Frank Williams, Jr. v. Lockheed Martin Corp, 958 F.3d 341

JAMES C. HO, Circuit Judge, joined by OWEN, Chief Judge, and JONES, STEWART, DENNIS, ELROD, HAYNES, GRAVES, HIGGINSON, and ENGELHARDT, Circuit Judges:

When a plaintiff sues multiple defendants, counsel may need to take certain steps to ensure the plaintiff‘s right to appeal. That is because courts of appeals have jurisdiction to review only certain types of district court decisions.

Under 28 U.S.C. § 1291, courts of appeals may review only “final decisions” of the district courts. Under our precedents, there is no final decision if a plaintiff voluntarily dismisses a defendant without prejudice, because the plaintiff “is entitled to bring a later suit on the same cause of action.” Ryan v. Occidental Petroleum Corp., 577 F.2d 298, 302 (5th Cir. 1978). And in a suit against multiple defendants, there is no final decision as to one defendant until there is a final decision as to all defendants. See FED. R. CIV. P. 54(b) (absent an order to the contrary, “any order or other decision, however designated, that adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties does not end the action as to any of the claims or parties“).

A potential complication arises when a case implicates both of those principles—that is, when a plaintiff sues two defendants, and then voluntarily dismisses one defendant without prejudice, while litigating against the other to conclusion. Some have expressed concern that the plaintiff may fall into a “finality trap“—unable to obtain an appealable final decision, despite having lost to the second defendant. See Terry W. Shackmann & Barry L. Pickens, The Finality Trap: Accidentally Losing Your Right to Appeal (Part I), 58 J. MO. B. 78, 78 (2002).

But established rules of civil procedure provide many tools to avoid that alleged “trap.” They include amendment of the complaint to remove claims or parties under Federal Rule of Civil Procedure 15(a); severance of parties under Rule 21; and entry of a partial final judgment under Rule 54(b).¹ A plaintiff can also voluntarily dismiss a defendant with prejudice.

In this case, Plaintiffs chose Rule 54(b). As a result, there is no need to address the other finality issues raised by the parties. Because we conclude that the district court properly entered partial final judgment under Rule 54(b), we have jurisdiction to hear these appeals and accordingly return them to the panel for a ruling on the merits.

I.

Plaintiffs Tarsia and Breck Williams sued twenty-four defendants after their father died from mesothelioma. After protracted litigation before a multi-district litigation court, several defendants, including Taylor Seidenbach, Inc., and McCarty Corp., obtained summary judgment.

The Williamses subsequently moved to dismiss the remaining defendants, including CSR, Ltd., Environmental Abatement Services, Inc., and The Gottfried Corp., pursuant to Rule 41(a). The district court granted the Rule 41(a) dismissal motions, but it did not specify whether the dismissals were with or without prejudice. The Williamses then appealed as to the several defendants who had previously obtained summary judgment, including Taylor Seidenbach and McCarty.

On appeal, this court held that CSR, Environmental Abatement Services, and Gottfried were dismissed without prejudice. Accordingly, we dismissed the appeal for want of a “final decision” under 28 U.S.C. § 1291. Williams v. Taylor-Seidenbach, Inc. (Williams I), 748 F. App‘x 584, 587–88 (5th Cir. 2018).

In response, the Williamses sought and obtained partial final judgment under Rule 54(b) as to various defendants, including Taylor Seidenbach and McCarty, and then appealed again. A panel of this court held that the district court lacked the power to enter partial final judgment under Rule 54(b), and therefore dismissed the appeal once again for want of a final decision. Williams v. Taylor Seidenbach, Inc. (Williams II), 935 F.3d 358, 360 (5th Cir. 2019), vacated on rehearing en banc, 941 F.3d 1183 (5th Cir. 2019).

We subsequently granted rehearing en banc. We now conclude that Rule 54(b) authorized the district court to enter partial final judgment following the dismissal of the remaining defendants under Rule 41(a), and that this appeal may therefore proceed.

II.

This case involves the intersection of two different Federal Rules of Civil Procedure—Rules 41(a) and 54(b). Accordingly, we address Rule 41(a) briefly, before turning to Rule 54(b).

Rule 41(a) allows plaintiffs to voluntarily dismiss “an action.” One could—as the dissent does—plausibly construe “action” under Rule 41(a) to refer only to the entire case and not to individual defendants. See Harvey Aluminum, Inc. v. Am. Cyanamid Co., 203 F.2d 105, 108 (2nd Cir. 1953). But our circuit precedents interpret “action” to cover individual defendants—thus allowing plaintiffs, like the Williamses, to use Rule 41(a) to dismiss individual defendants. See Nat‘l City Golf Fin. v. Scott, 899 F.3d 412, 415 n.3 (5th Cir. 2018) (“Rule 41(a) permits a plaintiff to dismiss just one defendant, ‘even though the action against another defendant would remain pending.‘“) (quoting Plains Growers, Inc. ex rel. Florists’ Mut. Ins. Co. v. Ickes-Braun Glasshouses, Inc., 474 F.2d 250, 253 (5th Cir. 1973)); Exxon Corp. v. Md. Cas. Co., 599 F.2d 659, 662–63 (5th Cir. 1979) (distinguishing between impermissible Rule 41(a) dismissals of individual claims and permissible Rule 41(a) dismissals of individual defendants).

In this en banc proceeding, the parties do not challenge the validity of our circuit precedents permitting the Williamses to voluntarily dismiss individual defendants under Rule 41(a). That is unsurprising. For even if we were to reconsider our precedent and to construe Rule 41(a) to permit voluntary dismissal only of an entire case, and not of individual defendants, we would still conclude that there is appellate jurisdiction over this appeal.

To see why, assume that Rule 41(a) only allows dismissals of the entire case. Then the Williamses’ Rule 41(a) dismissal, which sought to dismiss only some of the defendants, was invalid, and an invalid Rule 41(a) dismissal is a nullity. So the claims against the purportedly dismissed defendants would still be “pending in district court,” as circuit precedents confirm. Exxon, 599 F.2d at 663. See also Perry v. Schumacher Grp. of La., 891 F.3d 954, 958–59 (11th Cir. 2018) (“[T]he [Rule 41(a)(1)(A)] Stipulation, which purported to dismiss ‘Count III of the Fourth Amended Complaint . . . without prejudice,’ was invalid. By stroke of sheer good fortune for Dr. Perry, the Stipulation did not divest the District Court of its jurisdiction.“).

The upshot is this: If we accept the dissent‘s reading of Rule 41(a), then the claims against the purportedly dismissed defendants were in fact never dismissed, but instead were indisputably pending before the district court at the time it entered partial final judgment under Rule 54(b). It therefore follows that the partial final judgments under Rule 54(b) were valid—even under the dissent‘s understanding of Rules 41(a) and 54(b)—and that we therefore have jurisdiction over this appeal.

But as we shall explain, we can also reach that same result without upsetting circuit precedent. And because that is so, we have no occasion to reconsider our precedent. See United States v. Castillo-Rivera, 853 F.3d 218, 221 n.1 (5th Cir. 2017) (en banc) (refusing to revisit circuit precedent where it was “not necessary to our disposition of [the] case“).

III.

We turn now to the validity—and appealability—of the partial final judgments entered by the district court here under Rule 54(b).

The courts of appeals are courts of limited jurisdiction, “possess[ing] only that power authorized by Constitution and statute.” Kokkonen v. Guardian Life Ins. Co. of Am., 511 U.S. 375, 377 (1994). Under 28 U.S.C. § 1291, courts of appeals have jurisdiction to review “final decisions” of the district courts.

Congress has “empowered [the Supreme Court] to clarify when a decision qualifies as ‘final’ for appellate review purposes, and to expand the list of orders appealable on an interlocutory basis.” Swint v. Chambers Cty. Comm‘n, 514 U.S. 35, 48 (1995). “The procedure Congress ordered for such changes, however, is not expansion by court decision, but by rulemaking.” Id. (emphasis added). See 28 U.S.C. § 2072(a), (c) (“The Supreme Court shall have the power to prescribe general rules of practice and procedure . . . . Such rules may define when a ruling of a district court is final for the purposes of appeal under section 1291 of this title.“). See also Microsoft Corp. v. Baker, 137 S. Ct. 1702, 1712–13 (2017) (instructing courts not to “subvert[] the final-judgment rule and the process Congress has established for refining that rule“).

The Supreme Court has thus instructed that “Congress’ designation of the rulemaking process as the way to define or refine when a district court ruling is ‘final’ and when an interlocutory order is appealable warrants the Judiciary‘s full respect.” Swint, 514 U.S. at 48. So when a question arises about appealability in light of the final judgment rule, courts and litigants must adhere to a simple principle: follow the rules. That is what, after various twists and turns, the Williamses did here—they invoked Rule 54(b) to secure a “final judgment as to one or more, but fewer than all, claims or parties.”

A.

To begin, it is important to understand what the Williamses want: appellate review of the district court‘s entry of summary judgment as to certain defendants, without litigating all the claims they brought against all the defendants they sued.

There are many ways they could have achieved that under current precedent without running afoul of the final judgment rule. They could have dismissed the remaining defendants with prejudice under Rule 41(a). See, e.g., Marshall v. Kan. City S. Ry. Co., 378 F.3d 495, 500 (5th Cir. 2004). They could have amended their complaint to excise any remaining claims or parties under Rule 15(a). See, e.g., Rockwell Int‘l Corp. v. United States, 549 U.S. 457, 473–74 (2007) (“[W]hen a plaintiff . . . voluntarily amends the complaint, courts look to the amended complaint to determine jurisdiction.“); Nat‘l Broiler Mktg. Ass‘n v. United States, 436 U.S. 816, 819 n.5 (1978) (noting that a party used Rule 15(a) “to facilitate the appeal“). They could have requested severance of certain parties under Rule 21. See United States v. O‘Neil, 709 F.2d 361, 368 (5th Cir. 1983). Or they could do what they did here: seek a partial final judgment under Rule 54(b).

Federal Rule of Civil Procedure 54(b) states:

When an action presents more than one claim for relief—whether as a claim, counterclaim, crossclaim, or third-party claim—or when multiple parties are involved, the court may direct entry of a final judgment as to one or more, but fewer than all, claims or parties only if the court expressly determines that there is no just reason for delay. Otherwise, any order or other decision, however designated, that adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties does not end the action as to any of the claims or parties and may be revised at any time before the entry of a judgment adjudicating all the claims and all the parties’ rights and liabilities.

FED. R. CIV. P. 54(b).

Rule 54(b) thus sets forth the governing framework for determining finality in a suit against multiple defendants. Of course, once the district court has decided all claims against all parties, that decision is plainly final. By contrast, any order that resolves “the rights and liabilities of fewer than all the parties” is not final—because any such order “may be revised at any time” before final judgment. Id. So if a party wishes to appeal such an order, it must ask the district court to enter “a final judgment as to one or more, but fewer than all, claims or parties,” and the court should do so, provided “there is no just reason for delay.” Id.

By following the framework set forth in Rule 54(b) and obtaining a partial final judgment as to Taylor Seidenbach and McCarty, the Williamses preserved their right to appeal against those defendants.

B.

Rule 54(b) sets forth a straightforward path that the Williamses dutifully followed. The only wrinkle is whether a district court can enter partial final judgment where a plaintiff first litigates the case to conclusion as to various defendants, and then voluntarily dismisses all remaining defendants under Rule 41(a) at the end of the case—as the Williamses did here.

We see no reason why the order of events should alter the court‘s power under Rule 54(b). Indeed, our court allowed entry of partial final judgment under Rule 54(b) under substantially indistinguishable circumstances in Swope v. Columbian Chemicals Co., 281 F.3d 185 (5th Cir. 2002). In Swope, the district court granted summary judgment to one defendant, and then later granted summary judgment to the second defendant as to all but one claim. Id. at 190. In response, the plaintiff moved to dismiss the remaining claim under Rule 41(a) and to obtain an appealable partial final judgment as to the rest of the case under Rule 54(b). Id. The district court obliged, dismissing the remaining claim without prejudice and then immediately entering a Rule 54(b) judgment as to the summary judgment orders. Id. at 190–91 & nn.1–2. The plaintiff subsequently appealed, and our court accepted jurisdiction over that appeal. Id. at 194.

