Joseph Thomas v. Phil Bryant

Opinion

Case: 19-60133 Document: 00514887045 Page: 1 Date Filed: 03/25/2019

REVISED March 25, 2019

IN THE UNITED STATES COURT OF APPEALS FOR THE FIFTH CIRCUIT United States Court of Appeals Fifth Circuit

___________________ FILED March 22, 2019 No. 19-60133 Lyle W. Cayce ___________________ Clerk

JOSEPH THOMAS; VERNON AYERS; MELVIN LAWSON,

Plaintiffs - Appellees

v.

PHIL BRYANT, Governor of the State of Mississippi,; DELBERT HOSEMANN, Secretary of State of the State of Mississippi, both 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 DENNIS, CLEMENT, and COSTA, Circuit Judges. GREGG COSTA, Circuit Judge: A district court found that the boundaries for Mississippi State Senate District 22 dilute African-American voting strength and prevent those citizens from having the equal opportunity “to participate in the political process and to elect representatives of their choice” that the Voting Rights Act guarantees. 52 U.S.C. § 10301(b). To remedy the violation, the district court switched 28 Case: 19-60133 Document: 00514887045 Page: 2 Date Filed: 03/25/2019

No. 19-60133 precincts between District 22 and a bordering district. The Governor and Secretary of State seek a stay of the final judgment. I. Most of District 22 lies in the heart of the Mississippi Delta. Those unfamiliar with the state’s geography might think that the Mississippi Delta is in the southern portion of the state that is closer to the mouth of the Mississippi River. But the “Delta” refers not to the outlet of that mighty river into the Gulf of Mexico, but to the alluvial plain between the Mississippi River and the Yazoo River. John M. Barry, RISING TIDE: THE GREAT MISSISSIPPI FLOOD OF 1927 AND HOW IT CHANGED AMERICA 96 (1997). It is located in the northwest section of Mississippi and looks like an “elongated diamond.” Id. The district court recited this colorful description of its boundaries: “The Mississippi Delta begins in the lobby of the Peabody Hotel [in Memphis] and ends on Catfish Row in Vicksburg.” David L. Cohn, GOD SHAKES CREATION (1935). The sediment that the Mississippi River deposited in this region over the millennia made Delta soil some of the richest in the world. Barry, supra, at 97. And that soil made the Delta one of the leading cotton-producing areas in the world. One historian compared the influence the Delta once held over global cotton prices to the modern influence of Saudi Arabia over oil prices. Sven Beckert, EMPIRE OF COTTON: A GLOBAL HISTORY 113 (2014). But before agriculture could prosper in the Delta its forests of hardwood trees had to be cleared. Slaves were forced to begin that process, and emancipated slaves finished it in the latter part of the nineteenth century. Since that time, African-Americans have been a majority in the Delta. In the early twentieth century, they made up almost 90% of the region’s population. Barry, supra, at 125 (noting that in 1908 the Delta had “a black population of

2 Case: 19-60133 Document: 00514887045 Page: 3 Date Filed: 03/25/2019

No. 19-60133 at least 171,209” and “a white population of 24,137”). The Great Migration to the north reduced the African-American percentage, but they remain a sizeable majority in the region. Of course, for most of their time laboring in the Delta, African-Americans could not vote. The Voting Rights Act changed that. It has resulted in numerous African-Americans being elected to office in the Delta. Indeed, the State of Mississippi has more African-American elected officials than any other state. U.S. Census Bureau, STATISTICAL ABSTRACT OF THE UNITED STATES: 2011, 258 (130th ed.). But in places like the state legislature, African-Americans have still not achieved political power in Mississippi that comes close to their share of the population. This case involves a claim that one state senate district contributed to that underrepresentation by diluting the voting strength of African-Americans that Delta demographics and geography should otherwise support. District 22 includes all or part of five Delta counties: Bolivar, Humphreys, Sharkey, Washington, and Yazoo. It also includes parts of one non-Delta county: Madison. The addition of the white-majority Madison County precincts in the last redistricting reduced the African-American percentage in District 22. Overall the district still has an African-American majority, but just barely at 50.7%. 1 Without the non-Delta precincts from

1 Plaintiff Thomas told the Department of Justice that the 50.7% figure includes the population of the Yazoo City federal prison, the majority of whose inmates he says are African-American. He contended that without including that prison, African-Americans would not be a majority in the district. It is questionable whether disenfranchised people should be included in the BVAP (black voting age population) calculation for purposes of the Voting Rights Act. The focus is usually on those eligible to vote, thus the typical requirement in our circuit that the percentage focus on those of voting age who are citizens. See Reyes v. City of Farmers Branch, Tex., 586 F.3d 1019, 1023 (5th Cir. 2009) (noting that only citizens are included in population calculations). 3 Case: 19-60133 Document: 00514887045 Page: 4 Date Filed: 03/25/2019

No. 19-60133 Madison County, that percentage would be much higher. The candidate preferred by the black majority did not prevail in the 2015 election. The losing candidate those voters did prefer (Joseph Thomas), along with two voters (Vernon Ayers and Melvin Lawson), brought this lawsuit in 2018 alleging that the district violated of Section 2 of the Voting Rights Act. Last month, the district court held a bench trial and agreed. It concluded that the threshold Gingles factors for establishing vote dilution existed. The African-American population is “sufficiently large and geographically compact” to constitute a majority with electoral power in the district; that racial group is politically cohesive; but the white population voted as a bloc to prevent African-Americans from electing their chosen candidate. LULAC v. Perry, 548 U.S. 399, 425 (2006). The bloc voting by both racial groups was stark: in 10 elections within district 22, the candidate favored by African-Americans received between 82% and 93% of their vote but only between 8% and 19% of the white vote. The district court then followed the statutory directive to consider the “totality of circumstances” in determining whether African-Americans “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” 52 U.S.C. § 10301(b); LULAC, 548 U.S. at 425–26. It credited Plaintiffs’ expert’s analysis of voter participation as more rigorous and concluded that African-American turnout was depressed in off-year elections—that is, odd-year elections without federal races—which is when Mississippi elects its legislature. The district court also highlighted the persistent representation gap in the Mississippi Senate, as well as substantial socioeconomic differences between black voters and white voters in the district. After finding a Voting Rights Act violation, the court initially declined to

4 Case: 19-60133 Document: 00514887045 Page: 5 Date Filed: 03/25/2019

No. 19-60133 order a remedy so “the Mississippi Legislature [could] consider[] whether to redraw the District and extend the candidate qualification deadline.” Although the primary and general elections were months away, the deadline for candidate filing was March 1. A flurry of procedural moves followed at warp speed. Three days after the court made its trial findings, the Governor and Secretary of State (Defendants) 2 filed a motion a motion to stay with the district court and also filed an appeal, arguing that the ruling was an appealable interlocutory injunction. Less than a week later, they asked this court to stay the district court’s ruling. That same day, Plaintiffs asked the court to extend the qualification deadline in District 22 and bordering District 23 (the other district affected by the proposed remedy) by two weeks. Later that day, the district court ordered Defendants to update it on any progress, if any, in legislative redrawing of District 22. The next day, Defendants filed a brief opposing Plaintiffs’ motion to extend the qualification deadline, in which they stated that they had contacted the leadership of the legislative chambers. The Senate leaders had said that “should the stay motions pending before [the district court] and the Fifth Circuit be denied, the Senate desire[d] the opportunity to enact a new redistricting plan redrawing Senate District 22.” Late that same day, the district court ruled that, because the legislature was unwilling to act, the first map drawn by one of Plaintiffs’ experts would be adopted as the boundaries of Districts 22 and 23. The court also moved the qualification deadline to March

2 The third defendant, Attorney General Jim Hood, did not join these motions or this appeal. 5 Case: 19-60133 Document: 00514887045 Page: 6 Date Filed: 03/25/2019

No. 19-60133 15 for the two affected senate districts. The court also declined to enter a stay and entered final judgment. The final judgment caused our court to dismiss the appeal of the preliminary ruling. Thomas v. Bryant, 2019 WL 994034 (5th Cir. Feb. 28, 2019). But by that time the Defendants had already appealed from the final judgment, and not long after that they sought a stay pending the new appeal, which is the motion before this panel. 3 This is the effect of the final judgment: It switches some precincts between Districts 22 and 23. It extends the candidate filing deadline to March 15. In terms of other upcoming deadlines, the primary election is August 6, with a June 22 deadline to send ballots to troops overseas. The general election is November 5. II. The stay Defendants seek is an extraordinary remedy. Nken v. Holder, 556 U.S. 418, 437 (2009) (Kennedy, J., concurring). It is also an equitable one committed to this court’s discretion. Id. at 433; Ruiz v. Estelle, 666 F.2d 854, 856 & n.4 (5th Cir. 1982). Four factors guide the exercise of our discretion: 1) whether the applicant has made a strong showing of likelihood to succeed on the merits; 2) whether the movant will be irreparably harmed absent a stay; 3) whether issuance of a stay will substantially injure other interested parties; and 4) where the public interest lies. Nken, 556 U.S. at 434. The first two factors are usually the most important. Id.; ODonnell v. Goodhart, 900 F.3d 220, 223 (5th Cir. 2018). In this case, the likelihood of success ends up being dispositive. That is because Defendants can establish irreparable harm. A court order preventing

