Before JONES, ELROD, and HIGGINSON, Circuit Judges. EDITH H. JONES, Circuit Judge: Since 2013, Appellants (collectively, “Defense Distributed”) have been challenging publication restraints imposed by the U.S. State Department, federal courts, and the State of New Jersey after Defense Distributed published to the Internet computer assisted design (“CAD”) files for a single-round plastic pistol. Although Defense Distributed is still prevented from publishing,[1] the CAD files it published remain available to this day on countless other websites internationally. Defense Distributed has yet to secure a court decision condemning what appear to be flagrant prior restraints.[2] The instant combined appeal and motion for mandamus relief stems from a district court order severing the case against one defendant, the Attorney General of New Jersey (NJAG),[3] and transferring it to a federal court in New Jersey. We conclude that mandamus relief is appropriate in this unusual case. Accordingly, we direct the district court to request retransfer from its counterpart in New Jersey and order other relief in accordance herewith. BACKGROUND Two previous opinions of this court describe the litigation history surrounding Defense Distributed’s publication of then-novel CAD files for a pistol that can theoretically be “printed” from a computer affixed to the proper equipment. See Defense Distributed v. Grewal, 971 F.3d 485, 488-90 (5th Cir. 2020); Defense Distributed v. U.S. Dep’t of State, 838 F.3d 451, 454- 58 (5th Cir. 2016); id. at 461-66 (Jones, J., dissenting). We quote from the more recent opinion: Plaintiff Defense Distributed is a Texas company operated for the purpose of promoting popular access to firearms. To carry out this purpose, it produces and makes accessible information related to the 3D printing of firearms and publishes and distributes such information to the public. Plaintiff Second Amendment Foundation, Inc. (“SAF”) is a nationwide, non-profit membership organization that “promotes the right to keep and bear arms by supporting education, research, publications, and legal efforts about the Constitution’s right to privately own and possess firearms and the consequences of gun control.” Across the nation, SAF members seek the digital firearms information created by Defense Distributed, circulate their own digital firearms information by utilizing Defense Distributed’s facilities, and republish digital firearms information independently. Defense Distributed began distributing files related to the 3D printing of firearms in December 2012. It did so by publishing files to its defcad.org and defcad.com websites and letting visitors freely download them. It also distributed digital firearms information via mail and at a brick-and-mortar public library in Austin, Texas. Defense Distributed’s efforts were initially met with opposition from the United States Department of State.[4] But, after a period of litigation, the parties reached a settlement agreement that granted Defense Distributed a license to publish its files. Shortly thereafter, nine Attorneys General, including New Jersey Attorney General Grewal, filed suit on behalf of their respective states in the Western District of Washington to enjoin the State Department from authorizing the release of Defense Distributed’s files. They argued that the State Department’s license to Defense Distributed constituted an ultra vires about-face that violated the Administrative Procedure Act and jeopardized the states’ statutory and regulatory schemes for firearms. The Western District of Washington quickly issued a temporary restraining order, followed closely by a nationwide preliminary injunction.[5] Just before the Attorneys General sued in Washington, Defense Distributed and SAF brought the instant action in the Western District of Texas challenging select enforcement actions taken by the state Attorneys General. Of relevance to this appeal, Plaintiffs alleged these actions by Grewal: (1) sending a cease-and-desist letter threatening legal action if Defense Distributed published its files; (2) sending letters to third-party internet service providers based in California urging them to terminate their contracts with Defense Distributed; (3) initiating a civil lawsuit against Defense Distributed in New Jersey;[6] and (4) threatening Defense Distributed with criminal sanctions at a live press conference. Further, these actions, coupled with the injunctive orders issued in the Washington litigation, have caused Defense Distributed to cease publication of its materials. The Plaintiffs asserted, inter alia, that these actions infringed the exercise of their First Amendment freedoms and constituted tortious interference with the State Department’s settlement agreement. Grewal moved to dismiss for lack of personal jurisdiction.[7] The Plaintiffs, meanwhile, sought a preliminary injunction. After holding a hearing and considering the parties’ arguments, the court granted Grewal’s motion and dismissed the action without prejudice. Defense Distributed v. Grewal, 971 F.3d at 488-89 (footnotes in original). Defense Distributed appealed. This court held that the NJAG is amenable to personal jurisdiction in Texas courts. Id. at 488. Accordingly, the court reversed and remanded for further proceedings. Id. Following our remand to the district court, Defense Distributed amended its complaint to join the State Department as a defendant for its alleged failure to comply with a Settlement Agreement reached with Defense Distributed in 2018. Shortly after, the NJAG moved to sever Defense Distributed’s case against him and transfer that portion of the case to a New Jersey federal court. The State Department opposed severance, as did Defense Distributed. Nonetheless, the district court obliged the NJAG by written order both severing and transferring the case against him. Defense Distributed immediately noticed an appeal from the severance-and-transfer order and followed with an alternate request for mandamus relief against the district judge. This court imposed a temporary stay of the case pending appeal. The New Jersey district court also stayed all proceedings pending this appeal. DISCUSSION Appellate jurisdiction We pretermit Defense Distributed’s resort to the collateral order doctrine as a basis for appellate jurisdiction in this interlocutory appeal[8] because, in this circuit, mandamus is the prescribed vehicle for reviewing rulings on transfers of cases pursuant to 28 U.S.C. § 1404(a). See In re Volkswagen of Am., Inc., 545 F.3d 304, 309 (5th Cir. 2008) (en banc); In re Rolls Royce Corp., 775 F.3d 671, 676-77 (5th Cir. 2014). The twist in this case is the transfer to a district court outside the Fifth Circuit, a court over which this court exercises no control. This court lacks power to order a return of the case to our circuit. But In re Red Barn Motors, Inc., 794 F.3d 481 (5th Cir. 2015), explained the path to a cognizable mandamus remedy. Id. at 483-84. The All Writs Act, 28 U.S.C. § 1651, empowers courts to issue writs of mandamus if the courts also have appellate jurisdiction, “although no appeal has been perfected.” Roche v. Evaporated Milk Ass’n., 319 U.S. 21, 25, 63 S. Ct. 938, 941 (1943). Because of the transfer, the Texas transferor court can no longer enter an appealable order in the case. In re Red Barn, nevertheless, approved that “several circuits, where appropriate, have endorsed the method of directing the transferor district court to request that the transferee district court return the case.” Id. at 484 (collecting cases). To be sure, to avoid friction with sister circuits, the court held that “if we even have the power to reverse such a transfer, we should exercise it only if faced with ‘a very extreme case.’” Id. (citing In re Sw. Mobile Homes, Inc., 317 F.2d 65, 66-67 (5th Cir. 1963)). But intercircuit friction was reduced where the transferee court there, as here, stayed proceedings pending our appellate panel’s decision. Id. The court concluded by requiring a party seeking relief from a transfer order to exercise diligence. Id. at 485. Because the petitioner had waited three months to file its mandamus petition, the writ was denied. Id. In this case, of course, the notice of appeal was filed the day after the district court’s order and mandamus relief formally sought within thirty-nine days while briefing was underway. Because In re Red Barn sets the standard, we have jurisdiction to consider the mandamus petition. The NJAG asserts four objections to our jurisdiction. He asserts that In re Red Barn did not hold what we just stated it held; that other circuits would not allow a writ of mandamus where a case is transferred to another circuit; and that Defense Distributed was not diligent. Finally, he contends that because Defense Distributed may eventually challenge the transfer order in courts of the Third Circuit, the availability of appellate review disentitles him to an equitable writ. We find none of these arguments persuasive. First, In re Red Barn denied the writ only for lack of petitioner’s diligence. That conclusion would have been inappropriate had the panel concluded that mandamus is not available to challenge an out-of-circuit transfer order. The court’s entire discussion of the writ, and a limitation where intercircuit friction may occur, would have been unnecessary. Instead, the court weighed the approaches of other circuits, noting that in particular cases a circuit court exercised jurisdiction to vacate a completed intercircuit transfer. See In re Red Barn, 794 F.3d at 484 n.6. This discussion would also be extraneous had the court simply been uttering dicta rather than explaining its rationale. And of course, the court noted that several circuits have authorized mandamus relief in these circumstances. Id. Finally, there is no “assuming arguendo” language that couches the court’s ruling on mandamus in hypothetical terms.[9] In short, In re Red Barn’s procedural holding is not predicated on dicta. Second, other circuits have adopted the same approach. See, e.g., In re Howmedica Osteonics Corp., 867 F.3d 390, 399-400, 411 (3d Cir. 2017) (asserting jurisdiction to vacate intercircuit transfer order); In re Warrick, 70 F.3d 736, 739-40 (2d Cir. 1995) (asserting mandamus jurisdiction and distinguishing In re Drabik, 246 F.2d 408 (2d Cir. 1957), as not justifying denial of all relief “in these circumstances”); A.C. Nielsen Co. v. Hoffman, 270 F.2d 693, 695 (7th Cir. 1959) (concluding that the court has mandamus jurisdiction over intercircuit transfer order). Each of these circuits has applied mandamus to decide the propriety of intercircuit transfers. Although the Tenth Circuit held that a completed intercircuit transfer divested it of appellate jurisdiction, Chrysler Credit Corp. v. Country Chrysler, Inc., 928 F.2d 1509, 1516-17 (10th Cir. 1991), that court went on to recognize that “technically,” the district court should have deferred effectuating the transfer so that the parties could seek certification or “file a mandamus petition.” Id. at 1520 n.9. See also In re Nine Mile Ltd., 673 F.2d 242 (8th Cir. 1982)(mandamus ordered where district court erroneously transferred papers to consummate Section 1404 case transfer before appeal could be filed in transferor district). Thus, the balance of circuit court authority, in addition to the cases cited by In re Red Barn, favors jurisdiction in this case. Third, the AG questions the diligence of Defense Distributed because of its failure to seek an immediate stay in the district court or file an immediate mandamus petition. Nothing in applicable precedent, however, mandates the particular method by which a party disadvantaged by an out-of- circuit transfer must bring that issue to the circuit court. Defense Distributed filed its notice of appeal the day after the district court’s transfer order and promptly sought relief in this court. The defendants and district court were immediately placed on notice of Defense Distributed’s intent to challenge the transfer in this court, and the actual transfer order was not docketed until the following day. It sought mandamus relief in briefing to this court, filed thirty-nine days after the court’s order.[10] The Appellants’ diligence comports with this court’s emphasis on “diligence” in In re Red Barn and with the timing of filings in In re Howmedica, 867 F.3d at 400. Fourth, it is inaccurate to assert, as the NJAG does, that Defense Distributed has adequate available alternatives for appellate court review of the district court’s order, which render a mandamus remedy unavailable. Unlike In re Red Barn, this case encompasses severance as a necessary condition of the transfer.[11] On appeal, the Third Circuit could not review the severance order, much less attempt to consolidate the case against the NJAG with the case still pending in Texas against the State Department. Therefore, on its face, this tandem order is effectively unreviewable in the transferee circuit. Moreover, if the potential Third Circuit appeal were (arbitrarily) limited to the transfer order alone, a premier procedural treatise notes that review in the transferee court seems illusory given the application of comity and law-of-the-case principles. See 15 CHARLES ALAN WRIGHT, ARTHUR R. MILLER & EDWARD H. COOPER, FEDERAL PRACTICE AND PROCEDURE [HEREAFTER, WRIGHT & MILLER] § 3846 (4th ed. 2021); id. at § 3855 (similar). The likelihood of securing a retransfer by the transferee district court seems equally constrained. But even if these propositions are arguable, see SongByrd, Inc. v. Estate of Grossman, 206 F.3d 172, 178 n.7 (2d Cir. 2000) (collecting cases), the SongByrd decision, like that of other courts, acknowledges the availability of mandamus review in the Fifth Circuit. Id. at 176-77, 176 n.5.[12] Finally, the fact that the district court here granted the transfer, rather than denying it, makes no difference. In this circuit, neither In re Rolls Royce nor In re Volkswagen limited writs of mandamus according to whether the district courts granted or denied transfer.[13] And in Howmedica, the Third Circuit applied mandamus to vacate the order granting transfer to the Central District of California. In re Howmedica, 867 F.3d at 411. Whatever the possibility of review in the transferee circuit, this has not been regarded as sufficiently effective to preclude writs of mandamus in the transferor court.