Before JOLLY, COSTA, and ENGELHARDT, Circuit Judges. GREGG COSTA, Circuit Judge: In 1999, philanthropist, environmental activist, and haircare and liquor tycoon John Paul DeJoria was attempting to achieve yet another title: oil magnate. It did not go well. What started as a project that promised to provide Morocco with decades of energy independence ended with a Moroccan court’s levying a judgment north of $100 million against DeJoria and his business partner. Whether Texas should recognize that foreign judgment is now the centerpiece of this decades-long dispute. In fact, proving that it is often harder to collect a judgment than win one, this is the second time the question of the judgment’s validity has come before us. This time around we decide whether an interim change in the Texas recognition law violates the state’s constitutional ban on retroactive laws. If not, we must determine whether the district court properly followed this court’s 2015 mandate and whether it properly applied the new law. I. The facts of this case are littered across the pages of the Federal Reporter. See DeJoria v. Maghreb Petroleum Expl., S.A., 804 F.3d 373 (5th Cir. 2015); Skidmore Energy, Inc. v. Maghreb Petroleum Expl., S.A., 337 F. App’x 706 (9th Cir. 2009); Skidmore Energy, Inc. v. KPMG, 455 F.3d 564 (5th Cir. 2006). Because this court has already described the background of this corner of the dispute, we will do our best not to spill unnecessary ink. See DeJoria, 804 F.3d at 377-78. The winding path the case followed after our court’s 2015 remand will spill enough as it is. For now, suffice it to say that in 1999 DeJoria and his business partners started Lone Star Energy Corporation in Morocco with the help of King Mohammed Vl’s first cousin. The enterprise hoped to discover oil reserves in Northeastern Morocco. The prospects looked good—so good that the King took to Moroccan airwaves to announce that the country would soon be in possession of “copious and high-quality” oil that would allow Morocco to be self-sufficient for 30 years. The King’s announcement made the Moroccan stock market jump more than five percent in anticipation of the expected riches. But when the promised reserves did not materialize, the project quickly soured. DeJoria and his business partner were forced off Lone Star’s board, and, fearing for their lives because of an alleged death threat, fled Morocco, never to return. Not long after their ouster, DeJoria and his associates were sued in Moroccan commercial court by Lone Star’s new management (now called Maghreb Petroleum Exploration, S.A.) and its major investor, Mideast Fund for Morocco. Maghreb, the term we will use to collectively refer to those two entities, alleged that DeJoria and his partners mismanaged Lone Star and fraudulently induced investment in the doomed oil project. Seven years later, the Moroccan court returned a large judgment for Maghreb. It dismissed claims against five of the seven defendants, placing the blame—and the bill for 969,832,062.22 Moroccan dirhams[1]—squarely on DeJoria and his partner. Before going further, a little bit about the legal backdrop is helpful. In order to collect its winnings from DeJoria’s assets in the United States, Maghreb must convince an American court to recognize and enforce the Moroccan judgement.[2] Recognition of foreign-country judgments is a matter of state law and was once mostly governed by principles of comity. See Hilton v. Guyot, 159 U.S. 113, 163-64, 180-81 (1895). In some jurisdictions, comity is still the rule. See, e.g., Kwongyuen Hangkee Co., Ltd. v. Starr Fireworks, Inc., 634 N.W.2d 95, 96 (S.D. 2001). But most states have codified their recognition standards and procedures by enacting the 1962 Uniform Foreign Money Judgments Recognition Act[3] or its 2005 successor, the Uniform Foreign- Country Money Judgments Recognition Act.[4] Both acts make foreign judgments that are final and conclusive where rendered “enforceable” in the relevant state court just like another state’s judgment would be. Unif. Foreign- Country Money Judgments Recognition Act § 7(2), U.L.A. (2005) (West); Unif. Foreign Money Judgment Act § 3, U.L.A. (1962) (West). Although these acts presumptively treat properly filed foreign judgments as enforceable, exceptional circumstances can rebut that presumption. Some of those exceptions are mandatory, others discretionary. If the rendering court did not have personal jurisdiction over the judgment debtor, for instance, the state court (or federal court sitting in diversity) cannot recognize the foreign judgment. 2005 Unif. Act § 4(b)(2); 1962 Unif. Act § 4(a)(2). Other grounds for nonrecognition, like fraud in obtaining the judgment, instead give the American court the option of not recognizing the foreign judgment. 