What‘s more, our court in Swope expressly pointed out that the timing of the Rule 41(a) dismissal did not matter to its analysis—the appeal of the Rule 54(b) judgment was valid regardless. As the court went out of its way to observe, “[i]t is unnecessary to decide . . . whether the dismissal took effect upon filing . . . of the [Rule 41] motion“—and therefore, like in this case, well before the entry of partial final judgment under Rule 54(b)—“or upon the trial court‘s granting[] of the motion.” Id. at 192 n.15. Translation: Entry of a partial final judgment is proper under Rule 54(b) regardless of whether it occurs before or after the voluntary dismissal of any remaining defendants under Rule 41(a).

Swope is correct and we apply it here. Rule 54(b) authorizes district courts to “direct entry of a final judgment as to one or more, but fewer than all, claims or parties.” FED. R. CIV. P. 54(b). And a voluntary dismissal of some defendants under Rule 41(a) at the end of the case does not alter the court‘s Rule 54(b) authority in any way. A dismissed claim remains a part of the case, absent amendment of the complaint under Rule 15. And that is so regardless of when the Rule 41(a) dismissal occurs. Likewise, the fully litigated claims—such as the partial summary judgment claims that the Williamses hope to appeal here—naturally remain pending before the court as well. See, e.g., Nat‘l City Golf Fin., 899 F.3d at 415 n.3 (“Rule 41(a) permits a plaintiff to dismiss just one defendant, ‘even though the action against another defendant would remain pending.‘“) (quoting Plains Growers, 474 F.2d at 253).²

So there is no reason why a district court cannot enter a partial final judgment under Rule 54(b) when any remaining defendants are voluntarily dismissed at the end of a case under Rule 41(a).

Notably, a distinguished panel of the Seventh Circuit reached the same conclusion in ITOFCA, Inc. v. MegaTrans Logistics, Inc., 235 F.3d 360 (7th Cir. 2000). ITOFCA involved the same sequence of events as this case—entry of summary judgment as to some claims, followed by dismissal of the rest of the case without prejudice. See id. at 362–63. The Seventh Circuit concluded that Rule 54(b) would generally authorize entry of partial final judgment under those circumstances. Its reasoning was simple: A dismissal without prejudice is equivalent to no dismissal at all because the claim can be refiled at any time. Id. at 364. So we may—indeed, according to the Seventh Circuit, we “must,” id.—act as if the claim is still pending before the district court.

To begin with, the Seventh Circuit concluded, as we did in Swope, that the dismissal of certain claims without prejudice deprived plaintiffs of finality as to the other claims. “[W]e must view the counterclaims dismissed without prejudice as if they are still before the district court, which they could be at any moment. Since we would not have appellate jurisdiction under § 1291 if the court had not dismissed the counterclaims, we have no appellate jurisdiction under the circumstances as they now exist.” Id. at 364.

As a result, the Seventh Circuit concluded, the plaintiff must take some additional step, such as entry of partial final judgment under Rule 54(b), in order to appeal any of the other claims. That is because “Rule 54(b) authorizes the district court to make immediately appealable a judgment that disposes, with finality, of one or more (but not all) claims, even though other claims remain pending in the district court so that the suit as a whole has not been finally disposed of by that court.” Id. Thus, “[g]iven . . . that the dismissal of MegaTrans’ counterclaims without prejudice has not created a final decision that allows for Section 1291 review, Rule 54(b) would seem to be an adequate alternative means of gaining appellate jurisdiction.” Id.³

That conclusion is unsurprising given the plain text of Rule 54(b): A district court “may direct entry of a final judgment as to one or more, but fewer than all, claims or parties.” FED. R. CIV. P. 54(b). And a district court may do so in any case that “presents more than one claim for relief—whether as a claim, counterclaim, crossclaim, or third-party claim—or when multiple parties are involved.” Id. “Otherwise“—that is, in the absence of a partial final judgment under Rule 54(b)—“any order or other decision . . . that adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties does not end the action as to any of the claims or parties and may be revised at any time.” Id.

In other words, the text of Rule 54(b) identifies the problem that it is attempting to solve (adjudication as to only some parties or only some claims “does not end the action” and is thus not, by itself, a final judgment), and then provides the solution (district courts may enter partial final judgment to facilitate appeal).

The Supreme Court has made clear that the Federal Rules of Civil Procedure warrant our “full respect” as “the way to define or refine when a district court ruling is ‘final’ and when an interlocutory order is appealable.” Swint, 514 U.S. at 48. Consistent with that command, we conclude that Rule 54(b) provides a complete solution for plaintiffs who, like the Williamses, sue multiple defendants, but then later seek an appealable final judgment as to only a subset of those defendants.

C.

Plaintiffs have many options to preserve their right to appeal under these circumstances, and the Williamses have properly exercised one of those options here. As a result, we need not answer certain questions that have been raised in this en banc proceeding. For example, some have argued that voluntary dismissal of a defendant or claim without prejudice is a final decision and thus does not deprive this court of appellate jurisdiction—and that we should thus revisit Ryan. See Williams II, 935 F.3d at 361–62 (Haynes, J., concurring). Others have responded that a voluntary dismissal without prejudice is not a “final decision” because the dismissal decides nothing—the plaintiff can re-file—and at a minimum, stare decisis commands respect for that understanding of finality because it is not demonstrably erroneous. See Blue v. D.C. Pub. Sch., 764 F.3d 11, 17 (D.C. Cir. 2014) (noting that “[e]very circuit . . . appears to acknowledge a presumption against” treating a voluntary dismissal without prejudice as a final decision) (collecting cases). Our decision today avoids the need to resolve those issues.

* * *

The vortex of the finality trap lurks whenever a plaintiff sues a basket of defendants. But established rules of civil procedure offer various tools for avoiding that trap. The Williamses used one of those tools here—partial final judgment under Rule 54(b). As a result, we have jurisdiction to hear these appeals. We therefore return these cases to the panel for a decision on the merits.

JAMES C. HO, Circuit Judge, concurring:

The dissent urges a different interpretation of Rule 54(b) than the one adopted by the majority. The dissenters’ views are entitled to respect. Textualists can and sometimes do disagree, in good faith, about the proper interpretation of words and provisions.

But their criticisms of the majority are mistaken, as I write separately to explain. For example, in their introductory paragraph, the dissenters intimate that no federal court has ever adopted the majority‘s interpretation of Rule 54(b). It is a curious attack, considering that they do not provide one example of a court adopting their view that Rule 54(b) is unavailable in situations like this. It is even more curious, considering that the available circuit decisions to date all support the majority‘s view that Rule 54(b) is available here—as the dissent (grudgingly) concedes. See Swope v. Columbian Chems. Co., 281 F.3d 185, 192 & n.15 (5th Cir. 2002); ITOFCA, Inc. v. MegaTrans Logistics, Inc., 235 F.3d 360, 364 (7th Cir. 2000). The dissent simply dismisses the supportive passages in Swope and ITOFCA as mere “drive-by” “dicta” that the en banc court should reject. But as the majority explains, and as I further detail below, both Swope and our respected colleagues on the Seventh Circuit got it right.¹

I.

Before I respond to the dissent‘s effort to overturn our Rule 54(b) precedent, I will first address its effort to overturn our Rule 41(a) precedent.

To begin, the dissent makes three valid points. First, the dissent is correct that one important purpose of en banc rehearing is to reconsider our circuit precedent—for example, “to better align our precedents with the text and original understanding of the Constitution or the plain language of United States statutes” to the maximum extent that Supreme Court precedent permits. Alvarez v. City of Brownsville, 904 F.3d 382, 401 (5th Cir. 2018) (en banc) (Ho, J., concurring).² Second, the dissent is correct that, whether the parties dispute jurisdiction or not, we have an independent duty to confirm that we have jurisdiction to hear cases presented to us.³

Third, as the majority acknowledges, the dissent presents a plausible theory that our precedents construing Rule 41(a) are wrong as a matter of text. Section I of the dissent contends that Rule 41(a) only authorizes dismissals of an entire “action“—and not individual defendants, as the Williamses purported to do here. The dissent would therefore overrule our precedents in Plains Growers, Inc. ex rel. Florists’ Mutual Insurance Co. v. Ickes-Braun Glasshouses, Inc., 474 F.2d 250, 253 (5th Cir. 1973), and National City Golf Finance v. Scott, 899 F.3d 412, 415 n.3 (5th Cir. 2018)—not to mention Exxon Corp. v. Maryland Casualty Co., 599 F.2d 659, 662–63 (5th Cir. 1979) (distinguishing between impermissible Rule 41(a) dismissals of individual claims and permissible Rule 41(a) dismissals of individual defendants).

It is an interesting textual argument, and if it affected our jurisdiction in this appeal, I might seriously consider it. But the majority explains why overruling Plains Growers, National City Golf, and Exxon would not deprive us of jurisdiction. The reasoning is simple: If Rule 41(a) does not let the Williamses dismiss a subset of defendants, then their Rule 41(a) motion is invalid. Various circuits (including ours) have made that clear. See Perry v. Schumacher Grp. of La., 891 F.3d 954, 958–59 (11th Cir. 2018) (“[T]he [Rule 41(a)(1)(A)] Stipulation, which purported to dismiss ‘Count III of the Fourth Amended Complaint . . . without prejudice,’ was invalid. By stroke of sheer good fortune for Dr. Perry, the Stipulation did not divest the District Court of its jurisdiction.“); Exxon Corp., 599 F.2d at 663 (concluding that an invalid attempt to dismiss a claim under Rule 41(a) meant that the claim was “still pending in district court“). If the Williamses’ Rule 41(a) dismissal was invalid, then there were defendants and claims still before the district court. And if that is true, then the Rule 54(b) certification is indisputably valid.

So the dissent should be captioned as a concurrence in the judgment. Because if the dissent is right about Rule 41(a), then it should deny legal effect to the Williamses’ Rule 41(a) motion for invalidly attempting to dismiss only some of the defendants. It should not rewrite the motion to dismiss every defendant in the entire case.

II.

As for the dissent‘s attack on the majority‘s Rule 54(b) textual analysis and circuit precedent: Much of Section II of the dissent expressly relies on Section I. So the problems noted with Section I apply equally to Section II. I simply offer the following additional thoughts in response:

1. The dissenters suggest that the majority‘s reliance on Rule 54(b) is unprecedented. Two problems:

First, the available precedents support the majority—as the dissenters ultimately acknowledge. They do not deny that the Seventh Circuit agrees with the majority but simply dismiss its extensive and thoughtful analysis as mere “dicta.” See ITOFCA, 235 F.3d at 363–65. Likewise, the dissenters attempt to limit the force of Swope by referring to it as a “drive-by jurisdictional ruling.” But a “drive-by jurisdictional ruling” means there is no jurisdictional ruling when a court doesn‘t discuss jurisdiction—not when the dissent disagrees with the discussion. See, e.g., Steel Co. v. Citizens for a Better Env‘t, 523 U.S. 83, 91 (1998) (describing “drive-by jurisdictional rulings” as those where jurisdiction is “assumed without discussion“). Swope devotes an entire section to jurisdiction—spanning five pages of the Federal Reporter and including fourteen paragraphs and thirty footnotes. See Swope, 281 F.3d at 190–94. That is not a drive-by—that is walking in, sitting down, and staying for cocktails, dinner, and dessert.

In addition, it is Swope‘s jurisdictional analysis that the majority discusses at length, and that supports the majority‘s interpretation of Rule 54(b). As the dissent says, Swope involved the validity of a premature notice of appeal. Id. at 191. But for the notice to be valid, there had to be a subsequent, appealable Rule 54(b) certification. As the dissent notes, Swope said that “a premature notice of appeal is effective if Rule 54(b) certification is subsequently granted.” Id. (emphasis added) (quoting St. Paul Mercury Ins. Co. v. Fair Grounds Corp., 123 F.3d 336, 338 (5th Cir. 1997)). Swope further concluded that a valid Rule 54(b) certification existed notwithstanding the plaintiff‘s Rule 41(a) dismissal of the last remaining claim. See id. at 192 & n.15. So the proposition for which the majority invokes Swope—that a Rule 41(a) dismissal doesn‘t bar a subsequent Rule 54(b) certification—follows directly from the court‘s analysis.⁴

Second, it is the dissent‘s approach that lacks precedential support. The dissent does not identify any court that has adopted its view that Rule 54(b) is unavailable in cases like this—or even questioned the supportive passages from Swope and ITOFCA.⁵

2. The en banc majority explains why the text of Rule 54(b) authorizes district courts to enter partial final judgments as to certain claims and defendants—precisely what the district court did here. So why does the dissent disagree? The dissent appears to offer three theories—all of which, to my mind, are hard to square with the text of Rule 54(b).

