Before DAVIS, COSTA, and OLDHAM, Circuit Judges. GREGG COSTA, Circuit Judge: Environmental groups sued ExxonMobil under the Clean Air Act for thousands of unauthorized emissions from the company’s complex in Baytown, Texas. The first time we considered the case, we found Exxon liable for many of those violations and remanded for the district court to determine an appropriate penalty. When the case came to us again a few years later, we primarily addressed whether the plaintiffs have standing to seek redress for those violations. The case now returns to us after a limited remand for factfinding on traceability and Exxon’s affirmative defenses. Finding no error in the district court’s fact-intensive analysis of standing or penalty, we affirm. I This long-pending Clean Air Act suit stems from operations at ExxonMobil’s massive Baytown complex. The complex, which houses a refinery, a chemical plant, and an olefins plant, is heavily regulated by federal permits that are enforced jointly by the Texas Commission on Environmental Quality and the United States Environmental Protection Agency. Environment Texas Citizen Lobby, Inc. v. ExxonMobil Corp. (ETCL I), 824 F.3d 507, 512 (5th Cir. 2016). The permits require Exxon to document, and sometimes to report, certain instances of noncompliance. Exxon’s substantive obligations and reporting requirements are explained in detail in ETCL I, 824 F.3d at 512–22. Environment Texas Citizen Lobby and Sierra Club, on behalf of their members who live, work, and recreate near Baytown, sued Exxon under the Clean Air Act’s citizen suit provision, 42 U.S.C. § 7604(a)(1), for thousands of self-reported permit violations that occurred between October 2005 and September 2013. After some litigation, Exxon stipulated to 16,386 days of violations. Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp. (ETCL II), 968 F.3d 357, 363 (5th Cir. 2020); see also id. at 363 n.1 (explaining that “[i]f an emissions event released multiple pollutants, each with its own emissions standard, . . . each standard violat[ion] [counts] as a separate day of violation”). Those violations fall into five categories, including unplanned emissions, emissions exceeding authorized rates, and unsafe or unauthorized flaring. See id. at 363 (describing the five types of violations). After a bench trial, the district court found only a few of the violations actionable and declined to assess a penalty against the company. Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp., 66 F. Supp. 3d 875, 895–902, 911–12 (S.D. Tex. 2014). We agreed with the environmental groups that the district court erred in its analysis of Exxon’s substantive liability and abused its discretion in addressing three of the factors that courts consider in assessing civil penalties. ETCL I, 824 F.3d at 515–23 (liability), 524–33 (remedies); see also 42 U.S.C. § 7413(e)(1) (listing the penalty factors). On remand, the district court reconsidered the factors and fined Exxon $19.95 million dollars. Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp., 2017 WL 2331679, at*25-31 (S.D. Tex. Apr. 26, 2017). Then Exxon appealed. The company asserted that the plaintiffs only proved standing for a handful of violations and challenged the new penalty determination. This panel determined that the organizational plaintiffs established two out of the three requirements for Article III standing: injury- in-fact and redressability. ETCL II, 968 F.3d at 367–68 (injury); id. at 371– 72 (redressability). We further explained that the district court should analyze traceability by asking whether each violation (1) “causes or contributes to the kinds of injuries” alleged by the plaintiffs and (2) has a “‘specific geographical or other causative nexus’ such that the violation could have affected their members.” Id. at 369–70 (quoting Sierra Club, Lone Star Chapter v. Cedar Point Oil Co., Inc., 73 F.3d 546, 557, 558 n.24 (5th Cir. 1996) (internal quotation marks in first quotation omitted)). We remanded for the limited purpose of determining which violations are fairly traceable to Exxon’s actions[1] and reserved judgment on the appropriate penalty. Id. at 374–75. Our instructions had a significant impact on remand. Applying our guidance, the district court determined that plaintiffs proved traceability for only 3,651 of the 16,386 violation days. Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp., 524 F. Supp. 3d 547, 565 (S.D. Tex. 2021). It then revised its penalty calculation. It held that a penalty was appropriate because of the size, duration, and seriousness of the violations as well as Exxon’s economic benefit from noncompliance. Id. at 576. It ordered Exxon to pay $14.25 million dollars, lessening the penalty by more than five million dollars to reflect the reduced number of justiciable violations. Id. at 577. Because Exxon disagrees with both the standing and penalty determinations, we now weigh in for the third time. II Only those disputes that meet the “irreducible constitutional minimum” of standing can be heard in a federal forum. Lujan v. Defs. of Wildlife, 504 U.S. 555, 560 (1992). The three components of standing are familiar: injury-in-fact, traceability, and redressability. Id. at 560–61. Clean Air Act plaintiffs must prove these elements for each claimed violation. ETCL II, 968 F.3d at 365–67. We first consider whether the plaintiff organizations met this burden. A After our last remand, the district court made additional findings on traceability. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 555-65. Exxon does not challenge that factfinding. Instead, the company dedicates more than two thirds of its brief to asking us to revisit our approach to standing. Exxon takes two shots at our standing framework. First, it says that a recent decision from the Supreme Court abrogates our finding of injury-in-fact. Second, it argues that our traceability precedent is overly broad and risks exceeding the bounds of Article III. Neither reason compels us to redo our prior opinion. Nor could we; our prior opinion is law of the case. White v. Murtha, 377 F.2d 428, 431–32 (5th Cir. 1967). The reason for that rule ring true in this long-running case: Suits would never end “if a question, once considered and decided by [a court] were to be litigated anew in the same cases upon any and every subsequent appeal.” Id. at 431 (quoting General Am. Life Ins. Co. v. Anderson, 156 F.2d 615, 618 (6th Cir. 1946)). A prior ruling in a case thus can be disturbed only if new evidence is substantially different, controlling authority has changed, or maintaining the decision would result in manifest injustice. Id. at 432. None of those exceptions apply. 1 Exxon first takes aim at our finding of injury-in-fact. We previously determined that the plaintiffs “easily” met their burden of proving injury for each alleged violation because “throughout the claims period, [they] regularly saw flares, smoke, and haze coming from the complex; smelled chemical odors; suffered from allergy-like or respiratory problems; feared for their health; refrained from outdoor activities; or moved away.” ETCL II, 968 F.3d at 367–68. Exxon asserts that even if this holding was correct when decided, TransUnion LLC v. Ramirez, 141 S. Ct. 2190 (2021), is an intervening change in law that requires us to reconsider. Of course, Supreme Court rulings can overrule our precedent. But we cannot disregard our precedent simply because we think the Court might someday disagree with it. See United States v. Alcantar, 733 F.3d 143, 146 (5th Cir. 2013). Until the highest court “unequivocally” overrules our precedent, we are bound by it. United States v. Zuniga-Salinas, 945 F.2d 1302, 1306 (5th Cir. 1991). This aspect of the rule of orderliness rule promotes stability in the law. United States v. Longoria, 958 F.3d 372, 378 (5th Cir. 2020). In TransUnion, a class of consumers sued a credit reporting agency for failing to reasonably ensure the accuracy of their credit files. 141 S. Ct. at 2200. Although all of the plaintiffs had a cause of action under the Free Credit Reporting Act, the Court held that only some of them had Article III standing. Id. at 2201, 2209. The class members whose credit reports the agency disseminated to potential creditors suffered a constitutional injury because their reputations were harmed by the inaccurate disclosures. Id. at 2209. The others lacked standing because the mere inclusion of misleading information in their files, without publication to any third party, was not a concrete harm of the type “traditionally recognized as providing a basis for a lawsuit in American courts.” Id. at 2204, 2209–10. Distinguishing between the groups, the Court emphasized the “important difference” between a statutory cause of action and actual injury and reiterated that the latter is always required for federal jurisdiction. Id. at 2205. TransUnion did not unequivocally overrule ETCL II or any of the cases it relies on. The requirement that plaintiffs have a concrete stake in federal litigation is not new. TransUnion merely reaffirmed the well- established rule that a violation of a federal law alone is not an Article III injury. See, e.g., Spokeo, Inc. v. Robins, 578 U.S. 330, 341 (2016) (“[A plaintiff] could not . . . allege a bare procedural violation, . . . and satisfy the injury-in- fact requirement of Article III.”); Summers v. Earth Island Inst., 555 U.S. 488, 496 (2009) (“[D]eprivation of a procedural right without some concrete interest . . . is insufficient to create Article III standing.”). That rule was canon in our court long before ETCL II. See, e.g., Lee v. Verizon Commc’ns, Inc., 837 F.3d 523, 529–30 (5th Cir. 2016) (concluding that a “bare allegation of incursion on [a] purported statutory right” is not a constitutional injury absent an “allegation of a real risk” to the plaintiff). Our prior opinion faithfully applied the fundamental rule that there is no standing absent concrete injury. We distinguished between suits under the False Claims Act, whereby uninjured citizens can sue to redress an injury to the government, and environmental citizen suits, which are subject to the ordinary injury rule. ETCL II, 968 F.3d at 364 (“[U]nlike qui tam relators . . . citizens suing under the bountyless environmental statutes must meet the standing requirement in their own right.” (citations omitted)). We then explained that the injuries suffered by the plaintiffs—interference with recreation, breathing and smelling polluted air, and allergy-like or respiratory problems—are concrete harms that have long been a basis for constitutional standing. Id. at 368 (“Each of those experiences was an Article III injury.”). Rather than overruling ETCL II, TransUnion supports it. To illustrate the concrete harm requirement, TransUnion offered the example of two people, a Maine citizen and a Hawaii citizen, each hoping to sue a Maine factory for violations of a federal environmental law. 141 S. Ct. at 2205. The Maine citizen, whose property was directly affected by the factory’s unlawful pollution, could sue to redress that injury. Id. at 2205–06. But the Hawaii citizen, whose interest in abating the nuisance was largely conceptual, lacked the personal stake in the litigation our Constitution requires. Id. The present plaintiffs are the Maine citizen: They live, work, and recreate near Baytown and personally experience the effects of Exxon’s unauthorized emissions. ETCL II, 968 F.3d at 368. They are not, as Exxon asserts, “merely seeking to ensure [Exxon's] ‘compliance with regulatory law.’” TransUnion, 141 S. Ct. at 2206 (quoting Spokeo, 578 U.S. at 345 (Thomas, J., concurring)). TransUnion did not upset our approach to injury-in-fact,[2] so our prior holding controls: The plaintiffs satisfied the first requirement of standing. 