under the plan in effect in 2010 (the "benchmark plan"), Senate District 21 "was a 'squarely shaped' district located in the northwestern quadrant of Cumberland County." Covington I, 316 F.R.D. at 146. The version of Senate District 21 adopted in the 2011 plan was drawn, using Dr. Hofeller's VRA "exemplar," as a 50%-plus-one BVAP district and contained "multiple appendages, which [we]re so thin and oddly shaped that it [wa]s hard to see where the district beg[a]n and end[ed]." Id. This Court concluded that the district constituted a racial gerrymander because Dr. Hofeller drew the district's lines to comply with the Chairs' unconstitutional 50%-plus-one criterion and because the district was noncompact, "divide[d] traditional political boundaries on the basis of race," and divided 33 of the 41 precincts in Cumberland County. Id. at 147. We further concluded that compliance with the Voting Rights Act did not provide the General Assembly with the compelling interest necessary to justify its reliance on race, as the State presented no evidence that "racial bloc voting . . . would enable the majority usually to defeat the minority group's candidate of choice." Id. at 167.

The proposed remedial version of Senate District 21 reduced the district's BVAP from 51.53 percent to 47.51 percent. Add. Stats. on 2017 Sen. Redistricting Plan, Sept. 7, 2017, ECF No. 184-6. However, the remedial version's BVAP still exceeds the BVAP of the benchmark version (44.93%). Covington I, 316 F.R.D. at 146. Although the remedial version of the district no longer includes some of the former version's Cumberland County appendages and splits fewer precincts, the remedial version retains the core shape of the unconstitutional version of the district. In particular, the district still encompasses all of Hoke County and reaches into Cumberland County to include a horseshoe-shaped section of the city of Fayetteville. A comparison between Dr. Hofeller's Cumberland County exemplar and proposed remedial Senate District 21 supports the conclusion that the General Assembly's use of political data—which Legislative Defendants concede is closely correlated with race, Hr'g Tr. 115:8-15—to ensure the incumbents in Senate Districts 19 and 21 would prevail in their remedial districts served to perpetuate the unconstitutional design of the invalidated 2011 map. Most notably, the exemplar district for Senate District 21 contained a similar horseshoe-shaped section of the city of Fayetteville that includes Fayetteville's predominantly black VTDs and blocks and excludes Fayetteville's predominantly white VTDs and blocks. Tr. Ex. 3019-76. Although more compact than the previous version, the remedial district still performs poorly on statistical measures of compactness relative to other Senate districts. Senate District Compactness, Sept. 15, 2017, ECF No. 187-9.

Racial density maps prepared by Plaintiffs' expert Anthony Fairfax, which indicate the percentage of population in each census block that identified as any part black, reveal that, like the unconstitutional version of the district, the General Assembly's remedial version of the district "cuts through downtown Fayetteville and only includes the majority black VTDs as well as practically all of the majority black blocks." Decl. of Anthony E. Fairfax ("Fairfax Decl.") 4, apps. 2-5, Sept. 15, 2017, ECF No. 187-6; see also Covington I, 316 F.R.D. at 141. Large swaths of the majority-white sections of Fayetteville are drawn out of the district. Fairfax Decl. 4, apps. 2-5, Sept. 15, 2017, ECF No. 187-6. Legislative Defendants maintain that remedial Senate District 21's division of Fayetteville on racial lines reflects a legitimate effort to "preserve[] the heart of Fayetteville." Leg. Defs.' Objs. Resp. 37. But when confronted with the racial density maps, Legislative Defendants fail to provide any explanation or evidence as to why "preserv[ing] the heart of Fayetteville" required the exclusion of numerous majority-white precincts in downtown Fayetteville from the remedial district.

In addition to highlighting the similarities between the shape of the remedial district and the unconstitutional version, the lack of compactness, and the racial make-up of the district, Plaintiffs also submitted an analysis by an applied mathematics expert, Dr. Gregory Herschlag of Duke University, who used a computer to generate 78,485 hypothetical district maps for the Hoke/Cumberland County grouping. The computer drew the hypothetical district maps to conform to equal population requirements, maintain contiguity, preserve precincts, and, once those criteria are satisfied, maximize compactness according to the Polsby-Popper metric relied on by the General Assembly. Decl. of Dr. Gregory Herschlag ¶ 10, Sept. 14, 2017, ECF No. 187-10. Dr. Herschlag's analysis found that Senate District 21 "contain[ed] a significantly higher percentage in population that is African-American (46.5%) than any district in the 78,485 simulated districting plans." Id. at ¶ 8 (emphasis added). Legislative Defendants correctly note that the analysis has certain limitations—it relied on only one of the two principal measures of compactness embraced by the Joint Committee and did not account for one traditional districting criterion adopted by the Joint Committee, keeping municipalities whole. Nonetheless, Dr. Herschlag's analysis does provide additional evidence that the remedial version of the district perpetuates the race-based districting that rendered the earlier version unconstitutional, particularly in light of Legislative Defendants' failure to introduce any evidence explaining or justifying the remedial district's racial make-up.

Legislative Defendants' criticism of Dr. Herschlag's analysis for failing to keep municipalities whole is undermined by the fact that one indicium that the remedial district continues to constitute a racial gerrymander is that it divides the city of Fayetteville along racial lines. The proposed 2017 Senate Plan also divides the town of Spring Lake between Senate District 21 and Senate District 19, Fairfax Decl. at 17, further demonstrating that the General Assembly did not place significant weight on preserving municipal lines.

In conclusion, the district (1) preserves the core shape of the unconstitutional version of the district and Dr. Hofeller's VRA exemplar, (2) has a higher BVAP than its benchmark version, (3) divides the city of Fayetteville along racial lines, (4) has a low compactness score and is significantly less compact than the benchmark version, and (5) has a far greater percentage of African Americans than thousands of other districting plans that satisfy most traditional districting principles adopted by the Joint Committee. Based on this evidence, we conclude that the remedial version of Senate District 21 failed to eliminate the discriminatory aspects of the unconstitutional version, and therefore continues to constitute a racial gerrymander.

2. Senate District 28

The proposed remedial version of Senate District 28, which is shaped like a reverse "L," sits at the center of Guilford County. Dr. Hofeller drew the version of the district adopted in the 2011 redistricting as a 50%-plus-one BVAP district, and "[a]lthough the portion of the district in Greensboro [wa]s not particularly strange in its shape, an arm of the district protrude[d] west, then hook[ed] south, to capture part of the city of High Point." Covington I, 316 F.R.D. at 147. The northeast arm reached into predominantly black sections of Greensboro. This Court concluded that the district constituted a racial gerrymander because it was drawn, using Dr. Hofeller's VRA "exemplar," to be a 50%-plus-one district, was less compact than its benchmark district, added substantially more black voters and subtracted white voters from its benchmark, and split municipalities along racial lines. Id. at 147-48. The Court further concluded that compliance with the Voting Rights Act did not provide the General Assembly with the compelling interest necessary to justify its reliance on race, as the State lacked any evidence that "racial bloc voting" would allow the majority to usually to defeat black voters' candidate of choice. Id. at 167.

The proposed remedial version of Senate District 28 eliminates the "arm" into High Point included in the previous version, but otherwise tracks the shape of the version of the district held unconstitutional. Indeed, the proposed remedial version's contours more closely follow Dr. Hofeller's VRA "exemplar" than the unconstitutional version, taking on the exemplar's reverse "L" shape and capturing most of the precincts included in the exemplar. See Tr. Ex. 3019-71; Hr'g Pls.' Ex. PD-1. The General Assembly's remedial version reduced the district's BVAP from 56.49 percent to 50.52 percent. Add. Stats. on 2017 Sen. Redistricting Plan, Sept. 7, 2017, ECF No. 184-6. But the BVAP of the remedial version still exceeds that of the benchmark version (47.20%), Covington I, 316 F.R.D. at 147, and the 50%-plus-one threshold, establishing that the General Assembly's retention of the unconstitutional version's core and previous use of the majority-black target continues to shape the remedial district's racial make-up.

Whereas the benchmark version of the district had approximately 2,000 more black voters than white voters, the remedial version of the district has approximately 29,000 more black voters than white voters. Add. Stats. on 2017 Sen. Redistricting Plan, Sept. 7, 2017, ECF No. 184-6. Although the district encompasses only a portion of Greensboro, racial density maps reveal that the district encompasses all of the majority black VTDs within Greensboro. Fairfax Decl. 5. Notwithstanding that the district excludes predominantly white sections of Greensboro, it reaches out of Greensboro's city limits to capture predominantly African-American areas in eastern Guilford County. And the uncontradicted affidavit of Democratic Senator Gladys Robinson, who represents Senate District 28, avers that under the revisions to the district "the more heavily African-American precincts were included in the district while the predominantly white precinct was removed." Decl. of Sen. Gladys A. Robinson ("Robinson Decl.") 5-6, Sept. 14, 2017, ECF No. 187-5. Although more compact than the unconstitutional version, the remedial district is among the least compact senate districts in the state and is substantially less compact than its benchmark version. Sen. District Compactness, Sept. 15, 2017, ECF No. 187-9.