First, the dissent says Rule 54(b) cannot apply here because the entry of the Rule 41(a) dismissal as to the remaining defendants constituted the “end of [the] case.” But the text of Rule 54(b) expressly states that “any order or other decision, however designated, that adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties does not end the action as to any of the claims or parties and may be revised at any time before the entry of a judgment adjudicating all the claims and all the parties’ rights and liabilities.” It is unclear to me how a series of interlocutory orders ends the entire action when Rule 54(b) says the opposite.

Second, the dissent focuses on the opening phrase of Rule 54(b): “[w]hen an action presents more than one claim.” The dissent claims that this case does not satisfy that opening phrase because it concludes that there are “no claims pending” once the Rule 41(a) dismissal is entered. The dissent‘s conclusion is hardly compelled by the text—but I see another problem. The dissent‘s theory doesn‘t appear to account for the immediately following phrase “or when multiple parties are involved.” FED. R. CIV. P. 54(b). This case plainly satisfies that condition—and the dissent to its credit does not suggest otherwise. Parties remain “involved” in a case so long as they are “[c]onnected by participation or association” to it. Involved, AMERICAN HERITAGE DICTIONARY (5th ed. 2011). That accurately describes dismissed parties. They remain in the operative complaint. Moreover, as the dissent‘s own cases acknowledge, dismissed parties may continue to participate in the litigation in limited ways, see Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 398 (1990) (allowing for Rule 11 sanction proceedings after a Rule 41(a) dismissal of the complaint)—and their dismissal is subject to later reversal, see Yesh Music v. Lakewood Church, 727 F.3d 356, 363 (5th Cir. 2013) (holding that a “voluntary dismissal [under Rule 41(a)(1)(A)] is subject to vacatur under Rule 60(b)“).

Third and finally, the dissent claims that the majority‘s reading of Rule 54(b) would render surplusage the proviso that a court may only enter partial final judgment if it determines that “there is no just reason for delay.” After all, as the dissent explains, cases like this—where the parties are at the end of the entire case—will always satisfy the requirement that there is no just reason for delay. But there is no surplusage issue unless the majority‘s reading would render the delay proviso “wholly superfluous.” Conn. Nat‘l Bank v. Germain, 503 U.S. 249, 253 (1992) (emphasis added). And the dissent admits it would not, conceding that the delay proviso would still apply when a partial final judgment is entered at any other earlier stage in the litigation. Moreover, there is a common-sense concern with the dissent‘s surplusage argument: Because cases like this will always satisfy the Rule 54(b) delay proviso, cases like this can never satisfy Rule 54(b)? That is difficult to understand.

III.

Finally, the dissent predicts a parade of horribles that they fear will result if we follow Swope and ITOFCA in the admittedly rare circumstance that these issues even arise in the first place.

1. Section III.A of the dissent expresses a fear of “perpetual jurisdiction.” It worries that, under the majority‘s reading of Rule 54(b), cases will never end, and district courts will retain jurisdiction forever.

Two responses: First, as the dissent notes, all it takes to end a case like this is a Rule 54(b) judgment. That‘s not an unreasonable burden. There are usually steps parties or courts must take before they can perfect their appeal. For example, courts must enter judgment in a separate document per Rule 58(b), which is expressly “[s]ubject to Rule 54(b).” But see FED. R. CIV. P. 58(c)(2) (providing for when judgment is entered even if a court fails to comply with the separate document requirement).

Second, both the winning and losing side of a court order have an incentive to request partial final judgment under Rule 54(b), and thus end the district court litigation. Winning parties have an incentive to seek judgment in order to enforce their win. See, e.g., FDIC v. McFarland, 243 F.3d 876, 892–93 (5th Cir. 2001) (absent a Rule 54(b) order, a party who received a favorable, partial judgment could not register it); Redding & Co. v. Russwine Constr. Corp., 417 F.2d 721, 727 (D.C. Cir. 1969) (“We think the role Rule 54(b) plays

with reference to the finality of a judgment for purposes of appeal has implications as regards its finality for purposes of execution as well.”). Losing parties have an incentive to seek partial final judgment in order to appeal the district court’s decision—see, for example, this case.⁶

2. Section III.B.1 of the dissent claims that a dismissal for lack of subject matter jurisdiction is a dismissal without prejudice. And it suggests that this insight is somehow fatal to the majority. But I am not sure the premise is correct. If a claim is dismissed for lack of subject matter jurisdiction, the plaintiff cannot refile that claim unless it first cures the jurisdictional defect. If it cures the defect, res judicata will not bar that claim—but if it does not, it will. See Lopez v. Pompeo, 923 F.3d 444, 447 & n.2 (5th Cir. 2019) (collecting authorities). So as we have explained in another context: “Although plaintiff is not totally precluded from bringing a second suit, he must, nevertheless, prove his case preliminarily to the district court before being allowed the right to relitigate. This requirement, unlike those normally imposed with a Rule 41(a)(2) dismissal, involves prejudice in a legal sense.” LeCompte v. Mr. Chip, Inc., 528 F.2d 601, 604 (5th Cir. 1976).⁷

3. Finally, Section III.B.2 of the dissent expresses concern that the majority’s approach will result in rampant piecemeal appeals. But the whole point of Rule 54(b) is to permit piecemeal appeals so long as the district court approves, just like interlocutory appeals under 28 U.S.C. § 1292(b). See FED. R. CIV. P. 54(b) advisory committee’s note to 1946 amendment (“Rule 54(b) was originally adopted in view of the wide scope and possible content of the newly created ‘civil action’ in order to avoid the possible injustice of a delay in judgment of a distinctly separate claim to await adjudication of the entire case.”).

One other point: The dissent is concerned about piecemeal appeals when a plaintiff dismisses his last remaining claim before seeking a Rule 54(b) order. But what if a plaintiff seeks a piecemeal appeal after dismissing any other claim besides the last one? The dissent expresses no concern with that. Yet one appeal is just as piecemeal as the other.

* * *

We clearly disagree about the proper interpretation of certain terms and phrases. But that is all we disagree about. Textualists can and sometimes do disagree in good faith about the proper interpretation of words and provisions. See, e.g., Gamble v. United States, 139 S. Ct. 1960, 1986 (2019) (Thomas, J., concurring); JCB, Inc. v. Horsburgh & Scott Co., 912 F.3d 238, 242 (5th Cir. 2018) (Ho, J., concurring). But the majority obviously agrees with the dissent that courts cannot exercise jurisdiction where the law does not authorize it. And just as obviously, the dissenters agree with the majority that courts should exercise jurisdiction where the law authorizes it. This is a garden variety disagreement over statutory interpretation—not a constitutional crisis.

The dissenters end with Judge Smith’s dissent in United States v. Castillo-Rivera, 853 F.3d 218 (5th Cir. 2017) (en banc), and thus so will I. It may be “handy in the interest of expediency” to dismiss this appeal for want of jurisdiction—but it “undermines the supervisory responsibility of a federal court of appeals” to do so. Id. at 233 (Smith, J., dissenting).

I agree. That’s why I concur with my respected colleagues in the en banc majority and on the Seventh Circuit. There is no need to bend over backwards to create a loophole where the text imposes none—all so that we can spring a “trap” for litigants to be held captive by federal district courts. The majority faithfully construes the Rules to require greater superintendence, not greater immunity, of federal district courts. The plaintiffs are entitled to appeal.

DON R. WILLETT, Circuit Judge, joined by LESLIE H. SOUTHWICK, Circuit Judge, concurring in judgment:

Now and then a disconnect arises between the law and common sense. Decades ago in Texas, a lawmaker proposed requiring would-be criminals to give their future victims at least 24 hours’ notice, detailing the time/place of the planned offense and the specific constitutional rights they intended to violate.¹ Other legal oddities abound, including the so-called “finality trap,” which has plagued the federal circuits for decades.² How can a case be over yet not final—or, more specifically, not final enough for purposes of appeal yet too final for district court alteration?³ This case, which has commanded the attention of about 20 federal judges over a dozen years, presents an opportunity to remedy the finality trap’s “egregious mess.”⁴ ¡Ya basta!

The majority and dissenting opinions don’t lack for rigor (or vigor).⁵ I agree on remanding to the panel. But along the way, I offer a modest proposal for untangling our “muddled . . . wrong and illogical” precedent that leaves parties mired in litigation limbo.⁶ I would abjure any Miracle Max line-drawing between unappealable “mostly dead” judgments and appealable “all-dead” judgments and let plaintiffs convert the former into the latter.⁷ How? By bindingly disclaiming their right to reassert any dismissed-without-prejudice claims.

* * *

The Williamses’ judgment is stuck in finality-trap purgatory, unable to be fixed by the district court or appealed to us.⁸ The Federal Rules “were not adopted to set traps and pitfalls by way of technicalities for unwary litigants.”⁹ Admittedly, one could quibble over whether “unwary” is the right adjective here, as the Williamses had many tools (over many years) to achieve finality in the district court.¹⁰ Putting that aside, the Rules’ paramount aim is simply stated, if not achieved: “the securing of speedy and inexpensive justice in a uniform and well[-]ordered manner.”¹¹

Any finality-trap remedy must, of course, abide constraints on our jurisdiction. It must preserve § 1291’s finality requirement by disallowing piecemeal appeals and protracted, serial litigation of the same subject matter.¹² And it cannot be allowed to flout other applicable laws or Rules.

Applying these overarching principles:

First, overruling the Ryan rule to let all voluntary dismissals create § 1291 jurisdiction seems a non-starter. We ought not vanquish the finality trap by vanquishing finality.¹³ Second, for what it’s worth, allowing the district court to recapture jurisdiction (two years later) over previously entered voluntary dismissals and convert them from “without prejudice” to “with prejudice” is alien to any circuit.¹⁴ Third, the malleable, case-by-case approach of the Fourth, Sixth, Eighth, and Ninth Circuits enervates § 1291 finality, and predictability, by inviting inconsistent intra-circuit outcomes.¹⁵

As I see it, the litigant-disclaimer solution consistently checks every box:

✅ unsprings the trap;

✅ honors finality principles; and

✅ respects the text of the Rules.

A litigant who wants to appeal a finality-trapped judgment can affirmatively, bindingly, and permanently disclaim the right to reassert dismissed-without-prejudice claims.¹⁶ The Second, Third, and Seventh Circuits get this right.¹⁷ To avoid the trap’s procedural cul-de-sac, plaintiffs simply disavow to the circuit court their right to revive the dismissed claims, which become barred by judicial estoppel and effectively “dismissed with prejudice.”¹⁸ This approach—understandable, administrable, and practicable— is plug-and-play. No required action by the district court. No ad hoc subjectivity. No needless interpretive squabbles over governing statutes or Rules. Just satisfying finality, as the stipulation removes any specter of piecemeal appeals or protracted litigation.¹⁹ The litigant-disclaimer approach elegantly achieves what we set out to do by taking this case en banc.²⁰ It unsprings the trap—fully, formally, and faithfully.

The dissenting opinion frets that the litigant-disclaimer approach is: (1) technically unfaithful to our jurisdictional limitations, thus allowing litigants to “manufacture finality;”²¹ and (2) temporally open-ended, thus allowing litigants to wire around appellate deadlines.²² Respectfully, neither “worry” is genuinely worrisome.

Nobody disputes that finality—our jurisdiction precondition—is essential in every case, big and small. The rule of law is most certainly the law of rules.²³ And generally, the brighter the lines and sharper the corners the better. But for nearly two centuries, the Supreme Court has grasped the futility of a unified theory of finality.²⁴ Within “the ‘twilight zone’ of finality,”²⁵ the Supreme Court has repeatedly cautioned against “rigidity,” recognizing that questions of finality and appealability involve “considerations that always compete.”²⁶ The Court thus instructs us to approach the determination not prudishly but prudently. That is, our duty to ensure finality entails discretion. And since § 1291 often poses “a close question,”²⁷ we must assess finality neither too flexibly nor too inflexibly, but rather with “just right” Goldilocks pragmatism.