[14] Mandamus A writ of mandamus is “a drastic and extraordinary remedy reserved for really extraordinary cases.” In re Depuy Orthopaedics, Inc., 870 F.3d 345, 350 (5th Cir. 2017). This court will grant a petition for a writ of mandamus only if the petitioner satisfies three conditions. First, the petitioner must show that there are “no other adequate means to attain the relief he desires.” Cheney v. U.S. Dist. Court, 542 U.S. 367, 380, 124 S. Ct. 2576, 159 L.Ed.2d 459 (2004). Second, the court “must be satisfied that the writ is appropriate under the circumstances.” Id. at 381, 124 S. Ct. 2576. And third, the petitioner must show a “clear and indisputable right to the writ.” Id. Defense Distributed easily satisfies the first two conditions for mandamus relief. As we just explained, the Third Circuit transferee court cannot, under normal appellate standards, exercise appellate review over either the severance or transfer orders of this Western District of Texas. E.g., In re Red Barn, at 484-85. Second, granting the writ would be “especially appropriate where the issues implicated have importance beyond the immediate case.” In re JPMorgan Chase & Co., 916 F.3d 494, 499 (5th Cir. 2019) (quoting In re Lloyd’s Register N. Am., Inc., 780 F.3d 283, 294 (5th Cir. 2015); and In re Volkswagen, at 319 (internal quotation marks omitted). Before this court are “issues that implicate not only the parties’ interests but those of the judicial system itself.” United States v. Bertoli, 994 F.2d 1002, 1014 (3d Cir. 1993). Preeminent are questions about the abridgement of the Plaintiffs’ first amendment rights to publish their materials. Also critical, however, are tactics suggesting the abusive manipulation of federal court procedures in order to delay or altogether avoid meaningful merits consideration of Plaintiffs’ claims. What the parties most strenuously debate is whether Defense Distributed has shown a clear and indisputable right to the writ or a clear abuse of discretion by the district court. See Cheney at 281, 2576; see also In re Volkswagen, at 308. “A district court by definition abuses its discretion when it makes an error of law.” Koon v. United States, 518 U.S. 81, 100, 116 S. Ct. 2035, 2047 (1996) (citing Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 405, 110 S. Ct. 2447, 2461 (1990) (“a district court would necessarily abuse its discretion if it based its ruling on an erroneous view of the law”)). We conclude that the district court clearly abused its discretion by applying the wrong legal standard for evaluating the NJAG’s conjoined severance and transfer motions and by egregiously misinterpreting Defense Distributed’s claims. Moreover, even if the motions are evaluated wholly independently, the transfer motion cannot stand if the severance motion was wholly unjustified. See, e.g., Chrysler Credit Corp. v. Country Chrysler, 928 F.2d 1509 (10th Cir. 1991); In re Nine Mile Ltd., 673 F.2d at 242 (because transfer order was procedurally incorrect, mandamus ordered district court to request return of the case from the transferee court). Joint Severance and Transfer To explain this conclusion, we remind that the Federal Rules of Civil Procedure were enacted to ensure “the just, speedy and inexpensive determination of every action and proceeding.” Fed. R. Civ. P. 1. To be more precise, Fed. R. Civ. P. 21, which authorizes severance of parties, and 28 USC § 1404(a), governing transfers among federal district courts, “are [both] designed to facilitate just, convenient, efficient and less expensive determination.” In re Nintendo of America, Inc., 756 F.3d 1363, 1365 (Fed. Cir. 2014). Although a district court has broad discretion to sever parties that were otherwise properly joined by a Plaintiff,[15] “[u]nder the Rules, the impulse is toward entertaining the broadest possible scope of action consistent with fairness to the parties; joinder of claims, parties and remedies is strongly encouraged.” United Mine Workers v Gibbs, 383 U.S. 715, 724, 86 S. Ct. 1130, 1138 (1966). On two occasions, this court has discussed standards for joint severance and transfer motions brought by a defendant or defendants, the effect of which would require a Plaintiff to split its case into two different federal courts. In a well-known opinion, this court noted that “our jurisprudence suggests that the severance inquiry is different—and more focused on judicial efficiency—when it is combined with a section 1404 motion to transfer than when the severed case would remain in the original judicial district.” In re Rolls Royce, 775 F.3d at 680.[16] The court quoted at length Judge Alvin Rubin’s opinion explaining the problems that can arise in a case involving multiple defendants when claims against one or some defendants are severed and transferred while claims against the remaining defendant are retained in the original court. Id., (quoting Liaw Su Teng v. Skaarup Shipping Corp., 743 F.2d 1140, 1148-49 (5th Cir. 1984)). Judge Rubin’s list of parameters that should inform this type of proposed transfer bears repeating: But before thus putting asunder what the Plaintiff has joined, the court must weigh carefully whether the inconvenience of splitting the suit outweighs the advantages to be gained from the partial transfer. It should not sever if the defendant over whom jurisdiction is retained is so involved in the controversy to be transferred that partial transfer would require the same issues to be litigated in two places. [See Korbel, The Law of Federal Venue and Choice of the Most Convenient Forum, 15 Rutgers L.Rev. 607 (1961); Masington, Venue in the Federal Courts---The Problem of the Inconvenient Forum, 15 U.Miami L.Rev. 237 (1961); Note, 46 Cornell L.Q. 318 (1961)]. That being the situation here, the district court should not have severed the claims if there were any alternative. Manifestly, the Plaintiffs will suffer some inconvenience if they are forced to litigate their claims in two courts, half the world apart from each other, with not only the consequent added expense and inconvenience but also the possible detriment of inconsistent results. A single forum is also most suitable for determining possible counter- and cross-claims. The public also has an interest in facilitating a speedy and less-expensive determination in one forum of all of the issues arising out of one episode. Liaw Su Teng v. Skaarup Shipping Corp., 743 F.2d 1140, 1148–49 (5th Cir. 1984), overruled on other grounds, In re Air Crash Disaster Near New Orleans, La., on July 9, 1982, 821 F.2d 1147 (5th Cir. 1987). Several points made in Judge Rubin’s opinion are common to this species of joint severance and transfer motions. A court must “weigh carefully” the comparative inconvenience of splitting the suit versus the advantages to be gained from a partial transfer. It should not sever if the defendant not transferred is “so involved” in the controversy transferred to another court that partial transfer would require the same issues to be litigated in two places. “Manifestly,” Plaintiffs are disadvantaged by the expense and inconvenience of having to litigate in two disparate fora and by the possibility of inconsistent results. And the public has an interest in the comparative speediness and cost-savings