Legislative Defendants maintain that "the BVAP level in District 28 is naturally occurring as it is the result of the population residing in those whole precincts that were included in the district." Leg. Defs.' Objs. Resp. 32. But this argument begs—rather than answers—the relevant question: what was the General Assembly's predominant reason for including those particular whole precincts in the district? And the Special Master's Recommended Senate District 28, which significantly improves on the district's compactness and more closely tracks Greensboro's municipal lines, indicates that the district's lines, and therefore its BVAP, were not, in fact, "naturally occurring," but rather a consequence of the district's tracking of the core shape of Dr. Hofeller's VRA exemplar. See infra Part IV.B.2.

Legislative Defendants further argue that the district remedies the constitutional violation because a "district anchored in eastern Greensboro that tracks the city boundaries" could not be drawn with a lower BVAP without considering race. Leg. Defs.' Objs. Resp. 32. But Legislative Defendants failed to introduce any evidence, much less race-neutral evidence, establishing that the General Assembly had to "anchor[]" the remedial district in eastern Greensboro—the predominantly black portion of the city that served as the "anchor" of the unconstitutional version of the district. Indeed, by deciding to "anchor" the district in the same predominantly black area as the unconstitutional version of the district and Dr. Hofeller's exemplar, Dr. Hofeller ensured that the district would retain a high BVAP, thereby perpetuating the effects of the racial gerrymander.

When viewed in totality, the district (1) preserves much of the core shape of the unconstitutional version of the district and Dr. Hofeller's VRA exemplar, (2) continues to have a BVAP that exceeds fifty percent, (3) divides Greensboro's VTDs and precincts along racial lines, and (4) has a low compactness score and is significantly less compact than the benchmark version in the plan in effect in 2010. Based on this evidence, we conclude that the General Assembly carried forward constitutional deficiencies of the previous version of the district and therefore failed to remedy the racial gerrymander.

3. House District 21

Proposed remedial House District 21 runs along the northeast edge of Sampson County into southeast Wayne County. The version of the district that the General Assembly adopted in 2011 included portions of Sampson, Duplin, and Wayne Counties and was drawn to achieve the 50%-plus-one threshold. Covington I, 316 F.R.D. at 155. This Court concluded that the district constituted a racial gerrymander because it was "visually less compact" than its benchmark and performed poorly on statistical measures of compactness, it split municipalities and counties along racial lines, and its "racial density map . . . indicate[d] that areas with a high proportion of African-American voting-age population [we]re enveloped by the protrusion and contours of House District 21." Id. at 155-56. As with the unconstitutional versions of Senate Districts 21 and 28, we further concluded that compliance with the Voting Rights Act did not provide the General Assembly with the compelling interest necessary to justify its reliance on race, as the State lacked any evidence that "racial bloc voting . . . would enable the majority usually to defeat the minority group's candidate of choice." Id. at 167.

The proposed remedial version of House District 21 reduced the BVAP from 51.90 percent to 42.34 percent, whereas the benchmark version had a BVAP of 46.25 percent. Add. Stats. on 2017 House Redistricting Plan, Sept. 7, 2017, ECF No. 184-3; Covington I, 316 F.R.D. at 158. The district no longer includes any part of Duplin County, which had to be moved to a different county grouping in order to comply with the Whole County Provision, and the revised Wayne County section of the district is more compact. But the Sampson County section of the district conforms to the bizarre shape of the district previously held unconstitutional. To be sure, the unusual borders in Sampson County are attributable in large part to the unusual borders of the selected precincts. But although the Sampson County section generally runs along the eastern edge of the county, the proposed remedial version of the district continues to include a protrusion stretching into the center of the county to capture the disproportionately black sections of the city of Clinton. Fairfax Decl. 6-7, apps. 10-11. The district separates the predominantly black areas of Clinton from the predominantly white areas by splitting a precinct on racial lines. Id. When viewed as a whole, the remedial district continues to contain all but one "of the majority black VTDs within Sampson and Wayne Counties." Id. at 6. Although the proposed remedial version of the district is more compact than the previous version, it is the lowest among all 120 House districts on one statistical measure of compactness. House District Compactness, Sept. 15, 2017, ECF No. 187-11.

Considering this evidence as a whole, the district (1) preserves the core shape of the Sampson County section of the previously unconstitutional district, (2) includes all but one of the majority-black VTDs in the two counties through which it runs, (3) divides a municipality and precinct along racial lines, (4) has an irregular shape that corresponds to the racial make-up of the geographic area, and (5) has an extremely low compactness score and is significantly less compact than the benchmark version in the plan in effect in 2010. We find this to be strong evidence that the proposed remedial district fails to remedy the racial gerrymander.

To defend the remedial district's constitutionality, Legislative Defendants assert the district's shape and racial make-up are attributable to the need to "connect" the more compact Wayne County portion of the district to the Sampson County precinct where incumbent Democratic Representative Larry Bell resides and to ensure Representative Bell and Democratic Representative William Brisson, who represents House District 19, which abuts House District 21, would likely prevail in an election in their new districts. Leg. Defs.' Objs. Resp. 42-44. Put differently, according to Legislative Defendants, the district's contours and racial make-up reflect an allegedly legitimate effort by the General Assembly to engage in two forms of incumbency protection: (1) avoiding the "double-bunking" of incumbents and (2) using electoral data to ensure an incumbent is likely to prevail in his new district. We conclude that any interest the General Assembly had in engaging in these two forms of incumbency protection should have given way to the requirement that the remedial plan completely remedy the racial gerrymander. See supra Part III.A.

Although Representative Brisson was a member of the Democratic party at the time the House and Senate redistricting plans were enacted, he was the only Democratic House member to vote for both the adopted Senate and House plans on the second and third readings. Leg. Defs.' Objs. Resp. 44 n.9. Following the enactment of the remedial redistricting plans, he announced his intention to change his party registration and run for a seventh term as a Republican. Lynn Bonner, An NC House Democrat switches to the GOP, News & Observer (Oct. 26, 2017, 6:22 PM), http://www.newsobserver.com/news/politics-government/politics-columns-blogs/under-the-dome/article180794221.html.

In particular, in order to draw Representative Bell's residence into House District 21, the General Assembly retained much of the bizarre shape of the Sampson County portion of the district and divided a precinct and municipality along racial lines—the very problems that rendered the prior version of the district unconstitutional. Because the General Assembly's incumbency protection efforts served to "validate the very maneuvers that were a major cause of the unconstitutional districting," Abrams, 521 U.S. at 86, we find that House District 21 continues to be a racial gerrymander. That the General Assembly sought not only to avoid pairing incumbents, but also to engage in the more suspect practice of using political data to "exclud[e] . . . voters from the district simply because they are likely to vote against the officeholder," LULAC, 548 U.S. at 441, reinforces this conclusion, particularly since Legislative Defendants concede that race and political affiliation are highly correlated, Hr'g Tr. 115:8-15. Accordingly, we conclude that proposed House District 21 fails to remedy the racial gerrymander.

We further note that, as a factual matter, the General Assembly did not need to draw the district to protect Representative Bell. In particular, several months before Dr. Hofeller drew the remedial districts and the General Assembly enacted Dr. Hofeller's proposed maps, Representative Bell announced that he would not be seeking re-election. See Colin Campbell, NC Rep. Larry Bell to Step Down Next Year, News & Observer (Apr. 17, 2017, 5:28 PM), http://www.newsobserver.com/news/politics-government/state-politics/article145086079.html. During legislative debate regarding the proposed districting plans, at least one legislator expressed concern that the remedial plans were protecting incumbents who already had decided to retire. Statement of Senator Jackson, H. Redist. Comm. Tr. Aug. 25, 2017, at 62:21-24, ECF 184-18 (noting that mapdrawers "should not consider people who have announced their retirements" within the context of incumbency protection). Representative Bell has since confirmed under oath that he publicly announced his intention not to run for re-election in April 2017 and that he will not, in fact, run for re-election in 2018. Decl. of Rep. Larry Bell, Nov. 10, 2017, ECF No. 211-1.

4. House District 57

The General Assembly's proposed remedial House District 57 stands in the center of Guilford County. The version of the district adopted in 2011 was drawn to add a third majority black district in Guilford County. Covington I, 316 F.R.D. at 163. In order to create the third majority black district, the General Assembly "moved and reshaped significantly" the Guilford County house districts included in the benchmark plan. Id. Analyzing the 2011 version of House District 57 alongside the other two majority-black districts in Guilford County, we concluded that the district constituted a racial gerrymander because the three districts were unnecessarily drawn to create a third majority African-American district; were "visually less compact" than the Guilford County districts in the benchmark plan; required shifting thousands of African Americans into House District 57 and moving thousands of non-African-Americans out in order to turn it into a majority-black district; created a significant difference between the racial makeup of majority-black districts and the remaining districts in Guilford County; included numerous split precincts; were less compact than the Guilford County districts in the benchmark plan; and, as revealed by racial density maps, were drawn to "encompass areas with a high proportion of voting-age African Americans." Id. at 163- 64. We further concluded that compliance with the Voting Rights Act did not provide the General Assembly with the compelling interest necessary to justify its reliance on race, as the State presented no evidence that "racial bloc voting" would consistently prevent black voters from electing the candidate of their choice. Id. at 167.

