STATE OF NORTH CAROLINA IN THE GENERAL COURT OF JUSTICE SUPERIOR COURT DIVISION WAKE COUNTY 18 CVS 014001 Common Cause; et al. Plaintiffs, V. REPLY BRIEF IN RESPONSE TO OBJECTIONS Representative David R. Lewis, in his of?cial capacity as senior chairman of the House Select Committee on Redistricting, er Defendants. Plaintiffs agree that the Senate plan complies with governing law and the Court?s decree. Implicit in that position is agreement with the approach the Senate utilized?winnowing down a set of Dr. Chen?s simulated maps to the ?ve optimal group maps using a score Dr. Chen developed, picking one at random, and allowing minor changes to accommodate neutral goals. It should also follow that the House plan is valid. The House utilized the same approach as the Senate in all material respects: it winnowed down Dr. Chen?s simulated maps to the ?ve optimal group maps using Dr. Chen?s own criteria, picked one at random for each affected grouping, and allowed minor changes under highly restrictive criteria that were all executed in public and by non-partisan staff. Plaintiffs? objections are a case of selective outrage that smacks of partisan manipulation. They challenge only the House plan and focus primarily on only five groupings?which they ask the Court to redraw from scratch even though most lines are Dr. Chen?s own work. And the nature of Plaintiffs? criticismsraises more questions still, since they all ring hollow. As shown below, their contention that the House should have used Chen set 2 rather than set 1 ignores that this choice was a concession to Democratic members. Likewise, Plaintiffs? contention that the General Assembly had access to political data ignores that the information was deleted and never used. And the fact that Plaintiffs only attack ?ve groupings completely undercuts any remaining integrity of this argument. Plaintiffs complain that the House utilized experts to authenticate the information Dr. Chen provided to the Committee, but Rep. Lewis stated from the chair of the House Committee?when Plaintiffs? counsel were present?what was occurring. Separately, Plaintiffs? Counsel represented to multiple Senators in a discussion in the Senate Committee room that Legislative Defendants could use their outside experts for the limited purpose of verifying the authenticity of the Chen Data, in compliance with the Court order. The experts did not ?assist in the map?drawing process,? Court?s Decree 1] 9, but simply vetted Dr. Chen?s remedial-phase production to ensure that it matched his production at the liability phase. Under the Court?s order, Plaintiffs? assertion that the House Should have utilized a new and untested set of maps would have required far more expert involvement. Nor do Plaintiffs have a shred of evidence that the House?s outside legal counsel, who are also the outside counsel for the Senate, were providing political advice. They were not. They lack even the ability to do so. So Plaintiffs? complaints boil down to the assertion that ?ve county groupings are ?outliers,? a fact for which they cite a new set of simulated maps. But Dr. Chen made a critical mistake that dooms this ?Chen set he allowed his algorithm to redraw districts the House Committee considered frozen in place. As a result, set 3 says nothing of partisan purpose or effect?or anything else. And Plaintiffs? grouping-speci?c arguments are a contradictory mess: a grouping changed should have changed more, a grouping changed somewhat more should have changed only a bit, and a grouping not changed at all should have been totally redrawn. Moreover, as an objective measure, the remedial House plan is free from partisan bias as shown using multiple measures of partisan bias often used by Democratic redistricting plaintiffs. In short, Plaintiffs? objections fail and, worse, appear to be the product of the partisan goals of the North Carolina Democratic Party. The House process was fully compliant with the Court?s decree, deferential to Democratic input, and executed in a way to avoid being infected by hidden partisan bias. It would be counterproductive to adopt objections seemingly partisan in nature to oust the decisions that were non-partisan and almost entirely the work of Dr. Chen?s Computer. All but one grouping passed Committee by a unanimous vote, and all but one received supermaj ority, bipartisan support in the ?oor votes. The Court should decline to enjoin both plans or, at a minimum, return lines it rejects to the con?gurations created at random by Dr. Chen?s computer. ARGUMENT I. Plaintiffs? Objections Are Groundless It is (or appears to be) common ground between Plaintiffs and the General Assembly that the approach both chambers used constitutes a legitimate response to the Court?s judgment and decree. In fact, Plaintiffs agree (at 2) that the Senate? 3 remedial plan should not be enj oined.1 1 Nevertheless, Plaintiffs hint (at 3) that the Chen process might be invalid root and branch because (they say) ?the decision to use Dr. Chen?s simulated plans? was not made in public. This is incorrect: both House and Senate committees debated the idea and adopted the approach in public. What Plaintiffs appear to mean is that the idea did not ?rst emerge in public before it was conceived in private. But this is a simple consequence of individuals? minds and lives having a necessarily private core dimension. What matters is that the legislative the process by which it made its decisions?and all ?map drawing? occurred in public. Court?s Decree ii 8. Plaintiffs? failure to press this point in their argument section indicates that even they see it is meritless. To the extent it helps assuage this odd concern, Legislative Defendants represent that no consultants were utilized in discussing the idea of picking a Chen map at any point, no partisan analysis was performed, and the advice of legal counsel concerned whether the approach was legally valid g, is defensible under the Court?s decree and as a method of remediation)??a fair question to vet before proposing the approach publicly. With ?no reference whatsoever to the But, as explained in the General Assembly?s September 23 ?ling, both maps are predominantly the work of Dr. Chen?s method, which the Court already approved. In both chambers, the departures from those lines were limited, conducted in public, andmwith the exception of a single county grouping?enj oyed overwhelming, sometimes unanimous, bipartisan support. The resultingmaps are acts of the General Assembly and merit the highest deference from this Court. ?All doubts must be resolved in favor of the Act.? In re Housing Bonds, 296 281, 284 (N .C. 1982). The House map is an improbable candidate for invalidation as a gerrymander, and Plaintiffs? surprising objections are, unsurprisingly, meritless. A. The Base Map Complies With Law and the Court?s Order As Legislative Defendants? September 23 ?ling made clear, political data and partisan considerations were not part of the redistricting in either chamber. See Sept. 23 Br. 10?11 (disclaiming the use of political data and ?even the appearance of partisan motive?). Plaintiffs have no basis to contend otherwise, and they simply ask the Court to assume as much from inferences they draw from innocuous facts. Those inferences simply do not follow. This is precisely why acts of the General Assembly are entitled to a presumption of constitutionality and its proceedings a presumption of regularity.2 Wayne Cty. Citizens Ass ?n for Better Tax Control v. Wayne Cty. Bd. of Comm ?rs, 399 311, 315 (NC. 1991). attorney-client privilege,? the Court?s order cannotbe read to waive Legislative Defendants? right to private legal counsel. Dickson v. Racho, 737 362, 368 (NC. 2013). 2 Democracy North Carolina is wrong in contending that these acts of the General Assembly enjoy no presumption in their favor. The remedial posture does not change the plans? character as acts of the General Assembly. See Abbott v. Perez, 138 S. Ct. 2305, 2324?26 (2018). Stephenson v. Bartlett, 582 247, 251 (NC. 2003), did not hold otherwise, but rather faulted the State for failing ?to offer any evidence? of ?a compelling government interest? after a plan had been shown ?to violate the requirements of the North Carolina Constitution.? 1. The Choice of Chen Set 1 in the House Was Not Partisan Plaintiffs? assertion (at 3?4) that the House acted improperly by using Chen set 1 is incorrect, and their suggestion, with no evidence, that the House utilized set 1 because their counsel analyzed it and advised that set 1 is politically advantageous is absolutely false.3 Neither Legislative Defendants nor their counsel nor any agent used political data to assess any redistricting plan or proposal or any other element of the remedial process. In contrast, Plaintiffs? counsel were present at every hearing, they were assisted by Plaintiffs? non?testifying expert, and they were not hesitant to object when they believed there was even a whiff of political activity. They did not state any objection to the use of Chen set 1 at the time, and their post hoc objection should not be countenanced. Plaintiffs ignore the simple reality of what happened. Rep. Lewis proposed that Chen set 2 be used, Democratic members spent a committee session contesting the use of set 2, and the Republican members eventually conceded and agreed to begin with Chen set 1 and to make modi?cations to protect incumbents. The Committee chose set 1 unanimously. The timeline is as follows: On September 9, 2019, Rep. Lewis proposed to the House Redistricting Committee the structure of the redistricting by using a Chen ?base map? from Chen set 2. Tr. H. Redist Comm. Sept. 9, 2019 at 16:19?2219. He reasoned that incumbency protection is allowed and that set 2 would protect more incumbents than would set 1. Tr. H. Redist Comm. Sept. 9, 2019 at 1724?11. Rep. Lewis?s proposal was discussed in public at the hearings?in contradiction to Plaintiffs? suggestion (at 3) that the decision was made behind closed doors?and his proposal of set 2 drew significant criticism from Democratic members?again, in public with 3 It also bears reemphasizing that both Chen sets 1 and 2 were created by Dr. Chen?s algorithm, not the legislature, with no consideration of partisan advantage and were approved by the Court as non-partisan baselines. So, by definition, working within the range of Chen?s simulations of both sets is working within a range of non-partisan maps. Plaintiffs? counsel present. Democratic members asserted that using set 2 would be more favorable to Republican incumbents and, as a result, more favorable to Republicans. See, 3. g. Tr. H. Redist Comm. Sept. 10, 2019 at 36:24m38:20, 39:21?4l:2 (Rep. Reives); Id. at 26:15?23 (Rep. Queen). Initially, Rep. Lewis was resistant, arguing that ?the Chen report. . .found no statistical differences when the residence of the incumbent was considered.? Id at 4128?11. Likewise, Democratic members were persistent, insisting that ?the ?rst set is far and away the more pure and devoid of any partisan inference.? Id. at 50:11??13 (Rep. Butler). Rep. Queen called set 1 ?the best of the best.? Id. at 68:6?7. Ultimately, Democratic Rep. Reives made the proposal that was adopted: My position would be it would be best to start with Group 1 or Set 1, see how that?s affected by incumbency, because, again, Set 2 was not drawn based on present incumbency. And I know that there would be a motion coming up to try to include present incumbency. But since Set 2 already wasn?t drawn like that, we could start with Set 1, see what that does, and then we can use an incumbency overlay if we choose at a later time to try to ?x any problems that we may decide as a committee need to be fixed. So I would propose that we do Plan l?Set l, and start with those maps as our basis. Id. at 57:22?58:10. This was a compromise position; Rep. Butler proposed that incumbency protection be ignored entirely. See id. at 59223?6027. The discussion went on in public at length. Finally, on Sept. 11, after hearing out his colleagues concerns, Rep. Lewis stated as follows: I thought a lot about Representative Butler?s concerns from yesterday. And, while it may take a little more time to look at incumbents that are inadvertently paired, I think we will save that time by adopting the criteria that Representative Harrison brought up. So, I would, also, propose that we use the Chen Maps Set 1 instead of Set 2. Tr. H. Redist Comm. Sept. 11, 2019 at 3212?18. He also stated that his concession on this and other points was that he became convinced they were ?really good consensus ideas that could move us forward in a fast and fair way.? Id. at 2:7?8. And Rep. Lewis was right. The Committee unanimously adopted Chen set 1 and approved the base-map approach as described by Rep. Reives. Tr. H. Redist Comm. Sept. 11, 2019 at 17:3? 22. That is called compromise, and it happens in the legislature every day. The Court should applaud it, not penalize it. Plaintiffs could hardly be more off base, then, in alleging that the choice of Chen set 1 was a malicious conspiracy of Legislative Defendants and their counsel. Legislative Defendants and their counsel ?atly deny it. Plaintiffs have zero evidence?nonem?to support their absurd View. And their recklessness in leveling it is only unscored by the fact that?in public, before Plaintiffs? very eyes?Democratic members successfully lobbied for set 1. Plaintiffs? other arguments fall ?at. Their assertion (at 3) that House set 1 is better than set 2 for Republicans, even if true, only proves that the members were not behaving politically.4 Republican members wanted set 2 and Democratic members set 1. In Plaintiffs? View, both sides were advocating against their political interests. This disproves their unsupported conspiracy theory that legislators relied on partisan data. Further, Plaintiffs do not assert that the all?important choice of which con?guration to use had any partisan basis. Nor could they. The Committee chose Dr. Chen?s ranking system and used the State Lottery Commission?s personnel to pick the specific groupings adopted. So whatever 4 Moreover, it is unclear whether Plaintiffs are even right that set 1 was the more Republican- friendly set. They cite (at 3) to Dr. Chen?s report, page 27, but that page does not give partisan data for county groupings, and the House Committee was working at the grouping level, not the map-wide level. bene?t might in theory inhere in the choice of set 1 or set 2 was offset by the random draw, which could involve any Chen map. Plaintiffs also complain that the choice of Chen set 1 was not made in public, and that the persons involved were not disclosed. But, in fact, it was made in public, in the process described above, and the persons involved in the decision were the Committee membersw who debated and eventually approved the process.5 Plaintiffs" premise that there simply must be more than meets the is wrong. Like the Senate, the House made its choices in a public way after public debate. 