Whitford, William et al v. Nichol, Gerald et al
Filing
166
OPINION and ORDER. Signed by Judges Kenneth F. Ripple, Barbara B. Crabb and William C. Griesbach. Signed by District Judge Barbara B. Crabb on 11/21/2016. (voc)
IN THE UNITED STATES DISTRICT COURT
FOR THE WESTERN DISTRICT OF WISCONSIN
----------------------------------------------------------------WILLIAM WHITFORD, ROGER ANCLAIM,
EMILY BUNTING, MARY LUNNE
DONOHUE, HELEN HARRIS, WAYNE
JENSEN, WENDY SUE JOHNSON, JANET
MITCHELL, ALLISON SEATON, JAMES
SEATON, JEROME WALLACE AND
DONALD WINTER,
OPINION and ORDER
Plaintiffs,
15-cv-421-bbc
v.
BEVERLY R. GILL, JULIE M. GLANCEY,
ANN S. JACOBS, STEVE KING, DON MILLIS,
and MARK L. THOMSEN,
Defendants. 1
-----------------------------------------------------------------
As of June 30, 2016, the Government Accountability Board, the entity previously tasked with
administering and enforcing Wisconsin’s election laws, ceased to exist and its authority, as it relates to
this action, was transferred to the newly formed Elections Commission. See Wis. Act 118, 2015 Wis. Legis.
Serv. 1104 (West). Prior to this transfer, the parties filed a stipulated motion to substitute the members of
the Elections Commission for the members of the Government Accountability Board. We grant that
motion and have substituted as defendants the members of the Elections Commission in their official
capacity.
1
2
OPINION
Before RIPPLE, Circuit Judge, and CRABB and GRIESBACH, District Judges.
TABLE OF CONTENTS
I.
II.
III.
IV.
V.
VI.
VII.
Background .................................................................................................................3
A. Reapportionment in Wisconsin .........................................................................3
B. Drafting of Act 43 .................................................................................................7
C. Prior Court Challenges to Act 43 ..................................................................... 16
Procedural History .................................................................................................. 16
A. Allegations of the Complaint ........................................................................... 16
B. Motion to Dismiss .............................................................................................. 19
C. Motion for Summary Judgment....................................................................... 20
D. Witnesses Testifying at Trial ............................................................................ 21
E. Post-Trial Briefing .............................................................................................. 28
The Legal Landscape ............................................................................................... 29
A. The Foundational Case Law ............................................................................. 31
B. Present Supreme Court Precedent................................................................... 34
Elements of the Cause of Action ............................................................................ 55
A. Discriminatory Intent or Purpose .................................................................... 56
B. Discriminatory Effect of Act 43 ........................................................................ 74
Justification .............................................................................................................. 90
Standing ....................................................................................................................111
Order .........................................................................................................................115
A. Remedy ...............................................................................................................115
B. Evidentiary Matters ..........................................................................................116
Appendices
RIPPLE, Circuit Judge. The plaintiffs have brought this action alleging that Act 43,
the redistricting plan enacted by the Wisconsin Legislature in 2011, constitutes an
unconstitutional partisan gerrymander. Specifically, they maintain that the
Republican-controlled legislature drafted and enacted a redistricting plan that
systematically dilutes the voting strength of Democratic voters statewide. We find that
Act 43 was intended to burden the representational rights of Democratic voters
throughout the decennial period by impeding their ability to translate their votes into
legislative seats. Moreover, as demonstrated by the results of the 2012 and 2014
elections, among other evidence, we conclude that Act 43 has had its intended effect.
3
Finally, we find that the discriminatory effect is not explained by the political
geography of Wisconsin nor is it justified by a legitimate state interest. Consequently,
Act 43 constitutes an unconstitutional political gerrymander. This opinion constitutes
our findings of fact and conclusions of law pursuant to Federal Rule of Civil Procedure
52(a)(1).
I
BACKGROUND 2
We begin our consideration of the plaintiffs’ claims by examining Wisconsin’s
statutory requirements for redistricting as well as its recent redistricting history.
A.
Reapportionment in Wisconsin
1.
The State’s constitutional and statutory framework
Reapportionment of state legislative districts is a responsibility constitutionally
vested in the state government. See, e.g., Growe v. Emison, 507 U.S. 25, 34 (1993) (citing
U.S. Const. art I., § 2); Chapman v. Meier, 420 U.S. 1, 27 (1975). Although some states have
chosen to avoid the problem of partisan gerrymandering by vesting this power in a
neutral body designed specifically to perform that delicate function, see Arizona State
Legislature v. Arizona Indep. Redistricting Comm’n, 135 S. Ct. 2652, 2661–62 & n.6 (2015),
the people of Wisconsin have so far chosen to rely on its legislature to reapportion its
districts after the decennial census. They have vested responsibility in the bicameral
legislature composed of the Wisconsin State Senate and the Wisconsin State Assembly.
Wis. Const. art. IV, §§ 1, 3. According to Wisconsin law, “[t]he state is divided into 33
senate districts, each composed of 3 assembly districts. Each senate district shall be
entitled to elect one member of the senate. Each assembly district shall be entitled to
elect one representative to the assembly.” Wis. Stat. § 4.001.
The Wisconsin Constitution directs the Wisconsin legislature, “[a]t its first
session after each enumeration made by the authority of the United States,” to
“apportion and district anew the members of the senate and assembly, according to the
number of inhabitants.” Wis. Const. art. IV, § 3. The Wisconsin Constitution also
imposes specific requirements for reapportionment plans. Assembly districts are “to be
bounded by county, precinct, town or ward lines, to consist of contiguous territory and
be in as compact form as practicable.” Id. § 4. With respect to political subdivisions, a
The reader will notice on occasion some repetition and cross-reference. We think this is necessary to
ensure ease of comprehension to the first-time reader.
2
4
prior federal district court observed that, “[a]lthough avoiding the division of counties
is no longer an inviolable principle, respect for the prerogatives of the Wisconsin
Constitution dictate that wards and municipalities be kept whole where possible.”
Baumgart v. Wendelberger, Nos. 01-C-0121 & 02-C-0366, 2002 WL 34127471, at *3 (E.D.
Wis. May 30, 2002), amended by 2002 WL 34127473 (E.D. Wis. July 11, 2002). The
Wisconsin Constitution further requires that “no assembly district shall be divided in
the formation of a senate district.” Wis. Const. art. IV, § 5.
In addition to the state constitutional requirements, the Wisconsin legislature
must comply with federal law when redistricting. In particular, state legislatures must
ensure that districts are approximately equal in population, so that they do not violate
the “one-person, one-vote” principle embedded in the Equal Protection Clause of the
Fourteenth Amendment. See Reynolds v. Sims, 377 U.S. 533, 568 (1964) (“[T]he Equal
Protection Clause requires that the seats in both houses of a bicameral state legislature
must be apportioned on a population basis.”); see also Brown v. Thomson, 462 U.S. 835,
842–43 (1983) (holding “that an apportionment plan with a maximum population
deviation under 10%” is presumptively constitutional, while a population deviation
larger than 10% must be justified by the state); Harris v. Arizona Indep. Redistricting
Comm’n, 136 S. Ct. 1301, 1306–07 (2016) (same). Further, states also must comply with
§ 2 of the Voting Rights Act of 1965, which focuses on preserving the voting power of
minority groups. 52 U.S.C. § 10301; see also Thornburg v. Gingles, 478 U.S. 30, 47 (1986).
Redistricting laws in Wisconsin are enacted, in large measure, in the same
manner as other legislation, specifically, by way of bills originating in either house of
the legislature, see Wis. Const. art. IV, § 19. Tad Ottman, aide to the Senate Majority
Leader, explained in some detail this legislative process:
[L]egislators will work either on their own or with drafters or with a small
group of people to develop legislation. Usually it’s developed among
members of your own party, if not just the individual legislator. They
create a proposal with the assistance of the Legislative Reference Bureau.
At that point, the bill is often, but not always, circulated among other
legislators to see if anybody else would want to sign on ….[3]
The bill is then circulated. At some point it is introduced. … And
then once they are introduced, they are assigned to a committee. The
committee chairman or chairwoman can choose to hold a public hearing
on that piece of legislation. Most of the time a public hearing is held. …
Ottman noted that, regardless of the political party in power, “[t]ypically the first time the minority
party, and frankly most of the majority party sees legislation, is when a bill is circulated.” R.148 at 53.
3
5
And then that legislation is forwarded to the full body, either the Senate or
the Assembly, for debate and then it is passed over to the other House
where a similar process occurs.[4]
A bill must then “be presented to the governor,” who can sign or veto the bill. Wis.
Const. art. V, § 10.
The caucus system plays a significant role in the legislative process. 5 Caucus
meetings are held in the morning prior to the legislative session to vet legislation
internally before a vote on the floor.6 Professor William Whitford, a named plaintiff and
retired professor of law from the University of Wisconsin, testified that important
“debate and discussion,” as well as the “vote[] that matters,” occur within the caucus
meetings. 7 “Once the party caucuses come to a majority result, the other members of the
party are expected to follow the party line … .” 8 Thus, it is “extremely difficult” to pass
legislation through a bipartisan coalition. 9
2.
The modern history of reapportionment in Wisconsin
In the wake of the 1980 census, the plan that had been enacted in 1972 could no
longer satisfy the constitutional requirement of “one-person one-vote.” See Wis. State
AFL-CIO v. Elections Bd., 543 F. Supp. 630, 631 (E.D. Wis. 1982). In response to these
changes in population, a redistricting plan was drafted and enacted by the Wisconsin
legislature, which had a Democratic majority, but it was vetoed by the Republican
governor. Consequently, a federal district court was asked to devise a remedy. See id. at
632–33. Upon reviewing several plans submitted by legislators and interest groups, the
court “reluctantly concluded” that it could “be more faithful to the goals of
reapportionment” by drafting its own plan. Id. at 637. In doing so, the court focused on
ensuring population equality, avoiding the dilution of racial minority voting strength,
and keeping communities of interest together. Id. at 637–39. This “AFL-CIO Plan”
remained in effect for one election in 1982. As a result of that election, the Democratic
4
Id. at 51–52.
Pursuant to Federal Rule of Evidence 201(b)(2), we take judicial notice of the legislative process in
Wisconsin as set forth in Office of the Chief Clerk of the Wisconsin State Assembly, How a Bill Becomes
Law (2016), https://legis.wisconsin.gov/assembly/acc/media/1106/howabillbecomeslaw.pdf.
5
6
See id.
7
R.147 at 33.
8
Id.
9
Id.
6
Party held control of both houses of the Wisconsin legislature and also gained the
governor’s office. 10 The legislature passed, and the governor signed, a new
apportionment plan that lasted for the rest of the decennial period. See 1983 Wis. Sess.
Laws 633.
Following the 1990 election, the Wisconsin government again was divided
between two political parties. See Prosser v. Elections Bd., 793 F. Supp. 859, 862 (W.D.
Wis. 1992). The Democratic Party controlled both houses of the Wisconsin legislature
while the governor was a Republican. Id. “For that or other reasons, no bill to
reapportion the legislature had been enacted into law” by January 1992, leading several
Republican legislators to challenge the existing apportionment plan “as
unconstitutional and violative of the Voting Rights Act.” Id. As a result, the federal
court was asked to draft a new plan.
In an attempt to play a more limited role in the redistricting process, the court
“asked the parties at the outset whether they had any objection … to [the court’s]
selecting the best of the submitted plans rather than trying to create [its] own plan.” Id.
at 865 (emphasis removed). Upon receiving these submissions, however, the court
determined that the plans bore “the marks of their partisan origins.” Id. at 865. It
therefore used parts of one Republican plan and one Democratic plan. The court plan
preserved the strengths of the partisan plans, “primarily population equality and
contiguity and compactness,” while “avoid[ing] their weaknesses.” Id. at 870. The plan
remained in effect through the 2000 election.
Following the 2000 census, a divided Wisconsin legislature again was unable to
agree upon a redistricting plan. Arrington v. Elections Bd., 173 F. Supp. 2d 856, 862 (E.D.
Wis. 2001). In an ensuing law suit, the federal district court determined that “the
existing Wisconsin Assembly and Senate districts,” which had not been redrawn since
1992, were “violative of the ‘one person, one vote’ standard.” Baumgart, 2002 WL
34127471, at *1. A new plan was therefore necessary. The court considered sixteen plans
that had been submitted by legislators and other interest groups, but “found various
unredeemable flaws” in all of them. Id. at *6. The court therefore drew a plan “in the
most neutral way it could conceive—by taking the 1992 reapportionment plan as a
template and adjusting it for population deviations.” Id. at *7. In making these changes,
the court attempted to “maintain[] municipal boundaries and unit[e] communities of
interest.” Id. The “Baumgart Plan” was in effect from 2002 until 2010.
10
Tr. Ex. 211, at 3.
7
B.
Drafting of Act 43
In 2010, for the first time in over forty years, the voters of Wisconsin elected a
Republican majority in the Assembly, a Republican majority in the Senate, and a
Republican Governor. This uniformity in control led the Republican leadership to
conclude that a legislatively enacted redistricting plan was possible.
In January 2011, Scott Fitzgerald, Wisconsin Senate Majority Leader, and
Jeff Fitzgerald, Speaker of the Wisconsin Assembly, retained attorney Eric McLeod and
the law firm of Michael Best & Friedrich, LLP, to assist with the reapportionment of the
state legislative districts. 11 The firm supervised the work of Tad Ottman, staff member
to Senate Majority Leader Fitzgerald; Adam Foltz, staff member to Speaker Fitzgerald;
and Joseph Handrick, a consultant with the law firm Reinhart Boerner Van Deuren s.c.,
in planning, drafting, and negotiating the new districting plan. Ottman, Foltz, and later
Handrick, worked in a room located in the offices of Michael Best & Friedrich, which
they referred to as the “map room.” 12
Ottman, Foltz, and Handrick also received assistance from Professor
Ronald Keith Gaddie, a professor of political science at the University of Oklahoma.
Michael Best & Friedrich had retained Professor Gaddie “as an independent advisor on
the appropriate racial and/or political make-up of legislative and congressional districts
in Wisconsin.” 13 Professor Gaddie described his job as “devis[ing] measures and
consult[ing] … about measures” of partisanship, compactness, “the integrity of
counties, the integrity of city boundaries, the so-called good government principles of
redistricting.” 14 “Where [he] … spent most of [his] time was trying to disentangle the
performance of the majority/minority districts in Milwaukee County.” 15
A “significant part” of his work was “building a regression model to be able to
test the partisan makeup and performance of districts as they might be configured in
different ways.” 16 As explained by one of the plaintiffs’ experts, Professor
Kenneth Mayer, “[r]egression is a technique where we can seek to explain a dependent
variable, the variable that we’re trying to account for. … [W]e attempt to explain the
values that a dependent variable take[s] with what are called independent variables or
11
See R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 5, ¶ 20.
12
Id. at 7, ¶¶ 23, 26.
13
Tr. Ex. 169.
14
Tr. Ex. 161 (Gaddie Dep.), at 45.
15
Id. at 46.
16
Id.
8
underlying causal variables.” 17 In this instance, Professor Gaddie’s dependent variable
was the baseline partisanship of a unit of geography, which then could be aggregated
into different configurations of Assembly districts. In this way, Professor Gaddie was
able to assess the partisanship of the Assembly maps that the drafters passed on to him
for analysis. 18 Professor Mayer testified that “the political science literature is essentially
unanimous” that the approach taken by Professor Gaddie is “the appropriate
method,” 19 and Professor Mayer used the same methodology to construct his
Demonstration Plan. 20
Ottman, Foltz, and Handrick began drafting the map that would become Act 43
in April 2011, after they received census data from the Legislative Technology Services
Bureau (“LTSB”). 21 The LTSB also had provided them with computers loaded with the
redistricting software, autoBound. 22 Ottman described in detail how the software was
used:
[Y]ou would open up a plan that you’d been working on or label a new
plan and assign it the Assembly district that you wanted to work with and
then you could also pick a color that you wanted that Assembly district to
be. It’s sort of like a color-by-number exercise. …
You also determine what other layers that you want to look at on
the screen. There were a number of different overlays that you have,
anywhere from existing Senate and Assembly districts, … count[y]
boundaries, municipal boundaries, ward boundaries all the way down to
census block boundaries. As a practical matter what you tried to do is you
would zoom in the region of your screen to the area that you’re looking at
to the smallest amount that you could see and then have kind of the
fewest layers displayed that you would need because the more
information that you were requiring it to display slows down the
computer speed a lot and makes it really slow to render.
….
17
R.148 at 156–57.
18
See infra at 12–14.
19
R.148 at 152.
20
See infra at 107–08.
21
See R.148 at 68–69.
22
See id. at 62.
9
And then what you would do is there were a couple different ways
that you could add population to the district.[23]
Ottman further explained that, in more populated areas, the drafters worked more at
the ward level: “So you would have the wards displayed and you would literally draw
a circle, click on it, and it would assign it to the map and fill it in.” 24 “In other parts of
the state … you might do that at the county level because it’s so sparsely populated so
you’d grab three or four counties at [a] time.” 25
When the drafters would increase the area size of the districts that they were
drawing, autoBound provided demographic information for the area that the drafter
had included, such as the number of people in the district, the deviation from the ideal
population, voting-age population, and different minority group populations. 26 It also
allowed the user to include “customized … demographic data.” 27
One piece of “customized demographic data” employed by the drafters was a
composite partisan score. From the time that Ottman, Foltz, and Handrick received the
census data from the LTSB, they worked to develop a composite partisan score that
accurately reflected the political make-up of the population units. 28 Having this
measure was necessary so that, when they aggregated those units into new districts,
they could assess the partisan make-up of the new district they had drawn. On April 19,
2011, they developed a composite of “all statewide races from [20]04 to 2010” that
“seem[ed] to work well.” 29 They sent this composite measure to Professor Gaddie, who
tested it against his regression model. Professor Gaddie confirmed to Handrick that
“the partisanship proxy you are using (all races) is an almost perfect proxy for the open
23
Id. at 72–73.
24
Id. at 73.
Id. at 74. In the early stages, the drafters worked “almost exclusively in the City of Milwaukee.” Id. at
75. Ottman explained that
25
[w]e knew there were going to be more redistricting criteria, including … the
voting rights application that was going to apply there. … [W]e wanted to kind of get
those Milwaukee districts drawn in such a way that the lawyers advised us was kind of
in a good place and then we just kind of wanted to lock that in and leave it alone before
we drew the rest of the map.
Id. at 75.
26
See id. at 63–64.
27
Id. at 64.
28
See Tr. Ex. 175, at 1–2.
29
Id. at 2.
10
seat vote, and the best proxy you’ll come up with.” 30 Once Professor Gaddie confirmed
the usefulness of their composite measure, Ottman, Foltz, and Handrick could “assess
the partisan impact of the map[s] that [they] drew.” 31
Although Ottman, Foltz, and Handrick worked in the same room at Michael Best
& Friedrich, they worked independently on their own maps. They drew several
statewide maps, and even more regional maps from which the legislative leadership
eventually would choose. As they drew the maps, they would ensure that the districts
were “close-to-ideal population.” 32 They did an “eyeball test” for “compactness and
contiguousness.” 33 They “looked at … what the core of the existing district was
compared to the new district,” “looked at municipalities that were split,” whether the
new district had changed Senate districts, and “where incumbents lived.” 34
The drafters were attentive to traditional districting criteria like population
equality, compactness, and municipal splits throughout the drafting process. When the
drafters had created a statewide map with which they were satisfied, they would export
the district-by-district partisanship scores from autoBound into a spreadsheet for that
“finalized” “statewide” plan. 35
The drafters used their composite score to evaluate the statewide maps that they
had drawn based on the level of partisan advantage that they provided to Republicans.
In many instances, the names of the maps reflected the level of partisan advantage
achieved by the districting plan; for instance, there are maps labeled “Assertive” and
30
Id. at 1.
See R.147 at 61; R.148 at 16–17. The drafters’ “partisanship proxy” and Professor Gaddie’s “open seat
vote” measure of partisanship, Tr. Ex. 175, at 1, correlated almost “identical[ly]” with the “open-seat
baseline model” that Professor Mayer developed by way of a regression analysis and that he used to
construct his Demonstration Plan. R.148 at 191; see also infra at 24.
31
32
R.148 at 83.
33
Id.
Id. at 85–86. Ottman testified that where incumbents lived “matter[s] because in the end this was a map
that we were going to ask the Legislature to vote for and we knew that that was one of the considerations
that was going to be very important to the people being asked to vote for this.” Id. at 86.
34
R.147 at 162 (Foltz testifying that once “you get a statewide plan finalized, all 99 assembly districts,”
“you can then take that composite column from auto[B]ound and then move it over into those Excel
spreadsheets that I was talking about earlier”); R.148 at 14 (Ottman testifying that the partisan
performance spreadsheet was “one of the reports that was generated on any statewide map that we laid
down”).
35
11
“Aggressive.” 36 Foltz testified that “aggressive” in this context meant “probably that
[the map] was a more aggressive map with regard to GOP leaning.” 37
The drafters created spreadsheets which collected the partisan scores, by district,
for each of the statewide map alternatives. Each spreadsheet included a corresponding
table comparing the partisan performance of the draft plan to the prior map drawn by
the Baumgart court, which they called the “Current Map.” These performance
comparisons were made on the following criteria: “Safe” Republican seats, “Lean”
Republican seats, “Swing” seats, “Safe” Democratic seats, and “Lean” Democratic
seats. 38
The process of drafting and evaluating these alternative district maps spanned
several months. In early April 2011, the drafters produced a document comparing the
partisan performance of the Current Map to two early draft maps: Joe’s Basemap Basic
and Joe’s Basemap Assertive. 39 Under the Current Map, the drafters anticipated that the
Republicans would win 49 Assembly seats. 40 This number increased to 52 under the
Joe’s Basemap Basic map and to 56 under the Joe’s Basemap Assertive map. 41 The
number of safe and leaning Republican seats increased from 40 under the Current Map
to 45 under the Joe’s Basemap Basic map and 49 under the Joe’s Basemap Assertive
map; the number of swing seats decreased from 19 to 14 to 12. 42 The number of safe and
See, e.g., R.148 at 30. During the drafting process, Ottman met with individual senators to review with
them the census numbers for their district, to verify their addresses, and to ask general questions about
their districts, such as “are there areas you like, are there areas you don’t like, are there areas surrounding
your district that you like.” Id. at 81. Ottman also received a few requests from Senators concerning their
districts. Senator Vukmir provided specific suggestions on how her district could be re-drawn to take the
seat away from a Democratic member of the Assembly: “If you need a way to take the Staskunas seat, put
a little bit of my Senate seat into New Berlin (2–3 wards could make that a GOP Assembly seat).” Tr. Ex.
239. However, because Senator Vukmir’s district encompassed Milwaukee, the drafters could not
implement the suggestion because “there was simply less flexibility in how [they] could draw that district
than in some other areas of the state.” R.148 at 82.
36
37
R.147 at 65.
38
See, e.g., Tr. Ex. 364; see also R.148 at 15 (Ottman testifying about these criteria).
39
Tr. Ex. 465; see also Tr. Ex. 476 (sorting districts by Republican vote share).
40
Tr. Exs. 465, 467.
41
Tr. Exs. 465, 467.
42
Tr. Ex. 465.
12
leaning Democratic seats, however, remained roughly the same under all three maps,
hovering between 38 and 40. 43
The drafters prepared and evaluated the partisan performance of at least another
six statewide alternative maps. 44 Each of these maps improved upon the anticipated
pro-Republican advantage generated in the initial two draft plans. The total number of
safe and leaning Republican seats now ranged between 51 and 54, and the number of
swing seats was decreased to between 6 and 11. 45 The number of safe and leaning
Democratic seats again remained about the same under each draft map, ranging
between 37 and 39. 46
The drafters sent their completed draft maps to Professor Gaddie for further
analysis. For each map, Professor Gaddie created an “S” curve—a “visual aide[] to
demonstrate the partisan structure of Wisconsin politics.” 47 These “S” curves show how
each map would operate within an array of electoral outcomes.48
The “S” curves give a visual depiction of how each party’s vote share (on the
x axis), ranging from 40% to 60%, relates to the number of Assembly seats that party
likely will secure (on the y axis). Democratic seats are depicted by shades of blue, and
Republican seats by shades of red. 49 To produce the “S” curves, Professor Gaddie first
used his regression analysis to calculate the expected partisan vote shares for each new
district. 50 He then shifted the vote share of each district ten points in either direction,
from 40% to 60%, and assigned a color to districts that “tend[ed]” towards, or were
“safe” seats, for that party.51 The “S” curves—at least some of which were printed in
large format and kept in the map room—allowed a non-statistician, by mere visual
inspection, to assess the partisan performance of a particular map under all likely
43
Id.
These were: Milwaukee_Gaddie_4_16_11_V1_B (Tr. Ex. 172, at 1); Statewide2_Milwaukee_Gaddie_
4_16_V1_B (Tr. Ex. 172, at 2); Tad MayQandD (Tr. Exs. 364, 477); Joe Assertive (Tr. Exs. 366, 478); Tad
Aggressive (Tr. Ex. 283); and Adam Aggressive (Tr. Ex. 283). See generally Tr. Ex. 225 (containing
metadata from the drafters’ computers).
44
45
Tr. Exs. 172, 364, 366.
46
Tr. Exs. 172, 364, 366.
47
Tr. Ex. 134; see Tr. Exs. 263–82.
48
Tr. Ex. 161 (Gaddie Dep.), at 44–45.
49
A partial “S” curve for the “Team Map,” see infra at 69, is attached as Appendix 1 to this opinion.
50
Tr. Ex. 161 (Gaddie Dep.), at 44–47; see also supra at 7–8.
51
Tr. Ex. 161 (Gaddie Dep.), at 128.
13
electoral scenarios. On one occasion, Senator Fitzgerald came to the map room, and
Professor Gaddie showed him one of the large printouts of the “S” curves and
“basically explain[ed] how to interpret” them. 52
Not long after Professor Gaddie had performed his analyses, the Republican
legislative leadership contacted the drafters and indicated that they wanted to be
prepared to act on a redistricting plan. Over several days in early June, the drafters
presented a selection of regional maps drawn from their statewide drafts,
approximately three to four per region, to the Republican leadership. Along with these
regional alternatives, the leadership “saw the partisan scores for the maps that [the
drafters] presented to them in those alternatives.” 53 Foltz testified during his deposition
that, although he could not recall a particular example, he was sure that he was asked
by the leadership about the partisan performance of the various regional options. 54
Following this meeting, the drafters amalgamated the regional alternatives
chosen by the leadership. Foltz testified that “the draft map called team map emerged
as a result of the … leadership’s choices at those meetings.” 55 Under the Team Map,
which was also referred to as the “Final Map,” 56 the Republicans could expect to win 59
Assembly seats, with 38 safe Republican seats, 14 leaning Republican, 10 swing, 4
leaning Democratic, and 33 safe Democratic seats. 57 In a document bearing the heading
“Tale of the Tape,” 58 the drafters, among other things, compared the partisan
52
Id. at 75.
53
R.148 at 20.
54
Tr. Ex. 191, at 106.
55
R.147 at 80.
Foltz testified that if the “Team Map” “[wa]sn’t the final one that was pushed, put forward in the public
domain, it was very close to it, and it was the result of that mashing process of taking the various regional
alternatives and putting them all together.” Id. at 165; see also infra note 68 (discussing changes made after
the regional maps were amalgamated). He explained that the “Final Map” was the one “after the leaders
got together and made the regional decisions and they were then merged together.” R.147 at 62. If it was
not identical to the map that “ultimately became Act 43, it[ wa]s probably fairly close.” Id.; see also Tr. Ex.
172, at 3–4 (showing partisan performance of the Final Map).
56
The drafters in fact produced and evaluated several distinct versions of the Team Map, but each
rendition is virtually identical. See Tr. Ex. 172, at 3–4 (Final Map); Tr. Ex. 467, at 1 (Team Map (Joe
Aggressive)); id. at 2 (Team Map Ranking (Joe Aggressive 2)); id. at 3 (Team Map (6-15-11)).
57
It is unclear who titled this document “Tale of the Tape,” see R.148 at 33 (Ottman testifying that he “did
not create that title” and was “not sure what it signifies”); Ottman did testify, however, that he had
“heard the expression” as it “refer[red] … in boxing matches[] [to] pre-fight measurements of the boxer’s
reach,” id. at 34.
58
14
performance of the Team Map directly to the Current Map on each of these criteria. 59
They highlighted specifically that under the Current Map, 49 seats are “50% or better”
for Republicans, but under the Team Map, “59 Assembly seats are 50% or better.”60
The Team Map underwent even more intense partisan scrutiny in a document
identified as “summary.xlsx.” 61 The drafters divided the new Team Map districts into
six categories of partisan performance, listing beside each district its “new incumbent”
and its Republican vote share under the Current Map and the Team Map. 62 The drafters
considered five districts to be “Statistical Pick Up[s],” meaning they were currently held
by a Democratic incumbent but likely to become Republican; they grouped fourteen
districts under the heading “GOP seats strengthened a lot”; they designated eleven
districts “GOP seats strengthened a little”; they labeled three districts as “GOP seats
weakened a little”; they considered another three GOP districts “likely lost”; and,
finally, they identified four districts where the Democrats were “weakened.” 63 The
drafters also listed the twenty Republican Assembly members who, under the Team
Map, could be considered “GOP Donors to the Team”: “Incumbents with numbers
above 55% that donate[d] to the team.” 64 These representatives stood in contrast to
“GOP non-donors,” who were Republican incumbents with “over 55% who d[id] not
donate points.” 65
The Team Map was then sent to Professor Gaddie, who conducted an “S” curve
analysis. The Team Map demonstrated that Republicans would maintain a majority
under any likely voting scenario; indeed, they would maintain a 54 seat majority while
garnering only 48% of the statewide vote. The Democrats, by contrast, would need 54%
of the statewide vote to capture a majority. 66
Once the map had been finalized, Foltz presented each Republican member of
the Assembly with information on his or her new district. The memos prepared for the
Assembly members informed them whether the district number had changed, whether
adjustment to the district population was necessary based on the census numbers, and
59
Tr. Ex. 283.
60
Id.
61
Tr. Ex. 284, at 1.
62
Id.
63
Id.
64
Id.; see also infra note 221 (discussing meaning of “Donors to the Team”).
65
Tr. Ex. 283.
66
Tr. Ex. 282.
15
provided a “[c]omparison of [k]ey [r]aces” in the new district compared to the old.67
Specifically, the memorandum detailed what percentage of the population in the old
and new districts voted for Republican candidates in representative statewide and
national elections held since 2004. This information also was provided in terms of raw
votes. The memoranda did not provide the individual legislators with any information
about contiguity, compactness, or core population.
Ottman engaged in a similar process with Republican members of the State
Senate.68 For each meeting, he created a talking-points memo that included information
about population, where changes in the district’s population had occurred, and the
geography of the new district. 69 These also contained information on how the
re-configured district had voted in national and statewide elections. 70
Ottman also made a presentation to the Republican caucus. His notes for that
meeting state: “The maps we pass will determine who’s here 10 years from now,” and
“[w]e have an opportunity and an obligation to draw these maps that Republicans
haven’t had in decades.” 71
On July 11, 2011, the redistricting plan was introduced by the Committee on
Senate Organization. 72 On July 13, 2011, a public hearing was held, during which
Ottman and Foltz presented the plan and fielded questions. 73 The Senate and Assembly
passed the bill on July 19, 2011, and July 20, 2011, respectively. The Governor signed the
bill, and it was published as Wisconsin Act 43 on August 23, 2011. 74
67
Tr. Ex. 342.
Foltz and Ottman both testified that, as a result of these meetings, there were only slight changes made
to the Assembly map; specifically, in response to concerns articulated by Senator Mike Ellis, they “redrew
the Assembly boundaries a little bit” within his district. R.148 at 111; see also R.147 at 185–86. No other
changes were made in response to these meetings.
68
69
See, e.g., Tr. Ex. 242 at 1 (“Added East Troy and part of the town, as well as Mukwonago.”).
See id. at 1 (noting, for example, that “Scott Walker won this new seat with 64.2%,” “McCain won with
51.5%,” and “Van Hollen 06 won with 59.4%”).
70
71
Tr. Ex. 241, at 1.
72
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 24, ¶ 85.
73
Id. ¶ 86; Tr. Ex. 353.
74
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 24, ¶ 86.
Another important legislative measure, enacted around the time of the drafting and passage of
Act 43, bears mentioning. Act 39, enacted on July 25, 2011, and published on August 8, 2011, permits the
legislature to draw new districts before Wisconsin’s municipalities draw their ward lines. The
longstanding practice in Wisconsin had been that municipalities drew their ward lines first, and the
16
C.
Prior Court Challenges to Act 43
Even before Act 43 was passed, two actions were brought challenging the plan
on constitutional and statutory grounds, including under Section 2 of the Voting Rights
Act. See Baldus v. Members of the Wis. Gov’t Accountability Bd., 849 F. Supp. 2d 840, 846–47
(E.D. Wis. 2012). The court consolidated the actions for decision and concluded that the
plan did not violate the “one-person, one-vote” principle, nor did it violate the Equal
Protection Clause by “disenfranchise[ing]” voters who were moved to a new Senate
district and were unable to vote for their state senator for another two years. Id. at 849–
51, 852–53. However, the court did find that the plaintiffs were entitled to relief on their
claim that Act 43 violated the Voting Rights Act by diluting the voting power of Latino
voters in Milwaukee County, and it ordered the State to redraw these districts. Id. at
859. The remainder of Act 43, however, remained intact and governed the 2012 and
2014 Assembly elections.
In 2012, the Republican Party received 48.6% of the two-party statewide vote
share for Assembly candidates and won 60 of the 99 seats in the Wisconsin Assembly. 75
In 2014, the Republican Party received 52% of the two-party statewide vote share and
won 63 assembly seats. 76
II
PROCEDURAL HISTORY
A.
Allegations of the Complaint
We now turn to the dispute before this court. Plaintiffs William Whitford,
Roger Anclam, Emily Bunting, Mary Lynne Donohue, Helen Harris, Wayne Jensen,
Wendy Sue Johnson, Janet Mitchell, James Seaton, Allison Seaton, Jerome Wallace, and
Don Winter are United States citizens registered to vote in Wisconsin. They reside in
various counties and legislative districts throughout Wisconsin. All of them are
“supporters of the Democratic party and of Democratic candidates and they almost
legislature drew districts based on the new wards. See R.148 at 123–24 (Ottman testifying to same).
Following Act 39’s passage, wards are drawn in response to the districts, rather than the other way
around. In the absence of Act 39, the legislature would have had to postpone its drafting effort by several
months until the municipalities adopted their ward boundaries.
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 70, ¶ 289; id. at 69, ¶ 285.
These percentages are based on the calculations of plaintiffs’ expert, Professor Simon Jackman. The
defendants have not contested the accuracy of these figures, and we accept them as accurate.
75
76
Id. at 70, ¶ 290.
17
always vote for Democratic candidates in Wisconsin elections.” 77 Defendants are
Beverly R. Gill, Julie M. Glancey, Ann S. Jacobs, Steve King, Don Millis, and
Mark L. Thomsen, each in his or her official capacity as a member of the Wisconsin
Elections Commission.
According to the plaintiffs, in drafting Act 43, the Republicans employed two
gerrymandering techniques: “cracking”—“dividing a party’s supporters among
multiple districts so that they fall short of a majority in each one”—and “packing”—
“concentrating one party’s backers in a few districts that they win by overwhelming
margins,” 78 in order to dilute the votes of Democrats statewide. This “cracking and
packing result[ed] in ‘wasted’ votes: votes cast either for a losing candidate (in the case
of cracking) or for a winning candidate but in excess of what he or she needs to prevail
(in the case of packing).” 79 They therefore urge the court to adopt a new measure for
assessing the discriminatory effect of political gerrymanders—the efficiency gap (or
“EG”). “The efficiency gap is the difference between the parties’ respective wasted votes
in an election, divided by the total number of votes cast.” 80 When two parties waste
votes at an identical rate, a plan’s EG is equal to zero. An EG in favor of one party,
however, means that the party wasted votes at a lower rate than the opposing party. It
is in this sense that the EG arguably is a measure of efficiency: Because the party with a
favorable EG wasted fewer votes than its opponent, it was able to translate, with greater
ease, its share of the total votes cast in the election into legislative seats. In short, the
complaint alleges that Act 43 purposely distributed the predicted Republican vote share
with greater efficiency so that it translated into a greater number of seats, while
77
Id. at 3, ¶ 2.
78
R.1 at 3, ¶ 5; see also infra at 81.
R.1 at 3, ¶ 5. “Wasted” is merely a term of art used to describe votes cast for losing candidates and votes
cast for winning candidates in excess of 50% plus one; the term is not meant to suggest that the votes are
worthless. See infra note 267 and accompanying text.
79
80
R.1 at 3, ¶ 5; see also infra at 81. The plaintiffs provided the following example:
Suppose … that there are five districts in a plan with 100 voters each. Suppose also that
Party A wins three of the districts by a margin of 60 votes to 40, and that Party B wins
two of them by a margin of 80 votes to 20. Then Party A wastes 10 votes in each of the
three districts it wins and 20 votes in each of the two districts it loses, adding up to 70
wasted votes. Likewise, Party B wastes 30 votes in each of the two districts it wins and 40
votes in each of the three districts it loses, adding up to 180 wasted votes. The difference
between the parties’ respective wasted votes is 110, which, when divided by 500 total
votes, yields an efficiency gap of 22% in favor of Party A.
R.1 at 15, ¶ 50.
18
purposely distributing the Democratic vote share with less efficiency so that it would
translate into fewer seats.
The plaintiffs’ complaint incorporated the EG into a proposed three-part test for
partisan gerrymandering. First, plaintiffs would have to establish that a State had an
intent to gerrymander for partisan advantage. Second, the plaintiffs would need to
prove a partisan effect, by proving that the EG for a plan exceeds a certain numerical
threshold (which the plaintiffs proposed, based on historical analysis, to be 7%). 81 If a
plan exceeds that threshold, the plaintiffs asserted that it should be presumptively
unconstitutional. Third, and finally, the plaintiffs placed the burden on the defendants
to rebut the presumption by showing that the plan “is the necessary result of a
legitimate state policy, or inevitable given the state’s underlying political geography.” 82
If the state is unable to rebut the presumption, then the plan is unconstitutional.
The plaintiffs alleged that they had satisfied all of these elements. According to
the complaint, Act 43 “was drafted and enacted with the specific intent to maximize the
electoral advantage of Republicans and harm Democrats to the greatest possible
extent.”83 Additionally, Act 43 “produced a pro-Republican efficiency gap of 13% in
2012 and 10% in 2014.” 84 They further claimed that this EG is unjustified because one of
their experts, Professor Mayer, had crafted a “Demonstration Plan” with “an efficiency
gap of just 2% in 2012,” which “perform[ed] at least as well as [Act 43] on every other
relevant metric.” 85
For these reasons, plaintiffs claimed that Act 43 “treats voters unequally, diluting
their voting power based on their political beliefs, in violation of the Fourteenth
Amendment’s guarantee of equal protection,” and “unreasonably burdens their First
Amendment rights of association and free speech.” 86 They requested a declaration that
Act 43 is unconstitutional, an injunction prohibiting further elections under the map,
and the drawing of a new redistricting map. 87
The plaintiffs’ proposed threshold is based on the analysis of one of their experts, Professor Simon
Jackman. See infra at 25–27, 83–84.
81
82
R.1 at 25, ¶ 84.
83
Id. at 9, ¶ 31.
84
Id. at 16, ¶ 55.
85
Id. at 23–24, ¶¶ 78–79 (emphasis removed); see also infra at 24.
86
R.1 at 2, ¶ 2.
87
Id. at 29, ¶¶ 97–99.
19
B.
Motion to Dismiss
The defendants filed a motion to dismiss on August 18, 2015, which contended
that the court could not grant relief for three primary reasons. First, the defendants
argued that the EG was directly analogous to the proportional-representation standard
rejected by the Supreme Court in Vieth v. Jubelirer, 541 U.S. 267, 287–88 (2004). 88 Second,
the defendants argued that the EG failed to account for the impact of traditional
districting criteria like contiguity and compactness. Finally, the defendants argued that
the plaintiffs lacked the standing to challenge Act 43 on a statewide basis, and instead
could only challenge their individual districts.
In an order dated December 17, 2015, we denied defendants’ motion to dismiss.
We first noted that the claim was justiciable, and that, “[u]ntil a majority of the
Supreme Court rules otherwise, lower courts must continue to search for a judicially
manageable standard.” 89 We acknowledged the defendants’ argument that the EG was
analogous to a proportionality standard, but noted that the plaintiffs’ experts disagreed
with the defendants’ contention and that factfinding therefore was needed. We
concluded that “[a] determination whether plaintiffs’ proposed standard is judicially
manageable relies at least in part on the validity of plaintiffs’ expert opinions” and that
a more developed record would be necessary to resolve that question. 90 Finally, we
concluded that the plaintiffs had standing, explaining that “[b]ecause plaintiffs’ alleged
injury in this case relates to their statewide representation, it follows that they should be
permitted to bring a statewide claim.” 91 We noted, however, that the defendants were
“free to raise this issue again on a more developed record.” 92
The defendants pointed to Professor Jackman’s report, which employs a “simplified method” for
calculating the EG:
88
EG = S − .5 − 2(V − .5)
R.34 at 18. In this equation, “S” is the party’s expected seat share and “V” is the party’s expected vote
share. The “simplified method” implies that for 1% of the vote a party obtains above 50%, the party
would be expected to earn 2% more of the seats (what is called a “winner’s bonus”). It is this direct
correlation between seat and vote share that, the defendants maintained, ran afoul of Vieth v. Jubelirer,
541 U.S. 267, 287–88 (2004).
89
R.43 at 9.
90
Id. at 23.
91
Id. at 13.
92
Id. at 15.
20
C.
Motion for Summary Judgment
Defendants subsequently filed a motion for summary judgment, raising new
challenges to the plaintiffs’ claims. 93 In the motion, the defendants argued that the EG
metric was overinclusive and captured several plans—including court-drawn plans in
Wisconsin—that were not drawn with any partisan intent. Furthermore Democratic
voters tended to live in cities, which created a “natural packing” effect and distorted the
EG. 94
The defendants acknowledged the plaintiffs’ argument that a requirement of
partisan intent could remedy this over-inclusivity problem, but noted that the intent
element was not sufficiently demanding. The defendants contended that “[a]s long as
redistricting is done by a legislature, it should not be very difficult to prove that the
likely political consequences of the reapportionment were intended.” 95 The intent
element proposed by the plaintiffs was, therefore, “meaningless,” 96 and the Supreme
Court’s decision in Vieth already had ruled out the more demanding standard of
“predominant intent.” See 541 U.S. at 284–86 (plurality opinion); id. at 308 (Kennedy, J.,
concurring).
The defendants levied two additional criticisms of the plaintiffs’ test. First, they
noted that the plaintiffs’ “Demonstration Plan” was based on a counterfactual scenario
and therefore failed to address concerns raised by some Justices about a standard which
dealt with a “hypothetical state of affairs.” 97 Second, they alleged that the EG is highly
sensitive to “vote-switchers” in swing districts. 98 Had voters in close (or competitive)
elections voted for the other party, and had a few candidates of the other party won
those seats, then the EG might be dramatically different. In their view, a plan that
included such competitive districts could be found unconstitutional under the plaintiffs’
proposed standard.
We denied the motion for summary judgment. We explained that judgment “as a
matter of law would be premature because there [we]re factual disputes regarding the
93
See R.45; R.46.
94
R.46 at 38.
Id. at 41 (quoting Davis v. Bandemer, 478 U.S. 109, 129 (1986) (plurality opinion) (alteration in original)
(internal quotation marks omitted)).
95
96
Id.
Id. at 48 (quoting League of United Latin American Citizens v. Perry (“LULAC”), 548 U.S. 399, 420 (2006)
(opinion of Kennedy, J.) (internal quotation marks omitted)).
97
98
Id.
21
validity of plaintiffs’ proposed measurement.” 99 We also noted that there was
conflicting evidence on the “natural packing” of Democrats in Wisconsin. 100 We further
observed that the defendants’ arguments might serve as “a suggestion to alter the
threshold of the plaintiffs’ test and, perhaps, shift the burdens of production or
proof.”101 In particular, we left open the question of the requisite level of intent and
directed the plaintiffs to “be prepared to present the strongest evidence that they have
on this issue … in order to meet even the most demanding intent requirement.” 102 We
therefore set the case for trial.
D.
Witnesses Testifying at Trial
During the four-day trial, from May 24, 2016, through May 28, 2016, the parties
presented their cases through eight witnesses. Some of the testimony of the witnesses
involved in the passage of Act 43 has been set forth above, so it is not necessary to
summarize it again here. An overview of the remaining testimony is set forth below. 103
1.
William Whitford
First to testify was William Whitford, one of the plaintiffs in this litigation and a
resident of the 76th Assembly District. 104 Professor Whitford testified to his long-time
affiliation with the Democratic Party. He related that he consistently has voted for
Democratic candidates, has made donations to Democratic Assembly candidates
outside of his own district, has raised money on their behalf, and has donated to the
Assembly Democratic Campaign Committee. 105 According to Professor Whitford, given
Wisconsin’s caucus system, “[t]he only practical way to accomplish [his] policy
objectives is to get a majority of the Democrats in the Assembly and the Senate,” which
is “virtually impossible under this apportionment [plan].” 106
99
R.94 at 2.
100
Id. at 14–17.
101
Id. at 16.
102
Id. at 30.
103
A more complete treatment of the experts’ opinions are set forth in Parts IV and V.
104
R.147 at 29.
105
Id. at 31, 39.
106
Id. at 33–34.
22
2.
Ronald Keith Gaddie
Professor Gaddie was deposed by the plaintiffs on March 9, 2016, and a video of
that deposition was admitted into evidence and played at trial. As explained in some
detail above, 107 Professor Gaddie testified that he was retained by Michael Best &
Friedrich on April 11, 2011, to “serv[e] as an independent advisor on the appropriate
racial and/or political make-up of legislative and congressional districts in
Wisconsin.” 108 In particular, Professor Gaddie took “the electoral data … and
constructed a regression analysis … in order to create an estimate of the vote
performance of every district.” 109 He explained that this analysis “could be used to
create a set of visual aids to demonstrate the partisan structure of Wisconsin politics.” 110
As noted above, Professor Gaddie’s regression analysis was employed to confirm
the validity of the composite measure developed by Foltz, Ottman, and Handrick.
Professor Gaddie also used his regression analysis to assess each of the drafters’
proposed maps and to create “S” curves to illustrate how the Republican seat share
would change based on changes in the party’s statewide vote share. 111 In Professor
Gaddie’s words, the “S” curves were “designed to tease out a potential estimated vote
for the legislator in the district and then allow you to also look at that and say, okay,
what if the Democrats have a good year? What if the Republicans have a good year?
How does it shift?” 112 At least some of the “S” curves were printed and kept in the map
room at Michael Best & Friedrich; in print form, the “S” curves were large enough to
“cover half th[e] table.” 113
3.
Adam Foltz
Foltz worked as a legislative aide for Speaker Fitzgerald and served as one of the
primary drafters of Act 43. 114 One additional aspect of Foltz’s testimony at trial,
107
See supra at 7–10, 12–14.
108
Tr. Ex. 169, at 1; see also Tr. Ex. 161 (Gaddie Dep.), at 69–70.
109
Tr. Ex. 161 (Gaddie Dep.), at 98.
110
Id.
See id. at 101. Specifically, the “S” curves give a visual depiction of how each party’s vote share (on the
x axis) relates to the number of Assembly seats that party likely will secure (on the y axis). See supra at 12
(discussing “S” curves in detail) and infra at Appendix 1 (depicting partial “S” curve).