The proposed remedial version of House District 57 increased the district's BVAP from 50.69 percent to 60.75 percent, whereas the benchmark version had a BVAP of 29.93 percent. Add. Stats. on 2017 House Redistricting Plan, Sept. 7, 2017, ECF No. 184-3; Covington I, 316 F.R.D. at 163. Members of the General Assembly were informed of the significant increase in House District 57's BVAP during the legislative process, but did not alter the district in response to that information. Statement of Rep. Harrison, H. Comm. Redistricting Tr. 119:2-120:1, Aug. 25, 2017, ECF No. 184-18 ("The current African-American composition [of House District 57] is 47 percent and . . . . [t]he proposed district is now . . . 60 percent African American, which doesn't seem to cure the constitutional issue of racial gerrymandering.").

The shape of House District 57 does not follow the shape of the unconstitutional version or the shape of any Guilford County district in the benchmark plan. House District 57's reverse "L" shape does, however, encompass the core of the unconstitutional version of Senate District 28, and closely tracks Dr. Hofeller's VRA exemplar in Guilford County. See Tr. Ex. 3019-71; supra Part III.B.2. In particular, remedial House District 57 captures the same high BVAP blocks and VTDs in Greensboro included in unconstitutional remedial Senate District 28 and Dr. Hofeller's Guilford County exemplar. Fairfax Decl. 8, apps. 12-14. The vast majority of the VTDs in remedial House District 57 have BVAPs of at least 25 percent, with more than half of the VTDs having BVAPs exceeding 50 percent. Id. at 8, app. 12. And the district includes only five VTDs from the predominantly white sections of Greensboro.

The uncontradicted affidavit of State Senator Robinson, who represents Greensboro, averred that in redrawing the district the General Assembly removed a wealthy white neighborhood, Irving Park, and added a "densely populated, heavily African-American community" in Southeast Greensboro. Robinson Decl. 10-11. The district scores below the statewide mean on measures of compactness.

Similar to their arguments regarding proposed remedial Senate District 28, Legislative Defendants maintain that "the BVAP level in District 57 is naturally occurring as it is a result of the population residing in those whole precincts that were included in the district" and that a district "anchored" in eastern Greensboro and tracking city boundaries could not be drawn with a lower BVAP without considering race. Leg. Defs.' Objs. Resp. 39-40. But, as noted above, the General Assembly has provided no evidence as to why it needed to "anchor" the district in eastern Greensboro, the part of the city with a disproportionately large African-American population. And by tracking the shape of the Greensboro section of unconstitutional Senate District 28 and Dr. Hofeller's VRA exemplar, which included nearly all of the city's high BVAP VTDs, Dr. Hofeller ensured that the district would have a high BVAP, thereby carrying forward the effects of the racial gerrymander. Additionally, the Special Master's recommended reconfiguration of the Guilford County House districts reveals that the General Assembly could have drawn House districts in Guilford County that were more compact and more closely followed Greensboro's municipal lines without drawing House District 57 to mirror the shape of unconstitutional Senate District 28 and Dr. Hofeller's VRA exemplar. See infra Part IV.B.4.

Legislative Defendants further assert that we should reject Plaintiffs' objection because their alternative map would have "double-bunked" incumbents. Leg. Defs.' Objs. Resp. 41-42. But the General Assembly had an obligation to completely remedy the constitutional violation, regardless of whether Plaintiffs—or any other member of the public—provided it with a satisfactory map. And, more significantly, the Special Master's Recommended Plan demonstrates that the General Assembly could have drawn a remedial configuration of the Guilford County House Districts without double-bunking incumbents. See infra Part IV.B.4. Accordingly, we find proposed remedial House District 57 fails to completely remedy the racial gerrymander because it (1) encompasses the core of unconstitutional Senate District 28 and Dr. Hofeller's Guilford County VRA exemplar; (2) has an extremely high BVAP level—nearly 40 percent higher than its benchmark version and 10 percent higher than the unconstitutional version; (3) is almost entirely made up of high-BVAP VTDs and excludes predominantly non-black VTDs; and (4) divides the city of Greensboro along racial lines.

* * * * *

In sum, we find that proposed remedial Senate Districts 21 and 28 and House Districts 21 and 57 fail to completely remedy the constitutional violation. Because the General Assembly failed to enact "a constitutionally acceptable" remedial plan, "then the responsibility falls on th[is] Court" to reconfigure those infirm districts. Chapman, 420 U.S. at 27.

B.

Next, Plaintiffs assert that certain aspects of the remedial plan violate the North Carolina Constitution. In particular, Plaintiffs assert (1) that 2017 Enacted House Districts 36, 37, 40, 41, and 105 violate the constitutional prohibition on mid-decade redistricting, N.C. Const. art. II, §§ 3(4), 5(4); (2) that two groups of districts violate the North Carolina Constitution's so-called "Whole County Provision," id. art. II, §§ 3(3), 5(3); and (3) that one district is unconstitutionally noncompact. We address each of these objections in turn.

1.

The North Carolina Constitution provides that "[w]hen established, the [House and] [S]enate districts and the apportionment of [Representatives and] Senators shall remain unaltered until the return of another decennial census of population taken by order of Congress." Id. art. II, §§ 3(4), 5(4). Accordingly, the plain and unambiguous language of Sections 3(4) and 5(4) prohibits the General Assembly from engaging in mid-decade redistricting. Granville Cty. Comm'rs v. Ballard, 69 N.C. 18, 20-21 (1873) (holding that a state law altering a county boundary was invalid insofar as it would alter the House and Senate districts in violation of the state constitutional prohibition against mid-decade redistricting). Plaintiffs assert that five districts established by the plans (House Districts 36, 37, 40, 41, and 105 in Wake and Mecklenburg Counties) violate the constitutional prohibition on mid-decade redistricting because those districts did not violate the Constitution, did not abut a district violating the Constitution, and did not need to be altered in order to ensure compliance with the Whole County Provision. Pls.' Objs. 37.

The Supreme Court of North Carolina has not addressed the scope of the General Assembly's authority to engage in mid-decade redistricting when a decennial districting plan is found to violate the Constitution or federal law. However, when addressing an analogous question regarding the North Carolina Constitution's Whole County Provision, which immediately follows the constitutional prohibitions on mid-decade redistricting, the Supreme Court of North Carolina held that "[f]ederal law . . . preempts the State Constitution only to the extent that the [provision] actually conflicts with the VRA and other federal requirements relating to state legislative redistricting and apportionment." Stephenson I, 562 S.E.2d at 396. The North Carolina Supreme Court further held that because it has an obligation to follow the policies established by the people of North Carolina in their Constitution "whenever possible," the redistricting provisions in the North Carolina Constitution "must be enforced to the maximum extent possible." Id. at 396-97 (emphasis added). In light of this reasoning, we read Stephenson I as likewise requiring that the North Carolina Constitution's prohibition on mid-decade redistricting "be enforced to the maximum extent possible." Id. Therefore, unless required by federal law or a judicial order, Sections 3(4) and 5(4) preclude the General Assembly from engaging in mid-decade redistricting.

As explained above, the Supreme Court of the United States' decision in Upham requires that a federal district court's remedial order not unnecessarily interfere with state redistricting choices. 456 U.S. at 40-41; see also Johnson v. Miller, 922 F. Supp. 1556, 1559 (S.D. Ga. 1995) ("In fashioning a remedy in redistricting cases, courts are generally limited to correcting only those unconstitutional aspects of a state's plan."). When a court must draw remedial districts itself, this means that a court may redraw only those districts necessary to remedy the constitutional violation. Upham, 456 U.S. at 40-41; Personhuballah, 155 F. Supp. 3d at 563 (concluding that in order to comply with state policy "our chosen remedial plan should not alter any districts outside of the [racially gerrymandered district] and those abutting it"). Accordingly, our order did not—and could not—require the General Assembly to redraw districts that did not need to be redrawn to cure the constitutional violation.

Legislative Defendants did not put forward any evidence showing that revising any of the five Wake and Mecklenburg County House districts challenged by Plaintiffs was necessary to remedy the racially gerrymandered districts in those two counties. And both the Special Master's proposed map and Plaintiffs' alternative map establish that the racially gerrymandered House districts in Wake and Mecklenburg County could be remedied without redrawing those five districts. Accordingly, there is no "actual[] conflict" between this Court's order and the mid-decade redistricting prohibition. Stephenson I, 562 S.E.2d at 396. Therefore, we conclude the General Assembly exceeded its authority under our order by disregarding the mid-decade redistricting prohibition. See id. at 388 ("Because Congress has not preempted the entire field of state legislative redistricting and reapportionment, state provisions in this area of law not otherwise superseded by federal law must be accorded full force and effect." (citations omitted)); Cleveland Cty. Ass'n for Gov't by People v. Cleveland Cty. Bd. of Comm'rs, 142 F.3d 468, 477 (D.C. Cir. 1998) (holding that contravention of North Carolina state law governing the at-large election of county commissioners was not warranted as it was not necessary to remedy any violation of federal law or otherwise permitted by a special enactment by the state legislature).