from utilizing a single forum for the issues arising out of one episode. See also 15 WRIGHT & MILLER, § 3845, at 86 (echoing these factors). Other courts have agreed with Liaw Su Teng, and specified that in analyzing severance and transfer motions, “[b]efore effecting such a severance, a judge should weigh the convenience to the parties requesting transfer against the potential inefficiency of litigating the same facts in two separate forums.” White v. ABCO Eng’g Corp., 199 F.3d 140, 144 (3d Cir. 1999). Previously, the Third Circuit, explicitly endorsing Judge Rubin’s view, denied severance and transfer where the defendant not transferred was more than “indirectly connected” to the Plaintiff’s dispute with the defendant seeking transfer. Sunbelt Corp. v. Noble, Denton and Assoc., Inc., 5 F.3d 28, 34 (3d Cir. 1993). The Second Circuit, which pioneered using severance and transfer as a permissible strategy of case management, nonetheless held it appropriate only “where the administration of justice would be materially advanced….” and where a defendant in one district is only “indirectly connected” to the claims that will be transferred to the other district. Wyndham Assoc. v. Bintliff, 398 F.2d 614, 618-19 (2d Cir. 1968).[17] The district court here quoted Rolls Royce, but it misperceived and thus misapplied this court’s explanation about “judicial efficiency.” In the above cases, courts were not deciding ordinary Rule 21 severance motions, in which parties or claims may be split while the disputes are maintained before the same court. Severance and transfer requires two courts to engage in the work of one, prompting serious concerns about duplication of judicial resources, the consistency of rulings, and litigation costs. The circuit courts that discussed severance and transfer motions in this context have uniformly indicated an aversion to granting such motions at the expense of needless duplication of judicial effort. The rule emanating from Rolls Royce and Liaw Su Teng is that in most multidefendant cases—other than those involving forum selection clauses—severance and transfer makes sense only where the administration of justice would be materially advanced and a defendant in one district is not “so involved” in the transferred controversy that the same issues would have to be litigated twice.[18] Consequently, the district court erred legally in finding that the State Department is only indirectly connected to claims pled against the NJAG[19] and in failing to find that the administration of justice would not be materially advanced by transfer. The direct connection between the Plaintiffs’ claims against these two defendants contradicts the district court’s artful dissection of Defense Distributed’s Second Amended Complaint. The court ignored the major components of the lawsuit and the overarching connection alleged between the State Department and the NJAG. Both government entities have suppressed legal speech by prohibiting the publication of the company’s digital firearms files on the internet and into New Jersey. That the First Amendment protects all of the Plaintiffs’ publications underlies all of their other claims.[20] The Plaintiffs’ assertion of seventeen separate claims reflects the consequences of that suppression and the tortuous path they have been forced to pursue for redress. Yet, contrary to common sense and heedless of a potential for disparate constitutional rulings if the case against the NJAG remains transferred, the court belittled the First Amendment issues. According to the district court, not all of Plaintiffs’ claims involve the First Amendment. The court overlooked Plaintiffs’ claims that directly involve both Defendants in the settlement agreement breach and in censorship under the Second Amendment and Due Process clause. And the larger point is that none of Plaintiffs’ claims would exist if they had been allowed to publish the various digital firearms files continuously since 2013. The First Amendment is the sine qua non of this case. The other larger point ignored by the district court is that the principal claims against both defendants are temporally and factually intertwined to the extent that litigation in separate courts would largely overlap. A paraphrase of the facts alleged in the Second Amended Complaint is demonstrative. Just when the Plaintiffs, after several years of litigation, thought they had obtained relief from a signed settlement agreement with the State Department, the NJAG loudly led a bevy of states in opposition. Within a span of about six months, he filed suit in Washington state, even more remote from New Jersey than Texas, and forced the joinder of Defense Distributed as a defendant there. He evidently thought the parties’ fates were legally and factually connected when he sought and obtained an injunction expressly to prevent the State Department from completing the settlement with Plaintiffs. The NJAG kept up public pressure against the settlement with threats of civil and criminal punishment against Plaintiffs’ president Cody Wilson personally. At this time, the State Department opted not to complete the settlement. Accordingly, the State Department did not modify relevant federal regulations, disavowed the license and exemption from regulations it had promised Defense Distributed, and refused to appeal adverse decisions by the Washington federal court. The extent to which the NJAG’s campaign influenced the State Department’s alleged breaches is relevant to the claims against both defendants. At the same time, the NJAG’s seminal place in the litigation is a major, though far from the only facet of his conduct that was designed to derail the settlement and independently to restrain Plaintiffs’ exercise of free speech. Plaintiffs’ essential claims for First Amendment violations, breach of the settlement agreement, and the NJAG’s interferences cannot be understood by a factfinder without investigating and telling the whole story. Similarly, having to tell the same story in two courts would abuse Plaintiffs and the judicial process.[21] Also plain from the complaint is that the subsidiary claims are dependent upon the primary claims identified above. To cite one instance, Defense Distributed pleads that the State Department violated the Administrative Procedure Act in various aspects of its settlement agreement breach. The Washington State litigation was also based on alleged violations of the Administrative Procedure Act. The State Department’s noncompliance may have been caused by the mere fact of the Washington State litigation or by the imperative of complying with that court’s injunctive decrees or by the NJAG’s public campaign against the settlement. Sorting out cause and effect for these claims demands a single factfinder and judicial resolution. The district court also erred legally by failing to conclude that the administration of justice would not be materially advanced by transfer of the case against the NJAG to New Jersey. See Rolls Royce, 775 F.3d at 680. The burden rests on the NJAG as the movant to demonstrate that proposition, and he did not succeed. First, the public interest in achieving a single court’s ruling on Plaintiffs’ First Amendment