Searching over 5,500,000 cases.


searching
Buy This Entire Record For $7.95

Download the entire decision to receive the complete text, official citation,
docket number, dissents and concurrences, and footnotes for this case.

Learn more about what you receive with purchase of this case.

League of Women Voters of Michigan v. Benson

United States District Court, E.D. Michigan, Southern Division

April 25, 2019

LEAGUE OF WOMEN VOTERS OF MICHIGAN, et al., Plaintiffs,
v.
JOCELYN BENSON, in her official capacity as Michigan Secretary of State, et al., Defendants.

          BEFORE: CLAY, Circuit Judge; HOOD and QUIST, District Judges.

          OPINION AND ORDER

          HONORABLE ERIC L. CLAY UNITED STATES CIRCUIT JUDGE.

         The League of Women Voters of Michigan (“League”), numerous League members (“League Plaintiffs”), and several Democratic voters (“Individual Plaintiffs”) bring suit against Jocelyn Benson, the Michigan Secretary of State in her official capacity, under 42 U.S.C. §§ 1983 and 1988, alleging that Michigan's current legislative apportionment plan (the “Enacted Plan”), which the state legislature implemented as Michigan Public Acts 128 and 129 of 2011, violates Plaintiffs' Fourteenth Amendment equal protection rights and First Amendment free speech and association rights by deliberately discriminating against Democratic voters.[1] (See Compl., ECF No. 1.)

         After Plaintiffs filed suit, several parties moved to intervene. Ultimately, intervention was granted to several of the Republican members of Michigan's United States congressional delegation and two Republican state house members (together “Congressional and State House Intervenors”) (see ECF Nos. 103, 157) and to numerous Republican state senators and the Michigan Senate as a whole (together “Senate Intervenors”) (see ECF No. 237).[2]

         Plaintiffs initially sought to invalidate the entire Enacted Plan. (See Compl.) However, they have since narrowed their claims to 34 congressional, House, and Senate districts (the “Challenged Districts”).[3]

         The Court held a trial on Plaintiffs' claims. (See Trial Trs., ECF Nos. 248, 249, 250.) In addition to presenting witnesses at trial, the parties submitted hundreds of exhibits and deposition testimony from numerous witnesses in lieu of in-person testimony, pursuant to the Court's order, which reflected the parties' stipulation about the presentation and admissibility of evidence. (See Order Re: Parties' Partial Stipulations and Report, ECF No. 234.) The parties also filed post-trial briefs, including proposed findings of fact and proposed conclusions of law.[4] (See ECF Nos. 254, 255, 256, 257, 258.) The Court has carefully considered all the evidence.

         Today, this Court joins the growing chorus of federal courts that have, in recent years, held that partisan gerrymandering is unconstitutional. We find that the Enacted Plan violates Plaintiffs' First and Fourteenth Amendment rights because it deliberately dilutes the power of their votes by placing them in districts that were intentionally drawn to ensure a particular partisan outcome in each district. See Gill v. Whitford, 138 S.Ct. 1916, 1929-31 (2018). The Enacted Plan also injures Plaintiffs' First Amendment right to association by discriminating against them and their political party and subjecting them to “disfavored treatment by reason of their views.” Vieth v. Jubelirer, 541 U.S. 267, 314 (2004) (Kennedy, J., concurring in the judgment). Because we find that these constitutional violations will reoccur if future elections are held under the Enacted Plan, we HEREBY ENJOIN the use of the Challenged Districts in any future election.

         I. INTRODUCTION

         The term “partisan gerrymandering” describes “the drawing of legislative district lines to subordinate adherents of one political party and entrench a rival party in power.” Ariz. State Leg. v. Ariz. Indep. Redistricting Comm'n, 135 S.Ct. 2652, 2658 (2015). “By definition, partisan gerrymandering amounts to an effort to dictate electoral outcomes by favoring candidates of one party and disfavoring candidates of another.” Common Cause v. Rucho, 318 F.Supp.3d 777, 800 (M.D. N.C. 2018) (three-judge panel) (citing U.S. Term Limits, Inc. v. Thornton, 514 U.S. 779, 833-34 (1995)). Partisan gerrymandering thus violates the core purpose of legislative apportionment-providing “fair and effective representation for all citizens.” Reynolds v. Sims, 377 U.S. 533, 565 (1964).

         The Supreme Court has acknowledged that partisan gerrymandering is “incompatible . . . with democratic principles.” Ariz. State Leg., 135 S.Ct. at 2658 (quoting Vieth, 541 U.S. at 292 (plurality opinion)). It violates “the core principle of republican government . . . that the voters should choose their representatives, not the other way around.” Id. at 2652 (internal quotation marks and citation omitted). Lower federal courts have also noted that partisan gerrymandering diminishes our democracy, aptly describing it as a “noxious” practice that “has no place in a representative democracy[, ]” Shapiro v. McManus, 203 F.Supp.3d 579, 600 (D. Md. 2016) (Bredar, J. dissenting) (three-judge panel) (internal citation omitted); a “cancerous” problem that “undermin[es] the fundamental tenets of our form of democracy, ” Benisek v. Lamone, 266 F.Supp.3d 799, 818 (D. Md. 2017) (three-judge panel), aff'd, 138 S.Ct. 1942 (2018); and a phenomenon “widely considered to be repugnant to representative democracy.” Benisek v. Lamone, 348 F.Supp.3d 493, 511 (D. Md. 2018) (three-judge panel).

         Drawing district lines is an inherently political process. See Gaffney v. Cummings, 412 U.S. 735, 753 (1973) (“The reality is that districting inevitably has and is intended to have substantial political consequences.”) And “because the Constitution commits district apportionment to political departments [and] it is quintessentially a political process . . . courts cannot invalidate a redistricting map merely because its drafters took political considerations into account in some manner.” Benisek, 348 F.Supp.3d at 511 (citing Gaffney, 412 U.S. at 752-53). But, as a recent three-judge panel explained,

the political nature of redistricting does not “immunize state congressional apportionment laws which debase a citizen's right to vote from the power of courts to protect the constitutional rights of individuals from legislative destruction, a power recognized at least since our decision in Marbury v. Madison [1 Cranch 137, 2 L.Ed. 60 (1803)] . . . . The right to vote is too important in our free society to be stripped of judicial protection.”

Id. (quoting Wesberry v. Sanders, 376 U.S. 1, 6-7 (1964) (citations omitted)). As Justice Kagan recently observed, “the need for judicial review is at its most urgent in [partisan gerrymandering] cases” because “politicians' incentives conflict with voters' interests, leaving citizens without any political remedy for their constitutional harms.” Gill, 138 S.Ct. at 1941 (Kagan, J., concurring). Partisan gerrymandering “enables politicians to entrench themselves in power against the people's will. And only the courts can do anything to remedy the problem, because gerrymanders benefit those who control the political branches.” Id. at 1935 (Kagan, J, concurring).

         Federal courts must not abdicate their responsibility to protect American voters from this unconstitutional and pernicious practice that undermines our democracy. Federal courts' failure to protect marginalized voters' constitutional rights will only increase the citizenry's growing disenchantment with, and disillusionment in, our democracy, further weaken our democratic institutions, and threaten the credibility of the judicial branch. See Vieth, 541 U.S. at 310 (predicting that “[a] determination by the Court to deny all hopes of intervention” in partisan gerrymandering cases “could erode confidence in the courts”). Judges-and justices-must act in accordance with their obligation to vindicate the constitutional rights of those harmed by partisan gerrymandering.

         II. FACTS

         A. The Enacted Plan

         The Michigan Constitution provides that the Michigan legislature shall redraw Michigan's congressional and state legislative districts by November 1, 2001, and every ten years thereafter. See Mich. Comp. Laws §§ 3.62, 4.261. This directive ensures that Michigan legislators will have the benefit of the decennial federal census and population data when they redraw the legislative maps.

         As the release of the 2010 census data approached, the Republican State Leadership Committee (“RSLC”)[5] engaged in a national effort to ensure that states redrew their congressional lines during the 2011 redistricting cycle to favor Republican candidates and disadvantage Democrats. The RSLC appropriately named their initiative the “REDistrictng Majority Project, ” or “Project REDMAP.” (See Pls.' Trial Ex. 477.) According to a 2013 report from the RSLC, they raised $30 million towards Project REDMAP from 2009 to 2010. (Pls.' Trial Ex. 270 at 2.) The goal of Project REDMAP was simple: “[d]rawing new district lines in states with the most redistricting activity . . . to solidify conservative policymaking at the state level and maintain a Republican stronghold in the U.S. House of Representatives for the next decade.” (Id.) The report explained that drawing district lines favorable to Republicans was at the core of this strategy. (Id.) According to the report, “all components of a successful congressional race . . . rest in the congressional district lines, and this was an area where Republicans had an unquestioned advantage [in the 2012 elections].” (Id. at 1.)

         James Bolger, a Republican who at the time served as Speaker of the Michigan House of Representatives, worked closely with the RSLC during the 2011 redistricting cycle. Bolger stated that REDMAP materials were “ubiquitous” during the 2011 redistricting cycle and that he had seen redistricting materials from the RSLC “many times.” (Bolger Dep. at 294:24-295:14.) Bolger testified that he has attended approximately ten RSLC-related events. (Id. at 274:1-7.) He also served as Chairman of the Executive Committee of the Republican Legislative Campaign Committee, a caucus group of the RSLC. (Id. at 273:15-25.) According to Randy Richardville, a Republican who served as the Michigan Senate's Majority Leader during the 2011 redistricting cycle, Bolger was “pretty involved” in these groups. (See Richardville Dep. at 36:19-37:5.)

         Project REDMAP proved wildly successful both nationally and in Michigan. As the RSLC said in its 2013 report, in the 113th Congress, elected in November 2012 during the first elections after the 2011 redistricting cycle, “Republicans enjoy[ed] ¶ 33-seat margin” nationally despite the fact that “voters pulled the lever for Republicans only 49 percent of the time in congressional races, suggesting that 2012 could have been a repeat of 2008” when Democrats won the United States House of Representatives. (Pls.' Trial Ex. 270 at 1.) Further, the report found that “[t]he effectiveness of REDMAP is perhaps most clear in the state of Michigan.” (Id. at 3.) The REDMAP report explained that while “[t]he 2012 election was a huge success for Democrats at the statewide level in Michigan[, ]” with voters “elect[ing] a Democratic U.S. Senator by more than 20 points and reelect[ing] President Obama by almost 10 points . . . Republicans at the state level maintained majorities in both chambers of the legislature and voters elected a 9-5 Republican majority to represent them in Congress.” (Id. at 3-4.)

         Republicans enjoyed great success in Michigan's 2012 elections due in large part to the efforts of Republican legislators and map-drawers in the redistricting process in 2011. The passage of the Enacted Plan represented the culmination of a calculated initiative by Michigan's Republican legislators and map-makers, in the 2011 redistricting cycle, to deliberately draw Michigan's legislative districts to maximize Republican advantage and, consequently, disadvantage Democratic voters, Democratic candidates, and the Democratic Party. The partisan advantage that Michigan lawmakers achieved through the Enacted Plan persists to this day.

         During the 2011 redistricting cycle, Michigan Republican lawmakers engaged various political operatives to draw the congressional, Senate, and House maps to favor Republicans and disadvantage Democrats.

         For the congressional districts, Republican lawmakers engaged Robert LaBrant, who provided “advice and counsel” to “legislators and committee chairs . . . .” (LaBrant Dep. at 126:22-127:4.) At the time, LaBrant served as President for the Michigan Redistricting Resource Institute (“MRRI”), an organization that LaBrant had formed to “generate money to finance redistricting litigation” and defend Republican maps.[6] (LaBrant Dep. at 74:21-78:20.) LaBrant, through MRRI, in turn hired Sterling Corporation, a “Republican political consulting firm, ” and, specifically, Jeffrey Timmer, a Senior Counselor at Sterling, to draw the congressional map. (Timmer Dep. at 13:3-6; 16:14-17; 18:6-8; LaBrant Dep. at 140:23-141:8.) Timmer had extensive map-drawing experience; he had served as the Republicans' lead map-drawer in Michigan's 1991 and 2001 redistricting cycles. (Pls.' Trial Ex. 272 at 1.) He was also a well-known political figure. From 2005 to 2009, Timmer served as the Executive Director of the Michigan Republican Party. (Timmer Dep. at 20:11-16.) Timmer began his mapping and consulting work for the 2011 redistricting cycle in 2009, over a year before the 2010 census data was released. (Timmer Trial Tr., ECF No. 250 at PageID #9306.)