Legislative Defendants nevertheless argue that adopting a standard that permits changes only to those districts not directly impacted by the racial gerrymander—districts that violate the Constitution, abut a district violating the Constitution, or otherwise need to be altered in order to ensure compliance with federal law or state constitutional provisions—would perpetuate a racial gerrymander by "forcing a legislature to use the core of [a] racially gerrymandered district to draw the new district and those immediately surrounding it." Leg. Defs.' Objs. Resp. 52. In particular, for those districts not directly impacted by the racial gerrymander such a standard would "reduce or eliminate the legislature's ability to eliminate the hallmarks of gerrymanders by, for instance, eliminating split precincts, or changing surrounding districts to more closely follow municipal boundaries." Id.

But our opinion does not endorse a legislature's preservation of an unconstitutional district's "core" in drawing a remedial district. On the contrary, we find that several of the General Assembly's proposed districts failed to remedy the constitutional violation precisely because they preserved the "core" of the unconstitutional version of the districts. See supra Part III.A. And we do not hold that a state redistricting body tasked with drawing a remedial plan can never redraw districts that were not found to violate the Constitution or abut such a district. Indeed, if Legislative Defendants had put forward evidence establishing that redrawing additional districts was necessary to completely remedy the racial gerrymander, then our Order would have authorized the redrawing of such districts. Covington III, 2017 WL 3254098, at *3 (providing the General Assembly with the opportunity to "enact new House and Senate districting plans remedying the constitutional deficiencies"). Legislative Defendants, however, put forward no such evidence. And the Special Master's Recommended Plans for the Wake and Mecklenburg County House districts demonstrate that one can remedy the racial gerrymander—and not preserve the "cores" of the unconstitutional districts—without redrawing districts untainted by the constitutional violations. See infra Part IV.B.5-6.

Additionally, the Supreme Court has recognized that "racial gerrymandering claim[s] . . . appl[y] to the boundaries of individual districts." Alabama, 135 S. Ct. at 1265 (emphasis added). Accordingly, remedying a racial gerrymandering violation generally entails redrawing the "boundaries of [those] individual districts," id., not redrawing a districting plan as a whole, as Legislative Defendants' argument suggests. And regardless of whether splitting precincts or failing to follow municipal precinct lines is good from a policy perspective, the failure to follow such policies does not render a state redistricting plan unconstitutional. Bush v. Vera, 517 U.S. 952, 962 (1996) (opinion of O'Connor, J.) (explaining that "the neglect of traditional districting criteria is . . . not sufficient" to establish a racial gerrymandering claim); cf. Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 798 (2017) ("Race may predominate even when a reapportionment plan respects traditional principles."). Rather, a district amounts to a racial gerrymander if, in drawing the district, "race predominated over traditional race-neutral redistricting principles." Covington I, 316 F.R.D. at 129 (quoting Shaw v. Hunt, 517 U.S. 899, 908 (1996)).

2.

The North Carolina Constitution's Whole County Provision states that "[n]o county shall be divided in the formation of a [representative or] senate district." N.C. Const. art. II, §§ 3(3), 5(3). In Stephenson I, the Supreme Court of North Carolina recognized that the Whole County Provision must give way to federal law, including the Equal Protection Clause and VRA. 562 S.E.2d at 396 ("Although we discern no congressional intent, either express or implied, to preempt the WCP through the operation of the VRA, we also recognize that the WCP may not be interpreted literally because of the VRA and the 'one-person, one-vote' principles."). The North Carolina Supreme Court further held, however, that the Whole County Provision "should be adhered to by the General Assembly to the maximum extent possible." Id. at 391. To that end, the court identified a complex set of nine criteria governing the General Assembly's application of the Whole County Provision in redistricting.

Of particular relevance, one criterion provides that "[w]hen two or more non-VRA legislative districts may be created within a single county, which districts shall fall at or within plus or minus five percent deviation from the ideal population consistent with 'one-person, one-vote' requirements, single-member non-VRA districts shall be formed within said county." Id. at 397. And another criterion provides detailed guidance regarding the drawing of districts encompassing "counties having a non-VRA population pool which cannot support at least one legislative district at or within plus or minus five percent of the ideal population for a legislative district or, alternatively, counties having a non-VRA population pool which, if divided into districts, would not comply with the at or within plus or minus five percent 'one-person, one-vote' standard." Id. In such counties, the General Assembly must "combin[e] or group[] the minimum number of whole, contiguous counties necessary" to comply with one-person, one-vote. Id. In the county groupings, district lines must not traverse the "exterior" line of the county group. Id. "[I]nterior county lines created by any such groupings may be crossed or traversed in the creation of districts within said multi-county grouping but only to the extent necessary" to comply with one-person, one-vote. Id. (emphasis added). Moreover, because "the intent underlying the WCP must be enforced to the maximum extent possible[,] . . . only the smallest number of counties necessary to comply with the at or within plus or minus five percent 'one-person, one-vote' standard shall be combined." Id.

Plaintiffs argue that districts drawn in two county groupings violate these criteria. First, notwithstanding that "Cabarrus County has the population to justify more than two house districts," the remedial House plan includes only one district, House District 82, wholly within Cabarrus County. Pls.' Objs. 39-40. According to Plaintiffs, the plan's failure to draw two districts within Cabarrus County violates the requirement that the Whole County Provision be maximally enforced and "interior" county lines be traversed "only to the extent necessary." Id.

By contrast, Legislative Defendants argue that the Cabarrus County group complies with the Whole County Provision as construed in Stephenson I because although it does not maximize the number of districts wholly contained within a single county, it minimizes the number of traversals. Leg. Defs.' Objs. Resp. 54 ("[E]ach grouping must contain the fewest number of traversals possible in creating districts which comply with equal population requirements."). Put differently, according to Legislative Defendants, the Whole County Provision requires minimizing the number of traversals, not the number of multi-county districts in a grouping. To that end, Legislative Defendants also point out that within the relevant county cluster, the Plaintiffs' alternative plan has more traversals of county lines compared with the 2017 Enacted House Plan. Id. at 55. In addition, the Plaintiffs' proposed plan alters HD 67 to spread it across three separate counties.

Notwithstanding its extended discussion of the Whole County Provision in Stephenson I, the Supreme Court of North Carolina has not expounded on the proper application of that provision within a multi-county cluster, the issue here, much less whether the Whole County Provision requires maximizing the number of districts wholly contained within a single county or minimizing the number of county-line traversals in the grouping. Given that this is an unsettled question of state law and support exists for each party's position, we exercise our discretion not to exercise pendent jurisdiction over Plaintiffs' objection related to the Cabarrus County grouping. See supra Part II.C; Robertson, 148 F. Supp. 2d at 461-62.

Second, Plaintiffs argue that the county grouping including Greene County fails to comply with the Whole County Provision because it adds population from two counties (Johnston and Wayne) to a county with insufficient population to make a district (Greene), when it is only necessary to add population from one county (Wayne). In support of their position, Plaintiffs rely on Stephenson I's statement that in creating county groupings, the General Assembly must combine the "smallest number of counties necessary to comply with the . . . 'one-person, one-vote requirement.'" 562 S.E.2d at 396. That requirement, however, dealt with the creation of county groupings, not with the drawing of interior district lines within a county grouping, the relevant question. Id.

Legislative Defendants again argue that the Greene County configuration complies with the Whole County Provision because it minimizes the number of traversals in the multi-county group. Legislative Defendants further note that Plaintiffs' proposed plan fails to demonstrate that it would be feasible to implement an alternative plan that would minimize such traversals. The Supreme Court of North Carolina has not addressed whether, in the context of a multi-county grouping, the Whole County Provision requires minimizing the number of counties a particular district spans or minimizing the number of county-line traversals in the grouping as a whole. In light of the absence of such guidance from North Carolina courts, we again exercise our discretion not to exercise pendent jurisdiction over Plaintiffs' objection related to the Greene County grouping. See supra Part II.C; Robertson, 148 F. Supp. 2d at 461-62.

By adopting Plaintiffs' alternative plan, House District 28 would span 3 counties, whereas the version in the 2017 Plan spans only 2 counties, presumably in violation of the Plaintiffs' own purported constitutional rule.

Our decision not to exercise pendent jurisdiction over Plaintiffs' objections related to the Cabarrus and Greene County groupings is made without prejudice to Plaintiffs or other litigants asserting such arguments in separate proceedings. We note that there are ongoing proceedings in state court regarding North Carolina's legislative districting plans. See Dickson v. Rucho, 804 S.E.2d 184, 185 (N.C. 2017) (remanding case to trial court to determine whether (1) in light of Cooper v. Harris and North Carolina v. Covington, a controversy exists or if this matter is moot in whole or in part; (2) there are other remaining collateral state and or federal issues that require resolution; and (3) other relief may be proper").