111
112
Id.
113
Id. at 107.
114
See supra at 7–15.
23
however, is worthy of note. His testimony revealed a shortcoming in the drafters’
composite partisan measure. Specifically, the composite score likely was skewed to
show a greater Republican advantage because of an error in the data for the 2006
Governor’s race (one of the components of the composite score). As a result of this
error, the partisan estimates in the drafters’ spreadsheets were distorted and differed
from the estimates reached by Professor Gaddie in his “S” curves. Foltz testified that he
had not noticed this discrepancy at the time of drafting. He explained that, at the time,
he “didn’t spend a whole lot of time with” Professor Gaddie so he “[did]n’t really
understand the nuts and bolts” of the “S” curves. 115
4.
Tad Ottman
Ottman testified to his involvement in the drafting and passage of Act 43. 116
115
R.147 at 65–66.
See supra at 7–15. The third individual involved in the drafting process, Joseph Handrick, did not
testify at trial and was not deposed by counsel in this case. However, Handrick was deposed multiple
times in Baldus v. Members of the Wisconsin Government Accountability Board, 849 F. Supp. 2d 840 (E.D. Wis.
2012), and his depositions were admitted into the record here. Handrick is a former state legislator who
was involved in Wisconsin’s reapportionment in 1991, 2001, and 2011—first as a staffer to the
assemblyman chairing the committee, then as an independent consultant, and, in 2011, as an employee of
the law firm of Reinhart Boerner Van Dueren, which was retained by Michael Best & Friedrich.
116
Handrick’s deposition testimony largely conforms with the trial testimony of Foltz and Ottman,
with a few notable exceptions. Handrick described himself as a “nonpartisan consultant,” Tr. Ex. 311
(Handrick Dep. 2/1/12), at 351, and denied seeking to achieve any partisan advantage in the drafting
process, Tr. Ex. 290 (Handrick Dep. 12/20/11), at 125. Handrick did not recall being provided with any
data on voting results from past elections. Tr. Ex. 311, at 332. He testified instead that his role was limited
to evaluating completed maps solely on the bases of “[p]opulation[] deviation, municipal splits, [and]
contiguity.” Tr. Ex. 290, at 57.
Notably, Handrick was presented with an account of his role in previous Wisconsin redistricting
cycles in a book written by Professor Gaddie. The book described Handrick as a “talented artisan of
electoral maps” who “was contracted as an independent consultant, working through the law firm
representing the assembly in redistricting, to develop legislative maps that would stand up to a high
degree of scrutiny by the courts and that would also be favorable to Republicans.” Id. at 73–74. When
asked if he agreed with this description, Handrick responded, “I don’t disagree.” Id. at 74. Similarly,
Handrick was asked about a particular quote attributed to him in the book: “When they sat me down at
the [computer] terminal, I just had a knack for being able to see how to craft the kind of districts they
wanted, with the right political skew and in a fashion that would be attractive to a court.” Id. at 71.
Handrick was asked if this quotation was accurate, to which he responded, “I presume it is.” Id.
24
5.
Kenneth Mayer
Kenneth Mayer, a professor of political science at the University of Wisconsin,
served as an expert witness for the plaintiffs. His ultimate goal was to design an
alternative districting plan to Act 43 “that had an efficiency gap as low to zero as I could
get it” while also complying with traditional districting criteria to the same extent as
Act 43. 117 He first created a regression model that estimated partisanship for each
geographic area, so that he could compare his plan to Act 43. To ensure the model was
accurate, Professor Mayer compared the predictions made by his regression model to
the actual results in 2012. Once he was confident in his model, Professor Mayer “used a
GIS redistricting program called Maptitude … to … complete the task of actually
drawing the Assembly district map.” 118
Professor Mayer’s alternative “Demonstration Plan” yields a 2.2% EG in favor of
the Republicans, compared to an 11.69% EG yielded by Act 43. 119 According to
Professor Mayer, “[o]n all constitutional requirements, the Demonstration Plan is
comparable to Act 43.” 120 On cross-examination, however, the defendants pointed out
that Professor Mayer did not take account of incumbents when drawing the plan. 121 As
a result, his plan paired a greater number of incumbents than Act 43, including one
pairing in a majority-minority district. 122 Further, Professor Mayer had not drawn any
Senate districts, and therefore had not taken account of disenfranchisement. 123
117
R.148 at 146.
118
Id. at 151; see also infra at 107–08.
R.148 at 180–81. Professor Mayer and Professor Jackman calculate the EG in different ways. Professor
Jackman’s analysis is a “simplified method,” see supra note 88, which looks at the percentage of vote and
the percentage of seats. Alternatively, Professor Mayer calculates the EG by “generat[ing] estimates for
the number of Democratic and Republican votes that were cast in each district and … used that to
calculate the number of surplus and lost votes.” R.148 at 179. The differences in these calculations will be
discussed in more detail infra at IV.B.3.
119
120
Tr. Ex. 2, at 38.
121
R.149 at 112.
122
Id. at 113.
Id. at 118–19. As explained by the Court in Baldus v. Members of Wisconsin Government Accountability
Board, 849 F. Supp. 2d 840, 852 (E.D. Wis. 2012), disenfranchisement occurs because,
123
[p]ursuant to Wisconsin Constitution Article IV, section 5, state senators serve four-year,
staggered terms with half of the senators elected in presidential years and the other half
during midterm years. The redistricting plan shifts voters among senate districts in a
manner that causes certain voters who previously resided in an even-number district
(which votes in presidential years) to be moved to an odd-numbered district (which
25
In addition to discussing the Demonstration Plan, Professor Mayer responded to
points made by the defendants’ experts in their reports. Specifically, Professor Mayer
testified that he had conducted a sensitivity analysis to address concerns about the
effect of “wave” elections—elections that that dramatically favor one party—on the EG
calculations for both the Demonstration Plan and Act 43. He first looked over the last
twenty years of elections in Wisconsin and found the greatest and smallest statewide
vote shares for each party.124 Using these vote shares as the likely electoral spectrum, he
performed a swing analysis where the Democrats received an additional 3% of the
statewide vote (compared to their 2012 share) and the Republicans received an
additional 5% of the statewide vote (again compared to their 2012 share) “to see what
effect that would have on [his] efficiency gap calculations for the Demonstration
Plan.” 125 Professor Mayer’s analysis revealed that the Demonstration Plan’s EG
remained below 4%, regardless of the swing. 126 Act 43’s EG, however, increased during
a Democratic swing but significantly decreased during a Republican swing. Professor
Mayer noted that this is because “we’ve swung the Republican vote percentage up to 54
percent” but “[t]he number of [Republican] seats doesn’t change.” 127 In Professor
Mayer’s view, the result “is a confirmation that the bias in Act 43 is about the maximum
that you can get.” 128
6.
Simon Jackman
Simon Jackman, a professor of political science and statistics at Stanford
University, also served as an expert witness for the plaintiffs. Professor Jackman
primarily testified about the reliability and practicability of the EG. He conducted a
survey of 786 state legislative elections (under 206 different districting plans) in the
United States between 1972 and the present day, in order to ascertain whether there was
votes in mid-term years); this shift means that instead of voting for a state senator in
2012, as they would have done, they must wait until 2014 to have a voice in the
composition of the State Senate.
R.148 at 225. “[T]he largest statewide vote share that the Democrats received in Assembly elections was
in 2006 and it was 54.2 percent, 54—it was 54 and change. And the smallest statewide vote share that the
Democrats received was about 46 percent and we saw that in 2010.” Id.
124
125
Id.
126
Id. at 228.
127
Id. at 229.
128
Id.
26
a baseline EG which should “trigger scrutiny” and also to compare Act 43 to other
redistricting plans. 129
Professor Jackman sought to determine how much the EG varied from election
year to election year, and whether a districting plan had any impact on that EG.
Professor Jackman presented a “scatterplot,” which graphed the relationship between
the EG in the first election year of a redistricting plan (set forth on the x axis) and the
average EG over the lifetime of the plan (set forth on the y axis). 130 He found a
“relatively strong predictive relationship,” meaning that a high EG in the first year of a
redistricting plan likely means that the EG will remain high for the lifetime of the
plan. 131
Based on his research, Professor Jackman proposed that an EG of 7% or higher
should be legally significant:
I arrived at 7 percent because that seemed to be a reasonable threshold for
saying yes, if the first election under a plan produces an efficiency gap
score at least that big, then you can be confident now that you’ve seen not
just a one-off, but something that’s going to persist over the life of the plan
as a signal of — a reliable signal as to the set of efficiency gap scores and
the average efficiency gap score you might see if the plan were allowed to
run.[132]
In other words, an EG of 7% in favor of one party in the first election year of a plan
almost certainly means that the EG will favor that same party in each subsequent
election year under that plan.
Professor Jackman noted that the EGs for the 2012 and 2014 races in Wisconsin—
13% and 10%, respectively—were particularly high by historical levels. The EG in 2012
was, according to Professor Jackman, “among the largest scores we’ve seen anywhere”
and “in the top 3 percent in terms of magnitude.” 133 Act 43’s average EG ranked fifth
out of the 206 plans that Professor Jackman surveyed.134 He testified that he was
129
R.149 at 150, 200.
130
Id. at 209–10; Tr. Ex. 83, at 17.
131
R.149 at 210.
132
Id. at 209.
133
Id. at 225.
134
Id. at 227.
27
“virtually certain” that “Act 43 will exhibit a large and durable advantage in favor of
Republicans over the rest of the decade.” 135
7.
Sean Trende 136
Sean Trende, Senior Elections Analyst for the website RealClearPolitics, served as
an expert witness for the defendants. Mr. Trende primarily testified on the political
geography of Wisconsin and its potential effect on the EG.
Mr. Trende explained that, as a general matter, political geography of the United
States currently favors Republicans. In his view, the Democratic coalition has contracted
geographically and is now concentrated heavily in urban areas. This concentration, in
turn, has hurt the Democratic Party in congressional elections, which tend to favor
parties with wider geographic reach. 137
Mr. Trende also testified to the political geography of Wisconsin itself, which he
analyzed using a measure called the “partisan index” (“PI”). The purpose of the PI is
“to determine the partisan lean of political units,” 138 in order to “compare results across
elections.” 139 Mr. Trende explained that the county and ward PI values within
Wisconsin have shifted such that the Democratic Party’s influence was strengthening in
areas “that already leaned Democratic,” but was contracting geographically. 140
Mr. Trende then applied his PI analysis to Wisconsin’s wards in what he referred
to as a “nearest neighbor” analysis, which assessed the median distance between
heavily Democratic wards compared to the median distance between heavily
Republican wards. 141 From this analysis, Mr. Trende concluded that it has “become[]
progressively harder to draw … Democratic districts elsewhere in the state,” which in
135
Id. at 233.
Mr. Trende’s report and testimony was the subject of a motion in limine. We address this motion in
Part V of our opinion.
136
Mr. Trende demonstrated this theory using color-coded maps illustrating the 1996, 2004, and 2008
presidential vote results by county in Texas, Oklahoma, Arkansas, Louisiana, Mississippi, Alabama,
Tennessee, Kentucky, and Virginia. Tr. Ex. 547, at 17. Over the three election cycles, the number of
counties shaded blue (indicating that a majority of the county voted for the Democratic presidential
candidate) decreased, and the number of red counties increased. See id.
137
138
R.150 at 19.
139
Tr. Ex. 547, at 20.
140
Id. at 26.
141
R.150 at 60.
28
his view explained at least some of the EG. 142 However, he did not determine exactly
how much of the EG was attributable to geography. 143
8.
Nicholas Goedert
Nicholas Goedert, a visiting professor of political science at Lafayette College,
was retained by the defendants to offer opinions on using the EG to measure partisan
gerrymandering.
Professor Goedert’s main objection to the EG was its perceived volatility. In
Professor Goedert’s view, “wave elections are the norm,” meaning that “much more
often than not one party wins by 5 percent or more” of the vote. 144 Therefore, relying on
an EG from one election year, which might have taken place during a close election,
might not be reliable. Professor Goedert opined that, “at a very minimum, … you need
to have some sort of robust sensitivity testing that would be codified if you were going
to use the efficiency gap in any way.” 145
Professor Goedert also raised a series of policy concerns. First, he pointed out
that the EG measure arguably rests on a “2-to-1” vote-to-seats ratio and therefore a
certain standard of proportionality. 146 He also noted that there are “normatively good
reasons why a state might cho[ose] to draw a map in a certain way and even under
these normatively good reasons we could and actually do observe very high efficien[cy]
gaps.” 147 For example, Professor Goedert noted that some states may wish to create a
more proportional system or encourage competitive elections. 148 In his view, states
might be discouraged from pursuing these policy goals if the court adopted the EG as
the standard for partisan gerrymandering.
E.
Post-Trial Briefing
Both parties filed post-trial briefs, which summarized their views of the case in
light of the evidence presented at trial. The plaintiffs contended that they satisfied their
142
Id. at 64.
143
Id. at 98.
144
Id. at 175.
145
Id. at 176.
146
Id. at 163–64; see also infra at 82 (explaining 2-to-1 ratio).
147
Id. at 162–63.
148
Id. at 165–70.
29
proposed three-part test by proving discriminatory intent, discriminatory effect, and an
absence of a justification for that effect. On intent, the plaintiffs focused in particular on
the alternative maps that the drafters rejected, the “S” curves drawn by Professor
Gaddie, and memos written by Foltz and Ottman. On effect, the plaintiffs stressed that
the EG was not only likely to favor Republicans for the lifetime of the plan, but that it
also was likely to stay relatively high. The plaintiffs also highlighted the sensitivity
testing that had been conducted by Professors Jackman and Mayer. On justification, the
plaintiffs pointed out that the previous Assembly maps in Wisconsin, the alternative
plans drafted by the defendants, and Professor Mayer’s Demonstration Plan all
exhibited lower EGs while arguably complying as well with traditional districting
criteria. 149
In response, the defendants contended that “a plan that complies with all neutral
districting criteria, and whose efficiency gap is consistent with prior court-drawn plans”
cannot be unconstitutional. 150 The defendants noted that Act 43’s districts were
congruent, compact, and fairly equal in population. Further, much of the secrecy
surrounding Act 43’s enactment was consistent with how bills typically are enacted in
Wisconsin. The defendants also pointed to evidence that the political geography in
Wisconsin favors Republicans, which they contend explains the trend in EGs towards
that party over the past two decades. In the defendants’ view, this evidence also
illustrates the unreliability of the EG. The defendants concluded that the plaintiffs had
not presented enough of a reason for a court to intervene in the redistricting process.
We express our appreciation to both parties for their thorough and informative
presentation, and now turn to the legal principles that must guide our analysis of the
case.
III
THE LEGAL LANDSCAPE
The plaintiffs’ claim is that Act 43 violates their First and Fourteenth Amendment
rights because it discriminates against Democratic voters by diminishing the strength of
their votes in comparison to their Republican counterparts.
The plaintiffs also discussed alternative maps drawn by Professor Jowei Chen, who wrote an article
that the plaintiffs have moved to admit into evidence. The defendants have challenged the admissibility
of Professor Chen’s article on hearsay grounds. We shall address those arguments infra note 350.
149
150
R.153 at 1.
30
We note, as a prefatory matter, that we have acknowledged, throughout this
litigation, that the plaintiffs’ standing to maintain a cause of action is a threshold issue.
See, e.g., Tierney v. Advocate Health and Hosps. Corp., 797 F.3d 449, 450 (7th Cir. 2015).
Indeed, in our disposition of the defendants’ motion to dismiss, we addressed
extensively standing and “conclude[d] that plaintiffs’ alleged injury [wa]s sufficiently
concrete and particularized under current law to satisfy Lujan [v. Defenders of Wildlife,
504 U.S. 555 (1992),] with respect to a statewide challenge to the districting plan.” 151 “We
reach[ed] the same conclusion with respect to [Lujan’s] second and third elements of
standing, which are causation and redressability.” 152 We noted, though, that the
“defendants [we]re free to raise this issue again on a more developed record.” 153
Lujan explains that, because the elements of standing “are not mere pleading
requirements but rather an indispensable part of the plaintiff’s case, each element must
be supported in the same way as any other matter on which the plaintiff bears the
burden of proof, i.e., with the manner and degree of evidence required at the successive
stages of the litigation.” 504 U.S. at 561. Our assessment of the evidence, as well as our
elucidation of the political gerrymandering cause of action, therefore will inform our
standing analysis. Consequently, we postpone a plenary discussion of standing until we
fully have set forth the evidence as well as the constitutional standard. 154 As a precursor,
however, we conclude that the plaintiffs have established a concrete and particularized
injury: “[a]s a result of the statewide partisan gerrymandering, Democrats do not have
the same opportunity provided to Republicans to elect representatives of their choice to
the Assembly. As a result, the electoral influence of plaintiffs and other Democratic
voters statewide has been unfairly [and] disproportionately … reduced” for the life of
Act 43. 155 Additionally, the plaintiffs have shown causation: Act 43 was designed with
the purpose of solidifying Republican control of the legislature for the decennial period
and, indeed, has had that effect. Finally, the plaintiffs have established that their injury
is redressable: adopting a different statewide districting map would redress the
constitutional violation by removing the state-imposed impediment on Democratic
voters. 156
In resolving the plaintiffs’ claim, we face a significant analytical problem.
Although the Supreme Court’s political gerrymandering cases establish that “an
151
R.43 at 11.
152
Id. at 14.
153
Id. at 15.
154
See infra Part VI.
155
R.1 at 6–7, ¶ 16; see also infra at 112-13.
156
See infra at 113.
31
excessive injection of politics is unlawful,” Vieth, 541 U.S. at 293 (plurality opinion)
(emphasis removed), the Court has not come to rest on a single, judicially manageable
or discernible test for determining when the line between “acceptable” and “excessive”
has been crossed. Indeed, a signature feature of these cases is that no single opinion has
garnered a majority of the Court.
But the absence of a well-trodden path does not relieve us of the obligation to
render a decision. True, we cannot anticipate that the Court will alter course from the
decisional law, however sparse, that currently exists. See State Oil Co. v. Khan, 522 U.S. 3,
20 (1997) (noting that lower courts should apply outstanding precedent until explicitly
overruled by the Supreme Court). Nor can we cobble together the opinions of the
various Justices who have written on the matter and call the resulting amalgam binding
precedent. See Rodriguez de Quijas v. Shearson/American Exp., Inc., 490 U.S. 477, 484 (1989)
(noting that lower courts should follow precedent despite expressed dissatisfaction by
various members of the Court until the precedent is overruled explicitly). Nevertheless,
understanding that we are in an area where the navigational signs are not yet
well-placed, we must decide the case before us and satisfy our “duty … to say what the
law is,” Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803), or at least what we believe
it to be.
We begin by examining the cases that set forth the constitutional principles
which later informed the Court’s political gerrymandering decisions.
A.
The Foundational Case Law
1.
Over half a century ago, the Supreme Court recognized that the constitutionality
of legislative apportionments is governed by the Equal Protection Clause of the
Fourteenth Amendment. See Reynolds v. Sims, 377 U.S. 533 (1964). Reynolds was not a
political gerrymandering case, but addressed allegations that an outdated
apportionment scheme resulted in “serious discrimination with respect to the allocation
of legislative representation” in violation of the Equal Protection Clause. Id. at 540.
Nevertheless, the Supreme Court spoke to the importance and nature of the right to
vote in terms that also inform our consideration of the plaintiffs’ claims.
The Court first observed that the right to vote “is a fundamental matter in a free
and democratic society. Especially since the right to exercise the franchise in a free and
unimpaired manner is preservative of other basic civil and political rights, any alleged
infringement of the right of citizens to vote must be carefully and meticulously
scrutinized.” Id. at 561–62. The Court explained that “[m]ost citizens” exercise their
“inalienable right to full and effective participation in the political process” by voting
32
for their elected representatives. Id. at 565. “Full and effective participation by all
citizens in state government requires, therefore, that each citizen have an equally
effective voice in the election of members of his state legislature.” Id. Moreover,
the concept of equal protection has been traditionally viewed as requiring
the uniform treatment of persons standing in the same relation to the
governmental action questioned or challenged. With respect to the
allocation of legislative representation, all voters, as citizens of a State,
stand in the same relation regardless of where they live.
Id.
The Court explained, however, that the requirement of equal treatment was not
limited to where a voter resided. Instead, “[a]ny suggested criteria for the differentiation
of citizens are insufficient to justify any discrimination, as to the weight of their votes,
unless relevant to the permissible purposes of legislative apportionment.” Id. (emphasis
added). The Court therefore concluded that,
[s]ince the achieving of fair and effective representation for all citizens is
concededly the basic aim of legislative apportionment, … the Equal
Protection Clause guarantees the opportunity for equal participation by all
voters in the election of state legislators. Diluting the weight of votes
because of place of residence impairs basic constitutional rights under the
Fourteenth Amendment just as much as invidious discriminations based
upon factors such as race or economic status.
Id. at 565–66 (citations omitted). 157
Reynolds therefore establishes that, in electing state representatives, the votes of
citizens must be weighted equally. If an apportionment scheme violates the principle of
one-person, one-vote, it must be justified on the basis of other, permissible, legislative
considerations.
In Reynolds v. Sims, 377 U.S. 533, 577 (1964), the Court required “that a State make an honest and good
faith effort to construct districts … as nearly of equal population as is practicable.” Later cases set a 10%
threshold: an apportionment plan with a maximum population deviation between the largest and
smallest district of 10% is presumptively constitutional; larger disparities create a prima facie case of
discrimination, and the State must justify its plan. See Evenwel v. Abbott, 136 S. Ct. 1120, 1124 (2016);
Mahan v. Howell, 410 U.S. 315, 329 (1973) (approving an apportionment plan with a deviation of 16% in
light of the State’s interest in “maintaining the integrity of political subdivision lines”).
157
33
2.
The Court soon had the opportunity to apply the principles set forth in Reynolds
to allegations of vote-dilution brought by racial minorities. In Fortson v. Dorsey, 379 U.S.
433 (1965), the Court considered the constitutionality of an apportionment scheme
which included traditional single-member districts and multimember districts, where
citizens reside in a comparatively larger district and vote for multiple representatives.
Voters alleged that these multimember districts were “defective because county-wide
voting in multi-district counties could, as a matter of mathematics, result in the
nullification of the unanimous choice of the voters of a district.” Id. at 437. The district
court granted summary judgment to the plaintiffs, finding that the statute was
unconstitutional on its face.
The Supreme Court disagreed that such districts were unconstitutional per se,
and it declined to strike the plan. The Court acknowledged, however, that “[i]t might
well be that, designedly or otherwise, a multi-member constituency apportionment
scheme, under the circumstances of a particular case, would operate to minimize or
cancel out the voting strength of racial or political elements of the voting population.” Id. at 439
(emphasis added). The Court, therefore, remanded for factfinding to determine whether
the plaintiffs could meet this burden.
Following Fortson, the Court has held that multimember districts violate the
Constitution when the plaintiffs have produced evidence that an election was “not
equally open to participation by the group in question—that its members had less
opportunity than did other residents in the district to participate in the political
processes and to elect legislators of their choice.” White v. Regester, 412 U.S. 755, 766
(1973).
Later cases refined the methodology by which courts evaluate claims of vote
dilution. In Rogers v. Lodge, 458 U.S. 613 (1982), Burke County, Georgia, employed an
at-large system of elections to determine its Board of Commissioners, rather than
dividing the county into districts and allowing each district to choose a commissioner.
Id. at 615. African-American citizens in that county brought an action in which they
alleged that the county’s system of at-large elections violated their First, Thirteenth,
Fourteenth, and Fifteenth Amendment rights by diluting their voting power. The
district court held that, although the at-large electoral system was neutral in origin, it
was being maintained for invidious purposes and therefore ordered the county to be
divided into districts for purposes of electing commissioners.
The Supreme Court affirmed. It explained that districts violate the Equal
Protection Clause when “‘conceived or operated as purposeful devices to further racial
discrimination’ by minimizing, cancelling out or diluting the voting strength of”
minority populations. Id. at 617. These cases “are thus subject to the standard of proof
34
generally applicable to Equal Protection Clause cases,” specifically the “‘quality of a law
claimed to be racially discriminatory must ultimately be traced to a racially
discriminatory purpose.’” Id. (quoting Washington v. Davis, 426 U.S. 229, 240 (1976)).
Discriminatory intent, however, “need not be proved by direct evidence,” but may be
“‘inferred from the totality of the relevant facts.’” Id. at 618 (quoting Washington, 426
U.S. at 242).
Applying this standard, the Court “decline[d] to overturn the essential finding of
the District Court … that the at-large system … ha[d] been maintained for the purpose
of denying blacks equal access to the political processes in the county.” Id. at 627.
Evidence of discriminatory purpose included the fact that no African American ever
had been elected despite “overwhelming evidence of bloc voting along racial lines.” Id.
at 623–24. There also was evidence of historical discrimination in the form of literacy
tests, poll taxes, and school segregation, id. at 624–25; of a disparity in socio-economic
status that “result[ed] in part from the lingering effects of past discrimination,” id. at 626
(internal quotation marks omitted); and of county elected officials’ unresponsiveness
and insensitivity to African-American constituents, see id. at 625–26. 158
Although focused on racially discriminatory apportionment schemes, Fortson
and subsequent vote-dilution cases establish that Equal Protection concerns arise when
apportionment plans “minimize or cancel out the voting strength” either of racial
minorities or, as we have here, “political elements of the voting population.” 379 U.S. at
439 (emphasis added). Moreover, they instruct that vote-dilution cases are governed by
the same standards as other equal-protection claims, namely the plaintiffs must
establish both a discriminatory intent and a discriminatory effect.
B.
Present Supreme Court Precedent
1.
The Court drew heavily from the Fortson line of cases in resolving the political
gerrymandering claim asserted in Gaffney v. Cummings, 412 U.S. 735 (1973). In Gaffney,
Shortly after the Court’s decision in Rogers v. Lodge, 458 U.S. 613 (1982), Congress amended Section 2
of the Voting Rights Act to prohibit apportionment schemes that result in the dilution of the voting
strength of minorities, regardless of intent. 52 U.S.C. § 10301. Consequently, the Court now addresses
vote dilution through the Voting Rights Act rather than the Constitution. See, e.g., LULAC, 548 U.S. at
423–35; Voinovich v. Quilter, 507 U.S. 146, 155 (1993); Thornburg v. Gingles, 478 U.S. 30, 34 (1986). However,
the Court still recognizes the validity of Fortson and its progeny. See, e.g., Shaw v. Reno, 509 U.S. 630, 649
(1993) (noting that it has “considered the constitutionality of” multimember districting and at-large
voting systems and has “required plaintiffs to demonstrate that the challenged practice has the purpose
and effect of diluting a racial group’s voting strength”).
158
35
the Connecticut Apportionment Board created a redistricting plan designed to yield
Democratic and Republican seats in proportion to the statewide vote. A three-judge
district court invalidated the plan on the ground that the deviations from equality of
population in both houses were not “justified by any sufficient state interest,” “[m]ore
particularly, … that the policy of partisan political structuring … cannot be approved as
a legitimate reason for violating the requirement of numerical equality of population in
districting.” Id. at 740 (internal quotation marks omitted).
The Supreme Court reversed. In its analysis, the Supreme Court acknowledged
that “[s]tate legislative districts may be equal or substantially equal in population and
still be vulnerable under the Fourteenth Amendment”; it stated:
A districting plan may create multimember districts perfectly acceptable
under equal population standards, but invidiously discriminatory because
they are employed “to minimize or cancel out the voting strength of racial
or political elements of the voting population.” We must, therefore,
respond to appellees’ claims in this case that even if acceptable
populationwise, the Apportionment Board’s plan was invidiously
discriminatory because a “political fairness principle” was followed in
making up the districts in both the House and Senate.
Id. at 751–52 (citations omitted).
The Court, however, was “unconvinced” that the plan violated the Fourteenth
Amendment. Id. at 752. The Court observed that Connecticut’s Apportionment Board
had sought to “achieve a rough approximation of the statewide political strengths of the
Democratic and Republican parties,” by implementing a “political fairness” plan. Id.
(internal quotation marks omitted). The Court saw no constitutional impediment to the
State’s considering partisan interests in this way. Id. at 752–53.
The Court made clear, however, that the drawing of legislative districts along
political lines “is not wholly exempt from judicial scrutiny under the Fourteenth
Amendment.” Id. at 754. Relying on its vote-dilution cases, it gave as an example
“multimember districts [that] may be vulnerable” to constitutional challenges “if racial
or political groups have been fenced out of the political process and their voting
strength invidiously minimized.” Id. “Beyond this,” the Court continued, it had “not
ventured far or attempted the impossible task of extirpating politics from what are the
essentially political processes of the sovereign States.” Id.
In closing, however, the Court was careful to distinguish the plan before it,
which employed political classifications for benign—even salutary—purposes, with
plans that did not have proportional representation as their aim:
36
[N]either we nor the district courts have a constitutional warrant to
invalidate a state plan, otherwise within tolerable population limits,
because it undertakes, not to minimize or eliminate the political strength of any
group or party, but to recognize it and, through districting, provide a rough sort
of proportional representation in the legislative halls of the State.
Id. (emphasis added).
In sum, the Court reiterated that its concern was invidious discrimination by the
State; absent the plaintiffs’ establishing an intent to dilute the strength of a particular
group or party, the Equal Protection Clause was not offended.
2.
The Court next addressed partisan gerrymandering in Davis v. Bandemer, 478 U.S.
109 (1986). Because Bandemer was the first case in which a party directly raised, and the
Court squarely addressed, a claim that a legislative redistricting plan invidiously
discriminated against members of a political party, we treat it in some depth.
In Bandemer, Indiana Democrats challenged the 1981 state reapportionment plan
passed by a Republican-controlled legislature. Specifically, they alleged that the plan
was intended to disadvantage Democrats in electing representatives of their choosing,
in violation of the Equal Protection Clause under the Fourteenth Amendment. In
November 1982, before the case went to trial, elections were held under the new plan.
The district court had “sustained an equal protection challenge to Indiana’s 1981 state
apportionment on the basis that the law unconstitutionally diluted the votes of Indiana
Democrats,” id. at 113 (plurality opinion), but the Supreme Court reversed. A majority
of the Court 159 first concluded that the issue before the Court, like those in the
one-person, one-vote cases and in the vote-dilution cases, “is one of representation” and
“decline[d] to hold that such claims [we]re never justiciable.” Id. at 124. “As Gaffney
demonstrates,” the Court continued, the fact that a “claim is submitted by a political
group, rather than a racial group, does not distinguish it in terms of justiciability.” Id. at
125. That the complaining group does not share an “immutable” characteristic or
otherwise “has not been subject to the same historical stigma may be relevant to the
manner in which the case is adjudicated, but these differences do not justify a refusal to
entertain such a case.” Id.
The majority consisted of the Justices in the plurality (White, Brennan, Marshall, and Blackmun) and
Justices Powell and Stevens, who concurred in part and dissented in part.
159
37
Turning to the standard to be applied, a majority of the Court agreed that the
“plaintiffs were required to prove both intentional discrimination against an identifiable
political group and an actual discriminatory effect on that group.” Id. at 127. 160 A
majority of the Court also believed that the first requirement—intentional
discrimination against an identifiable group—had been met. See id. (citing Mobile v.
Bolden, 446 U.S. 55, 67–68 (1980)). 161 Indeed, it observed that, “[a]s long as redistricting is
done by a legislature, it should not be very difficult to prove that the likely political
consequences of the reapportionment were intended.” Id. at 129.
The plurality, however, rejected “the District Court’s legal and factual bases for
concluding that the 1981 Act visited a sufficiently adverse effect on the appellees’
constitutionally protected rights to make out a violation of the Equal Protection
Clause.” Id. at 129. It was not the case that “any apportionment scheme that purposely
prevents proportional representation is unconstitutional.” Id. at 129–30 (emphasis
added). Indeed, the plurality noted that precedent “clearly foreclose[d] any claim that
the Constitution requires proportional representation or that legislatures in
reapportioning must draw district lines to come as near as possible to allocating seats to
the contending parties in proportion to what their anticipated statewide vote will be.”
Id. at 130 (first citing Whitcomb v. Chavis, 403 U.S. 124, 153 (1971); then citing White, 412
U.S. at 765–68).
Moreover, the plurality held “that a particular apportionment scheme makes it
more difficult for a particular group in a particular district to elect the representatives of
Justices Powell and Stevens concurred in this aspect of the plurality’s opinion. See Davis v. Bandemer,
478 U.S. 109, 161 (1986) (Powell, J., concurring in part and dissenting in part).
160
161
Justice Powell described the process that led the Court to this conclusion:
In 1981, the Republican Party controlled both houses of the Indiana General Assembly,
and its candidate held the Governor’s seat. Pursuant to the requirements of the State
Constitution, the General Assembly undertook legislative redistricting based on 1980
census data. A Conference Committee, all of whose members were Republicans, was
assigned the task of drawing district maps with the assistance of a private computer firm.
The information fed into the computer primarily concerned the political complexion of
the State’s precincts. The redistricting process was conducted in secret. Democratic
legislators were not afforded any participation in designing the district maps that were
adopted. There were no hearings where members of the public were invited to express
their views. The Republican Committee revealed its proposed redistricting plan two days
before the end of the legislative session, and the Democrats hurriedly presented an
alternative plan. On the last day of the session, the Republican plan was adopted by
party line vote in both Houses of the General Assembly. The Governor signed the plan
into law.
Id. at 162–63 (Powell, J., concurring in part and dissenting in part).
38
its choice” also did “not render that scheme constitutionally infirm.” Id. at 131. In
reaching this conclusion, it noted that the Court had refused to approve the use of
multimember districts “[o]nly where there [wa]s evidence that excluded groups ha[d]
‘less opportunity to participate in the political processes and to elect candidates of their
choice.’” Id. (quoting Rogers, 458 U.S. at 624). It emphasized that “unconstitutional
discrimination occurs only when the electoral system is arranged in a manner that will
consistently degrade a voter’s or a group of voters’ influence on the political process as
a whole”:
[A]n equal protection violation may be found only where the electoral
system substantially disadvantages certain voters in their opportunity to
influence the political process effectively. In this context, such a finding of
unconstitutionality must be supported by evidence of continued
frustration of the will of a majority of the voters or effective denial to a
minority of voters of a fair chance to influence the political process.
Id. at 132–33.
Applying this standard to the facts before them, the plurality concluded that
“this threshold condition” had not been met. Id. at 134. It observed that the district court
had relied “primarily on the results of the 1982 elections” in which Democratic
candidates had garnered “51.9% of the votes cast statewide,” but secured only 43 seats.
Id. Republicans, however, had received only “48.1% … yet, of the 100 seats to be filled,
Republican candidates won 57.” Id. 162 “Relying on a single election to prove
unconstitutional discrimination,” however, was “unsatisfactory.” Id. at 135. The
plurality specifically noted a lack of evidence that (1) the 1981 Act prevented the
Democrats from “secur[ing] … sufficient vote[s] to take control of the assembly”;
(2) “the 1981 reapportionment would consign the Democrats to a minority status in the
Assembly throughout the 1980’s”; or (3) “the Democrats would have no hope of doing
any better in the reapportionment that would occur after the 1990 census.” Id. at 135–36.
“Without findings of this nature,” the plurality stated, “the District Court erred in
concluding that the 1981 Act violated the Equal Protection Clause.” Id. at 136.
The plurality then addressed a few aspects of Justice Powell’s opinion. “[T]he
crux of [his] analysis” was that—“at least in some cases—the intentional drawing of
district boundaries for partisan ends and for no other reason violates the Equal
Protection Clause.” Id. at 138. It disagreed that “the specific intention of disadvantaging
“In the Senate, 53.1% of the votes were cast for Democratic candidates and 46.9% for Republicans; of
the 25 Senate seats to be filled, Republicans won 12 and Democrats 13.” Id. at 134 (plurality opinion). The
district court also had “relied upon the use of multimember districts in Marion and Allen Counties,
where Democrats or those inclined to vote Democratic in 1982 amounted to 46.6% of the population of
those counties but Republicans won 86%–18 of 21–seats allocated to the districts in those counties.” Id.
162
39
one political party’s election prospects,” standing alone, established a constitutional
violation. Id. at 139. Instead, invidious intent must be coupled with evidence that “the
redistricting d[id] in fact disadvantage [a party] at the polls,” and the disadvantage
must be more than “a mere lack of proportionate results in one election.” Id. The
plurality, however, acknowledged that “election results” were “relevant to a showing of
the effects required to prove a political gerrymandering claim under our view. And the
district configurations may be combined with vote projections to predict future election
results,” which also would be relevant to showing discriminatory effects. Id. at 141.
The plurality recognized that its own test “may be difficult of application.” Id. at
142. “Nevertheless,” it concluded, the test “recognizes the delicacy of intruding on this
most political of legislative functions and is at the same time consistent with our prior
cases regarding individual multimember districts, which have formulated a parallel
standard.” Id. at 143.
Justice O’Connor, joined by Chief Justice Burger and Justice Rehnquist,
concurred in the judgment, but wrote separately. Justice O’Connor took issue with the
plurality’s reliance on both the “one-person, one-vote” principle and the Court’s
vote-dilution cases. Id. at 146–55 (O’Connor, J., concurring). In her view,
Reynolds makes plain that the one person, one vote principle safeguards
the individual’s right to vote, not the interests of political groups: “To the
extent that a citizen’s right to vote is debased, he is that much less a
citizen. The fact that an individual lives here or there is not a legitimate
reason for overweighting or diluting the efficacy of his vote.”
Id. at 149–50 (quoting Reynolds, 377 U.S. at 567) (emphasis added). Justice O’Connor also
viewed political gerrymandering as distinct from racial gerrymandering. She explained
that, “where a racial minority group is characterized by ‘the traditional indicia of
suspectness’ and is vulnerable to exclusion from the political process, individual voters
who belong to that group enjoy some measure of protection against intentional dilution
of their group voting strength by means of racial gerrymandering.” Id. at 151 (citations
omitted). “[M]embers of the Democratic and Republican Parties,” however, did not
constitute “a discrete and insular group vulnerable to exclusion from the political
process by some dominant group: these political parties are the dominant groups, and
the Court has offered no reason to believe that they are incapable of fending for
themselves through the political process.” Id. at 152 (emphasis in original).
In an opinion concurring in part and dissenting in part, Justice Powell, joined by
Justice Stevens, concluded that a redistricting plan violated the Constitution when it
served “no purpose other than to favor one segment—whether racial, ethnic, religious,
economic, or political—that may occupy a position of strength at a particular time, or to
disadvantage a politically weak segment of the community.” Id. at 164 (internal
40
quotation marks omitted) (quoting Karcher v. Daggett, 462 U.S. 725, 748 (1983)
(Stevens, J., concurring in part and dissenting in part)). He believed that this conclusion
followed from the principles articulated in Reynolds, namely “that equal protection
encompasses a guarantee of equal representation, requiring a State to seek to achieve
through redistricting ‘fair and effective representation for all citizens.’” Id. at 166
(quoting Reynolds, 377 U.S. at 565–66). He further explained that
[t]he concept of “representation” necessarily applies to groups: groups of
voters elect representatives, individual voters do not. Gross population
disparities violate the mandate of equal representation by denying voters
residing in heavily populated districts, as a group, the opportunity to elect
the number of representatives to which their voting strength otherwise
would entitle them. While population disparities do dilute the weight of
individual votes, their discriminatory effect is felt only when those
individual votes are combined. Thus, the fact that individual voters in
heavily populated districts are free to cast their ballot has no bearing on a
claim of malapportionment.
Id. at 167 (emphasis in original).
Applying these standards, Justice Powell believed that the “case present[ed] a
paradigm example of unconstitutional discrimination against the members of a political
party that happened to be out of power” and would have found that Indiana’s
redistricting plan violated the Equal Protection Clause. Id. at 185.
Although history would establish that the plurality correctly predicted that its
test for political gerrymandering was, in fact, “difficult of application,” id. at 142
(plurality opinion), Bandemer nevertheless provides some meaningful guidance. First,
the Court’s one-person, one-vote and vote-dilution cases provide the foundation for
evaluating claims of political gerrymandering. Second, that a “claim is submitted by a
political group rather than a racial group, does not distinguish it in terms of
justiciability.” Id. at 125. And, third, a successful political gerrymandering claim must
include a showing of both discriminatory intent and discriminatory effect.
3.
The Court revisited the issue of political gerrymandering in Vieth v. Jubelirer, 541
U.S. 267 (2004). In Vieth, the Court addressed an action filed by Democratic voters in
Pennsylvania that challenged the state legislature’s new congressional districting plan.
Justice Scalia, writing for a plurality, began with a critique of the standard articulated in
Bandemer:
41
Over the dissent of three Justices, the Court held in Davis v.
Bandemer that, since it was “not persuaded that there are no judicially
discernible and manageable standards by which political gerrymander
cases are to be decided,” 478 U.S., at 123, such cases were justiciable. …
There was no majority on that point. Four of the Justices finding
justiciability believed that the standard was one thing, see id., at 127
(plurality opinion of White, J., joined by Brennan, Marshall, and
Blackmun, JJ.); two believed it was something else, see id., at 161 (Powell,
J., joined by STEVENS, J., concurring in part and dissenting in part). The
lower courts have lived with that assurance of a standard (or more
precisely, lack of assurance that there is no standard), coupled with that
inability to specify a standard, for the past 18 years.
Id. at 278–79 (plurality opinion) (emphasis in original). In the plurality’s view,
“[e]ighteen years of judicial effort with virtually nothing to show for it justif[ied] …
revisiting the question whether the standard promised by Bandemer exists.” Id. at 281. It
concluded that “no judicially discernible and manageable standards for adjudicating
political gerrymandering claims have emerged. Lacking [such standards],” it
concluded, “political gerrymandering claims are nonjusticiable and … Bandemer was
wrongly decided.” Id.
The plurality turned first to the shortcomings of the test proposed by the
plaintiffs:
To satisfy appellants’ intent standard, a plaintiff must “show that
the mapmakers acted with a predominant intent to achieve partisan
advantage,” which can be shown “by direct evidence or by circumstantial
evidence that other neutral and legitimate redistricting criteria were
subordinated to the goal of achieving partisan advantage.” … As
compared with the Bandemer plurality’s test of mere intent to
disadvantage the plaintiff’s group, this proposal seemingly makes the
standard more difficult to meet—but only at the expense of making the
standard more indeterminate.
Id. at 284. The plurality determined that, in a statewide plan, there was no principled
way to discern predominant intent.
The test also included an “effects” prong: “The requisite effect is established
when ‘(1) the plaintiffs show that the districts systematically “pack” and “crack” the
rival party’s voters, and (2) the court’s examination of the “totality of circumstances”
confirms that the map can thwart the plaintiffs’ ability to translate a majority of votes
into a majority of seats.’” Id. at 286–87 (footnote omitted). According to the plurality,
this aspect of the test also was not judicially discernible because there is no
42
constitutional right to proportional representation: the Constitution “guarantees equal
protection of the law to persons, not equal representation in government to equivalently
sized groups.” Id. at 288. Nor, in the plurality’s opinion, was the proposed test judicially
manageable because there was no reliable method to establish “a party’s majority
status” or for “ensur[ing] that that party wins a majority of seats—unless we radically
revise the States’ traditional structure for elections.” Id. at 288–89.
The plurality then critiqued the standards proposed by the dissenting Justices.
Contrary to the view held by other members of the Court, the plurality did not believe
that the “one-person, one-vote cases” had any “bearing upon this question,” either “in
principle” or “in practicality.” Id. at 290 (first citing Reynolds, 377 U.S. 533; then citing
Wesberry v. Sanders, 376 U.S. 1 (1964)).
Not in principle, because to say that each individual must have an equal
say in the selection of representatives, and hence that a majority of
individuals must have a majority say, is not at all to say that each
discernible group, whether farmers or urban dwellers or political parties,
must have representation equivalent to its numbers. And not in
practicality, because the easily administrable standard of population
equality adopted by Wesberry and Reynolds enables judges to decide
whether a violation has occurred (and to remedy it) essentially on the
basis of three readily determined factors—where the plaintiff lives, how
many voters are in his district, and how many voters are in other districts;
whereas requiring judges to decide whether a districting system will
produce a statewide majority for a majority party casts them forth upon a
sea of imponderables, and asks them to make determinations that not
even election experts can agree upon.
Id. at 290.
Turning first to Justice Stevens’s view, the plurality agreed that “severe partisan
gerrymanders” were “incompatib[le] … with democratic principles.” Id. at 292. It could
not agree, however, that political gerrymandering should be treated equivalently to
racial gerrymandering. Id. at 293–95. In the plurality’s view, “[a] purpose to
discriminate on the basis of race receives the strictest scrutiny under the Equal
Protection Clause, while a similar purpose to discriminate on the basis of politics does
not.” Id. at 293. The plurality was unpersuaded by Justice Stevens’s reference to political
patronage cases, contending that “the underlying rights, and consequently
constitutional harms, are not comparable.” Id. at 294.
The plurality also rejected Justice Souter’s multi-factor test, which was “loosely
based in form on [the Court’s] Title VII cases.” Id. at 295. According to the plurality, this
test was “doomed to failure” because “[n]o test—yea, not even a five-part test—can
43
possibly be successful unless one knows what he is testing for. In the present context,
the test ought to identify deprivation of that minimal degree of representation or
influence to which a political group is constitutionally entitled.” Id. at 297. Although
Justice Souter “vaguely describe[d] the harm he is concerned with as vote dilution, a
term which usually implies some actual effect on the weight of a vote,” no element of
his test measured this effect. Id. Consequently, the plurality was unsure of “the precise
constitutional deprivation his test [wa]s designed to identify and prevent.” Id.
Addressing Justice Breyer’s dissent, the plurality agreed “that our Constitution
sought to create a basically democratic form of government,” but found that this was “a
long and impassable distance away from the conclusion that the Judiciary may assess
whether a group (somehow defined) has achieved a level of political power (somehow
defined) commensurate with that to which they would be entitled absent unjustified
political machinations (whatever that means).” Id. at 299 (internal quotation marks
omitted) (citations omitted).
The plurality concluded, therefore, that the Equal Protection Clause did not
“provide[] a judicially enforceable limit on the political considerations that the States
and Congress may take into account when districting.” Id. at 305.
Justice Kennedy concurred in the judgment. He agreed that “[a] decision
ordering the correction of all election district lines drawn for partisan reasons would
commit federal and state courts to unprecedented intervention in the American political
process.” Id. at 306 (Kennedy, J., concurring in the judgment). “The Court,” he stated,
was “correct to refrain from directing this substantial intrusion into the Nation’s
political life.” Id. Furthermore, “[w]hile agreeing with the plurality that the complaint
the appellants filed in the District Court must be dismissed, and while understanding
that great caution is necessary when approaching this subject, [he] would not foreclose
all possibility of judicial relief if some limited and precise rationale were found to
correct an established violation of the Constitution in some redistricting cases.” Id.
Justice Kennedy believed that
[a] determination that a gerrymander violates the law must rest on
something more than the conclusion that political classifications were
applied. It must rest instead on a conclusion that the classifications, though
generally permissible, were applied in an invidious manner or in a way
unrelated to any legitimate legislative objective.
Id. at 307 (emphasis added). In this case, Justice Kennedy explained, the plaintiffs had
not overcome the dual hurdles of discernibility and manageability:
The fairness principle appellants propose is that a majority of voters in the
Commonwealth should be able to elect a majority of the Commonwealth’s
44
congressional delegation. There is no authority for this precept. Even if the
novelty of the proposed principle were accompanied by a convincing
rationale for its adoption, there is no obvious way to draw a satisfactory
standard from it for measuring an alleged burden on representational
rights. The plurality demonstrates the shortcomings of the other standards
that have been considered to date. See ante, at Parts III and IV
(demonstrating that the standards proposed in Davis v. Bandemer, 478 U.S.