2. No Expert of the House or Senate Was Involved in the Map-Drawing Process Plaintiffs? objection to the use of two experts, Dr. Thornton and Clark Bensen, to authenticate Dr. Chen?s submission to the legislature is baf?ing. These experts performed a Single task that was not part of the map-drawing process and was disclosed to Plaintiffs? counsel, and they approved. It was discussed at the Committee in public, and it resulted in no change of position on any issue, since the experts simply con?rmed that Dr. Chen provided the correct information. All of this arose from a simple problem that warranted a simple solution. When the House and Senate Committees chose to work off maps from Dr. Chen, the Committees needed Dr. Chen?s data to conduct the act of picking simulated county groupings. To that end, Plaintiffs? counsel provided the data they believed the General Assembly should use, but the data needed to be veri?ed as being the authentic data presented to the Court at the liability phase. Needless to say, the Committees were justified in not relying, without independent vetting, on the representations of an adverse lawyer and expert. 5 This objection, again, raises the problem identified above (see supra note 2) that ideas happen in private before they are communicated in public. What matters for present purposes is that the ?process? of legislative review and approval all occurred, without exception, in public. So, to solve the problem, and consistent with the representation of Stanton Jones, counsel to Plaintiffs, that con?rming the accuracy of the data by Legislative Defendants? experts would not violate the Court order, Legislative Defendants? counsel forwarded the data to Thornton and Bensen for authentication. Moreover, Rep. Lewis stated in public in the Committee hearing that this process was occurring, and no objection was lodged?indeed, no question was even asked. Tr. Redist. Comm., Sept. 9, 2019 at 4:9?5:14 defendants? counsel have asked for a chance to review this and make sure, indeed, that this is the same information that was before the Court because that?s what this committee has elected to consider?). Plaintiffs? counsel were in the room when Rep. Lewis made this repreSentation and raised no objection. A similar announcement was made by Senator Hise in public committee, and no objections were lodged. Tr. S. Redist. Comm. Sept. 10, 2019 at 50:10-15 (?We as Chairs are having outside counsel look at the data to verify the data is what it claims to be. Assuming there are no issues with that, I think we will?I think it?s prudent to proceed as if it Since Legislative Defendants? counsel lack the skill to vet the data, Legislative Defendants asked Mr. Bensen and Dr. Thornton to do so. These experts, in turn, reviewed the data and informed counsel for Legislative Defendants that it indeed matched the data presented at the liability phase and that nothing more need be done. That was the full extent of their involvement. Affidavit of Bensen6 6, 18?19; Af?davit of Thornton7 2?7. They did the same analysis for both the House and Senate, as Legislative Defendants? Sept. 23 ?ling disclosed (at 27), the only difference between the chambers was the timing of the adoption of the base map in comparison to the timing of the completion of the expert analysis. Tr. S. Redist. Comm. Sept. 10, 2019 at 49:19?? 6 The Af?davit of Mr. Clark Bensen is attached hereto as Exhibit 1. 7 The Affidavit of Dr. Janet Thornton is attached hereto as Exhibit 2. 24, Chairman Newton or the purposes of your work to choose the base map, let?s assume that the data is accurate and uncorrupted. Let?s go on with that work. During this process if we ?nd that the data is corrupt, then we?ll have to drop that?). None of Plaintiffs? objections to this involvement hold water. First, the Court?s decree does not prevent it. The Court?s order requires approval of agents only to the extent they will ?assist in the map?drawing process.? Court?s Decree ?l 9. Mr. Bensen and Dr. Thornton did not assist in the map-drawing process. They identi?ed no lines or con?gurations that might be drawn or map that might be picked or even a method of redistricting that might be used. All they did was review whether the data was the same data produced at the liability phase simply because it was not submitted to the Committee in the same format that it was submitted to the Court during the liability phase. This had no more impact on the map?drawing process than did the role of whoever at the State Lottery Commission ensured that the lottery equipment was compliant with best practices to effectuate a random draw or whoever at IBM ensured that the computers eventually used were functional. Most importantly, these experts reviewed no political data, gave no advice of any kind, and had no communication with any legislator. Aff. of Bensen ii 18; Aff. of Thornton at 7. They simply received instructions from Legislative Defendants? counsel on ensuring authenticity and relayed their view that the data was authentic. Second, Plaintiffs object to Mr. Bensen?s experience in giving political advice in redistricting. But this is irrelevant where, as here, Mr. Bensen did not draw on that experience for the simple task he performed. Aff. of Bensen 20?23. By the same token, various lawyers for Legislative Defendants have experience in, among other things, employment, tax, bankruptcy, and federal administrative-procedure law. But they did not practice that type of law here. Whatever 10 Mr. Bensen might be able to do is irrelevant; what he did here was simply check Dr. Chen?s production. Third, Plaintiffs observe that it took two days for this to occur, but that was simply the result of Dr. Chen?s having scrambled the order of maps in his simulation sets?apparently at Plaintiffs? counsels? request. Although Plaintiffs are correct (at 4) that there are ways to match the renumbered maps to the liability-phase maps?-and, indeed, they describe precisely what Mr. Bensen and Dr. Thornton did?they are wrong in implying that this can be done instantaneously. The process takes time. Moreover, Dr. Thornton and Mr. Bensen were not anticipating work on this matter at the time and were not immediately available to begin. Aff. of Bensen 7; Aff. of Thornton 2. Besides, the delay in the House Committee at this time was due in part to a highly contentious budget battle playing out as the redistricting process unfolded. Time was taken out of Committee hearings, not to await the experts? vetting, but to handle one of the most hotly contested legislative matters in over a decade.8 8 Finally, the use of these experts for this reason would not be an independent basis to strike down the plan. The Court lacks jurisdiction to police the General Assembly?s internal affairs, absent a violation of law. Again and again, coUrts have held that judicial efforts to control the internal processes of legislatures present non-justiciable questions that, if entertained, would violation of separation-of-powers doctrine. See, e. Abood v. League of Women Voters of Alaska, 743 P.2d 333, 336 (Alaska 1987); Mo?itt v. Willis, 459 So. 2d 1018, 1021 (Fla. 1984); Baines v. New Hampshire Sen. Pres. 876 A.2d 768, 776 (NH. 2005); State ex rel. La Follez?te v. Stilt, 114 Wis.2d 358, 338 684, 68687 (Wis. 1983); Bd. of Trustees of Jud. Form Ret. System v. Atty. Gen. of Corn, 132 770, 776?77 (Ky. 2003). The Court?s prerogative to assess whether the product of legislative process complies with the State Constitution does not entail the right to control the legislative process itself. The General Assembly is not a second-class constitutional citizen; it is co?equal with the State Supreme Court. Because the Constitution does not restrict the General Assembly?s ability to hire agents to assist it in the redistricting process, there is no basis for the Court to strike down legislation on this basis. The General Assembly has made every effort to comply with the Court?s decree for the purpose of proving beyond doubt that its end product is compliant with the new legal theory this Court has identi?ed. But, if the Court disagrees, the 11 3. The General Assembly Did Not Use Political Data Plaintiffs have attacked the integrity of Legislative Defendants and the redistricting process because of data transferred to Central Staff members and members of General Assembly by Legislative Defendants? counsel. Plaintiffs contend that Legislative Defendants intentionally ?outed the directive of the Court because the transmitted data included political information used by Dr. Chen. But after Plaintiffs objected to this transmission, Legislative Defendants investigated Plaintiffs? objections, and cancelled the link that included any such data. Plaintiffs have offered no evidence (and there is none) that any of the Central Staff or Republican members reviewed this information or that Republican members had the expertise to derive political informationfrom the data. Plaintiffs? argument is without any substantive foundation and is made only to distract the Court from the reality that their objections are meritless. The data in question came from Dr. Chen and was produced with his April 8th expert report. This material was forwarded by Legislative Defendants? counsel only after they received an email from Plaintiffs? counsel, Stanton Jones, on September 9, 2019. More speci?cally, in this email Mr. Jones stated ?plaintiffs will send the requested data tonight.? Mr. Jones? sent his email to Republican and Democratic members of the General Assembly, members of Central Staff, and several of the Legislative Defendants? attorneys. Mr. Jones? email was in response to a request by staff at the General Assembly, who had been instructed to contact counsel for both the Legislative Defendants and Plaintiffs to obtain the Chen data. The Committees needed the data in order to use Dr. Chen?s randomly drawn maps. It was, therefore, Legislative Defendants? understanding that the ?requested data? included all of the information provided by Dr. Chen with his opening report. General Assembly does not concede that there is any basis for the Court to strike down legislation duly enacted by the General Assembly. 12 Because Legislative Defendants understood that Plaintiffs? counsel would be sending all of the Chen data to everyone copied on Mr. Jones? email, Legislative Defendants transmitted an email stating their counsel would also send the Chen data to the persons listed on Mr. Jones? email. They also asked that Plaintiffs copy everyone on what was eventually transmitted by them. Legislative Defendants believed that sending the Chen data they received with Chen?s original report was necessary for the staff to be able to confirm that any data eventually forwarded by Mr. Jones was identical and not new data generated by Dr. Chen. Time was of the essence given the. timeframe ordered by the Court to complete redistricting. Because counsel for Legislative Defendants received no objections from Plaintiffs to their email stating they would forward the Chen data, Legislative Defendants? counsel then sent the Chen data. In order to be transparent, counsel for Legislative Defendants indicated that the data provided was the data received with Dr. Chen?s original report. Counsel for Legislative Defendants sent the entire set of Chen data because (1) they understood that Plaintiffs would be doing the same; (2) the Central Staff needed Chen?s original data to con?rm that the data sent by Plaintiffs and Legislative Defendants matched; and (3) they wanted to avoid accusations that Legislative Defendants cherry picked from the Chen data and only sent information favorable to Legislative Defendants. Plaintiffs objected to the transmission on September 9th at 4:45 PM. Upon receipt of this objection, counsel for Legislative Defendants investigated the claim. Legislative Defendants? investigation into the matter revealed that the first download was not attempted until nearly half an hour later. In order to avoid further controversy, Legislative D-efendants? counsel terminated the link within approximately an hour after receipt of Plaintiffs? objection. Legislative Defendants I con?rmed that the link was terminated in a communication to Plaintiffs? counsel later that evening l3 at 7:09 P.M. No member of Central Staff (the persons who drew the maps) completed the download. Because the Plaintiffs only attack the House, it is worth noting that only the offices for two House Members completed the download. One was a staff member in Democratic Rep. Pricey Harrison?s of?ce. At a minimum, Rep. Harrison and her staff certainly did not conspire with Republican members of the committee, and there is no evidence of a successful attempt on her or her of?ce to inject partisan changes into the maps. The other is Rep. Torbett?s Legislative Assistant, Mrs. Torbett, who began the download as part of her ordinary duties of saving committee documents for Rep. Torbett. However, it appears Mrs. Torbett never even opened the zip-file. Rep. Torbett never reviewed its contents. Neither Rep. nor Mrs. Torbett shared or discussed the contents of the data with any other House members, members of outside counsel, experts, or staff during the course of the two-week remedial period. It was certainly not used in the only district Rep. Torbett was directly involved in working on, the single unpairing in Gaston and Cleveland Counties, which passed unanimously. The complex files are entirely unusable for a non-expert in political data analysis. No other person connected to the House downloaded the data?no member, committee staff member, central staff member, or legislative office staff member. Legislative Defendants categorically deny that the virtually?unusable data in these files in?uenced the redistricting process at any point. There is simply no evidence that the persons responsible for drawing the maps, Central Staff, or any Republican member ever considered any political data or even knew how to locate or 14 use the political information buried somewhere within the thirteen plus gigabytes of Chen information included in the transmission made by Legislative Defendants? counsel. Much more concerning is the constant attempt by Plaintiffs? counsel and experts to advise Democratic members during the process of changing the remedial maps. They repeatedly conferred with Democratic legislators throughout the process, with open computer screens and map drawing experts. At one point, Chairman Lewis had to warn the committee Democrats of this constant strategizing with plaintiffs? counsel. B. Departures From the Base Map Are Compliant With Law Plaintiffs lodge two types of objections to the House?s departures from the base maps. They, ?rst, object to the procedure used. They, second, object to the resulting districts. Both sets of objections should be rejected. 