3.

Finally, Plaintiffs argue that Senate District 41 violates the Whole County Provision because it is "grossly non-compact." Pls.' Objs. 41. As noted above, the Whole County Provision provides that "[n]o county shall be divided in the formation of a [representative or] senate district." N.C. Const. art. II, §§ 3(3), 5(3). Accordingly, the plain language of that provision does not address compactness. And in its most recent discussion of the Whole County Provision, the Supreme Court of North Carolina stated that lack of compactness does not "constitut[e] an independent basis for finding a violation, and we are unaware of any justiciable standard by which to measure [lack of compactness]." Dickson II, 781 S.E.2d at 440. Given that the Whole County Provision does not mention compactness and the Supreme Court of North Carolina has stated that lack of compactness is not an "independent" basis for striking down an otherwise legal district, we reject Plaintiffs' objection to Senate District 41.

* * * * *

In conclusion, we sustain Plaintiffs' state-law objections as to remedial House Districts 36, 37, 40, 41, and 105, decline to consider Plaintiffs' state-law objections related to the Cabarrus and Greene County groupings, and reject Plaintiffs' state law objection related to proposed remedial Senate District 41.

IV.

Having sustained Plaintiffs' objections to the Subject Districts, this Court now must assume the "unwelcome obligation" of drawing remedial districting configurations for the Subject Districts. Perry, 565 U.S. at 392 (quoting Connor v. Finch, 431 U.S. 407, 415 (1977)). To that end, we now consider whether the Special Master's Recommended Plans remedy both the 2011 Plan's constitutional violations and the aspects of the 2017 Plan that render the Subject Districts legally unacceptable; comply with governing law; and adhere, to the extent possible, with the General Assembly's legitimate redistricting objectives. See Personhuballah, 155 F. Supp. 3d at 561-65 (examining whether remedial plan prepared by Special Master (1) complied with one-person, one-vote requirement; (2) remedied the identified racial gerrymander; (3) conformed, to the extent possible, with legislative policies embraced in the existing plan; and (4) otherwise complied with governing law); Johnson, 922 F. Supp. at 1561-69 (same).

Legislative Defendants reassert their argument that the General Assembly is entitled to a second opportunity to redraw the the Subject Districts. As this Court previously explained in rejecting that argument, "[t]he State is not entitled to multiple opportunities to remedy its unconstitutional districts." Appointment Order 4 (citing Reynolds, 377 U.S. at 585-87). To that end, numerous courts have imposed their own remedial redistricting plan after a proposed governmental plan failed to remedy the identified violation or was otherwise legally unacceptable. See, e.g., Large, 670 F.3d at 1148-49 ("[W]e AFFIRM the district court's order that rejected the County's proffered Section 2 remedial plan and implemented a plan of its own design."); Jeffers, 756 F. Supp. at 1200; Osceola Cty., 474 F. Supp. 2d at 1256. Legislative Defendants identify no authority to the contrary. That providing the General Assembly with a second bite at the apple would further draw out these proceedings and potentially interfere with the 2018 election cycle further militates against providing the General Assembly with such an opportunity.

A.

We first examine the Recommended Plans as a whole and find no deficiencies in—and instead, many marked improvements over—the related districts in the 2017 Plan. The Special Master's Recommended Plans comply with one-person one-vote requirements, i.e., all population deviations are within the restrictions imposed by the Equal Protection Clause. See Alabama, 135 S. Ct. at 1263 ("[A] 5% deviation from ideal[—(i.e., perfectly equipopulous districts)—is] generally permissible." (citing Brown v. Thomson, 462 U.S. 835, 842 (1983))). The recommended districts are consistently more compact under the compactness measures preferred by the General Assembly, with an average increase—as compared to the 2017 Plan—of 13.5 percent in the Reock metric and 11.5 percent in the Polsby-Popper metric. See Rec. Plan & Rep. 26. Further, the revised districts in the Recommended Plan split 5 fewer precincts and 2 fewer municipalities than their counterparts in the 2017 Plan. Id. at 22, 24, 29. The Recommended Plans also cure the constitutional violation by not tracking the contours of their racially gerrymandered versions, and not dividing municipalities and counties along racial lines. See id. at 21-22, 31, 34, 40-41, 45-47. And the recommended reconfigurations of the Wake and Mecklenburg County House districts remedy the racial gerrymanders in the 2011 Plans, while preserving those districts from the 2011 Plan untainted by the unconstitutional districts and retaining the features of the 2017 Plan as much as possible. Id. at 56-68.

Generally, courts must strive to draw remedial plans that are as close to equipopulous as possible. See Abrams, 521 U.S. at 98 ("Court-ordered districts are held to higher standards of population equality than legislative ones."). Some of the districts in the Recommended Plans hew closely to the 5 percent maximum population deviation selected by the General Assembly and authorized in the Court's Appointment Order. Rec. Plan & Rep. 18. These larger deviations results from the fact that "the Whole County Provision of the State Constitution requires working within a county grouping to achieve equipopulous districts." Id. No party takes issue with the population deviations in the Special Master's Recommended Plans.

Before examining the Recommended Plans' performance on a district-by-district basis, we first address three objections by Legislative Defendants to the Recommended Plan as a whole: (1) that, in drawing the Recommended Plan, the Special Master impermissibly sought to achieve a specific BVAP quota by "systematically reduc[ing] the [BVAP] in each district"; (2) that the Recommended Plan fails to advance several of the General Assembly's stated or revealed political objectives; and (3) that the Special Master impermissibly drew the plan to favor the Democratic party.

1.

First, Legislative Defendants contend the Special Master "single-minded[ly] focus[ed] on race" and that "the special master's fixation on a racial 'residuum' was used to lower the BVAP of each district to an undisclosed target level." Leg. Defs.' Rec. Plan Resp. 4, 6. This argument wholly disregards the instructions this Court provided to the Special Master—and the Special Master's careful adherence to those instructions—and amounts to a baseless attack on the Special Master's integrity and credibility.

This Court's Appointment Order governing the drawing of the remedial districts did not direct the Special Master to pursue any BVAP target in drawing the remedial districts. Appointment Order ¶ 2. Rather, it stated that the Special Master could "consider data identifying the race of individuals or voters to the extent necessary to ensure that his plan cures the unconstitutional racial gerrymanders and otherwise complies with federal law." Id. at ¶ 2(i).

The Special Master credibly and unambiguously stated that, in drawing the Recommended Plans, "no racial targets were sought or achieved." Special Master's Rec. Plan for the N.C. Sen. & House of Rep. ("Special Master Hr'g Pres.") 37, Jan. 5, 2018, ECF No. 239; Hr'g Tr. 26:8-9. Likewise, the Special Master averred that in accordance with the Court's instructions, "the remedial districts were drawn not with any racial target in mind, but in order to maximize compactness, preserve precinct boundaries, and respect political subdivision lines." Rec. Plan & Rep. 21. To that end, the "Special Master's Plan removes the racial predominance of the [racially gerrymandered districts in the 2017 Plan] by replacing the constitutionally tainted districts with others that adhere to explicitly race-neutral criteria." Id. at 21. In particular, the Recommended Plans "do[] not preserve the core shape of the unconstitutional version of the district[s], avoid[] dividing counties and municipalities, and attempt[] to enhance compactness," the Special Master explained. Id. at 22. The Recommended Plans achieved those goals, more effectively respecting precinct and municipal lines than the 2017 Plan's versions and improving on the measures of compactness embraced by the General Assembly. See supra Part IV.A; infra Part IV.B. Accordingly, Legislative Defendants' BVAP targeting argument amounts to a claim that the Special Master made false representations to the Court regarding the approach he followed in drawing the Recommended Plans.

In support of their attack on the Special Master's plans, Legislative Defendants rely on a report and opinion by their proffered expert in census data and geography in redistricting, Dr. Douglas Johnson, who Legislative Defendants retained after they had already filed their Response asserting that the Special Master impermissibly pursued racial targets. Hr'g Tr. 78:19-21, 90:7-8, 104:19-22. Dr. Johnson opined as to the Special Master's "[a]pparent [p]redominant [u]se of [r]ace [d]ata" and that "certain racial quotas were targeted by the Special Master when drawing the districts" or "dictated the configuration" of the districts. Expert Rep. of Douglas Johnson, Ph.D. ("Johnson Rep.") 13, 15, 20, Dec. 27, 2017, ECF No. 234-1; see also Hr'g Tr. 78:17-19 (opining as to the Special Master's "apparent quota of the African-American percentage of the voting-age population").