109 (1986), by the parties before us, and by our dissenting colleagues are
either unmanageable or inconsistent with precedent, or both).
Id. at 308.
However, Justice Kennedy was not willing to go so far as the plurality and hold
partisan gerrymanders nonjusticiable. Although agreeing that there were “weighty
arguments for holding cases like these to be nonjusticiable” and acknowledging that
“those arguments may prevail in the long run,” it was Justice Kennedy’s view that “the
arguments [we]re not so compelling that they require us now to bar all future claims of
injury from a partisan gerrymander.” Id. at 309. According to Justice Kennedy, the
Court’s “willingness to enter the political thicket of the apportionment process with
respect to one-person, one-vote claims ma[de] it particularly difficult to justify a
categorical refusal to entertain claims against this other type of gerrymandering.” Id. at
310.
Justice Kennedy noted specifically that, in the end, it may be the First
Amendment, not the Equal Protection Clause, which provides the framework within
which political gerrymandering claims should be analyzed. See id. at 314. “After all,” he
explained, “these allegations involve the First Amendment interest of not burdening or
penalizing citizens because of their participation in the electoral process, their voting
history, their association with a political party, or their expression of political views.
Under general First Amendment principles those burdens in other contexts are
unconstitutional absent a compelling government interest.” Id. (citing Elrod v. Burns, 427
U.S. 347 (1976) (plurality opinion)). Moreover, a “‘[r]epresentative democracy … is
unimaginable without the ability of citizens to band together in promoting among the
electorate candidates who espouse their political views.’” Id. (quoting California
Democratic Party v. Jones, 530 U.S. 567, 574 (2000)). According to Justice Kennedy, these
precedents demonstrate that
First Amendment concerns arise where a State enacts a law that has the
purpose and effect of subjecting a group of voters or their party to
disfavored treatment by reason of their views. In the context of partisan
gerrymandering, that means that First Amendment concerns arise where
45
an apportionment has the purpose and effect of burdening a group of
voters’ representational rights.
Id.
Justice Kennedy disagreed with the plurality that application of a First
Amendment standard would render invalid “all consideration of political interests in an
apportionment.” Id. at 315. He explained:
The inquiry is not whether political classifications were used. The inquiry
instead is whether political classifications were used to burden a group’s
representational rights. If a court were to find that a State did impose
burdens and restrictions on groups or persons by reason of their views,
there would likely be a First Amendment violation, unless the State shows
some compelling interest.
Id. Because “[t]he First Amendment analysis concentrates on whether the legislation
burdens the representational rights of the complaining party’s voters for reasons of
ideology, beliefs, or political association,” Justice Kennedy suggested that “[t]he
analysis allows a pragmatic or functional assessment that accords some latitude to the
States.” Id.
Justice Stevens dissented. Drawing both on the Court’s racial gerrymandering
cases, see id. at 322–23 (Stevens, J., dissenting) (citing, among other authorities, Shaw v.
Reno, 509 U.S. 630 (1993)), and the Court’s political patronage cases, see id. at 324 (citing
Elrod, 427 U.S. 347)), Justice Stevens believed that the plaintiffs had standing, presented
a redressable claim, and were entitled to relief. Specifically, he observed that “political
belief and association constitute the core of those activities protected by the First
Amendment” and that government employment decisions that burden these interests
are subject to strict scrutiny. Id. (quoting Elrod, 427 U.S. at 356 (plurality opinion)).
“Thus,” he continued, “unless party affiliation is an appropriate requirement for the
position in question, government officials may not base a decision to hire, promote,
transfer, recall, discharge, or retaliate against an employee, or to terminate a contract, on
the individual’s partisan affiliation or speech.” Id. (citing, among other sources, O'Hare
Truck Service, Inc. v. City of Northlake, 518 U.S. 712, 716–17 (1996)). Justice Stevens
concluded that “[i]t follows” therefore “that political affiliation is not an appropriate
standard for excluding voters from a congressional district.” Id. at 325. 163
Justice Stevens made it clear in his dissent that “purpose [w]as the ultimate inquiry.” Vieth, 541
U.S. at 321 (Stevens, J., dissenting). He noted that there have been “ready standards for testing
the lawfulness of a gerrymander,” id.; included among these were “configurations of the
districts,” id. at 322 (quoting Bandemer, 478 U.S. at 165 (Powell, J., dissenting)). Among other
163
46
Justice Souter wrote a dissenting opinion, joined by Justice Ginsburg, which
rested on the “one-person, one-vote” principle. Id. at 343 (Souter, J., dissenting) (citing
Reynolds, 377 U.S. 533). According to Justice Souter:
Creating unequally populous districts is not, however, the only
way to skew political results by setting district lines. The choice to draw a
district line one way, not another, always carries some consequence for
politics, save in a mythical State with voters of every political identity
distributed in an absolutely gray uniformity. The spectrum of opportunity
runs from cracking a group into impotent fractions, to packing its
members into one district for the sake of marginalizing them in another.
However equal districts may be in population as a formal matter, the
consequence of a vote cast can be minimized or maximized, and if
unfairness is sufficiently demonstrable, the guarantee of equal protection
condemns it as a denial of substantial equality.
Id. (citation omitted). Justice Souter acknowledged the Court’s prior struggles in
articulating a workable test for political gerrymandering. Accordingly, he suggested
preserving the holding in Bandemer that political gerrymandering was justiciable, but
“otherwise start[ing] anew.” Id. at 346. Specifically, he suggested using a
burden-shifting test similar to that in McDonnell Douglas Corp. v. Green, 411 U.S. 792
(1973), “calling for a plaintiff to satisfy elements of a prima facie cause of action, at
which point the State would have the opportunity not only to rebut the evidence
supporting the plaintiff’s case, but to offer an affirmative justification for the districting
choices, even assuming the proof of the plaintiff’s allegations.” Vieth, 541 U.S. at 346.164
indicators of intent were “contemporaneous statements and press accounts, demonstrating that
the architects of the districts were motivated solely by partisan considerations.” Id. (internal
quotation marks omitted). Thus, to Justice Stevens, irregular shapes were not the sine qua non of a
gerrymander, see Dissent at 129, but only one possible indicator.
The factors proposed by Justice Souter were: 1) the plaintiff belonged to a cohesive political
group; 2) the plaintiff’s district of residence “paid little or no heed to … traditional districting
principles”; 3) there were “specific correlations between the district’s deviations from traditional
districting principles and the distribution of the population of his group”; 4) there is a
“hypothetical district including [the plaintiff’s] residence, one in which the proportion of the
plaintiff’s group was lower (in a packing claim) or higher (in a cracking one) and which at the
same time deviated less from traditional districting principles than the actual district; and 5) “the
defendants acted intentionally to manipulate the shape of the district in order to pack or crack his
group.” Vieth, 541 U.S. at 347–50 (Souter, J., dissenting). The goal of these factors was to discern
whether “the defendant had chosen either to pack the group … or to crack it …, the ordinary
methods of vote dilution.” Id. at 349. Although a “bizarre shape,” Dissent at 129, would be
evidence of the second factor, Justice Souter did not propose it as a requirement.
164
47
Justice Breyer, also in dissent, opined that “the workable democracy that the
Constitution foresees” must include “a method for transforming the will of the majority
into effective government.” Id. at 356 (Breyer, J., dissenting). In his view, this method
could be harmed by “the unjustified use of political factors to entrench a minority in
power.” Id. at 360. Justice Breyer quoted extensively from Reynolds to support his view
that “[t]he democratic harm of unjustified entrenchment is obvious”:
“Logically, in a society ostensibly grounded on
representative government, it would seem reasonable that a
majority of the people of a State could elect a majority of that
State’s legislators. … Since legislatures are responsible for
enacting laws by which all citizens are to be governed, they
should be bodies which are collectively responsive to the
popular will.” Reynolds, 377 U.S. at 565.
Where unjustified entrenchment takes place, voters find it far more
difficult to remove those responsible for a government they do not want;
and these democratic values are dishonored.
Id. at 361. Consequently, “gerrymandering that leads to entrenchment amounts to an
abuse that violates the Constitution’s Equal Protection Clause.” Id. at 362.
Although the test articulated in Bandemer proved unworkable, Vieth has placed
district courts in an even greater quandary. For all its shortcomings, the Bandemer
decision at least set forth a test for district courts to apply. In Vieth, however, the
members of the Court were unanimous only in their willingness to jettison the test set
forth in Bandemer. We conclude, therefore, that the specific test for political
gerrymandering set forth in Bandemer no longer is good law. Moreover, any attempt to
craft a new test ought to avoid those shortcomings in the Bandemer test specifically
identified by the members of the Court.
4.
The Supreme Court’s most recent case on partisan gerrymandering, League of
United Latin American Citizens v. Perry (“LULAC”), 548 U.S. 399 (2006), gives little more
in the way of guidance. Nevertheless, we set forth those aspects of the decision that may
be useful in evaluating the plaintiffs’ claims.
In the 1990s, the Democrats controlled both houses of the Texas legislature and
the statehouse and enacted what was “later described as the shrewdest gerrymander of
the 1990s.” Id. at 410 (internal quotation marks omitted). Following the 2000 census,
Texas was entitled to two additional congressional seats. However, the legislature now
was split politically between a Republican Senate and a Democratic House of
48
Representatives. “As so constituted, the legislature was unable to pass a redistricting
scheme,” resulting in a court-ordered plan which left “[t]he 1991 Democratic Party
gerrymander largely in place as a ‘legal’ plan.” Id. at 411–12 (alteration in original). In
2002, however, Republicans gained control of both houses of the legislature and enacted
legislation that re-drew congressional districting lines; these new districts resulted in
the Republicans securing 21 seats with 58% of the vote in statewide races, compared to
the Democrats’ 11 seats with 41% of the vote.
Shortly after the plan was enacted, some Texas voters mounted both statutory
and constitutional challenges to it. In the constitutional challenge, the plaintiffs claimed
that a decision to enact a new redistricting plan mid-decade, “when solely motivated by
partisan objectives, violates equal protection and the First Amendment because it serves
no legitimate public purpose and burdens one group because of its political opinions
and affiliation.” Id. at 416–17. The Supreme Court disagreed.
Justice Kennedy, joined by Justices Souter and Ginsburg, opined that “a
successful claim attempting to identify unconstitutional acts of partisan
gerrymandering must do what appellants’ sole-motivation theory explicitly disavows:
show a burden, as measured by a reliable standard, on the complainants’
representational rights.” Id. at 418 (opinion of Kennedy, J.). Moreover, Justice Kennedy
was concerned that the plaintiffs’ proposed test would exempt from constitutional
scrutiny other, more serious examples of partisan gerrymandering:
The text and structure of the Constitution and our case law indicate there
is nothing inherently suspect about a legislature’s decision to replace
mid-decade a court-ordered plan with one of its own. And even if there
were, the fact of mid-decade redistricting alone is no sure indication of
unlawful political gerrymanders. Under appellants’ theory, a highly
effective partisan gerrymander that coincided with decennial redistricting
would receive less scrutiny than a bumbling, yet solely partisan,
mid-decade redistricting. More concretely, the test would leave untouched
the 1991 Texas redistricting, which entrenched a party on the verge of
minority status, while striking down the 2003 redistricting plan, which
resulted in the majority Republican Party capturing a larger share of the
seats. A test that treats these two similarly effective power plays in such
different ways does not have the reliability appellants ascribe to it.
Id. at 418–19. Justice Kennedy also noted that the current Texas map could “be seen as
making the party balance more congruent to statewide party power.” Id. at 419. “To be
sure,” Justice Kennedy continued,
there is no constitutional requirement of proportional representation, and
equating a party’s statewide share of the vote with its portion of the
49
congressional delegation is a rough measure at best. Nevertheless, a
congressional plan that more closely reflects the distribution of state party
power seems a less likely vehicle for partisan discrimination than one that
entrenches an electoral minority.
Id. at 419 (emphasis added).
Justice Kennedy also commented on a submission by an amicus which
“propose[d] a symmetry standard that would measure partisan bias by ‘compar[ing]
how both parties would fare hypothetically if they each (in turn) had received a given
percentage of the vote.’” Id. at 419. He stated:
Amici’s proposed standard does not compensate for appellants’ failure to
provide a reliable measure of fairness. The existence or degree of
asymmetry may in large part depend on conjecture about where possible
vote-switchers will reside. Even assuming a court could choose reliably
among different models of shifting voter preferences, we are wary of
adopting a constitutional standard that invalidates a map based on unfair
results that would occur in a hypothetical state of affairs. Presumably such
a challenge could be litigated if and when the feared inequity arose. More
fundamentally, the counterfactual plaintiff would face the same problem
as the present, actual appellants: providing a standard for deciding how
much partisan dominance is too much. Without altogether discounting its
utility in redistricting planning and litigation, I would conclude
asymmetry alone is not a reliable measure of unconstitutional partisanship.
Id. at 420 (citation omitted) (emphasis added). Justice Kennedy thus concluded that “a
legislature’s decision to override a valid, court-drawn plan mid-decade” is not
“sufficiently suspect to give shape to a reliable standard for identifying unconstitutional
political gerrymanders.” Id. at 423. Consequently, he concluded that the petitioners had
not established a “legally impermissible use of political classifications” and had not
stated a claim on which relief could be granted. Id.
Justice Stevens, in a separate opinion joined by Justice Breyer, reiterated the view
of impartiality that he had articulated in Vieth. He observed that “the Fourteenth
Amendment’s prohibition against invidious discrimination[] and the First
Amendment’s protection of citizens from official retaliation based on their political
affiliation” “limit the State’s power to rely exclusively on partisan preference in
drawing district lines.” Id. at 461 (Stevens, J., concurring in part and dissenting in part).
He explained:
The equal protection component of the Fourteenth Amendment requires
actions taken by the sovereign to be supported by some legitimate
50
interest, and further establishes that a bare desire to harm a politically
disfavored group is not a legitimate interest. Similarly, the freedom of
political belief and association guaranteed by the First Amendment
prevents the State, absent a compelling interest, from “penalizing citizens
because of their participation in the electoral process, … their association
with a political party, or their expression of political views.” These
protections embodied in the First and Fourteenth Amendments reflect the
fundamental duty of the sovereign to govern impartially.
Id. at 461–62 (citations omitted) (quoting Vieth, 541 U.S. at 314 (Kennedy, J.,
concurring in the judgment)). Justice Stevens also set forth some of the
representational harms engendered by political gerrymanders. Specifically, he
noted that, “in addition to the possibility that a representative may believe her
job is only to represent the interests of a dominant constituency, a representative
may feel more beholden to the cartographers who drew her district than to the
constituents who live there.” Id. at 470.
Justice Breyer, in addition to joining Justice Stevens’s opinion, wrote separately
to describe why he believed that the plan violated the Constitution:
[B]ecause the plan entrenches the Republican Party, the State cannot
successfully defend it as an effort simply to neutralize the Democratic
Party’s previous political gerrymander. Nor has the State tried to justify
the plan on nonpartisan grounds, either as an effort to achieve legislative
stability by avoiding legislative exaggeration of small shifts in party
preferences or in any other way.
In sum, “the risk of entrenchment is demonstrated,” “partisan
considerations [have] render[ed] the traditional district-drawing
compromises irrelevant,” and “no justification other than party advantage
can be found.” The record reveals a plan that overwhelmingly relies upon
the unjustified use of purely partisan line-drawing considerations and
which will likely have seriously harmful electoral consequences. For these
reasons, I believe the plan in its entirety violates the Equal Protection
Clause.
Id. at 492 (Breyer, J., concurring in part and dissenting in part) (quoting Vieth, 541 U.S. at
359, 367 (Breyer, J., dissenting)) (emphasis in original) (citations omitted).
Justices Souter and Ginsburg adhered to their view, set forth in Vieth, as to the
proper test for political gerrymandering, but concluded that there was “nothing to be
gained by working through these cases on th[at] standard” because, like in Vieth, the
Court “ha[d] no majority for any single criterion of impermissible gerrymander.” Id. at
51
483 (Souter, J., concurring in part and dissenting in part). Chief Justice Roberts, joined
by Justice Alito, agreed with Justice Kennedy “that appellants ha[d] not provided a
reliable standard for identifying unconstitutional political gerrymanders,” but took no
position as to “whether appellants ha[d] failed to state a claim on which relief can be
granted, or ha[d] failed to present a justiciable controversy.” Id. at 492–93 (Roberts, C.J.,
concurring in part, concurring in the judgment in part, and dissenting in part) (internal
quotation marks omitted). Finally, Justices Scalia and Thomas reiterated their view that
the voters’ political gerrymandering claims were nonjusticiable. See id. at 511 (Scalia, J.,
concurring in the judgment in part and dissenting in part).
5.
In its consideration of the reapportionment issue, the Court has acknowledged
that the appropriate analysis is grounded not only in its jurisprudence of equal
protection, but also its jurisprudence of associational rights under the First Amendment.
The gravamen of an equal protection claim is that a state has burdened artificially a
voter’s ballot so that it has less weight than another person’s vote. A year after Reynolds,
the Court again articulated this concept in Fortson v. Dorsey, 379 U.S. 433 (1965), when it
evaluated whether multimember legislative districts had a constitutionally
impermissible impact on the weight of African-American voters. There, the Court
reiterated its concern that voters’ ability to participate in the electoral process was
unequal. While declining to hold multimember districts were unconstitutional per se, it
noted that “designedly or otherwise, a multi-member constituency apportionment
scheme, under the circumstances of a particular case, [might] operate to minimize or
cancel out the voting strength of racial or political elements of the voting population.”
Id. at 439. Again, in White v. Register, 412 U.S. 755 (1973), the Court held that certain
multimember districts were violative of the Constitution when the plaintiffs produced
evidence that an election was not “equally open to participation by the group in
question—that its members had less opportunity than did other residents in the district
to participate in the political processes and to elect legislators of their choice.” Id. at 766
(emphasis added). In Gaffney, 412 U.S. at 754, the Court again noted that apportionment
plans that “invidiously minimize[]” the voting strength of “political groups” “may be
vulnerable” to constitutional challenges.
In these cases, the Court’s emphasis on ensuring that an individual’s vote receive
the same weight as every other person’s vote necessarily implicates that individual’s
associational rights. The Court previously has observed the link between the right to
vote and the right to associate in its ballot-access cases. One of the foundational
ballot-access cases, Anderson v. Celebrezze, 460 U.S. 780 (1983), involved a challenge to a
state law which required independent candidates to file their nominating petitions
52
seventy-five days before the primary election in order to qualify for the general election
ballot. Id. at 804–06. The Court observed that the statute in question implicated both the
“right to vote” and “freedom of association”: “Each provision of these schemes,
whether it governs the registration and qualifications of voters, the selection and
eligibility of candidates, or the voting process itself, inevitably affects—at least to some
degree—the individual’s right to vote and his right to associate with others for political
ends.” Id. at 788 (emphasis added).
The Court then outlined the analysis a court must undertake in considering a
challenge to a state’s election law:
It must first consider the character and magnitude of the asserted injury to
the rights protected by the First and Fourteenth Amendments that the
plaintiff seeks to vindicate. It then must identify and evaluate the precise
interests put forward by the State as justifications for the burden imposed
by its rule. In passing judgment, the Court must not only determine the
legitimacy and strength of each of those interests, it also must consider the
extent to which those interests make it necessary to burden the plaintiff’s
rights. Only after weighing all these factors is the reviewing court in a
position to decide whether the challenged provision is unconstitutional.
Id. at 789. Applying these steps, the Court determined that the early filing deadline at
issue in Anderson placed a burden on independent parties and that “it is especially
difficult for the State to justify a restriction that limits political participation by an
identifiable political group.” Id. at 793. After considering the state’s interests in keeping
voters well-educated about the candidates, being fair to the parties who hold primaries,
and ensuring political stability, the Court held that there was an unconstitutional
burden on “the interests of the voters who chose to associate together to express their support
for [an independent’s] candidacy and the views he espoused.” Id. at 806 (emphasis added).
The Court also noted that, in reaching its conclusion, it was relying “directly on the First
and Fourteenth Amendments” and was “not engag[ing] in a separate Equal Protection
Clause analysis.” Id. at 786–87 n.7. It had relied, however,
on the analysis in a number of our prior election cases resting on the Equal
Protection Clause of the Fourteenth Amendment. These cases, applying
the “fundamental rights” strand of equal protection analysis, have
identified the First and Fourteenth Amendment rights implicated by
restrictions on the eligibility of voters and candidates, and have
considered the degree to which the State’s restrictions further legitimate
state interests.
Id.
53
Since Anderson, the Court has continued to assess election laws through the lens
of the First and Fourteenth Amendments, without explicit reference to the Equal
Protection Clause. In evaluating election laws, the Court employs a multi-step process
that looks at the totality of the circumstances:
When deciding whether a state election law violates First and
Fourteenth Amendment associational rights, we weigh the character and
magnitude of the burden the State’s rule imposes on those rights against
the interests the State contends justify that burden, and consider the extent
to which the State’s concerns make the burden necessary. Regulations
imposing severe burdens on plaintiffs’ rights must be narrowly tailored
and advance a compelling state interest. Lesser burdens, however, trigger
less exacting review, and a State’s important regulatory interests will
usually be enough to justify reasonable, nondiscriminatory restrictions.
No bright line separates permissible election-related regulation from
unconstitutional infringements on First Amendment freedoms.
Timmons v. Twin Cities Area New Party, 520 U.S. 351, 358–59 (1997) (internal quotation
marks omitted) (citations omitted).
Nevertheless, the close relationship between equal protection and associational
rights is clear. For example, in Williams v. Rhodes, 393 U.S. 23 (1968), one of the equal
protection cases relied upon in Anderson, the Court considered the constitutionality of a
law which required new political parties to obtain the signatures of electors equaling
15% of the number of ballots cast in the preceding gubernatorial election. It stated:
[W]e have … held many times that “invidious” distinctions cannot be
enacted without a violation of the Equal Protection Clause. In determining
whether or not a state law violates the Equal Protection Clause, we must
consider the facts and circumstances behind the law, the interests which
the State claims to be protecting, and the interests of those who are
disadvantaged by the classification. In the present situation the state laws
place burdens on two different, although overlapping, kinds of rights—the
right of individuals to associate for the advancement of political beliefs, and the
right of qualified voters, regardless of their political persuasion, to cast their votes
effectively. Both of these rights, of course, rank among our most precious
freedoms. We have repeatedly held that freedom of association is
protected by the First Amendment. And of course this freedom protected
against federal encroachment by the First Amendment is entitled under
the Fourteenth Amendment to the same protection from infringement by
the States. Similarly we have said with reference to the right to vote: “No
right is more precious in a free country than that of having a voice in the
54
election of those who make the laws under which, as good citizens, we
must live. Other rights, even the most basic, are illusory if the right to vote
is undermined.”
Id. at 30–31 (emphasis added) (citations omitted). The Court held that the law in
question was unconstitutionally burdensome on new political parties. Id. at 34. 165
We therefore believe that there is a solid basis for considering the associational
aspect of voting in assessing the gravamen of the harm allegedly suffered by the
plaintiffs. Indeed, in this case, the associational harm is especially important to the
analysis because the testimony of the defendants’ witnesses as well as the plaintiffs’
demonstrate that, given the legislative practice and custom of Wisconsin, legislative
action is controlled, as a practical matter, solely by the majority caucus. In such a
circumstance, when the state places an artificial burden on the ability of voters of a
certain political persuasion to form a legislative majority, it necessarily diminishes the
weight of the vote of each of those voters when compared to the votes of individuals
favoring another view. The burdened voter simply has a diminished or even no
opportunity to effect a legislative majority. That voter is, in essence, an unequal
participant in the decisions of the body politic.
On the facts presented in past cases, some members of the Supreme Court have
expressed the view that judicial enforcement of the principle that each voter has a right
to have his vote treated equally must be limited to situations where the dilution is based
on classifications such as race and population. These reservations have been grounded
in the concern that distinguishing between legitimate and illegitimate political
motivations is not a task to be undertaken by judges. In their view, moreover, there are
insurmountable problems in formulating manageable standards. See Bandemer, 478 U.S.
at 147 (O’Connor, J., concurring in the judgment); Vieth, 541 U.S. at 288 (plurality
opinion). Other Justices have not accepted such a limitation. See, e.g., Vieth, 541 U.S. at
306–17 (Kennedy, J., concurring in the judgment). As we shall discuss at greater length
later, however, this case does not present these conundrums. We are not presented with
the problem of distinguishing between permissible and impermissible political
motivations. We have a far more straight-forward situation. The plaintiffs have
established, on this record, that the defendants intended and accomplished an
In subsequent cases, the Court similarly assessed claims under the Equal Protection Clause. See, e.g.,
Am. Party of Texas v. White, 415 U.S. 767, 788–89 (1974) (holding that a requirement that minor parties
obtain signatures equivalent to 1% of the votes in the previous election was not unconstitutional); Storer v.
Brown, 415 U.S. 724, 733–34 (1974) (holding that a law which required an independent candidate to not
have been affiliated with a political party for a year for before the party “involves no discrimination” and
was not unconstitutional); Bullock v. Carter, 405 U.S. 134, 149 (1972) (holding that the imposition of filing
fees in order to seek the nomination of a party constituted a constitutional violation).
165
55
entrenchment of the Republican Party likely to endure for the entire decennial period.
They did so when the legitimate redistricting considerations neither required nor
warranted the implementation of such a plan.
IV
ELEMENTS OF THE CAUSE OF ACTION
As our description of the case law reveals, the law governing political
gerrymandering, still in its incipient stages, is in a state of considerable flux. We must,
however, accept that situation and seek in these authorities a solution to the case before
us. Therefore, while not discounting the difficulty of the task before us, we now identify
the guideposts available to us.
We begin with a principle that is beyond dispute. State legislative apportionment
is the prerogative and therefore a duty of the political branches of the state government.
We must “recognize[] the delicacy of intruding on this most political of legislative
functions.” Bandemer, 478 U.S. at 143. 166 We also know that we cannot rely on the simple
finding “that political classifications were applied.” Vieth, 541 U.S. at 307 (Kennedy, J.,
concurring in the judgment). Similarly, “the mere lack of proportional representation
will not be sufficient to prove unconstitutional discrimination.” Bandemer, 478 U.S. at
132 (plurality opinion).
It is clear that the First Amendment and the Equal Protection Clause protect a
citizen against state discrimination as to the weight of his or her vote when that
discrimination is based on the political preferences of the voter.167 This principle applies
not simply to disparities in raw population, but also to other aspects of districting that
“operate to minimize or cancel out the voting strength of racial or political elements of
the voting population.” Fortson, 379 U.S. at 439. Specifically, apportionment plans that
“invidiously minimize[]” the voting strength of “political groups” “may be vulnerable”
to constitutional challenges, Gaffney, 412 U.S. at 754, because “each political group in a
State should have the same chance to elect representatives of its choice as any other
political group,” Bandemer, 478 U.S. at 124.
Cf. Vieth, 541 U.S. at 306 (Kennedy, J., concurring in the judgment) (cautioning against “the correction
of all election district lines drawn for partisan reasons” because that course “would commit federal and
state courts to unprecedented intervention in the American political process”).
166
Cf. Reynolds, 377 U.S. at 565 (“Any suggested criteria for the differentiation of citizens are insufficient to
justify any discrimination, as to the weight of their votes, unless relevant to the permissible purpose of
legislative apportionment.”).
167
56
We conclude, therefore, that the First Amendment and the Equal Protection
clause prohibit a redistricting scheme which (1) is intended to place a severe
impediment on the effectiveness of the votes of individual citizens on the basis of their
political affiliation, (2) has that effect, and (3) cannot be justified on other, legitimate
legislative grounds.
A.
Discriminatory Intent or Purpose
The Supreme Court has stressed the “basic equal protection principle that the
invidious quality of a law … must ultimately be traced to a discriminatory purpose.”
Washington v. Davis, 426 U.S. 229, 240 (1976); see also Vill. of Arlington Heights v. Metro.
Hous. Dev. Corp., 429 U.S. 252, 265 (1977) (“Proof of … discriminatory intent or purpose
is required to show a violation of the Equal Protection Clause.”). A legislature’s
discriminatory intent also factors into a First Amendment analysis. Timmons, 520 U.S. at
358–59 (considering whether a state has imposed “reasonable, nondiscriminatory
restrictions” on First Amendment associational rights (emphasis added)); see also Wash.
State Grange v. Wash. State Republican Party, 552 U.S. 442, 452 (2008) (same); Ashcroft v.
Iqbal, 556 U.S. 662, 676 (2009) (“Where the claim is invidious discrimination in
contravention of the First … Amendment[], our decisions make clear that the plaintiff
must plead and prove that the defendant acted with discriminatory purpose.”).
The Court explicitly has held that equal protection challenges to redistricting
plans require a showing of discriminatory purpose or intent. See Rogers, 458 U.S. at 617
(explaining that cases involving allegations of vote-dilution on the basis of race “are …
subject to the standard of proof generally applicable to Equal Protection Clause cases”
including a showing of a “‘a racially discriminatory purpose’” (quoting Washington, 426
U.S. at 240)). This requirement applies with equal force to cases involving political
gerrymanders. See Bandemer, 478 U.S. at 127 (stating that plaintiffs who bring a claim of
partisan gerrymandering “[a]re required to prove … intentional discrimination against
an identifiable political group” (emphasis added)).
1.
When considering the level of partisan intent necessary to establish a
political-gerrymandering claim, our first task is to determine what kind of partisan
intent offends the Constitution. The plurality in Bandemer simply required a plaintiff to
show any level of “intentional discrimination against an identifiable political group.”
478 U.S. at 127; see also Vieth, 541 U.S. at 284 (plurality opinion) (describing the Bandemer
plurality’s standards as “mere intent to disadvantage the plaintiff’s group”). It
suggested that “[a]s long as redistricting is done by a legislature, it should not be very
57
difficult to prove that the likely political consequences of the reapportionment were
intended.” Bandemer, 478 U.S. at 129. A majority of the Court in Vieth, however, rejected
the Bandemer plurality’s test, which included this standard of intent. Vieth, 541 U.S. at
284 (plurality opinion) (“declin[ing] to affirm [the Bandemer test] as a constitutional
requirement”); id. at 308 (Kennedy, J., concurring in the judgment) (noting that “[t]he
plurality demonstrates the shortcomings of the other standards that have been
considered to date” and specifically identifying “the standards proposed in Davis v.
Bandemer”).
At the outset, we note that the Court recently has acknowledged that the
constitutionality of partisan favoritism in redistricting is an open question. See Harris v.
Arizona Indep. Redistricting Comm'n, 136 S. Ct. 1301, 1310 (2016) (“assuming, without
deciding, that partisanship is an illegitimate redistricting factor”). Nevertheless, we
know that legislatures may employ some political considerations when making
redistricting decisions; considerations such as achieving “political fairness,” Gaffney, 412
U.S. at 752, and “avoiding contests between incumbent[s],” Bush v. Vera, 517 U.S. 952,
964 (1996) (internal quotation marks omitted) (quoting Karcher v. Daggett, 462 U.S. 725,
740 (1983)) (alteration in original), are permissible.
That some political considerations may intrude into the redistricting process
without running afoul of the Constitution, however, does not answer the question
whether partisan favoritism is permissible. The Court’s members appear to acknowledge
that some level of partisanship is permissible, or at least inevitable, in redistricting
legislation. The plurality in Vieth, for instance, noted that “partisan districting is a
lawful and common practice.” 541 U.S. at 286. In his opinion, Justice Kennedy observed
that political classifications are “generally permissible.” Id. at 307 (Kennedy, J.,
concurring in the judgment). Justices Souter and Breyer, dissenting in Vieth, expressed
the view that partisan favoritism in some form was inevitable, if not necessarily
desirable. See id. at 344 (Souter, J., dissenting) (“[S]ome intent to gain political advantage
is inescapable whenever political bodies devise a district plan ….”); id. at 360 (Breyer, J.,
dissenting) (“[T]raditional or historically based boundaries are not, and should not be,
‘politics free.’ … They … represent an uneasy truce, sanctioned by tradition, among
different parties seeking political advantage.” (emphasis added)).
Other justices, however, have not acknowledged that political affiliation is “an
appropriate standard for excluding voters from a congressional district.” Id. at 325
(Stevens, J., dissenting). Even so, these justices have proposed tests that “cover only a
58
few meritorious claims” and “preclude extreme abuses” of the districting process. Id. at
339. 168
As a starting point, it is safe to say that this concept of abuse of power seems at
the core of the Court’s approach to partisan gerrymandering. In Arizona State Legislature
v. Arizona Independent Redistricting Commission, 135 S. Ct. 2652, 2658 (2015), the Court
defined partisan gerrymandering as “the drawing of legislative district lines to
subordinate adherents of one political party and entrench a rival party in power.”
Justice Kennedy noted in Vieth that a claim of partisan gerrymandering “must rest … on
a conclusion that [political] classifications … were applied in an invidious manner or in a
way unrelated to any legitimate legislative objective.” 541 U.S. at 307 (Kennedy, J.,
concurring in the judgment) (emphasis added). The plurality, as well, acknowledged
that “an excessive injection of politics is unlawful.” Id. at 293 (plurality opinion). And
Justice Breyer in dissent observed that there was “at least one circumstance where use
of purely political boundary-drawing factors can amount to a serious, and remediable,
abuse, namely, the unjustified use of political factors to entrench a minority in power.”
Id. at 360 (Breyer, J., dissenting) (emphasis in original). 169
When “acceptable”—or at least tolerable—crosses a line to become “excessive,”
however, remains unclear. Moreover, as Justice Kennedy warns, a standard of
excessiveness has its drawbacks:
[C]ourts must be cautious about adopting a standard that turns on
whether the partisan interests in the redistricting process were excessive.
Excessiveness is not easily determined. Consider these apportionment
schemes: In one State, Party X controls the apportionment process and
draws the lines so it captures every congressional seat. In three other
States, Party Y controls the apportionment process. It is not so blatant or
egregious, but proceeds by a more subtle effort, capturing less than all the
seats in each State. Still, the total effect of party Y’s effort is to capture
more new seats than Party X captured. Party X’s gerrymander was more
To address the inevitability of partisan favoritism, Justice Souter, like Justice Stevens, proposed a more
rigorous “effects” analysis. Vieth, 541 U.S. at 347–50 (Souter, J., dissenting) (observing that “under a plan
devised by a single major party, proving intent” under his test “should not be hard, … politicians not
being politically disinterested or characteristically naïve.”). Alternatively, Justice Breyer proposed a
standard that addressed “circumstance[s] where use of purely political boundary-drawing factors can
amount to a serious, and remediable abuse.” Id. at 360 (Breyer, J., dissenting).
168
See also LULAC, 548 U.S. at 419 (opinion of Kennedy, J.) (observing that “a congressional plan that
more closely reflects the distribution of state party power seems a less likely vehicle for partisan
discrimination than one that entrenches an electoral minority”).
169
59
egregious. Party Y’s gerrymander was more subtle. In my view, however,
each is culpable.
Id. at 316 (Kennedy, J., concurring in the judgment).
“Excessiveness” does not need to be defined simply in terms of raw seat tallies.
The danger with extreme partisan gerrymanders is that they entrench a political party in
power, making that party—and therefore the state government—impervious to the
interests of citizens affiliated with other political parties. This imperviousness may be
achieved by manipulating a map to achieve a supermajority. But it also may be
achieved by “lock[ing]-in” or creating the requisite “safe seats” such that legislators
“elected from such safe districts need not worry much about the possibility of shifting
majorities” and “have little reason to be responsive to the political minorities within
their district.” LULAC, 548 at 470–71 (Stevens, J., concurring in part and dissenting in
part).
When a party is “locked-in” through the intentional manipulation of legislative
districts, “representational harms” to those affiliated with the “out”-party necessarily
ensue. See id. at 470. Specifically, “in addition to the possibility that a representative
may believe her job is only to represent the interests of a dominant constituency, a
representative may feel more beholden to the cartographers who drew her district than
to the constituents who live there.” Id. The result is a system that assigns different
weights to the votes of citizens and accords to those citizens different levels of
legislative responsiveness based on the party with which they associate. See Reynolds,
377 U.S. at 565.
Whatever gray may span the area between acceptable and excessive, an intent to
entrench a political party in power signals an excessive injection of politics into the
redistricting process that impinges on the representational rights of those associated
with the party out of power. Such a showing, therefore, satisfies the intent requirement
for an equal protection violation. 170
2.
A “‘discriminatory purpose’ … implies more than intent as volition or intent as
awareness of consequences. It implies that the decisionmaker … selected or reaffirmed a
particular course of action at least in part, ‘because of,’ not merely ‘in spite of,’ its
The intent we require, therefore, is not simply an “intent to act for political purposes,” Dissent
at 120, but an intent to make the political system systematically unresponsive to a particular
segment of the voters based on their political preference.
170
60
adverse effects upon an identifiable group.” Pers. Admin. of Mass. v. Feeney, 442 U.S. 256,
279 (1979); see also Chavez v. Ill. State Police, 251 F.3d 612, 645 (7th Cir. 2001) (quoting
same). The plaintiffs therefore must show that the intent to entrench the Republican
Party in power was “a motivating factor in the decision.” Arlington Heights, 429 U.S. at
265–66. It need not be the “sole[]” intent or even “the ‘dominant’ or ‘primary’ one.” Id.
at 265. 171 Indeed, it rarely can “be said that a legislature or administrative body
operating under a broad mandate made a decision motivated by a single concern.” Id.
This is certainly true in redistricting legislation where the Court has identified
“traditional districting principles such as compactness, contiguity, and respect for
In an “analytically distinct” line of cases, the Supreme Court has required that plaintiffs establish that
the discriminatory motive be the legislature’s “predominant” intent. Miller v. Johnson, 515 U.S. 900, 911,
917 (1995). These cases, beginning with Shaw v. Reno, 509 U.S. 630 (1993), concern the use of racial
classifications in the drawing of district lines. Specifically, in Shaw, the plaintiffs had “alleged that the
General Assembly deliberately ‘create[d] two Congressional Districts in which a majority of black voters
was concentrated along racial lines’” and “to assure the election of two black representatives to
Congress.” Id. at 637. The Court held that such classifications were subject to strict scrutiny. Although
these voting schemes did not dilute the voting strength of racial minorities, they nevertheless resulted in
“special harms that are not present in … vote-dilution cases,” which “warrant[ed] [a] different analysis.”
Id. at 649–50. In the Court’s view, classifying voters on the basis of race “reinforce[d] racial stereotypes
and threaten[ed] to undermine our system of representative democracy by signaling to elected officials
that they represent a particular racial group rather than their constituency as a whole.” Id. at 650.
171
In Miller, the Court reiterated the special harms in such cases:
Just as a State may not … segregate citizens on the basis of race in its public parks, so …
it may not separate its citizens into different voting districts on the basis of race. …
Race-based assignments “embody stereotypes that treat individuals as the product of
their race, evaluating their thoughts and efforts—their very worth as citizens—according
to a criterion barred to the Government by history and the Constitution.”
515 U.S. at 911–12 (citations omitted). To establish this kind of equal protection claim, the Court
continued, the “plaintiff’s burden is to show … that race was the predominant factor motivating the
legislature’s decision to place a significant number of voters within or without a particular district.” Id. at
916.
The Shaw line of cases does not speak directly to the political gerrymandering case before us. In
those cases, the Court particularly was concerned about “racial stereotypes,” Shaw, 509 U.S. at 647–48;
Miller, 515 U.S. at 911–12, which are not present here. Relatedly, applying a “special harms” analysis to
the partisan gerrymandering context would call into question bipartisan districting plans designed to
create parity between the parties; the Court, however, clearly has held that “partisan fairness” is a
legitimate consideration in crafting legislative districts. See supra at 34–36 (discussing Gaffney, 412 U.S.
735). Finally, the Court has rejected the “predominant intent” standard in the context of political
gerrymandering claims. See supra at 40–42 (discussing plurality opinion in Vieth, 541 U.S. at 284–86) and
43–45 (discussing Justice Kennedy’s concurrence in Vieth, 541 U.S. at 306–08).
61
political subdivisions” that legitimately may inform drafters in the drawing of district
lines. Shaw v. Reno, 509 U.S. 630, 647 (1993).
Relying on traditional districting principles, defendants propose a novel rule: a
redistricting plan that “is consistent with, and not a radical departure from, prior plans
with respect to traditional districting principles” cannot, as a matter of law, evince an
unconstitutional intent. 172 In other words, compliance with traditional districting
principles necessarily creates a constitutional “safe harbor” for state legislatures.
The defendants’ approach finds no support in the law. It is entirely possible to
conform to legitimate redistricting purposes but still violate the Fourteenth Amendment
because the discriminatory action is an operative factor in choosing the plan. Indeed,
the Court rejected a similar claim in Fortson: while acknowledging that there was no
“mathematical disparity” that violated the principle of “one-person, one-vote,” it did
not rule out the possibility that a districting plan, which included multimember
districts, could “operate to minimize or cancel out the voting strength of racial or
political elements of the voting population.” 379 U.S. at 439. Similarly, in Gaffney, the
Court observed that “[s]tate legislative districts may be equal or substantially equal in
population and still be vulnerable under the Fourteenth Amendment.” 412 U.S. at 751.
Moreover, the Court has made clear that “traditional districting principles” are
not synonymous with equal protection requirements. Instead, they “are objective
factors that may serve to defeat a claim that a district has been gerrymandered.” Shaw,
509 U.S. at 647 (citing Gaffney, 412 U.S. at 752 n.18). In other words, they are
constitutionally permissible, but not “constitutionally required.” Id. Individual Justices
also have noted that a map’s compliance with traditional districting principles does not
necessarily speak to whether a map constitutes a partisan gerrymander:
[E]ven those criteria that might seem promising at the outset (e.g.,
contiguity and compactness) are not altogether sound as independent
judicial standards for measuring a burden on representational rights. They
cannot promise political neutrality when used as the basis for relief.
Vieth, 541 U.S. at 308 (Kennedy, J., concurring in the judgment); see also id. at 366
(Breyer, J., dissenting) (opining that a map where “no radical departure from traditional
districting criteria is alleged” but an unjustified partisan result occurs in two elections
“would be sufficient to support a claim of unconstitutional entrenchment”). Highly
sophisticated mapping software now allows lawmakers to pursue partisan advantage
without sacrificing compliance with traditional districting criteria. A map that appears
congruent and compact to the naked eye may in fact be an intentional and highly
R.153 at 5; see also R.156 at 1 (“[A] democratically-enacted districting plan … is entirely lawful when it
complies with traditional districting principles.”).
172
62
effective partisan gerrymander. When reviewing intent, therefore, we cannot simply ask
whether a plan complied with traditional districting principles. Therefore, the
defendants’ contention—that, having adhered to traditional districting principles, they
have satisfied the requirements of equal protection—is without merit. 173
We therefore must confront the question of how we are to discern whether, in
creating the map that became Act 43, the drafters employed an impermissible intent—
cutting out for the longterm those of a particular political affiliation. In assuming this
task, we are mindful that “[i]nquiries into congressional [and other legislative bodies’]
motives or purposes are a hazardous matter.” United States v. O’Brien, 391 U.S. 367, 383
(1968). When the issue is one of “mixed intent” as it is here, “[e]valuating the legality of
acts … can be complex … . When the actor is a legislature and the act is a composite of
manifold choices, the task can be even more daunting.” LULAC, 548 U.S. at 418 (opinion
of Kennedy, J.). “Determining whether invidious discriminatory purpose was a
motivating factor demands a sensitive inquiry into such circumstantial and direct
evidence of intent as may be available,” including (1) “[t]he impact of the official
action” as “an important starting point”; (2) “the historical background of the decision”;
(3) “[t]he specific sequence of events leading up to the challenged decision”; (4)
“[d]epartures from the normal procedural sequence”; (5) “legislative or administrative
history … , especially … contemporary statements by members of the decisionmaking
The Dissent relies on the opinion of Justice Stevens concurring in the summary affirmance in Cox v.
Larios, 542 U.S. 947 (2004), to “reinforce[],” Dissent at 131, its conclusion that oddly shaped districts are a
necessary component of a claim that a partisan gerrymander violates the Constitution. Respectfully, that
reliance is misplaced. Cox is, as the Dissent notes, a malapportionment case, and it was affirmed because,
as the district found, and Justice Stevens repeated, “[t]he numbers largely speak for themselves.” 542 U.S.
at 948 (alteration in original) (internal quotation marks omitted). The unusual shapes in the map
“supplied further evidence,” id., but were by no means essential to the result.
173
Moreover, the two Justices concurring in the summary affirmance went on to note that the map’s
“selective incumbent protection” and related incumbent pairings, done for partisan gain, would have
violated any partisan gerrymandering standard the Court could have adopted in Vieth, where the
gerrymander was “visible to the judicial eye.” Id. at 949–950. Read in context, we believe this language
refers clearly to the concerning feature of intentional incumbent pairings, not the shape of the districts.
In any event, the Justices continued, “[d]rawing district lines that have no neutral justification in
order to place two incumbents of the opposite party in the same district is probative of the same
impermissible intent” as prior case involving oddly shaped districts. Id. at 950. Plainly, this language does
not make odd shapes a necessary part of a claim; it merely shows that it is a permissible way for a
plaintiff to show intent. Indeed, we read this passage not to confirm, as the Dissent does, a shape-based
analysis, but to confirm a separate point disputed by the Dissent: that intent is a requirement of a
unconstitutional gerrymandering claim. True enough, a case involving odd shapes presents an easier
claim, both to prove and to adjudicate. But the complexities of proving a case without these shapes are
not fatal to the claim.
63
body, minutes of its meetings, or reports.” Arlington Heights, 429 U.S. at 266–68; see also
Miller, 515 U.S. at 913–14.
However, discerning the intent of a legislative body can be less daunting in some
cases than in others. In some cases, the legislature is aware that a distinction is
constitutionally impermissible and surreptitiously attempts to create legislation on the
basis of that distinction. These cases require that we engage in a careful inquiry of
circumstantial evidence, because the drafters’ intent often is hidden from the casual
observer. 174 In other cases, a legislature seems unaware that a distinction is
constitutionally impermissible and deliberately enacts legislation on the basis of that
distinction. This situation typically arises in periods before the Supreme Court has
illuminated the full meaning of a constitutional right. 175 In these cases, courts are able to
discern the legislature’s intent more easily and less intrusively because the evidence is
far more direct.
This case falls more in the latter category. The Court never has invalidated a
redistricting plan on the ground of partisan gerrymandering, and the Court’s recent
pronouncements have caused some district courts to question the viability of the cause
of action. 176 Here, the record demonstrates that, although the drafters were aware of
some constitutional limits on the degree to which they could neutralize the political
power of the opposition party, those limits were not firmly established.
We therefore turn to the sequence of events that led to the enactment of Act 43 to
discern whether one purpose behind the legislation was to entrench a political party in
power.
3.
a.
Evidence of intent
The evidence at trial establishes that one purpose of Act 43 was to secure the
See, e.g., Rogers, 458 U.S. at 623–25 (relying on circumstantial evidence of intent in a case of racial voting
dilution and noting that “the evidence shows that discriminatory practices were commonly utilized, that
they were abandoned when enjoined by courts or made illegal by civil rights litigation, and that they
were replaced by laws and practices which, though neutral on their face, serve to maintain the status
quo”).
174
See, e.g., Miller, 515 U.S. at 918 (noting that the State had conceded that it had drawn lines on the basis
of race and concluding that the district court therefore “had little difficulty” determining the legislature’s
intent).
175
See, e.g., Radogno v. Illinois St. Bd. of Elections, No. 1:11-cv-04884, 2011 WL 5025251, at *4–6 (N.D. Ill.
Oct. 21, 2011) (recounting district courts’ approaches to political gerrymandering claims).