3 Before seeking the stay pending appeal for the second time in our court after entry of final judgment, Defendants also sought a stay in the district court for the second time. The district court denied that motion. 6 Case: 19-60133 Document: 00514887045 Page: 7 Date Filed: 03/25/2019

No. 19-60133 enforcement of a state law, including the drawing of legislative lines, constitutes that harm. Abbott v. Perez, 138 S. Ct. 2305, 2324 n.17 (2018). And that injury may be irreparable in light of the meaningful possibility (but not certainty) that a full appeal cannot be decided in time to provide Defendants relief before this year’s senate elections. But Plaintiffs face the same risk that the appellate ruling will prove futile should this court grant a stay. And the injury they seek to prevent—holding an election under an unlawful plan with discriminatory effects—is also, it should go without saying, a serious one. Veasey v. Abbott, 830 F.3d 216, 270 (5th Cir. 2016) (en banc) (“It would be untenable to permit a law with a discriminatory effect to remain in operation for that election.”); cf. Reynolds v. Sims, 377 U.S. 533, 585 (1964) (“[I]t would be the unusual case in which a court would be justified in not taking appropriate action to insure that no further elections are conducted under the invalid plan.”). So Defendants’ entitlement to a stay turns on whether they have a strong likelihood of success. III. We will first address Defendants’ attempt to make that showing with a jurisdictional argument. They contend that a three-judge court should have decided this case. The statute says that 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 U.S.C. § 2284(a). The Voting Rights Act does not require three-judge courts for Section 2 cases (notably it does for Section 5 cases, 52 U.S.C. § 10304(a)), so Defendants focus on the latter part of the statute. They contend that “constitutionality” modifies only challenges to apportionment of congressional districts, not challenges to apportionment of state legislatures. On their

7 Case: 19-60133 Document: 00514887045 Page: 8 Date Filed: 03/25/2019

No. 19-60133 reading, any challenge to state legislative redistricting—including the statutory Voting Rights Act one asserted here—requires a three-judge panel. Defendants requested a three-judge panel late in the litigation, a week before trial. Failing to request such a panel in their answer may have forfeited the issue unless it is a jurisdictional requirement. The three-judge statute has conflicting language on whether it is one. It begins by saying “a district court of three judges shall be convened” for specified cases. 28 U.S.C. § 2284(a). That mandatory language sounds jurisdictional. So does the statute’s prohibition on a single judge taking certain actions, including granting injunctive relief. Id. § 2284(b)(3). But the statute also says that the procedure for convening a three-judge court kicks in “[u]pon the filing of a request for three judges.” Id. § 2284(b)(1). That does not sound jurisdictional. If the requirement is jurisdictional, then the single district judge first assigned the case would seemingly have an obligation to convene a three-judge panel even if a party does not ask for it. Recognizing these conflicting statutory signals, the leading federal procedure treatise is ambivalent about whether Section 2284 is jurisdictional. See 17A Charles Alan Wright et al., FED. PRAC. & PROC. § 4235, at 206–08 (3d ed. 2007). But apparently the courts that have considered the question— including ours in a nonbinding decision—have treated Section 2284 as jurisdictional. See Kalson v. Patterson, 542 F.3d 281, 287 (2nd Cir. 2008); Armour v. Ohio, 925 F.2d 987, 989 (6th Cir. 1991) (en banc); LULAC of Texas v. Texas, 318 F. App’x 261, 264 (5th Cir. 2009). That consensus in the caselaw means that forfeiture is an obstacle Defendants will likely be able to overcome on appeal. Defendants’ delay in raising this issue does, however, say something about its merits. At the outset of this lawsuit, it was not obvious to Defendants

8 Case: 19-60133 Document: 00514887045 Page: 9 Date Filed: 03/25/2019

No. 19-60133 (or Plaintiffs or the district judge for that matter) that its exclusively statutory claims required a three-judge court. Indeed, no reported case has ever used a three-judge panel for a case challenging district lines only under Section 2 of the Voting Rights Act. See Rural W. Tenn. African-American Affairs Council v. Sundquist, 209 F.3d 835, 838 (6th Cir. 2000) (noting case reassigned to single judge after dismissal of constitutional and Section 5 claims); Chestnut v. Merrill, 2019 WL 338909 (N.D. Ala. 2019) (rejecting argument that a single judge could not hear Section 2 challenge); Bone Shirt v. Hazeltine, 336 F. Supp. 2d 976, 980 (D. S.D. 2004) (same as Rural West); Old Person v. Brown, 182 F. Supp. 2d 1002, 1003 (D. Mont. 2002) (single judge hearing Section 2 challenge); see also Armour, 925 F.2d at 989 (stating the jurisdictional test for Section 2284 as whether “there exists a non-frivolous constitutional challenge to the apportionment of a statewide legislative body” (emphasis added)). 4 That no court has adopted Defendants’ reading of the three-judge statute in the more than four decades it has been on the books alone seems enough to prevent them from showing a strong likelihood of succeeding on this issue. To be sure, only the Chestnut opinion from earlier this year addressed an argument like the one Defendants make. But that is again telling: Before this year, it apparently had never dawned on a judge or party in a Section 2-only state redistricting case

4 Though uncited by Defendants or their amicus, the dissent finds Page v. Bartels, 248 F.3d 175 (3d Cir. 2001), “most significant[].” Dissenting Op. at 15. But unlike the cases cited above, Page does not involve a situation in which plaintiffs only brought Section 2 claims. Instead, it involved both constitutional and statutory claims. Yet a single district court judge attempted to spin off the statutory claims and deny an injunction based on them. Id. at 187. The Third Circuit said that when plaintiffs bring both claims, both must be sent to a three- judge panel. Id. at 191. That is consistent with the common practice when both constitutional and statutory challenges to reapportionment are brought—the constitutional hook for three- judge jurisdiction sweeps in the statutory claims. While the dissent tries to find support for its position in some of Page’s discussion of legislative history and policy, notably the Third Circuit does not adopt the textual argument the dissent and amicus push.

9 Case: 19-60133 Document: 00514887045 Page: 10 Date Filed: 03/25/2019

No. 19-60133 that a three-judge panel might be required. And that is because the most straightforward reading of the three-judge statute is that it applies only when the “constitutionality” of apportionment is being challenged. So why does the dissent, adopting the argument of an amicus, think the three-judge statute applies to what is only a statutory challenge when no court has ever taken that view? One word—the “the” that precedes “apportionment of any statewide legislative body.” 28 U.S.C. § 2284(a). Inclusion of that article means, in their view, that “constitutionality” applies only to what immediately follows it: challenges to the “apportionment of congressional districts.” If “constitutionality” does not also modify the “apportionment of any statewide legislative body,” id., then any challenges to those districts, including statutory ones, require a three-judge panel. The textual argument is a dispute about application of the “series modifier” canon of construction. It normally means that a modifier (“constitutionality” in Section 2284) applies to an entire series of parallel terms. See Scalia & Garner, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 147 (2012). The district court concluded that principle supports reading the statute as applying to constitutional challenges to apportionment of both congressional and state legislative seats. But an amicus supporting the defendants argues the “the” coming before “apportionment of any statewide legislative body” breaks the series modifier. The amicus notes that a “determiner” word, such as the “the” in Section 2284(a), is an indication that a modifier should not reach the second element. 5 See id. at 148–49.

5 Defendants make a different textual argument in support of their position that this case belongs with a three-judge court. They focus not just on the “the,” but on “the apportionment” that precedes “of any statewide legislative body.” They contend it is superfluous if it does not break off the “constitutionality” modifier. But it may be “appropriate to tolerate a degree of surplusage.” United States v. Atl. Research Corp., 551 10 Case: 19-60133 Document: 00514887045 Page: 11 Date Filed: 03/25/2019

No. 19-60133 But inferring that “the” was meant to cut off the “constitutionality” modifier ignores what may be the most important teaching about the series modifier canon from the Scalia and Garner book on statutory interpretation: “Perhaps more than most of the other canons, this one is highly sensitive to context.” Id. at 150. That context supports the natural reading that courts have long given it: that “constitutionality” modifies both “the apportionment of congressional districts” and “the apportionment of any statewide legislative body.” That Congress wanted the two phrases to sound the same—“the apportionment of congressional districts” and “the apportionment of any statewide legislative body” are word-for-word identical, with the exception of the type of district that they reference—is consistent with how ordinary people speak and write. See Lockhart v. United States, 136 S. Ct. 958, 970 (2016) (Kagan, J., dissenting). To illustrate this, consider what a reader would think after seeing the following in the newspaper: “The NCAA is investigating the recruiting practices of the football program and the basketball program.” As with the three-judge statute, the final “the” may not be necessary. But would it make the reader think the investigation into the basketball program is not limited to recruiting violations, but also might include point shaving or ticket scalping violations? Of course not. The dissent’s textual argument about the three-judge statute is just as detached from common usage. What is more, giving so much significance to “the” runs counter to another point Justice Scalia made about statutory interpretation: that Congress “does not . . . hide elephants in mouseholes.” Whitman v. Am. Trucking Ass’n, 531 U.S. 457, 468 (2001). The historical development of the three-judge statute—that is, statutory history as opposed to the more