In support of his opinion, Dr. Johnson (a) points to "the remarkable similarity in the African-American percentages of the Voting Age Population in the districts drawn by the Speical Master"; (b) highlights that the Recommended Plans reduced the BVAP in all of the racially gerrymandered districts in the 2017 Plan; and (c) notes that, for several of the racially gerrymandered districts, Dr. Johnson was able to draw a remedial configuration that, he maintained, more effectively advanced the General Assembly's objectives without bringing the district's BVAP "into the Special Master's remarkably consistent [BVAP] range for his adjusted districts." Id. at 13-25. For several reasons, we find Dr. Johnson's analysis and opinion as to the alleged racial targeting in the Recommended Plans unreliable and not persuasive.

To begin, we fail to see how the alleged "remarkable similar[ity]" in the BVAP for districts redrawn in the Special Master's Recommended Plan proves that the Special Master drew his Recommended Plans to achieve a specific target BVAP. Dr. Johnson notes that Recommended Senate Districts 21 and 28 and House Districts 21 and 57 have BVAPs ranging from 38 percent to 44 percent, Johnson Rep. 14—a range Legislative Defendants characterize as "narrow," Leg. Defs.' Rec. Plan Resp. 7. But Dr. Johnson conceded that the fact that several districts' BVAPs fall in a particular range does not prove that "a racial quota was being employed." Hr'g Tr. 98:24-99:6.

Additionally, "correlation [is] not evidence of causation." Brown v. Entm't Merchants Ass'n, 564 U.S. 786, 800 (2011). To the extent the BVAPs of those four districts are "remarkabl[y] similar[]"—and Dr. Johnson provides no empirical basis for determining whether the BVAPs of two districts are "similar" from a statistical perspective—any such similarity may be attributable to the underlying demographic make-up of the geographic areas in which the districts are drawn or other non-discriminatory districting considerations, not racial targeting. See Tagatz v. Marquette Univ., 861 F.2d 1040, 1044 (7th Cir. 1988) (Posner, J.); Ste. Marie v. E. R.R. Ass'n, 650 F.2d 395, 400 (2d Cir. 1981) (Friendly, J.). And neither Legislative Defendants nor Dr. Johnson offer any controlled statistical analysis ruling out non-discriminatory explanations for the four districts' BVAPs. Absent such evidence, we find that the BVAPs themselves do not prove that the Special Master, contrary to his unambiguous statements to the Court, engaged in racial targeting.

The Special Master credibly explained why BVAPs decreased in Senate Districts 21 and 28 and House Districts 21 and 57. As he stated in his report, "[t]he fact that the districts happen to reduce the [BVAP] in the redrawn districts, while increasing it in adjoining districts, is to be expected whenever a plan replaces racial predominance with other redistricting principles." Rec. Plan & Rep. 19. Additionally, the Special Master noted that House District 33, which was a racial gerrymander in the 2011 Plan, had a slightly higher BVAP in the Recommended Plan, meaning that, contrary to Dr. Johnson's analysis, the Recommended Plan did not universally decrease the BVAP in redrawn districts that were previously racially gerrymandered. Accordingly, we find that the reduced BVAP in the four districts fails to demonstrate that the Special Master engaged in racial targeting.

Finally, Dr. Johnson provided one alternative configuration for several of the districts in the Recommended Plan, which, according Dr. Johnson, have lower BVAPs and somewhat more effectively adhere to several traditional redistricting criteria, like compactness and population equality. Even assuming Dr. Johnson is correct that his configurations more effectively advance these criteria—and reasonable minds could differ as to that conclusion—Legislative Defendants cite no legal authority for the proposition that being able to produce a single alternative districting configuration that somewhat improves on certain districting considerations, while reducing a district's BVAP, establishes that that a mapdrawer intentionally engaged in racial targeting. On the contrary, the Supreme Court has recognized that "a State could construct a plethora of potential maps that look consistent with traditional, race-neutral principles," some of which may involve impermissible racial targeting, and others of which may not. See Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 799 (2017); Vera, 517 U.S. at 967 ("If, as may commonly happen, traditional districting principles are substantially followed without much conscious thought, they cannot be said to have been 'subordinated to race.'"). Likewise, Dr. Johnson conceded that minor differences between two proposed maps do not signal that one version is legally unacceptable or better achieves traditional redistricting goals. Hr'g Tr. 92:23-93:3.

Legislative Defendants did not offer these alternative configurations as a potential replacement for either the related Subject District or for the Recommended Plans. Rather, Legislative Defendants solely offered these alternative configurations to criticize the Recommended Plans. See Hr'g Tr. 87:1-88:12

For example, Dr. Johnson's rendering of Senate District 28 in Guilford County less closely tracks Greensboro's municipal boundaries than the Recommended Plan's version. Compare Johnson Rep. 23, with Rec. Plan & Rep. 39.

Beyond the alleged similarities in the districts' BVAPs and Dr. Johnson's alternative maps, Legislative Defendants offer no other direct or circumstantial evidence indicating that the Special Master used racial targets in drawing the districts' lines. Legislative Defendants' failure to put forward such evidence is particularly notable when compared with the extensive direct, circumstantial, and expert evidence that this Court relied upon both to find the that 2011 Plan relied unjustified race-based districting, Covington I, 316 F.R.D. at 130-65, and to find that the 2016 Plan failed to remedy the identified racial gerrymanders in Senate Districts 21 and 28 and House Districts 21 and 57, see supra Part III.A.1-4.

In sum, Dr. Johnson's report and testimony do not in any way call into question the Special Master's repeated, credible, and unambiguous statements—made in his capacity as an officer of the Court—that he did not engage in racial targeting, and that any changes to the BVAP of districts in his Recommended Plan are attributable to his efforts to achieve the non-discriminatory redistricting objectives set forth in this Court's Appointment Order.

2.

Legislative Defendants next contend that the Recommended Plans fail to achieve several of the General Assembly's statewide or district-specific political objectives. In particular, Legislative Defendants assert that certain districts in the Recommended Plan fail to accomplish the legislature's goals of ensuring that a Republican candidate had an opportunity to prevail in a particular district or that a particular incumbent would win in his new district. Leg. Defs.' Rec. Plan Resp. 9, 15.

But the Supreme Court long has held that courts lack "political authoritativeness" and, therefore, must act "in a manner free from any taint of arbitrariness or discrimination" in drawing remedial districts. Wise v. Lipscomb, 437 U.S. 535, 541 (1978) (quoting Connor v. Finch, 431 U.S. 408, 417 (1977)). To that end, in drawing a remedial plan, a court may not draw district lines solely to advance partisan or political objectives, even when the state redistricting body expressly adopted such objectives. See, e.g., Wyche v. Madison Par. Police Jury, 769 F.2d 265, 268 (5th Cir. 1985) ("Many factors, such as the protection of incumbents, that are appropriate in the legislative development of an apportionment plan have no place in a plan formulated by the courts."); Wyche v. Madison Par. Police Jury, 635 F.2d 1151, 1160 (5th Cir. 1981) (noting that "a court is forbidden to take into account the purely political considerations that might be appropriate for legislative bodies"); Essex v. Kobach, 874 F. Supp. 2d 1069, 1093 (D. Kan. 2012) (declining to unpair certain incumbents in remedial district plan because "any efforts to protect [such] incumbents would require our choosing among incumbents, an inherently political exercise we are neither able nor inclined to undertake"); Colleton Cty. Council v. McConnell, 201 F. Supp. 2d 618, 668 (D.S.C. 2002) ("[E]ven were we to agree that [a proposed change to a district configuration] had some political benefit, such an important change to the core of an existing district in a [court-drawn] redistricting plan, based on nothing more than our determination that one election official will do a better job than another, is clearly beyond the scope of our remedial authority."). Accordingly, the Special Master's alleged failure to achieve the General Assembly's partisan objectives in no way calls into question the legal adequacy of the Recommended Plans.

3.

Finally, Legislative Defendants maintain that the Special Master drew the Recommended Plans to favor Democrats. Leg. Defs.' Rec. Plan Resp. 2. The only support Legislative Defendants provide for this assertion is an article in the Raleigh News & Observer, which opined that Democratic candidates had a better chance of prevailing in several of the districts in the Recommended Plans than in such districts' counterparts in the 2017 Plans. See id. (citing Colin Campbell & Bruce Henderson, Redrawn Election Maps Would Help Democrats, News and Observer, Nov. 28, 2017, at 2A). Even assuming that the reporters are correct that the Recommended Plans are more favorable to Democratic candidates than the 2017 Plans—and Legislative Defendants introduced no analysis of their own showing that that is in fact the case—that does not establish that the Special Master drew the districts to favor Democrats. See Brown, 564 U.S. at 800 ("[C]orrelation [is] not evidence of causation."). Rather, any adverse consequences on the electoral prospects of Republican candidates may simply derive from the Special Master's duty to draw plans that completely eliminate the vestiges of the racial gerrymanders, rather than an intentional effort to benefit any candidate of either political party. And Legislative Defendants present no evidence, much less a rigorous empirical analysis, demonstrating that the Special Master could have drawn districts that completely remedied the racial gerrymander that were more favorable to Republican candidates.