         Terry Marquardt drew the Senate map. (Marquardt Dep. at 31:6.) Marquardt had previously served as the Political Director of the Michigan Republican Party during the 2001 redistricting cycle. (Marquardt Dep. at 27:4-6.)

         Daniel McMaster had the primary responsibility for drawing the House map. From early 2009 to January 2011, McMaster served as the Director of the House Republican Campaign Committee. (McMaster Dep. at 29:19-23.) Following the 2010 elections, the Michigan House Republican Caucus hired McMaster as a Senior Policy Advisor to draw the House map. (Id. at 36:1-5; 49:12-14.) After McMaster informed Pete Lund, a Republican House member who had recently been appointed as Chairman of the Redistricting and Elections Committee, that he needed assistance, Lund hired Brian Began to assist McMaster in drawing the House map. (Id. at 50:20-51:19; Lund Dep. 23:15-23.) McMaster appreciated the importance of drawing districts that favored Republicans and disadvantaged Democrats. He candidly stated that “flipping the House in 2010 was more important than in any other year” because “there was just a belief or assumption that the party that wins . . . the majority in the last race of the decade will have a role in the redistricting process the following decade.” (McMaster Dep. at 33:18-34:2.)

         By statute, redistricting in Michigan must follow a set of criteria known as the “Apol” criteria. The Apol criteria contain a hierarchical set of requirements and provide, among other things, that districts be contiguous, contain either a population within 5% of the ideal district size (for the Senate and House districts) or exactly equal population (for the congressional districts), and minimize county and municipal breaks. See Mich. Comp. Laws § 3.63 (Senate and House districts); Mich. Comp. Laws § 4.261 (congressional districts). The Apol statutes provide that redistricting plans shall be created using “only” the Apol guidelines. See Mich. Comp. Laws § 3.63 (Senate and House districts); Mich. Comp. Laws § 4.261 (congressional districts). The Apol criteria do not provide that legislators may take political considerations into account when drawing maps. See generally, Mich. Comp. Laws § 3.63; Mich. Comp. Laws § 4.261.

         The map-makers claim that they drew their respective maps to simply comply with the Apol criteria.[7] But the evidence tells a much different story. The evidence points to only one conclusion: partisan considerations played a central role in every aspect of the redistricting process.

         Political data served as the foundation upon which the map-makers constructed their maps. They used this political data to deliberately draw districts that advantaged Republicans and disadvantaged Democrats. For example, to draw congressional maps that provided maximum Republican advantage, Timmer purchased historical voter lists and election data from a company called Combat Data. (Timmer Dep. at 16:21-17:10.) This political data consisted of historical election data from a number of historical statewide elections: the 2000, 2004, and 2008 presidential elections; the 2004 and 2008 gubernatorial elections; and the Michigan Board of Education elections in 2000, 2002, 2004, 2006, and 2008.[8] (Timmer Trial Tr., ECF No. 250 at PageID #9323:9-9324:15; Pls.' Trial Ex. 149 at 2.) This political data consisted of the number of Democratic and Republican votes cast in each of these historical elections down to the census block level.[9] (Timmer Trial Tr., ECF No. 250 at PageID #9316:19-25.) Timmer inputted this data into Maptitude, the map-drawing software he used to draw the congressional map. (Id.; Timmer Dep. at 28:17-20.) Relying on this granular political data, Timmer began drawing the congressional map in January 2011, approximately two months before the federal government released the actual 2010 census data, in March 2011. (Timmer Trial Tr., ECF No. 250 at PageID #9317:1-2; id. at PageID #9344:14-18.) Using this political data in conjunction with Maptitude, Timmer monitored the partisan composition of the congressional districts as he drew them, obtained instantaneous updates about the political composition of a district as he altered its boundaries, and deliberately drew districts that packed or cracked Democratic voters.[10] (See Timmer Dep. at 60:14-61:9; 95:1-5.)

         Marquardt used a program called AutoBound to draw the Senate map. (Marquardt Dep. at 40:5-15.) Marquardt used political data from a database compiled by the Michigan government that was provided to Michigan's Republican and Democratic caucuses for use in the 2011 redistricting cycle. (Marquardt Dep. at 42:8-43:12; Vatter Trial Tr., ECF No. 250 at PageID #8986:21-8987:16.) This data consisted of “[p]ast election results” from a variety of statewide elections in Michigan, and it showed how Michigan voters voted “down to [the] block level.” (Marquardt Dep. at 42:20-43:12; 69:11-22; 102:13-104:15.) Marquardt paid closest attention to past presidential results, as they “give[] you the best gauge on” the political makeup of a district. (Marquardt Dep. at 70:2-20.) Marquardt elected to have AutoBound display a matrix of political data next to his Senate maps as he drew them; the figures in the matrix updated automatically as he altered the shape and contours of a given district. (Marquardt Dep. at 195:16-198:7.) Marquardt was aware of the political makeup of the new districts as he drew them and could tell whether they were more advantageous for Republicans than the maps that had been used in the previous decade. (Marquardt Dep. at 148:13-21; 158:9-15; 174:8-175:19.) Marquardt drew districts that he expected would produce better Republican outcomes compared to the existing map. (Marquardt Dep. at 114:7-11.)

         When McMaster began drawing the House maps, he used magic markers to draw district boundaries because he had still not received map-drawing computer software. (McMaster Dep. at 87:16-25.) However, McMaster eventually received AutoBound and used it to draw the House maps. (McMaster Dep at 87:13-15.) While McMaster believes that he did not initially have access to political data, he eventually received “the election numbers” from previous elections and “plugged in, double-checked” the numbers in AutoBound. (McMaster Dep. at 88:15-24.)

         The map-makers believed that they could draw maps to secure a partisan advantage even if doing so required violating the traditional non-partisan redistricting criteria enumerated in the Apol standards. (See, e.g., Timmer Trial Tr., ECF No. 250 at PageID #9301:18-19; PageID #9304:24-9305:2.) Timmer, who drew the congressional districts, stated that map-makers and legislators may consider purportedly “neutral considerations” not listed in the Apol criteria, such as “incumbency considerations” and “achieving support sufficient for passage.” (Timmer Report at 6.) However, these factors were not politically neutral given that Republicans dominated both chambers of the state legislature and held the governorship during the 2011 redistricting cycle. Protecting incumbents generally meant protecting Republicans. And securing enough votes for passage did not necessarily require securing a single vote from a Democratic legislator in either chamber. At trial, Timmer went a step further, testifying that legislators may consider “whatever factors they deem necessary” and may “take into account any political consideration they like in drawing these maps.” (Timmer Trial Tr., ECF No. 250 at PageID #9301:18-19; PageID #9304:24-9305:2.)

         The map-makers were intimately aware of the political consequences of their map-drawing efforts. For instance, McMaster and Began drew the House maps in a secure location to which nobody else had access because, according to McMaster, “everyone was honing us. Every staffer's political future, every representative's political future, depended on how the lines turned out . . . .” (McMaster Dep. at 63:18-24.) Marquardt drew the Senate maps to ensure that they would garner enough support among Senators to win a majority of votes in the Senate-where the Republicans held 26 seats compared to Democrats' 12-because “if [the maps] don't pass the legislature, you've . . . done your work for nothing . . . you have to have at least 20 people happy with the draft.” (Marquardt Dep. at 56:23-57:13.) Marquardt explained that “sitting, you know, representatives or senators, you know, obviously in many cases want to be re-elected, so that was probably the major consideration as far as getting the vote-the 20 votes that we needed in the Senate to pass the bill.” (Marquardt Dep. at 63:9-14.)

         During the spring of 2011, Michigan's Republican leadership held weekly meetings on Thursday mornings at the Dickinson Wright law firm to discuss their redistricting efforts. (Pls.' Trial Exs. 463, 384; Richardville Dep. at 92:14-93:12.) Bobby Schostak, Chairman of the Michigan Republican Party, attended between three and eight of these redistricting meetings, even though he did not serve in the Michigan legislature and would therefore not cast a vote on any of the maps at issue. (Schostak Dep. at 13:12-16; 43:1-18.) Congressional map-drawer Jeff Timmer also attended some of these leadership meetings, even though he was not a legislator and even though he was working on the congressional maps, not the Senate or House maps containing districts whose legislators would vote on the redistricting legislation. (Id. at 15:13.) James Bolger, the Republican House member who at the time served as Speaker of the Michigan House, attended “some or all” of these weekly redistricting meetings. (Bolger Dep. at 120:11- 122:14.) Randy Richardville, a Republican who served as Michigan's Senate Majority Leader, attended between three and six of these meetings. (Richardville Dep. 85:2-14.) Joe Hune, the Republican Senator who served as Chairman of the Senate Redistricting Committee, attended more than five of these redistricting meetings. (Hune Dep. at 19:2-5; 142:2-5.) Pete Lund, the Republican House member who served as Chairman of the Redistricting and Elections Committee, also attended. (Lund Dep. at 93:7-95:20.)

         The Republican leadership took several steps to ensure that these weekly redistricting meetings remained secret. Members of the Republican leadership and their staffs often used personal-rather than governmental-email addresses to communicate about the redistricting meetings. (See, e.g., Pls.' Trial Ex. 462 (email confirming March 10, 2011 meeting sent to the personal email addresses of Suzanne Miller, Chief of Staff to Speaker Bolger (Bolger Dep. at 138:20-25); Laura Dubreuil, Executive Assistant to Speaker Bolger (Bolger Dep. at 149:7-23); Jordan Hankwitz, aide to Senate Majority Leader Richardville (Bolger Dep. at 127:20-22); Keli Saunders, a staffer in Republican Governor Rick Snyder's administration (Bolger Dep. at 126:6- 18); and Laurie Rospond, a staffer for Lieutenant Governor Dick Posthumus (Bolger Dep. at 126:19-127:4)); Pls.' Trial Ex. 463 (Speaker Bolger using personal email address to communicate with his top assistant about the April 19, 2011 redistricting meeting).)

         Majority Leader Richardville explained the significance of communicating over personal email; he and his staff “were very sensitive” to “doing something with or on government property if there are potential political discussions going on[, ]” and he and his staff used nongovernmental email addresses for any conversation that was “going to be political.” (Richardville Dep. at 99:10-100:5; see also id. at 169:11-170:7.) Furthermore, the agendas for each of these meetings were labeled “confidential.” (Pls.' Trial Ex. 274.) Democrats were not invited to attend the meetings until June 2011, i.e., after the maps had already been voted out of committee by the Michigan legislature. (Timmer Trial Tr., ECF No. 250 at PageID #9341:1-12; see Schostak Dep. at 48:25-49:23.)

         In addition to the leadership meetings, map-drawer meetings regularly took place at Dickinson Wright. (Timmer Dep. at 56:2-25; 256:18-257:2.) Map-drawer meetings occurred “[p]erhaps as often as weekly.” (Id. at 56:17-18.) Timmer, Marquardt, McMaster, and Began frequently attended. (McMaster Dep, at 52:9-54:8.) Robert LaBrant also attended some in his capacity as president of MRRI. (Timmer Dep. at 258:11-20.) Attorneys from Dickinson Wright would also attend as legal counsel. (McMaster Dep. at 53:14-23; Marquardt Dep. at 88:12- 89:1.) The map-drawers shared maps with each other, gave “report[s] of where we are, what problems we're having[, ]” strategized about the map-making process, discussed technical questions, and worked with counsel to confirm that their maps complied with what they believed to be the relevant legal requirements. (McMaster Dep. at 52:8-19; Marquardt Dep. at 82:18-22.) Like the leadership meetings, the map-drawer meetings were “confidential.” (Timmer Dep. at 256:21-23.) No representative from the Democratic Party, or member of any Democratic interest group, ever attended. (Id. at 56:19-25.)