2005 Unif. Act § 4(c)(2); 1962 Unif. Act § 4(b)(2). So, in 2013, Maghreb came to the United States seeking recognition of the Moroccan judgment.[5] DeJoria resisted in several ways. At the time, Texas had adopted (with slight modification) the 1962 Uniform Recognition Act. See TEX. CIV. Prac. & REM. Code § 36.001-08 (Vernon’s 2015). That law included ten nonrecognition grounds. DeJoria pressed seven of them. The district court focused on only one avenue to nonrecognition. It determined that the Moroccan judgment was “rendered under a system that does not provide impartial tribunals or procedures compatible with the requirements of due process of law.” Id. § 36.005(a)(1). Because this was a mandatory nonrecognition ground, the district court refused to recognize the Moroccan judgment and dismissed the case. We reversed. DeJoria, 804 F.3d at 389. The panel held that, under Texas’s version of the 1962 Uniform Recognition Act, DeJoria could not obtain nonrecognition by showing he was denied due process or impartial tribunals in his case, but instead had the much greater burden of showing that Morocco’s legal “system as a whole” was so deficient that no Texas court should ever recognize a Moroccan judgment. Id. at 381.[6] And although the prior panel’s inquiry focused on whether the Moroccan judicial system could provide Americans fair proceedings, it remarked that “the record does not establish that the King actually exerted any improper influence on the Moroccan court in this case.” Id. at 382 n.9. The case was remanded. Back before the district court, and in front of the magistrate judge to whom the matter was referred, the parties immediately began to squabble over the scope of that court’s power on remand.[7] DeJoria was adamant that he should still be allowed to push for nonrecognition on grounds not addressed by the Fifth Circuit. Maghreb disagreed and moved for entry of judgment. The district court denied Maghreb’s motion, agreeing with DeJoria that he could still attempt to establish other grounds for nonrecognition. While the sound and fury continued apace in the trial court, a second front in this dispute opened, this time in the Texas legislature. With the testimonial aid of one of DeJoria’s lawyers, the 2017 legislative session was considering updating the Recognition Act to the 2005 uniform act. Among other changes, the new law would add two discretionary grounds for nonrecognition: a court would be able to deny recognition if “the judgment was rendered in circumstances that raise substantial doubt about the integrity of the rendering court with respect to the judgment” or, more importantly in this case, if “the specific proceeding in the foreign court leading to the judgment was not compatible with the requirements of due process of law.” 2005 Unif. Act § 4(c)(7)-(8). These substantive differences between the old and new law were not the focus of hearings on the bill. Instead, a change not found in the new Uniform Law nor in the versions of that law passed by other states drew the most attention. The drafters had made the law retroactive to pending cases. The only pending case the legislators were told about was this one. Despite the concern of at least one legislator that the law was going to change the outcome of this case midstream, the law was adopted with the retroactivity provision. 2017 Tex. Sess. Law Serv. Ch. 390 (S.B. 944) (Vernon’s), codified at Tex. CIV. Prac. & Rem. Code § 36A.001-11. With his legislative victory in hand, DeJoria returned to the district court to inform it of the change in Texas law. Although he argued that nonrecognition was warranted on multiple grounds, the district court again focused on only one. Finding the new law did not run afoul of the Texas Constitution’s prohibition of retroactive laws, this time the court granted DeJoria’s motion for nonrecognition after determining that the specific proceedings leading to the judgment against him were incompatible with the requirements of due process.[8] To reach that decision, the district court readopted many of the case-specific findings underlying the order this court had reversed. But it also made new findings: that DeJoria was unable to attend the Moroccan proceedings, that he was unable to obtain counsel to represent him in those proceedings, and that, although the Moroccan court relied on an expert’s opinion to determine damages, that expert lacked independence. The court again dismissed the case. Maghreb again appealed. II. We have jurisdiction over this case owing to the diversity of the parties, so we apply Texas substantive law. Erie R. Co. v. Tompkins, 304 U.S. 64, 78 (1938). In doing so, we are bound by the decisions of the Supreme Court of Texas. Comm’r v. Bosch’s Estate, 387 U.S. 456, 465 (1967). But when no decision of that court directly addresses the case before us, we are forced to make an Erie guess, doing our best to write the opinion the Texas high court would if it had the chance.