More significantly, this Court's Appointment Order barred the Special Master from taking into account political considerations in drawing his remedial plans, except for the purpose of preventing the pairing of incumbents. Appointment Order 7. Legislative Defendants offer no evidence that the Special Master disregarded this instruction. On the contrary, the Special Master repeatedly averred that he complied with all of the Court's instructions set forth in the Appointment Order, including the instruction that he take a nonpartisan approach in drawing his Recommended Plans. Hr'g Tr. 8:23-9:16; Rec. Plan & Rep. 11 (stating that the Special Master's "nonpartisan approach . . . is absolutely critical to bolstering the legitimacy of the Special Master's Plan"). And the Special Master took a number of steps "[t]o avoid even the appearance of partisanship," including rejecting Plaintiffs' proposed plans as unduly favorable to Democratic candidates and unpairing incumbents of both parties, notwithstanding that Legislative Defendants never requested that the Special Master unpair Republican incumbents. Rec. Plan & Rep. 12-14, 30. Accordingly, Legislative Defendants' assertion that the Recommended Plans were drawn to favor Democratic candidates finds no record support.

Legislative Defendants also take issue with what Dr. Johnson describes as the Special Master's "bewildering[]" labeling of municipality splits as "Municipalities (CDPs)," Johnson Report 6—a critique they failed to raise in commenting on the Special Master's draft plan. "Municipalities" and "CDPs" differ insofar as municipalities are officially recognized local governments within a particular state, whereas CDPs are "settled concentrations of population that are identifiable by name but are not legally incorporated under the laws of the state in which they are located." Id. at 5. Dr. Johnson asserts that the Special Master's data labeling indicates that he potentially conflated the two terms and therefore may have failed to correctly ascertain the number of municipal splits.
Dr. Johnson conceded, however, that he "d[id] not have sufficient time . . . to rerun the [Special Master's 'Municipalities (CDPs)'] tables using only municipalities." Id. at 6. Therefore, his opinion that "the tables would show different results if only the 533 municipalities are analyzed instead of [what] . . . the Special Master appears to have used in his analysis" lacks any empirical basis. Id. at 6. The Special Master responded directly to Dr. Johnson's criticism at the hearing, credibly explaining that although there are differences between CDPand municipalitybased boundaries, the few minor differences in the relevant North Carolina districts in no way materially affected the boundaries and municipalitysplit calculations in the Recommended Plan. Hr'g Tr. 10:711:8; see also Special Master Hr'g Pres. 58. Accordingly, we find this alleged deficiency in the Special Master's Recommended Plan to be without merit.

B.

Having rejected Legislative Defendants' objections to the Recommended Plans as a whole, we now examine the Recommended Plans on a district-by-district basis.

1. Senate District 21

Like the version of Senate District 21 in the 2011 and 2017 Plans, the Special Master's Recommended Senate District 21 encompasses all of Hoke County and a portion of Cumberland County. Rec. Plan & Rep. 31. But unlike the General Assembly's proposed remedial version in the 2017 Plan, the Special Master's recommended version no longer retains the core shape of the Cumberland County portion of the unconstitutional 2011 version of the district. Id. at 32. Most notably, Recommended Senate District 21 no longer includes proposed remedial Senate District 21's "long extension into Fayetteville that seems surgically designed to capture heavily African American precincts, while evading heavily white precincts." Id. at 31; see supra Part III.A.1.

Recommended Senate District 21 and its partner in the Hoke-Cumberland grouping, Recommended Senate District 19, satisfy the Constitution's one-person, one-vote requirement. Id. at 33. Both recommended districts improve on the compactness of their counterparts in the 2017 Plan under the measures of compactness adopted by the General Assembly. Id. And the Recommended Plan's configuration reduces the number of split precincts and municipalities in both districts, in accordance with the Adopted Criteria. Id.

Legislative Defendants object to Recommended Senate District 21 on two grounds. First, they claim that its lines are the product of intentional racial targeting, Leg. Defs.' Rec. Plan Resp. 8-10—a contention we already have rejected, see supra Part IV.A.1. Second, Legislative Defendants assert that the Recommended Plan violates the General Assembly's political decision to "place the Fort Bragg precinct in [Senate District] 19 . . . in order to provide the Republican incumbent . . . with an opportunity to win that district." Leg. Defs.' Rec. Plan Resp. 9. However, as noted previously, a court—or a special master acting on a court's behalf—is barred from considering partisan or political objectives in drawing a remedial districting plan. See supra Part IV.A.2. And even if a court tasked with drawing a remedial districting plan was entitled to give effect to partisan objectives—like ensuring the Republican incumbent would prevail in his new district—any legislative interest in protecting an incumbent must yield to remedying the unconstitutional racial gerrymander if necessary. See supra Section III.A. Therefore, we reject Legislative Defendants' objections and approve the Recommended Plan's reconfiguration of Senate District 21.

2. Senate District 28

Like the General Assembly's proposed remedial version, Recommended Senate District 28 lies in the center of Guilford County. Unlike the General Assembly's proposed remedial version of the district, Recommended Senate District 28—which takes on a highly compact circular shape almost wholly within the municipal boundaries of Greensboro—no longer divides Greensboro along racial lines, nor does it track the contours of Dr. Hofeller's VRA exemplar. Rec. Plan & Rep. 34-36; see supra Part III.A.2.

Recommended Senate District 28 abuts Senate Districts 24 and 27; however, the Special Master's Recommended Plan leaves the version of Senate District 24 in the 2017 Plan largely unchanged. See Rec. Plan & Rep. 35. The recommended configuration decreases Senate Districts 27's population deviation by 2.0 percentage points, and increases Senate District 28's population deviation by 0.5 percent. Id. at 36. Both Recommended Senate District 27 and 28 improve on their counterparts in the 2017 Plan in terms of the compactness measures included in the Adopted Criteria. Id. And in accordance with the Adopted Criteria, the recommended districts split fewer municipalities and precincts than their counterparts in the 2017 Plans—Senate District 27 would split one fewer precinct and Senate District 28 would split two fewer precincts and one fewer municipality. Id.

In addition to reasserting their unsupported contention that Recommended Senate District 28 was the product of racial targeting, Legislative Defendants also object to the recommended configuration because two incumbents are paired in Recommended Senate District 27. Leg. Defs.' Rec. Plan Resp. 11-12. But neither Legislative Defendants nor Plaintiffs asked the Special Master to unpair the incumbents—one of whom is a Democrat and one of whom is a Republican—notwithstanding that the Special Master expressly provided them an opportunity to suggest approaches for unpairing the incumbents. Rec. Plan & Rep. 37. And we find that the Special Master reasonably recommended against unpairing the incumbents because doing so "would require significant restructuring of the district" and that potential alternative plans for the districts would either take both incumbents "out of the territory that comprises most of their present districts" or significantly reduce the district's compactness. Id. at 37-38. Finding that the Recommended Senate District 28 cures the racial gerrymander and that Legislative Defendants' objections are without merit, we approve the Recommended Plan's reconfiguration of Senate District 28.

3. House District 21

Like its predecessor in the 2017 Plan, Recommended House District 21 spans a portion of Wayne County and the eastern edge of Sampson County. Rec. Plan & Rep. 42. Unlike the unconstitutional version of the district and version of the district in the 2017 Plan, Recommended House District 21 no longer includes a protrusion into central Sampson County to take in the majority-black sections of the City of Clinton, while excluding the city's majority-white sections. Id. at 40.

The recommended configuration of House District 21 and its neighbor, House District 22, satisfies the one-person, one-vote requirement. Id. at 43. And the recommended configuration also, on average, improves on the two districts' compactness, as measured by the General Assembly's two preferred metrics. Id. Recommended House District 21 has the same number of municipality or precinct splits as the version in the 2017 Plan, whereas Recommended House District 22 splits one fewer municipality than its counterpart in the 2017 Plan. Id.

Legislative Defendants again argue that Recommended House District 21's reduced BVAP relative to the version in the 2017 Plan is a product of BVAP targeting—a contention which finds no support in the record. See supra Part IV.A.1. Legislative Defendants further argue that Recommended House District 21 does not protect its incumbent as effectively as the version of the district in the 2017 Plan. But the Special Master was not authorized to draw a district to ensure an incumbent will prevail. See supra Part IV.B.2. Accordingly, we reject Legislative Defendants objections and approve the Special Master's remedial configuration of House District 21.

Even if the Special Master had been so authorized, the incumbent in House District 21 has stated under oath that he will not run for re-election in 2018, Decl. of Rep. Larry Bell, Nov. 10, 2017, ECF No. 211-1, meaning that there was no need for the Special Master to consider incumbent protection in redrawing the district. --------

4. House District 57

As with the version in the 2017 Plan, Recommended House District 57 lies wholly within Guilford County. However, unlike the 2017 Plan version, the Special Master's recommended version of the district no longer includes virtually all of the heavily black precincts in eastern Greensboro, which were included in Dr. Hofeller's Guilford County VRA exemplar. Rec. Plan & Rep. 45. And whereas the version of House District 57 in the 2017 Plan had a BVAP exceeding 60 percent—a substantially higher BVAP than its unconstitutional version—by no longer dividing Greensboro's precincts along racial lines, Recommended House District 57 has a BVAP of 38.4 percent. Id. at 50.

