Tennessee State Conference of the NAACP v. Lee

 No. 3:23-cv-00832 IN THE UNITED STATES DISTRICT COURT FOR THE MIDDLE DISTRICT OF TENNESSEE NASHVILLE DIVISION TENNESSEE STATE CONFERENCE OF ) THE NAACP; et al., ) ) Plaintiffs, ) ) v. ) ORDER ) WILLIAM B. LEE, in his official capacity as ) Governor of the State of Tennessee, et al., ) ) Defendants. ) BEFORE: RICHARDSON, District Judge, MURPHY, Circuit Judge, and PEARSON, District Judge. PER CURIAM. This redistricting case requires us to consider what a complaint must allege to plausibly suggest that a legislature relied on unlawful racial considerations—rather than lawful political ones—to draw a legislative map. Our response must incorporate a mix of constitutional and pleading rules. As for the constitutional rules, the Supreme Court recently made clear that plaintiffs who claim that a legislature relied on race “must rule out the possibility that politics drove the districting process” whenever race and politics are highly correlated. Alexander v. S.C. State Conf. of the NAACP, 144 S. Ct. 1221, 1243 (2024). And the presumption of legislative good faith requires courts to uphold a map if the record could support either a finding that the legislature relied on race or a finding that it relied on politics. Id. at 1235–36. Yet Alexander arose after a trial. This case, by contrast, remains at the pleading stage. That fact takes us to the pleading rules. We must dismiss a complaint if its “well-pleaded facts do not permit the court to infer more than the mere possibility of misconduct[.]” Ashcroft v. Iqbal, 556 U.S. 662, 679 (2009). Such a “complaint has alleged—but it has not ‘show[n]’—‘that the pleader is entitled to relief.’” Id. (quoting Fed. R. Civ. P. Rule 8(a)(2)). And a complaint will sometimes fail this test if its well-pleaded facts “have an obvious alternative explanation” that would establish no wrongdoing. Bell Atl. Corp. v. Twombly, 550 U.S. 544, 567 (2007). This combination of constitutional and pleading rules requires us to dismiss the Complaint in this case—at least as currently pleaded. In 2022, Tennessee’s legislature redrew the lines that divide the State’s congressional and state senate districts. The Plaintiffs—a group of civil-rights organizations and voters whom we will collectively call “the Challengers”—assert that the legislature enacted a racial gerrymander and discriminated against minority voters in its changes to three congressional districts and one state senate district. The Complaint alleges that the changes split these minority voters across several districts and so diluted their power to influence the election in any one district. It adds that the changes divided counties and communities of interest in violation of traditional redistricting criteria. On the Complaint’s own terms, however, a political gerrymander amounts to an “obvious alternative explanation” for these effects. Id. The Complaint alleges that minority voters prefer Democratic candidates. It also alleges that the changes (by a Republican-controlled legislature) flipped a congressional seat long held by a Democratic representative and shored up a state senate seat that a Republican senator barely won in a recent election. In light of this partisanship explanation for the changes, the Complaint fails to allege “more than the mere possibility” of racial discrimination. Iqbal, 556 U.S. at 679. That said, we will give the Challengers the opportunity to replead their claims because the Supreme Court issued Alexander after they brought this lawsuit. We also reject Defendants’ alternative request to dismiss the Complaint on the ground that the Challengers waited too long to sue. But we dismiss Tennessee’s governor from the lawsuit on sovereign-immunity grounds. I The Constitution requires the federal government to conduct an “Enumeration” of the population every ten years. U.S. Const. art. I, § 2, cl. 3. The government must use this census data to determine the number of each State’s representatives in the House of Representatives. See id. And the Supreme Court has interpreted the Constitution to contain a “one person, one vote” requirement for federal representatives and state legislators. See Evenwel v. Abbott, 578 U.S. 54, 59 (2016). In other words, each federal representative or state legislator within a State “must be accountable to (approximately) the same number of constituents” as the other representatives and legislators in that State. Rucho v. Common Cause, 588 U.S. 684, 709 (2019). Each State must create federal congressional districts with populations that are as close to equal “as possible.” Evenwel, 578 U.S. at 59. In comparison, the States presumptively satisfy this “one person, one vote” requirement for state legislative districts if “the maximum population deviation between the largest and smallest district is less than 10%[.]” Id. at 60. Given these requirements, state legislatures traditionally redraw their legislative maps after each census to account for population changes over the last decade. See Alexander, 144 S. Ct. at 1242. In Tennessee, the state constitution instructs its General Assembly to undertake this redistricting for its state senators and representatives. See Tenn. Const. art. 2, § 4. And a state statute instructs the General Assembly to do the same for its federal representatives. See Tenn. Code Ann. § 2-16-102. This case concerns the Tennessee General Assembly’s efforts to redraw the maps for its congressional districts and its state senate districts after the 2020 census. Compl., R.1, PageID 17. That census showed that the State’s population had increased by about 564,735—a brisk 8.9% growth rate. Id., PageID 15. But this growth had occurred unevenly. Id., PageID 16. A larger amount had taken place in the City of Nashville and its surrounding county, Davidson County. Id. The General Assembly thus needed to equalize the populations in the outdated 2010 maps. The General Assembly began its efforts to redraw these maps after receiving the 2020 census data in August 2021. Id., PageID 17–18. According to the Challengers, various legislative committees created the maps in a way that “lacked transparency,” and the General Assembly “rapidly” voted them into law. Id., PageID 17, 24. On January 12, 2022, the House Select Committee approved a new congressional map without previously “disclos[ing it] to the public.” Id., PageID 21–23. The next day, the Senate Ad Hoc Committee approved a congressional map and a state senate map. Id., PageID 20–21, 23. On January 18, the Senate Judiciary Committee discussed and approved these maps. Id., PageID 23–25. Two days later, the Senate approved the maps “along party lines” with all Republicans in favor and all Democrats opposed. Id., PageID 24–25. The House followed suit a short time later. Id. And Governor Lee signed them into law on February 6. Id., PageID 24, 26. The Challengers criticize specific aspects of both maps. Congressional Map. For years, Tennessee has had nine congressional districts. Id., PageID 29. Since 1940, District 5 has included all of Nashville. Id. It thus came to be known as “the Nashville/Davidson County” district. Id. Under the old map for the 2012–2020 elections, District 5 combined all of Davidson County with most of two neighboring counties: No. 3:23-cv-00832, Tenn. State Conf. of the NAACP, et. al. v. William B. Lee, et al. PSEA ee Montgomery ‘Sumner mo | PR ee ry... pe Se) Humphreys _ 05 | 06 | ~" a 08) Dekalb | white cumberiand aa Crockett’ 07 ) Hickman Butnerrord [Cannon} rat [Roan Heywood | wadison yt eae Bedford § Coffee n “oa 3 Marshall Jay ead oe % EPA? LincoIn Po Id. The district’s voters had elected Jim Cooper, a Democratic candidate, since 2002. Id. After the 2020 census, the “ideal” size of each of Tennessee’s nine congressional districts became 735,463 under the “one person, one vote” rule. /d., PageID 31. As of that time, Davidson County contained about 715,884 people. /d. To create a new District 5 of ideal size with Davidson County kept intact, then, the General Assembly needed to add only 19,579 people from a neighboring county in the old District 5. But the General Assembly took a more complex approach. It split Davidson County into three different districts: Districts 5, 6, and 7. /d. In the following map of these three new districts, Davidson County comprises some of the yellow (District 5), some of the dark blue (District 6), and some of the orange (District 7): ye Id. The next map shows a closer view of the new borders of these three districts in that county: No. 3:23-cv-00832, Tenn. State Conf. of the NAACP, et. al. v. William B. Lee, et al. sy oo th, ee GoefrO'™ = fe A car } > J Den % i : - 2 Oy Ba - ¥ a “x8le i py Lay 8 eet Ne Haier Id. This map left 346,457 residents of Davidson County in new District 5. /d. It moved 188,668 residents of the county to new District 6 and 180,759 residents of the county to new District 7. □□□ The Complaint alleges that these changes had racial and political impacts. As for race, the Complaint asserts that the General Assembly split up old District 5’s population of racial minorities (in particular, African Americans and Hispanic Americans) by moving some of these minority voters into other districts. /d., PageID 32-34. In 2020, old District 5 had a “Black and Hispanic Voting Age Population” or “BHVAP” of “around 30%” and a “White Voting Age Population” or “WVAP” of around 61%. /d., PageID 32-33. Old Districts 6 and 7 had a much smaller BHVAP (about 9% and 14%) and a much larger WVAP (about 86% and 79%) because they included “white populations” in rural counties. /d. Under the new map, however, new District 5 has a much smaller BHVAP (about 20%) and a much larger WVAP (about 72%). /d., PageID 33-34. And new Districts 6 and 7 have a larger BHVAP (about 15% and 22%) and a smaller WVAP (about 80% and 72%). Id. As for politics, the Complaint alleges some factual claims that the changes harmed Democrats and helped Republicans. It alleges that “voters of color” had “coalesced behind the same candidate of choice” in old District 5: the Democratic candidate, Jim Cooper. /d., PageID No. 3:23-cv-00832, Tenn. State Conf. of the NAACP, et. al. v. William B. Lee, et al. 29, 32, 34. Under the old map, by comparison, the “rural, white populations” in old Districts 6 and 7 had elected “Republican candidates” by almost 40% margins. /d., PageID 33. The new map decreased these margins in new Districts 6 and 7 while still allowing Republican candidates to win. /d., PageID 34. But the new map now allowed a Republican candidate to win new District 5. Id. In fact, Representative Cooper, District 5’s longstanding representative, chose not to run in 2022 because the “gerrymandering” had left “no way” for him to win. /d., PageID 35. The eventual Democratic nominee lost by 13.7%. /d., PageID 36. State Senate Map. The Challengers raise a similar claim against the revised state senate districts in Shelby County, which encompasses the City of Memphis. /d., PageID 41. Under the old map that covered the 2012-2020 elections, the State Senate contained four seats completely within Shelby County (old Senate Districts 29, 30, 31, and 33) and one senate seat partially within the county (old Senate District 32): Tipton 32, 29) Id., PageID 37-38. This map anchored old Senate District 31 in the “Black and Hispanic neighborhood of Cordova in Memphis”: No. 3:23-cv-00832, Tenn. State Conf. of the NAACP, et. al. v. William B. Lee, et al. ton TA d > @ phis. Memphis ae a | Southaven” a @ Id., PageID 39, 41. Old Senate District 31 had a majority “White Voting Age Population” (again, “WVAP”) of 57.51%, and had traditionally elected a Republican candidate. /d., PageID 39. By the decade’s end, however, its “Black and Hispanic Voting Age Population” (again, “BHVAP”) had grown to about 35%. /d. And the District’s racial minorities (centered in Cordova) had “come extremely close” (within 1.8%) to electing their preferred candidate—a Democrat—in 2018. □□□ The 2020 census showed that the State Senate’s