U.S. 128, 137 (2007). Indeed, one could make the case that even the first “the,” the one before “apportionment of congressional districts,” is unnecessary. 11 Case: 19-60133 Document: 00514887045 Page: 12 Date Filed: 03/25/2019

No. 19-60133 controversial legislative history Defendants rely on—shows why it would have been so significant and anomalous to require three-judge panels for statutory claims. The statute was first enacted in the aftermath of Ex parte Young to require three judges to hear what Congress thought would be an increasing number of suits challenging state laws “upon the ground of the unconstitutionality of such statute.” 28 U.S.C. § 2281 (1970); see generally David P. Currie, The Three-Judge District Court in Constitutional Litigation, 32 U. CHI. L. REV. 1, 5–8 (1964). When courts later struck down many New Deal reforms, one of the only aspects of President Roosevelt’s court-packing plan to become law was a measure also requiring three-judge panels for suits seeking to enjoin “any Act of Congress for repugnance to the Constitution of the United States.” 28 U.S.C. § 2282 (1970); see Wright, FED. PRAC. & PROC. § 4234, at 194–95. Both laws’ focus on only constitutional challenges made sense as striking down democratically enacted laws is “the gravest and most delicate duty” courts are “called on to perform.” Nw. Austin Mun. Util. Dist. No. One v. Holder, 557 U.S. 193, 204 (2009) (referring to striking down an Act of Congress (quoting Blodgett v. Holden, 275 U.S. 142, 147–48 (1927) (Holmes, J., concurring))). The idea of requiring three judges for this “class of cases of special importance” was to “assure more weight and greater deliberation by not leaving the fate of such litigation to a single judge.” Phillips v. United States, 312 U.S. 246, 249–50 (1941) (first quotation from Ex parte Collins, 277 U.S. 565, 567 (1928)). But in the mid-1970s, Congress decided to scrap just about all of the three-judge regime because it was burdening the Supreme Court and lower courts and had resulted in procedural complexities. See Wright, FED. PRAC. & PROC. § 4234, at 195–98; Kalson, 542 F.3d at 287 (noting that the 1976 Act “vastly reduced the category of cases for which a three-judge court is mandated”). It nonetheless decided to retain the procedure for a small

12 Case: 19-60133 Document: 00514887045 Page: 13 Date Filed: 03/25/2019

No. 19-60133 set of important cases: constitutional challenges to redistricting for congressional and state legislative seats, then-recent phenomena in the aftermath of the revolutionary one person, one vote line of cases. See, e.g., Reynolds v. Sims, 377 U.S. 533 (1964). If Congress had intended to expand the statute it was otherwise contracting by applying it for the first time to statutory challenges there were certainly better ways to do it than by stealthily inserting an extra “the.” Reading the statute in the way Defendants suggest does not make sense for another reason: Why would Congress require three judges to hear statutory claims challenging state legislative redistricting but not congressional redistricting? The dissent posits, without citing any authority, that perhaps federalism concerns would explain such a difference. But there are also strong federalism concerns with how a state chooses to divvy up its citizens into congressional districts. In fact, contemporary critics of the initial judicial foray into review of legislative apportionment viewed courts’ redrawing of congressional district as even more intrusive on traditional state prerogatives than judicial redrawing of state legislative districts. That is because of the view that Article I, Section II of the Constitution grants “States . . . plenary power to select their allotted Representatives in accordance with any method of popular election they please, subject only to the supervisory power of Congress.” Wesberry v. Sanders, 376 U.S. 1, 23 (1964) (Harlan, J., dissenting). The other problem with the dissent’s theory—that Congress in 1976 had a special concern with statutory challenges to the drawing of state legislative districts—is that it fails to grapple with the rarity of Section 2 challenges to redistricting at that time. It was then not even clear that Section 2 provided a private right of action. City of Mobile v. Bolden, 446 U.S. 55, 60 & n.8 (1980) (assuming without deciding such a suit could exist). A few years later the

13 Case: 19-60133 Document: 00514887045 Page: 14 Date Filed: 03/25/2019

No. 19-60133 Supreme Court held Section 2 did not prohibit discriminatory effects which meant it provided no guarantee beyond what the Constitution already did. See id. And the typical pre-1982 Section 2 claim that courts did consider was a challenge to at-large electoral systems, not the line drawing of single member districts. See, e.g., id. In other words, there was no practice of statutory challenges to state legislative apportionment that Congress needed to address in 1976. To sum up, the dissent would give this much weight to the “the” that comes before “reapportionment of any statewide legislative body”: Insertion of that article would require three-judge panels for exclusively statutory claims— followed by direct appeal to the Supreme Court, 28 U.S.C. § 1253—when the three-judge regime Congress was contracting in 1976 never did. It would require those three-judge panels only for statutory challenges to apportionment of state legislative seats, not congressional ones. And it would do all this to address statutory challenges to apportionment of state legislatures when those claims hardly existed in 1976. An elephant indeed. The dissent contends that there might have been clearer ways to emphasize that the statute is limited to constitutional challenges. Dissenting Op. at 12. But if Congress intended to include statutory-only challenges within the reach of the general three-judge statute for the first time, there were many easy ways to make that clear. This brings up a final point on the jurisdictional question. To the extent there is ambiguity, the longstanding principle that “congressional enactments providing for the convening of three-judge courts must be strictly construed” comes into play. Allen v. State Bd. of Elections, 393 U.S. 544, 561 (1969); Phillips, 312 U.S. at 250 (noting the three-judge statue should be read narrowly because of its impact on the Supreme Court docket and the burdens of convening three-judge courts).

14 Case: 19-60133 Document: 00514887045 Page: 15 Date Filed: 03/25/2019

No. 19-60133 That rule of construction, the text of the three-judge statute, its lineage, and the caselaw applying it all favor the district court’s view that three judges are not required for a claim raising only statutory challenges to state legislative redistricting. Defendants thus face a steep climb in convincing this court to be the first to ever read the statute as applying to a case challenging a district’s lines only under Section Two of the Voting Rights Act. IV. A. Defendants also seek a stay on the ground that we are likely to reverse the district court’s ruling that District 22 dilutes the voting power of African- Americans. The standard of review poses an obstacle to undoing the factual findings that led the court to conclude that the district violates Section 2. An appellate court “will not set aside” Section 2 factual findings “unless [it is] left with the definite and firm conviction that a mistake has been committed.” Magnolia Bar Ass’n, Inc. v. Lee, 994 F.2d 1143, 1147 (5th Cir. 1993) (quotation omitted); see NAACP v. Fordice, 252 F.3d 361, 365 (5th Cir. 2001) (noting that clear error review applies). In an attempt to avoid the clear-error standard, Defendants frame their primary challenge to liability as a legal one. They contend that a court cannot find vote dilution of a racial group that makes up a majority of the challenged district (recall that here that is just barely the case for District 22). This per se rule they advocate—a bar on vote dilution claims whenever the racial group crosses the 50% threshold—is at odds with the reason we give such deference to factual findings in Section 2 cases: the question of vote dilution is “peculiarly dependent upon the facts of each case,” Thornburg v. Gingles, 478 U.S. 30, 79 (1986) (quoting Nevett v. Sides, 571 F.2d 209, 224 (5th Cir. 1978)), and “requires an ‘intensely local appraisal’ of the challenged district,” LULAC, 548

15 Case: 19-60133 Document: 00514887045 Page: 16 Date Filed: 03/25/2019

No. 19-60133 U.S. at 437 (quoting Gingles, 478 U.S. at 79). The numerous local factors courts consider—including the strength of bloc voting, demographics, and voter turnout—may mean that a minority group has the equal electoral opportunity the Voting Rights Act guarantees when its members make up only 40% of a district, but not when they make up 51%. Applying Defendants’ absolute bar on Section 2 scrutiny of 50%-plus districts to this case shows how it would arbitrarily override the numerous local considerations that Congress and the Supreme Court have said courts need to consider in this important area. If there were a few hundred fewer African-Americans out of the nearly 44,000 voting-age citizens in District 22, Defendants would allow courts to engage in the regular multifaceted vote dilution inquiry. But, to Defendants, just over 300 additional African-Americans makes the difference between that ordinary inquiry and removing Section 2’s protections for tens of thousands of Mississippi voters. Given the statutory mandate to focus on the “totality of circumstances” in Section 2 cases, 52 U.S.C. § 10301(b), it is not surprising that numerous courts have found dilution of the voting power of a racial group in districts where they make up a majority of the voting population. Most importantly, our court has long rejected the argument Defendants advance. In an early vote dilution case—also out of the Mississippi Delta—we explained that the “mere existence of a black population majority does not preclude a finding of dilution.” Moore v. Leflore Cty. Bd. of Election Comm’rs, 502 F.2d 621, 624 (5th Cir. 1974). We echoed that more recently, stating that “a protected class that is also a registered voter majority is not foreclosed, as a matter of law, from raising a vote dilution claim.” Salas v. Sw. Tex. Jr. Coll. Dist., 964 F.2d 1542, 1550 (5th Cir. 1992). We have even gone so far so to describe the authority on this point as “unimpeachable.” Monroe v. City of Woodville, 881 F.2d 1327, 1333 (5th

