18 F.4th 806

5th Cir.

2021

George RAYMOND WILLIAMS, MEDICAL DOCTOR, ORTHOPAEDIC SURGERY, A PROFESSIONAL MEDICAL, L.L.C. v. HOMELAND INSURANCE COMPANY OF NEW YORK; MED-COMP USA, INCORPORATED

No. 20-30196

United States Court of Appeals for the Fifth Circuit

November 30, 2021

Before JONES, HAYNES, and HO,¹ Circuit Judges.

HAYNES, Circuit Judge:

A class of Louisiana medical providers (the “Class“) sued a number of Louisiana Preferred Provider Organizations (the “PPOs“) in Louisiana state court on state law causes of action approximately ten years ago. The Class amended its decade-old state lawsuit to assert newly assigned bad faith insurance claims against Homeland Insurance Company, an out-of-state defendant. After Homeland removed the case to federal court, the district court concluded the claims were barred by an earlier Delaware judgment and dismissed the suit. The Class contends that the district court lacked jurisdiction because a non-diverse defendant remained in the case from the original lawsuit, and, alternatively, that the court erred in granting preclusive effect based on the statute of limitations dismissal in Delaware. We agree that the district court lacked jurisdiction. Accordingly, the district court‘s jurisdictional holding is REVERSED, the district court‘s judgment and other orders are VACATED, and the case is REMANDED to the district court with instructions to remand the entire case to state court.

I. BACKGROUND

With a long procedural history, this case returns on appeal for the third time.² In 2009, Plaintiff George Raymond Williams filed a putative class action in Louisiana state court on behalf of the Class against three Louisiana defendants, including Med-Comp USA. Med-Comp operated a PPO network, which contracted with the proposed class of medical providers for discounted rates. The Class alleged that the PPOs violated the Louisiana Preferred Provider Organization Act (“PPO Act“) by discounting the Class‘s bills without prior notice. See LA. STAT. ANN. § 40:2203.1(G). In 2011, the Class amended its complaint to add claims against additional non-Louisiana defendants: CorVel Corporation, Homeland Insurance, and Executive Risk Specialty Insurance.

Executive Risk and Homeland removed the case to federal court, asserting both ordinary diversity jurisdiction under 28 U.S.C. § 1332(a) and jurisdiction under the Class Action Fairness Act‘s (“CAFA“) minimal diversity jurisdiction under 28 U.S.C. § 1332(d)(2). CorVel and the Class sought to remand pursuant to CAFA‘s local controversy exception. 28 U.S.C. § 1332(d)(4). The district court found the criteria of the exception satisfied and remanded the case to Louisiana state court. Homeland appealed, but we affirmed the remand order. Williams I, 657 F.3d at 293.

On remand, after receiving class certification, the Class settled with Executive Risk and all of the Louisiana defendants except Med-Comp. The Class also prevailed on its direct-action claims against Homeland in state court—leaving Med-Comp as the sole remaining defendant. See Williams v. SIF Consultants of La., Inc., 209 So. 3d 903, 906, 912 (La. Ct. App. 2016) (affirming summary judgment in favor of the Class on its claims against Homeland).

As part of their settlement agreement, CorVel (Homeland‘s insured) assigned to the Class its insurance coverage claims against Homeland. This assignment underlies the present dispute. The assignment, however, did not initially include the bad faith claim CorVel was pursuing against Homeland in Delaware state court.³ See Homeland Ins. Co. of N.Y. v. CorVel Corp., 197 A.3d 1042 (Del. 2018). Although CorVel initially prevailed against Homeland, the Delaware Supreme Court reversed, holding that the claim was barred by the statute of limitations. Id. at 1044. CorVel then assigned the entirety of its claims against Homeland to the Class.

The Class then amended its still unresolved complaint against Med-Comp⁴ in Louisiana state court to assert the bad faith claim against Homeland that CorVel had just lost on in Delaware. The litigation then consisted of the Class bringing state law PPO Act claims as a class against one non-diverse defendant (Med-Comp) and a state law bad faith insurance claim as an assignee against one diverse defendant (Homeland).

After being joined again in state court, Homeland removed the case to federal court on the basis of ordinary diversity jurisdiction and CAFA minimal diversity jurisdiction. Homeland justified removal by arguing that Med-Comp‘s non-diverse Louisiana citizenship could be disregarded because the PPO Act claims against Med-Comp were “improperly and egregiously misjoined” with the assignment-based bad faith claim against Homeland. The Class moved to remand.

Adopting a magistrate judge‘s report and recommendation, the district court denied the motion to remand in part and granted it in part. The district court retained jurisdiction over the bad faith claim against Homeland and remanded the PPO Act claims against Med-Comp to state court.⁵

Bypassing CAFA, the district court reasoned that it had ordinary diversity jurisdiction over the case because Homeland was diverse from the Class and because Med-Comp, the non-diverse defendant, was improperly joined. Without addressing the viability of the Class‘s PPO Act claims against Med-Comp, the district court reasoned that improper joinder applied because, in its capacity as CorVel‘s assignee, the Class lacked a reasonable basis for recovery against Med-Comp.

The Class moved for leave to appeal under CAFA‘s interlocutory appeal mechanism. 28 U.S.C. § 1453(c)(1); see Williams II, 788 F. App‘x at 298. A prior panel of this court denied the Class‘s motion. We held that the Class did not raise a CAFA issue because the district court based its jurisdictional holding entirely on ordinary diversity jurisdiction. Id. at 298-99.

Homeland then moved to dismiss the bad faith claim, which the Delaware Supreme Court had resolved in its favor based on the statute of limitations. The district court approved a magistrate judge‘s report and recommendation concerning the motion and held that under the Full Faith and Credit Act, 28 U.S.C. § 1738, the court must look to Delaware law for applicable preclusion principles. Concluding that Delaware would apply claim preclusion to a case that had been dismissed for failure to file within the statute of limitations, the court dismissed. The Class timely appealed.

II. DISCUSSION

The district court concluded that it had ordinary diversity jurisdiction under 28 U.S.C. § 1332(a)(1). We have jurisdiction to evaluate whether the district court was correct. See In re Transtexas Gas Corp., 303 F.3d 571, 576-77 (5th Cir. 2002). “We review questions of subject matter jurisdiction de novo.” Wagner v. United States, 545 F.3d 298, 300 (5th Cir. 2008) (quotation omitted). When analyzing whether the proponent of jurisdiction has established improper joinder, we examine all factual allegations and ambiguities of state law in the light most favorable to the party resisting jurisdiction. See Campbell v. Stone Ins., Inc., 509 F.3d 665, 669 (5th Cir. 2007).

The Class asserts that jurisdiction in federal district court was improper because Med-Comp‘s presence destroyed complete diversity of the parties and barred removal.⁶ The Class alternatively contends that the district court erred in dismissing the bad faith claim on claim preclusion grounds. Because we conclude that the district court lacked jurisdiction in this case, we do not reach the merits of the claim preclusion issue.

A. Jurisdiction

A defendant can remove “any civil action brought in a State court of which the district courts of the United States have original jurisdiction.” 28 U.S.C. § 1441(a). Two types of original jurisdiction are at issue here: (1) ordinary diversity jurisdiction under 28 U.S.C. § 1332(a)(1) and (2) CAFA minimal diversity jurisdiction under 28 U.S.C. § 1332(d).⁷ We discuss both below.