2 Exxon takes a second shot at ETCL II‘s standing framework, now aiming for our holding on traceability. Again, it misses. Our prior opinion provided the lower court detailed directions on how to determine which of the 16,386 alleged days of violations are traceable to Exxon’s actions. We explained that because injury must only be fairly traceable to the challenged conduct, plaintiffs “need not ‘connect the exact time of their injuries with the exact time of an alleged violation.’” ETCL II, 968 F.3d at 368 (quoting Texans United for a Safe Econ. Educ. Fund v. Crown Petrol. Corp., 207 F.3d 789, 793 (5th Cir. 2000)). Consequently, we held, Article III is satisfied if Exxon’s violations were “of a type that ’causes or contributes to the kinds of injuries alleged by the plaintiffs.’” Id. (quoting Cedar Point, 73 F.3d at 557). Translating these principles to the Clean Air Act context, we explained that the plaintiffs must show that each alleged violation (1) “causes or contributes to the kinds of injuries” they allege and (2) has a “‘specific geographic or other causative nexus’ such that the violation could have affected their members.” Id. at 369–70 (quoting Cedar Point, 73 F.3d at 557, 558 n.24 (internal quotation marks omitted from first quotation)). Exxon takes issue with this test. In the prior appeal, however, the company objected to the district court’s traceability findings but did not question our court’s approach to traceability. See id. at 368 n.4 (“Exxon does not question the vitality of Cedar Point or our other decisions applying this [traceability] standard “). Now it objects to ETCL II‘s reading of Cedar Point. But on this issue, Exxon does not even try to invoke an exception to the law-of-the-case doctrine. In any event, Exxon’s position is unconvincing. The company’s view is that Article III requires plaintiffs to show that each challenged emission is a but-for cause of their injuries. But none of the cases Exxon cites support its position. Although a but-for causal connection is sufficient to establish traceability, see Duke Power Co. v. Carolina Env’t Study Grp., Inc., 438 U.S. 59, 74–78 (1978), the Supreme Court has never said such proof is required. Consider Friends of the Earth, Inc. v. Laidlaw Env’t Servs. (TOC), Inc., 528 U.S. 167 (2000). There, citizen-suit plaintiffs had constitutional standing to challenge 489 Clean Water Act permit violations that occurred between 1987 and 1995. Id. at 176, 180–88. The Court did not conduct a separate standing inquiry for each violation, nor did it require the plaintiffs to connect their injuries to specific unlawful discharges. Instead, it credited testimony that the plaintiffs’ members no longer recreated near or waded in a river because of their concerns about pollutants. Id. at 182–83. Exxon’s position cannot be reconciled with Laidlaw. And Lujan v. Defenders of Wildlife, which Exxon says is its best case, did not address the issue at all. See generally 504 U.S. 555. Exxon’s position is also at odds with more than three decades of law from this court holding that traceability “requires less of a causal connection than tort law.” ETCL II, 968 F.3d at 368; Texans United, 207 F.3d at 793 (“No relevant case law supports [the] argument that [plaintiffs] must connect the exact time of their injuries with the exact time of an alleged violation “); Cedar Point, 73 F.3d at 557, 558 n.24 (adopting the two-part test described above); Save Our Cmty. v. EPA, 971 F.2d 1155, 1161 (5th Cir. 1992) (noting that traceability does not require plaintiffs to connect their harms to the defendant’s actions by a “scientific certainty”) (quotation omitted). Other circuits agree. As then-Judge Alito explained, “Article III standing demands ‘a causal relationship,’ but neither the Supreme Court nor our Court has ever held that but-for causation is always needed.” Khodara Env’t, Inc. v. Blakey, 376 F.3d 187, 195 (3d Cir. 2004); see also, e.g., Webb as next friend of K.S. v. Smith, 936 F.3d 808, 814 (8th Cir. 2019) (“[T]he fairly- traceable inquiry is much more forgiving that the merits-based, tort-causation inquiry.”); Friends of the Earth, Inc. v. Gaston Copper Recycling Corp., 204 F.3d 149, 161 (4th Cir. 2000) (en banc) (Wilkinson, C.J.) (“[T]he ‘fairly traceable’ standard is ‘not equivalent to a requirement of tort causation.’” (quoting Pub. Int. Rsch. Grp. of N.J. v. Powell Duffryn Terminals, Inc., 913 F.2d 64, 72 (3d Cir. 1990))). Even Exxon realizes that its position is an outlier. It acknowledged at oral argument, acknowledging that but-for causation is not always appropriate, even in the tort context, and ceded that it is enough that the defendant’s conduct provide “at least some contribution” to the plaintiff’s injury. The “causes or contributes to” test that we outlined in ETCL II requires just that. We are bound by our prior articulation of the test for traceability, and we stand by it. B Having confirmed that the standing framework from ETCL II remains good law, we now consider whether the district court applied it correctly on remand. Exxon does not challenge any of the findings underlying the district court’s justiciability determination. The company’s challenge is more conceptual: It says that our prior opinion impermissibly restricted the district court’s factfinding ability. It did not. Initially, the district court found standing for all 16,386 violation days. 2017 WL 2331679, at *11. We affirmed its conclusions across- the-board for injury and redressability. For traceability, however, we reached a mixed result. As we explained: For any violation that could cause or contribute to flaring, smoke, or haze, the district court’s findings have established traceability. The district court need only decide which violations fall within this category. For violations that could not contribute to flaring, smoke, or haze, the district court should first consider whether the pollutant emitted could cause or contribute either to (a) chemical odors or (b) allergy-like or