16 Case: 19-60133 Document: 00514887045 Page: 17 Date Filed: 03/25/2019

No. 19-60133 Cir. 1989) (describing the caselaw from our circuit that “has rejected any per se rule that a racial minority that is a majority in a political subdivision cannot experience vote dilution”). We are not alone. Many courts of appeals, including one last year, have rejected the argument that a “racial minority cannot prevail on a Section 2 claim when it constitutes a bare numerical majority within the district.” Mo. St. Conf. of the NAACP v. Ferguson-Florissant Sch. Dist., 894 F.3d 924, 933 (8th Cir. 2018) (collecting cases from the Second, Fifth, Eleventh, and D.C. Circuits). Defendants try to distinguish these cases and others by arguing that they addressed at-large elections rather than single-member districts. But nothing in these cases turned on that feature. And one court of appeals case directly confronting the question rejected a bar on vote dilution claims for groups that make up more than 50% of the challenged single-member district. Kingman Park Civic Ass’n v. Williams, 348 F.3d 1033, 1041 (D.C. Cir. 2003) (agreeing with Salas and Moore in a single-member district vote dilution case). So did a recent three-judge panel in this circuit. See Perez v. Abbott, 253 F. Supp. 3d 864, 879–90 (W.D. Tex. 2017) (rejecting Defendants’ position when finding vote dilution in a congressional district with a Hispanic voting-age population of 58.5%). Against all this authority of both old and recent vintage, all Defendants can point to is Justice Kennedy’s opinion in Bartlett v. Strickland, 556 U.S. 1 (2009) (opinion of Kennedy, J.). 6 Here is the statement they contend supports a categorical rule that there can be no Voting Rights Act violation when a racial group exceeds 50% of the district’s population: “Where an election district could

6The defendants also cite Jeffers v. Beebe, 895 F. Supp. 2d 920, 932 (E.D. Ark. 2012), as an example of a court employing their 50%+ bar on Section 2 claims. Even if they read that case correctly, that holding is no longer good law in its circuit. See Mo. St. Conf. of the NAACP, 894 F.3d at 933. 17 Case: 19-60133 Document: 00514887045 Page: 18 Date Filed: 03/25/2019

No. 19-60133 be drawn in which minority voters form a majority but such a district is not drawn,” a Section 2 claim is possible. This is an “objective, numerical test.” Id. at 18. Assuming this opinion is the controlling one under Marks 7 (Justice Kennedy wrote for only three judges), it is addressing a different question. He was rejecting the concept of crossover districts and instead would require that a vote-dilution plaintiff show that a district could be drawn in which the racial group is a majority. Id. at 15-20. In fact, that is the first showing Gingles requires. LULAC, 548 U.S. at 425 (noting the first “condition[] for establishing a § 2 violation” is that “the racial group is sufficiently large and geographically compact to constitute a majority in a single-member district” (cleaned up)). From that requirement it does not follow that vote dilution cannot occur when the protected group already makes up a majority in the contested district. It is thus not surprising that no court has read Bartlett to override the widely accepted view that the voting strength of a protected group can still be diluted when that group is a majority of a challenged district. Indeed, if any opinion from Justice Kennedy addresses the question before us, it is his recognition in LULAC that “it may be possible for a citizen voting-age majority to lack real electoral opportunity.” Id. at 428. Recognizing this robust caselaw rejecting Defendants’ per se rule (including in single-member cases), the dissent relies on an argument the Defendants have never made: that there needed to be a finding of cracking or packing minority voters. We typically will not consider arguments not advanced by a party. State Indus. Prods. Corp. v. Beta Tech Inc., 575 F.3d 450, 456 (5th Cir. 2009). That principle makes even more sense in the stay context.

7See Marks v. United States, 430 U.S. 188, 193 (1977) (“When a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.” (quotation omitted)). 18 Case: 19-60133 Document: 00514887045 Page: 19 Date Filed: 03/25/2019

No. 19-60133 If a “party is not entitled to a stay as a matter of right” even based on objections it raised before the district court and in requesting a stay, Campaign for Southern Equality v. Bryant, 773 F.3d 55 (5th Cir. 2014), it should not be entitled to that extraordinary remedy based on arguments the district court never had the chance to consider. Even if we could grant a stay based on a challenge to the judgment Defendants do not raise, the dissent’s position misses the marks on the merits. It argues that vote dilution can occur in a 50%-plus district only if the dilution is the result of cracking or packing minority voters. The terms cracking (dividing minority voters among various districts to weaken their strength) and packing (concentrating minority voters in one or a small number of districts where they make up an excessive majority and thus reduce that group’s influence in other districts) describe gerrymandering techniques. Gill v. Whitford, 138 S. Ct. 1916, 1924 (2018) (describing the practice in challenge alleging partisan gerrymander); Shaw v. Reno, 509 U.S. 630, 670 (1993) (White, J., dissenting) (same in challenge alleging racial gerrymander). They explain why vote dilution is occurring. Gill, 138 S. Ct. at 1931. The terms can thus be used to describe types of Section 2 vote dilution claims. See, e.g., Voinovich v. Quilter, 507 U.S. 146, 153 (1993). But the “cracking” or “packing” description does not result in a different Section 2 test or finding. Whether the theory for dilution is cracking or packing or something else, Section 2 claims are evaluated by first looking at the three Gingles factors to determine whether demographics, geography, and voting behavior would allow the creation of a district in which the minority group has electoral power. LULAC, 548 U.S. at 425. If the plaintiffs satisfy that threshold, then a court considers the “totality of the circumstances” to determine if “members of a racial group have less

19 Case: 19-60133 Document: 00514887045 Page: 20 Date Filed: 03/25/2019

No. 19-60133 opportunity than do other members of the electorate.” Id. at 425–26. The district court faithfully followed this path. 8 Although cracking and packing are not official parts of the Section 2 analysis, as a matter of arithmetic vote dilution will often exist because of cracking or packing. Put another way, minority voting strength is often weaker than demographics would allow because those voters have been crammed into one district with a very high minority percentage and thus diluted in other districts (packing) or spread among various districts when they could have formed an electoral majority in at least one (cracking). Indeed, the district court’s findings demonstrate cracking occurred here. 9 District 22 has a BVAP of 50.77% and District 23 has a BVAP of 42.00% when, as the remedy adopted by the district court demonstrates, African-Americans could easily form sufficient electoral strength to elect their representative of choice in at least one of these two districts. The district court correctly focused on the Section 2 factors as prescribed by the Supreme Court and the Voting Rights Act. Part of the reason it found that the <51% majority of African-American voters in the district did not result in an ability to elect a representative of that group’s choosing was because of turnout in off-year elections. So the “50%-plus” issue is largely an attempt to package Defendants’ challenge to the factual finding in a more favorable standard of review. But as with any other factual finding, the clearly

8 It is thus not true that the district court thought that “simply because a district can

be drawn in a way that will guarantee a minority an election win, the Voting Rights Act compels that such a district be drawn.” Dissenting Op. at 7–8. The district court considered all of the statutorily prescribed factors in finding vote dilution. The dissent is also wrong in saying that the district court thought a 62% African-American district was required. See id. That is just the demographics of the remedy it adopted that changed only 28 precincts. To say that the court thought that percentage was a requirement is to say that any time a district court remedies a Section 2 violation it is creating a new minimum. 9 There was no need to use the cracking label because Defendants never made the argument

the dissent advances. 20 Case: 19-60133 Document: 00514887045 Page: 21 Date Filed: 03/25/2019

No. 19-60133 erroneous standard means it is unlikely to be disturbed on appeal. That is particularly true when the factfinder gives a sound reason for crediting one side’s evidence over the other’s. The district court did that in explaining why it relied on Plaintiffs’ expert for turnout data rather than Census Bureau figures. The Census Bureau’s election data is for even-year elections, while Mississippi conducts odd-year state senate elections. Defendants have not shown a high likelihood of overturning the finding of vote dilution because their legal argument is at odds with “unimpeachable authority” from our court and their factual challenges must overcome deferential standards of review. B.

Defendants also argue that the district court erred by not dismissing the lawsuit on laches grounds, focusing entirely on Thomas’s delay in bringing this lawsuit. Even if he were the only plaintiff, the standard of review would again pose a substantial obstacle for Defendants to overcome on appeal. A district court’s equitable ruling on laches is reviewed for abuse of discretion, and the factual findings of delay, inexcusability, and prejudice that underlie it are reviewed for clear error. See Retractable Tech., Inc. v. Becton Dickinson & Co., 842 F.3d 883, 898 (5th Cir. 2016). More problematic than that for Defendants, Thomas is not the only person who brought this suit. The stay motion makes no argument for why laches should bar the lawsuit as to Plaintiffs Ayers and Lawson. As with all affirmative defenses, the party asserting a laches defense bears the burden of proving it. See City of El Paso, Tex. v. El Paso Entm’t, Inc., 382 F. App’x 361, 366 (5th Cir. 2010). Defendants did not even attempt to meet this burden for Ayers and Lawson. Nor is there, despite the dissent’s attempt to salvage a laches defense, an exception for the “main plaintiff,” whatever that means. Cf.