1. Ordinary Diversity Jurisdiction & Improper Joinder Doctrine

A case exhibits diversity of citizenship when the parties are citizens of different states and the amount in controversy exceeds $75,000. 28 U.S.C. § 1332(a)(1). But removal based on diversity is precluded “if any of the parties in interest properly joined and served as defendants is a citizen of the State in which such action is brought.” § 1441(b)(2). Moreover, ordinary diversity jurisdiction requires complete diversity—all of the plaintiffs must be citizens of different states than all of the defendants. Flagg v. Stryker Corp., 819 F.3d 132, 136 (5th Cir. 2016) (en banc). If a non-diverse defendant is improperly joined, however, a district court can disregard the citizenship of that defendant for the purposes of evaluating its jurisdiction. Id. Our precedents in this area are not based upon the text of the statute but rather on the concept that fake or fraudulent joinder of a party to defeat federal jurisdiction should not be approved.

The district court purported to apply our established improper joinder doctrine—namely, that a defendant is improperly joined if there is no possibility of recovery against it—but its reasoning and conclusions cannot be squared with this circuit‘s improper joinder jurisprudence. Rather, the district court functionally applied the fraudulent misjoinder doctrine, which we have never adopted and do not adopt now.

i. Traditional Improper Joinder Doctrine

In this circuit, a non-diverse defendant is improperly joined such that its citizenship can be ignored for purposes of evaluating diversity jurisdiction if the removing party shows either that: (1) there was actual fraud in the pleading of jurisdictional facts; or (2) the plaintiff is unable to establish a cause of action against the non-diverse defendant in state court. Smallwood v. Ill. Cent. R.R. Co., 385 F.3d 568, 573 (5th Cir. 2004) (en banc).⁸ In other words, committing fraud in the facts or faking a defendant is not a basis for denying diversity jurisdiction. That said, it is the removing party‘s burden to establish improper joinder. See Id. at 574. Because Homeland does not allege that there was actual fraud in the pleading of jurisdictional facts, we assess only whether Homeland has demonstrated that the Class lacks a viable state law cause of action against the non-diverse defendant, Med-Comp. We typically evaluate this question by evaluating whether the plaintiff had any possibility of recovery against the non-diverse defendant in state court at the time of removal. See Flagg, 819 F.3d at 136. The inquiry is virtually identical to the inquiry on a motion to dismiss for failure to state a claim: “Ordinarily, if a plaintiff can survive a Rule 12(b)(6) challenge, there is no improper joinder.” Id. (quoting Smallwood, 385 F.3d at 573); see also Int‘l Energy Ventures Mgmt., L.L.C. v. United Energy Grp., 818 F.3d 193, 208 (5th Cir. 2016). In limited circumstances, however, we permit district courts to perform a more detailed factual analysis: if it appears that the plaintiff has misstated or omitted discrete facts that would determine the propriety of joinder, a district court can, in its discretion, pierce the pleadings and conduct a limited “summary inquiry.” Flagg, 819 F.3d at 136.

Here, the district court concluded that the Class lacked any possibility of recovery against Med-Comp. But the district court did not actually assess whether the pleadings or other facts showed that the Class lacked a possibility of recovery on its PPO Act claims against Med-Comp. Instead, it proceeded down a new path of reviewing the Class in two different capacities in the case: (1) the Class-as-the Class, who had the PPO Act claims against Med-Comp; and (2) the Class-as-CorVel, who had the assigned bad faith claim against Homeland.

The district court reasoned that because the Class‘s litigation was under two different capacities, the capacity in which the Class had a suit against Homeland, was different from the capacity against Med-Comp and, therefore, lacked a possibility of recovery. Thus, the district court concluded that Med-Comp was improperly joined such that ordinary diversity jurisdiction was appropriate. In reaching that conclusion, the only authority the district court cited was our suggestion, in an opinion denying mandamus relief, that “it might be concluded that misjoinder of plaintiffs should not be allowed to defeat diversity jurisdiction.” In re Benjamin Moore & Co., 309 F.3d 296, 298 (5th Cir. 2002) (emphasis added).⁹

The district court‘s conclusion to split the Class into two different capacities is inconsistent with this circuit‘s traditional improper joinder analysis and has absolutely no basis in the text of the relevant statutes. We have never held that a capacity-by-capacity analysis is warranted in evaluating improper joinder.¹⁰ See generally Smallwood, 385 F.3d at 573 (identifying only actual fraud and impossibility of recovery as constituting improper joinder). Indeed, we apply the traditional no-possibility-of-recovery test even when an assigned claim destroys complete diversity; so long as that claim is colorable, ordinary diversity jurisdiction is inappropriate. See Grassi v. Ciba-Geigy, Ltd., 894 F.2d 181, 182-85 (5th Cir. 1990) (collecting cases and stating that, under Supreme Court precedent, “the colorable assignment of a complete cause of action to defeat removal [is] effective to give the state court exclusive jurisdiction“).

Applying the traditional analysis here, we conclude that Med-Comp was not improperly joined because the Class has a possibility of recovery against Med-Comp (a non-diverse defendant) on the PPO Act claims. The complaint alleged that Med-Comp failed to provide the Class with benefit cards identifying the applicable contractual discount applied to the Class‘s services and failed to give thirty days written notice of contractual alternative rates of pay. Accordingly, the allegations in the complaint plausibly stated that Med-Comp violated the PPO Act. See LA. STAT. ANN. § 40:2203.1(B). Indeed, the Class secured favorable settlements on virtually identical claims against other PPOs and won summary judgment in state court on essentially the same claims against Homeland, see SIF Consultants, 209 So. 3d at 912. Viewing the Class‘s allegations in the light most favorable to it, Campbell, 509 F.3d at 669, we conclude that the Class‘s complaint provides a possibility of recovery against Med-Comp on its PPO Act claims.

Homeland does not appear to dispute that the complaint plausibly stated a claim. Rather, it argues that we should pierce the pleadings and evaluate language in a settlement agreement between the Class and CorVel, which, Homeland argues, resulted in the Class releasing some or all of its claims against Med-Comp. The district court did not engage in this analysis, but even if we construed that it implicitly did, the settlement agreement identified by Homeland does not establish that the Class has no possibility of recovery against Med-Comp. Indeed, the settlement agreement released only the Class‘s claims against CorVel.¹¹ In sum, none of the above facts, even if we were to consider them, eliminate all possibility that the Class could recover against Med-Comp. At a minimum, the Class could prevail in Louisiana state court in the face of the settlement agreement and recover notwithstanding Med-Comp‘s financial situation. Thus, Med-Comp was properly joined. There is nothing fraudulent or fake about its presence in this lawsuit.

ii. Fraudulent Misjoinder

Homeland urges us to follow the fraudulent misjoinder doctrine, which it claims is consistent with our case law. It is not. That doctrine has been adopted only by the Eleventh Circuit, see Tapscott v. MS Dealer Serv. Corp., 77 F.3d 1353, 1360 (11th Cir. 1996), abrogated on other grounds by Cohen v. Office Depot, Inc., 204 F.3d 1069 (11th Cir. 2000), and a number of district courts. See generally 14C CHARLES ALAN WRIGHT & ARTHUR R. MILLER, FEDERAL PRACTICE & PROCEDURE § 3723.1 & nn.27, 28 (Rev. 4th ed. 2020) (describing fraudulent misjoinder as a “newer doctrine” related to improper joinder and identifying the split in authority on whether to adopt it). In Tapscott, the Eleventh Circuit extended the improper joinder doctrine to procedural questions concerning party joinder, holding that the “egregious” misjoinder of parties “constitute[s] fraudulent joinder” that permits a defendant to remove to federal court. 77 F.3d at 1360. Homeland argues that we should do the same and conclude that the Class‘s purportedly unrelated—but undisputedly viable—claims against the two defendants justify removal. Although we have cited Tapscott without disapproving it,¹² this circuit has not adopted it or any other form of the fraudulent misjoinder doctrine, and we refuse the invitation to do so now.