True, litigants cannot concoct finality by stipulating away fatal jurisdictional defects.²⁸ “[E]very federal appellate court has a special obligation to ‘satisfy itself . . . of its own jurisdiction . . .’ even though the parties are prepared to concede it.”²⁹ But in fulfilling that obligation, we enjoy a measure of real-world leeway. Section 1291 proscribes our review of non-final decisions but doesn’t prescribe what constitutes finality. So, while I share the dissent’s unease with whims and ipse dixits, such concern is misplaced here. A litigant’s binding disclaimer of voluntarily dismissed claims induces no trickery or gamesmanship—just the opposite. No reassertion means no recapturing of the merits. No serial litigation. No piecemeal appeals. None of the oft-cited debilitating burdens on judicial administration. The litigant-disclaimer approach in no way upends finality; it upholds it.³⁰

As for the dissent’s temporal concern—that there’s no deadline by which a party must disclaim its dismissed-without-prejudice claims—this seems rooted in a belief that “final decision” under § 1291 necessarily means a decision that was final at the time it was made. Respectfully, this collides with the Supreme Court’s centuries-old insistence—first spoken through Chief Justice Marshall³¹ and recently reaffirmed in Microsoft³²—that “finality is to be given a practical rather than a technical construction.”³³ As to timing specifically, the Court stressed 80 years ago that “finality . . . is not a technical concept of temporal or physical termination. It is the means for achieving a healthy legal system.”³⁴ Our focus must always be, the High Court admonishes, on “the competing considerations underlying all questions of finality,”³⁵ specifically, as Justice Jackson wrote for the Court, “the inconvenience and costs of piecemeal review on the one hand and the danger of denying justice on the other.”³⁶ The Court’s direction is plain: Carefully weigh the underlying considerations, and approach the finality determination practically—not technically.³⁷

Indeed, our own cases respect the Supreme Court’s preference for pragmatism. We commonsensibly consider events that occur post-appeal-filing when making “final decision” decisions. Under our settled precedent, for example, where a voluntarily dismissed claim would be time-barred upon refiling, “[w]e treat the dismissal without prejudice as a dismissal with prejudice.”³⁸ That is, where limitations would bar reassertion of a dismissed-without-prejudice claim, we have deemed the without-prejudice dismissal as, functionally, a with-prejudice final decision—the “prejudice” being extrinsic to the court’s order. Take Berry, where we treated a dismissal without prejudice as final and reviewable “as if it were a dismissal with prejudice” because “further litigation of [the dismissed-without-prejudice] claim [would] be time-barred.”³⁹

In short, we have dutifully heeded that Court’s command to tackle finality determinations with levelheaded workability, “a practical rather than a technical” approach that is, hopefully, less sophistic than sophisticated. So too would be a binding disclaimer from a finality-trapped litigant.

* * *

Summing up: I concur in the judgment to remand Williams II to the panel, where I would let the Williamses bindingly disclaim any right to reassert their voluntarily dismissed claims—if they choose to do so.

ANDREW S. OLDHAM, Circuit Judge, joined by SMITH, COSTA, and DUNCAN, Circuit Judges, dissenting:

We took this case en banc to address the so-called “finality trap.” We needed to do so to exorcise the “ghostly magic” that prevents certain dismissals from becoming appealable. Williams v. Taylor Seidenbach, Inc. (Williams II), 935 F.3d 358, 361 (5th Cir. 2019) (Haynes, J., concurring), vacated on rehearing en banc, 941 F.3d 1183 (5th Cir. 2019). For whatever reason, the majority gave up that ghost. Instead, it authorizes a district court to grant a Rule 54(b) motion filed after the conclusion of a case. We didn’t do that in Swope. The Seventh Circuit didn’t do it in ITOFCA either. No federal court has ever done that before—not one time since the adoption of Rule 54(b) in 1939. With greatest respect for my learned colleagues, I dissent.

I.

Today’s decision conflicts with foundational rules and statutes governing our jurisdiction. Let’s start with Rule 41(a). The majority ignores that Rule’s text. And in so doing, it exercises jurisdiction in a way that Article III prohibits.

A.

Rule 41(a) provides that “an action may be dismissed [by court order] at the plaintiff’s request.” FED. R. CIV. P. 41(a)(2) (emphasis added); see also FED. R. CIV. P. 41(a)(1)(A) (setting out the circumstances in which a “plaintiff may dismiss an action”). The Williamses and the district court purported to rely on Rule 41(a) “to dismiss the remaining claim” in the case. Ante, at 7 (majority op.) (emphasis added). That was the first mistake—an action is not synonymous with a claim.

Rule 41(a) allows a plaintiff to dismiss an “action.” FED. R. CIV. P. 41(a)(2). An “action” is the entire lawsuit, not just one part of it. See Action, BLACK’S LAW DICTIONARY (11th ed. 2019) (defs. 3 & 4). A “claim,” by contrast, is a subpart of the action. Id. at Claim (def. 3) (defining “claim” as a “part of a complaint in a civil action”); see also Ryan v. Occidental Petroleum Corp., 577 F.2d 298, 302 n.2 (5th Cir. 1978) (describing a “claim” as “a fragment of an action”).

The Rules employ this distinction. For example, the subsection that immediately follows in Rule 41 gives the defendant the choice between “mov[ing] to dismiss the action or any claim against it” in the event “plaintiff fails to prosecute.” FED. R. CIV. P. 41(b) (emphasis added). And Rule 54 contemplates an instance “[w]hen an action presents more than one claim for relief.” FED R. CIV. P. 54(b); see also Exxon Corp. v. Md. Cas. Co., 599 F.2d 659, 662 (5th Cir. 1979) (“When Rule 41(a) refers to dismissal of an ‘action,’ there is no reason to suppose that the term is intended to include the separate claims which make up an action. When dismissal of a claim is intended, as in Rule 41(b), that concept is spelled out in plain language.” (quotation omitted)).

Rule 41(a) therefore does not allow a plaintiff to pick and choose among claims as candidates for dismissal: Under the Rule, it’s dismissal of the action, or of nothing at all. See Perry v. Schumacher Grp. of La., 891 F.3d 954, 958 (11th Cir. 2018) (“It is clear from the text that only an ‘action’ may be dismissed. There is no mention in the Rule of the option to [dismiss] a portion of a plaintiff’s lawsuit—e.g., a particular claim—while leaving a different part of the lawsuit pending before the trial court.”); Berthold Types Ltd. v. Adobe Sys. Inc., 242 F.3d 772, 777 (7th Cir. 2001) (Easterbrook, J.) (Rule [41(a)] does not speak of dismissing one claim in a suit; it speaks of dismissing ‘an action’—which is to say, the whole case.”); State Treasurer of Mich. v. Barry, 168 F.3d 8, 15 (11th Cir. 1999) (in a case where a party attempted to use Rule 41(a) to dismiss an individual claim, noting that “Rule 41 is not meant for the use the parties in this case and others like it have put it: the rule speaks of voluntary dismissal of ‘an action,’ not a claim”) (quotation omitted)); Harvey Aluminum, Inc. v. Am. Cyanamid Co., 203 F.2d 105, 108 (2d Cir. 1953) (“Rule 41(a) . . . provides for the voluntary dismissal of an ‘action’ not a ‘claim’; the word ‘action’ as used in the Rules denotes the entire controversy, whereas ‘claim’ refers to what has traditionally been termed ‘cause of action.’”).

This inability to dismiss individual claims under Rule 41(a) is not as harsh as it may at first seem. As the majority notes, ante, at 7, the Rules provide other routes to a plaintiff looking to dismiss undesirable claims or parties. Rule 15 explains how a party may “amend its pleading.” FED. R. CIV. P. 15(a)(1); see also Perry, 891 F.3d at 958 (“There are multiple ways to dismiss a single claim without dismissing an entire action. The easiest and most obvious is to seek and obtain leave to amend the complaint to eliminate the remaining claim, pursuant to Rule 15.”); Ryan, 577 F.2d at 302 n.2 (noting that “the plaintiff’s elimination of a fragment of an action . . . is more appropriately considered to be an amendment to the complaint under Rule 15”). And Rule 21 lets the district court, “[o]n a motion or on its own, . . . add or drop a party.” FED. R. CIV. P. 21; see also Harvey Aluminum, 203 F.2d at 108 (suggesting that plaintiffs use Rule 21, rather than Rule 41, to eliminate a party from a multiparty action). We need not stretch Rule 41(a) beyond its text to do the work of other Rules.

B.

When a plaintiff sidesteps these other routes and chooses to proceed via Rule 41(a), he chooses to dismiss the entire action—every single claim in the case. That choice has consequences: It deprives us of jurisdiction to hear the appeal.

For a case or controversy to engage our Article III jurisdiction, there must be adversity, “extant at all stages of review, not merely at the time the complaint is filed.” Arizonans for Official English v. Arizona, 520 U.S. 43, 67 (1997) (quotation omitted). A case becomes moot, and non-adverse, “when the issues are no longer live . . . [n]o matter how vehemently the parties continue to dispute the lawfulness of the conduct that precipitated the lawsuit . . . .” Already, LLC v. Nike, Inc., 568 U.S. 85, 91 (2013) (quotation omitted).

One way a party can moot a case is to agree not to proceed with it. That’s why a settlement terminates Article III jurisdiction. See Summers v. Earth Island Inst., 555 U.S. 488, 494 (2009). So does the issuance of “a covenant not to sue.” Already, LLC, 568 U.S. at 90. Under this rubric, a voluntary dismissal with prejudice plainly moots a case: “When the plaintiffs ask[] the District Court to dismiss their claims, they consent[] to the judgment against them and disavow[] any right to relief . . . .” Microsoft Corp. v. Baker, 137 S. Ct. 1702, 1717 (2017) (Thomas, J., concurring in the judgment).

So too for voluntary dismissals without prejudice. Although the plaintiff “is entitled to bring a later suit on the same cause of action,” he has agreed that this one should go no further. Ryan, 577 F.2d at 302. It follows, then, that a Rule 41(a) dismissal represents a voluntary abandonment of the entire action, and deprives us of jurisdiction over any element of the dispute. If, as the parties and the majority assume, we are dealing with an appeal from a Rule 41(a) dismissal, we have no constitutional authority to hear this case.

C.

The majority does not disagree with this interpretation of Rule 41(a). It merely feigns surprise that I would raise an argument not raised by the parties themselves. But a correct understanding of Rule 41(a) has jurisdictional consequences, so we have an independent obligation to get it right, regardless of the parties’ views. As the Supreme Court has repeatedly reminded us: “Every federal appellate court has a special obligation to satisfy itself not only of its own jurisdiction, but also that of the lower courts in a cause under review, even though the parties are prepared to concede it.” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 95 (1998) (quotation omitted). So the majority cannot ignore the problem just because the parties did so.

Nor does it matter that we’ve previously misread Rule 41(a). See, e.g., Plains Growers, Inc. v. Ickes-Braun Glasshouses, Inc., 474 F.2d 250 (5th Cir. 1973). In Plains Growers, we rejected a textualist approach to Rule 41(a)(1) and pursued instead “[t]he purpose of the Rule.” Id. at 254. At least in 1973, we found “[t]he purpose” of the Rule was “to permit the plaintiff voluntarily to take the case out of court if no other party will be prejudiced.” Ibid.¹ So we allowed a plaintiff to use Rule 41 to dismiss a single defendant without dismissing the entire “action” because, in our view, that was the result “intended by the rule-makers.” Id. at 255. Such purposivist-cum-intent-of-the-framers reasoning was quite common in the 1970s. See, e.g., Train v. Colo. Pub. Interest Research Grp., Inc., 426 U.S. 1, 10 (1976) (reversing the court of appeals because it “excluded reference to the legislative history of the [law] in discerning its meaning”); Philbrook v Glodgett, 421 US 707, 713 (1975) (“Our objective . . . is to ascertain the congressional intent and give effect to the legislative will.”); ANTONIN SCALIA & BRYAN GARNER, READING LAW 383 (2012) (describing the 1970s and 1980s as the “heyday of legislative history”). But even in those days, when we adopted a purpose-first approach to Rule 41, we found the textualist argument “persuasive.” Plains Growers, 474 F.2d at 254.

