NLRB v. Constellium Rolled Products

43 F.4th 395

Opinion

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 20-2140

NATIONAL LABOR RELATIONS BOARD, Petitioner,

v.

CONSTELLIUM ROLLED PRODUCTS RAVENSWOOD, LLC, Respondent.

------------------------------ GILBERT CHARLES DICKEY, Court-Assigned Amicus Counsel.

On Petition to Enforce an Order of the National Labor Relations Board. (09-CA-255275; 09-CA-257508; 09-CA-257510; 09-CA-257889)

Argued: January 25, 2022 Decided: August 5, 2022

Before HARRIS, RICHARDSON, and QUATTLEBAUM, Circuit Judges.

Dismissed by published opinion. Judge Richardson wrote the opinion, in which Judge Quattlebaum joined. Judge Harris wrote a separate opinion, dissenting.

ARGUED: Gilbert Charles Dickey, MCGUIREWOODS, LLP, Charlotte, North Carolina, for Court-Assigned Amicus Counsel. Gregoire Frederic Sauter, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for Petitioner. Dallas Floyd Kratzer III,

2

STEPTOE & JOHNSON PLLC, Columbus, Ohio, for Respondent. ON BRIEF: Jennifer A. Abruzzo, General Counsel, Peter Sung Ohr, Deputy General Counsel, Ruth E. Burdick, Deputy Associate General Counsel, David Habenstreit, Assistant General Counsel, Amy H. Ginn, Supervisory Attorney, NATIONAL LABOR RELATIONS BOARD, Washington, D.C., for Petitioner. Anne L. Doherty, MCGUIREWOODS LLP, Charlotte, North Carolina, for Court-Assigned Amicus Counsel.

3

RICHARDSON, Circuit Judge:

The National Labor Relations Board petitions this Court to enforce its order imposing obligations on an employer. The charged employer, Constellium Rolled Products Ravenswood, LLC, consented in a stipulated settlement agreement to the enforcement of the order, skipping a process of agency prosecution and adjudication. Constellium agreed to a factual statement, waived any defenses, and now dutifully agrees that this Court should enter a judgment against it. We questioned our jurisdiction, concerned that this petition does not present a case or controversy fit for judicial resolution because the parties lack adverseness. We now hold that we lack jurisdiction to exercise judicial power when it would have no real consequences for the parties and would only rubberstamp an agreement the parties memorialized in writing and consummated before ever arriving on a federal court’s doorstep. So the petition must be dismissed. I. Background Constellium Rolled Products Ravenswood employs members of the United Steel, Paper and Forestry, Rubber, Manufacturing, Energy, Allied Industrial and Service Workers International Union, Local 5668 (“United Steelworkers”). When a labor dispute boiled over, United Steelworkers filed four charges with the Board alleging that Constellium committed unfair labor practices. In 2019 and 2020, United Steelworkers requested information from Constellium that it believed would be relevant to collective bargaining— surveillance footage from a dock at the plant, the number of outside contractors working on various projects, and written agreements about those outside contractors, among other things. United Steelworkers says that Constellium refused to provide the requested

4

information. Believing the allegations had merit, the Board’s General Counsel issued an agency complaint against Constellium for violating the National Labor Relations Act, 29 U.S.C. § 151, et seq. And Constellium answered the complaint. Typically, the parties would proceed to a trial before an administrative law judge, and eventually an appeal to the Board.

But rather than proceed through agency adjudication, Constellium and United Steelworkers decided to settle their labor dispute. The parties entered a Formal Settlement Stipulation. By signing the agreement, Constellium effectively withdrew its answer to the complaint before the Board and agreed to stipulated facts—that it had wrongly withheld the security footage, the number of contractors, contracting agreements, and so on. The Stipulation also included proposed terms for a Board order. The Formal Settlement Stipulation was “subject to the approval of the Board” and would “not become effective until the Board has approved it.” J.A. 8. Constellium agreed, upon entry of the Board’s order, to “immediately comply with the provisions of the order.” J.A. 8. Constellium also agreed that when the Board sought a judgment in federal court enforcing its order, Constellium would waive all defenses and consent to the entry of that judgment.

The Board approved the Formal Settlement Stipulation and issued an order reflecting its terms. A week later the Board petitioned this Court under 29 U.S.C. § 160(e) to enter a consent judgment against Constellium reflecting the order’s terms. 1 In its petition

1

Under § 160(e) the Board has the “power to petition any court of appeals of the United States . . . wherein the unfair labor practice in question occurred or wherein such person resides or transacts business, for the enforcement of such order and for appropriate (Continued)

5

to this Court the Board did not argue that Constellium had violated—or threatened to violate—the terms of the Board’s order. Instead, it asserted that it was “entitled to enforcement because Respondent has expressly consented to this judgment in a stipulation that Respondent entered into during the proceedings before the Board.” Appl. for Enforcement of Order of NLRB Upon Stipulation of Parties for Consent J. 1, ECF No. 2- 1. Constellium responded by following through on its promise, stating in a one-sentence response that “it consents to the entry of judgment and enforcement of the Board’s order.” Respondent’s Answer to NLRB’s Enforcement Appl. 1, ECF No. 9. So the parties together, hand-in-hand, ask this Court to enter a judgment binding Constellium to the promises it made in the settlement—without disagreement or a violation of the Board’s order. 2 II. Discussion When concerns about our jurisdiction arise, we must zealously ensure that we do not exercise judicial power outside the Constitution’s bounds. Constantine v. Rectors & Visitors of George Mason Univ., 411 F.3d 474, 480 (4th Cir. 2005). “No action of the parties can confer subject-matter jurisdiction upon a federal court, and ordinary principles

temporary relief or restraining order.” The court of appeals “shall have jurisdiction of the proceeding and of the question determined therein, and shall have power to grant such temporary relief or restraining order as it deems just and proper, and to make and enter a decree enforcing, modifying and enforcing as so modified, or setting aside in whole or in part the order of the Board.” § 160(e).

Because the parties agree that we have jurisdiction and should enter judgment, we

2

appointed Gilbert C. Dickey as amicus curiae to argue the opposite position. He has ably discharged his responsibilities, and we thank him.

6

of consent, waiver, and estoppel do not apply.” Id. (cleaned up). Likewise, Congress cannot extend the judicial power to matters that are neither cases nor controversies. Raines v. Byrd, 521 U.S. 811, 820 n.3 (1997). The Board petitioned this Court for a consent judgment against Constellium that would enforce the same terms as the Board’s consent order. We questioned whether we have the power to grant its request.

Under our Constitution, “[t]he judicial power” extends only to “Cases” and “Controversies.” U.S. Const. art. III, § 2. 3 This limitation cabins federal courts to “the proper—and properly limited—role of the courts in a democratic society.” Allen v. Wright, 468 U.S. 737, 750 (1984). And it ensures that federal courts only exercise power over “those disputes which are appropriately resolved through the judicial process.” Lujan v. Defs. of Wildlife, 504 U.S. 555, 560 (1992). Unlike the “legislative Powers” or the “executive Power,” U.S. Const. arts. I & II, our Power “is legitimate only in the last resort, and as a necessity in the determination of real, earnest, and vital controversy between individuals,” Chi. & Grand Trunk Ry. v. Wellman, 143 U.S. 339, 345 (1892).

A case or controversy’s hallmark has long been “the existence of present or possible adverse parties, whose contentions are submitted to the court for adjudication.” Muskrat v. United States, 219 U.S. 346, 357 (1911). That means “there must be an actual controversy, and adverse interests” between the parties. Lord v. Veazie, 49 U.S. (8 How.)

3

The Supreme Court has treated the terms “Cases” and “Controversies” as coextensive. See Aetna Life Ins. Co. v. Haworth, 300 U.S. 227, 239–40 (1937). “The term ‘controversies,’ if distinguishable at all from ‘cases,’ is so in that it is less comprehensive than the latter, and includes only suits of a civil nature.” Id. at 239 (quoting In re Pac. Ry. Comm’n, 32 F. 241, 255 (C.N.D. Cal. 1887)). One might question this treatment, see infra note 17, but it is the Court’s current position.