         Throughout the redistricting cycle, the map-makers regularly met with legislative leaders and incumbent Republican legislators. During these meetings, the map-makers presented incumbent Republicans with drafts of each legislator's proposed district, discussed the boundaries and partisan makeup of each legislator's proposed district, and solicited the incumbent's preferences and/or desired changes to his or her district. Map-makers held these conversations even though they-and the Republican leadership for whom they worked-knew that protecting incumbents was not a permissible standard under the Apol criteria. (See Pls.' Trial Ex. 274.)

         Marquardt, drawer of the Senate maps, held “meetings [with] and called” each Republican senator between April and June 2011 to show them the partisan composition of their new district compared to their existing district. (Marquardt Dep. at 82:23-84:20.) Marquardt never met with Democrats. (Id. at 84:17-20.) Marquardt brought two maps with him to every meeting, which he presented side-by-side on a single piece of paper: one that depicted the senator's “Current Senate District” and one that depicted the senator's “Proposed Senate District.” (Id. at 100:21-106:15.) Marquardt displayed two sets of political data underneath the map of each senator's current and proposed district: the percentage of votes cast within that district for the Republican gubernatorial candidates in the last three elections and the percentage of votes cast within that district for Republican candidates in the last three elections for the Michigan Board of Education.[11] Marquardt displayed this political data on each senator's current and proposed district map “[b]ecause the senators obviously would be interested in knowing whether their district got better or worse[, ]” better meaning more Republican and “worse being either less Republican or more Democrat.” (Id. at 103:23-104:15; see, e.g., Pls.' Trial Exs. 331- 56 (side-by-side maps Marquardt prepared for each Republican senator showing the political composition of each senator's current and proposed districts); see also Richardville Dep. at 232:19-235:14 (discussing meetings with Marquardt and incumbent Republican senators).)

         Like Marquardt, McMaster, the principal architect of the House map, showed incumbent legislators the boundaries and political composition of their proposed districts before the maps were introduced into the legislature. (McMaster Dep. at 65:19-67:13; 71:9-19; Began Dep. at 45:17-25.) McMaster met with Republican House members in May 2011. (McMaster Dep. at 65:19-67:13.) During these meetings, McMaster presented Republican House members with various “choices” for the boundaries of their districts. (Id. at 65:19-67:13.) For example, if McMaster had identified multiple ways of splitting a county into two House districts, he would ask, “So, what do you guys want? Do you want this one? Do you want this one? Do you want to flip a coin?” (Id. at 66:11-13.) The maps that McMaster brought to the meetings contained historical election data from approximately ten previous elections and allowed incumbent Republican House members to gauge the partisan composition of their proposed districts. (Id. at 71:9-19.) After one of these meetings, McMaster made changes to the House map to accommodate the concerns that an incumbent Republican House member expressed about the “political makeup” of his proposed district and, specifically, “how hard he thought it would be for him to win” in the new district. (Id. at 86:1-89:22.)

         In late February 2011, Timmer traveled to Washington, D.C., to meet with the Republican members of Michigan's congressional delegation and present them with draft maps of Michigan's congressional districts. (Pls.' Trial Ex. 398; Timmer Trial Tr., ECF No. 250 at PageID #9324:16-9325:6.) Prior to his trip, Timmer sent the proposed congressional districts to LaBrant, the president of MRRI and the person who hired Timmer, and the two discussed the “potential options” that Timmer planned to present to the congresspersons with whom he was scheduled to meet. (See Pls.' Trial Ex. 398.) Timmer did not meet with any Democrats during his trip or provide any proposed maps to Democrats during the 2011 redistricting cycle. (Timmer Dep. at 110:2-25.)

         When Timmer arrived in Washington, D.C., he met with the Republican members of Michigan's congressional delegation and presented several different configurations of Michigan's statewide congressional map. (See Pls.' Trial Exs. 398-99.) Each iteration estimated the political composition of each congressional district based on historical election results from ten recent statewide races. (See id.) In this way, each congressperson could evaluate the partisanship of his or her proposed district in each possible configuration, as well as the partisanship of each congressional district in the overall map. (See id.) Timmer discussed the potential districts with the incumbent Republican congresspersons and asked them what portions of their current districts they would most like to continue to represent. (Timmer Dep. at 169:7- 11.)

         After returning from Washington, D.C., Timmer revised the congressional map to accommodate the preferences expressed by the incumbent Republican congresspersons with whom he had met. (Timmer Trial Tr., ECF No. 250 at PageID #9325:19-9327:17; see Pls.' Trial Ex. 399.) On March 6, 2011, he emailed LaBrant an updated version of the congressional map. (See Pls.' Trial Ex. 399.) The map contained political data for each proposed district. (Id.) Timmer told LaBrant that the map “captures the bulk of what we heard the [congressional delegation] indicate they'd like to see.” (Id.) Two days later, Timmer emailed LaBrant “the map we discussed yesterday with separate maps for the individual districts.” (Pls' Trial Ex. 400.) The map for each district contained the same set of political data that Timmer had included in his statewide maps. (Id.)

         The map-makers also shared draft maps with each other and strategized about how to best draw maps that advantaged Republicans. For example, in one email exchange from May 9, 2011, Marquardt asked Timmer to send him a plan that Timmer had proposed for Senate districts covering Oakland and Genesee counties. (Pls.' Trial Ex. 358.) After Timmer replied, Marquardt said that one of the proposed Senate districts “would be a little vulnerable under this plan.” (Id.) By “a little more vulnerable, ” Marquardt meant that the district was “more competitive, maybe a little more Democratic.” (Marquardt Dep. at 137:22-138:6.) Marquardt asked Timmer for the “McCain numbers” for the proposed district because Marquardt wanted to know where the district “landed on the political scale.” (Id. at 140:4-5.)

         Emails between the map-makers and Republican leaders, legislators, and legislative staff further reveal that the motivating factor behind the map-making process was a desire to construct districts that favored Republicans and disadvantaged Democrats.

         For example, on May 17, 2011, Jim Brandell, Chief of Staff to Republican Congressman Dave Camp, used his personal yahoo email address to email Timmer and LaBrant about Camp's concerns with the “[l]atest [d]raft” of Camp's district.[12] (Pls.' Trial Ex. 579.) LaBrant-who had hired Timmer to draw the congressional maps-replied that “[w]e will accommodate whatever Dave wants in his district . . . .” (Id.) LaBrant also assured Brandell that “[w]e've spent a lot of time providing options to ensure we have a solid 9-5 delegation in 2012 and beyond[, ]” (id.), meaning a congressional delegation of nine Republicans and five Democrats. (Timmer Trial Tr., ECF No. 250 at PageID #9318:2-7.)

         Another example is an email exchange from May 16, 2011, where LaBrant told Timmer that he had just received a call from Jack Daly, Chief of Staff to incumbent Republican Congressman Thad McCotter. (Pls.' Trial Ex. 409.) LaBrant had “walked [Daly] through Thad's proposed district.” (Id.) LaBrant reported that, according to Daly, “Dale at the RNC has drawn a 10-4 map.” (Id.) LaBrant told Daly “I suppose you could [draw a map that likely produces a 10-4 Republican/Democratic split] if you broke a [sic] sorts of counties and MCD's . . . .” (Id.) But LaBrant told Daly that “we need for legal and PR purposes a good looking map that [does] not look like an obvious gerrymander.” (Id.) LaBrant complimented Timmer on his latest draft of the congressional map, stating that it “protects all nine [Republican] incumbents and it looks good.” (Id.)

         In a separate email exchange, Daly requested that Timmer evaluate additional district configurations using newly-available population and elections data.[13] (See Pls.' Trial Ex. 401.) Timmer responded, “[i]nteresting numbers overall. Detroit being 150k less than projected shakes things up.” (Id.) Daly replied that it was “a glorious way that makes it easier to cram ALL of the Dem garbage in Wayne, Washtenaw, Oakland, and Macomb counties into only four districts.” (Id.) Daly also asked, “Is there anyone on our side who doesn't recognize that dynamic?” (Id.) Timmer admitted that the only way to interpret Daly's comments was as a request that Timmer pack Democratic voters. (Timmer Trial Tr., ECF No. 250 at PageID #9327:18-9329:11.)

         Yet another example is a June 22, 2011 email exchange that began with a Republican operative informing Timmer that he was “[w]orking on damage control” in the Third Congressional District because incumbent Republican Congressman Justin Amash had concerns about the political makeup of his new district. (Pls.' Trial Ex. 432.) Timmer responded that, under the new maps, the Third Congressional District “is a bit less GOP, but not so much less so that it is in jeopardy of going south on us.” (Id.) Timmer further explained that “the new 3rd would become slightly less Republican” to allow the district held by incumbent Republican Congressman Tim Walberg “to become slightly more so.” (Id.)

         In a different email, Jamie Roe, a staffer for incumbent Republican Congresswoman Candice Miller, expressed his approval of one of Timmer's proposed maps, saying it was “perfect” because “it's giving the finger to [S]andy [L]evin, ” a long-time Democratic United States congressman.[14] (Pls.' Trial Ex. 426; Timmer Dep. 127:2-10.)

         Schostak served as a liaison between incumbent Republican congresspersons and Timmer throughout the 2011 redistricting cycle. Congresspersons contacted Schostak-the Chairman of the Michigan Republican Party-and expressed preferences or concerns about the contours of their districts, and Schostak conveyed these concerns to Timmer, who revised the congressional maps accordingly. (Schostak Dep. at 89:1-94:20.) For example, in June 2011, Timmer emailed Schostak and explained changes that he had made to Congressional District 2 and Congressional District 3 based on concerns that Schostak had relayed from Bill Huizenga and Justin Amash, the incumbent Republican congressmen that represented those districts. (Pls.' Trial Ex. 387.) In another email, Timmer informed Stu Sandler, an outside consultant working for Schostak, that despite incumbent Republican congressmen Thad McCotter's disapproval of his new district, “it's hard to envision a Dem winning this seat even in a year like 2008.” (Ex. 65 to Timmer Dep.)

         Republican donors also inserted themselves into the map-drawing process through Schostak. For example, Schostak forwarded Timmer an email in which a Republican donor had asked Schostak if the map-drawers could “draw a favorable Republican district with the Pointes all together[, ]” an apparent reference to the five communities in the Grosse Pointes area. (Pls.' Trial Ex 389.) A few weeks later, Timmer sent Schostak a new set of maps. (See Pls.' Trial Ex. 438.) Timmer told Schostak that “as part of the maps I've redrawn, I've come up with a configuration which likely also satisfies the Grosse Pointes wrinkle you've been dealing with.” (Id.) Timmer told Schostak to “[k]eep this all under hat” given the “possibility that this fizzles.” (Id.) But Timmer reassured Schostak that “you may be able to deliver a vistory [sic] to the G[rosse] P[ointes] $$ folks.” (Id.)

         Schostak did not merely act as an intermediary through which Republican donors and Timmer would communicate. On one occasion, in response to a request from a Republican donor, Schostak had Timmer speak directly to the donor to “[t]ell them [sic] what's possible.” (Schostak Dep. at 109:1-13.)

         Even after they formally introduced the redistricting legislation, the Republican-controlled legislature concealed the contents of the redistricting plan and expedited its progression through the legislative process to prevent it from being subject to meaningful public scrutiny.

         Shortly after the proposed district maps were introduced in June 2011, the House Elections Committee held a public hearing on redistricting. (See Smith Trial Tr., ECF No. 248 at PageID #8771:1-10.) Susan Smith attended the hearing in her capacity as president of the League. (Id. at PageID #8771:10-15.) When Smith arrived, she picked up a copy of the bill so that she could follow along during the hearing. (Id. at PageID #8771:10-24.) When Smith opened up the bill, she realized that it was a “shell bill”-a “cover and a back sheet, and ending sheet[, ] but nothing in between.” (Id. at PageID #8771:19-24.)