respiratory symptoms. If so, the district court will conduct the geographic nexus inquiry described [previously], finding it satisfied if the emission (i) violated a nonzero emissions standard, (ii) had to be reported under Texas regulations, or (iii) is otherwise proven to be of sufficient magnitude to reach Baytown neighborhoods outside the Exxon complex in quantities sufficient to cause chemical odors, allergy-like symptoms, or respiratory symptoms. ETCL II, 968 F.3d at 371. We thus concluded that the then-existing “findings have established traceability” for some categories of violations. Id. But for other categories, we could not sustain the factfinding on the current record and provided a framework for the district court to apply on remand. That is not appellate court factfinding, it is ordinary appellate review of factfinding that reached different conclusions for different types of violations. Plus, the outcome on remand belies Exxon’s contention that we left the district court with a “vanishingly small” window for factfinding. The district court reduced the number of justiciable violations by 12,735, which is over 75% percent. See Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 565. As a fallback, Exxon says that the district court erred by describing the criteria it used to identify the traceable violations rather than listing each justiciable violation individually. Following a bench trial, the district court must “find the facts specially.” FED. R. CIV. P. 52(a)(1). This show-your- work requirement ensures that we have a meaningful record to review. Garner v. Kennedy, 713 F.3d 237, 242–43 (5th Cir. 2013). But Rule 52 does not require “punctilious detail” or “slavish tracing of the claims issue by issue.” Ratliff v. Governor’s Highway Safety Program, 791 F.2d 394, 400 (1986). The district court need only include “sufficient detail to enable the appellate court to consider the findings under the applicable reviewing standard.” Burma Navigation Corp. v. Reliant Seahorse MV, 99 F.3d 652, 657 (5th Cir. 1996). The district court’s 10-page traceability analysis was thorough and sufficiently explained. It conducted separate analyses for violations based on category (flaring, smoke, haze, chemical odor, and asthma-like symptoms) and explained how it evaluated traceability for each. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 555–65. Take flaring, for example. The plaintiffs testified that they could see flares from their homes and other part of Baytown, so we found the geographic nexus requirement met for violations “that could cause or contribute to flaring” and instructed the district court to identify those instances. ETCL II, 968 F.3d at 371. The district court reviewed the evidence and found 1,801 traceable flaring violations. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 557-58. It explained that plaintiffs did not prove that every emission from a flare stack causes visible flaring. Id. at 557. But it credited trial testimony that flaring occurs when a compressor trips or shuts down. Id. It thus found traceability for the subset of flaring violations that occurred at a flare stack and were caused by the failure of a compressor. Id. One need only compare this description to the stipulated spreadsheet of violations to determine which of the flaring events qualify. The district court’s analysis of the other categories is similarly detailed and rigorous. See id. at 558-65. It did not have to list all sixteen thousand alleged violations and state whether each is justiciable or not. In fact, we specifically said that it did not have to make “line-by-line findings for the thousands of violations” and could instead “group violations by type and magnitude.” ETCL II, 968 F.3d at 371. The lower court faithfully followed this instruction. Rule 52 is satisfied. Our standing framework is legally sound. So are the district court’s findings. The environmental organizations could sue for 3,651 days of violations. III Several thousand violations are justiciable. And Exxon accepts substantive liability for those claims. See ETCL II, 968 F.3d at 364 (explaining that Exxon appealed only standing, affirmative defenses, and penalty). The remaining question, then, is whether the district court abused its discretion in ordering a $14.25 million dollar penalty. Courts may, but are not required to, assess civil penalties for Clean Air Act violations. 42 U.S.C. § 7413(e)(2); see also ETCL I, 824 F.3d at 524. In making that determination, they have a statutory obligation to consider: the size of the business, the economic impact of the penalty on the business, the violator’s full compliance history and good faith efforts to comply, the duration of the violation as established by any credible evidence (including evidence other than the applicable test method), payment by the violator of penalties previously assessed for the same violation, the economic benefit of noncompliance, and the seriousness of the violation. 42 U.S.C. § 7413(e)(1). They must also account for “such other factors as justice may require.” Id. The district court found that almost all of the statutory factors favor a penalty. On the first two, it noted that because of Exxon’s large size and profitability, only a large penalty would be meaningful. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 568. It also found that the duration and seriousness of the many violations, as well as the significant economic benefit Exxon received from not meeting its Clean Air Act obligations, counseled against the company. Id. at 570–76. One factor, however, weighed against a penalty. The district court credited Exxon’s good faith and “substantial efforts to improve environmental performance and compliance.” Id. at 569–71. Balancing those considerations and deducting the $1.42 million dollars Exxon already paid regulators for the violations, the court concluded that a $14.25 million penalty is appropriate. Id. at 571, 576-77. Exxon challenges the district court’s findings on the economic benefit, duration, and seriousness factors. We are highly deferential to the district court’s penalty assessment. We review its application of the penalty factors for abuse of discretion and its underlying findings for clear error. Cedar Point, 73 F.3d at 573; see also Tull v. United States, 481 U.S. 412, 427 (1987) (noting the “highly discretionary” nature of weighing the similar Clean Water Act penalty factors). The district court did not err on either front, so we affirm the $14.25 million dollar penalty. A We begin with the economic benefit factor. This consideration seeks to prevent polluters from gaining a competitive advantage through noncompliance with environmental laws. Benefit can be calculated in two ways: by determining the cost of capital (what it would cost the polluter to fund pollution prevention), or by determining the return on capital (what the polluter earned on the funds it should have spent on pollution control but instead invested elsewhere). ETCL I, 824 F.3d at 527 (citing United States ex rel. EPA v. CITGO Petrol. Corp., 723 F.3d 547, 552 (5th Cir. 2013)). Improvements that are “necessary to correct” the violations alleged in the suit are benefits of noncompliance. ETCL I, 824 F.3d at 530 (quoting CITGO, 723 F.3d at 552); see also United States v. Allegheny Ludlam Corp., 366 F.3d 164, 178 (3d Cir. 2004) (stating that the economic benefit calculation “starts with the costs spent or that should have been spent to achieve compliance [with the Clean Water Act]“). Plaintiffs need not tie the projects to prevention of specific violations. ETCL I, 824 F.3d at 530 n.19. Rather, the inquiry centers on “whether the projects will ameliorate the kinds of general problems that have resulted in at least some of the permit violations upon which Plaintiffs have sued.”[3] Id. The district court valued Exxon’s benefit of noncompliance at more than fourteen million dollars ($11.75 million dollars at the time of the expert’s report plus $61,000 per month after that) because the company delayed implementation of four emission-reducing projects mandated by a 2012 agreement between Exxon and state regulators. Env’t Tex. Citizen Lobby, Inc., 524 F. Supp. 3d at 573. Its conclusion that the projects are necessary to correct the violations was not clear error. The projects represent “an effort to reduce emissions and unauthorized emissions events” at the Baytown complex. ETCL I, 824 F.3d at 530 (quoting Env’t Tex. Citizen Lobby, Inc., 66 F. Supp. 3d at 908). Such unauthorized emissions are the heart of the violations alleged in this suit. The district court carefully compared the goal of each of the four projects to the remaining violations and concluded that each project “addressed the types of violations found traceable by this Court.” Env’t Tex. Citizen Lobby, Inc.,524 F. Supp. 3d at 573–74. Exxon does not contest those determinations. The company would have us reduce the value of its noncompliance because the number of justiciable violation days is a fraction of the number of violations previoulsy found. But that approach would give Exxon an unwarranted discount. Exxon needed to invest $11.75 million dollars in improvements to comply with its Clean Air Act obligations. That figure does not depend on how many times Exxon violated its permits. Say, for example, a plumber quotes you $250 to fix a leaky faucet. The repair cost would be the same regardless whether you called the plumber after the first few drops or waited for multiple buckets to fill. And the district court properly accounted for the reduced number of violations in its final balancing of the statutory factors, reducing the penalty multiplier from 50% of the value of noncompliance to 10%. Id. at 577. The district court’s conclusion on economic benefit stands. B Courts must also factor “the duration of the violation as established by any credible evidence” into the penalty assessment. 42 U.S.C. § 7413(e)(1). We previously reserved judgment on whether this penalty factor “requires scrutiny of the length of each individual violation or allows for assessment of an overall violation period.” ETCL I, 824 F.3d at 531. The district court took the latter approach. Finding that Exxon averaged more than one violation per day across the eight years covered by the suit, it concluded that the duration of the violation weighed in favor of a penalty. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 571. We see no error in the district court’s analysis. Courts across the country agree that “when multiple ‘intermittent’ violations over a span of time are at issue, a court may consider the overall length of the period during which the violations occurred.” ETCL I, 824 F.3d at 531; see also, e.g., Pound v. Airosol Co., Inc., 498 F.3d 1089, 1098 (10th Cir. 2007) (remanding for the district court to consider that the defendant’s “violations lasted more than a decade”); United States v. Vista Paint Corp., 1996 WL 477053, at *15 (C.D. Cal. Apr. 16, 1996); United States v. B & W Inv. Props., Inc., 1994 WL 53781, at *4 (N.D. Ill. Feb. 18, 1994); United States v. Midwest Suspension & Brake, 824 F. Supp. 713, 736–37 (E.D. Mich. 1993); United States v. A.A. Mactal Const. Co., Inc., 1992 WL 245690, at *3 (D. Kan. Apr. 10, 1992) (all considering the length of the violation period). No court has rejected the overall-violation-period approach. See Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 570 n.5 (“Exxon contends the Court should continue to look to duration of the violations standing alone in analyzing this factor. However, Exxon cites no case law to support this proposition.”). And it makes sense in the context of this suit. Exxon’s unlawful omissions occurred regularly for many years. Considering the length of only select few of those thousands of violations would not fully reflect the extent of Exxon’s