1. The Court Should Reject Plaintiffs? PrOcedural Objections Plaintiffs? scattershot objections to the process of unpairing incumbents fails. a. Their objection (at 11?12) that the ?entire approach to incumbency protection. . .was unreasonable? asks too much of court intervention in the legislative process. It is true that the House could have chosen a different path, but it does not follow that different was legally required. As noted, the choice of Chen set 1 was selected because it earned the trust of Democratic members who proposed it. Tr. H. Redist. Comm. Sept. ll, 2019 at 215?3118. Any suggestion that the House was required to start from set 2 simply because the Senate did is self-evidently wrong. The House is allowed to strike a different compromise. Equally wrong is Plaintiffs? position (at 1) that the House was legally obligated to ?ask[] Dr. Chen to run a new version of his Simulation Set 2 that avoided pairing the current incumbents.? Even if that option was available, the House was not obligated to use it. One very good reason not to use it was that Dr. Chen?s liability-phase maps were the ones subject to adversarial proceedings 15 and court review. Using a set presented at that stage allowed the House (and the Senate) to utilize Chen maps without extensive expert review. As described, the experts simply needed to ensure that the maps were the same as those presented in court. Plaintiffs? objection to even that limited role only underscores the wisdom of not going back for further expert work. And the implication of Plaintiffs? theory is that the House was legally obligated to place blind trust in an adverse expert?to trust, inter alia, that he did not change his algorithm or otherwise try to achieve his own partisan ends. The Court?s judgment reaf?rms that the General Assembly has the right to redistrict, and it was not required to delegate that role to Dr. Chen. b. Plaintiffs? assertion (at 11?12) that the House process ?opened the door to partisan manipulation? ignores the many safeguards against precisely that. Every line was drawn in public. The lines were, by and large, implemented by non-partisan Central Staff members whose neutrality Plaintiffs do not attack. Bipartisan teams of legislators supervised the line drawing. With two exceptions, the Committee unanimously adopted the alterations, and Plaintiffs do not challenge one of these two groupings. Since the House complied with the Court?s safeguards, Plaintiffs? cannot be heard to propose the additional requirement of choosing one of two equally non?partisan Chen simulation sets. Equally important?and equally ignored in Plaintiffs? obj ections?wis the independent requirement the House Committee unanimously adopted that alterations be restricted to the minimum amount of changes to the speci?c districts affected by an incumbency pairing. Tr. H. Redist. Comm., Sept. 12, 2019, at 12:25-l6z4, 4327?12. So, for example, if two incumbents were paired in one district, a second adjacent district was left unrepresented, and a third adjacent district had a single incumbent, the Committee?s rules required that only the ?rst two districts, and not the third, be changed. Id; see also, Tr. H. Redist. Comm, Sept. 12, 2019, at 8211?17, 12:44. The 16 approach ensured that the incumbency?protection efforts remained ?reasonable.? Court?s Decree 11 The process did not, because it could not, turn into a freewheeling prerogative to redraw groupings. In fact, Dr. Chen?s ?set 3? does not adhere to this rule, claiming a license to redraw entire groupings, so the House gave itself far less discretion than Dr. Chen gave his set 3 algorithm. 0. Plaintiffs? assertion that the House ?improperly sought to preserve ?communities of interest? actually describes a consideration used by the Senate in at least one county grouping process Plainti?fs do not challenge. Whereas the House ?adopted a general approach of altering the base map only to unpair incumbents drawn into the same districts,? Sept. 23 Filing at 16, the Senate allowed itself more discretion to, for example, allow Sen. Blue ?to keep whole a community that Sen. Blue had represented for decades,? Id. at 22. As described above, the House?s ability to protect communities of interest was severely restricted because it could only change unrepresented districts and those with two or more members. Tellingly, Plaintiffs rely (at 12) on a Senate hearing transcript for their view that the House improperly protected communities of interest. A Senate hearing is a dubious source of information on the House, whose members were not all present and unable to object to representations they might dispute. Indeed, the transcript says more of Democratic Senators? concern for protecting communities of interest than anything else. In the exchange Plaintiffs? cite, Democratic Sen. McKissick raised the concern that the House might not have tried to protect communities of interest, Tr. S. Redist. Comm., Sept. 17, 2019 at 1726?10, and Rep. Hall simply responded that the ?members worked that out in what I believe to be a bipartisan and of course nonpartisan manner in coming to a consensus,? Id. at 17:25?18:6. Sen. McKissick then commended the House on 17 ?really examining closely communities of interest? (which was not quite what Rep. Hall stated), which he called an ?important[]? goal (which Rep. Hall also did not state). Id. at 18:14?19:5. In other words, the a Senator valued communities of interest, and Rep. Hall, in selling the plan to the Senate, was not-inclined to completely disclaim that goal. Plaintiffs? demand that the House be penalized for the Senate?s priorities is inexplicable. As their silence on the Senate concedes, neither chamber should be penalized: communities of interest, recognized on a non- partisan basis, are fair considerations and not prohibited by law or the Court?s order. d. Plaintiffs are simply wrong (at 13) that the House ?entirely ignored compactness in protecting incumbents.? First, the Committee began from Chen simulations that qualified as among the top ?ve scoring for their respective groupings, based on Dr. Chen?s compactness formula and scoring mechanism. The compactness scores of all districts are above the threshold set in the criteria. See Court?s Decree ?l see also, Cromartz'e v. Hunt, 133 F. Supp. 2d 407, 415?167 (E.D.N.C. 2000).9 And any drop in compactness score (as compared to the base map) is due, not to partisan intent, but to the severe restrictions the House imposed. By allowing only districts with zero or multiple incumbents?but not districts with one incumbent?to be redrawn, the House created an extremely limited number options for unpairing incumbents. It might have been possible to create more compact districts by altering more of the groupings (or by dividing more cities or VTDs), but any subordination of compactness was to the equally valid goal of ensuring that incumbency?protection efforts were ?reasonable.? Court?s Decree 1] 9 Compactness reports have already been previously submitted to the Court. However, the compactness report at issue here can also be found at ://webservices.ncle g.net/V ieWBillDocument/ 201 9/ 65 5 6/ 020,%202nd%20Edition 10 In any case, any difference between compactness scores discussed by Dr. Chen are less than slight. See Af?davit of Dr. Thomas Brunell W26, attached hereto as Exhibit 3. Moreover, Chen?s 18 Plaintiffs do not show otherwise. Their assertion (at 13) that compactness levels of the redrawn districts are of lower than in Chen sets 1 and 2 fails for reasons stated above: Chen sets 1 and 2 did not freeze into place the districts in the base-map con?gurations with one incumbent and change only the districts with zero or many incumbents. Thus, they do not follow the House?s approach to ensuring reasonable incumbency protection. Dr. Chen? 5 comparison counts as partisan the entirely non-partisan goal, adopted at Democratic request, of restricting alterations from the base map to districts where it was essential for the incumbency-protection purpose. Moreover, as the remedial report of Dr. Brunell shows, Plaintiffs? criticism, even taken at face value, is over remarkably minor differences. Aff. of Brunell fl 2-6. The districts are more compact than both the 2011 and the 2017 plans. Id. 2. The Court Should Reject Plaintiffs? Objections to Five House Groupings The defects in Plaintiffs? position become even more apparent when applied to the ?ve groupings they choose for their ?focus.? Pls? Br. 16. Four of the ?ve groupings were adopted by unanimous Committee votes, and those four were also adopted by a bipartisan, supermaj ority vote of the entire North Carolina House. Plaintiffs misstate the record in asserting that the entire plan passed ?on straight party-line votes.? Pls? Br. 7. To the contrary, the House utilized a procedure, originally suggested by Democratic Rep. Hawkins, whereby the Fender?Columbus grouping was separated from the others and voted on separately. That grouping passed on a party-line vote; the others passed on a bi-partisan supermaj ority vote of 68 to 42 with nearly twenty percent of present Democrats voting in favor.11 Tr. H. Floor, Sept. 13, 2019, Vol. II at 591:1?12. Thus, only one focus on slight differences in compactness scores completely ignores the marked improvement for all of these districts regarding divided and cities. Id. at 11119-10 11 Democrat, R. Smith did not vote. Five other Democrats had excused absences. See l9 grouping even caused political controversy, yet Plaintiffs demand Court intrusion into ?ve. That does not ?limit the scope of relief,? Pls? Br. 16, but unnecessarily expands it where the Court should decline any ?relief? at all. a. Brunswick-New Hanover Plaintiffs object to the Brunswick-New Hanover grouping, even though they concede it is 100% drawn by Dr. Chen?s computer. No changes were made to the base map. By de?nition, the Chen grouping complies with the criteria. The Committee unanimously adopted the grouping. That should end the matter. Yet Plaintiffs demand?in a curious twist for self?professed redistricting reform advocates?that the General Assembly protect the incumbency of someone who does not intend to seek reelection to the North Carolina House. Plaintiffs? objection to the HOuse?s choice not to change the district to unpair Representatives Holly Grange and Ted Davis is meritless. Rep. Grange announced her intent to run in the upcoming gubernatorial race several months ago?, and North Carolina law prohibits her from simultaneously seeking reelection to the House. Plaintiffs? assertion (at 39) that not protecting a member not seeking to run failed to protect incumbents ?even handedly? ignores that the public-policy purpose of incumbency protection is to protect the incumbent?s relation to a constituency, not to protect a random dot on a map. There was no point in seeking to protect Rep. Grange after she chose to pursue higher office; she was, in effect, no longer an incumbent. This exact point was made by Democrat Rep. Butler, who interrupted the readings of incumbency pairings in the September 12th House Committee Meeting to point out that Representative Grange had announced another candidacy, thus there was ?probably not a double? bunk.? Tr. H. Redist Comm. Sept. 12, 2019 at 1122?17. 12 In fact, Rep. Grange?s decision to enter the 2020 North Carolina governor?s race was Widely reported. See 3474ca6ab423499eba2a5aa 20 Nor was this choice unusual: the Senate did exactly the same thing. When Republican Senator Alexander announced his intent not to run again, the remaining exclusively Democratic senatorial incumbents in Wake County redrew the districts to better preserve communities of interest Without the burden of preserving his seat. Sept. 23 Filing at 22. Plaintiffs do not object to this choice in the Senate, presumably because objecting to the choice not to protect a retiring incumbent would be nonsensical. So is their objection here. Undeterred, Plaintiffs invent a preposterous scenario in which (they say) outside counsel and experts told a sitting legislator to retire to help Republicans?and she obeyed. This is so outlandish as to hardly merit a response, but Legislative Defendants have no choice but do so under these seemingly unreal circumstances. For starters, outside counsel deny this allegation absolutely and unequivocally. Legislative Defendants? outside counsel had no contact with Rep. Grange in connection with the redistricting. No one made any assessment of the partisan composition of her district and grouping, and gave no political advice regarding it. Next, Plaintiffs? argument is implausible. The idea that agents of the legislature could demand the retirement of a member is so beyond the realm of ordinary experience that Plaintiffs should be required to present real evidence before lodging such an allegation.13 Rep. Grange?s candidacy for Governor had been announced for months. Suf?ce it to say that elected representatives, almost by de?nition, have attained a certain level of self-con?dence so as not to be ordered around by persons that the legislature is free to ?re at will. 13 Plaintiffs? Willingness to make these accusations based on bald speculation does not help the standards of civility and decency in practice at the bar before this Court. 