21 Case: 19-60133 Document: 00514887045 Page: 22 Date Filed: 03/25/2019

No. 19-60133 Massachusetts v. EPA, 549 U.S. 497, 518 (2007) (observing that only one petitioner needed to have standing to bring suit). The Voting Rights Act protects the rights of voters, not politicians. As to voters Ayers and Lawson, Defendants point to nothing showing that a laches defense is likely to succeed on appeal. V. A. The final ground on which Defendants seek a stay is the remedy the district court ordered. The standard of review again poses a challenge: the district court’s remedy is reviewed for abuse of discretion. Rodriguez v. Bexar Cnty., Tex., 385 F.3d 853, 870 (5th Cir. 2004). But “whenever practicable . . . a reasonable opportunity [should be afforded] the legislature to meet constitutional requirements by adopting a substituted measure rather than for the federal court to devise and order into effect its own plan.” Veasey, 830 F.3d at 270 (quoting Wise v. Lipscomb, 437 U.S. 535, 540 (1978) (opinion of White, J.)) (emphasis original). As recounted above, the district court first honored this principle in noting it would allow time for a legislative fix before it would redraw the district. When the court was informed that the legislature would not act until it saw the outcome of the stay requests, the district court went ahead and redrew the lines given concerns about the candidate qualifying deadline. It noted, however, that the legislature always has the ability to redraw districts. Plaintiffs also point this out in their response to the stay motion. The legislature’s desire to wait to see the outcome of the stay ruling is understandable. It would have been inefficient for it to remedy a violation only to have this court grant a stay of the district court’s entire judgment on the ground that Defendants were likely to overturn the finding of vote dilution.

22 Case: 19-60133 Document: 00514887045 Page: 23 Date Filed: 03/25/2019

No. 19-60133 But now that we have concluded to the contrary, the legislature should have the initial opportunity to draw new lines for District 22 that comply with the Voting Rights Act. The only reason not to give it that first shot at crafting a remedy is the candidate filing deadline, but the Secretary of State is requesting the legislative fix and has not indicated that a further extension of that deadline for the affected districts is problematic. And while the legislature is always able to redraw lines, the remedial preference is for it to have the first chance to remedy a Section 2 violation, not for it to act only after the court draws a new district, which would result in three different districts in a short time period. United States v. Brown, 561 F.3d 420, 435 (5th Cir. 2009). We thus issued an order on March 15 granting a temporary stay to allow the legislature until April 3 to remedy the Section 2 violation. The order also extended the candidate filing deadline for any affected districts until April 12. B. Defendants otherwise are not likely to surmount the abuse of discretion standard in challenging the court’s remedy. If the legislature does not act and the judicial remedy goes into effect, it must still respect the legislature’s policy objectives as much as possible. See Perry v. Perez, 565 U.S. 38, 393 (2012). The remedy is a narrowly-tailored one that does that. While redistricting cases often result in redrawing the entire state or at least several districts, this remedy moves just 28 precincts, less than 2% of the precincts in the state. The remedy affects only 2 of Mississippi’s 52 Senate districts. The court found that the new lines met traditional redistricting criteria: compactness, contiguity, preservation of political subdivisions like counties and cities, and avoiding pairing incumbents. It adopted the proposed map of an expert it found credible who also determined that the new map honored traditional criteria of compactness and contiguity. In fact, the new district is more consistent with

23 Case: 19-60133 Document: 00514887045 Page: 24 Date Filed: 03/25/2019

No. 19-60133 traditional geographic divisions in one important respect: it removes the suburban Madison county precincts and thus limits the district to counties in the rural and agricultural Delta. The slightly altered districts are, if anything, also more compact than they were before. The following map of the new and old districts show the limited impact of the court’s remedy. The orange and pink highlighting shows the new districts; the blue lines show the old ones.

24 Case: 19-60133 Document: 00514887045 Page: 25 Date Filed: 03/25/2019

No. 19-60133 Defendants’ main critique is that the new lines split Vicksburg between the districts. But some splitting of counties or cities is inevitable given equal population requirements. A look at the statewide senate districts reveals this. The lines the legislature drew split at least eight cities with populations between 23,000 and 27,000 residents (Vicksburg had a population of 26,407 when the 2012 maps were drawn) into different senate districts, including Clinton, whose 25,216 citizens are split among 5 districts. Having to split Vicksburg thus does not lack fidelity to the legislature’s redistricting policy which often results in the same situation. Defendants have not demonstrated a high likelihood of showing that that district court’s narrow redraw was an abuse of discretion. C. The dissent also contends that the district court’s remedy interferes with ongoing election preparation, though the Defendants—including the Secretary of State—do not cite this as a reason for a stay. In a recent case denying a stay in a voting rights case, we did not consider this issue when government officials did not argue that the district court’s plan would substantially disturb the election process. Patino v. City of Pasadena, 677 F. App’x 950, 951–52 (5th Cir. 2017). The election is even further away in this case than it was in Patino. Id. (denying stay sought during the candidate registration period with election three months away); see also Patino v. City of Pasadena, 229 F. Supp. 3d 582, 585 (S.D. Tex. 2017) (providing the relevant dates). Primary voting is more than four months away; the general election is more than seven months away. That distance to the election means the potential for voter confusion that

25 Case: 19-60133 Document: 00514887045 Page: 26 Date Filed: 03/25/2019

No. 19-60133 animates this concern, Purcell v. Gonzalez, 549 U.S. 1, 4 (2006), is absent. 10 Mississippi voters in the 28 affected precincts will go to the same voting location they always have, face the same voting requirements, and have months to research the candidates in advance of the election. Indeed, the effect of redrawing some of the lines for just two state senate districts with elections months away pales in comparison to the impact of statewide redistricting that has been implemented in a similar timeframe. See, e.g., Covington v. North Carolina, 2018 WL 604732 (M.D. N.C. 2018); Common Cause v. Rucho, 284 F. Supp. 3d 780 (M.D. N.C. 2018); Larios v. Cox, 305 F. Supp. 2d 1335 (N.D. Ga. 2004). With no risk of voter confusion, and no outcry from state officials that implementing the district court’s remedy substantially disturbs its election process, this is not a basis for a stay. D. One final comment is necessary to set the record straight in response to the dissent’s attack on the remedy. It irresponsibly insinuates that the district judge—whom it takes the unusual step of repeatedly naming—drew new lines with the aim of eliminating partisan opposition to Plaintiff Thomas. That is not true. Importantly, the incumbent state senator in District 22 had decided not to run for reelection before the district court ruled in this case. The dissent’s conjecture also ignores that primary candidates file with the political party (not the state), with the parties notifying state officials of the candidates

10 As most of the cases cited by the dissent show, the voter confusion concern is typically greatest when a court alters election procedures (as opposed to districts) with an election quickly approaching. See Veasey v. Perry, 769 F.3d 890, 892 (5th Cir. 2014) (staying district court injunction striking down voter ID laws); Purcell, 549 U.S. at 1 (same); North Carolina v. League of Women Voters of N.C., 135 S. Ct. 6 (2014) (staying district court order that changed voter ID requirements, voting hours, voting locations, and voter registration); Husted v. Ohio State Conference of NAACP, 135 S. Ct. 42 (2014) (staying district court order that changed voting hours). 26 Case: 19-60133 Document: 00514887045 Page: 27 Date Filed: 03/25/2019

No. 19-60133 only after the filing deadline. That had not yet occurred when the court issued final judgment. In light of these circumstances, there is no basis for believing that the Plaintiffs’ expert who drew the proposed remedy back in mid-January, or the district court that later adopted it, knew who the new candidates would be. No one brought the names or residences of the candidates to the court’s attention prior to its remedy ruling. And the district court issued its remedy before the initial candidate qualifying deadline, so there may have been additional candidates. Not until after the court issued its remedy, in a motion for stay, did Defendants bring to the court’s attention the fact that two candidates would no longer reside in District 22. On learning this, the district court lamented that it was not told prior to its ruling. As for the competition in the new district, we do not know who has filed nor who will file prior to the new qualifying deadline. *** Per the order issued last week, the motion for stay is GRANTED IN PART and DENIED IN PART. A majority of the panel concludes Defendants have satisfied the stay factors only based on the need to give the legislature and Governor an opportunity to remedy the Section 2 violation. A stay is entered for that purpose until April 3, 2019. The candidate filing deadline for any districts whose lines are redrawn (whether by the legislature, or if it does not act, per the district court’s remedy) is extended to April 12, 2019. The district court has jurisdiction to consider any challenges to the adequacy of a legislative remedy.

27 Case: 19-60133 Document: 00514887045 Page: 28 Date Filed: 03/25/2019

No. 19-60133 EDITH BROWN CLEMENT, Circuit Judge, dissenting in part.

Joseph Thomas is a black Democrat and resident of Yazoo County, Mississippi. A one-time State Senator who lost in 2007 and sat out the 2011 election cycle, he is trying hard to get his seat back. Thomas ran in the 2015 State Senate election in District 22 and lost to the Republican incumbent by a thin margin. Three years later, Thomas and two voters sued the Mississippi Governor, Secretary of State, and Attorney General, accusing them of gerrymandering District 22 in violation of Section 2 of the Voting Rights Act. The plaintiffs argued that black voters in District 22 are politically cohesive, but that white bloc voting prevents them from electing their candidate of choice. Although black voters already comprised a majority of the voting age population in District 22 at 51%, the plaintiffs argued that this majority was too narrow for them to elect a candidate. They say they need at least a 62% majority to do that. The district court agreed that 62% was necessary. And because achieving that percentage was possible by shifting black voters from a neighboring district into District 22, Judge Reeves ordered it done. But he also managed to move something important out of District 22 in the process. Namely, the two Republican challengers to Thomas for the 2019 election.