         The members of the House Elections Committee entered the room. A staff member brought an additional stack of bills and placed them on a back table. Smith retrieved a copy of the bill, and returned to her seat to read it, “expecting to see a description of the maps[.]” (Id. at PageID #8772:6-8.) But the bill did not contain any maps, only rows of census data, “like you see when you get your tax bill . . . .” (Id. at PageID #8772:8-10.) It was impossible for Smith to figure out what district she lived in. (Id. at PageID #8772:10-11.) Shortly thereafter, a staff person set up three easels, one for each map. (Id. at PageID #8772:11-15.) Each map was approximately 30 by 36 inches and portrayed the entire state of Michigan. (Id. at PageID #8772:17-19.) Smith approached the maps for a closer examination but, like many other attendees, could not tell with any particularity what districts she lived in. (Id. at PageID #8772:20-8773:10.) Despite many attendees at the hearing voicing their complaints about the lack of transparency and public involvement, the bill made it out of committee after a vote along party lines, with Democrats voting against the bill and Republicans voting for it. (Id. at PageID #8772:20-8773:10.)

         Within approximately two weeks of the hearing, both chambers of Michigan's Republican-controlled legislature had voted to pass bills containing the redistricting legislation. The bills went to Governor Snyder, a Republican, who signed the legislation into law on August 9, 2011.[15]

         Democrats played no meaningful role in the 2011 redistricting process. On one occasion, Marquardt showed Mike Vatter, the Democratic map-drawer, a Senate map that Marquardt had drawn for the Detroit area. But according to Vatter, Marquardt was not soliciting his input but rather telling him that “these were the districts that were going to be drawn in the City of Detroit, ” a Democratic stronghold. (Vatter Trial Tr., ECF No. 249 at PageID #8994:20- 8995:15.) While Richardville, the Republican Senate Majority Leader, once met with Vatter and then-Democratic Senate Minority Leader Gretchen Whitmer, this meeting did not occur until after the Republican maps were already published. (Timmer Trial Tr., ECF No. 250 at PageID #9320:22-9321:12.) And while Republicans briefly entered into negotiations with the legislative black caucus, these negotiations took place after the Michigan legislature had already approved the Enacted Plan and did not result in any change to the maps. (Timmer Dep. at 149:20-150:1.)

         The Enacted Plan proved tremendously successful in advantaging Republicans and disadvantaging Democrats throughout several election cycles. In each of the three statewide elections held under the Enacted Plan between 2012 and 2016, Republicans won 64% of Michigan's congressional seats (i.e., 9 of 14) even though they never earned more than 50.5% of the statewide vote.[16] During this same period, Republicans won at least 53.6% of Michigan's House Districts while never earning more than 50.3% of the total vote.[17] In Michigan's 2014 Senate election, Republicans earned only 50.4% of the vote but won 71.1% of the seats.[18]

         The partisan bias in Michigan's legislative maps continued in the 2018 midterm elections. Democrats earned approximately 55.8% of the vote in congressional elections but gained only 50% of the congressional seats;[19] 52.6% of the vote in the House but only 47% of the House seats, [20] and over 50% of the vote in the Senate but only 42% of the Senate seats.[21] In other words, despite earning a sizable majority of aggregate votes for congressional candidates, Democrats merely pulled even with Republicans in terms of seats won. And despite earning a majority of the votes cast in the House and Senate elections, Democrats remained decidedly in the minority in both chambers of the state legislature.

         B. Plaintiffs' Expert Evidence

         The Court finds that Plaintiffs have proven that the Enacted Plan is a partisan gerrymander. Plaintiffs' expert evidence overwhelmingly supports this conclusion. Plaintiffs introduced testimony from three witnesses, each of whom employed several different statistical analyses and metrics to evaluate the partisan outcomes resulting from the Enacted Plan, clearly demonstrates that the Enacted Plan strongly and systematically advantages Republicans and disfavors Democrats. In reaching this conclusion, the Court does not rely primarily on any individual expert's testimony or on any particular statistical measure; rather, the Court reaches this determination after considering the totality of Plaintiffs' wide-ranging and extensive expert evidence. All of this evidence points to the same conclusion: the Enacted Plan gives Republicans a strong, systematic, and durable structural advantage in Michigan's elections and decidedly discriminates against Democrats.

         1. Dr. Jowei Chen[22]

         Plaintiffs' first expert, Jowei Chen, Ph.D. (“Dr. Chen”), analyzed Michigan's congressional, Senate, and House districts to determine whether the Enacted Plan produced “an extreme partisan outcome that diverges from possible alternative maps.” (Chen Report at 2, 6.) Dr. Chen compared the Enacted Plan to computer-simulated districting plans drawn without partisan intent that adhere to “traditional districting criteria, ” such as “equalizing population, maximizing geographic compactness, and preserving county, municipal, and ward boundaries.” (Id. at 3.) Dr. Chen created an algorithm that used these traditional, non-partisan criteria to produce randomly-generated districting plans. His algorithm generated 1, 000 alternative congressional districting plans, 1, 000 alternative Senate districting plans, and 1, 000 alternative House districting plans. (Id.)

         Next, using past election results, Dr. Chen measured the partisanship of each legislative district in the Enacted Plan and in his simulated districting plans. (Id. at 5.) Specifically, Dr. Chen assessed block-level election results for all of Michigan's 40 statewide elections from 2006-2010 and 2012-2016, and calculated “the vote totals across these statewide elections within every district.” (Id. at 6-7.) Dr. Chen then compared the partisan composition of the districts under the Enacted Plan to the partisan composition of the alternative simulated districts. In this way, Dr. Chen was able to deduce whether a district in the Enacted Plan favors Republican or Democratic candidates compared to the corresponding simulated districts drawn without partisan intent. (Id. at 7.)

         To calculate the partisanship of each district, Dr. Chen used three different measurements. As the Court describes below, each of Dr. Chen's metrics indicates that the Enacted Plan strongly and systematically advantages Republicans and discriminates against Democrats.

         a. Number of Republican and Democratic Districts

          First, Dr. Chen simply counted the number of Republican and Democratic districts in the Enacted Plan and each alternative simulated plan based on a district's partisan composition. (Id. at 9.) This method is considered the most basic and commonly-used method for measuring the partisanship of a districting plan. (Id.) By using this method, Dr. Chen directly and precisely quantified the difference in partisanship between the Enacted Plan and the simulated plans drawn without partisan intent. (Id.)

         Dr. Chen also determined whether the districts in the Enacted Plan constitute partisan outliers. (Id. at 11.) Dr. Chen defined a “partisan outlier” as a district whose partisanship falls “outside the middle 95% range [of partisanship] of the simulated geographically overlapping districts;” he considered a simulated district to be “geographically overlapping” if it contained at least 50% of the same population as the corresponding district under the Enacted Plan. (Id. at 55- 56.) Therefore, if Dr. Chen deemed a district in the Enacted Plan a “partisan outlier, ” the partisan composition of that district falls outside the middle 95% of the simulated, geographically-overlapping districts drawn without partisan intent. (Id.)

         i. Congressional Map

          Michigan currently has 14 congressional districts. Using the 2012-2016 election results, the Enacted Plan contains 9 Republican congressional districts and only 5 Democratic congressional districts. (Id. at 14.) The vast majority of the Dr. Chen's simulated plans, drawn without partisan intent, created 7 Republican congressional districts and 7 Democratic congressional districts. (Id.) Not a single one of Dr. Chen's 1, 000 simulated plans created 9 Republican congressional districts and only 5 Democratic congressional districts. (Id.) Dr. Chen concluded with overwhelming certainty that the pro-Republican bias in the Enacted Plan's congressional map did not occur because of chance but was rather the product of a deliberate effort to tilt the scales to favor Republicans. (Id. at 14-15.)

         ii. Senate Map

         There are 38 Senate districts in Michigan. Using the 2012-2016 election results, the Enacted Plan contains 24 Republican Senate districts and only 14 Democratic Senate districts. (Id. at 26.) Dr. Chen's simulated plans all created between 18 and 22 Republican Senate districts, and the vast majority of his simulated plans created 19 or 20 Republican Senate districts. (Id.) Not a single one of Dr. Chen's 1, 000 simulated plans created 24 Republican Senate districts. (Id.) Dr. Chen concluded with overwhelming certainty that the Enacted Plan's Senate map is a partisan outlier and was intentionally created to produce a pro-Republican outcome. (Id.)

         iii. House Map

          There are 110 House districts in Michigan. (Id. at 39.) Using the 2012-2016 election results, the Enacted Plan contains 61 Republican House districts and only 49 Democratic House districts. (Id.) Dr. Chen's simulated plans produced between 56 and 60 Republican House districts, and the vast majority of his simulated plans created 58 Republican House districts. (Id.) Not a single one of Dr. Chen's simulated plans created 61 Republican House districts. (Id.) Dr. Chen determined that the Enacted Plan's House map is a partisan outlier and was intentionally created to produce a pro-Republican partisan outcome. (Id.)

         b. Median-mean Difference

          Second, Dr. Chen analyzed the partisanship of the maps in the Enacted Plan and the simulated plans using median-mean difference analysis. Scholars commonly use this method to compare the partisan bias of different districting plans. (Id. at 11.) As Dr. Chen explained, “[f]or any districting plan, the mean is simply calculated as [the] average of the Republican vote share across all districts, and the median is the Republican vote share in the district where the Republicans performed the middle-best.” (Id.) Higher, positive values signify that the median district's Republican vote share is higher than the mean district-level Republican vote share, which indicates that the plan favors Republicans and disadvantages Democrats. (Id.) By calculating the median-mean difference of the Enacted Plan, Dr. Chen assessed whether a Republican-favoring skew in the median-mean difference resulted from an intentional partisan effort to favor one party over another. (Id. at 12.)

         Dr. Chen's median-mean analysis demonstrates that the Enacted Plan strongly advantages Republicans. Using the 2012-2016 statewide election results, the congressional districts in the Enacted Plan have a median-mean difference of 7.55%. (Id. at 18.) In other words, the median congressional district is 7.55% more Republican than the Enacted Plan's “average” district. (Id.) Dr. Chen found that this large pro-Republican bias also manifests in the Senate and House maps; the median-mean difference for the current Senate plan is 5.97% and the median-mean difference for the current House plan is 6.86%. (Id. at 33, 46.) These findings demonstrate that across the congressional, Senate, and House maps, the median district under the Enacted Plan is significantly more Republican than the “average” district. Dr. Chen found that these large median-mean differences give Republicans a strong, systematic advantage in elections held under Michigan's congressional, Senate, and House maps. (Id. at 21, 33, 46.)

         c. Efficiency Gap

         Third, Dr. Chen measured the partisanship of the Enacted Plan using the efficiency gap. This measurement is calculated by first determining the partisan composition of each district in the Enacted Plan and the simulated maps, which Dr. Chen measured using the 2012-2016 statewide election results. (Id. at 12.) A district is considered a Democratic district if the number of Democratic votes in that district exceeds the number of Republican votes. (Id.) If a district is not designated as a Democratic district, Dr. Chen classified it as being a Republican district. (Id.) Dr. Chen then calculated “the total sum of surplus votes in districts [each] party won and lost votes in districts where [each] party lost.” (Id.) In the districts that a party lost, all the votes cast for that party are considered “lost votes.” (Id.) Conversely, in the districts that a party won, only the party's votes that exceed the 50% threshold needed to win those districts are considered “surplus votes.” (Id.) The sum of a party's lost and surplus votes is that party's total number of “wasted votes.” (Id.) The efficiency gap is calculated from “the total wasted Republican votes minus total wasted Democratic votes, divided by the total number of two-party votes cast statewide across all seven[23] elections.” (Id.)