21 For another thing, Legislative Defendants? lawyers have no ability to know that a randomly drawn Chen map the base map) is more politically advantageous to Republicans than a hypothetical map redrawn to protect Rep. Grange?s non-existent incumbency interest. Chen set 3 did not exist at the time of the redistricting, Dr. Chen is Plaintif?" expert, Legislative Defendants have no simulation expert, and Plaintiffs? counsel seem to have a highly exaggerated view of Legislative Defendants? lawyers? ability to interpret partisan data and predict election results. Moreover, as noted above, no political consultants or experts were engaged at the remedial phase to do anything of the sort. Far more plausible than Plaintiffs? black?helicopter theory is the truth: Rep. Grange can be seen on video entering the committee room only after Chairman Lewis announced that only incumbents who intended to run again would be unpaired. Tr. H. Redist Comm. Sept. 12, 2019 at 11:13-22. Democratic members discussed whether Reps. Grange and Davis?s pairing even constituted a double?bunking given that Rep. Grange had announced her candidacy for another of?ce. See Tr. H. Redist Comm. Sept. 12, 2019 at 11:2?17. Rep. Grange then enters the committee room in time for the discussion about this county grouping to begin. Id at 37:18?25. Having not heard the incumbency?protection explanation of Rep. Lewis, Rep. Butler, and that it should be used only for members who were running again?and being under the mistaken impression that the committee was utilizing the Chen set 2 that incorporated incumbency in her county grouping frOm 2011, Rep. Grange asserts that her incumbency should be considered but does not commit to running again for her House seat?in fact, she states that she is an announced candidate for higher of?ce. Id. at 37:2?23. Reps. Grange and Davis begin disputing how to redraw a county grouping that, under the committee?s criteria, could not have been redrawn at all. Id. at 37:18?25. 22 Members of the House pulled her aside to discuss whether she could commit to running again for the House given the Committee?s criteria. The House, after all, expressed a policy of narrowly tailoring departures from the base map. So, without a good reason to manually adjust - districts, it was fair for Rep. Grange?s colleagues to ask her about her intentions to seek reelection. (One would think Plaintiffs would approve, given their distaste for even small manual line adjustments.) Once the criteria was explained to her, Rep. Grange withdrew her request to be unpaired. Tr. H. Floor., Sept. 13, 2019, Vol. II at 560115?561 She is currently running a very active campaign in pursuit of the Republican gubernatorial nomination and has not asserted any interest in returning to her House seat. The idea that this was coordinated with Rep. Davis is also patently absurd. Legislative Defendants? outside counsel and experts had no contact with Rep. Davis at all during the process. Again, Legislative Defendants and their agents ?atly deny discussing or even knowing the partisan implications of the Chen base map with Rep. Davis. Nor was this an attempt to preserve the ?core? of his old seat. It could not possibly have been so because Dr. Chen?s algorithm did not preserve cores; by de?nition, the base map does not do so. Predictably, a mere visual inspection reveals that the districts in this grouping are nothing alike from their old version. Rep. Davis?s seat sheds all of the southern portions of Wilmington and the southern beach communities of New Hanover County for an entirely new, compact seat based entirely in the northeastern corner of the county. One signi?cant purpose of using Dr. Chen?s maps was to achieve this goal of avoiding core retention. Likewise, Plaintiffs? demand that discussions about Rep. Grange?s career choices occur in publicm?even though this was not directly part of the remedial process, fairly understood?is 23 unrealistic. There comes a point at which judicial involvement in the political thicket should be too much by anyone ?3 standard, and this is clearly that point. Rep. Grange?s mere reference to the Covington litigation?where a federal district court held that, ?as a factual matter, the General Assembly did not need to draw the district to protect Representative Bell? because he announced his retirement, Covingion v. North Carolina, 283 F. Supp. 3d 410, 441 n.7 (M.D.N.C. 2018), is also not nefarious. Rep. Grange, in addition to being a West Point graduate, veteran, and successful businesswoman, has a law degree and has practiced law, although she does not practice in the state of North Carolina. The fact that Rep. Grange was able to cite this precedent and apply it to her particular incident merely shows that, as an attorney herself, she understands the importance of complying with applicable law. Even setting aside that Chen?s set 3 maps for this group are based on the false premise that Rep. Grange is running for re-election, Dr. Chen?s purported showing that a grouping 100% drawn by Dr. Chen?s computer is an outlier only proves how ?awed Dr. Chen?s analysis is. See Pls? Br. at 37?3 8. The grouping cannot be an outlier because it is one of the maps Dr. Chen previously simulated and, by consequence, de?ned the very concept of what constituted an outlier at the liability phase.14 That Dr. Chen has now purported to ?nd his own map an outlier proves only that, at will, Dr. Chen can make an outlier of anything. To the extent the Court adopts his findings with no adversarial process (and, indeed, no ability of Legislative Defendants to run simulations, since only a few people in the world have that skill) it makes itself beholden to whatever result Dr. Chen desires. Here, that means a Chen map, drawn with no partisan intent, is an outlier as compared to another Chen map, drawn with no partisan intent. 14 This is discussed at Legislative Defendants? post-trial brief (at 31?35). Needless to say, the very flaws that brief discusses have been revealed in the remedial process, and this is the latest example. 24 That aside, the Court can see the problem with Chen set 3 here as elsewhere. His simulations (at Br. 38) redraw every district in the grouping, giving Chen leeway to redraw HD17, HD18, and HD20. But if Rep. Grange was planning on running and the House had chosen to redraw the grouping it would only have redrawn HD20 (where Reps. Grange and Davis were paired) and HD19, the empty district. HD17 and HD18 were off limits because they were occupied by Reps. Her and Butler, respectively. Dr. Chen cannot prove anything about partisan motive because he assumes a far broader range of leeway than the House committee gave itself-w? not to mention that all of set 3 maps purposefully unpair an announced gubernatorial candidate. Finally, Dr. Chen reports that he drew 1000 simulated maps for Brunswick- New Hanover in Set 3. In truth, there are only 36 unique maps in Chen?s set 3 for this county group. (Aff. of Dr. Janet Thornton 119) Moreover, Dr. Chen performs a statewide comparison of predicted House seats under HB1020 version of his simulated plans, he does not conduct such a comparison of the challenged groups to the corresponding groups from his simulations. Perhaps. this is because 99% of Chen?s simulations for this group result in three Republican districts and one Democratic districts, the exact number he predicts for the enacted version of this group. (Thornton Aff. Fig In short, a grouping drawn by Dr. Chen is not an outlier, complies with the criteria, and was a valid option for the House to select. 15 Dr. Chen?s statewide analyses have little value because the General Assembly Was not directed to redraw for the entire state, but instead, only the 14 county groups identi?ed by the Court. In any case, Dr. Chen?s prediction of 44 Democratic seats under HB 1020 con?ict with the 49 seats predicted under the same plan by noted Plaintiffs? expert Nicholas Stephanopoulos and Defendants? expert Dr. Brunell. (Aff. of Brunell. 1T7) 25 b. Guilford Plaintiffs? objections to the Guilford grouping hardly fare better. Whereas they complain that the House changed the base map for Cleveland-Gaston and Columbus?Fender too much, here they complain that the House changed the base map too minimally. Incredibly, Plaintiffs complain that a bipartisan group of legislators supervised the shifting of a single VTD in Guilford County - that (Plaintiffs also concede) contained a paired member. Pls? Br. at 41-42. See also Tr. H. Redist Comm. Sept. 12, 2019 at 99:19-100216. But What could be more tailored than shifting one VTD containing an incumbent from one district to the next? Erika Churchill of non-partisan Central Staff con?rmed on the record that this was the most expedient way to correct the grouping. (Tr. H. Redist. Comm. Sept. 12, 2019 at 100:12-16. Besides, Plaintiffs concede that the district is not an outlier, even by their ?awed approach. So whatever concerns Plaintiffs have about miniscule compactness differences are irrelevant; this is a partisan?gerrymandering case, not an abstract compactness case. Thus, the compactness scores are not significant independent of partisan motive and impact. The Committee adopted the grouping unanimously. Tr. H. Redist Comm. Sept. 12, 2019 at 99:16n10121 No Republican or Democratic member discussed the grouping on the ?oor or in committee in a negative light after one of the fastest and easiest county grouping exercises the committee undertook. See Tr. H. Redist. Cemm. Sept 12, 2019 at 100:17-21. That should resolve the matter. In truth, Plaintiffs? objection is not to that one shift but to Dr. Chen?s base map, which they complain is very similar to the enacted version of HD58.16 Pls? Br. at 41. This is hardly surprising, 16 Although the 86% similarity score counts the one VTD shifted, that single VTD does not appear to move the needle in the overall similarity of the districts; the complaint would seem to be the same with or without the shift. 26 however, since, HD58 is surrounded by districts on three sides that were redrawn in the Covington case. Further, analysis by Dr. Thornton reveals that due to the frozen districts, there were only 19 unique maps for Guilford County to pick from. See Aff. of Thornton at 11917. Nor can Plaintiffs credibly call the base map partisan when it is overwhelmingly the work of Dr. Chen. The problem here is, as Legislative Defendants said in their post-trial brief (at 33), that Plaintiffs? experts were measuring ?minor? differences as ?stark.? Here, it is now apparent that this was correct: because three surrounding districts in Guilford were frozen by the court?s order and Chen?s own simulations, a Chen map had enormous overlap with HDS 8, even though Dr. Chen concluded that HD58 was an ?outlier??-indeed that zero of his simulated maps were as Republican friendly. See at 106 (identifying HD58 as to the right of every simulated version of the district). Now, at the remedial phase, Plaintiffs complain that the Chen map is too much like the enacted map, but they cannot seriously contend that it is not lawful under the Court?s order, which treated the Chen maps as a valid, non-partisan base line. In any event, the Plaintiffs? own maps show that that HD59, the allegedly outlying Republican?friendly seat in the 2017 plan, had substantial changes made to it in the 2019 remedial plan. No longer does the district stretch to take in rural, Republican-friendly precincts in Southern Guilford County that split High Point and Jamestown. Instead, more rural areas are added to the two other seats surrounding it, those municipal splits are eliminated, and HD 59 picks up deep-blue territory on the edges of Greensboro. None of this was intentional?it was simply the work of Dr. Chen?s algorithm. 17 This reality, not disclosed by Dr. Chen, shows that there are few options for revising the remaining three districts in Guilford County when the other three districts are frozen. Chen?s core analysis testimony could only be potentially useful if he had provided the same analysis for his other unique maps for this group. 27 c. Cleveland-Gaston Yet another grouping Plaintiffs challenge even though it received unanimous Committee support is the Cleveland-Gaston con?guration. (Tr. H. Redist Comm. Sept. 12, 2019 at 5627?22). As noted, their reliance on Chen set 3 is misplaced because it redraws all four districts where the Committee viewed changes to 10 (Chen and HD108 as the only allowable maneuvers.18 This presented the committee with dif?cult choicesm?how to manage splits of Gastonia to a reasonable level while not splitting the many smaller municipalities that are in northern and western Gaston County. Ultimately, the effort was successful, as High Shoals, Bessemer City, Stanley, Dallas, Ranlo, and other smaller cities were not split at all. In fact, not one additional municipality was split during the unpairing process. The only other feasible option would have been to place Rep. Hastings in the empty district by splitting Cherryville and at least one VTD, leaving Rep. Torbett representing an elongated, barely contiguous seat that split additional municipalities. 18 The districts were renumbered between the base map and enacted plan. The districts changed under the base map numbering were 11 and HD108. 28 Chen House Base Map Districts for Cleveland and Gaston Counties la. a? Mk.? 'h ?r . .sma I 0:14 a w? Hum-'4? - 11% ulcny 5? 