The text of Rule 41(a) is more than persuasive; it’s determinative. Purposivism has been out of fashion for a long time. And the whole reason we went en banc in this case was to revisit our jurisdictional precedents—as we have many times in the past. See, e.g., Williams v. Catoe, 946 F.3d 278 (5th Cir. 2020) (en banc) (overturning a 1985 panel opinion on appellate jurisdiction over district court orders denying counsel to civil plaintiffs); Smallwood v. Illinois Cent. R.R. Co., 385 F.3d 568 (5th Cir. 2004) (en banc) (reconsidering panel precedent on district court jurisdiction over certain in-state defendants in diversity cases); Marathon Oil Co. v. Ruhrgas, 145 F.3d 211, 215 (5th Cir. 1998) (en banc) (using en banc review to “reconcile the conflicting circuit precedent” regarding the adjudication of motions to dismiss for lack of subject-matter and personal jurisdiction), rev’d 526 U.S. 574 (1999); Newpark Shipbuilding & Repair, Inc. v. Roundtree, 723 F.2d 399 (5th Cir. 1984) (en banc) (harmonizing circuit precedent on the finality requirement for review of Benefits Review Board orders); United States v. Mendoza, 581 F.2d 89 (5th Cir. 1978) (en banc) (going en banc to review panel precedent concerning district court jurisdiction over motions to reduce sentence).

We have an independent obligation to right our jurisdictional wrongs. This case provides us with an opportunity to do so. It is therefore of no moment that the parties have nothing to say about the issue, or that outmoded panel precedent goes the other way.

II.

Next, let’s talk about Rule 54(b). The majority’s approach finds no support in that Rule’s text, its 81-year history, or precedent.

A.

The text of Rule 54(b) says: “When an action presents more than one claim for relief—whether as a claim, counterclaim, crossclaim, or third-party claim—or when multiple parties are involved, the court may direct entry of a final judgment as to one or more, but fewer than all, claims or parties . . . .” FED. R. CIV. P. 54(b) (emphases added). In the majority’s view, a Rule 54(b) certification is available even after every single claim pending before the district court has been dismissed.

No. Since the adoption of Rule 54(b) in 1939,² that Rule has never authorized an appeal at the end of a case. As Justice Jackson explained, Rule 54(b) allows an appeal before “the determination of all issues as to all parties” and “before a final judgment can be had.” Dickinson v. Petroleum Conversion Corp., 338 U.S. 507, 511 (1950) (emphasis added).

The italicized phrases in Rule 54(b) above tell you everything you need to know. First, “[w]hen an action presents more than one claim.” That is, the action must—in the present tense—present more than one claim. It’s insufficient that the action might have presented more than one claim in the past. It must present more than one claim now. Of course, if the plaintiff has dismissed the action under Rule 41(a), then the action presents zero claims—there isn’t an action at all.

Let’s imagine, then, that the Williamses instead used Rule 15 or 21 to modify their complaint and remove the remaining, unresolved claims from the action. Then what? Rule 54(b) kicks in only when the court can direct entry of final judgment “as to one or more, but fewer than all” the claims. Obviously, the district court can do that only if other claims remain pending. When the plaintiff instead uses Rules 15 and 21 to strip out the unresolved parties and claims from the complaint,³ there are no other claims “present[ed]” in the action. No Rule 54(b) order is permissible, or even necessary: At the end of the case, when all the claims have been dismissed, the district court is no longer directing entry of judgment on “fewer than all claims”; it’s just entering judgment. See FED. R. CIV. P. 58; 10 JAMES W. MOORE ET AL., MOORE’S FEDERAL PRACTICE § 54.22[2][c], at 73 (3d ed. 2020) [hereinafter MOORE’S FEDERAL PRACTICE] (“[I]f the plaintiff voluntarily dismisses the unadjudicated claims against the other defendants, an earlier order as to one or more but fewer than all defendants becomes appealable despite the absence of a Rule 54(b) judgment.”).

Perhaps the most damning piece of text is the one thing the district court must find in every Rule 54(b) certification: “that there is no just reason for delay.” FED. R. CIV. P. 54(b). Now that makes sense when Rule 54(b) is applied—as it always has been—to partial appeals. But what possible work

does the no-just-reason-for-delay requirement serve when all the claims in the case have already been decided? Of course there‘s no reason to delay the appeal then—because there‘s nothing left to do but appeal. Imagine getting to the end of a movie, watching the credits run to the end, and then seeing a screen permitting you to leave the theater “if there‘s no just reason for delay.” What else were you going to do?

After today, it‘s unclear how to apply the current no-just-reason-for-delay caselaw, all of which is predicated on the existence of undismissed claims pending in the district court. Consider:

• The Supreme Court says that the district court, while mulling over a Rule 54(b) motion, may consider the amount of time “the rest of the litigation would be expected to continue.” Curtiss-Wright Corp. v. Gen. Elec. Co., 446 U.S. 1, 11 (1980). Of course, that makes no sense if there‘s nothing left to litigate in the district court.
• Courts of appeals should “scrutinize the district court‘s evaluation of such factors as the interrelationship” between the adjudicated claims and the unadjudicated claims. Id. at 10. How can they, when there are no unadjudicated claims outstanding?
• Some courts ask whether “the need to review the issues currently appealed may be mooted by future developments in the district court.” Sussex Drug Prods. v. Kanasco, Ltd., 920 F.2d 1150, 1156 (3d Cir. 1990); see also Consol. Rail Corp. v. Fore River Ry. Co., 861 F.2d 322, 326 (1st Cir. 1988) (similar). How does that work when there‘s nothing left to develop in the district court?

I suppose we‘ll have to wait for another en banc petition to figure out the answers to these questions.

B.

The majority attempts to square its position with the text of the Rule. It can‘t.

The majority‘s textualist argument for its use of Rule 54(b) goes something like this: (1) a dismissed claim technically remains part of the complaint, “absent amendment,” ante, at 9; (2) so even when all claims have been dismissed, an action may still “present[] more than one claim for relief,” FED. R. CIV. P. 54(b); and (3) that means, even when the entire case is dismissed, “fewer than all the claims” have been “adjudicate[d]“—so Rule 54(b) certification is available, ibid.; see ante, at 9 (“[A] voluntary dismissal of some defendants under Rule 41(a) at the end of the case does not alter the court‘s Rule 54(b) authority in any way.“). Every single proposition in that argument is wrong.

Proposition (1) is wrong because a Rule 41(a) dismissal gets rid of the entire action, so there‘s no operative complaint to speak of. See supra Part I.A. Proposition (2) is wrong because, when all claims have been dismissed, the action no longer “presents more than one claim for relief“—it presents none at all. See supra Part II.A. Still, even if (1) and (2) are correct, Proposition (3) is wrong: a court-ordered dismissal under Rule 41(a) of all remaining claims “adjudicates” those claims, foreclosing Rule 54(b) relief.

Rule 54(b) is clear: “any order or other decision . . . that adjudicates fewer than all the claims or the rights and liabilities of fewer than all the parties does not end the action . . . and may be revised at any time before the entry of a judgment . . . .” FED. R. CIV. P. 54(b). So if an order does adjudicate whatever claims remain, it ends the action and may not be revised. To “adjudicate” simply means “[t]o rule on judicially.” Adjudicate, BLACK‘S LAW DICTIONARY (11th ed. 2019). That word, as it is used by the Federal Rules of Civil Procedure, requires only a decision by the court—it does not require a decision on the merits. That‘s why the Rules take pains to indicate when an adjudication is also a merits decision. See, e.g., FED. R. CIV. P. 41(a)(1)(B) (providing that a plaintiff‘s first voluntary dismissal of a claim is “without prejudice,” but a second dismissal of that claim “operates as an adjudication on the merits” (emphasis added)).

In this case, the district court dismissed the action under Rule 41(a). That obviously constituted a judicial ruling. It follows, then, that the Rule 41(a) dismissal “adjudicated” the plaintiffs’ claims. And because there were no claims pending after that adjudication, Rule 54(b) was (and still is) completely irrelevant. See 10 CHARLES ALAN WRIGHT ET AL., FEDERAL PRACTICE AND PROCEDURE § 2656 (4th ed.) [hereinafter WRIGHT & MILLER] (observing that “no [Rule 54(b)] certificate is needed if no claims remain pending“).

C.

Precedent does not save the majority, either. The Supreme Court has told us: Once all claims have been dismissed, the “action is no longer pending“—even if the dismissals are voluntary and without prejudice. Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 395 (1990). And the termination of the action through voluntary dismissal means that the district court loses the power to tinker with the merits of the case. Id. at 396.

This unambiguous description of district court jurisdiction makes swift work of the majority‘s view that a voluntarily dismissed claim remains pending. Compare ante, at 9 (“A dismissed claim remains a part of the case . . . and that is so regardless of when the Rule 41(a) dismissal occurs.“), with Yesh Music v. Lakewood Church, 727 F.3d 356, 360 & n.5 (5th Cir. 2013) (observing that a “voluntary dismissal terminates, closes, and ends [plaintiff‘s] cause of action,” and that “it ends the pending action” (quotation omitted)), and DIRECTV, Inc. v. Budden, 420 F.3d 521, 524–26 (5th Cir. 2005) (considering Rule 54(b) inapplicable when summary judgment had been granted on some claims and “the remaining claims ha[d] already been abandoned“).

Against the text of Rule 54(b) and an unbroken line of cases spanning the Rule‘s 81-year lifetime, the majority offers two authorities: our decision in Swope v. Columbian Chems. Co., 281 F.3d 185 (5th Cir. 2002), and the Seventh Circuit‘s decision in ITOFCA, Inc. v. MegaTrans Logistics, Inc., 235 F.3d 360 (7th Cir. 2000). But the very most that can be said about Swope is that it was a drive-by jurisdictional ruling on its way to an altogether different question about premature notices of appeal. And ITOFCA was not a Rule 54(b) case at all.

Take Swope. In that case, the district court granted summary judgment on fewer than all claims and left one claim pending. 281 F.3d at 190. The plaintiffs then filed three things: (1) a notice of appeal from the summary-judgment rulings, (2) a request that the district court certify those rulings as final judgments under Rule 54(b), and (3) a Rule 41(a) motion to dismiss the one undismissed claim. Id. at 190–91. The district court granted (2) and (3). Ibid. But defendants moved to dismiss the appeal based on (1): They argued we lacked jurisdiction “because appellants’ notice of appeal preceded the trial court‘s designation of its summary judgment decisions as final.” Id. at 191.

Our decision likewise focused on (1)—the premature notice of appeal. We recognized that a notice of appeal should ordinarily be filed only after a final decision, not before it. Ibid.; see FED. R. APP. P. 4(a)(1)(A) (“[T]he notice of appeal . . . must be filed with the district clerk within 30 days after entry of the judgment or order appealed from.” (emphasis added)). But the panel permitted the appeal to proceed anyway, applying a narrow exception “that a premature notice of appeal is effective if Rule 54(b) certification is subsequently granted.” Swope, 281 F.3d at 191 (quotation omitted). The rest of Swope‘s analysis was the sort of “drive-by jurisdictional ruling[]” that has “no precedential effect.” Steel Co., 523 U.S. at 91.

And as drive-by jurisdictional decisions go, this one undercuts the majority in three important ways. First, the Rule 54(b) motion in Swope was filed before the Rule 41(a) motion was granted. That timing matters. Before the district court granted the Rule 41(a) motion, it still had jurisdiction over the case. See supra Part I. And before the district court granted the Rule 41(a) motion, it had “adjudicate[d] fewer than all the claims” under Rule 54(b).

Second, the Swope court agreed with my reading of Rule 41(a). See 281 F.3d at 192 n.15. It recognized that the text of ”Rule 41(a) contemplates dismissal of an ‘action’ rather than a ‘claim’ or ‘claims.‘” Ibid. And it recognized that our court previously interpreted Rule 41(a) to permit only dismissal of actions, not claims. See ibid. (citing, inter alia, Ryan, 577 F.2d at 302 n.2). The Swope court thought it needn‘t address this problem because the district court granted its Rule 54(b) certification and its Rule 41(a) dismissal on the same day, and the former sufficed to create finality. See 281 F.3d at 192 & n.15. That obviously says nothing about our problem—where the Williamses filed their Rule 54(b) motion years after dismissing the entire “action” under Rule 41(a). And in all events, Swope‘s agreement with my reading of Rule 41(a) makes the majority‘s embrace of the decision quite awkward.