7

251, 255 (1850). Although adverseness is an abstract concept that defies straightforward definition, typical adverseness is easy enough to describe: It is where one party “asserts its right” and the other party “is resisting.” Old Colony Tr. Co. v. Comm’r, 279 U.S. 716, 724 (1929). Classic adverseness is the push and pull of parties with opposing interests who offer disagreements to the court. 4 This case is not classically adverse. The parties agree about everything before this court, and there will be none of the push and pull of typical adverseness. And despite Constellium’s capitulation, it does not clearly lack adverseness as a feigned or collusive case or one brought by “fraud or trickery.” Chi. & Grank Trunk Ry., 143 U.S. at 344. 5 So we must dig deeper.

4

Many cases that work out this adverseness principle are concerned with the misuse of judicial review, the federal courts’ power to say what the law is. See, e.g., Chi. & Grand Trunk Ry., 143 U.S. at 345; Cleveland v. Chamberlain, 66 U.S. (1 Black) 419, 426 (1861); Muskrat, 219 U.S. at 357–61. This case may not seem to fit that mold because it resolves a less grandiose dispute, a single labor dispute’s settlement without reaching any binding legal issues or striking down any laws. But “[e]ven in a litigation where only private rights are involved,” courts should guard against issuing judgments where the parties are not adverse. United States v. Johnson, 319 U.S. 302, 304 (1943).

5

There is a motif in the Supreme Court’s adverseness decisions that considers situations in which one party has come to control the litigation on both sides of the “V,” where one party becomes the dominus litis (master of the suit) on both sides. See, e.g., S. Spring Hill Gold-Mining v. Amador Medean Gold-Mining, 145 U.S. 300, 301 (1892). This configuration is evidence that the suit lacks the necessary adverseness. See Cleveland, 66 U.S. (1 Black) at 425; Johnson, 319 U.S. at 303–05.

This case repeats the motif. Before ever arriving in federal court, Constellium signed a contract dictating precisely how it would act. And that agreement dictated what it would say before this Court. The Board’s order, which Constellium signed, said, in essence, you will consent to the judgment, you will make no argument against the judgment, and that will be that; there will be no arguing in front of the judges. And Constellium played that part, presenting a one-line answer to the application for judgment, (Continued)

8

The parties before the court need not disagree about everything to be adverse enough to capture our jurisdiction. Id. at 345–46; Pope v. United States, 323 U.S. 1, 11–12 (1944). For example, parties may agree on liability but dispute the damages owed. At one time, though, agreement about both the merits and the remedy deprived courts of Article III adverseness. Moore v. Charlotte-Mecklenburg Bd. of Educ., 402 U.S. 47, 47–48 (1971) (finding “no case or controversy within the meaning of Art. III” where both parties agreed the law was constitutional and required setting aside a district court order). But now, in at least some circumstances, the parties can apparently agree about essentially everything— both the merits of the legal arguments and the appropriate relief—and still be legally adverse, so long as they retain adverse interests in the litigation’s outcome. See United States v. Windsor, 570 U.S. 744, 756–59 (2013).

In Windsor, the plaintiff sued the government to recover extra taxes she paid because the Defense of Marriage Act prevented her from claiming an estate-tax exemption as the surviving spouse of her same-sex partner. Id. at 750–51. Both she and the government agreed that the Defense of Marriage Act was unconstitutional. Id. at 756. They also agreed on the relief that would be afforded if the Act was held unconstitutional— payment of Windsor’s tax refund. But the government nonetheless continued enforcing the Act and refusing to pay the refund. Id. at 754.

simply consenting. Perhaps more revealing is Constellium’s presentation during oral argument in this case: “We completely agree with the legal analysis of the NLRB, and we adopt that position and have nothing to add to it.” Oral Arg. at 44:52–45:06. It is a red flag that we lack adverseness when, from the word go, one side has dominated the other side’s arguments and presentation.

9

The Executive’s decision to not defend the Act’s constitutionality concerned the Court because the case lacked adverse arguments from the parties. The Court recognized that adverse arguments serve a useful purpose because they “sharpen[] the presentation of issues.” Id. at 760 (quoting Baker v. Carr, 369 U.S. 186, 204 (1962)). But the Court held that adversarial presentation of issues, which aids federal courts’ decision-making, is only a prudential consideration. Id. According to Windsor, the push and pull of opposing legal argument is not the constitutional heart of adverseness.

Still, Windsor did not abandon the longstanding principle that Article III requires adverseness. Windsor instead reaffirmed that Article III requires “sufficient adverseness” to confer an “adequate basis for jurisdiction.” Id. at 759. There remained a constitutional floor. But it did not require adverse arguments. Article III adverseness instead required the government, as the defendant, “retain[] a stake sufficient to support Article III jurisdiction.” Id. at 757. That meant that the parties had to have adverse interests when federal jurisdiction was invoked. Id. at 755, 757–58; see Old Colony Tr. Co., 279 U.S. at 722–24; see also Aetna Life Ins. Co. v. Haworth, 300 U.S. 227, 240–41 (1937) (satisfying the case-or-controversy requirement requires that a declaratory judgment action be “definite and concrete, touching the legal relations of parties having adverse legal interests” (emphasis added)).

In Windsor, the parties retained the required adverse interests because the government had not acquiesced to the remedy: While refusing to defend the Act’s constitutionality in court and conceding that invalidating the Act would require it to pay the tax refund, the government continued to enforce the Act by refusing to pay the refund.

10

Windsor, 570 U.S. at 757–58. As a result, the court judgment directing the Treasury to pay the tax refund—a step the Executive otherwise refused to take—had a real and immediate economic effect on the Treasury’s interest: “Windsor’s ongoing claim for funds that the United States refuses to pay thus establishes a controversy sufficient for Article III jurisdiction.” Id. at 758. In concluding “sufficient adverseness” existed, the Court relied on its prior decision in Chadha as holding that “the refusal of the Executive to provide the relief sought suffices to preserve a justiciable dispute as required by Article III.” Id. at 759 (citing Immigr. & Naturalization Serv. v. Chadha, 462 U.S. 919, 939 (1983)). 6 Adverse interests—that minimum adverseness threshold required by Windsor—

exist only when judicial action would have “real-world consequences” and “‘real meaning’ for the parties.” Seila Law LLC v. CFPB, 140 S. Ct. 2183, 2196–97 (2020). 7 To have meaning in the real world, Article III requires that a judgment issued by this court have some effect in the world, that it will cause real and immediate action or inaction by one of

6

In so holding, the Court acknowledged that Article III adverseness would be lacking if the government had instead “taken the further step of paying Windsor the refund to which she was entitled.” Windsor, 570 U.S. at 758; accord Seila Law LLC v. CFPB, 140 S. Ct. 2183, 2196–97 (2020) (noting adverseness would be lacking if the government agreed to withdraw the investigative demand).

7

The conclusion we reach here is not “foreclosed by” Windsor. See Dissenting Op.

28 n.3 (quoting Seila, 140 S. Ct. at 2196). What Seila described as “foreclosed” is the notion that agreement “on the merits of the constitutional question” deprives the dispute of “adverseness.” 140 S. Ct. at 2196. Seila never suggests that this “adverseness” is no longer an Article III concern. See id. Instead, Seila describes Windsor as saying precisely what we’ve been saying: “real-world consequences for the Government and its adversary suffices to support Article III jurisdiction—even if ‘the Executive may welcome’ an adverse order.” Id. (quoting Windsor, 570 U.S. at 758). In Seila, there was real meaning to the litigation because real-world consequences hung in the balance: the difference between ignoring the investigative demand and complying with it. Id. at 2197.

