IN THE UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT
United States Court of Appeals
Fifth Circuit
FILED
No. 19-60133 September 3, 2019
Lyle W. Cayce
JOSEPH THOMAS; VERNON AYERS; MELVIN LAWSON, Clerk
Plaintiffs–Appellees
v.
PHIL BRYANT, Governor of the State of Mississippi, all in the official
capacities of their own offices and in their official capacities as members of
the State Board of Election Commissioners; DELBERT HOSEMANN,
Secretary of State of the State of Mississippi, all in the official capacities of
their own offices and in their official capacities as members of the State
Board of Election Commissioners,
Defendants–Appellants.
Appeal from the United States District Court
for the Southern District of Mississippi
Before DAVIS, HIGGINSON, and WILLETT, Circuit Judges.
W. EUGENE DAVIS, Circuit Judge:
Section 2 of the Voting Rights Act of 1965, as amended in 1982, prohibits
state and political subdivisions from redistricting “in a manner which results”
in members of a protected class of racial and language minorities “hav[ing] less
opportunity than other members of the electorate to participate in the political
process and to elect representatives of their choice.” 1 Plaintiffs challenged the
legislative boundaries for Mississippi State Senate District 22, arguing that
1 52 U.S.C. § 10301(b).
No. 19-60133
the district, as drawn in 2012, diluted African-American voting strength.
Following a two-day bench trial, the district court concluded that the evidence
established a § 2 violation under the standards set forth in Thornburg v.
Gingles. 2 We AFFIRM the district court’s judgment declaring District 22
violative of § 2 of the Voting Rights Act. The State’s appeal of the district
court’s judgment granting injunctive relief is DISMISSED AS MOOT.
I.
A.
At least once every decade, the State of Mississippi redraws its
legislative boundaries based in part on the decennial census. In 2012, the
Mississippi Legislature realigned the contours of Senate District 22 to stretch
over 100 miles, from north to south. The district then consisted of all or part
of five counties in the Mississippi Delta—Bolivar, Humphreys, Sharkey,
Washington, and Yazoo—as well as parts of one non-Delta county, Madison. 3
Since early times, concentrations of African Americans have occupied the Delta
area, which has a unique history and character. 4
District 22 contains some very poor areas predominantly populated by
African Americans. Based on the 2012 redistricting plan, the black voting age
population (BVAP) in District 22 was 50.8%. In this district, almost 41.2% of
African Americans lived in poverty; their median household income was
$23,741. But the district also contained some very wealthy areas, mainly in
the southern portion of Madison County, inhabited primarily by white
residents. The poverty rate for white residents in District 22 was
approximately 8.8%; their median household income was $66,736.
2 478 U.S. 30 (1986).
3 See APPENDIX for a depiction of Mississippi Senate District 22.
4 The uniqueness of the Delta has been described in detail by our court and the district
court. See Thomas v. Bryant, 919 F.3d 298, 301 (5th Cir. 2019); Thomas v. Bryant, 366 F.
Supp. 3d 786, 790–91 (S.D. Miss. 2019).
2
No. 19-60133
Since 2003, Senate District 22 has been represented by Eugene Clark,
who is white. The candidate, whether African-American or not, preferred by
African-American voters has not occupied this office in that time. 5
B.
On July 9, 2018, three African-American voters of District 22, Joseph
Thomas, Melvin Lawson, and Vernon Ayers, filed a complaint under § 2 of the
Voting Rights Act (VRA) against Mississippi Governor Phil Bryant, Secretary
of State Delbert Hosemann, and Attorney General Jim Hood, who constitute
the State Board of Election Commissioners (collectively, the “State”). Plaintiffs
alleged that the boundary lines of District 22, as drawn under the 2012
legislative redistricting plan, diluted African-American voting strength and
deprived African-American voters of an equal opportunity to elect candidates
of their choice. According to Plaintiffs, the Mississippi Legislature’s “addition
of th[e] predominantly white areas from Madison County . . . limit[ed] the
district’s black voting age population to the present level of 50.8%, which
combine[d] with white bloc voting and lower African-American turnout . . .
dilute[d] African-American voting strength in the district.” Plaintiffs
contended that “[p]ast elections in Mississippi, particularly in District 22 and
its surrounding area, have been marked by a clear pattern of racially polarized
voting.” Consequently, as Plaintiffs averred, no African-American or African-
American-preferred candidate has represented District 22 in its current or
previous form, including in 2015, when Plaintiff Thomas ran for the seat but
lost, despite garnering a vast majority of African-American votes.
In their complaint, Plaintiffs pointed out that while the State of
Mississippi is at least 35% African-American in voting age population, only
5 The record does not contain evidence of the district lines or the election results for
any election earlier than 2003.
3
No. 19-60133
25% of the members of the Mississippi Senate are African-American. They
maintained that though African Americans made up a majority of the voting
age population in District 22, that majority lacked “real electoral opportunity”
under Supreme Court precedent interpreting § 2 of the VRA. Plaintiffs
contended that District 22 could be redrawn to cure the § 2 violation with
changes to only one or two adjacent districts; that the resulting districts would
be more compact than in the 2012 plan; and that the alteration would not
create Voting Rights Act violations in any other districts. Plaintiffs therefore
asked the district court to declare that the 2012 districting plan violated § 2 of
the VRA and enjoin the State from conducting any future elections in District
22—including the upcoming November 2019 election—under that plan.
C.
The district court conducted a bench trial from February 6–7, 2019.
Under 28 U.S.C. § 2284(a), the State had requested a district court of three
judges to hear the case, 6 but the court found that provision inapplicable to a
standalone § 2 statutory claim.
At trial, Plaintiffs provided demographic analyses and results from
elections in 2003, 2007, 2011, and 2015, which showed that Clark (and white-
preferred candidates for statewide offices, e.g., Governor and Secretary of
State) consistently beat out candidates preferred by African Americans in
District 22. Utilizing ecological inference (EI), Plaintiffs’ expert, Dr. Maxwell
Palmer, 7 highlighted consistent patterns of bloc voting by white residents and
lower election turnout of African Americans. Plaintiffs further proffered
6 The State first raised this argument in late January 2019, on the eve of trial. The
State’s answer, filed in August 2018, had acknowledged that the district court had subject-
matter jurisdiction. The State’s motion for summary judgment, filed in September 2018, did
not discuss jurisdiction.
7 Dr. Palmer, a political science professor at Boston University, was qualified by the
district court as an expert in data analysis pertaining to redistricting.
4
No. 19-60133
evidence about the bleak socioeconomic realities of African Americans in
District 22. Plaintiff Thomas, who ran for state senate in 2015, and Plaintiff
Lawson also testified about the lack of transportation options to assist African-
American voters in traveling to the polls to vote during odd-year elections.
Plaintiffs’ other expert, William Cooper, 8 presented the district court with
three plans that would, in his opinion, increase the BVAP of District 22 while
honoring traditional redistricting criteria.
On the other hand, the State’s expert, Dr. Peter Morrison, testified that
African Americans have performed well in county and municipal races within
District 22 and that African-American voters statewide have higher election
turnout than white voters. 9 The State argued that District 22, as then drawn,
was already a majority-minority district, so no relief was warranted.
Applying the familiar test set forth in Gingles, the district court
determined that District 22, as drawn in 2012, did in fact violate § 2 of the
VRA. On February 13, 2019, the district court issued an order preliminarily
advising the Mississippi Legislature of its conclusion that District 22 did not
afford African Americans “an equal opportunity to participate in the political
processes and to elect candidates of their choice.” The district court noted that
Plaintiffs had presented three different plans for redrawing District 22 that
would remedy the § 2 violation, but the court acknowledged that the
Mississippi Legislature was entitled to the first opportunity to redraw the
district.
On February 16, 2019, the district court issued a lengthy opinion,
denying the State’s laches defense and concluding that Plaintiffs had proven
8 Mr. Cooper is a geographic information system (GIS) consultant who conducts
computer mapping for redistricting plans; he was qualified as an expert in that field.
9 Dr. Morrison, an applied demographer retired from the Rand Corporation, was
qualified by the district court as an expert in that field.
5
No. 19-60133
by a preponderance of the evidence that the boundaries of Mississippi Senate
District 22 violated § 2. 10 The court, however, declined to order any specific
relief “while the Mississippi Legislature consider[ed] whether to redraw the
District and extend the qualification deadline.” 11
The State thereafter appealed the district court’s order and filed a
motion for stay pending appeal in this court pursuant to 28 U.S.C. § 1292(a)(1),
which provides appellate jurisdiction over a district court’s interlocutory order
granting an injunction.
On February 21, 2019, the district court conducted a teleconference,
joined by Secretary of State Hosemann; the parties “convinced the Court that
no further hearings were necessary” regarding remedies. 12 On February 26,
2019, the State further informed the district court that the Mississippi
Legislature “would take up the Court’s suggestion to redraw District 22 only if
the motions for stay were denied [at the district court] and in the Fifth
Circuit.” 13 On that same day, however, noting that the Mississippi Legislature
had not redrawn the boundaries of District 22, the district court granted
Plaintiffs’ motion to extend the qualification deadline from March 1, 2019 to
March 15, 2019. The district court also redrew the boundaries of District 22.
Specifically, the district court adopted Plaintiffs’ “Illustrative Plan 1,” which
affected the boundaries of Districts 22 and 23. Finally, the district court issued
a final judgment “resolv[ing] all of the claims and defenses in th[e] case” and
rendering judgment in favor of Plaintiffs and against the State. 14
10 Thomas, 366 F. Supp. 3d at 810.
11 Id.
12 ECF Dkt. No. 85 at 2.
13 Id.
14 Because the district court’s final judgment on February 26, 2019 varied from its
interlocutory order, we dismissed the appeal of the interlocutory order as moot. Thomas v.
Bryant, 755 F. App’x 421 (5th Cir. 2019).
6
No. 19-60133
D.
On February 27, 2019, the State filed a notice of appeal of the district
court’s final judgment. 15 On March 8, 2019, the State filed in this court an
emergency motion for stay of judgment. On March 15, 2019, a divided motions
panel of this court granted in part and denied in part the State’s motion and
issued an opinion one week later providing its reasoning. 16 The panel first
determined that the State did not have a strong likelihood of overturning the
district court’s judgment. Nevertheless, noting that the Mississippi
Legislature should be afforded an additional opportunity to redraw District 22
and comply with the Voting Rights Act, the panel issued a temporary stay of
the district court’s judgment until April 3, 2019, and extended the candidate
filing deadline for any affected districts until April 12, 2019.
E.
On March 26, 2019, the Mississippi Senate exercised its prerogative and
redrew the legislative map, which the Mississippi House of Representatives
approved the next day. The resulting legislation adopted new district
boundaries—different from those of the court-mandated plan—affecting
Senate Districts 22 and 13. The legislation provides:
[T]his resolution shall take effect and be in force from and after its
passage; however, in the event that the appellants prevail in the
appeal of the case of Joseph Thomas, et al. v. Phil Bryant, et al.,
No. 19-60133 (5th Cir. Mar. 15, 2019), this resolution shall be
repealed and the districts as originally configured in Chapter 2234,
Laws of 2012, shall take effect. 17
15 Attorney General Jim Hood did not appeal the district court’s judgment and
consequently is not a party to this appeal.
16 See Thomas, 919 F.3d at 298.
17 MISS. LEGIS., J.R. 202 (2019).
7
No. 19-60133
On April 4, 2019, Plaintiffs informed the district court that they would
not challenge the State’s newly drawn plan; thus, this is the operative plan for
the 2019 state senate elections.
F.
We held oral argument in this matter on June 11, 2019. On August 1,
2019, we announced our holding, noting that detailed opinions would follow.
Because we affirmed the district court’s judgment, the Legislature’s plan was
used in the primary election on August 6, 2019. The general election is
scheduled for November 5, 2019.
II.
Before addressing the merits of Plaintiffs’ § 2 claim, we must examine de
novo whether we have jurisdiction on appeal and whether the district court
had jurisdiction to hear this case with a single judge notwithstanding the
three-judge provision contained in 28 U.S.C. § 2284(a). We discuss each in
turn.
A.
Article III, section 2 of the Constitution confines federal courts to the
decision of “Cases” or “Controversies.” 18 “[A]n actual controversy must be
extant at all stages of review, not merely at the time the complaint is filed.” 19
A case becomes moot “if an event occurs while a case is pending on appeal that
makes it impossible for the court to grant ‘any effectual relief what[so]ever’ to
a prevailing party.” 20 A legislative remedy to a challenged law may moot a
case pending appeal because courts can no longer enjoin the enforcement of a
18U.S. CONST. art. III, § 2.
19Arizonans for Official English v. Arizona, 520 U.S. 43, 67 (1997) (quoting Preiser v.
Newkirk, 422 U.S. 395, 401 (1975)).
20 Church of Scientology of Cal. v. United States, 506 U.S. 9, 12 (1992) (quoting Mills
v. Green, 159 U.S. 651, 653 (1895)).
8
No. 19-60133
repealed law that has no effect. 21 A case is moot when “(1) it can be said with
assurance that there is no reasonable expectation that the alleged violation
will recur, and (2) interim relief or events have completely and irrevocably
eradicated the effects of the alleged violation.” 22
Prompted by our request for supplemental briefing on mootness,
Plaintiffs assert that the State’s appeal is moot because the Mississippi
Legislature exercised its prerogative and adopted new boundaries for District
22—different from the district court’s mandated electoral plan—to be used for
the upcoming election. Plaintiffs do not challenge the Legislature’s newly
reconfigured boundaries for District 22. So, according to Plaintiffs, the
Legislature’s amendment to the 2012 redistricting plan mooted this case while
our review was pending. We disagree.
We find that the Legislature’s revised boundaries for District 22 did not
moot the State’s appeal because the State asserts that it will restore the
challenged boundaries if it “prevails” and we reverse the district court’s
judgment. We have previously declined to find that a legislature’s responsive
fix to an election statute mooted the appeal of the district court ruling that
triggered the fix. 23 When the Mississippi Legislature redrew its district
boundaries in March 2019, it specifically provided in the legislation that the
amendments for Districts 22 and 13 “shall be repealed and the districts as
originally configured” shall be used if the State prevails on this appeal. 24 So
21 See Fantasy Ranch Inc. v. City of Arlington, 459 F.3d 546, 564 (5th Cir. 2006)
(citation omitted) (“[S]tatutory changes that discontinue a challenged practice are ‘usually
enough to render a case moot, even if the legislature possesses the power to reenact the
statute after the lawsuit is dismissed.’”) (quoting Valero Terrestrial Corp. v. Paige, 211 F.3d
112, 116 (4th Cir. 2000)).
22 Los Angeles Cty. v. Davis, 440 U.S. 625, 631 (1979) (internal quotation marks
omitted) (citations omitted).
23 See, e.g., Veasey v. Abbott, 888 F.3d 792, 799 (5th Cir. 2018); Miss. State Chapter,
Operation Push v. Mabus, 932 F.2d 400, 404, 409 (5th Cir. 1991).
24 MISS. LEGIS., J.R. 202.
9
No. 19-60133
even though the primary election has passed, the State maintains that, if it
prevails on appeal, it could then revert to using its original map and conduct
special elections for Senate Districts 22 and 13. Because the State could gain
effectual relief as a result of this appeal, we hold that this case is not moot with
respect to the district court’s declaratory judgment; 25 appellate jurisdiction
therefore exists.
B.
1.
We next determine whether 28 U.S.C. § 2284(a), as amended in 1976,
required the district court to convene a three-judge panel to decide the merits
of this case. 26 “When the plain language of a statute is unambiguous and does
not ‘lead[] to an absurd result,’ ‘our inquiry begins and ends with the plain
meaning of that language.’” 27
Section 2284(a) provides: “A district court of three judges shall be
convened when otherwise required by Act of Congress, or when an action is
filed challenging the constitutionality of the apportionment of congressional
districts or the apportionment of any statewide legislative body.” 28 The
question presented is whether “constitutionality of” modifies only “the
apportionment of congressional districts” (as the State argues) or whether it
also modifies “the apportionment of any statewide legislative body” (as
Plaintiffs assert). In other words, pertinent here, the question is whether
25 As explained infra at SECTION V, the State’s appeal is moot with respect to the
district court’s order selecting and adopting a remedial plan because the Legislature’s
enactment of a new plan superseded the district court’s order.
26 See 21st Mortg. Corp. v. Glenn (In re Glenn), 900 F.3d 187, 189 (5th Cir. 2018)
(holding questions of statutory interpretation are reviewed de novo).
27 Sealed Appellee 1 v. Sealed Appellant 1, 767 F.3d 418, 421 (5th Cir. 2013) (alteration
in original) (quoting United States v. Dison, 573 F.3d 204, 207 (5th Cir. 2009)); Duncan v.
Walker, 533 U.S. 167, 172 (2001).
28 28 U.S.C. § 2284(a).
10
No. 19-60133
§ 2284(a) requires a “district court of three judges” only when a plaintiff brings
a constitutional—as opposed to a statutory—challenge to state legislative
apportionment.
The text of § 2284(a) is clear, so we apply the statute as written. 29 We
read § 2284(a) to require a district court to convene as a three-judge district
court “in cases challenging the constitutionality of state or federal legislative
apportionment.” 30 Leading civil procedure scholars share this view. 31 Indeed,
for decades, district courts nationwide have routinely handled § 2 challenges
to state legislative maps as single-judge matters—and circuit courts have
reviewed their work without a hint of jurisdictional doubt. 32 This reflects the
settled understanding of § 2284(a) that has prevailed since its enactment. 33
Moreover, in Harris v. Arizona Independent Redistricting Commission, a
29 Cf. Lamie v. U.S. Tr., 540 U.S. 526, 534 (2004) (“The statute is awkward, and even
ungrammatical; but that does not make it ambiguous . . . .”).
30 22 JAMES WM. MOORE ET AL., MOORE’S FEDERAL PRACTICE § 404.02[1] (3d ed. 1997)
(emphasis added).
31 See id.; 17 CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE
§ 4235 (3d ed. 2007) (“[A] three-judge court for reapportionment cases . . . [is convened for]
all federal constitutional challenges that could result in a reapportionment.” (emphasis
added)).
32 See, e.g., Bone Shirt v. Hazeltine, 336 F. Supp. 2d 976 (D.S.D. 2004), aff’d, 461 F.3d
1011 (8th Cir. 2006) (South Dakota House and Senate); Metts v. Almond, 217 F. Supp. 2d 252
(D.R.I. 2002), rev’d en banc sub nom. Metts v. Murphy, 363 F.3d 8 (1st Cir. 2004) (Rhode
Island Senate); Old Person v. Brown, 182 F. Supp. 2d 1002 (D. Mont. 2002), aff’d, 312 F.3d
1036 (9th Cir. 2002) (Montana House and Senate); Rural W. Tenn. African Am. Affairs
Council v. Sundquist, 29 F. Supp. 2d 448 (W.D. Tenn. 1998), aff’d, 209 F.3d 835 (6th Cir.
2000) (Tennessee House); Sanchez v. Colo., 861 F. Supp. 1516 (D. Colo. 1994), rev’d, 97 F.3d
1303 (10th Cir. 1996) (Colorado House); Dickinson v. Ind. State Election Bd., 740 F. Supp.
1376 (S.D. Ind. 1990), rev’d in part, vacated in part, 933 F.2d 497 (7th Cir. 1991) (Indiana
House); Mirrione v. Anderson, 717 F.2d 743 (2d Cir. 1983) (New York Assembly).
33 See, e.g., Cavanagh v. Brock, 577 F. Supp. 176, 180 n.3 (E.D.N.C. 1983) (three-judge
court) (“A three-judge court is mandated pursuant to § 2284 when an action challenges the
apportionment of a statewide legislative body on substantial federal constitutional grounds.”