176
64
Republican Party’s control of the state legislature for the decennial period. The drafters’
concern with the durable partisan complexion of the new Assembly map was present
from the outset of the legislative process. Ottman, Foltz, 177 and Handrick began drafting
the map that would become Act 43 in April 2011. 178 One of their first orders of business
was to develop a composite partisan score that accurately reflected the political makeup of population units, which would allow them to assess the partisan make-up of the
new districts. 179 When they came up with a composite of “all statewide races from
[20]04 to 2010” that “seem[ed] to work well,” they sent it to Professor Gaddie. 180
Professor Gaddie, 181 the “advisor on the appropriate racial and/or political
make-up of legislative … districts,” 182 “buil[t] a regression model … to test the partisan
makeup and performance of districts as they might be configured in different ways.” 183
Professor Gaddie then tested the drafters’ composite measure against his model and
confirmed that their measure was “almost a perfect proxy for the open seat vote, and
the best proxy you’ll come up with.” 184 Professor Mayer testified that the drafters’
composite measure correlated very strongly with his own measure of partisanship,
which led him to conclude that “they knew exactly what they were doing, that they had
a very accurate estimate of the underlying partisanship of the Act 43 maps.” 185
Once Ottman, Foltz, and Handrick received Professor Gaddie’s imprimatur on
their composite measure, they employed this measure “to assess the partisan impact of
the map[s] that [they] drew.” 186 We find that the maps the drafters generated, as well as
the statistical comparisons made of the various maps, reveal that a focal point of the
drafters’ efforts was a map that would solidify Republican control. The maps often bore
With some limited exceptions, we find Ottman to be a credible witness. We have less confidence in
Foltz’s testimony, which appeared to us rehearsed and guarded. Throughout our discussion, we will note
those areas of testimony which we find unworthy of credence.
177
178
See R.148 at 68.
179
See Tr. Ex. 175, at 1–2.
180
Id. at 2.
As noted earlier, Professor Gaddie’s testimony was offered through a video deposition. We find his
testimony credible.
181
182
Tr. Ex. 169 (emphasis added).
183
Tr. Ex. 161 (Gaddie Dep.), at 46.
184
Tr. Ex. 175, at 1.
185
R.148 at 192; see also id. at 207–09.
186
See R.147 at 61; R.148 at 15–16.
65
names that reflected the level of partisan advantage achieved. For instance, maps
labeled “aggressive” referenced “a more aggressive map with regard to GOP
leaning.” 187 When producing these more advantageous maps, the drafters did not
abandon traditional districting criteria; 188 to the contrary, the maps complied with
traditional districting criteria while also ensuring a significant partisan advantage.
The drafters also created spreadsheets that collected the partisan scores, by
district, for each of the map alternatives. For each spreadsheet, there was a
corresponding table that listed the number of “Safe” Republican seats, “Lean”
Republican seats, “Swing” seats, “Safe” Democratic seats, and “Lean” Democratic seats;
these figures also were compared to the number of seats in each category under the
Current Map, the map drawn by the court in Baumgart v. Wendelberger, Nos. 01-C-0121
& 02-C-0366, 2002 WL 34127471 (E.D. Wis. May 30, 2002), amended by 2002 WL 34127473
(E.D. Wis. July 11, 2002). 189
The process of drafting and evaluating these alternative district maps spanned
several months. In April, the drafters produced a document comparing the partisan
performance of the Current Map to two early draft maps: Joe’s Basemap Basic and Joe’s
Basemap Assertive. 190 Under the Current Map, the drafters anticipated that the
Republicans would win 49 191 Assembly seats. 192 This number increased to 52 under the
Joe’s Basemap Basic map and to 56 under the Joe’s Basemap Assertive map. 193 The
number of safe and leaning Republican seats increased from 40 under the Current Map
to 45 under the Joe’s Basemap Basic map and 49 under the Joe’s Basemap Assertive
map; the number of swing seats decreased from 19 to 14 to 12. 194 The number of safe
R.147 at 65. Also during the drafting process, Ottman met with individual senators to review with
them the census numbers and to obtain general information about their districts. One senator suggested
to Ottman how her district could be re-drawn to take the seat away from a Democratic member of the
Assembly: “If you need a way to take the Staskunas seat, put a little bit of my Senate seat into New Berlin
(2–3 wards could make that a GOP Assembly seat).” Tr. Ex. 239.
187
188
See supra at 10.
189
See, e.g., Tr. Ex. 364.
190
Tr. Ex. 465.
Id. These consisted of the “strong gop” seats, the “lean gop” seats, and approximately one-half of the
“toss up” seats.
191
192
Id.
193
Id.
194
Id.
66
and leaning Democratic seats, however, remained roughly the same under all three
maps, hovering between 38 and 40. 195
The drafters prepared and evaluated the partisan performance of at least another
six statewide alternative maps. 196 Each of these maps improved upon the anticipated
pro-Republican advantage generated in the initial two draft plans. The total number of
expected Republican seats now ranged between 57 and 60, and the number of swing
seats was diminished to between 6 and 11. 197 The number of Democratic seats again
remained about the same under each draft map. 198
The drafters sent their completed draft maps to Professor Gaddie, who created a
visual “S” curve for each map. 199 These “S” curves show how each map would operate
within an array of electoral outcomes. 200 To produce the “S” curves, Professor Gaddie
calculated the expected partisan vote shares for each district. 201 He then shifted the vote
share of each district ten points in either direction, from 40% to 60%, and assigned a
color to districts that “lean[ed]” towards, or were “safe” seats for that party. 202 Professor
Gaddie explained that his analysis “was designed to tease out a potential estimated
vote” under a range of electoral scenarios, when either “the Democrats have a good
year” or “the Republicans have a good year.” 203 At bottom, the “S” curves—at least
some of which were printed in large format and kept in the map room—allowed a
non-statistician, by mere visual inspection, to assess the partisan performance of a
particular map under all likely electoral scenarios. On one occasion, Professor Gaddie
195
Id. On the “Tale of the Tape,” see Tr. Ex. 283, the drafters did note that “Criteria to Monitor” included
total population deviation, split municipalities, split counties, incumbent pairings, and senate
disenfranchisement. Additionally, Ottman created some spreadsheets that looked at disenfranchisement.
See Tr. Ex. 225 (WRK32587 Responsive Spreadsheets). However, the defendants have not pointed to any
evidence in the record that suggests that measures of traditional districting criteria were being scrutinized
on a regular basis or with the intensity that partisan scores were being evaluated.
These were: Milwaukee_Gaddie_4_16_11_V1_B (Tr. Ex. 172, at 1); Statewide2_Milwaukee_Gaddie_
4_16_V1_B (Tr. Ex. 172, at 2); Tad MayQandD (Tr. Exs. 364, 477); Joe Assertive (Tr. Exs. 366, 478); Tad
Aggressive (Tr. Ex. 283); and Adam Aggressive (Tr. Ex. 283).
196
197
Tr. Exs. 172, 364.
198
Tr. Exs. 364, 366.
199
Tr. Ex. 134; see Tr. Exs. 263–82.
200
Tr. Ex. 161 (Gaddie Dep.), at 45.
201
Id. at 44.
202
Id. at 150–51.
203
Id. at 101.
67
showed the “S” curves to Senator Fitzgerald and explained to the Senator “how to
interpret” them. 204
Over several days in early June, the drafters presented a selection of regional
maps drawn from their statewide drafts, approximately three to four per region, to the
Republican leadership. Along with these regional alternatives, the leadership “saw the
partisan scores for the maps that [the drafters] presented to them in those
alternatives.” 205 Foltz testified during his deposition that although he could not recall a
particular example, he was sure that he was asked by the leadership about the partisan
performance of the various regional options. 206
Following this meeting, the drafters amalgamated the regional alternatives
chosen by the leadership. Foltz testified that “the draft map called team map emerged
as a result of the … leadership’s choices at those meetings.” 207 Under the Team Map,
which was also referred to as the “Final Map,” 208 the Republicans could expect to win 59
Assembly seats, with 38 safe Republican seats, 14 leaning Republican, 10 swing, 4
leaning Democratic, and 33 safe Democratic seats. 209 In the Tale of the Tape, the drafters
compared the partisan performance of the Team Map directly to the Current Map on
each of these criteria. 210 They highlighted specifically that under the Current Map, “49
seats are 50% or better,” but under the Team Map, “59 Assembly seats are 50% or
better.” 211
The Team Map underwent even more intense partisan scrutiny in a document
204
Id. at 75.
205
R.148 at 20.
206
Tr. Ex. 191 (Foltz Dep.), at 106.
207
R.147 at 80.
As we noted earlier, Foltz testified that if the “Team Map” was not “the final one that was pushed, put
forward in the public domain, it was very close to it, and it was the result of that mashing process of
taking the various regional alternatives and putting them all together.” Id. at 165. He further explained
that the “Final Map” was the one “after the leaders got together and made the regional decisions and they
were then merged together.” Id. at 62. If it was not identical to the map that “ultimately became Act 43,
it[ wa]s probably fairly close.” Id.; see also Tr. Ex. 172, at 3–4; supra note 56.
208
The drafters in fact produced and evaluated several distinct versions of the Team Map, but each
rendition is virtually identical. See Tr. Ex. 172, at 3–4 (Final Map); Tr. Ex. 467, at 1 (Team Map (Joe
Aggressive)); id. at 2 (Team Map Ranking (Joe Aggressive 2)); id. at 3 (Team Map (6-15-11)).
209
210
Tr. Ex. 283.
211
Id.
68
identified as “summary.xlsx.” 212 The drafters divided the new Team Map districts into
six categories of partisan performance, listing beside each district its “new incumbent”
and its Republican vote share under the Current Map and the Team Map; the change in
Republican vote share was the district’s “improvement” under the new plan. 213 The
drafters considered five districts to be “Statistical Pick Up[s],” meaning they were
currently held by a Democratic incumbent but “move[d] to 55% or better” in
Republican vote share under the new Team Map. 214 Fourteen districts were grouped
under the heading “GOP seats strengthened a lot,” meaning they were “[c]urrently held
GOP seats that start[ed] at 55% or below that improve[d] by at least 1%” in Republican
vote share. 215 Eleven districts were “GOP seats strengthened a little,” meaning they
“improve[d] less than 1%.” 216 Only three districts were labeled “GOP seats weakened a
little,” meaning they had “start[ed] at 55% or below” but “decline[d]” slightly in
Republican vote share. 217 Another three districts were “GOP seats likely lost,” meaning
they had “drop[ped] below 45%” Republican vote share under the Team Map. 218
Finally, the drafters noted four districts where Democrats were “weakened,” which
were districts with “45% or better” Democratic vote share “that bec[a]me more GOP”
under the Team Map. 219 The drafters also identified twenty Republican Assembly
members who enjoyed sufficiently comfortable partisan scores such that they could
become “GOP donors to the team.”220 These were members of the Assembly who had
partisan scores of 55% or greater and, therefore, could spread their partisan voting
strength to politically weaker colleagues. 221
212
Tr. Ex. 284, at 1.
213
Id.
214
Id.
215
Id.
216
Id.
217
Id.
218
Id.
219
Id. at 1–2.
Id. at 1. In his testimony concerning his post-drafting meetings with individual senators, see supra at 15,
Ottman did not identify any senators who were reluctant to be “donors to the team.”
220
See supra at 14. In his testimony, Ottman stated that the “GOP donor” designation “simply indicate[d] a
seat that had a lower percentage under that partisan metric than it started with.” R.148 at 29. We do not
believe that this answer can be reconciled fully with the information on the spreadsheet. Specifically, the
spreadsheet states that donors are “[i]ncumbents with numbers above 55% that donate to the team.”
Tr. Ex. 284, at 1. The inclusion, on the spreadsheet, of the strength of the donors’ numbers, strongly
suggest that they had political strength to spare and to share with other, perhaps more vulnerable,
221
69
The Team Map also was sent to Professor Gaddie. The “S” curve demonstrates
that this map would allow the Republicans to maintain a comfortable majority under
likely voting scenarios; their statewide vote share could fall to 48%, and they still would
preserve a 54 seat majority in the Assembly. The Democrats, by contrast, would need
54% of the statewide vote to capture a simple majority of Assembly seats. 222
Once the map had been finalized, Foltz presented each Republican member of
the Assembly with information on his or her new district. These memos provided a
“[c]omparison of [k]ey [r]aces” in the new districts compared to the old. 223 Specifically,
the memoranda detailed what percentage of the population in the old and new districts
voted for Republican candidates in representative statewide and national elections held
since 2004. Importantly, the memoranda did not provide the individual legislators with
any information about contiguity, compactness, or core population.
Additionally, Ottman made a presentation to the Republican caucus that
highlighted the long-term effects of Act 43, as reflected in his prepared notes: “The
maps we pass will determine who’s here 10 years from now,” and “[w]e have an opportunity
and an obligation to draw these maps that Republicans haven’t had in decades.” 224
districts.
See Tr. Ex. 282. Professor Mayer also conducted a swing analysis that evaluated the outcome of Act 43
under likely electoral scenarios. He, like Professor Gaddie, concluded that, under Act 43, “even when the
Democrats receive 54 percent of the statewide vote, they still aren’t even close to a majority of the
Assembly.” R.148 at 229.
222
223
Tr. Ex. 342.
Tr. Ex. 241, at 1 (emphasis added). Similarly, Ottman created talking-points memos for his meetings
with Republican members of the Senate. These memos included information about population, where
changes in the district’s population had occurred, and the geography of the new district, see, e.g., Tr. Ex.
242, at 1 (“Added East Troy and part of the town, as well as Mukwonago.”). Importantly, these also
contained information on how the re-configured district had voted in national and statewide elections.
See id. (noting, for example, that “Scott Walker won this new seat with 64.2%,” “McCain won with
51.5%,” and “Van Hollen 06 won with 59.4%”).
224
At trial, counsel for the plaintiffs cross-examined Ottman on statements that he had made during
the joint public hearing on Act 43, which was held on July 13, 2011. See R.148 at 44–45. Plaintiffs’ counsel
subsequently offered the transcript of the public hearing, see Tr. Ex. 353, into evidence, see R.148 at 45. The
transcript includes testimony by Ottman and Foltz (although, in the transcript, he is identified as Holtz),
as well as the statements and questions of several members of the Wisconsin Assembly and Senate.
Counsel for the defendants made no objection to the admission of Ottman’s testimony from the public
hearing, and we initially admitted that transcript for that limited purpose. Counsel for the plaintiffs,
however, asked that the entire transcript be admitted; counsel for the defendants objected to its
admission on the ground that it contained numerous statements from members of the Wisconsin
70
In sum, from the outset of the redistricting process, the drafters sought to
understand the partisan effects of the maps they were drawing. They designed a
measure of partisanship and confirmed the accuracy of this measure with Professor
Gaddie. They used this measure to evaluate regional and statewide maps that they
drew. They labeled their maps by reference to their partisanship scores, they evaluated
partisan outcomes of the maps, and they compared the partisanship scores and partisan
outcomes of the various maps. When they completed a statewide map, they submitted
it to Professor Gaddie to assess the fortitude of the partisan design in the wake of
various electoral outcomes.
The map that emerged from this process reduced markedly the possibility that
the Democrats could regain control of the Assembly even with a majority of the
statewide vote. The map that would become Act 43 had a pickup of 10 Assembly seats
compared to the Current Map. 225 As well, if their statewide vote fell below 48%, the
design of Act 43 ensured that the Republicans would maintain a comfortable
majority. 226
Finally, it is clear that the drafters were concerned with, and convinced of, the
durability of their plan. Professor Gaddie confirmed the staying power of the Republican
majority under the plan, and Ottman emphasized to the Republican caucus the
long-term consequences of enacting the plan. 227
legislature that were hearsay. See id. at 45. In response, plaintiffs’ counsel asserted that “it’s a public
record. It’s an exception to the hearsay rule. It’s part of the legislative history of Act 43.” Id. at 46.
The transcript does not fall neatly within the public record exception to hearsay set forth in
Federal Rule of Evidence 803(8). Namely, it is not the “record or statement of a public office,” and it does
not set forth “the office’s activities” or “a matter observed while under a legal duty.” Id. (emphasis
added).
The second possible basis for its admission—that the transcript is “part of the legislative history
of Act 43”—is somewhat more persuasive. The transcript provides useful background information on Act
43’s path to enactment and on the types of concerns voiced by the legislators. In this way, it is not being
offered “to prove the truth of the matter[s] asserted in the statement[s]” of the individuals participating in
the hearing. Consequently, it falls outside the definition of hearsay set forth in Federal Rule of Evidence
801(c).
225
See Tr. Ex. 283.
226
See Tr. Ex. 282 (Gaddie “S” curve predicting Republicans would win 54 seats with 48% of the vote).
The plaintiffs argue that the “[s]ecret [d]rafting” of Act 43 and the “[e]xclusion of Democrats” from the
drafting process are further evidence of illicit intent. See R.155 at 4–5. We find this evidence less probative
of whether Act 43 was intended to entrench the Republicans in power. Witnesses for both the plaintiffs
and the defendants testified concerning the strength and operation of the caucus system in Wisconsin,
and there appears to be very little effort to woo colleagues from “across the aisle” either to sponsor or to
227
71
We conclude, therefore, that the evidence establishes that one of the purposes of
Act 43 was to secure Republican control of the Assembly under any likely future
electoral scenario for the remainder of the decade, in other words to entrench the
Republican Party in power.
b.
Alleged shortcomings in the evidence
The defendants point to the miscalculation of the composite measure, to
limitations of the composite measure itself, and to the drafters’ lack of reliance on
Professor Gaddie’s analysis as evidence that they did not have the requisite intent to
subjugate the voting strength of Democrats. The defendants first note that the drafters’
partisan score “was not even correct.” 228 Because of an error in the data for the 2006
Governor’s race—one of the components for their composite measure—the drafters’
numbers were skewed, and the resulting partisan scores were more pro-Republican
than if the scores had been calculated with the correct data. 229 However, as the plaintiffs
note, these errors may diminish the reliability of the composite measure, but they are
irrelevant to the drafters’ intent. 230
support legislation originating with the other party. See, e.g., R.147 at 33 (Whitford explaining that “it’s
extremely difficult to put together a bipartisan coalition to pass something in either … the Assembly or
the Senate”); R.148 at 51 (Ottman describing the process of drafting legislation and noting that “[u]sually
it’s developed among members of your own party”). Although we might find the Wisconsin legislature’s
procedures to be counterproductive, the actions on which the plaintiffs rely appear simply to be par for
the legislative course. We do not discount the possibility, however, that, in some other states, these
actions may suggest a deviation from regular procedures from which an inference of discriminatory
intent may arise.
Finally, the plaintiffs believe that the defendants’ actions in “requiring municipalities to design
wards that followed the new districts’ boundaries” is further evidence of an unconstitutional motive.
R.155 at 5. Although Wisconsin never has passed legislation reversing the order in which wards are
drawn, this idea is not a new one. At trial, the defendants presented undisputed evidence that, following
the 2000 census, Democratic Senate Majority Leader Chvala “drafted a bill that … made changes that
would allow the state to act earlier [to draw wards] or put a deadline for municipalities to act.” R.148 at
94.
228
R.153 at 8; see supra at 9–10.
229
See R.153 at 8.
See R.155 at 10. Professor Mayer also testified that, regardless of the drafters’ calculation errors, the
partisan measure still correlated highly with Professor Gaddie’s regression model. See R.148 at 209.
230
After Professor Mayer had offered this testimony, counsel for the defendants interposed an
objection that Professor Mayer’s testimony was not “based on firsthand knowledge and [it was] not in his
report.” Id. at 210. We reserved ruling and allowed counsel for the plaintiff to continue this line of
72
The defendants also disparage the notion that “the partisan scores were a crystal
ball with predictive powers ensuring that Act 43 would lock Democrats out from seats
that leaned Republican.” 231 They contend that their composite did not have a
“forward-looking component,” but was simply “an average of past elections applied to
the new districts.” 232 We reject as not worthy of belief the assertion that the drafters
would have expended the time to calculate a composite score for each district on the
statewide maps simply to gain an historical understanding of voting behavior. Their
measure was only useful to them—and the exercise of calculating the composite was
only worth the effort—if it helped them assess how Republican representatives in the
newly created districts likely would fare in future elections.
Moreover, each completed map was submitted to Professor Gaddie, who then
generated an “S” curve. The “S” curves were designed to discern “the political potential
of the district.” 233 Professor Gaddie explained that, when he used the term “potential,”
he meant “[i]f you had an election in the future, how might it turn out. So when I say
potential … this is our best estimate of what a non-incumbent election would look like
given a particular set of circumstances, depending on whether one party is stronger or
weaker.” 234
According to the defendants, however, Professor Gaddie’s “S” curves are
irrelevant to the issue of intent because the drafters “didn’t look at them much.” 235 We
cannot accept that estimation of the importance of Professor Gaddie’s work to the
drafters. The record makes clear that the drafters sent Professor Gaddie their completed
maps for which he produced “S” curves. Both Ottman and Foltz testified that, when the
“S” curves were generated, Professor Gaddie provided an explanation of what they
showed. 236 That Ottman may not have used the “S” curves much once they were
questioning. Professor Mayer answered only two additional questions on the subject following the
objection. At the time counsel objected, he admitted that he already had let related questioning “go on for
a while,” id.; indeed, Professor Mayer had given four pages of testimony on the subject prior to counsel’s
objection. Because the bulk of Professor Mayer’s testimony on the calculation errors was offered prior to
counsel’s objection, we now overrule counsel’s objection as untimely.
231
R.153 at 8.
232
R.147 at 47.
233
Tr. Ex. 161 (Gaddie Dep.), at 100.
234
Id. at 100–01.
235
R.147 at 73.
236
See R.148 at 18; R.147 at 73.
73
generated,237 or that Foltz was not able to explain their full significance at trial, five
years later, 238 does not diminish the fact that the drafters sought, and received,
Professor Gaddie’s expert analysis on how each map would behave under the range of
likely electoral scenarios. 239
Finally, the defendants contend that the partisan intent shown by the evidence in
this case cannot be considered invidious because Act 43’s districts are consistent with
traditional districting principles. However, as we have explained earlier, a plan that
adheres to those principles can violate the Equal Protection Clause. Here, the evidence
shows that one purpose of enacting Act 43 was to secure Republican control of the
Wisconsin Assembly. In particular, the history of Act 43 reveals that the drafters created
several alternatives that resulted in a less severe partisan outcome. Of the maps
presented to them, the Republican leadership opted for a map that significantly
increased the number of Republican-leaning districts compared to the Current Map.
Further, the memos prepared for the Assembly members informed them whether the
district number had changed, whether adjustment to the district population was
necessary based on the census numbers, and provided a “[c]omparison of [k]ey [r]aces”
in the new districts compared to the old, but provided little information regarding
traditional districting factors. 240
These facts, in tandem with the overwhelming number of reports and
memoranda addressing the partisan outcomes of the various maps, lead us to conclude
that, although Act 43 complied with traditional redistricting principles, it nevertheless
237
See R.148 at 19.
238
See R.147 at 73.
In their post-trial reply brief, the defendants also attempt to discount the importance of Professor
Gaddie’s “S” curves by referencing his testimony that his “S” curves do not “provide any information on
the durability of the districts over time.” R.156 at 7 (quoting Tr. Ex. 161 (Gaddie Dep.), at 182). Defendants
interpret this answer to mean that the “S” curves do not speak to the likely voting behavior, over time, of
the newly created districts. We do not believe that this interpretation can be reconciled with the other,
detailed testimony that Professor Gaddie provided concerning the purpose of the “S” curves. We believe
a better reading—and one consistent with Professor Gaddie’s other testimony—is that his “S” curves do
not speak to how the districts’ constituencies may change over time.
239
See, e.g., Tr. Ex. 342, at 1. In the memos Foltz provided to members of the Assembly, he attached the
maps of the new districts. See, e.g., id. at 2. In Ottman’s talking-points memos for his meetings with
members of the Senate, he sometimes, but not always, included a brief description of how the district had
changed; for instance, the memo for Senate District 11 states: “Added East Troy and part of the town, as
well as Mukwonago.” Tr. Ex. 242, at 1. The legislators were not given compactness scores, core
population numbers, or the number of municipal and county splits.
240
74
had as one of its objectives entrenching the Republicans’ control of the Assembly. 241
B.
Discriminatory Effect of Act 43
Act 43 also achieved the intended effect: it secured for Republicans a lasting
Assembly majority. It did so by allocating votes among the newly created districts in
such a way that, in any likely electoral scenario, the number of Republican seats would
not drop below 50%. Through the combination of the actual election results for 2012 and
2014, the swing analyses performed by Professors Gaddie and Mayer, as well as the
plaintiffs’ proposed measure of asymmetry, the efficiency gap (or “EG”), the plaintiffs
have “show[n] a burden, as measured by a reliable standard, on [their] representational
rights.” LULAC, 548 U.S. at 418 (opinion of Kennedy, J.).
1.
It is clear that the drafters got what they intended to get. There is no question
that Act 43 was designed to make it more difficult for Democrats, compared to
Republicans, to translate their votes into seats. In the Tale of the Tape, the drafters
compared the partisan performance of the Team Map directly to the Current Map. 242
Where the Current Map had only “49 [Assembly] seats” that were “50% or better” for
Republicans, the Team Map increased that number by ten so that “59 Assembly seats”
were designated as “50% or better” for Republicans. 243 Moreover, under the Team Map
that became Act 43, Republicans expected the following seat distribution: 38 safe
Republican seats, 14 leaning Republican, 10 swing, 4 leaning Democratic, and 33 safe
Democratic seats. 244
Professor Mayer explained the significance of this distribution at trial. 245 Using
the baseline partisan measure that he used to create his Demonstration Plan, 246 Professor
Mayer created a histogram that graphed the predicted percentage of Republican vote of
each district (by 5% increments) on the x axis, and the number of districts that fell into
We also do not believe that the record supports a conclusion that the drafters only wanted to improve
their position incrementally. See Dissent at 125–26. Had this, indeed, been their purpose, they could have
settled on one of the maps that provided a pickup of a smaller number of Republican seats. See supra at
65–66.
241
242
Tr. Ex. 283.
243
Id.
244
Id.
245
See R.148 at 183–85.
246
See supra at 24–25; see also infra at 107–08.
75
each 5% increment on the y axis. 247 The graph reveals that Act 43 includes 42 districts
with predicted Republican vote percentages of between 50 and 60%; only seventeen
districts have predicted Democratic vote percentages of between 50 and 60%. 248 This
demonstrates that, under Act 43, Republican voters are distributed over a larger number
of districts so that they can secure a greater number of seats; in short, “Republicans are
distributed in a much more efficient manner than Democrats.” 249 Professor Mayer’s
graph also reveals that there are only 15 districts with a predicted Republican vote
percentage of 60% or greater; this is compared to 25 districts that have a predicted
Democratic vote percentage of 60% or greater. In other words, Democrats have been
packed into “safe” Democratic districts.
The 2012 and 2014 election results reveal that the drafters’ design in distributing
Republican voters to secure a legislative majority was, in fact, a success. In 2012,
Republicans garnered 48.6% of the vote, but secured 60 seats in the Assembly. 250 In 2014,
Republicans increased their vote percentage to 52 and secured 63 Assembly seats. 251
Moreover, Professors Gaddie and Mayer testified that, consistent with what
actually occurred in 2012 and 2014, under any likely electoral scenario, the Republicans
would maintain a legislative majority. After Professors Gaddie and Mayer developed
their regression models to measure baseline partisanship, 252 each conducted a separate
swing analysis to demonstrate this outcome. “What a swing analysis does,” Professor
Mayer explained, “is ask the question … what might happen” under different electoral
conditions. 253 To determine this, “the statewide vote percentage” is altered by a fixed
amount, typically in one-percentage-point increments, across all districts. 254 “It’s a way
of, generally speaking, estimating what is a plausible outcome given a change in the
statewide vote, which in this case a change in the statewide vote is a proxy for a
different election environment, what might happen if there’s a pro-Democratic swing or
a pro-Republican swing.” 255
247
See Tr. Ex. 15 (attached as Appendix 2 to this opinion); see also Tr. Ex. 107.
248
See Tr. Ex. 15.
249
R.148 at 184.
250
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 70, ¶ 289; id. at 69, ¶ 285.
251
Id. at 70, ¶ 290.
252
See supra at 7–8.
253
R.148 at 222.
254
See id.
255
Id. at 223. There was consensus among the experts—Professors Gaddie, Mayer, Jackman, and
76
Professor Gaddie’s swing analysis is contained in his “S” curves. His “S” curves
include the electoral outcome for each map based on Republican statewide vote
percentage ranging from 40% to 60%. The “S” curve for the Team Map demonstrates
that, to maintain a comfortable majority (54 of 99 seats), Republicans only had to
maintain their statewide vote share at 48%. 256 The Democrats, by contrast, would need
more than 54% of the statewide vote to obtain that many seats. 257
Professor Mayer’s swing analysis did not include the wide-ranging electoral
scenarios set forth in Professor Gaddie’s “S” curves. Instead, Professor Mayer included
only likely electoral scenarios in his analysis. He looked at the electoral outcomes dating
back to 1992 and determined that the maximum statewide vote share the Democrats
had received was 54% in 2006, or roughly 3% more than they had received in 2012. 258
The minimum statewide vote share Democrats had received was 46% in 2010, or
roughly 5% less than they had received in 2012. 259 Professor Mayer’s swing analysis,
therefore, looked at how Act 43 would fare under these two scenarios—the Democrats
receiving 46% of the vote, and the Democrats receiving 54% of the vote. Adjusting the
Goedert—that some type of swing analysis was the accepted method of testing how a particular map
would fare under different electoral conditions. See, e.g., R.149 at 216–17 (Professor Jackman testifying
concerning the application of a uniform swing analysis); R.150 at 181 (Professor Goedert employing a
uniform swing analysis).
256
See Tr. Ex. 282; R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 21, ¶ 70.
See Tr. Ex. 282. The following chart summarizes how the “S” curve for the Team Map predicted how
each party would fare under different electoral outcomes:
257
% vote received (D)
seats won (D)
% vote received (R)
seats won (R)
47
33
47
50
48
35
48
54
49
39
49
56
50
41
50
58
51
43
51
60
52
45
52
64
53
49
53
66
54
53
54
67
Id.; R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 70, ¶ 289; id. at 69, ¶ 285.
258
R.148 at 225.
259
See id.
77
Democratic vote share in each district by these amounts, 260 Professor Mayer predicted
that a 5% decrease in Democratic vote share would have no effect on the allocation of
legislative seats; the Republicans would keep the 60 seats they had, but would not
increase their numbers. 261 When Democratic vote share increased by 3% to 54%,
Professor Mayer predicted that the Democrats would secure only 45 seats. 262
However, both Professor Gaddie and Professor Mayer underestimated the
strength of Act 43 when it came to securing and maintaining Republican control. When
the Republican vote share dropped in 2012 to 48.6%, Republicans still secured 60 seats—
10 more than what Professor Gaddie’s “S” curve predicted. 263 Additionally, when the
Republican vote share increased in 2014 to 52%, the Republicans increased the number
Professor Mayer’s goal was to “make a prospective estimate of what would happen in the subsequent
election” to the 2012 races. R.149 at 77. He therefore based his analysis on the observed results in 2012,
rather than a partisan baseline measure. R.148 at 226. He assumed that all members of the Assembly
would run for re-election, because “we don’t know where incumbents will or will not run and … this is a
uniform way” of accounting for an incumbency effect. R.149 at 91–92. To “calculate the incumbency
advantage,” Professor Mayer “us[ed] the underlying data” in each district, so that the effect was “not …
identical in every district.” Id. at 87.
260
Professor Mayer’s inference from the chart was “that the way in which Act 43 has been drawn has
already secured what in practice amounts to the most you can practically do.” R.148 at 229. Of course, in
2014, the Republican statewide vote percentage increased to 52%, and the number of seats that they
secured was not stagnant.
261
See Tr. Ex. 117. Professor Jackman also presented a swing analysis that was specific to Wisconsin. R.149
at 243–48; Tr. Ex. 495. Professor Jackman did not provide this analysis during discovery, but we admitted
the evidence after the defendants conceded that they had not been prejudiced by the delay. R.149 at 292.
262
Professor Jackman relied on the actual results from 2012 in each district in Wisconsin and then
adjusted the vote in each district based on a 5% swing in each party’s vote share. R.149 at 243–49. He then
calculated the EG for each of these vote-share levels. Professor Jackman observed the same trend as
Professor Mayer: as the Republican vote share went down, the Republicans would not lose many seats; as
the Republican vote share went up, the Republicans did not pick up many more seats (suggesting that the
Republicans discovered a way to maximize the seats they had any potential of winning with the smallest
possible percentage of the vote). Tr. Ex. 495.
After trial, the plaintiffs brought to our attention some discrepancies between our list of trial
exhibits, see R.146, and the rulings that we had made during the course of trial. See R.151 (Motion to
Admit Certain Trial Exhibits). For clarification, the following exhibits were admitted during trial: Tr. Ex.
122, see R. 150 at 291; Tr. Ex. 125, see R. 150 at 291; Tr. Ex. 486, see R.148 at 199; Tr. Ex. 487, see R.149 at 24;
Tr. Ex. 488, see R.159 at 293; Tr. Exs. 492–495, see R. 149 at 293; and Tr. Ex. 581, see R.150 at 255.
With respect to the Democratic wave election, therefore, it would seem that Professor Mayer’s swing
analysis correctly predicted that, even with 54% of the statewide vote share, the Democrats would not
secure a majority in the Assembly.
263
78
of seats they held by 3, as opposed to their seat share being stagnant, as predicted by
Professor Mayer. 264 In other words, the actual election results suggest that Act 43 is
more resilient in the face of an increase in the statewide Democratic vote share, and is
more responsive to an increase in the statewide Republican vote share, than either
Professor Gaddie or Professor Mayer anticipated.
The fact that Democrats and Republicans were treated differently under Act 43
becomes even more stark when we examine the number of seats secured when the
parties obtain roughly equivalent statewide vote shares. In 2012, the Democrats
received 51.4% of the statewide vote, but that percentage translated into only 39
Assembly seats. A roughly equivalent vote share for Republicans (52% in 2014),
however, translated into 63 seats—a 24 seat disparity. Moreover, when Democrats’ vote
share fell to 48% in 2014, that percentage translated into 36 Assembly seats. Again, a
roughly equivalent vote share for Republicans (48.6% in 2012) translated into 60 seats—
again a 24 seat disparity. 265 The evidence establishes, therefore, that, even when
Republicans are an electoral minority, their legislative power remains secure. 266
2.
The record here is not plagued by the infirmities that have precluded the Court,
in previous cases, from concluding that a discriminatory effect has been established. In
Bandemer, the Court made clear that plaintiffs could not establish a constitutional
At this end of the spectrum, Professor Gaddie’s prediction was more accurate; his “S” curve predicted
that a 52% vote share would translate into 64 Republican seats.
264
At trial, Foltz testified that the drafters’ calculation of the composite partisanship measure, at least for
some districts, was flawed because of data errors related to the 2006 Governor’s race. We agree that these
errors reduce the composite’s reliability as a measure. However, in reaching our conclusion that the
plaintiffs have met their evidentiary burden, we have not relied on the drafters’ composite measure of
partisanship, but on actual election results and analyses performed by Professors Gaddie and Mayer,
which were not infected by the faulty data. Moreover, as explained in supra note 230, Professor Mayer
testified that, regardless of the drafters’ calculation errors, the partisan measure still correlated highly
with Professor Gaddie’s regression model. See R.148 at 209.
265
The Dissent questions whether the Republicans actually can entrench themselves in power given that a
popularly elected Democratic governor could prevent the Republicans from enacting their agenda. See
Dissent at 123. Although the governorship may be a check on Republican legislative efforts, it also cannot
secure for Democrats the opportunity to pass an agenda consistent with their policy objectives.
266
The Dissent also doubts whether the plaintiffs have been damaged by their inability to secure a
political majority. See Dissent at 145–46. According to the Dissent, Republican legislators who win by
slimmer margins will be more receptive to the needs of their Democratic constituents. Although this
argument might have some intuitive appeal in other political contexts, it is not supported by the record
here, where there is evidence of a strong caucus system. See supra at 54; infra at 115.
79
violation based “on a single election.” 478 U.S. at 135 (plurality opinion). This was
because
Indiana is a swing State. Voters sometimes prefer Democratic candidates,
and sometimes Republican. The District Court did not find that because of
the 1981 Act the Democrats could not in one of the next few elections
secure a sufficient vote to take control of the assembly. … The District
Court did not ask by what percentage the statewide Democratic vote
would have had to increase to control either the House or the Senate. The
appellants argue here, without a persuasive response from the appellees,
that had the Democratic candidates received an additional few percentage
points of the votes cast statewide, they would have obtained a majority of
the seats in both houses. Nor was there any finding that the 1981
reapportionment would consign the Democrats to a minority status in the
Assembly throughout the 1980’s or that the Democrats would have no
hope of doing any better in the reapportionment that would occur after
the 1990 census. Without findings of this nature, the District Court erred
in concluding that the 1981 Act violated the Equal Protection Clause.
Id. at 135–36.
The record here answers the shortcomings that the Bandemer plurality identified.
First, we now have two elections under Act 43. In 2012, the Democrats garnered 51.4%
of the vote, but secured only 39 seats in the Assembly—or 39.3% of the seats. 267 In 2014,
the Democrats garnered 48% of the vote and won only 36 seats—or 36.4% of the seats. 268
If it is true that a redistricting “plan that more closely reflects the distribution of state
party power seems a less likely vehicle for partisan discrimination,” LULAC, 548 U.S. at
419 (opinion of Kennedy, J.), then a plan that deviates this strongly from the
distribution of statewide power suggests the opposite.
Moreover, as described in some detail above, Professor Gaddie’s “S” curve and
Professor Mayer’s swing analysis reveal that the Democrats are unlikely to regain
control of the Assembly. And Act 43 has proven even more resistant to increases in
Democratic vote share, and more responsive to increases in Republican vote share, than
was predicted. Consequently, it is not the case that “an additional few percentage
points of the votes cast statewide” for the Democrats will yield an Assembly majority.
Bandemer, 478 U.S. at 135 (plurality opinion). 269
267
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 52, ¶ 257.
268
Id., ¶ 258.
269
The Dissent notes that, in 2012, the seats-to-vote ratio under Act 43 was similar to that under the
80
Furthermore, because we have the actual election results to confirm the reliability
of Professor Gaddie’s model and “S”-curve analysis, we are not operating only in the
realm of hypotheticals—a prospect that at least one member of the Court in LULAC
found troubling. In LULAC, Justice Kennedy commented on a proposal by one of the
amici to adopt a partisan-bias standard, which would compare how the two major
parties “would fare hypothetically if they each (in turn) had received a given percentage
of the vote.” 548 U.S. at 419 (opinion of Kennedy, J.) (internal quotation marks omitted).
Justice Kennedy explained that,
[e]ven assuming a court could choose reliably among different models of
shifting voter preferences, we are wary of adopting a constitutional
standard that invalidates a map based on unfair results that would occur
in a hypothetical state of affairs. Presumably such a challenge could be
litigated if and when the feared inequity arose.
Id. at 420. Professor Gaddie’s “S” curves and Professor Mayer’s swing analysis, like a
partisan-bias analysis, depend upon a hypothetical state of affairs: they assume a
uniform increase or decrease in vote share across all districts—something that does not
occur in actual elections. Here, however, the predictive work of the professors is
combined with the results of two actual elections in which the feared inequity did arise.
3.
While the evidence we have just described certainly makes a firm case on the
question of discriminatory effect, that evidence is further bolstered by the plaintiffs’ use
of the “efficiency gap,” or EG for short, to demonstrate that, under the circumstances
presented here, their representational rights have been burdened. We begin with an
explanation of the EG. Because the EG is a new measure and was the focus of extensive
testimony at trial, we believe it appropriate to examine its value and shortcomings in
detail.
a.
apportionment scheme in Bandemer and concludes, therefore, that our case cannot be distinguished from
Bandemer. See Dissent at 119. As we have demonstrated in the above discussion, however, the Court’s
primary concern in Bandemer was not that the numbers were not sufficiently egregious, but that there was
no evidence that the gerrymander was durable. Here we have two elections under Act 43, as well as
swing analyses conducted by three experts, all of which support the conclusion that Act 43’s partisan
effects will survive all likely electoral scenarios, throughout the decennial period.
81
The allegations in this case are that Act 43’s drafters employed two of the
traditional methods of gerrymandering in order to diminish the electoral power of
Democratic voters in Wisconsin: “packing” and “cracking.” Packing refers to the
concentration of a party’s voters in a limited number of districts; as a result, the party
wins these packed districts by large margins. 270 Cracking, on the other hand, is the
division of a party’s voters across a number of districts such that the party is unable to
achieve a majority in any. 271 The EG is a measure of the degree of both cracking and
packing of a particular party’s voters that exists in a given district plan, based on an
observed electoral result. 272
The EG calculation is relatively simple. First, it requires totaling, for each party,
statewide, (1) the number of votes cast for the losing candidates in district races (as a
measure of cracked voters), along with (2) the number of votes cast for the winning
candidates in excess of the 50% plus one votes necessary to secure the candidate’s
victory (as a measure of packed voters). 273 The resulting figure is the total number of
“wasted” votes for each party. 274 These wasted vote totals are not, of themselves,
independently significant for EG purposes; rather, it is the comparative relationship of
one party’s wasted votes to another’s that yields the EG measure. 275 The EG is the
difference between the wasted votes cast for each party, divided by the overall number
of votes cast in the election. 276 When the two parties waste votes at an identical rate, the
plan’s EG is equal to zero.277 An EG in favor of one party (Party A), however, means
that Party A wasted votes at a lower rate than the opposing party (Party B).278 It is in
270
R.1 at 3, ¶ 5; see supra at 17.
271
R.1 at 3, ¶ 5.
272
Id. at 15, ¶¶ 49–50; R.149 at 170–71.
273
R.149 at 181–82.
The votes are “wasted” in the sense that votes cast for losing candidates do not help to generate seats
for the party and that votes cast for winning candidates in excess of 50% plus one could have been
deployed elsewhere to greater effect. Id. at 182; see also supra note 79.
274
275
R.149 at 171–72.
276
This can be expressed mathematically in the following formula:
EG =
WB – WA
n
n
R.34 at 16; R.149 at 181–82.
277
R.149 at 181–82.
In some documents in the record, negative EG values indicate higher wasted vote rates for Democrats
compared to Republicans (i.e., a pro-Republican EG), and positive EG values indicate the opposite.
278
82
this sense that the EG is a measure of efficiency: because Party A wasted fewer votes
than Party B, Party A was able to translate, with greater ease, its share of the total votes
cast in the election into legislative seats. Put simply, an EG in Party A’s favor means it
carried less electoral dead weight; its votes were, statistically, more necessary to the
victories of its candidates, and, consequently, it secured a greater proportion of the
legislative seats than it would have secured had Party A and Party B wasted votes at the
same rate.
In a related sense, the EG can be viewed as a measure of the proportion of
“excess” seats that a party secured in an election beyond what the party would be
expected to obtain with a given share of the vote. 279 In a purely proportional
representation system, a party would be expected to pick up votes and seats at a
one-to-one ratio, i.e., for every additional percentage of the statewide vote the party
gains, it should also gain a percentage in the share of the seats. 280 Based on decades of
observed historical data, however, the parties’ experts agreed that with single-member,
simple-plurality systems like Wisconsin’s, we can expect that for every 1% increase in a
party’s vote share, its seat share will increase by roughly 2%.281 Thus, a party that gets
52% of the statewide vote should be expected to secure 54% of the legislative seats. If
the party instead translates its 52% of the vote into 58% of the seats, the district plan has
demonstrated an EG of 4% in favor of that party (the difference between the expected
seat share and the actual seat share).
Both Professors Mayer and Jackman calculated the EG for the 2012 Assembly
elections in Wisconsin. In his analysis, Professor Mayer employed the “full method,”
which requires aggregating, district-by-district, the wasted votes cast for each party.
Applying this methodology, he determined that Act 43 yielded a pro-Republican EG of
11.69%. 282 Professor Jackman, however, used the “simplified method,” 283 that assumes
equal voter turnout at the district level. His calculations estimated a pro-Republican EG
of 13% for the 2012 election. Professor Jackman also calculated an EG for the 2014
election; that calculation resulted in a pro-Republican EG of 10%. 284
279
Tr. Ex. 34, at 5.
280
Id. at 11.
281
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 28, ¶¶ 105–06.
282
See Tr. Ex. 2, at 46.
283
See supra note 88.
284
Tr. Ex. 34, at 5–6.
83
Professor Jackman also conducted an historical analysis of redistricting plans
which compared the trends in efficiency gaps across a wide variety of states over the
last forty years (a total of 786 state legislative elections). 285 He observed that an EG in the
first year after a districting plan is enacted bears a relatively strong relationship to the
efficiency gap over the life of a plan. 286 The party that “wastes” more votes in the first
election year is likely to continue “wasting” more votes in future elections.
Relatedly, Professor Jackman conducted two additional analyses which suggest
that an efficiency gap above 7% in any districting plan’s first election year will continue
to favor that party for the life of the plan. First, Professor Jackman compared districting
plans across a wide variety of states, and determined that over 95% of plans with an EG
of at least 7% will never have an EG that favors the opposite party.287 Second, Professor
Jackman conducted a “swing analysis” of all redistricting plans since 2010 and
determined that nearly all plans that resulted in a 7% efficiency gap favoring one party
in the first election year will retain an efficiency gap that favors that same party, even
when one adjusts a party’s statewide vote share by five points. 288
Professor Jackman then compared his EG estimates for Act 43 with the historical
EG estimates from other states. Given historical trends and averages, he opined that
Wisconsin’s plan would have an average pro-Republican efficiency gap of 9.5% for the
entire decennial period. 289 Therefore, in his expert opinion, Wisconsin Democrats would
continue to have a less effective vote for the life of the plan. 290 Barring an
“unprecedented political earthquake,” Democrats would be at an electoral
disadvantage for the duration of Act 43. 291
Professor Jackman also presented a swing analysis that was specific to
Wisconsin. 292 He relied on the actual results from 2012 in each district in Wisconsin and
285
R.149 at 229–33.
Id. at 233. In his rebuttal report, Dr. Jackman notes that a plan with an initial pro-Republican efficiency
gap of 7% will have a plan-average efficiency gap of approximately 5.3%. See Tr. Ex. 83 at 16.
286
287
R.149 at 224–40.
Tr. Ex. 93; R.149 at 215–24. This was the basis for the plaintiffs’ proposed threshold for liability. See
supra notes 81 and 132, and accompanying text.
288
Tr. Ex. 83, at 15–17; R.149 at 232 (describing Act 43’s lifetime average efficiency gap as “in the
neighborhood of negative ten percent”).
289
290
R.149 at 233 (opining that he was “[v]irtually certain” of this outcome, “[v]irtually 100 percent”).
291
Id. at 232.
292
R.149 at 243–48; Tr. Ex. 495.
84
then adjusted the vote in each district based on a 5% swing in each party’s vote share.293
He then calculated the EG for each of these vote-share levels. Professor Jackman
observed that, even with a 5% swing in the Democrats’ favor, the EG would not drop
below 7%. 294
As we already have seen, this more efficient distribution of Republican voters
has allowed the Republican Party to translate its votes into seats with significantly
greater ease and to achieve—and preserve—control of the Wisconsin legislature. In both
elections held under Act 43, the Republicans obtained a far greater proportion of the
Assembly’s 99 seats than they would have without the leverage of a considerable and
favorable EG. In 2012, the Republicans won 61% of Assembly seats with only 48.6% of
the statewide vote, resulting in a 13% EG in their favor. In 2014, the Republicans
garnered 52% of the statewide vote but secured 64% of Assembly seats, resulting in a
pro-Republican EG of 10%. 295 Thus, the Republican Party in 2012 won about 13
Assembly seats in excess of what a party would be expected to win with 49% of the
statewide vote, and in 2014 it won about 10 more Assembly seats than would be
expected with 52% of the vote.