11

the parties that otherwise would not occur. Id.; Windsor, 570 U.S. at 758; Chadha, 462 U.S. at 939. 8 For example, once a federal judgment issues, money must change hands, see Windsor, 570 U.S. at 758 (“Windsor’s ongoing claim for funds that the United States refuses to pay thus establishes a controversy sufficient for Article III jurisdiction.”), or a party must be bound to stop doing something it otherwise would do or keep doing, see Chadha, 462 U.S. at 939–40 (“Chadha has asserted a concrete controversy, and our decision will have real meaning: if we rule for Chadha, he will not be deported; if we uphold § 244(c)(2), the INS will execute its order and deport him.”). In contrast, judicial intervention cannot be justified when the court’s decision would have no tangible effect on the state of the world. Windsor, 570 U.S. at 758 (noting that the Court would lack Article III jurisdiction “if the Executive had taken the further step of paying Windsor the refund to which she was entitled under the District Court’s ruling”).

8

The Court’s treatment of a demand for funds owed in In re Metro. Ry. Receivership, 208 U.S. 90, 107–08 (1908), illustrates the focus on real consequences. The petitioners argued that the circuit court lacked jurisdiction because “the defendant admitted the indebtedness and other allegations of the bill of complaint, and consented to and united in the application for the appointment of receivers.” Id. at 107. But the Court rejected that argument and held that there was a “controversy between these parties.” Id. The plaintiff made an unsatisfied demand for funds owed, and the defendant could not avoid the court’s jurisdiction merely by “admit[ing] his liability and the amount thereof as claimed, although not paying or satisfying the debt.” Id. at 108. The defendant’s acquiescence to the allegations did not deprive the court of jurisdiction because it did not change the fact that the court’s judgment would still have real-world consequences on the defendant—a requirement to pay the funds owed, which did not mirror any preexisting legal obligation. It is the real-world consequences stemming from a court’s judgment, and not the parties’ agreement on any particular matter in the litigation, which supplies the necessary adverseness for federal court jurisdiction.

12

This idea is not new. In 1890, for example, the Supreme Court held that an adverse controversy was “extinguished” when a disputed tax was paid by the losing party below. Little v. Bowers, 134 U.S. 547, 556–58 (1890) (“Neither the affirmance nor the reversal of that judgment would make any difference as regards the controversy brought here by this writ of error.”). A few years later, the Court pushed that idea a little further. When California tried to refuse payment of money owed to keep a suit alive, the railroad company offered to pay in full and placed the money in escrow with a bank. California v. San Pablo & Tulare R.R., 149 U.S. 308, 313–14 (1893). Under the relevant statutes, placing the money in escrow had the same legal effect as payment, which deprived the court of jurisdiction of the no-longer-adverse suit. Id. When injunctive relief is sought, the same idea holds. “If the prosecutor expressly agrees not to prosecute, a suit against him for declaratory and injunctive relief [to not prosecute] is not such an adversary case as will be reviewed here.” Poe v. Ullman, 367 U.S. 497, 507 (1961) (citing C.I.O. v. McAdory, 325 U.S. 472, 475 (1945)).

With that said, there was adverseness between these parties at some point before the Board. But that adverseness was extinguished before the case got to federal court. See Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 109 (1998) (“Past exposure to illegal conduct does not in itself show a present case or controversy regarding injunctive relief . . . if unaccompanied by any continuing, present adverse effects.”). A case must be justiciable when it reaches federal court. And adverseness that existed before the agency cannot confer jurisdiction upon this court. See Old Colony Tr. Co., 279 U.S. at 722–24; In re Pac. Ry. Comm’n, 32 F. 241, 255 (C.N.D. Cal. 1887). That is the issue here.

13

When the Board issued its order, Constellium became subject to a legal obligation to “[c]ease and desist from” certain conduct and to take other “affirmative action.” J.A. 2. The consent judgment the Board seeks only maintains this status quo by requiring Constellium to do what it is already bound to do, which all parties agree Constellium is already doing. Under those circumstances, this Court’s judgment would have no “real consequences” on the parties; it would merely reiterate Constellium’s obligations under the Board’s order. If this court entered a judgment, no money would change hands, Constellium would take no new actions, and Constellium would not stop any existing conduct. If (and when) this court refuses to enter a judgment, no money will change hands, Constellium will have to take no new actions, and Constellium will not stop any planned or ongoing conduct, bound as it is by the stipulation and order. Under these circumstances, the Board has no interest adverse to Constellium that our judgment would resolve.

None of this is to say that the Board lacks any interest in enforcing federal labor laws. But that interest must be adverse to Constellium. This requirement— that some real- world consequences ride on the litigation that pits the parties against one another—was not changed by Windsor’s shift away from requiring parties to make adverse arguments. See 570 U.S. at 759–60. In Windsor, when the Court says, “the United States retains a stake sufficient to support Article III jurisdiction,” it does not discuss some generalized interest in getting the “ruling it wants” or in properly enforcing federal law but rather that the order at issue required something consequential: “The judgment in question orders the United States to pay Windsor the refund she seeks.” 570 U.S. at 757–58. What matters is that Windsor asked for the funds and the United States refused to pay them even though it

14

agreed with the legal arguments. Id. at 758. That is what the Court calls “sufficient adverseness.” Id. at 759. Again, the Court makes clear that “it would be a different case if the Executive had taken the further step of paying Windsor the refund to which she was entitled under the District Court’s ruling.” Id. It would be different because then the parties would no longer have an adverse interest. In that situation, the litigation has no real meaning and adverseness is lacking.

The entering of a judgment itself could be described as having “real meaning,” at least colloquially. A judgment would allow one party the future use of the court’s contempt power. Contempt proceedings might then result in money damages and maybe even jail time, which seems meaningful indeed. But the “real meaning” required by Windsor demands more than just an enforceable judgment that might, in some future proceeding, lead to real consequences. Parties cannot borrow against future justiciability like that. The court’s judgment in this case must require a party to act—pay money, not deport, comply with investigative demand, and so on. See Windsor, 570 U.S. at 758–59; Chadha, 462 U.S. at 939–40; Seila, 140 S. Ct. at 2197.

Yet the Dissenting Opinion argues that entering a judgment should suffice to create real-world consequences here. Real meaning exists, it argues, because under the statutory scheme the Board’s orders are not “self-executing.” Dissenting Op. 24 (quoting NLRB v. Thill, Inc., 980 F.2d 1137, 1142 (7th Cir. 1992)). And a non-self-executing order cannot be judicially enforced by contempt in a future judicial proceeding should the order be violated. See Thill, 980 F.2d at 1142 (noting Board orders lack “teeth” and are “not judicially enforceable”). So, the Dissenting Opinion posits, the real-world consequence is

15

“consummating the administrative process” to give bite to a Board order so that a future violation will subject Constellium to the “pain of contempt.” Dissenting Op. 33. 9 But potential consequences in a potential proceeding cannot create real-world consequences in this one.

There are surely practical consequences for the Board in lacking a more powerful remedy for a future violation in a future proceeding. See Dish Network Corp., 953 F.3d at 375 n.2 (“Congress has long limited the Board’s powers”). But the lack of a future judicial sanction does not eliminate the obligations under the Board’s order. An obligation is distinct from a judicial remedy. See Perry v. United States, 294 U.S. 330, 354 (1935) (separating the binding obligation from a judicial remedy); Ogden v Saunders, 25 U.S. (12 Wheat.) 213, 349–54 (1824) (Marshall, C.J., dissenting) (distinguishing a contractual obligation from a remedy for violating it); H.L.A. Hart, The Concept of Law 18–25, 34 (2d ed. 1994) (distinguishing between the conduct a rule prohibits and the sanction for violating it).

9

The Dissenting Opinion supports this conclusion with a Third Circuit case holding that jurisdiction exists to confirm an unchallenged arbitration award because the entry of a confirmation order was necessary to provide a full remedy. Teamsters Local 177 v. United Parcel Serv., 966 F.3d 245, 250–51 (3d Cir. 2020). But we have questioned “whether a federal court may confirm a labor arbitration award where there is no live controversy between the parties regarding the award”; a topic that has divided the circuits. Brown & Pipkins, LLC v. Serv. Employees Int’l Union, 846 F.3d 716, 729 n.2 (4th Cir. 2017). We need not weigh in on that question, which is not presented here.