(emphasis added)); Sullivan v. Crowell, 444 F. Supp. 606, 615 n.6 (W.D. Tenn. 1978) (three-
judge court) (“Congress did not intend to require a three-judge court in the state legislative
apportionment context except where a constitutional challenge is made to a statute involving
‘the apportionment of (a) statewide legislative body.’” (emphasis added)).
11
No. 19-60133
unanimous Supreme Court noted in a parenthetical explanation that § 2284(a)
“provid[es] for the convention of [a three-judge district court] whenever an
action is filed challenging the constitutionality of apportionment of legislative
districts.” 34 Both courts and preeminent treatises support this longstanding
understanding of § 2284(a). The State identifies no caselaw or commentary to
the contrary. 35
Furthermore, the “series modifier” canon of construction, even if
applicable, confirms our reading of § 2284(a). A modifier ordinarily applies to
an entire series of parallel terms. 36 So naturally, here, “constitutionality of”
applies to both “the apportionment of congressional districts” and “the
apportionment of any statewide legislative body.” The State, supported by an
amicus brief, argues the “the” coming before “apportionment of any statewide
34 136 S. Ct. 1301, 1306 (2016) (emphasis added). The Supreme Court offered this
parenthetical explanation in the context of a constitutional challenge to a state legislative
map, in the course of summarizing the case’s procedural history. See id. A three-judge court
had been convened because the case involved a Fourteenth Amendment challenge to
Arizona’s legislative districts. Id. Though this parenthetical explanation is dicta, the
Supreme Court’s view confirms our conclusion that a straightforward reading of § 2284(a)
requires a three-judge district court only in constitutional challenges to the apportionment
of a state legislative district.
35 Page v. Bartels, 248 F.3d 175 (3d Cir. 2001), though cited by the State, hurts its
argument. In Page, a case involving both constitutional and § 2 challenges to the New Jersey
legislature’s apportionment, the Third Circuit considered whether a single district judge
erred by ruling on a request for preliminary relief on the statutory claim before convening
the three-judge court. Id. at 181–82. After careful analysis of the “inextricably intertwined”
statutory and constitutional claims, the court concluded that the single judge should have
left the matter of preliminary relief to the three-judge court. Id. at 190. This careful analysis
was necessary, however, only because the Third Circuit viewed the § 2 claim as outside the
scope of § 2284(a). Id. at 181 (describing § 2284 as “a special procedural mechanism for
constitutional challenges to statewide legislative apportionment schemes” (emphasis added)).
If the § 2 claim were within the scope of § 2284(a) and required a three-judge district court,
the question would have been easy for the Third Circuit; the single judge’s action on
preliminary relief would have been clearly wrong.
36 See ANTONIN SCALIA & BRYAN GARNER, READING LAW: THE INTERPRETATION OF
LEGAL TEXTS 147 (2012).
12
No. 19-60133
legislative body” interrupts the series modifier. 37 It asserts that this
“determiner”—the “the” in § 2284(a)—breaks off the modifier to the
“apportionment of any statewide legislative body.” So the State concludes that
a district court of three judges is required for all claims challenging state
legislative seats. We disagree. The modifying clause in § 2284(a) “is applicable
as much to the first . . . as [it is] to the last.” 38 As a leading statutory
interpretation guide recognizes, “[p]erhaps more than most of the other canons,
[the series modifier canon] is highly sensitive to context.” 39 It makes little
sense to say that three judges are required to hear statutory claims challenging
state legislative redistricting but not statutory claims challenging
congressional redistricting. 40 The text of § 2284(a) gives no indication that
Congress was “hid[ing] elephants in mouseholes.” 41 Accordingly, we hold that
the natural reading of § 2284(a) requires that “constitutionality of” be applied
to both “the apportionment of congressional districts” and “the apportionment
of any statewide legislative body.” 42
37Amicus Judicial Watch also argues that “the presence of a determiner before the
phrase ‘any statewide legislative body’ in § 2284(a) means that the earlier phrase ‘the
constitutionality of’ was not meant to modify that second phrase.”
38 See Paroline v. United States, 572 U.S. 434, 447 (2014) (quoting Porto Rico Railway,
Light & Power Co. v. Mor, 253 U.S. 345, 348 (1920)).
39 See SCALIA & GARNER, supra note 38, at 150.
40 To further illustrate, we note, as an example, that the judicial review section of the
Bipartisan Campaign Reform Act of 2002 (BCRA) provides:
If any action is brought for declaratory or injunctive relief to challenge the
constitutionality of any provision of this Act or any amendment made by this
Act, the following rules shall apply: (1) The action shall be filed in the United
States District Court for the District of Columbia and shall be heard by a 3-
judge court convened pursuant to section 2284 of title 28, United States Code.
52 U.S.C. § 30110(a). Under the State’s interpretation and application of the “series modifier”
canon, a three-judge district court would be required for statutory claims challenging “any
amendment” to the BCRA, but not to “any provision” of the BCRA itself. Such a reading, and
application thereof, would be absurd.
41 Whitman v. Am. Trucking Ass’n, 531 U.S. 457, 468 (2001).
42 Because this statute is unambiguous, we need not address the State’s legislative
history argument. See United States v. Key, 599 F.3d 469, 480 (5th Cir. 2010) (noting
“repairing to legislative history would be unnecessary and improper” if statute is clear).
13
No. 19-60133
2.
In the alternative, the State argues that a § 2 challenge is a
constitutional one because Congress passed the VRA to enforce the Fifteenth
Amendment. 43 This argument is unavailing. As the Supreme Court noted in
Gingles, Congress amended the VRA in 1982 to reject the Court’s prior
determination in City of Mobile v. Bolden. 44 The Bolden Court held that the
prior version of § 2 “was intended to have an effect no different from that of the
Fifteenth Amendment itself,” such that proof of intentional discrimination was
required to prove a § 2 vote dilution claim. 45 But Congress disagreed—
amending the VRA to provide a scheme for plaintiffs to establish a § 2 case by
“showing discriminatory effect alone.” 46 Henceforth, constitutional challenges
under, inter alia, the Fifteenth Amendment and statutory challenges under § 2
adhere to different legal precedents and evidentiary standards. 47 Indeed, as
noted, courts have repeatedly treated § 2 claims as non-constitutional
challenges. 48
43 The Fifteenth Amendment provides:
Section 1. The right of citizens of the United States to vote shall not be
denied or abridged by the United States or by any State on account of race,
color, or previous condition of servitude.
Section 2. The Congress shall have power to enforce this article by
appropriate legislation.
U.S. CONST. amend. XV.
44 478 U.S. at 35 (citing City of Mobile v. Bolden, 446 U.S. 55 (1980)).
45 Bolden, 446 U.S. at 61–62.
46 Gingles, 478 U.S. at 47; accord Dickinson v. Ind. State Election Bd., 933 F.2d 497,
502 (7th Cir. 1991) (“[P]laintiffs here base their claim on [§ 2] alone, not on constitutional
principles, implying that the district court’s reliance on pre-1982 cases or those dealing with
constitutional challenges to apportionment schemes is less relevant.”).
47 See, e.g., City of Carrollton Branch of the NAACP v. Stallings, 829 F.2d 1547, 1553
(11th Cir. 1987) (“Proof that a contested electoral practice or mechanism was adopted or
maintained with the intent to discriminate against minority voters . . . is not required under
Section 2 of the Voting Rights Act.”); see also League of United Latin Am. Citizens v. Perry
(“LULAC”), 548 U.S. 399, 425, 442 (2006) (declining to apply First Amendment and equal
protection standards to a district because it had already been found to violate § 2).
48 See supra notes 32-33.
14
No. 19-60133
The State’s theory has a more basic problem. Every valid congressional
enactment is based on some provision of the Constitution. Under the State’s
theory, federal criminal law and tax law—typically enacted via Congress’s
commerce, taxing, and necessary and proper powers—would all be
constitutional matters. We find no legal or logical support for the State’s
position.
Because § 2284(a) does not apply, and this suit only concerns a § 2
statutory challenge, we hold that the district court, sitting as a single judge,
had proper authority to decide this case. 49
III.
We now consider the State’s laches defense. The State contends that this
lawsuit comes too late—the challenged map was enacted in 2012, Plaintiff
Thomas ran and lost the state senate race for District 22 in 2015, and the three
plaintiffs here did not commence legal action until July 2018. The State
asserts that Plaintiffs inexcusably delayed the filing of this suit, and the State
contends that it suffered undue prejudice. It argues that the district court
therefore erred by not dismissing this case based on laches. We review the
district court’s ruling on laches for abuse of discretion and its findings of fact
underpinning that ruling for clear error. 50
Laches “is a defense developed by courts of equity,” 51 which applies
where a plaintiff “unreasonably delays in filing a suit and as a result harms
the defendant.” 52 To successfully establish a laches defense, the defendant
49 Plaintiffs argue that even if § 2284(a) applies, the State has waived it because the
State did not request a three-judge panel until two weeks before trial. The State, however,
asserts that § 2284(a) is jurisdictional and thus cannot be waived. Because we conclude that
§ 2284(a) does not apply in this case, we need not decide whether this statute poses a
jurisdictional rule.
50 Retractable Techs., Inc. v. Becton Dickinson & Co., 842 F.3d 883, 900 (5th Cir. 2016).
51 Petrella v. Metro-Goldwyn-Mayer, Inc., 134 S. Ct. 1962, 1973 (2014).
52 Nat’l R.R. Passenger Corp. v. Morgan, 536 U.S. 101, 121 (2002).
15
No. 19-60133
must prove “(1) a delay asserting a right or claim; (2) that the delay was
inexcusable; [and] (3) that undue prejudice resulted from the delay.” 53 As set
forth below, we hold that the district court did not abuse its discretion in
rejecting the State’s laches defense.
A.
We first consider whether Plaintiffs unreasonably and inexcusably
delayed in filing this suit. The State has proffered two different time periods
in which it alleges Plaintiffs should have commenced legal action. In the
district court, the State argued that Plaintiffs’ claim accrued in September
2012 when the U.S. Department of Justice granted preclearance to the State’s
redistricting plan under § 5 of the VRA. The State argued that Mississippi’s
general three-year “catch-all” statute of limitations 54 applied and thus barred
Plaintiffs’ action because they delayed this action by six years. But now, on
appeal, the State argues that Plaintiffs’ claim accrued with the November 2015
election in District 22, when, according to the State, “all of the necessary facts
were in place.” The State concedes that “[t]here is no question that if this suit
had been brought in 2015, when all the facts necessary to plaintiffs’ case were
known, orderly review and orderly deliberation could have taken place.”
As a threshold matter, “[t]he trial court cannot have erred as to matters
which were not presented to it.” 55 Because the State’s laches argument at
summary judgment identified September 2012 as the accrual date, we cannot
53 Retractable Techs., 842 F.3d at 899–900 (alteration in original) (quoting Elvis
Presley Enters., Inc. v. Capece, 141 F.3d 188, 205 (5th Cir. 1998)); see also City of El Paso v.
El Paso Entm’t, Inc., 382 F. App’x 361, 366 (5th Cir. 2010) (holding the party asserting a
laches defense bears the burden of proving it).
54 Mississippi’s “catch-all” statute of limitation provides: “All actions for which no
other period of limitation is prescribed shall be commenced within three (3) years next after
the cause of such action accrued, and not after.” MISS. CODE ANN. § 15-1-49(1).
55 Gabel v. Lynaugh, 835 F.2d 124, 125 (5th Cir. 1988).
16
No. 19-60133
fault the district court for not ruling on “arguments not presented below.” 56
We agree, nonetheless, with the State’s argument on appeal that the Plaintiffs’
clock, for laches purposes, began to run after the November 2015 election.
“Although laches is an equitable defense, it is usually applied with
reference to the limitations period for the analogous action at law, which may
be state law if no federal limitation law exists.” 57 In the district court, the
State argued that because the VRA does not contain a limitations period, the
general three-year “catch-all” limitations period of Mississippi Code § 15-1-49
should bar Plaintiffs’ § 2 claim. We use the state statute of limitations here as
a reference point. 58 The first time Plaintiffs should have known of their claim
is in November 2015. Significantly, the 2015 election was the first time the
challenged map was implemented and used for District 22. To Plaintiffs,
Thomas’s loss confirmed their basis for an actionable § 2 claim; it bolstered
their case showing a Voting Rights Act violation. It follows that under
Mississippi’s three-year “catch-all” period, Plaintiffs’ suit—filed in July 2018—
would be considered timely. 59
Furthermore, we note that resolving a laches defense is a fact-specific
inquiry, and the State points to no evidence that Plaintiffs unreasonably
delayed in filing this suit given the three-year analogous prescriptive period.
For election cases, circuit courts have generally approved laches defenses only
in cases filed much closer in time to the general election than this case. 60
56 McDonald v. Bd. of Miss. Levee Comm’rs, 832 F.2d 901, 909 (5th Cir. 1987).
57 Retractable Techs., 842 F.3d at 900 (citation omitted).
58 The district court also used the three-year “catch-all” limitations period as
reference, which the State does not challenge on appeal.
59 Cf. Holman v. Gulf Ref. Co. of La., 76 F.2d 94, 99 (5th Cir. 1935) (“[L]aches is not a
mere matter of the lapse of time, but rather of neglect and delay that is hurtful to the other
party and to a fair trial, because of which the extraordinary powers of equity will remain
passive. It may exist though no statute of limitation bars.”).
60 See, e.g., Dickinson, 933 F.2d at 499–502 (suit filed eight months before general
election); Simkins v. Gressette, 631 F.2d 287, 289 (4th Cir. 1980) (suit filed two days before
17
No. 19-60133
Moreover, one circuit court has rejected a laches defense to § 2 relief ordered
closer in time to the general election than the relief ordered here. 61 The State
cites no binding Supreme Court authority on laches. Its lone circuit authority
is White v. Daniel, a Fourth Circuit case in which plaintiffs challenged the map
for a Virginia county’s board of supervisors. 62 That case is easily
distinguishable. There, the map was drawn in 1971 and retained after the
1980 census. 63 The last election under the map before the 1990 census was in
1987. 64 The plaintiffs sued after that election, in September 1988. 65 The
plaintiffs’ suit thus came too late to affect any election under the map and
would have saddled the county with two redistricting processes in a short
period. 66
Here, by contrast, Plaintiffs filed their complaint well before the
elections: sixteen months before the general election and thirteen months
before the primary election. To our knowledge, no circuit court has disturbed
a district court’s denial of a laches defense in any case where the suit was filed
more than eight months before the election. Consequently, if we were to
reverse the district court based on the doctrine of laches, we would be the
circuit outlier. Accordingly, we find that the district court acted within its
discretion. 67
the start of the candidate-qualifying period and eight months before the general election);
Md. Citizens for a Representative Gen. Assembly v. Governor of Md., 429 F.3d 606, 609 (4th
Cir. 1970) (suit filed thirteen weeks before filing deadline, five months before primaries, and
seven months before general election).
61 See Garza v. Cty. of Los Angeles, 756 F. Supp. 1298, 1303 (C.D. Cal. 1990), aff’d in
relevant part, 918 F.2d 763, 772 (9th Cir. 1990) (rejecting laches in suit that went to trial six
months before general election and resulted in relief four months before election).
62 909 F.2d 99, 100–02 (4th Cir. 1990).
63 Id. at 101.
64 Id. at 103–04.
65 Id.
66 Id. at 104.
67 We understand that changes to upcoming elections will inevitably cause disruption.
Nonetheless, we do not find that the Legislature’s operative plan—adopted over seven
18
No. 19-60133
B.
Laches is not a doctrine concerned solely with timing. A defendant must
also establish undue prejudice as an independent element of its defense. 68
“Measuring prejudice entails balancing equities.” 69 Plaintiffs understandably
consider their case of great importance, and it carries significant public
implications. The State, however, argues that Plaintiffs’ suit injects confusion
and uncertainty into the November 2019 election. It argues that candidates
do not know where to campaign or fundraise, and voters do not know their
electoral options. Moreover, the State complains of the tight deadline the
district court imposed for the Legislature to create a remedy.
months before the general election for two of fifty-two districts—is unduly late or disruptive.
For instance, in Reynolds v. Sims, the epochal decision establishing the “one-person, one-
vote” doctrine, the Supreme Court reviewed a district court decision issued in late-July 1962
that set a new map for use in Alabama’s elections later that year. 377 U.S. 533 (1964). The
Supreme Court endorsed that action as “an appropriate and well-considered exercise of
judicial power,” done “in a most proper and commendable manner.” Id. at 586–87. The
dissent approvingly cites Reynolds’s guidance on the timing of voting remedies without
acknowledging that the district court there revised a state’s entire legislative map four
months closer to the general election than the two-district change imposed here.
Purcell v. Gonzalez is an instructive contrast. 549 U.S. 1 (2006). There, the suit was
filed in May 2006 challenging an Arizona law governing voter registration and identification.
Id. at 2–3. In September 2006, the district court declined to issue a preliminary injunction
affecting the election that November. Id. at 3. A motions panel of the Ninth Circuit then
stayed the law without oral argument or explanation on October 5, affecting the election
starting a few weeks later. Id. The Supreme Court then stepped in on October 20 and lifted
the Ninth Circuit’s injunction, citing the need to avoid disrupting the looming election. Id.
at 4–5. The dissent’s favored cases on the timing of election remedies all featured comparably
truncated timetables. See Husted v. Ohio State Conf. of N.A.A.C.P., 573 U.S. 988 (2014)
(Mem.); North Carolina v. League of Women Voters of N.C., 135 S. Ct. 6 (2014) (Mem.); Veasey
v. Perry, 769 F.3d 890, 892 (5th Cir. 2014).
Admittedly not laches cases, Reynolds and Purcell nonetheless provide some useful
guidance to lower courts as we decide what actions to take with an election nearing: Late-
July relief was permissible in Reynolds; October relief was not in Purcell. This indicates, to
us, that the Legislature’s March relief, in response to the court’s order, is not unduly
disruptive.
68 Envtl. Def. Fund, Inc. v. Alexander, 614 F.2d 474, 479 (5th Cir. 1980) (treating
prejudice as “prerequisite” to laches, distinct from delay).
69 Id.
19
No. 19-60133
For purposes of a laches defense, forms of prejudice generally cognizable
by courts in the election context include the burden of redistricting twice in a
short span of time—that is, when a court orders redistricting shortly before the
regular, decennial redistricting 70—or upsetting the results of an election
already conducted (e.g., a corresponding primary election). 71 The State
principally relies on the cases discussing these sources of prejudice, but neither
scenario applies. In this case, Plaintiffs filed their voting rights case sixteen
months before the general election and thirteen months before the primaries;
the parties had until January 18, 2019 (over six months) to conduct discovery;
the district court conducted a bench trial in February 2019, nine months before
the general election. Moreover, this suit was filed four years before 2022, the
Mississippi Legislature’s deadline to redistrict after the 2020 census. 72
We are satisfied that the district court did not clearly err in finding no
undue prejudice. 73 Following the trial and the district court’s judgment, the
70 See, e.g., White, 909 F.2d at 104; Fouts v. Harris, 88 F. Supp. 2d 1351, 1354–55 (S.D.
Fla. 1999).
71 See, e.g., Lopez v. Hale Cty., 797 F. Supp. 547, 550 (N.D. Tex. 1992) (three-judge
court) (addressing suit filed after primary elections had already occurred under challenged
plan).
72 See MISS. CONST. art. XIII, § 254; Miss. State Conf. of NAACP v. Barbour, 2011 WL
1870222, at *7 (S.D. Miss. May 16, 2011), aff’d, 565 U.S. 972 (2011).
73 The State makes the conclusory argument that the expedited timetable impaired
its ability to litigate effectively. But it points to no specific problem such as inadequate time
to conduct discovery or hire experts. Additionally, we have held that in making scheduling
decisions, a court’s “judgment range is exceedingly wide,” for it “must consider not only the
facts of the particular case but also all of the demands on counsel’s time and the court’s.” HC
Gun & Knife Shows, Inc. v. City of Houston, 201 F.3d 544, 549–50 (5th Cir. 2000).