Finally, as the majority concedes, the Swope panel expressly declined to consider how or why a valid Rule 41(a) dismissal might deprive the district court of jurisdiction and thereby render a subsequent Rule 54(b) certification a nullity. See ante, at 9 (quoting Swope, 281 F.3d at 192 n.15). That‘s understandable—the Swope panel was focused on the interplay between Rule 54(b) and notices of appeal, not between Rules 41(a) and 54(b). But it also means that Swope does not support, much less dictate, today‘s result.

Even less relevant is ITOFCA. See ante, at 10 (relying on the decision). The district court in that case did not enter a Rule 54(b) certification at all, so the Seventh Circuit had no occasion to exercise jurisdiction over such a certification. In fact, the Seventh Circuit held it had no jurisdiction at all and dismissed the appeal of a non-final decision under § 1291. So the only part of the ITOFCA decision that today‘s majority invokes is its dicta about some other hypothetical case that might be appealable under § 1291. And even that dicta was equivocal: The Seventh Circuit said ”Rule 54(b) would seem to be an adequate alternative means of gaining appellate jurisdiction” in some other hypothesized case. 235 F.3d at 364 (emphasis added). That subjunctive aside is hardly a firm foundation for today‘s decision.

All this is to say: If we went en banc to ensure that our court‘s decisions are uniform and correct, see FED. R. APP. P. 35(a), uncritical reliance on Swope and ITOFCA isn‘t going to cut it.

III.

If the majority opinion is remarkable for what it does not do—follow the plain meaning of Rule 54(b), as recognized in an unbroken line of decisions spanning eight decades—it is more remarkable for what it does do. The new rules of civil procedure announced today will confer on district courts perpetual jurisdiction over long-dismissed claims. It also permits an end-run around the limits of our appellate jurisdiction set by Congress.

A.

Until today, the contours of district court jurisdiction were well-defined: While claims are pending, a court may exercise its authority over an action. See Cooter & Gell, 496 U.S. at 395. Once all claims are dismissed, however, the court may only “consider collateral issues,” such as Rule 11 sanctions and requests for attorney‘s fees. Ibid. No more. The majority‘s regime imbues district courts with perpetual jurisdiction over long-dead claims.

Imagine, as the majority does, that Rule 41(a) permits piecemeal dismissals of individual claims within a multi-claim action. And suppose a plaintiff brings an action with three claims. Claim 1 and Claim 2 are adjudicated at summary judgment, but Claim 3 survives and remains pending. If the parties stipulate to dismiss Claim 3 under FED. R. CIV. P. 41(a)(1)(A)(ii), it is well-settled that, “[a]fter the dismissal, the action no longer is pending in the district court and no further proceedings in the action are proper.” 9 WRIGHT & MILLER § 2367 (3d ed.); see also ibid. (collecting cases). That‘s because “[a] stipulation of dismissal under that rule ordinarily—and automatically—strips the district court of subject-matter jurisdiction.” Nat‘l City Golf Fin. v. Scott, 899 F.3d 412, 415–16 (5th Cir. 2018). This rule is so airtight that “any action by the district court” on the merits of the case “after the filing of such a stipulation can have no force or effect . . . .” SmallBizPros, Inc. v. MacDonald, 618 F.3d 458, 463 (5th Cir. 2010). So, under settled, near-unanimous precedent, there‘s nothing more to be done. The case is over.

Now consider the majority‘s approach. Under that understanding of Rule 54(b), the case is never final because there‘s no “final judgment,” all claims remain pending before the district court, and that court retains jurisdiction over those claims forever. What‘s to stop a plaintiff from moving for a Rule 54(b) order years after dismissal of the entire action? Or a district court from granting it? That‘s precisely what happened in this case, and that‘s precisely the sort of never-ending litigation the majority opinion condones.

It‘s unclear how appeals are even supposed to work in this brave new world. In our hypothetical case, once Claim 3 is dismissed, when does the 30-day clock for filing a notice of appeal start to run? It should be obvious that the answer is when Count 3 is dismissed—the whole reason we went en banc was because current Fifth Circuit precedent (unmodified by today‘s opinion) tells us that the voluntary dismissal of the last remaining claim makes the action “final” within the district court. See Williams II, 935 F.3d at 361 (Haynes, J., concurring). But, again, under the majority‘s approach, because there‘s no “final judgment,” the action never becomes final in the district court, and that court retains jurisdiction in perpetuity—at least until it grants a Rule 54(b) order.

B.

These problems hint at another unacceptable result of the majority opinion: The greatly diminished role of 28 U.S.C. § 1291, a congressionally imposed limit to our appellate jurisdiction.

1.

The majority‘s theory is that only “final judgments” may proceed through § 1291, while other final orders need Rule 54(b) certification to be appealable. Wrong again.

Congress gave us appellate jurisdiction over the “final decisions of the district courts of the United States.” 28 U.S.C. § 1291 (emphasis added). As a general matter, “[d]ismissals without prejudice are canonically non-final and hence not appealable under 28 U.S.C. § 1291.” Am. States Ins. Co. v. Capital Assocs. of Jackson Cty., Inc., 392 F.3d 939, 940 (7th Cir. 2004) (Easterbrook, J.). That is so because “without prejudice” is shorthand for “without prejudice to refiling.” See Dismissal, BLACK‘S LAW DICTIONARY (11th ed. 2019) (defining “dismissal without prejudice” as “[a] dismissal that does not bar the plaintiff from refiling the lawsuit . . . .“). And the very potentiality of refiling makes the decision not “final.” See Ryan, 577 F.2d at 302.

As with all general rules, however, this one has an exception. Some without-prejudice dismissals are final. Take subject-matter jurisdiction, for example. If a party fails to allege subject-matter jurisdiction, it fails to properly invoke the machinery of the Article III courts. And when that happens, the plaintiff is not entitled to a judgment on the merits or a with-prejudice dismissal.⁴ It‘s entitled to nothing. That‘s why we merely announce the fact and dismiss. Ex parte McCardle, 74 U.S. (7 Wall.) 506, 514 (1868). That decision is obviously “final” because the district court is forever done with it. So the plaintiff can appeal under § 1291. See, e.g., 16 Front Street, LLC v. Miss. Silicon, LLC, 886 F.3d 549, 561 (5th Cir. 2018).

The text of § 1291 demands the distinction between “final decisions” and “final judgments.” Congress could have written § 1291 to give us appellate jurisdiction only over “final judgments.” That obviously would prevent appeals where the district court dismissed the claim for lack of jurisdiction and hence for lack of judicial power to enter a judgment. Of course, that‘s not the statute Congress wrote. It used the broader term “final decisions“—and with it, gave us judicial power over some without-prejudice dismissals, and some orders that aren‘t dismissals at all. See Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546 (holding that § 1291 permits appeals from certain decisions “collateral to[] rights asserted in the action,” even when there‘s no final judgment); Mitchell v. Forsyth, 472 U.S. 511, 530 (1985) (holding “that a district court‘s denial of a claim of qualified immunity . . . is an appealable ‘final decision’ within the meaning of 28 U.S.C. § 1291 notwithstanding the absence of a final judgment“).

The line between final and non-final dismissals can be difficult to draw. See, e.g., 15A WRIGHT & MILLER § 3914.6 (2d ed.). And this difficulty can lead to mistakes in the district court. Often, for example, a district court will announce its lack of jurisdiction, dismiss the case without prejudice, and then purport to enter a “final judgment” under Rule 58. See, e.g., Warren v. Am. Bankers Ins. of Fla., 507 F.3d 1239, 1243 (10th Cir. 2007) (opining that the district court should have entered a Rule 58 judgment after dismissing the case for lack of subject-matter jurisdiction). That last step is obviously wrong because, again, the lack of jurisdiction means the lack of federal judicial power to do anything beyond announce the fact and dismiss. See Am. Nat‘l Bank of Jacksonville v. FDIC, 710 F.2d 1528, 1535–36 (11th Cir. 1983) (observing that final judgments are unnecessary when an action is dismissed for lack of subject-matter jurisdiction). True, the entry of a purported “judgment” clearly communicates to the parties and to us that the district court is done and that its decision is “final” for purposes of § 1291. The important point for present purposes, however, is that the existence of a without-prejudice dismissal—standing alone—tells us nothing about whether it‘s appealable under § 1291.

The fact that some without-prejudice dismissals may create appealable final decisions under § 1291 poses another problem for the majority. The majority‘s theory is that without-prejudice dismissals are never final or appealable until “final judgment” has been entered, see ante, at 8, hence the need for Rule 54(b) certification. But if dismissals for lack of subject-matter jurisdiction don‘t need an accompanying final judgment to be final and appealable, how can the majority be correct? It can‘t and it‘s not. A dismissal that gets rid of all claims pending before the district court strips away the district court‘s power to enter a Rule 54(b) order, judgment or no judgment.

2.

All this is to say: The majority‘s misreading vastly expands Rule 54(b)‘s reach. In so doing, our court impermissibly undermines the policy decisions Congress made when it enacted 28 U.S.C. § 1291.

The existence of our Court is constitutionally permitted but not constitutionally required. See U.S. CONST. art. III, § 1. By virtue of the Madisonian Compromise, “the creation of lower federal courts” was “left to the wisdom of Congress.” Haywood v. Drown, 556 U.S. 729, 746 (2009) (Thomas, J., dissenting). And “it [is] entirely discretionary with Congress to what extent it . . . vest[s] Federal judicial power in them.” Ibid. (quotation omitted).

“The jurisdiction of the courts of appeals over decisions of the district courts is controlled primarily by 28 U.S.C. § 1291.” 10 MOORE‘S FEDERAL PRACTICE § 54.20[2], at 8. In limiting our appellate review to “final decisions” of district courts, 28 U.S.C. § 1291, Congress codified the longstanding “general rule . . . that the whole case and every matter in controversy in it must be decided in a single appeal,” Microsoft, 137 S. Ct. at 1712 (quotation omitted). Section 1291 therefore “preserves the proper balance between trial and appellate courts, minimizes the harassment and delay that would result from repeated interlocutory appeals, and promotes the efficient administration of justice.” Ibid.

Rule 54(b) should do the same thing. Recognizing that multi-party, multi-claim actions had become more common in federal court, Congress and the Supreme Court promulgated Rule 54(b) to allow for appeals from some final judgments—even as litigation continues in district court. Sears, Roebuck & Co. v. Mackey, 351 U.S. 427, 431–33 (1956). Still, because piecemeal appeals “undermine[] efficient judicial administration,” Mohawk Indus., Inc. v. Carpenter, 558 U.S. 100, 106 (2009) (quotation omitted), the Supreme Court has “repeatedly” warned against attempts by plaintiffs and lower courts to “erode the finality principle and disserve its objectives,” Microsoft, 137 S. Ct. at 1712; see also id. at 1714 (“These changes are to come from rulemaking, . . . not judicial decisions . . . or inventive litigation ploys.“).

For example, in Microsoft v. Baker, the plaintiffs, who had been unsuccessful in procuring class certification, voluntarily “dismiss[ed] their claims with prejudice—subject . . . to the right to ‘revive’ those claims if the denial of class certification [wa]s reversed on appeal.” 137 S. Ct. at 1715. In other words, plaintiffs wanted the benefits of finality (an appeal of right under § 1291) without any of the burdens that come with a with-prejudice dismissal (res judicata and the related preclusion doctrines). No dice, the Supreme Court said: That‘s not a final decision at all. Ibid. If it were, “Congress’ final decision rule would end up a pretty puny one.” Ibid. (quotation omitted).

Microsoft tells us that if a plaintiff dismisses his remaining claims and brings litigation to an end in the district court, § 1291 is his only door to an appeal. And that door is only open if the dismissal truly creates a “final decision.” That is the balance Congress chose to strike. We have no power to strike a different one.

The rule that voluntary dismissals without prejudice are ordinarily not appealable therefore effectuates Congress‘s choice. That‘s because the existence of a without-prejudice dismissal does not stop the plaintiff from filing the exact same claim again, undermining the one-appeal rule. See, e.g., State Treasurer of Mich., 168 F.3d at 13–14 (holding the court lacked jurisdiction over the appeal because “it [wa]s possible that the claim dismissed without prejudice w[ould] be refiled,” so exercising jurisdiction “would undermine the policies of judicial efficiency [and] avoiding piecemeal litigation” (quotation omitted)). With-prejudice dismissals, on the other hand, have res judicata implications and therefore prevent gamesmanship by plaintiffs and their lawyers. Cf. Ryan, 577 F.2d at 301–02. Again: If a party wants to dismiss its remaining claims and then appeal, Congress and precedent require those dismissals to be with-prejudice, and the appeal to be taken through § 1291.