various districts had an “ideal population” of 209,419 with a “10% overall deviation range” between 198,948 and 219,890 under the Supreme Court’s “one person, one vote” precedent. /d., PageID 38. So the population of Senate District 31 (209,168) remained within this acceptable 10% range and close to the ideal number. /d. But the General Assembly chose to redraw the five Shelby County districts: No. 3:23-cv-00832, Tenn. State Conf. of the NAACP, et. al. v. William B. Lee, et al. Id., PageID 38. The new Senate District 31 “no longer centers around Cordova, but instead around Germantown—a much whiter neighborhood”: | + "Or vena District 31 Ce fd | 4 2 LoL : p= |p ) tS Aah a = oe {SAY Bie + 14 Pins eS TS | / Eye cbimantown mM.) Of aS A lh iA mo 4 LE ‘| ~ southaven ~ aT □ er | a NN ah am BP ED Ty Id., PageID 40-41. This change also allegedly had racial and political impacts. New District 31’s BHVAP decreased from about 35% to about 20% and its WVAP increased from 57.51% to 70.29%. Id., PageID 39, 41. The change also led the Republican candidate to win easily in 2022. /d., PageID 40. Minority voters thus could not “elect candidates of choice” in the new district. Jd. * Ok OK The Challengers originally included civil-rights organizations (the Tennessee State Conference of the NAACP, the League of Women Voters of Tennessee, the Equity Alliance, the African American Clergy Collective of Tennessee, and Memphis A. Philip Randolph Institute) and individuals who live in the affected districts under the new map (Judy Cummings and Brenda Gilmore of Congressional District 7, Ophelia Doe of Congressional District 5, Freda Player of Congressional District 6, and Ruby Powell-Dennis of Senate District 31). Id., PageID 5–14. They filed this suit under 42 U.S.C. § 1983 against several Tennessee officials in their official capacities: Governor William Lee, Secretary of State Tre Hargett, Coordinator of Elections Mark Goins, and the State Election Commission and its members (collectively, “Tennessee”). Id., PageID 14–15. We have since granted the Challengers’ unopposed motion to voluntarily dismiss the claims of two plaintiffs (Memphis A. Philip Randolph Institute and the Equity Alliance) without prejudice. The other claims remain pending. The Challengers assert a total of four claims—two types of claims against the new Congressional Districts 5, 6, and 7, and the same two types of claims against the new Senate District 31. Id., PageID 43–48. In the first set of claims (against the new Congressional Districts 5, 6, and 7, and the new Senate District 31), the Challengers assert that the General Assembly engaged in racial gerrymandering when drawing the districts in violation of the Fourteenth and Fifteenth Amendments. Id., PageID 43–44. In the second set of claims (again against the new Congressional Districts 5, 6, and 7, and the new Senate District 31), the Challengers assert that the General Assembly intentionally drew the districts in a way that diluted the power of minority voters in violation of the Fourteenth Amendment. Id., PageID 45–46. Tennessee moved to dismiss all four claims, but it asked us to delay our decision until after the Supreme Court decided Alexander. We consented to that requested delay. The motion is now ripe for our review. II. Racial-Gerrymandering Claims Tennessee argues that the Complaint fails to plausibly plead that the General Assembly used race to draw the districts because the Complaint does not rebut the possible alternative explanation of politics. While acknowledging that they must plausibly allege that race drove the legislature’s decisionmaking, the Challengers respond both that their Complaint need not rebut Tennessee’s “politics” defense at the pleading stage and that the Complaint did so anyway. A This case requires us to decide how two lines of Supreme Court decisions coalesce. The first concerns the specific substantive requirements for proving a racial-gerrymandering claim. The second concerns the general procedural requirements for pleading any type of claim. 1 The Equal Protection Clause bars state legislatures from relying on race to draw the lines that will divide legislative districts. See Alexander, 144 S. Ct. at 1234; Cooper v. Harris, 581 U.S. 285, 291 (2017); Shaw v. Reno, 509 U.S. 630, 642–49 (1993) (Shaw I). This prohibition follows from the bedrock principle that courts must view all race-based classifications with suspicion. See Miller v. Johnson, 515 U.S. 900, 904 (1995). Just as a State should not use race to identify the schools that children may attend, so too it should not use race to determine the districts in which citizens should vote. See id. (citing Brown v. Bd. of Educ., 347 U.S. 483 (1954)). Yet our country has traditionally reserved redistricting for state legislatures. Alexander, 144 S. Ct. at 1233. And legislatures need “discretion” to balance the many considerations that might point in different directions over how best to redraw maps. Miller, 515 U.S. at 915. Consider a few questions that a legislature must confront when drawing districts: Which county lines should it “respect” and which counties should it split apart to satisfy the Supreme Court’s one-person, one-vote principle? Shaw I, 509 U.S. at 647. Should it strive to keep communities of interest together even if they live across a broader geographic range or should it strive to create compact districts? Cf. Cooper, 581 U.S. at 308. Which districts should it keep largely intact and which should it completely redraw to account for population changes? Cf. Allen v. Milligan, 599 U.S. 1, 21–22 (2023). Which existing legislators should get “safer” seats and which should have competitive ones? Cf. Bush v. Vera, 517 U.S. 952, 964–65 (1996) (plurality opinion). Given the many delicate issues and the “presumption of good faith” that the legislature receives, courts must “exercise extraordinary caution” when considering a racial-gerrymandering claim. Miller, 515 U.S. at 916; see Hunt v. Cromartie, 526 U.S. 541, 553 (1999) (Cromartie I). As all Justices agree, therefore, the governing test for evaluating racial-gerrymandering claims should protect against unlawful racial discrimination while giving legislatures enough room to perform their work. See Alexander, 144 S. Ct. at 1233–34; see also id. at 1271–72 (Kagan, J., dissenting). To satisfy these competing concerns, the Court has adopted a test that requires plaintiffs who assert racial-gerrymandering claims to prove that racial considerations played the “predominant” role in the redistricting. Miller, 515 U.S. at 916. This test sets a “high bar.” Alexander, 144 S. Ct. at 1235; see Easley v. Cromartie, 532 U.S. 234, 241 (2001) (Cromartie II). Plaintiffs must prove that a legislature intentionally subordinated traditional (and race-neutral) redistricting criteria (“such as compactness, contiguity, and core preservation”) to divide citizens by race. Alexander, 144 S. Ct. at 1235. Stated the other way, the legislature must have refused to “compromise[]” its use of race in order to advance these other race-neutral redistricting criteria. Id. at 1234 (quoting Shaw v. Hunt, 517 U.S. 899, 907 (1996) (Shaw II)). Only if plaintiffs meet this “predominance” test does the burden of persuasion shift to the State to prove that the legislature’s use of race satisfies strict scrutiny. See Miller, 515 U.S. at 920. Plaintiffs can use direct evidence or circumstantial evidence (or a combination of both) to prove that a legislature relied primarily on race to draw a legislative district. See Cooper, 581 U.S. at 291; Bethune-Hill v. Va. State Bd. of Elections, 580 U.S. 178, 190–91 (2017). They will meet this test more easily if they present direct evidence of racial gerrymandering. This type of evidence is not out of the question and sometimes can “be smoked out over the course of litigation.” Alexander, 144 S. Ct. at 1234. Most commonly, legislators will admit that they relied on race to comply with the Voting Rights Act—say, because they believed that they needed to create a “majority-minority” district. See Cooper, 581 U.S. at 299–301; Vera, 517 U.S. at 969–70, 975 (plurality opinion); Miller, 515 U.S. at 917–19; see also Alexander, 144 S. Ct. at 1286 n.1 (Thomas, J., concurring). Other times, the legislature’s documentary and electronic records and similar evidence may reveal that legislators focused on racial considerations during the redistricting. Cooper, 581 U.S. at 318; see Alexander, 144 S. Ct. at 1234. If direct evidence does not exist, however, plaintiffs will have to travel a “much more difficult” path by proving a circumstantial case. Alexander, 144 S. Ct. at 1234. “[I]n theory,” a new map could look so strange and its borders could have such a strong correlation with race that its shape alone might establish that race predominated in the redistricting. Id.; see also Shaw I, 509 U.S. at 643−47. But if this type of circumstantial evidence could “plausibly support” the conclusion that a race-neutral factor drove the legislature’s decisionmaking, the presumption of good faith still looms, requiring federal courts to “draw” that competing “inference” in favor of the map’s legality. Alexander, 144 S. Ct. at 1236. So the Court in Alexander could identify just a single case that would meet its demanding circumstantial-evidence requirements—one that addressed “the most grotesque racial gerrymander in the U.S. Reports[.]” Id. at 1273 (Kagan, J., dissenting); see id. at 1250 (discussing Gomillion v. Lightfoot, 364 U.S. 339, 364 (1960)). This rule dooms many racial-gerrymandering claims. Why? As the Supreme Court has stated, a strong correlation has long existed between race and politics. See id. at 1235; Cromartie II, 532 U.S. at 242; Vera, 517 U.S. at 968 (plurality opinion). In recent decades, for example, a much larger percentage of minority voters have voted for Democratic candidates than for Republican candidates. See Alexander, 144 S. Ct. at 1236; Cromartie II, 532 U.S. at 239, 245, 252. So a district that appeared to be drawn on racial lines instead might have been drawn to further a political objective—whether to create a “safe” Democratic seat, see Cromartie II, 532 U.S. at 239, or a “safe” Republican one, see Alexander, 144 S. Ct. at 1238, 1241–42. And “[i]f either politics or race could explain a district’s contours,” the claim fails as a matter of law. Id. at 1235; Cromartie II, 532 U.S. at 243, 257. When race and politics are highly correlated, then, plaintiffs must “disentangle” the two by showing that race (not politics) “drove a district’s lines.” Cooper, 581 U.S. at 308. How might plaintiffs do so? They might establish that those who drew the maps lacked good data on a region’s political makeup but had excellent data on its racial makeup. See id. at 318. If the map drawers used race data “as a proxy” for political data, race would predominate in an illegal way even if they sought to achieve a race-neutral goal (partisan gain) in drawing the lines. Miller, 515 U.S. at 914; Cooper, 581 U.S. at 291 n.1, 318. Or plaintiffs might establish that the legislature “could have achieved its legitimate political objectives” (the creation of a safe Democratic seat or a safe Republican seat) through an “alternative” map that had less stark racial impacts. Cromartie II, 532 U.S. at 258; see Alexander, 144 S. Ct. at 1235–36, 1249–50. Because alternative maps “can be designed with ease” using modern technology, the Court has taken the plaintiffs’ failure to produce one “as an implicit concession” that no such map exists. Alexander, 144 S. Ct. at 1250. And this failure triggers an “adverse inference against” any claim that race rather than politics predominated in the mapmaking. Id. at 1249. Alexander shows the “demanding” nature of this test. Id. at 1240 (citation omitted). That case concerned South Carolina’s congressional map. South Carolina historically had a “6-to-1 Republican-Democratic delegation,” but its District 1 had become competitive. Id. at 1236–37. When redrawing the map, the Republican-leaning legislature sought to make District 1 a safer Republican seat. Id. at 1238. The plaintiffs countered that the legislature had wrongly used race to achieve this objective. After a bench trial, a district court agreed. Id. But the Supreme Court held that the court committed clear error. Id. at 1240. The challengers lacked direct evidence that the legislature used race. Id. at 1241. And their circumstantial case failed to rebut the “possibility” that the legislature drew the map based on political considerations alone. Id. at 1241–43. For example, the desire to create a safer Republican seat could have plausibly explained the decision to violate traditional redistricting criteria such as “the avoidance of county splits.” Id. at 1242. And the plaintiffs’ failure to produce an alternative map that would have created this safe seat with less of a racially disparate impact undercut the challengers’ claim even more. Id. at 1249–50. 