         The efficiency gap may appear to be a rather convoluted statistical measure. But its explanatory power is simple and clear: it indicates whether, and to what extent, votes cast for a particular political party are “wasted” in elections across a given districting plan. (Id.) As Dr. Chen explains, “[a] significant positive efficiency gap indicates that there are more Republican wasted votes, while a significantly negative efficiency gap indicates far more Democratic wasted votes.” (Id.) By comparing the efficiency gap under the Enacted Plan with the efficiency gaps under his simulated plans, Dr. Chen was able to determine whether the Enacted Plan causes Democrats to frequently waste more votes than they would waste under alternative maps drawn without partisan intent. (Id.)

         Dr. Chen's analysis of Michigan's 2012-2016 statewide elections demonstrates that the Enacted Plan consistently results in more wasted Democratic votes than wasted Republican votes. The current congressional plan has an efficiency gap of-19.8%. (Id. at 25.) This figure indicates that the Enacted Plan “consistently results in significantly more wasted Democratic votes than Republican votes.” (Id.) Furthermore, this efficiency gap “is far more biased than even the most biased” of Dr. Chen's 1, 000 simulated congressional maps. (Id.) In fact, more than half of Dr. Chen's simulation maps resulted in an efficiency gap within ±5% of 0, meaning that “it is clearly not difficult to create a map that is relatively unbiased according to the efficiency gap” and that complies with Michigan's statutory non-partisan districting criteria. (Id.) Dr. Chen concluded that to create a congressional map with an efficiency gap score greater than ±15% “would require extraordinary and deliberate partisan map-drawing efforts.” (Id.)

         The Senate and House districts show similarly staggering pro-Republican efficiency gap scores. The current Senate plan has an efficiency gap of-16.6%. (Id. at 38.) The current House plan creates an efficiency gap of-12.1%. (Id. at 51.) These heavily pro-Republican efficiency gaps are “entirely outside of the range produced by the simulated plans.” (Id. at 38, 51.) In other words, according to the efficiency gap metric, the Senate and House districts under the Enacted Plan are “far more biased than even the most biased” simulated maps drawn without partisan intent. (Id.)

         d. Durability

         After Dr. Chen concluded that the current legislative plans were partisan outliers compared to the simulated plans using the three metrics described above, he then assessed whether the partisan bias in the Enacted Plan is “politically durable.” According to Dr. Chen, the “partisan durability of a districting plan refers to whether a plan would allow a particular political party to preserve its majority control over a chamber or congressional delegation under a reasonable range of alternative electoral conditions.” (Id. at 52.) Dr. Chen's sought to ascertain if Republicans would still win the majority of Michigan's districts during an election in which the overall Republican electoral performance was worse than normal. (Id.) As Dr. Chen explained, if the Enacted Plan “gives Republicans control over a majority of all districts, and only a significant pro-Democratic [] swing would allow Democrats to ever win a majority of districts in a single election, then the Republican's control” is “durable.” (Id.)

         To analyze the partisan durability of the legislative plans, Dr. Chen used the “uniform swing analysis.” (See id.) This metric begins with the assumption that a political party's performance in a given election is generally not confined to a single district. When a political party performs worse than usual, its vote share typically decreases by a comparable degree in all legislative districts across the state. (Id.) Conversely, when a political party performs better than usual, its vote share generally rises in all districts across that state. (Id.) The uniform swing analysis involves “simulating a uniform increase (or decrease) in a party's vote share across all districts within a state.” (Id.)

         Using this method, Dr. Chen evaluated how Republicans and Democrats in Michigan “would perform under alternative electoral conditions.” (Id.) Dr. Chen evaluated the results of Michigan's seven statewide elections that took place between the adoption of the Enacted Plan and when Plaintiffs filed suit. (Id. at 53.) Republicans won a majority of the seats in each of these elections. For each election, Dr. Chen calculated the smallest pro-Democratic uniform swing that would have allowed Democrats to win one-half of the seats. (Id.)

         Dr. Chen's found that, in each of these seven statewide elections, Republicans would have maintained a majority under any reasonable range of alternative electoral conditions. (Id.) Specifically, in November 2012, Republicans won 9 of 14 congressional districts and the Republican vote share in the seventh-most Democratic district was 53.37%. (Id.) Therefore, a uniform swing of -3.37% would have been required for Republicans to lose their majority control over the congressional districts and for Democrats to win 7 of the 14 congressional seats. (Id.) In 2014 and 2016, Republicans also won 9 of 14 congressional districts. (Id.) A uniform swing of -6.45% in November 2014 and -7.79% in November 2016 would have been required for the Democrats to win one-half of the congressional districts in those elections. (Id.) In the November 2014 Senate election, Republicans won 27 of 38 districts, and it would have taken a uniform swing of-6.4% for Democrats to win one-half of all Senate districts. (Id.) Finally, in the 2012, 2014, and 2016 House elections, the Republicans won 59, 63, and 63 of Michigan's House districts, respectively. (Id.) For the Democrats to win one-half of all House districts, it would have taken a uniform swing of -1.04% in November 2012, -2.25% in November 2014, and -4.14% in November 2016. (Id.)

         Based on the large uniform swings required for Democrats to win a majority of the seats in these seven state-wide elections that took place over three separate election cycles, Dr. Chen concluded that “Republican majority control” was “durable” across “a reasonable range of alternative electoral conditions.” (Id.) In other words, given the large pro-Republican bias in the Enacted Plan, it would take an extraordinarily strong showing by Democrats to unseat the Republican majority.

         e. Single District Comparisons

         Dr. Chen also compared the partisanship of the districts in the Enacted Plan to the partisanship of the corresponding districts in his alternative, simulated maps. By comparing the partisanship of the actual and simulated districts, Dr. Chen determined whether the districts in the Enacted Plan constitute “partisan outliers.” Specifically, he evaluated whether the districts in the Enacted Plan are “packed” and/or “cracked” to favor Republicans. “Packing” involves filling a district with a supermajority of members of a disfavored political party. Vieth, 541 U.S. at 287 n.7. “Cracking” involves dispersing members of a disfavored party among several districts to deny that party a majority in any of those districts. Id.

         At first glance, it might appear perplexing that a dominant party would deliberately pack a large majority of members of the disfavored party into a small number of districts, thereby guaranteeing that the disfavored party would win those districts handily. However, upon closer examination, that strategy can prove quite effective for the party in power, if it chooses to use its power for nefarious ends; by packing a large proportion of the disfavored party's voters into a small number of districts, the dominant party can crack (i.e., disperse) the remaining members of the disfavored party across many districts such that the disfavored party can never obtain a majority of votes in those districts. While the disfavored party will consistently prevail by large margins in the small number of districts where its members are packed, it will inevitably lose, albeit by somewhat lesser margins, across the larger number of districts where its members are cracked. Through packing and cracking, a dominant party can prevent a disfavored party from ever obtaining a majority of the seats in a legislature or congressional delegation.

         Dr. Chen compared the individual districts under the Enacted Plan and the simulated plans in two ways. First, Dr. Chen aligned the districts from least to most Republican. (Id. at 54.) Then, he compared the partisanship of the most Republican congressional district from the Enacted Plan to the most Republican district from each of the 1, 000 simulated congressional plans. (Id.) Dr. Chen subsequently compared the second-most Republican congressional district in the Enacted Plan to the second-most Republican congressional district in the simulated plans, and so forth. (Id.) Dr. Chen also applied the same procedure to the Senate and House districts. (Id.)

         Second, Dr. Chen compared the districts under the Enacted Plan and the simulated maps using district geography. (Id.) Dr. Chen compared each district under the Enacted Plan to the district from the simulated plans that overlaps with it the most. (Id.) Comparing geographically-overlapping districts in this way allowed Dr. Chen to “identify partisan differences” between the Enacted Plan and the simulated districts “in terms of how each region in Michigan is districted.” (Id.) When a district under the Enacted Plan is a partisan outlier compared to the simulated districts that cover the same geographic area, Dr. Chen inferred that the district's boundaries were manipulated in violation of the Apol criteria, Michigan's non-partisan statutory redistricting guidelines. See Mich. Comp. Laws § 3.63; Mich. Comp. Laws § 4.261.

         Using these methods to compare individual districts, Dr Chen determined whether the districts in the Enacted Plan constitute partisan outliers. (Chen Report at 55.) As discussed above, Dr. Chen labeled a district a “partisan outlier” if the partisan composition of that district falls outside the middle 95% of the simulated, geographically-overlapping districts drawn without partisan intent. (Id.) Dr. Chen concluded that Congressional Districts 1, 4, 5, 8, 9, 10, 11, and 12; Senate Districts 8, 9, 18, 22, 24, 27, and 32; and House Districts 11, 12, 14, 16, 19, 20, 21, 30, 31, 32, 36, 43, 44, 45, 51, 52, 53, 55, 57, 60, 62, 63, 65, 69, 75, 76, 80, 87, 91, 92, 94, 98, 103, 105, 106, and 107 are partisan outliers. (Id. at 56.) Dr. Chen determined that these districts “are the most effectively cracked and packed districts” in the Enacted Plan. (Id.)

         f. Criticisms

         Congressional and State House Intervenors argue that Dr. Chen's expert evidence is unreliable because he misapplied the Apol criteria in two respects. (Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #10986.) First, they argue that when he conducted his simulations, Dr. Chen considered the Apol criteria to be absolute, which they contend is contrary to the Michigan State Supreme Court's ruling in LeRoux v. Secretary of State, 465 Mich. 594, 615 (2002) (Id.). We are skeptical about the validity of LeRoux's holding. In LeRoux, the Michigan Supreme Court decided that the 2001 Michigan legislature was not required to follow the guidelines set forth in Mich. Comp. Laws § 3.63 that were adopted by the 1999 Michigan legislature. (Id.) This holding conflicts with the plain language of Mich. Comp. Laws § 3.63, which provides that Michigan's redistricting plans can “only” be drawn using the guidelines set forth in that statute. However, regardless of the validity of LeRoux's holding, LeRoux, as a state court case, does not insulate the Enacted Plan from Plaintiffs' claims alleging a violation of the United States Constitution. See U.S. Const. art. VI, cl. 2. (“This Constitution . . . shall be the supreme Law of the Land; and the Judges in every State shall be bound thereby, any Thing in the Constitution or Laws of any State to the Contrary notwithstanding.”)

         Intervenors also assert that Dr. Chen's methodology and data is flawed because he improperly applied the Apol criteria in creating his simulations. They claim that Dr. Chen's algorithm failed to follow the Apol requirement that, when choosing between two townships or municipalities to shift, the one with the lesser population shall be shifted, and that his algorithm improperly always favored compactness, not just in the limited circumstances required by the Apol criteria. (Cong. and State House Intervenors' Proposed Conclusions of Law, ECF No. 258 at PageID #11107-11; Senate Intervenors' Proposed Conclusions of Law, ECF No. 254 at PageID #10373-76.) Simultaneously to arguing that Dr. Chen's algorithm applied the wrong Apol criteria, Senate Intervenors argue that Dr. Chen's simulations are flawed because his algorithm applied the Apol criteria too strictly and failed to appreciate that they “are only guidance, not mandatory.” (Senate Intervenors' Proposed Conclusions of Law at PageID #10376.) In other words, they contend that Dr. Chen's simulations are irrelevant because they failed to account for map-maker discretion and the political considerations that factored into the creation of the Enacted Plan. (Id.)

         Intervenors cannot have it both ways. They cannot convincingly argue both that Dr. Chen's algorithm is flawed because it failed to properly apply the mandatory Apol criteria, and because it erroneously considered the Apol criteria as binding when, in fact, they are merely non-binding guidelines that the map-drawers may discard whenever they desire. Further, the map-makers did not scrupulously follow the Apol criteria when drawing the districts in the Enacted Plan. Instead, they drew districts with the predominant purpose of advantaging Republicans and disadvantaging Democrats. The map-makers regularly discarded the Apol criteria to achieve their aim of entrenching Republicans in power. In fact, Timmer-the congressional map-drawer-testified that “a legislator can take into account any political consideration they like in drawing these maps.” (Timmer Trial Tr., ECF No. 250 at PageID #9304:24-9305:2) (emphasis added). Further, Intervenors have not presented any evidence that Dr. Chen's conclusions would be any different if his algorithm perfectly applied the Apol criteria. Under these circumstances, the fact that Dr. Chen's algorithm may not have flawlessly followed the Apol criteria does not diminish the value of Dr. Chen's simulations or undermine the validity of his conclusion that the Enacted Plan is a partisan gerrymander.