The majority‘s decision cannot be squared with these rules. Under today‘s new approach to Rule 54(b), a plaintiff need not be concerned that his last remaining claim is dismissed without prejudice. Instead of appealing under § 1291, he can just secure a Rule 54(b) order from the district court, despite the absence of any pending litigation in the district court, and despite the absence of district court jurisdiction now that the litigation is finished. If the plaintiff fares well on appeal, he can re-file his voluntarily dismissed claim without any fear of claim preclusion. If the appeal fails, he can re-file the voluntarily dismissed claims anyway. Either way, the plaintiff has lost nothing from his dismissal, but the courts are burdened with piecemeal litigation and appeals. And the plaintiff gets all this without ever having to satisfy the finality requirement of § 1291. That result was impermissible in Microsoft. It‘s equally impermissible here.

IV.

I also worry about the “litigant-disclaimer solution.” Ante, at 27 (Willett, J., concurring). As I understand it, the concurrence would allow parties to manufacture finality by “disclaim[ing] the right to reassert dismissed-without-prejudice claims.” Ibid. But I cannot find anything in the Federal Rules or § 1291 that allows a party‘s say-so to turn a non-final decision into a final one. Rather, we must determine our judicial power in every case, irrespective of what the parties might say or do:

The parties cannot confer on a federal court jurisdiction that has not been vested in that court by the Constitution and Congress. This means that the parties cannot waive lack of subject matter jurisdiction by express consent, or by conduct, or even by estoppel. The subject matter jurisdiction of the federal courts is too fundamental a concern to be left to the whims and tactical concerns of the litigants.

13 WRIGHT & MILLER § 3522 (3d ed.) (emphasis added; footnotes omitted). That‘s why we don‘t let parties stipulate away jurisdictional defects. See, e.g., Frank v. Gaos, 139 S. Ct. 1041, 1046 (2019) (per curiam) (noting that, even when parties agree to settle, a court may not exercise the judicial power while questions of standing remain unresolved); Nike, 568 U.S. at 91 (explaining that a plaintiff may not overcome mootness merely by asserting that the case isn‘t moot); Steel Co., 523 U.S. at 95 (requiring courts to ignore the parties’ jurisdictional concessions and make an independent determination as to jurisdiction). Yet the “litigant-disclaimer solution” would convert a decision over which we have no jurisdiction into one over which we do, through nothing more than an appellant‘s ipse dixit.

Moreover, I worry that the “litigant-disclaimer” proposal causes the problem it seeks to avoid: “vanquish[ing] the finality trap by vanquishing finality.” Ante, at 26 (Willett, J., concurring). When is the deadline for a party to transmute-by-disclaimer a without-prejudice dismissal into a with-prejudice final decision? If this case is anything to go by, the answer is “never“: The without-prejudice dismissal at issue here was entered in November 2016. Yet the concurrence proposes allowing the Williamses to “bindingly disclaim any right to reassert their voluntarily dismissed claims” on remand to a three-judge panel of this court—no matter that this disclaimer may arrive almost four years after the dismissal. Id. at 32.

Alternatively, if the normal rules apply and the plaintiff must disclaim and appeal within the 30-day deadline to file a notice of appeal from the dismissal, see FED. R. APP. P. 4(a)(1)(A) & 28 U.S.C. § 2107(a), that time came and went three years ago—without any mention of a disclaimer, see Williams v. Lockheed Martin, Corp., No. 16-31238 (5th Cir. Feb. 15, 2017) (per curiam) (dismissing appeal from the Rule 41(a) dismissal after the Williamses “concede[d] that a Rule 54 order is necessary to invoke this court‘s jurisdiction“). That was the time to determine the finality of the district court‘s decision. And the Williamses failure to hew to that 30-day deadline carries jurisdictional consequences today. See Bowles v. Russell, 551 U.S. 205, 213 (2007).

* * *

We can never forget that the rule of law is the law of rules. See Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. CHI. L. REV. 1175 (1989). And few rules are more important than the ones that limit our jurisdiction and hence legitimize our exercise of it. This case gave us a chance to address the inconsistencies in our jurisdictional precedents, but instead we‘ve made them worse. I hope that we‘ll come back and fix these errors. Cf. United States v. Castillo-Rivera, 853 F.3d 218, 237 (5th Cir. 2017) (en banc) (Smith, J., dissenting) (observing, of an en banc court‘s refusal to tackle an issue presented for en banc consideration, that “[r]efusing to take this fork in the road is the easy way, but not the right one“). Until then, I respectfully dissent.

Notes

1

Technically, Rule 54(b) allows a district court to “direct entry” of judgment—the actual entry of judgment occurs under Rule 58. But since district courts must “mechanically appl[y]” Rule 58, United States v. Indrelunas, 411 U.S. 216, 222 (1973) (per curiam), we use some variation of the shorthand “entry of judgment under Rule 54(b).” The dissent makes special effort to point out when it cites an opinion associated with Judge Easterbrook. It may be notable to them, then, that Judge Easterbrook joined ITOFCA. H.B. 110, 63rd Reg. Sess. (Tex. 1973). Even that framing of the Rule’s “purpose” requires the plaintiff to “take the case”—not just part of the case—“out of court.” Plains Growers, 474 F.2d at 254 (emphasis added). And even on its own terms, it’s unclear how Plains Growers understood the Rule’s purpose to conflict with its text. The court suggested that “action” must mean the same thing in Rule 41(a)(1) (which allows the plaintiff voluntarily to dismiss the “action” without leave of court at the outset of the case) and in Rule 41(a)(2) (which requires leave of court to dismiss the “action” on the plaintiff’s motion after the outset of the case). Ibid. That’s obviously correct; “action” means the same thing in both subsections. But then the court said that if the plaintiff can use Rule 41(a)(1) to dismiss only the entire “action,” then Rule 41(a)(2) somehow would prevent the court from dismissing less-than-the-entire “action.” See ibid. It’s true that a court can dismiss only an entire “action” under Rule 41(a)(2), but it’s obviously not true (as Plains Growers appeared to worry) that courts are therefore powerless to “dismiss the action as to less than all defendants upon motion.” Ibid. The court has that power under all sorts of other Rules, including Rules 12, 21, 54, and 56—just not Rule 41.

2

Of course, if a plaintiff voluntarily dismisses an entire case under Rule 41(a), then the case is over and there is nothing left pending in the district court. See SmallBizPros, Inc. v. MacDonald, 618 F.3d 458, 461 (5th Cir. 2010) (per curiam). See also Edmo v. Corizon, Inc., 949 F.3d 489, 506 (9th Cir. 2020) (Bumatay, J., dissenting from the denial of rehearing en banc) (“As inferior court judges, we are bound by Supreme Court precedent[s]. . . . [But] ‘[w]e should resolve questions about the scope of those precedents in light of and in the direction of the constitutional text and constitutional history.‘“) (fourth alteration in the original) (quoting Free Enter. Fund v. Public Co. Accounting Oversight Bd., 537 F.3d 667, 698 (D.C. Cir. 2008) (Kavanaugh, J., dissenting)), aff‘d in part, rev‘d in part, 561 U.S. 477 (2010)). Terry W. Schackmann & Barry L. Pickens, The Finality Trap: Accidentally Losing Your Right to Appeal (Part II), 58 J. MO. B. 138, 138 (2002) (describing the “‘finality trap,’ the rule that dismissal of claims without prejudice . . . may undermine the finality of any resulting judgment and thereby bar appeal at the litigation’s end”). Rule 54(b) was first adopted in 1939, along with the rest of the Federal Rules of Civil Procedure. See Sears, Roebuck & Co. v. Mackey, 351 U.S. 427, 432–33 (1956). The text was not a model of clarity, and courts had a hard time figuring out what sorts of orders could be appealed under the Rule. See id. at 434. So the Rule “was amended, in 1946, to take effect in 1948.” Ibid. Those amendments included the clarification that a Rule 54(b) certification would create a “final judgment,” and the addition of the requirement that the district court “express[ly] determin[e] that there is no just reason for delay.” Id. at 434–35 (quotation omitted). In 1961, the Rule was updated to answer affirmatively the question of whether it applied to multi-party actions just as it applied to multi-claim cases. See FED. R. CIV. P. 54(b) advisory committee’s note to 1961 amendment. The 2007 amendments made some stylistic changes, but otherwise modified none of the substance of Rule 54(b). See FED. R. CIV. P. 54(b) advisory committee’s note to 2007 amendment.

3

The Seventh Circuit ultimately denied appellate jurisdiction, but only because the parties never requested a Rule 54(b) partial final judgment. Id. at 365. The court also identified a quirk in that case not present here—namely, the complete overlap between the summary judgment claim and the dismissed counterclaim in ITOFCA. Id. at 364–65. Notwithstanding our independent duty to confirm that we have jurisdiction, the dissenting opinion (in both its introduction and conclusion) criticizes the majority for deciding this case on grounds not presented by the parties. But there‘s nothing wrong with what the majority is doing—the dissenters would likewise decide this case based on grounds not presented by the parties (namely, their effort to overturn our Rule 41(a) circuit precedent). Terry W. Schackmann & Barry L. Pickens, The Finality Trap: Accidentally Losing Your Right to Appeal (Part I) (Finality Trap I), 58 J. MO. B. 78, 84 (2002) (“Concluding an action in the trial court without final disposition of voluntarily-dismissed claims creates the prospect of banishment from the appellate court, without any action left pending in the trial court to which to return to correct the matter. The error by counsel forecloses both any appeal from decisions during the litigation and also any further litigation in the trial court.”). Or Rule 41(a), insofar as the majority understands that Rule to permit dismissal of individual claims.

4

The dissenters suggest “timing matters” as to the filing of the Rule 41(a) and 54(b) motions—ignoring Swope‘s express language that timing does not matter (as the majority also explains). See Swope, 281 F.3d at 192 n.15 (calling the analysis “unnecessary“). They also claim Swope “agreed with [their] reading of Rule 41(a).” Actually, Swope says “[i]t is unnecessary to decide whether the stipulated motion properly came under Rule 41(a).” Id. (emphasis added). And the reason it was unnecessary is because the Rule 54(b) certification was valid under any reading of Rule 41(a)—as the majority and this concurrence point out. Williams v. Taylor Seidenbach, Inc. (Williams II), 935 F.3d 358, 361 (5th Cir.) (Haynes, J., concurring), reh’g en banc granted, 941 F.3d 1183 (5th Cir. 2019). ⁴ A with-prejudice dismissal ordinarily is an adjudication on the merits that carries with it, among other things, the preclusive effects of the law of judgments. See, e.g., Fernandez-Montes v. Allied Pilots Ass‘n, 987 F.2d 278, 284 n.8 (5th Cir. 1993). Obviously, if the plaintiff has failed to properly invoke the powers of Article III courts, the federal courts cannot adjudicate a claim on the merits and trigger the various preclusion doctrines. See FED. R. CIV. P. 41(b) (providing that a dismissal “for lack of jurisdiction” does not “operate[] as an adjudication on the merits“); 18A WRIGHT & MILLER § 4436 (3d ed.) (“[A] dismissal for lack of jurisdiction does not bar a second action as a matter of claim preclusion . . . .“).

5

Both Cooter & Gell v. Hartmarx Corp., 496 U.S. 384 (1990), and Yesh Music v. Lakewood Church, 727 F.3d 356 (5th Cir. 2013), support the majority‘s reading of Rule 54(b), as I shall discuss. In DIRECTV, Inc. v. Budden, 420 F.3d 521 (5th Cir. 2005), the plaintiff did not voluntarily dismiss any claim—instead, the court determined that the plaintiff failed to prosecute, and thus abandoned, some claims while litigating the remaining claims to final judgment. Id. at 525–26. As today’s interpretive scuffle proves, good-faith disagreements among principled textualists are not uncommon. Interpreting language is a human enterprise, not an algorithmic one.