Recognizing the upcoming election timeline, the district court did not abuse its discretion in
expediting its administration of this case. See id. (noting “[a] trial judge’s control of discovery
is granted great deference” and “[w]e will reverse a discovery ruling only if it is ‘arbitrary or
clearly unreasonable.’” (citations omitted)).
Even if this litigation timeline were relevant to laches, it is unpersuasive. The State
had sufficient time to prepare this case. As previously noted, this lawsuit was filed sixteen
months before the November 2019 election, and the district court conducted the bench trial
in February 2019. Here, both parties had enough time for discovery and filed numerous
motions. Plaintiffs’ experts, Dr. Palmer and Mr. Cooper, completed their reports on
20
No. 19-60133
Legislature was able to redraw District 22. Under the state constitution, the
Mississippi Legislature meets no more than ninety days annually. 74 This
year’s legislative session started on January 8, 2019 and concluded on March
29, 2019. 75 On February 13, 2019, six days after the bench trial concluded, the
district court “advise[d] the Mississippi Legislature that the evidence supports
the plaintiffs’ allegations” and “[a]s [then] drawn, District 22 [did] not afford
the plaintiffs ‘an equal opportunity to participate in the political processes and
to elect candidates of their choice.’” At that point, the Legislature had fifty-
three days left of its regular session. When the district court issued its findings
of fact and conclusions of law on February 16, it invited the Legislature to
redraw District 22 if the State wished to do so. The district court correctly
recognized that a “legislative plan is unequivocally to be preferred over a court-
ordered plan.” 76
On February 21, 2019, the district court held a teleconference, personally
joined by Secretary of State Hosemann. In that meeting, based on the district
court record, the parties “convinced the Court that no further hearings were
necessary.” 77
December 10, 2018 and January 18, 2019, respectively. The State’s expert, Dr. Morrison,
also completed his report on January 18, 2019. The State had about two months to review
Dr. Palmer’s analysis and about three weeks to review Mr. Cooper’s findings. Though the
State complains that it only learned several days before trial that Dr. Palmer’s analysis did
not capture 2,000 voters who mistakenly voted outside the district in 2015, the State properly
cross-examined him about this point at trial. Based on our careful review of the record, this
was a well-tried case. The district court promptly decided this case and issued its findings of
fact and conclusions of law on February 16, 2019.
74 MISS. CONST. art. IV, § 36.
75 Under the Mississippi Constitution, the Legislature could have stayed in session
until April 7, 2019, without two-thirds approval from both legislative chambers. See id.
76 See Operation Push, 932 F.2d at 406 (citations omitted) (“In the reapportionment
cases, courts clearly defer to the legislature in the first instance to undertake remedies for
violations of § 2.”).
77 Given this teleconference, the dissent’s criticism of the district court for not
conducting a “sufficient” evidentiary hearing on remedies is unfounded. See Dissenting Op.
at 4. See SECTION V(B) for discussion on the district court’s remedial plan.
21
No. 19-60133
Then, on February 26, 2019, the State informed the district court that
the Mississippi Legislature intended to redraw District 22 “should the stay
motions pending before [the district court] and the Fifth Circuit be denied.”
But recognizing the fast-approaching election qualification deadline of March
1, the district court implemented a remedial plan for District 22 on February
26, 2019, and extended the qualification deadline to March 15, 2019. A motions
panel of this Court later stayed the court-mandated plan until April 3, 2019,
and provided the Legislature with another opportunity to redraw the
boundaries for District 22. The Legislature then acted, and created new lines
for Districts 22 and 13 on March 27, 2019, which Plaintiffs do not challenge.
In short, the Legislature was able to redistrict District 22. It was
unnecessary for the State, for instance, to extend the legislative calendar or
call a special session to accomplish this task. This suit was about one district
out of fifty-two in the Senate, and only one other district had to be changed in
producing the new map. At base, just eight precincts were shifted. We find
unpersuasive the State’s claim of clear error as to undue prejudice here.
Finally, we consider any prejudice to the candidates and voters.
According to Plaintiffs, 78 the Legislature’s new plan moves only five Bolivar
County precincts previously in Senate District 22 to Senate District 13 and
shifts three Sunflower County precincts from Senate District 13 to Senate
District 22. The vast majority of District 22 remains unaffected: The newly
drawn map shifts eight precincts. The Legislature’s plan makes District 22
much more compact and reduces its geographic size, from over 100 miles to 87
miles from north to south. Candidates have time to continue fundraising and
78 The following description of the Legislature’s plan is taken from the uncontroverted
affidavits of Plaintiffs’ experts filed in the district court record after the Legislature’s plan
was adopted.
22
No. 19-60133
campaigning; voters still go to the same polls. We see no clear error by the
district court as to undue prejudice here. 79
The dissent relies on one pending district court case, Chestnut v.
Merrill, 80 for support that laches should bar Plaintiffs’ request for injunctive
relief. 81 Yet Chestnut is distinguishable: The delay and prejudice in that case
were great. In Chestnut, the plaintiffs asked the court to redraw a majority of
the congressional districts in Alabama—four out of seven. 82 Three elections
(in 2012, 2014, and 2016) had already taken place under the challenged
congressional district boundaries. In contrast, the instant case involves only
one of fifty-two state senate districts, and the remedy affected just one other.
And no election took place with the challenged Mississippi senate map until
2015. In sum, Chestnut is not helpful to the State and does not support the
dissent’s position. 83
Accordingly, we conclude that the district court did not abuse its
discretion in denying the State’s laches defense. The State did not meet its
ultimate burden.
79 To the extent the State is arguing that Plaintiffs injected uncertainty into the
November 2019 election, we note that such confusion, if any, is compounded by the
Legislature’s resolution conditioning the effect of its newly drawn district lines on the instant
appeal.
80 377 F. Supp. 3d 1308 (N.D. Ala. 2019).
81 Dissenting Op. at 9–10.
82 Chestnut, 377 F. Supp. 3d at 1311.
83 Even if Chestnut did help the State, it would not end this litigation. Though the
district court denied injunctive relief, it allowed the plaintiffs to pursue declaratory relief.
Chestnut, 377 F. Supp. 3d at 1317–18 (“[W]hile the defendant may be prejudiced by the
plaintiff’s delay in bringing a voting rights claim for injunctive relief, the same prejudice does
not apply to a claim for declaratory relief. . . . Even if the court were to find that Defendant
would be prejudiced as to injunctive relief, declaratory relief would still be available to
Plaintiffs.”) (citing Sanders v. Dooly Cty., 245 F.3d 1289, 1291–92 (11th Cir. 2001)); see also
Dickinson, 933 F.2d at 503 (Seventh Circuit decision taking same approach).
23
No. 19-60133
IV.
Turning to the merits, Plaintiffs asserted that the redistricting plan
adopted by the Mississippi Legislature in 2012 violated § 2 of the VRA in that
the boundary lines of District 22 diluted African-American voting strength and
deprived African-American voters of an equal opportunity to elect candidates
of their choice. They maintained that though African Americans made up a
majority of the voting age population in District 22, that majority lacked “real
electoral opportunity” under Supreme Court precedent interpreting § 2 of the
VRA. Plaintiffs contended that District 22 could be redrawn to cure the § 2
violation with changes to only one or two adjacent districts; that the resulting
districts would be more compact than in the 2012 plan; and that the alteration
would not create Voting Rights Act violations in any other districts. After
conducting a two-day bench trial, the district court determined that Plaintiffs
met their burden of proof as to their § 2 vote-dilution claim and rendered
judgment in their favor.
A.
Before we address the State’s appellate challenges to the district court’s
decision, we set forth the law applicable to a § 2 vote-dilution claim and our
standard of review. Subsection 2(a) of the VRA prohibits all states and political
subdivisions from imposing “any standards, practices, or procedures which
result in the denial or abridgement of the right to vote of any citizen who is a
member of a protected class of racial and language minorities.” 84 Subsection
2(b) provides that a violation of § 2(a) is established if:
[B]ased on the totality of circumstances, it is shown that the
political processes leading to nomination or election in the State or
political subdivision are not equally open to participation by
members of a [protected] class . . . in that its members have less
84 Gingles, 478 U.S. at 43 (citing 42 U.S.C. § 1973(a) (current version at 52 U.S.C.
§ 10301(a))).
24
No. 19-60133
opportunity than other members of the electorate to participate in
the political process and to elect representatives of their choice. 85
Subsection 2(b) specifically provides that “one circumstance” which may be
considered in determining whether § 2(a) has been violated is “[t]he extent to
which members of a protected class have been elected to office in the State or
political subdivision.” However, it cautions “[t]hat nothing in [§ 2] establishes
a right to have members of a protected class elected in numbers equal to their
proportion in the population.” 86
In Thornburg v. Gingles, the seminal vote-dilution case, the Supreme
Court explained that “[t]he essence of a § 2 claim is that a certain electoral law,
practice, or structure interacts with social and historical conditions to cause an
inequality in the opportunities enjoyed by black and white voters to elect their
preferred representatives.” 87 The Court further explained that Congress
amended § 2 of the VRA in 1982 to “make clear that a [§ 2] violation could be
proved by showing discriminatory effect alone” and that proof of discriminatory
intent was not required. 88 The Court additionally noted that “[t]he Senate
Judiciary Committee majority Report accompanying the bill that amended § 2
85 52 U.S.C. § 10301(b).
86 Id. While the dissent emphasizes this cautionary provision, it ignores the preceding
sentence specifically stating that the district court may consider, as “one circumstance” in
the totality of the circumstances establishing a violation of § 2(a), “[t]he extent to which
members of a protected class have been elected to office in the State or political subdivision.”
Here, the district court properly applied the entirety of § 2(b) in its analysis. It noted the
Supreme Court precedent “instruct[ing] district courts to be cautious about finding a § 2
violation where the ‘districting scheme’ features ‘majority-minority districts in substantial
proportion to the minority’s share of voting age population.’” Thomas, 366 F. Supp. 3d at 809
(quoting Johnson v. DeGrandy, 512 U.S. 997, 1013 (1994)). But then determined that such
caution was not warranted in this case because the number of majority-minority districts in
substantial proportion to the minority’s share of voting age population would be 21, and the
number under the 2012 redistricting plan was only 15. Id.
87 478 U.S. at 47.
88 Id. at 35.
25
No. 19-60133
elaborate[d] on the circumstances that might be probative of a § 2 violation.” 89
Those circumstances are frequently referred to as the “Senate factors” and
include the following:
(1) the history of voting-related discrimination in the State or
political subdivision;
(2) the extent to which voting in the elections of the State or
political subdivision is racially polarized;
(3) the extent to which the State or political subdivision has
used voting practices or procedures that tend to enhance the
opportunity for discrimination against the minority group,
such as unusually large election districts, majority vote
requirements, and prohibitions against bullet voting;
(4) the exclusion of members of the minority group from
candidate slating processes;
(5) the extent to which minority group members bear the effects
of past discrimination in areas such as education,
employment, and health, which hinder their ability to
participate effectively in the political process;
(6) the use of overt or subtle racial appeals in political
campaigns; and
(7) the extent to which members of the minority group have
been elected to public office in the jurisdiction. 90
The Senate Report stressed that the above list of factors was “neither
comprehensive nor exclusive” and that there was “no requirement that any
particular number of factors be proved, or that a majority of them point one
way or the other.” 91
89 Id. at 36.
90 Id. at 44–45.
91 Id. at 45 (internal quotation marks and citation omitted).
26
No. 19-60133
The Gingles Court held that prior to establishing the Senate factors,
minority voters asserting a § 2 vote-dilution claim must first prove three
“necessary preconditions.” 92 First, the minority group must show “that it is
sufficiently large and geographically compact to constitute a majority in a
single-member district.” 93 Second, the minority group must demonstrate “that
it is politically cohesive.” 94 And third, the minority group must establish “that
the white majority votes sufficiently as a bloc to enable it—in the absence of
special circumstances, such as the minority candidate running unopposed—
usually to defeat the minority’s preferred candidate.” 95
Finally, the Court “reaffirm[ed] its view that the clearly-erroneous test
of Rule 52(a) is the appropriate standard for appellate review of a finding of
vote dilution.” 96 In reaffirming the proper standard of review, the Court noted
that a district court’s finding of vote dilution is “peculiarly dependent upon the
facts of each case” and “requires an intensely local appraisal of the design and
impact of the contested electoral mechanisms.” 97 The Court additionally
observed that the clearly-erroneous standard of “Rule 52(a) does not inhibit an
appellate court’s power to correct errors of law . . . or a finding of fact that is
predicated on a misunderstanding of the governing rule of law.” 98 Rather,
“application of the clearly-erroneous standard to ultimate findings of vote
92 Id. at 50–51.
93 Id. at 50.
94 Id. at 51.
95 Id. (citations omitted). The dissent contends that the Gingles factors do not
constitute the test for a vote-dilution claim, and that the “totality of the circumstances” is not
“the entirety” of the statute’s text. Dissenting Op. at 14. In its apparent attempt to disavow
Supreme Court jurisprudence interpreting § 2, the dissent does not favor us with the proper
standard to be applied when reviewing the district court’s declaratory judgment, which is the
only matter reviewable in this appeal. Instead, the dissent focuses mostly on issues not
before us.
96 Id. at 79.
97 Id. (internal quotation marks and citations omitted).
98 Id. (internal quotation marks and citations omitted).
27
No. 19-60133
dilution preserves the benefit of the trial court’s particular familiarity with the
indigenous political reality without endangering the rule of law.” 99
B.
The State first asserts that, as a matter of law, Plaintiffs were not
entitled to assert a § 2 challenge to District 22 as drawn by the 2012
redistricting plan because African Americans already made up a majority of
the voting age population (50.8%) in the district. The State contends that its
argument is supported by Supreme Court precedent. We disagree.
“Unimpeachable authority from our circuit has rejected any per se rule
that a racial minority that is a majority in a political subdivision cannot
experience vote dilution.” 100 Other circuits have also rejected such a rule. 101
As interpreted by the Supreme Court in Gingles, § 2 of the Voting Rights Act
protects the voting rights of “member[s] of a protected class of racial and
language minorities.” 102 We have held that § 2 “is concerned with protecting
the minority in its capacity as a national racial or language group,” and not
with protecting “a numerical minority of voters in the jurisdiction at issue.” 103
Importantly, the Supreme Court has specifically stated that “it may be possible
for a citizen voting-age majority to lack real electoral opportunity,” and thereby
require the protection of § 2. 104
In asserting that Plaintiffs’ § 2 claim is not legally cognizable at all
because they represent a BVAP majority in District 22, the State relies on a
99 Id.
100 Monroe v. City of Woodville, 881 F.2d 1327, 1333 (5th Cir. 1989) (citation omitted).
101 See Mo. St. Conf. of the NAACP v. Ferguson-Florissant Sch. Dist., 894 F.3d 924,
934 (8th Cir. 2018) (holding that “minority voters do not lose VRA protection simply because
they represent a bare numerical majority within that district” and collecting cases from the
Second, Fifth, Eleventh, and D.C. Circuits holding same).
102 478 U.S. at 43.
103 Salas v. Sw. Tex. Jr. Coll. Dist., 964 F.2d 1542, 1547–48 (5th Cir. 1992) (providing
thorough discussion of the “kind of ‘minority’ the [VRA] protects”).
104 LULAC, 548 U.S. at 428.
28
No. 19-60133
Supreme Court case, Bartlett v. Strickland, 105 that involved a fundamentally
different issue: crossover districts. The plaintiffs in Bartlett were county
officials who contended that state officials had violated a provision of the state
constitution prohibiting the partition of counties between legislative
districts. 106 The defendants responded that § 2 of the VRA required them to
preserve a district in which African-American voters consistently elected their
preferred candidate with “crossover” voters from the white majority, which
necessitated splitting the plaintiffs’ county. 107 The Court rejected the
defendants’ argument, holding that “[n]othing in § 2 grants special protection
to a minority group’s right to form political coalitions.” 108 The Court further
observed: “There is a difference between a racial minority group’s ‘own choice’
and the choice made by a coalition.” 109
In rendering its holding, the Court clarified the meaning of the first
Gingles precondition: A party asserting a § 2 claim must show by a
preponderance of the evidence that the voting-age minority population in the
potential election district is greater than 50%. 110 The Court did not hold, as
the State incorrectly suggests, that if the district being challenged already
contains a majority-minority population, then a § 2 claim is precluded. That
situation did not arise in Bartlett, so the Court’s holding did not address it. As
observed by the Second Circuit, while the “creation of [majority-minority
districts] with simple majority representation does not necessarily close the
gate to a potential finding of vote dilution, plaintiffs need not show more than
105 556 U.S. 1, 13 (2009).
106 Id. at 7.
107 Id. at 7–9.
108 Id. at 14–15.
109 Id. at 15.
110 Id. at 19–20.
29
No. 19-60133
a simple majority at the first step of the Gingles analysis to open the gate to
further Section 2 inquiry.” 111
If anything, a close reading of Bartlett shows that it supports Plaintiffs’
assertion of a § 2 claim here. The Court began its opinion in Bartlett by
“review[ing] the terminology often used to describe various features of election
districts in relation to the requirements of the [VRA].” 112 The Court then
stated: “In majority-minority districts, a minority group composes a
numerical, working majority of the voting-age population. Under present
doctrine, § 2 can require the creation of these districts.” 113 Here, although
African Americans composed a bare numerical voting-age majority in District
22 under the 2012 plan, Plaintiffs alleged (and ultimately proved) that African
Americans were not a working majority and that § 2 required creation of a
working majority-minority district. The State’s reliance on Bartlett, thus, is
misplaced.
Contrary to the State’s contentions, the fact that Plaintiffs’ § 2 challenge
involves a single-member district, and not an at-large electoral district, does
not render the above principles inapplicable. The Supreme Court has held that
the Gingles framework is applicable in § 2 challenges to single-member
districts. 114 And, in the Court’s decision in LULAC v. Perry, the Court applied
111 Pope v. Cty. of Albany, 687 F.3d 565, 575 n.8 (2d Cir. 2012). Although the State
relies on a decision from the Eastern District of Arkansas, Jeffers v. Beebe, 895 F. Supp. 2d
920 (E.D. Ark. 2012), which applied Bartlett to preclude a plaintiff from asserting a § 2
challenge to a majority-minority district, that decision was overruled by the Eighth Circuit’s
contrary holding in Ferguson-Florissant. See 894 F.3d at 934.
112 Bartlett, 556 U.S. at 13.
113 Id. (emphasis added) (citations omitted). In the recent case, Abbott v. Perez, the
Court reaffirmed these principles: “In a series of cases tracing back to Thornburg v. Gingles,
[] we have interpreted [§ 2’s] standard to mean that, under certain circumstances, States
must draw ‘opportunity’ districts in which minority groups form ‘effective majorit[ies].’” 138
S. Ct. 2305, 2315 (2018) (citations omitted).
114 Growe v. Emison, 507 U.S. 25, 40–41 (1993) (holding that although Gingles involved
multimember districts, the Gingles threshold requirements should be applied in § 2 vote-
dilution challenges to single-member districts).
30
No. 19-60133
the Gingles framework and the above principles when considering a § 2
challenge to a single-member congressional district in Texas. 115 Specifically,
the Court determined that although Latinos were a “bare majority of the
voting-age population” in the challenged district, that majority was a “hollow”
one, and plaintiffs had established that they “could have had an opportunity
district in [the challenged single-member district] had its lines not been
altered.” 116 The Court ultimately determined that the plaintiffs were entitled
to § 2 relief. 117 The State’s argument in this case that Plaintiffs’ § 2 claim is
not legally cognizable because it challenges a majority-minority, single-
member district has no merit. 118
Finally, we disagree with the State’s assertion, which was never raised
to the district court, that Plaintiffs’ § 2 claim is “novel” because Plaintiffs’
allegations did not contain the terms “cracking” (fragmentation of minority
group among various districts so that it is a majority in none) or “packing”
(concentration of minority group in a few number of districts where it
constitutes an excessive majority). We find no statute or caselaw mandating
115 548 U.S. at 425–43.