16

Our practice of reviewing petitions under § 160(f) by persons aggrieved by finalized Board orders confirms that those Board orders are legal obligations. 10 Like anything else, challenges to agency action require standing in federal court. See Lujan, 504 U.S. at 559– 561. An employer challenging a Board order therefore must have an “injury in fact” to invoke our jurisdiction. See id. But if Board orders are not legal obligations but meaningless as the Dissenting opinion suggests, being subject to a Board order would not be the kind of injury that confers standing for a party to seek our review. The remedies for violating the agreed-upon order here may be limited to those remedial powers that Congress gave the executive, 11 but the obligation remains. Cf. Sturges v. Crowninshield, 17 U.S. (4 Wheat) 122, 200 (1819) (noting the “distinction between the obligation of a contract, and the remedy given by the legislature to enforce that obligation”).

But this talk of agreement evokes an obvious question: What about consent decrees? The Supreme Court has told us that “a judgment upon consent is ‘a judicial act.’” Pope, 323 U.S. at 12. And we know that there is a long history in American and English practice of courts blessing consent decrees. See Swift & Co. v. United States, 276 U.S. 311, 323–24 (1928) (citing Bradish v. Gee (1754) 27 Eng. Rep. 152, 152, 1 Amb. 229, 229, and Webb v. Webb (1676) 36 Eng. Rep. 1011, 1012, 3 Swans. 658, 659 (“[T]here can be no

10

See, e.g., Sinai Hosp. of Balt. v. NLRB, 33 F.4th 715, 719 (4th Cir. 2022); U.S.

Postal Serv. v. NLRB, 614 F.2d 384, 385 (4th Cir. 1980); Va. Ferry Corp. v. NLRB, 101 F.2d 103, 104 (4th Cir. 1939).

11

See generally 29 U.S.C. § 160(c); NLRB, GC Memo. No. 21-07, Full Remedies in Settlement Agreements (Sep. 15, 2021); NLRB, GC Memo. No. 21-06, Seeking Full Remedies (Sep. 8, 2021).

17

injustice in a decree by consent[.]”)); Pac. R.R. v. Ketchum, 101 U.S. 289, 289–97 (1880). And that history can matter in understanding Article III. See Tutun v. United States, 270 U.S. 568, 576 (1926) (appealing to history to support jurisdiction in naturalization proceedings); William Baude, Constitutional Liquidation, 71 Stan. L. Rev. 1, 32–35, 42– 44 (2019). Taking the history and precedent together suggests that parties may seek a consent decree to resolve a dispute without forfeiting jurisdiction. See Loc. No. 93, Int’l Ass’n of Firefighters v. City of Cleveland, 478 U.S. 501, 525 (1986) (describing limitations on consent decrees); United States v. North Carolina, 180 F.3d 574, 578 (4th Cir. 1999).12 But the Board’s petition here meaningfully differs from the long-standing practice of entering consent decrees. In cases approving a consent decree, the plaintiff files a complaint, the defendant answers denying material allegations, and the parties submit a proposed consent decree that would become a binding obligation only upon the court’s

12

Swift is often cited as support for finding jurisdiction to enter consent decrees. In Swift, the United States brought a petition to enjoin several defendants for violating antitrust law. 276 U.S. at 319. Simultaneously, the defendants filed answers and a stipulation that proposed a consent decree that “the court might, without finding any fact, enter the proposed decree therein set forth.” Id. at 320. Without finding any facts, the court entered a consent decree based on the stipulation. Id. at 321. And no party appealed. Id. at 321. Nearly five years later, several defendants sought to vacate the decree based on the issuing court’s lack of jurisdiction; that motion eventually worked its way to the Supreme Court of the United States. Id. at 325–26.

Among its many rulings, Swift held that whether a case or controversy existed was not an issue that could be raised for the first time on collateral review. Id. at 326 (“On a motion to vacate, the determination . . . that a case or controversy existed is not open to attack.”). But it also rejected an argument that the existence of a case under Article III was eliminated when the government proposed the consent decree. As the Court explained, that “argument ignores the fact that a suit for an injunction deals primarily, not with past violations, but with threatened future ones; and that an injunction may issue to prevent future wrong, although no right has yet been violated.” Id.

18

approval. See North Carolina, 180 F.3d at 578; Swift, 276 U.S. at 320. 13 If the consent decree is adopted by the court, it ends an ongoing controversy in federal court between the parties. If the court rejects the consent decree, the case continues. But not so here. No dispute existed when the Board filed its petition. Constellium did not deny any allegations. The controversy between them ended when the Board issued its order. At that point, Constellium agreed in writing to “immediately comply with the provisions of the order as set forth below.” J.A. 8. And it dutifully did not contest the Board’s petition in federal court. It makes no difference to Constellium if we enter judgment or not. Either way, the case is over and its obligations under the Board’s order remains. Because the Board’s order created Constellium’s obligations, our judgment cannot create “real meaning for the parties” by regurgitating the same obligations from the order—the only thing both parties ask us to do. See Seila, 140 S. Ct. at 2197 (internal citation marks omitted).

Finally, the Board contends that none of this should matter because the Supreme Court has already sanctioned our jurisdiction to enforce Board consent orders like the one

13

While it may be practically true in the consent-decree context that parties may come to court having settled their disputes, “with the proposed judgment in hand,” Dissenting Op. 35 (quoting SEC v. Randolph, 736 F.2d 525, 528 (9th Cir. 1984)), those parties have not formally resolved their dispute; they came with only a “contingent” decree that depends on “court approval,” Randolph, 736 F.2d at 528. And while there may well be a case in which the defendant admitted all the allegations independent of the judicially approved consent decree, the Dissenting Opinion’s cited cases are not such an example. See In re Musical Instruments & Equip. Antitrust Litig., 798 F.3d 1186, 1190 (9th Cir. 2015) (discussing allegations made by the FTC in its own proceeding to enter an administrative consent decree, not in the context of a judicial consent decree); Stinson Canning Co. v. United States, 170 F.2d 764, 765 (4th Cir. 1948) (recounting—in the unusual context of a bond proceeding regarding cans of allegedly diseased sardines which had been forfeited under the Pure Food, Drug, and Cosmetic Act—that the judicially approved decree admitted the allegations).

19

here. The Supreme Court has reviewed Board orders over the years, approving many and never mentioning this adverseness problem. See, e.g., NLRB v. Pa. Greyhound Lines, 303 U.S. 261 (1938); NLRB v. Mexia Textile Mills, 339 U.S. 563 (1950). 14 The Court has even reviewed a Board consent order just like this one.

14

Some history of Board orders at the Supreme Court helps set the stage. Going on a century ago, in Pennsylvania Greyhound Lines, the Supreme Court said that “an order of the character made by the Board, lawful when made, does not become moot because it is obeyed or because changing circumstances indicate that the need for it may be less than when made.” 303 U.S. at 271. There, the Court dealt with a Board order that was gotten without consent at the agency level. From there, the Board applied to the Third Circuit, with the company opposing it, but by the time the Supreme Court was considering the case, the company had come into full compliance with the order. The Court rejected the company’s mootness argument.

A similar pattern emerges in several other Board cases. See, e.g., NLRB v. E.C.

Atkins & Co., 331 U.S. 398, 402 (1947); NLRB v. Raytheon Co., 398 U.S. 25, 27 (1970) (citing Mexia Textile Mills, 339 U.S. at 567–68); NLRB v. Crompton-Highland Mills, 337 U.S. 217, 225 n.7 (1949). We here in the Fourth Circuit have echoed this holding: “that where the possibility of continuing violations exist, usually in the context of employers’ refusals to bargain in good faith, compliance with an NLRB order will not moot a claim.” NLRB v. Greensboro News & Rec., Inc., 843 F.2d 795, 797 n.3 (4th Cir. 1988) (citing Mexia Textile Mills, 339 U.S. at 567); but see NLRB v. Fourco Glass Co., 646 F.2d 863, 864 (4th Cir.1981) (holding that Board orders can sometimes moot out). Beyond any given holding, the Board reaches for this run of cases to support the idea that we have jurisdiction over Board consent orders like this one.

