League of Women Voters of PA v. Cmwlth

OPINION

, It is a core principle of our republican form of government “that the. voters should choose their representatives, not the other way around.”1 In this case, Petitioners allege that the Pennsylvania Congressional Redistricting Act of 20112 (the “2011 Plan”) does the latter, infringing upon that most central of democratic rights — the right to vote. Specifically, they contend that the 2011 Plan is an unconstitutional partisan gerrymander. While federal courts have, to date, been unable to settle on a workable standard' by which to assess such claims under the federal Constitution, we find no such barriers under our great Pennsylvania charter. The .people of this Commonwealth should never lose sight of the fact that, in its protection of essential rights, our founding document is the ancestor, not the offspring, of the federal Constitution, We conclude that, in this matter, it provides a constitutional standard, and remedy, even if the federal charter does not. Specifically, we hold that the 2011 Plan violates Article I, Section 5 — -the Free and Equal Elections Clause— of the Pennsylvania Constitution.

The challenge herein was brought in June 2017 by Petitioners, the League of Women Voters3 and 18 voters — all registered Democrats, one from each of our state’s congressional districts — against Governor . Thomas W. Wolf, Lieutenant Governor Michael J. Stack, III, Secretary Robert Torres, and Commissioner Jonathan M. Marks (collectively, “Executive Respondents”), and the General Assembly, Senate President Pro Tempore Joseph B. Scarnati, III, and House Speaker Michael C. Turzai (collectively, “Legislative Respondents”).4 5 Petitioners alleged that the 20Í1 Plan violated .several provisions of our state Constitution.

. On January 22, 2018, this Court entered a per curiam order6 agreeing with Petitioners, and deeming the 2011 Plan to “clearly, plainly and palpably violate[]” our state Constitution, and so enjoined its further use.7 See Order, 1/22/18. We further provided .that, if the' General Assembly and the Governor did not enact a remedial plan by February 15, 2018, this Court would choose a remedial plan. For those endeavors, we set forth the criteria to be applied in measuring the constitutionality of any remedial plan, holding that:

any congressional districting plan shall consist of: congressional districts composed of compact and contiguous territory; as nearly equal in population as practicable; and which do not divide any county, city, incorporated town, borough, township, or ward, except where necessary to ensure equality of population.

Order, 1/22/18, ¶ “Fourth.”8 Our Order indicated that an opinion would follow. This is that Opinion, and we emphasize that, while explicating our rationale, nothing in this Opinion is intended to conflict with, or in any way alter, the mandate set forth in our Order of January 22, 2018.9

I. Background

A. Redistricting Mandate

Article I, Section 2 of the United States Constitution requires that a census be taken every 10 years for the purpose of apportioning the United States House of Representatives. Following the 2010 federal census, Pennsylvania’s share in the House was reduced from 19 to 18 members.10 As a result, the Commonwealth was required to redraw its congressional district map.

Pennsylvania’s congressional districts are drawn by the state legislature as a regular statute, subject to veto by the Governor.11 While this process is dictated by federal law, it is delegated to the states. The federal Constitution’s Elections Clause provides that “[t]he Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof,” unless Congress should “make or alter such Regulations.” U.S. Const, art. I, § 4, cl. 1. Pursuant to the Elections Clause, Congress passed 2 U.S.C. § 2a, which provides that, following the decennial census and reapportionment, the Clerk of the House of Representatives shall “send to the executive of each State a certificate of the number of Representatives to which such State is entitled” and the state shall be redistricted “in the manner provided by the law thereof.” 2 U.S.C. § 2a. If the state does not do so, Representatives are to be elected as further provided in Section 2a.12

B. Plan Passage

The 2011 Plan, Senate Bill 1249, was enacted on December 22, 2011, setting forth Pennsylvania’s 18 congressional districts.13 In the November 2010 general election, voters elected Republicans to majorities in both houses of the General Assembly and elected a Republican, Tom Corbett, as Governor. Thus, in 2011, the Republican-led General Assembly was tasked with reconstituting Pennsylvania’s congressional districts, reducing their number by one, and adjusting their borders in light of population changes reflected by the 2010 Census. On May 11, June 9, and June 14, 2011, the Pennsylvania House and Senate State Government Committees held hearings on the subject of redistricting, for the ostensible purpose of receiving testimony and public comment on the subject of redistricting generally. On September 14, 2011, Senate Bill 1249, Printer’s Number 1520, principally sponsored by the Republican leadership, was introduced, but contained absolutely no information concerning the boundaries of any congressional districts. On December 7, 2011, the bill was brought up for first consideration, and, on December 11, 2011, for second consideration.

