THE CITY OF GREENSBORO et al v. GUILFORD COUNTY BOARD OF ELECTIONS, No. 1:2015cv00559 - Document 136 (M.D.N.C. 2017)

Court Description: MEMORANDUM OPINION AND ORDER signed by Judge Catherine C. Eagles on 04/03/2017. Plaintiff's request for a permanent injunction will be GRANTED, and a judgment and permanent injunction will be entered as time permits. In the meantime, the preliminary injunction 36 remains in force. (Coyne, Michelle)
Download PDF
THE CITY OF GREENSBORO et al v. GUILFORD COUNTY BOARD OF ELECTIONS Doc. 136 IN THE UNITED STATES DISTRICT COURT FOR THE MIDDLE DISTRICT OF NORTH CAROLINA CITY OF GREENSBORO, et al., Plaintiffs, v. GUILFORD COUNTY BOARD OF ELECTIONS, Defendant. ) ) ) ) ) ) ) ) ) ) 1:15-CV-559 MEMORANDUM OPINION AND ORDER Catherine C. Eagles, District Judge. The Equal Protection Clause of the Fourteenth Amendment to the United States Constitution requires that the votes of all citizens have equal weight. A state violates this rule, known as the one-person, one-vote principle, when it places voters into electoral districts of materially different population size for no legitimate reason. Even small deviations from equally populated districts violate the Equal Protection Clause if those deviations are driven by illegitimate factors. The Fourth Circuit recently held that a districting plan with a population deviation under ten percent violated the Equal Protection Clause because the unequal districts were created in an attempt to guarantee electoral success to one political party.1 As in that case, the evidence here establishes that the North Carolina General Assembly drew Greensboro City Council districts with materially unequal populations in an attempt to Raleigh Wake Citizens Ass’n v. Wake Cty. Bd. of Elections, 827 F.3d 333, 345, 351 (4th Cir. 2016) (hereinafter cited as RWCA). 1 Dockets.Justia.com maximize success for Republican candidates. Neither the State nor any legislative leaders defended this law in court or disputed the plaintiffs’ evidence, and the primary legislative sponsor refused to testify. Consistent with the holding in RWCA, the Court finds and concludes that the General Assembly’s 2015 City Council redistricting plan violates the Equal Protection Clause and must be permanently enjoined. I. BACKGROUND FACTS2 AND PROCEDURAL HISTORY Since 1983, the Greensboro City Council has been composed of nine members. Three council members and the mayor are elected at-large in citywide elections and five council members are elected from single-member districts.3 The City Council last redrew its districts in 2011, after the 2010 census.4 After the 2011 redistricting, the population in each district was almost equal, with only a 3.86 percent maximum population deviation5 in the plan as a whole; the largest district was only 2.34 percent above the ideal 2 Throughout this opinion, the facts stated are those the Court finds by a preponderance of the evidence. The Court has weighed the testimony at trial, the stipulations, and the documentary exhibits. See RWCA, 827 F.3d at 342 (noting preponderance of the evidence standard). The historical and demographic facts are undisputed, as reflected in the amended joint stipulations. See Doc. 126. The Board of Elections did not cross-examine witnesses, offer evidence, or make a substantive closing argument. The witnesses were largely credible and the testimony contained only minor discrepancies on a few non-material matters. The Court provides citations to some of the evidence that supports its factual findings but has not attempted to cite all the supporting evidence. The Court has disregarded evidence of non-material matters. 3 Doc. 126 at ¶ 13. 4 Id. at ¶ 17; see also id. at ¶ 18 (showing map of the five pre-existing districts). 5 The population deviation of an individual district is the percentage by which that district varies from the “ideal” population, i.e., the population of each district if all districts in the plan had equal populations. A district’s population deviation can be positive or negative, depending on whether the district is overpopulated or underpopulated. The “maximum population deviation” is the difference between the population deviations of the most overpopulated and most underpopulated districts. See Daly v. Hunt, 93 F.3d 1212, 1215 n.2 (4th Cir. 1996). 2 population size and the smallest district was only 1.53 percent below ideal population size. 6 In 2015, the North Carolina General Assembly passed a bill (“the Act”) that changed the City Council from this three-at-large, five-district council to an eight-district council and drew district lines for the new districts.7 According to the General Assembly’s numbers, the smallest of these new districts was 3.68 percent below the ideal population, the largest district was 4.57 percent above ideal, and the maximum population deviation in the new plan was 8.24 percent.8 This Act prohibited changes to the city’s government by the City Council; it also prohibited changes by citizen referendums and initiatives, which were otherwise available by statute to municipal citizens statewide.9 The Act also made other changes to the government of the City of Greensboro and the City of Trinity that are not at issue in this lawsuit.10 Almost immediately, several individual citizens of Greensboro filed suit, challenging both the population deviations between the new districts and the ban on citizen-initiated referendums and initiatives.11 The City of Greensboro joined in the latter Pls.’ Ex. 145 at 7 tbls.3-1 & 3-2. The Court’s calculations show that the evidence rounded these numbers to the second decimal place. 7 2015 N.C. Sess. Laws 138 sec. 2.(c), available at Pls.’ Ex. 33 and Doc. 1-1. Relevant parts of the Act would have been codified at Greensboro, N.C., Charter ch. III, subch. A § 3.01, subch. B § 3.23(b), subch. E § 3.81. 8 See Pls.’ Ex. 3 at 1; Pls.’ Ex. 145 at 7 tbls.3-1 & 3-2. The Court’s calculations show that the evidence rounded these numbers to the second decimal place. 9 See 2015 N.C. Sess. Laws 138 sec. 2.(b) (prohibiting changes under N.C. Gen. Stat. §§ 160A-101 to -111). 10 E.g., id. at sec. 2.(c) (setting four-year terms for council members). 11 See Doc. 1. 