Cases like Pennsylvania Greyhound Lines are distinguishable because they did not involve a consent order. Constellium’s consent shows that our decree would not have real- world consequences because, at least for now, the parties formally agree that Constellium does not intend to violate the Board’s order. See Steel Co., 523 U.S. at 109 (describing the difference between “the allegation of present or threatened injury upon which initial standing must be based,” which is a jurisdictional requirement, and the “presumption of future injury when the defendant has voluntarily ceased its illegal activity in response to litigation,” which does not deprive the court of jurisdiction (cleaned up)); Renne v. Geary, 501 U.S. 312, 320 (1991) (“[T]he mootness exception for disputes capable of repetition yet evading review . . . will not revive a dispute which became moot before the action commenced.”). So far as we can tell, neither this Court nor the Supreme Court has ever discussed this adverseness issue, one way or another.

20

In National Labor Relations Board v. Ochoa Fertilizer Corp., 368 U.S. 318, 322 (1961), the Court considered the appellate court’s power to modify Board orders when entering a consent judgment under § 160(e). As here, the parties before the Board entered into a stipulated settlement agreement that agreed to the form and content of a Board order and “waive[d] all defenses to the entry of the decree.” See id. at 318–19. After the Board issued the order as agreed by the parties, it petitioned the First Circuit to enforce the order. Id. at 320. Rather than enter the order as submitted, the First Circuit cut several portions of the order, and enforced the order as modified. Id. at 321. The Supreme Court reversed the First Circuit’s modification of the Board’s order, holding that “[a] decree entered by consent ‘is always affirmed, without considering the merits of the cause.’” Id. at 323 (quoting Nashville, Chattanooga & St. Louis Ry. v. United States, 113 U.S. 261, 266 (1885)). 15 While the Court’s opinion focuses almost entirely on when a consent order can be modified (not whether a Board consent order is reviewable at all), the Court cited Swift for the proposition that “[t]here are not here applicable any of the exceptions [to nonreviewability], such as a claim of lack of actual consent, or of fraud in the procurement of the order, or of lack of federal jurisdiction.” Id. at 323 (emphasis added) (citing Swift, 276 U.S. at 324). The Board latches on to that passing reference to Swift and jurisdiction.

15

Because Ochoa holds that courts are forbidden from looking at the merits of a Board consent order, judicial review of these orders is extremely limited, resembling a rubber stamp. See 368 U.S. at 323. This limited review only accentuates the defects of our jurisdiction here. Constellium is hamstrung by the consent order, and we too are hamstrung by the consent order, permitting that we may only nod in approval at what the parties have already bound themselves to and are pursuing in full. Unlike the review of proposed consent decrees, our substantively constrained blessing here is a far cry from the use of judicial power.

21

The Board argues that Ochoa’s citation to Swift is a holding that petitions to enforce a Board order are sufficiently adverse regardless of the parties’ consent to the judgment or whether the Board alleges a breach or threatened breach of the order. But this passing citation to Swift should not be read so broadly. “[D]rive-by jurisdictional rulings” are not precedential. Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 91 (1998). Apart from arguably addressing adverseness, the Swift Court also resolved seven other jurisdictional challenges to the consent decree. 276 U.S. at 324 (“[E]ight reasons are relied on as showing that, in whole or in part, it was beyond the jurisdiction of the court.”). 16 Ochoa’s mere passing reference to eight possible issues of jurisdiction, without explanation or rationale, see 368 U.S. at 323, does not resolve the jurisdictional question before this Court and does not change our analysis.

In sum, the parties agree on every relevant question potentially before this Court.

That agreement led the parties to resolve this dispute among themselves before ever coming to federal court, leaving nothing for this Court to do that would have real consequences in the world. And the Board agrees that Constellium has complied with the order and

16

Notably, Ochoa does not cite Swift’s relevant discussion of Article III adverseness. See Ochoa, 368 U.S. at 323 (citing Swift, 276 U.S. at 324). In the portion of Swift cited by Ochoa, the Swift Court’s discussion of jurisdiction is limited to acknowledging that “[d]ecrees entered by consent have been reviewed on appeal [where] there was lack of federal jurisdiction because of the citizenship of the parties” and a discussion of whether the Supreme Court of the District of Columbia had subject-matter jurisdiction under the Sherman Anti-Trust Act. See Swift, 276 U.S. at 324. Swift’s pertinent discussion of Article III does not come until the next page, see id. at 325, which Ochoa did not cite.

22

continues to do so. There is nothing meaningful for this Court to do. Because this suit lacks adverseness, we lack jurisdiction. 17

* * *

“It is the office of courts of justice to decide the rights of persons and of property, when the persons interested cannot adjust them by agreement between themselves.” Lord, 49 U.S. at 255. The parties here came to such an agreement, resolving their dispute themselves. Because the Board’s petition to this Court seeks to merely reaffirm Constellium’s preexisting obligations under the Board consent order, this is not a case or controversy subject to the judicial power. We lack jurisdiction, and therefore the petition is DISMISSED.

17

Although we reach this result by applying precedent, recent originalist scholarship has challenged the foundation of existing adverseness doctrine (and indeed standing more broadly). See generally James E. Pfander, Cases Without Controversies: Uncontested Adjudication in Article III Courts (2021). But such broad revisions to the Supreme Court’s existing justiciability doctrine are not for this Court to decide. Accepting our place, we do our level best to apply the law as it stands, not as it might one day become.

23

PAMELA HARRIS, Circuit Judge, dissenting:

Today’s majority forbids, as outside the scope of Article III, a practice that the National Labor Relations Board (“NLRB” or “Board”) has used for decades to resolve labor disputes. Until now, neither we nor any other court has ever questioned our jurisdiction to enforce consent orders like this one. Indeed, year in and out, we – like every other court of appeals – routinely enforce orders exactly like this one, in a practice that the majority now says has persistently exceeded our powers. 1 Because I cannot agree that the Constitution forbids this familiar practice, I respectfully dissent.

This case began as an unremarkable labor dispute between Constellium and the labor union representing employees at its Ravenswood, West Virginia aluminum plant. See Maj. Op. at 3–4. The union filed charges with the Board alleging that Constellium had denied it information to which it was entitled; the Board issued a complaint against Constellium based on those charges; and Constellium responded by filing an answer denying the Board’s allegations. See 29 U.S.C. § 160(b).

1

See, e.g., NLRB v. USPS, No. 21-1202 (4th Cir. 2021); see also, e.g., NLRB v.

Empresas Velazquez, Inc., No. 21-1059 (1st Cir. 2021); NLRB v. ADT, LLC, No. 21-974 (2d Cir. 2021); NLRB v. United Gov’t Sec. Officers of Am., Loc. 171, No. 20-2795 (3d Cir. 2020); NLRB v. USPS, No. 22-60039 (5th Cir. 2022); NLRB v. Bridgestone Ams. Tire Operations, LLC, No. 20-2010 (6th Cir. 2021); NLRB v. Golden Mile Hotels, LLC, No. 21- 3262 (7th Cir. 2022); NLRB v. Richfield Hosp., Inc., No. 18-1262 (8th Cir. 2018); NLRB v. Noel Canning, No. 17-71893 (9th Cir. 2017); NLRB v. USPS, No. 18-9590 (10th Cir. 2019); NLRB v. Int’l Longshoremen’s Ass’n, Loc. Union No. 1402, No. 20-10266 (11th Cir. 2020). Fortunately, whatever unsettling effect today’s ruling might have going forward, it should not disturb the finality of these prior cases. See Swift & Co. v. United States, 276 U.S. 311, 326 (1928) (jurisdictional objection to consent decree is “not open on a motion to vacate”; “[i]f [a court] erred in deciding that there was a case or controversy, the error is one which could have been corrected only” on direct review).