unlawfulness. We will not disturb the district court’s conclusion that the duration factor weighs for a penalty. C Exxon next challenges the district court’s determination that its violations were serious. See 42 U.S.C. § 7413(e)(1) (commanding consideration of “the seriousness of the violation”). We previously observed that other courts assess seriousness by looking at the risk the emissions potentially pose to the environment as well as “the overall number and quantitative severity of emissions or discharges.” ETCL I, 824 F.3d at 532 (citing Pound, 498 F.3d at 1099; Powell Duffryn, 913 F.2d at 79). The district considered both. It first determined that Exxon’s violations posed a low risk to the environment because the plaintiffs did not show that any individual violation was concentrated enough to harm people or the planet. Env’t Tex. Citizen Lobby, 524 F. Supp. 3d at 575 n.121. But it concluded that the quantity of emissions weighed the other way because the traceable violations emitted over 1.5 million pounds of pollutants. Id. at 575–76. Exxon does not challenge this finding, and we see no reason to find it clearly erroneous. Exxon argues, as it did for the duration factor, that the court must consider the seriousness of each violation individually. But the district court did consider each violation; it found that the traceable violations involved relatively high levels of emissions and necessarily considered the amount of each violation when it added them up to reach the 1.5-million-pound figure. Exxon does not offer any alternative definitions of “seriousness” that the district court could have applied instead. There was no abuse of discretion on the seriousness factor. D Finally, Exxon faults the district court for not assessing whether its ruling might deter the industry from negotiating with regulators in the future, a consideration the company believes is required in the interest of justice. See 42 U.S.C. § 7413(a) (allowing consideration of “other factors as justice may require”). Any error on this factor lies with Exxon, not the district court. Exxon did not offer this argument in its initial proposed findings to the district court,[4] so it forfeited the ability to raise it on appeal. See Rollins v. Home Depot USA, 8 F.4th 393, 397 (5th Cir. 2021). *** This time, the district court got it right. It properly considered each of the penalty factors and found that the many factors favoring a penalty outweigh the one that does not. It also subtracted more than five million dollars from Exxon’s bill in recognition of the reduced number of justiciable violations. The district court’s penalty determination was well within its wide discretion. *** AFFIRMED. ANDREW S. OLDHAM, Circuit Judge, dissenting: This case is a jurisdictional mess. In fairness to my esteemed colleagues in the majority, some of the mess predates our panel’s first decision in this case. See Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp. (“ETCL I“), 824 F.3d 507 (5th Cir. 2016). And in fairness to the esteemed district court judge, most of the mess stems from our decisions, not the district court’s. See ibid.; Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp. (“ETCL II“), 968 F.3d 357, 375 (5th Cir. 2020) (Oldham, J., concurring in part, dissenting in part, and concurring in the judgment). What’s clear is that only our en banc court can clean this up. I. Some basics first. To invoke our jurisdiction, plaintiffs must satisfy the familiar tripartite test for Article III standing by showing: (1) an injury in fact; (2) that’s fairly traceable to the defendant’s conduct; and (3) that’s likely redressable by a favorable decision. See Lujan v. Defs. of Wildlife, 504 U.S. 555, 560–61 (1992). Plaintiffs must show standing “with the manner and degree of evidence required at the successive stages of the litigation.” Id. at 561. So “in a case like this that proceeds to trial, the specific facts set forth by the plaintiff to support standing must be supported adequately by the evidence adduced at trial.” TransUnion LLC v. Ramirez, 141 S. Ct. 2190, 2208 (2021) (quotation omitted). Plaintiffs thus must “prove standing by a preponderance of the evidence.” ETCL II, 968 F.3d at 367; see also E.T. v. Paxton, 41 F.4th 709, 714 (5th Cir. 2022). The traceability requirement is particularly important here. The Supreme Court has explained that a plaintiff can establish traceability without establishing the tort requirement of proximate causation. See Lexmark Int’l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 134 n.6 (2014) (“Proximate causation is not a requirement of Article III standing, which requires only that the plaintiff’s injury be fairly traceable to the defendant’s conduct.”). The Court has also said that a plaintiff need not show but-for causation. See Khodara Env’t, Inc. v. Blakey, 376 F.3d 187, 195 (3d Cir. 2004) (Alito, J.) (“Article III standing demands ‘a causal relationship,’ but neither the Supreme Court nor our Court has ever held that but-for causation is always needed.”). It’s nonetheless true that a plaintiff must establish, at a minimum, causation in fact. See, e.g., Dep’t of Com. v. New York, 139 S. Ct. 2551, 2566 (2019) (“Because Article III ‘requires no more than de facto causality,’ traceability is satisfied here.” (quoting Block v. Meese, 793 F.2d 1303, 1309 (D.C. Cir. 1986) (Scalia, J.))). That is, the plaintiff must show that the defendant’s action X did in fact cause the plaintiff’s injury Y.[5] Our sister circuits have dutifully followed those instructions.[6] The Supreme Court also recently reaffirmed that “standing is not dispensed in gross; rather, plaintiffs must demonstrate standing for each claim that they press and for each form of relief that they seek (for example, injunctive relief and damages).” TransUnion, 141 S. Ct. at 2208; see also, e.g., Town of Chester v. Laroe Ests., Inc., 137 S. Ct. 1645, 1650 (2017) (“Our standing decisions make clear that standing is not dispensed in gross.” (quotation omitted)). Trying to align with that principle, the ETCL II panel explained that in the context of the Clean Air Act, we must do “a separate standing inquiry for each violation asserted as part of that claim.” 