Moreover, the expert testimony before us indicates that the Republican Party’s
comparative electoral advantage under Act 43 will persist throughout the decennial
period; Democratic voters will continue to find it more difficult to affect district-level
outcomes, and, as a result, Republicans will continue to enjoy a substantial advantage in
converting their votes into seats and in securing and maintaining control of the
Assembly.
b.
The defendants have made a number of legal, methodological, and policy-based
attacks against judicial use of the EG as a measure of a district plan’s partisan effect. We
begin with their claim that use of the EG is foreclosed by Supreme Court precedent. The
Supreme Court has made clear that the Constitution does not require that a map result
in each party gaining a share of the legislative seats in proportion to their share of the
statewide vote. LULAC, 548 U.S. at 419 (opinion of Kennedy, J.) (“To be sure, there is no
293
R.149 at 243–49.
294
Tr. Ex. 495.
By way of comparison, if the EG had been 0, that is, the Republicans and Democrats had been wasting
votes at the same rate, the Republicans would have secured approximately 47 seats with 48.6% of the vote
and would have secured 53 seats with 52% of the vote.
295
85
constitutional requirement of proportional representation ….”). 296 The defendants have
argued throughout this action that this precept forecloses the use of any metric that
employs a votes-to-seats relationship as its starting point to measure a plan’s partisan
effect. The EG, they say, is rooted in a baseline requirement that a district plan deliver
hyper-proportional representation in the form of the 2-to-1 seats-to-votes ratio
described above and is therefore unavailable for use as a measure of discriminatory
effect. 297
We cannot accept this argument. To say that the Constitution does not require
proportional representation is not to say that highly disproportional representation may
not be evidence of a discriminatory effect. Indeed, acknowledging that the Constitution
does not require proportionality, Justice Kennedy observed in LULAC that “a
congressional plan that more closely reflects the distribution of state party power seems
a less likely vehicle for partisan discrimination than one that entrenches an electoral
minority.” 548 U.S. at 419 (opinion of Kennedy, J.). 298 We do not believe, therefore, that
the Constitution precludes us from looking at the ratio of votes to seats in assessing a
plan’s partisan effect. 299
See also Vieth, 541 U.S. at 286–88 (plurality opinion) (stating that the proposed effects standard that a
map not be able to “thwart [a party’s] ability to translate a majority of votes into a majority of seats”
“rests upon the principle that groups … have a right to proportional representation” and observing that
“the Constitution contains no such principle”); id. at 308 (Kennedy, J., concurring in the judgment) (“The
fairness principle appellants propose is that a majority of voters in the Commonwealth should be able to
elect a majority of the Commonwealth’s congressional delegation. There is no authority for this
precept.”).
296
297
See supra note 88 and at 82.
According to the Dissent, saying that highly disproportional representation may be evidence of
discriminatory intent is tantamount to making proportional representation a constitutional requirement.
See Dissent at 138. The Dissent, however, is conflating the evidence of a constitutional violation with a
violation itself. As we already have explained, in order to establish a constitutional violation a plaintiff
must show that the drafters “place[d] a severe impediment on the effectiveness of the votes of individual
citizens on the basis of their political affiliation.” See supra at 56. Thus, a districting map that produces
results significantly out of proportion with a party’s voting strength is evidence that the drafters of the
map have erected such an impediment.
298
This reading finds a constitutional analogue in the malapportionment context. See Brown v. Thomson,
462 U.S. 835, 842–43 (1983) (explaining that “minor deviations from mathematical equality among state
legislative districts are insufficient to make out a prima facie case of invidious discrimination,” but
finding discriminatory effects in “plan[s] with larger disparities in population” (internal quotation marks
omitted)). That population deviation is measured in relation to equal apportionment as a baseline—an
outcome that the Constitution does not require in the state legislative context—does not make it any less a
measure of discriminatory effect.
299
86
As it has been presented here, the EG does not impermissibly require that each
party receive a share of the seats in proportion to its vote share. Rather, the EG
measures the magnitude of a plan’s deviation from the relationship we would expect to
observe between votes and seats. We do not believe Vieth or LULAC preclude our
consideration of the EG measure. 300
We turn next to what are best described as methodological and operational
critiques of the EG measure. First, the defendants point out that the plaintiffs have
proposed two distinct methods for calculating the EG. The differing approaches can
yield materially different EG values, which, in turn, will produce uncertainty in the
maps that should be subject to judicial scrutiny. As explained previously, Professor
Mayer employed the “full method,” which included aggregating every district’s wasted
votes for each party. Professor Jackman used the “simplified method” that assumes
equal voter turnout at the district level. 301 These two methods produce identical results
when voter turnout is equal across districts; however, where voter turnout varies, as it
does in Wisconsin, the EG measure will differ depending on the method used.
Although we view the full method as preferable because it accounts for the
reality that voters do not go to the polls at equal rates across districts, we do not believe
that this calls into question Professor Jackman’s use of the simplified method in his
analysis. Professor Goedert in his expert report described the simplified method as “an
appropriate and useful summary measure” for calculating the EG, 302 and the parties
have stipulated that the shortcut’s implied 2-to-1 votes-to-seats relationship reflects the
“observed average seat/votes curve in historical U.S. congressional and legislative
elections.” 303 Were there record evidence indicating that Professor Jackman’s shortcut
did not correlate highly with both the full method and electoral reality, we would have
reason to doubt its validity. Because this is not the case here, we are not troubled by the
We note that in LULAC a majority of the Justices discussed, with varying degrees of skepticism,
another measure of asymmetry called “partisan bias.” Justice Kennedy was particularly reluctant to
endorse this measure because it might be used to invalidate a plan “based on unfair results … in a
hypothetical state of affairs” rather than in an observed electoral result. LULAC, 548 U.S. at 420 (opinion
of Kennedy, J.). The EG, which is calculated using the results of actual elections, does not suffer from this
drawback and, we conclude, does not raise the same concern articulated by Justice Kennedy.
300
301
See supra at 82–83.
302
Tr. Ex. 546, at 5.
See R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 28, ¶ 105; Tr. Ex. 546, at
6. The plaintiffs attribute the high correlation between the full and simplified methods to the fact that
districts must be equal or nearly equal in population, which they say results in generally small and
nonpartisan turnout deviations across districts. R.134 at 75.
303
87
existence of distinct methods of calculating the EG. Moreover, we are not addressing a
legislative plan that is at the statistical margins. In this case, both methods yield an
historically large, pro-Republican EG.
The defendants also contend that the EG, as an indicator of partisan
gerrymandering, is both overinclusive and underinclusive. They presented evidence
that districting plans, which had been put in place by courts, commissions, or divided
governments, sometimes register high EG values. 304 Conversely, the defendants pointed
to several congressional districting plans that are commonly understood as partisan
gerrymanders but registered low EG values or even EG values favoring the party that
did not create the districting map. We do not share this particular concern. If a
nonpartisan or bipartisan plan displays a high EG, the remaining components of the
analysis will prevent a finding of a constitutional violation. See, e.g., Arlington Heights,
429 U.S. at 264–65 (“[O]fficial action will not be held unconstitutional solely because it
results in a racially disproportionate impact.”); Washington, 426 U.S. at 240 (stating the
“basic equal protection principle that the invidious quality of a law … must ultimately
be traced to a racially discriminatory purpose”). For example, if a claim of partisan
gerrymandering is brought against a court- or commission-drawn district plan with a
high EG, it will stall when the plaintiffs attempt to make the necessary showing of
discriminatory intent. In the same way, a challenge to a map enacted with egregious
partisan intent but demonstrating a low EG also will fail because the plaintiffs cannot
demonstrate the required discriminatory effect. The present case, of course, does not
present either of these situations. Here, the plaintiffs have put forward sufficient
evidence showing both that Act 43 was enacted with impermissible intent and that it
demonstrates a large and durable EG value.
Lastly, the defendants argue that the EG measure is overly sensitive to small
changes in voter preferences. At trial, Mr. Trende testified that the EG will vary
depending on whether there is a national wave in the electorate favoring one party or
the other. He described a hypothetical scenario in which a national pro-Republican
wave resulted in an increase in Republican vote share in every district of two points
above the otherwise expected Republican vote share. This slight change, Mr. Trende
explained, could alter the outcomes in particularly close races and thus produce a
significantly different EG value than if the national wave had not occurred. Professor
Goedert raised a related point. He suggested that assessing a given plan based on the
results of the first observed election under the plan is arbitrary and may yield
problematic results if that first election happens to be a national wave election.
304
Tr. Ex. 34, at 55.
88
We acknowledge these as legitimate criticisms of the EG measure generally;
however, they are less compelling in the context of this case. Both concerns are rooted in
an EG being drawn from only a single election, which, for any number of reasons, may
represent an electoral aberration. 305 Here we have the results of two elections under Act
43, one in which the Republicans failed to garner a majority of the statewide vote (2012),
and one in which they exceeded it by two percentage points. Under both electoral
scenarios, there was a sizeable pro-Republican efficiency gap: 13% in 2012 and 10% in
2014. 306
Even in the absence of these results, however, there is evidence in the record that
establishes the durability of Act 43’s pro-Republican efficiency gap. Professor Jackman
conducted an historical analysis of redistricting plans which compared the trends in
efficiency gaps across a wide variety of states over the last forty years (totaling 786 state
legislative elections). 307 Based on this analysis, Professor Jackman estimated that
Wisconsin’s plan, with an initial pro-Republican efficiency gap of 13.3%, would have a
plan average pro-Republican efficiency gap of 9.5%. 308 In other words, the Republicans’
ability to translate their votes into seats will continue at a significantly advantageous
rate through the decennial period. 309
Moreover, Mr. Trende himself attested to the durability of Act 43’s EG in the face
of a wave election. In his expert report, Mr. Trende observed that if the Democrats
engaged in a “modestly better effort” to get out the vote, and secured just 600 more
votes in Districts 1 and 94, the “EG falls by more than two points off these modest shifts,
to 9.466.”310 Nevertheless, Mr. Trende conceded that, although such a shift might affect
the EG’s applications in other contexts, it “would not make a difference in terms of
whether the Wisconsin map invited Court scrutiny” because the EG still was above the
plaintiffs’ proposed threshold of 7%. 311
For instance, the Dissent explicitly makes the point that 2012 was an electoral anomaly. See Dissent at
153–55.
305
By way of comparison, the average pro-Republican efficiency gap over the prior decade was 7.6%,
R.125 (J. Final Pretrial Report containing J. Statement of Stipulated Facts) at 42, ¶ 194; in the two elections
immediately prior to adoption of Act 43, however, the efficiency gaps were 5% (2008), and 4% (2010), id.
at 51, ¶¶ 255–56.
306
307
R.149 at 229–33.
308
Tr. Ex. 83, at 16; R.149 at 232.
309
R.149 at 233 (opining that he was “[v]irtually certain” of this outcome, “[v]irtually 100 percent”).
310
Tr. Ex. 547, at ¶ 147.
311
Id., ¶ 148; see supra note 84 and at 26. The Dissent takes issue with the plaintiffs’ proposed 7% threshold
89
The defendants also raise policy-based objections to the EG as a measure of
discriminatory effect. First, they claim that the creation of many competitive districts,
which may be a desirable and non-partisan policy choice, will result in a highly
sensitive map in which the EG could swing rather wildly with even mild electoral
shifts. We do not doubt this is the case. 312 However, as with some of the criticisms that
we already have discussed, this concern is ameliorated by other aspects of the equal
protection analysis. It would be difficult to establish that drafters who designed a map
with many competitive districts had the requisite partisan intent to show a
constitutional violation.
The defendants similarly claim that identifying an EG of zero as the baseline or
ideal would discourage states from enacting systems of proportional representation.
See Gaffney, 412 U.S. at 752 (upholding plan that sought to “achieve a rough
approximation of the statewide political strengths of the Democratic and Republican
parties”). Professor Goedert in particular noted that if a state successfully achieved
proportional representation, the plan might fail an EG analysis because it fails to give a
hyperproportional share to the party winning the majority of the statewide vote. Again,
however, drafters who had the intent to create a proportional system hardly could be
accused of harboring a discriminatory intent. Moreover, the defendants have offered no
evidence that Act 43’s drafters had any interest in hewing closely to proportional
representation; indeed, the evidence is directly to the contrary. For these reasons, we are
not persuaded that the policy objections to the EG bear any relationship to this case. 313
We further emphasize, in any event, that we have not determined that a particular
measure of EG establishes presumptive unconstitutionality, which itself diminishes all
of the defendants’ policy-based arguments. Instead, we acknowledge that the expert
opinions in this case have persuaded us that, on the facts before us, the EG is
corroborative evidence of an aggressive partisan gerrymander that was both intended
and likely to persist for the life of the plan. 314
for liability as being too easy to meet. See Dissent at 157–59. Here, the efficiency gap for Act 43 in 2012
exceeds this baseline by 6%. Therefore, we need not reach the propriety of the 7% number.
The Dissent makes a related point that, the more close races the Republicans win, the more Democratic
votes are wasted, resulting in a large efficiency gap. See Dissent at 149.
312
The Dissent also notes limitations in the EG’s usefulness in evaluating a gerrymander where there is a
75–25 vote split between the candidates, see Dissent at 151; this also is not the case before us.
313
The Dissent also takes issue with the EG measure because it will not be known until after the first
election, and, therefore, it is impossible for mapmakers to know in advance whether their plan will pass
muster. See Dissent at 133–34. This is somewhat of a red herring. The level of the EG only will become an
issue if the drafters have evinced an intent to entrench their party in power. Moreover, the drafters can
assess the durability of their partisan maps, even absent an actual electoral outcome, by employing a
314
90
In sum, we conclude that the plaintiffs have established, by a preponderance of
the evidence, that Act 43 burdens the representational rights of Democratic voters in
Wisconsin by impeding their ability to translate their votes into legislative seats, not
simply for one election but throughout the life of Act 43. We therefore turn our
attention to whether the burden is justified by some legitimate state interest.
V
JUSTIFICATION
In the initial stages of this litigation, the plaintiffs took the view that, should they
successfully establish the intent and effects elements of their constitutional claim, the
burden should then shift to the defendants to show that Act 43’s unlawful effects were
“‘unavoidable’ in light of the state’s political geography and legitimate districting
objectives.” 315 In our summary judgment order, we noted that “some type of burdenshifting is appropriate,” adding that “to the extent that plaintiffs have an initial burden
to show that [Act 43] cannot be justified using neutral criteria,” it was met at the
summary judgment stage by their presentation of the Demonstration Plan. 316 We left
open the question of which party ultimately should bear the burden of proving Act 43’s
legitimacy. However, we rejected definitively the plaintiffs’ “unavoidable” standard as
an “overstate[ment]” of the degree of the burden. 317
In response, the plaintiffs reformulated the third step of their test to allow the
defendants to avoid liability if they can justify Act 43’s effects on the basis of legitimate
districting goals or Wisconsin’s natural political geography. They maintain, however,
that it is the State’s burden ultimately to prove that Act 43’s effect is justified and not
their burden to prove that it is not.
The defendants maintain that even this lesser showing is too demanding. They
argue that because Act 43 complies with traditional districting objectives, its partisan
effect is necessarily excusable as a matter of law and need not be explained by neutral
considerations. We already have considered this argument in detail in our evaluation of
the intent element of the plaintiffs’ claim, and so we do not repeat that discussion
here. 318
swing analysis. See supra note 255.
315
R.94 at 30.
316
Id. at 35.
317
Id. at 32, 35.
318
See supra at 61–62.
91
In the absence of explicit guidance from the Supreme Court, we think that the
most appropriate course in this context is to evaluate whether a plan’s partisan effect is
justifiable, i.e., whether it can be explained by the legitimate state prerogatives and
neutral factors that are implicated in the districting process. This approach allows us to
hew as closely as possible to the Supreme Court’s approach in analogous areas. As we
observed in our summary judgment order, members of the Court have applied this
formulation at several points throughout its political gerrymandering case law. See
Vieth, 541 U.S. at 307 (Kennedy, J., concurring in the judgment) (“A determination that a
gerrymander violates the law must rest on something more than the conclusion that
political classifications were applied. It must rest instead on a conclusion that [political]
classifications … were applied in … a way unrelated to any legitimate legislative
objective.”); id. at 351 (Souter, J., concurring in part and dissenting in part) (stating that,
after the plaintiff has made a prima facie case, “I would then shift the burden to the
defendants to justify their decision by reference to objectives other than naked partisan
advantage”); Bandemer, 478 U.S. at 141 (plurality opinion) (“The equal protection
argument would proceed along the following lines: If there were a discriminatory effect
and a discriminatory intent, then the legislation would be examined for valid
underpinnings.”). It is also consistent with the Court’s approach in the state legislative
malapportionment context. See Voinovich v. Quilter, 507 U.S. 146, 161 (1993)
(“[A]ppellees established a prima facie case of discrimination, and appellants were
required to justify the deviation.”); Brown v. Thomsen, 462 U.S. 835, 842–43 (1983)
(explaining that a plan with “large[] disparities in population … creates a prima facie
case of discrimination and therefore must be justified by the State”).
The record before us does not require us to anticipate how the Supreme Court
will resolve the allocations of proof on this issue. It is clear that the parties, recognizing
the present ambiguity on this point, placed before us all the evidence they could in
support of their respective positions. Assuming the plaintiffs have the ultimate burden
of proof on the issue, they have carried that burden.
The evidence further makes clear that, although Wisconsin’s natural political
geography plays some role in the apportionment process, it simply does not explain
adequately the sizeable disparate effect seen in 2012 and 2014 under Act 43. Indeed, as
we already noted and will discuss again, the defendants’ own witnesses produced the
most crucial evidence against justifying the plan on the basis of political geography.
Their testimony credibly established that Act 43’s drafters produced multiple
alternative plans that would have achieved the legislature’s valid districting goals while
generating a substantially smaller partisan advantage. We therefore must conclude that,
regardless where the burden lies, Act 43’s partisan effect cannot be justified by the
legitimate state concerns and neutral factors that traditionally bear on the
reapportionment process.
92
A.
The defendants’ primary argument is that Wisconsin’s political geography
naturally favors Republicans because Democratic voters reside in more geographically
concentrated areas, particularly in urban centers like Milwaukee and Madison. For this
reason, they submit, any districting plan in Wisconsin necessarily will result in an
advantageous distribution of Republican voters statewide just as Act 43 does.
The plaintiffs have stressed, as a general matter throughout this litigation, that
even if there were some inherent pro-Republican bias in Wisconsin, there is no evidence
that such a bias could explain Act 43’s large EG measures. They maintain that without
such evidence, political geography cannot justify the burden that Act 43 places on
Democratic voters in Wisconsin.
The bulk of evidentiary support for the defendants’ political geography
argument was presented through the testimony of Mr. Trende. 319 His overarching
Prior to trial, the plaintiffs moved in limine to exclude Mr. Trende’s report and testimony pursuant to
Federal Rule of Evidence 702 and Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). See R.71.
The defendants filed a response, see R.78, and we heard argument on the motion at the summary
judgment hearing, see R.89 at 35–42, 60–61. We reserved our ruling and permitted Mr. Trende to testify at
trial. After carefully reviewing Mr. Trende’s qualifications and submissions, and with the benefit of his
trial testimony, we deny the plaintiffs’ request to exclude Mr. Trende as an expert.
319
To begin, Mr. Trende is qualified to give expert testimony in this case. The plaintiffs maintain that
Mr. Trende is not an expert because “he is neither a Ph.D. nor a political scientist, has no particular
training in the kinds of issues involved in this case, and has never written a peer-reviewed article in
political science or any other field ….” R.71 at 9–10. The plaintiffs further attack Mr. Trende’s experience,
skills, and knowledge. They emphasize his unfamiliarity with “the relevant literature regarding partisan
gerrymandering and geographic clustering,” as well as his lack of Wisconsin-specific experience. Id. at
11–12. We have explained, however, that neither Daubert nor Rule 702 “require[] particular credentials”
or “require that expert witnesses be academics or PhDs.” Tuf Racing Prods., Inc. v. Am. Suzuki Motor Corp.,
223 F.3d 585, 591 (7th Cir. 2000). Indeed, “[a]nyone with relevant expertise enabling him to offer
responsible opinion testimony helpful to judge or jury may qualify as an expert witness.” Id. Although
not a social scientist, Mr. Trende has studied, written on, and analyzed voting trends and political
geography throughout the United States. He has developed an expertise in this area, and his opinions are
informative to the issues before us and are helpful in conducting our analysis.
We further conclude that the principles and methodologies employed by Mr. Trende are
sufficiently reliable. See Fuesting v. Zimmer, Inc., 421 F.3d 528, 535–36 (7th Cir. 2005), vacated in part on other
grounds, 448 F.3d 936 (7th Cir. 2006). In their motion in limine, the plaintiffs identified several purported
flaws in the methodologies that Mr. Trende used to form his opinions, particularly those pertaining to his
“partisan index” (“PI”) and “nearest neighbor” analyses. R.71 at 13–27. As is evident throughout our
discussion of Mr. Trende’s testimony, see infra at 93–96, we believe that these criticisms, although valid,
go to the weight of his opinions rather than to their admissibility. Moreover, having allowed Mr. Trende
to testify at trial, we are able to consider his opinions with the benefits of “[v]igorous cross-examination”
and the “presentation of contrary evidence,” both of which “are … traditional and appropriate means of
93
theory is that the Democratic coalition nationwide has become more liberal over the last
several decades; as a result, it has contracted geographically and is now concentrated
heavily in urban areas. 320 This concentration, in turn, has hurt the Democratic Party in
congressional elections, which tend to favor parties with wider geographic reach. 321
Mr. Trende first demonstrated this theory using color-coded maps illustrating the 1996,
2004, and 2008 presidential vote results by county in Texas, Oklahoma, Arkansas,
Louisiana, Mississippi, Alabama, Tennessee, Kentucky, and Virginia. 322 Over the three
election cycles, the number of counties shaded blue (indicating that a majority of the
county’s votes in the presidential election were cast for the Democratic candidate)
decreased, and the number of red counties (indicating that a majority of the county’s
votes in the presidential election were cast for the Republican candidate) increased.
Mr. Trende testified that these maps supported his hypothesis that the Democratic
coalition has shrunk over time. 323
We are skeptical that presidential voting trends at the county level in states other
than Wisconsin bear directly on the determination that we must make about
Wisconsin’s political geography. Moreover, the color-coding of Mr. Trende’s maps,
although a useful demonstrative, purported to serve as a substitute for quantitative data
on the margin of victory in each county. Without this information, we cannot know
whether, for example, a county won by a Republican presidential candidate was deeply
or narrowly Republican. Nor can we tell how the partisan breakdown of that county
may have changed over time; as long as the county retained the same partisan majority,
it remained the same color. 324 In our view, this evidence is worthy of little, if any,
weight.
The remainder of Mr. Trende’s testimony concerned the political geography of
Wisconsin itself, which he analyzed using a measure called the “partisan index” (“PI”).
The PI, he explained, is the difference between a party’s vote share at one electoral level
and its vote share at a larger electoral level. For example, the Republican PI for the State
attacking shaky but admissible evidence.” Daubert, 509 U.S. at 596. Thus, we deny the plaintiffs’ motion in
limine to exclude Mr. Trende’s testimony.
320
R.150 at 12.
321
Id.
322
Id. at 45–47.
323
Id. at 46.
On cross-examination, Mr. Trende admitted that he did not know if any peer-reviewed study had ever
attempted to analyze geographic clustering by studying trends in the counties won by presidential
candidates.
324
94
of Wisconsin is “computed by subtracting the share of the state that voted for the
Republican presidential candidate from the share of the nation that voted for the
Republican presidential candidate.” 325 The purpose of the PI is “to determine the
partisan lean of political units” 326 in order to “compare results across elections.” 327
Mr. Trende explained that Wisconsin’s statewide PI, as compared to the national
electorate, has remained stable since the 1980s; however, the county and ward PI values
have shifted. He presented color-coded maps illustrating Wisconsin’s presidential vote
results by county in 1996, 2004, and 2012. Each county was colored a shade of blue or
red depending on its degree of partisanship, e.g., counties with large Democratic or
Republican PI values were shaded dark blue or dark red, respectively. Although the
maps did not contain the actual county PI values, Mr. Trende testified that the
pro-Democratic PI values of Dane and Milwaukee Counties increased significantly
between 1996 and 2012. 328 He also testified that the combined PI values of three of
Wisconsin’s reliably Democratic counties—Dane, Milwaukee, and Rock—nearly
doubled between 1996 and 2012, despite the statewide Democratic vote share actually
decreasing over that time. 329 On cross-examination, Mr. Trende conceded that the
heavily Republican Ozaukee, Washington, and Waukesha Counties had Republican PI
values as large as the Democratic PI values in Dane and Milwaukee Counties. However,
the trial evidence also showed that the total number of votes cast for major-party
candidates in the Republican counties were significantly smaller than their Democratic
counterparts.
Mr. Trende then applied the PI to Wisconsin’s wards in what he referred to as a
“nearest neighbor” analysis. 330 First, he calculated ward-level PI values in order to
determine the average partisan lean of Wisconsin’s wards from 2002 to 2014.
Mr. Trende testified that, based on his analysis, “over time, the average Democratic
ward had become about two-and-a-half percent more Democratic than it was in
2002”; 331 he did not, however, observe the same trend in Republican wards. Mr. Trende
then grouped the wards into quantiles based on their degree of partisanship—the more
heavily Democratic wards together with similarly Democratic wards and the same for
325
Tr. Ex. 547, at 19.
326
R.150 at 19.
327
Tr. Ex. 547, at 20.
328
R.150 at 27–38.
329
Id. at 40.
330
Tr. Ex. 547, at 30.
331
R.150 at 51.
95
Republican wards—and used a computer program to determine, for each ward in each
grouping, the median distance between that ward and a ward of similar partisanship.
Mr. Trende concluded that, over time, Democratic-leaning wards in each quantile had
grown closer together but Republican-leaning wards actually had grown farther apart.
In his view, this made it more difficult to draw a neutral districting plan that did not
favor Republicans. 332
Although Mr. Trende’s report and testimony provides some helpful background
information on political trends and political geography generally, they do not provide
the level of analytical detail necessary to conclude that political geography explains
Act 43’s disparate partisan effects. 333 Mr. Trende’s conclusions regarding the PI values
of Wisconsin’s counties were based largely on the shaded maps rather than quantitative
data analysis. And although Mr. Trende did provide PI values for particular
pro-Democratic counties, he conceded on cross-examination that several counties had
pro-Republican PI values as large as the pro-Democratic numbers observed in Dane and
Milwaukee counties.
Additionally, we question how useful Mr. Trende’s nearest neighbor analysis is
in the context of this case. The significance of the distance between wards of similar
partisanship is not clear given the restraints placed on the districting process in
Wisconsin. Under the Wisconsin Constitution, Assembly districts must “be bounded by
At trial, the defendants proffered several exhibits, Tr. Exs. 576–579, that contained revisions to
Mr. Trende’s ward-level analysis. The plaintiffs objected to these exhibits as untimely amendments to
Mr. Trende’s expert report. R.150 at 53–54, 72.
332
The revisions were prompted by criticisms levied by Professor Mayer in his own report,
specifically that Mr. Trende should have used the governor’s race, as opposed to the senator’s race, in
calculating the PI for 2006, and that, in his nearest neighbor analysis, Mr. Trende should have taken into
account the fact that wards vary in size across the State of Wisconsin. See id. at 52–53, 63. The revisions
also corrected an error that Mr. Trende had made in writing the computer program that yielded his PI
values. Id. at 56. At no time prior to trial did the defendants file (with or without leave of court) a revised
expert report for Mr. Trende containing these revisions. Nor did the plaintiffs have notice of Mr. Trende’s
revisions prior to trial. Moreover, the criticism and the error that prompted Mr. Trende’s revisions did
not come to light for the first time at trial, which may have justified their admission despite their lack of
timeliness. We therefore sustain the plaintiffs’ objections to these documents and we have not considered
them in our analysis. Even if we had, they would not have affected our decision on liability.
The Dissent concedes this point, but observes that, even if political geography plays “a ‘modest’ role—
for example three to six percent—it would seriously undermine the notion that the Republicans in this
case engaged in a partisan gerrymander of historical proportions.” Dissent at 157. However, on
cross-examination, Mr. Trende could not give a precise estimate of the effect of Wisconsin’s natural
political geography on the efficiency gap; he thought it was “more than 0, but as far as … putting it on the
1-to-100 percent spectrum, I haven’t done that.” R.150 at 98.
333
96
county, precinct, town or ward lines, to consist of contiguous territory and be in as
compact form as practicable.” Wis. Const. art. IV, § 4. Accordingly, the distance
between wards of similar partisanship is relevant to reapportionment only to the extent
that it is feasible that those wards be grouped together in one contiguous district. The
nearest neighbor analysis, however, does not differentiate between those wards that
realistically could be aggregated to form a lawful assembly district—wards that are
physically adjacent (or at least near one another) and not separated by legally
significant boundaries—and those that are not.
This problem is further compounded by Mr. Trende’s use of the median distance
between wards rather than the mean distance. Although the average Republican ward
is twice the size of the average Democratic ward, the undisputed trial evidence was that
the median Republican ward is six times the size of the median Democratic ward. When
the mean is used, however, Professor Mayer demonstrates that the distance between
Democratic and Republican wards of similar partisanship “are exactly parallel,” and the
disparity between Republican- and Democratic-leaning wards and their closest
neighboring ward of similar partisanship substantially decreases. 334
Like Mr. Trende, Professor Goedert testified that Wisconsin’s political geography
inherently favors Republicans. Using Wisconsin’s 2012 Presidential election results,
Professor Goedert employed a uniform swing to adjust the vote share in each ward and
anticipate the results in an election where each party garnered 50% of the total
statewide vote. He then assembled the wards into ten different groups based on this
adjusted percentage of the Democratic vote share. 335 Professor Goedert’s analysis
showed that between seven and eight percent of Wisconsin’s wards had a very high
concentration of Democrats (more than eighty percent), while fewer than one percent of
wards demonstrated a similar strength in Republican vote. 336 He testified that because
significantly more wards in Wisconsin are narrowly Republican than are narrowly
Democratic, it is “fairly easy” “to try to pack Democrats into a small number of
districts,” because “there are so many wards that are already so heavily packed.” 337 For
the same reasons, he explained, it is “easy” to “disperse” Republican voters. 338
334
R.148 at 294–95; see Tr. Ex. 106.
335
See Tr. Ex. 546, at 21–22.
336
R.150 at 184–85.
337
Id. at 185.
Id. at 185–86. At trial, the Court inquired whether data, specifically the Democratic vote share by ward,
was “part of the record here.” Id. at 253. Professor Goedert responded that “[t]hey should be publicly
available.” Id. The court then inquired whether counsel had “any objection to our taking notice of them if
338
97
The persuasiveness of Professor Goedert’s ward-level analysis was called into
question at trial. To begin, the evidence showed that in the 2010 redistricting cycle
Wisconsin’s wards were, for the first time in the state’s modern history, drawn after the
Assembly district lines were created under Act 43. Professor Goedert admitted that he
was unaware of this chronology when he conducted his analysis. The partisan
imbalance in Act 43’s district configuration therefore may have affected Professor
Goedert’s ward-level analysis. Furthermore, Professor Mayer testified that, in this
context, the relevant geographic unit is not the ward but rather the district because, to
create a district plan, wards ultimately must be aggregated into districts, at which point
their biases may disappear. He also presented his own analysis illustrating that
Wisconsin’s ward distribution, although “not perfectly symmetrical,” resembles a
normal distribution (i.e., a bell curve). 339 He testified that such a distribution is closer to
what would be expected given a neutral political geography. When Professor Mayer
aggregated the wards into Act 43’s districts, however, the resulting distribution was
skewed due to “an unusually large number of districts where the Democrats will
receive between 40 and 50 percent” of the district vote. 340 In Professor Mayer’s opinion,
this incongruity between the distributions of Wisconsin’s wards and its districts
demonstrates that Act 43’s partisan imbalance is caused by its district configuration;
indeed, he characterized this distribution of districts as “the fingerprint of a
gerrymander … the absolute DNA of cracking.” 341
Professor Mayer also presented his own analysis of Wisconsin’s political
geography. 342 Specifically, he testified at length about measures known as the “Isolation
they’re publicly available” because, as the court had “these graphic descriptions of actual votes,” “[i]t
might be helpful for us to see what the wards actually look like.” Id. at 253–54. The plaintiffs then filed an
unopposed stipulation regarding the 2008 and 2012 presidential vote totals by ward. See R.152. By
submitting this data, the plaintiffs fulfilled the Court’s request.
The plaintiffs subsequently sought leave for Professor Mayer to file a second declaration which,
analytically and graphically, “compare[d] the vote distribution in 2000s wards to vote distributions in the
post-Act 43 wards.” R.154 at 4–5. The plaintiffs’ submission is beyond the scope of the court’s inquiries
and is tantamount to additional testimony that has not been subject to the rigors of cross-examination.
We therefore deny the plaintiffs’ motion.
339
R.148 at 242.
Id. at 244; see Tr. Ex. 107 (figure demonstrating distribution of wards compared to districts, arranged by
Democratic vote share); see also supra note 247 and accompanying text (discussing Tr. Ex. 15, a histogram
demonstrating distribution of Act 43 districts, arranged by Republican vote share).
340
341
R.148 at 243.
At trial, plaintiffs sought to introduce highlighted sections of twenty-three articles, ranging in length
from two to fifty pages, as “Learned Treatises” under Federal Rule of Evidence 803(18). See R.148 at 142;
342
98
R.149 at 175. Counsel established that the Professors Mayer and Jackman considered these materials
reliable and also had relied upon them, specifically the highlighted portions, for their expert opinions.
See R.148 at 141; R.149 at 175. The highlighted portions of the articles were not read into evidence, and, in
many cases, plaintiffs’ counsel did not elicit further explanation of these articles during direct
examination. Defense counsel objected to the admission of the highlighted sections of the articles on the
ground that, under Rule 803(18), “documents don’t actually come into the evidence. The witness has to
testify to the statement.” R.149 at 175. We agree.
Federal Rule of Evidence 803 provides in relevant part:
The following are not excluded by the rule against hearsay, regardless of whether the
declarant is available as a witness:
…
(18) Statements in Learned Treatises, Periodicals, or Pamphlets. A statement contained
in a treatise, periodical, or pamphlet if:
(A) the statement is called to the attention of an expert witness on
cross-examination or relied on by the expert on direct examination; and
(B) the publication is established as a reliable authority by the expert’s
admission or testimony, by another expert’s testimony, or by judicial notice.
If admitted, the statement may be read into evidence but not received as an exhibit.
Fed. R. Evid. 803. The rule is straightforward. In order for this exception to apply, counsel first must
either call the statement to the attention of the expert witness on cross or establish that the expert has
relied upon the statement in his or her direct. Second, counsel must establish that the publication from
which the statement came is reliable. When these requirements are met, the statement may be read into
evidence; however, it may not be received as an exhibit. Every authority that we have located has
confirmed this reading of the rule: when the prerequisites are met, the document containing the
statement may not be admitted into evidence; only the statement, on which the expert is relying, may be
read into evidence. See Finchum v. Ford Motor Co., 57 F.3d 526, 532 (7th Cir. 1995) (“Under Rule 803(18),
statements contained in a published periodical which are relied upon by an expert witness may be
admitted, but they must be read into evidence rather than received as exhibits.” (emphasis added)); Graham
ex rel. Graham v. Wyeth Labs., 906 F.2d 1399, 1414 (10th Cir. 1990) (quoting J. Weinstein & M. Berger, 4
Weinstein’s Evidence ¶803(18)[2] for the proposition that “the last paragraph of Rule 803(18) bars the
admission of treatises as exhibits”); Fisher v. United States, 78 Fed. Cl. 710, 714 (2007) (sustaining objections
to plaintiffs’ proposed use of books and articles as exhibits and stating that, “[i]f plaintiff wishes to
introduce at trial relevant statements from those learned treatises, plaintiff may do so, provided and to
the extent they have been relied on by an expert witness in the formulation of his or her direct testimony,
by instructing her witnesses to read the statements into the record” and further noting that “the treatises
themselves may not be admitted into evidence as exhibits”); see also Jack B. Weinstein and Margaret
Berger, 5 Weinstein’s Federal Evidence §803.20[1] (2d ed. 2016) (“Moreover, information that qualifies for
this exception ‘may be read into evidence but not received as an exhibit.’ This limitation ensures that the
jurors will not be unduly impressed by the treatise, and that they will not use the text as a starting point
for conclusions untested by expert testimony.” (footnote omitted)); Michael H. Graham, Handbook of
Federal Evidence §803(18) at 472 (7th ed. 2012) (“A safeguard against jury misuse of the published
99
Index” and “Global Moran’s I,” which he said are far more common in this area of
academic study than the methods employed by the defendants’ experts. 343 According to
Professor Mayer, he used the Isolation Index to measure the extent to which the average
Republican or Democratic voter lives in a ward that leans more heavily Republican or
Democratic than the state as a whole. 344 Global Moran’s I, he explained, was used to
measure the likelihood that a Republican- or Democratic-leaning ward is adjacent to a
similarly Republican- or Democratic-leaning ward. 345 Professor Mayer testified that both
of these measures show that Wisconsin’s political geography is neutral and does not
inherently favor one party or the other. 346
We do not find these methods reliable as they have been applied in this context.
Professor Mayer acknowledged on cross-examination that he had not heard of the
authority is found in the final sentence of Rule 803(18) which provides that statements may be read into
evidence, but not received as an exhibit and thus cannot [be] taken to the jury room.”).
The plaintiffs maintain that the defendants’ “interpretation and proposed application … of Rule
803(18),” which matches our own, “is directly contradicted by the text of the Rule itself, defies common
sense, and would displace the Court’s discretion over the admission of evidence and how best to achieve
the ‘just, speedy, and inexpensive determination’ of this action.” R.161 at 2. They submit that that the
defendants are attempting “to graft an additional requirement for the admissibility of statements in
learned treatises that Rule 803(18) does not contain.” Id. at 5. In plaintiffs’ view, once the prerequisites set
forth in subsections (A) and (B) are met, the statement is admitted for all purposes; the last statement
simply indicates the proffering party’s “option of reading it into the record.” Id. (emphasis in original).
We do not believe that the plaintiffs’ approach can be squared with the blanket prohibition, set
forth explicitly in Rule 803(18), that the statements in learned treatises may “not [be] received as an
exhibit.” Moreover, as we already have explained, their interpretation is at odds with the case law and
commentary. Their position does not even find support in the one case that they cited in their submission,
DaGraca v. Laing, 672 A.2d 247 (N.J. Super. Ct. App. Div. 1996). DaGraca stands for the unexceptional
proposition that learned-treatise statements may be introduced on cross-examination as long as they are
established as reliable through some accepted means. Id. at 299–300. It does not speak to, and therefore
does not support, the plaintiffs’ contention that learned-treatise statements may be offered through
documentary evidence.
For these reasons, we sustain the defendants’ objections to the admission of the highlighted
portions of exhibits 98–100, 102, 118–119, 131, 141, 148, 150–152, 333, 391, 394, 405–406, 408, 414–415, 417,
and 498. We have considered, however, all statements from these authorities included within the
testimony of Professors Mayer and Jackman.
343
R.149 at 5–23.
344
Id. at 15, 17–18.
345
Id. at 6, 11–13.
346
Id. at 13–14, 21–22.
100
Isolation Index before he was retained as an expert in this case. 347 Similarly, Professor
Mayer testified that he had never calculated the Global Moran’s I measure before he
was retained for this litigation. 348 Moreover, the defendants emphasized during trial
that Professor Mayer relied on scholarly articles that either used a related measure
known as Local Moran’s I, or used Global Moran’s I to study demographic groups. 349
He could not point to any peer-reviewed, scholarly article that had used either measure
specifically on partisanship. 350
347
Id. at 28.
Id. at 40. During his testimony, Professor Goedert also critiqued Professor Mayer’s use of these
measures. The plaintiffs objected to much of this testimony as outside the scope of Professor Goedert’s
expert report. R.150 at 191–92. We agree, sustain the objection, and therefore do not rely on
Professor Goedert’s testimony in reaching our conclusions about the infirmity of Professor Mayer’s
conclusions.
348
349
See R.149 at 41 (acknowledging that Chen and Rodden use “the Local Moran’s I”).
Id. at 39. (“Q. Okay. You say that the Glaeser article, that’s the one example that we saw of this being
used to determine the distribution of partisans? A. That’s the example that I cited. … Q. I believe you said
this wasn’t peer reviewed, was it? A. Not as far as I know.”).
350
In addition to the testimony of their experts, the plaintiffs also ask us to consider a forthcoming
article by Professor Jowei Chen analyzing Wisconsin’s political geography. We decline to do so.
The defendants and their experts relied on previously published articles by Professor Chen,
which included randomly simulated district maps for multiple states other than Wisconsin, to argue that
Wisconsin’s natural political geography favors Republican voters. See R.46 at 27; Tr. Ex. 547, at ¶¶ 89–90,
126; Tr. Ex. 136, at 18, 21; R.150 at 111–12, 243–44. On March 17, 2016, one week before the summary
judgment hearing in this case, Professor Chen filed a motion for leave to participate as an amicus curiae,
contending that the defendants and their experts had “misinterpreted and misapplied” his work to the
facts of the present case. R.82-1 at 3. Attached to his motion, Professor Chen included an analysis
applying the simulation methodology that he used in his published work to Wisconsin. R.82-2. We
denied Professor Chen’s request to participate because the timing left “the parties insufficient time to
respond.” R.85.
Professor Chen subsequently submitted his analysis of Wisconsin’s political geography as an
article for publication to the Election Law Journal, where it was accepted and is forthcoming in 2017. The
plaintiffs requested that the article be admitted into evidence. At the conclusion of trial, we requested that
the parties address the admissibility of Professor Chen’s article in their post-trial briefs. In their post-trial
brief, the plaintiffs maintain that we should admit Professor Chen’s article to “correct[] defendants’
misrepresentations of Professor Chen’s work.” R.155 at 26. As we have not relied on any of these
“misrepresentations” in our analysis of the issues before us, we find it unnecessary to consider how
Professor Chen’s later scholarship might alter our views of either his original work or the defendants’
interpretation of his work.
The plaintiffs also argue that the article should be admitted because “Professor Mayer relied on
Professor Chen’s article in formulating his own expert opinions.” Id. at 33–34. There is no support for this
101
Having carefully examined the evidence bearing on this issue, we find that
substantial portions of the record indicate, at least circumstantially, that Wisconsin’s
political geography affords Republicans a modest natural advantage in districting.
Indeed, the plaintiffs conceded as much in their closing argument when counsel stated
that “there likely is some natural packing” of Democratic voters, “especially of minority
assertion in the record. During his deposition, Professor Mayer responded “I did,” to the following
question: “Dr. Mayer, subsequent to you preparing your rebuttal report, did you receive and did you review a
document entitled Dr. Joey Chen’s Analysis of Wisconsin’s Act 43?” R.99 at 36 (Mayer Dep. at 138)
(emphasis added). Professor Mayer then stated that the article was “additional confirmation of my own
analysis that indicated that there was no geographic clustering of … Democrats and Republicans that
would produce a natural pro-Republican gerrymander.” Id. In sum, the article played no part in Professor
Mayer’s analysis and merely confirmed, after the fact, the analysis that he had conducted.
The timing and nature of Professor Chen’s submission counsel against admitting it into evidence
in this case. Professor Chen’s analysis is highly technical in both methodology and substance; it is, in
effect, an expert report prepared specifically for this litigation. Under the Federal Rules of Civil
Procedure, parties must “disclose to the other parties the identity of any witness it may use at trial to
present” expert testimony. Fed. R. Civ. P. 26 (a)(2)(A). Accompanying the disclosure of their identity,
Rule 26 further mandates, unless the court orders otherwise, that retained experts prepare and sign a
written report stating and supporting their opinions. Fed. R. Civ. P. 26(a)(2)(B). Here, we ordered that the
plaintiffs disclose their experts and their reports by October 23, 2015, and the defendants by December 2,
2015. R.33 at 2. We also permitted rebuttal reports to be filed by December 16, 2015. Id. The plaintiffs did
not disclose Professor Chen as an expert at either of these times. Indeed, the court was not made aware of
Professor Chen’s interest in this case until he filed his amicus brief on March 17, 2016, one week before
the summary judgment hearing.
Moreover, because Professor Chen was not identified as an expert, he was not deposed and did
not testify at trial. The admissibility of expert testimony in federal court is governed by Federal Rule of
Evidence 702 and Daubert, 509 U.S. 579. These authorities set forth guideposts designed to assist district
courts, as “the gatekeeper[s] of expert testimony,” in assessing and ensuring the reliability of an expert’s
principles and methods. C.W. ex rel. Wood v. Textron, Inc., 807 F.3d 827, 834–35 (7th Cir. 2015); see also
Fed. R. Evid. 702. In addition to our admissibility determination, “the normal adversarial process of
‘[v]igorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of
proof,’” Lees v. Carthage Coll., 714 F.3d 516, 526 (7th Cir. 2013) (quoting Daubert, 509 U.S. at 596), is
designed to test “[t]he reliability of data and assumptions used in applying [the expert’s] methodology,”
Manpower, Inc. v. Ins. Co. of Penn., 732 F.3d 796, 808 (7th Cir. 2013). Here, we are unable to examine
properly the reliability of Professor Chen’s methodologies, and we are without the benefits of adversarial
scrutiny. We therefore cannot consider his submissions as part of the record before us. Cf. Kitzmiller v.
Dover Area Sch. Dist., No. 04CV2688, 2005 WL 2736500, at *1–2 (M.D. Pa. Oct. 24, 2005) (striking amicus
brief because it was a “‘back door’ attempt to insert expert testimony into the record free of the crucible of
trial and cross-examination”). We therefore sustain the defendants’ objections to the admission of all
exhibits related to Professor Chen’s analysis of Wisconsin districting, Tr. Exs. 154–160. For the same
reasons, we also disregard Professor Mayer’s trial testimony regarding Professor Chen’s
Wisconsin-specific analysis, R.148 at 256–68; R.149 at 22–23.
102
voters in places like Milwaukee.” 351 Several pieces of evidence lead us to this
conclusion. The first, and most compelling, is Professor Mayer’s analysis comparing the
distributions of Wisconsin’s wards and Act 43’s districts by Democratic vote share. As
Professor Mayer himself testified, the ward-level distribution is “not perfectly
symmetrical.” 352 In fact, the mean ward in the distribution—the highest point on the
curve—is located left of the fifty percent line, which indicates that the average ward in
Wisconsin leans slightly Republican. His analysis also shows that there are a substantial
number of wards that are over eighty percent Democratic, but virtually no wards that
are similarly Republican. We find these facts to be consistent with the notion that
Democratic voters are uniquely packed in urban centers like Milwaukee and Madison.
Moreover, Mr. Trende’s testimony establishes that the counties with the highest
Democratic PI values are far larger in population than counties with equivalent
Republican PI values. This fact indicates that some of the most heavily Democratic areas
in Wisconsin are more densely populated than their equally Republican counterparts.
Again, we find this to be consistent with a modest Republican advantage in the State’s
political geography.