6 3 claim, and the Court issued a preliminary injunction blocking implementation of the Act.12 In September 2015, as part of a “technical corrections” bill, the General Assembly amended the Act to allow changes to Greensboro’s government by the City Council and by citizen initiatives and referendums after the 2020 census.13 This amendment did not change the Act’s district map.14 Under the injunction, the 2015 municipal elections took place under the pre-existing City Council format.15 The individual plaintiffs later added a racial gerrymandering challenge to one of the new districts.16 No one has substantively defended the constitutionality of the Act, directly or indirectly. A group of citizens who initially intervened, alleging that the Act was constitutional, decided to withdraw after concluding that any defense would be “futile.”17 The Attorney General decided not to participate on behalf of the State,18 and legislative leaders who had the statutory right to intervene did not seek to do so.19 The primary legislative sponsor of the Act invoked legislative privilege and refused to testify. 20 The State of North Carolina, the Governor, and legislative leaders in their official capacities have not defended the Act, and they appear to have immunity from suit under 12 Doc. 36. 2015 N.C. Sess. Laws 264 sec. 85.5, available at Pls.’ Ex. 36. 14 See id.; Pls.’ Ex. 35 at 3:4-:22. 15 See Doc. 36 at 2. See generally Pls.’ Ex. 78 (showing election results). 16 See Doc. 65 at ¶¶ 93-97. 17 Doc. 103 at 2; see Doc. 53; Text Order 12/07/2016. 18 Doc. 31-2 at 2. See generally Katherine Shaw, Constitutional Nondefense in the States, 114 Colum. L. Rev. 213 (2014). 19 See N.C. Gen. Stat. § 1-72.2. 20 See Doc. 79 (motion to quash); Doc. 125 (order granting motion to quash and adopting magistrate judge’s recommendation at Doc. 111). 13 4 the Eleventh Amendment to the Constitution.21 The only defendant is the Guilford County Board of Elections, the entity responsible for conducting elections for the City of Greensboro. The Board of Elections acknowledges that it is a necessary party to this litigation because of its responsibility for Greensboro elections.22 However, it had nothing to do with passing the Act and possesses no evidence about the redistricting process. It believes that its duty is to fairly and impartially administer whatever election laws validly apply, not to determine whether those laws are constitutional or to advocate on behalf of the State.23 The Board of Elections has offered no evidence or legal authority in support of the Act. II. ONE-PERSON, ONE-VOTE CLAIM The plaintiffs contend that the General Assembly drew the lines for the eight City Council districts in an attempt to guarantee a partisan advantage for Republicans.24 Specifically, they contend that the legislature packed Democratic-leaning voters into 21 See Wright v. North Carolina, 787 F.3d 256, 261-63 (4th Cir. 2015); see also City of Greensboro v. Guilford Cty. Bd. of Elections, No. 1:15-CV-559, 2016 WL 6810965 (M.D.N.C. Mar. 23, 2016), available at Doc. 72 (denying defendant-intervenors’ motion to dismiss or join parties). 22 See, e.g., Trial Tr. Vol. II, Doc. 134 at 336:12-341:23 (defendant’s closing argument). Citations to Doc. 134 use the transcript page numbers at the top-right of each page. 23 E.g., Doc. 100-1 at 5 (stating that the Board of Elections “performs only a ministerial role regarding elections” and “cannot[] function in a policy-making role”). 24 See RWCA, 827 F.3d at 346 (holding that “an attempt to guaranty Republican victory through the intentional packing of Democratic districts” is an illegitimate factor). The plaintiffs contend that there were other illegitimate factors, including race, that help prove their oneperson, one-vote claim, e.g., Doc. 132 at ¶ 4, but the Court need not address those factors in view of its resolution of the plaintiffs’ primary argument about partisan motives. 5 overpopulated districts and created underpopulated districts that lean Republican. The credible evidence is uncontroverted that this is indeed what happened. A. Legal Principles “The right to vote is ‘fundamental,’ and once that right ‘is granted to the electorate, lines may not be drawn which are inconsistent with the Equal Protection Clause of the Fourteenth Amendment.’”25 It is easy to see how this principle is violated when relatively small round numbers are used: If the District A representative is elected by 100 voters and the District B representative is elected by 1,000 voters, it is obvious that each of the voters in District A have significantly more power and that the votes of each District A and District B voter are not equal. Although one-person, one-vote cases typically involve larger total populations and smaller relative differences in populations than in this example, the principle is the same. This kind of “unequal apportionment” of voters violates the Equal Protection Clause.26 Districts that make votes “worth more in one district than in another would run counter to our fundamental ideas of democratic government.”27 The requirement that each vote have equal weight applies to all levels of elected government, from the United States Congress to county commissioners and school 25 RWCA, 827 F.3d at 340 (quoting Bush v. Gore, 531 U.S. 98, 104-05 (2000)). Analysis under the equal protection clause in the North Carolina Constitution “generally follows the analysis of . . . the corresponding federal clause.” Id. at 352. 26 E.g., Gray v. Sanders, 372 U.S. 368, 379 (1963) (stating that the Equal Protection Clause requires that “all who participate in the election are to have an equal vote”). 27 RWCA, 827 F.3d at 340 (quoting Reynolds v. Sims, 377 U.S. 533, 563 (1964)). 6 boards.28 The principle applies no matter the size of the population being divided into districts. It is difficult to draw districts that are exactly equal in population, and courts have recognized that “mathematical exactness or precision is hardly a workable constitutional requirement.”29 Nonetheless, the Constitution requires that governments “make an honest and good faith effort” to create state and local legislative districts “as close to equal population ‘as is practicable.’”30 District size may deviate from the ideal population based on “legitimate considerations.”31 Those legitimate considerations are “numerous and malleable,”32 but they at least include compactness, contiguity, and respect for political subdivisions.33 While a district plan with a maximum population deviation under ten percent will not, by itself, establish an equal protection violation,34 neither will such a plan get an automatic stamp of constitutional approval. Instead, the plaintiffs can prove an equal protection violation by showing that it is more probable than not that the deviation 28 Id. (citing Avery v. Midland Cty., 390 U.S. 474, 480 (1968)). Id. (quoting Reynolds, 377 U.S. at 577). 