I feel confident that the district court’s decision will ultimately be reversed. At least, it should be. Laches may well bar the challenge. There is a jurisdictional issue. The theory of liability is unheard-of. The evidence is suspect. And the remedy (as even the majority acknowledges) is faulty. A likelihood of success on any one of these cascading objections is enough to entitle the defendants to a stay of the district court’s redistricting plan. After all, as the majority explains, the stakes are high here. When the majority says there is a “meaningful possibility” that a full appeal could come too late, that is quite the understatement. It is like saying there is a “meaningful possibility” 28 Case: 19-60133 Document: 00514887045 Page: 29 Date Filed: 03/25/2019

No. 19-60133 that a Republican could win an election in the new District 22. Possible in theory, perhaps, but contrary to common understanding and experience.

I dissent from the panel’s refusal to grant the stay in its entirety. Unless we act now, the November election in Senate District 22 is all but decided. True, predictable election results are not uncommon. What is uncommon, however, is for a federal district judge to be the one to decide them.

I. The Remedy Problem

I begin where the district court ended, because even the majority has trouble defending that. Even if Judge Reeves was permitted to decide this case by himself and even if he was correct that District 22 violates Section 2, it remains obvious that his remedy should be stayed and, ultimately, reversed. It is not even a close call.

“Federal-court review of districting legislation represents a serious intrusion on the most vital of local functions.” Miller v. Johnson, 515 U.S. 900, 915 (1995). It is well settled that “reapportionment is primarily the duty and responsibility of the State.” Id. Appropriately, therefore, our case law is clear that “whenever practicable, . . . a reasonable opportunity [should be afforded] for the legislature to meet constitutional requirements by adopting a substitute measure rather than for the federal court to devise and order into effect its own plan.” Veasey v. Abbott, 830 F.3d 216, 270 (5th Cir. 2016) (en banc) (quotation omitted).

That standard undoubtedly leaves room for a district court to act before the legislature when necessary. Given that the original candidate qualifying deadline has come and gone, that the ballots must be sent to overseas troops by the end of June, and that the primary is scheduled for early August, I can

29 Case: 19-60133 Document: 00514887045 Page: 30 Date Filed: 03/25/2019

No. 19-60133 appreciate that we are operating on an abbreviated timetable. But that is all the more reason to preserve the status quo. Federal courts rarely have the right to issue orders the effect of which is to dramatically interfere with ongoing election preparation. See Veasey v. Perry, 769 F.3d 890, 892 (5th Cir. 2014) (“The Supreme Court has repeatedly instructed courts to carefully consider the importance of preserving the status quo on the eve of an election.”). “[U]nder certain circumstances, such as where an impending election is imminent and a State’s election machinery is already in progress, equitable considerations might justify a court in withholding the granting of immediately effective relief in a legislative apportionment case, even though the existing apportionment scheme was found invalid.” Reynolds v. Sims, 377 U.S. 533, 585 (1964).

When issues are hotly contested or when the district court’s reasoning or remedy are flawed, when time to act is short, and when elections hang in the balance, it is almost always better to wait than to interfere. Our court has stayed district court orders in such situations in the past. See, e.g., Veasey, 769 F.3d 890. And when we or other appellate courts have been unwilling to do so, the Supreme Court has stepped in to preserve the status quo. See, e.g., North Carolina v. League of Women Voters of N.C., 135 S. Ct. 6 (2014) (overturning Fourth Circuit injunction which barred implementation of the state’s voting reform laws); Husted v. Ohio State Conference of NAACP, 135 S. Ct. 42 (2014) (staying district court’s injunction); Perry v. Perez, 565 U.S. 1090 (2011) (same); Purcell v. Gonzalez, 549 U.S. 1, 4–5 (2006) (per curiam) (vacating Ninth Circuit injunction, and observing that “[c]ourt orders affecting elections, especially conflicting orders, can themselves result in voter confusion and consequent incentive to remain away from the polls”).

30 Case: 19-60133 Document: 00514887045 Page: 31 Date Filed: 03/25/2019

No. 19-60133 Curiously, the majority believes that because the Supreme Court has mostly done so in cases involving “election procedures,” not redistricting, the Court’s admonitions carry less weight here. But they have it backwards. Altering which candidates appear on a voter’s ballot is a much greater imposition on him than changing the location of his poll or the form of ID he must have in his pocket. Indeed, I have not found a case in which a court altered district boundaries during or after a candidate qualification period. We should maintain the status quo here, or else expect the country’s highest court to once again do it for us.

But even if Judge Reeves had little choice but to implement a remedy himself, I am deeply concerned with the remedy he chose and the way that he chose it. As the majority explains, Judge Reeves adopted the first of three plans submitted by the plaintiffs. He adopted it without making factual findings and without a hearing. In his single-page order adopting the plan, Judge Reeves did not explain why “Plan 1” was the narrowest remedy or clarify why he had rejected “Plan 2” and “Plan 3.” The only time Judge Reeves discussed the details of the plans was in his liability opinion, where he devoted three pages to describing them.

The three plans are similar. All require moving several thousand individuals (27,000 for Plan 1) either into or out of District 22 from one or two adjacent districts. Plans 1 and 2 only affect only Districts 22 and 23, whereas Plan 3 also changes the boundary of District 13. All three plans split the city of Vicksburg in half, though only Plan 2 and Plan 3 offset that splitting by reuniting other cities that are currently split. Under Plan 1, the black voting age population of District 22 would increase to 62%. Under Plan 2, the population would increase to 61%. And under Plan 3, it would increase to 66%. Apart from stating generically that none of the plans “dilute minority voting 31 Case: 19-60133 Document: 00514887045 Page: 32 Date Filed: 03/25/2019

No. 19-60133 strength,” Judge Reeves did not discuss the effect each of the plans would have on the black voting age population of District 23 or District 13. He simply noted that Plan 1 would decrease the black voting age population in neighboring District 23 from 42% to 31%.

Looking at the record, I have no idea why Judge Reeves preferred one plan over another. I do not understand why he selected the one plan in which the splitting of Vicksburg was not offset in some other way. I do not understand how Judge Reeves could endorse increasing the black voting age population of one district without discussing the effect that the change would have on the black voting age population of adjacent districts. And I do not understand why Judge Reeves did not hold a hearing before making his decision. The fact remains that his order—perhaps, inadvertently—eliminated meaningful competition for Thomas in the upcoming election. The majority says that Judge Reeves’s plan is “narrowly tailored.” Agreed—it is “narrowly tailored” to win Thomas the election. 1 In my view, our circuit’s strong preference for the legislature to have the first say on redistricting, combined with the obvious defects in Judge Reeves’s remedy, make it likely that his redistricting plan will ultimately be reversed.

II. The Merits Problem

Another problem with the remedy is that there shouldn’t be one. Indeed, I also think the district court erred in finding a Voting Rights Act violation. I agree with the majority that, at least as a matter of theory, it may be possible

1 The majority thinks this observation “irresponsibl[e].” But even though the slate of candidates had not and could not have been finalized at the time of the district court’s order, it does not require a crystal ball to foresee the result of Judge Reeves’s remedy. He eliminated the Republican base in District 22, essentially making the primary election dispositive. As far as we know, Thomas is the only previously-elected State Senator running in that primary and as far as we know, his only two Republican challengers are now in another district. 32 Case: 19-60133 Document: 00514887045 Page: 33 Date Filed: 03/25/2019

No. 19-60133 for a state to violate Section 2 of the Act even when a protected group forms a majority of the voting population. That option is left open by our precedent. See Salas v. Southwest Texas Jr. College Dist., 964 F.2d 1542, 1550 (5th Cir. 1992); Moore v. Leflore County Bd. of Election Com’rs, 502 F.2d 621, 624 (5th Cir. 1974). But even accepting that possibility, there is a strong likelihood that Judge Reeves committed clear error by finding a Section 2 violation here. Where “the ultimate finding of dilution” is based on “a misreading of the governing law,” there is reversible error. Johnson v. De Grandy, 512 U.S. 997, 1022 (1994).

As the Supreme Court has explained, “a plaintiff may allege a § 2 violation in a single-member district if the manipulation of districting lines fragments politically cohesive minority voters among several districts or packs them into one district or a small number of districts, and thereby dilutes the voting strength of members of the minority population.” Shaw v. Hunt, 517 U.S. 899, 914 (1996). In other words, a state can violate Section 2 by “cracking” the minority vote into separate districts so that the minority is unable to elect its chosen representatives, or by “packing” the minority vote into supermajority-minority districts—thereby reducing the overall number of representatives that the minority receives. See, e.g., Pope v. Cty. of Albany, 687 F.3d 565, 567 (2d Cir. 2012) (minority alleging that their population size and geographical distribution warranted five majority-minority districts, not four); Kingman Park Civic Ass’n v. Williams, 348 F.3d 1033, 1038 (D.C. Cir. 2003) (minority alleging that redistricting diluted the minority vote overall by packing it into some districts and cracking it in others).