claims cannot be overstated. Government instigated censorship of constitutionally protected speech is abhorrent to self-government; courts have a duty to prevent illegal censorship. But severance and transfer enhances the risk of conflicting rulings, which would seriously injure the Plaintiffs and throw the law in this important constitutional area into national disarray for several years. Second, because of the complex factual interactions among Plaintiffs and these Defendants, discovery and trial for the principal claims will require many of the same witnesses. No efficiency exists from conducting two fact- based litigations half a continent apart. On the other hand, no efficiency is gained by having two courts decide the separate subsidiary claims against each Defendant. Such claims are largely legal in nature, overlapping in several instances, and best resolved on briefing to a single court that has gained familiarity with the intricacies of Plaintiffs’ dozens of digital firearms files and background materials that could be published to the internet and in New Jersey. There is no obvious efficiency advantage, much less materially enhanced judicial economy from forcing Defense Distributed’s case into two separate cases. The necessity of denying severance conjoined with transfer is confirmed by proper application of Rolls Royce and Liaw Su Teng. In addition, the district court’s separate and independent Rule 21 severance and Section 1404 transfer analyses are plagued with error and therefore alone justify the rebuke of mandamus. We turn first to severance. Severance In the Fifth Circuit, the accepted basis for Rule 21 severance analysis considers five factors: (1) whether claims arise out of the same transaction, occurrence, or series of transactions or occurrence; (2) whether the claims present common questions of law or fact; (3) whether settlement of the claims or judicial economy would be facilitated; (4) whether prejudice would be avoided if severance were granted; and (5) whether different witnesses and documentary proof are required for the separate claims.” See Rolls Royce, 775 F.3d at 680; Paragon Office Servs., LLC v. UnitedHealthcare Ins. Co., 2012 WL 4442368, (N.D. Tex. Sept. 26, 2012); Parker v. Louisiana Dep’t of Pub. Safety & Corr., 2019 WL 5103811 (M.D. La. Oct. 11, 2019); Calhoun v. WA DHS Child Support Div., 2018 WL 2865315 (N.D. Tex. June 11, 2018); Pouncie v. Dlorah, Inc., 2015 WL 5178401 (N.D. Tex. Sept. 4, 2015); Cty. of Travis v. Purdue Pharma, LP, 2018 WL 1518848 (W.D. Tex. Mar. 28, 2018). First, as has been explained, the district court misconstrued the Second Amended Complaint. Relevant to the severance factors, it erroneously found that there was no common transaction or occurrence or series of transactions or occurrences giving rise to the Plaintiffs’ claims against the Defendants. The district court compounded its error by stating that “the parties all acknowledge that the Washington Suit” had nothing to do with the settlement agreement between Defense Distributed and the State Department. This assertion is contradicted by the record. As the NJAG put it in his complaint in the Washington lawsuit, “in sum, the Government’s covert agreement to deregulate the CAD files by way of the Settlement Agreement—which culminated in the enactment of the “temporary modification” on July 27, 2018—are final agency decisions that not only failed to comply with procedural requirements, but that have far-reaching implications for national security and the safety and security of the State and people of Washington.” The district court for the Western District of Washington also stated that at the heart of its decision to grant a TRO, as requested by the NJAG, was the manner in which the State Department entered into and followed through on its settlement agreement with Defense Distributed. State v. United States Dep’t of State, 315 F. Supp. 3d 1202, 1205 (W.D. Wash. 2018). Defense Distributed’s claims against both Defendants fundamentally concern the State Department’s settlement agreement and the NJAG’s efforts after it was signed to prevent its meaningful implementation. The interference claims against the NJAG cannot be proven without initial proof that the settlement agreement was breached. These transactions and occurrences both perpetuated the violation of Plaintiffs’ First Amendment rights. The occurrences and transactions involved in Plaintiffs’ principal claims are thus logically, factually, and temporally inextricable.[22] The district court also misapplied the second factor of the severance analysis in determining that Plaintiffs’ claims do not involve common issues of law or fact. The court put the matter exactly backward, as it stated that although “many of Plaintiff’s claims” arise from alleged First Amendment violations, “this is not the case for all of Plaintiffs’ claims….” To the contrary, a court is required to determine whether “there is at least one common question of law,” not whether all of the claims in a case involve the same question of law. Acevedo v. Allsup’s Convenience Stores, Inc., 600 F.3d 516, 521 (5th Cir. 2010). Further, the district court incorrectly analyzed “judicial economy” as applied to joint severance and transfer motions: according to Rolls Royce and Liaw Su Teng, severance is inappropriate where two parties are as inextricably entwined in the litigation as these defendants. The NJAG’s and district court’s other assertions about the efficiency benefits from severance ring particularly hollow in light of the course of the litigation. For example, bifurcation and transfer of the case, the district court states, would be an economical act since there is already a case in New Jersey dealing with separate claims against the NJAG by Defense Distributed. But Defense Distributed brought that case as a self-protective measure only after the district court below had (erroneously) dismissed the NJAG from suit in the Western District of Texas. The district court cannot claim judicial efficiency is furthered when its decision directly caused the inefficient filing of that second suit. Nor is the timing of the district court’s severance/transfer order persuasive, as it followed directly on the heels of this court’s decision upholding jurisdiction in the Western District of Texas. Judicial economy is also less than credibly asserted by the NJAG, who chose to travel to the West Coast to pursue its campaign against Defense Distributed while simultaneously defending this case and pursuing litigation and other measures in New Jersey. The multiplication of venues has disserved judicial economy. At this point, after essentially seven years of litigation in Texas and at least two other venues over Defense Distributed’s publications, to split this case into two parallel litigation tracks before two courts is beyond inefficient.[23] The above discussion also shows why the court erred in regard to the final relevant factor involved in its severance analysis: finding no prejudice to Defense Distributed by splitting its case.