2 In most of the Supreme Court’s cases addressing racial gerrymandering, the evidence at a trial confirmed that a correlation existed between race and politics in the relevant geographic area. See Alexander, 144 S. Ct. at 1235, 1238; id. at 1268 (Kagan, J., dissenting); Cromartie II, 532 U.S. at 239, 242. This case arises in a far different posture: Tennessee’s motion to dismiss reaches us at the pleading stage. That difference raises a key question: What allegations must a plaintiff plead at this initial stage of the case to state a racial-gerrymandering claim? Our answer begins with basic pleading rules. Courts must evaluate a motion to dismiss in two steps. See Iqbal, 556 U.S. at 678–79. A court should first identify the allegations that it must accept as true. Id. This presumption of truth extends only to well-pleaded facts—not to statements of law or conclusory recitations of a claim’s elements. See id. at 678. Once the court identifies the well-pleaded facts, it should ask whether they “plausibly” suggest the existence of all the claim’s elements. Twombly, 550 U.S. at 557. The complaint meets this test if the facts allow the court to reasonably infer that the defendant violated the law. Iqbal, 556 U.S. at 678. But it does not meet the test if the facts are merely “consistent with” the defendant’s liability, Twombly, 550 U.S. at 557, or reveal only a “possibility of misconduct,” Iqbal, 556 U.S. at 679. Critically, a complaint will fall short of this plausibility standard if the well-pleaded facts that purport to establish a defendant’s misconduct are just as consistent “with an obvious alternative explanation” that would preclude a showing of a legal violation. Id. at 682 (quoting Twombly, 550 U.S. at 567). Twombly illustrates this rule. There, the complaint alleged that telecommunications companies had violated the antitrust laws by conspiring not to compete in each other’s old territories after federal deregulation. 550 U.S. at 549–51. As its well-pleaded facts to support the existence of this agreement, the complaint asserted that each company had not expanded outside its territory. Id. at 551, 567. The Court held that this allegation did not plausibly plead an illegal agreement. Id. at 564–70. It noted that the complaint’s express allegations of an illegal agreement qualified as “legal conclusions” that it need not accept. Id. at 564. So the plaintiffs were left with a circumstantial case tied to the mere fact that the companies had not competed with each other by seeking to expand outside their territories. But this lack of competition had a lawful “alternative explanation”: that the companies independently chose to remain in their own territories. Id. at 567–69. And that type of “conscious parallelism” (in which firms in a concentrated market each rationally engage in the same conduct independent of the other competitors) would not violate the antitrust laws. Id. at 553–54, 567–70. Since Twombly, countless courts have dismissed complaints because they failed to rebut an “obvious alternative explanation”—that is, an explanation for complained-of conduct that involved no wrongdoing. Iqbal, 556 U.S. at 682 (citation omitted). Take Iqbal. There, a detainee alleged that high-level executive officials had ordered his harsh detention because of his race and religion (Arab and Muslim, respectively) after the terrorist attacks on September 11, 2001. 556 U.S. at 667–69. The Supreme Court held that his complaint failed to plausibly plead intentional discrimination. Id. at 680–84. It treated the “bare assertions” that the officials had adopted a discriminatory policy as legal conclusions that it should disregard. Id. at 680–81. And the alleged fact that the government had primarily detained Arab Muslim men after the attacks could be explained by an “obvious alternative explanation”: the government was pursuing those with a “suspected link to the attacks,” which had been “perpetrated” by a group largely composed of the “Arab Muslim disciples” of Osama bin Laden. Id. at 682 (citation omitted). Similar examples are not hard to find. As the Sixth Circuit has said, when an obvious (and lawful) explanation exists for factual allegations of claimed illegality, the allegations have not moved the needle from “possible and conceivable” to “plausible and cognizable.” Smith v. CommonSpirit Health, 37 F.4th 1160, 1167 (6th Cir. 2022). So that court has held that a complaint failed to plausibly plead that a bank racially discriminated against a borrower struggling to pay off a loan, given the bank’s “obvious alternative explanation”: a “concern about repayment.” 16630 Southfield Ltd. P’ship v. Flagstar Bank, F.S.B., 727 F.3d 502, 505–06 (6th Cir. 2013). Similarly, the Third Circuit has held that a complaint failed to plausibly plead that airport security officials detained a traveler for possessing “Arabic-English flashcards and a book critical of American interventionism,” given the officials’ “obvious alternative explanation”: a concern that he “might pose a threat to airline security.” George v. Rehiel, 738 F.3d 562, 567, 586 (3d Cir. 2013). And the Seventh Circuit has held that a complaint failed to plausibly plead that a city discriminated against female domestic-violence victims, given the city’s “obvious alternative explanation” for the failure to protect them: “limited police resources.” McCauley v. City of Chicago, 671 F.3d 611, 616, 619 (7th Cir. 2011) (citation omitted); see Doe v. Samford Univ., 29 F.4th 675, 689 (11th Cir. 2022). Conversely, the Supreme Court has disavowed any notion that Twombly’s plausibility test requires plaintiffs to satisfy evidentiary requirements they must meet later in the case. Twombly, 550 U.S. at 570 (distinguishing Swierkiewicz v. Sorema N. A., 534 U.S. 506 (2002)). Before Twombly, for example, the Court rejected the view that Title VII plaintiffs must plead factual allegations that show a prima facie case of employment discrimination under the burden-shifting framework from McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973). See Swierkiewicz, 534 U.S. at 510–15. Among other reasons, the Court noted that plaintiffs need to show a prima facie case under McDonnell Douglas only at the summary-judgment stage (as opposed to the pleading stage) and only if they rely on circumstantial evidence of racial discrimination. Id. at 511–12. Yet some plaintiffs instead attempt to prove their claims only with direct evidence. Id. 3 Should Twombly’s pleading rules apply to racial-gerrymandering claims? We think so. Generally, the Supreme Court has left no doubt that Twombly’s plausibility test applies across the board. See Iqbal, 556 U.S. at 684. Specifically in this redistricting context, the Court has stated that its constitutional “principles”—including its sensitivity to “the intrusive potential of judicial intervention into the legislative realm”—do not just “inform the plaintiff’s burden of proof at trial.” Miller, 515 U.S. at 916. They also go into a judicial evaluation of the “adequacy of a plaintiff’s showing at the various stages of litigation[.]” Id. at 916–17. Here, then, courts must recognize this potential for intrusion at the pleading stage when deciding whether plaintiffs should get to take “discovery” on their racial-gerrymandering claim. Id. at 917 (citing Fed. R. Civ. P. 12(b)). Alexander’s requirements for racial-gerrymandering claims also fit within the Court’s pleading framework. We start with Twombly’s standard. It clarifies that a complaint must allege well-pleaded facts that plausibly rebut an “obvious alternative explanation” to the claim that the legislature engaged in racial gerrymandering. 550 U.S. at 567, 569–70. So if a complaint suggests that race and politics are “highly correlated” in a region, it must plead facts that plausibly “rule out the possibility that politics [as opposed to race] drove the districting process.” Alexander, 144 S. Ct. at 1233, 1243. And a complaint will not meet this test if it alleges only that a redrawn map had a “disparate, incidental impact” along racial lines or that the map violated traditional redistricting criteria. Iqbal, 556 U.S. at 682. Both alleged facts would represent the “side effect” of a partisan gerrymander (as opposed to the intentional effect of a racial gerrymander) when race and politics are highly correlated. Alexander, 144 S. Ct. at 1241–42; see Cromartie II, 532 U.S. at 243. These facts alone thus “shed[] no light” on whether politics was the cause or race was the cause. Cromartie I, 526 U.S. at 555 (Stevens, J., concurring in the judgment). An analogy to Twombly proves this point. Allegations of a racially disparate impact and of the disregard of traditional redistricting criteria resemble the allegations that fell short in that case: that the companies refrained from competing in each other’s territories. See Twombly, 550 U.S. at 567. Both sets of allegations (that companies were not competing or that a legislative map violated traditional redistricting criteria and had a racially disparate impact) may be “consistent with” illegality: an agreement in violation of the antitrust laws or a racial gerrymander. Iqbal, 556 U.S. at 681. But they do not “plausibly establish” this illegality, given the existence of “more likely explanations”: conscious parallelism or a partisan gerrymander, respectively. Id. Alexander’s clarification of the “presumption of legislative good faith” also supports this view that plaintiffs must allege facts beyond the existence of a racially disparate impact or the disregard of traditional redistricting criteria. 144 S. Ct. at 1235. The presumption compels courts to “draw the inference” that favors upholding a map if the evidence would permit a court to reach competing “conclusions” about the map’s lawfulness. Id. at 1235–36. In this respect, the presumption does not resemble the McDonnell Douglas burden-shifting approach to make out a prima facie case of discrimination—an approach that applies only if and when the plaintiff chooses to rely on circumstantial evidence at the summary-judgment stage. Rather, the presumption of legislative good faith comprises part of the constitutional test that invariably applies at the “various stages of litigation” in redistricting cases, including the pleading stage. Miller, 515 U.S. at 916– 17. This presumption thus directs district courts to “draw the inference that cuts in the legislature’s favor when confronted with [a complaint’s allegations] that could plausibly support multiple conclusions.” Alexander, 144 S. Ct. at 1235–36. At the same time, Swierkiewicz shows that complaints need not take any specific route to plead the required facts. 534 U.S. at 510–15. Alexander’s suggestion that plaintiffs produce an “alternative map,” for example, does appear to be an evidentiary tool that applies at the summary- judgment stage if the plaintiff tries to make out a circumstantial case of racial gerrymandering. 