It is crucial to understand this context. No court has ever found that a majority-minority single-member district violates Section 2 by itself. The idea that a majority-minority district violates Section 2 only makes sense if the 33 Case: 19-60133 Document: 00514887045 Page: 34 Date Filed: 03/25/2019

No. 19-60133 allegation is that packing or cracking has occurred. See Thornburg v. Gingles, 478 U.S. 30, 46 n.11 (1986) (“Dilution of racial minority group voting strength may be caused 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 they constitute an excessive majority.”). Both kinds of violations span district lines. A plaintiff cannot prove packing or cracking without analyzing minority population and voter statistics for multiple districts. See, e.g., Perez v. Abbott, 253 F. Supp. 3d 864, 879 (W.D. Tex. 2017) (analyzing statistics for numerous districts and concluding that “while Plaintiffs may not have submitted sufficient proof that they are entitled to eight [Hispanic citizen voting age population]-majority districts in South/West Texas, they have shown that they are entitled to seven such districts, and they may assert claims under . . . § 2 against the districts in [the current plan]”).

There was no such evidence in this case. Indeed, the complaint nowhere mentions “packing” or “cracking,” and the evidence focused on District 22 alone. I do not understand how a majority-minority district could violate Section 2 unless the district is the result of packing or cracking the minority vote among a geographical area that is larger than the district itself. The allegation that the minority’s majority is not large enough to overcome white bloc voting would only become actionable if the minority’s majority is being artificially lowered to dilute the minority vote.

The majority opinion suggests that Section 2 vote dilution requires that a minority group be deprived of a district in which it has electoral strength to elect a representative of its choice when such a district is possible. But nothing in Section 2 “establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.” 52 U.S.C. § 10301(b). “[T]he ultimate right of § 2 is equality of opportunity, not a guarantee of 34 Case: 19-60133 Document: 00514887045 Page: 35 Date Filed: 03/25/2019

No. 19-60133 electoral success for minority-preferred candidates of whatever race.” De Grandy, 512 U.S. at 1014 n.11; see also Gingles, 478 U.S. at96 (O’Connor, J., concurring in the judgment).

The district court and the majority seem to think that simply because a district can be drawn in a way that will guarantee a minority an election win, the Voting Rights Act compels that such a district be drawn. Not so. Section 2 does not grant the plaintiffs a right to a 62% majority to overcome low black turnout or white bloc voting in the abstract. It is only when the drawing of district lines along traditional methods should have resulted in a different electoral scheme that Section 2 is violated. There must be evidence that packing or cracking has taken place (whether or not it is expressly referred to as such). 2 That is the difference between guaranteeing equal opportunity and guaranteeing electoral success.

The majority would avoid this reality by saying the defendants do not raise the issue. But challenging the viability of Voting Rights Act claims in majority-minority single-member districts and also challenging the plaintiffs’ evidence in this case—both of which defendants do—places the district court’s merits ruling in contention. It is telling that the majority, apparently recognizing the conceptual difficulty of finding vote dilution in the absence of packing or cracking, ultimately attempts to recast the district court’s finding in that vein. But because there was no such evidence—indeed, neither packing

2 Putting these fundamental concerns to the side, I also believe the plaintiffs did not offer sufficient evidence to support their claim. I am particularly troubled by Judge Reeves’s decision to disregard Census Bureau statistics concluding that black voter turnout generally exceeds white voter turnout. Even the cases relied on by the plaintiffs acknowledge that Census Bureau statistics are usually dispositive. See Mo. St. Conf. of the NAACP v. Ferguson- Florissant Sch. Dist., 894 F.3d 924, 932–33 (8th Cir. 2018). 35 Case: 19-60133 Document: 00514887045 Page: 36 Date Filed: 03/25/2019

No. 19-60133 nor cracking was even alleged—the plaintiffs have not shown a Section 2 violation.

III. The Laches Problem

The plaintiffs are entirely to blame for the haste with which we must resolve this case. The challenged redistricting occurred seven years ago. Disliking what he saw from the start, Thomas contacted the Department of Justice in 2012 and asked them to “look hard” at newly drawn Senate Districts 21, 22, and 34. He questioned whether the new redistricting plan as a whole reduced black voting strength in Mississippi. Rejecting Thomas’s concerns, the DOJ precleared the plan in September 2012.

Thomas waited. He did not file a Voting Rights Act challenge in 2012, 2013, or 2014. He chose instead to run in the 2015 election, losing 54% to 46% to the incumbent chairman of the State Senate Appropriations Committee. Thomas was “real disappointed” that despite his efforts to appeal to white voters in the 2015 election, he garnered little of the white vote. See Thomas v. Bryant, 2019 WL 654314, at *1 (S.D. Miss. Feb. 16, 2019).

Still, Thomas did nothing. He did not file a Voting Rights Act challenge in 2015, 2016, or 2017. Thomas has explained that he was unaware that an individual could file a Voting Rights Act challenge until his lawyer told him in 2018. But “laches does not depend on subjective awareness of the legal basis on which a claim can be made.” Envtl. Def. Fund, Inc. v. Alexander, 614 F.2d 474, 479 (5th Cir. 1980). The letter to the DOJ shows that Thomas had concerns about District 22 as early as 2012, yet he waited until halfway through 2018 to act.

36 Case: 19-60133 Document: 00514887045 Page: 37 Date Filed: 03/25/2019

No. 19-60133 That delay prejudiced the defendants and the public. Laches is an inexcusable delay on the part of the plaintiff that results in prejudice to the defendant. Conan Props., Inc. v. Conans Pizza, Inc., 752 F.2d 145, 153 (5th Cir. 1985). It has been applied by the Supreme Court to bar untimely Voting Rights Act challenges. Lopez v. Hale Cty., Tex., 506 U.S. 1042 (1993) (affirming Lopez v. Hale Cty., Tex., 797 F. Supp. 547, 550 (N.D. Tex. 1992)); Cf. White v. Daniel, 909 F.2d 99, 102 (4th Cir. 1990). By waiting until now to challenge the district, Thomas has injected needless uncertainty into the November 2019 election. Mere months before Election Day, Mississippi voters went to bed in one district and woke up in another. Candidates suddenly find themselves running for office in a district they do not know, appealing to a public that does not know them. And the Republican Party finds itself without a horse in the race, moments before the starting gun is fired.

The delay has inured to the benefit only of Thomas, whose prospects have brightened noticeably since the district court redrew the political landscape. His postponement of legal action has—unless the stay is granted—all but ensured that there will be no time for reconsideration of the district court’s opinion before the election. Judge Reeves’s decision, made on an accelerated timeline following expedited discovery, is likely to be the end of the matter. The balance of the equities, in my view, is not on Thomas’s side. 3

3 The other plaintiffs, Vernon Ayers and Melvin Lawson, do not undermine this view. Lawson is a District 22 voter who worked on Thomas’s Senate campaign and then joined Thomas’s lawsuit. No one seems to know anything about Ayers besides the fact that he is a registered voter in District 22. Thomas is the main plaintiff (even the district court’s opinion focused almost exclusively on Thomas). In any event, all the plaintiffs are challenging district lines which were implemented seven years ago. It is that unnecessary delay, common to all the plaintiffs, which should bar the suit. 37 Case: 19-60133 Document: 00514887045 Page: 38 Date Filed: 03/25/2019

No. 19-60133 IV. The Jurisdiction Problem

Assuming Thomas did not unduly delay, however, was it permissible for the district court to decide this Voting Rights Act case without convening a three-judge panel? As the majority notes, there is no need to get into this unless 28 U.S.C. § 2284(a) is jurisdictional. That statute 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.”

The majority is right that the statute certainly “sounds jurisdictional.” Maybe that is why every court to consider the question has said that it is. See Kalson v. Patterson, 542 F.3d 281, 287 (2nd Cir. 2008); Armour v. State of Ohio, 925 F.2d 987, 988–89 (6th Cir. 1991) (en banc); LULAC of Texas v. Texas, 318 F. App’x 261, 264 (5th Cir. 2009) (per curiam). I think so too. And since the majority ultimately concedes that forfeiture is an obstacle the defendants “will likely be able to overcome on appeal,” I will get straight to the heart of the question.

This lawsuit is indisputably an action challenging the apportionment of a statewide legislative body. The only question is whether the “constitutional” modifier in § 2284(a) applies to the second phrase in the sentence. The district court ruled that it does, giving three reasons. Thomas v. Bryant, No. 0190219C, 2019 WL 454598, at *2–3 (S.D. Miss. Feb. 5, 2019). First, that there was no reason to treat Voting Rights Act claims and constitutional claims as identical in the statewide context. Second, that the plain language of the text and statutory canons of construction compelled the court’s interpretation. Third,

38 Case: 19-60133 Document: 00514887045 Page: 39 Date Filed: 03/25/2019

No. 19-60133 that the court’s decision was buttressed by other persuasive decisions. The first two reasons are totally incorrect, and the third is weak.