30 Id. at 340-41 (quoting Reynolds, 377 U.S. at 577). 31 Id. at 341 (quoting Harris v. Ariz. Indep. Redistricting Comm’n, 136 S. Ct. 1301, 1306 (2016)). Congressional districts have a more demanding standard: Even small deviations are allowed only in “unavoidable” instances. See, e.g., White v. Weiser, 412 U.S. 783, 790 (1973). 32 See Bethune-Hill v. Va. State Bd. of Elections, 137 S. Ct. 788, 799 (2017). 33 See Bethune-Hill v. Va. State Bd. of Elections, 141 F. Supp. 3d 505, 535-38 (E.D. Va. 2015), vacated in part on other grounds, 137 S. Ct. 788; see also RWCA, 827 F.3d at 341. 34 RWCA, 827 F.3d at 341 (quoting Wright, 787 F.3d at 264). 29 7 “reflects the predominance of illegitimate reapportionment factors” rather than legitimate redistricting considerations.35 B. Greensboro’s Districts and Voting Patterns The General Assembly’s numbers show that when a city the size of Greensboro— approximately 269,000 people36—is divided into eight City Council districts, the ideal population for each district is 33,631 people.37 Under the 2015 Redistricting Act, the eight districts range in population from 32,395 (District 2) to 35,167 (District 6).38 Thus District 2 is 3.68 percent below the ideal population and District 6 is 4.57 percent above ideal, while the plan as a whole has a maximum population deviation of 8.24 percent. 39 These deviations are significantly larger than the deviations in the pre-existing five-district plan. The maximum population deviation of that 2011 plan was only 3.86 percent.40 35 Id. (quoting Harris, 136 S. Ct. at 1307). Legislative intent and predominance can be difficult to determine. See Vill. of Arlington Heights v. Metro. Hous. Redev. Corp., 429 U.S. 252, 265 (1977) (“Rarely can it be said that a legislature . . . made a decision motivated solely by a single concern, or even that a particular purpose was the ‘dominant’ or ‘primary’ one.”). Testimony about an individual legislator’s motive may or may not be helpful, and it is particularly problematic when legislators who opposed a bill speculate about the motives of those who supported the bill. 36 See Pls.’ Ex. 3 at 2 (269,044). Other exhibits give slightly different population figures. See Doc. 126 at ¶ 5 (2010 census showing 269,666); Pls.’ Ex. 6 at 2 (269,231). 37 This figure is from the population figures in the “Stat. Pack Report” available to the General Assembly during the 2015 redistricting process. Pls.’ Ex. 3 (indicating total population is 269,044, or 8 x 33,630.5); see also Pls.’ Ex. 148 at 42. 38 Pls.’ Ex. 3 at 1. 39 Pls.’ Ex. 145 at 7 tbls.3-1 & 3-2. As explained supra note 8, the numbers were rounded to the second decimal place. The seven-district plan originally proposed, see discussion infra at notes 77, 86, had a maximum population deviation of approximately 9.6 percent. See Pls.’ Ex. 6 at 1 (indicating the largest and smallest districts had deviations of +4.77 and -4.87 percent). 40 Pls.’ Ex. 145 at 7 tbl.3-1. 8 In the General Assembly, Republicans held a majority in both the House and the Senate when the Act was passed.41 According to the information before the legislature in 2015, 56 percent of Greensboro’s voters are registered as Democrat, 23 percent as Republican, 21 percent as unaffiliated, and less than 1 percent as Libertarian.42 In the last statewide election for which results were available to the legislature in 2015, the 2010 U.S. Senate race, Greensboro voters favored the Democratic candidate by sixteen percentage points.43 Greensboro’s City Council elections are non-partisan and would remain so under the 2015 Redistricting Act; while there is crossover voting, the partisan registration of council members is well-known.44 Since 1999, the number of City Council members registered as Republicans has varied from zero to six; registered Republicans have usually, but not always, been in the minority.45 At the time the 2015 Redistricting Act passed, all three at-large council members and the mayor were registered Democrats, four of the five district council members were registered Democrats, and one district council member was a registered Republican.46 41 See Doc. 126 at ¶¶ 39, 60. Id. at ¶ 9; see also Trial Tr. Vol. I, Doc. 133 at 104:19-:23 (Mayor Nancy Vaughan, testifying that “Greensboro leans Democratic”). 43 The Greensboro results in the 2010 U.S. Senate election showed 41,462 votes for the Democratic candidate and 29,642 votes for the Republican candidate. Pls.’ Ex. 253. Dr. Jowei Chen credibly testified that the 2010 U.S. Senate race was the most recent election with data available to the legislature in 2015. Trial Tr. Vol. II, Doc. 134 at 234:14-235:3. 44 See Doc. 126 at ¶ 11; Trial Tr. Vol. I, Doc. 133 at 105:4-:11 (Mayor Vaughan, testifying that “I have noticed in City Council elections there is some crossover”), 106:2-:12. 45 See Doc. 126 at ¶ 21. 46 Id. at ¶ 19. 42 9 For the eight individual districts drawn by the General Assembly in the 2015 Redistricting Act, the population deviations and partisan leanings are as follows,47 with all figures in percentages: District 1 2 3 4 5 6 7 8 Deviation Democrat48 Republican +4.37 80.9 17.5 -3.68 76.2 22.4 -3.03 47.4 50.5 +2.87 86.9 11.9 -0.09 48.1 50.0 +4.57 68.5 29.4 -3.56 37.8 60.0 -1.46 38.3 59.4 Under the 2015 Redistricting Act, four of the eight districts lean Republican: two by a small percentage and two by over twenty percentage points.49 Approximately eighty-six percent of all Republican voters are in underpopulated districts.50 Pls.’ Ex. 145 at p. 7 tbl.3-2, p. 15 tbl.6-1. The last two columns are taken from results in the 2010 U.S. Senate race between Democrat Elaine Marshall and Republican Richard Burr. Pls.’ Ex. 145 at 15 tbl.6-1. According to plaintiffs’ expert witness Anthony Fairfax, these midterm results are better for evaluating City Council races than a presidential election because the midterm results better mimic the relatively low turnout in a City Council election. Trial Tr. Vol. I, Doc. 133 at 90:2-:11. The 2010 U.S. Senate race was also the most recent election with data available to the legislature. See supra note 43. Dr. Chen also testified that he would have reached “exactly the same conclusions” if he had used data from the 2012 presidential race between Mitt Romney and Barack Obama instead of the 2010 results. Trial Tr. Vol. II, Doc. 134 at 244:17-245:14, 251:5-252:6. No evidence indicated that the 2010 U.S. Senate race did not fairly represent partisan tendencies. Party registration is generally not helpful to this analysis because a large portion (21 percent) of Greensboro voters are unaffiliated. Doc. 126 at ¶ 9. See generally Pls.’ Ex. 3 at 4 (showing party registration by district). 49 Pls.’ Ex. 145 at 15 tbl.6-1; see Pls.’ Ex. 133 at 13. 50 Pls.’ Ex. 133 at 3-4. 47 48 10 C. Analysis If the maximum deviation of a district plan is less than ten percent, the plaintiffs must show that it is more probable than not that the deviations in the plan “‘reflect[] the predominance of illegitimate reapportionment factors rather than’ legitimate considerations.”51 Here, the maximum population deviation of the 2015 plan was 8.24 percent.52 All of the credible evidence establishes that the deviations occurred as the result of an intentional attempt to guarantee Republicans a partisan advantage in City Council elections. Moreover, this evidence is uncontroverted. First, credible evidence based on computer simulations by Dr. Jowei Chen establishes that it is highly unlikely for a Greensboro redistricting process to result in four Republican-leaning districts absent an intentional effort to draw lines giving Republicans an advantage.53 Specifically, these simulations show that when a plan is drawn with a maximum deviation of four percent or less, only two or three out of eight districts lean Republican.54 When the maximum deviation is allowed to increase to between eight and ten percent, simulations occasionally create plans with four districts leaning Republican, but even then most simulations result in only two or three Republican-leaning districts.55 51 RWCA, 827 F.3d at 341 (quoting Harris, 136 S. Ct. at 1307). Pls.’ Ex. 145 at 7 tbl.3-1. 53 See RWCA, 827 F.3d at 344, 347 (crediting similar computer simulation evidence from Dr. Chen); see also Trial Tr. Vol. I, Doc. 133 at 134:19-135:3 (testimony of Councilmember Marikay Abuzuaiter that the Act’s map would strongly tilt election results in favor of Republicans), 27:24-28:11 (testimony of Senator Gladys Robinson to same effect). 54 Pls.’ Ex. 133 at 13-14 & fig.3a. 55 Id. at 13, 15 fig.3b. 52 11 Second, all four of the Republican-leaning districts in the 2015 Redistricting Act were underpopulated, two to a significant degree.56 By comparison, only one Democratleaning district was underpopulated, and all of the overpopulated districts leaned Democratic.57 In a neutral redistricting plan, voters of either party are equally likely to be placed in an underpopulated or overpopulated district.58 Yet under the 2015 Redistricting Act, 86 percent of all Republican voters59 were in districts that are underpopulated while only 57 percent of Democratic voters were in an underpopulated district.60 Credible evidence based on computer simulations run by Dr. Chen again establishes that this was “very, very unlikely” to happen by chance,61 and that this “partisan skew” resulted from an intent “to significantly favor” Republican voters.62 Plaintiffs’ expert Anthony Fairfax also identified this pattern of overpopulation in Democratic-leaning districts and underpopulation in Republican-leaning districts.63 Third, six of the seven incumbent Democratic council members’ residences were assigned into districts with another incumbent Democrat.64 Due to this “double- See Pls.’ Ex. 145 at p. 7 tbl.3-2 (indicating that Districts 3 and 7 have deviations of -3.03 percent and -3.56 percent), p. 15 tbl.6-1. 57 Id. at p. 7 tbl.3.2 (indicating District 2 has a deviation of -3.68 percent), p. 15 tbl.6-1. 58 Pls.’ Ex. 133 at 3. 59 “Republican voters” refers to voters’ choices in the 2010 U.S. Senate election, not to voter registration. See supra note 48. 60 See Pls.’ Ex. 253 (indicating that 23,835 of 41,462 Democratic voters are in underpopulated districts). 61 Trial Tr. Vol. II, Doc. 134 at 237:14-238:2. 62 Pls.’ Ex. 133 at 10. 63 Pls.’ Ex. 145 at 16; see RWCA, 827 F.3d at 346 (crediting similar evidence from Mr. Fairfax). 64 Pls.’ Ex. 145 at 12 tbl.5-2. 56 12 bunking,” at least three Democratic incumbents would necessarily lose their seats in the next election. One of those incumbents assigned to new District 3, Nancy Hoffman, lives in a precinct that, aside from a small area around her residence, was assigned to District 7; her residence was essentially carved out of District 7 and placed in District 3.65 As a result of this split-precinct carve-out, District 7 had no incumbent while District 3 had two incumbents.66 Carving out Councilmember Hoffman’s residence did not reduce population disparities because both Districts 3 and 7 are significantly underpopulated.67 The sole Democrat who was not double-bunked, Councilmember Marikay Abuzuaiter, was assigned to a district that strongly leans Republican.68 The sole Republican, Councilmember Tony Wilkins, was not double-bunked and was placed in a district that leans Republican.69 This evidence shows that the legislature intended to unseat half the incumbents registered in one political party and corroborates partisan motivation in the drawing of district lines.70 Fourth, the 2015 districts split more precincts71 than would be expected from a partisan-neutral plan. Dr. Chen credibly testified that “the vast majority” of neutral plans resulting from his simulations had five or fewer split precincts, and that “very often” 65 See id. at 12-13 & fig.5-6. See id. at 12 tbl.5-2. 67 Id. at 7 tbl.3-2. 68 Id. at p. 14 tbl.5-3, p. 15 tbl.6-1. 69 Id. 70 See id. at 16. The earlier seven-district plan also would have double-bunked Democratic incumbents. See Pls.’ Ex. 18 at 8:2-:17. 71 Greensboro precincts are nearly identical with voter tabulation districts (“VTDs”), such that the terms can be used interchangeably. Trial Tr. Vol. II, Doc. 134 at 230:22-231:13 (testimony of Dr. Chen); Trial Tr. Vol. I, Doc. 133 at 76:2-:21 (testimony of Mr. Fairfax). 66 13 there were no split precincts.72 The 2015 Redistricting Act split eight precincts.73 The purpose of one of those splits was, obviously, to double-bunk two Democratic incumbents.74 Fifth, the legislative path of the 2015 Redistricting Act departed from normal legislative procedures.75 In North Carolina, “local bills,” such as municipal redistricting bills, are generally only brought forth at the request of the municipality or collectively by the local delegation.76 The Greensboro changes originally appeared in a local bill, Senate Bill 36,77 which the City of Greensboro and many in the Greensboro legislative delegation opposed.78 Before introducing SB 36, which would have changed the City Council to a seven-district system, the Act’s primary sponsor, Republican Senator Trudy Wade, did not discuss the details of her bill or share maps showing the new districts with 72 Id. at 230:8-:18. Mr. Fairfax also credibly concluded that an eight-district plan could have been developed with “significantly fewer” precinct splits. Pls.’ Ex. 145 at 15. 73 Pls.’ Ex. 145 at 8 tbl.4-1. The 2011 five-district plan split only one precinct. Id. 74 See supra pp. 12-13; Pls.’ Ex. 145 at 15. 75 Cf. Arlington Heights, 429 U.S. at 267-68 (holding that, inter alia, “[d]epartures from the normal procedural sequence” can be evidence of an improper racially discriminatory intent). Hardball political tactics do not ordinarily show partisan bias raising equal protection concerns, but here the circumstances corroborate other persuasive evidence about legislators’ partisan motivations in drawing the district lines. 76 See Doc. 126 at ¶¶ 23-25; Trial Tr. Vol. I, Doc. 133 at 20:11-21:24 (testimony of Senator Robinson), 112:12-:24 (testimony of Mayor Vaughan). 77 S.B. 36, 2015 Gen. Assemb., Reg. Sess. (N.C. 2015), available at Pls.’ Ex. 4; see Doc. 126 at ¶¶ 31-33. As noted supra note 39 the maximum population deviation in SB 36 was even larger. 78 Trial Tr. Vol. I, Doc. 133 at 28:15-:17 (testimony of Senator Robinson), 49:13-:14 (testimony of Representative Pricey Harrison); Pls.’ Ex. 17 at 4:15-:19 (reflecting opposition of City Council); Pls.’ Ex. 217 at 131 (same); see Pls.’ Ex. 226 (reflecting Representative John Blust’s and Representative Jon Hardister’s opposition to House Bill 263). 14 other Greensboro legislators or with any members of the City Council.79 After passing in the Senate, SB 36 stalled in the House,80 so Senator Wade asked the Senate Redistricting Committee to add the Greensboro changes to House Bill 263, a bill about the City of Trinity that had already passed the House, and the Committee complied.81 Before amending HB 263, she again gave no notice to other legislators from Greensboro or members of the City Council.82 The amended HB 263, with the Greensboro changes, then passed the Senate.83 When the Senate’s amended bill with the Greensboro changes was brought back to the House, the House did not then concur.84 Irregularities continued when four out of five House members appointed to the conference committee supported the bill with the Greensboro City Council changes, in contravention of House rules requiring a majority of conference committee members to have “generally supported the House position.”85 The conference bill eventually presented to the House and Senate differed from the original 79 E.g., Trial Tr. Vol. I, Doc. 133 at 24:10-25:8 (testimony of Senator Robinson), 110:24112:16 (testimony of Mayor Vaughan). 80 See Doc. 126 at ¶¶ 38, 40-41. 81 See id. at ¶¶ 48-50; Pls.’ Ex. 22 at 8:4-:6 (indicating that Senator Wade made the changes to HB 263). 82 Trial Tr. Vol. I, Doc. 133 at 56:7-:13 (testimony of Representative Harrison), 115:22116:19 (testimony of Mayor Vaughan). Senator Robinson testified the changes to HB 263 occurred in “the dark of night.” Id. at 40:11-:17. 83 Doc. 126 at ¶ 54. 84 Id. at ¶ 59. 85 Pls.’ Ex. 26 r. 44; Trial Tr. Vol. I, Doc. 133 at 60:7-:15 & 61:8-:13 (testimony of Representative Harrison); see Doc. 126 at ¶¶ 29, 61-62; Pls.’ Ex. 27. Conference committees are used to resolve differences when the Senate and House pass different versions of the same bill. A conference committee includes members from both the House and the Senate. If the conference committee agrees on a single compromise version of the bill, that version then returns to both the House and the Senate for an up-or-down vote. Pls.’ Ex. 26 r. 44. 15 bill, as it drew eight new districts rather than seven.86 Despite this significant difference, the maps showing the lines for the eight new districts were made available to the rest of the legislators and the public only 24 hours before the bill was called for a vote.87 There is even evidence that the conference bill presented to the House was not the bill that conference committee members approved.88 Finally, this new eight-district bill was initially rejected by the House,89 and it only passed after a closed caucus by the Republicans resulted in several Republicans changing their votes.90 The Senate then also adopted the conference committee bill, enacting it into law.91 D. Stated Justifications for the Act Do Not Explain the Deviations There is no direct evidence before the Court about the criteria or process used to draw either the seven- or eight-district plans.92 Senator Wade, the primary sponsor of the Pls.’ Ex. 30 sec. 2.(c); Doc. 126 at ¶ 64. Doc. 126 at ¶¶ 63, 68; Trial Tr. Vol. I, Doc. 133 at 59:2-:11 (testimony of Representative Harrison). 88 See Pls.’ Ex. 31 at 18:14-:19 (statement by Representative John Blust: “Well, Representative [Pat] Hurley, you should have told me that—if you want to go here, you should have told me that—what was said about four—after four years the city council can go back to being done locally was not—was not in the bill.”); Trial Tr. Vol. I, Doc. 133 at 63:14-64:5 (testimony of Representative Harrison that tone of conversation between Representatives Blust and Hurley, both members of the conference committee, was that the bill had been changed after the conference committee approved it). 89 Doc. 126 at ¶ 68. 90 See id. at ¶¶ 69-75; Trial Tr. Vol. I, Doc. 133 at 64:23-66:16 (testimony of Representative Harrison). 91 Doc. 126 at ¶¶ 76, 78. 92 Counsel for the plaintiffs represented to the Court that neither Senator Wade, nor any other legislator, nor legislative staff produced any written materials during discovery showing or explaining the process used to draw either the seven- or eight-district plans, despite the plaintiffs’ request for such information. See Trial Tr. Vol II, Doc. 134 at 323:19-325:16 (representation by Allison Riggs, counsel for plaintiff); see also, e.g., Doc. 74-1 at 74, 85 ¶ 7 (subpoena to Senator 86 87 16 Act, refused to testify and invoked her legislative privilege,93 and the legislative history does not contain any explanation from any legislator about the criteria used to draw the districts. Indeed, the legislative history does not provide even indirect evidence of how the district lines were drawn. For example, when Senator Wade was asked during a floor debate to explain the large population deviations in the plan, she said, “I don’t believe I can answer,” and blamed “the professional staff” and unnamed “House members.”94 Representative Pat Hurley, the House member who sponsored HB 263 and allowed the Greensboro changes to be added to her pre-existing Trinity bill, said “I don’t really know” why the population deviations were so large or why so many precincts were split.95 During the legislative process, supporters of SB 36 and HB 263 offered a number of reasons the proposed changes were good ideas. These justifications fall into several categories: (1) smaller districts are better;96 (2) the City Council needs geographic diversity;97 (3) the public supports the Act;98 (4) the Act will provide fairer minority Wade requesting all documents related to “rationale(s) for the proposed new electoral district lines”). The few emails in the record about drafting the bill do not explain the method used to draw the lines. See, e.g., Pls.’ Ex. 227 at 2 (email asking staff to draft bill and referencing “redistricting plan” as attached, but no plan is attached to the email in evidence). 93 See Doc. 79 (motion to quash). The Court finds it unnecessary to decide whether the invocation of legislative privilege could result in drawing an inference that the privileged testimony and documents would support the plaintiffs’ case. See Doc. 121 at 27-28 (plaintiffs’ request for an adverse inference). The Court mentions the invocation of the privilege merely to explain why Senator Wade’s testimony is not before the Court. 94 Pls.’ Ex. 32 at 2:23-3:13. 95 Pls.’ Ex. 31 at 5:9-:18. 96 E.g., Pls.’ Ex. 13 at 3:2-:13; see also Pls.’ Ex. 18 at 3:20-4:3, 5:19-:22, 6:8-:19. 97 Pls.’ Ex. 17 at 32:23-33:19, 43:23-44:14; Pls.’ Ex. 18 at 2:9-:23. 98 Pls.’ Ex. 18 at 2:24-3:5; Pls.’ Ex. 22 at 13:2-:5; Pls.’ Ex. 24 at 8:23-9:3. 17 representation;99 and (5) the Act aligns Greensboro’s system of government with that of other cities.100 These justifications were consistent, whether legislators were referring to the original seven-district plan or the final eight-district plan.101 These reasons explain a move from at-large council members to a district-only system102 but they do not speak to why the General Assembly drew these specific boundaries for the new districts. There is no evidence that meeting these goals required a maximum population deviation over eight percent, required putting 86 percent of Republicans in underpopulated districts, or required any particular line to be drawn in any particular place. None of these justifications rebut or undermine the plaintiffs’ evidence that the deviations were motivated by illegitimate partisanship.103 Pls.’ Ex. 17 at 2:16-3:13; Pls.’ Ex. 18 at 3:6-:10. E.g., Pls.’ Ex. 167 (stating that bill “[m]odels” Greensboro after “many other municipalities in our state”). 101 E.g., Pls.’ Ex. 31 at 4:6-:15. 102 Even though the plaintiffs do not contend that the change to an all-district system violates their equal protection rights, see generally Doc. 109, their witnesses spent a good bit of trial time criticizing the part of the Act that did away with the three at-large council members. E.g., Trial Tr. Vol. I, Doc. 133 at 119:9-120:9 (testimony of Mayor Vaughan), 137:11-138:17 (testimony of Councilmember Abuzuaiter). The Court has not considered this evidence, as the question of whether that change was right or good from a policy standpoint is not an appropriate question for judicial review. Cf. Ferguson v. Skrupa, 372 U.S. 726, 730 (1963) (noting that courts are “not concerned with the wisdom, need, or appropriateness of the legislation” (quotation omitted)). Nor has the Court given any weight to opinion testimony that Greensboro voters overwhelmingly opposed the bill. E.g., Trial Tr. Vol. I, Doc. 133 at 123:24-124:5 (testimony of Mayor Vaughan), 158:2-:8 (testimony of Anna Fesmire). The Constitution does not require legislatures to pass only those bills that have public support, and anecdotal evidence of public opinion is immaterial to constitutional analysis. 103 See RWCA, 827 F.3d at 349-50. 99 100 18 In an early floor debate, Senator Wade also made passing mention of compliance with the Voting Rights Act.104 Voting Rights Act compliance may well have been a legitimate redistricting criteria at the time,105 but no party has made any argument or offered any evidence that VRA compliance required unequal district populations.106 There is no credible evidence that the deviations predominantly reflected efforts to achieve compliance with the federal Voting Rights Act. E. Legitimate Criteria Do Not Explain the Population Disparity No credible evidence before the Court indicates that legitimate redistricting criteria predominated when the lines were drawn for the eight new districts. In fact, the evidence shows that these legitimate considerations did not predominate. Legitimate criteria include compactness, contiguity, and the integrity of political subdivisions.107 Fairness or “balance” between political parties108 and protection of incumbents have also been recognized in some circumstances as legitimate criteria.109 Nothing in the legislative history indicates that the deviations were required to achieve compact districts. To the contrary, every one of Dr. Chen’s simulations was at See Pls.’ Ex. 18 at 4:11-:14. See Harris, 136 S. Ct. at 1306-07. 106 See RWCA, 827 F.3d at 350 n.11 (finding it was unnecessary to address VRA compliance as a possible legitimate consideration when no party addressed it). 107 Bethune-Hill, 141 F. Supp. 3d at 535-38, vacated in part on other grounds, 137 S. Ct. 788; see also RWCA, 827 F.3d at 341. 108 RWCA, 827 F.3d at 341. 109 Vieth v. Jubelirer, 541 U.S. 267, 284 (2004) (plurality opinion of Scalia, J.); see also Harris, 136 S. Ct. at 1307 (citing Vieth’s list of criteria); cf. Bush v. Vera, 517 U.S. 952, 964 (1996) (principal opinion of O’Connor, J.) (in racial gerrymandering case, noting that the Supreme Court has “recognized incumbency protection, at least in the limited form of avoiding contests between incumbents, as a legitimate state goal” (quotation omitted)). 104 105 19 least as compact as the Act’s districts, as measured by the Reock test, and his algorithm produced scores of district plans with population deviations under four percent.110 For reasons unspecified in the legislative history, the Act increased the number of split precincts to eight,111 yet eight districts could have been created with significantly fewer precinct splits if the district plan had been partisan-neutral.112 The Act also split Councilmember Hoffman’s precinct to double-bunk incumbent Democratic council members.113 Instead of protecting incumbents, the Act targeted Democratic incumbents by pitting them against each other.114 Finally, the legislative history does not support that district lines were drawn based on the legitimate factor of competitive balance among political parties,115 alternately known as political fairness. No legislator mentioned political fairness or competitive balance as a reason for drawing the lines in a particular way. Moreover, political fairness refers to representation that “reflect[s] the relative strength of the parties.”116 There is no See Pls.’ Ex. 133 at 5, 7. The Reock test is the only compactness measurement in evidence here. 111 As noted supra p. 14; Pls.’ Ex. 145 at 8 tbl.4-1. 112 Trial Tr. Vol. II, Doc. 134 at 230:8-:18 (Dr. Chen testifying that “the vast majority” of his neutral simulations had five or fewer split precincts and that “[v]ery often” there were zero split precincts); see also Pls.’ Ex. 145 at 15 (Mr. Fairfax concluding that the plan “could have been developed with significantly fewer” split precincts). 113 See discussion supra p. 13; Pls.’ Ex. 145 at pp. 12-13 & fig.5-6, p. 15. 114 See Pls.’ Ex. 145 at 12 tbl.5-2. Where incumbency protection is applied in a “blatantly partisan and discriminatory manner,” it is not a legitimate state interest. Larios v. Cox, 300 F. Supp. 2d 1320, 1347-49 (N.D. Ga.) (per curiam), aff’d, 542 U.S. 947 (2004). 115 See Harris, 136 S. Ct. at 1306. 116 Gaffney v. Cummings, 412 U.S. 735, 752 (1973); cf. Vera, 517 U.S. at 964-65 (in a racial gerrymandering case, citing Gaffney for the holding that a state “may draw irregular district lines in order to allocate seats proportionately to major political parties”). 110 20 evidence that half of Greensboro’s voters typically vote Republican and indeed, the evidence is otherwise.117 As Dr. Chen testified, “four Republican districts is an outcome that never occurs” in simulations with low population deviations.118 The move to an alldistrict system may have enhanced political fairness by doing away with at-large council seats, but the lines drawn by the Act undermine this goal by disproportionately inflating the weight of votes cast by Republican-leaning voters and decreasing the weight of votes cast by Democratic-leaning voters. The legislative history does not explain the reasons behind the deviations or offer any legitimate reason for them.119 The available evidence supports the conclusion that the Act largely ignored legitimate redistricting criteria along the way to achieving a partisan goal. F. Conclusion The evidence in this case is quite similar to the evidence in Raleigh Wake Citizens Association v. Wake County Board of Elections.120 In both cases, the North Carolina legislature redrew districts for a local government entity using a plan with a maximum 117 See supra p. 9 & note 42. Trial Tr. Vol. II, Doc. 134 at 248:15-249:1. 119 In view of burden of proof, the Court has reviewed that legislative history—offered by the plaintiffs as relevant to whether the statute proceeded through an unusual legislative process—to see if it provides obvious evidence of any non-discriminatory purposes for the deviations. As this section shows, it does not. The Court has not attempted to create arguments in support of the constitutionality of the Act sua sponte, as it would be inappropriate for the Court to depart from its role as impartial arbiter and act as an advocate for the Act merely because the State has chosen not to defend it. See generally City of Greensboro v. Guilford Cty. Bd. of Elections, No. 1:15-CV-559, Doc.135 at 18 n.93 (M.D.N.C. Apr. 3, 2017) (granting partial summary judgment for plaintiffs). 120 827 F.3d 333. 118 21 population deviation between eight and ten percent.121 In both cases, the legislature redistricted the local entities through a “truncated” legislative process without soliciting input from the affected parties or the local delegations.122 In both cases, there was a pattern of overpopulation in Democratic-leaning districts and underpopulation in Republican-leaning districts.123 In both cases, credible computer simulation evidence showed that the partisan benefits from the redistricting were “completely outside the range of outcomes that are possible under a nonpartisan districting process that creates equally populated districts” and follows traditional criteria.124 In both cases, the General Assembly did not come to court to defend its redistricting.125 This is not a case where it is difficult to discern legislative motivation.126 As in RWCA, all of the credible evidence points in one direction: a “skewed, unequal redistricting” intentionally designed to create a partisan advantage by increasing the weight of votes of Republican-leaning voters and decreasing the weight of votes of Democratic-leaning voters.127 This evidence is unchallenged and uncontroverted. On this record, as in RWCA, the evidence compels a decision for the plaintiffs: The districting plan in the Act violates the equal protection rights of the plaintiffs and all Greensboro voters. 121 See id. at 338-39. See id. at 346. 123 See id. 124 See id. at 347. 125 See id. at 345. 126 See supra note 35. 127 827 F.3d at 346. 122 22 III. PROHIBITION ON REFERENDUMS AND INITIATIVES Plaintiffs also sought judgment on the pleadings for their claim that the Act unconstitutionally prohibits Greensboro citizens from bringing referendums or initiatives to change their form of government by singling out Greensboro, alone among municipalities in the state.128 As set forth in a more detailed order entered today at Doc. 135, the Court grants summary judgment on that claim and finds that the prohibition on referendums and initiatives by Greensboro citizens violates the Equal Protection Clause because that prohibition intentionally treats Greensboro voters differently and lacks any legitimate governmental purpose. IV. REMEDY AND SEVERABILITY The appropriate remedy for a law that violates the one-person, one-vote principle is an injunction against elections conducted under the Act’s unconstitutional redistricting.129 The Court will therefore enjoin the Guilford County Board of Elections from conducting any elections under the eight-district plan established by Section Two of the Act130 and direct that future elections shall use the pre-existing statutory and city charter system with five single-member districts and three at-large members,131 unless 128 See Doc. 95; Doc. 109 ¶¶ 77-83. See RWCA, 827 F.3d at 353-54 & n.13. 