144 S. Ct. at 1249–50; see Bethune-Hill, 580 U.S. at 188–91. So this alternative-map option does resemble the McDonnell Douglas burden-shifting test that plaintiffs need not satisfy at the pleading stage. See Swierkiewicz, 534 U.S. at 510–11. The option also does not look like a substantive element of the claim (the satisfaction of which, according to Iqbal, the complaint must plausibly allege). Although alleging such a map might plausibly establish a racial-gerrymandering claim, plaintiffs need not do so. Rather, they can plausibly plead a racial-gerrymandering claim in other ways (so long as they can allege the factual matter in good faith). To give a few examples, a complaint might allege that the legislature’s new map treated minority voters of one party worse than white voters of the same party—something that could “undercut the possibility that partisan politics were to blame for the decision.” Christian Ministerial All. v. Thurston, __ F. Supp. 3d __, 2024 WL 398428, at *3 (E.D. Ark. Feb. 2, 2024). Or it might allege that the mapmakers did not have good political data and so relied on racial data “as a proxy” for partisan preference. Miller, 515 U.S. at 914. Or it might allege that legislators announced that they considered race in an effort to comply with the Voting Rights Act. Cf. Shaw I, 509 U.S. at 634–35. In sum, when a complaint makes clear that race and politics are highly correlated, it must allege facts that plausibly “disentangle race from politics[.]” Cooper, 581 U.S. at 308; see Twombly, 550 U.S. at 570. But challengers have flexibility in how they plausibly allege these facts, including, for example, by alleging facts that suggest racial discrimination among voters of the same political party. See Christian Ministerial All., 2024 WL 398428, at *3. B The Challengers’ two racial-gerrymandering claims suffer from the same problem under these pleading rules. To show that race predominated in the redistricting plan, the Complaint relies on allegations that the changes had a racially disparate impact and violated traditional redistricting criteria. But the Complaint also pleads allegations suggesting that race and politics are highly correlated. And it does not allege any additional facts that would plausibly rule out the possibility that politics (rather than race) drove the redistricting. We will discuss each claim in turn. 1 Congressional Districts 5, 6, and 7. The Complaint alleges that “[r]ace predominated” in the legislature’s decision to split up Nashville and Davidson County into three congressional districts. Compl., R.1, PageID 43. We start by identifying the well-pleaded facts that the Complaint alleges to support this general conclusion. See Iqbal, 556 U.S. at 678–79, 680–81. As its main support, the Complaint asserts two sets of facts. For one thing, it pleads statistics alleging that the new map’s changes had disparate “racial effects” on minority voters in Davidson County. Alexander, 144 S. Ct. at 1241. The new map moved to new Districts 6 and 7 many minority voters who had been in old District 5 and so decreased the Black and Hispanic Voting Age Population (BHVAP) in new District 5 by about 11%. Compl., R.1, PageID 32–33. Conversely, it shifted many white voters to new District 5 from old Districts 6 and 7 and so increased the White Voting Age Population (WVAP) in new District 5 by about 10%. Id., PageID 32–34. Minority voters in the old District 5 thus lost their ability to elect their preferred candidate by combining with white voters who held similar political views. Id., PageID 29–30, 36; cf. Bartlett v. Strickland, 556 U.S. 1, 13 (2009) (plurality opinion). For another thing, the Complaint pleads facts suggesting that the new map’s changes disregarded “traditional redistricting objective[s].” Alexander, 144 S. Ct. at 1242. Among other things, the new map engaged in a “drastic split of Nashville and Davidson County[.]” Compl., R.1, PageID 36. It separated many “communities of interest” in Davidson County. Id., PageID 35. And it redrew the “core” of the old District 5, which had centered in Nashville for decades. Id., PageID 30. Lastly, new District 5 is “far less compact” than the old one. Id., PageID 35. These two sets of well-pleaded allegations may be “consistent with” the claim that race predominated in the redistricting. Twombly, 550 U.S. at 557. But they fail to “plausibly suggest[]” such illegal use of race because the Complaint also alleges facts suggesting that race and party preference are correlated in the Nashville area. Id.; see Cromartie II, 532 U.S. at 243. According to the Complaint, Jim Cooper—a member of the Democratic Party—long represented the old District 5 that contained all of Davidson County. Compl., R.1, PageID 29. And the Complaint adds that this congressman was the “candidate of choice for most voters of color.” Id. It then suggests that the “candidate of choice of voters of color” (another Democrat candidate) lost the 2022 election in the new District 5. Id., PageID 36. It similarly alleges that the “candidate[s] of choice of voters of color” in Districts 6 and 7 (the Democratic candidates) also lost by wide margins during the 2022 election (but by much less than “40 percentage points” that they had lost to Republicans under prior maps). Id., PageID 33–34, 36. In short, the Complaint universally treats Democratic candidates as the “candidates of choice” for minority voters. Given these allegations, the racially disparate impact and departure from traditional redistricting criteria had an “obvious alternative explanation”: a partisan gerrymander to flip District 5. The Complaint, for example, alleges that the House Select Committee that created the initial version of the map was “dominated and chaired by Republicans” and that these “Republican members” approved the map. Id., PageID 18, 22. It adds that the State Senate voted on the map along “party lines” by a 26-5 vote. Id., PageID 24–25. And Jim Cooper, District 5’s congressman, described the new map as the product of “gerrymandering” that made it impossible for him to win. Id., PageID 35. Lastly, the Republican candidate won in 2022. Id., PageID 36. In sum, the Complaint alleges facts that are consistent with a racial gerrymander. But the facts are also consistent with a political gerrymander. And the Complaint alleges nothing else that might plausibly “rule out the possibility that politics drove the districting process.” Alexander, 144 S. Ct. at 1243. Senate District 31. The same problem exists for the Complaint’s allegations that “[r]ace predominated” in the drawing of new Senate District 31. Compl., R.1, PageID 44. We start a second time by identifying the well-pleaded facts that the Complaint alleges to support this general conclusion. See Iqbal, 556 U.S. at 678–79, 680–81. The Complaint again alleges that the changes disparately affected minority voters. The new district’s BHVAP decreased by about 15% and its WVAP increased by about 13%. Id., PageID 39, 41. And the Complaint again alleges that the changes disregarded traditional redistricting criteria. It suggests that the new district is “more sprawling” than before and divided communities of interest in Memphis. Id., PageID 41. Yet the Complaint again alleges facts suggesting “that racial identification is highly correlated with political affiliation” in Memphis. Cromartie II, 532 U.S. at 243. It states that minority voters in the old Senate District 31 sought to elect a Democratic candidate. Compl., R.1, PageID 39. And it suggests that “voters of color” “supported” “a Black Democrat,” not a “white Republican,” under the new map. Id., PageID 40. So while the allegations may be “consistent with” a claim that race predominated, they fail to “plausibly suggest[]” this conclusion because they do not account for the possibility that the legislature adopted a partisan gerrymander to shore up a teetering Republican district. Twombly, 550 U.S. at 557. Indeed, the Complaint alleges that a Democrat came within 1.8% of winning old Senate District 31 but that the Republican won easily after the redistricting. Id., PageID 39–40. Because the Complaint includes no facts to “rule out the possibility that politics drove the districting process,” it does not suffice. Alexander, 144 S. Ct. at 1243. 2 The Challengers’ arguments do not convince us otherwise. First, they criticize Tennessee for relying on outside-the-complaint sources in its motion to dismiss. Resp., R.46, PageID 413– 14. They have a point. The motion cites everything from recorded legislative hearings, to news articles, to a census official’s memo. See Mot., R.43, PageID 285 n.1, 287 n.10. Tennessee responds that we can take judicial notice of these materials. We are not so sure. But rather than enter that debate, we will simply disregard the materials. See Bates v. Green Farms Condo. Ass’n, 958 F.3d 470, 484 (6th Cir. 2020); Max Arnold & Sons, LLC v. W.L. Hailey & Co., 452 F.3d 494, 502–04 (6th Cir. 2006). Evaluated alone, the Complaint fails to plausibly plead that race predominated in the redistricting. Second, the Challengers criticize Tennessee for arguing that they must allege that “an alternative map” existed in which the legislature could have achieved its political goals (to turn Congressional District 5 into a Republican seat and to make Senate District 31 a safer Republican seat) with less of a racially disparate impact. Resp., R.46, PageID 424. We agree that the Challengers do not have to satisfy any alternative-map obligation at this stage. See Swierkiewicz, 534 U.S. at 510–15. Still, they did have to plead some factual allegations “plausibly suggesting” that the General Assembly redrew the districts based on racial classifications rather than political ones. Twombly, 550 U.S. at 557. They did not meet this pleading standard. Third, the Challengers suggest that they rebutted the argument that politics drove the redistricting. They first contend that they merely needed to allege that “race—not any other factor—predominated in the drawing of the maps” to satisfy this burden. Resp., R.46, PageID 423–24. But we need not assume the truth of this allegation. In fact, the Iqbal complaint likewise alleged that the executive officers had devised a policy that discriminated based on race and religion. 556 U.S. at 680–81. And the Twombly complaint alleged that the companies had entered a conspiracy that violated the antitrust laws. 550 U.S. at 564 & n.9. Yet the Court in both cases refused to assume the truth of “these bald allegations” because of their “conclusory nature[.]” Iqbal, 550 U.S. at 681; see Twombly, 550 U.S. at 564–65. The same logic covers the conclusory allegations here that “[r]ace predominated” in the redistricting. Compl., R.1, PageID 43–44. In support of this third argument, then, the Challengers alternatively argue that the Complaint included enough factual matter because it alleged that the redistricting had a racially disparate impact and disregarded traditional redistricting criteria. Resp., R.20, PageID 421–22, 424–27. But these allegations do not suffice for the reasons we have already identified. Indeed, the Complaint fails to address partisanship at all as a possible alternative reason for the racially disparate impact and for the disregard of traditional redistricting criteria even though those two allegations themselves suggest that possibility. As their contrary support, the Challengers cite one Supreme Court case that arose at the pleading stage: Shaw I. That decision dates to 1993—a time when the Supreme Court would not dismiss a complaint unless “it appear[ed] beyond doubt that the plaintiff [could] prove no set of facts in support of his claim which would entitle him to relief.” Conley v. Gibson, 355 U.S. 41, 45–46 (1957). Shaw I thus says little about this issue because the Court later overruled Conley in favor of its “plausibility” test. Twombly, 550 U.S. at 562–64. Besides, the defendants in Shaw I nowhere asserted that politics predominated. See 509 U.S. at 634–58. Rather, the state legislature used race to create a second majority-minority district under the Voting Rights Act. See id. at 634–37; see also Alexander, 144 S. Ct. at 1234. The Challengers allege nothing similar, instead merely citing other cases in which district courts have denied motions to dismiss. See Resp., R.46, PageID 423. But they fail to explain why this case is like those ones—especially after Alexander clarified the governing racial-gerrymandering test. Fourth, the Challengers assert that we must draw “all reasonable inferences” in their favor at this stage. Resp., R.46, PageID 414. They mistake the substantive constitutional test for a procedural pleading rule. True, the pleading rules require us to draw every “reasonable inference” in a plaintiff’s favor. Iqbal, 556 U.S. at 678. But Alexander’s “presumption of legislative good faith” means that courts must uphold a map unless challengers “rule out the possibility that politics drove the districting process.” 