Thereafter, the bill was referred to the Senate State Government Committee, where, on December 14, 2011, it was amended and reprinted as Senate Bill 1249, Printer’s Number 1862, now providing proposed boundaries for each of Pennsylvania’s 18 congressional districts, before being reported out of committee. The same day, the bill was referred to the Senate Appropriations Committee, where it was again amended and reprinted as Senate Bill 1249, Printer’s Number 1869, and reported out of committee to the floor. There, Democratic Senator Jay Cos-ta introduced an amendment to the bill he indicated would modify it to create 8 Republican-favorable districts, 4 Democrat-favorable districts, and 6 swing districts, but the Senate declined to adopt the amendment and passed Senate Bill 1249, Printer’s Number 1869, in a 26-24 vote, with all Democrats voting against passage. The same day, Senate Bill 1249, Printer’s Number 1869, proceeded to the House of Representatives, where it was referred to the House State Government Committee,' and reported out of committee. The next day, on December 15, 2011, Senate Bill 1249, Printer’s Number 1869, was brought up for first consideration, and, on December 19, 2011, second consideration. On December 20, 2011, the bill was referred to the House Appropriations Committee, re-' ported out of the committee, and passed in. a 136-61 vote, with 36 Democrats voting in favor of passage,14 On December 22, 2011, Senate Bill 1249, Printer’s Number 1869, proceeded to the governor’s desk where then-Governor Corbett signed, it into law as Act 131 of 2011, the 2011 Plan.

C. The 2011 Plan

A description of the 2011 Plan and some of its characteristics is appropriate.15 A map of the entire 2011 Plan is attached as Appendix A.

1. The Districts

a. 1st Congressional District

The 1st Congressional District is composed of parts of Delaware and Philadelphia Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 6.

b. 2nd Congressional District

The 2nd Congressional District is composed of parts of Montgomery and Philadelphia Counties, and. appears as follows:

[[Image here]]

See Joint Exhibit 7.

c. 3rd Congressional District

The 3rd Congressional District is composed of Armstrong, Butler, and Mercer Counties, together with parts of Clarion, Crawford, Erie, and Lawrence Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 8.

d. 4th Congressional District

The 4th Congressional District is composed of Adams and York Counties, together with parts of Cumberland and Dauphin Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 9.

e. 5th Congressional District

The 5th Congressional District is composed of Cameron, Centre, Clearfield, Clinton, Elk, Forest, Jefferson, McKean, Potter, Venango, and Warren Counties, together with parts of Clarion, Crawford, Erie, Huntingdon, and Tioga Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 10.

f. 6th Congressional District

The 6th Congressional District is composed of parts of Berks, Chester, Lebanon, and Montgomery Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 11.

g. 7th Congressional District

The 7th Congressional District is composed of parts of Berks, Chester, Delaware, Lancaster, and Montgomery Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 12.

h. 8th Congressional District

The 8th Congressional District is composed of Bucks County, together with parts of Montgomery County, and appears as follows:

[[Image here]]

See Joint Exhibit 13.

i. 9th Congressional District

The 9th Congressional District is composed of Bedford, Blair, Fayette, Franklin, Fulton, and Indiana Counties, together with parts of Cambria, Greene, Hunting-don, Somerset, Washington, and West-moreland Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 14.

j. 10th Congressional District

The 10th Congressional District is composed of Bradford, Juniata, Lycoming, Mifflin, Pike, Snyder, Sullivan, Susquehanna, Union, and Wayne Counties, together with parts of Lackawanna, Monroe, Nor-thumberland, Perry, and Tioga Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 15.

k. 11th Congressional District

The 11th Congressional District is composed of Columbia, Montour, and Wyoming Counties, together with parts of Carbon, Cumberland, Dauphin, Luzerne, Northumberland, and Perry Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 16,

l. 12th- Congressional District

The 12th Congressional District is composed of Beaver County, together with parts of Allegheny, Cambria, Lawrence, Somerset, and Westmoreland Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 17.

m. 13th Congressional District

The 13th Congressional District is composed of parts of. Montgomery and Philadelphia Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 18.

n. 14th Congressional District

The 14th Congressional District is composed of parts of Allegheny and West-moreland Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 19.

o. 15th Congressional District

The 15th Congressional District is composed of Lehigh County and parts of Berks, Dauphin, Lebanon, and Northampton Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 20.

p. 20th Congressional District

The 16th Congressional District is composed of parts of Berks, Chester, and Lancaster Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 21.

q. 17th Congressional District

The 17th Congressional District is composed of Schuylkill County and parts of Carbon, Lackawanna, Luzerne, Monroe, and Northampton Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 22.

r. 18th Congressional District

Finally, the 18th Congressional District is composed of parts of Allegheny, Greene, Washington, and Westmoreland Counties, and appears as follows:

[[Image here]]

See Joint Exhibit 23.