In order to reconfigure House District 57 to remedy the racial gerrymander, the Special Master had to reconfigure several other House districts in Guilford County (House Districts 59, 61, and 62). The Special Master's reconfiguration of those districts more effectively respects municipal boundaries than the 2017 Plan, containing three districts that lie almost entirely within Greensboro's city limits. Id. at 46-48. Additionally, pursuant to his obligation to respect the General Assembly's redistricting decisions to the extent possible, the Special Master maintained the shape of House Districts 58 and 60, as they were drawn in the 2017 Plan. Id. at 45 ("The Special Master's Recommended Plan redraws House District 57, but keeps intact the other "Subject Districts" (House Districts 58 and 60) as redrawn in the 2017 Plan."). Each of the reconfigured districts satisfies the one-person, one-vote requirement. Id. at 49. The Recommended Plan's configuration is as compact as the 2017 Plan, and more compact than the 2011 Plan, in accordance with the Adopted Criteria. Id. Further, the recommended configuration does not pair any incumbents, and each incumbent retains a majority of his or her constituency from the 2017 Plan. Id. at 51.

Legislative Defendants again argue that "the most significant difference in these two versions of [district] 57 is the BVAP," and that the "shape difference" between the two versions is "explained by policy decisions which had nothing to do with race." Leg. Defs.' Rec. Plan Resp. 16-17. However, Legislative Defendants nowhere identify the nature of these alleged "policy decisions" (stating only, "The 2017 district is based upon whole precincts located primarily in eastern Greensboro."), id., making it impossible for this Court to determine both whether the Special Master's recommended configuration in fact failed to advance those objectives and whether the Special Masters should have—or legally could have—advanced those objectives. Legislative Defendants also characterize Recommended District 61's increase in BVAP "from 11.5% to . . . 40.3%" as "astonishing," maintaining that the district "would have been labeled a racial gerrymander" if the General Assembly had recommended such a configuration. Id. at 17. But the Special Master did not target any BVAP percentage in drawing the Recommended Plans. See supra Part IV.A.1; Rec. Plan & Rep. 53. The increase in Recommended District 61's BVAP is attributable to shift of voters from the General Assembly's proposed House District 57, which had a BVAP exceeding 60 percent, into Recommended House District 61, and was therefore a consequence of the Special Master's obligation to remedy the racial gerrymander. Id. at 50. Thus, we reject Legislative Defendants' objections and approve the Special Master reconfiguration of the Guilford County House districts.

5. Wake County House Districts

As the Special Master correctly recognized, the problem with the Wake County House district configuration in the 2017 Plan—that the General Assembly violated the North Carolina Constitution by redrawing districts untainted by the constitutional violation—is "characteristically different" than the four districts in the 2017 Plan that failed to remedy the racial gerrymander. Id. at 56. The Special Master, therefore, took a different approach to reconfiguring the Wake County districts. Id. at 56-57. In particular, the Special Master first "reinstate[d]" the four untainted Wake County districts from the 2011 Plan that the General Assembly altered in the 2017 Plan. Id. at 57. Then, he reconfigured some of the remaining Wake County districts so as to cure the racial gerrymander, satisfy the one-person, one-vote requirement, and improve on the districts' compactness and adherence to precinct and municipal lines, as required by the Adopted Criteria. Id. at 57-58. The Special Master left intact two 2017 Plan districts, which he did not need to change to remedy the violation and made only minor changes to a third. Id. at 57.

The Recommended Wake County House plan satisfies the one-person, one-vote rrequirement. Id. at 60. The districts in the Special Master's recommended Wake County configuration are uniformly more compact and split fewer municipalities and precincts than those in the 2011 Plan configuration, in accordance with the Adopted Criteria. Id. at 60-61. The Special Master's configuration is slightly less compact, on average, than the 2017 Plan, and splits more municipalities and precincts. Id. These differences are attributable to the Special Master's obligation to reinstate the untainted districts in the 2011 Plan, which were less compact and split more municipalities and precincts than their counterparts in the 2017 Plan. Id. The Special Master's Recommended Plan does not pair any incumbents in Wake County. Id.

Legislative Defendants object to the Special Master's reconfiguration of the Wake County districts in his Recommended Plan on grounds that it unpaired two Democratic incumbents that were paired in his draft plan. Leg. Defs.' Rec. Plan Resp. 19-20. But the General Assembly's incumbency criterion expressed a preference for not pairing incumbents of "either party" in a district. Adopted Criteria for House and Senate Plans, Sept. 7, 2017, ECF No. 184-37. And in accordance with that legislative policy preference, this Court directed the Special Master to unpair incumbents if doing so would "not interfere with remedying the constitutional violations and otherwise complying with federal and state law." Appointment Order 7. The Special Master reasonably concluded that unpairing the Democratic incumbents—which required moving six precincts between the two districts and did not materially impact the Recommended Plan's compactness or respect for municipal and precinct boundaries—did not undermine the integrity of his plan. Rec. Plan & Rep. 62. Therefore, we again reject Legislative Defendants' objections and approve the Special Master's recommended reconfiguration of the Wake County House districts.

6. Mecklenburg County House Districts

Like the Wake County House district configuration, the Mecklenburg County House district configuration in the 2017 Plan unnecessarily, and therefore unconstitutionally, altered the version of House District 105 in the 2011 Plan, which was not impacted by the identified constitutional violation. Id. at 64. In redrawing the Mecklenburg County configuration, the Special Master restored the lines of House District 105 to those in the 2011 Plan and, as a result, had to somewhat alter only three adjoining districts (House Districts 92, 103, and 104). Id. at 64. In doing so, the Special Master sought "to keep precincts whole (outside of those already split by [the] 2011 [Plan's] District 105), to keep the districts in the area relatively compact and contiguous, and to make only the changes necessary to remedy the constitutional violation." Id. The Special Master's configuration is slightly less compact, on average, than that of the 2017 Plan, and splits more precincts. Id. These differences are attributable to the Special Master's obligation to reinstate the version of House District 105 in the 2011 Plan, which was noncompact and split a number of municipalities and precincts. Id. at 65-67. No party asserts any specific objection to the Special Master's reconfiguration. Therefore, we approve the Special Master's Recommended Plan for the Mecklenburg County House districts.

V.

Finally, we consider the remaining districts of the 2017 Plan unaffected by our decision today. We earlier found the following additional districts unconstitutional gerrymanders: Senate Districts 4, 5, 14, 20, 32, 38, and 40; and House Districts 5, 7, 12, 24, 29, 31, 32, 38, 42, 43, 48, 58, 60, 99, 102, 107. The General Assembly enacted the 2017 Plans to remedy the constitutional violations related to each of these districts. The Supreme Court has provided that "[t]he new legislative plan, if forthcoming, will then be the governing law unless it, too, is challenged and found to violate the Constitution." Wise, 437 U.S. at 540.

No party has raised a substantive challenge to any of these districts, and therefore no party has provided this Court with evidence that the 2017 Plans fail to remedy the constitutional violations we identified. In the absence of any finding that the remedial districts offend the Constitution or Voting Rights Act, these districts are entitled to the presumption of constitutionality afforded an enactment of a duly elected legislature. Upham, 456 U.S. at 43; Wise, 437 U.S. at 540. Under these circumstances, our district-by-district review cannot discern any apparent failure to adequately remedy the specific constitutional violation this Court identified. Therefore, the Court will approve and adopt the remaining remedial districts of the 2017 Plan for use in future elections in the State. See Shaw v. Hunt, No. 92-202-CIV-5-BR, slip op. at 8 (E.D.N.C. Sept. 12, 1997) (three-judge court approving remedial legislative plan enacted to remedy racial gerrymander in the absence of challenge by any party).

VI.

In conclusion, for the reasons stated above, we sustain Plaintiffs' objections to the Subject Districts and approve and adopt the State's 2017 Plan, as modified by the Special Master's Recommended Plan, for use in future North Carolina legislative elections. Accordingly, this Court's previous injunction against the State from conducting any elections for State House and State Senate offices, Order and Judgment, Aug. 15, 2016, ECF No. 125, is dissolved. We direct the Defendants to implement the Special Master's Recommended Plans.

SO ORDERED


Summaries of

Covington v. North Carolina

UNITED STATES DISTRICT COURT FOR THE MIDDLE DISTRICT OF NORTH CAROLINA
Jan 19, 2018
1:15CV399 (M.D.N.C. Jan. 19, 2018)
Case details for

Covington v. North Carolina

Case Details

Full title:SANDRA LITTLE COVINGTON, et al., Plaintiffs, v. THE STATE OF NORTH…

Court:UNITED STATES DISTRICT COURT FOR THE MIDDLE DISTRICT OF NORTH CAROLINA

Date published: Jan 19, 2018

Citations

1:15CV399 (M.D.N.C. Jan. 19, 2018)