         Intervenors additionally contend that Dr. Chen's data is flawed because he erroneously used Voter Tabulation Districts (“VTDs”) as the “building blocks” of his simulations, while the map-makers used census tracks and census blocks. (Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #10989; Senate Intervenors' Proposed Findings of Fact, ECF No. 255 at PageID #10423-24.) Census block data changes every ten years following the release of the U.S. census data, while precinct level data is adjusted every two years by the local election clerks within their respective jurisdictions. (Timmer Trial Tr., ECF No. 250 at PageID #9282:9-25.) According to Timmer, the census block data is more static and neutral than the precinct level data. (Id. at PageID #9283:12.) However, Timmer testified that he has never “actually calculated what percentage difference” using VTDs instead of census data “actually makes.” (Id. at PageID #9299:23-25.) Further, Intervenors have not presented any evidence to support their assertion that Dr. Chen's using VTDs rendered his simulations inaccurate. Nor have they provided any reason for this Court to believe that Dr. Chen's algorithm would have produced materially different results had he had used census block data in his simulations. Accordingly, we reject Intervenors' argument that Dr. Chen's using VTDs instead of census data undermines his findings.

         Yan Liu, Ph.D., [24] (“Dr. Liu”), Defendants' expert, articulated several criticisms of Dr. Chen's findings, which the Congressional and State House Intervenors separate into three categories. (See Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #10990.) First, Dr. Liu criticizes Dr. Chen's entire simulation methodology.[25] (Liu Report at 27.) Second, Dr. Liu claims that Dr. Chen's compactness analysis is flawed because he uses the Reock measure of compactness, which is not the compactness measure required by Michigan law, and because it is unclear whether there is a substantively meaningful difference between the compactness of the Enacted Plan and Dr. Chen's simulated maps. (Id. at 11-14.) Third, Dr. Liu asserts that Dr. Chen failed to provide him with a copy of the source code used to create his simulations, which would have allowed Dr. Liu to evaluate Dr. Chen's algorithm and “critique additional flaws as to his methodology.” (Id. at 25.)

         The Court finds that Dr. Chen's data is reliable notwithstanding Dr. Liu's criticisms. Dr. Chen's data has been peer reviewed, published in academic journals, and deemed credible and admissible in several other redistricting cases. Further, Dr. Liu has not shown that Dr. Chen's data would produce alternate results if Dr. Chen's methodology were altered in the manner that Dr. Liu suggests. Moreover, Dr. Liu has neither personally generated any simulated maps nor evaluated the partisanship of Michigan's maps under the Enacted Plan. Dr. Liu's examination of Dr. Chen's simulated maps merely consisted of an “eyeball assessment.” (Id. at 12.) Regarding Dr. Chen's source code, Dr. Chen testified that he provided the final computer-readable code that he used to run his simulations and a near-final version of the source code, which was structurally identical and substantially the same as the source code that was compiled to create the simulated maps. (Chen Dep. at 50:4-52:7.) Dr. Liu contends that if he had access to a version of Dr. Chen's source code that was more easily readable, he could have found other “flaws” in Dr. Chen's methodology. But there is no evidence that Dr. Chen's findings were erroneous in any way, or that Dr. Liu would have found that Dr. Chen's methodology lacked credibility. Dr. Chen's findings are not excludable merely because Dr. Liu did not receive Dr. Chen's source code in the exact format that would have been ideal for Dr. Liu.

         2. Dr. Christopher Warshaw[26]

         Plaintiffs next expert, Christopher Warshaw, Ph.D. (“Dr. Warshaw”), evaluated the partisanship of the Enacted Plan using three different statistical measures: the efficiency gap, median-mean difference, and declination. (Warshaw Report at 6-12.) He also used a historical lens to evaluate the degree of partisan bias in the Enacted Plan. Like Dr. Chen, Dr. Warshaw concluded that the Enacted Plan strongly and systematically advantages Republicans and disadvantages Democrats. (Id. at 4-5.)

         a. Efficiency Gap

         i. Congressional Plan

         Dr. Warshaw found that, in recent congressional elections, Michigan had a pro-Republican efficiency gap that is extreme compared to its own historical efficiency gaps and the historical efficiency gaps in other states. (Id. at 16-19.) Dr. Warshaw reported that the efficiency gaps in Michigan's past three congressional elections “were among the most Republican-leaning efficiency gaps the nation has ever seen.” (Id. at 17.)

         For example, in the 2012 Michigan congressional election, there was a pro-Republican efficiency gap of approximately -19.7%. (Id. at 17.) Democrats wasted over 1.5 million votes, while Republicans wasted only 650, 000 votes. (Id.) Republicans' greater efficiency at translating their votes into seats resulted in their enjoying 64.3% of the congressional seats despite earning only 47.3% of the vote. (Id.)

         The 2014 and 2016 Michigan congressional elections also had efficiency gaps that greatly advantaged Republicans. These elections had efficiency gaps of approximately -16% and -13.2%, respectively. (Id. at 18) In these elections, Republicans won 64.3% of Michigan's congressional seats, even though they lost the statewide vote in 2014 and only narrowly earned the statewide vote in 2016. (Id.) Dr. Warshaw explained that these efficiency gaps “imply that Republicans in Michigan won 2-3 more seats in these elections than they would have won if Michigan had no partisan bias in its [e]fficiency [g]ap.” (Id.)

         Overall, Dr. Warshaw found that the efficiency gaps in Michigan were similar to the efficiency gaps of other states until the Enacted Plan came into effect. (Id. at 19.) However, “[a]fter the most recent redistricting, Michigan had more extreme pro-Republican [e]fficiency [g]aps than it has ever had before.” (Id.) According to Dr. Warshaw, the 2012 efficiency gap in Michigan's congressional elections was “more extreme than 95% of previous plans in states with more than six seats over the past 45 years, and it was more Republican-leaning than 98% of previous congressional districting plans.” (Id.)

         Using the efficiency gap measure, Dr. Warshaw concluded that the Enacted Plan's congressional maps have one of the largest partisan biases of any congressional districting plan in history. (Id.) He further found that the extreme pro-Republican advantage in Michigan's congressional districts, which manifested immediately after the Enacted Plan was implemented, is unlikely to have been caused by political geography. (Id.)

         ii. Senate and House Plans

         Dr. Warshaw found that the efficiency gaps in Michigan's House and Senate districts during the 2012, 2014, and 2016 elections were among the most pro-Republican efficiency gaps in history. (Id. at 33.) For example, in the 2012 House elections, Democratic candidates earned 54% of the votes but only gained 46% of Michigan's House seats, which yielded a pro-Republican efficiency gap of approximately -12.3%. (Id. at 34.) In the 2014 Senate elections, Democrats received about 49% of the votes but only 29% of the Senate seats, which yielded a pro-Republican efficiency gap of almost -20%. (Id. at 33-34.)

         When considered in a historical context, the pro-Republican efficiency gaps in Michigan's Senate and House plans are as problematic as the pro-Republican efficiency gaps in Michigan's congressional plan. Dr. Warshaw explained that Michigan's 2014 Senate election had a larger pro-Republican efficiency gap than 99.7% of the of the state senate elections over the past fifty years. (Id. at 36.) He also found that Michigan's House election in 2012 had a larger pro-Republican efficiency gap than 98% of the state house elections over the past 50 years and had a “larger absolute bias” than 91% of previous plans. (Id. at 35.)

         b. Median-mean Difference[27]

         Dr. Warshaw's median-mean analysis also demonstrates that the Enacted Plan strongly favors Republicans. Dr. Warshaw found that, in the 2012 congressional election, there was a 6.9% pro-Republican bias in the partisan composition of the median congressional district compared to the “average” congressional district. (Id., at 19.) This was “more extreme than the median-mean difference in 78% of previous elections and more pro-Republican than the median-mean difference in 89% of previous [congressional] elections.” (Id.) Furthermore, like Dr. Chen, Dr. Warshaw found that the Enacted Plan packed Democratic voters into 5 congressional districts, which Democratic candidates generally won by “overwhelming margins, ” and cracked the remaining Democratic voters across the other 9 congressional districts. (Id. at 9-10.)

         Dr. Warshaw concluded that the median-mean differences in Michigan's Senate and House districts are similarly extreme. For example, the median-mean difference in Michigan's 2014 Senate election was more extreme than in 95% of previous state senate elections nationwide. (Id. at 36.) And in the three House elections that took place between 2012 and 2016, the median-mean difference was greater than in 97% of previous state house elections. (Id.)

         c. Declination

         As Dr. Warshaw explains, the declination metric assumes “that a plan drawn with the intent to advantage one party will arrange the distribution of district vote shares in a way that treats the 50 percent threshold for victory differently than other vote values.” (Id. at 10.) If all the districts in a plan drawn without partisan intent are lined up from the least Democratic to the most Democratic, “then the mid-point of the line formed by one party's seats should be about as far from 50 percent on average as the other party's.” (Id.) If a plan is not deliberately drawn to favor one party over the other, the angles of the lines representing each parties' mean vote share in each district should be roughly equal. (Id. at 11.) When the lines deviate from each other, the smaller angle will “generally identify the favored party.” (Id.) In Michigan, the line representing Republicans' mean vote share across congressional elections has a much smaller angle than that representing Democrats' mean vote share, indicating that the Enacted Plan's congressional districts favor Republicans. (Id.)

         Dr. Warshaw notes that one weakness of the declination approach is that it lacks a “clear interpretation in terms of the number of seats that a party gains through gerrymandering.” (Id. at 15.) However, Dr. Warshaw also explains that some scholars claim that it represents a better measure of intent in the gerrymandering process than the efficiency gap. (Id.) Additionally, declination is arguably less sensitive to the outcome of close elections than the efficiency gap or the median-mean difference. (Id.)

         The declination measure demonstrates that the Enacted Plan favors Republicans. Michigan's 2012 congressional election had a declination score that was more extreme than 91% of previous congressional elections and more pro-Republican than 96% of previous congressional elections over the past 45 years. (Id. at 20.) Michigan's 2014 Senate election had a more extreme declination value than 96% of previous state senate elections and a larger pro-Republican declination value than 99% of the previous state senate elections. (Id. at 36.) Similarly, Michigan's 2012 House election had a more extreme declination value than 90% of previous state house elections and a larger pro-Republican declination value than 97% of the previous state house elections. (Id.) Therefore, like the efficiency gap and the median-mean difference, the declination metric indicates that the Enacted Plan strongly advantages Republicans and disadvantages Democrats.

         d. Review of Dr. Chen's Simulations

         In addition to conducting his own analysis, Dr. Warshaw reviewed Dr. Chen's findings. Dr. Warshaw created charts comparing the partisanship of each Individual Plaintiff and League Member's district under the Enacted Plan to the partisanship of that same district under Dr. Chen's simulations. (See Pls.' Trial Ex. 278.) The charts contain red X's showing the partisanship of each current district and gray bars denoting Dr. Chen's simulated districts for each Individual Plaintiff and League Member's current address under each type of map, congressional, Senate, and House. (Warshaw Trial Tr., ECF No. 248 at PageID #8824:20-23.) By comparing the red X's to the gray bars, Dr. Warshaw determined whether a district under the Enacted Plan falls within the range of simulated districts drawn without partisan intent that include that same Individual or League Plaintiff's address. (Id. at PageID #8824:24-8825:2.)

         Dr. Warshaw explained that if a district under the Enacted Plan falls outside the range of simulated districts, its partisan composition is extremely unlikely to have occurred by chance. (Id. at PageID #8825:11-14.) However, the mere fact that a district falls within Dr. Chen's simulations does not indicate that the district was drawn without partisan intent. For example, if a district falls within only a small number of Dr. Chen's simulations, it could still exhibit “more extreme partisanship than 99 percent of Dr. Chen's simulations.” (Id. at PageID #8918:14-19.) Thus, if a current district falls along the outer range of Dr. Chen's simulated districts, it is extremely unlikely that that the district's partisan composition occurred by chance. (Id. at PageID #8918:2-5.)