[9] Meador v. Apple, Inc., 911 F.3d 260, 264 (5th Cir. 2018). We must make such a guess to determine which of the Uniform Recognition Acts applies. The Texas Constitution provides that “[n]o bill of attainder, ex post facto law, retroactive law, or any law impairing the obligation of contracts, shall be made.” Tex. Const. art. I, § 16. Although phrased as an absolute prohibition, “[m]ere retroactivity is not sufficient to invalidate a statute.” Robinson v. Crown Cork & Seal Co., Inc., 335 S.W.3d 126, 139 (Tex. 2010) (quotation omitted). Texas courts have tailored the scope of the prohibition to “protect[] settled expectations and prevent[] abuse of legislative power.” Id. Three factors determinine whether a law runs afoul of those objectives: “the nature and strength of the public interest served by the statute as evidenced by the Legislature’s factual findings; the nature of the prior right impaired by the statute; and the extent of the impairment.” Id. at 145. The nature and extent of the interference with a party’s rights loom particularly large. For that reason, “changes in the law that merely affect remedies or procedure, or that otherwise have little impact on prior rights, are usually not unconstitutionally retroactive.” Id. at 146; see also Univ. of Tex. Sw. Med. Ctr. at Dallas v. Estate of Arancibia, 324 S.W.3d 544, 548 (Tex. 2010) (“Statutes . . . that do not deprive the parties of a substantive right . . . may be applied to cases pending at the time of enactment.”). The new law’s limited interference with Maghreb’s legitimate rights resolves the question before us. Unlike Robinson—the seminal Texas case on retroactivity—this is not a case in which a law that allowed a party’s recovery was changed to “abrogate their claim.” Robinson, 335 S.W.3d at 148. It is not even certain that the law as it stood before the adoption of the updated act would have led to recognition of the Moroccan judgment. As we mentioned, the district court agreed to allow DeJoria to press several arguments for nonrecognition after this court returned the case to its hands.[10] Because the passage of the new act made it unnecessary to address those claims, we do not know how likely they were to succeed. Maghreb’s expectation that it would prevail was, in other words, not yet settled. See Union Carbide Corp. v. Synatzske, 438 S.W.3d 39, 58 (Tex. 2014) (upholding retroactive application of a law because the plaintiff’s “recovery was not yet predictable” at the time the law went into effect). The bigger point, though, is that the retroactive law does not abrogate Maghreb’s claim. It does not strip Maghreb of the ability to seek recognition of the Moroccan judgment. It just gives a district court the ability to deny recognition if it finds the judgment was obtained in proceedings that were incompatible with the requirements of due process. So the only right that has been impinged is the right to automatic recognition of a judgment obtained in proceedings that denied the judgment debtor fundamental fairness. To state that “right” is to show why we cannot recognize it, let alone allow its protection to sink a state statute. Robinson, 335 S.W.3d at 146 (“[C]ourts must be mindful that statutes are not to be set aside lightly.”). Indeed, the absurdity of lending a court’s power to the vindication of fundamentally unfair proceedings is why the 2005 Uniform Act recognizes an absence of due process as one of the rare situations when an American court may not recognize a foreign judgment. It is also noteworthy that the Supreme Court of Texas has only upheld challenges to the retroactive application of a law on four occasions, all of which dealt with laws that revived expired claims or fully extinguished vested rights. Tenet Hospitals Ltd. v. Rivera, 445 S.W.3d 698, 708 (Tex. 2014) (collecting cases). The updated recognition act does neither. We are mindful that the whiff of home cooking also pervades the Texas side of this case. There is a deep irony in allowing DeJoria to contend he was denied due process in Morocco when it was his lobbying efforts that changed the rules of the game midway through the proceedings in the United States. Indeed, the Supreme Court of Texas has been suspicious of retroactive laws that inure to the benefit of only one company or individual.