First, and most importantly, § 1441(b)(2) does not provide a textual basis for expanding federal jurisdiction to include fraudulent misjoinder, nor is there any other statutory authority (or case law interpreting statutory authority in this context) that does so. That statute, which does not allow removal “if any of the parties in interest properly joined and served as defendants is a citizen of the State in which such action is brought,” does not reach the question of “properly joined and served” unless the “civil action [is] otherwise removable solely on the basis of the jurisdiction under section 1332(a).” Thus, it does not support the improper joinder argument Homeland asserts.

Indeed, Homeland‘s requested expansion would invite district courts to evaluate procedural questions regarding misjoinder that are better resolved in state courts prior to removal. Put another way, if there is a possibility of recovery against both defendants but one defendant believes the case should be severed, there is nothing to prevent seeking that severance in state court. If it is granted, then the removal would be straightforward; if not, then clearly not appropriate. Additionally, the sought expansion is completely inconsistent with our emphasis that improper joinder is a “narrow exception” to § 1332(a)‘s complete diversity requirement. Campbell, 509 F.3d at 669 (quotation omitted). Homeland‘s approach would greatly broaden it.

Second, expanding this circuit‘s improper joinder jurisprudence to include fraudulent misjoinder is foreclosed by precedent. Our court has gone en banc twice on precisely what is needed to remove a case from state to federal court on the basis of diversity jurisdiction notwithstanding a lack of complete diversity between the parties. Flagg, 819 F.3d at 137; Smallwood, 385 F.3d at 573. As articulated above, our established approach is straightforward. Our case law emphasizes substantive viability—not procedural questions like party joinder. Smallwood, 385 F.3d at 575. We have never held that a defendant can remove a case based on party joinder issues in state court. Indeed, we have directed that any viable cause of action against a diversity-destroying party requires the entire case to be remanded: “the existence of even a single valid cause of action against in-state defendants (despite the pleading of several unavailing claims) requires remand of the entire case to state court.” Gray ex rel. Rudd v. Beverley Enters.-Miss., Inc., 390 F.3d 400, 412 (5th Cir. 2004) (emphasis added); cf. Grassi, 894 F.2d at 183-85 (recognizing that the complete assignment of a colorable claim to a non-diverse plaintiff generally destroys complete diversity). Permitting a party to remove a case and then sever a diversity-destroying defendant notwithstanding viable claims against that defendant fundamentally conflicts with that statement.

The severance mechanism for omitting a non-diverse defendant also upends our rules on removal. Our precedent makes clear that we look at jurisdiction at the time of removal, not after a federal court severance: “Crucially, jurisdictional facts are determined at the time of removal, not by subsequent events.” Flagg, 819 F.3d at 137 (quotation and brackets omitted). Thus, federal court severance does not (and cannot) create jurisdiction that otherwise would not exist. Indeed, otherwise a diverse defendant could almost always remove and then simply sever the non-diverse defendant to remain in federal court. Applying fraudulent misjoinder turns our precedent around by effectively allowing a federal district court to engage in significant procedural wrangling in order to create jurisdiction. Contra id. at 139. No Fifth Circuit precedent supports this concept.

Our conclusion not to extend improper joinder doctrine to procedural questions like party joinder is also practically sound. The response to Tapscott and its underlying reasoning has been unfavorable. No other circuit court has yet adopted the approach from that case, and it has received, at best, mixed reviews from district courts.¹³ For good reason: it‘s a usurpation of the traditional ability of state courts to resolve procedural questions central to the administration of their cases.

It is also a well-known rule that federal courts are courts of limited jurisdiction. Adopting the fraudulent misjoinder doctrine will dramatically expand federal jurisdiction, putting the federal district courts in this circuit in the position of resolving procedural matters that are more appropriately resolved in state court—all without a clear statutory hook. See In re Prempro Prods. Liab. Litig., 591 F.3d 613, 621-22 (8th Cir. 2010) (summarizing courts’ criticism of Tapscott as creating an “unpredictable and complex jurisdictional rule” that contravenes the principle that federal jurisdiction be “narrowly construed“). State courts (courts of general jurisdiction) are certainly able to address these issues, and respecting state court resolution of state law issues is a bedrock principle of our federal system. See, e.g., Davis v. Mich. Dep‘t of Treas., 489 U.S. 803, 818 (1989) (noting that state courts have “special expertise” on state law and “are in the best position” to adjudicate such issues); Erie R.R. Co. v. Tompkins, 304 U.S. 64, 78-79 (1938) (describing the dangers of federal courts failing to respect the function of state courts). Accordingly, we will not usurp a state court‘s ability to decide whether parties have been misjoined under a state court‘s own procedural rules. See Felder v. Casey, 487 U.S. 131, 138 (1988). Indeed, if the Class‘s claims against Med-Comp really were misjoined, Homeland could have (and should have) moved to sever the claims in state court and then removed.¹⁴

Moreover, the fraudulent misjoinder doctrine poses a challenging question as to the correct source of misjoinder law—federal or state. Some courts that have adopted the doctrine determine whether a party has been misjoined by looking to the party joinder framework of Federal Rule of Civil Procedure 20, see, e.g., Goodwin v. Kojian, No. SACV 13-325-JST (JPRx), 2013 WL 1528966, at *3 (C.D. Cal. Apr. 12, 2013), but doing so puts the cart before the horse: as with all federal rules, Rule 20 applies after a federal court has jurisdiction, FED. R. CIV. P. 81(c)(1). So it would be odd to use the impropriety of joinder under those rules to establish jurisdiction. See FED. R. CIV. P. 82 (“These rules do not extend . . . the jurisdiction of the district court.“). Other courts look to state procedural rules, but doing so conflicts with Rule 81‘s direction that federal procedural rules apply after removal. See Centaurus Unity v. Lexington Ins. Co., 766 F. Supp. 2d 780, 789 n.27 (S.D. Tex. 2011) (noting that most courts look to state court joinder rules). Our holding that misjoinder issues should be addressed in state court prior to removal resolves these inconsistencies.