         Dr. Warshaw determined that many of the Enacted Plan's districts contained pro-Republican partisan compositions that fall completely outside the range of Dr. Chen's simulated districts. (Pls.' Trial Ex. 278.) At trial, Dr. Warshaw used Plaintiff Rosa Holliday, who resides in Congressional District 5, as an example of this phenomenon. Dr. Warshaw explained that none of the 1, 000 simulated congressional districts containing Holliday's address were as heavily Democratic as the district in which she lives, indicating that Congressional District 5 “packs” Democratic voters. (Warshaw Trial Tr., ECF No. 248 at PageID #8826:14-18.)

         Dr. Warshaw also found that the partisanship of some districts lies within, but at the extreme outer ranges of, the partisanship of Dr. Chen's simulated districts. At trial, Dr. Warshaw used Plaintiff Karen Sherwood, who lives in Congressional District 4, as an example of this phenomenon. Dr. Warshaw explained that while Congressional District 4, as currently drawn, is within the range of Dr. Chen's simulations for districts containing Sherwood's address, Congressional District 4's partisanship is “certainly more extreme than the vast majority” of simulated districts. (Id. at PageID #8827:17-8828:7.) Dr. Warshaw testified that the 4th Congressional District's partisanship is more extreme than 95% of the simulated congressional districts, indicating that its partisan composition is very unlikely to have occurred by chance. (Id. at PageID #8828:8-14.)

         e. Criticisms

         The Congressional and Senate Intervenors attempt to discredit Dr. Warshaw's findings, but the Court is not convinced that his results are unreliable. Their first critique is that the efficiency gap is not a dependable measure because it has been criticized by some political scientists. (Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #11008-9; Senate Intervenors' Proposed Findings of Fact, ECF No. 255 at PageID #10421.) The Court is not swayed by this argument. Contrary to Intervenors' assertion, recent literature has characterized the efficiency gap metric as a “powerful way” to evaluate the existence of partisan gerrymandering.[28] Furthermore, as noted above, Dr. Warshaw does not rely on the efficiency gap in isolation. He compares the efficiency gap of the Enacted Plan to historical measures of the efficiency gap. He also uses the median-mean difference and the declination metric. Each comparison and metric strongly supports his conclusion that the Enacted Plan advantages Republicans and disfavors Democrats.

         Next, the Congressional and State House Intervenors challenge the legitimacy of Dr. Warshaw's two other measures-median-mean difference and declination. The Congressional and State House Intervenors argue that these measures have not been widely accepted in the political science community and urge the Court to consider them flawed. (Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #11018-20.) In support of their contention, Congressional and State House Intervenors cite to sections of Dr. Warshaw's deposition testimony where he discusses criticisms of these two measures that have been voiced in some scientific publications. (Warshaw Dep. at 171:5-7; 177:3-8.) However, Congressional and State House Intervenors conveniently fail to cite other parts of Dr. Warshaw's deposition testimony where he explains why the criticisms concerning the median-mean difference and declination measures do not undermine his conclusions in this case. Dr. Warshaw explains that “it's important to remember that the measure you use is relatively unimportant . . . at the end of the day all these measures of gerrymandering are extremely highly correlated, particularly in states like Michigan with competitive elections.” (Id. at 167:2-12.)

         Dr. Warshaw's point, with which the Court agrees, is that while one could reasonably criticize certain aspects of each measure used to evaluate partisan gerrymandering, these criticisms do not diminish Plaintiffs' experts' findings given that all of the measures they use uniformly and unequivocally point to the same conclusion: the Enacted Plan strongly, and systematically, advantages Republicans and disadvantages Democrats.

         3. Dr. Kenneth Mayer[29]

         Plaintiffs' last expert, Kenneth Mayer, Ph.D., (“Dr. Mayer”), evaluated whether the Enacted Plan constituted an extreme partisan gerrymander. (Mayer Report at 3.) Dr. Mayer used various metrics to analyze the partisanship of Enacted Plan, namely: (1) partisan bias;[30] (2) partisan symmetry;[31] (3) efficiency gap; (4) median-mean difference; and (5) declination.[32] Like the evidence presented by Plaintiffs' other experts, Dr. Mayer's statistical analyses all indicate that the Enacted Plan constitutes a pro-Republican partisan gerrymander.

         a. Partisan Bias and Partisan Symmetry i. Congressional Map

         Dr. Mayer found that in the congressional elections that took place from 2012 to 2016, the partisan bias was -16.6%, which when adjusted to 50% of the statewide vote using a uniform swing analysis, becomes -14.3%. (Mayer Report at 30-31.) In other words, if Republicans won 50% of the statewide vote under the Enacted Plan, they would be estimated to win approximately 64.3% of the congressional seats. (Id.) Dr. Mayer further found that, to win a majority of the congressional seats under the Enacted Plan, Democrats would need to win 57.5% of the statewide vote. (Id. at 30.) This is in stark contrast to Republicans, who won 9 of the 14 congressional seats (64.3%) in each election between 2012 and 2016 despite never securing more than 50.5% of the statewide vote. Dr. Mayer also determined that if the Republicans received the same percentage of the statewide vote as Democrats across these congressional elections, they would have won 9 seats compared to the 5 seats that the Democrats won. (Id.)

         ii. Senate Map

         Dr. Mayer found that during the 2014 Senate elections, the partisan bias of the Senate map was -15.5%, which when adjusted to 50% of the statewide vote using a uniform swing analysis, becomes -15.8%. (Id. at 50.) In other words, even if Republicans only won 50% of the vote, they were estimated to win approximately 65.8% of the Senate seats. (Id.) Dr. Mayer also found that, if Republicans received the same percentage of the aggregate vote as Democrats received in the 2014 Senate election, they would have won 27 seats compared to the 14 that Democrats won. (Id.)

         iii. House Map

         Dr. Mayer found that during the House elections that took place between 2012 and 2016, the partisan bias of the House map was -7.8%, which when adjusted to 50% of the statewide vote using a uniform swing, becomes -9.1%. (Id. at 40.) In other words, even if Republicans only won 50% of the vote, they would be estimated to win 59.1% of the House seats. (Id.) Dr. Mayer also found that if Republicans received the same percentage of the aggregate vote as Democrats, they would have won 70 seats compared to the 49 that Democrats won on average. (Id.)

         b. Efficiency Gap

         Dr. Mayer determined that the efficiency gap for the congressional elections held between 2012 and 2016 was “extraordinarily large and negative, ” meaning that it strongly favored Republicans. (Id. at 31-32.) Specifically, the efficiency gap for these elections was -19.7%, which translated to Republicans winning approximately 3 additional seats in each congressional election. (Id. at 32.) Dr. Mayer similarly found that efficiency gap for the Senate map in the 2014 elections was -19.3%. (Id. at 51.) He further found that efficiency gap for the House map between 2012 and 2016 was -11.9%, which indicates a pro-Republican bias that “endur[ed] over multiple election cycles.” (Id. at 41.)

         c. Median-mean Difference

          Like the other statistical metrics used by Dr. Mayer, the median-mean difference demonstrates a strong pro-Republican bias across all three maps. Dr. Mayer found that the median-mean measure for the congressional, Senate, and House maps is -7.7%, -6.1%, and -6.9%, respectively. (Id. at 32; 42; 51.) These results indicate that the Enacted Plan weighs Republican and Democratic votes unequally. (Id.)

         d. Declination

         Dr. Mayer also found that the declination measure demonstrates the Enacted Plan's strong pro-Republican bias. According to Dr. Mayer, the declination metric for Michigan's current congressional, Senate, and House plans for 2012-2016 is 0.398, 0.380, and 0.243, respectively. (Id. at 35; 44; 54.) These values suggest that the Enacted Plan cracks and packs Democrats and thus advantages Republicans. (Id.)

         e. Criticisms

         Intervenors' main criticism of Dr. Mayer's findings is that they are based on Dr. Chen's data, which Intervenors claim is unreliable. (See, e.g., Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #11002-3.) However, as previously noted, Dr. Chen is widely-renowned as an expert in his field and he has provided admissible expert testimony in several redistricting cases. Furthermore, the Court has determined that Dr. Chen's data and expert findings are reliable.

         Congressional and State House Intervenors also argue that the Court should disregard Dr. Mayer's findings because he relied on the efficiency gap, which they contend is an inherently flawed measure, and because Dr. Mayer's efficiency gap scores did not perfectly match those calculated by Dr. Chen. (Cong. and State House Intervenors' Proposed Findings of Fact, ECF No. 258 at PageID #11062.) But, as the Court noted above, the efficiency gap is a widely-accepted measure of partisan gerrymandering. While reasonable criticisms of the measure exist, these criticisms fail to diminish the power of Plaintiffs' experts' findings given that all the metrics unvaryingly point towards the Enacted Plan being a partisan gerrymander. The same is true with the discrepancies in Dr. Chen and Dr. Mayer's efficiency gap scores; while their scores do not perfectly match, both experts' analyses show an extremely large and durable pro-Republican efficiency gap across Michigan's congressional, Senate, and House maps. Minor inconsistencies between their data do not diminish the power of their findings, particularly given that all of Plaintiffs' experts' evidence unequivocally supports the same conclusion: the Enacted Plan profoundly, and systematically, advantages Republicans and disadvantages Democrats.

         C. Summary

         Based on the robust qualitative evidence of discriminatory partisan intent discussed in Part II.A, and the equally powerful quantitative evidence of discriminatory partisan intent and discriminatory partisan effects discussed in Part II.B, the Court concludes that the Enacted Plan was designed with the predominant purpose of advantaging Republicans and discriminating against Democrats. The Court further finds that the Enacted Plan achieved its intended effects because it discriminated against Democratic voters in numerous elections across multiple election cycles. Therefore, the Enacted Plan constitutes a durable partisan gerrymander.

         Having made its factual findings above, the Court will now turn to the relevant legal issues.

         III. LACHES

         A. Introduction

         Intervenors argue that the equitable doctrine of laches bars Plaintiffs claims because Plaintiffs unreasonably delayed in asserting their rights. Intervenors observe that Plaintiffs did not file suit until December 2017, over six years after passage of the Enacted Plan and after three election cycles had already been held under the Enacted Plan. (See Senate Intervenors' Proposed Conclusions of Law, ECF No. 254 at PageID #10382-88; Cong. and State House Intervenors' Proposed Conclusions of Law, ECF No. 258 at PageID #11148-49.) Intervenors argue that Plaintiffs should have known that the Enacted Plan constituted an alleged partisan gerrymander in June 2011 and, at the very latest, Plaintiffs knew that the Enacted Plan was an alleged partisan gerrymander in 2015 when the League hired expert witnesses to examine the issue. (Cong. and State House Intervenors' Proposed Conclusions of Law at PageID #11148-49.) They also argue that laches bars Plaintiffs' claims because any relief would require drawing new maps using census data from 2010 which is no longer accurate. (Senate Intervenors' Proposed Conclusions of Law at PageID #10383-86.)

         In response, Plaintiffs contend that laches does not apply to Plaintiffs' partisan gerrymandering claims as a matter of law because Plaintiffs are entitled to seek injunctive relief to prevent continuing constitutional harms and because the Enacted Plan causes ongoing violations of their constitutional rights. (Pls.' Proposed Conclusions of Law, ECF No. 260-1 at PageID #11480-85.) Plaintiffs alternatively argue that, even if laches applies to partisan gerrymandering claims in theory, it does not bar their claims here, because Plaintiffs acted diligently to assert their rights and because Intervenors have not demonstrated that they have suffered any prejudice because of any alleged delay. (Id. at PageID #11485-87.)

         The Court holds that laches does not apply to Plaintiffs' partisan gerrymandering claims as a matter of law. In the alternative, the Court holds that even if laches applies to these types of claims, Intervenors have failed to establish that laches bars Plaintiffs claims in this case.