[11] Robinson, 335 S.W.3d at 149. But in the retroactivity context as in others, “unfair does not always equal unconstitutional.” Id. at 160 (Willett, J. concurring). And it cannot be said that a state’s desire to provide immediate protection to the due process rights of its citizens is not compelling. When balanced against the slight imposition on a right of dubious provenance, retroactive application of the updated Recognition Act does not violate the Texas Constitution. III. A. Having decided that Texas’s choice to apply its new Recognition Act to pending cases was proper, we now must review the district court’s application of that law. And to do that we must determine how closely we should scrutinize that court’s work. Reciting the standard of review in an appellate opinion is often a rote exercise. Not here. Recognizing that the appeal’s outcome largely turns on this question, the parties have spent considerable energy contesting whether we owe deference to certain district court rulings. Maghreb insists that we should review all aspects of the district court’s denial of recognition de novo, likening the inquiry to a review for legal sufficiency. DeJoria counters that we should review the court’s factual findings only for clear error. Much of the confusion surrounding the standard of review arises from this case’s odd posture. The district court did not rule on a motion for summary judgment or conduct a bench trial, but instead resolved a “motion for nonrecognition.” That motion is a creature of state law.[12] Regardless of the styling of the motion on which the district court ruled, however, our appellate standard of review is governed by federal law, even in this diversity case. See Goodner v. Hyundai Motor Co., Ltd., 650 F.3d 1034, 1040 (5th Cir. 2011); Tax Track Sys. Corp. v. New Investor World, Inc., 478 F.3d 783, 789 (7th Cir. 2007); Hershon v. Gibraltar Bldg. & Loan Ass’n, Inc., 864 F.2d 848, 852 (D.C. Cir. 1989).[13] The prior panel explained that “[w]hether the judgment debtor established that [a] non-recognition provision[] applies is a question of law reviewed de novo.” DeJoria, 804 F.3d at 379. We agree. But the panel had no cause to determine the proper standard of review for the factual findings that underpin the district court’s legal decision. After all, the issues in that appeal—whether the Moroccan system provides procedures compatible with due process, whether Moroccan law provides a mechanism to reciprocate recognition of Texas judgments, and whether the Moroccan court had personal jurisdiction over DeJoria—were all legal determinations. See FED. R. CIV. P. 44.1 (“The court’s determination [of foreign law] must be treated as a ruling on a question of law.”); In re DePuy Orthopaedics, Inc., Pinnacle Hip Implant Product Liability Litig., 888 F.3d 753, 778 (5th Cir. 2018) (“This court reviews [the] district court’s exercise of personal jurisdiction de novo.” (quotation omitted)). But it is a venerable principle that a district court’s factual findings are reversed only if clearly erroneous. FED. R. CIV. P. 52(a)(6) (standard for bench trials); see also Steven Alan Childress & Martha S. Davis, 1 FEDERAL Standards of review § 2.03[8] 2-32-33 (4th ed. 2010) (explaining that “[m]any courts . . . have assumed that [the] clearly erroneous rule applies to findings made on motions in addition to trial findings”). Even when an appellate court considers a legal question de novo, that plenary power of review does not extend to subsidiary factual findings. See Allstate Ins. Co. v. Abbott, 495 F.3d 151, 160 (5th Cir. 2007) (“[S]ubsidiary facts are reviewed for clear error.”) {citingMaine v. Taylor, 477 U.S. 131, 144-45 (1986)). To take just one example, jurisdiction is a legal question. But the facts that underlie a jurisdictional determination are still reviewed only for clear error. See, e.g., id.; DePuy Orthopaedics, 888 F.3d at 778 (applying clear error review to “underlying jurisdictional findings of fact” and de novo review to ultimate personal jurisdiction holding (quotation omitted)); Pederson v. La. State Univ., 213 F.3d 858, 869 (5th Cir. 2000) (“If the district court resolves any factual disputes in making its jurisdictional findings,” those resolutions are overturned only if “clearly erroneous.” (quotation omitted)). The same must be true for factfinding that underpins the legal conclusion of nonrecognition. Appellate court deference to district court factfinding is grounded in concerns of both expertise and efficiency. Maghreb points out that one of the strongest justifications for deference—the trial court’s ability to assess the credibility of live testimony, Anderson v. City of Bessemer City, 470 U.S. 564, 575 (1985)—is not present because the testimony of the foreign witnesses was presented on paper. But we defer even when the trial court’s findings are “based . . . on physical or documentary evidence or inferences from other facts.” Id. at 574. That is because “[t]he trial judge’s major role is the determination of fact, and with experience in fulfilling that role comes expertise.” Id. Clear error review also promotes judicial efficiency. Id. at 574-75 (“Duplication of the trial judge’s efforts in the court of appeals would very likely contribute only negligibly to the accuracy of fact determination at a huge cost in diversion of judicial resources.”). District court judges, who do the lion’s share of the work in our federal system, do not dig through voluminous records only to have courts like this one restart the factfinding from scratch. Instead of redoing their work, we defer to their findings so long as they take a permissible view of the evidence. Id. at 574. Although the standard of review is a federal issue, like the prior panel we “look to Texas law” governing recognition to see if anything counsels in a different direction.