[24] See Acevedo, 600 F.3d at 521. The court’s legal and factual errors so permeate its severance order as to render it a clear abuse of discretion. Transfer Order Because the severance order was a clear abuse of discretion, the district court likewise “lacked authority . . . to transfer a portion of the single action . . . for one purpose while retaining jurisdiction over the remainder.” Chrysler Credit Corp., 928 F.2d at 1519; In re Nine Mile Ltd., 673 F.2d at 244. But even setting aside severance as impermissible, any transfer of this case, in whole or in part, constitutes an abuse of the district court’s discretion. A party seeking a transfer under Section 1404(a) “must show good cause” by “clearly demonstrat[ing] that a transfer is ‘[f]or the convenience of parties and witnesses, in the interest of justice.’” In re Volkswagen of Am., Inc., 545 F.3d 304, 315 (5th Cir. 2008) (quoting 28 U.S.C. § 1404(a)) (emphasis added). “When the transferee venue is not clearly more convenient than the venue chosen by the Plaintiff, the Plaintiff’s choice should be respected.” Id. When a defendant is haled into court, some inconvenience is expected and acceptable. Assuming that jurisdiction exists and venue is proper, the fact that litigating would be more convenient for the defendant elsewhere is not enough to justify transfer. In other words, the standard is not met by showing one forum is more likely than not to be more convenient, but instead the party must adduce evidence and arguments that clearly establish good cause for transfer based on convenience and justice. See e.g. id. 314-315. Courts are required to assess four private interest factors and four public interest factors pertinent to a transfer motion. See In re Volkswagen, 545 F.3d at 315. The private interest factors are: “(1) the relative ease of access to sources of proof; (2) the availability of compulsory process to secure the attendance of witnesses; (3) the cost of attendance for willing witnesses; and (4) all other practical problems that make trial of a case easy, expeditious and inexpensive.” Id. The public interest factors are discussed below. As with its discussion of Rule 21 severance, the district court abused its discretion factually and legally in weighing the transfer factors against the standard. Regarding the first private interest factor, the movant has the burden to establish good cause, which requires an actual showing of the existence of relevant sources of proof, not merely an expression that some sources likely exist in the prospective forum. In re Volkswagen, 545 F.3d at 315. Here, however, the district court erred by uncritically accepting the NJAG’s conclusory assertions that “the sources of proof relevant to these issues (including any non-party witnesses) are all in New Jersey.” Defense Distributed, 1:18-cv-00637, Dkt. 121, at 15–16 (W.D. Tex.); See e.g. Hammers v. Mayea-Chang, 2019 WL 6728446, at *7 (E.D. Tex. Dec. 11, 2019). The NJAG’s conclusory assertions lack that necessary proof, while the Plaintiffs identified proof, documents, and witnesses that are located in Texas and support maintaining Texas as the forum.[25] Weighing the first factor as “neutral” in the face of the NJAG’s lack of proof and Plaintiffs’ proffer abused both logic and the court’s discretion. Additionally, the district court legally erred by introducing a “prejudice” consideration into the first factor. The first private interest factor does not ask whether a transfer would “prejudice” the non-moving party, but which forum provides easier access to sources of proof. Volkswagen, 545 F.3d at 315. Access to compulsory process for non-party witnesses is the gravamen of the second private interest factor. See e.g. Garrett v. Hanson, 429 F. Supp. 3d 311, 318 (E.D. Tex. 2019). Neither the NJAG nor the district court identified any witness who is subject to compulsory process in New Jersey but not in Texas. Yet the court inexplicably weighed this factor not as neutral but in favor of transfer. “To show good cause[, however] means that a moving party . . . clearly demonstrates” the appropriateness of transfer. Where there is no demonstration by the movant, let alone a clear one, the court cannot weigh a factor against the non-movant and in favor of transfer. Besides lacking support in the record, the district court’s weighing of this factor in favor of New Jersey does not “reflect[] the appropriate deference to which the Plaintiff’s choice of venue is entitled,” In re Volkswagen, 545 F.3d at 315.[26] The fourth private interest factor addresses “all other practical problems that make trial of a case easy, expeditious and inexpensive.” In re Volkswagen, at 315. The district court weighed this factor in favor of the NJAG because absent transfer, it would have to decide whether it has personal jurisdiction over the NJAG to adjudicate Defense Distributed’s allegedly newly raised claim challenging the constitutionality of NJ Stat. § 2C:39-9(l)(2). But as we discussed above that issue is not so novel as to require a separate analysis of personal jurisdiction. The district court itself previously recognized that “the instant action concern[ed]” The NJAG’s “ criminal enforcement actions,” such as “threatening to enforce a criminal law against Defense Distributed.” 364 F.Supp. 3d, at 686.[27] The public interest factors bearing on transfer are: “(1) the administrative difficulties flowing from court congestion; (2) the local interest in having localized interests decided at home; (3) the familiarity of the forum with the law that will govern the case; and (4) the avoidance of unnecessary problems of conflict of laws [or in] the application of foreign law.” See In re Volkswagen, at 315. No party disputes that the first factor is neutral. As an initial matter, the district court erroneously treated the second and the third factors together. The district court reasoned that because some of Plaintiffs’ claims against the NJAG implicate New Jersey’s criminal law, § 2C:39-9(l)(2), the District of New Jersey has a greater interest in testing the constitutionality of that statute and is better equipped than Texas courts to evaluate it. The two factors, however, are distinct in the law for good reason. The second public interest factor, which focuses on the local interest in having localized interests decided at home, “most notably regards not merely the parties’ significant connections to each forum writ large, but rather the significant connections between a particular venue and the events that gave rise to a suit.” In re Apple Inc., 979 F.3d 1332, 1345 (Fed. Cir. 2020). Important considerations include the location of the injury, witnesses, and the Plaintiff’s residence. See Volkswagen, 545 F.3d at 317–18; Zurich Am. Ins. Co. v. Tejas Concrete & Materials Inc., 982 F. Supp. 2d 714, 727 (W.D. Tex. 2013). Indeed, “[t]he place of the alleged wrong is one of the most important factors in venue determinations.” Watson v. Fieldwood Energy Offshore, LLC, 181 F. Supp. 3d 402, 412 (S.D. Tex. 2016). Texas’s “local interest in having [the] localized interests” this case implicates “decided at home” cannot be overstated. Volkswagen, 545 F.3d at 315. The controversy over New Jersey’s statute is not “localized” to New Jersey. The AG has “projected himself across state lines and asserted a pseudo-national executive authority” in Texas by seeking “to bar Defense Distributed from