Declining to adopt the fraudulent misjoinder doctrine as a basis for exercising diversity jurisdiction and holding that misjoinder issues should be addressed in state court prior to removal also brings greater clarity to the line between removable and non-removable cases. A defendant would have a clear framework for determining the removability of a case involving plausible but potentially misjoined claims between non-diverse parties. The defendant would move to sever those claims on misjoinder grounds before removal, and either the state court grants the motion and the case becomes removable or it denies the motion and the case remains non-removable. See Crockett v. R.J. Reynolds Tobacco Co., 436 F.3d 529, 532-33 (5th Cir. 2006) (recognizing that state court severance to address misjoinder can make a case removable notwithstanding the general rule that cases only become removable based on the voluntary acts of the plaintiff).

2. CAFA Minimal Diversity Jurisdiction

Homeland alternatively argues that the district court appropriately exercised jurisdiction because of CAFA minimal diversity jurisdiction. This argument, however, is barred by the law of the case doctrine. Under that doctrine, we will “follow the prior decisions in a case as the law of that case.” Reeves v. AcroMed Corp., 103 F.3d 442, 448 (5th Cir. 1997) (quotation omitted). A subsequent panel will reexamine issues of law resolved by a prior panel opinion only if: “(i) the evidence on a subsequent trial was substantially different, (ii) controlling authority has since made a contrary decision of the law applicable to such issues, or (iii) the decision was clearly erroneous and would work a manifest injustice.” Id. (quotation omitted).

Here, a prior panel of this court concluded that the district court‘s remand order was not based on CAFA minimal diversity jurisdiction. See Williams II, 788 F. App‘x at 298. None of the three exceptions to the law of the case doctrine apply here. See Reeves, 103 F.3d at 448. First, there has been no subsequent evidence on the point: the district court has not, for example, reconsidered its order and concluded that it had jurisdiction under CAFA. See id. Second, there has been no change in controlling authority. See id. Third, the prior panel‘s decision was not clearly erroneous; the district court was clear that its decision to exercise jurisdiction was premised on ordinary diversity jurisdiction—not CAFA. See id. To address CAFA minimal diversity jurisdiction as an alternative basis for jurisdiction now would undermine the prior panel‘s conclusion, and we thus decline to do so.¹⁵

III. CONCLUSION

For the foregoing reasons, we REVERSE the district court‘s denial of the Class‘s motion to remand to state court, VACATE the district court‘s subsequent orders, and REMAND to the district court with instructions to remand the entire case to state court.

JAMES C. HO, Circuit Judge, concurring:

The question presented in this appeal is an arcane issue of federal civil procedure. But the proper answer to that question implicates a fundamental principle of judicial methodology: When faced with a conflict between text and precedent, we should maximize the former—and minimize the latter.

“[J]udges swear an oath to uphold the Constitution, consistent of course with a judicial system based on precedent. That should mean that we decide every case faithful to the text and original understanding of the Constitution, to the maximum extent permitted by a faithful reading of binding precedent.” Texas v. Rettig, 993 F.3d 408, 409 (5th Cir. 2021) (Ho, J., dissenting from denial of rehearing en banc). “So if we are forced to choose between upholding the Constitution and extending precedent in direct conflict with the Constitution, the choice should be clear: ‘Our duty is to apply the Constitution—not extend precedent.‘” Id. at 417 (quoting NLRB v. Int‘l Ass‘n of Bridge, Structural, Ornamental, & Reinforcing Iron Workers, Local 229, AFL-CIO, 974 F.3d 1106, 1116 (9th Cir. 2020) (Bumatay, J., dissenting from denial of rehearing en banc)) (cleaned up).¹

To borrow from a noted opinion by Justice Thomas, “my rule is simple“—when legal text conflicts with judicial precedent, we should follow the text, not “demonstrably erroneous precedent,” to the maximum permissible extent. Gamble v. United States, 139 S. Ct. 1960, 1984 (2019) (Thomas, J., concurring). “[F]idelity to original meaning counsels against further extension of [] suspect precedents.” Hester v. United States, 139 S. Ct. 509, 509 (2019) (Alito, J., concurring in the denial of certiorari).

And the same logic applies when judicial precedent conflicts with legal texts other than the Constitution. See, e.g., Williams, 958 F.3d at 350 (Ho, J., concurring) (same principle requires adhering to “the plain language of United States statutes to the maximum extent that Supreme Court precedent permits“) (quotations omitted).

Faithful application of these principles requires me to reverse the district court here. For that reason, I concur.

I.

Diversity jurisdiction generally extends to civil actions where the amount in controversy exceeds $75,000—but only if the suit is between “citizens of different states.” 28 U.S.C. § 1332(a)(1). It is well established that § 1332(a)(1) “require[s] complete diversity between all plaintiffs and all defendants.” Lincoln Prop. Co. v. Roche, 546 U.S. 81, 89 (2005).

This case plainly fails the complete diversity requirement. As the majority details, this case has undergone a series of procedural twists and turns. But all that matters for this appeal is this undisputed fact: Both the plaintiff and one of the defendants, Med-Comp USA, Inc., are citizens of

Louisiana, the forum state. See (en banc) (“jurisdictional facts are determined at the time of removal“) (quotations omitted). As a result, the case is not completely diverse and thus cannot be removed to federal court under 28 U.S.C. § 1441(a).

As a textual matter, that should be the end of the analysis.

II.

Homeland Insurance Company nevertheless contends that removal jurisdiction is appropriate, because Med-Comp was not properly joined and therefore cannot be used to defeat complete diversity in this case. But the only possible textual basis for Homeland‘s assertion of removal jurisdiction is 28 U.S.C. § 1441(b)(2). Homeland does not claim otherwise. And § 1441(b)(2) does not authorize removal jurisdiction here.

Under 28 U.S.C. § 1441(b)(2), “[a] civil action otherwise removable solely on the basis of the jurisdiction under section 1332(a) of this title may not be removed if any of the parties in interest properly joined and served as defendants is a citizen of the State in which such action is brought.”

This provision is commonly known as the “forum-defendant rule,” because it prevents defendants from removing diversity cases to federal court if one of the defendants resides in the forum state. In re 1994 Exxon Chemical Fire, 558 F.3d 378, 391 (5th Cir. 2009).

Notably, the forum-defendant rule applies only to local defendants who are “properly joined.” 28 U.S.C. § 1441(b)(2). So if only one of the defendants is local, but that defendant was not “properly joined,” the case is still removable under § 1441(b)(2).

But the case must still satisfy the complete diversity requirement of § 1332. And the complete diversity requirement must be met at the outset of removal—before any suggestion of improper joinder. The text of § 1441(b)(2) makes this clear: Courts must first determine whether an action is ”otherwise removable solely on the basis of [diversity jurisdiction],” and only if so do courts then apply the forum-defendant rule (and its provision concerning properly joined local defendants). 28 U.S.C. § 1441(b)(2) (emphasis added).

The net effect is this: If a case is completely diverse, and one of the defendants is local but improperly joined, the case can still be removed pursuant to § 1441(b)(2). But if a case is not completely diverse when a notice of removal is filed, then nothing in § 1441(b)(2) authorizes the exercise of federal removal jurisdiction, period. Even if the case might become completely diverse if the federal court were to dismiss an improperly joined defendant, that would not satisfy the terms of § 1441(b)(2), because the case was not “otherwise removable” at the outset of removal.

III.