116 Id. at 429.
117 Id. at 442.
118 The dissent acknowledges these legal principles but insists that no court has ever
found that a majority-minority district violated § 2. That is wrong. Some claims, of course,
involve districts where the protected class is needlessly far above a majority. See, e.g., Bone
Shirt v. Hazeltine, 336 F. Supp. 2d 976, 980 (D.S.D. 2004), aff’d, 461 F.3d 1011 (8th Cir. 2006)
(finding a district with a 90-percent Native American supermajority violated §2). Courts
have also found violations where the protected class’s voting strength was too weak in a
district despite being a numerical majority. LULAC was such a case. See also Perez v. Abbott,
253 F. Supp. 3d 864, 879–90 (W.D. Tex. 2017) (three-judge court) (finding, in a ruling that
state defendants did not appeal, that the 2011 version of Texas Congressional District 23
violated § 2 in both intent and effect despite having a Hispanic citizen voting-age majority of
58.5%); Baldus v. Members of Wis. Gov’t Accountability Bd., 849 F. Supp. 2d 840, 854–58
(E.D. Wis. 2012) (finding that two state assembly districts diluted Latino voting strength
despite having 54% and 61% Latino voting-age population).
31
No. 19-60133
that a plaintiff’s allegations under § 2 include those terms. 119 Moreover, we
note that specific allegations of “cracking” or “packing” were not made in
LULAC v. Perry, yet the Court still determined that the plaintiffs were entitled
to relief because the legislature’s redrawing of a single-member district
violated § 2. 120
Based on the foregoing, the district court did not err in determining that
Plaintiffs’ § 2 challenge to a majority-minority, single-member district was
legally cognizable.
C.
Under the guidance of Gingles, this court uses a two-part framework for
analyzing § 2 claims. First, plaintiffs must satisfy the three Gingles
preconditions; “[f]ailure to establish all three elements defeats a § 2 claim.” 121
Second, if the three preconditions are proved, plaintiffs must then show that
based on the “totality of the circumstances” (as informed by the Senate factors),
the challenged practice or structure results in plaintiffs’ having “less
opportunity than other members of the electorate to participate in the political
119 Allegations of “cracking” or “packing” can figure prominently in racial
gerrymandering cases, in which evidence that a legislature moved large numbers of one race
in or out of a district may be strong circumstantial evidence that “racial considerations
predominated” in the legislature’s decisions. See Cooper v. Harris, 137 S. Ct. 1455, 1472–78
(2017). But this case is governed by § 2’s results test, not racial gerrymandering caselaw, so
it does not entail that potentially sensitive inquiry into the Mississippi Legislature’s intent.
Cf. Veasey v. Abbott, 830 F.3d 216, 280–81 (5th Cir. 2016) (en banc) (Jones, J., concurring in
part and dissenting in part). This difference between racial gerrymandering claims, which
focus on intentional discrimination under the Equal Protection Clause, and vote-dilution
claims, which focus on discriminatory effect under § 2 of the VRA, was explained by the Court
in Abbott v. Perez, 138 S. Ct. 2305, 2314-15, 2330-35 (2018). See discussion infra. p. 45. In
any event, Plaintiffs’ theory—that District 22 lacked adequate African-American voting
strength, remediable by adding precincts with African-American voters from adjoining
districts—is functionally equivalent to asserting “cracking.”
120 548 U.S. at 423–43, 447.
121 Teague v. Attala Cty., 92 F.3d 283, 287 (5th Cir. 1996).
32
No. 19-60133
process and to elect representatives of their choice.” 122 We have noted that “it
will be only the very unusual case in which the plaintiffs can establish the
existence of the three Gingles factors but still have failed to establish a
violation of § 2 under the totality of the circumstances.” 123
The State asserts that even if Plaintiffs can assert a legally cognizable
§ 2 challenge to a majority-minority, single-member district, they failed to
prove the first Gingles precondition—that the minority group is sufficiently
large and geographically compact to constitute a voting age majority in a
single-member district—because District 22 already contained a majority-
minority voting age population. It is an incontestable fact that African
Americans can constitute a majority of the district; they already do, if only
barely. The State’s argument is simply a reassertion of its argument
(discussed above) that Plaintiffs’ § 2 claim is not legally cognizable. The State’s
attempt to reframe its argument in the context of the Gingles preconditions is
unavailing. For the reasons previously stated, the State’s argument
challenging the first precondition has no merit.
Regarding the second Gingles precondition—that the minority group is
“politically cohesive”—the State does not directly challenge the district court’s
factual finding of this precondition as clearly erroneous. In fact, the State
never disputed Plaintiffs’ proof of the second Gingles precondition in the
district court. 124
122 Clark v. Calhoun Cty., 21 F.3d 92, 94 (5th Cir. 1994) (citing 42 U.S.C. § 1973(b)
(current version at 52 U.S.C. § 10301(a))).
123 Id. at 97 (internal quotation marks and citation omitted).
124 See Thomas, 366 F. Supp. 3d at 805 (“It is undisputed that African-American voters
in District 22 are politically cohesive.”). To the extent the State now attempts to dispute
political cohesiveness, we find the issue waived based on the State’s failure to brief. See
Yohey v. Collins, 985 F.2d 222, 225 (5th Cir. 1993) (“[A]rguments must be briefed to be
preserved.”).
33
No. 19-60133
As to the third Gingles precondition—whether white bloc voting usually
defeats the African-American community’s preferred candidate—the State
asserts primarily that Plaintiffs’ evidence was flawed because it was based in
part on the results of the 2015 senate election in District 22, which was not
“properly conducted.” Plaintiffs explain that the State is referring to an
election error that occurred in Bolivar County during the 2015 senate election
when most voters in two of District 22’s fifty-two precincts were given ballots
that did not include the District 22 election.
The district court specifically addressed this issue in its decision. The
district court noted that in the 2015 state senate election for District 22,
“approximately 1,500 voters in Bolivar County received ballots for the wrong
Senate race.” 125 Plaintiffs’ expert, Dr. Palmer, agreed that this was a
“significant election administration error,” which was why he decided to
exclude those precincts in that race from his ecological inference analysis. 126
Dr. Palmer explained that his analysis “remain[ed] valid because EI identifies
the pattern of behavior running through a series of elections over time.” 127 The
2015 election in District 22 still furnished fifty precincts from which to infer
voter behavior, and other elections also furnished usable data. Furthermore,
the State “presented no evidence indicating that Dr. Palmer’s approach was in
error or would cast any shadow on his conclusions.” 128
Dr. Palmer testified that EI “is the process of extracting clues about
individual behavior from information reported at the group or aggregate level.”
125 Thomas, 366 F. Supp. 3d at 806.
126 As noted by the district court, EI analysis is widely recognized and accepted in
voting cases. See, e.g., Terrebonne Par. Branch NAACP v. Jindal, 274 F. Supp. 3d 395, 433–
34 (M.D. La. 2017); Jamison v. Tupelo, Miss., 471 F. Supp. 2d 706, 710 (N.D. Miss. 2007);
Houston v. Lafayette Cty., 20 F. Supp. 2d 996, 1002 n.4 (N.D. Miss. 1998).
127 Thomas, 366 F. Supp. 3d at 806.
128 Id.
34
No. 19-60133
He contended it is useful in voting cases because “the secret ballot hinders the
[research] process and surveys in racially polarized contexts are known to be
of little value.” Dr. Palmer further explained that EI “estimates the underlying
propensity of each group to turn out for an election and to vote for a particular
candidate using the estimation technique of maximum likelihood.” Dr. Palmer
believed that EI was a superior statistical method to use in this case because
it allowed him to run 100,000 simulations of each election in the sample, and
provided valuable statistical checks, such as confidence intervals, on the
results.
Dr. Palmer also testified that he used precinct-level voting and census
data to analyze ten elections in District 22. The elections consisted of three
“endogenous” elections—the 2003, 2007, and 2015 Senate District 22 elections,
and seven “exogenous” elections—the 2003 Lieutenant Governor and
Treasurer elections, the 2007 Insurance Commissioner election, the 2011
Governor election, and the 2015 Governor, Secretary of State, and Agriculture
Commissioner elections. The goal of the endogenous and exogenous
comparisons was to see if findings were consistent between the Senate races
and statewide races also held in odd years in District 22. All ten elections
featured contests between white and African-American candidates. In all ten
elections, the candidate preferred by African-American voters lost. Based on
his analysis, Dr. Palmer appropriately concluded that white bloc voting usually
defeats the African-American community’s preferred candidate in District 22.
The district court found “Dr. Palmer’s thorough and largely unrebutted
analysis to be persuasive.” 129
By contrast, the State’s expert, Dr. Morrison, did not analyze any
District 22 elections or the results of statewide elections among the voters of
129 Id. at 807.
35
No. 19-60133
District 22. He looked only at county and municipal elections within District
22’s territory, without attending to the demographic differences between those
smaller jurisdictions and District 22. Basic errors also infected the data that
he analyzed. 130 Moreover, Dr. Morrison offered no opinion, critical or
otherwise, on Dr. Palmer’s analysis. Dr. Morrison’s testimony thus did not
rebut Dr. Palmer’s report in any meaningful respect.
On appeal, the State simply highlights the election administration error
that occurred in the 2015 senate election for District 22. The State does not
explain why Dr. Palmer’s method of accounting for the error—deleting the
affected precincts from his analysis—was incorrect or undermined his analysis
in any way. The State fails to show any clear error in the district court’s finding
that District 22’s white majority votes sufficiently as a bloc to enable it usually
to defeat the African-American preferred candidate.
Based on the foregoing, the district court did not clearly err in
determining that Plaintiffs met their burden of proving the three Gingles
preconditions.
D.
We next address the Senate factors, which are probative of the “totality
of the circumstances” demonstrating a vote-dilution claim under § 2. The
district court found persuasive the Plaintiffs’ evidence regarding Senate
factors one (the history of voting-related discrimination in the State or political
subdivision), two (the extent to which voting in the elections of the State or
political subdivision is racially polarized), five (the extent to which minority
group members bear the effects of past discrimination in areas such as
education, employment, and health, which hinder their ability to participate
130 See id. at 805.
36
No. 19-60133
effectively in the political process), and seven (the extent to which members of
the minority group have been elected to public office in the jurisdiction). 131
As to factor one, the district court noted that “Mississippi plainly has a
long history of official discrimination against African-Americans seeking to
vote.” The district court credited the State for acknowledging this fact. As to
factor two, the district court noted that Plaintiffs presented evidence, through
Dr. Palmer’s testimony, that voting in District 22 features a “high level of racial
polarization.” The district court further noted that the State’s expert also did
not dispute this evidence. 132
The fifth Senate factor focuses on the extent to which minority group
members bear the effects of past discrimination in areas such as education,
employment, and health, which hinder their ability to participate effectively in
the political process. As described above, the socioeconomic disparities
between the white and African-American residents in District 22 are stark.
They differ markedly by poverty rate, household income, educational
attainment, and other measures of privation and status. Considering the
unique history of the Delta, there can be no question that these disparities
reflect the effects of slavery and segregation.
The State argues that although Plaintiffs presented evidence that
African Americans in District 22 trail white residents in socioeconomic
131 The district court determined that factor three (voting practices or procedures that
tend to enhance the opportunity for discrimination against the minority group) and factor
four (the exclusion of members of the minority group from candidate slating processes) were
not relevant as neither side presented any evidence regarding those factors. Regarding factor
six (the use of overt or subtle racial appeals in political campaigns), the district court noted
that the Plaintiffs did not present evidence of any recent racial appeals or of any racial
appeals that would have had an effect on any District 22 election within the timeframe of the
Plaintiffs’ case.
132 It is unclear whether the State disputes these factors on appeal. However, to the
extent the State now attempts to dispute the first and second Senate factors, we find those
issues waived based on the State’s failure to brief. See Yohey, 985 F.2d at 225.
37
No. 19-60133
categories, Plaintiffs “offered no proof to connect that fact to past
discrimination.” The State further asserts that Plaintiffs did not prove that
the socioeconomic circumstances of African Americans in District 22 “hinder
their ability to participate effectively in the political process.” As the district
court observed, however, we have held that where socioeconomic disparities
are shown and where the level of African-American political participation is
depressed, § 2 plaintiffs “need not prove any further causal nexus between
their disparate socioeconomic status and the depressed level of political
participation.” 133 At trial, the State acknowledged this law, both when moving
for judgment at the close of Plaintiffs’ case and in closing arguments. The point
the State now makes, that Plaintiffs did not offer evidence linking
discrimination to depressed participation, thus is new on appeal, legally
insignificant, and––in any event––belied by the record.
In this case, Plaintiffs went beyond the requirements of our law and were
able to show a causal nexus between the disparate socioeconomic
circumstances of African Americans in District 22 and their depressed level of
political participation. Specifically, Plaintiff Lawson testified that in the Delta,
many African-American voters are uneducated, reside in the rural areas, and
have no transportation to travel to the polls. The district court noted that this
evidence “is consistent with the Supreme Court’s recognition ‘that political
participation by minorities tends to be depressed where minority group
members suffer effects of prior discrimination such as inferior education, poor
employment opportunities, and low incomes.’” 134
The State acknowledges that Plaintiffs’ expert, Dr. Palmer, presented a
statistical estimate that African-American voter turnout in the 2015 Senate
133 Teague, 92 F.3d at 294 (quoting League of United Latin Am. Citizens, Council No.
4434 v. Clements, 999 F.2d 831, 867 (5th Cir. 1993) (en banc)).
134 Thomas, 366 F. Supp. 3d at 808 (quoting Gingles, 478 U.S. at 69).
38
No. 19-60133
election was 29.6%, while the white turnout was 36.9%. But the State asserts
that the estimate was inaccurate because Dr. Palmer admitted that he had
excluded from his analysis the precincts in Bolivar County in which voters
were given the wrong ballots. This is the same challenge to Dr. Palmer’s
analysis raised in the State’s challenge to the third Gingles precondition. As
discussed above, the State presented no evidence indicating that Dr. Palmer’s
approach was erroneous or would have cast any doubt on his conclusions. But
in the context of voter turnout, the State argues that exclusion of the precincts
was significant because the precincts included Delta State University, a
predominantly white institution, and that college students are part of the
voting age population in the census but notoriously unlikely to register and
vote. The State submits that had those non-voting white students been
considered in the turnout estimates, then white participation throughout
District 22 would have fallen, making the difference between the percentage
levels of white and African-American voter turnout not as great. But here
again, the State can point to no evidence in the record supporting its
supposition or in any other respect its challenge to Dr. Palmer’s conclusion of
depressed African-American voter turnout. 135
The State additionally asserts that Census Bureau statistics show that
the statewide African-American voter turnout has exceeded white voter
turnout by one to ten percent in even-numbered years since at least 2004.
Based largely on this information, the State’s expert, Dr. Morrison, opined that
the existing socioeconomic differences no longer diminish African-American
Of course, attorney argument is not itself evidence. See Foradori v. Harris, 523
135
F.3d 477, 513 (5th Cir. 2008); United States v. Garza, 608 F.2d 659, 662–63 (5th Cir. 1979).
The State asserts that it requested the district court to take judicial notice of the information
on Delta State University, but the record shows the State never clearly did so. We thus need
not conclusively resolve our doubt that the imputation of extra-record data into a complex
quantitative analysis would be an appropriate exercise of judicial notice.
39
No. 19-60133
participation in the political process as they did in the past. However, the data
used by Dr. Morrison had serious problems. It was self-reported, rather than
based on official state records. It came from statewide elections, as opposed to
District 22 elections, and from even-numbered years for federal elections
(including presidential election years), as opposed to odd-numbered years for
state elections. Accordingly, the district court, assessing this trial testimony
in the first instance, held that Dr. Morrison’s “data points fail to persuade”
because “[t]hey look at the wrong jurisdiction, the wrong election years, and
rely upon known issues with self-reported voting surveys—issues that
[ecological inference], in contrast, seeks to overcome.”
The district court’s conclusion that Dr. Palmer’s analysis of voter turnout
held more probative value than the data used by Dr. Morrison is consistent
with our previous direction. 136 It is also consistent with the Supreme Court’s
pronouncements in Gingles that § 2 vote dilution determinations “require[] an
intensely local appraisal of the design and impact of the contested electoral
mechanisms.” 137
Finally, the seventh Senate factor—the extent to which members of the
minority group have been elected to public office in the jurisdiction—strongly
favors Plaintiffs. No African American has been elected to the Mississippi
Senate from District 22 in its previous or current form.
We find that the district court did not clearly err in finding that Senate
factors one, two, five, and seven demonstrated that District 22, as drawn in the
2012 legislative redistricting plan, resulted in African Americans’ having less
136 See Operation Push, 932 F.2d at 412 (affirming district court’s decision to favor
expert analysis of voter registration rates, based on official state records, over Census Bureau
figures, due to unreliability of “uncorroborated self-reporting” and to “evidence
demonstrating that black respondents to the [census] questionnaire overreported voter
registration at a higher rate than white respondents”).
137 Gingles, 478 U.S. at 79 (internal quotations omitted).
40
No. 19-60133
opportunity than other members of the electorate to participate in the political
process and to elect representatives of their choice. Plaintiffs proved the
Gingles preconditions and presented persuasive evidence regarding several of
the Senate factors, while the State scarcely entered any evidence to the
contrary.
The dissent acknowledges that § 2 of the VRA requires relief when a
protected class has less opportunity to participate in the political process. Yet
it turns a blind eye to the evidence produced by Plaintiffs showing the stark
socioeconomic circumstances of African Americans in District 22 and how those
circumstances have resulted in depressed voter turnout. Moreover, though the
dissent states that “there are major disconnects among the district court’s
findings,” and provides as an example “the minority turnout figures as well as
minority electoral success in other races throughout the district,” the dissent
provides no explanation for its vague contention. Instead, the dissent
challenges the district court’s judgment in generalities and provides no
analysis of the district court’s extensive factual findings and legal analysis
supporting its judgment declaring District 22 violative of § 2. 138
Accordingly, we hold that the district court did not clearly err in its
ultimate finding of vote dilution. We further hold that the district court’s
138 The dissent raises a similarly hazy contention about the district court considering
District 22 “in a vacuum.” The nature of the criticism is unclear, partly because the dissent
cites no caselaw that might identify a requirement the district court failed to meet. In any
event, the trial record shows extensive testimony on District 23, the adjoining district from
which Plaintiffs’ illustrative plans and the district court’s remedial plan drew voters. Indeed,
all three experts testified about the effects on District 23 of changing District 22. The district
court also followed the Supreme Court’s guidance that the existence of majority-minority
districts “in substantial proportion to the minority’s share of voting age-population” should
caution courts against creating more majority-minority districts. Thomas, 366 F. Supp. 3d
at 808–09 (quoting Johnson v. DeGrandy, 512 U.S. 997, 1013 (1994)). The district court
properly considered how changes to District 22 would affect Mississippi’s senate map as a
whole.
41
No. 19-60133
conclusion that Plaintiffs were entitled to § 2 relief is fully supported by the
record and not clearly erroneous.
V.
A.
The State argues that the district court erred in adopting Plaintiffs’
Illustrative Plan 1 to remedy the § 2 violation and requests that we vacate the
district court’s injunction mandating its use. Plaintiffs contend that the State’s
appeal as to this claim of error is moot. We agree. The unlawful 2012
legislative plan, which necessitated the district court’s injunctive relief, was
repealed by the Legislature in March 2019. Plaintiffs do not challenge the new
plan. Although there is no question that the new plan governs the upcoming
election, the State asserts that the district court’s injunction should be vacated
because, if not, the State could be forced to operate under the court-ordered
plan. That is simply inaccurate. The Mississippi Legislature fashioned the
2019 plan in such a way that the only other redistricting plan that could
possibly ever become effective is the original 2012 redistricting plan. In other
words, no matter the resolution of the State’s appeal, the court-ordered plan
will never become operative. The State’s request for relief from the district
court’s injunction is therefore moot.