publishing its materials anywhere,” chilling its speech, and reducing “Texans’ access to [its] materials.” Grewal, 971 F.3d at 492-95. In these circumstances, the aggressor state’s interest is considerably diminished because the Texas court’s ruling will have no direct effect on New Jersey’s citizens. If § 2C:39-9(l)(2) were declared unconstitutional in this litigation, that ruling would preclude the NJAG’s enforcing the statute against these Plaintiffs. At least the NJAG would not necessarily be prevented from enforcing the law in New Jersey. Thus, the strength of New Jersey’s interest in having this case decided at home is considerably less than that of Texas citizens whose primary recourse, as targets of this litigation in Texas, is a suit in Texas. And Texas courts have a significant interest in assessing the constitutionality and extraterritorial impact of New Jersey’s criminal law[28], especially a law that criminalizes speech. The familiarity of the forum with the law that will govern the case was also improperly weighed in favor of New Jersey. “Federal judges routinely apply the law of a State other than the State in which they sit….” This court is “not aware of any exceptionally arcane features of Texas, [New Jersey, or constitutional law,] that are likely to defy comprehension by a federal judge sitting in [Texas].” Atl. Marine Const. Co. v. U.S. Dist. Ct. for W. Dist. of Texas, 571 U.S. 49, 67–68, 134 S. Ct. 568, 584 (2013). Moreover, New Jersey courts’ familiarity with New Jersey law produces no meaningful efficiency rendering New Jersey a more convenient forum. Defense Distributed pursues no claims arising under New Jersey law. Plaintiffs’ challenge to Section 2C:39-9(l)(2) is founded in the First Amendment and the extreme breadth of this criminal law, which purports to render illegal any speech that reaches the state from other states. Plaintiffs lodge other claims against the NJAG predicated on other federal constitutional provisions. Thus, the Texas court’s superior familiarity with Defense Distributed’s Texas law claims, and the fact that the New Jersey court would be bound to Texas law concerning such claims, see Country Chrysler, 928 F.2d at 1516, more than offsets any efficiencies that might be gained from New Jersey courts’ familiarity with New Jersey law. The last public interest factor seeks to avoid “unnecessary problems of conflict of laws [or in] the application of foreign law.” Volkswagen, 545 F.3d at 315. There are no risks of such a conflict here. Defense Distributed brought primarily federal law claims; and the state law claims involve Texas law, not New Jersey law. The district court repeatedly weighed the transfer factors against Plaintiffs by asserting that they could not be disadvantaged by transfer because they had “voluntarily” instituted additional litigation in New Jersey covering similar issues against the NJAG. This is incredible. Plaintiffs filed suit in New Jersey only because this district court had erroneously closed the door to their suit in Western District of Texas. The currently stayed litigation in New Jersey has no legitimate bearing against Plaintiffs’ original choice of forum. Correctly assessed, the NJAG did not carry its burden to clearly demonstrate that transfer is clearly more appropriate than the Plaintiffs’ choice of forum. The district court erred legally and factually in virtually every aspect of this issue, and its decision, which has unnecessarily lengthened this litigation even more, represents a clear abuse of discretion for which mandamus is an appropriate remedy. CONCLUSION For the foregoing reasons, we conclude that the district court’s order severing and transferring of the claims against the NJAG to the District of New Jersey was a clear abuse of discretion giving rise to an appropriate exercise of the court’s mandamus power. A writ of mandamus shall issue herein directing the district court to: Vacate its order dated April 19, 2021 that severed Defense Distributed’s claims against the NJAG and transferred them to the United States District Court for the District of New Jersey; Request the District of New Jersey to return the transferred case to the Western District of Texas, Austin Division; and, After return, to reconsolidate Defense Distributed’s case against the NJAG back into the case still pending against the State Department. The petition for writ of mandamus is hereby GRANTED. STEPHEN A. HIGGINSON, Circuit Judge, dissenting: “Countless expressions can be found in the jurisprudence to support the black-letter proposition that mandamus is an extraordinary remedy for extraordinary causes.” United States v. Denson, 603 F.2d 1143, 1146 (5th Cir. 1979) (Rubin, J.). The Supreme Court cautioned long ago that “[m]andamus, prohibition and injunction against judges are drastic and extraordinary remedies . . . reserved for really extraordinary causes.” Ex parte Fahey, 332 U.S. 258, 259-60 (1947). As we have said, In recognition of the extraordinary nature of the writ, we require more than showing that the court misinterpreted the law, misapplied it to the facts, or otherwise engaged in an abuse of discretion. And even reversible error by itself is not enough to obtain mandamus. Rather, we limit mandamus to only “clear abuses of discretion that produce patently erroneous results.” In re Lloyd’s Register N. Am., Inc., 780 F.3d 283, 290 (5th Cir. 2015) (quoting In re Volkswagen of Am., Inc., 545 F.3d 304, 310 (5th Cir. 2008)). Today, the majority uses that extraordinary remedy to reverse a district court’s discretionary transfer order, ruling that it was a clear abuse of discretion.[29] This district court’s interest-of-justice grant of transfer is not extraordinary cause, and compelling it to be undone is our indisputable error, for many reasons. First, Plaintiffs-Appellants Defense Distributed, the Second Amendment Foundation and Conn Williamson (hereafter DDSA) never properly requested the mandamus relief the majority has issued. Compare FED. R. APP. P. 21(a), with Notice of Appeal, Defense Distributed, No. 21-50327 (5th Cir. Apr. 23, 2021), and Brief of Appellants, Defense Distributed, No. 21-50327 (5th Cir. June 17, 2021). And as DDSA acknowledged in oral argument, they could not point to any reviewing court that has ever, before today, deployed mandamus authority to compel a trial court to undo a Section 1404(a) discretionary, interest-of-justice, joint severance and transfer grant. Oral Argument at 4:48, Defense Distributed, No. 21-50327 (5th Cir. Aug. 3, 2021).