We also find it significant that Republican-leaning wards in Wisconsin tend to be
twice the size of Democratic-leaning wards. Indeed, when Professor Mayer conducted
his own nearest neighbor analysis using the mean distances between wards, it became
clear that this size differential exists at every level of partisanship. 353 We recognize that
the impact of this disparity on the districting process arguably is negligible because
districts must be approximately equal in population; ward size, therefore, does not
directly bear on the creation of districts. Still, the tendency of Republican wards to be
much larger than Democratic wards is consistent with the notion that Democratic voters
on the whole are more likely than Republican voters to live in geographically
concentrated areas. This, in turn, increases the prospect that heavily Democratic wards
will exist within the same political boundary such that it is, at least somewhat, more
difficult to draw politically competitive districts in that part of the state.
Finally, it is undisputed that Professor Mayer’s Demonstration Plan itself
exhibited a slight pro-Republican bias despite his stated objective, reiterated at trial, of
drawing an alternative to Act 43 that performed comparably on traditional districting
objectives but “had an efficiency gap as low to zero as [he] could get it.” 354 Under the
351
R.150 at 267.
352
R.148 at 242; see supra at 97.
353
See Tr. Ex. 106.
354
R.148 at 146. Professor Mayer also testified at trial that he “probably could have” achieved a lower EG
103
Demonstration Plan, when the Republicans secure 48% of the statewide vote as they did
in 2012, the plan still yields an EG of 2.2% in favor of the Republicans. This certainly is a
far smaller advantage than the 11.69% pro-Republican EG generated under Act 43 in
2012, 355 but it nevertheless illustrates that even a neutrally drawn plan, crafted under
conditions unimpeded by politics, imposes a slight burden on Democratic voters.
For these reasons, we find that Wisconsin’s political geography, particularly the
high concentration of Democratic voters in urban centers like Milwaukee and Madison,
affords the Republican Party a natural, but modest, advantage in the districting process.
B.
Because the evidence at trial establishes that Wisconsin has a modestly
pro-Republican political geography, we now examine whether this inherent advantage
explains Act 43’s partisan effect. We conclude that it does not.
The record reveals that, before the legislature enacted Act 43, its drafters had
produced several alternative district plans that performed satisfactorily on traditional
districting criteria but secured a materially smaller partisan advantage when compared
to the advantage produced by Act 43. Foltz and Ottman testified that, while drafting a
particular map, they would remain attentive to various districting criteria—population
equality, compactness, contiguity, and municipal and county splits—as well as where
incumbents lived and levels of disenfranchisement. 356 When the drafters finalized a
statewide map, they were able to generate various reports through the autoBound
in creating the Demonstration Plan but “when [he] got to the point where [he] had an efficiency gap of 2.2
and a map that was equivalent to Act 43, [he] stopped.” Id. at 185–86. This statement, of course,
contradicts his previous trial testimony that he was pursuing an EG of zero. It also contradicts his expert
report, in which he stated that his objective was to design an alternative plan “that has an efficiency gap
as close to zero as possible while complying with” traditional districting criteria as well as Act 43. R.54 at
2. In light of these statements, we discount Professor Mayer’s attempt at trial to minimize the significance
of the Demonstration Plan’s 2.2% pro-Republican EG. We believe that the Demonstration Plan reflects
Professor Mayer’s stated goal of achieving an EG of zero and, therefore, that his failure to achieve that
goal is a material fact in our evaluation of Wisconsin’s political geography.
The Dissent pegs the EG for the Demonstration Plan, when adjusted for incumbency, as 4%. See
Dissent at 136. This figure comes from Professor Mayer’s swing analysis. To conduct that analysis,
Professor Mayer accounted for incumbency, and, after that adjustment, the Demonstration Plan’s EG rose
to 3.89%. See Tr. Ex. 116. When he applied the same factor to Act 43, the EG for Act 43 rose to 14.15%. See
Tr. Ex. 117.
355
356
See R.147 at 154–60; R.148 at 83–90; see also supra at 10.
104
software that evaluated the plan on these different districting criteria. 357 In particular,
once the drafters had “a statewide plan finalized, all 99 assembly districts,” they would
“take th[e] [partisan] composite column from auto[B]ound and then move it over into
… Excel spreadsheets.” 358 These spreadsheets evaluated a plan’s expected
district-by-district partisan performance, and the drafters exported and saved them for
numerous statewide draft plans.
Although the autoBound software also enabled the drafters to generate reports
on other districting criteria that they were considering, the defendants have not pointed
us to any documents in the record that compare the various maps under consideration
according to traditional district criteria. 359 It therefore is unclear precisely how the
drafters’ statewide maps performed on other districting criteria. Nevertheless, Foltz
testified that the drafters “would pull regional alternatives from” the statewide maps
they had finalized and evaluated. 360 These regional maps were then presented to the
Republican leadership with the expectation that they ultimately would be a part of a
final district plan. 361 Neither Foltz nor Ottman testified, and nothing in the record
indicates, that any of these statewide plans performed unsatisfactorily on any other
districting criteria. Indeed, had these maps demonstrated, for instance, insufficiently
compact districts or an unacceptable number of municipal splits, the drafters would not
have pulled regional alternatives from them to present to the legislative leadership. We
therefore can infer that the finalized statewide plans for which we have partisan
performance spreadsheets in the record complied satisfactorily with the other
districting criteria that the drafters considered.
The evidence also revealed that as the reapportionment process progressed and
the drafters finalized and evaluated these statewide draft plans, the magnitude of the
expected partisan advantage increased. In many instances, the names of these plan
357
R.147 at 155–56; R.148 at 84, 102–05.
R.147 at 162; see also id. at 51 (Foltz testifying that after a map was completed he “would export or copy
and paste … the data out of the auto[B]ound matrix, and then put it into an Excel file that summarized
the partisan scores”).
358
359
Ottman did create some spreadsheets on disenfranchisement. See supra note 195.
360
R.147 at 163.
Id. at 163–64; id. at 176 (Foltz testifying that the drafters would “pull[] regional alternatives from within
a broader statewide plan and present[] [them] to the leadership”); R.148 at 94–95 (Ottman testifying that
“[a]fter [they] had made a number of draft maps and set up meetings for legislative leadership to come
over, [the drafters] … discussed how to kind of break up the state[] in regions to discuss with the
legislative leadership, and then [they] each kind of printed off maps that [they] had been working on for
those different regions”).
361
105
alternatives reflected the degree of partisan advantage that could be anticipated in the
map, e.g., “Assertive” or “Aggressive.” 362 Each of the drafters’ partisan score
spreadsheets included a corresponding table comparing the partisan performance of the
draft plan to the Current Map. These performance comparisons were made on the
following criteria: “Safe” Republican seats, “Lean” Republican seats, “Swing” seats,
“Safe” Democratic seats, and “Lean” Democratic seats. 363 Under the Current Map, the
drafters anticipated that the Republicans would secure 49 Assembly seats, 364 with 40
districts safe or leaning Republican, 40 districts safe or leaning Democratic, and 19
swing districts. 365 However, by the time the drafters had solicited the preferences of the
Republican legislative leadership and pieced together the Team Map—the closest
version in the record to Act 43—the expected Republican seats had ballooned to 59.366
The number of safe or leaning Republican districts had grown from 40 to 52, apparently
at the expense of swing districts, which decreased from 19 to 10. 367 The number of safe
or leaning Democratic districts also were reduced from 40 to 37. 368
Careful review of the record convinces us that benign factors cannot explain this
substantial increase in Republican advantage between the Current Map and the plan
that would become Act 43. Rather, it is evident that the drafters achieved this end by
making incremental “improvements” to their plan alternatives throughout the drafting
process. For example, the Republican advantages expected in the drafters’ initial two
draft plans, produced in early April 2011, were significantly smaller than the advantage
anticipated in the Team Map. Under these draft plans—Joe’s Basemap Basic and Joe’s
Basemap Assertive—the drafters expected Republican candidates to win 52 and 56
seats, respectively, compared to the 49 expected under the Current Map. 369 The Current
Plan’s 40 safe and leaning Republican districts improved to 45 and then to 49, while the
362
See R.148 at 19–21.
363
See, e.g., Tr. Ex. 364; see also R.148 at 15 (Ottman testifying about these criteria).
364
See Tr. Ex. 283.
365
Tr. Ex. 364.
366
Tr. Ex. 283.
367
Id.
368
Id.
369
Tr. Exs. 465, 476.
106
number of swing districts dwindled from 19 to 14 to 12. 370 The number of
pro-Democratic districts, however, remained relatively constant. 371
Apparently not satisfied with the political performance of these early plans, the
drafters produced and evaluated at least another six statewide maps prior to their
meeting with the Republican leadership in early June 2011. 372 Each of these maps
improved upon the anticipated pro-Republican advantage generated in the initial two
draft plans. The total number of expected Republican seats in these drafts ranged
between 57 and 60, and the number of swing seats ranged between 6 and 11. 373 The
number of Democratic seats again remained about the same under each draft map. 374
The Team Map, as an amalgamation of several statewide plan alternatives,
reflects the drafters’ iterative efforts throughout the drafting process to achieve a
substantial, if not maximal, partisan advantage. That these efforts were highly
successful is obvious with the benefit of hindsight. But the drafters themselves took
pains to gauge their success at the time, taking stock of the degree to which they had
improved upon the Current Map. In their Tale of the Tape, the drafters compared the
partisan performance of the Team Map directly to the Current Map.375 They highlighted
specifically that under the Current Map, “49 seats are 50% or better,” but under the
Team Map, “59 Assembly seats are 50% or better.” 376 In a second document, they
categorized each of Wisconsin’s Assembly districts according to its partisan
“improvement” from the Current Map to the Team Map. 377 For example, five districts
were “Statistical Pick Up[s],” held by a Democratic incumbent who would now face a
“55% or better” Republican vote share. 378 Another fourteen districts were “strengthened
a lot”: “Currently held GOP seats that start[ed] at 55% or below [and] improve[d] by at
370
Tr. Ex. 465.
371
Id. (showing “strong” and “lean” Democratic seats of 40, 40, and 38.)
These were: Milwaukee_Gaddie_4_16_11_V1_B (Tr. Ex. 172, at 1); Statewide2_Milwaukee_Gaddie_
4_16_V1_B (Tr. Ex. 172, at 2); Tad MayQandD (Tr. Exs. 364, 477); Joe Assertive (Tr. Exs. 366, 478); Tad
Aggressive (Tr. Ex. 283); and Adam Aggressive (Tr. Ex. 283).
372
373
Tr. Exs. 172, 364.
374
Tr. Exs. 172, 364, 366 (showing safe and leaning Democratic seats ranging from 38–40).
375
Tr. Ex. 283.
376
Id.
377
Tr. Ex. 284, at 1.
378
Id.
107
least 1%” in Republican vote share. 379 The drafters also made particular note of which
Republican Assembly members had contributed to the achievement of their partisan
goals, the 20 so-called “GOP donors to the team.” 380
The substantial record evidence of the multiple statewide plan alternatives
produced during the drafting process convinces us that Wisconsin’s modest,
pro-Republican political geography cannot explain the burden that Act 43 imposes on
Democratic voters in Wisconsin. The drafters themselves disproved any argument to
the contrary each time they produced a statewide draft plan that performed
satisfactorily on legitimate districting criteria without attaining an expected partisan
advantage as drastic as that demonstrated in the Team Map and, ultimately, in Act 43.
In reaching this conclusion, we emphasize that we did not require, as the plaintiffs
initially proposed, that the defendants show that Act 43’s partisan effect was necessary
or unavoidable. Rather, our task at trial was to determine whether the burden that Act 43
imposes is justifiable in light of legitimate districting considerations and neutral
circumstances. The defendants offered Wisconsin’s natural political geography as one
such neutral circumstance. Because we find that a Republican advantage in political
geography, although it exists, cannot explain the magnitude of Act 43’s partisan effect,
and because we find that the plan’s drafters created and passed on several less
burdensome plans that would have achieved their lawful objectives in equal measure,
we must conclude that the burden imposed by Act 43 is not justifiable.
Professor Mayer’s Demonstration Plan provides additional evidence that the
legislative imbalance resulting from Act 43 is not attributable to political geography.
Professor Mayer attempted to draw an alternative districting plan to Act 43 “that had
an efficiency gap as low to zero as I could get it” while also complying with traditional
districting criteria as well as Act 43. 381 He first created a regression model that estimated
partisanship for each geographic area, so that he could compare his plan to Act 43. 382 To
ensure the model was accurate, Professor Mayer compared the predictions made by his
regression model to the actual results in 2012. He concluded that the results aligned
almost perfectly. 383
379
Id.
380
Id.
381
R.148 at 146.
382
Id. at 151.
To assess “the overall accuracy of the model,” Professor Mayer used what he characterized as “one of
the most important diagnostics … called the R squared,” which is “a measure that tells … what percentage
of the variation in the dependent variable [a regression] model is picking up.” Id. at 162–63. Professor
383
108
Once he was confident in his model, Professor Mayer “used a GIS redistricting
program called Maptitude for redistricting to go ahead and complete the task of actually
drawing the Assembly district map.” 384 Proceeding along this course, Professor Mayer
was able to draw a districting map that would have yielded a pro-Republican EG of
only 2.2% for 2012, and “is comparable to Act 43” with respect to “all constitutional
requirements.” 385 Specifically his plan has a population deviation of .86% whereas Act
43 has a population deviation of .76%. 386 He also noted that his plan keeps the same
number of majority-minority districts. 387 The plan is also slightly more compact, based
on the “Reock score,” than Act 43. 388 Finally, it had three fewer county splits but two
more municipal splits than Act 43. 389
The defendants argue that we should discount the evidentiary value of the
Demonstration Plan for several reasons. First, they maintain that the Demonstration
Plan “achieved its EG through 20/20 hindsight” and that the low EG will “hold only for
those specific election conditions” that occurred in 2012. 390 Specifically, the defendants
contend that if the Republicans had a good electoral outcome like the one they saw in
2014, they would have received 63 seats under the Demonstration Plan and ended up
with the same EG as Act 43. 391 Consequently, from the standpoint of partisan effects, the
Demonstration Plan is just as problematic as Act 43.
Mayer testified that his model and the actual results had an “R squared” of .9903, which he characterized
as “ridiculously high. It’s the kind of number you almost never see in social science research.” Id. at 163.
Professor Mayer also compared the results of his analysis to Professor Gaddie’s “open-seat baseline
partisanship measure.” Tr. Ex. 2, at 29–30. Professor Mayer found that “[t]he r-squared for this regression
[comparing the results] is 0.96, indicating that the two measures are almost perfectly related, and are both
capturing the same underlying partisanship.” Id. at 30.
384
R.148 at 151.
385
Tr. Ex. 2, at 37.
386
R.148 at 177.
Id. The Demonstration Plan retains the same majority Latino district that the federal court drew in
Baldus. However, the six majority African-American districts have boundaries that differ from Act 43. See
Tr. Ex. 2, at 37–38.
387
388
R.148 at 178.
389
Id. at 178–79.
390
R.153 at 25.
391
See id.; R.149 at 94–101.
109
Although this evidence shows the need to test how the Demonstration Plan fares
under likely electoral scenarios, 392 it does not render the Demonstration Plan useless for
our purposes. Under Professor Mayer’s Demonstration Plan, the EG would be
significantly pro-Republican had the Republicans received a high vote share in the first
election year of the plan. However, had the opposite happened, and Democrats
received a higher vote share in the first election year, the EG would have skewed
towards the Democrats. This is because the Demonstration Plan was designed to have
competitive districts, and the EG will be reactive to such districts. By contrast, as
Professor Gaddie’s and Professor Jackman’s sensitivity analyses show, Act 43 will
remain pro-Republican regardless of the electoral outcome. Consequently, the
Demonstration Plan and Act 43 do not suffer from the same infirmities.
The defendants also contend that Professor Mayer’s Demonstration Plan fails to
account for core retention, i.e., it does not try to keep districts from the previous
districting plan in a similar form. 393 Although there is testimony by Ottman that the
drafters “looked at kind of what the core of the existing district was compared to the
new district,” 394 we question how much this consideration actually factored into the
drafting process. In Baldus, the court noted that “[o]nly 323,026 people needed to be
moved from one assembly district to another in order to equalize the populations
numerically,” but that “Act 43 moves more than seven times that number – 2,357,592
people”—a number that the court found to be “striking.” 849 F. Supp. 2d at 849
(emphasis added).
On a similar note, the defendants point out that Professor Mayer did not draw
Senate districts and therefore did not account for how many voters would be
disenfranchised by moving into a Senate district where they would not get a vote for
another two years. 395 Ottman testified that, because he worked for Senator Fitzgerald
and “was familiar with the Senate seats,” he “was able to eyeball [disenfranchisement] a
little bit.” 396 Foltz testified that “you can notice [disenfranchisement] when you’re
drawing individual districts. But I think it’s another one of those metrics where the
392
See supra at 25, 28 (discussing sensitivity testing).
R.153 at 24. Professor Mayer admitted during his testimony that he did not take account of core
retention. R.149 at 117–19. That being said, we do not know whether Professor Mayer inadvertently
retained several of the old districts when drawing the Demonstration Plan, or how it compares to Act 43
on core retention. We simply know that Professor Mayer did not consider core retention.
393
394
R.148 at 85.
See R.153 at 24; R.149 at 117–19. As with core retention, we do not know whether the Demonstration
Plan disenfranchised a significant number of voters. We simply know that the issue was not considered.
395
396
R.148 at 86.
110
back-end report is really where you get a sense for where you’re sitting.” 397 Although
Foltz ran “disenfranchisement reports on [his] plans,” he did not testify as to specific
numerical targets he was aiming for, nor did he testify that any of his maps were
changed in response to the reports that were generated. 398
The defendants also urge that the Demonstration Plan incorporates districts
around Fond du Lac that are not compact. 399 There may be individual districts in the
Demonstration Plan that are not as compact as their equivalent districts in Act 43.
Nevertheless, the Demonstration Plan has a better overall “Reock score”—the measure
of compactness utilized by the drafters—than Act 43 has. 400 We cannot conclude,
therefore, that, overall, the Demonstration Plan was less compact than Act 43.
Finally, the defendants argue that the Demonstration Plan fails to protect
incumbents to the same degree as Act 43. Professor Mayer testified that he “didn’t pay
attention to where incumbents resided.” 401 The defendants contend that the number of
paired incumbents in the Demonstration Plan was so great that such a plan would not
have passed in the legislature. 402 According to the defendants, the Demonstration Plan
paired 37 incumbents, 403 13 more than were paired in Act 43. 404
There is no question that, unlike Act 43, the Demonstration Plan does not take
into account incumbency concerns. This infirmity does not negate entirely the value of
the Demonstration Plan. Notably, the defendants have not argued that the location of
incumbents hampered them in their efforts to draw a non-partisan plan or otherwise
accounts for the electoral imbalance resulting from Act 43. Nevertheless, Professor
397
R.147 at 157–58.
Id. at 158. As noted previously, Ottman created some spreadsheets on disenfranchisement. See supra
notes 195, 359. However, Ottman did not provide any explanation of these in his testimony. We therefore
do not know how frequently they were generated or whether he made changes to any maps in response
to those numbers.
398
399
R.153 at 24; see also R.149 at 106–08.
400
See R.148 at 178.
401
R.149 at 84.
For example, District 53 of the Demonstration Plan paired three Republican incumbents. R.149 at 111.
The Demonstration Plan also paired incumbents in the area where Senator Mike Ellis actually had
complained about incumbent pairings during the drafting of Act 43. Id. at 112. Third, the plan paired two
incumbents in a majority-minority district. Id. at 113.
402
403
R.156 at 5.
See Tr. Ex. 192. Ottman testified that, “[i]f he could recall correctly, … there were 22 legislators paired
in the final map. R.148 at 87. Regardless whether the number is 22 or 24, the Demonstration Plan
represents a significant increase in incumbent pairings.
404
111
Mayer’s lack of attentiveness to this concern well might diminish the Demonstration
Plan’s worth as a viable, legislative alternative. The Demonstration Plan still shows,
however, that it is very possible to draw a map with much less of a partisan bent than
Act 43 and, therefore, that Act 43’s large partisan effect is not due to Wisconsin’s natural
political geography.
The evidence of multiple statewide plan alternatives produced during the
drafting process, coupled with Professor Mayer’s Demonstration Plan, convinces us that
Wisconsin’s modest, pro-Republican political geography cannot explain the burden that
Act 43 imposes on Democratic voters in Wisconsin. The drafters established this finding
themselves; they produced several statewide draft plans that performed satisfactorily
on legitimate districting criteria without attaining the drastic partisan advantage
demonstrated in the Team Map and, ultimately, in Act 43. Professor Mayer’s
Demonstration Plan further dispels the defendants’ claim. As we have noted in our
discussion, the evidence in support of a larger effect of political geography simply
lacked specificity and careful analysis and, consequently, was less convincing.
VI
STANDING
Lujan v. Defenders of Wildlife, 504 U.S. 555, 561 (1992), makes clear that we must
assess the issue of standing at all stages of the proceedings. Therefore, now that we
have set forth the factual record and the elements of a political gerrymandering cause of
action, we revisit the issue of standing. 405 The standing requirement is meant to ensure
that the plaintiffs have “alleged such a personal stake in the outcome of the controversy
as to assure that concrete adverseness which sharpens the presentation of issues upon
which the court so largely depends for illumination of difficult constitutional
questions.” Baker v. Carr, 369 U.S. 186, 204 (1962). The party invoking federal
jurisdiction, here the plaintiffs, bears the burden of establishing Article III standing.
DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 342 (2006).
The constitutional requirements for standing are well-established:
First, the plaintiff must have suffered an “injury in fact”—an invasion of a
legally protected interest which is (a) concrete and particularized, and (b)
“actual or imminent, not ‘conjectural’ or ‘hypothetical.’” Second, there
must be a causal connection between the injury and the conduct
complained of—the injury has to be “fairly … trace[able] to the challenged
action of the defendant, and not … th[e] result [of] the independent action
405
See supra at 30.
112
of some third party not before the court.” Third, it must be “likely,” as
opposed to merely “speculative,” that the injury will be “redressed by a
favorable decision.”
Lujan, 504 U.S. at 560–61 (alteration in original) (citations omitted) (footnote omitted).
We turn first to the question whether the plaintiffs have established the invasion
of a legally protected interest. Although the proposition is not settled in Supreme Court
jurisprudence, we hold, for the reasons set forth in this opinion, that state legislatures
cannot, consistent with the Equal Protection Clause, adopt a districting plan that is
intended to, and does in fact, entrench a political party in power over the decennial
period. The plaintiffs have established that, “[a]s a result of the statewide partisan
gerrymandering, Democrats do not have the same opportunity provided to Republicans
to elect representatives of their choice to the Assembly. As a result, the electoral
influence of plaintiffs and other Democratic voters statewide has been unfairly [and]
disproportionately … reduced” for the life of Act 43. 406 Professor Whitford testified to
the impact of political gerrymandering on individual voters in Wisconsin where it is
“extremely difficult” to pass legislation through a bipartisan coalition. 407 Wisconsin’s
strict caucus system means that all of the important “debate and discussion” of
proposed legislation takes place in the party caucus meeting, and the party’s vote, yea
or nay, is the one “that matters.” 408 Consequently, erecting a barrier that prevents the
plaintiffs’ party of choice from commanding a legislative majority diminishes the value
of the plaintiffs’ votes in a very significant way.
We believe the situation here is very close to that presented in Baker v. Carr, 369
U.S. 186. In Baker, the plaintiffs’ constitutional claim was that a decades-old districting
statute
constitute[d] arbitrary and capricious state action, offensive to the
Fourteenth Amendment in its irrational disregard of the standard of
apportionment prescribed by the State’s Constitution or of any standard,
effecting a gross disproportion of representation to voting population. The
injury which appellants assert is that this classification disfavors the
voters in the counties in which they reside, placing them in a position of
406
R.1 at 6–7, ¶ 16.
407
R.147 at 33.
Id. Professor Whitford’s description is consistent with the testimony given by Foltz and Ottman
concerning the drafting and passage of legislation generally, as well as the path of Act 43. We therefore
credit Professor Whitford’s description of the legislative process in Wisconsin.
408
113
constitutionally unjustifiable inequality vis-à-vis voters in irrationally
favored counties.
Id. at 207–08. The Court explained that, “[i]f such impairment does produce a legally
cognizable injury, [the appellants] are among those who have sustained it.” Id. at 208.
As noted above, today we recognize a cognizable equal protection right against stateimposed barriers on one’s ability to vote effectively for the party of one’s choice.
Moreover, Act 43 did, in fact, prevent Wisconsin Democrats from being able to translate
their votes into seats as effectively as Wisconsin Republicans. Wisconsin Democrats,
therefore, have suffered a personal injury to their Equal Protection rights—akin to that
suffered by the plaintiffs in Baker—that is both concrete and particularized.
Moreover, there can be no dispute that a causal connection exists between Act 43
and the plaintiffs’ inability to translate their votes into seats as efficiently as
Republicans. The evidence has established that one of the purposes behind Act 43 was
solidifying Republican control of the legislature for the decennial period. Indeed, the
drafters had drawn other statewide maps that, their own analysis showed, would
secure fewer Republican seats. 409 Finally, adopting a different statewide districting map,
perhaps one of those earlier maps or a map as proposed in Professor Mayer’s
Demonstration Plan, would redress the constitutional violation by removing the
state-imposed impediment on Democratic voters.
Defendants nevertheless contend that the plaintiffs lack standing for several
reasons. First, they assert that “[a] majority of Justices in Vieth properly recognized that
a statewide challenge to a redistricting plan was not justiciable.” 410 This view, however,
is not the equivalent of holding that the plaintiffs did not have standing to pursue their
cause of action. Standing is just one aspect of justiciability, which also includes ripeness,
mootness, and the political question doctrine. See Allen v. Wright, 468 U.S. 737, 750
(1984), abrogated on other grounds by Lexmark Int’l Inc. v. Static Control Components, 134 S.
Ct. 1377, 1388 (2014); Charles Alan Wright et al., Federal Practice & Procedure § 3529 (3d
ed. 2008). The Vieth plurality held that that the plaintiffs’ claim for political
gerrymandering presented a nonjusticiable political question, Vieth, 541 U.S. 277–81;
only one Justice opined that the plaintiffs lacked standing to bring a statewide political
gerrymandering claim, id. at 328 (Stevens, J., dissenting). 411
409
See supra at 104–07.
410
R.39 at 7 (emphasis added).
In Vieth, Justice Souter, joined by Justice Ginsburg, affirmed his belief “that political gerrymandering is
a justiciable issue,” but explained that he would “otherwise start anew” in fashioning a test. 541 U.S. at
346 (Souter, J., dissenting). He stated that it was his “own judgment … that we would have better luck at
411
114
The defendants also claim that in recognizing the plaintiffs’ standing to challenge
a statewide map, we are at odds with the Court’s holding in United States v. Hays, 515
U.S. 737 (1995). Hays, like its predecessor, Shaw v. Reno, 509 U.S. 630 (1993), involved
allegations that a districting map constituted “an effort to segregate voters into separate
voting districts because of their race.” Hays, 575 U.S. at 738 (internal quotation marks
omitted). This particular cause of action is limited, therefore, to individuals who
“reside[] in a racially gerrymandered district” because they are the ones who “ha[ve]
been denied equal treatment because of the legislature’s reliance on racial criteria.” Id.
at 745. 412
The rationale and holding of Hays have no application here. As we already have
discussed, 413 the harm in such cases is not that the racial group’s voting strength has
been diluted, but that race has been used “as a basis for separating voters into districts.”
The district lines, therefore, “embody stereotypes that treat individuals as the product
of their race, evaluating their thoughts and efforts—their very worth as citizens—
according to a criterion barred to the Government by history and the Constitution.”
Miller v. Johnson, 515 U.S. 900, 912 (1995) (internal quotation marks omitted) (quoting
Metro Broadcasting, Inc. v. FCC, 497 U.S. 547, 604 (1990) (O’Connor, J., dissenting)). The
concern here is a very different one: it is the effect of a statewide districting map on the
ability of Democrats to translate their votes into seats. The harm is the result of the
entire map, not simply the configuration of a particular district. It follows, therefore,
that an individual Democrat has standing to assert a challenge to the statewide map.
The defendants also argue that the wrong alleged by the plaintiffs is not
sufficiently “particularized” to satisfy the standing requirement. According to the
defendants, “the plaintiffs are asserting an injury that is not personal to any one of
them, but instead is common to anyone who supports the Democratic Party.” 414 We
cannot take the defendants’ arguments at face value. If, for instance, Congress should
pass a law that imposed income taxes only on Democrats, surely an individual
Democrat could bring a constitutional challenge to the law even though the harm was
devising a workable prima facie case if we concentrated as much as possible on suspect characteristics of
individual districts instead of statewide patterns.” Id. He did not foreclose the concept of a challenge to a
statewide map, nor did he discuss specifically whether an individual voter would have standing to
challenge a statewide map.
The Court in Hays did not foreclose the possibility that a plaintiff living outside a racially
gerrymandered district could present evidence that he “ha[d] personally been subjected to a racial
classification” in the drawing of district lines. 515 U.S. at 745. Under such circumstances, that individual
as well would have standing to pursue a Shaw cause of action. Id.
412
413
See supra note 171.
414
R.39 at 5.
115
shared by so many. Moreover, an injury is not sufficiently particularized only if it is a
wrong shared by the “public at large”:
We have consistently held that a plaintiff raising only a generally available
grievance about government—claiming only harm to his and every
citizen’s interest in proper application of the Constitution and laws, and
seeking relief that no more directly and tangibly benefits him than it does
the public at large—does not state an Article III case or controversy.
Lujan, 504 U.S. at 573–74. The harm that the plaintiffs have experienced is not one
shared by the public at large. It is one shared by Democratic voters in the State of
Wisconsin. The dilution of their votes is both personal and acute. Consequently, the
plaintiffs have satisfied the requirement of a particularized injury.
The defendants finally maintain that that “[t]here is no reliable causal connection
between re-doing statewide districts and what the Plaintiffs themselves are involved in,
namely localized elections.” 415 We believe that this claim is belied by the evidence. The
plaintiffs have established that, given Wisconsin’s caucus system, the efficacy of their
vote in securing a political voice depends on the efficacy of the votes of Democrats
statewide. Moreover, the drafters themselves drew maps that would have resulted in
significantly greater partisan balance than that obtained by Act 43. In short, there is no
question that Act 43 imposed a disability on Democratic voters and that redrawing a
district map—indeed, perhaps employing one of the drafters’ earlier efforts—would
remove that disability.
VII
ORDER
A.
Remedy
In their complaint, the plaintiffs request three types of relief: (1) that we declare
the Assembly districts established by Act 43 unconstitutional; (2) that, “[i]n the absence
of a state law establishing a constitutional district plan for the Assembly districts,
adopted by the Legislature and signed by the Governor in a timely fashion, [we]
establish a redistricting plan that meets the requirements of the U.S. Constitution and
federal statutes …”; and (3) that we enjoin the defendants from “administering,
preparing for, and in any way permitting the nomination or election of members of the
State Assembly from the unconstitutional districts that now exist.” 416
415
R.39 at 3.
416
R.1 at 29, ¶¶ 97–99.
116
We defer, at this time, a ruling on the appropriate remedy. The parties have not
had an opportunity to brief fully the timing and propriety of remedial measures. We
therefore order briefing on the appropriate remedy according to the following schedule:
1.
The parties shall file simultaneous briefs on the nature and timing of all
appropriate remedial measures in 30 days’ time;
2.
Simultaneous response briefs are due 15 days thereafter.
The parties will provide the court with all evidentiary and legal support they
believe is required for the court to make its ruling. If the parties do not believe that the
court can rule on the appropriate remedy without the benefit of additional testimony,
they should inform the court of the nature and extent of the testimony they believe is
required.
B.
Evidentiary Matters
For the reasons set forth in this opinion, the motions set forth in our docket
numbers 151 (with respect to the admission of exhibits 98–100, 102, 118–119, 131, 141,
148, 150–152, 333, 391, 394, 405–406, 408, 414–415, 417, and 498) and 154 are DENIED.
The motions set forth in our docket numbers 152 and 158 are GRANTED.
IT IS SO ORDERED.
Entered this 21st day of November, 2016.
BY THE COURT:
/s/
KENNETH F. RIPPLE
Circuit Judge
/s/
BARBARA B. CRABB
District Judge
117
Appendix 1 – Partial “S” Curve for the Team Map
District
16
9
10
8
76
17
75
77
18
15
74
46
47
64
78
7
45
11
43
44
19
65
90
79
48
71
94
72
73
91
66
42
89
93
14
63
80
70
6
69
95
49
92
88
41
12
68
20
67
85
50
5
55
54
28
62
1
29
84
86
87
53
27
34
3
2
52
25
4
35
51
26
40
21
33
30
81
13
61
36
56
39
83
24
32
23
31
37
38
22
82
60
57
96
59
99
97
58
98
Observed
Index_46
Index_47
Index_48
Index_49
Index_50
Index_51
Index_52
Index_53
Index_54
0.1555
0.1245
0.1345
0.1445
0.1545
0.1645
0.1745
0.1845
0.1945
0.2045
0.1844
0.1534
0.1634
0.1734
0.1834
0.1934
0.2034
0.2134
0.2234
0.2334
0.1963
0.1653
0.1753
0.1853
0.1953
0.2053
0.2153
0.2253
0.2353
0.2453
0.2068
0.1758
0.1858
0.1958
0.2058
0.2158
0.2258
0.2358
0.2458
0.2558
0.2098
0.1788
0.1888
0.1988
0.2088
0.2188
0.2288
0.2388
0.2488
0.2588
0.2183
0.1873
0.1973
0.2073
0.2173
0.2273
0.2373
0.2473
0.2573
0.2673
0.2572
0.2262
0.2362
0.2462
0.2562
0.2662
0.2762
0.2862
0.2962
0.3062
0.2984
0.2674
0.2774
0.2874
0.2974
0.3074
0.3174
0.3274
0.3374
0.3474
0.2995
0.2685
0.2785
0.2885
0.2985
0.3085
0.3185
0.3285
0.3385
0.3485
0.3141
0.2831
0.2931
0.3031
0.3131
0.3231
0.3331
0.3431
0.3531
0.3631
0.3368
0.3058
0.3158
0.3258
0.3358
0.3458
0.3558
0.3658
0.3758
0.3858
0.3474
0.3164
0.3264
0.3364
0.3464
0.3564
0.3664
0.3764
0.3864
0.3964
0.3562
0.3252
0.3352
0.3452
0.3552
0.3652
0.3752
0.3852
0.3952
0.4052
0.3697
0.3387
0.3487
0.3587
0.3687
0.3787
0.3887
0.3987
0.4087
0.4187
0.3725
0.3415
0.3515
0.3615
0.3715
0.3815
0.3915
0.4015
0.4115
0.4215
0.3812
0.3502
0.3602
0.3702
0.3802
0.3902
0.4002
0.4102
0.4202
0.4302
0.3852
0.3542
0.3642
0.3742
0.3842
0.3942
0.4042
0.4142
0.4242
0.4342
0.3895
0.3585
0.3685
0.3785
0.3885
0.3985
0.4085
0.4185
0.4285
0.4385
0.3937
0.3627
0.3727
0.3827
0.3927
0.4027
0.4127
0.4227
0.4327
0.4427
0.4047
0.3737
0.3837
0.3937
0.4037
0.4137
0.4237
0.4337
0.4437
0.4537
0.4053
0.3743
0.3843
0.3943
0.4043
0.4143
0.4243
0.4343
0.4443
0.4543
0.4063
0.3753
0.3853
0.3953
0.4053
0.4153
0.4253
0.4353
0.4453
0.4553
0.4138
0.3828
0.3928
0.4028
0.4128
0.4228
0.4328
0.4428
0.4528
0.4628
0.4156
0.3846
0.3946
0.4046
0.4146
0.4246
0.4346
0.4446
0.4546
0.4646
0.4158
0.3848
0.3948
0.4048
0.4148
0.4248
0.4348
0.4448
0.4548
0.4648
0.4412
0.4102
0.4202
0.4302
0.4402
0.4502
0.4602
0.4702
0.4802
0.4902
0.4463
0.4153
0.4253
0.4353
0.4453
0.4553
0.4653
0.4753
0.4853
0.4953
0.4483
0.4173
0.4273
0.4373
0.4473
0.4573
0.4673
0.4773
0.4873
0.4973
0.4488
0.4178
0.4278
0.4378
0.4478
0.4578
0.4678
0.4778
0.4878
0.4978
0.4494
0.4184
0.4284
0.4384
0.4484
0.4584
0.4684
0.4784
0.4884
0.4984
0.4503
0.4193
0.4293
0.4393
0.4493
0.4593
0.4693
0.4793
0.4893
0.4993
0.4504
0.4194
0.4294
0.4394
0.4494
0.4594
0.4694
0.4794
0.4894
0.4994
0.4543
0.4233
0.4333
0.4433
0.4533
0.4633
0.4733
0.4833
0.4933
0.5033
0.4624
0.4314
0.4414
0.4514
0.4614
0.4714
0.4814
0.4914
0.5014
0.5114
0.4649
0.4339
0.4439
0.4539
0.4639
0.4739
0.4839
0.4939
0.5039
0.5139
0.4721
0.4411
0.4511
0.4611
0.4711
0.4811
0.4911
0.5011
0.5111
0.5211
0.4737
0.4427
0.4527
0.4627
0.4727
0.4827
0.4927
0.5027
0.5127
0.5227
0.4751
0.4441
0.4541
0.4641
0.4741
0.4841
0.4941
0.5041
0.5141
0.5241
0.4788
0.4478
0.4578
0.4678
0.4778
0.4878
0.4978
0.5078
0.5178
0.5278
0.4819
0.4509
0.4609
0.4709
0.4809
0.4909
0.5009
0.5109
0.5209
0.5309
0.4854
0.4544
0.4644
0.4744
0.4844
0.4944
0.5044
0.5144
0.5244
0.5344
0.4923
0.4613
0.4713
0.4813
0.4913
0.5013
0.5113
0.5213
0.5313
0.5413
0.4926
0.4616
0.4716
0.4816
0.4916
0.5016
0.5116
0.5216
0.5316
0.5416
0.5055
0.4745
0.4845
0.4945
0.5045
0.5145
0.5245
0.5345
0.5445
0.5545
0.5056
0.4746
0.4846
0.4946
0.5046
0.5146
0.5246
0.5346
0.5446
0.5546
0.5122
0.4812
0.4912
0.5012
0.5112
0.5212
0.5312
0.5412
0.5512
0.5612
0.5143
0.4833
0.4933
0.5033
0.5133
0.5233
0.5333
0.5433
0.5533
0.5633
0.5162
0.4852
0.4952
0.5052
0.5152
0.5252
0.5352
0.5452
0.5552
0.5652
0.5175
0.4865
0.4965
0.5065
0.5165
0.5265
0.5365
0.5465
0.5565
0.5665
0.5247
0.4937
0.5037
0.5137
0.5237
0.5337
0.5437
0.5537
0.5637
0.5737
0.5253
0.4943
0.5043
0.5143
0.5243
0.5343
0.5443
0.5543
0.5643
0.5743
0.5256
0.4946
0.5046
0.5146
0.5246
0.5346
0.5446
0.5546
0.5646
0.5746
0.5302
0.4992
0.5092
0.5192
0.5292
0.5392
0.5492
0.5592
0.5692
0.5792
0.531
0.5
0.51
0.52
0.53
0.54
0.55
0.56
0.57
0.58
0.5332
0.5022
0.5122
0.5222
0.5322
0.5422
0.5522
0.5622
0.5722
0.5822
0.5372
0.5062
0.5162
0.5262
0.5362
0.5462
0.5562
0.5662
0.5762
0.5862
0.5373
0.5063
0.5163
0.5263
0.5363
0.5463
0.5563
0.5663
0.5763
0.5863
0.5382
0.5072
0.5172
0.5272
0.5372
0.5472
0.5572
0.5672
0.5772
0.5872
0.5384
0.5074
0.5174
0.5274
0.5374
0.5474
0.5574
0.5674
0.5774
0.5874
0.5402
0.5092
0.5192
0.5292
0.5392
0.5492
0.5592
0.5692
0.5792
0.5892
0.5422
0.5112
0.5212
0.5312
0.5412
0.5512
0.5612
0.5712
0.5812
0.5912
0.5424
0.5114
0.5214
0.5314
0.5414
0.5514
0.5614
0.5714
0.5814
0.5914
0.5426
0.5116
0.5216
0.5316
0.5416
0.5516
0.5616
0.5716
0.5816
0.5916
0.5439
0.5129
0.5229
0.5329
0.5429
0.5529
0.5629
0.5729
0.5829
0.5929
0.544
0.513
0.523
0.533
0.543
0.553
0.563
0.573
0.583
0.593
0.5458
0.5148
0.5248
0.5348
0.5448
0.5548
0.5648
0.5748
0.5848
0.5948
0.5501
0.5191
0.5291
0.5391
0.5491
0.5591
0.5691
0.5791
0.5891
0.5991
0.5519
0.5209
0.5309
0.5409
0.5509
0.5609
0.5709
0.5809
0.5909
0.6009
0.5523
0.5213
0.5313
0.5413
0.5513
0.5613
0.5713
0.5813
0.5913
0.6013
0.5528
0.5218
0.5318
0.5418
0.5518
0.5618
0.5718
0.5818
0.5918
0.6018
0.5572
0.5262
0.5362
0.5462
0.5562
0.5662
0.5762
0.5862
0.5962
0.6062
0.5576
0.5266
0.5366
0.5466
0.5566
0.5666
0.5766
0.5866
0.5966
0.6066
0.5602
0.5292
0.5392
0.5492
0.5592
0.5692
0.5792
0.5892
0.5992
0.6092
0.5663
0.5353
0.5453
0.5553
0.5653
0.5753
0.5853
0.5953
0.6053
0.6153
0.5664
0.5354
0.5454
0.5554
0.5654
0.5754
0.5854
0.5954
0.6054
0.6154
0.5731
0.5421
0.5521
0.5621
0.5721
0.5821
0.5921
0.6021
0.6121
0.6221
0.5752
0.5442
0.5542
0.5642
0.5742
0.5842
0.5942
0.6042
0.6142
0.6242
0.5772
0.5462
0.5562
0.5662
0.5762
0.5862
0.5962
0.6062
0.6162
0.6262
0.5777
0.5467
0.5567
0.5667
0.5767
0.5867
0.5967
0.6067
0.6167
0.6267
0.5791
0.5481
0.5581
0.5681
0.5781
0.5881
0.5981
0.6081
0.6181
0.6281
0.5797
0.5487
0.5587
0.5687
0.5787
0.5887
0.5987
0.6087
0.6187
0.6287
0.5839
0.5529
0.5629
0.5729
0.5829
0.5929
0.6029
0.6129
0.6229
0.6329
0.587
0.556
0.566
0.576
0.586
0.596
0.606
0.616
0.626
0.636
0.5924
0.5614
0.5714
0.5814
0.5914
0.6014
0.6114
0.6214
0.6314
0.6414
0.5934
0.5624
0.5724
0.5824
0.5924
0.6024
0.6124
0.6224
0.6324
0.6424
0.5996
0.5686
0.5786
0.5886
0.5986
0.6086
0.6186
0.6286
0.6386
0.6486
0.6014
0.5704
0.5804
0.5904
0.6004
0.6104
0.6204
0.6304
0.6404
0.6504
0.6017
0.5707
0.5807
0.5907
0.6007
0.6107
0.6207
0.6307
0.6407
0.6507
0.6018
0.5708
0.5808
0.5908
0.6008
0.6108
0.6208
0.6308
0.6408
0.6508
0.6088
0.5778
0.5878
0.5978
0.6078
0.6178
0.6278
0.6378
0.6478
0.6578
0.6095
0.5785
0.5885
0.5985
0.6085
0.6185
0.6285
0.6385
0.6485
0.6585
0.6148
0.5838
0.5938
0.6038
0.6138
0.6238
0.6338
0.6438
0.6538
0.6638
0.6163
0.5853
0.5953
0.6053
0.6153
0.6253
0.6353
0.6453
0.6553
0.6653
0.6275
0.5965
0.6065
0.6165
0.6265
0.6365
0.6465
0.6565
0.6665
0.6765
0.6561
0.6251
0.6351
0.6451
0.6551
0.6651
0.6751
0.6851
0.6951
0.7051
0.6702
0.6392
0.6492
0.6592
0.6692
0.6792
0.6892
0.6992
0.7092
0.7192
0.686
0.655
0.665
0.675
0.685
0.695
0.705
0.715
0.725
0.735
0.6946
0.6636
0.6736
0.6836
0.6936
0.7036
0.7136
0.7236
0.7336
0.7436
0.7109
0.6799
0.6899
0.6999
0.7099
0.7199
0.7299
0.7399
0.7499
0.7599
118
Appendix 2
119
GRIESBACH, District Judge, dissenting. Through a secretive and one-sided
process, the state Republican leaders who controlled the legislature used the latest
computer software and political consultants to draw up legislative district maps with
the unashamedly partisan goal of winning as many seats as possible. The maps they
drew gave short shrift to traditional districting principles, often producing districts with
unusual and suspicious shapes. The governor, also a Republican, quickly signed the
act. The Republicans’ efforts were rewarded when, in the very next election, they won
more than twenty percent more seats in the legislature than their statewide vote totals
would have suggested.
The state in question is Indiana, not Wisconsin. The procedure used to draw the
map in Indiana is identical to what led up to the enactment of Wisconsin’s Act 43: in
short, it was crafted in secret by Republicans who, at least in Indiana, conceded that
naked political gain was their overwhelming purpose. It also allowed the Republicans
to win far more seats than their statewide vote totals would warrant—in the Plaintiffs’
parlance, a historically-high efficiency gap of eleven percent. Thirty years ago,
however, the Supreme Court upheld the districts drawn by Indiana Republicans, with a
plurality of the Court concluding that the Democrats had not shown they were
sufficiently injured. 1 Despite these similarities, and despite the Court’s clear reluctance
to intervene in what are essentially political cases, the Plaintiffs ask this court to find
that Wisconsin’s Act 43 is an unconstitutional partisan gerrymander.
In fact, Wisconsin’s Act 43 differs from Indiana’s upheld plan in one key fashion:
unlike Indiana’s plan, Act 43 pays heed to all of the principles that have traditionally
governed the districting process, such as contiguity, compactness and respect for
political subdivisions like counties and cities. And unlike Indiana’s plan, there is no
allegation that the Republicans drew any of the many kinds of unusually-shaped
districts that are traditionally seen in gerrymandering cases. (The term “gerrymander”
arises from a district shaped like a salamander that was drawn during the term of
Massachusetts Governor Elbridge Gerry.) Thus, although Wisconsin’s plan, like
Indiana’s, was politically motivated, but unlike Indiana’s, complies with traditional
redistricting principles, and though it has the same partisan impact as the plan upheld
in Bandemer, the Plaintiffs nevertheless ask the court to intervene, claiming to have
Davis v. Bandemer, 478 U.S. 109, 134 (1986) (plurality opinion). In Indiana, Democratic candidates
received 51.9% of the vote. Only 43 Democrats, however, were elected to the 100-member House. Under
the Plaintiffs’ efficiency gap calculations, winning 52% of the vote would entitle a party to receive 54% of
the seats. Since they only received 43 seats, that results in an efficiency gap of around 11%. The
Republicans won 57 seats despite winning only 48% of the statewide vote. The Plaintiffs’ proposed norm
would have the Republicans winning only 46 seats. By winning 57, the Republicans won almost 24%
more seats than their statewide totals would suggest. This very nearly mirrors the results of Wisconsin’s
election in 2012, the first election conducted under Act 43.
1
120
discovered the long sought-after “judicially discernable and manageable standard[]” for
deciding such cases that a majority of the Court has thought might exist. 478 U.S. at
123; Vieth v. Jubelirer, 541 U.S. 267, 278–79 (2004) (plurality opinion).