Homeland does not dispute this textual analysis. Instead, it relies on a series of precedents that permit courts to ignore the citizenship of improperly joined defendants even in cases that lack complete diversity at the outset of removal. See Wecker v. Nat‘l Enameling & Stamping Co., 204 U.S. 176, 185–86 (1907); see also Smallwood v. Ill. Cent. R.R. Co., 385 F.3d 568, 572–73 (5th Cir. 2004) (en banc); Flagg, 819 F.3d at 136.

But these precedents are not based on statutory text—as both our circuit and our sister circuits have observed. See, e.g., Green v. Amerada Hess Corp., 707 F.2d 201, 206 (5th Cir. 1983) (describing fraudulent joinder as a “judicially created doctrine“); Hoyt v. Lane Constr. Corp., 927 F.3d 287, 295 (5th Cir. 2019) (same); Triggs v. John Crump Toyota, Inc., 154 F.3d 1284, 1287 (11th Cir. 1998) (same); Coyne v. Am. Tobacco Co., 183 F.3d 488, 493 (6th Cir. 1999) (same); Mayes v. Rapoport, 198 F.3d 457, 461 (4th Cir. 1999) (same).

A number of our colleagues have criticized the improper joinder doctrine for conflicting with statutory text—not to mention basic principles of federalism—calling it “a legal rule fashioned entirely from judges’ imaginations,” forcing federal courts to “exercise jurisdiction where none exists over questions of state law that the state courts are better suited to address themselves.” Murriel-Don Coal Co. v. Aspen Ins. UK Ltd., 790 F. Supp. 2d 590, 596–97 (E.D. Ky. 2011).

These critics have acknowledged, of course, that the doctrine is “well-established.” Id. at 594. “Like other well-established doctrines, though, fraudulent joinder suffers from a common problem—courts rarely stop and think about whether the doctrine makes sense.” Id. “A closer look at fraudulent joinder reveals a doctrine resting on a rickety foundation,” for “it is unclear where federal courts get the authority to decide whether a defendant has been fraudulently joined.” Id. at 594–95. “The better course is for the court to remand the case and allow the state court to dismiss the claims against the non-diverse defendants if those claims truly present no hope of success. Once the state court dismisses the non-diverse defendants from the case, the remaining diverse defendants would have thirty days to file a notice of removal under 28 U.S.C. § 1446(b).” Id. at 597. See also Flores-Duenas v. Briones, 2013 WL 6503537, *29 (D.N.M. Dec. 1, 2013) (agreeing with concerns expressed in Murriel-Don); Gaither v. Beam Partners, LLC, 2017 WL 1217166, *3–4 (E.D. Ky. Mar. 31, 2017) (same); Howard v. Pearl Interactive Network Inc., 2018 WL 3421824, *4 n.3 (E.D. Ky. July 12, 2018) (noting “extensive, foundational criticism of the fraudulent joinder doctrine,” citing Murriel-Don).

As a panel, we are of course bound to follow our circuit precedent. And that requires reading precedent faithfully. “Lower court judges don‘t have license to adopt a cramped reading of a case in order to functionally overrule it.” Int‘l Ass‘n, 974 F.3d at 1116 (Bumatay, J., dissenting from denial of rehearing en banc) (quotations omitted). See also Blackman, 13 NYU J.L. & LIBERTY at 51 (“Of course, judges can always draw razor-thin distinctions and contend that a particular issue is not governed by a non-originalist precedent. But judges should resist this temptation.“).

But our precedent has applied the improper joinder doctrine only in cases where a defendant was improperly joined as a matter of substantive law. See Flagg, 819 F.3d at 136 (“the test for improper joinder ‘is whether the defendant has demonstrated that there is no possibility of recovery by the plaintiff against an in-state defendant‘“) (quoting Smallwood, 385 F.3d at 573). None of our precedents involved a defendant who was improperly joined as a matter of procedure, as is the case here.

I see no reason why we should be compelled to extend our erroneous precedents to fit this case. This is not a case, after all, in which logic demands that we extend an atextual body of precedent in order to preserve rationality or consistency in the law. Cf. Leocal v. Ashcroft, 543 U.S. 1, 11 n.8 (2004) (courts must interpret statutes “consistently“); Clark v. Martinez, 543 U.S. 371, 380 (2005) (same).²

Moreover, courts often distinguish between matters of substance and procedure. See, e.g., Gasperini v. Ctr. for Humanities, Inc., 518 U.S. 415, 427 (1996) (“Under the Erie doctrine, federal courts sitting in diversity apply state substantive law and federal procedural law.“) (emphasis added). And there is good reason to draw such a distinction here: It is one thing to invoke improper joinder to dismiss claims that are destined to fail on the merits in any event—it is quite another to invoke improper joinder to dismiss claims that could very well be meritorious.

***

We should decide every case by adhering to the governing legal text to the maximum extent permitted by a faithful reading of binding precedent. That is what the majority does today. I accordingly concur.

EDITH H. JONES, Circuit Judge, dissenting:

A class of Louisiana medical providers (“the Class“) amended a decade-old state lawsuit to assert newly assigned bad faith insurance claims against Homeland Insurance Company, an out-of-state defendant. After Homeland removed the case to federal court, the district court concluded the claims were barred by an earlier Delaware judgment and dismissed the suit. The majority holds that the district court lacked jurisdiction because a non-diverse defendant remained from the original lawsuit. According to the majority, the sky will fall on removal jurisdiction unless this case against Homeland proceeds only in Louisiana state court. The majority‘s arguments, in my view, prove too much. I would affirm the district court‘s jurisdiction.

BACKGROUND

After various procedural twists in state court, the Class of medical providers was certified for a variety of insurance coverage claims (Class PPO Act claims) and eventually settled with CorVel and all of the Louisiana defendants except Med-Comp, which became the sole remaining defendant. As part of their settlement agreement, CorVel assigned to the Class its insurance coverage claims against Homeland. Later, this assignment was expanded to include a bad faith claim made by CorVel against Homeland that went to the Delaware Supreme Court (Class-as-CorVel assignee claim). See Homeland Ins. Co. of N.Y. v. CorVel Corp., 197 A.3d 1042 (Del. 2018).

With its newly assigned claim in hand, the Class amended its never-resolved complaint against Med-Comp in Louisiana state court to assert against its Class-as-CorVel assignee claim against Homeland. The litigation then consisted of a Class PPO claim under state law against a non-diverse defendant (MedComp) and a Class-as-CorVel assignee claim against the diverse defendant (Homeland).

Homeland removed the refashioned case to federal court on the basis of diversity jurisdiction and CAFA minimal diversity jurisdiction.¹ Homeland argued that Med-Comp‘s non-diverse Louisiana citizenship should be disregarded because the Class PPO Act claims against Med-Comp were “improperly and egregiously misjoined” with the assignment-based bad faith claim against Homeland. The Class moved to remand.

Adopting a magistrate judge‘s report and recommendation, the district court retained jurisdiction over the Class-as-CorVel assignee claim against Homeland and remanded to state court the Class PPO Act claims against Med-Comp. The court reasoned that it had ordinary diversity jurisdiction over the case because Homeland‘s citizenship was diverse from the Class, while Med-Comp, the non-diverse defendant, was improperly joined. Without addressing the viability of the Class‘s PPO Act claims against Med-Comp, the district court explained that improper joinder applied because, in its new capacity as CorVel‘s assignee, the Class qua assignee lacked a reasonable basis for recovery against Med-Comp.² The Class appealed an unfavorable judgment on jurisdiction and on the merits.