144 S. Ct. at 1243. At the trial stage, challengers have not met this burden if the evidence could point either way. Id. at 1235–36. That is why Alexander could hold that a district court committed clear error by finding that race predominated even though the record contained conflicting evidence on this race-versus-politics question. While the clear-error standard generally requires an appellate court to draw an inference in favor of a district court’s findings in that situation, see Cooper, 581 U.S. at 293, the presumption of legislative good faith required the district court to rule for the State, see Alexander, 144 S. Ct. at 1241–42 & 1249 n.11. Applying this presumption at the pleading stage, a complaint does not plausibly plead a racial- gerrymandering claim if its allegations leave open the additional “possibility” that politics played the key role. Id. at 1241. The Challengers’ Complaint leaves open this possibility. Fifth, the Challengers attempt to meet this test with a few other allegations apart from their claims of a racially disparate impact and the disregard of traditional redistricting criteria. For starters, they point to the Complaint’s allegation that no legislator offered a “justification” for the maps’ changes other than the need to equalize populations across districts under the “one person, one vote” requirement. Compl., R.1, PageID 36, 43; Resp., R.46, PageID 423. Yet we find this equal-population requirement neutral in the debate between whether political considerations or racial considerations predominated. As the Supreme Court has explained, this requirement represents a “background rule against which redistricting takes place” because the legislature must always attempt to achieve population equality between districts. Ala. Legislative Black Caucus v. Alabama, 575 U.S. 254, 273 (2015). But this command says little about the specific reasons why the legislature chose to place certain voters in one (equalized) district and other voters in another (equalized) district. See id. So we must “put” this background principle “to the side” and ask whether race predominated on that subsidiary choice in how to draw those equalized districts. Id.; see Christian Ministerial All., 2024 WL 39842, at *4. Yet the Complaint does not plausibly “rule out the possibility” that politics drove this choice. Alexander, 144 S. Ct. at 1243 Nor can the Challengers plausibly “rule out [this] possibility” simply because legislators did not publicly announce a partisan gerrymander during the redistricting. Id. The challengers in Alexander similarly alleged that many legislators had disavowed a partisan motive, and the state defendants had not even raised this theory until trial. Id. at 1275 (Kagan, J., dissenting). But the Court found it “implausible” that the plaintiffs had not known of any “partisan concerns” until the trial given that the legislators had been accused of gerrymandering during the redistricting process itself. Id. at 1244–45, 1251. And here, the Complaint alleges that Cooper accused the legislature of “gerrymandering” on his Twitter page as the reason why he could not win reelection as a Democratic candidate in the new District 5: “I explored every possible way, including lawsuits, to stop the gerrymandering and to win one of the three new Congressional districts that now divide Nashville. There’s no way, at least for me in this election cycle, but there may be a path for other worthy candidates.” Compl., R.1, PageID 35. Next, the Challengers separately invoke their vote-dilution allegations. That the legislature acted with a discriminatory animus, they argue, bolsters their claim that race predominated. Resp., R.46, PageID 427. We agree with their major premise: Discriminatory animus can help plausibly suggest that a legislature used race during a redistricting. But we disagree with the minor premise: As we will next explain, the Complaint does not plausibly allege intentional discrimination. III. Vote-Dilution Claims The Challengers alternatively raise what they call “discriminatory purpose” claims against the districts. Because the Supreme Court has referred to these claims as “vote-dilution” claims, we will use that term. See Alexander, 144 S. Ct. at 1251. The Supreme Court has referred to vote- dilution claims as “analytically distinct” from racial-gerrymandering claims. Miller, 515 U.S. at 911 (quoting Shaw I, 509 U.S. at 652); see also Alexander, 144 S. Ct. at 1252. Apart from barring the predominant use of race in redistricting, the Equal Protection Clause bars legislatures from designing districts with the invidious intent to dilute the voting power of voters of particular races. See Alexander, 144 S. Ct. at 1251–52. This rule follows from the bedrock principle that a race- neutral law can violate equal protection if the legislature passed the law with the purpose to harm a racial group and if the law has a harmful impact on this group. See Hunter v. Underwood, 471 U.S. 222, 227–28 (1985); City of Mobile v. Bolden, 446 U.S. 55, 66–67 (1980) (plurality opinion); Vill. of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 264–66 (1977); see also Pers. Adm’r of Mass. v. Feeney, 442 U.S. 256, 272–74 (1979); Washington v. Davis, 426 U.S. 229, 239– 42 (1976). A vote-dilution claim thus requires plaintiffs to prove two elements: a discriminatory purpose and a discriminatory effect. See Alexander, 144 S. Ct. at 1252. Here, legal uncertainty makes it difficult to say whether the Challengers plausibly pleaded a discriminatory effect. But we need not decide that question because they failed to plausibly plead a discriminatory purpose. A. Discriminatory Effect Outside the redistricting context, plaintiffs typically have little difficulty alleging a disparate effect. They must simply show that a law has harmed members of one race more than members of another one. Consider the Alabama law in Hunter that disenfranchised those who committed certain crimes. That law had a disparate effect because it disenfranchised “ten times as many” African Americans as whites. 471 U.S. at 227 (citation omitted). Yet the test is not so simple for challenges to legislative districts. The original vote-dilution claims attacked the use of “multimember” districts (which combine many voters into one large district that votes on two or more legislators) rather than single-member districts (which separate voters into smaller districts that vote on one legislator each). See Bolden, 446 U.S. at 65–66; White v. Regester, 412 U.S. 755, 765–67 (1973); Whitcomb v. Chavis, 403 U.S. 124, 141–43 (1971). Plaintiffs alleged that a legislature had intentionally adopted this type of multimember-district structure to “cancel out” the voting strength of racial minorities. Fortson v. Dorsey, 379 U.S. 433, 439 (1965). But the Supreme Court held that a racial group seeking to prove a discriminatory effect from this structure had to do more than show that it could not elect the number of its preferred legislators “in proportion” to the group’s percentage of the district’s overall population. White, 412 U.S. at 765–66; Whitcomb, 403 U.S. at 149. Rather, a racial group could prove the required effect only by establishing “that the political processes leading to nomination and election were not equally open to participation by the group,” such “that its members had less opportunity than did other residents in the district to participate in the political processes and to elect legislators of their choice.” White, 412 U.S. at 766; see Rogers v. Lodge, 458 U.S. 613, 624 (1982). Before the Supreme Court decided how White’s constitutional “effect” test should apply to challenges to single-member districts like the districts at issue in this case, Congress changed the law. In 1982, it amended § 2 of the Voting Rights Act to allow statutory challenges to legislative districts based on a discriminatory effect alone without the need to prove a discriminatory purpose. See Allen, 599 U.S. at 11–14. Congress codified—nearly word for word—the language (quoted above) that White used to articulate the “effect” test for establishing constitutional vote-dilution claims. Compare 52 U.S.C. § 10301(b), with White, 412 U.S. at 766. Since then, most vote- dilution claims have arisen under the statute because § 2 (unlike the Equal Protection Clause) does not require proof of discriminatory purpose. See Allen, 599 U.S. at 19 (citing cases). For plaintiffs to show that a legislative district’s structure has a discriminatory “effect” under § 2’s language, the Supreme Court has since held that they must meet three “preconditions.” Thornburg v. Gingles, 478 U.S. 30, 50 (1986). And the Court has extended these so-called “Gingles preconditions” to single-member districts. See Allen, 599 U.S. at 38; Bartlett, 556 U.S. at 12 (plurality opinion); Growe v. Emison, 507 U.S. 25, 40–41 (1993). The first precondition requires plaintiffs to show that a minority group “is sufficiently large and geographically compact to constitute a majority” in a hypothetical district with different borders than the challenged one. Growe, 507 U.S. at 40 (quoting Gingles, 478 U.S. at 50). If a minority group could not form a majority under another district design, the Court has reasoned, the existing structure does not have the effect of depriving the group of “the potential to elect a representative of its own choice in some single-member district.” Id.; see Bartlett, 556 U.S. at 12–20 (plurality opinion). Recall, moreover, that § 2’s “effect” test merely adopted the constitutional “effect” test that the Supreme Court applied in White. See 412 U.S. at 765–66. So some courts have suggested that plaintiffs who assert constitutional vote-dilution claims must both establish intentional discrimination and meet the same Gingles “effect” test that they must satisfy to allege a § 2 claim. See Georgia State Conf. of NAACP v. State, 269 F. Supp. 3d 1266, 1278 (N.D. Ga. 2017); Lopez v. City of Houston, 2009 WL 1456487, at *17–18 (S.D. Tex. May 22, 2009), aff’d on other grounds, 617 F.3d 336 (5th Cir. 2010); see also Johnson v. DeSoto Cnty. Bd. of Comm’rs, 204 F.3d 1335, 1344–46 (11th Cir. 2000); Wesley v. Collins, 791 F.2d 1255, 1263 (6th Cir. 1986). This approach would pose a problem for the Challengers’ vote-dilution claims. They do not assert that any of the challenged districts qualified as a “majority-minority” district before the redistricting. Compl., R.1, PageID 32, 39–41. And they do not assert that minority voters could make up a majority of a district with differently configured borders. So they have not pleaded the first Gingles factor (which perhaps explains why they have not brought Voting Rights Act claims). That said, other courts have rejected the notion that plaintiffs who assert intentional- discrimination claims must establish the same “effect” test that they must meet to satisfy § 2 (which lacks any intent element). One court reasoned, for example, that the Gingles test was not “clearly rooted” in the text of the Voting Rights Act and was “even further removed from the text of the Constitution.” League of United Latin Am. Citizens v. Abbott, 601 F. Supp. 3d 147, 163–64 (W.D. Tex. 2022); see also Cano v. Davis, 211 F. Supp. 2d 1208, 1249 (C.D. Cal. 2002). The Challengers also find support for this view in Bartlett. When refusing to allow plaintiffs to lessen Gingles’s first precondition (that a racial minority would qualify as a majority in some other district), the plurality opinion noted in dicta that its holding did not reach cases alleging “intentional discrimination against a racial minority.” 