B.
Even though the district court’s remedial plan will not be used in any
election because it has been replaced by the Legislature’s plan, the dissent
objects to the court’s plan at length. 139 The dissent is particularly “alarmed”
that the district court did not conduct an evidentiary hearing before adopting
Plan 1. These concerns are moot, because the Legislature’s plan is in place. 140
139See Dissenting Op. at 4, 15–16, 19–20.
140At most, these criticisms would get the State a more thorough remedial process or
an opportunity to craft its own map. The State eventually obtained this relief anyway after
42
No. 19-60133
We nevertheless clarify that there is nothing alarming or unprecedented about
what the district court did.
1.
We must set the facts straight—they bear repeating. In its February 16,
2019 opinion following the bench trial, the district court unequivocally
“decline[d] to order any specific relief while the Mississippi Legislature
considers whether to redraw the District and extend the candidate
qualification deadline.” 141 The court added: “A hearing will be set for the near
future.” 142 Later, as the record shows, at the February 21 teleconference, the
district court asked “whether the parties sought additional hearings and
whether the defendants would be submitting their own Plan.” 143 In response,
“[t]he parties convinced the Court that no further hearings were necessary.” 144
And the “defendants said they would confer to be sure, then made no request
at all.” 145 So the dissent’s focus on this moot point is incorrect: The State did
not want another hearing.
Furthermore, the bench trial in this case in fact involved both the merits
and remedies. The State “had ample opportunity to present evidence at the
remed[ies]” portion of the trial. 146 The State extensively cross-examined
Plaintiffs’ expert, William Cooper, about his proposed plans during discovery
and at trial. In fact, Cooper incorporated the State’s concerns and arguments
after discovery by proposing two additional plans. 147 In Cooper’s opinion,
the stay panel of our court gave the State another chance to redistrict. On this issue, it is not
possible to grant the State “any effectual relief” beyond what it has already received. See
Church of Scientology, 506 U.S. at 12.
141 Thomas, 366 F. Supp. 3d at 810.
142 Id. (emphasis added).
143 ECF Dkt. No. 85 at 1–2.
144 Id. at 2 (emphasis added).
145 Id.
146 Cf. Abrams v. Johnson, 521 U.S. 74, 94–95 (1997).
147 Thomas, 366 F. Supp. 3d at 795.
43
No. 19-60133
“District 22’s boundaries could be reconfigured to increase the BVAP while
honoring traditional redistricting criteria and minimizing disruption to
adjacent Districts” and while “increas[ing] the BVAP by at least 10 additional
percentage points.” 148 As a result of the trial evidence, the district court
concluded that “[a]ll of Cooper’s illustrative plans satisfy traditional
redistricting criteria. They are contiguous, reasonably compact, reasonably
shaped, satisfy one-person one-vote, and do not dilute minority voting
strength.” 149 Moreover, the district court noted that the “incumbent Senator
in District 23 remains in the same District.” 150 Because the State later
declined the district court’s invitation for an additional hearing and had not
presented a remedial plan, the district court adopted Cooper’s Plan 1 as the
best plan presented.
In approving the district court’s course of conduct, we draw guidance
from Abrams v. Johnson. There, the Supreme Court upheld a district court’s
remedial plan, which resulted from a joint bench trial and remedial hearing,
against parties’ arguments that the district court should have held a further
hearing on the court-drawn plan’s compliance with § 2. 151 As the Court said
there, “neither our precedents nor the [Voting Rights] Act require the court to
hold a separate hearing on the adequacy under § 2 of a remedial plan.” 152
148 Id. at 794.
149 Id. at 796.
150 Id.
151 521 U.S. at 94–95. The remedial map at issue in Abrams was court-drawn,
following the Supreme Court’s invalidation of a Georgia congressional district as a racial
gerrymander. See Miller v. Johnson, 515 U.S. 900 (1995). On remand, the district court held
a bench trial to consider whether another district was a racial gerrymander as well, and it
also considered remedial plans submitted by the parties. See Abrams, 521 U.S. at 82–85.
The district court then created its own plan and declined to hold a hearing on its plan’s
compliance with § 2 when intervenors raised the issue. Id. at 84–85, 94–95.
152 Id. at 94–95. The dissent relies on dicta in Jones v. City of Lubbock, 727 F.2d 364
(5th Cir. 1984) for authority that a hearing is required. Dissenting Op. at 15–16. To the
extent that Jones stood for that proposition, we note that it predated the Supreme Court’s
Abrams decision. Regardless, we do not read Jones the same way. Rather, its principal
44
No. 19-60133
In sum, the district court did not pull a remedial plan out of thin air. The
district court offered the Legislature the first opportunity to redraw District
22. Yet the Legislature waited until after our motion panel denied a stay before
it responded to the district court’s request for a remedial plan, which is now
the operative plan. Therefore, the court-ordered plan is moot—and so is the
State’s argument.
2.
Finally, the dissent suggests that a race-conscious remedy to a Voting
Rights Act violation infringes upon the Fourteenth Amendment’s Equal
Protection Clause. 153 The dissent, however, is once again asserting arguments
never raised by the State 154 and challenging the propriety of the district court’s
remedial plan, which is not before us. In any event, the Supreme Court in
Abbott v. Perez has explained the interplay between the Equal Protection
Clause and the Voting Rights Act as follows:
At the same time that the Equal Protection Clause restricts the
consideration of race in the districting process, compliance with
the Voting Rights Act pulls in the opposite direction: It often
insists that districts be created precisely because of race. . . . In an
effort to harmonize these conflicting demands, we have assumed
that compliance with the VRA may justify the consideration of race
in a way that would not otherwise be allowed. 155
concern was that the district court had imposed its own remedy before allowing the relevant
legislative body a chance. Id. at 387. Our subsequent citations of Jones bear this out. See
United States v. Brown, 561 F.3d 420, 435 (5th Cir. 2009); Rodriguez v. Bexar Cty., 385 F.3d
853, 870 (5th Cir. 2004). The district court and the stay panel of our court gave the
Mississippi Legislature that opportunity, which it took, so Jones poses no problem. Jones’s
secondary concern was that district courts produce adequate records to enable appellate
review. 727 F.2d at 387. As we explain, the district court saw to that here.
153 See Dissenting Op. at 5, 20–21.
154 Arguments not raised before the district court and in an appellant’s opening brief
are waived. Valle v. City of Houston, 613 F.3d 536, 544 (5th Cir. 2010); LeMaire v. La. Dep’t
of Transp. & Dev., 480 F.3d 383, 387 (5th Cir. 2007).
155 Abbott, 138 S. Ct. at 2314–15 (citations omitted).
45
No. 19-60133
The Court further explains that in “technical terms,” it has assumed that
compliance with the VRA is a “compelling state interest.” 156 Therefore,
contrary to the dissent’s assertion, race can be considered for remedying a § 2
violation.
VI.
Based on the foregoing, the district court’s judgment declaring that the
redistricting plan adopted by the Mississippi Legislature in 2012 violates § 2
of the Voting Rights Act, in that the boundary lines of District 22 dilute
African-American voting strength and deprive African-American voters of an
equal opportunity to elect candidates of their choice, is AFFIRMED. The
State’s appeal of the district court’s judgment granting injunctive relief is
DISMISSED AS MOOT.
156 Id.
46
No. 19-60133
STEPHEN A. HIGGINSON, Circuit Judge, joined by W. EUGENE DAVIS,
Circuit Judge, concurring:
I join fully in Judge Davis’s majority opinion. I write separately to
elaborate on the law governing plaintiffs’ causal burden when the protected
class they represent is a numerical majority in the challenged jurisdiction.
Some ambiguity seemed to me to exist in our law, stemming primarily from
our decision in Salas v. Southwest Texas Junior College District, 964 F.2d 1542
(5th Cir. 1992). 1
In Salas, our court placed a “difficult burden” on plaintiffs who represent
a protected class comprising a majority in their district to explain why, despite
their numbers, their preferred candidates lost elections and why the plaintiffs
still warranted relief under § 2. Id. at 1555. This enhanced burden varies from
longstanding § 2 caselaw, indeed, notably from our subsequent en banc
decision in LULAC, Council No. 4434 v. Clements, 999 F.2d 831, 866–67 (5th
Cir. 1993), which specifies a less demanding approach.
On further inspection, there is neither ambiguity nor tension, however.
Salas’s “difficult burden” discussion is limited to the problem that its facts
presented: “a registered voter majority” in the challenged jurisdiction. 964 F.2d
at 1555. A protected class that has navigated the practical obstacles of voter
registration in enough numbers to be a majority naturally has a harder time
explaining why it continues to lose elections. Here, however, it was not
established that Appellees represent such a majority, because registered-voter
data for Mississippi State Senate District 22 was not in the record.
1I offer this inquiry into Salas in a separate concurrence because Appellants clearly
and repeatedly urged our court’s settled § 2 framework to the district court; indeed, they
never briefed an argument on Salas to us and disclaimed its application at oral argument.
47
No. 19-60133
***
The Voting Rights Act is an exercise of Congress’s power to enact
“appropriate legislation” enforcing the guarantees of the Fifteenth
Amendment. See Shelby County, Ala. v. Holder, 570 U.S. 529, 536 (2013); Lopez
v. Monterey County, 525 U.S. 266, 269 (1999); City of Rome v. United States,
446 U.S. 156, 173 (1980); Georgia v. United States, 411 U.S. 526, 535 (1973);
South Carolina v. Katzenbach, 383 U.S. 301, 327 (1966). Section 2 of the Voting
Rights Act prohibits any election practice or procedure “which results in a
denial or abridgment of the right of any citizen of the United States to vote on
account of race or color” or membership in a language minority group. 52
U.S.C. § 10301(a). Denial or abridgment occurs if the “totality of the
circumstances” shows that “the political processes leading to nomination or
election . . . are not equally open to participation by members of [a protected
class] in that its members have less opportunity than other members of the
electorate to participate in the political process and to elect representatives of
their choice.” Id. § 10301(b).
As explained in the majority opinion, Congress codified the § 2 results
test in 1982, abrogating a Supreme Court decision two years earlier requiring
a showing of discriminatory purpose. Thornburg v. Gingles, 478 U.S. 30, 35
(1986) (citing City of Mobile v. Bolden, 446 U.S. 55 (1980)); see THE LAW OF
DEMOCRACY: LEGAL STRUCTURE OF THE POLITICAL PROCESS 603–54 (Samuel
Issacharoff et al. eds., 4th ed., 2012).
In Gingles, the Court supplied the framework for redistricting challenges
we still use today. There, it explained the “essence” of a § 2 claim: “that a
certain electoral law, practice, or structure interacts with social and historical
conditions to cause an inequality in the opportunities enjoyed by black and
white voters to elect their preferred representatives.” 478 U.S. at 47. As a guide
to this inquiry, the Court endorsed a non-exclusive set of factors listed in the
48
No. 19-60133
Senate committee report to the 1982 amendments and found in an earlier
decision of this court. Id. at 36–37 & n.4 (citing Zimmer v. McKeithen, 485 F.2d
1297 (5th Cir. 1973) (en banc), aff’d sub nom. E. Carroll Parish Sch. Bd. v.
Marshall, 424 U.S. 636 (1976)).
The fifth of these factors is “the extent to which members of the minority
group in the state or political subdivision bear the effects of discrimination in
such areas as education, employment and health, which hinder their ability to
participate effectively in the political process.” Gingles, 478 U.S. at 37 (citing
S. Rep. No. 97-417, at 28–29 (1982), reprinted in 1982 U.S.C.C.A.N. at 206–
07). Following the Supreme Court’s guidance, we developed a framework for
plaintiffs to show the interaction between a challenged election practice, social
and historical conditions, and unequal political opportunity. Plaintiffs must
make three showings: past discrimination; socioeconomic disparities arising
therefrom; and depressed political participation. See LULAC, Council No. 4434
v. Clements, 999 F.2d 831, 866–67 (5th Cir. 1993) (en banc) (citing S. Rep. No.
97-417, at 29 n.114). Plaintiffs may show depressed political participation in
the form of reduced voter registration, lower turnout, or “any other factor
tending to show that past discrimination has affected [the protected class’s]
ability to participate in the political process.” Id. at 867. Once discrimination,
disparity, and depressed participation are shown, “plaintiffs need not prove
any further causal nexus between their disparate socio-economic status and
the depressed level of political participation.” Id. (quoting S. Rep. No. 97-417,
at 29 n.114).
We have used this framework consistently since then. See Veasey v.
Abbott, 830 F.3d 216, 258–61 (5th Cir. 2016) (en banc); N.A.A.C.P. v. Fordice,
252 F.3d 361, 367–68 (5th Cir. 2001); Teague v. Attala County, Miss., 92 F.3d
283, 293–94 (5th Cir. 1996); Clark v. Calhoun County, Miss., 88 F.3d 1393,
1399 (5th Cir. 1996). Under this framework, the fifth Senate factor does not
49
No. 19-60133
favor plaintiffs when, for instance, the protected class’s voter turnout is no less
than white voter turnout, Fordice, 252 F.3d at 368, or when plaintiffs fail to
put on evidence of depressed political participation at all, LULAC, 999 F.2d at
867.
Appellants clearly and repeatedly urged this framework to the district
court. At the close of Appellees’ case, Appellants moved for judgment on partial
findings, challenging the sufficiency of Appellees’ evidence. 2 Responding to the
motion, counsel for Appellees cited the portion of Veasey v. Abbott explaining
the framework. 3 In rebuttal, counsel for Appellants concurred that the
framework governed and conceded two of its three parts:
The constitutional chain of events that must be proven in order for
Section 2 to be constitutionally applied is that the state did
something unconstitutional. And you have evidence before you
that the state has done a few unconstitutional things over the
years and that those things led to socioeconomic disadvantages for
black folks. And you have evidence before you that black folks do
suffer socioeconomic advantages in that case. And then what
Veasey says, I believe, is that if you have those two things proven,
then you may conclude that those things have caused the political
participation disparities that you see. You don’t have to have proof
of the causation; you may infer that causation.
Counsel then proceeded to the factual argument, reiterated to us on appeal,
that Appellees’ evidence did not adequately show depressed political
participation in the form of lower black turnout. At closing argument the next
day, counsel for Appellants urged the same framework to the district court,
again to set up an attack on Appellees’ evidence of depressed turnout. 4
2 See FED. R. CIV. P. 52(c).
3 “[T]he case of Veasey v. Abbott . . . is a Fifth Circuit . . . en banc decision that says: If
you have socioeconomic disparities and if you have lower African-American or lower minority
turnout, that factor of the Senate Report factors is counted in favor of the plaintiffs in
analyzing the proof of whether there is vote dilution.”
4 “They absolutely have to prove that black political participation in the relevant area
is depressed. And it’s certainly true that Veasey v. Abbott says that you may presume that a
50
No. 19-60133
Our framework for establishing depressed political participation matters
here because it is a compelling way to explain why a protected class
constituting a bare majority in the challenged jurisdiction may nevertheless
win relief under § 2. The facts in evidence showed that District 22 had a black
voting-age majority just over 50% as of the 2010 Census, a turnout gap between
black and white voters of roughly 10%, and a consistent single-digit margin of
defeat for candidates preferred by black voters. Mississippi’s history of
discrimination, recognized by our court, 5 was conceded. The uncontroverted
record showed stark socioeconomic disparity between black and white
residents of District 22: five times the poverty rate; one-third the median
household income; three times the high school dropout rate; three times the
uninsured rate; and so on. Under both our law and common sense, and as the
district court concluded, that record suffices to explain why the lines making
up District 22 “interact[] with social and historical conditions to cause an
inequality in the opportunities enjoyed by black and white voters to elect their
preferred representatives.” Gingles, 478 U.S. at 47.
This conclusion comports with our longstanding recognition that a
protected class may constitute a numerical majority in the challenged
jurisdiction but nevertheless obtain relief under § 2. See Salas, 964 F.2d at
1547; Monroe v. City of Woodville, Miss., 881 F.2d 1327, 1333 (5th Cir. 1989);
Zimmer, 485 F.2d at 1303. Other circuit courts share that recognition. See, e.g.,
Mo. St. Conf. of NAACP v. Ferguson-Florissant Sch. Dist., 894 F.3d 924, 934
(8th Cir. 2018) (“[M]inority voters do not lose VRA protection simply because
depressed political participation derives from past unconstitutional behavior, but you have
to prove that political participation is presently depressed.” In keeping with the parties’
arguments, the district court appropriately applied this framework in its ruling on the merits.
See Thomas v. Bryant, 366 F. Supp. 3d 786, 807–08 (S.D. Miss. 2019) (citing Teague, 92 F.3d
at 294).
5 See, e.g., Fordice, 252 F.3d at 367; Teague, 92 F.3d at 293–94.
51
No. 19-60133
they represent a bare numerical majority within the district.”); Pope v. County
of Albany, 687 F.3d 565, 575 n.8 (2nd Cir. 2012) (“[T]he law allows plaintiffs to
challenge legislatively created bare majority-minority districts on the ground
that they do not present the ‘real electoral opportunity’ protected by Section
2.”) (quoting LULAC v. Perry, 548 U.S. 399, 428 (2006)).
This conclusion comports also with LULAC v. Perry, in which the
Supreme Court recognized that “it may be possible for a citizen voting-age
majority to lack real electoral opportunity.” 548 U.S. at 428. Notably, the Court
found there that a Texas congressional district violated § 2 where Latinos were
“a bare majority of the voting-age population . . . but only in a hollow sense,”
owing to the legal restriction of voting rights to citizens. Id. at 429.
Finding a § 2 violation despite a bare majority is likewise consistent with
the Supreme Court’s admonition that we are to “take a ‘functional’ view of the
political process and conduct a searching and practical evaluation of reality.”
Gingles, 479 U.S. at 66 (quoting S. Rep. No. 97-417, at 30 n.119). In that vein,
when assessing how large a protected class must be to have equal political
opportunity, the Court has spoken variously of “effective majorities,” Abbott v.
Perez, 138 S. Ct. 2305, 2315 (2018) (quotation omitted); “a numerical, working
majority,” Bartlett v. Strickland, 556 U.S. 1, 13 (2009); and “an effective voting
majority,” Johnson v. DeGrandy, 512 U.S. 997, 1014 (1994). 6 Finding a § 2
violation despite a bare majority is consistent also with Abrams v. Johnson,
6 It is useful here to distinguish the present case from two situations previously
addressed by the Supreme Court: first, a protected class numerous enough to elect its
preferred candidate despite not being a majority; and second, a protected class so numerous
in one district that its voting strength in other districts is diluted. See Bartlett, 556 U.S. at
14 (describing the first); Voinovich v. Quilter, 507 U.S. 146, 153–54 (1993) (describing the
second). We are interested here in determining when a protected class becomes so numerous
that it can no longer claim it has inadequate voting strength in the challenged district.
Bartlett concerned the inverse––adequate voting strength despite low numbers. Voinovich
concerned effects on voting strength in other districts, rather than the challenged one.
52
No. 19-60133
521 U.S. 74 (1997), in which the Court affirmed a district court’s remedial plan
setting a congressional district’s black voting-age population (BVAP) “as close
to 55% as possible” to ensure compliance with § 2 because, at 50% BVAP, black
voters’ preferred candidates were more likely than not to lose. Id. at 94. 7
New questions arise when registered-voter data enters the picture. In
previous cases, it has often been true that a protected class made up a smaller
share of registered voters than of voting-age population. See, e.g., Gingles, 478
U.S. at 39 (fourteen-point gap between black and white voter registration
rates); Teague, 92 F.3d at 294 (twenty-point gap). It cannot be assumed that a
protected class comprising a majority of voting-age population also comprises
a majority of registered voters. In the present case, as counsel for Appellants
conceded, no voter-registration data for District 22 was in the record. 8
But when voter-registration data is available, and it shows that the
protected class is a registered-voter majority in the challenged jurisdiction, it
becomes harder to contend that the protected class lacks political opportunity.