DISCUSSION

The majority treats this case as if a ruling favoring federal jurisdiction will fundamentally undermine settled principles of diversity-based removal jurisdiction and its corollary, the “improper joinder” doctrine as articulated in this court‘s decisions. See, e.g., Smallwood v. Ill. Cent. R.R. Co., 385 F.3d 568, 573 (5th Cir. 2004) (en banc). I respectfully disagree. As this case was framed in state court before the most recent removal, it did not fit the classic diversity mold: Louisiana plaintiffs, the PPO Class and the Class-as-CorVel assignee, in effect two separate plaintiffs, sued, respectively, the non-diverse defendant Med-Comp, a citizen of Louisiana, and the diverse defendant Homeland. There are two separate plaintiffs and two entirely distinct legal claims.

My position is that this unconventional alignment could yield federal jurisdiction because a non-diverse, in-state defendant‘s citizenship destroys diversity-based removal jurisdiction only if that defendant is “properly joined.” 28 U.S.C. § 1441(b) (emphasis added). “[I]f the plaintiff improperly joins a non-diverse defendant, then the court may disregard the citizenship of that defendant, dismiss the non-diverse defendant from the case, and exercise subject matter jurisdiction over the remaining diverse defendant.” Flagg v. Stryker Corp., 819 F.3d 132, 136 (5th Cir. 2016) (en banc). This court holds that improper joinder can be established by the “inability of the plaintiff to establish a cause of action against the non-diverse party in state court.” Smallwood v. Ill. Cent. R.R. Co., 385 F.3d 568, 573 (5th Cir. 2004) (en banc) (quoting Travis v. Irby, 326 F.3d 644, 646–47 (5th Cir. 2003)). And the critical question is “whether the defendant has demonstrated that there is no possibility of recovery by the plaintiff against an in-state defendant.” Id. The motive or purpose of the joinder is irrelevant. Id. at 574. Instead, the test for improper joinder is founded on the Federal Rules, whereby a court “conduct[s] a Rule 12(b)(6)-type analysis, looking initially at the allegations of the complaint to determine whether the complaint states a claim under state law against the in-state defendant. Ordinarily, if a plaintiff can survive a Rule 12(b)(6) challenge, there is no improper joinder.”³ Id. at 573. If not, the defendant is improperly joined.

The majority asserts that Smallwood sets forth the exclusive tests for improper joinder, but it ignores that when this court clarified the improper joinder test in Smallwood, there was no occasion to consider how the test might apply to a situation involving the improper joinder of plaintiffs’ unrelated claims against an out-of-state and an in-state defendant. See id. at 571–72. Nor has this court subsequently addressed the situation. What I contend is that Smallwood‘s logic suggests that if removal jurisdiction exists notwithstanding that one plaintiff “improperly” asserted a claim against a non-diverse defendant, then jurisdiction also exists if multiple plaintiffs have “improperly” joined in-state and out-of-state defendants using claims wholly unrelated to the other plaintiffs’ claims. Federal Rule of Civil Procedure 20(a) places limitations on the joinder of defendants and plaintiffs. The Rule provides that either plaintiffs or defendants may be joined in an action if “they assert any right to relief . . . with respect to or arising out of the same transaction, occurrence, or series of transactions or occurrences.” FED. R. CIV. P. 20(a)(1)(A), (2)(A).⁴ In this case, there is no relationship between the claims of the Class based on the Louisiana PPO statute and the Class‘s claim as the assignee of CorVel‘s insurance coverage dispute with Homeland. The Class‘s original PPO claims were filed because they, as medical providers, believed they had not been paid properly by PPO-related entities.⁵ The assigned claim against Homeland, in contrast, seeks reimbursement for CorVel‘s settlement with the Class based on the CorVel-Homeland insurance policy and violation of Louisiana law prescribing insurers’ duties. As a result, the required relationship between Med-Comp and Homeland as defendants is lacking

The majority contends that the Smallwood test, applying Rule 12(b)(6), is easily met because the “Class” has a viable claim against Med-Comp in addition to its assigned claim against Homeland. This is incorrect. As the district court correctly recognized, the claim against Homeland is brought in the Class‘s capacity as CorVel‘s assignee—meaning the Class has “stepped into the shoes of [CorVel] for the purposes of this lawsuit.” Woodfield v. Bowman, 193 F.3d 354, 359 (5th Cir. 1999) (discussing the effect of an assignment under Louisiana state law). Functionally, this suit consists of two plaintiffs, the Class and the Class qua “CorVel-assignee,” and each asserts a claim against a separate defendant. More important, the Class (with the PPO Act claims) asserts no claim against Homeland, and the Class qua “CorVel-assignee” asserts no claim against Med-Comp. Independently, the “CorVel-assignee” claim against Homeland satisfies complete diversity and thus falls within federal jurisdiction, but the PPO Act claim against Med-Comp does not.

The majority simply ignores the nature of an assignment under Louisiana law. The Class-as-CorVel‘s assignee claim against Homeland is delimited by the rights that CorVel possessed. In asserting that claim, the Class takes on the separate and distinct persona of the assignor, CorVel.⁶ See Wagoner v. Chevron USA, Inc., 48,119 (La. App. 2 Cir. 7/24/13), 121 So. 3d 727, 734 (”Wagoner I and Wagoner II do not include the ‘same parties,’ because the Wagoners are appearing in a different capacity in Wagoner II than they did in Wagoner I.“). Thus, the Class acting as CorVel‘s assignee cannot accomplish what CorVel in the first instance could not. Ultimately, the Class wishes to have it both ways: it wants to assume CorVel‘s identity on the merits of the Homeland claim but discard it while navigating any procedural hurdles. I find no support for this tactic in our precedent or that of Louisiana.

Retaining federal jurisdiction in the face of improper joinder is premised on the idea that defendants may remove an entire lawsuit to federal court, and the district court can sever the portion that does not belong under the Federal Rules. Whether Rule 12(b)(6), i.e. the “no possibility” test, or the joinder Rules underlie the severance makes no difference. The court here severed the improperly joined claim. The district court is permitted to “add or drop a party . . . [and] sever any claim against a party.” FED. R. CIV. P. 21. The court‘s exercise of authority in severing the Med-Comp claim is in accord with Smallwood, which permits the court, post-removal, to dismiss a claim against the non-diverse defendant under Rule 12(b)(6).

It is implied, though the majority does not really contend, that Rule 21 cannot be used to “extend” federal jurisdiction. See FED. R. CIV. P. 82. But improper joinder cannot deprive the federal courts of jurisdiction. Diversity jurisdiction exists for the suit between the Class-as-CorVel assignee and Homeland. Thus, using Rule 21 to sever the improperly joined parties and claims does not extend or create federal jurisdiction, it merely reinforces the statutorily prescribed balance between state and federal courts. This result discourages artful joinder schemes devised around anchor plaintiffs. This is not different from using a Rule 12(b)(6)-type analysis to determine that there is “no possibility of recovery” against an in-state defendant who is therefore improperly joined. “[T]he Federal courts should not sanction devices intended to prevent the removal . . . where one has that right, and should be equally vigilant to protect the right to proceed in the Federal court as to permit the state courts, in proper cases, to retain their own jurisdiction.” Smallwood, 385 F.3d at 573 (quoting 14 CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE & PROCEDURE § 3641, at 173 (3d ed. 1998)).