2. Other Characteristics

Of the 67 counties in Pennsylvania, the 2011 Plan divides a total of 28 counties between at least two different congressional districts:16 Montgomery County is divided among five congressional districts; Berks and Westmoreland Counties are each divided among four congressional districts; 17 Allegheny, Chester,18 and Philadelphia Counties are each divided among three congressional districts; and Cambria, Carbon, Clarion, 'Crawford, Cumberland, Delaware, Erie,19 Greene,. Huntingdon, Lackawanna, Lancaster, Lawrence, Lebanon, Luzerne, Monroe, Northampton,20 Northumberland, Perry, Somerset, Tioga, and Washington Counties -áre each split between two congressional districts.21 Additionally, whereas, prior to 1992, no municipalities in Pennsylvania were divided among multiple congressional districts, the 2011 Plan divides 68, or 2.66%, of Pennsylvania’s municipalities between at least two Congressional districts.22

Finally, as noted above, the General Assembly was tasked with reducing the number of Pennsylvania’s congressional districts from 19 to 18, necessitating the placement of at least two congressional incumbents into the same district. The 2011 Plan placed then-Democratic Congressman for the 12th Congressional District Mark Critz and then-Democratic Congressman for the 4th Congressional District Jason Altmire into the same district. Notably, the two faced off in an ensuing primary election, in which Critz prevailed. He subsequently lost the general election to now-Congressman Keith Rothfus, who has prevailed in each biannual election thereafter.

D. Electoral History

As grounding for the parties’ claims and evidentiary presentations, we briefly review the Commonwealth’s electoral history before and after the 2011 Plan was enacted.23 As noted above, the map for the 2011 Plan is attached at Appendix A. The parties have provided copies of prior congressional district maps — for 1943, 1951, 1962, 1972, 1982, 1992, and 2002 — -which were procured from the Pennsylvania Manual.24 They are attached as Joint Exhibit 26 to the Joint Stipulations of Fact. See Joint Stipulation of Facts, 12/8/17, at ¶ 93.

The distribution of seats in Pennsylvania from 1966 to 2010 is shown below:

Year Districts Democratic Seats Republican Seats

1966 27 14 13

1968 27 14 13

1970 27 14 13

1972 25 13 12

1974 25 14 11

1976 25 17 8

1978 25 15 10

1980 25 12[25] 12

1982 23 13 10

1984 23 13 10

1986 23 12 11

1988 2? 12 11

1990 23 11 12

1992 21 11 10

1994 21 11 10

1996 21 11 10

1998 21 11 10

2000 21 10 11

2002 1Ó 7 12

2004 19 . 7 12

2006 19 11 8

2008 19 12 7

2010 19 7 12

Joint Stipulation of Facts, 12/8/17, at ¶ 70.

In the three elections since the 2011 Plan was enacted, Democrats have won the same five districts, and Republicans have won the same 13 districts. In the 2012 election, Democrats won five congressional districts with an average of 76.4% of the vote in each, whereas Republicans won the remaining 13 congressional districts with an average 59.5% of the vote in each, and, notably, Democrats earned a statewide share of 50.8% of the vote, 'an average of 50.4% per district, with a median of 42.8% of the vote, whereas Republicans earned only a statewide share of 49,2% of the vote.26

In the 2014 election, Democratic candidates again won five congressional races, ■with an average of 73.6% of the vote in each, whereas Republicans again won 13 congressional districts, with an average of 63.4% Of the vote in each.27 In 2014, Democrats earned a 44.5% statewide vote share in contested races, whereas Republicans earned a 55.5% statewide vote share in contested races, with a 54.1% statewide share vote in the aggregate.

In the 2016 election, Democrats again won those same five congressional districts, with an average of 75.2% of the vote in each and a statewide vote share of 45.9%, whereas Republicans won those same 13 districts with an average of 61.8% in each and a statewide vote share of 54,1%.28 29

In short, in the last three election cycles, the partisan distribution has been as follows:

Year Districts Democratic Seats Republican Seats Democratic Vote Percentage Republican Vote Percentage

2012 18 5 13 50.8% 49.2%

2014 18 5 13 44.5% 55,5%

2016 18 5 13 45.9% 54,1%

Joint Stipulation of Facts, 12/8/18, at ¶ 102.