6

To the extent the dissent fears district courts may hold cases hostage by declining to take the steps required to perfect an appeal, that problem exists no matter how we decide this case. If a party loses an adverse jury verdict, the party cannot immediately appeal but must wait for the district court to enter judgment on the verdict. See In re Deepwater Horizon, 785 F.3d 986, 998 (5th Cir. 2015) (“[U]ndocketed orders cannot be appealed.”). See also Jung v. K. & D. Mining Co., 356 U.S. 335, 337–38 (1958) (per curiam) (holding that a district court order dismissing the complaint with leave to amend within twenty days was not a final decision until it formally entered an order—two years later—dismissing the case and entering judgment under Rule 58). It is also no secret among experienced civil litigators that a district judge who hopes to evade appellate review can refuse to enter final judgment to pressure the parties to settle. Similarly, a district court can refuse to set a trial date in order to pressure the parties to settle or to withdraw their claims. Those are, of course, abuses of judicial power. But they have happened—and will continue to happen—even under the dissent’s approach to Rule 54(b). Williams II, 935 F.3d at 361 (Haynes, J., concurring).

7

Section III.B.1 also appears to rely on the premise that “[t]he text of § 1291 demands [a] distinction between ‘final decisions’ and ‘final judgments.’” But the Supreme Court has suggested the opposite. See Bankers Trust Co. v. Mallis, 435 U.S. 381, 384 n.4 (1978) (per curiam) (“A ‘judgment’ for purposes of the Federal Rules of Civil Procedure would appear to be equivalent to a ‘final decision’ as that term is used in 28 U.S.C. § 1291.”). THE PRINCESS BRIDE (20th Century Fox 1987).

8

Williams v. Taylor-Seidenbach, Inc. (Williams I), 748 F. App’x 584, 587 (5th Cir. 2018); Williams II, 935 F.3d at 359–61; cf. JTC Petroleum Co. v. Piasa Motor Fuels, Inc., 190 F.3d 775, 776 (7th Cir. 1999) (noting “a problem of appellate jurisdiction, namely that two claims against one of the applicator defendants were dismissed without prejudice . . . . [S]uch a form of dismissal does not terminate the litigation in the district court in any realistic sense and so is not a final decision within the meaning of 28 U.S.C. § 1291” (citations omitted)); Erie Cty. Retirees Ass’n v. Cty. of Erie, 220 F.3d 193, 201 (3d Cir. 2000) (same).

9

Hernandez v. Thaler, 630 F.3d 420, 425 (5th Cir. 2011) (quotation marks omitted); see also, e.g., Soley v. Star & Herald Co., 390 F.2d 364, 368 (5th Cir. 1968) (“Federal rules . . . were not promulgated to trap unwary litigants.”); U.S. ex rel Erskine v. Baker, 213 F.3d 638, *1 n.2 (5th Cir. 2000) (per curiam) (restating “the general policy of reading the Federal Rules in a manner which avoids ‘traps for the unwary’” (citation omitted)); Brennan’s Inc. v. Dickie Brennan & Co. Inc., 376 F.3d 356, 366 (5th Cir. 2004) (lamenting that legal traps are undesirable); Ellison v. Conoco, Inc., 950 F.2d 1196, 1201 (5th Cir. 1992) (detailing how Rule 58 was designed “not to create traps for the unwary, but rather to afford litigants greater certainty”); 8 WRIGHT & MILLER, FED. PRAC. & PROC. § 2049.1 & n.4 (3d ed. 2010) (stating that “it is important to avoid traps for the unwary or innocently ignorant”).

10

Majority at 7 (detailing how the Williamses could have avoided the trap through “dismiss[ing] the remaining defendants with prejudice under Rule 41(a) . . . . [A]mend[ing] their complaint to excise any remaining claims or parties under Rule 15(a) . . . . [R]equest[ing] severance of certain parties under Rule 21 . . . Or . . . seek[ing] a partial final judgment under Rule 54(b).” (citations omitted)).

11

Des Isles v. Evans, 225 F.2d 235, 236 (5th Cir. 1955); U.S. ex rel. Russell v. Epic Healthcare Mgmt. Grp., 193 F.3d 304, 307 (5th Cir. 1999), abrogated by U.S. ex rel. Eisenstein v. City of N.Y., 556 U.S. 928 (2009) (“Rule 1 . . . provides that the Rules ‘shall be construed and administered to secure the just, speedy, and inexpensive determination of every action.’ This is a charge to resist reading the Rules in a manner that lays traps for the unwary.”).

12

Microsoft Corp. v. Baker, 137 S. Ct. 1702, 1713 (2017) (“Repeatedly we have resisted efforts to stretch § 1291 to permit appeals . . . that would erode the finality principle and disserve its objectives.”); Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546 (1949).

13

See Swope v. Columbian Chems. Co., 281 F.3d 185, 193; Finality Trap I, 58 J. MO. B. at 79 (“[Eliminating the trap] . . . would enable a plaintiff ‘conceivably [to] appeal as many times as he has claims . . . .’”); 9 WRIGHT & MILLER, FED. PRAC. & PROC. § 3914.8 (“[A] dismissal without prejudice that leaves the plaintiff free to reopen the entire case by filing a new action should the challenged ruling be affirmed . . . [is] an undue threat to finality principles . . . .”).

14

For a sampling of other circuits’ views, see Santiago v. Victim Servs. Agency of Metro. Assistance Corp., 753 F.2d 219, 221 (2d Cir. 1985) (“Once the plaintiff has dismissed the action under [Rule 41(a)], the court loses all jurisdiction over the action.” (citation omitted)); State Nat’l Ins. Co. v. County of Camden, 824 F.3d 399, 407 (3d Cir. 2016) (“A voluntary dismissal deprives the District Court of jurisdiction over the action.” (citation omitted)); Commercial Space Mgmt. Co. v. Boeing Co., 193 F.3d 1074, 1079–80 (9th Cir. 1999) (“It follows . . . that the district court has no role to play once a notice of dismissal under Rule 41(a)(1) is filed. The action is terminated at that point, as if no action had ever been filed.”); Anago Franchising, Inc. v. Shaz, LLC, 677 F.3d 1272, 1278 (11th Cir. 2012) (“[C]ourts need not and may not take action after the [Rule 41(a)] stipulation becomes effective because the stipulation dismisses the case and divests the district court of jurisdiction.”); Versata Software, Inc. v. Callidus Software, Inc., 780 F.3d 1134, 1136 (Fed. Cir. 2015) (vacating its own opinion because it was issued after the parties voluntarily dismissed their case under Rule 41(a)); see also McCall-Bey v. Franzen, 777 F.2d 1178, 1190 (7th Cir. 1985) (“An unconditional dismissal terminates federal jurisdiction except for the limited purpose of reopening and setting aside the judgment of dismissal within the scope allowed by Rule 60(b).”); Smith v. Phillips, 881 F.2d 902, 904 (10th Cir. 1989) (same); Adams v. USAA Cas. Ins. Co., 863 F.3d 1069, 1078 n.9 (8th Cir. 2017) (collecting relevant cases).

15

See, e.g., Sneller v. City of Bainbridge Island, 606 F.3d 636, 638 (9th Cir. 2010) (exercising jurisdiction if there is no “evidence of intent to manipulate [] appellate jurisdiction” (citation omitted)); Zayed v. United States, 368 F.3d 902, 905–06 (6th Cir. 2004) (stating that if an “action” is clearly terminated “the order . . . is final and appealable”); Goode v. Cent. Va. Legal Aid Soc’y, Inc., 807 F.3d 619, 623–24 (4th Cir. 2015) (similar); Hope v. Klabal, 457 F.3d 784, 790 (8th Cir. 2006) (similar).

16

This disclaimer may be made through appellate briefing, at oral argument, or in a post-oral-argument stipulation. Nat’l Inspection & Repairs, Inc. v. George S. May Int’l Co., 600 F.3d 878, 883 (7th Cir. 2010).

17

Jewish People for the Betterment of Westhampton Beach v. Vill. of Westhampton Beach (Jewish People), 778 F.3d 390, 394 (2d Cir. 2015) (“[A] plaintiff may cure such a defect in appellate jurisdiction by disclaiming an intent to revive the dismissed claim [which is then binding under estoppel principles] . . . .”); JTC Petro. Co., 190 F.3d at 776–77 (“[T]he [lawyer] . . . agreed that we could treat the dismissal of the two claims as . . . with prejudice, thus winding up the litigation and eliminating the bar to our jurisdiction.” (citation omitted)); Erie Cty. Retirees Ass’n, 220 F.3d at 201–02 (similar).

18

Jethroe v. Omnova Sols., Inc., 412 F.3d 598, 600 (5th Cir. 2005) (“A court should apply judicial estoppel if (1) the position of the party against which estoppel is sought is plainly inconsistent with its prior legal position; (2) the party against which estoppel is sought convinced a court to accept the prior position; and (3) the party [acted intentionally].”).

19

Microsoft Corp., 137 S. Ct. at 1712–13; 9 WRIGHT & MILLER, FED. PRAC. & PROC. § 3911.1 (“There is much to be said for a rule that routinely permits a plaintiff to manufacture finality by abandoning all remaining parts of a case but that forbids any attempt at recapture.”).

20

United States v. Castillo-Rivera, 853 F.3d 218, 236 (5th Cir. 2017) (Smith, J. dissenting) (“By declining to take the fork in the road—that is, by refusing to address and decide one of the two questions on which the court decided to rehear this case en banc—the court [erred] . . . .”).

21

Dissent at 53–54.

22

Dissent at 54–55.

23

See Antonin Scalia, The Rule of Law as a Law of Rules, 56 U. CHI. L. REV. 1175 (1989).

24

“No verbal formula yet devised can explain prior finality decisions with unerring accuracy or provide an utterly reliable guide for the future.” Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 170 (1974) (footnote omitted). The Supreme Court put it forthrightly a half-century ago: “[W]hether a ruling is ‘final’ within the meaning of [§] 1291 is frequently so close a question that decision of that issue either way can be supported with equally forceful arguments, and . . . it is impossible to devise a formula to resolve all marginal cases coming within what might well be called the ‘twilight zone’ of finality.” Gillespie v. U.S. Steel Corp., 379 U.S. 148, 152 (1964).

25

Gillespie, 379 U.S. at 152.

26

Dickinson v. Petroleum Conversion Corp., 338 U.S. 507, 511 (1950).

27

Eisen, 417 U.S. at 170; McGourkey v. Toledo & O.C. Ry. Co., 146 U.S. 536, 544–45 (1892) (“Probably no question of equity practice has been the subject of more frequent discussion in this court than the finality of decrees . . . . The cases, it must be conceded, are not altogether harmonious.”).

28

Dissent at 53–54.

29

Bender v. Williamsport Area Sch. Dist., 475 U.S. 534, 541 (1986) (quoting Mitchell v. Maurer, 293 U.S. 237, 244 (1934)).

30

Microsoft Corp., 137 S. Ct. at 1712 (recounting how finality “preserves the proper balance between trial and appellate courts, minimizes the harassment and delay that would result from repeated interlocutory appeals, and promotes the efficient administration of justice”).

31

Bank of Columbia v. Sweeny, 26 U.S. (1 Pet.) 567, 569 (1828).

32

137 S. Ct. at 1712.

33

Id.; see Cohen, 337 U.S. at 546; see also United States v. River Rouge Improvement Co., 269 U.S. 411, 414 (1926); Cobbledick v. United States, 309 U.S. 323, 328 (1940).

34

Cobbledick, 309 U.S. at 326.

35

Eisen, 417 U.S. at 171.

36

Dickinson, 338 U.S. at 511 (footnote omitted).

37

Microsoft Corp., 137 S. Ct. at 1712–13; Eisen, 417 U.S. at 170–71.

38

Sealed Appellant v. Sealed Appellee, 452 F.3d 415, 417 (5th Cir. 2006); see also, e.g., Berry v. Cigna/RSI-Cigna, 975 F.2d 1188, 1191–92 (5th Cir. 1992).

39

975 F.2d at 1191–92.

Case Details

Case Name: Frank Williams, Jr. v. Lockheed Martin Corp

Court Name: Court of Appeals for the Fifth Circuit

Date Published: May 4, 2020

Citations: 958 F.3d 341; 18-31161

Docket Number: 18-31161

Court Abbreviation: 5th Cir.