968 F.3d at 365. That is, plaintiffs must show—for each violation, not just each claim— an injury in fact that is fairly traceable to the violation and that is likely to be redressed by a favorable judicial decision. And here, plaintiffs must do so with trial-appropriate evidence and by the preponderance of the evidence. II. The majority opinion conflicts with these basic principles. I first (A) explain how the majority dispenses standing in gross. I then (B) explain how the majority overlooks plaintiffs’ failures to establish traceability. A. In recognition of our inability to dispense standing in gross, the ETCL II panel concluded that a plaintiff must establish standing “for each violation asserted as part of that claim.” 968 F.3d at 365. That conclusion was right. But what the panel concluded next was wrong. It proceeded to define the injury generically and create per se rules for when the district court must irrebutably presume that the generic injury is traceable to a specific violation. See id. at 371. In so doing, the panel allowed the very thing it sought to forbid—standing in gross. To show an injury in fact, plaintiffs must state “specific facts” and back those facts up with adequate evidence. Lujan, 504 U.S. at 561. The injury-in-fact requirement thus includes two different parts: The first governs specification (or identification) of the injury, and the second governs the level (or type) of proof required for that specified injury. The ETCL II majority misunderstood both. First, specification. The plaintiff must specify an injury that’s fairly traceable and redressable. Here, given that plaintiffs must show standing for each violation, traceability requires plaintiffs to specify their injuries with some granularity. For example, they must trace their particular injuries to particular violations on particular days. Otherwise, a court could not determine whether a particular violation in fact caused a particular injury, or instead whether plaintiffs seek to use one injury for standing in gross to challenge violations that never injured them. Consider, for example, plaintiffs’ alleged aesthetic injury: They say they saw flares. The aesthetic injury of seeing a flare cannot be fairly traced to the defendants without specifying: I saw a flare on Day A, which was during Violation X; or I was in town on Day B, which was during Violation Y. If plaintiffs do not show their injury in fact at that level of specificity, then we cannot be sure that the traceability requirement is met—that is, plaintiffs have not met their burden of showing that the violation “likely caused” their injury. TransUnion, 141 S. Ct. at 2203. If plaintiffs instead specify their injury as “aesthetic harms more generally,” then their injury could’ve occurred on Day C, which was not during any violation, or on Day D, when plaintiffs were out of town. Friends of the Earth, Inc. v. Laidlaw Environmental Services (TOC), Inc., 528 U.S. 167 (2000), is not to the contrary. See ante, at 9. First, traceability wasn’t at issue in Laidlaw, so the Court had no occasion to address how an injury must be specified. See ETCL II, 968 F.3d at 377–78 (opinion of Oldham, J.). That effectively makes Laidlaw a drive-by traceability ruling, which doesn’t get much weight. Cf. Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 91 (1998) (“drive-by jurisdictional rulings of this sort . . . have no precedential effect”). Second, the Supreme Court subsequently narrowed Laidlaw. In Clapper v. Amnesty International USA, 568 U.S. 398 (2013), for example, the Court construed Laidlaw to support standing only where the plaintiffs took “preventive measures” to avoid “concededly ongoing” violations of the Clean Water Act. Id. at 419 (quotation omitted). Here, by contrast, there’s no adequate contention, backed by trial-appropriate evidence, that any of plaintiffs’ members took reasonable preventive measures to avoid Exxon’s violations, let alone measures that they would non-speculatively stop if the violations did. Nor are there concededly ongoing violations for each type of violation. At no point did plaintiffs clearly spell out, in this appeal or before it, that each violation repeatedly reoccurred after the complaint was filed, even though it has always been plaintiffs’ burden to do so. See E.T., 41 F.4th at 718 n.2. Second, plaintiffs’ burden. Traceability requires not just specification but also non-speculative proof of causation. The majority contends it would be impossible to prove causation on a per-violation basis. See ante, at 8–9; ETCL II, 968 F.3d at 368 (“Requiring proof that specific is not consistent with the traceability requirement, which requires less of a causal connection than tort law (and even tort causation would not require such specific proof ).”). That may or may not be true, but it’s irrelevant. The Supreme Court has been clear that standing is no “mere pleading requirement[] but rather an indispensable part of [plaintiffs'] case” and that plaintiffs’ burden is the preponderance of the evidence. Lujan, 504 U.S. at 561; see also ETCL II, 968 F.3d at 367 (recognizing as much). And as we’ve held, “analyzing standing at this level of granularity can be tedious in a sweeping challenge like this one. But it’s what Article III requires.” In re Gee, 941 F.3d 153, 165 (5th Cir. 2019) (per curiam). TransUnion illustrates the point. In that case, the plaintiffs argued that they met their burden of showing their credit reports were sent to third parties (the injury) by pointing to a stipulation governing other similarly situated individuals. 