II. Petitioners’ Action

Petitioners filed this lawsuit on June 15, 2017, in 'the Commonwealth Court. In Count I of their petition for review, Petitioners alleged that the 2011 Plan30 violates their rights to free expression and association under Article I, Sections 731 and 2032 of the Pennsylvania Constitution. More specifically, Petitioners alleged that the General Assembly created the 2011 Plan by “expressly and deliberately consider[ing] the political views, voting histories, and party affiliations of Petitioners and other Democratic voters” with the intent to burden and disfavor Petitioners’ and other Democratic voters’ rights to free expression and association. Petition for Review, 6/15/17, at ¶¶ 105. Petitioners further alleged that the 2011 Plan had the effect of burdening and disfavoring. Petitioners’ and other Democratic voters’ rights to free expression and association because the 2011 Plan “prevented Democratic voters from electing the representatives of their choice and from influencing the' legislative process” and suppressed “the political views and expression of Democratic voters.” Id. at ¶ 107. They contended the Plan “also violates the Pennsylvania Constitution’s prohibition against retaliation against individuals who exercise their rights under” these articles. Id. at ¶ 108. Specifically, Petitioners alleged that the General Assembly’s “cracking” of congressional districts in the 2011 Plan has resulted in their inability “to elect representatives of their choice or to influence the political process.” Id. at ¶ 112.

In Count II, Petitioners alleged the Plan violates the equal protection provisions of Article 1, Sections 1 and 2633 of the Pennsylvania Constitution, and the Free and Equal Elections Clause of Article I, Section 534 of the Pennsylvania Constitution. More specifically, Petitioners alleged that the Plan intentionally discriminates against Petitioners and other Democratic voters by using “redistricting to maximize Republican seats in Congress and entrench [those] Republican members in power.” Id. at ¶ 116. Petitioners further alleged that the Plan has an actual discriminatory effect, because it “disadvantages Petitioners and other Democratic voters at the polls and severely burdens their representational rights,” Id. at ¶ 117. They contended that “computer modeling and statistical tests demonstrate that Democrats receive far fewer congressional seats than they would absent the gerrymander, and that Republicans’ advantage is nearly impossible to overcome.” Id. at ¶ 118. Petitioners claimed that individuals who live in cracked districts under the 2011 Plan are essentially excluded from the political process and have been denied any “realistic opportunity to elect representatives of their choice,” and any “meaningful opportunity to influence legislative outcomes.” Id. at ¶ 119. Finally, Petitioners claimed that, with regard to individuals living in “packed” Democratic districts under the Plan, the weight of their votes has been “substantially diluted,” and their votes have no “impact on election outcomes.” Id. at ¶ 120.

In response to Respondents’ application, on October 16, 2017, Judge Dan Pellegrini granted a stay of the Commonwealth Court proceedings pending the United States Supreme Court’s decision in Gill v. Whitford, No. 16-1161 (U.S. argued Oct. 3, 2017). However, thereafter, Petitioners filed with this Court an application for extraordinary relief, asking that we exercise extraordinary jurisdiction over the matter.35 On November 9, 2017, we granted the application and assumed plenary jurisdiction over the matter, but, while retaining jurisdiction, remanded the matter to the Commonwealth Court to “conduct all necessary and appropriate discovery, pre-trial and trial proceedings so as to create an evidentiary record on which Petitioners’ claims may be decided.” Supreme Court Order, 11/9/17, at 2. We ordered the court to do so on an expedited basis, and to submit to us findings of fact and conclusions of law no later than December 31, 2017. Id. Finally, we directed that the matter be assigned to a commissioned judge of that court.

The Commonwealth Court, by the Honorable P. Kevin Brobson, responded with commendable speed, thoroughness, and efficiency, conducting a nonjury trial from December 11 through 15, and submitting to us its recommended findings of fact and conclusions of law on December 29, 2017, two days prior to our deadline.36 Thereafter, we ordered expedited briefing, and held oral argument on January 17, 2018.

III. Commonwealth Court Proceedings

In the proceedings before the Commonwealth Court, that court initially disposed of various pretrial matters. Most notably, the court ruled on Petitioners’ discovery requests, and Legislative Respondents’ objections thereto, directed to gleaning the legislators’ intent behind the passage of the 2011 Plan. By order and opinion dated November 22, 2017, the court concluded that, under the Speech and Debate Clause of the Pennsylvania Constitution,37 the court “lack[ed] the authority 'to compel testimony or the production of documents relative to the intentions, motivations, and activities of state legislators and their staff with respect to the consideration and passage of’ the 2011 Plan, Commonwealth Court Opinion, 11/22/17, at 7, and so quashed those requests 38

In addition, Petitioners sought to admit, and Legislative Respondents sought to exclude, certain materials produced by-House Speaker Mike Turzai in the federal litigation in Agre v. Wolf, supra, in response to permitted discovery in that case, along with Petitioners’ expert Dr. Jowei Chen’s expert reports and testimony based on those materials. (As noted, similar discovery was denied in this case, per the Commonwealth Court’s Speech and Debate Clause ruling.) These materials include redistricting maps revealing partisan scoring down to the precinct level, demonstrating that some legislators designing the 2011 Plan relied upon such partisan considerations. Ultimately, the court permitted Dr. Chen’s testimony about these materials, but refused to admit the materials themselves, refused to make any findings about them, see Findings of Fact at ¶ 307, and submitted a. portion to this Court under seal, see Petitioners’ Exhibit 140. Notably, that sealing order required Petitioners to submit both a “Public” and a “Sealed” version of their brief in order to discuss Exhibit 140.39 Given our disposition of this matter, we do not further address these materials or the court’s evidentiary rulings with respect to them.