556 U.S. at 20 (plurality opinion). Regardless, this debate may be purely academic. Tennessee’s counsel conceded that the intentional drawing of a single-member district’s lines to harm racial minorities would likely show that the legislature predominately used race during the redistricting. Tr., R.68, PageID 1653. So whether or not plaintiffs could establish the “effect” element of a vote-dilution claim, they likely could prove a racial-gerrymandering claim if they could establish the “purpose” element of that vote-dilution claim. In the end, then, perhaps constitutional vote-dilution claims and racial- gerrymandering claims largely overlap in this context of single-member districts. At day’s end, though, we opt not to resolve these vote-dilution claims on this ground. The parties’ briefing spends little time on this complex topic. And since the Challengers have failed to plausibly allege discriminatory purpose, their claims fail on that alternative ground alone. B. Discriminatory Purpose A vote-dilution claim requires plaintiffs to prove that the legislature drew the maps for a specific purpose (or with a specific intent): “to minimize or cancel out the voting potential of” voters of a certain race or ethnicity. Miller, 515 U.S. at 911 (quoting Bolden, 446 U.S. at 66 (plurality opinion)); see Abbott v. Perez, 585 U.S. 579, 586 (2018); Rogers, 458 U.S. at 617. The Challengers’ Complaint fails to allege facts that plausibly show this invidious purpose or intent— again, because of the obvious partisanship explanation for the maps. 1 When compared to the Supreme Court’s test for racial-gerrymandering claims, its test for vote-dilution claims establishes a more demanding “purpose” element in one sense and an arguably more lenient “purpose” element in another. Start with the more demanding element. A racial- gerrymandering claim requires plaintiffs to show that the legislature relied primarily on race when drawing the maps—no matter the reason for doing so. See Cooper, 581 U.S. at 291 n.1. A vote- dilution claim, by contrast, requires plaintiffs to show that the legislature relied on race for an invidious reason: to harm a racial group’s ability to elect the group’s preferred candidates. See Miller, 515 U.S. at 911. And plaintiffs cannot prove this invidious reason merely by showing that the legislature knew that the revised map would have such harmful effects on the racial group. See Iqbal, 556 U.S. at 676. Rather, the legislature must have drawn the map “‘because of,’ not merely ‘in spite of,’ [those] adverse effects[.]” Id. at 677 (quoting Feeney, 442 U.S. at 279). Two examples demonstrate this test. Suppose that a legislature knows that its capital- punishment law disparately affects minority defendants, who receive a death sentence more often than do non-minority defendants. See McCleskey v. Kemp, 481 U.S. 279, 286, 297–98 (1987). To prove intentional discrimination, a challenger would have to show that the legislature passed the law “because of” this racial effect. Id. at 298. If the legislature enacted the law to deter crime, the law would pass muster. See id. Or suppose that a legislature knows that a veteran’s preference in employment disparately affects women because over 98% of veterans are men. Feeney, 442 U.S. at 270–71. To prove intentional discrimination, a challenger would have to show that the legislature enacted this preference to harm women’s employment options. See id. at 279. If the legislature enacted the law to help veterans, the law would pass muster. Id. at 279–80. Turn to the arguably more lenient standard. A racial-gerrymandering claim requires plaintiffs to show that the legislature predominately used race when drawing the maps. Miller, 515 U.S. at 916. A vote-dilution claim, by contrast, requires plaintiffs to show only that racial discrimination was a “‘substantial’ or ‘motivating’ factor” in the redistricting. Hunter, 471 U.S. at 228 (quoting Mt. Healthy City Bd. of Educ. v. Doyle, 429 U.S. 274, 287 (1977)) (emphases added). So they need not establish that the challenged action “rested solely on racially discriminatory purposes.” Arlington Heights, 429 U.S. at 265. Indeed, they need not even establish that discrimination was “the ‘dominant’ or ‘primary’” purpose. Id. If plaintiffs satisfy this motivating-factor test, though, the defendants can still escape liability. To do so, they must show that the legislature would have drawn the same map “even had the impermissible purpose not been considered.” Id. at 270 n.21. That is, the defendants must show that the challenged map “would have been enacted” even if the legislators had not been motivated by racial animus. Hunter, 471 U.S. at 228. Or to put this concept in the language of the law of torts, the defendant’s ultimate liability requires a “but-for” causal relationship between the unconstitutional motive and the challenged map. Nieves v. Bartlett, 587 U.S. 391, 399 (2019). How can plaintiffs prove that a legislature enacted a law to harm voters of a certain race? The Court has long recognized the “hazards” of trying to divine the “intent” of a collective body made up of many legislators with many motivations. Palmer v. Thompson, 403 U.S. 217, 224–25 (1971); see United States v. O’Brien, 391 U.S. 367, 383–84 (1968); Fletcher v. Peck, 10 U.S. 87, 130–31 (1810). Should courts try to deduce a single objective intent of the whole body based on all the circumstances surrounding a law’s enactment? Or should they ask about the subjective intent of each legislator and inquire into whether a majority of those who voted on a law harbored racial animus? Cf. Va. Uranium, Inc. v. Warren, 587 U.S. 761, 775–77 (2019) (lead opinion). The Court has yet to provide a test unmistakably answering these questions. Despite what the Court has called the “problematic” nature of this endeavor, Hunter, 471 U.S. at 228, it has simultaneously recognized that plaintiffs can seek to identify a legislative body’s reasons for a law through any direct and circumstantial evidence that exists on the question. See Arlington Heights, 429 U.S. at 266. Direct evidence may include statements by those who voted for the challenged law that they did so for a racist reason. See Hunter, 471 U.S. at 229. Nowadays, however, legislators rarely provide this type of direct evidence of their invidious motives. So plaintiffs typically rely on a circumstantial case. A challenged law’s harmful impact on racial minorities itself offers some circumstantial evidence of a discriminatory purpose, especially if the law has no rational explanation. Arlington Heights, 429 U.S. at 266. But it will fall far short of establishing that purpose when the law is “readily explainable on grounds apart from race[.]” Bolden, 446 U.S. at 70 (plurality opinion). So the Court has told us to conduct a “sensitive inquiry” into the totality of the circumstances relevant to whether legislators passed the challenged law to harm minorities. Arlington Heights, 429 U.S. at 266–68; see also Rogers, 458 U.S. at 618. It has also identified relevant factors to guide the analysis. Courts might consider, for example, whether a law’s “historical background” suggests that legislators passed it to achieve a discriminatory objective. Arlington Heights, 429 U.S. at 267 (citing, among others, Lane v. Wilson, 307 U.S. 268, 275–76 (1939)). This backdrop might reveal racial animus if the legislature “commonly utilized” “discriminatory practices” in the past, abandoned the practices because of court injunctions or “civil rights legislation,” and attempted to reimplement them through laws that, “though neutral on their face, serve[d] to maintain the status quo.” Rogers, 458 U.S. at 625; see Lane, 307 U.S. at 276. Courts also might ask whether the “sequence of events” leading up to the law’s passage suggests something invidious. Arlington Heights, 429 U.S. at 267. “Departures” from a body’s usual rules might imply hidden animus. Id. So too could contemporaneous statements by legislators recorded in the “legislative . . . history” of a law. Id. at 268. Courts have extended this totality-of-the-circumstances approach to the vote-dilution context. See, e.g., Common Cause Fla. v. Byrd, __ F. Supp. 3d __, 2024 WL 1308119, at *29 (N.D. Fla. Mar. 27, 2024); League of United Latin Am. Citizens v. Abbott, 617 F. Supp. 3d 622, 632 (W.D. Tex. 2022). As for a history of discrimination, they have considered, among other things, whether courts have enjoined the legislature in prior redistricting efforts. See League of United Latin Am. Citizens, 617 F. Supp. 3d at 632. As for the sequence of events, they have considered, among other things, whether the legislature drew the map “in secret” and “shut out” “minorities[] and certain representatives” from participating. See id. When evaluating these types of circumstantial factors, though, courts must remember that the presumption of legislative good faith extends to vote-dilution claims alleging that legislators acted with racial animus. See Abbott, 585 U.S. at 603. That is because this presumption rests in part on the notion that neutral federal courts should act cautiously before we “hurl such accusations at the political branches.” Alexander, 144 S. Ct. at 1236. On the other hand, if a plaintiff proves that racial discrimination was a “motivating factor” for the challenged map, that fact would suffice to rebut the presumption of good faith. Arlington Heights, 429 U.S. at 265–66. 2 The Challengers’ Complaint does not meet these standards. Keep in mind that the Complaint must do more than plausibly allege that Tennessee’s legislators knew that their Republican-friendly map would harm voters who preferred Democratic candidates—including the higher percentage of minority voters who preferred those candidates. Compl., R.1, PageID 17; see Iqbal, 556 U.S. at 676; Cromartie I, 526 U.S. at 558 (Stevens, J., concurring in the judgment). The Complaint instead must plausibly allege that the legislators enacted the map in part “because of” its disproportionate effect on minority voters. McCleskey, 481 U.S. at 298. As with their political- gerrymandering claims, however, the well-pleaded facts in the Complaint fail to satisfy this plausibility test. See Iqbal, 556 U.S. at 682–83; see also Alexander, 144 S. Ct. at 1234–36. Two initial points set the stage. For starters, the Challengers do not argue that this case resembles Hunter—in which the decisionmakers expressly stated their “zeal for white supremacy” when enacting the challenged provision. 