24

Rather than proceed to a full-blown adjudication, however, the parties – consistent with the Board and Congress’s longstanding policy favoring “the peaceful, nonlitigious resolution of disputes” – reached an agreement they hoped would settle the matter. See Indep. Stave Co., 287 N.L.R.B. 740, 741 (1987); id. (discussing Congress’s recognition that “settlements constitute the life blood of the administrative process, especially in labor relations” (internal quotation marks omitted)). In that agreement, the parties stipulated to the relevant facts, Constellium admitted the allegations against it, and the parties agreed that “the Board may immediately enter an order” directing Constellium to comply with its obligations under federal labor law. J.A. 12. Nothing in the agreement required Constellium to take any actions unless the Board approved it. But after reviewing the matter, the Board did so, ordering Constellium to provide the union with the information it had been denied and to post a notice informing employees at the plant of their rights.

Importantly – though one might not know it from the majority’s opinion – the Board’s order, when issued, had no binding legal effect. “Unlike the orders of other agencies, the Board’s orders are not self-executing.” NLRB v. Thill, Inc., 980 F.2d 1137, 1142 (7th Cir. 1992); see Dish Network Corp. v. NLRB, 953 F.3d 370, 375 n.2 (5th Cir. 2020). For a Board order to take effect, the NLRB must “petition [a] court of appeals of the United States . . . for the enforcement of such order.” 29 U.S.C. § 160(e). Until then, “[c]ompliance is not obligatory,” In re NLRB, 304 U.S. 486, 495 (1938), and parties “can violate [a board order] with impunity until a court of appeals issues an order enforcing it,” NLRB v. P*I*E Nationwide, Inc., 894 F.2d 887, 890 (7th Cir. 1990); see id. at 892 (noting that the NLRB’s “curious impotence” in this regard originally “was a swap for [the]

25

procedural informality” of its adjudications, and likely has persisted due to “[d]istrust” of, or disagreement with, the Board’s decision-making). So, as the parties had agreed, the Board soon applied to us for enforcement of its order, and Constellium consented to our doing so.

That brings us to the majority’s opinion, which dismisses the Board’s application.

As best as I can tell, the court holds that the parties here have insufficiently “adverse” interests because (1) they settled this dispute before the agency and (2) their settlement included a promise, fulfilled by Constellium, not to contest enforcement before us. In other words, we lack jurisdiction, and therefore purportedly must dismiss the NLRB’s enforcement application, because given the parties’ settlement, any judicial action here “would have no real consequences” and “no real meaning.” Maj. Op. at 10–13 (internal quotation marks omitted).

This is a novel conclusion. As far as I am aware, no court before ours has questioned, let alone rejected, its jurisdiction to hear a case like this one. See, e.g., supra note 1. And while I agree that no case has specifically held to the contrary on these exact facts, 2 essentially all of the relevant authority points in favor of finding that we have jurisdiction to enforce the Board order here.

2

To be sure, and as the majority recognizes, the Supreme Court has considered a case in a posture much like this one, where it concluded (in passing) that there was no jurisdictional defect with the proceeding and required enforcement of an NLRB order. See Maj. Op. at 19–21 (discussing NLRB v. Ochoa Fertilizer Corp., 368 U.S. 318 (1961), which required enforcement of a Board consent order even after the employer had “waive[d] all defenses to the entry of the decree,” id. at 319 (internal quotation marks omitted)). I agree with the majority, largely for the reasons it gives, that Ochoa’s passing (Continued)

26

To see why, it might help to begin with the important ways in which this case differs from many of the authorities on which the majority relies. The majority concedes that this is not a “feigned or collusive” case – no one doubts that Constellium would rather not be here at all, and rather views settlement and compliance with the Board’s order as the best way to resolve its conflict with the Board. Maj. Op. at 6–7 (citing Chi. & Grand Trunk Ry. v. Wellman, 143 U.S. 339, 344–45 (1892); and Lord v. Veazie, 49 U.S. (8 How.) 251, 255 (1850)). Similarly, the majority seems to accept that there is no real “misuse of judicial review,” or improper attempt to commandeer “the federal courts’ power to say what the law is,” to worry about here. Id. at 7 n.4 (collecting cases). And though it finds a “motif” in some cases declining jurisdiction where one party “has come to control the litigation on both sides of the ‘V,’” id. at 7 n.5, those cases, unlike this one, raise obvious concerns about friendly suits, see, e.g., Cleveland v. Chamberlain, 66 U.S. (1 Black) 419, 425–26 (dismissing case where one party bought out the litigating interests of the other side and brought suit “for the evident purpose of obtaining a decision injurious to the rights and interests of third parties”), and even the majority does not suggest that they actually control in this case. So to the extent Article III contains a freestanding adverseness requirement, none of the paradigm concerns animating that doctrine carry much weight here. See 13 Charles Allan Wright & Arthur R. Miller, Federal Practice & Procedure § 3530 (3d ed.

reference to jurisdiction is not, strictly speaking, binding precedent on the issue before us. But the reality that the Court “in fact reviewed an identical order,” Shoop v. Twyford, No. 21-511, slip op. at 5 n.1 (U.S. June 21, 2022), and compelled its enforcement – despite identifying the lack of jurisdiction as a potential reason not to do so and finding it “significant” that the order had been obtained on consent, Ochoa, 368 U.S. at 323 – should give us substantial pause here.

27

Apr. 2022 update) (“Cases involving genuinely adversary interests, but lacking any dispute as to facts or remedy, must be sharply distinguished from the decisions rested on the fact of common interests and a shared desire to affect nonparties. There may be a very real need to secure a judicial decree to establish status or rights, or assist in their enforcement, even though there is no present dispute.”).

More importantly, after the Supreme Court’s decision in United States v. Windsor, 570 U.S. 744, 756–59 (2013), it appears that Article III does not require adverseness at all. Instead, as the majority explains, “the parties can apparently agree about essentially everything” without calling our power to hear the case into question. Maj. Op. at 8; see id. at 7–11 (discussing Windsor, 570 U.S. at 756–59). Under Windsor, that is, the parties can agree on the facts, view the law the same way, and seek exactly the same outcome for exactly the same reasons – but whatever prudential concerns we may have in that scenario, Article III has nothing to say about our power to decide the case. See 570 U.S. at 759–60 (fact that parties agreed on case’s outcome, and resulting risk of a “friendly, non-adversary, proceeding” rather than a “real, earnest and vital controversy,” went only to prudential, not jurisdictional, considerations (internal quotation marks omitted)).

As a result of all this, the majority falls back on a different Article III theory. In light of the above, it acknowledges, it is now clear that the Constitution “d[oes] not require adverse arguments.” Maj. Op. at 9. But still, it pivots, we cannot exercise jurisdiction unless there are “adverse interests” at stake. Id. And this case lacks this distinct variety of adverseness, the majority reasons, because nothing we can do would change the result the parties have already agreed upon: The Board has already issued its order, Constellium

28

intends to abide by it, and with or without our imprimatur, nothing will change on the ground in the real world. See id. at 10–13.

The problem with this reframing, however, is that it has little to do with adverseness.