In all, the court heard oral argument and ruled on .eight motions in limine.40

A. Findings of Fact of the Commonwealth Court

Prior to the introduction of testimony, the parties and Interveners stipulated to certain background faets, much of which we have discussed above, and to the introduction of certain portions of deposition and/or prior trial testimony as exhibits.41

1. Voter Testimony

Initially} several Petitioners testified at trial. They testified as to their belief that, under the 2011 Plan, their ability to elect a candidate who represents their interests and point of view has been compromised. William Marx, a resident of Delmont in Westmoreland County, testified that he is a registered Democrat, and that, under the 2011 Plan, he lives in the 12th Congressional District, which is represented by Congressman Keith Rothfus, a Republican. Marx testified that Congressman Rothfus does not represent his views on, inter alia, taxes, healthcare, the environment, and legislation regarding violence against women, and he stated that he has been unable to communicate with him. Marx believes that the 2011 Plan precludes the possibility of having a Democrat elected in his district. N.T. Trial, 12/11/17, at 113-14.

■ Another Petitioner, Mary Elizabeth Lawn, testified that she is a Democrat who lives, in the city of Chester. Under the 2011 Plan, Chester is in the 7th Congressional District, which is represented by Congressman Patrick Meehan, a Republican.'42 Id. at 134, 137-39. According to Lawn, Chester is a “heavily African-American” city, and,, prior to the enactment of the 2011 Plan, was a part of the 1st Congressional District, which is represented by Congressman Bob' Brady, a Democrat.43 Id. at 135, 138-39. According to Lawn, since the enactment of the 2011 Plan, she has voted for the Democratic candidate in three state elections, and her candidate did not win any of the elections. Id. at 140. Lawn believes that the 2011 Plan has affected her ability to participate in the political process because she was placed in a largely Republican district where the Democratic candidate “doesn’t really have a chance.” Id. Like Marx, Lawn testified that her congressman does not represent her views on many issues, and that she found her exchanges with his office unsatisfying. Id. at 140-44.

Finally, Thomas Rentschler, a resident of Exeter Township, testified that he is a registered Democrat. N.T. Trial, 12/12/17, at 669. Rentschler testified that he lives two miles from the City of Reading, and that he has a clear “community of interest” in that city. Id. at 682. Under the 2011 Plan, however, Reading is in the" 16th Congressional District, and Rentschler is in the 6th Congressional District, which is represented by Congressman Ryan Costello, a Republican. Id. at 670-71, 677. Rentschler testified that, while he voted for the Democratic candidate in the last three state elections, all three contests were won by the Republican candidate. Id. at 673. In Rentschler’s view, the 2011 Plan “has unfairly eliminated [his] chance of getting to vote and actually elect a Democratic candidate just by the shape and the design of the district.” Id. at 674.

2. Expert Testimony

Petitioners presented the testimony of four expert witnesses, and the Legislative Respondents sought to rebut this testimony through two experts of their own. We address this testimony seriatim⅛.

Dr. Jowei Chen

Petitioners presented the testimony of Dr. Jowei Chen, an expert in the areas of redistricting and political geography who holds research positions at the University of Michigan, Stanford University, and Willamette University.44 Dr. Chen testified that he evaluated the 2011 Plan, focusing on three specific questions: (1) whether partisan intent was the predominant factor in the drawing of the Plan; (2) if so, what was the effect of the Plan on the number of congressional Democrats and Republicans elected from Pennsylvania; and (3) the effect of the Plan on the ability of the 18 individual Petitioners to elect a Democrat or Republican candidate for congress from their respective districts. N.T. Trial, 12/11/17, at 165.

In order to evaluate the 2011 plan, Dr. Chen testified that he used a computer algorithm to create two sets, each with 500 plans, of computer-simulated redistricting plans for Pennsylvania’s congressional districts. Id. at 170. The computer algorithm used to create the first set of simulated plans (“Simulation Set 1”) utilized traditional Pennsylvania districting criteria, specifically: population equality; contiguity; compactness; absence of splits within municipalities, unless necessary; and absence of splits within counties, unless necessary. Id. at 167. The computer algorithm used to create the second set of simulated plans (“Simulation Set 2”) utilized the aforementioned criteria, but incorporated the additional criteria of protecting 17 incumbents,45 which, according to Dr. Chen, is not a “traditional districting criterion.” Id. at 206. Dr. Chen testified that the purpose of adding incumbent protection to the criteria for the second set of computer-simulated plans was to determine whether “a hypothetical goal by the General Assembly of protecting incumbents in a nonpartisan manner might somehow explain or account for the extreme partisan bias” of the 2011 Plan. Id.