[14] DeJoria, 804 F.3d at 379. Nothing does. We see no Texas recognition case that conflicts with the principles of federal appellate review outlined above. When a trial court is presented with conflicting evidence in recognition proceedings, Texas courts “defer to the trial court’s . . . resolution of those conflicts.” Mariles v. Hector, No. 05-16-00814-CV, 2018 WL 3723104 *6 (Tex. App.—Dallas Aug. 6, 2018, pet. denied). Maghreb cites some Texas cases that explain what we have acknowledged: that review of the district court’s ultimate determination of the application of a nonrecognition factor should be de novo. See, e.g., Sanchez v. Palau, 317 S.W.3d 780, 785 (Tex. App.—Houston [1st Dist.] 2010, pet. denied) (“[W]e review de novo a trial court’s ruling on recognition of a foreign country judgment.”); The Courage Co., L.L.C. v. The Chemshare Corp., 93 S.W.3d 323, 331 (Tex. App.—Houston [14th Dist.] 2002, no pet.) (same).[15] But they have pointed to no case that instructs a court of appeals to start on a blank slate in determining the facts. That is not surprising. Consistent with the standard practice, Texas courts also generally defer to trial court factfinding. In re I.I.G.T., 412 S.W.3d 803, 806 (Tex. App.—Dallas 2013, no pet.) (explaining that an appellate court should not normally “disturb the [trial] court’s resolution of evidentiary conflicts that turn on . . . the weight of the evidence”). We thus can disturb the district court’s findings only if they are not “plausible in light of the record viewed in its entirety.” Anderson, 470 U.S. at 574. B. Maghreb’s primary argument on appeal—that DeJoria lost his opportunity to complain about the Moroccan proceedings because he failed to participate in them—must overcome this deference to the district court’s factfinding. Maghreb notes that we have “flatly reject[ed]” the due process objections of judgment debtors who were “ given, and waived, the opportunity of making [an] adequate presentation” in the foreign tribunal. Society of Lloyd’s v. Turner, 303 F.3d 325, 331 n.20 (5th Cir. 2002) (quotation omitted); see also Dart, 953 S.W.2d at 480 (“Grounds for nonrecognition may be waived if a party had the right to assert that ground as an objection or defense in the foreign country but failed to do so.”). But our limited authority when it comes to facts makes short work of that argument. The district court made three major findings to support nonrecognition: 1) DeJoria’s fear for his safety should he return to Morocco to litigate was credible and arose directly from his involvement in the Moroccan lawsuit because DeJoria’s position in the Moroccan lawsuit was directly adverse to the interests of the royal family, he was unable to retain a lawyer to appear for him in the initial proceedings or to bring an appeal although the determination of damages was based on expert opinion, the Moroccan court manipulated that process when it went through four experts before finding one that would deliver its preferred recommendation. Taken together, the first and second findings mean that DeJoria was never “given . . . the opportunity of making [an] adequate presentation” in Moroccan court and the third means his case did not otherwise receive fair treatment. Turner, 303 F.3d at 331 n.20. So unless those findings were clearly erroneous, Maghreb’s “waiver” argument fails. To be sure, Maghreb points to substantial evidence that could support contrary findings. Its problem is that there is evidence on both sides of these disputes. Even if Maghreb can convince us that its evidence is stronger, that is not enough to establish that the district court’s crediting of DeJoria’s evidence is implausible. Theriot v. Par. of Jefferson, 185 F.3d 477, 490 (5th Cir. 1999) (“Where the evidence can support findings either way, a choice by the trial judge between two permissible views of the weight of the evidence is not clearly erroneous.”) Take for instance the finding that DeJoria credibly feared for his life and so was unable to attend the Moroccan proceedings in person. Michael Gustin, DeJoria’s business partner, described receiving a death threat and explained that it was directed at both him and DeJoria. DeJoria himself declared that Gustin communicated