As suggested by the majority’s shift from discussing “the push and pull” of adverse contestation, id. at 7, to demanding “real-world consequences,” a “tangible effect,” and “real meaning for the parties,” id. at 10–11 (internal quotation marks omitted), we have strayed quite a ways from any freestanding adverseness doctrine. Now, we seem instead to be dealing with a much more familiar Article III requirement: that a party needs a concrete interest in the outcome of a case to invoke federal jurisdiction. This idea is sometimes expressed as a component of our mootness case law. See, e.g., Eden, LLC v. Justice, 36 F.4th 166, 169 (4th Cir. 2022) (case moot where “the parties lack a legally cognizable interest in the outcome” (internal quotation marks omitted)). Other times, it relates to a party’s standing. See, e.g., Windsor, 570 U.S. at 757–58. But in all cases, the core question is this: Consistent with Article III’s demand for “cases” or “controversies,” will a judicial decision matter for real people in the real world? Cf. Allen v. Wright, 468 U.S. 737, 750 (1984) (discussing commonalities running across “[a]ll of the doctrines that cluster about Article III” (internal quotation marks omitted)). 3

3

The majority resists this conclusion, trying to salvage from Windsor some remnants of a “constitutional floor” for adverseness distinct from the more general requirements of a concrete interest. Maj. Op. at 9. But the Court has already held that this effort is simply “foreclosed by” Windsor. See Seila L. LLC v. Consumer Fin. Prot. Bureau, 140 S. Ct. 2183, 2196 (2020). As long as a decision would have “real-world consequences,” the parties’ agreement will not defeat federal jurisdiction. Id. Indeed, Justice Scalia’s dissenting opinion in Windsor would seem to confirm this reading of the (Continued)

29

And once we have identified the right question – whether the NLRB has a sufficient interest in the enforcement of its order to support our jurisdiction – the answer becomes quite clear: It does. That follows first from the long line of cases holding that compliance with a Board order does not deprive the Board of an interest in enforcing that order. By 1950, it was “plain from the cases that the employer’s compliance with an order of the Board does not render the cause moot, depriving the Board of its opportunity to secure enforcement from an appropriate court.” NLRB v. Mexia Textile Mills, Inc., 339 U.S. 563, 567 & n.4 (1950) (collecting cases). That was because “[a] Board order imposes a continuing obligation” and, even without an ongoing dispute, “the Board is entitled to have the resumption of the unfair practice barred by an enforcement decree.” Id. at 567; see NLRB v. Raytheon Co., 398 U.S. 25, 27 (1970) (Board’s unique role in enforcing federal labor law prevented enforcement action from becoming moot).

True, in Mexia, the employer had originally contested the matter before the Board, and complied only after the Board had ruled against it. See Maj. Op. at 9, 19 n.10. But I do not see how that could matter to our assessment of whether the Board has an interest here, now that we are at the enforcement stage of the case. Both here and in Mexia, by the time the NLRB had sought enforcement, it had issued its order and the employer had complied with it. In either case, it could equally be said that the facts on the ground would remain unchanged by the judicial enforcement of that order. Yet in Mexia and the cases on

case. See Windsor, 570 U.S. at 784–85 (Scalia, J., dissenting) (criticizing the Court for converting what he viewed to be the “separate Article III requirement of adverseness between the parties” to “an element (which it then pronounces a ‘prudential’ element) of standing,” and relying on many of the same cases that the majority cites here).

30

which it relied, the Board remained “entitled to have the resumption of the unfair practice barred by an enforcement decree.” 339 U.S. at 567. I cannot see why we must reach a different result here.

It is true, of course, that these precedents address mootness, and not the specific “adverseness problem” the court has posed to itself. Maj. Op. at 18–19 & n.10. 4 But for reasons I have already explained, once we clarify that we are talking about adverse “interests” instead of adverse “arguments,” this becomes a distinction without a difference. The Board has an interest in the outcome of this case because it has an interest in preventing further violations of federal labor law, like the ones Constellium has admitted to in this case. See Raytheon, 398 U.S. at 27–28. That interest gives the Board standing to bring enforcement actions even once all parties have already complied with its orders; it prevents those actions from becoming moot even if compliance begins after it seeks enforcement; and, all told, it gives rise to a sufficiently concrete “case” or “controversy” throughout the process.

4

The majority suggests an additional distinction: The parties in this case “formally agree that Constellium does not intend to violate the Board’s order,” whereas in these other cases the employer simply complied with the Board’s order without any contractual obligations. Maj. Op. at 19 n.14. But whatever the legal significance of Constellium’s agreement, it does not change the fact that the Board remains “entitled to have the resumption of the unfair practice barred by an enforcement decree,” which will have legal consequences for the parties to this case beyond those provided by contract. See Mexia, 339 U.S. at 567 (reasoning that Board’s entitlement to enforcement prevents cases from becoming moot because it “adds to existing sanctions that of punishment for contempt” (emphasis added) (internal quotation marks omitted)).

31

The majority also seems to rely on a distinction between standing and mootness, noting that, while the need to deter future misconduct may save a case, once brought, from becoming moot, it cannot create the injury necessary to support standing at the outset. See Maj. Op. at 12, 19 n.14 (citing Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 109 (1998)). This may be true in general civil litigation. But in the context of the Board’s enforcement of federal labor law, the Supreme Court has already told us that the Board’s interests transcend remedying “particular” incidents of misconduct because its role is more broadly “to protect employees in the exercise of their organizational rights.” Raytheon, 398 U.S. at 27. Indeed, the lead case that Mexia relied on, see 339 U.S. at 567–68, rejected a claim that the court lacked jurisdiction to enforce a Board order due to a settlement that had occurred before the Board had issued that order, see NLRB v. Gen. Motors Corp., 179 F.2d 221, 222 (2d Cir. 1950) (per curiam) (“[I]t is well settled in Labor Board cases that the Board need not dismiss the proceeding and that the courts need not refuse to enforce the Board’s order, because, pendente lite, the original dispute has been settled.” (footnote omitted)). None of these cases purported to turn on whether the employer had complied with the Board’s order before or after the NLRB had sought enforcement in the court of appeals.

All this is reinforced by the details of the statutory scheme here. In concluding that this proceeding can have no “real-world consequences,” Maj. Op. at 13, the majority repeatedly asserts that Constellium is already “bound” by the Board’s order, and that its “obligations under the Board’s order” will “remain[]” unchanged no matter what we do, id. at 13, 15–16. But that is not so. As explained above, the NLRB is distinct among

32

federal agencies in that its final orders lack any effect unless they are judicially enforced. See In re NLRB, 304 U.S. at 495; Dish Network Corp., 953 F.3d at 375 n.2; Thill, Inc., 980 F.2d at 1142. Far from “bound” by the Board’s order, Constellium “can violate it with impunity until” we enforce it, at which point “violations [will] expose [Constellium] to proceedings for contempt.” P*I*E Nationwide, Inc., 894 F.2d at 890. 5 As a result, the adversary process that begins before the Board essentially stays alive up through the enforcement of a Board order. In the analogous context of arbitral confirmation, the Third Circuit has held similarly. See Teamsters Loc. 177 v. UPS, 966 F.3d 245, 248 (2020). There, the party that lost in arbitration sought to dismiss the winner’s

5

The majority responds that Board orders are not “meaningless” before they are enforced, distinguishing “binding obligation[s]” from “judicial remed[ies]” and noting that even without the latter, these orders could impose some legal obligation that is binding in the abstract. Maj. Op. at 14–16. The first problem with this argument is that the Supreme Court has told us the exact opposite: Until enforcement of a Board order, “[c]ompliance is not obligatory.” In re NLRB, 304 U.S. at 495. Whatever conceptual distinctions might exist between a legal obligation and remedies for violations of that obligation, see Maj. Op. at 15 (citing H.L.A. Hart, The Concept of Law 18–25, 34 (2d ed. 1994)), we know that, until enforcement, the Board cannot count on either. In any event, my point is not that the Board’s orders, pre-enforcement, are “meaningless”; it is simply that, whatever their exact status at that time, our enforcement of them is not meaningless because doing so “imposes a continuing obligation” that does not otherwise exist and “adds to existing sanctions that of punishment for contempt.” Mexia, 339 U.S. at 567.

The majority also relies on “[o]ur practice of reviewing petitions under § 160(f) by persons aggrieved by finalized Board orders,” reasoning that our failure to note any jurisdictional defects in cases like the ones it cites shows that employers challenging Board orders “must have an ‘injury in fact,’” which implies that the Board’s orders must in turn impose some legal obligation. Maj. Op. at 16. It is unclear to me why this “practice” deserves weight in this analysis, but not our (and every other circuit’s) long-established practice of enforcing consent orders just like this one. See supra note 1; see also Ochoa, 368 U.S. at 319–23. But either way, as I have said, the fact that a Board order has some legal significance does not mean that our enforcement of it cannot also have real meaning for the parties.