130 Most redistricting cases involve post-census redistricting, so if constitutional violations are found, the districts must affirmatively be redrawn. That is not the case here, where the districts were drawn after the 2010 census in a way that appears to comply with one-person, onevote requirements. See Doc. 126 at ¶¶ 17-18. There is thus no constitutional need for the Court to hold the matter open for the legislature or the City Council to draw new districts, or for the Court to draw new districts. 131 This is how the City Council was elected in 2015, after the Court’s preliminary injunction. See City of Greensboro v. Guilford Cty. Bd. of Elections, 120 F. Supp. 3d 479, 492 (M.D.N.C. 129 23 and until that system or those district lines are lawfully changed.132 The Court will also enjoin enforcement of the part of Section Two of the Act that prohibits Greensboro citizens from bringing referendums and initiatives to alter the form of their municipal government, for reasons set forth in a separate order.133 The plaintiffs contend that all provisions of Section Two of the Act should be enjoined.134 The Court concludes that the provisions of Section Two are not severable from each other and that plaintiffs’ request should be granted. Under North Carolina law, “where the various clauses of a statute are so interrelated and mutually dependent that one clause cannot be enforced without reference to another, the statute must stand or fall as a whole.”135 If, however, the parts of a statute are independent, or separable, “the invalid part may be rejected and the valid part may stand, provided it is complete in itself and capable of enforcement.”136 2015), available at Doc. 35 at 20 (granting motion for preliminary injunction, and finding that Greensboro should meanwhile “return to the previous system, which has been in place for some years and has not been challenged on constitutional grounds”). 132 All such provisions are subject to change by the City Council, and all changes except district lines are subject to change by referendum or by initiative, as allowed by N.C. Gen. Stat. §§ 160A-101 to -111. Further, nothing in this order prohibits the General Assembly from making future constitutional changes to Greensboro’s municipal government structure, district numbers and boundaries, or electoral systems. 133 See 2015 N.C. Sess. Laws 138 sec. 2.(b). Section Three of the Act sets the effective dates of the other sections. Id. at sec. 3. 134 See Doc. 132 at ¶ 27 (plaintiffs’ proposed supplemental conclusions of law, requesting an injunction of “the Greensboro provisions” of the Act). 135 Flippin v. Jarrell, 301 N.C. 108, 118, 270 S.E.2d 482, 488-89 (1980); see also Fulton Corp. v. Faulkner, 345 N.C. 419, 421-22, 481 S.E.2d 8, 9 (1997) (citing Flippin). 136 Flippin, 301 N.C. at 118, 270 S.E.2d at 489 (quotation omitted). 24 The Act is divided into two main sections. HB 263 had earlier passed the House as a bill to modify the municipal government of the City of Trinity.137 That earlier part of the bill was preserved in Section One of the Act, which reduces the size of the Trinity’s City Council and shortens the terms of Trinity council members.138 Section Two of the bill, titled “City of Greensboro Elections,” makes changes to Greensboro’s municipal government and municipal elections139 and applies only to Greensboro.140 The plaintiffs have not challenged the constitutionality of any portion of Section One or any of the changes to Trinity’s elections,141 and all of the statutory provisions at issue in this lawsuit are in Section Two. Section One is severable from Section Two. Section Two addresses a number of aspects of Greensboro city governance, in provisions that are comprehensive and interrelated. The Act does not have a severability clause,142 which tends to indicate that the General Assembly saw Section Two as all of a piece.143 No party has asked the Court to limit injunctive relief to only parts of Section See Pls.’ Ex. 19. 2015 N.C. Sess. Laws 138 sec. 1.(a). 139 See id. at sec. 2.(a)-(b) (setting permanent form of government), 2.(c) (drawing districts and setting four-year terms), 2.(d)-(e) (weakening role of mayor). 140 Id. at sec. 2.(g). 141 See generally Doc. 109. Nor do the plaintiffs ask for an injunction against the provisions of Section One. The Trinity potion of the bill is wholly independent of the Greensboro portion. See 2015 N.C. Sess. Laws 138 sec. 1.(c) (stating that Section One applies “only to the City of Trinity”). The Court will not enjoin Section One. 142 See 2015 N.C. Sess. Laws 138. There was no severability clause in the Act as originally passed, and the amendment in the technical corrections bill did not add a severability clause to the Act. See 2015 N.C. Sess. Laws 264 sec. 85.5. 143 In re Springmoor, Inc., 348 N.C. 1, 14, 498 S.E.2d 177, 185 (1998) (“[W]hile the absence of a severability clause is not necessarily conclusive, it does provide evidence of legislative intent.”). 137 138 25 Two. The Court has found two provisions of Section Two to be unconstitutional, and those provisions cannot be separated from the remainder of the Act. Because the provisions of Section Two of Session Law 2015-138 are not severable from each other, the Court will enjoin implementation of Section Two in its entirety. V. RACIAL GERRYMANDERING The plaintiffs contend that District 2 was a racial gerrymander and that citizens were placed in that district based on race.144 Given the Court’s findings and conclusions on the one-person, one-vote claim and on severability, the Court need not decide this issue. Standing alone, a violation of equal protection based on the one-person, one-vote principle requires a remedy.145 VI. CONCLUSION The United States Constitution does not allow an electoral system which makes one person’s vote more powerful than another’s. Nor does the Constitution allow a system which gives governance and electoral rights to one group of citizens while prohibiting another group of citizens from exercising those same rights, with no legitimate governmental purpose. The plaintiffs have shown, by a preponderance of the evidence, that the Act violates the Equal Protection Clause in both of these ways. The appropriate remedy is to enjoin enforcement of the Act and to preserve the City’s preexisting election system unless and until it is lawfully changed. 144 145 Doc. 109 at ¶¶ 93-97. See discussion supra pp. 23-25 & note 129. 26 It is ORDERED that the plaintiffs’ request for a permanent injunction will be GRANTED, and a judgment and permanent injunction will be entered as time permits. In the meantime, the preliminary injunction, Doc. 36, remains in force. This the 3rd day of April, 2017. __________________________________ UNITED STATES DISTRICT JUDGE 27