         B. Legal Standard

         “Where a plaintiff seeks solely equitable relief, his action may be barred by the equitable defense of laches if (1) the plaintiff delayed unreasonably in asserting his rights and (2) the defendant was prejudiced by this delay.” Am. Civil Liberties Union of Ohio, Inc. v. Taft, 385 F.3d 641, 647 (6th Cir. 2004) (citing Brown-Graves Co. v. Cent. States, Se. and Sw. Areas Pension Fund, 206 F.3d 680, 684 (6th Cir. 2000)). The Sixth Circuit has defined laches as the “‘negligent and unintentional failure to protect one's rights.'” Nartron Corp. v. STMicroelectronics, Inc., 305 F.3d 397, 408 (6th Cir. 2002) (quoting Elvis Presley Enter., Inc. v. Elvisly Yours, Inc., 936 F.2d 889, 894 (6th Cir. 1991)).

         “Laches only bars damages that occurred before the filing date of the lawsuit.” Id. at 412 (internal citation omitted). “It does not prevent plaintiff[s] from obtaining injunctive relief or post-filing damages.” Id.; see Kellogg Co. v. Exxon Corp., 209 F.3d 562, 568 (6th Cir. 2000) (holding that laches “does not bar injunctive relief”) (citing TWM Mfg. Co., Inc. v. Dura Corp., 592 F.2d 346, 349-50 (6th Cir. 1979)). Laches does not apply to ongoing or recurring harms because while “[l]aches stems from prejudice to the defendant occasioned by the plaintiff's past delay . . . almost by definition, the plaintiff's past dilatoriness is unrelated to a defendant's ongoing behavior that threatens future harm.” Danjaq LLC v. Sony Corp., 263 F.3d 942, 959-60 (9th Cir. 2001) (citing Lyons P'ship, L.P. v. Morris Costumes, Inc., 243 F.3d 789, 799 (4th Cir. 2001)); see also Kuhnle Bros., Inc. v. Cty. of Geauga, 103 F.3d 516, 522 (6th Cir. 1997) (holding that “a law that works an ongoing violation of constitutional rights does not become immunized from legal challenge” merely because the plaintiff failed to sue within the applicable statute of limitations).

         Notably, a recent three-judge panel in the Sixth Circuit held that laches does not apply as a matter of law to partisan gerrymandering claims. See Ohio A. Philip Randolph Inst. v. Smith, 335 F.Supp.3d 988, 1002 (S.D. Ohio 2018) (three-judge panel).

         C. Laches Does Not Apply as a Matter of Law Because Plaintiffs Seek to Redress Ongoing Harms and Recurring Violations of Their Constitutional Rights

         The Court finds that laches does not bar Plaintiffs' claims as a matter of law. Plaintiffs seek declaratory and injunctive relief. The Sixth Circuit has held that laches does not apply to claims for prospective relief, see Nartron Corp., 305 F.3d at 412 and Kellogg Co., 209 F.3d at 568, and a three-judge panel in the Sixth Circuit has held that laches does not apply to allegations of partisan gerrymandering, see Smith, 335 F.Supp.3d at 1002. In this case, Plaintiffs assert that the Enacted Plan has injured, and will continue to harm, their First and Fourteenth Amendment rights. They ask this Court to declare the Challenged Districts unconstitutional and enjoin their use in future elections to prevent further harm to their constitutional rights. Laches does not apply to Plaintiffs' claims for declaratory and injunctive relief. See Nartron Corp., 305 F.3d at 412; Kellogg Co., 209 F.3d at 568; Smith, 335 F.Supp.3d at 1002.

         Our holding-that laches does not apply as a matter of law to partisan gerrymandering claims-is consistent with the Supreme Court's pronouncement, in Bandemer, that

an equal protection violation may be found only where the electoral system substantially disadvantages certain voters in their opportunity to influence the political process effectively. In this context, such a finding of unconstitutionality must be supported by evidence of continued frustration of the will of a majority of the voters or effective denial to a minority of voters of a fair chance to influence the political process.

Davis v. Bandemer, 478 U.S. 109, 133 (1986). In Bandemer, the Supreme Court held that “[r]elying on a single election cycle to prove unconstitutional discrimination is unsatisfactory” to establish partisan gerrymandering. Id. Given that relying on a single election cycle is “unsatisfactory” and that plaintiffs must demonstrate a “continued frustration” to prevail on their partisan gerrymandering claims, laches cannot bar Plaintiffs' claims on the basis that they waited three election cycles to sue. See generally, id. To hold otherwise would force plaintiffs to sue after exactly two elections have been held under a challenged districting plan. This Court is not aware of any authority supporting such a rigid temporal requirement.

         To argue that laches applies to Plaintiffs' claims, Intervenors unconvincingly cite cases that this Court has already distinguished in our opinion on summary judgment. For example, Congressional and State House Intervenors rely on Benisek v. Lamone, 138 S.Ct. 1942, 1944 (2018) (see Cong. and State House Intervenors' Proposed Conclusions of Law at PageID #11147). But as we previously stated, Benisek involved a preliminary injunction and did not directly address laches. See League of Women Voters of Michigan v. Johnson, 352 F.Supp.3d 777, 808 (E.D. Mich. 2018) (three-judge panel). Congressional and State House Intervenors also rely on Block v. N. Dakota ex rel. Bd. of Univ. & Sch. Lands, 461 U.S. 273 (1983). (Cong. and State House Intervenors' Proposed Conclusions of Law at PageID #11147.) However, as we have already observed, Block “is completely inapposite because it involved title to the bed of a river, not gerrymandering.” Id. at 809. These misguided analogies to Benisek and Block failed to convince us at the summary judgment stage and similarly fail to persuade us now.

         D. In the Alternative, Intervenors Have Failed to Prove Their Affirmative Defense of Laches

         Even if the doctrine of laches were applicable to partisan gerrymandering claims, it would not bar Plaintiffs' claims in this case for two independently sufficient reasons. First, Plaintiffs did not unreasonably delay in asserting their rights. Second, Intervenors did not suffer prejudice from any alleged delay. See, e.g., Taft, 385 F.3d at 647. Therefore, Intervenors have failed to establish their affirmative defense of laches. See generally E.E.O.C. v. Watkins Motor Lines, Inc., 463 F.3d 436, 439 (6th Cir. 2006) (“As laches is an affirmative defense, the burden of establishing both of these elements is on the party raising the defense.”)

         Plaintiffs did not unreasonably delay in asserting their rights. See Taft, 385 F.3d at 647. The three election cycles that occurred under the Enacted Plan before Plaintiffs filed their lawsuit solidified Plaintiffs' concerns that the Enacted Plan is a partisan gerrymander and added legitimacy and support to Plaintiffs' legal claims. Given that the Supreme Court has held that relying on a single election cycle is “unsatisfactory” to prove partisan gerrymandering, Plaintiffs did not act unreasonably by waiting until three elections had been held to sue. Furthermore, this is an evolving area of the law that has experienced significant developments in recent years; it was not unreasonable for Plaintiffs to wait to sue until the law in this area had developed sufficiently to allow Plaintiffs to articulate and support their partisan gerrymandering claims. Intervenors have therefore failed to satisfy their burden of proving that Plaintiffs unreasonably delayed in asserting their rights. Id.

         Additionally, Intervenors have not demonstrated that they suffered prejudice because of any delay in Plaintiffs' decision to sue. See id. In support of their assertions of prejudice, Intervenors posit that, because of Plaintiffs' alleged delay, witnesses might have failed to remember certain events and may have lost emails and other documents. (Cong. and State House Intervenors' Proposed Conclusions of Law at PageID #11148; Senate Intervenors' Proposed Findings of Fact, ECF No. 255 at PageID #10447-48.) However, Intervenors have failed to substantiate these amorphous allegations. They argue, in a speculative and conclusory fashion, that “they were unable to potentially mount as vigorous [a] defense” as they would have had Plaintiffs sued sooner (Senate Intervenors' Proposed Findings of Fact at PageID #11148), but they have not explained with any particularity which emails, documents, and conversations were lost to the passage of time or how access to these materials would have helped them defend against Plaintiffs' claims. Further, Intervenors undoubtedly benefitted from the fact that Plaintiffs did not sue until 2016 insofar as, to date, four election cycles have been held under the Enacted Plan. Intervenors have therefore failed to satisfy their burden of proving that they suffered prejudice because of Plaintiffs' alleged delay. Id.

         Intervenors have failed to prove their affirmative defense of laches. Even if laches applied to partisan gerrymandering claims, it would not bar Plaintiffs' claims in this case.

         IV. JUSTICIABILITY AND SUBSTANTIVE STANDARDS

         The Supreme Court has held that partisan gerrymandering claims are justiciable. See Bandemer, 478 U.S. at 127 (holding that partisan gerrymandering claims present justiciable controversies under the Equal Protection Clause of the Fourteenth Amendment). It is true that, in Bandemer, the Supreme Court “could not . . . settle on a standard for what constitutes an unconstitutional partisan gerrymander.” Gill, 138 S.Ct. at 1927 (discussing Bandemer, 478 U.S. at 125, 127). Nonetheless, the Supreme Court has never overturned Bandemer's holding that political gerrymandering claims are justiciable. In Vieth, a four-justice plurality would have overturned Bandemer because they believed that a judicially-manageable standard does not exist to analyze claims of partisan gerrymandering. Vieth, 541 U.S. at 288-91; 305-07. However, Justice Kennedy, who concurred in the judgment, refused to hold that partisan gerrymandering claims are categorically non-justiciable and reaffirmed that Bandemer constitutes “controlling precedent on the question of justiciability.” Id. at 310 (Kennedy, J., concurring in the judgment) (stating that “[o]ur willingness to enter the political thicket of the apportionment process with respect to one-person, one-vote claims makes it particularly difficult to justify a categorical refusal to entertain claims against this other type of gerrymandering”). The Supreme Court declined to reconsider Bandemer's holding in its most recent decisions involving partisan gerrymandering, League of United Latin Am. Citizens v. Perry, 548 U.S. 399 (2006) and Gill, 138 S.Ct. 1916.

         In recent years, several three-judge panels have held that partisan gerrymandering claims are justiciable. See, e.g., Ohio A. Philip Randolph Inst. v. Householder, No. 1:18-CV-357, 2019 WL 652980, at *2-7 (S.D. Ohio Feb. 15, 2019) (three-judge panel); Rucho, 318 F.Supp.3d at 838; see Benisek, 348 F.Supp.3d at 513; see also Shapiro, 203 F.Supp.3d at 594. And lower federal courts have formulated judicially-manageable standards for adjudicating partisan gerrymandering claims. Householder, 2019 WL 652980, at *3-4 (explaining that federal courts have “converged considerably on common ground in establishing standards for determining whether a partisan gerrymander is unconstitutional”). In these cases, federal courts have largely agreed on a three-part framework for evaluating the constitutionality of alleged partisan gerrymanders whereby, to prevail, Plaintiffs must demonstrate: (1) discriminatory partisan intent, (2) discriminatory partisan effects, and (3) causation and/or a lack of justification. See Id. at *4 (applying three-part discriminatory partisan intent, discriminatory partisan effects, and lack of legitimate justification test for Fourteenth Amendment and First Amendment claims); Rucho, 318 F.Supp.3d at 860-68 (articulating three-part discriminatory intent, discriminatory effects, and lack of justification test for Fourteenth Amendment claims and a similar three-part test for First Amendment claims); Benisek, 348 F.Supp.3d at 515 (formulating a three-part discriminatory intent, discriminatory effects, and causation test for First Amendment claims); Shapiro, 203 F.Supp.3d at 598 (using a three-part intent, effects, and causation standard for First Amendment claims).

         In keeping with these other federal courts, we previously determined, at the summary judgment stage, that judicially-manageable standards exist to adjudicate Plaintiffs' partisan gerrymandering claims under the three-part discriminatory intent, discriminatory ...


Buy This Entire Record For $7.95

Download the entire decision to receive the complete text, official citation,
docket number, dissents and concurrences, and footnotes for this case.

Learn more about what you receive with purchase of this case.