5km .539 - - 3 Chanel?: - Democrat [j ]Dislrict I: or [Town :?lji Sources: NC General Assembly. NC State '5 31:: i . Republican l:l Precinct County 0 5 10 Printed by the North Carullna Genoa! Assembly. 2019 MIIES S.L. 2019-220 Districts for Cleveland and Gaston Counties 6% Cleveland r" 1:75;. . I. I: Sources: NC GeneralAssambly. NC State Democrat l:l CHWTOWH '5 {Ali} Board ofElec?ons. US Census Bureau . Republican Precinct County 0 5 ?10 Pr'nlad by um North Calohna General Asseirbly. October 2. 2019 ?"35 Thus, while they criticize the con?guration, they have provided no alternative that complies with the Committee?s rule on incumbency protection. There appears to be no way to bring one incumbent into HD108 without either creating a non-contiguous district or making 29 alterations to a district with a single incumbent. See Pls? Br. 30. The effort gained the support of the Democratic member of the committee present for the redrawing, and Plaintiffs can point to no evidence of partisan intent. Moreover, even assuming the validity of set 3, it is hard to see what incentive there would be to gerrymander these districts, when Dr. Chen?s chart shows that all are heavily Republican in all events. Pls? Br. 32.19 And the most Democratic of the four-min fact a district more Democratic than 93.3% of Dr. Chen?s simulation sets, 1, belongs to the Republican Speaker of the North Carolina House of Representatives, Tim Moore. Pls? Br. 32. That district and HD109 were entirely unchanged from the Chen base map. Changes were executed by Central Staff Member Erika Churchill, a respected member of the non?partisan legislative staff and supervised by Democratic member Representative Hawkins. Tellingly, the discussion of this grouping was short and non-controversial, and no member raised the possibility that the district had been gerrymandered. See Tr. H. Redist. Comm. Sept. 12, 2019 at 54: 1 5 -5 6:22. Plaintiffs? counsel can be seen on the livestream conferring with Democratic members of the committee during the redraw process for this grouping, yet no objection was made to this grouping at any point until the briefing ?led last week. Id. The unanimous adoption of this grouping speaks louder than Plaintiffs? guesswork on partisan motive. d. County Plaintiffs? objections to the configuration fare no better. Like the groupings discussed above, the House Committee unanimously adopted the configuration Plaintiffs call a 19 Instead of 1,000 maps for this group, Chen?s set 3 only produced 80 unique maps. Aff. of Thornton 1] 9. Moreover, all of Chen?s simulations predict four Republican districts, the same number predicted for HB 1020. Aff. of Thornton. Figure 3. 30 partisan gerrymander. Tr. H. Redist. Comm. Sept. 12, 2019 at 69:13?7022. What?s more, a bipartisan team, including the deputy minority leader, was involved in the unpairing. See Tr. H. Redist. Comm., Sept. 12, 2019 at 66:3?13. First, Central Staff worked under their supervision, to devise as solution that involved moving a single VTD to unpair Reps. Montgomery and Conrad. 10?. Plaintiffs do not concentrate their objections to this unpairing, nor of course on the entirely unchanged HD73, yet ask for the entire grouping to be redrawn. For the pairing involving Reps. Lambeth and Terry, the bipartisan delegation (and Deputy Leader Reives) faced a choice. Both Terry and Lambeth were located two VTDs away from the empty district in eastern County. All four VTDs were heavily populated, and there was no correct or incorrect methodology for selecting who would be send to the empty seat. The fact that Rep. Lambeth expressed an interest in representing Kemersville because he had in the past represented part of the town, does not betray any partisan intent or the desire to preserve the core of his old district, it was simply the easiest way to decide who would represent the empty seat.20 The solution was amenable to all parties. This unpairing caused a population imbalance, and three VTDs were selected to correct this under-population. Plaintiffs attack the selection of the three VTDs in central but can produce no evidence that it was Rep. Lambeth, rather than the delegation as a whole with or without Rep. Reives, Rep. Terry, or Central Staff themselves, who selected those three VTDs. Despite the odd shapes of the municipal lines, the approved unpairing of these two members kept 99.98% of Kemersville whole, roughly two-thirds of Walkertown whole, moved only 5 VTDS, split none, and attracted the unanimous support of the committee. 20 Rep. Lambeth?s new seat in the remedial plan bears very little similarity to his district under the 2017 election districts. Instead of wrapping around the southern edge of the county, headed towards Clemmons, the district stays entirely on the eastern side of the County. 31 S.L. 2019-220 for and Yadkln Counties 6? Hi" Wal utC vo Ellun . .J 3 'm 9 an E?Oak Yadk 1 "?55 nn- - . L?in nu rut dlu Ile nu ?th5m. Kern R- 1" In tourism? .1 rm Mammary "m "n 1 2r 1 TONY a1." Fair Inlh Glam 9. m: rm 19? Lam .9 Bermuda 71 llbul ?dill: rm mu Mldway ?lit a a. .. Sources: NC General Assembly, NC State . Democrat District I: City/Town wig-E 1in Board of Elecllons. US Census Bureau . Republican I: Precinct I: County 0 5 10 Primed by the North Carolina General Assembly. October 2. 2019 Miles Chen House Base Map for and Yadkln Counties ?l 9-- - I: a Hill Elm Ha?z .. .. .., .. L, Forsyt? 7.4 i 5 ma E. 0.1er ?Yadkin "3 . an: my I Vadluudu ll uzr 4 all I 12 Zachary - npdtill m- I Terry run Lamholn A BMmudaR Ibulg 13:; Midway Hang a u. Sources: NC General Assembly. NC Stale . Democrat District '3 City/Town ?hf? Sling," Board ofEleclions. US Census Bureau . Republican I: Precinct County . 0 5 7 10 Primed by lha Carolina General Assembly. October 2, 2019 was At the end of the process, nonpartisan Central Staff con?rmed that, to the best of their abilities, only minimal changes were made in order to accommodate incumbents. See Tr. H. Redist. 32 Comm., Sept. 12, 2019 at Republican members could achieve an ingenious partisan gerrymander under those conditions is implausible. And, once again, Dr. Chen?s set 3 does not provide an appropriate point of comparison; the House Committee viewed only HD75 and as amenable to alterations to unpair these two members, but Dr. Chen?s algorithm redraws HD72, HD73, and HD74 as well. The Committee?s non-partisan restrictions are being counted in his set as partisan considerations. Plaintiffs? label ?obvious gerrymander? does not follow from their own maps, which show that most of the lines dividing Winston?Salem existed in the base map, as does the core structure of the grouping. Pls? Br. 24. Plaintiffs neither present nor simulate an alternative con?guration that would ?t the House?s restrictive criteria, which disallowed changes to HD73. Plaintiffs do nothing to show that the partisan impact they purport to divine, if it even exists or is meaningful, was anything but the incidental consequence of the criteria.21 e. Columbus-Pender-Robeson This group was the only group which was approved by the House based upon a party line vote. Even here, in committee, the ?nal version was supported by Rep. Brockman, a Democrat. (Tr. H. Redist. Comm. Sept. 13, 2019, at The dispute concerned the preferences of the Democratic members for plans that would divide one or more cities, or one or more VTDs, than the enacted plan. The legislative members supported the enacted version of this group because it divided fewer cities or VTDs than the versions proposed by the Democrats (Id. at 18:20-19:13; Tr. H. Floor, Sept. 13, 2019, Vol. II at 54321-646215). Members of both parties spent considerable 21 Chen?s set 3 includes 572 unique maps for this group. Aff. of Thornton 1i 9. However, 71 percent of Chen?s Simulations predict 2 Democratic seats and 3 republican seats. Id. at 10-1 1; Figure 2. Only 24.8% of Chen?s simulations predict 3 Democratic seats making them ?political outliers? under Chen?s de?nition of the term. Id. 33 time looking for compromise, but were unable to achieve it. Ultimately, Republican members of the Committee, along with Rep. Brockman, selected a map that did not split any precincts or municipalities. The House and Senate member for this grouping felt strongly that creating new splits of Columbus County municipalities or VTDs where none existed in the base map would be a poor policy choice during the process of unpairing Reps. Smith and Jones.22This was a reasonable policy choice, not one representing an intent to ?maximize? republican advantage. In any case, Chen?s simulations only produce 276 unique maps, not 1,000. Thornton Aff 1i 9. Moreover, 99.5% of Chen?s simulations predict 2 democratic seats, just like HB 1020 (Thornton Aff. {?ll 0-1 1; Fig, 1). But, Chen?s simulations also produce 5 maps with only 1 Democratic seat. Id. The fact that Chen?s simulations provide for 5 lawful maps with two Republican districts, and that the Legislature did not select these 5 maps, or amend the randomly drawn version, to create a second republican seat, proves that there was no intent to place partisanship over criteria. In all events, district 47 is entirely unaltered from the Chen base map, and is not featured at all in the Plaintiffs? complaints, which center around how best to trade precincts within Columbus county. The Plaintiffs? lack of consistency is on full display in their challenge to this grouping. While in other counties they criticize minor, noncontroversial municipal or VTD splits, in Columbus county, complete compliance with those traditional nonpartisan principles of redistricting is blasted as a partisan plot. The minor deviations to the base map in Columbus County completely conform to the court?s criteria and should be upheld, while Robeson County?s district 47 is an unaltered Chen district which of course should be upheld as well. 22 It is worth noting that one of the precincts in Tabor City, North Carolina selected for splitting by Rep. Jackson had faced severe election administration issues in 2018, and would have been a bad choice for splitting due to potential voter problems. com/201 8/1 precinct/. 34 C. The Court Should Return Any Grouping Alterations It Enjoins to the Base Map As Legislative Defendants? September 23 submission explained, any districts the Court rejects as improper amendments to the base map should be remedied by returning the relevant districts to their base-map con?gurations. At a minimum, the non?partisan process of picking a random map from the ?ve top-scoring maps that Plaintijfv presented in their case in chief is an unimpeachable solution to the very dif?cult remedial problems involved in this case. And it represents a legitimate policy choice that this Court is bound to respect. As re?ected in the incident surrounding the Brunswick?New Hanover grouping, the House Committee viewed adherence to the base map as the default rule and required a justi?cation for departing from it. As a result, any ?nding that the House erred in manually adjusting lines would mean that no justi?cation was warranted and that a return the base map best honors legitimate legislative policy.23 1. Plaintiffs? Demand for a ?Blank Slate? Is Untenable Given Plaintiffs? objection (at 11) to ?incumbents from each grouping. . .amend[ing] their own districts?m?and their presumed con?dence in the work of their own expert?one would have thought they would agree. Instead, they vehemently protest that approach and insist that ?the Court should direct the Referee to draw from a blank slate all ?ve of the House groupings described above.? Pls? Br. 45. That is a remarkable position. The Court is obliged to tailor its remedy to any violations it ?nds. As described above, the ?ve groupings were overwhelmingly the product of Dr. Chen? non- partisan simulation exercise. Why would a new expert need to draw from a ?blank Slate? to remedy the supposed errors re?ected in a few small redistricting moves? At the same time, Plaintiffs 23 To be clear, Legislative Defendants do not believe the Court should ?nd this or that the policy standard is legally enforceable. The point here is that, in the event the Court disagrees, this is the best way to honor state policy in tailoring its remedy, which it is obligated to do. 35 apparently propose that the other House groupings can be left alone (along with all Senate groupings) meaning that Plaintiffs? approach would, against their own advice, ?result in different criteria being applied in different groupings.? Pls? Br. 45. Given that the base map represents a legislative prerogative, it would make no sense to draw from a blank slate when the contemplated error resulted from departures ?om the base map.24 Plaintiffs? arguments for this approach are unpersuasive. First, their position that returning these groupings to the base map would apply different criteria to different groupings is simply wrong. As Legislative Defendants? September 23 filing explained, many House and Senate groupings saw no changes from the base map. Thus, returning groupings that were changed to the base map simply applies the criteria applied in all other groupings. Again, it bears emphasizing that having a Referee take the pen would result in this very inconsistency.25 Thus, it seems that Legislative Defendants? proposal is the only way to apply the same standards across the board evenly. Second, Plaintiffs? argument (at 45) that ?the base maps themselves are infected? is ?at wrong for reasons stated above. But even if it were right, it proves too much and too little. On the one hand, if the Court is not inclined to jettison the entire base?map approach in both chambers (and, of course, that would be remarkable and entirely unjusti?ed), then it will have already disagreed that the so-called ?procedural violations? merit that drastic act. There would, then, be no more reason to apply it in these groupings than anywhere else. On the other hand, if the Court 24 For example, Plaintiffs contend (at 17) that ?