[30] A second obstacle is that our court has no authority to compel a federal district court in New Jersey to return a case it received and docketed on April 20, 2021, above all when, over seven weeks later, unopposed by DDSA, that district court took the additional step of consolidating the case it received with its pre-existing, parallel litigation, which DDSA itself filed in 2019. See In re Red Barn Motors, Inc., 794 F.3d 481, 484-85 (5th Cir. 2015). Here is the pertinent chronology: After the district court granted the motion to transfer, DDSA did not, in district court, seek a stay or reconsideration. Nor did DDSA immediately seek mandamus of the district court’s transfer order, see Notice of Appeal, Defense Distributed v. Grewal, No. 21-50327 (5th Cir. Apr. 23, 2021), and, in its initial filings in this court, DDSA explicitly disclaimed intention to “enjoin anything about the interlocutory transfer or severance rulings below,” Appellants’ Reply in Support of Appellants’ Motion for a Preliminary Injunction, Defense Distributed, No. 21-50327 (5th Cir. June 21, 2021). DDSA did not ask the New Jersey district court to return the case to the Western District of Texas, nor did they oppose consolidation with the pre-existing New Jersey case. Hearing no opposition from DDSA, the New Jersey district court then consolidated the cases in that court. That was over half a year ago. Third, even if all we order of our district court is a request to transfer the case back to Texas—so that we can resolve a constitutional showdown between New Jersey law and 3D-printed weaponry—and assuming our court can sua sponte reconfigure an appellant’s notice of appeal into a non- conforming mandamus petition, still, here, the district judge painstakingly applied the Volkswagen transfer factors, which we set forth with the (short- lived) promise (“we stress”) that “in no case will we replace a district court’s discretion with our own.” 545 F.3d at 312. The district court’s Volkswagen analysis is so comprehensive and well-reasoned as to show that had transfer been refused, a writ for mandamus to compel transfer would be well supported by the settled precedent of this court, see Lloyds, 780 F.3d 283, In re Rolls Royce Corp., 775 F.3d 671 (5th Cir. 2014); In re Radmax, Ltd., 729 F.3d 285 (5th Cir. 2013); Volkswagen, 545 F.3d 304, which other courts follow and scholars credit, see In re McGraw-Hill Glob. Educ. Holdings LLC, 909 F.3d 48, 57 (3d Cir. 2018) (citing Lloyds); In re Hudson, 710 F.3d 716, 717-18 (7th Cir. 2013) (Posner, J.) (citing Volkswagen); In re Apple, Inc., 602 F.3d 909, 911-12 (8th Cir. 2010) (citing Volkswagen); In re: Apple, No. 2021-181, 2021 WL 5291804, at *2 (Fed. Cir. Nov. 15, 2021); In re: Hulu, LLC, No. 2021-142, 2021 WL 3278194, at *2 (Fed. Cir. Aug. 2, 2021); 55 JOHN BOURDEAU ET AL., C.J.S. MANDAMUS § 97 (2021). Notably, severance is not only discretionary and tied to case management prerogatives, but it is not itself subject to review through a writ of mandamus. At all. See In re Rolls Royce, 775 F.3d at 676. And though Rolls Royce held that a district court’s ruling on a joint transfer and severance motion may be reviewed through mandamus, neither DDSA nor the majority opinion cites any mandamus case, ever, reversing a grant of a joint severance and transfer motion. Constructed for DDSA, the novel argument — opposite to indisputable — appears to be that New Jersey is a necessary party because the United States Department of State somehow was its pull-toy, allegedly breaching a settlement agreement whose provisions plaintiffs could not specify at oral argument as having been breached, much less that New Jersey caused the federal government to breach. Manifestly, New Jersey has no role in United States weapons export regulations. Several additional points highlight our usurpation. First, the district court here transferred to a court whose intercession these same plaintiffs had already sought, had argued on appeal in the Third Circuit to maintain, and thereafter had kept pending for over two years. In other words, the transfer was to a receiving court which had pre-existing jurisdiction over these plaintiffs’ claims, at their initiation and insistence. Second, that transfer was back to a court within the only federal circuit authorized to certify questions of interpretation of New Jersey law to the Supreme Court of New Jersey, which could interpret that law to avoid, not force, a novel and difficult constitutional showdown. See N.J. COURT RULES 2:12A-1. Notably, the Supreme Court recently reversed our court when we tried to force resolution of a different First Amendment collision without deferring first to one of our states’ highest court’s prerogative to interpret its own law. McKesson v. Doe, 141 S. Ct. 48 (2020).[31] A third related, especially serious point is that removing discretion to grant transfer undermines our court’s mandamus transfer precedent, which, hitherto, we have used to require comity, not defy it. And it does so imprudently, forcing a constitutional collision that may well be avoidable through certification of questions of state law, which we cannot do but the receiving circuit can. Hereafter, as to other transfer grants, our mandamus, discordantly, will reduce or remove trial judge discretion to avoid adjudicating the constitutionality of other states’ laws. Finally, we commit not just an error we told our district courts in Volkswagen would not occur, but another one we took pains earlier in this litigation to cabin against. See Defense Distributed v. Grewal, 971 F.3d 485, 496 (5th Cir. 2020). Hereafter, state attorneys general, including those within our own circuit, may not only be hailed into federal courts across the country to defend their state laws, but then, if a trial judge assesses interests of justice to favor transfer—say for consolidation with related claims already pending in a co-equal court whose circuit has certification authority—that is indisputable error.[32] Consistent with the New Jersey Attorney General’s express and emphatic denials, twice to our panel, that it seeks to enforce New Jersey law in Texas, the record is devoid of enforcement measures, such as a takedown order or criminal prosecution, against DDSA. Indeed, the parties point to no prosecution even in New Jersey using New Jersey’s new law, yet the majority rules today that mandamus lies to force New Jersey to defend its new law for the first time in Texas. Because this precedent, now binding as indisputable, does injury to state sovereignty and comity, not to mention district court case management discretion, it is small comfort that there is a pending dismissal motion filed by the Department of State which, if granted, may lead to re- transfer, albeit after further, up-down delay protracted by us. Also specific to this case, it is small comfort that the most we compel is a request, which the New Jersey federal district court no doubt will seriously consider, informed by comity and state law priorities set forth by the Supreme Court, as well as by its own assessment of whether litigation resolving New Jersey law should be decided in Texas as well as delayed and encumbered by the national security interests implicated in the federal government’s overseas export munitions restrictions, which are at the heart of the litigation in the Western District of Texas. Mandamus rulings announce law inflexibly. Here, without precedent, our court, seemingly impatient for the last half decade, to force a difficult First (Second?) Amendment clash over government regulation of 3D printable “ghost” weapons, supplants district courts’ long-standing, fact- specific case management discretion to transfer litigation for consolidation with an existing case in another circuit, in derogation of state sovereignty, comity and constitutional avoidance principles, contrary to instruction given to us by the Supreme Court just last year.