The Plaintiffs have made that standard—the efficiency gap—the center piece of
their case and asked this court to adopt it as a matter of constitutional law. (ECF No. 1
at ¶¶ 5, 44–53.) Despite the central role the efficiency gap has played in the case from
the beginning, however, the majority has declined the Plaintiffs’ invitation to adopt
their standard and uses it only as confirming evidence of a constitutional violation it
has found based on its own newly created test: whether the State’s redistricting plan
had the intent and effect of entrenching the Republican party in power over the life of
the plan. For the following reasons, I part ways with my colleagues.
First, I am unable to accept proof of intent to act for political purposes as a
significant part of any test for whether a task constitutionally entrusted to the political
branches of government is unconstitutional. If political motivation is improper, then
the task of redistricting should be constitutionally assigned to some other body, a
change in law we lack any authority to effect. Second, to the extent the majority’s
“intent to entrench themselves in power” standard is intended to mean what those
Justices who have used that language in previous cases intended, I am not convinced
that the plaintiffs have met this standard. Third, of the small majority of Justices who
would even entertain political gerrymandering cases, several of them would require
plaintiffs to establish that the challenged plan failed to follow traditional principles of
redistricting. Because the Plaintiffs do not even attempt to argue that Act 43 violates
traditional redistricting principles, I would enter judgment in favor of the Defendants
on that basis alone. Fourth, it is very likely that the Republicans would have won
control of the legislature in 2012 and 2014 even without the alleged gerrymandering, and so
this case presents a poor vehicle for the remedying of any grave injustice.
In addition, the efficiency gap concept that the Plaintiffs have offered as the
“judicially discernable and manageable standard[ ] by which political gerrymander
cases are to be decided,” Bandemer, 478 U.S. at 123, appears to have substantial
theoretical and practical limitations that render it unsuitable for the task at hand. First,
the efficiency gap—or any measure that simply compares statewide votes to seats—is
little more than an enshrinement of a phantom constitutional right, namely, the idea
that voters for one party are entitled to some given level of representation proportional
to how many votes that party’s candidates win in every assembly district throughout
the state as a whole. Second, the efficiency gap simply measures each party’s ability to
win more assembly seats, but winning more assembly seats does not usually translate
into any significant additional power, and thus does not cause material political
injury—unless of course it is the seat that turns over control of the legislature to the
121
gerrymandering party. Third, the efficiency gap essentially begs the ultimate question
of whether a partisan gerrymander occurred, and it fails to capture the essence of what
it means to vote since it presupposes that voters are voting for a statewide party rather
than simply for an individual candidate.
In addition to these theoretical problems, the efficiency gap suffers from practical
issues as well. First, the Plaintiff’s efficiency gap calculation, which is based on tallying
“wasted” votes, appears to ignore a large number of wasted votes attributable to
winning candidates, thereby undermining its reliability as a tool for measuring even
what it purports to capture. Second, the test Plaintiffs propose does not adequately
account for Wisconsin’s political geography, which naturally “packs” large numbers of
Democrats into urban areas like Madison and Milwaukee, resulting in hundreds of
thousands of “wasted” votes in inevitable landslide Democratic victories for assembly
candidates. Finally, the efficiency gap is highly volatile and could easily trigger judicial
intervention when no intervention is warranted. For all of these additional reasons, I
would enter judgment in favor of the Defendants.
I. Partisan Intent and Effect
I begin with a point upon which I agree with my colleagues. It is almost beyond
question that the Republican staff members who drew the Act 43 maps intended to
benefit Republican candidates. They accumulated substantial historical knowledge
about the political tendencies of every part of the state and consulted with Dr. Ronald
Gaddie to confirm their predictions about voting patterns. Though they denied the
suggestion that such information was used to project future voting tendencies, my
colleagues rightly conclude that when political staffers compile historical voting
information about potential districts, their claim that they did not intend to use that
information to predict future voting patterns is hardly worthy of belief. After all, these
individuals are not operating under even the pretense that they are nonpartisan: they
are employed by Republicans in leadership and draft district maps at their direction.
That they would resort to partisan considerations in drawing the maps is therefore
anything but surprising.
This alone does not make it wrong, however. The majority cites Rogers v. Lodge,
458 U.S. 613, 617 (1982), for the proposition that “equal protection challenges to
redistricting plans require a showing of discriminatory purpose or intent.” But Rogers is
a race discrimination case challenging an electoral system on the ground that it was
intended to dilute voting strength of the black population. The intent to weaken a racial
group’s political power in drawing district lines is always and everywhere wrongful.
The same is not true for political motivations. The Supreme Court has long
acknowledged partisan considerations are inevitable when partisan politicians draw
122
maps. “The Constitution clearly contemplates districting by political entities, see
Article I, § 4, and unsurprisingly that turns out to be root-and-branch a matter of
politics.” Vieth, 541 U.S. at 285 (plurality opinion) (citing Miller v. Johnson, 515 U.S. 900,
914 (1995) (“[R]edistricting in most cases will implicate a political calculus in which
various interests compete for recognition . . . .”); Shaw v. Reno, 509 U.S. 630, 662 (1993)
(White, J., dissenting) (“[D]istricting inevitably is the expression of interest group
politics . . . .”); Gaffney v. Cummings, 412 U.S. 735, 753 (1973) (“The reality is that
districting inevitably has and is intended to have substantial political consequences.”)).
In other words, so long as it is deemed acceptable for politicians to draw district
maps—and it is—we cannot pretend to be shocked that legislators so engaged will act
like the politicians they are. As Justice Stevens put it, “Legislators are, after all,
politicians.” Karcher v. Daggett, 462 U.S. 725, 753 (1983) (Stevens, J., concurring).
“[S]ome intent to gain political advantage is inescapable whenever political bodies
devise a district plan, and some effect results from the intent.” Vieth, 541 U.S. at 344
(Souter, J., dissenting). “That courts can grant relief in districting cases where race is
involved does not answer our need for fairness principles here. Those controversies
implicate a different inquiry. They involve sorting permissible classifications in the
redistricting context from impermissible ones. Race is an impermissible classification. . .
. Politics is quite a different matter.” Id. at 307 (Kennedy, J., concurring). 2
The majority opinion wrestles with the “how much intent is too much” question,
a question that has bedeviled the courts for decades and caused several members of the
Supreme Court to give up on finding an answer. But whose intent are we talking about
and how does one go about measuring it? The Republican leadership clearly wanted a
plan that would give them a majority of seats, but some of their members had to be
talked into accepting less safe districts—the so-called donors—in the hope that they
could still win their seat and the party would win a majority of seats as well. They
more or less acquiesced. The more difficult question is how do you measure intent? A
person either intends a result or he does not. Making gradations of intent a standard is
a recipe for interminable litigation. Vieth, 541 U.S. at 286 (plurality opinion)
Notwithstanding the acknowledgement by almost every Justice to address the issue that partisan intent
is to be expected in redistricting, the majority, citing Harris v. Arizona Independent Redistricting Commission,
136 S. Ct. 1301 (2016), suggests that it is an open question whether partisanship is an illegitimate
redistricting factor. But the issue in Harris was whether a deviation of less than 10% from the equal
population requirement of Reynolds v Sims, 377 U.S. 533 (1964), could be justified by partisan
considerations. The Court stated, “[E]ven assuming, without deciding, that partisanship is an illegitimate
redistricting factor, appellants have not carried their burden.” Harris, 136 S. Ct. at 1310. Taken in context
and in light of the Court’s repeated acknowledgement that partisan considerations are to be expected, I
read Harris as leaving open the question whether partisan intent could legitimately justify an
underpopulated district; not whether it is illegitimate in itself. See Cox v. Larios, 542 U.S. 947 (2004). Here,
there is no allegation that any district was underpopulated.
2
123
(“Moreover, the fact that partisan districting is a lawful and common practice means
that there is almost always room for an election-impeding lawsuit contending that
partisan advantage was the predominant motivation; not so for claims of racial
gerrymandering.”).
My colleagues attempt to limit the potential for unending litigation such an
intent element might encourage by holding that the level of partisanship may be
deemed “too much” when the map-drawers intend to entrench their party in power for
the life of the plan and achieves that effect. Slip Op. at 58, 71. Adding the qualification
that the intent and effect be to entrench the party in power for the life of the plan,
however, does not help. How is that intent different from intending to benefit the
party? We are talking about redistricting plans, after all, not a bill to name the State
mascot. Redistricting plans, by their very nature, affect future elections for the life of
the plan. And what does “entrench their party in power” mean in this context?
The plurality in Bandemer sought to limit court intervention to cases where “a
particular group has been unconstitutionally denied its chance to effectively influence
the political process.” 478 U.S. at 132–33 (plurality opinion). On the statewide level, the
plurality said, “such finding of unconstitutionality must be supported by evidence of
continued frustration of the will of a majority of voters or effective denial to a minority
of voters of a fair chance to influence the political process.” Id. at 133. The standard
adopted by the majority in this case is equally opaque, but less demanding. Plaintiffs
have challenged the redistricting plan for Assembly seats, but the Assembly, by itself,
can do little more than hold things up. Every four years, Wisconsin voters elect a
governor. If plaintiff’s party is able to convince a majority of Wisconsin voters that their
policies are better for the State, nothing the Republicans have done will prevent them
from winning the governor’s office and not only stopping the Republicans from
enacting their agenda at that point, but also denying them control over the next
redistricting process. See Vieth, 541 U.S. at 362 (Breyer, J., dissenting) (“Where a State
has improperly gerrymandered legislative or congressional districts to the majority's
disadvantage, the majority should be able to elect officials in statewide races—
particularly the Governor—who may help to undo the harm that districting has caused
the majority’s party, in the next round of districting if not sooner.”).
Indeed, nothing will prevent a candidate from Plaintiffs’ party from convincing
the voters in a district Republican staff members drew, believing it would elect a
Republican candidate, from electing a Democrat instead. The assumption underlying
Plaintiffs’ entire case is that party affiliation is a readily discernable characteristic in
voters and that it matters above all else in an election. Voters are placed either in one
124
party or the other based on their last vote.3 But party affiliation is not set in stone or in a
voter’s genes:
[A] person’s politics is rarely as readily discernible—and never as
permanently discernible—as a person's race. Political affiliation is not an
immutable characteristic, but may shift from one election to the next; and
even within a given election, not all voters follow the party line. We dare
say (and hope) that the political party which puts forward an utterly
incompetent candidate will lose even in its registration stronghold.
Vieth, 541 U.S. at 287 (plurality opinion) (citing Bandemer, 478 U.S. at 156 (O’Connor, J.,
concurring in judgment)). True, many voters, perhaps most, vote for the brand; but
many make their decision based on the person and his or her position on the issues that
matter most to them at the time. Moreover, candidates for state offices run on different
issues than candidates for national offices, which presumably explains the difference in
voter turn-out and results in the recent Wisconsin presidential and gubernatorial
elections. For all the confidence political experts may have in their predictions of future
election results, Vieth itself stands as a stark reminder that they can be wrong. The
plaintiffs in that case alleged that the Pennsylvania congressional plan was “rigged to
guarantee that thirteen of Pennsylvania’s nineteen congressional representatives will be
Republican.” Vieth v. Pennsylvania, 188 F. Supp. 2d 532, 546 (M.D. Pa. 2002). Yet, as
Professor Nicholas Goedert testified and pointed out in his report, Democrats won a
majority of Pennsylvania’s congressional seats in the two elections following the
Supreme Court’s 2004 decision, including twelve of nineteen in 2008. (ECF Nos. 50-1 at
13; 150 at 150:8–18.)
There are additional problems with the majority’s proposed standard. To the
extent the term has been used by members of the Supreme Court, “entrenchment” has
often referred to a minority party rigging the system so much that it could win a
majority of seats even while consistently garnering only a minority of the statewide
vote. For example, Justice Kennedy has noted that a plan “that entrenches an electoral
minority” is more likely to be a vehicle for partisan discrimination. League of United
Latin Am. Citizens v. Perry, 548 U.S. 399, 419 (2006) [hereinafter LULAC]. Justice Breyer’s
Vieth opinion is more explicit: he explains that “[b]y entrenchment I mean a situation in
which a party that enjoys only minority support among the populace has nonetheless
contrived to take, and hold, legislative power.” 541 U.S. at 360 (Breyer, J., dissenting).
In Justice Breyer’s view, “gerrymandering that leads to entrenchment amounts to an
abuse that violates the Constitution’s Equal Protection Clause.” Id. at 362. These
Plaintiffs offered no evidence as to actual party membership in Wisconsin. Because of its open primary
system, voters in Wisconsin are not required to join a party in order to vote in that party’s primary
election. Democratic Party of U.S. v. Wisconsin, 450 U.S. 107, 110–11 (1981).
3
125
Justices’ concerns about entrenchment thus appear to be focused on the problem
whereby a majority of voters in a state are consistently deprived the opportunity to
control a branch of government. In our case, however, the Republican Party is not a
minority party in Wisconsin. In statewide elections, the state has elected a Republican
governor in the last two general elections (plus a recall election, in 2012). In 2010 GOP
members of the assembly received 53.5% of the statewide popular vote, while they
obtained 52% of the vote in 2014. (ECF No. 125 at ¶¶ 286, 290.) Thus, in this case we
are not dealing with a minority party entrenching itself in power, which means the
majority of the citizens of Wisconsin are not consistently deprived of the right to control
the legislature.
The notion that Republicans took drastic steps to entrench themselves in power
in this sense is also undermined by recent history. When mapmakers sit down to
redraw district maps, it is not as though they are drawing on a blank slate—the 99
districts then in existence will necessarily play a role in how the new districts will look.
The majority opinion glosses over the fact that Republicans enjoyed very healthy
efficiency gaps during the previous decade, despite the fact that the district maps then
in effect were produced through plans created by federal courts, not a partisan
legislature. As the Plaintiffs’ expert Simon Jackman concluded, the plan in effect during
the previous decennial period favored Republicans with an average 7.6% efficiency gap,
including a gap as large as 11.8% in 2006. (ECF No. 125 at ¶¶ 190, 192, 194, 242.) When
one considers that the pre-existing maps were already quite favorable to Republicans, it
is hardly surprising that the maps they ultimately created increased their advantage
somewhat.
In fact, under the Plaintiffs’ proposed test the Republicans were obligated not
only to draw fairer maps, but to engage in heroic levels of nonpartisan statesmanship.
The Plaintiffs are evidently of the view that the Republicans, having achieved the oncein-a-lifetime feat of controlling both branches of the legislature and the governorship
during a redistricting year, should have used that unique opportunity not for selfadvantage but instead to draw a map that was less favorable to them than even the
court-drawn plan that governed the previous decade. Ironically, even if the
Republicans had said to themselves, “let’s stick with a plan like the one drawn by the
federal courts—it helps us enough already,” the Plaintiffs would still take umbrage at
the resulting map and call it an impermissible partisan gerrymander, assuming the
efficiency gaps continued to follow the pattern of the previous decade. Any test that
requires heroic levels of nonpartisanship does not square with the courts’ recognition of
the reality that legislators tasked with drawing maps will always seek to advantage
their own party. Under these circumstances, it is difficult to credit the Plaintiffs’
assertions that the Republicans exhibited “too much” partisanship when they drew a
126
map that was only somewhat more favorable to the GOP than maps drawn by a federal
court the previous decade.
The Republicans’ control of the districting process appears to have been little
different than the Republicans’ conduct in Davis v. Bandemer, 478 U.S. 109 (1986). There,
the district court described the Republican-controlled process as “contrived,” after the
Republicans enacted dummy bills and named Democratic “advisors” who in actuality
had no input and “no access to the mapmaking process that ensued.” Bandemer v.
Davis, 603 F. Supp. 1479, 1483 (S. D. Ind. 1984). The Republicans spent a quarter million
dollars on a research firm, which used the latest computer equipment, while the
Democrats had no such support. Id. at 1484. One Republican senator admitted that
though the Democrats could offer their own map proposals, they would never be
accepted. Id. This “unashamedly partisan” process resulted in party-line approval of
the plan in both houses of the legislature and the prompt signature by the Republican
governor. Id. And yet the plan drawn in Indiana was upheld, despite a nearly identical
partisan effect as the current plan.
None of this is to extol the process whereby the district maps were drawn, and
neither do I intend to espouse the cynical conclusion that politics must always be onesided and bare-knuckle. Indeed, the very accusation and at least the appearance of
heavy-handed unfairness may itself be made a political issue and lead a significant
number of less committed or independent voters to change their views about which
party they wish to support. By the same token, I believe it is largely true that
individuals who attempt to gain political advantage through map-drawing are not
engaged in foul play or dirty tricks, but are merely using the power the voters have
granted them to enact the policies they favor. They are not intending to “burden the
representational rights of Democratic voters” by “impeding their ability to translate
their votes into legislative seats.” Majority op. at 2. These are legal concepts that do not
translate easily into the world of politics. Imagine a congressman facing President
Johnson’s demand that he vote for the Civil Rights Act of 1964 or lose a key federal
project in his district claiming that his constituents were deprived of their
representational rights. The political process does not operate by the same rules that
govern judges and courts. By and large, whether it is the Democrats or Republicans
doing the gerrymandering, they try to create partisan majorities not to suppress
opposing viewpoints but because they honestly believe they will then be able to enact
the policies that in their view are best for the state, or nation. 4
Notably, although the Democrat-Plaintiffs express outrage at the maps the Republicans drew, the
Democrats are hardly immune to map-drawing chicanery of their own. For example, the plans they
4
127
In sum, partisan intent is not illegal, but is simply the consequence of assigning
the task of redistricting to the political branches of government. 5 The standard
proposed by the majority offers no improvement over the tests that have already been
rejected by the Supreme Court. Moreover, even if I accepted the majority’s standard, I
am unconvinced that Republicans intended to or could entrench themselves in power in
the sense understood by those members of the Court that have addressed it. Given the
fact that Republicans already enjoyed significant advantages under court-drawn
districting plans then in effect, it should hardly surprise anyone that, when afforded the
rare opportunity to draw their own maps, they extended their electoral advantage
somewhat. I am therefore unable to conclude that Act 43's passage was anything other
than the kind of “politics as usual” that courts have routinely either tolerated or
acquiesced in.
II. A Gerrymander without Gerrymandering
Justices Souter and Ginsburg counseled in Vieth that statewide districting
challenges are “a function of claims that individual districts are illegitimately drawn.”
Vieth, 541 U.S. at 347 (Souter, J., dissenting). Therefore, it makes sense to “concentrate[]
as much as possible on suspect characteristics of individual districts instead of
statewide patterns.” Id. Surprisingly, the Plaintiffs in this action did exactly the
opposite. Instead of pointing to specific districts that had been gerrymandered, they
relied on statewide data and calculations, as well as spreadsheets, metadata, graphs and
charts, all without referring to any actual maps or lines drawn by the Defendants. The
Plaintiffs purported to show the “DNA” of gerrymandering in a graph comparing
proposed following the 2000 census reflect the same partisan intent as Act 43 and were “riddled with
their own partisan marks.” As described by the three-judge panel that heard that case:
Leg Dem B and Leg Dem C divide the City of Madison into six districts radiating out from
the Capitol in pizza slice fashion. The Leg Dem plans have higher levels of population
deviation, lower levels of core retention, higher levels of disenfranchisement, and lower
levels of compactness than the Alt A and Alt C plans, in part because they renumber the
Senate districts in Milwaukee County (again for presumed partisan advantage).
Baumgart v. Wendelberger, Nos. 01-C-0121, 02-C-0366, 2002 WL 34127471, at *4 (E.D. Wis. May 30, 2002).
Because Wisconsin government was divided at the time, the Democrats were unable to enact their
proposed plan into law, but there is no reason to believe they would not attempt to do so now if the
circumstances were reversed. Importantly, there is no evidence that Act 43 violated any of the traditional
redistricting principles cited by the Baumgartner court in rejecting the Democratic proposal.
It was only a term ago that the Court held by a 5 to 4 vote that it was constitutionally permissible to
remove redistricting from the political branches. Ariz. State Legislature v. Ariz. Indep. Redistricting Comm’n,
135 S. Ct. 2652 (2015). Adoption of the majority’s standard may well compel States to do so.
5
128
wards to districts, but, like a prosecutor trying to prove a murder without a body, not
once did they actually show any district maps demonstrating the gerrymander they
alleged occurred.
This was not an oversight. The reason for the absence of any discussion of
individual district lines is that Act 43 does not violate any of the redistricting principles
that traditionally govern the districting process, including compactness, contiguity and
respect for political boundaries like counties and cities. In other words, unlike every
other gerrymandering case to come before the courts, the plaintiffs did not argue that
Act 43 created any districts with unusual lines or shapes. Nor were there appreciable
problems with contiguity, compactness, or regard for political boundaries. Act 43’s
districts split more counties than previous plans, but the plan splits fewer
municipalities than the 1990s map. The current plan’s compactness scores are
comparable to previous plans, and there is no indication that any districts had problems
with contiguity. At trial, it was undisputed that the drafting of the current plan placed
the correct number of citizens into each district and also took into account other more
practical (and legitimate) concerns, such as the number of voters who would be
disenfranchised in upcoming senate elections, 6 as well as the residences of the actual
legislators whose district boundaries were changing—factors none of the theoretical
plans considered. In short, although the Plaintiffs argued that their own demonstration
map created similarly compact and contiguous districts with less partisan effect, they
conceded that the districts drawn by Act 43 are sufficiently compact, contiguous and
respectful of political boundaries.
Gerrymandering, as the term’s etymology suggests, has traditionally been
understood as the drawing of unusually-shaped districts in order to achieve a political
advantage. Gerrymandering is “the deliberate and arbitrary distortion of district
boundaries and populations for partisan or personal political purposes.” Bandemer, 478
U.S. at 164 (Powell, J., concurring in part and dissenting in part) (quoting Kirkpatrick v.
Preisler, 394 U.S. 526, 538 (1969) (Fortas, J., concurring)) (emphasis added)); see also Vieth,
541 U.S. at 323 (Stevens, J., dissenting, noting “outlandish district shapes” in traditional
gerrymanders). Without evidence of any distortion of otherwise legitimate district
boundaries, there is no gerrymander, at least as the term is traditionally understood.
The Plaintiffs’ belief that gerrymandering can occur without distortions of
district boundaries is not just a definitional novelty, it flies in the face of Vieth. In Vieth,
four justices found political gerrymandering claims non-justiciable, meaning that they
believed courts should not even get involved in such cases. Of the remaining five
Because senate elections are staggered, the possibility arises that some voters who are moved to a
different district would have no vote for a senate candidate for two consecutive election cycles.
6
129
justices who would consider such claims, three of them (a majority) explicitly would
require a failure to follow traditional redistricting principles as part of any
gerrymandering test. Justice Stevens noted that “an uncouth or bizarre shape can easily
identify a district designed for a single-minded, nonneutral purpose.” Vieth, 541 U.S. at
321 (Stevens, J., dissenting). Citing Justice Powell’s Bandemer opinion, Justice Stevens
noted that “configurations of the districts [and] the observance of political subdivision
lines . . . have independent relevance to the fairness of redistricting.” Id. at 322 (citing
Bandemer, 478 U.S. at 165 (Powell, J., concurring in part and dissenting in part)). Justice
Stevens observed that in Bandemer, Justice Powell had made the irregularity of district
shapes part of his proposed test, remarking on the “strange shape of districts that
conspicuously ignored traditional districting principles.” Id. Any test should “properly
focus[] on whether the boundaries of the voting districts have been distorted
deliberately and arbitrarily to achieve illegitimate ends. . . . Under this definition, the
merits of a gerrymandering claim must be determined by reference to the
configurations of the districts, the observance of political subdivision lines, and other
criteria that have independent relevance to the fairness of redistricting.” Id. Justice
Stevens noted that the Court had used Justice Powell’s test in racial gerrymandering
cases, and he believed it appropriate to do so in a political gerrymandering context as
well. Id.
Citing the Vieth complaint, Justice Stevens observed that one challenged district
“looms like a dragon descending on Philadelphia from the west, splitting up towns and
communities throughout Montgomery and Berks Counties.” Id. at 340. The plan “is so
irregular on its face that it rationally can be viewed only as an effort . . . to advance the
interests of one political party, without regard for traditional redistricting principles
and without any legitimate or compelling justification.” Id. Ultimately, under Justice
Stevens’ proposed test, “if the only possible explanation for a district’s bizarre shape is a
naked desire to increase partisan strength, then no rational basis exists to save the
district from an equal protection challenge. Such a narrow test would cover only a few
meritorious claims, but it would preclude extreme abuses . . . .” Id. at 339.
A “bizarre shape” was also a factor in the test proposed by Justices Souter and
Ginsburg. As part of their proposed analysis, they would require a plaintiff “to show
that the district of his residence . . . paid little or no heed to those traditional districting
principles whose disregard can be shown straightforwardly: contiguity, compactness,
respect for political subdivisions, and conformity with geographic features like rivers
and mountains.” Id. at 348 (Souter, J., dissenting). Because courts are already
experienced at applying these standards, they argued, “a test relying on these standards
would fall within judicial competence.” Id. Thus, of the bare majority of the Court that
would even consider political gerrymandering claims, at least three members of the
130
Vieth court would require a plaintiff to demonstrate that the challenged plan or district
failed to adhere to traditional districting principles.
The Plaintiffs suggest that any test relying on traditional districting principles is
foreclosed by precedent. Strangely, for that premise they rely on the Vieth plurality,
which, it is true, criticized any standard based on district shapes as being difficult to
manage: “Justice SOUTER would require lower courts to assess whether mapmakers
paid ‘little or no heed to . . . traditional districting principles.’ What is a lower court to
do when, as will often be the case, the district adheres to some traditional criteria but
not others?” Id. at 296 (plurality opinion). While it is true that the Vieth plurality
criticized reliance on traditional criteria, that hardly helps the Plaintiffs’ cause, since the
same plurality opinion would reject their claim altogether on justiciability grounds.
My point is not that all Justices would require unusually shaped districts before
considering a partisan gerrymander; the point is that of the Justices who would even
entertain a partisan gerrymandering claim, a majority would require adherence to
traditional districting principles as part of any test. Here, it is clear that seven of the
nine Justices in Vieth would have ruled against the Plaintiffs, either on justiciability
grounds or because they have not identified any violation of traditional districting
principles. No other conclusion can be drawn from the Justices’ separate opinions.
And, as discussed earlier, Justice Breyer would not find an unconstitutional
gerrymander here because this case does not involve a minority party “entrenching”
itself in power.
That leaves Justice Kennedy, whose Vieth concurrence expressed a grudging
willingness to consider political gerrymandering challenges, but did not give any
indication as to whether respect for traditional districting principles would play a role
in any test he might find appropriate. Even so, he remarked that a legal violation
would only arise if the line-drawers acted in an “invidious manner or in a way
unrelated to any legitimate legislative objective.” Id. at 307 (Kennedy, J., concurring).
Since respecting political subdivisions and following standards of compactness and
contiguity are “legitimate legislative objectives,” it would be impossible to say that Act
43, which actually achieved those objectives, was “unrelated to” those very objectives.
Id. That it achieved those objectives, as well as other legitimate objectives, including
consideration of the residence of the legislators themselves and Voting Rights Act
requirements, would seem to preclude the finding of any violation under whatever test
Justice Kennedy might entertain. The fact that the map-drawers chose to adopt plans
that were more “assertive” or “aggressive” than others (a mistake of nomenclature they
surely will not repeat) does not mean the maps they drew were “unrelated to”
legitimate traditional districting principles.
131
Indeed, Justice Kennedy’s view of the importance of traditional districting
principles can be gleaned from Miller v. Johnson, a racial gerrymandering case, where his
majority opinion found that “a plaintiff must prove that the legislature subordinated
traditional race-neutral districting principles, including but not limited to compactness,
contiguity, and respect for political subdivisions or communities defined by actual
shared interests, to racial considerations.” 515 U.S. 900, 916 (1995). Justice Kennedy
went on to note the district court’s finding that it was “‘exceedingly obvious’ from the
shape of the Eleventh District, together with the relevant racial demographics, that the
drawing of narrow land bridges to incorporate within the district outlying appendages
containing nearly 80% of the district's total black population was a deliberate attempt to
bring black populations into the district.” Id. at 917. Given the centrality of traditional
districting principles to racial gerrymandering cases, there is every reason to believe
that any political gerrymandering test Justice Kennedy might adopt would include the
plan’s adherence to such principles as part of its analysis.
The majority addresses these concerns by concluding that following traditional
districting principles should provide no “safe harbor” for an Equal Protection violation.
It is possible to see the argument in such a light if all one is concerned with is raw
numbers, or translating votes into seats. In my view, however, the Defendants are not
asking for a safe harbor, they are asking the court to conclude that the drawing of
bizarrely shaped districts is part of the very definition of unconstitutional
gerrymandering itself—to the extent such a claim exists. Looked at from the voter’s
perspective, living in a district that looks like some type of amphibian is itself a
component of any gerrymandering injury that voter might suffer. Thus, I do not view
the following of traditional districting principles as a “safe harbor” that would
whitewash any Equal Protection violation; instead it is evidence that the map-drawers
were not committing a violation at all. Without gerrymandered districts, there is no
unconstitutional gerrymander.
This conclusion is reinforced by Cox v. Larios, a one-person, one-vote case. There,
Democratic mapmakers in Georgia drew maps designed to pit large numbers of
Republican incumbents against each other, resulting in nearly half of the Republican
delegation losing their seats. The Supreme Court summarily affirmed the three-judge
district court’s decision finding an Equal Protection violation. 542 U.S. at 947–50. Key
to the district court’s conclusion was its finding that, in drawing the maps that
contained many “oddly shaped” districts, the Georgia legislators paid no heed to
traditional districting principles like compactness or contiguity. Larios v. Cox, 300 F.
Supp. 2d 1320, 1330 (N.D. Ga. 2004).
[O]ne can easily discern [an absence of compactness] just by looking at the
maps themselves . . . . Moreover, as we have noted, a more sophisticated
132
analysis of district compactness, calculated by the perimeter-to-area
measure or the smallest circle measure, also establishes that compactness
was not a factor here. Indeed, quite a few of the districts have shapes that
defy Euclidean geometry. The drafters of the House and Senate Plans
made no effort to keep districts compact and certainly did not create
deviations for the purpose of improving compactness.
Id. at 1350.
In concurring with the Supreme Court’s summary affirmance, Justices Stevens
and Breyer wrote that Georgia’s
partisan gerrymander is visible to the judicial eye . . . . Drawing district
lines that have no neutral justification in order to place two incumbents of
the opposite party in the same district is probative of the same
impermissible intent as the “uncouth twenty-eight-sided figure” that
defined the boundary of Tuskegee, Alabama, in Gomillion v. Lightfoot, 364
U.S. 339, 340 (1960), or the “dragon descending on Philadelphia from the
west” that defined Pennsylvania's District 6 in Vieth, 541 U.S., at 340
(STEVENS, J., dissenting) (internal quotation marks omitted).
Cox, 542 U.S. at 950 (Stevens, J., concurring).
Thus, contrary to the majority’s view, traditional districting criteria—the shape,
size and other physical characteristics of a district—are part and parcel of an Equal
Protection analysis because deviations from those norms are offensive wholly
independent from any partisan effect they might occasion.
The Plaintiffs and the majority also suggest that advances in computer
technology make it easy for map-drawers to produce pleasing-looking districts that
stealthily mask a partisan purpose, and so merely following traditional principles and
producing unsuspicious maps cannot be enough to pass muster. The idea of some kind
of high-tech stealth gerrymander is nothing more than a bugaboo, however. Computer
technology was advanced in 2004, when Vieth was decided. The Justices’ opinions cited
above would all require a plaintiff to demonstrate districts with unusual shapes,
without any apparent concern about computer technology. The Plaintiffs have offered
no evidence that the technology that existed in 2011, when the Republicans drew the
Act 43 maps, was somehow more sophisticated than what existed a mere seven years
earlier when Vieth was decided. 7
In fact, the Justices have been remarking on the use of technology in gerrymandering cases for decades.
In Bandemer, for example, Justice Powell noted that “[c]omputer technology now enables gerrymanderers
to achieve their purpose while adhering perfectly to the requirement that districts be of equal
7
133
It may be worth pointing out that the Justices’ desire for normal-looking district
lines is not a purely aesthetic conceit, or a “beauty contest.” Bush v. Vera, 517 U.S. 952,
977 (1996). As stated above, living in a bizarrely-shaped district is part of the injury a
voter suffers in an unconstitutional gerrymander. Geographic lines that everyone can
understand lend legitimacy to a district, minimize voter confusion, and suggest that
voters are being treated similarly based on where they live rather than how they have
voted in the past. As Justices Stevens and Powell have noted, “[C]onfigurations of the
districts [and] the observance of political subdivision lines . . . have independent
relevance to the fairness of redistricting.” Vieth, 541 U.S. at 322 (Stevens, J., dissenting)
(citing Bandemer, 478 U.S. at 165 (Powell, J., concurring in part and dissenting in part)).
And, as one commentator has noted:
Disregard of compactness facilitates gerrymandering by making it easier
to include reliable voters in a particular district and avoid those who
might be unreliable. It also destroys some of the advantages of singlemember districts, including a sense of community and an awareness of
what areas a district includes. Disregard of compactness also substantially
impairs the ability of potential candidates to organize on a grass-roots
basis. . . . A district that is non-compact, such as the infamous 160 mile
long “I-85” district in North Carolina, creates enormous difficulties in this
situation.
Paul L. McKaskle, Of Wasted Votes and No Influence: An Essay on Voting Systems in the
United States, 35 Hous. L. Rev. 1119, 1144–45 (1998).
Just as importantly, perhaps, part of the Justices’ interest in policing the
redistricting process is not merely in detecting invidious gerrymandering after the fact,
but in preventing it from happening in the first place. As demonstrated at trial, the
individuals drawing the lines will not know what their map’s efficiency gap will be
until after the first election—typically, more than a year later—making it impossible for
legislators to know in advance whether their plan will pass muster. In contrast, the
population.” 478 U.S. at 174 (Powell, J., concurring in part and dissenting in part). And in Vieth, Justice
Kennedy noted that “[c]omputer assisted districting has become so routine and sophisticated that
legislatures, experts, and courts can use databases to map electoral districts in a matter of hours, not
months.” 541 U.S. at 312 (Kennedy, J., concurring). Justice Breyer, too, observed that “[t]he availability of
enhanced computer technology allows the parties to redraw boundaries in ways that target individual
neighborhoods and homes, carving out safe but slim victory margins in the maximum number of
districts, with little risk of cutting their margins too thin.” Id. at 364 (Breyer, J., dissenting). Computer
technology was well-advanced in the 1980’s, and certainly by 2004, and the Justices were clearly aware of
its benefits and dangers. When Justices Souter, Stevens and Ginsburg (three-fifths of the Justices who
would consider political gerrymandering challenges) say that a test should include adherence to
traditional districting principles, we cannot simply ignore those opinions on the Plaintiffs’ say-so.
134
mapmakers (and their critics) will immediately be able to detect when their efforts have
produced unusual and suspicious visual results—dragons in flight, salamanders, sick
chickens, or any other of the flamboyantly monikered chimeras that creative
cartographers have conjured up over the decades. Unlike most witnesses who testified
at trial in this action, the line drawers will not require advanced graduate training in
statistics, regression analysis, or political science, but merely a respect for traditional
political boundaries and an affinity for relatively straight lines. Constitutional law need
not become the province of a cottage industry of Ph.D.’s and statisticians.
Another benefit of reliance on traditional districting factors is that the public and
other legislators, when presented with the proposed maps, will be able to identify
unusual shapes, and litigation may commence immediately to prevent unlawful
discrimination from affecting even a single election. Vieth, 541 U.S. at 339 (Stevens, J.,
dissenting, expressing hope that any test would “shorten the time period in which the
pernicious effects of such a gerrymander are felt.”) As noted earlier, the Plaintiffs’ test
will never even be triggered until after the first election under a new plan, which would
allow legislators a free bite at what the Plaintiffs describe as the forbidden fruit. If a
typical plan is only in force for five state assembly elections, a test that would guarantee
that an entire election cycle must occur before any challenge would seem inadequate to
the task of curbing the serious abuses the Plaintiffs allege. This is especially true in light
of the incumbency effect. Allowing an election to take place under an unconstitutional
gerrymander would allow political newcomers from the gerrymandering party to win
an election on an unfair playing field, but then run as incumbents in the next election,
thus preserving most or all of their ill-gotten gains even though the gerrymandered
plan has ostensibly been fixed. This is exactly what happened in Texas, after a courtdrawn plan remedied a pro-Democratic map: “in the 2002 congressional elections,
however, Republicans were not able to capitalize on the advantage that the Balderas
Plan had provided them. A number of Democratic incumbents were able to attract the
votes of ticket-splitters . . . and thus won elections in some districts that [now] favored
Republicans. As a result, Republicans carried only 15 of the districts drawn by the
Balderas court.” LULAC, 548 U.S. at 452 (Stevens, J., concurring in part and dissenting in
part).
These are surely among the reasons that Justices Souter and Stevens both
observed that constitutional violations should be easily detectible: Justice Souter (joined
by Justice Ginsburg) believed such violations “can be shown straightforwardly” when
traditional districting principles are violated, 541 U.S. at 348 (Souter, J., dissenting),
while Justice Stevens noted that an offending plan would be “irregular on its face,” id.
at 339 (Stevens, J., dissenting)—so obviously a gerrymander that the plan’s invidious
purpose would be immediately detectable. As noted above, this would alert the
drawers themselves that their plan was suspect, and if they failed to correct the problem
135
it would allow quicker litigation in order to prevent the offending plan from affecting
an election. In addition, requiring a violation of traditional districting principles would
serve as a check on court intervention into the inherently political process of mapdrawing. As this court recognized in its summary judgment decision, no member of the
Supreme Court has expressed a desire to involve the court in gerrymandering cases as a
matter of course. Justice Stevens suggested that his “narrow test would cover only a
few meritorious claims, but it would preclude extreme abuses,” such as those described
in the California case of Badham v. Eu, 694 F. Supp. 664 (N.D. Cal. 1988), which involved
“a large number of districts with highly irregular shapes, all designed . . . to dilute
Republican voting strength throughout the State.” 541 U.S. at 339 n.34 (Stevens, J.,
dissenting). As Justice Souter suggested, courts are eminently capable of assessing
traditional districting principles, Vieth, 541 U.S. at 348 (Souter, J., dissenting), whereas it
is not clear that they are equipped to undertake the complex statistical and political
science inquiries the Plaintiffs press in this case. 8
In sum, this is hardly fertile ground for the kind of test Plaintiffs propose. Every
Justice who has expressed an opinion on the subject would reject the Plaintiffs’ claim
either because it is non-justiciable; because the challenged plan did not involve minority
party entrenchment; or because the Plaintiffs failed to show that the Defendants
violated traditional districting principles in some meaningful way. If this case were
before the Vieth Justices, the Plaintiffs would likely lose 9-0.
III. The Republicans Would Control the Legislature Even Without a Gerrymander
Given courts’ historical reluctance to involve themselves in political
gerrymandering cases, it would seem that this case presents a particularly poor
candidate for court intervention. A key reason is that the Republicans would have won
control of the legislature in both elections under Act 43 even without a gerrymander. In
2014, the most recent election, they won a majority of the statewide vote, and so
naturally they would have won control of the chamber. And in 2012, the first election
under Act 43, they won close to 49% of the statewide vote. 9 Here, too, they would have
retained control of the legislature. My colleagues and I are in agreement that, based on
The Plaintiffs also suggest that their proposed test does account for traditional districting factors. For
example, if the Defendants can show that traditional districting criteria required them to draw the maps as
they did, then that would excuse the large efficiency gap. But being able to cite traditional principles as
some kind of defense is a far cry from the tests described above, which would require a failure to follow
such principles as part of the burden plaintiffs must show.
8
It appears from the Plaintiffs’ calculations that the Republicans won something on the order of 48.6% of
the statewide vote in 2012. (ECF No. 125 at ¶ 257.)
9
136
Wisconsin’s political geography and the large efficiency gaps that have existed even
under neutrally-drawn plans, Republicans enjoy some degree of natural advantage. (I
address geography below, but in a nutshell it comes down to Democratic voters’
tendency to live in closely compacted areas in Milwaukee and Madison, whereas
Republicans are more efficiently dispersed.) In fact, even the Plaintiffs’ own
demonstration map, when adjusted to include the effect of incumbency, produced an
efficiency gap of nearly 4% in favor of Republicans—and recall that this was a map
drafted by a political science professor, hired by the Democratic-voting Plaintiffs, whose
entire goal was to try to produce the smallest efficiency gap possible. (ECF No. 149 at
65:3.) Accordingly, it is very likely that Republicans, despite receiving less than 49% of
the statewide vote in 2012, would have won control of the legislature even without any
gerrymandering whatsoever, because they would have enjoyed a substantial advantage
even under a neutrally drawn plan.
This is a major obstacle to the Plaintiffs’ argument because their case, as
explained below, is based solely on an injury they describe as an inability to convert
statewide vote totals into seats in the legislature; in other words, they blame the
Republican gerrymander for their inability to control that branch of government. The
fact that their inability to control the legislature is due not to Republican
gerrymandering but to Republican statewide strength combined with certain natural
advantages means, at a minimum, that this case is hardly the kind of outrageous
partisan iniquity the Plaintiffs portray it to be. Many of the Justices who would
entertain political gerrymandering challenges have expressed only a grudging
willingness to do so, leaving the door open for review of only those most egregious
partisan injustices. See, e.g., Vieth, 541 U.S. at 339 (Stevens, J., dissenting) (Justice
Stevens’ narrow test “would cover only a few meritorious claims, but it would preclude
extreme abuses.”). Here, it is difficult to perceive an extreme abuse when the
gerrymandering party would have won control of the legislature even without
gerrymandering.
IV. Theoretical Problems with the Efficiency Gap and other Votes/Seats Measures
In this court’s decision denying the Defendants’ motion to dismiss, the panel
observed that the justices had expressed some support for the concept of partisan
symmetry, a doctrinal cousin of the efficiency gap. Whitford v. Nichol, No. 15-CV-421BBC, 2015 WL 9239016, at *9 (W.D. Wis. Dec. 17, 2015). However, the court correctly
noted that Justice Kennedy’s support was “tepid, at best,” and at the time we could also
have rightly observed that the support of the other Justices was hardly a ringing
endorsement of the symmetry theory. Id. (citing LULAC, 548 U.S. at 419–20 (Kennedy,
J., concurring in the judgment)). Despite this faint praise, the court now is being asked
137
to elevate the efficiency gap theory from the annals of a single, non-peer-reviewed law
review article to the linchpin of constitutional elections jurisprudence. This request is
made despite the efficiency gap’s significant, and likely insurmountable, limitations, as
detailed below.
A.
The Plaintiffs’ Case is Premised on a Right to Proportional Representation
The concepts of efficiency and waste are inherently normative ones, requiring us
to consider the proper role of a vote, as opposed to a vote being “wasted.” If we say
something is efficient, that implies knowledge of an ultimate purpose or goal: if a
furnace is 90% efficient, that is a measure of how well it converts fuel into heat, with
heat being the goal. According to the Plaintiffs, the goal of voting—voting’s only
purpose, in fact— is to convert votes into additional seats in the assembly, and so one’s
vote is only efficient insofar as it translates into more seats. Any other result is wasted
and inefficient, like heat escaping from a leaky furnace.
Whether the argument is premised on the efficiency gap or on other measures
comparing legislative seats to statewide votes, it is clear that the Plaintiffs’ case is really
premised on a right to proportional representation, that is, the right to translate one
party’s statewide vote totals into a given number of seats in the legislature. If Party A
has a large statewide total of votes, say 60%, but has only received 51% of the seats,
there is a large efficiency gap reflecting the disproportionality of that party’s
representation: the number of seats they won was disproportionally small compared to
their statewide vote totals. Any injury premised on such a comparison is an injury
based on an absence of proportionality. As the parties have recognized, however, there
is no constitutional requirement that groups of voters must enjoy political strength
proportionate to their numbers. The Bandemer court recognized that “the mere fact that
a particular apportionment scheme makes it more difficult for a particular group in a
particular district to elect the representatives of its choice does not render that scheme
constitutionally infirm.” 478 U.S. at 131–32 (plurality opinion). “Our cases . . . clearly
foreclose any claim that the Constitution requires proportional representation or that
legislatures in reapportioning must draw district lines to come as near as possible to
allocating seats to the contending parties in proportion to what their anticipated
statewide vote will be.” Id. at 130.
This principle was reiterated a decade later in Vieth:
Deny it as appellants may (and do), this standard rests upon the principle
that groups (or at least political-action groups) have a right to
proportional representation. But the Constitution contains no such
principle. It guarantees equal protection of the law to persons, not equal
138
representation in government to equivalently sized groups. It nowhere
says that farmers or urban dwellers, Christian fundamentalists or Jews,
Republicans or Democrats, must be accorded political strength
proportionate to their numbers.
541 U.S. at 288 (plurality opinion); see also id. at 338 (Stevens, J., concurring) (“The
Constitution does not, of course, require proportional representation of racial, ethnic, or
political groups.”).
My colleagues concede, as they must, that there is no constitutional right to
proportional representation. In their view, however, the fact that there is no right to
proportional representation does not foreclose looking to disproportional representation
as evidence of a discriminatory effect. Yet it is unclear to me how that statement differs
in practical terms from establishing a covert right to proportional representation itself: if
there is no constitutional right to something, then why look to the absence of that thing
as evidence of constitutional injury? Saying that there is a right to not have
disproportional representation is tantamount to saying there is a right to have
proportional representation. Suppose a plaintiff incarcerated in prison claimed injury
because his meals tasted bad; in particular, he complained that the prison refused to
serve him filet mignon and lobster for dinner every night. Of course there is no
constitutional right to have steak and lobster in prison, and so a court would summarily
reject the claim on that basis and move on. No court in the land would say that,
“although there is no right to eat steak in prison, we see no reason we can’t consider the
absence of steak and lobster as evidence that the prison’s food is so poor that it violates
the Eighth Amendment.” If something is not a constitutional right, then its absence
cannot cause constitutional injury. Here, the majority appears to be saying in one
breath that there is no right to proportional representation but then in the next that the
absence of proportional representation may constitute the entire basis of a cause of
action. Disproportionality cannot be viewed merely as evidence of a partisan effect—
the absence of proportionality is the signature feature of the Plaintiff’s entire case.
In denying that the Plaintiffs’ theory is based on a right to proportional
representation, the majority also relies on an opinion of Justice Kennedy, who observed
in LULAC that “a congressional plan that more closely reflects the distribution of state
party power seems a less likely vehicle for partisan discrimination than one that
entrenches an electoral minority.” 548 U.S. at 419. From this, the majority appears to
extrapolate the principle that when the number of seats a party wins deviates from how
many we would “expect” it to receive, such a scenario could prove an unconstitutional
partisan gerrymander. Again, however, the notion that we would “expect” a given
number of seats requires imputing the normative judgment that a party’s seats won
139
must be proportional to the party’s statewide vote totals. The fuller context of Justice
Kennedy’s statement is as follows:
[C]ompared to the map challenged in Vieth, which led to a Republican
majority in the congressional delegation despite a Democratic majority in
the statewide vote, Plan 1374C can be seen as making the party balance
more congruent to statewide party power. To be sure, there is no
constitutional requirement of proportional representation, and equating a
party's statewide share of the vote with its portion of the congressional
delegation is a rough measure at best. Nevertheless, a congressional plan
that more closely reflects the distribution of state party power seems a less
likely vehicle for partisan discrimination than one that entrenches an
electoral minority.
Id.