With regard to Simulation Set 1, the set of computer-simulated plans utilizing only traditional districting criteria, Dr. Chen noted that one of those plans, specifically, “Chen Figure 1: Example of a Simulated Districting Plan from Simulation Set 1 (Adhering to Traditional Districting Criteria)” (hereinafter “Simulated Plan 1”), which was introduced as Petitioners’ Exhibit 3, results in only 14 counties being split into multiple congressional districts, as compared to the 28 counties that are split into multiple districts under the 2011 Plan. Id. at 173-74. Indeed, referring to a chart titled “Chen Figure 3: Simulation Set 1: 500 Simulated Plans Following Only Traditional Districting Criteria (No Consideration of Incumbent Protection),” which was introduced as Petitioners’ Exhibit 4, Dr. Chen explained that the maximum number of split counties in any of the 500 Simulation Set 1 plans is 16, and, in several instances, is as few as 11. Id. at 179. The vast majority of the Simulation Set 1 plans have 12 to 14 split counties. Id.

With respect to splits between municipalities, Dr. Chen observed that, under the 2011 Plan, there are 68 splits, whereas the range of splits under the Simulation Set 1 plans is 40 to 58. Id. at 180; Petitioners’ Exhibit 4. Based on the data contained in Petitioners’ Exhibit 4, Dr. Chen noted that the 2011 Plan “splits significantly more municipalities than would have resulted from the simulated plans following teadi-tional districting criteria, and [it] also split significantly more counties.” N.T. Trial, 12/11/17, at 180. He concluded that the evidence demonstrates that the 2011 Plan “significantly subordinated the traditional districting criteria of avoiding county splits and avoiding municipal splits. It shows us that the [2011 Plan] split far more counties, as well as more municipalities, than the sorts of plans that would have arisen under a districting process following traditional districting principles in Pennsylvania.” Id. at 181.

In terms of geographic compactness, Dr. Chen explained that he compared Simulated Plan 1 to the 2011 Plan utilizing two separate and widely-accepted standards. First, Dr. Chen calculated the Reock Compactness Score, which is a ratio of a particular district’s area to the area of the smallest bounding circle that can be drawn to completely contain the district — the higher the score, the more compact the district. Id. at 175. The range of Reock Compactness Scores for the congressional districts in Simulated Set 1 was “about .38 to about .46,” id. at 182, and Simulated Plan 1 had an average Reock Compactness Score range of .442, as compared to the 2011 Plan’s score of .278, revealing that, according to Dr. Chen, the 2011 Plan “is significantly less compact” than Simulated Plan 1. Id. at 175.

Dr. Chen also calculated the Popper-Polsby Compactness Score of both plans. The Popper-Polsby Compactness Score is calculated by first measuring each district’s perimeter and comparing it to the area of a hypothetical circle with that same perimeter. The ratio of the particular district’s area to the area of the hypothetical circle is its Popper-Polsby Compactness Score — the higher the score, the greater the geographic compactness. Id. at 176-77. The range of Popper-Polsby Compactness Scores for congressional districts in the Simulated Set 1 plans was “about .29 up to about .35,” id. at 183, and Simulated Plan 1 -had an average Popper-Polsby Score of .310, as compared to the 2011 Plan’s score of .164, again leading Dr. Chen to conclude that “the enacted map is significantly far less geographically compact” than Simulated Plan 1. Id. at 177.

Utilizing a, chart showing the mean Popper-Polsby Compactness Score- and the mean Reock Compactness Score for each of the 500 Simulation Set 1 plans, as compared to the 2011 Plan, see Petitioners’ Exhibit 5 (“Chen Figure 4: Simulation Set 1: 500 Simulated Plans Following Only Traditional Districting Criteria (No Consideration of Incumbent Protection)”), Dr. Chen opined that “no matter which measure of compactness you use, it’s very clear that the [2011 Plan] significantly and completely sacrifice[s] the traditional district-ing principle of geographic compactness compared to the sorts of plans that would have emerged under traditional districting principles.” N.T. Trial, 12/11/17, at 184.