33

bid for confirmation of the arbitral award, arguing that because it (the arbitral loser) had not violated the award and did not challenge its validity, the winner lacked a concrete interest in its confirmation and therefore lacked standing. Id. at 251. Relying on the structure of the Federal Arbitration Act (“FAA”) and the fact that Congress had left arbitral awards unenforceable until confirmed by a court, the circuit rejected this argument. “[I]t is confirmation under § 9 [of the FAA] that converts the award into a judgment of the court and completes the arbitration process under the FAA framework.” Id. at 251–52. As a result, “the dispute the parties went to arbitration to resolve is ‘live’ until the arbitration award is confirmed and the parties have an enforceable judgment in hand.” Id. at 252.

So too here: Enforcing the Board’s order would have the real-world effect of consummating the administrative process that Congress has carefully set up, entitling the Board and complaining union to compel Constellium’s compliance upon pain of contempt. Absent our enforcement, Constellium would be free to violate the order at its discretion without facing any consequences. Given the NLRB’s distinct role in enforcing our nation’s labor law – and its historical commitment to doing so through the “peaceful, nonlitigious resolution of disputes,” Indep. Stave Co., 287 N.L.R.B. at 741 – I would hold that this is a sufficient interest to support our jurisdiction here. 6

6

The majority argues that the entry of judgment alone can never be enough of a real-world consequence to support jurisdiction under Windsor. See Maj. Op. at 14. Again, that may be the case in the mine run of civil cases. But in the labor context, we already know that the Board’s distinct role and interests in the enforcement of the nation’s labor law lead to a different result. See, e.g., Mexia, 339 U.S. at 567.

34

Finally, as the majority acknowledges, the long line of cases approving the entry of consent decrees outside the labor context also supports our jurisdiction in this case. Generally, a settlement ends a case, precluding further jurisdiction over the matter. See, e.g., Pressley Ridge Schs. v. Shimer, 134 F.3d 1218, 1220 (4th Cir. 1998). But even after settlement, the majority accepts, “there is a long history . . . of courts blessing consent decrees,” to which we ought to defer. Maj. Op. at 16; see, e.g., Pope v. United States, 323 U.S. 1, 12 (1944) (“It is a judicial function and an exercise of the judicial power to render judgment on consent. A judgment upon consent is a judicial act.” (internal quotation marks omitted)); Swift & Co. v. United States, 276 U.S. 311, 323–24 (1928).

The majority does not offer any serious way to distinguish these cases. It claims that a defendant seeking a consent decree usually “answers denying material allegations,” contrasting that with Constellium’s admissions here. Maj. Op. at 17. But that is not always the case, see, e.g., In re Musical Instruments & Equip. Antitrust Litig., 798 F.3d 1186, 1190 (9th Cir. 2015) (defendant “neither admitted nor denied the FTC’s allegations”); Stinson Canning Co. v. United States, 170 F.2d 764, 765 (4th Cir. 1948) (noting entry of consent decrees “reciting the claimant’s admission of the allegations . . . and its consent to judgments”); cf. In re Metro. Ry. Receivership, 208 U.S. 90, 108 (1908) (“We do not doubt the jurisdiction of the Circuit Court, although the facts were admitted, and the defendant joined with the complainants in a request that receivers should be appointed.”), and the majority tells us nothing about how its decision squares with these authorities. The majority also suggests that consent decrees are different because they “become a binding obligation only upon the court’s approval.” Maj. Op. at 17–18 & n.13. But that is just so

35

here: As I have explained, because the Board’s order is not self-executing, it, too, can only become binding upon our approval. See, e.g., In re NLRB, 304 U.S. at 495. Finally, the majority seems to believe that courts can only enter consent decrees if the parties resolve their dispute after coming to court, emphasizing the timing of the settlement here. See Maj. Op. at 12, 18. But it is common practice for parties seeking consent decrees to “arrive[] in court” having already settled, “with the proposed judgment in hand”; and the majority cites no cases disapproving of that practice. E.g., SEC v. Randolph, 736 F.2d 525, 528 (9th Cir. 1984) (rejecting that the court lacked jurisdiction to enter consent decree just because the parties “could rely on their [prior] settlement agreement”).

* * *

In sum: We all agree that Article III does not require parties before us to press

adverse arguments. The Supreme Court has told us that the NLRB has a sufficient interest in giving bite to its otherwise toothless orders to secure their enforcement even after any active dispute about them has ended. And courts have long enforced consent decrees in circumstances much like this one, in a historical practice the majority accepts is entitled to substantial weight in our analysis. For these reasons, I would conclude that we have jurisdiction and grant the Board’s request to enforce its order. 7

7

To be sure, I recognize and readily accept that, even where we have jurisdiction to enforce a Board order, there may be prudential reasons not to do so in specific cases in which such enforcement would be gratuitous, harmful, or otherwise unwise. See, e.g., NLRB v. Reeves Bros., 851 F.2d 356, 1988 WL 70402, at *1 (4th Cir. July 8, 1988) (unpublished table decision); NLRB v. Fourco Glass Co., 646 F.2d 863, 864–65 (4th Cir. 1981). Nothing here, however, appears to counsel such hesitation.

36

Perhaps today’s decision, standing alone, will cause little trouble for the NLRB, grounded as it is in the distinct facts before us. On my read, the court’s holding (though it does not clearly set out what factors it considers dispositive) is limited to cases in which (1) the parties settle before the Board issues an order; (2) that settlement includes a promise not to contest enforcement of the resulting order; (3) the parties abide by that promise; (4) the parties do not otherwise violate the order and have no dispute about its terms; and (5) the agreement does not depend on our enforcement of the order. 8 As a result, going forward the Board may well be able to amend the terms of its settlements to avoid these pitfalls – for example, by no longer having employers preemptively waive their defenses to enforcement, or by making the agreement conditional not only on Board approval, but also on judicial enforcement. 9 Even so, I worry that, in the majority’s zeal to manufacture a jurisdictional problem here, it cannot avoid creating

8

The majority also suggests an additional limitation on its holding, relying at least in part on our “extremely limited” review of Board consent orders. Maj. Op. at 20 n.15. I welcome any and all guardrails on the reach of today’s ruling. Even so, this is a curious point to make given that the limited review the majority decries comes straight from the Supreme Court’s decision in Ochoa – which, remember, considered the enforcement of a consent order just like this one and, finding no jurisdictional defects, compelled enforcement. See 368 U.S. at 323. To me, this seems to support our jurisdiction to enforce the order before us, notwithstanding the limits on our power of review. See supra note 2.

9

If not, the majority’s opinion may lead to absurd results. Since the Board would be unable to secure enforcement in settled cases, parties with no real dispute over the facts or law, who therefore wish to reliably resolve their conflict with the Board without costly litigation, might be forced to march through the motions of hollow “adverse” proceedings just to secure a final, durable resolution. Perversely, this would likely waste Board, worker, and employer resources mostly in cases where liability is most obvious, or where the cost of full-blown adversary proceedings most heavily outweighs the value of the matter to the parties involved.

37

costly doubt about the permissibility of a whole swath of long-accepted judicial practices, which parties often rely on to resolve disputes, minimize risk, and organize their relations. See, e.g., James E. Pfander & Daniel D. Birk, Article III Judicial Power, the Adverse-Party Requirement, and Non-Contentious Jurisdiction, 124 Yale L.J. 1346, 1359–91 (2015) (cataloging scores of non-adverse proceedings over which courts have long maintained jurisdiction). The majority does not tell us exactly how far its holding goes, or what other seemingly well-established types of cases might now be in jeopardy. Rather than inviting this wave of uncertainty, I would defer to the precedent and history summarized above, hold that we have jurisdiction, and enforce the Board’s order accordingly. For all these reasons, I respectfully dissent.