141 S. Ct. at 2212. The Supreme Court rejected that argument. Instead, the Court held “[t]he plaintiffs had the burden to prove at trial that their reports were actually sent to third-party businesses. The inferences on which the argument rests are too weak to demonstrate that the reports of any particular number of the 6,332 class members were sent to third-party businesses.” Ibid. The inferences were especially weak given that the plaintiffs “presumably could have” put forth stronger evidence. See ibid. (citing Interstate Cir., Inc. v. United States, 306 U.S. 208, 226 (1939) (“The production of weak evidence when strong is available can lead only to the conclusion that the strong would have been adverse.”)). The takeaway is that the Court required the plaintiffs to prove standing—regardless of how difficult that might be to do. In my view, we should’ve done the same thing in this case. B. If we’d applied the rules properly, plaintiffs would have standing to challenge violations on approximately 40 days, not 3,651. The principal source of our legal error lies in ETCL II, which created rigid per se rules that allowed plaintiffs to bypass the strictures of Article III. This hypothetical explains why: To illustrate, consider a hypothetical plaintiff Bob who lives in Baytown. Bob has asthma—that is, an injury. The question is whether his asthma injury is traceable to Exxon’s illegal emissions. From January 1 through January 10, Bob was visiting his sister in France. Meanwhile: On January 2, Exxon emitted pollutants that could have reached beyond the Exxon complex into the offsite areas of Baytown where Plaintiffs’ members lived and recreated. On January 5, Exxon released pollutants in excess of nonzero emissions limits or that constituted a reportable quantity under state regulations. On January 8, Exxon emitted pollutants that could have caused or contributed to flaring, smoke, or haze, even if the emission was of a small magnitude. ETCL II, 968 F.3d at 378 (opinion of Oldham, J.) (quotation omitted). That hypothetical should have yielded an obvious result: no standing. Bob obviously wasn’t harmed by any of the violations when he was breathing French air. So one would think that’d squelch any standing to recover for such violations. But alas, today’s majority says the proverbial Bob was injured, has standing, and can recover. The relevant time period here is “October 2005 through September 2013.” Id. at 363. And the time after September 2012 is critically similar to my hypothetical. Plaintiffs put on testimony from four of their members. The only members (Marilyn Kingman and Diane Aguirre Dominguez) who testified to suffering injuries after September 2012 did not live in Baytown, and one of those members stopped visiting Baytown regularly after March 2013 (Diane Aguirre Dominguez). See id. at 367; see also Env’t Tex. Citizen Lobby, Inc. v. ExxonMobil Corp., No. CV H-10-4969, 2017 WL 2331679, at *7–8 (S.D. Tex. Apr. 26, 2017). And Kingman “was not able to correlate any of her experiences or concerns to specific [violations].” Id. at *7. She also testified that from her home in Mont Belvieu, she could not smell odors related to the chemical releases. Nor does she have any medical problems or conditions related to the alleged violations. She at best “heads to Baytown a few times a week to run errands, recreate, and go to church.” ETCL II, 968 F.3d at 367. Plaintiffs provided no evidence showing which days Kingman went into Baytown, even though such evidence surely existed. So how can we say that Kingman more likely than not suffered from a particular violation? See TransUnion, 141 S. Ct. at 2212 (recognizing that “[t]he production of weak evidence when strong is available can lead only to the conclusion that the strong would have been adverse” (quotation omitted)). Yet under the majority’s view, every violation after September 2012 (and even March 2013) is fairly traceable to some injury. That’s so even if there is a violation every day from September 2012 to September 2013 and, concededly, no member of plaintiffs was in Baytown every day. That’s hardly a reasonable inference from the evidence. See Ctr. for Biological Diversity v. EPA, 937 F.3d 533, 545 (5th Cir. 2019) (“At some point this common sense observation becomes little more than surmise. At that point certainly the requirements of Article III are not met.” (quotation omitted)); In re Gee, 941 F.3d at 164 (“Article III requires more than theoretical possibilities.”). If anything, it’s equivalent to the very “conjecture” the majority previously conceded was insufficient to establish standing. See ETCL II, 968 F.3d at 368 (“Traceability instead requires something more than conjecture (‘The Exxon complex in Baytown emits pollutants, and I live in Baytown’).”). In sum, the ETCL II majority sanctioned defining the injury in fact at too high level of generality, making it impossible to properly assess traceability. The majority then created per se rules and irrebuttable presumptions for traceable injuries that, in fact, did not exist. The majority, for example, allowed a plaintiff who only suffered injuries three days of the week to get relief for all seven. See Spokeo, Inc. v. Robins, 578 U.S. 330, 340 (2016) (“A ‘concrete’ injury must be ‘de facto‘; that is, it must actually exist.”). In so doing, the majority effectively “eliminate[d] traceability altogether.” ETCL II, 968 F.3d at 375 (opinion of Oldham, J.). And it dispensed standing in gross, in violation of TransUnion.[7] * * * The implications of the majority’s approach are alarming. “By permitting citizens to pursue civil penalties payable to the Federal Treasury, the [Clean Air Act] . . . turns over to private citizens the function of enforcing the law.” Laidlaw, 528 U.S. at 209 (Scalia, J., dissenting). And by easing or eliminating the Article III minima for standing, today’s majority all but erases the distinction between private citizens and the government agencies that otherwise enforce the Clean Air Act. I respectfully dissent.