Dr. Chen next addressed the 500 Simulation Set 2 Plans, which, as noted above, included the additional criteria of protecting the 17 incumbents. Dr. Chen stated that, in establishing the additional criteria, no consideration was given to the identities or party affiliations of the incumbents. Id. at 208. One of the Simulation Set 2 plans, “Chen Figure 1A: Example of a Simulated Districting Plan from Simulation Set 2 (Adhering to Traditional Districting Criteria And Protecting 17 Incumbents)” (hereinafter "“Simulated Plan l'A”), which was introduced’ as Petitioners’ Exhibit 7, resulted in only 15 counties being split into multiple congressional districts, ■ as compared to the 28 counties that are split into multiple districts under the 2011 Plan. Id. at 213. Referring to Petitioners’ Exhibit 8, titled “Chen Figure- 6: Simulation Set 2: 500 Simulated Plans Following Traditional Districting Criteria and Protecting 17 Incumbents/’ Dr. Chen further observed that the 2011 Plan split more municipalities (68) than any of the Simulated Set 2 plans, which resulted in a range of splits between 50 and 66. Based on this data, Dr. Chen opined:

We’re able to conclude from [Petitioners’ Exhibit 8] that the [2011 Plan] subordinate[s] the traditional districting criteria of avoiding county splits and avoiding municipal splits and the subordination of those criteria was not somehow justified or explained or warranted by an effort to protect 17 incumbents in an nonpartisan manner. To put that in -layman’s terms, an effort to protect incumbents would not have justified splitting up as many counties and as many municipalities as we saw split up, in the [2011 Plan].

Id. at 217.

With respect to geographic compactness, Dr. Chen explained that Simulated Plan 1A had an average Reock Compactness Score of .396, as compared to the 2011 Plan’s score of .278, and Simulated Plan 1A had a Popper-Polsby Compactness Score of .273, as compared to the 2011 Plan’s score of .164. Id. at 214; Petitioners’ Exhibit 7. Based on an illustration of the mean Popper-Polsby Compactness Score and the mean Reock Compactness Score for each of the 500 Simulation Set 2 plans, as compared to the 2011 Plan, see Petitioners’ Exhibit 9 (“Chen Figure 7: Simulation Set 2: 500 Simulated Plans Following Tra-ditio'nal'Districting Criteria and Protecting 17 Incumbents”), Dr. Chen concluded that the 2011 Plan “significantly subordinated' [the] traditional districting criteria of geographic compactness and that subordination of geographic compactness of districts was not somehow justified or necessitated or explained by a hypothetical effort to protect 17 incumbents.” N.T. Trial, 12/11/17, at 220.

Dr. Chen also testified regarding the partisan breakdown of the 2011 Plan. Dr. Chen explained that he requested and obtained from the Department of State the actual election data for each voting precinct in Pennsylvania for the six 2008 and 2010 statewide ■ elections. Id. at 185-86.

Those elections included the elections for the President, Attorney General, Auditor General, and State Treasurer in 2008, and the United States Senate election and the state gubernatorial election in 2010. Id. at 187. The election data obtained by Dr. Chen indicated how many votes were cast for each party candidate. Id. at 189. By overlaying the precinct-level election results on top of the geographic boundaries as shown on a particular map, he was able to determine whether a particular district had more Republican or Democratic votes during the elections. Id. at 196-97. Those districts that had more Republican votes would, naturally, be classified as Republican.

Dr. Chen.observed that, under the 2011 Plan, 13 of the 18 congressional districts are classified as Republican. Id. at 198. However, when Dr. Chen overlaid the precinct-level election results on Simulated Plan 1, only 9 of the 18 congressional districts would be classified as Republican. Id. at 197. Indeed, in the 500 Simulation Set 1 plans, the highest number of classified Republican districts was 10, and in none of the simulated plans would 13 of the congressional districts be classified as Republican. Id. at 200. Based on this data, Dr. Chen stated “I’m able to conclude with well-over 99.9 percent statistical certainty that the [2011 Plan’s] creation of a 13-5 Republican advantage in Pennsylvania’s Congressional delegation is an outcome that would never have emerged from a districting process adhering to and following traditional districting principles.” Id. at 203-04.

Moreover, Dr. Chen testified that, even under the Simulation Set 2 plans, which took into account preservation of incumbent candidates, none of the 500 plans resulted in a Republican District/Democratic District ratio of more than 10 to 8. Id, at 221-22; Petitioners’ .Exhibit 10. Based on a comparison of the 2011 Plan and .his simulated redistricting plans, Dr. Chen determined that “partisan intent predominated the drawing of the [2011 Plan] ... and the [2011 Plan] was drawn with a partisan intent to create a 13-5 Republican advantage and that this partisan intent subordinated traditional districting principles in the drawing of the enacted plan.” Id. at 166.