that threat to him and that he believed it was credible. And the record contains evidence that their unsuccessful attempts to obtain representation in Morocco may have only heightened their fear. A French attorney with some Moroccan experience told them that it was not only unsafe for DeJoria and Gustin to return to Morocco, but it would be “unsafe and unwise for any lawyer” or “any sane person,” for that matter, to participate in a case that so closely touched the royal family’s interests. Nearly a decade later, that attorney repeated his concerns. The general counsel for Skidmore, DeJoria’s company that spearheaded the Moroccan project, also says he was told to stay out of the country by a Moroccan attorney who had been hired to handle various clerical tasks as the Moroccan lawsuit proceeded. She warned that “any appearance by Skidmore or any personal representative of Skidmore in the Moroccan lawsuit would be dangerous.” Of course, these assertions all come from individuals who may have an axe to grind in this case. And we are not told much about the circumstances or content of the death threat because Gustin maintains that he “cannot reveal [the] details . . . without compromising the safety of innocent people still in Morocco.” Bias and lack of detail are classic impeachment evidence. But impeachment usually goes to the weight of the evidence. Arguing about the weight of the evidence is not the terrain an appellant wants to be on. See La Day v. Catalyst Tech., Inc., 302 F.3d 474, 480 (5th Cir. 2002) (noting that it is the factfinder who “ultimately . . . decide[s] which side has the greater weight of the evidence”). Nor does Maghreb get over the clearly-erroneous hurdle because it presented testimony that DeJoria could have appeared and obtained counsel in the Moroccan litigation. Choosing between conflicting testimony is the province of the factfinder. See Anderson, 470 U.S. at 575 (concluding it “can virtually never be clear error” when a trial court “credit[s] the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence”). And while Maghreb emphasizes the testimony of its expert on Moroccan law, expert testimony does not automatically trump lay testimony. Breland v. United States, 372 F.2d 629, 633 (5th Cir. 1967) (“[L]ay testimony can be sufficient to satisfy [a party's] burden even though there is expert testimony to the contrary.”); see also FIFTH CIRCUIT PATTERN JURY INSTRUCTIONS 3.5 (Civil) (2014) (explaining that, for expert witness testimony, “[a]s with any other witness, it is up to [the factfinder] to decide whether to rely on it”). To undo factual determinations on appeal, Maghreb must convince us not that it has the more compelling evidence, but that the other side’s testimony is not “plausible.” Anderson, 470 U.S. at 574. Maghreb’s expert witness, a Moroccan attorney, contends that DeJoria’s worries were “baseless and reflect[] his poor understanding of Morocco.” And they also point to several instances in which Moroccan courts have ruled against royal interests. But that a trier of fact could plausibly infer that the death threat was fabricated does not mean it is implausible to find that the threat was real. Id. at 574. The same may be said for the other two key findings. For instance, although DeJoria was able to retain Moroccan attorneys as experts in proceedings stateside after the Moroccan trial court handed down its judgment, there was evidence that two of his attempts to obtain representation in the Moroccan proceedings were rebuffed. And though there was no smoking gun, it was not clear error for the district court to conclude that the Moroccan court went fishing for an expert who would determine DeJoria and his partner had caused Maghreb substantial damages. After all, the expert who found those damages was the fifth appointed by the Moroccan court—the first three “concluded that they could not provide any firm opinion on the matter” and the fourth was replaced for reasons that remain unclear. Although the district court’s assessment of the evidence may be subject to vigorous debate, it is the district court’s job to resolve evidentiary disputes, not ours. Guzman v. Hacienda Records & Recording Studio, Inc., 808 F.3d 1031, 1036 (5th Cir. 2015) (explaining that, even when “there are two permissible views of the evidence,” the trial court’s choice between them is typically owed “great deference”). Maghreb has not shown clear error.