471 U.S. at 229. The Complaint cites no “contemporary statements” in the public “legislative” “history” that would indicate that legislators voted on the maps for this racist reason. Arlington Heights, 429 U.S. at 268. Next, the Challengers do not suggest that their allegations make this case like the “rare” one in which the alleged disparate effect alone is “unexplainable on grounds other than race[.]” Id. at 266. The maps are “readily explainable on [political] grounds apart from race” due to the correlation between the two. Bolden, 446 U.S. at 70 (plurality opinion). So the Complaint must plausibly allege a circumstantial case of racism. The Challengers attempt to do so both with claims about the way that the legislature created the maps and with claims about other laws passed near the same time. Yet these allegations do not suffice to plausibly rebut the more straightforward “explanation”: naked partisanship. Twombly, 550 U.S. at 567. Sequence of Events. The Challengers first argue that the “sequence of events” leading up to the legislature’s passage of the maps implies racial animus. Arlington Heights, 429 U.S. at 267. The Complaint alleges that the General Assembly passed the maps “very quickly”—in less than a month—“without much debate” or transparency. Compl., R.1, PageID 3; see id., PageID 17–23. But these allegations do not plausibly “rule out the possibility” that the legislators acted for political—not racist—reasons. Alexander, 144 S. Ct. at 1243. In fact, the Supreme Court has previously held that the “brevity of the legislative process” in creating maps does not “overcome the presumption of legislative good faith”—at least when the legislature had a good reason for the shortened process. Abbott, 585 U.S. at 610–11. Besides, the Complaint did not suggest that the purportedly quick process departed from any “procedural” rules that the General Assembly must follow when considering legislation. Arlington Heights, 429 U.S. at 267. In addition, the Complaint repeatedly alleges that the General Assembly deliberated long enough for many individuals and groups—including many Challengers—to air their concerns about the “importance of keeping all of Davidson County” together so that minorities could elect their preferred candidates. Compl. R.1, PageID 2, 6–10. According to the Complaint, for example, agents of the NAACP testified before “the Senate and House redistricting committees” and submitted a proposed “Congressional concept map” for those committees to consider. Id., PageID 6. Likewise, agents of the League of Women Voters testified about the “impact” that the proposed maps would have on “Black voters and other voters of color.” Id., PageID 7. Did prior redistricting cycles include more deliberation? Did they take longer or shorter? Did any of these groups have more time to raise their concerns during those past cycles? The Complaint leaves us in the dark. The “sequence of events” as currently pled does not plausibly raise any racism red flags. Arlington Heights, 429 U.S. at 267. History of Discrimination. The Challengers also argue that Tennessee’s legislature “continued to show its racial animus” with other legislation. Compl., R.1, PageID 26. They correctly note that a law’s “historical background” can provide evidence of a discriminatory motive, especially “if it reveals a series of official actions taken for invidious purposes.” Arlington Heights, 429 U.S. at 267. But they do not identify any “[p]ast discrimination” that has prohibited racial minorities from presently registering and voting in equal numbers as other groups. See Rogers, 458 U.S. at 625. Nor do they suggest that the General Assembly specifically adopted its map to keep any such “status quo” of ongoing voting discrimination. Id.; see Greater Birmingham Ministries v. Sec’y of State for State of Alabama, 992 F.3d 1299, 1325 (11th Cir. 2021). Instead, the Challengers point to a mix of laws, bills, or statements that have little to do with redistricting. We will highlight a few examples. The Complaint alleges that the legislature sought to harm minority voters through a race-neutral law regulating entities that conduct voter- registration drives. Compl., R.1, PageID 26; see Tenn. State Conf. of NAACP v. Hargett, 420 F. Supp. 3d 683, 691–92 (M.D. Tenn. 2019). But this conclusory allegation of racial animus (articulated in one paragraph) does not warrant the presumption of truth. See Iqbal, 556 U.S. at 680–81. And although a district court preliminarily enjoined the law, it did so on speech grounds rather than discrimination grounds. See NAACP, 420 F. Supp. 3d at 698–711. The legislature has since repealed the law anyway. Similarly, the Complaint alleges that the legislature passed a law barring public-school teachers from “engaging their students in academic discussions regarding structural racism and unconscious bias[.]” Compl., R.1, PageID 27. But this law, by its terms, bars instructional materials conveying, among other things, that one “race” is “inherently superior to” another or that individuals “should be discriminated against” based on their race. Tenn. Code §§ 49-6-1019(a)(1), (3); 49-7-1902(1)(A), (C). In the face of this race-neutral language, the conclusory allegation that the law furthers a discriminatory purpose does not suffice. See Iqbal, 556 U.S. at 680–81. The same problem underlies the Complaint’s claims about a race-neutral law that reduced the number of seats on the Davidson County Metro Council. Compl., R.1., PageID 27. The Complaint again makes only a conclusory allegation (one not entitled to the presumption of truth) that the legislature passed this law with an invidious motive. See Iqbal, 556 U.S. at 680–81. And although a state court enjoined the law under Tennessee’s Home Rule Amendment, the court did so without identifying any hidden animus behind the law. See Op., R.43-1, PageID 332–38. Apart from conclusory insinuations about various laws, the Complaint also relies on the votes or statements of individual legislators about equally unrelated matters. It points out, for example, that a “handful” of legislators voted against a constitutional amendment that successfully repealed a provision permitting “slavery and involuntary servitude” as criminal punishments. Compl., R.1, PageID 27. Also according to the Complaint, when a Criminal Justice Committee debated a bill adopting alternative methods for implementing the death penalty, a legislator offensively “suggested adding ‘hanging by a tree’ as a method of execution—a clear reference to lynchings.” Id., PageID 28. The Complaint adds that the Tennessee Black Caucus’s resolution to remove this member for this “clear reference to lynchings” failed. Id. Lastly, the Complaint alleges that, several months after the legislature enacted the new maps, it passed a “resolution to expel two Black representatives” (Justin Jones and Justin Pearson) for participating in a protest at the capitol—even though the legislature failed to garner the votes to pass a resolution to expel a white representative (Gloria Johnson). Id. It further notes that some unidentified legislators criticized Representative Pearson for “wearing West African attire on the floor of the House” even though no official dress code prohibited him from doing so. Id. All told, these statements and votes of a “handful” of legislators on unrelated topics do not plausibly suggest that the whole legislature passed the legislative maps to discriminate against racial minorities. O’Brien, 391 U.S. at 384. While they might suggest the “possibility of misconduct,” Iqbal, 556 U.S. at 679, they do not cross “the line between possibility and plausibility,” Twombly, 550 U.S. at 557. In sum, “discrimination is not a plausible conclusion” to draw from the Complaint’s speculative claims of “purposeful, invidious discrimination” based on tangential matters, especially given the “obvious alternative explanation” that the legislature drew the maps for a purely partisan reason. Iqbal, 556 U.S. at 682 (quoting Twombly, 550 U.S. at 567). IV. Laches and Sovereign Immunity This conclusion leaves two final issues. Tennessee separately argues that laches should bar this suit in its entirety. And Governor Lee separately argues that we should dismiss him on sovereign-immunity and standing grounds. We agree with the second argument but not the first. A. Laches Tennessee asserts that the Challengers filed this suit too late. The revised maps became law in February 2022. Compl. R.1, PageID 24. But the Challengers did not sue until August 2023. Id., PageID 49. As a result, they let the 2022 election proceed under the new maps and do not seek relief ahead of the 2024 election. This litigation thus could not affect any election until 2026— over halfway through the lifespan of the maps. So Tennessee says “laches” should bar the suit. Courts of equity originally created the doctrine of laches—or “unreasonable, prejudicial delay in commencing a suit”—as a “gap-filling” measure to bar belated claims when the legislature had not adopted a statute of limitations to cover them. Petrella v. Metro-Goldwyn-Mayer, Inc., 572 U.S. 663, 678, 680 (2014). The defense now generally applies only if a plaintiff unreasonably delayed in suing and the delay prejudiced the defendant. See ACLU of Ohio v. Taft, 385 F.3d 641, 647 (6th Cir. 2004); Am. Addiction Ctrs., Inc. v. Nat’l Ass’n of Addiction Treatment Providers, 515 F. Supp. 3d 820, 838 (M.D. Tenn. 2021). Given the doctrine’s roots in equity, though, district courts retain residual discretion to reject a laches defense even if a claim satisfies both elements. See Memphis A. Philip Randolph Inst. v. Hargett, 473 F. Supp. 3d 789, 793 (M.D. Tenn. 2020). Even more important at this stage, laches falls within the list of “affirmative defenses” in Federal Rule of Civil Procedure 8(c)(1). See Petrella, 572 U.S. at 680. A defendant thus bears the burden of proving its elements. See EEOC v. Watkins Motor Lines, Inc., 463 F.3d 436, 439 (6th Cir. 2006). And unlike with a partisan-gerrymandering defense, the complaint need not plead facts showing that a plaintiff sued in a reasonable time. See Cataldo v. U.S. Steel Corp., 676 F.3d 542, 547 (6th Cir. 2012); Am. Addiction Ctrs., 515 F. Supp. 3d at 839. Rather, a district court may dismiss a complaint based on an affirmative defense only if the complaint’s allegations “affirmatively” trigger the defense. Cataldo, 676 F.3d at 547. So a district court may rely on laches at the pleading stage only in the rare case in which plaintiffs plead themselves out of court. This is not one of the rare cases. To begin with, the parties’ briefing leaves us unsure how the defense should apply, if at all, in this § 1983 context. The Supreme Court looks to state law to identify the statute of limitations and tolling rules for a § 1983 claim, but it looks to federal law to determine when such a claim accrues and starts the running of the limitations period. See Wallace v. Kato, 549 U.S. 384, 387–88, 395 (2007). Under this dichotomy, should the applicability of laches to a § 1983 claim depend on federal law or state law? Cf. Herman v. City of Chicago, 870 F.2d 400, 403 (7th Cir. 1989). If a matter of federal law, the Supreme Court clarified that laches has a narrow domain for claims otherwise subject to a statute of limitations. See SCA Hygiene Prods. Aktiebolag v. First Quality Baby Prods., LLC, 580 U.S. 328, 334–35 (2017); Petrella, 572 U.S. at 678. In Tennessee, § 1983 claims are subject to a one-year statute of limitations. See Dibrell v. City of Knoxville, 984 F.3d 1156, 1161 (6th Cir. 2021). If a plaintiff files a § 1983 claim within this limitations period, laches might not apply at all to bar legal relief (such as damages). See Petrella, 572 U.S. at 678– 79; Ivani Contracting Corp. v. City of New York, 103 F.3d 257, 259–62 (2d Cir. 1997). And while the Challengers seek equitable relief (an injunction), laches may well bar that relief only in “extraordinary circumstances” if their § 1983 claims are filed within the applicable limitations period. Petrella, 572 U.S. at 685. Yet we need not decide any of these issues now. At the least, Tennessee has not shown that the Complaint “affirmatively” proves the two requirements to establish its laches defense. Cataldo, 676 F.3d at 547. For starters, Tennessee cites no case that has held at the pleading stage that a redistricting plaintiff acted unreasonably by waiting a single election before suing. Indeed, a Supreme Court plurality once suggested that data from a single election alone might not even suffice to show a map’s political effects. Davis v. Bandemer, 478 U.S. 109, 135 (1986) (plurality opinion). Given the Challengers’ relatively modest delay, then, they should have the right to offer “facts” about the delay’s alleged “unreasonableness” and their “excuse” for it. Am. Addiction Ctrs., 515 F. Supp. 3d at 839. In the cases that Tennessee cites, by contrast, the plaintiffs waited much longer to sue. See, e.g., White v. Daniel, 909 F.2d 99, 102–03 (4th Cir. 1990) (17-year delay); Sanders v. Dooly County, 245 F.3d 1289, 1291 (11th Cir. 2001) (per curiam) (8-year delay). Likewise, the