The majority criticize my analysis for several additional reasons, none of which are persuasive. First, I allegedly extend our current authorities on “improper joinder” beyond the “no possibility of recovery” rule. Arguably true, but I plead logic, and the unusual circumstances here, which bespeak obvious joinder machinations undertaken to avoid federal court.⁷ Second, the majority asserts that the “improper joinder” analysis of Smallwood and other cases is completely confined to the “no possibility of recovery” rule.

Textually, the term “properly joined” and its negative inference “improperly joined” both envelop “joinder,” an action in no way semantically limited, nor limited by the Federal Rules, to 12(b)(6) deficiencies.

Third, the majority deride application of Tapscott v. MS Dealer Serv. Corp., 77 F.3d 1353, 1360 (11th Cir. 1996), abrogated on other grounds by Cohen v. Office Depot, Inc., 204 F.3d 1069 (11th Cir. 2000), which enunciated a “fraudulent misjoinder” doctrine. Of course, Tapscott has received mixed reviews in lower courts, but, as the majority recognize, its analysis is not criticized by Wright & Miller. Further, the logic of Tapscott makes sense where the basic aim of improper joinder analysis is to prevent parties from manipulating their pleadings to deprive diverse defendants of the federal forum to which they are statutorily entitled. This court, before today‘s opinion, favorably cited Tapscott on two occasions. See In re Benjamin Moore, 318 F.3d 626, 630-31 (5th Cir. 2002); Crockett v. R.J. Reynolds Tobacco Co., 436 F.3d 529, 533 (5th Cir. 2006). In Crockett, moreover, the court approved “recognizing an exception to the voluntary-involuntary rule where defendants are improperly, though not fraudulently, joined.” Id. In so doing, this court noted that its “conclusion finds support in the text of Sec. 1441(b), which does not refer to ‘nonfraudulently joined’ parties. Rather, it blocks removal only where ‘properly joined’ defendants are citizens of the state in which the action is brought.” Id. n. 7.⁸

Finally, the majority assert that federal courts interfere with state court case management if they conduct post-removal joinder assessments. This is pure speculation, because the case before us is plain. The Class asserting the PPO claim against a non-diverse defendant is wholly separate, legally and juridically, from the Class-as-CorVel assignee claim against the diverse defendant Homeland. Misjoinder under the Federal Rules is not debatable, even if the majority refuses to debate.⁹

In sum, if the Class-as-CorVel assignee wished to preclude Homeland from gaining access to the federal courts, there is a simple solution. It could have filed suit in the state courts of New York or Minnesota, where the defendant is incorporated or has its principal place of business. But having chosen to pursue litigation in Louisiana, the Class should have been bound by Homeland‘s right to proceed in federal court.

CONCLUSION

The only in-state defendant was improperly joined in a claim that the Class-as-CorVel assignee had no legal right to pursue. The plaintiffs’ mischievous use of their PPO Class claim cannot cure this defect because the PPO Class had no ability or right to maintain suit against Homeland. Federal removal jurisdiction is proper under 28 U.S.C. Sec. 1442(b) because, under the logic of our authorities, Med-Comp was not “properly joined” in the separate suit by the Class-as-CorVel assignee against Homeland. I respectfully dissent.

Notes

Judge HO fully joins in this opinion and also files a concurrence; Judge JONES dissents. See also, e.g., (en banc) (Ho, J., concurring) (judges should follow legal texts “to the maximum extent that Supreme Court precedent permits“) (citing (en banc) (Ho, J., concurring)); (en banc) (Bush, J., concurring) (same); (Bumatay, J., dissenting from denial of rehearing en banc) (same); (en banc) (Jordan, J., dissenting) (when it comes to precedent with a “shaky originalist foundation . . . there is always the option of declining to broaden it—of refusing to extend it one inch beyond its previous contours“); (Viviano, J., dissenting) (judges should not extend precedent “unless the extension is required by the Constitution‘s original meaning“); see generally Josh Blackman, Originalism and Stare Decisis in the Lower Courts, 13 NYU J.L. & LIBERTY 44, 51 (2019) (“[A] judge should only extend a Supreme Court precedent if the original meaning of the Constitution can support that extension.“). I do not contend that CAFA conferred jurisdiction in the district court.

See (Williams I); (Williams II). Both Leocal and Clark involved the interpretation of a statute. In both cases, the Supreme Court refused to give the same statutory text a different meaning based on the context of the case (in Leocal, for example, the Court refused to construe a statute differently based on whether the case involved civil or criminal penalties). That principle does not apply here, because improper joinder is a judge-made doctrine and therefore subject to judicially created limitations and exceptions. Cf. (“The good-faith exception is a judicially created exception to [a] judicially created rule,” so “we could . . . recognize . . . exception[s] to the good-faith exception.“). Homeland moved to prevent relitigation of the Delaware claim based on the Full Faith and Credit Act, 28 U.S.C. § 1738. The district court held that it must look to Delaware‘s preclusion principles, that Delaware would apply claim preclusion to a case that had been dismissed for failure to file within the statute of limitations, and that the claim against Homestead must be dismissed. Further discussion of the merits is inappropriate here.

This claim was brought under Louisiana‘s Bad Faith Statute. LA. STAT. ANN. § 22:1973. There may be other instances where the complaint appears to state a claim, but pertinent facts have been omitted or misstated. . The district court may then “pierce the pleadings and conduct a summary inquiry.” This is not one of those cases.

It appears that, by the time Williams amended the PPO Act complaint to bring the bad faith claim against Homeland in 2019 (approximately ten years after filing it in the first place), Med-Comp had essentially gone defunct. It filed its last annual report with the State of Louisiana in 2013, and the Louisiana Department of State terminated its corporate registration in 2016. Under Louisiana law, however, corporations with terminated registrations can still be sued, LA. CODE CIV. PROC. ANN. art. 85, and even corporate dissolution does not “abate or suspend” any proceedings that were pending at the time of dissolution, LA. STAT. ANN. § 12:1-1405(B)(6). Accordingly, despite Med-Comp‘s financial condition and the dissenting opinion‘s conclusion that “logic” compels otherwise, Med-Comp remained in the case. So much for the majority‘s complaint about uncertainty as to whether federal or state joinder rules should apply in this situation. Federal law should govern post-removal joinder just as Rule 12(b)(6)-type analysis governs the “no possibility of recovery” determination.

The district court indicated that it “denied” the motion to remand, but it is clear that it granted the motion in part: the order “remanded” all of the “pre-existing claims” against Med-Comp. That is also improper, as the proper mechanism of handling an improperly joined party is to dismiss it, not send it to another court to decide the merits. (en banc) (“[I]f the plaintiff improperly joins a non-diverse defendant, then the court may disregard the citizenship of that defendant, dismiss the non-diverse defendant from the case, and exercise subject matter jurisdiction over the remaining diverse defendant.“) Homeland was joined originally pursuant to Louisiana‘s direct action statute, L.A. STAT. § 22:1269(B), not on the basis of Homeland‘s alleged violation of its agreement to insure CorVel.