[16] C. Perhaps realizing that its argument founders on the district court’s difficult-to-undo findings, Maghreb’s primary challenge to those findings is that they should not have been made in the first place. Each of the pertinent findings, it argues, was precluded by the prior panel’s opinion. Under the law-of-the-case doctrine—and its corollary, the mandate rule—when a district court receives a case on remand, it may not reexamine the legal or factual determinations of this court or otherwise disobey our mandate. See Tollett v. City of Kemah, 285 F.3d 357, 363-64 (5th Cir. 2002). The reach of those related doctrines extends only to matters decided expressly or by necessary implication. In re Felt, 255 F.3d 220, 225 (5th Cir. 2001). And an issue is tacitly decided only when its disposition is a “necessary predicate[] to the ability to address the issue or issues specifically discussed” in the appellate court’s opinion. Id. The prior panel’s opinion did not preclude the findings the district court made on remand. First and foremost, the prior appeal was decided under a different law. United States v. Matthews, 312 F.3d 652, 657 (5th Cir. 2002) (describing an exception to the law-of-the-case doctrine when “there has been an intervening change of law by a controlling authority”). That law did not require the prior panel to determine whether DeJoria’s “specific proceeding[s]” were “ compatible with the requirements of due process of law.” TEX. CIV. PRAC. & REM. CODE § 36A.004(c)(8). So that panel had no cause to determine whether DeJoria could in fact safely return to Morocco or whether DeJoria could in fact retain representation. In determining whether the Moroccan legal system made fair proceedings impossible, whether Moroccan courts would reciprocate recognition, and whether the Moroccan commercial court had personal jurisdiction over DeJoria, the prior panel’s analysis was focused on legal questions. The fact-intensive inquiry demanded by Texas’s updated Recognition Act put the case on a new playing field. But even if the district court were operating in the same legal landscape, nothing in the prior panel’s opinion forecloses the district court’s findings. The panel held that the Moroccan court’s exercise of jurisdiction over DeJoria did not violate traditional notions of “fair play and substantial justice” because, despite any burden litigating in Morocco might place on DeJoria, “Moroccan courts do not require that the defendant appear personally, and DeJoria could have litigated entirely through counsel without returning to Morocco.” DeJoria, 804 F.3d at 389. And, relying on testimony from a Moroccan attorney acting as Maghreb’s expert, the court pointed out that “it is ‘not at all uncommon’ for Moroccan attorneys to represent unpopular figures in Moroccan courts.” Id. at 383. But these general statements about usual Moroccan practices did not address whether DeJoria could have found a willing attorney in Morocco in his high-profile case.[17] Nor does it avail Maghreb to draw our attention to the previous panel’s aside that, “[a]lthough our inquiry focuses on Morocco’s judicial system, we also observe that the record does not establish that the King actually exerted any improper influence on the Moroccan court in this case.” Id. at 382 n.9 (emphasis in original). For one, the comment is admittedly dicta—the footnote could have been erased from the opinion without disrupting its systemwide holding in the slightest. Pegues v. Morehouse Parish Sch. Bd., 706 F.2d 735, 738 (5th Cir. 1983) (explaining that law of the case does not apply to dicta). And in any case, the question under the Texas statute is not whether the King actively undermined the proceedings, but whether DeJoria was afforded a fundamentally fair hearing. The prior panel’s general observations did not foreclose the more searching factual inquiry now required under Texas law.[18] D. Our holding that the district court did not clearly err in its factfinding nor adopt those findings in the face of a contrary mandate from this court leaves us little left to do. Maghreb does not dispute the nonrecognition conclusion if we uphold the findings that DeJoria could neither appear personally nor find a lawyer to appear for him. That is a sensible stance. Recognition of a foreign-country judgment does not require the foreign court to “comply with the traditional rigors of American due process.” Turner, 303 F.3d at 330. But the opportunity to present one’s case is no minor twist or turn of modern due process jurisprudence: “The fundamental requirement of due process is the opportunity to be heard at a meaningful time and in a meaningful manner.” Mathews v. Eldridge, 424 U.S. 319, 333 (1976) (quotation omitted). In light of the facts as found by the district court, it properly determined that DeJoria was denied due process in Morocco. The district court thus had and properly exercised discretion to deny recognition to the Moroccan judgment.[19] *** So despite the seeming complexity of this case—royal intrigue, a foreign proceeding, almost a billion dirhams at stake—it ends up being resolved on one of the most basic principles of appellate law: deference to the factfinder. The judgment is AFFIRMED.