See Salas, 964 F.2d at 1544. Salas concerned a junior college that conducted
7 Similarly, decisions remedying § 2 violations regularly produce districts well above
the 50% threshold. See, e.g., Large v. Fremont County, Wyo., 2010 WL 11508507, at *14 (D.
Wyo. Aug. 10, 2010) (adopting remedial plan including district with 70% Native American
voting-age population), aff’d, 670 F.3d 1133 (10th Cir. 2012); Bone Shirt v. Hazeltine, 387 F.
Supp. 2d 1035, 1041 (D.S.D. 2005) (creating remedial districts with 66% and 74% Native
American voting-age population), aff’d, 461 F.3d 1011 (8th Cir. 2006); Jeffers v. Clinton, 756
F. Supp. 1195, 1198–1201 (E.D. Ark. 1990) (three-judge court) (creating remedial districts
with 64%, 63%, 58%, and 55% BVAP), aff’d, 498 U.S. 1019 (1991). The same is true when
courts remedy other districting violations and craft their remedial plans in view of § 2’s
requirements. See, e.g., Navajo Nation v. San Juan County, 2017 WL 6547635, at *13 (D.
Utah Dec. 21, 2017) (imposing remedial plan creating districts with Native American voting-
age populations of 63% and 64%, among others), aff’d, 929 F.3d 1270 (10th Cir. 2019);
Adamson v. Clayton County Elections & Registration Bd., 876 F. Supp. 2d 1347, 1357–58 &
n.7 (N.D. Ga. 2012) (drawing districts in remedial plan with BVAP’s from 59% to 77%).
8 Such data evidently is not publicly available in Mississippi. See Oral Argument at
8:10, Thomas v. Bryant (5th Cir. June 11, 2019) (No. 19-60133),
http://www.ca5.uscourts.gov/OralArgRecordings/19/19-60133_6-11-2019.mp3. The record
contained only self-reported statewide Census data of voter registration, collected in even
years, on which the district court appropriately did not rely, as the majority opinion explains.
53
No. 19-60133
at-large elections for its governing board. Id. The Hispanic residents of the
district were 63% of the population and 57% of the voting-age population, and
the Spanish-surname voter-registration rate was 53%. Id. The parties
evidently agreed that “Hispanics constitute[d] a slight majority of the
registered voters in the District.” Id.
We said in Salas that an “inquiry into causation” underlies the Gingles
analysis. 964 F.2d at 1154. Is it the challenged district or election practice that
“explain[s] the lack of electoral success of voters within the protected class,” or
is it something else not cognizable under the Voting Rights Act? Id. This
question is easier for plaintiffs to answer when, for instance, the protected class
could be a working majority but is cracked apart. It is harder for them to
answer when they have managed to register enough voters to constitute a
majority and yet continue losing elections. The voter registration process
presents one cluster of obstacles to the political participation of a protected
class that bears the effects of discrimination. See Miss. State Chapter,
Operation Push, Inc. v. Mabus, 932 F.2d 400 (5th Cir. 1991). 9 How to explain
a lack of electoral success when a majority in the jurisdiction has managed to
clear those obstacles?
In Salas, we suggested ways that plaintiffs might prove that their
“registered voter majority is illusory”: evidence of inaccurate, or “soft,” voter
rolls; evidence of other “practical impediments to voting,” like migratory work
patterns at election time; or “low turnout at elections [that] was the result of
9 In Operation Push, a § 2 challenge to Mississippi voter-registration statutes enacted
in 1972, we noted Mississippi’s “long history of using voter qualifications and registration
procedures to impede black citizens’ participation in the political process.” 932 F.2d at 402.
We affirmed the district court’s finding of a twenty-five-point gap between black and white
voter-registration rates in the state, “[b]ecause a significant percentage of Mississippi’s black
citizens do not have (1) access to the transportation necessary to travel to the county
courthouse to register, or (2) the type of jobs that would allow them to leave work during
business hours to register to vote.” Id. at 403, 411–12.
54
No. 19-60133
prior official discrimination.” Id. at 1555–56. We then refined the plaintiffs’
required causal showing: plaintiffs should show “evidence directly linking this
low turnout with past official discrimination.” Id. at 1556 (emphasis added).
This might seem to be in tension with our holdings in LULAC and other cases
that a causal nexus between discrimination, socioeconomic disparity, and
depressed participation can be inferred once those three phenomena are
shown. But reconciling Salas is simple. The protected class’s success at
navigating the voter-registration process calls the inference of a causal nexus
into question; hence, plaintiffs must produce more evidence to regain the
preponderance.
Counsel for Appellees suggested at oral argument that our en banc
decision in LULAC v. Clements overruled Salas. 10 We certainly did not do so
explicitly, and LULAC did not present the special problem of a registered-voter
majority, nor has any case in our court ever since. Neither does this one. As
noted, registered-voter data for District 22 was not in the record, so the
problem confronted by Salas did not arise here. Moreover, Appellants passed
on the many opportunities they had to argue that Salas should apply,
acknowledging, instead, that the Veasey and LULAC framework controls. 11
***
A suit under the Voting Rights Act is a complicated undertaking,
drawing together long and often painful local histories, tightly contested
stakes, intricate legal inquiries, and a deep desire for electoral fairness.
Whatever addition to that complexity might have come from the question of
Salas’s place in our law, I am hopeful this separate opinion may clarify it.
10Oral Argument at 29:00, Thomas v. Bryant (5th Cir. June 11, 2019).
11Appellants did not cite Salas in their briefing to us, and, when asked at oral
argument to explain that decision, counsel for Appellants said only that it “has good language
and bad language for both sides, which is why neither side brought it up.” Oral Argument at
49:21, Thomas v. Bryant (5th Cir. June 11, 2019).
55
No. 19-60133
APPENDIX
56
No. 19-60133
DON R. WILLETT, Circuit Judge, dissenting:
When President Lincoln issued the Emancipation Proclamation,
declaring that slaves in rebel states “are, and henceforward shall be free,” 1 he
remarked, “If my name ever goes into history, it will be for this act, and my
whole soul is in it.” 2 The war over America’s original sin waged another two
years. The Great Emancipator lived to see the Union preserved, but just barely.
Appomattox was Palm Sunday, and Ford’s Theatre was Good Friday.
Earlier that week, President Lincoln had delivered a speech declaring
his plans for peace and reconstruction. And perhaps healing would have taken
hold under a second Lincoln term. But as it happened, the post-Lincoln
American South was appallingly inhospitable to African-Americans.
In 1870 the Fifteenth Amendment was ratified, commanding that “[t]he
right of citizens of the United States to vote shall not be denied or abridged by
the United States or by any State on account of race, color, or previous
condition of servitude,” and empowering Congress “to enforce this article by
appropriate legislation.” But Congress’s first century of enforcement “can only
be regarded as a failure” 3 as southern states erected barriers galore to
disenfranchise African-American citizens. The institution of slavery—
America’s foundational failing—had ended, but the legacy of slavery had
endured. For black Americans in the South, it seemed ineradicable, and racial
injustice inescapable.
1 The Emancipation Proclamation, NAT’L ARCHIVES,
https://www.archives.gov/exhibits/featured-documents/emancipation-proclamation.
2 See Ronald G. Shafer, Lincoln Moved to End Slavery on New Year’s Day 1863. It
Went On for Three More Years., WASH. POST (Jan. 1, 2019),
https://www.washingtonpost.com/history/2019/01/01/lincoln-declared-an-end-slavery-new-
years-day-it-went-two-more-years/ (attributing this quotation to the moment before Lincoln
signed the Emancipation Proclamation).
3 Nw. Austin Mun. Util. Dist. No. One v. Holder, 557 U.S. 193, 197 (2009).
57
No. 19-60133
Finally, in 1965, a full century after the Civil War ended, Congress
passed the Voting Rights Act to make good on the promise of the Fifteenth
Amendment by counteracting entrenched racial discrimination in voting. As
the Supreme Court explained in upholding the VRA the year after it was
enacted, Congress was aiming to combat “an insidious and pervasive evil which
had been perpetrated in certain parts of our country through unremitting and
ingenious defiance of the Constitution.” 4 Section 2 of the VRA imposes a
nationwide ban on voting discrimination, forbidding any “standard, practice,
or procedure” that “results in a denial or abridgement of the right of any citizen
of the United States to vote on account of race or color.” 5
Nondiscrimination in voting is thus enshrined in the Constitution and
laws of the United States. In my judgment, the Voting Rights Act is perhaps
the noblest, most transformative legislation in American history. 6 And it “has
proved immensely successful at redressing racial discrimination and
integrating the voting process.” 7
4 South Carolina v. Katzenbach, 383 U.S. 301, 309 (1966).
5 42 U.S.C. § 1973(a).
6 See Holder v. Hall, 512 U.S. 874, 895 (1994) (Thomas, J., concurring) (“The Act was
immediately and notably successful in removing barriers to registration and ensuring access
to the ballot. For example, in Mississippi, black registration levels skyrocketed from 6.7% to
59.8% in a mere two years; in Alabama the increase was from 19.3% to 51.6% in the same
time period. By the end of 1967, black voter registration had reached at least 50% in every
covered State.” (citations omitted)); see also Nw. Austin Mun. Util. Dist. No. One, 557 U.S. at
227 (Thomas, J., concurring) (“[A]nd in Mississippi, black voter registration actually exceeded
white voter registration by 1.5 percentage points.”).
7 Shelby County v. Holder, 570 U.S. 529, 548 (2013); see also CHARLES S. BULLOCK III
& RONALD KEITH GADDIE, THE TRIUMPH OF VOTING RIGHTS IN THE SOUTH 323 (2009) (“The
consensus is that the [VRA] has been inordinately successful.”); SAMUEL ISSACHAROFF ET AL.,
THE LAW OF DEMOCRACY: LEGAL STRUCTURE OF THE POLITICAL PROCESS 514–16 (4th ed.
2012) (stating that “guarantees of the Fourteenth and Fifteenth Amendments were
essentially disregarded in many states” before the VRA, which “transformed American
politics in a variety of ways”).
58
No. 19-60133
Also enshrined in American law is the broad power of States, “as
provided in the Tenth Amendment . . . to regulate elections,” 8 including the
power to shape congressional and state legislative districts. 9 Time and again,
the Supreme Court has reaffirmed that electoral map-drawing “is primarily
the duty and responsibility of the State through its legislature or other body,
rather than of a federal court.” 10 Why? Because drawing district lines is, to
quote the Supreme Court, “a most difficult subject,” 11 which is a G-rated way
of describing a quintessentially political thicket besmeared with partisan
point-scoring and one-upsmanship. Districting is the politics of politics. And
the Framers, well acquainted with electoral maneuvering, wisely entrusted
districting to political actors. Given the “complex interplay” of value-laden
political forces at work, the Supreme Court says “the States must have
discretion to exercise the political judgment necessary to balance competing
interests.” 12 This “allocation of powers in our federal system preserves the
integrity, dignity, and residual sovereignty of the States.” 13 States are
sovereign political entities in their own right, and when it comes to political
map-making, “[f]ederal-court review of districting legislation represents a
serious intrusion on the most vital of local functions.” 14 This is particularly
true when elections are fast approaching. In such cases, a wealth of election-
8 Gregory v. Ashcroft, 501 U.S. 452, 461–62 (1991).
9 See Fitzgerald v. Green, 134 U.S. 377, 380 (1890) (“Congress has never undertaken
to interfere with the manner of appointing [members of the electoral college], or, where
(according to the now general usage) the mode of appointment prescribed by the law of the
state is election by the people, to regulate the conduct of such election, or to punish any fraud
in voting for electors, but has left these matters to the control of the states.”).
10 Chapman v. Meier, 420 U.S. 1, 27 (1975); see also Thomas, 919 F.3d at 318–19
(Clement, J., dissenting).
11 Miller v. Johnson, 515 U.S. 900, 915 (1995).
12 Id. at 915–16.
13 Bond v. United Sates, 564 U.S. 211, 221 (2011).
14 Miller, 515 U.S. at 915.
59
No. 19-60133
law precedent counsels that it is often wiser for courts to push pause instead
of fast-forward.
To be sure, state prerogatives cannot override legal imperatives. As the
Supreme Court recently recognized, “voting discrimination still exists; no one
doubts that.” 15 Maps must of course comply with both the Voting Rights Act
and the Constitution, and courts undeniably have room to act swiftly when
necessary. The issue in this case, though, is whether the district court’s order
transgressed settled principles of federal election law. To my mind, it did—
alarmingly so.
* * *
Respectfully, but fervently, I dissent—for four main reasons:
1. The Supreme Court disapproves meddling in imminent
elections.
2. The equitable defense of laches bars this suit.
3. The majority opinion’s novel § 2 theory contradicts Supreme
Court precedent, the VRA’s plain text, and quite possibly the
Fourteenth Amendment.
4. The jurisdictional issue may itself be dispositive, though that
argument, admittedly, is more confounding than convincing.
In sum, I believe it offends ordinary principles of election law for a single
federal district judge—with jarring haste, just days before the qualifying
deadline, and without conducting a sufficient evidentiary hearing—to order
redistricting after studying a majority-minority district, isolating its precise
Black Voting Age Population (50.8%), and ordering a State to take explicitly
race-conscious steps to increase that percentage (to 60.9%) . . . adding minority
15 Shelby County, 570 U.S. at 536.
60
No. 19-60133
voters seemingly to ensure a minority winner and elevating outcome over
opportunity. And the district court did all this absent any evidence (or even an
allegation) of “cracking” or “packing” of the majority-minority population—the
usual, Supreme Court-dictated ways of showing vote dilution under § 2.
The Equal Protection Clause’s “central mandate is racial neutrality in
governmental decisionmaking,” including in the “drawing of [legislative]
districts.” 16 Thus, “race-based districting by our state legislatures demands
close judicial scrutiny.” 17 True, the Mississippi Legislature reconvened and
reconfigured the district. But make no mistake—it did so under the Damoclean
sword of a stay denial from a divided motions panel, all the while specifying,
repeatedly, that it would very much prefer its original map.
Respectfully, the majority opinion defies the High Court’s federalism-
based warnings in short-fuse election cases, endorses an inflated view of
judicial power, and flouts the Constitution’s demand for racial neutrality.
I
The Supreme Court has repeatedly warned against tinkering with
looming elections. In Miller, the Court described judicial review of legislative
line-drawing as “a serious intrusion on the most vital of local functions.” 18
Twenty years earlier in Chapman the Court stressed that “reapportionment is
primarily the duty and responsibility of the State through its legislature or
other body, rather than of a federal court.” 19 And about a decade before that,
the Court in Reynolds laid the foundation for short-runway election cases:
“[W]here an impending election is imminent and a State’s machinery is already
in progress, equitable considerations might justify a court in withholding the
16 Miller, 515 U.S. at 904–05.
17 Shaw v. Reno, 509 U.S. 630, 657 (1993).
18 Miller, 515 U.S. at 915.
19 Chapman, 420 U.S. at 27.
61
No. 19-60133
granting of immediately effective relief in a legislative apportionment case,
even though the existing apportionment scheme was found invalid.” 20
We have heeded the Supreme Court’s admonition. In Veasey, we
underscored the need “to carefully consider the importance of preserving the
status quo on the eve of an election.” 21 Concurring, Judge Costa registered his
“extreme[] reluctan[ce] to have [the] election take place under a law that a
district court has found, and that our court may find, is discriminatory.” 22 But
even so, he continued, we “must follow the dictates of the Supreme Court.” 23
And “the only constant principle that can be discerned from the Supreme
Court’s recent decisions in this area is that its concern about confusion
resulting from court changes to election laws close in time to the election
should carry the day in the stay analysis.” 24
Indeed, last-minute changes to district boundaries may cause more voter
confusion than they cure. In vacating a Ninth Circuit injunction, the Supreme
Court observed, “Court orders affecting elections, especially conflicting orders,
can themselves result in voter confusion and consequent incentive to remain
away from the polls. As an election draws closer, that risk will increase.” 25
Here, the district court found a VRA violation in mid-February 2019. At
that time, a slew of election deadlines was fast approaching. Most notably, the
qualification deadline was March 1. Roughly a week after its original
decision—days before the qualifying period ended—the court imposed
redistricting and moved the qualifying deadline back only two weeks. And in a
20 Reynolds v. Sims, 377 U.S. 533, 585 (1964).
21 Veasey v. Perry, 769 F.3d 890, 892 (5th Cir. 2014).
22 Id. at 896 (Costa, J., concurring).
23 Id.
24 Id. at 897 (citing North Carolina v. League of Women Voters of N.C., 135 S. Ct. 6
(2014) (Mem.); Husted v. Ohio State Conference of NAACP, 573 U.S. 988 (2014) (Mem.)).
25 Purcell v. Gonzalez, 549 U.S. 1, 4–5 (2006) (per curiam).
62
No. 19-60133
separate order, the court created further uncertainty about district lines by
encouraging the parties to collaborate on a new plan. This was days away from
the qualifying deadline and mere months before the primary and general
elections. And lest you think that the election date is the only date that
matters, the Supreme Court has cautioned that qualifying deadlines (or other
preliminary deadlines) “cannot be considered in isolation from the election of
which [they] form[] a part.” 26
This severely interfered with candidate registration, with the political
process, and with the Legislature. Political candidates don’t jumpstart
campaigns at the drop of a hat. Nor are legislatures designed to legislate with
lickety-split haste—even in response to district court diktats. 27 The status quo
was preferable to federal-court intervention at the eleventh hour.
II
Laches bars this suit too. As the Supreme Court recently put it, laches
“protect[s] defendants against ‘unreasonable, prejudicial delay in commencing
suit.’ ” 28 To assert laches, a defendant must show “(1) a delay in asserting a
right or claim; (2) that the delay was not excusable; and (3) that there was
undue prejudice to the party against whom the claim is asserted.” 29 Here, every
26 NAACP v. Hampton Cty. Election Comm’n, 470 U.S. 166, 177 (1985).
27 The district court observed that “[a] stay would probably . . . be warranted if the
Court was ruling on this case in September or October.” Given the nature of this case and
the potential for en banc review, ongoing consideration of this matter into those months
seems quite possible. Even if a stay was not appropriate in February (which I believe it was),
it certainly seems appropriate now.
28 SCA Hygiene Prods. Aktiebolag v. First Quality Baby Prods., LLC, 137 S. Ct. 954,
960 (2017) (quoting Petrella v. Metro-Goldwyn-Mayer, Inc., 572 U.S. 663, 667 (2014)); see also
Am. Rice, Inc. v. Producers Rice Mill, Inc., 518 F.3d 321, 334 (5th Cir. 2008).
29 Save Our Wetlands, Inc. v. U.S. Army Corps of Eng’rs, 549 F.2d 1021, 1026 (5th
Cir.), cert. denied, 434 U.S. 836 (1977); see also Lopez v. Hale Cty., 797 F. Supp. 547 (N.D.
Tex. 1992), aff’d, 506 U.S. 1042 (1993).