“The task of statutory interpretation begins and, if possible, ends with the language of the statute.” United States v. Lauderdale Cty., Mississippi, 914 F.3d 960, 964 (5th Cir. 2019). The district court said that the “series-qualifier” canon compelled its conclusion that the constitutionality modifier applied to both phrases. Thomas, 2019 WL 454598 at *2–3. The book “Reading Law” explains the canon as follows: “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.” ANTONIN SCALIA & BRYAN A. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 147 (2012). A typical example is the phrase “Forcibly assaults, resists, opposes, impedes, intimidates, or interferes with,” in which the modifier “forcibly modifies each verb in the list.” Id. at 148. Crucially for our purposes, however, the authors also explain when the modifier does not apply:

The typical way in which syntax would suggest no carryover modification is that a determiner (a, the, some, etc.) will be repeated before the second element: [for example,] The charitable institutions or the societies (the presence of the second the suggests that the societies need not be charitable).

Id.

That is exactly the situation we have here. The statute provides for a three-judge court “when an action is filed challenging the constitutionality of the apportionment of congressional districts or the apportionment of any statewide legislative body.” 28 U.S.C. § 2284(a) (emphasis added). The determiner “the” (or the determining phrase “the apportionment”) cuts off the

39 Case: 19-60133 Document: 00514887045 Page: 40 Date Filed: 03/25/2019

No. 19-60133 continued application of the word “constitutionality” to the second phrase. That the sentence has a parallel structure, as the majority harps on, is not dispositive. After all, if the statute did not have a parallel structure, we would not be discussing the series-qualifier canon in the first place. The whole point of the canon is to describe when a modifier applies to a parallel series, and when it does not. The canon and the context both answer that question in the negative.

If Congress intended the majority’s interpretation, there would have been much clearer ways to say so. Here are just three examples:

• 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.

• A district court of three judges shall be convened . . . when an action is filed challenging the constitutionality of [either] the apportionment of congressional districts or the apportionment of any statewide legislative body.

• A district court of three judges shall be convened . . . when an action is filed challenging the constitutionality of the apportionment of congressional districts or [of] the apportionment of any statewide legislative body.

Notably, Congress said none of those things. At the risk of quoting someone other than the late Justice Scalia, the Supreme Court has observed time and again that we “must presume that a legislature says in a statute what it means and means in a statute what it says.” Connecticut Nat. Bank v. Germain, 503 U.S. 249, 253–54 (1992); see also, e.g., Arlington Cent. Sch. Dist. Bd. of Educ.

40 Case: 19-60133 Document: 00514887045 Page: 41 Date Filed: 03/25/2019

No. 19-60133 v. Murphy, 548 U.S. 291, 296 (2006); In re Camp, 631 F.3d 757, 759 (5th Cir. 2011).

This may seem overly technical. But the district court explicitly noted that “[d]espite its ‘fancy name,’ the series-qualifier canon ‘reflects the completely ordinary way that people speak and listen, write and read.’” Thomas, 2019 WL 454598 at *2 n.5 (quoting Lockhart v. United States, 136 S. Ct. 958, 970 (2016) (Kagan, J., dissenting)). And the canon supports the opposite interpretation from the one adopted by the district court. It is this reading, and not the majority’s reading, which reflects the plain meaning of the statute. Further, the district court admitted the defendants had a “fair point” that the court’s interpretation resulted in the statute containing needless words. Id. at *2. “If ‘the constitutionality of’ is indeed carried over to all following phrases, the second use of ‘the apportionment of’ is rendered unnecessary.” Id. The correct application of the series-qualifier canon has the added benefit of avoiding that contradiction and giving meaning to every word in the statute. This is consistent with yet another canon: the surplusage canon. See SCALIA & GARNER, at 174–79.

The district court could not understand why Congress would treat Voting Rights Act and constitutional claims as identical, focusing on the fact that they have different evidentiary requirements and that a plaintiff is master of his complaint. Thomas, 2019 WL 454598 at *2. But the differing burdens of proof and the plaintiff’s undenied ability to select the claims he wishes to bring do not undermine the rationale for treating all challenges to statewide legislative apportionment the same. Both constitutional and statutory attacks on legislative redistricting implicate the same fundamental concern: balancing judicial respect for a state’s decision regarding the best manner of selecting its representatives with the need to ensure equal opportunity within the 41 Case: 19-60133 Document: 00514887045 Page: 42 Date Filed: 03/25/2019

No. 19-60133 electorate. Whether court-ordered redistricting is based on the Constitution or the Voting Rights Act makes little difference to the state whose actions have been curtailed.

Indeed, a more relevant question than the one asked by the district court is why the statute would treat challenges to congressional redistricting differently from challenges to statewide legislative redistricting. The answer to that challenge may well lie in federalism. It is entirely plausible that Congress wanted federal courts to show more deference to state reapportionment plans that only affect state interests than to state reapportionment plans which affect national interests. The majority suggests that federalism would actually support the opposite inference, considering that a state’s power to select its congressional representatives stems directly from Article I, Section II of the Constitution. But that line of reasoning is undermined by Supreme Court precedent, which has allowed the states greater latitude in creating state legislative districts than in creating congressional districts. See Gaffney v. Cummings, 412 U.S. 735, 742–44 (1973).

In any event, I am not arguing that federalism compels the statutory interpretation for which I advocate. I simply offer a plausible explanation for why the plain text says what it does. The truth is that most of the time, the opposing interpretations of § 2284(a) will not matter, as Voting Rights Act claims and constitutional claims are usually asserted together. That is why— far from hiding an elephant in a mousehole—the correct interpretation merely ensures equal treatment for stand-alone Voting Rights Act claims whose implications can be quite as severe as constitutional ones. My view has the additional benefit of avoiding the forum-shopping concerns implicit in the majority’s reading.

42 Case: 19-60133 Document: 00514887045 Page: 43 Date Filed: 03/25/2019

No. 19-60133 Finally, the district court’s conclusion that non-binding cases “buttress[]” its decision is weak. See Thomas, 2019 WL 454598 at *3. The district court cited a Sixth Circuit decision, a Third Circuit decision, and a decision from the Northern District of Georgia. Id. The majority adds a citation from the Northern District of Alabama, the District of South Dakota, and the District of Montana. The sum of persuasive authority on this point is two circuit court opinions and four district court opinions. And as the majority admits, these cases really aren’t on point to begin with. The Sixth Circuit opinion does not mention § 2284(a)—it simply affirmed a single-judge decision. Rural West Tennessee African-American Affairs Council v. Sundquist, 209 F.3d 835 (6th Cir. 2000). Of the district court opinions, only the Northern District of Alabama addresses the question head-on, and it fails to mention the series-qualifier canon or the superfluous language that all parties admit results from the district court’s reading. See Chestnut v. Merrill, 2019 WL 338909 (N.D. Ala. 2019).

Most significantly, the Third Circuit decision supports the opposite conclusion from that reached by the district court. In Page, the court was considering whether a single judge may decide a Voting Rights Act claim when statutory and constitutional claims are brought together. See Page v. Bartels, 248 F.3d 175, 186–94 (3d Cir. 2001). While the opinion says that § 2284(a) “only requires a three-judge district court for certain constitutionally-based apportionment challenges,” the court further stated that Congress likely did not make a “deliberate choice to distinguish between constitutional apportionment challenges and apportionment challenges brought under § 2 of the Voting Rights Act.” Id. at 189.

Indeed, most redistricting challenges at the time of the 1976 amendment were constitutional, meaning the distinction carried less importance in 1976 43 Case: 19-60133 Document: 00514887045 Page: 44 Date Filed: 03/25/2019

No. 19-60133 than it does today. And “the legislative history of the 1976 revisions to 28 U.S.C. § 2284 clearly demonstrates that Congress was concerned less with the source of the law on which an apportionment challenge was based than on the unique importance of apportionment cases generally.” Id. at 190. Because all cases attacking “the legitimacy of the state legislative apportionment” are “highly sensitive matters,” the Third Circuit concluded that practical and policy considerations weighed in favor of three-judge review irrespective of the nature of the claims asserted. Id. So in its survey of the caselaw, the majority seems to have missed that the only circuit court opinion to address the issue in any depth goes the other way. 4

Of course, none of these opinions (and there are not very many of them) are binding on this court. As noted earlier, the issue does not come up often since plaintiffs normally bring suit using all the potential weapons in their arsenal. To the extent that the opinions that have considered this question cut in any direction, however, I’d submit they cut in favor of the three-judge panel for Voting Rights Act claims. The district court was wrong to read § 2284(a) in the manner it did. A three-judge panel should have decided this case.

* * *

This case presents several extraordinary issues. Unfortunately, this court’s usual procedures do not appear to permit en banc review of this denial of a stay even if a majority of the active judges would otherwise grant it. I am afraid defendants have simply had the poor luck of drawing a majority- minority panel. I trust that in light of this, the State will pursue a stay in the

4 The majority acknowledges Page in a footnote, noting that it “does not adopt the textual argument” I am advocating here. That is unsurprising, considering that no one advocated for such an interpretation in Page. It is the reasoning from the opinion which supports my view—not the opinion’s different answer to a different question. 44 Case: 19-60133 Document: 00514887045 Page: 45 Date Filed: 03/25/2019

No. 19-60133 Supreme Court because of the injustice that results from the joint efforts of the district judge and the motions panel majority. I also encourage the State to move for an expedited appellate process before this court, preferably seeking an April or May sitting—it might yet be possible for this court to undo its own mistake.