[t]he base map that Legislative Defendants selected from Dr. Chen?s simulations cured this cracking? in the Columbus?Pender?Robeson grouping. It is unclear why they would object to using a map that does precisely what they say it should do. 25 Plaintiffs? assertion that the Referee can simply follow the Court?s Decree 1i 5 is undermined by the fact that there are possibly an in?nite number of ways to do so. The House adopted a particular approach, first, by reference to the base map and, second, by reference to incumbent residences. 36 credited Plaintiffs? position and found these supposed violations (somehow) signi?cant enough to jettison all lines in these ?ve groupings, then it would logically be hard to defend maintaining the base map in other groupings. Either way, Plaintiffs? objections to the base map for some purposes and not others (and not in the Senate) are puzzling. Third, Plaintiffs? position (at 45?46) that ?adopting the base map would not remedy the violation in Brunswick-New Hanover? simply illustrates how strange their objection to that grouping is. What criterion should a Referee follow to protect an incumbent who is no longer running? And why would it be necessary to recon?gure the entire grouping for that purpose? Even if there were some partisan motive in retaining the base map (there was not), that motive could not justify striking down and otherwise objectively non-partisan grouping. 2. Plaintiffs? Objections Smack 0f Partisan Manipulation At base, Plaintiffs stake out a contradictory and incoherent position that is dif?cult to defend as anything but partisanship in action. The Plaintiffs concede that the Senate process followed the Court?s order and challenge none of the adopted Senate groupings. Yet the House process was remarkably similar in all material aspects?and even there the Plaintiffs only challenge ?ve of the fourteen redrawn groupings. Even within those groupings, they only attack the limited changes made to HD58 and HD59 in Guilford County, HD108 and 10 in Gaston, apparently only and HD75 in and HD16 and HD46 in Columbus County. In New Hanover, only HD19 and HD2O could have been changed if Rep. Grange was not an announced candidate for Governor. Then, as to those groupings, they ask for a total blank slate, not a return to a non-partisan baseline or even tailored adjustment. Many of the seats they asked to be redrawn were entirely unchanged from the base map. From there, Plaintiffs lodge a host of inconsistencies: a grouping with no change, they say (at 35?40), should have changed signi?cantly; a grouping with signi?cant change, they say (at 17?21) should have seen only two VTD shifts; and a grouping 37 with one VTD shift, they say (at 41?45) should have been totally redrawn. It is hard to discern any coherence in their position. From all this, it begins to seem like the objections have very little to do with facts or law but with the perception that unchallenged groupings in both plans were a win for the Democratic Party, and that these ?ve groupings might be drawn in a more Democratic-friendly manner by the Referee. Legislative Defendants have no way to know what partisan data Plaintiffs (who include the Democratic Party of North Carolina) have or why they are so concerned about these groupings?to the point of wanting a blank slate rather than their own expert plan?weir even whether their view about partisan impact is factually accurate. But it bears recalling that redistricting litigation is no less political than redistricting itself, and it would be naive to think the Democratic Party of North Carolina went to Court with no concern of electoral advantage as a result. Indeed, Dr. Chen?s latest report reveals that he has access to and has been using partisan data, so it is entirely possible that Plaintiffs are making remedial objections based on the very considerations the Court sought to bar from the process. It would be remarkably ironic and counterproductive if the Court adopted objections predicated on partisan motive and effect to a process that was publicly shown to be both non?partisan in motive and bi?partisan in support and cooperation. And this is precisely the concern Legislative Defendants raised, that a Court-drawn map cannot help but be political since the litigants before the Court?Plaintiffs included?are political and have political goals in mind. Courts must be careful to avoid buying into those goals or aiding 38 them intentionally or unintentionally. This is the core of Legislative Defendants? concern about where Plaintiffs want the Court to take this process.26 3. Any Redrawing Effort Should Comply With All Restrictions Imposed 0n the General Assembly If the Court concludes that the Referee should undertake to redraw districts (it should not), then simple fairness requires parity. And that means the Court should comply with ?the speci?c criteria to govern the drawing of remedial districts.? Pls? Br. 46. This means, at a minimum: 0 The Court should adopt a remedy in two weeks, Court?s Decree 4; It should comply with the criteria it imposed on the General Assembly, including the prohibition on the use of any partisan or political data, id. 1] 5, which means the Referee shall give zero consideration to politics and not have access to partisan data; a The Court and its agents should conduct the entire remedial process in full public view, id. 8, which means, at a minimum, that all the Referee?s map drawing must occur in public, with any relevant computer screen visible to the public, and neither the Court nor its agents shall undertake any steps to draw or revise the new districts outside of public view, id. 0 Likewise, the Members of the Court should deliberate over any proposed plans and objections in full public view, see Pls? Br. 7 (complaining of any relevant discussions occurring ?outside of public earshot?). 26 Plaintiffs? reference to these concerns (at 46) as ?a thinly-veiled threat? is baseless. Parties are entitled to raise concerns about the partisan motive of their litigation opponents and the possibility of it impacting judicial relief. 39 It would, in short, be entirely counter?productive for the Court to jettison a public process? perhaps the mo st transparent ever conducted?and replace it with a black-box judicial process. If the Court is unable or unwilling to comply with its own order, it should reconsider Plaintiffs? demand for a secret redistricting behind closed doors. Should the Court do anything other than summarily reject plaintiffs objections, Legislative Defendants request that the court hold a hearing. Prior to any hearing, Legislative Defendants request an opportunity to obtain all of Dr. Chen?s code and backup data for his most recent report and take his deposition. Legislative Defendants further request an opportunity to issue such subpoenas and take such depositions that may be necessary for Legislative Defendants to prepare for the hearing. II. The Objections of the Amici Present an Already-Disclaimed Demand for Proportional Representation The Court can safely ignore the objections of the amicz? participants, since they sound entirely in a theory of proportional representation that Plaintiffs long ago disclaimed. It is telling that the objections of the amici and those of Plaintiffs bear no resemblance in substance or approach. The amici want to litigate entirely new issues and inject unvetted expert analysis into this case at this late hour. The Princeton Gerrymandering Project complains (at 6) that Dr. Chen?s algorithm prioritized compactness and maintain whole counties, which the Project insists has an impact of partisan bias. Be that as it may, maintaining Whole counties is required by the North Carolina Constitution, and compactness was required by the Court?s order. That these may have a partisan impact does not create a cognizable claim. 40 In striking down the 2017 plans, the Court emphasized that ?Plaintiffs do not seek proportional representation? and ?that nonpartisan plans that do not intentionally discriminate against Democratic voters may well not provide for proportional representation.? Judgment Conclusions of Law 11 675. Speci?cally, the Court found that, ?[u]nder Dr. Chen?s and Dr. Mattingly?s simulations, there are scenarios where Democrats would win 50% of the statewide vote but less than 50% of the seats in either chamber.? Id. All the Project has shown is that this is correct. But the Court is in no position to swap legal theories at the remedy phase and to seek to remedy a violation that Plaintiffs disclaimed as even being viable.27 Democracy North Carolina makes a similar error in contending (at 6) that Dr. Chen?s simulated maps are insuf?cient to ?ensure that the discriminatory effect is fully addressed.? That cannot possibly be true when Dr. Chen?s maps did not begin with the 2017 plans, but rather began at random points within county groupings. No ?effect? was carried forward because no lines at all were carried forward. The Princeton Gerrymandering Project and Democracy North Carolina therefore both err in creating some measure of ?partisan bias,? applying it to the new enacted plans, and ?nding them insuf?ciently different from the supposed bias in the 2017 plans. The premise of this argument is that political parties are entitled to something beyond a principle of non-discrimination; they should have, the argument goes, an af?rmative right to ?fair electoral districts.? Democracy NC Br. 10. And this too is a way of asking for proportional representation, an af?rmative duty on the 27 Needless to say, the Project?s objection to the General Assembly?s failure to consider communities of interest is ironic as the General Assembly is fending off the (baseless) allegation that it violated the law by doing precisely that. If nothing else, the amici have shown how quickly partisan-gerrymandering litigation turns the legislature into a pi?ata, being attacked from every conceivable angle with contradictory arguments. 41 General Assembly to (somehow) match votes to seats. That has been disclaimed as a valid theory by Plaintiffs and the Court and should be rejected again here. CONCLUSION The Court should not enjoin the House or Senate plans. 42 This 4th day of October, 2019 BAKER HOSTETLER, LLP E. Mark Braden* (DC Bar #419915) Katherine L. McKnight* (DC Bar #994456) Trevor M. Stanley* (VA Bar 77351) Richard B. Raile* (VA Bar 84340) mbraden@bakerlaw.com tstanley@bakerlaw.com rraile@bakerlaw.com Washington Square, Suite 1100 1050 Connecticut Avenue, NW. Washington, DC 20036-5403 Telephone: (202) 861 -1500 Facsimile: (202) 861-1783 Patrick T. Lewis* (OH Bar 0078314) 127 Public Square, Suite 2000 plewise@bakerlaw.com Cleveland, OH 441 14-1214 Telephone: (216) 62 1 -02 00 Facsimile: (216) 696-0740 Robert J. Tucker* (OH Bar 0082205) rtucker@bakerlaw.com 200 Civic Center Drive, Suite 1200 Columbus, OH 43215-413 8 Telephone: (614) 228-1541 Facsimile (614) 462?26 1 6 Admitted pro hac vice Respectfully submitted, OGLETREE, DEAKINS, NASH, SMOAK STEWART, PC. I) 4 i I By: 711/ 3; (f 4., t, w/ Phillip J. Strach NC. State Bar No. 29456 Thomas A. Farr NC. State Bar No. 10871 Michael McKnight NC. State Bar No. 36932 torn.farr@ogletreedeakins.com 4208 Six Forks Road, Suite 1100 Raleigh, North Carolina 27609 Telephone: (919) 787-9700 Facsimile: (919) 783-9412 Counsel for the Legislative Defendants 43 CERTIFICATE OF SERVICE This is to certify that the undersigned has this day served the foregoing in the above titled action upon all other parties to this cause by email transmittal to the following: Edwin M. Speas, Jr. Caroline P. Mackie PO. Box 1 801 Raleigh, NC 27602-1801 (919) 783-6400 espeas@poynerspruill.com Counsel for Common Cause, the North Carolina Democratic Party, and the Individual Plainti?fs John E. Branch, 111 Nate Pencook 128 E. Hargett St, Suite 300 Raleigh, NC 27601 (919)856?9494 jbranch@shanahanlawgroup.com Counsel for the Intervenor-Defendants This 4th day of October, 2019. R. Stanton Jones David P. Gersch Elisabeth S. Theodore Daniel F. Jacobson 601 Massachusetts Ave. NW Washington, DC 20001-3761 (202) 942-5000 Stantonjones@arnoldporter. Com Marc E. Elias Aria C. Branch 700 13 th Street NW Washington, DC 20005-3960 (202) 654-6200 melias@perkinscoie.com Abha Khanna 1201 Third A venue Suite 4900 Seattle, WA 98101?3099 (206) 359-8000 Counsel for Common and the Individual Plaintif?' . f} .I . By; Phillip J. Strach, Bar No. 29456 44 Exhibit 1 STATE OF NORTH CAROLINA IN THE GENERAL COURT OF JUSTICE SUPERIOR COURT DIVISION WAKE COUNTY 18 CVS 014001 Common Cause; et Plaintiffs, V. Representative David R. Lewis, in his of?cial capacity as senior chairman of the House Select Committee on Redistricting, et a1. Defendants. AFFIDAVIT OF CLARK BENSEN State of Vermont ss. County of Orange Clark H. Bensen, af?ant, af?rms under oath as follows: 1. I am an attorney licensed in Vermont. 2. I served as a consulting expert for the North Carolina General Assembly in the above-captioned case. 3. However, I have had an extremely limited role since the date of the Court?s judgment, September 3. 2019. 4. I understand there was a redistricting process in the North Carolina General Assembly after the date of that judgment. But I know very little of the details. 5. Other than a brief discussion with counsel I had no input with anyone involved in the remedial proposals on how the 2019 redistricting should ocCur, how districts should be drawn or otherwise chosen, or anything of that nature. I am only vaguely aware that the General Assembly somehow worked with maps simulated by Dr. Chen at the liability phase. 6. My sole role was restricted to a single exercise designed to ensure that data I understood to be provided from Dr. Chen during the remedial phase was the same data produced by Dr. Chen at the liability phase. This was a mere authentication exercise to make sure that one set of information was an identical match with another set of information. 7. I received an email from counsel late on Monday, September 9, 2019 inquiring of my availability to potentially analyze 1,000 districting plans in a very short period of time. Before that time, I had not conducted any work on the 2019 North Carolina redistricting process, and I had not anticipated any role. I informed counsel, on the next day, Tuesday, September 10, that I had limited availability to process that many ?les expeditiously, since I had not blocked out time to work on the North Carolina redistricting. 