63
No. 19-60133
box is checked. The plaintiffs unreasonably delayed. And that delay prejudiced
the State. The district court’s ruling otherwise was clear error. 30
A
At bottom, the plaintiffs believed District 22 violated the law in 2012. As
we explained in our 1998 Capece decision, “[t]he period for laches begins when
the plaintiff knew or should have known of the infringement.” 31 True, a district
court has considerable discretion in applying laches. But it’s not unfettered
discretion. That’s why our 1982 Armco decision reversed a district court for
denying laches when it miscalculated the plaintiff’s delay as six years shorter
than it actually was. 32 Again, we emphasized that delay should be counted
from when the plaintiff should’ve known of the alleged violation. 33
Here, the plaintiffs could have sued in 2012, when the Department of
Justice granted preclearance to the State’s districts. Indeed, Thomas
challenged District 22 in a letter to the Department of Justice in August 2012:
This letter is a request for the Department of Justice to look hard
at the Mississippi Senate Redistricting plan. This plan has
violated section 5 and 2 of the Voting Right Bill. Please look at
District[s] 21, 22, [and] 34 . . . .
So Thomas himself admits that he knew every element of his claim in 2012.
This actually goes beyond the laches standard, which is when the plaintiff
should have known of the violation. 34
30 Retractable Techs., Inc. v. Becton Dickinson & Co., 842 F.3d 883, 898 (5th Cir. 2016)
(explaining that we review the district court’s findings of delay, unreasonableness, and
prejudice for abuse of discretion—overturning only if there was clear error).
31 Elvis Presley Enters., Inc. v. Capece, 141 F.3d 188, 205 (5th Cir. 1998); see also White
v. Daniel, 909 F.2d 99, 102 (4th Cir. 1990) (stating that an inexcusable delay occurs when a
party “discovers or with reasonable diligence could have discovered the facts giving rise to
his cause of action”).
32 Armco, Inc. v. Armco Burglar Alarm Co., 693 F.2d 1155, 1161–62 (5th Cir. 1982).
33 Id. at 1161.
34 See Envtl. Def. Fund, Inc. v. Alexander, 614 F.2d 474, 479 (5th Cir. 1980) (“[L]aches
does not depend on subjective awareness of the legal basis on which a claim can be made.”).
64
No. 19-60133
The majority suggests that the State abandoned this timeframe on
appeal, arguing that 2015, not 2012, was the operative date. Not true. That
was merely an argument in the alternative. But all along, the State has
contended that 2012 was the key date. Consider these excerpts from the State’s
briefs on appeal:
• “Rejecting Thomas’s concerns, the DOJ precleared the [State’s]
plan in September 2012. []Thomas waited. He did not file a
Voting Rights Act challenge in 2012, 2013, or 2014.”
• “[T]he length of delay in this case is at least three years . . . .”
• “Instead of bringing suit in 2012 when the plan was adopted
and precleared by DOJ or after the 2015 election, Thomas . . .
waited almost three additional years to commence this action.”
• “Under the ‘should have known’ test, they all ‘should have
known’ in 2015. There is no doubt that, if not in 2012, then by
the time of the 2015 election, any reasonable person would have
known of the present cause of action.”
The 2015 window merely confirmed that, as in 2012, the plaintiffs had
everything necessary for their § 2 claim. Waiting three years, much less six,
was inexplicable for Thomas. By then, the district had progressed through
preclearance and an election.
Take another Voting Rights Act case currently moving through the
courts: Chestnut v. Merrill, a recently decided case out of the Northern District
of Alabama. 35 The court held that laches barred the plaintiffs’ request for
injunctive relief. The plaintiffs’ claims first became available when the state
35 Chestnut v. Merrill, 377 F. Supp. 3d 1308 (N.D. Ala. 2019).
65
No. 19-60133
redistricted, and based on that timeline, the court found that the plaintiffs had
unreasonably delayed in suing Alabama. 36
B
So have the plaintiffs here. They dawdled—then dawdled some more.
And this was highly prejudicial. To the State, who was thwarted from
redistricting according to the wishes of its elected representatives. And also to
voters. Even the majority concedes that eleventh-hour changes “will inevitably
cause disruption.”
Indeed, this suit has injected pointless confusion into the electoral
process. Voters should know what district they live in and who they are voting
for. Candidates should know their potential constituency. Instead, the district
court strong-armed the State into a slap-dash redistricting plan. This suit
began in 2018 and wasn’t decided until 2019. And next year there’s a census.
So the State will doubtless have to redistrict again.
Returning to the Alabama district court in Chestnut, the court there
sensibly barred the VRA claim since “forc[ing] the state of Alabama to
redistrict twice in two years—once based on nine-year-old census data—would
result in prejudice.” 37 A Florida district court held likewise in 1999, applying
laches. 38 The defendants asserted that they were prejudiced by delay because
“voters [had] come to know their districts and candidates, and [would] be
confused by change”; and because “requiring redistricting [then], before the
2000 census [would] result in two redistrictings within a two year period, with
36 Id. at 1314–18.
37 Id. at 1317.
38 Fouts v. Harris, 88 F. Supp. 2d 1351, 1354 (S.D. Fla. 1999), aff’d sub nom. Chandler
v. Harris, 529 U.S. 1084 (2000) (Mem.).
66
No. 19-60133
resulting voter confusion, instability, dislocation, and financial and logistical
burden on the state.” 39
Clearly, as the Fourth Circuit has pointed out, “there is large potential
for disruption in reapportioning with undue frequency.” 40 But that’s not the
only prejudice. The plaintiffs hamstrung the State by creating a nigh-
impossible timetable for effectively litigating. The plaintiffs’ late filing date
forced the State, at the taxpayers’ expense, to expedite each facet of this case.
The Legislature had to quickly enact a new district map, without the
substantial deliberation that redistricting requires. This urgency stymied the
State’s ability to fully litigate its own preferred, democratically enacted plan,
which inherently prejudiced the State. Laches should bar this case.
For the same reason, this case—as the majority opinion correctly holds—
is not moot. As the Supreme Court explained seven years ago in Knox, a case
“becomes moot only when it is impossible for a court to grant ‘“any effectual
relief” whatever to the prevailing party.’ ” 41 Here, Mississippi wants its old
districts back.
III
Turning to the § 2 merits analysis:
Nothing in the Voting Rights Act requires proportional representation.
The statute states so unequivocally: “[N]othing in [§ 2] establishes a right to
have members of a protected class elected in numbers equal to their proportion
to the population.” 42 The VRA explicitly disavows the maximization of minority
electoral success and explicitly emphasizes electoral participation and
39 Id.
40 White v. Daniel, 909 F.2d 99, 104 (4th Cir. 1990).
41 Knox v. Serv. Emps. Int’l Union, Local 1000, 567 U.S. 298, 307 (2012) (quoting Erie
v. Pap’s A.M., 529 U.S. 277, 287 (2000)).
42 52 U.S.C. § 10301(b).
67
No. 19-60133
opportunity. Section 2 tells us to look to “the totality of the circumstances” to
determine if “members of a racial group . . . have less opportunity . . . to
participate in the political process and to elect representatives of their
choice.” 43
Both Congress and the Supreme Court have emphasized this. Congress
requires plaintiffs to show that the political process isn’t “equally open to
participation by members of a [protected] class . . . [because] members have
less opportunity than other members of the electorate to participate . . . .” 44 And
the Supreme Court noted 25 years ago in De Grandy that “the ultimate right
of § 2 is equality of opportunity, not a guarantee of electoral success for
minority-preferred candidates of whatever race.” 45 Thus, a viable § 2 claim
must show unequal opportunity. The Supreme Court explained how to do this
only a year later in Shaw. 46 A § 2 plaintiff can allege that a state either
“fragment[ed] politically cohesive minority voters among several districts or
pack[ed] them into one district or a small number of districts.” 47 That is, § 2
can be violated if a state “cracks” racial minorities among many districts
(making them an ineffective minority of voters) or “packs” them into few
districts (making them an excessive majority of voters)—weakening their
overall voting power.
The State argues that plaintiffs from a majority-minority population
cannot bring a § 2 claim at all. Our precedent, as Judge Higginson notes in his
43 League of United Latin Am. Citizens v. Perry (LULAC), 548 U.S. 399, 425 (2006)
(quoting 42 U.S.C. § 1973(b)).
44 52 U.S.C. § 10301(b) (emphasis added).
45 Johnson v. De Grandy, 512 U.S. 997, 1014 n.11 (1994).
46 Shaw v. Hunt, 517 U.S. 899, 914 (1996).
47 Id.; see also Thomas, 919 F.3d at 318–19 (Clement, J., dissenting).
68
No. 19-60133
concurrence, suggests they can. 48 We held in Salas that members of a
registered voter majority class are not barred, “as a matter of law,” from
bringing a vote-dilution claim. 49 Obtaining relief, however, turns on proof. 50
And that’s the real question: the standard of proof. The law lacks clarity as to
what exactly a protected-class plaintiff must show in order to prevail on a vote-
dilution claim when they are already part of the majority in the district.
This much is clear, however: No court has ever ruled that a majority-
minority district violates § 2 in isolation. The two dilutive tactics recognized by
the Supreme Court—packing and cracking—necessarily cross district lines,
which necessarily requires courts to examine electoral ripple effects, which
necessarily demands painstaking statistical analysis spanning multiple
districts to ferret out the siphoning of minorities’ voting power. 51 How can a
court conclude that a specific district violates § 2 without scrutinizing whether
and how the minority vote has been manipulated across various districts? I am
unaware of any court decision holding that a majority-minority district can
violate § 2 in a vacuum, all by itself, unaccompanied by evidence—or even an
allegation—of packing or cracking. 52
48 Salas v. Sw. Tex. Jr. Coll. Dist., 964 F.2d 1542, 1547 (5th Cir. 1992) (“We first
consider whether plaintiffs, as members of a registered voter majority class, are precluded,
as a matter of law, from bringing a vote dilution claim. We hold that they are not.”).
49 Id. at 1547. Judge Higginson’s concurrence forthrightly and correctly identifies an
unresolved tension in our caselaw that bears directly on the unusual situation presented
here: How must a majority-minority population in a single-member district show § 2
causation in the absence of packing or cracking of the majority-minority population?
50 See id.
51 See, e.g., Voinovich v. Quilter, 507 U.S. 146, 154 (1993) (“[W]e have recognized that
‘[d]ilution of racial minority group voting strength may be caused’ either ‘by the dispersal of
blacks into districts in which they constitute an ineffective minority of voters or from the
concentration of blacks into districts where the constitute an excessive majority.’ ” (quoting
Thornburg v. Gingles, 478 U.S. 30, 46 n.11 (1986))).
52 The majority opinion, in footnote 118, attributes to me this assertion: “The dissent
. . . insists that no court has ever found that a majority-minority district violated § 2.” Not
exactly. My precise point is that no court I’m aware of has ever found such a violation absent
allegations and evidence of packing and cracking and attendant multi-district statistical
69
No. 19-60133
No such evidence was presented here. Indeed, the complaint (neither
original nor amended) never mentions “packing” or “cracking.” But it did
mention that District 22 needed “a significantly increased black voting age
population,” specifically “60% rather than the existing 50.8%.” And the
evidence, such as it was, focused solely on District 22, not on any adjoining
districts.
Yes, the majority opinion is right to emphasize the Supreme Court’s
decision in Gingles. 53 But the Gingles factors are mere thresholds for a dilution
claim. 54 They’re not the test itself. Nor is “totality of the circumstances” the
entirety of the text from the statute. 55 It’s true that since Gingles, district-court
findings profoundly influence redistricting law. District courts must engage in
a close, fact-specific analysis of voting patterns. Yet in this case, there are
major disconnects among the district court’s findings. These disconnects make
analysis. See discussion infra n.51 & accompanying text. For support, the majority opinion
points to several cases. Respectfully, the cited cases underscore, rather than undermine, my
point.
Every single case includes discussion of packing, cracking, and multi-district electoral
consequences. First, Bone Shirt v. Hazeltine, 336 F. Supp. 2d 976 (D.S.D. 2004). Literally, the
first sentence in that opinion: “Plaintiffs contend that South Dakota’s 2001 legislative
redistricting plan dilutes . . . by packing . . . in violation of § 2 of the Voting Rights Act of
1965.” Id. (emphasis added). Second, there’s LULAC, discussed later in this dissent, which
had to do with an illusory majority, not an actual one. See discussion infra nn.64–66 &
accompanying text. Third, there’s Perez v. Abbott, 253 F. Supp. 3d 864 (W.D. Tex. 2017). True,
one sliver of that order did not discuss packing or cracking. But the overall case was a
sprawling challenge to Texas’s entire congressional redistricting plan. As such, it
necessitated intensive statewide and multi-regional analysis (“South/West Texas area,”
“Dallas–Ft. Worth area,” and “Houston area”). The order is chock-full of cross-district
number-crunching and discussion of plaintiffs’ claims of packing and cracking. Indeed,
variations of the word “pack” appear 44 times, and “crack” 42 times. Finally, fourth, there’s
Baldus v. Members of Wis. Gov’t Accountability Bd., 849 F. Supp. 2d 840, 854–58 (E.D. Wis.
2012). And here too, the decision analyzes whether there had been packing and cracking. See
id. at 848, 856.
53 Thornburg, 478 U.S. at 48–49.
54 Id.
55 Contra Thomas, 919 F.3d at 309 (discussing the “totality of the circumstances”
language from § 2 but neglecting to mention the rest of the sentence that the phrase is from).
70
No. 19-60133
the decision hard to understand—for example, the minority turnout figures as
well as minority electoral success in other races throughout the district. Plus,
under the district court’s adopted plan, the only declared opposition candidate
to Thomas was redistricted out of the district.
It’s also troubling that the district court imposed Plan 1—the first
plaintiff-proposed redistricting plan—without conducting a sufficient
evidentiary hearing or even explaining itself. 56 The court’s cursory, one-page
order adopting the plan never justifies why Plan 1 was superior to Plans 2 and
3; or why Plan 1 was the narrowest remedy. The only discussion of the plans’
details appears in the district court’s liability opinion, which devotes just three
pages to describing them. Compare that to our 1984 decision in Jones. 57 There,
we noted that even in an expedited situation, a district court must conduct an
evidentiary hearing, make specific factual findings, or otherwise produce a
record sufficient for appellate review. 58
The Supreme Court’s later decision in Abrams v. Johnson in no way
absolves the district court from that obligation. 59 In that case, the district court
conducted a full-dress remedial hearing, and the Supreme Court considered
what evidence was presented there and subject to cross-examination. 60 The
appellants complained because the district court, having already held a
remedial hearing, did not hold yet another hearing about whether the court’s
own plan satisfied § 2. 61
56 See Thomas, 919 F.3d 298 at 318 (Clement, J., dissenting).
57 Jones v. City of Lubbock, 727 F.2d 364, 387 (5th Cir. 1984).
58 Id.
59 521 U.S. 74 (1997).
60 Id. at 89 (“The ACLU plans were introduced at the remedial hearing . . . . [T]he
Justice Department submitted the plan after the close of evidence, and in consequence its
demographer could not be cross-examined on the question of racial motivation . . . .”); id. at
92–95.
61 Id. at 94–95.
71
No. 19-60133
It also bears mentioning: In its remedy order, the district court in
Abrams contributed 28 pages to the Federal Supplement, 62 whereas the
district court here mustered only one. Reliance on Abrams thus seems
misplaced.
Nor did representations during the February 21, 2019 telephone
conference obviate a hearing. Although the call transcript doesn’t seem to be
in the record, the district court later summarized that the State didn’t ask for
a hearing. But what matters isn’t what the State desires but what the law
requires. A court poised to impose a remedial redistricting plan has its own
obligations when deciding time-sensitive, consequential elections cases,
including producing for the appellate court a factually supported evidentiary
basis for its decision. 63 The two-day liability trial, one-page order, and three-
page liability opinion nowhere explain why the court chose Plan 1.
Yet the majority opinion ignores these red flags. Instead, it cherry-picks
quotations to create a new rule for determining § 2 violations: whether a
protected class has an “effective” or “working” majority. This essentially means
that if a minority population is roughly 50–59% of the district population,
district lines must afford them more than a mere “opportunity” to elect their
preferred candidate. They’re entitled to something close to recurring success.
62 Johnson v. Miller, 922 F. Supp. 1556, 1556–86 (S.D. Ga. Dec. 13, 1995) (totaling 28
pages—of which 11 explained the court’s reasoning and 17 contained tables and data backing
up that reasoning).
63 See Jones, 727 F.2d at 387 (“While it may be clear to us that the district court—with
at least the acquiescence, if not the connivance, of the parties—believed that the proceedings
merited expedited treatment, the procedures, if challenged, would have required that we
vacate this order. For the sake of future courts, we reiterate briefly some of the principles
that the district court should bear in mind. . . . For us to pass on the propriety of the district
court’s action, we must have either specific fact findings or, at least, a record sufficient to
allow review. Without hearings, and without findings addressed to the government body's
plan, we would not be in a position to determine whether the district court properly exercised
its discretion in rejecting the City’s plan.” (emphasis added)).
72
No. 19-60133
To support this new rule, the majority hangs its hat on the Supreme
Court’s 2006 LULAC decision. There, the Court recognized that “it may be
possible for a citizen voting-age majority to lack real electoral opportunity.” 64
But that case is drastically different. LULAC dealt with a Latino-majority
district in which many Latinos weren’t citizens and thus couldn’t vote. 65 So the
Court granted § 2 relief since the Latino majority was a “hollow” one. 66
Yet what the majority opinion holds today is far beyond that. Its results-
oriented focus transforms the Voting Rights Act landscape. To see how, first
consider the existing caselaw. On the one hand, the Supreme Court’s 2009
Bartlett decision tells us that if a protected class is a numerical minority but
politically successful, courts may not step in. 67 On the other hand, the Court’s
1993 Voinovich decision tells us that if a protected class is “so numerous” in
one district that it dilutes minority voting power in another district, that’s
“packing” which requires judicial correction. 68 Add to that this decision: A
minority voting-age population slightly north of 60% is necessary to secure § 2
rights.
So putting all this together—a minority voting age population under 50%
is okay if the protected class is politically successful; but not okay if the
protected class is too much below 50% and politically unsuccessful; and still
not okay if the unsuccessful majority-minority VAP is above 50% but below
60%.
This blinkered focus on outcomes rather than opportunity clashes with
the VRA’s express text and relevant caselaw, both of which underscore
64 LULAC, 548 U.S. at 428.
65 Id. at 429.
66 Id.
67 Bartlett v. Strickland, 556 U.S. 1, 13 (2009).
68 Voinovich, 507 U.S. at 153.
73
No. 19-60133
electoral participation and opportunity—not electoral success. Again, § 2 offers
relief where it is shown that the political process preceding the election is “not
equally open to participation by members of a [protected class] . . . in that its
members have less opportunity than other members of the electorate to
participate in the political process and to elect representatives of their
choice.” 69 The majority opinion effectively requires electoral success. But the
statutes and caselaw focus on genuine opportunity. Not guaranteed victory.
For contrast, let’s revisit the Alabama district court case Merrill. The
plaintiffs adequately alleged vote dilution by claiming improper “cracking” and
“packing.” 70 They argued that increasing the Black Voting Age Population in
one district from 58.33% to 60.91% was “packing.” 71 Instead, the Legislature
should’ve created a second majority-minority district. 72
Now reconsider this case. Rather than argue that the BVAP is unfairly
diffused across several districts or unfairly packed into one, the plaintiffs
simply say that their district could be drawn to afford them a bigger majority
than they currently have. Ironically, under this logic, dividing an African-
American population into as many majority-minority districts as possible
would violate the VRA. Instead, the plaintiffs’ demands are tantamount to
guarantees of electoral wins in each such district.
The project of judicial oversight in § 2 cases is one of ensuring equal
opportunity, not dictating winners. It is unfitting for the judiciary to intervene
and put a thumb (or anvil) on the electoral scale where nondiscriminatory
factors may play a role in preventing a majority of a protected-class minority
from electing their preferred candidate. Absent evidence that the voting-age,
69 52 U.S.C. § 10301(b).
70 Merrill, 377 F. Supp. 3d at 1311.
71 Id.
72 Id.
74
No. 19-60133
protected-class majority status is hollow (e.g., LULAC 73) or illusory (e.g.,
Salas 74)—that is, the majority isn’t actually a majority at all—nothing in the
VRA confers special solicitude on a majority that happens to be a protected
class.