My reading of the above paragraph is that Justice Kennedy would probably
agree with Justice Breyer that a map that allowed a statewide minority party to
consistently win a majority of seats would be constitutionally suspicious. Justice
Kennedy notes that the map reviewed in LULAC did not do this, however, because
Republican congressional candidates won 58% of the statewide vote in Texas and
received a healthy majority of 21 of the 32 available seats. Id. at 413. In other words,
because the majority party received a majority of seats, LULAC was not a case where a
plan “entrenches an electoral minority.” Id. at 419. This observation, modest as it is,
does not suggest that disproportionality might be injurious on its own; instead, it merely
means that it could prove problematic when the disproportionality is what allows a
minority party to win a majority of seats—the entrenched minorities also described by
Justice Breyer. In short, from Justice Kennedy’s opinion I am unable to glean a principle
that would treat disproportional representation per se as a constitutional injury. If
anything, it suggests a more stringent threshold for plaintiffs, requiring them to show
that an established minority party has managed to rig the system to entrench itself in
power despite the evident will of a majority of voters.
The Plaintiffs also argue that they are not insisting on using exact proportional
representation as their benchmark. For example, they do not say that winning 48% of
the statewide vote entitles them to 48% of the seats. But no one in Bandemer, or in any
other case brought to the court’s attention, had insisted on strict 1:1 proportionality
either, and so when they rejected gerrymandering challenges on that basis the courts do
not appear to have had “strict” proportional representation in mind; they were rejecting
the concept of proportionality more broadly. This is clearest in Vieth, where the
plaintiffs argued for a loose proportionality standard that would entitle a party who
won a majority of the statewide vote “to translate a majority of votes into a majority of
140
seats.” 541 U.S. at 287 (plurality opinion). The plaintiffs were not arguing their 51% of
the statewide vote entitled them to 51% of the seats, but merely that a statewide victory
entitled them to control of the legislature—any percentage greater than 50%. The court
rejected that test on the ground that the Constitution does not require that political
parties “must be accorded political strength proportionate to their numbers.” Id. at 288.
Thus, the court rejected that test not because a political party had no entitlement to a
strict proportion of seats to votes; it rejected it because parties are not entitled to any
proportion at all.
Here, the Plaintiffs’ claim is even more specific than the argument posited in
Vieth: not only do the Plaintiffs insist on receiving a majority of seats for a majority of
the vote (as in Vieth), they propose a linear 2:1 relationship between additional votes
and seats. 51% of the statewide votes should garner 52% of the seats, while 54% of the
votes would win 58% of the seats, and so on. Any significant deviation from that
predetermined proportion must be justified in court. Such a scheme, of course, is the
essence of proportionality. 10
At this point it might be worth exploring why proportional representation is not a
constitutional right. A key reason is that each election in each district is a separate
affair. Wisconsin’s constitution, like that of the nation, did not create a form of
government in which the party, or coalition of parties, that wins the majority of the
statewide vote is given all of the tools needed to enact and implement its legislative
program. Instead, we elect our representatives on a district-by-district basis. Some
candidates will win in landslides while others squeak out narrow victories. There is no
inherent reason to draw statewide inferences about the number of seats a given party
“should” win based on either scenario. “[O]ne implication of the districting system is
that voters cast votes for candidates in their districts, not for a statewide slate of
legislative candidates put forward by the parties. Consequently, efforts to determine
party voting strength presuppose a norm that does not exist–statewide elections for
representatives along party lines.” Bandemer, 478 U.S. at 159 (O’Connor, J., concurring).
Particularly at the assembly level, candidates are close to their voters. Responsiveness
and personalities matter. The Plaintiffs have provided no reason to assume that each
vote for a given candidate should be transformed into a vote for a state-wide party, nor
why the total votes received by a group of candidates in 99 different districts should
Notably, the Plaintiffs’ proposal also produces unusual results. For example, under the Plaintiffs’ test,
if a party received 60% of the statewide votes and won the same 60% of the seats, that would produce a
10% efficiency gap simply because it deviates from their preordained 2:1 relationship—with 60% of the
vote, the victorious party “should” have received a seat bonus of 70% of the seats. Obviously, the need
for court intervention in such a case would be completely absent, because the statewide majority party
has won a large majority of seats. Yet under the Plaintiffs’ test such a plan must be thrown out (assuming
intent were present) as an unconstitutional gerrymander.
10
141
play some kind of prescriptive role in determining how many districts that party
“should” win.
Another reason proportionality is not a right is that disproportionality is built in,
and in fact even assumed, in winner-take-all systems of voting. “District-based elections
hardly ever produce a perfect fit between votes and representation.” Id. at 133
(plurality opinion). On the federal level, the nationwide popular vote does not
determine the presidency, and neither does it determine the House of Representatives
or the Senate, both of which are voted on individual districts or separate states. If there
is an anomaly in wasted votes between the parties, we do not rejigger the seats to grant
one side more seats: wasted votes are just wasted votes. The same is true in any
assembly district. A candidate could lose by a single vote, and yet none of the votes cast
for him will translate into any additional power for his party. This is simply the nature
of any system where the winner gets everything and the loser receives nothing. Early in
our nation’s history, we experimented with a kind of proportional representation by
allowing the second-place presidential candidate to become vice-president, giving
something of a consolation prize to all of those voters whose votes would otherwise be
“wasted.” But soon enough, after Thomas Jefferson became vice-president under
President John Adams, that system was abandoned in favor of a winner-take-all
paradigm.
Many other countries, including many of the countries in Western Europe,
require some fashion of proportional representation, for example, by allowing voters to
vote for a list of candidates. “If properly implemented, [proportional representation]
allows all significant groups (political, racial, or otherwise) of the electorate to be
represented in proportion to their population, it eliminates the evils of gerrymandering,
and it eliminates the need to use race-conscious criteria in creating legislative districts.”
McKaskle, Of Wasted Votes and No Influence: An Essay on Voting Systems in the United
States, supra at 1126. But that is not the system of government the people who drafted
and ratified the constitutions for the State of Wisconsin and the nation chose.
The point is that proportional representation is one possible way of electing
legislators, governors or presidents, but it is not the only way. When states opt for
winner-take-all districts, disproportionality is simply a side-effect of that decision:
If all or most of the districts are competitive . . . even a narrow statewide
preference for either party would produce an overwhelming majority for
the winning party in the state legislature. This consequence, however, is
inherent in winner-take-all, district-based elections, and we cannot hold
that such a reapportionment law would violate the Equal Protection
Clause because the voters in the losing party do not have representation in
142
the legislature in proportion to the statewide vote received by their party
candidates.
Bandemer, 478 U.S. at 130 (plurality opinion). This inherent disproportionality is more
pronounced in states where the voters of one party are naturally clustered, or “packed”
in relatively small geographic regions, like Wisconsin’s Democratic voters are in
Milwaukee and Madison, as is explained by the majority opinion and below. In
essence, adoption of the efficiency gap (or any other “gap” between statewide vote
totals and seats) in such states would undermine the districting system itself. “If there
is a constitutional preference for proportionality, the legitimacy of districting itself is
called into question: the voting strength of less evenly distributed groups will
invariably be diminished by districting as compared to at-large proportional systems
for electing representatives.” Id. at 159 (O’Connor, J., concurring).
In fact, the only way to counter the adverse effect of the natural packing of one
party’s voters in a few discrete geographic areas in pursuit of the goal of proportional
representation is to “reverse” gerrymander districts in an attempt to more evenly
distribute that party’s voters. Id. at 160. That is precisely what the Plaintiffs’ expert, Dr.
Mayer, did with his demonstration plan. It wasn’t that the Defendants considered
partisan voting patterns in designing their plan and Dr. Mayer did not. Indeed, Dr.
Mayer considered actual votes, an advantage Defendants’ map-drawers did not have,
and assumed that each vote would be for the same party’s candidate even if voting in
different districts with different candidates. Regardless of whether that assumption is a
reasonable one, the larger point is that requiring some kind of statewide votes-to-seats
proportionality in a system where elections are for representatives in winner-take-all
districts does not eliminate partisan gerrymandering, if by partisan gerrymandering
one means drawing districts based on past voting history. Instead, it would
constitutionally mandate gerrymandering in order to offset the effects of natural
packing.
It follows that the number of votes a party receives in an entire state should have
no relevance to any gerrymandering injury alleged by a voter in a single district,
because any reference to statewide strength is the essence of proportionality. Id. at 130
(plurality opinion) (defining proportional representation as drawing district lines that
are “in proportion to what their [party’s] anticipated statewide vote will be”). The
premise of any test that merely compares statewide votes to seats is that there is
something constitutionally wrong with disproportional representation. This is nothing
short of a claim that voters of one party “must be accorded political strength
proportionate to their numbers.” Vieth, 541 U.S. at 288 (plurality opinion). Because
there is no such constitutional right, I would enter judgment for the Defendants.
143
B. The Efficiency Gap Incorrectly Treats Seats Won as a Measure of Political Power
It seems intuitive to consider a party’s number of assembly seats as an adequate
measure of political power in the assembly. The efficiency gap merely measures each
party’s ability to win more seats, and so the efficiency gap also has a basic intuitive
appeal. But upon even a cursory examination, it becomes clear that a party’s number of
seats is often a poor measure of political strength. For example, if the Republicans had
51 members to the Democrats’ 48, only a political neophyte might think the two parties
enjoyed about equal strength. The reality, of course, is that the Republicans have
tremendously more power simply by virtue of the few extra seats that give them the
majority in the legislature. Conversely, compare a Republican majority of 60-39 to a
majority of 70-29. In the 60-39 case, the Republicans have a 21-seat edge, or 54% more
seats than the Democrats. In the 70-29 assembly, the Republicans enjoy a massive
advantage with more than double the Democrats’ number of seats. And yet no one
with any experience in politics would think there was much practical difference
between the two majorities. Once a majority is comfortable (however defined), the
party in control has the ability to pass whatever bills it wants, and therefore winning (or
losing) additional seats will often provide no practical increase (or decrease) in a party’s
political power. 11 The point is that every seat gained (or lost) does not represent an
equivalent increase (or decrease) in political power—what is crucial is usually only the
seats necessary for one party to secure a comfortable majority.
In 2014, the Republicans won 52% of the statewide vote and took 63 seats. The
Democrats won 48% of the vote and took the remaining 36 seats. This resulted in an
efficiency gap of around 10% in favor of the Republicans. (ECF No. 125 at ¶¶ 258, 290.)
If the efficiency gap were zero (the Plaintiffs’ ideal), the Republicans would have won
only 54% of the assembly seats (53 or 54 seats), while the Democrats would have won
46% (45 or 46 seats). So, instead of enjoying a 54-45 majority, the purported
gerrymander (allegedly) allows the Republicans to enjoy a more robust 63-36 majority.
The problem is that the Plaintiffs never even attempted to identify a single practical
difference in their political power between the actual 63-36 Republican majority and the
“ideal” 54-45 majority that would exist under a zero efficiency gap. Whether the
Republicans have a majority of 9, 15, or 27 is not likely to impact anyone in any material
sense: either way, the Republicans are in charge (by a comfortable margin) and able to
pass whatever bills they want to pass. Not surprisingly, the Plaintiffs have identified no
legislation that passed only because the Republican majority was larger than it
otherwise would have been. (Ironically, the most controversial of the Republicans’ bills,
It is true that a two-thirds majority will have the power to override gubernatorial vetoes, but no one has
suggested that would be relevant here.
11
144
Act 10, was passed by the 2011 legislature, which was elected under a court-drawn
district plan.)
This demonstrates at least three things. First, it is difficult to perceive any injury
worthy of court intervention when a party that wins a majority of the statewide vote
has merely increased its number of seats beyond what a zero efficiency gap would
mandate.
A second obvious implication of the above is that any measure that treats all
seats as being of equal value cannot be a reliable measurement of political harm. The
efficiency gap is all about increasing seats, treating every seat as equal and the gaining
of more seats as the only efficient use of a vote. But in many elections, including 2014,
the additional seats the majority party gained, allegedly through their gerrymander, do
not appear to have any discernible impact on their power. In fact, as long as the
Republicans maintain statewide vote totals above 48 or 49 percent (as in 2012), we
would expect (based on history and even the demonstration plan) that even under a
neutrally-drawn plan they would enjoy comfortable control of the assembly. Thus, a
measure that is based solely on the number of seats one party wins does not seem up to
the task of measuring, or even identifying, the kinds of partisan gerrymanders that
might cry out for court intervention. Because all seats are not alike, neither are all
efficiency gaps alike. A 10% gap, as seen in 2014, will be of almost no practical import
because it merely increased seats for a party that would have maintained comfortable
control of the chamber even without gerrymandering. This gives the lie to the
Plaintiffs’ hyperbolic claim that the instant case represents “one of the worst partisan
gerrymanders in modern American history.” (ECF No. 1 at ¶ 1.)
In sum, as a general principle, the efficiency gap oversimplifies political injury by
assuming that any gain or loss of seats equates to a proportional gain or loss of political
power, when in fact the raw number of seats is often irrelevant. By reducing political
power to gaining seats—regardless of how many seats the gerrymandering party would
otherwise have—the efficiency gap does not adequately measure, or even detect,
political gerrymandering injuries. Accordingly, I would not rely on the efficiency gap,
or any other measure comparing statewide votes to seats, to find a partisan
gerrymander in this case.
C. Votes are Meaningful, even if “Inefficient”
In addition to oversimplifying the analysis by treating all seats equally, the
Plaintiffs’ analysis ignores the fact that votes “count” even if they do not lead to
additional seats. “[O]ur system of representative democracy is premised on the
assumption that elected officials will seek to represent their constituency as a whole,
145
rather than any dominant faction within that constituency.” LULAC, 548 U.S. at 469–70
(Stevens, J., concurring in part and dissenting in part) (citing Shaw v. Reno, 509 U.S. 630,
648 (1993)). It is of course undeniable that one of the central purposes in voting is to try
to have one’s political party win additional seats, especially if that means taking control
of a branch of government. But “the power to influence the political process is not
limited to winning elections.” Bandemer, 478 U.S. at 132 (plurality opinion). In short, it
is not accurate to say that votes are “wasted” merely because they fail to increase seats
for one’s political party. The Plaintiffs’ reliance on the efficiency gap is ultimately a
reductionist exercise that fails accurately to account for the influence of lost votes and
exaggerates the role of winning seats in the voting process.
1. Votes are not “Wasted” Simply Because they do not Produce Additional Seats
The Plaintiffs have presented this as a cracking case, meaning that they allege the
Republicans drew the maps in order to allow themselves to win a large number of close
(but not too close) elections in districts that skewed slightly Republican. This enabled
the Republicans to efficiently win narrow victories, while the Democrats squandered
hundreds of thousands of votes in landslide wins in their own districts. Even though
the Plaintiffs would no doubt prefer that the Democrats had won some of those seats, it
is not as though those lost votes are completely “wasted.” Plaintiffs ignore the fact that
Republicans and Democrats are not fungible: the (R) next to a candidate’s name does
not mean he will vote the same as the Republican candidate in the next district. “The
two major political parties are both big tents that contain within them people of
significantly different viewpoints.” Baldus v. Members of Wis. Gov't Accountability Bd.,
849 F. Supp. 2d 840, 851 (E.D. Wis. 2012). Thus, a Republican who has won with only
51% of the vote will very likely govern differently than one who has a safe seat, just as a
Republican in Massachusetts will be different from one in Utah. It is exceptionally
likely that legislators in swing districts will adopt more moderate, centrist positions
than some of their colleagues, and they will of necessity be more responsive to the 49%
of the electorate that did not vote for them. If that is true, then the losing candidate’s
votes were not “wasted” at all. They serve as an unsubtle reminder that the legislator
may ignore the views of the minority party at his own risk. The same, of course, is true
of those legislators whose seats are so safe that they routinely win in landslides or
seldom face opposition. It would not be surprising if legislators from Milwaukee
Democratic districts or suburban Waukesha County Republican districts, for example,
represented viewpoints further from the center of their respective parties’ ideologies,
being more concerned about a primary challenge from within their own party than any
threat from a candidate from the other party. The fact that thousands of votes in those
districts do not translate into seats does not mean that they have no impact on the
146
individuals who represent those districts. Instead, they provide cover to legislators on
both sides of the aisle and give voice to the more liberal and conservative views their
respective parties espouse. As a general principle, legislators from safe seats behave
differently: “the Constitution does not answer the question whether it is better for
Democratic voters to have their State’s congressional delegation include 10 wishywashy Democrats (because Democratic voters are “effectively” distributed so as to
constitute bare majorities in many districts), or 5 hardcore Democrats (because
Democratic voters are tightly packed in a few districts).” Vieth, 541 U.S. at 288 (plurality
opinion). Since it is the excess of wasted votes that make those seats safe in the first
place, these excess votes cannot be said to be wasted: they shape the larger political
debate, even if they do not translate into additional seats in the legislature. As the
Bandemer plurality explained:
the power to influence the political process is not limited to winning
elections. An individual or a group of individuals who votes for a losing
candidate is usually deemed to be adequately represented by the winning
candidate and to have as much opportunity to influence that candidate as
other voters in the district. We cannot presume in such a situation,
without actual proof to the contrary, that the candidate elected will
entirely ignore the interests of those voters. This is true even in a safe
district where the losing group loses election after election. Thus, a
group’s electoral power is not unconstitutionally diminished by the
simple fact of an apportionment scheme that makes winning elections
more difficult, and a failure of proportional representation alone does not
constitute impermissible discrimination under the Equal Protection
Clause.
478 U.S. at 132 (plurality opinion).
Finally, it should go without saying that because the ballot is secret, a minorityparty voter in a given district will have as much access to his legislator as any other
voter—to seek help in dealing with a government agency, to express a view about
pending legislation, or to request help in securing funds for repairing a local bridge or
extending a state bike trail. The bills passed by a legislature get all of the attention, but
the behind-the-scenes, day-to-day work of a legislator involves countless services for
constituents, none of which depend on which party holds a majority in the assembly.
Focusing solely on translating votes into seats ignores the fact that winning additional
seats is not the only purpose in voting.
147
2. Voting is Simply a Choice For One’s Own District
In addition, reliance on the efficiency gap ignores what actually occurs at the
ballot box and how voters likely perceive what they are doing by voting. Simply put,
many voters do not think in terms of efficiency or wasted votes or, more generally,
about translating votes made in individual districts into a statewide phenomenon.
Imagine a voter in one of the state’s heavily partisan districts in which the assembly
candidates routinely run without opposition. For example, in 2014 Democratic
incumbent Rep. Leon D. Young won District 16 with 16,183 votes compared to just 261
votes for unspecified write-in candidates, a landslide win with more than 98% of the
vote.12 When those 16,183 voters placed their vote for Rep. Young—the only name on
the ballot—they very likely knew that Young would win in a landslide and that their
vote was an exercise in futility, at most a symbolic gesture. They could have left that
spot blank, or stayed home that day, and Rep. Young would have won anyway, since
he was unopposed. The same is true of voters in District 58, where Republican Bob
Gannon ran unopposed and won 22,087 votes to just 483 votes for unregistered
candidates. Surely most of these candidates’ voters knew their votes were “wasted” in
the Plaintiffs’ sense of the term—that they were unnecessary to winning any additional
seats for their candidate’s party. But it is unlikely that such voters perceived some sort
of injustice arising out of the fact that Young and Gannon—the candidates they
supported—would win by such large margins. In other words, voters in such
circumstances expect that their votes will not gain additional seats for their party (on a
statewide basis), and, to the extent they consider the question at all, they likely believe
that such a phenomenon (“inefficiency”) is simply part of the democratic process.
The larger point is that, in voting, a citizen is simply expressing a choice about
who he believes is a better candidate to represent his own district, which of course is the
only question the ballot asks the voter to answer. The Plaintiffs presented no evidence
that voters view their vote as an exercise in maximizing the number of seats their party
wins in the assembly, nor is it plausible that voters believe their vote in a single district
should be calculated in assessing whether the number of seats their party won, on a
statewide basis, is fair. “[O]ne implication of the districting system is that voters cast
votes for candidates in their districts, not for a statewide slate of legislative candidates
put forward by the parties. Consequently, efforts to determine party voting strength
presuppose a norm that does not exist– statewide elections for representatives along
party lines.” Bandemer, 478 U.S. at 159 (O’Connor, J., concurring).
G.A.B. CANVASS REPORTING SYSTEM COUNTY BY COUNTY REPORT, 2014 GENERAL ELECTION (Nov. 26, 2014,
2:12 PM), http://www.gab.wi.gov/sites/default/files/11.4.14%20Election%20Results%20-%20all%20officesc%20x%20c%20report.pdf.
12
148
In fact, it is not difficult to imagine some voters preferring a result opposite of the
Plaintiffs’ assumption. Although there are thousands of die-hard party members like
Plaintiff Whitford in both parties, many voters are not quite so committed. A given
voter might like an incumbent Republican in his own district, even if that voter leans
Democratic in other respects, and so such a voter will vote for the Republican assembly
candidate even while preferring that his vote does not translate into additional
Republican seats in the assembly. Such a voter would be surprised if his wasted
Republican vote were used in some sort of ex post facto calculus to determine whether
the Republican candidates won “enough” seats that year. And what of ticket-splitters
and independents? Imagine a voter who votes for a Democratic assemblyman, a
Republican state senator and a Republican governor. What are we to make of such a
ballot, except to conclude that the voter is expressing individual preferences about
individual races, rather than some kind of global desire to increase seats for a given
party?
In sum, reliance on the efficiency gap ignores what the Bandemer court pointed
out, which is that there is more to politics than winning seats, and even the winning of
more seats often has little practical impact on one party’s power. In addition, it
overlooks the reality that individual voters do not perceive winning additional seats as
the overwhelming purpose of voting, either. Because the efficiency gap (as well as
Professor Gaddie’s S-curves) are measures only of translating statewide vote totals into
legislative seats, it is difficult to see how they could adequately measure any
unconstitutional level of partisan gerrymandering.
D. The Efficiency Gap Begs the Ultimate Question
An additional problem with the Plaintiffs’ reliance on the efficiency gap is that
the theory relies on circular logic to prove its point. Specifically, in this case the
efficiency gap is merely a somewhat more sophisticated way of saying that the
Republicans won a large number of close elections. This is because winning close
elections is the surest way to make sure the other side racks up lots of wasted votes—
every losing vote is wasted, whereas only a few winning votes are wasted. For example,
if A defeats B 5,200 to 4,800, A has wasted only 199 votes while B has wasted a
whopping 4,800—an eye-popping efficiency gap of 46%! This adds up, of course, any
time there is a statewide trend, and so any time one party wins a lot of close elections,
the efficiency gap will necessarily be high. That is simply and unavoidably how the
Plaintiffs’ math works. But simply stating that there is a high gap does not tell us
anything about gerrymandering, however, even if partisan intent is present; it simply
means one side won significantly more close elections than the other. And the
efficiency gap presumes that every lost vote in every election is a “cracked” vote, i.e.,
149
evidence of gerrymandering. Under the Plaintiffs’ theory, any time one side wins a lot
of close elections, the map must have been gerrymandered (assuming one side
controlled the process). 13
The second problem resulting from reliance on the efficiency gap is that the
Plaintiffs would use the Republicans’ own electoral success against them: under their
logic, the more close races the Republicans win, the more votes the Democrats waste,
which produces a large efficiency gap and therefore means the Republicans’ wins must
have been the result of an invidious gerrymander—a self-fulfilling prophecy. It thus
should be clear that using the efficiency gap simply begs the question of whether there
was a gerrymander by answering “yes” any time one party wins significantly more
close elections than the other. Without addressing why one party might have won more
close races than the other, and without evidence of specific districts that were
gerrymandered, we are left only to guess that the result must have been caused by
gerrymandering.
This reinforces my view, set forth above, that it is dangerous, and even
misleading, to find unconstitutional gerrymandering on the basis of statewide vote
totals rather than looking at actual maps to detect suspiciously-drawn districts that are
non-contiguous or compact. In this case, there was no evidence of an actual
gerrymandered district, no map that looked bizarre, and not even a suggestion as to
how the map-drawers moved lines here and there to achieve their allegedly
unconstitutional ends. Instead, the evidence of the effects of gerrymandering is simply
that one party won a lot of close elections. It should be obvious that winning close
elections is not unconstitutional, and yet that is all the efficiency gap shows—that a
party who loses lots of close races will have far more wasted votes, producing the high
efficiency gap seen in this case. Thus, without any actual evidence of gerrymandering, I
would find in favor of the defendants. 14
The Plaintiffs would object that this analysis ignores the fact that their test also requires evidence of
partisan intent. Thus, it is triggered not merely by the existence of a certain efficiency gap but also the
presence of intent. But as this case demonstrates, it will be easy enough to show intent whenever one
side controls the process. The fact that there will always be some partisan intent in cases like this will
enshrine the efficiency gap analysis as the decisive factor.
13
A final concern is that the Plaintiffs’ test presumes that parties should have the same number of wasted
votes, even if there are different numbers of voters for each party. But what happens when the two
parties have very different numbers of voters? In Massachusetts, for example, three of every four voters
who registered a party affiliation registered as a Democrat. Massachusetts Registered Voter Enrollment:
1948–2016, WILLIAM FRANCIS GALVIN, SECRETARY OF THE COMMONWEALTH OF MASS.,
https://www.sec.state.ma.us/ele/eleenr/enridx.htm (last visited Nov. 3, 2016). This is reflected in the fact
that the Massachusetts House of Representatives consists of 122 Democrats and only 34 Republicans.
14
150
V. Practical Problems with the Efficiency Gap
A. The Efficiency Gap’s “Wasted Votes” Metric Appears Incomplete
In addition to the more abstract problems with the efficiency gap and other
votes/seats measures noted above, more practical ones are evident as well. I begin with
what appears to be the Plaintiffs’ method of calculating wasted votes. To recall, a
“wasted” vote falls into one of two categories: a vote in excess of 50%+1 for the winning
candidate (“surplus” votes), and any vote for a candidate who has lost (“lost votes”). It
is easy enough to understand how to calculate a party’s lost votes, but it remains
opaque why a party’s surplus votes should be calculated based on a “50% plus one”
standard. (In fact, the Stephanopoulos and McGhee article, in which the theory is
propounded, ignores the “plus one” requirement entirely, but that is beside the point. 15)
The theory is that the winning party needed 50% plus one of the total votes cast in order
to win the seat, and so any votes in excess of that amount are deemed “surplus” and
therefore wasted.
But reliance on one-half (plus one) of the total votes produces unexpected results,
primarily because winning elections is not an exercise in division but in addition: in
reality, all you need to win an election in a two-candidate race is one more vote than the
other candidate, not 50%-plus-one of the total votes. For example, if the Indians defeat
the Cubs 8 to 2, any fan might say that the Indians “wasted” 5 runs, because they only
needed 3 to win yet scored 8. Under the Plaintiff’s theory, however, the Indians needed
5 runs to beat the Cubs that day: 4 runs to reach 50% of the total runs, plus one to win.
That, of course, is absurd.
The central flaw is that when discerning how many votes it takes to win an
election, we should not care what the total votes are, because that is an abstraction that
factors in how many votes the winning candidate receives. Since every vote cast for the
winning candidate increases the total number of votes (the denominator of the
percentage), it also necessarily increases the number that candidate needs to reach 50%
plus one. This reduces, by half, the winning candidate’s number of wasted votes. The
key point is that there is no reason to believe the number of votes needed to win should
be determined by how many votes the winning candidate receives. Just as a baseball
Under the Plaintiffs’ efficiency gap analysis, a perfectly symmetrical map (efficiency gap of zero) would
require equal numbers of wasted votes on both sides. Yet such a result is impossible to imagine in a state
where the number of voters in each party is so unequal at the outset. On what constitutional principle
would one rely to expect that the Republicans, who are vastly outnumbered, could ever produce similar
numbers of wasted votes as the Democrats?
Nicholas Stephanopoulos & Eric McGhee, Partisan Gerrymandering and the Efficiency Gap, 82 U. Chi. L.
Rev. 831 (2015).
15
151
game is not decided by reference to total runs, an election is not decided by a fraction of
total votes. Instead, the number of votes needed to win is simply the number one more
than the losing candidate won, and therefore anything beyond that should be counted as
a “wasted” vote, using Plaintiff’s terminology.
This defect is not just a quibble because it exposes the oddity of a scenario the
Plaintiffs described. In an effort to downplay the influence of naturally packed
Democratic voters in Milwaukee on the efficiency gap (a phenomenon discussed
below), Plaintiffs asserted that in a 75-25 district, wasted votes for each party would be
a “wash.” Under their math, if the Democratic candidate received 7,500 votes and the
Republican received 2,500, then the Republicans would waste 2,500 votes and the
Democrats would waste 2,499 (7,500 minus 5,001, which is 50% plus one of the total
votes cast). Since the wasted votes were virtually equal, they explained, the naturally
packed Democratic votes in such districts did not have any impact on the overall
efficiency gap. This, of course, fit very well with the Plaintiffs’ overarching theory of
the gerrymander, which was that Republicans had cracked large numbers of Democrats
out of several districts in order to create many districts that now leaned Republican.
The efficiency gap, in their view, was due to this intentional cracking rather than to the
“natural” packing that exists in several heavily-Democratic districts. 16
But it is counterintuitive to believe that wasted votes would be equal in a 75-25
district, when one party wins by a landslide. Suppose the Republicans had drawn lines
designed to pack thousands of Democratic voters into new 75%-25% districts. Under
Plaintiffs’ logic, such heavily slanted districts would have no impact on the efficiency
gap, despite the explicit packing of thousands—or hundreds of thousands—of voters.
Plaintiffs never explained why a 75-25 district should be viewed as some kind of
magical “neutral” district, when in reality it could be a deliberate, and even extreme,
gerrymander, full of wasted votes. Instead of relying on a 50%-plus-one standard, it
would make much more sense to count all the wasted votes, i.e., those in excess of what
the Democrats actually required to win. In a 7,500 to 2,500 election, the Republicans still
waste all 2,500 losing votes, but the Democrats waste 4,999 votes: 7,500 minus the 2,501
they needed to win. Now, instead of pretending that the district is a wash, the
Democrats are properly counted as having wasted twice as many votes as the
Republicans, and this would serve as evidence of the gerrymander that actually
occurred.
The Plaintiffs did not argue that the Republicans had intentionally packed Democrats in Milwaukee or
elsewhere. This is likely because many of the Milwaukee districts were drawn, with the help of lawyers
and Dr. Gaddie, with an eye to Voting Rights Act concerns, and the Republican operatives who drew the
rest of the map did not touch those districts’ boundaries.
16
152
Conversely, suppose a district were drawn by a neutral party with the intent of
making it competitive, or 50-50. In such a district, one candidate will necessarily lose—
maybe only by a few votes—and yet such a result would produce massive numbers of
wasted votes (and thus inefficiency) for the loser. For example, if A wins with 5,100
votes to B’s 4,900, B has wasted 4,900 votes and A only 99—producing a colossal
efficiency gap. Under the Plaintiff’s theory, the result from a 50-50 district—a district
designed to give each side a fair chance of winning—would be the strongest evidence of a
gerrymander, despite the opposite intent. This discrepancy would seem to render the
efficiency gap, as calculated by the Plaintiffs, an unhelpful and dangerously misleading
metric for gauging actual electoral disparities. Counting all wasted votes, as described
above, would alleviate part of this problem by doubling the number of votes wasted by
winners, thus mitigating the outsized role close elections play in the Plaintiffs’ efficiency
gap analysis. Because the efficiency gap, which the Plaintiffs made the centerpiece of
their case, does not appear to adequately count wasted votes, I would find in favor of
the Defendants.
B. The Efficiency Gap is Highly Volatile
1. Volatility in General
Immediately above I have attempted to demonstrate how one side’s losses in
close elections can produce large efficiency gaps due to the fact that every vote for a
losing candidate is considered wasted. Notably, massive efficiency gaps necessarily
arise even in districts that are designed to be tossups. Given how easy it is to produce
such large gaps, it should not be surprising that efficiency gaps are volatile. The
Defendants’ expert, Professor Nicholas Goedert, credibly testified that wave elections
were relatively common, and experience teaches that in some years the Republicans did
well across the board, while Democrats performed well in others. In a good Republican
year, it will not be surprising if the GOP’s candidates win a large number of swing
elections, racking up lots of efficient victories and causing the Democrats to waste
hundreds of thousands of futile votes. In such a year, the resulting efficiency gap
would suggest a historic gerrymander, even under a perfectly neutral map. This effect is
exaggerated when the Democrats’ voters are more closely packed than the Republicans,
because then Democratic losses and wins both produce massive numbers of wasted
votes. The wins tend to be landslides, producing large numbers of surplus votes, while
the losses are close calls, resulting in huge pileups of lost votes. The Bandemer court
predicted how volatile a measure like the efficiency gap could be:
If all or most of the districts are competitive-defined by the District Court
in this case as districts in which the anticipated split in the party vote is
within the range of 45% to 55%-even a narrow statewide preference for either
153
party would produce an overwhelming majority for the winning party in the state
legislature. This consequence, however, is inherent in winner-take-all,
district-based elections, and we cannot hold that such a reapportionment
law would violate the Equal Protection Clause because the voters in the
losing party do not have representation in the legislature in proportion to
the statewide vote received by their party candidates.
478 U.S. at 130 (plurality opinion) (emphasis added).
The Supreme Court thus recognized thirty years ago that even just a “narrow
statewide preference” for a single party could produce a large majority of seats, and
thus a large efficiency gap (a 51% statewide majority could easily produce 60% of the
seats). Rather than evidence of some kind of constitutional problem, however, such a
result is simply “inherent” in the system whenever a state (1) has winner-take-all
districts and (2) experiences a “mild statewide preference” for one party. Id. This
underscores the point about question-begging: when the Plaintiffs say there is a large
efficiency gap, all they are saying is that one side won a lot of close elections in winnertake-all districts. As such, the efficiency gap appears to be of little utility in measuring
constitutional injury.
2. The 2012 Election was Historic, Nationally and in Wisconsin
In addition to these general volatility concerns, it would appear problematic to
rely on 2012—the first election after Act 43— as a benchmark for measuring wasted
votes. As the Defendants’ expert Sean Trende pointed out, President Obama was
hugely successful in a few, traditional bastions of Democratic voters—even more
successful than in 2008. But in the rest of the state, his support declined. President
Obama’s landslide wins in the Cities of Milwaukee and Madison resulted in hundreds of
thousands of wasted votes—not wasted for the President, of course, but for the downticket assembly candidates who either won in landslide victories or, more commonly,
were unopposed entirely. Many of these are wasted votes that would not otherwise
exist but for the particular attraction of Obama’s candidacy in urban areas. A brief
review of the difference in turnout for Democratic voters in a few of the Milwaukee and
Madison wards will make the point.
154
Ward 17
Obama 2012 Votes
Dem Gov. Votes 2014
Drop
Milwaukee 105
716
493
31%
Milwaukee 116
715
466
35%
Milwaukee 143
843
573
32%
Madison 1
323
264
18%
Madison 16
1,894
1,685
11%
Madison 29
2,150
2,000
7%
In this small sample of the most heavily Democratic inner-city wards in
Milwaukee (which voted over 99% for President Obama), the drop in turnout between
the presidential election and the 2014 governor’s race was about one-third, reflecting a
significantly higher level of interest in the 2012 presidential election. 18 By contrast,
Madison districts that went heavily (85-95%) for Democrats in those same years saw a
much smaller decline in turnout in 2014. In fact, the Madison wards line up with the
turnout seen in some of the staunchest Republican areas:
Ward
Romney 2012 Votes
Walker Votes 2014
Drop
Chenequa
327
288
12%
Cedar Grove
950
849
11%
Brookfield Ward 20
733
638
13%
1515
1427
6%
Oostburg
The point is that Republicans and non-Milwaukee Democrats were similarly
energized in both elections, with turnout for the 2012 presidential election somewhat
higher for both, as expected. By contrast, the numbers reflect that 2012 was a historic
year for the African-American electorate, with turnout in those wards much higher than
Election data compiled by the Government Accountability Board are available at 2012 Fall General
Election, WIS. ELECTIONS COMMISSION, http://elections.wi.gov/elections-voting/results/2012/fall-general
(last visited Nov. 3, 2016).
17
According to City of Milwaukee data, these wards are between 93% and 95% Black. CITY OF
MILWAUKEE VOTING AGE POPULATIONS BY PROPOSED ADJUSTMENTS TO 2011 VOTING WARDS (Sept.6, 2011),
http://city.milwaukee.gov/ImageLibrary/Groups/ccCouncil/2011-PDF/
2012VotingWardsDemographics9-2.pdf.
18
155
it was two years later. But historic numbers do not create a reliable benchmark by
which gerrymandering should be measured. In some districts, President Obama was
winning by landslides of 85 or 90%, resulting in large —historically large—numbers of
wasted votes that the Republicans do not match anywhere else in the state. It should
thus be clear that President Obama’s presence on the 2012 ticket exaggerates the
efficiency gap, attributing the cause to partisan bias rather than the historic urban voter
turnout that year, which gave rise to historic numbers of wasted votes for Democratic
assembly candidates. (Not surprisingly, the efficiency gap dropped in 2014.)
3. Act 10
The 2012 election also came at a dramatic time in this state’s political history.
The legislature passed Act 10 in June 2011 and the Republican governor quickly signed
it. The Act required government employees to increase their contributions to their
health insurance and retirement benefits, and significantly reduced the power of public
employee unions by abolishing mandatory membership dues and capping wage
increases to a percentage based on the consumer price index. Madison Teachers, Inc. v.
Walker, 2014 WI 99, ¶ 1, 358 Wis. 2d 1, 19, 851 N.W.2d 337, 346 (2014). Prior to the Act’s
passage, however, in an unprecedented move, all 14 Democrats in the state senate fled
to Illinois to prevent passage of the bill, preventing a Republican quorum. Eventually
the Republicans found a way around the quorum requirements and passed the bill,
which was immediately subjected to court challenges and historic protests at the
Capitol, often receiving national news coverage. Also unprecedented was the number
of state senators almost immediately targeted for recall elections. Some Democrats
were challenged for leaving the state during the Act’s consideration, while some
Republicans were targeted by those who viewed Act 10's collective bargaining changes
unfavorably. The next year, after organizers collected nearly a million signatures,
Governor Walker was subjected to his own recall election, which he survived.
Whatever one’s views of Act 10 or the responses it generated, or of President
Obama’s reelection, the point is that 2012 was hardly the kind of “normal” year one
would expect to use as a basis of reference. The experts in this action testified at some
length about the sometimes complex methods they used to ensure accuracy and
engender confidence in their models, but none of that matters if the baseline election
used in their analysis is such a historical outlier. Just as we would not rush out to buy
flood insurance after a single, historic rainstorm, we should not have much confidence
in a measure whose central data point is an unusual political year.
156
C. Wisconsin’s Political Geography
It should go without saying that urban, more Democratic, voters are more closely
packed together than suburbanites and farmsteaders, who lean more Republican but
who are interspersed with lots of Democrats nonetheless. It is undeniable that voters
may group together in the heavily Republican “collar counties” of Washington,
Ozaukee, and Waukesha, which surround Milwaukee, while Democrats in Madison or
Milwaukee often group more densely in duplexes or apartment buildings, or at least
homes with much smaller lots. There are also far more residents in Milwaukee than in
the more suburban counties. All things being equal, two individuals in Marathon
County who supported Mitt Romney are likely to be spaced farther apart than two
Barack Obama-supporting neighbors in Madison or Milwaukee. This phenomenon is
taken as a given by the Vieth court: “Consider, for example, a legislature that draws
district lines with no objectives in mind except compactness and respect for the lines of
political subdivisions. Under that system, political groups that tend to cluster (as is the
case with Democratic voters in cities) would be systematically affected by what might
be called a ‘natural’ packing effect.” 541 U.S. at 289–90 (plurality opinion). At trial,
Professor Stephanopoulos acknowledged some natural packing of Democrats, and his
own law review article acknowledges this effect as well. In addition, it is notable that
the average efficiency gap during the 1980s (under a court-drawn plan, followed by
amendments when Democrats gained full control) was 1.9% in favor of Republicans. In
the 1990s (court-drawn plan) it was 2.4%, while the average gap during the 2000s
(another court-drawn plan) was 7.6%. (ECF No. 125 at ¶¶ 190, 192, 194.) Thus,
Republican-favoring efficiency gaps have been part of Wisconsin’s political landscape
for more than three decades, long before Republicans had the ability to draw the lines in
2011.
As the Defendants pointed out at trial, the most lopsided Republican assembly
win predicted even under the Plaintiffs’ demonstration plan favored the GOP candidate
by a margin of about 75%-25%, but there were nine other districts that favored
Democrats by even more than that, with winning tallies in excess of eighty percent. (Ex.
561.) In real-world terms, in 2012 President Obama won Assembly District 16 with
more than 90% of the vote and, not surprisingly, the incumbent Democratic candidate
ran unopposed. There simply are no districts that have comparable margins for
Republicans. For example, Rep. Duey Stroebel beat the Democratic challenger in his
Ozaukee and Washington County district with 23,905 votes to 9,682, or 71% of the
votes. 19 Waukesha County’s District 99 saw Chris Kapenga win 76% of the votes.
Notably, even though heavily Republican, these districts were considered competitive
enough to draw Democratic challengers, whereas there simply are no Republican
19
2012 Fall General Election, supra note 17.
157
challengers in the more staunchly Democratic Milwaukee districts. That no one even
runs as a Republican in several districts is itself highly suggestive of geographic
packing. The Plaintiffs provided no evidence that this natural packing effect could
somehow have been avoided, since Democratic voters remain tightly packed no matter
how the lines are drawn.
It is true, as the Plaintiffs have noted, that counties like Waukesha County are
every bit as Republican as Milwaukee County or Dane County are Democratic. Voters
vote by district, however, not by county, and so the relevance of that point is unclear.
Even so, if one looked at a red-blue map, one would clearly see the heavily red areas
surrounding Milwaukee, which the Plaintiffs point to as evidence that the Republicans
are also heavily clustered. But that does not mean the numbers somehow even out.
The colors on the maps are a reflection of partisanship (intensity), not of raw numbers
of partisan voters. At trial, it was shown that the number of Obama voters in
Milwaukee County was 332,438, while Dane County had 216,071, for a total of 548,509.
(ECF No. 150 at 135.) By contrast, Mitt Romney won the heavily Republican suburban
counties with only 36,077 (Ozaukee), 162,798 (Waukesha) and 54,765 (Washington)
votes, totaling 253,640—less than half the number of Obama voters in Milwaukee and
Dane Counties. (Id.) Thus, these heavily Republican counties do not come close to
balancing out the high concentration of Democratic voters in other counties.
None of the above is to suggest that natural geographic factors explain the
entirety of the efficiency gap seen under Act 43, as the majority rightly concludes. Still,
when pro-Republican efficiency gaps have existed in neutral court-drawn plans going
back decades, and when they exist even in the Plaintiffs’ own demonstration plan,
geography cannot and should not be ignored. Even if geography does not explain the
entire gap, and even if it plays only a “modest” role—for example, three to six percent—
it would seriously undermine the notion that the Republicans in this case engaged in a
partisan gerrymander of historical proportions.
D. “Sign-Flipping” Should Not be the Standard for Court Intervention
Efficiency gaps are measured at every election, and these measures change every
election based on a number of factors, including the issues raised, quality of local
candidates, waves (as discussed above), turnout, and other natural phenomena such as
shifts in demographics. Because any challenge will be based solely on the first election
under a challenged plan, the Plaintiffs have attempted to create a standard for
measuring the durability of the gap that is observed in that first election, that is, the
tendency of an efficiency gap to persist throughout the remaining years of a plan. The
Plaintiffs’ expert, Professor Jackman, presented credible evidence that efficiency gaps
greater than 7% have a strong tendency to remain on the same side of zero over the
158
course of a plan (especially for Republicans). For example, according to Professor
Jackman, an initial efficiency gap of -10% has only a very small chance of turning
positive (“flipping signs”) over a ten-year period. The theory is that efficiency gaps of
that size invite court intervention because there is almost no chance that the gap will
disappear through the normal course of politics.
Assuming Professor Jackman’s general analysis is correct, I can perceive no
intuitive reason to believe that the likelihood of “sign-flipping” should play such an
outsized role in determining when court intervention is appropriate. Plaintiffs’
threshold of -7% is based on the fact that such an efficiency gap is unlikely to disappear
(flip signs), but this ignores the fact that the efficiency gap may become much smaller
during its natural life even if it does not disappear entirely. For example, a plan could
move from an efficiency gap of -8% to -2% in the next election cycle, meaning the map
had become almost an even playing field. Such a plan would hardly be a good
candidate for court intervention. In fact, we know that in Wisconsin, under the last
court-drawn plan, the gap jumped around between -4% and -12% (always favoring
Republicans) throughout the 2000s. That is, the gap in the highest year was more than
triple the gap in the lowest year. It is thus not difficult to envision a plan having an
initial gap of 7 or 8% that would drop down to 2 or 3% purely through natural
phenomena. And when we know that a state’s political geography explains at least
some portion of any efficiency gap, the entirety of any lingering efficiency gap could be
explained through geography rather than partisan gerrymandering. Thus, even if the
gap did not disappear entirely, any remaining gap traceable to gerrymandering has
been all but eliminated without court intervention. And yet the Plaintiffs’ test would
demand that a court intervene to fix a problem that might largely ameliorate itself
naturally.
Given the Justices’ reluctance to involve the courts in the review of
gerrymandering claims, the sign-flipping metric seems far too easy to meet, since
according to Professor Jackman every gap larger than 7% will meet that standard.
Instead of gauging the likelihood that any efficiency gap would persist, a more robust
test would demand a strong likelihood that a large efficiency gap would persist
throughout the life of the plan. That is, a court would ask whether the gerrymandering
party has created a map that will ensure a strong likelihood that large, historically
significant efficiency gaps will persist—not just that some efficiency gap will persist. If a
plaintiff could demonstrate that efficiency gaps larger than 6% or 7% would likely
persist throughout a plan’s life, judicial intervention would be more appropriate
because the minority party would have much greater difficulty remedying the problem
through the political process. Here, however, the evidence is simply that the efficiency
gap is unlikely to disappear entirely, without any acknowledgment of the possibility
that the gap could be significantly reduced without any court intervention at all.
159
Accordingly, assuming the efficiency gap played some role in a gerrymandering test, I
would require the plaintiffs to demonstrate a significant likelihood that large efficiency
gaps—not just non-zero efficiency gaps—would be a feature of the challenged plan
throughout its operative life.
VI. Conclusion
The Supreme Court heard this same story in 1986. It was unmoved. In 2004 the
Court rejected a similar claim, and the reasons the Justices cited only twelve years ago
apply with equal force now. What made this case different is the Plaintiffs’ claim that
they had discovered the holy grail of election law jurisprudence—the long sought after
“judicially discernable and manageable standard” by which political gerrymander cases
are to be decided. Yet, even the majority has declined Plaintiffs’ request that the
efficiency gap standard be adopted as the presumptive test, choosing instead to use it
merely as corroborative evidence of its own entrenchment test. Slip Op. at 86. As I
have attempted to show above, however, the majority’s entrenchment test offers no
improvement over the tests that have already been rejected by the Supreme Court. And
the efficiency gap theory on which the Plaintiffs founded their case fatally relies on
premises the courts have already rejected, including proportional representation, and it
suffers from a number of practical problems as well. Simply put, I do not believe the
Supreme Court would direct courts to meddle in a state districting plan when that plan
adequately hews to traditional and legitimate districting principles; contains no
“gerrymander,” as traditionally understood; and when the plan only modestly extends
the map-drawing party’s electoral advantage beyond what would exist naturally. This
is particularly true given that the gerrymandering party very likely would have won
both elections conducted under the challenged plan even without gerrymandering. Under
these circumstances, and given the Justices’ reluctance to review gerrymandering
claims, the Plaintiffs’ theory does not persuade me that a majority of the Supreme Court
would find an unconstitutional gerrymander in this case. Accordingly, I would find in
favor of the Defendants and therefore respectfully dissent.
/s/
WILLIAM C. GRIESBACH
District Judge
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?