8. was subsequently requested to simply compare a sample of the two sets of 1,000 plans with two of the sets of 1,000 plans that Dr. Chen had provided during the liability stage, for the sole purpose of verifying that the plans submitted by Dr. Chen during the remedial stage appeared to be the same plans that had been submitted previously. 9. This task required a certain number of steps to complete because the plan identi?cation of the plans had changed from the liability-phase plans to the remedial-phase plans. 10. Previously they had simply numbered sequentially by their location in the output ?le provided by Dr. Chen?s counsel. 1 1. The new version of plan summary ?les included a pre?x or and then a number, from 1 to 999. This ?le was broken out into geographic subsets but also had a statewide value. 12. I tried to compare this with the information in the summary ?le for the previous sets of plan but the information that had been recently provided did not match exactly as to the ?elds, or variables, provided. 13. I used a combination of these selected factors to compare with all possible plans in the previous file to see if I could ?nd one or more plans that could match the values. 14. A bit of trial and error was required because not only did the content of the ?les differ, the precision of numeric variables sometimes did as well. 15. Due to the extremely short time period I had to ful?ll this request, I focused on the two compactness values, Reock, and Polsby-Popper, to estimate the degree to which it looked like, regardless of the plan identi?cation, the sets of plans were mostly identical. 16. After some more testing it appeared to me ?like almost all of the old plans are included?. I informed counsel of this assessment at 2:30pm on Tuesday. 17. The procedure described above summarizes all of the review I undertook for the sets of remedial plans. I had neither the time nor the instructions to undertake what plaintiffs? counsel alleges. 18. The exercise I performed did not include any review of the potential partisan performance of any simulated districts in the data set I provided. I gave no information to counsel or anyone else about partisan or political information at all. As described, the sole exercise I performed was ensuring, as best I could under the time constraints, that one set of information was identical to another set. 19. The only conclusion I offered was that, as best as I could tell in the limited time frame, that, indeed, the two sets of information matched. 20. The company I own, Polidata, has been involved in the development and analysis of information used in redistricting for several decades. Polidata has been retained by numerous types of clients who have varying needs for services. 22. Counsel for the General Assembly provided to me the objections Plaintiffs ?led in the remedial phase. That brie?ng includes portions of biographical information prepared by me and references multiple capabilities that are available to clients, including the capability of providing political advice and analysis. Indeed, I have provided political information to clients of all types, including state legislatures. Other clients include entities like the Cook Political Report that conduct election analyses. 23. None of those capabilities have anything to do with the General Assembly?s 2019 redistricting with respect to my involvement. Plaintiffs incorrectly infer that because these capabilities exist they were used in connection with that redistricting. As I have described above, they were not. I have read the foregoing statement consisting of 23 paragraphs and swear that it is true and accurate to the best of my knowledge and belief. Clark H. Bensen CALLAHAN blic. State of Vermont g? ion No. 1510011255 0 [its ion Expires Jan. 2021 This fday of October 2019. . A mtary Subscribed and sworn to before me Coin Exhibit 2 STATE OF NORTH CAROLINA IN THE GENERAL COURT OF JUSTICE SUPERIOR COURT DIVISION COUNTY OF WAKE Common Cause, et al., Plaintiffs, Docket No. 18 CVS 014001 V. Representative David R. Lewis, et al., Defendants. AFFIDAVIT OF JANET R. THORNTON, STATE OF FLORIDA ss. COUNTY OF LEON Dr. Janet R. Thornton, affiant, affirms under oath as follows: 1. Counsel for the Defendants in the above captioned matter asked me to describe the work that I was asked to perform for the remedial plans resulting from the September 3rd Decision. In addition, I was asked to review the September 27th report submitted by Dr. Jowei Chen. Work Performed for Remedial Plans From September 3rd Decision 2. On September 9th, I was informed that Dr. Chen would be providing information regarding his simulations. Upon receipt, I was asked to con?rm that the information came from the simulations that he prepared for his report dated April 8, 2019. Given that this was an unanticipated request, my schedule and that of my staff had to be rearranged. I had other deliverables from clients that needed to be ful?lled, as well. I was also asked to review the Declaration that Dr. Chen provided and which I received at 3:39 pm. on September 10'Eh (the time of the email sent to Counsel from Daniel Jacobson was 3:24 3. i received an Excel workbook containing four worksheets the morning of September 10th, 2019. Because Dr. Chen randomized the information rather than retaining the original order of the simulations, one could not make a simple comparison of the information by lining up the original statewide ?les received from Dr. Chen in support of his April report (?les labelled In addition, we had to determine which of the ?elds in Dr. Chen?s April ?les were being reported in the September 10th ?le. 4. To determine if the September 10*h information was from the same (April) simulations, it was necessary to match the ?elds common between the ?les received in April and the Excel ?le received on September 10th. Using the software package SAS, we read in the statewide information from the September 10th Excel ?le and the information from the ?les received for the April report and attempted to match the ?les by Reock score, Polsby?Popper score, number of VTD splits and the number of municipal splits. Unfortunately, they did not readily match because of the many digits after the decimal point in some of the ?elds. Consequently, we manually compared the simulations that did not match pro grammatically to ensure that they were the same. 5. We also attempted to determine if the September information from the county groupings was from the same simulations as the April report. However, because Dr. Chen did not provide the programming logic for the county grouping information that he provided in the Excel ?le received on September it would have taken at least another day to fuliy vet the information. As a consequence, we ended our attempt to replicate the county cluster information. 6. At 5:33 pm. on September 10th, I had identi?ed the simulation number in the information from Dr. Chen?s April report that corresponded to the randomized simulation number in the September 10th Excel ?le by con?rming that the values of the Reock score, Polsby-Popper score, the number of VTD splits and the number of municipal splits were the same between the two sources of information. 7. From September 3rd through the 29th, I did not review the partisan make-up of the simulated maps or review political information for the county groups for the remedial plan. Until I prepared this af?davit, I had not examined the number of Democratic or Republican seats for the House and Senate county groups for the remedial plan. In addition, I do not know any member of the General Assembly and, to the best of my knowledge, I have not spoken with a member of the General Assembly about the political implications of the data that Dr. Chen produced or about any other subject. Further, I was not asked by Counsel to do my own political analysis of any speci?c district or county grouping from Dr. Chen?s simulated maps until Counsel asked me to determine the predicted number of Democratic seats from Dr. Chen?s Set 3 simulations for the ?ve House county groupings in question as compared to the predicted number of Democratic seats from HB 1020 for those same ?ve county groupings, which is outlined in the paragraphs that follow. Work Related to My Review of Dr. Chen?s September 27th Report 8. I was asked to determine the number of unique maps that Dr. Chen generated from his simulations for the ?ve House county groupings, which are the focus of his report.1 In addition, I was asked to determine the predicted number of Democratic seats from his simulations compared to the number as estimated from HB 1020 for each of these ?ve county groupings. 9. With respect to each of the ?ve House county groupings, Dr. Chen did not generate 1,000 unique maps from his simulations. Instead, as shown below, his algorithm generated 1 Dr. Chen did not provide the programming logic used to produce the information for his September 27th Report. However, he did provide ?les for each county grouping and simulation (1,000 ?les per county grouping) containing the district to which his simulation assigned a precinct/VTD. In addition, he provided a summary ?le (G9.txt for Brunswick?New Hanover; Gl9.txt for Cleveland?Gaston; G20.txt for Columbus?Pender-Robeson; G26.txt for and G30.txt for Guilford) for each county grouping containing the number of Republican seats predicted from the simulation based on the 2010-2016 election composite. substantially fewer than 1,000 unique maps. For example, Dr. Chen?s simulation algorithm produced 36 unique maps for the Brunswick?New Hanover county grouping, not 1,000. Connty Grouping Unique Maps Columbus?Pender-Robeson 276 572 Cleveland-Gaston 80 Brunswick-New Hanover 36 Guilford l9 10. At Figures 1-4 of his report, Dr. Chen provides charts depicting the number of Democratic seats estimated from his simulations relative to the number of Democratic seats estimated based on the HB 1020 plan. Dr. Chen does not provide the same information for the ?ve House county groupings. For each of the ?ve county grouping simulations, Dr. Chen provided the estimated Republican seats for each simulation.2 From this information, I can calculate the number of Democratic seats. In addition, based on the information in his report, I can estimate the number of Democratic seats based on the HB 1020 plan.3 11. The following charts provide the estimated number of Democratic seats (districts) for each of the five House county groupings based on Dr. Chen?s Set 3 simulations, which I received on September 30, 2019. In each instance, the number of Democratic seats (districts) based on the HB 1020 plan falls within the most frequently occurring number of Democratic seats (districts) resulting from Dr. Chen?s 1,000 simulations. The one exception is the Guilford county 2 The information can be ascertained ?om the column from the ?les G9.txt for Brunswick-New Hanover; Gl9.txt for Cleveland-Gaston; G20.txt for Columbus-Pender?Robeson; G26.txt for and G30.txt for Guilford. The information is consistent with Figure 5 Figure 11 Figure 17 (Cleveland-Gaston); and Figure 23 (Brunswick?New Hanover). Dr. Chen did not provide the same ?gure for Guilford. The data which he provided for Guilford are not consistent with the information that he provided at Table 7a. His simulations estimate fewer Democratic seats than he reports at Table 7a. 3 Dr. Chen provided the Democratic seat share based on the same 2010?2016 composite for HB plan 1020 at Table 2a Table 321 (F Table 5a (Cleveland?Gaston), and Table 7a (Guilford). Dr. Chen did not provide the same table for Brunswick-New Hanover; but the information can be ascertained from Figure 23 at page 46. grouping for which the majority of Dr. Chen?s 1,000 simulations estimate one Democratic seat While the HB 1020 plan estimates two Democratic seats.4 Figure 1?Dr. Chen?s House Simulations for Columbus-Pender-Robeson County Grouping Democratic?Favoring Districts in HE 1020 House Plan Versus 1,000 Simulated Plans (Measured Using 2010?2016 Election Composite) *0 UI 1,000 900 800 BB 1020 Plan .4 7'00 600 500 400 .300 200 100 Fre que ucy Among Simulate (1 Districtng Plans (1 ,000 Total Simulate (1 Plans) 5 0.5% 99.5% of Plans of Plans Number of Districts with More Democratic than Republican Votes (Out of 3 total Districts for Columbus?Pendeer-Robeson) Gleasured Using Votes Summed Across 201042016 Statewide Elections) 4 While there are six House districts in the Guilford county grouping (see Expert Report of Jowei Chen, April 8, 2019 at page 90), the remedial plan required examining three of the six districts (Common Cause V. Lewis Judgment, September 3, 2019, page 353). Figure 2?Dr. Chen?s House Simulations for County Grouping Democratic-Favoring Districts in HE 1020 House Plan Versus 1,000 Simulated Plans (Measured Using 2010-2016 Election Composite) que ucy Among Simulate Districting Plans (1 ,0 00 Total Simulate [1 Plans) 1,000 900 800 711 700 71.1% 24.8% 4.1% of Plans of Plans of Plans Number of Districts with Llore Democratic than Republican Votes (Out of 5 total Districts for (l-Ieasnred Using Votes Summed Across 2010-2016 Statewide Elections) Figure 3?Dr. Chen?s House Simulations for Cleveland-Gaston County Grouping Democratic-Favoring Districts in HR 1020 House Plan Versus 1,000 Simulated Plans (Measured Using 2010?2016 Election Composite) 1,000 900 800 400 300 200 100 Fre que ucyAmong Simulate c1 Districh'ng Plans (1,000 Total SimulatedPluus) 1:000 HE 1020Plnu 0 1 1000/0 of Plans J- Number of Districts nith L'Iore Democratic than Republican Votes (Out of 4 total Districts for Cleveland-Gaston) Gleasured Using Votes Summed Across 2010?2016 Statewide Elections) Figure 4?Dr. Chen?s House Simulations for Hanover-New Brunswick County Grouping Democratic-Favoring Districts in HE 1020 House Plan Versus 1,000 Simulated Plans (Measured Using 2010-2016 Election Composite) 999 1,000 99.9% 0.1% of Plans of Plans Number of Districts with LIore Democratic than Republican Votes (Out of 4 total Districts for Hanover-I Tew Brunswick) (Ideasured Using Votes Summed Across 2010?2016 Statewide Elections) Figure 5?Dr. Chen?s House Simulations for Guilford County Grouping Democratic-Favoring Districts in HB 1020 House Plan Versus 1,000 Simulated Plans (Measured Using 2010-2016 Election Composite) 1000 D: 900 '3 826 '5 a 300 33.. 