Article III judges should take Article I lawmakers at their word:
“[N]othing in [§ 2] establishes a right to have members of a protected class
elected in numbers equal to their proportion in the population.” 75 Every
American citizen, including members of protected classes, must compete for
electoral success. Yet the district court and today’s majority opinion favor
something approaching proportional representation, which the text of the VRA
categorically rejects. And it implies that any minority candidate who loses an
election in a majority-minority district can sue under § 2—before a single
judge, without an evidentiary hearing, and on the eve of key election
deadlines—to get the district lines redrawn to his advantage.
This misstep is compounded by the fact that the district court never
considered or addressed the impact of its actions on the neighboring district at
all. Redistricting requires a holistic view, including inter-district ripple effects.
The dissent from the motions-panel decision ably demonstrates why these
types of § 2 claims must examine the impact, including potential vote dilution,
on neighboring districts. 76 It is virtually impossible to discern why the district
court preferred one plan over another. Given the district court’s failure to
develop and explain the facts relevant to its adoption of Plan 1, including the
effect that boosting the BVAP of one district would have on the BVAP of
LULAC, 548 U.S. at 428–29.
73
Salas, 964 F.2d at 1547.
74
75 52 U.S.C. § 10301(b).
76 See Thomas v. Bryant, 919 F.3d 298, 316, 318 (5th Cir. 2019) (Clement, J.,
dissenting).
75
No. 19-60133
adjacent districts, how can there be meaningful appellate review? The district
court’s cursory statement that none of the plans “dilute minority voting
strength” cannot substitute for an actual, real-world assessment of the effect
of choosing this plan over others.
All in all, I share the substantive objections explained in the dissent to
the motions-panel’s refusal to stay the map ordered by the district court. 77 And
where “the ultimate finding of dilution” is based on “a misreading of the
governing law,” there is reversible error. 78
Beyond these § 2-related qualms, I have one more: the district court’s
overt race-conscious remedy. The district court—again, without a sufficient
evidentiary hearing, with no analysis, and just days before the qualifying
deadline—engaged in unconcealed race-based districting to boost the BVAP
from 50.8% to 60.91%. Such a remedy would seem to run headlong into the
Fourteenth Amendment.
A federal judge cannot command what the Constitution condemns. As
noted above, the Equal Protection Clause’s “central mandate is racial
neutrality in governmental decisionmaking,” including in the “drawing of
[legislative] districts.” 79 When the Department of Justice adopted a policy
requiring states to design districts based on race, the Supreme Court in Miller
v. Johnson excoriated it as not only “quite far removed from [the VRA’s]
purpose” but also potentially unconstitutional. 80 The Department of Justice’s
“implicit command that States engage in presumptively unconstitutional race-
based districting brings the [VRA] . . . into tension with the Fourteenth
77 See generally id. (Clement, J., dissenting).
78 De Grandy, 512 U.S. at 1022.
79 Miller v. Johnson, 515 U.S. 900, 904–05 (1995).
80 Id. at 926.
76
No. 19-60133
Amendment.” 81 “It takes a shortsighted and unauthorized view of the Voting
Rights Act to invoke that statute, which has played a decisive role in redressing
some of our worst forms of discrimination, to demand the very racial
stereotyping the Fourteenth Amendment forbids.” 82
The district court ignored these reprimands. It was concerned that
District 22’s BVAP was too low. Apparently believing that 61.98% is better
than 50.77%, the district court redrew the district. Even if the VRA required
race-based districting in this case, the district court didn’t explain why so much
racial discrimination was necessary. Nor did the district court explain how its
explicitly race-conscious remedy could comport with the Fourteenth
Amendment and the narrow tailoring commanded by Miller. This is hardly the
stuff of “close judicial scrutiny.” 83
IV
Finally, a vexing jurisdictional issue lingers: whether 28 U.S.C.
§ 2284(a), which requires the convening of a three-judge district court, applies
when a plaintiff confines his suit to § 2 alone. The statute reads:
A district court of three judges shall be convened . . . when an
action is filed challenging the constitutionality of the
apportionment of congressional districts or the apportionment of
any statewide legislative body. 84
There are two ways to interpret this:
• (1) the constitutionality of the apportionment of congressional
districts; or (2) the apportionment of any statewide legislative
body.
81 Id. at 927.
82 Id. at 927–28.
83 Shaw, 509 U.S. at 657.
84 28 U.S.C. § 2284(a) (emphasis added).
77
No. 19-60133
or
• (1) the constitutionality of the apportionment of congressional
districts; or (2) the constitutionality of the apportionment of
any statewide legislative body.
Citing recent Supreme Court dicta, 85 the majority opinion opts for the latter
interpretation.
The Supreme Court indeed suggested in dicta that § 2284(a) requires a
three-judge panel only for constitutional challenges. 86 But that dictum was a
mere explanatory parenthetical in a citation to the proceedings below. So
regardless of which is the better interpretation, there seems to be a live
statutory question. Yet the majority opinion deals the competing argument
short shrift.
The competing argument invokes the series-qualifier canon. 87 The
general rule under that canon: “When there is a straightforward, parallel
construction that involves all nouns or verbs in a series, a prepositive or
postpositive modifier normally applies to the entire series.” 88 The exception is
that a determiner repeated before the next element in the series generally
means no carryover modification.
The State contends that the exception applies. Why? The statute says:
the constitutionality of congressional districts or the state legislative bodies.
Not: the constitutionality of congressional districts or state legislative bodies.
85 Harris v. Ariz. Indep. Redistricting Comm’n, 136 S. Ct. 1301, 1306 (2016).
86 Id. (characterizing, in a single parenthetical, § 2284(a) as “providing for the
convention of such a court whenever an action is filed challenging the constitutionality of
apportionment of legislative districts”).
87 See generally ANTONIN SCALIA & BRYAN A. GARNER, READING LAW § 19 (2012).
88 Id. § 19, at 147.
78
No. 19-60133
Here, “the” is a determiner. It’s a subtle but important distinction. Consider
this excerpt from Reading Law:
To make certain that the postpositive modifier does not apply to each
item, the competent drafter will position it earlier:
• Societies that are charitable in nature or institutions.
• A fence that is solid or a wall.
• A partnership registered in Delaware or a corporation. 89
Now contrast that with these weaker yet synonymous examples:
• An institution or a society that is charitable in nature (any
institution probably qualifies, not just a charitable one).
• A wall or a fence that is solid (the wall may probably have
gaps).
• A corporation or a partnership registered in Delaware (the
corporation may probably be registered anywhere). 90
Then return to our statute, § 2284(a): “the constitutionality of the
apportionment of congressional districts or the apportionment of any statewide
legislative body.”
You’ll notice that the modifier—“constitutionality”—not only has an
intervening determiner, it’s also at the very beginning of the series. This
suggests that § 2284(a) might require the convening of three-judge district
courts even for purely statutory challenges to the apportionment of any
statewide legislative body.
That said, the series-qualifier canon is “highly sensitive to context”—
“[o]ften the sense of the matter prevails.” 91 It’s possible that the drafters may
89 Id. § 19, at 149.
90 Id.
91 Id. § 19, at 150. To muddy the linguistic waters further, perhaps the series-qualifier
canon doesn’t apply at all. Some readers might find it odd to describe the statute as creating
a “series” for which “constitutionality” is a modifier. Their instinct may be to ask instead
whether the “or” separates:
79
No. 19-60133
well have understood “constitutionality” to modify the entire sentence. It would
make sense, then, that no court has gone the other way on this issue in the
years since the law’s passage. 92 All to say, this is a nettlesome issue that
warrants closer study.
V
As noted above, I share the substantive critiques expressed in the dissent
to the motions-panel denial in March of the State’s stay request. 93 But one final
issue merits discussion: this court’s seeming inability to reconsider that stay
denial en banc. The plaintiffs regard the supposed impossibility of en banc
review as a feature, not a bug. I disagree. Indeed, this case illustrates the point.
The State has now lost in front of the merits panel, and neither this court
sitting en banc nor the Supreme Court ever had an opportunity to review the
divided motions-panel decision. In my view, this inflicts lasting harm upon
Article III and our federal structure.
Recap:
A single federal district judge purported to redraw the State of
Mississippi’s voting maps. Then a motions panel of this court denied the State’s
request for a stay. On March 26, 2019, the State introduced Joint Resolution
• two alternative objects of the preposition “of ” (“challenging the
constitutionality of [1] the apportionment of congressional districts or [2]
the apportionment of any statewide legislative body”), or
• two alternative objects of the participle “challenging” (“challenging [1] the
constitutionality of the apportionment of congressional districts or [2] the
apportionment of any statewide legislative body”).
92 Another factor complicating the interpretation of § 2284(a) is that it isn’t clear
anyone originally thought there could be non-constitutional challenges to apportionment. See
Holder v. Hall, 512 U.S. 874, 923 (1994) (Thomas, J., concurring in the judgment) (concluding
that §2 protects ballot access, not ballot weight, and that “vote dilution” claims fall outside
the text’s focus on “voting qualification or prerequisite to voting, or standard, practice, or
procedure”).
93 See generally Thomas, 919 F.3d at 316 (Clement, J., dissenting)
80
No. 19-60133
202. 94 That resolution expressly references the stay order three times. 95 The
State recognized that our order compelled it to conduct emergency
redistricting. It dutifully completed that task on March 29. Yet in passing the
joint resolution, Mississippi specified—repeatedly—that it would like to have
its old districts back if it prevailed on appeal. 96
One might ask, “Well, if a majority of the court thought the order denying
the stay was wrong, why didn’t they take it en banc and fix it?” Answer:
because a majority of our court thinks we are legally prohibited from taking an
“interim” decision en banc.
Where might such a prohibition come from? It’s certainly not in the
Federal Rules of Appellate Procedure. Rule 35 says a majority of a circuit’s
judges “may order that an appeal or other proceeding be heard or reheard by
the court of appeals en banc.” 97 And our operating procedures say “[a]ny active
member of [our] court . . . may request a poll,” whether or not the parties ask
for one themselves. 98 It’s hard to imagine how the proceedings before the panel,
which as a matter of constitutional necessity comprised a case or controversy, 99
could not also be a “proceeding” under Rule 35. And they’re certainly part of
the larger “appeal” currently before our court. 100
Perhaps that’s why our fellow circuit courts routinely entertain en banc
petitions over appeals or proceedings like this one. 101 If the Federal Rules
94 S.J. Res. 202, 2019 Reg. Sess. (Miss. 2019).
95 Id. at 1, 24, 25.
96 See id. at 24–25.
97 FED. R. APP. P. 35(a) (emphasis added).
98 5TH CIR. I.O.P. foll. 35.6.
99 U.S. CONST. art. III, § 2, cl. 1.
100 See Thomas v. Bryant, No. 19-60133, ECF No. 1 (5th Cir. Mar. 1, 2019).
101 See Feldman v. Ariz. Sec’y of State’s Office, 843 F.3d 366, 367 (9th Cir. 2016) (en
banc) (motion to stay pending appeal); In re Lombardi, 741 F.3d 888, 892–93 (8th Cir. 2014)
(en banc) (motion to vacate discovery orders); Teshome-Gebreegziabher v. Mukasey, 545 F.3d
285, 286 (4th Cir. 2008) (en banc) (Shedd, J., concurring in the denial of rehearing en banc)
81
No. 19-60133
prohibited en banc consideration of such motions, the prohibition would apply
in our sister circuits as much as it would in ours. But it doesn’t.
One thing that might differentiate our court from others is our non-
public “Internal Court Policies.” Might those prohibit us from reconsidering an
“interim” decision en banc? Again, no. I cannot quote our internal policies
because they are stamped: “NOT FOR PUBLIC DISTRIBUTION.” But I don’t
need to quote them to prove nonpublic “rules” aren’t rules at all. That’s partly
why the Supreme Court once vacated a Ninth Circuit decision “refus[ing]
completely to consider the merits of [an] en banc request.” 102 Because the Ninth
Circuit’s en banc procedures were not “clearly explained” for “the members of
the court and litigants,” the Supreme Court reinstated the party’s en banc
petition. 103 If our private rule really is the rule, it’s not explained at all—much
less clearly. A rule we apply and yet never disclose is worse than a rule posted
“high on the columns so that [it] would be harder to read and easier to
transgress.” 104
Even if our private policy purported to prevent us from reconsidering an
“interim” decision en banc, we’d be prohibited from following it. The other half
of the Supreme Court’s holding in Western Pacific shows why: Although the
circuit courts have wide latitude in establishing their en banc procedures, at
an irreducible minimum it is “essential that litigants be left free to suggest to
the court . . . that a particular case is appropriate for consideration by all the
judges.” 105 Thus, completely independent of the Ninth Circuit’s obligation to
(motion to stay pending removal); Rakovich v. Wade, 834 F.2d 673, 673–74 (7th Cir. 1987)
(en banc) (motion to stay pending appeal).
102 W. Pac. R.R. Corp. v. W. Pac. R.R. Co., 345 U.S. 247, 265–67 (1953).
103 Id. at 266–67.
104 Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. CHI. L. REV. 1175, 1179
(1989) (describing “Nero’s nasty practices”).
105 345 U.S. at 261 (emphasis added).
82
No. 19-60133
establish clear rules, whatever rules that court adopted could not prevent the
mere suggestion that a given case “is an appropriate one for the exercise of the
[en banc] power.” 106 So too here. 107
All of this makes complete sense if we consider the premise that
underlies review by something less than the full court. Whether sitting as a
three-judge panel or as a single judge, whenever the judges of this court act
(other than when we’re sitting en banc), we act as agents of the full court. 108 A
panel’s decision is “no more than an initial exercise of the power that
ultimately resides only with the whole court.” 109
That accords with the notion that our panels can issue binding
precedent. Three judges, who act on behalf of all of us, can issue a decision that
binds all of us precisely because there is a mechanism that allows the full Court
to revisit that decision. 110 As wrong as it is, the order denying a stay in this
case is now published in the Federal Reporter, and it establishes Fifth Circuit
precedent on the workings of stay applications. 111 But if the full court cannot
take it en banc, then perhaps the other members of this court—who never had
the opportunity to reject or bless what the panel did here—should not be bound
106 Id. at 268.
107 See BRYAN A. GARNER ET AL., THE LAW OF JUDICIAL PRECEDENT 496 (2016)
(“Although a court of appeals is almost never required to grant en banc consideration, it
cannot eliminate en banc review.” (emphasis added)).
108 See Douglas H. Ginsburg & Donald Falk, The Court En Banc: 1981–1990, 59 GEO.
WASH. L. REV. 1008, 1011–13 (1991); Tom S. Clark, A Principal-Agent Theory of En Banc
Review, 25 J.L. ECON. & ORG. 55, 76 (2008).
109 16 CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE & PROCEDURE § 3981.3 n.7
(4th ed. 2018).
110 See United States v. Am.-Foreign S.S. Corp., 364 U.S. 685, 689–90 (1960); 28 U.S.C.
§§ 43(b), 46(c), 451.
111 See generally Thomas, 919 F.3d 298.
83
No. 19-60133
by it in a future case. 112 Which, of course, invites chaos as each motions panel
becomes a law unto itself.
If that weren’t enough, consider the source of the panel’s authority to do
what it did here. Appellate Rule 8 does not expressly confer power; it regulates
a power conferred somewhere else. The power to stay a district court decision
pending appellate review derives from the All Writs Act. 113 As an en banc court,
we have revisited panel decisions exercising (or refusing to exercise) our All-
Writs power before. 114 How odd if we lack the ability to consider, as an en banc
court, exercises of the same power—simply because they involve a motion for
a stay.
I can imagine two responses. First, some might worry that permitting
review of “interim” decisions opens the floodgates for en banc review of trivial
matters. I agree with half of that. Most of the procedural matters the judges of
this court decide on a daily basis do not merit en banc rehearing. But that
proves too much. There is no serious risk that we will begin granting en banc
petitions to rehear a decision denying an extension of time to file a brief. Self-
discipline still counts for something.
Second, some might wonder if this is really that big of a deal given the
State chose emergency redistricting over petitioning our court for rehearing.
We should hardly be surprised by that. As the dissent from the stay denial
112 See N.C. Utils. Comm’n v. FCC., 552 F.2d 1036, 1045 (4th Cir. 1977) (concluding
three-judge panel could ignore prior panel’s opinion where en banc review was impossible);
see also id. at 1056 n.1 (Widener, J., dissenting) (agreeing because “[o]therwise, many
subjects would be relatively frozen”).
113 See Nken v. Holder, 556 U.S. 418, 426–27 (2009) (citing 28 U.S.C. § 1651(a)); United
States v. Lynd, 301 F.2d 818, 819 (5th Cir. 1962); FED. R. APP. P. 8 advisory committee note
(1967).
114 See, e.g., In re Unknown, 701 F.3d 749 (5th Cir. 2012) (en banc) (considering
whether to grant writ of mandamus); In re Volkswagen of Am., Inc., 545 F.3d 304 (5th Cir.
2008) (en banc) (same); United States v. Denson, 603 F.3d 1143 (5th Cir. 1979) (en banc)
(same).
84
No. 19-60133
pointed out, we apparently told the parties they cannot seek en banc
reconsideration of a motion panel’s denial of a stay. (Though we’ve never
explained why.) That left only one option—conducting an emergency
redistricting proceeding under the Damoclean sword of a federal court order.
In my view, that’s a shabby way to treat a sovereign State. And in all events,
even if the State didn’t want us to take this case en banc, we have an
independent obligation to get the law right. Every time. That’s why, at the
dawn of en banc rehearing practice, judges routinely initiated the call for en
banc rehearing. 115
I firmly believe every one of my colleagues strives to get the right answer
when he or she sits on a panel or rules on a single-judge motion. But we all
make mistakes. By not taking the stay denial en banc, we shirked our duty to
correct a serious mistake. Not because this case is unimportant, and not
because we’re unconvinced a grievous mistake was committed. But rather
because we’ve convinced ourselves that our hands are tied. I hope, in time, we’ll
recognize that’s a mistake too.
* * *
I deeply admire my colleagues’ solicitude for the Voting Rights Act. More,
I share it. But respectfully, I believe the majority opinion oversteps (our
judicial power in election cases), overlooks (the VRA’s opportunity-not-
outcomes emphasis on racial neutrality), and overrides (the State of
Mississippi’s broad sovereign power to regulate its elections within the bounds
of the law).
Allowing the district court’s legally imaginative analysis to take
precedential root invites more of the same. Lawyers are nothing if not
entrepreneurial. In the wake of the Supreme Court’s Shelby County decision
115 See Ginsburg & Falk, supra note 107, at 1013.
85
No. 19-60133
in 2013, scholars swiftly urged “courts, in partnership with the Department of
Justice, [to] reform section 2 so that it fills much of the gap left by the Supreme
Court’s evisceration of section 5.” 116 Certainly, this district court’s novel take
on § 2 will, unless checked, induce other last-minute—and thus perhaps
unreviewable—redistrictings. The specter of such gamesmanship is precisely
why the Supreme Court has warned against eleventh-hour interference with
time-sensitive election processes.
Summing up:
The district court should never have heard this case in the first place.
The Supreme Court consistently denounces such last-minute tinkering, and
laches is itself determinative. As for the knotty jurisdictional question, I
concede uncertainty. But not on the § 2 merits, where the district court, rather
clearly, prescribed what is proscribed.
Respectfully, I dissent.
116Christopher S. Elmendorf & Douglas M. Spencer, Administering Section 2 of the
Voting Rights Act After Shelby County, 115 COLUM. L. REV. 2143, abstract (2015). This article
urges courts to give § 2 “special bite” such that “section 2 can be made to function like
erstwhile section 5 in the post-Shelby County world.” Id. at 2147.
86