Dr. Chen was asked to consider, whether the partisan breakdown of the 2011 Plan might be the result of a “hypothetical effort to produce a certain racial threshold of having one district of over a 56.8 percent African-American voting-age population.” Id. at 245,46 To answer this question, Dr. .Chen explained that he analyzed the 259 computer-simulated plans from Simulation Sets 1 and 2 that included a congressional voting district with an African-American voting age population of at least 56.8%. Dr. Chen testified that, of those 259 simulated plans, none resulted in a Republican-Democrat congressional district ratio of 13 to 5. Id. at 244-45, 250. Indeed, of the Simulated Set 1 plans, which did not take into account protection of incumbents, the maximum ratio was 9 to 9, and of the Simulated Set 2 plans, which did protect incumbents, the maximum ratio was 11 to 8, and, in one case, was as low as 8 to 11. Id.; Petitioners’ Exhibit Í5 (“Chen Figure 10”). Dr. Chen concluded “the 13-5 Republican advantage of the enacted map is an outcome that is not plausible, even if one is only interested in plans that create one district with over 56.8 percent African-American voting-age population.” N.T. Trial, 12/11/17, at 245.

Dr. Chen also was asked whether the 13-5 Republican advantage in the 2011 Plan could be explained by political geography — that is, the geographic patterns of political behavior. Id. at 251. Dr. Chen explained that political geography can create natural advantages for one party over another; for example, he observed that, in Florida, Democratic voters are often “far more geographically clustered in urban areas,” whereas Republicans “are much more geographically spaced out in rural parts” of the state, resulting in a Republican advantage in control over districts and seats in the state legislature. Id. at 252-53.

In considering the impact of Pennsylvania’s political geography on the 2011 Plan, Dr. Chen explained that he measured the partisan bias of the 2011 Plan by utilizing a common scientific measurement referred to as the mean-median gap. Id. at 257. To calculate the mean, one looks at the average vote share per party in a particular district. Id. To calculate the median, one “line[s] up” the districts from the lowest to the highest vote share; the “middle best district” is the median. Id. at 258. The median district is the district that either party has to win in order to win the election. Id. Dr. Chen testified that, under the 2011 Plan, the Republican Party has a mean vote share of 47.5%, and a median vote share of 53.4%. Id. at 261; Petitioners’ Exhibit 1, at 20. This results in a mean-median gap of 5.9%, which, according to Dr. Chen, indicates that, under the 2011 Plan, “Republican votes ... are spread out in a very advantageous manner so as to allow — in a way that would allow the Republicans to more easily win that median district.” N.T. Trial, 12/11/17, at 259. The converse of this mean-median gap result is that Democratic voters “are very packed into a minority of the districts, which they win by probably more comfortable margins,” which makes it “much harder for Democrats under that scenario to be able to win the median district. So, in effect, what that means is it’s much harder for the Democrats to be able to win a majority of the Congressional delegation.” Id. at 260.

Dr. Chen recognized that “Republicans clearly enjoy a small natural geographic advantage in Pennsylvania because of the way that Democratic voters are clustered and Republican voters are a bit more spread out across different geographies of Pennsylvania.” Id. at 255, However, Dr. Chen observed that the range of mean/median gaps created in any of the Simulated Set 1 plans was between “a little over 0 percent to the vast majority of them being under 3 percent,” with a maximum of 4 percent. Id. at 262-63; Petitioners’ Exhibit 16 (“Chen Figure 5”). Dr. Chen explained that this is a “normal range,” and that a 6% gap “is a very statistically extreme outcome that cannot be explained by voter geography or by traditional districting principles alone.” N.T. Trial, 12/11/17, at 263-64. Dr. Chen noted that the range of mean/median gaps created by any of the Simulated Set 2 plans also did not approach 6%, and, thus, that the 2011 Plan’s “extreme partisan skew of voters is not an outcome that naturally emerges from Pennsylvania’s voter geography combined with traditional districting principles and an effort to protect 17 incumbents in a nonpartisan manner. It’s not a plausible outcome given those conditions.” Id. at 266; Petitioners’ Exhibit 17 (“Chen Figure 9”).

In sum, Dr. Chen “statistically conclude[d] with extremely high certainty ... that, certainly, there is a small geographic advantage for the Republicans, but it does not come close to explaining the extreme 13-5 Republican advantage in the [2011 Plan].” N.T. Trial, 12/11/17, at 255-56.

Ultimately, the Commonwealth Court found Dr. Chen’s testimony credible; specifically, the court held that Dr. Chen’s testimony “established that the General Assembly included factors other than nonpartisan traditional districting criteria in creating the 2011 Plan in order to increase the number of Republican-leaning congressional voting districts.” Findings of Fact at ¶ 309. The court noted, however, that Dr. Chen’s testimony “failed to take into account the communities of interest when creating districting plans,” and “failed to account for the fact that courts have held that a legislature may engage in some level of partisan intent when creating redistricting plans.” Id. at ¶¶ 310, 311.

Dr. John Kennedy

Petitioners next presented the testimony of Dr. John Kennedy, an expert in the area of political science, specializing in the pol