Complaint does not “affirmatively” show any prejudice to Tennessee. Cataldo, 676 F.3d at 547. The State argues that the delay has prejudiced it because a potential witness has died. Mem., R.43, PageID 293−94; cf. Nartron Corp. v. STMicroelectronics, Inc., 305 F.3d 397, 412 (6th Cir. 2002). But the Complaint did not plead this fact. So Tennessee’s argument confirms that it raised this laches defense prematurely. See Cataldo, 676 F.3d at 547. Without discovery, we have no idea what this witness would have said about the redistricting. Tennessee also argues that belated changes to legislative maps burden election officials (who must implement new maps) and the public (who must learn new districts). Here again, however, Tennessee cites no case that has found this prejudice sufficient at the pleading stage. It instead cites cases that found prejudice when plaintiffs waited to seek preliminary relief until just before an election. These cases applied laches narrowly to bar that initial relief—not the entire suit. See, e.g., Crookston v. Johnson, 841 F.3d 396, 398–99 (6th Cir. 2016); Memphis A. Philip Randolph Inst., 473 F. Supp. 3d at 795–801. They thus relied on what today more commonly goes by the “Purcell principle,” which bars last-minute requests for election changes. See Crookston, 841 F.3d at 398 (citing Purcell v. Gonzalez, 549 U.S. 1, 5–6 (2006) (per curiam)). Here, however, the Challengers forthrightly disavow seeking relief ahead of the 2024 election. So these cases do not help Tennessee. In short, this laches defense does not belong in a motion to dismiss. B. Governor Lee Governor Lee raises a better argument. He contends that we should dismiss him from this suit on either standing or sovereign-immunity grounds. We usually must assure ourselves of our subject-matter jurisdiction (including a plaintiff’s standing to sue, either generally or with respect to a particular defendant) before deciding a claim on its merits. See Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 94–102 (1998). But we can dismiss a claim on non-merits grounds other than subject-matter jurisdiction (such as personal jurisdiction or forum non conveniens) in order to avoid a complicated question of subject-matter jurisdiction. See Sinochem Int’l Co. v. Malaysia Int’l Shipping Corp., 549 U.S. 422, 431–32 (2007). And sovereign-immunity qualifies as this type of threshold issue that we may address ahead of standing. See United States v. Abbott, 85 F.4th 328, 333 (5th Cir. 2023); Meyers v. Oneida Tribe of Indians of Wis., 836 F.3d 818, 821– 23 (7th Cir. 2016); Rockefeller v. Bingaman, 234 F. App’x 852, 854 (10th Cir. 2007) (order); Galvan v. Fed. Prison Indus., Inc., 199 F.3d 461, 463 (D.C. Cir. 1999). We thus opt to bypass standing and will dismiss the claim against Governor Lee on sovereign-immunity grounds alone. States have sovereign immunity from suits filed by private parties in federal court. See Seminole Tribe of Fla. v. Florida, 517 U.S. 44, 57–73 (1996). A suit against Governor Lee in his “official capacity” also counts as a suit against Tennessee that usually triggers its sovereign immunity. Will v. Mich. Dep’t of State Police, 491 U.S. 58, 71 (1989). Yet the Supreme Court has carved out a “narrow exception” to this rule. Whole Woman’s Health v. Jackson, 595 U.S. 30, 39 (2021). An “official capacity” suit against an officer does not qualify as a suit against the State if it seeks to stop the officer from enforcing a state law that conflicts with federal law. See id. (citing Ex parte Young, 209 U.S. 123, 159–60 (1908)); Will, 491 U.S. at 71 n.10. To fall within this so-called “Ex Parte Young exception” to sovereign immunity, though, a plaintiff may not name just any state actor as the defendant. The plaintiff may sue only those officials that the State tasks with enforcing the challenged law. See Russell v. Lundergan-Grimes, 784 F.3d 1037, 1047 (6th Cir. 2015); see Young, 209 U.S. at 157. This part of Ex Parte Young follows from the general principle that courts lack the power to invalidate state laws in the abstract; rather, courts have the power to bar state actors from enforcing those laws against the plaintiffs. See Whole Woman’s Health, 595 U.S. at 44; California v. Texas, 593 U.S. 659, 672–73 (2021). To sue a state official under Ex Parte Young, then, plaintiffs must identify the “legal or administrative actions” of the official that they seek to enjoin. Russell, 784 F.3d at 1048; see Children’s Healthcare Is a Legal Duty, Inc. v. Deters, 92 F.3d 1412, 1416 (6th Cir. 1996). This framework poses a problem for the Challengers’ claims against Governor Lee. The Challengers cite nothing in Tennessee law that gives the governor any “enforcement” authority to carry out elections under the maps that they challenge. Children’s Healthcare, 92 F.3d at 1416. To the contrary, Tennessee law identifies the coordinator of elections as the “chief administrative election officer” with the duty to “maintain uniformity in the application, operation and interpretation of the election code.” Tenn. Code Ann. § 2-11-201(b). It also gives the secretary of state (not the governor) the power to appoint and terminate this elections administrator. Id. § 2- 11-201(a). Tennessee’s Constitution, in turn, places the authority to appoint the secretary of state in the legislature (not the governor). Tenn. Const. art. 3, § 17. So what do the Challengers rely on as their hook to sue Governor Lee? They point to his general authority to enforce Tennessee law and his act of signing the legislation that enacted the maps. Both theories fall short. Starting with their first theory, the Challengers argue that Tennessee’s Constitution vests the “Supreme Executive power” in the “Governor,” who must “take care that the laws be faithfully executed.” Tenn. Const. art. III, §§ 1, 10. But the overwhelming weight of precedent holds that this “general duty to enforce state law” does not suffice to allow a plaintiff to sue the governor under Ex Parte Young. 13 Charles Alan Wright & Arthur R. Miller, Federal Practice & Procedure § 3524.3, at 384–85 (3d ed. 2008); see Abbott, 85 F.4th at 334–35; Hendrickson v. AFSCME Council 18, 992 F.3d 950, 967 (10th Cir. 2021); Church v. Missouri, 913 F.3d 736, 749 (8th Cir. 2019); Osterback v. Scott, 782 F. App’x 856, 859 (11th Cir. 2019) (per curiam); Doe v. Holcomb, 883 F.3d 971, 976 (7th Cir. 2018); Waste Mgmt. Holdings, Inc. v. Gilmore, 252 F.3d 316, 331 (4th Cir. 2001). As the Court in Ex Parte Young itself cautioned, the contrary rule would allow plaintiffs to sue a governor over every state law on the books. 209 U.S. at 157 (quoting Fitts v. McGhee, 172 U.S. 516, 530 (1899)). To be sure, the Challengers point to several Sixth Circuit cases that—if read broadly— might sit in tension with this rule. See League of Women Voters of Ohio v. Brunner, 548 F.3d 463, 475 n.16 (6th Cir. 2008); Lawson v. Shelby County, 211 F.3d 331, 335 (6th Cir. 2000); Allied Artists Picture Corp. v. Rhodes, 679 F.2d 656, 665 n.5 (6th Cir. 1982). But the Sixth Circuit has also held that an official’s “[g]eneral authority to enforce” a law—without more—does not trigger the Ex Parte Young exception. Block v. Canepa, 74 F.4th 400, 412 (6th Cir. 2023) (quoting Children’s Healthcare, 92 F.3d at 1416). And none of the Challengers’ cited cases says otherwise. In League of Women Voters, the court suggested that Ohio’s governor had the specific power to “control” county boards of election. 548 F.3d at 475 n.16. The Challengers allege no similar control. Lawson, by comparison, lumped all the defendants together. See Lawson, 211 F.3d at 335. It did not recognize Ex Parte Young’s requirement that a particular defendant must have a “connection with the enforcement of the act” in order to be suable. Children’s Healthcare, 92 F.3d at 1416 (quoting Ex Parte Young, 209 U.S. at 157). So any statements in Lawson on this topic amount to nonbinding dicta. See Wright v. Spaulding, 939 F.3d 695, 701–02 (6th Cir. 2019). The footnote from Allied Artists with discussion on this issue also does the Challengers no good. There, the major motion-picture companies had sued Ohio’s governor to enjoin the enforcement of a state law regulating their contracts with theaters. See 679 F.2d at 659, 665 n.6. Although the state law did not itself contain “specific state enforcement provisions,” the court interpreted Ohio’s statutory scheme as allowing the governor to bring an equitable quo warranto action against a motion-picture company that violated the law’s terms. Id. at 665 n.6 (citing Ohio Rev. Code § 2733.02). The Challengers do not claim that Governor Lee has any similar enforcement authority in this case. Turning to their second theory, the Challengers argue that Tennessee’s Constitution gives the governor the power to approve or veto bills. Tenn. Const. art. III, § 18. Because Governor Lee signed the legislative maps into law, this argument goes, his “duty sufficiently connects him with the enforcement of the challenged legislative enactment” to enable him to be sued over the maps under Ex Parte Young. Resp., R.46, PageID 434. But this theory is a non-starter because a Tennessee governor does not have a duty (as opposed to a prerogative) to sign legislative maps (or, for that matter, anything else passed by the General Assembly) into law. The Challengers conflate the governor’s (non-mandatory) act of signing legislation into law with the governor’s duty (such as it is) to generally see to the enforcement of the law once passed. And the Challengers identify no case holding that a decision to sign a bill into law triggers the Ex Parte Young exception. That theory conflicts with Ex Parte Young in two ways. The Challengers cannot seek “prospective relief” against the governor’s act of signing this law because he did so in the past. Russell, 784 F.3d at 1047 (citation omitted; emphasis added). Indeed, they sued Governor Lee in his official capacity, so his successor would automatically become a part of this suit. See Lewis v. Clarke, 581 U.S. 155, 162 (2017). The Challengers’ view, then, would force a later governor to litigate a prior governor’s act of signing a bill into a law. Not only that, Governor Lee’s conduct looks more like a legislative act to pass a law than an “executive” act to “enforc[e]” it. Whole Woman’s Health, 595 U.S. at 39; see Ariz. State Legis. v. Ariz. Indep. Redistricting Comm’n, 576 U.S. 787, 808 (2015). The Challengers conflate these two very different types of governmental powers. They thus would read Ex Parte Young in an unprecedented way by seemingly allowing federal courts to enjoin state legislatures from passing laws. Cf. California, 593 U.S. at 673. That view takes the exception too far. * * * All told, we deny Tennessee’s motion to dismiss the Complaint on laches ground. We grant the motion to dismiss Governor Lee on sovereign-immunity grounds. And we grant the motion to dismiss the Complaint for failing to state plausible claims of racial gerrymandering and vote dilution. But we grant this last relief without prejudice. A district court “should freely give leave [to amend] when justice so requires.” Fed. R. Civ. P. 15(a)(2). We find this standard met here because the Supreme Court in Alexander clarified the applicable law only after the parties completed their motion-to-dismiss briefing and only one day before oral argument on that motion. And the Challengers may be able to allege facts that plausibly “disentangle race from politics” now that they know our view of the law after Alexander. 144 S. Ct. at 1233. We thus should give them a chance to amend since “a more carefully drafted complaint might state a claim[.]” Walker v. Massey, 2023 WL 28435, at *7 (M.D. Tenn. Jan. 3, 2023) (citation omitted); see 5B Charles A. Wright et al., Federal Practice & Procedure § 1357, at 501 (2024). The Challengers will have thirty days to seek leave to file an amended complaint curing the pleading defects that we have identified in the original one. The Court retains jurisdiction in the meantime.