The Class also argues in its reply brief that Homeland waived or forfeited its arguments regarding improper joinder because Homeland did not specifically assert improper joinder in its notice of removal. To preserve an argument on appeal, a party must “press” — “not merely intimate” — that argument before the district court. . This argument is unavailing. In the district court, Homeland asserted that diversity jurisdiction existed because Med-Comp “was not properly joined as a defendant,” and argued in resistance to the Class‘s motion to remand that the Class had “no reasonable basis for recovery” against Med-Comp. Accordingly, we conclude that Homeland did not waive or forfeit its improper joinder argument. It is this distinction that also empowers courts to sever unrelated claims asserted in different capacities when they would otherwise be permissible under Federal Rule of Civil Procedure 18(a). See 6A CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE CIVIL § 1585 (3d ed. 2020 update) (“[W]hen a party asserts or defends against claims that have been properly asserted in more than one capacity, the rules relating to the joinder of parties rather than those dealing with the joinder of claims probably will apply. For example, a party may assert one claim as an individual and another claim as the administrator of an estate. In these situations there are essentially two different claimants, and although joinder under Rule 18(a) may be proper, the court may declare a misjoinder of parties if the claims asserted in each capacity are totally unrelated and do not meet the standard established by the rules governing party joinder.“).

The district court discussed both possible bases for original jurisdiction, but its decision to exercise jurisdiction was based on ordinary diversity jurisdiction. See . On appeal, Homeland also argues that the district court had CAFA minimal diversity jurisdiction over the case. What possible reason was there to revive a case against MedComp, the only remaining local defendant, which is defunct? And why rev up the Class PPO claim against such a defendant, although plaintiffs’ counsel might perhaps have found a way to extract fees at the end of litigation? Little did the plaintiff Class-as-Corvel assignee know that, had my position prevailed, the favorable judgment for Homeland here might well have been reversed.

This circuit has occasionally referred to improper joinder as “fraudulent joinder.” See . We refer to the doctrine as improper joinder and note that the improper joinder doctrine is distinct from the fraudulent misjoinder doctrine discussed below. To be sure, in Crockett, the state court had severed the “improperly joined” in-state defendants, and defendants may try to do that before removal. But this court went farther and explicitly tied the state court‘s severance to the Fed. R. Civ. Pro. 20(a) as well.

The dissenting opinion reaches that conclusion. Dissenting Op. at 5-6 & n.7. However, as explained more fulsomely below, no Fifth Circuit precedent supports this concept. Neither do the dissenting opinion‘s two cited authorities. The cited treatise indicates only that a party‘s different capacities can matter for federal party joinder under Federal Rules of Civil Procedure 20 and 21, not removal based on improper joinder. See 6A CHARLES ALAN WRIGHT, ARTHUR R. MILLER, & MARY KAY KANE, FEDERAL PRACTICE AND PROCEDURE § 1585 (3d ed. 2010 & Apr. 2021 Update) (noting that a “court may declare a misjoinder of parties if the claims asserted in each capacity are totally unrelated and do not meet the standard established by the rules governing party joinder“). The cited state court case, , resolves only the question of whether a Louisiana state court can appropriately sever misjoined claims brought in different capacities. In fact, the majority view means that the peculiarities of state procedure can dictate federal court jurisdiction—a result the Supreme Court has rejected in a similar context. See (rejecting removal on the basis that the removing parties, although denominated defendants under state practice, were plaintiffs under federal law); (“For the purpose of removal, the federal law determines who is plaintiff and who is defendant. It is a question of the construction of the federal statute on removal, and not the state statute. The latter‘s procedural provisions cannot control the privilege of removal granted by the federal statute.“) (citation omitted).

There are some contexts in which a party‘s different capacities can affect the improper joinder analysis—but using claim assignment to defeat diversity jurisdiction is not one of them. For instance, a party cannot collusively assign claims to create diversity jurisdiction. See 28 U.S.C. § 1359 (“A district court shall not have jurisdiction of a civil action in which any party, by assignment or otherwise, has been improperly or collusively made or joined to invoke the jurisdiction of such court.“); but see (“[N]o authority has as yet been given [to the federal courts] to take jurisdiction of a case by removal from a state court when a colorable assignment has been made to prevent such a removal.” (emphasis added)). Additionally, a party sometimes takes on a different citizenship when it acts in certain representative capacities not at issue here—specifically, a party adopts the citizenship of a represented person when acting as the representative of a decedent, an infant, or an incompetent person. See 28 U.S.C. § 1332(c)(2).

The settlement agreement defines “the Released Parties” as “CorVel and all Affiliates thereof,” excludes Med-Comp from the definition of “Affiliates,” and describes that the Class released claims against “the Released Parties” relating to an “Episode.”

Homeland (along with a number of district courts in our circuit and the dissenting opinion here) seems to believe that we endorsed some sort of Tapscott-based fraudulent misjoinder doctrine in , and . We did not. Specifically, in neither of those cases did we permit removal based on party misjoinder, the issue here. Indeed, our decision in Crockett is more properly read for the exact opposite proposition: that a defendant contemplating removal can (and should) ask the state court to resolve party misjoinder questions and then remove. .

See, e.g., (noting that fraudulent misjoinder doctrine is a “solution without a problem“); (“[I]n light of the . . . numerous unsettled doctrinal questions, the Court agrees with the other district courts that have left the whole enterprise to the state courts.“); (“In the Court‘s view, the Tapscott doctrine is an improper expansion of the scope of federal diversity jurisdiction by the federal courts.“); (“[T]he last thing the federal courts need is more procedural complexity.... [T]he better rule would require [the defendant] to resolve the claimed misjoinder in state court, and then . . . seek removal of the cause to federal court.“); see also (declining to adopt the doctrine and noting that most district courts in the Sixth Circuit have declined to do so); (noting the widespread rejection of the doctrine by district courts in the Ninth Circuit).

Homeland suggests that the answer is actually more complicated because, it claims, requiring defendants to resolve such procedural issues in state court prior to removal would force them to “risk missing the 30-day removal deadline under Section 1446.” Homeland is wrong, at least under our established case law. In a world without the fraudulent misjoinder doctrine, the 30-day removal deadline does not start running until after the state court concludes the claims are misjoined and severs all non-diverse parties. Only at that point does the case “become removable“—thereby starting the removal deadline—because only then are the parties completely diverse. 28 U.S.C. § 1446(b)(3); 14C CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE & PROCEDURE § 3723.1 (Rev. 4th ed. 2018 & Apr. 2021 Update) (“The 30-day time limit for removal . . . would not begin to run until any misjoined, or fraudulently joined, non-diverse party was dropped from the action.“); see also (concluding that an “initially nonremovable” case became removable “upon the severance” of non-diverse defendants in state court).

In particular, our prior decision that CAFA was not in play effectively foreclosed review of the district court‘s decision to remand the Class‘s PPO Act claims to state court. See 28 U.S.C. § 1447(d); cf. 28 U.S.C. § 1453(c)(1) (allowing appellate review of remand orders made under CAFA). To evaluate CAFA now would essentially reopen that issue by requiring this court to determine whether that remand was appropriate, something the law of the case doctrine prohibits.

Read the detailed case summary