7?82.6% 17.4% of Plans of Plans Number of Districts with hiore Democratic than Republican Votes (Out of 3 total Districts for Guilford) (.Measured Using Votes Summed Across 2010?2016 Statewide Elections) I have read the foregoing statement consisting of 11 paragraphs and swear that it is true and accurate to the best of my knowledge and belief. 0W Janet R. Thornton, Subscribed and sworn to before me g: - . this day of October, 2019. J??tw - mm?.m Exhibit 3 AFFIDAVIT OF THOMAS L. BRUNELL, 1. Counsel for the Defendants asked me to read the September 27, 2019 report by Prof. owei Chen and respond to it. I would like to note for the record that I did no work for the Defendants between my trial testimony and the passage of the revised maps. So between Friday July 26th (when I ?nished my testimony) and September 19th (when Tom Farr contacted my about doing this af?davit), I did no work and played no role in drawing districts nor conducting political analysis on Prof. Chen?s maps or any speci?c simulated group or district. 2. Columbus grouping?- The chief complaint for this county grouping appears to be that district 46 goes from 53.3% to 51.37% Democrat, and district 16 is more Democratic than most simulations. Overall the county grouping is still 2?1 Democratic to Republican districts. It is worth noting that there are ?ve of Chen?s simulated maps that end up with two Republican seats and just one Democratic seat. Prof. Chen reports that 40.2% of simulated plans have a higher Reock score, but this also means that 59.8% of the simulated maps are less compact. Similarly, he argues that 36.5% have higher Polsby?Popper scores, which means the enacted districts are more compact that 63.5% of the simulated districts. The enacted districts split no municipalities or VTDs there is no room for improvement on those two variables. 3. grouping The base map (A737) has 3 safe Republican districts and 2 safe Democratic districts. The minor changes made to the districts in HE 1020 have no substantive effect on the partisanship of any district. The biggest change is to District 75 in which a safe Republican district is made even safer (this is generally bene?cial to Democrats as it is effectively wasting more Republican votes). The enacted map does split more municipalities that Chen?s base map, but so do 87.3% of his simulated maps for this county grouping (Figure 15). Prof. Chen argues the small decreases in compactness scores of HB 1020 relative to the base map is another reason to redraw the districts in this county grouping. This is paying far too much attention to the metrics and the small differences. I think it is best to actually look at the districts themselves to evaluate their compactness or lack thereof. It is also useful to visually inspect districts to understand how small differences in these metrics matter very little. For instance, look at the shape of district 17 on page 46 of Prof. Chen?s report and compare it to district 47 on page 15 of the same report. One of these districts has a Polsby?Popper score of .3 and the other is scored .42. I think both look perfectly reasonable and relatively compact. I might have even guessed that district 17 is more compact that 47, but that?s not the case. It is to recall that both Reock and Polsby?Popper use a circle as the best shape of a legislative district, when in reality we do not draw-circular shaped districts. This fact, however, is the main reason that district 47 is scored higher on this metric it is closer in shape to a circle than district 17, which is morerectangular and elongated. But the shape of district 17, and any district for that matter is determined, in part, by the shape of the county grouping. 4. Cleveland grouping All four of these districts are safe Republican districts. None of Prof. Chen?s simulations draw even one district with 50% predicted Democratic vote share. There are some trivial differences in average levels of compactness and the enacted map does split one VTD, as do nearly half of Prof. Chen?s simulated districts (Figure 22). If reducing partisan gerrymandering is the goal, I am uncertain as to how not splitting that VTD or making districts more circular in shape helps advance the goal of eliminating a partisan gerrymanderi 5. Brunswick grouping This grouping has three Republican districts and one Democratic district. Every one of Prof. Chen?s simulations draw the same 3?1 split, except for one simulation that draws four Republican districts. Compactness measured by Reock scores is higher in the enacted map than any one of the 1,000 simulations. The enacted map splits the lowest number of municipalities as possible (one) and it splits no VTDs while more than three quarters of Prof. Chen?s simulations splits one VTD. 6. Guildford grouping The two main points that Prof. Chen makes in his objection to this county grouping is that District 59 is reduced from a 45.97% Democratic district to a 45.46% Democratic district. This change is very minor and the election data Prof. Chen uses is dated and averaged across many elections. In terms of compactness, the differences that Chen points out in Figures 29 and 30 are really and truly trivial somewhere between .01 and .02 on each of the metrics. I would be stunned if anyone could consistently tell the difference between a district with a .385 Reock score and a district with a .39 Reock score. Moreover, Prof. Chen?s graph arti?cially exaggerates the differences between the enacted map and the simulated maps by restricting the use of categories (or ?bins? as we call them for graphs). Notice on Figure 29 there are only three categories: .38, .39, and .40. HB 1020 is somewhere between .38 and .39. But Prof. Chen rounds his estimates up and down to get them into just three categories, so the ?gure is misleading. This is true for all his ?gures on compactness. 7. Comparison to Chen?s Simulations for Partisan Outcomes On page 27 of Prof. Chen?s report (Table 5) he has a range of outcome for House elections using 2010-2016 statewide elections composite for his simulations. The outcomes for Set 1 (with no incumbent protection) is as follows: Number of Democratic Number of Simulations Leaning Seats with this Outcome submitted the shape ?les for HB1020 to planscore.org and the reports for this map indicate that for the whole state 49 of the districts favor the Democrats. In Chen?s simulations only 2.2% of his 1,000 maps resulted in more seats for the Democrats at 50 seats, and 0.2% at 51 seats). Prof. Chen reports that using his older election data HB 1020 only has 44 districts that lean Democratic. Prof. Nicholas Stephanopoulos gets the same results as I do (his blog post is in the Appendix A of this report). Moreover, as Prof. Stephanopoulos notes in his blog post, the elections used by planscoreorg for North Carolina average 48 percent Democratic while Prof. Chen?s elections composite has an average of 47.92% (See page 31 of Prof. Chen?s April 8, 2019 report). 8. It is worth noting statewide partisan gerrymandering metrics include many areas in the state that are not impacted by the court?s decision. It is more appropriate to analyze either the 14 county groupings ordered to be redrawn, or even better, just the ?ve county groupings that the Plaintiffs are objecting to. That notwithstanding, Prof. Chen?s own analysis indicates that Democrats in North Carolina are naturally disadvantaged given the distribution of voters across the state (page 36, Chen Report of April 8, 2019). 9. Split Municipalities in House by County Groupings 2017 2019 Brunswick-New Hanover 7 3 Cleveland?Gaston 11 5 Columbus-?Pender?Robeson 2 0 13 9 Guilford 13 11 Totals 46 28 Among the ?ve contested county groupings, the number of split municipalities declined from 46 to 28.1L 1 House districts in each county grouping: Brunswick-New Hanover 17, 18, 19, and 20; Cleveland?Gaston: 108, 109, 110, and 111: Columbus-Pender~Robeson: 16, 46, and 47; 71, 72, 73, 74, and 75; Guilford: 57, 58, 59, 60, 61, and 62. 10. Split in House by County Grouping County Group 2017 2019 Brunswick?New Hanover 8 Cleveland-Gaston 2 2 Columbus- Ponder?Robeson 0 Guilford 0 0 Totals 10 2 Split VTDs decline signi?cantly among the affected districts in the county groupings falling from 10 before to just 2 after the new districts were drawn. I have read the foregoing statement consisting of ten paragraphs and swear that it is true and accurate to the best of my knowledge and belief. FURTHER THE AFFLANT SAYETH NOT. Thomas Brunell, Subscri and sworn to before me I This day of October, 2019 garage .75- A ?341.130,, SAFET VERLASEVIC .93. a '35? Notary Public, State of Taxes I 65,; {3?8 f; Appendix A Stephanopoulos Blog Post 8 Earlier today, the North Carolina legislature approved remedial maps to replace the state house and state senate plans that were recently struck down as partisan gerrymanders. PlanScore assessed the remedial maps, and here are the results. For the state house, the old plan had an efficiency gap of a partisan bias of and a mean-median difference of 5% (all in a Republican direction, and based on a model using 2016/2018 data). On the other hand, the new map has an efficiency gap of a partisan bias of and a mean-median difference of 3% (again all pro?Republican). So the new map is about half as skewed as the old plan. For the state senate, the old plan had an efficiency gap of 11%, a partisan bias of and a mean?median difference of 4% (ail pro~Republican). By comparison, the ge_w map has an efficiency gap of a partisan bias of and a mean?median difference of 3% (all pro-Republican). So the new map is approximately one?third as skewed as the old plan. it?s also interesting to compare the remedial plans to the distributions of maps randomly generated by the plaintiffs? expert, Jowei Chen. (Especially since each remedial plan used one of Chen?s maps as its starting point.) At the state house level, Chen?s maps contained anywhere from 43 to 51 Democratic districts with a median of 46 (assuming a 48% Democratic statewide vote share). The remedial plan has 49 Democratic districts in that electoral environment. At the state senate level, Chen?s maps contained anywhere from 19 to 22 Democratic districts with a median of 20. The remedial plan has 22 Democratic districts. i should note that this isn?t a perfect apples?to-apples comparison. Chen analyzed partisanship using an aggregate of ten statewide elections from 2010 to 2016. PlanScore relies on a model using 2016/2018 data in which legislative vote share is a function of presidential vote share and incumbency. Still, PlanScore assumes about the same electoral environment as Chen (48% Democratic) so the estimates for the remedial plans are at'least roughly comparable to Chen?s figures. Another caveat is that, because of North Carolina's whole county provision, its maps aren?t drawn in one statewide swoop. Rather, districts are drawn within a large number of county groupings. It?s thus perfectly possible for certain county groupings to be outside the distributions of simulations for those groupings, even if the maps as a whole are within the statewide distributions. That said, one?s conclusion about the remedial plans plainly depends on the baseline. Given a baseline of perfect symmetry, one would find the remedial plans better than their predecessors but still reasonably far from treating both parties equally (especially the House plan). But given a baseline of randomly generated maps, one would find the remedial plans satisfactory. Both plans fall within the corresponding distributions of simulated maps?and indeed on the Democratic side of the distributions? medians. Appendix Below is a list of district predictions from planscore.org?s analysis of 2019 House plan. Predicted Democratic Vote Predicted Republican Vote District Share Share 1 47.5% 52.5% 2 42.3% 57.7% 3 41.2% 58.8% 63.5% Fm w. 'le a 55.7% 44.3% WI 6 35.3% 64.7% '7 42.5% 57.5% 3 61.2% 8.8% 50.3% 49.7% 34.3% 65.7% 1'1 67.3% 32.7% 12 47.2% 52.8% 1.1% 68.9% [mu?L EH 5 La.) ti? . Predicted Democratic Vote District Share 16 21? Ni Ln}: 24% 27 37.9% 3169/0 38.1% 36.5% 42.1% 38.6% 60.4% Predicted Republican Vote Share 62.1% 68.2% 61.9% 63.5% 43.5% 57.9% 44.7% 56.2% District 33 33 4G 41 Predicted Democratic Vote Share 34.1% 84.3% 70.1%) 62.1% 68.7% 57.3% 50.6% 79.0% 64.7% 54.0% 47.1% Predicted Republican Vote Share 65.9% 15.7% 29.9" 20.5% 37.9% 31.3% 42.7% 49.4% 48.0% 52.9% 21.0% 35.3% 46.0% 40.9% 10 Predicted Democratic Vote Predicted Republican Vote District Share Share 42 64.3% 35.7% 44 70.0% 30.0% 45 49.0% 51.0% 4.5 42.9% 57.1% 47 48.4% 51.6% 48 55.3% 44.7% 65.6% :35. ?43 4.4% DJ 5% 56.4% 43.6% 43.5% 56.5% 52 37.7% 62.3% 38.6% 61 14% Um mm 58.8% 41.2% U1 1 55 44.2% 55.8% 11 Predicted Democratic Vote District Share 56 83.3% 57 33% 72.1% 5% 45.6% 68 62.4% "571.5%. 62 42.1% 63 49.2% ?at- 45.1% 373% 40.5% 6?7 73.6% 35.3924. 69 35.5% Share 16.7% 53.792, 27.9% 54.4% 37.6% 62.8% 59.5% Predicted Repubiican Vote 12 District 79 '72 76 77 7? Predicted Democratic Vote Share 25.8% 70.5% 70.4% 45.5% 37.6% 41.5% 26.8% 30.2% 4-4.3 41.3% Predicted Republican Vote Share 74.2% 29.5% 29.6% 64.3% LII o\ 62.4% 58.5% 74.0% 76.7% 62.l% 73.2% 69.8% 55.7% 58.2% 13 District 81% C16 . GD ?42: 91 96 9?7 Predicted Democratic Vote Share 32.i% 24.2% 26.9% 26.5% 25.2% Predicted Republican Vote Share 67.9% 75.8% 66.5% 73.1% 36.2% 69.9% 73.5% 74.5% 29.6% 74.8% 66.4% 71.6% 62f39/0 58.3% 14 District 13%} ?93 H34 16? 31% Predicted Democratic Vote Share 78.7% 79.9% 53.4% 37.9% 37.3% P0 03 a? . 30.9% Predicted Republican Vote Share 54.2% DJ :4 27.8% 29.3% 21.0% 44.44 46.6% 46.6% 20.3% 26.7% 62.3% 15 Predicted Democratic Vote Predicted Republican Vote District Share Share $12 27.3% 72.7% 113 . 628% 1714 56.1% 43.9% ?1'15 581% 41.9% 116 56.9% 43.1% 118 35.1% 64.9% 119 42.6% - 57.4% 126 . .. 263% . 737% 40234026.1 16