concurring:
I join the majority opinion in its entirety. I write separately to emphasize a few brief points.
The dissent spins a tale of corruption and conspiracy, in which the plaintiffs and the Government were complicit in bilking the nation’s taxpayers to pay a political ransom. While this narrative may have been advanced in news accounts and scholarly articles, most of those statements and opinions have not been validated by the solemnity of the oath and “testing in the crucible of cross-examination,” Crawford v. Washington, 541 U.S. 36, 61, 124 S.Ct. 1354, 158 L.Ed.2d 177 (2004); nor are they found in the record, and it is the record upon which our decision must be based. Fed. R. App. P. 10(a).
What the record shows is that the District Court expressly found that the settlement “was attained following an extensive investigation of the facts and the law ... [and] resulted from vigorous arms’-length negotiations, which were undertaken in good faith.” Order, Keepseagle v. Vilsack, No. 99-cv-3119 (D.D.C. Apr. 28, 2011), JA 589-91. Unless we find clear error in the *1057District Court’s conclusion—and even the dissent does not claim to do so—that finding stands.
It is true that more than half of the settlement fund was not distributed through the claims process and is now poised to be distributed via the cy-pres provision. But this was an unanticipated state of affairs, not an intended result. See Keepseagle v. Vilsack, 118 F.Supp.3d 98, 102 (D.D.C. 2015) (“[N]o one anticipated such a large amount of excess funds.”). As represented to the District Court, “the parties contemplated that no more than several million dollars in settlement funds would be unclaimed,” based on an expectation that “over ten thousand class members would likely file ... claims, and that most of those claims would be successful.” J.A. 718 & n.2. Instead of 10,000 claims, only 5,191 were received. J.A. 596.
The reason for this discrepancy is “unclear,” Keepseagle, 118 F.Supp.3d at 102, but the difference in the number of actual, versus estimated, claimants correlates closely with the amount of surplus settlement funds. The parties have offered several possible explanations, including that the deaths of eligible claimants (the claims process began 30 years after the first year covered by the settlement) left heirs with insufficient information to complete claim forms or that, perhaps, there were “simply fewer people with claims than Plaintiffs originally argued.” Id. at 108 n.3. In addition to falling short of the expected number of claims, a large number of submitted claims were unsuccessful. Most strikingly, out of 146 Track B claimants, only fourteen were successful. Compare J.A. 596, with J.A. 716.
Regardless of the cause of this “monumental” failure in the claims process, Keepseagle, 118 F.Supp.3d at 102, there is no occasion for considering the newly-resurrected claim that the cy-pres provision— a feature of the Settlement Agreement since it was first unveiled seven years ago—violates the Appropriations Clause. The dissent apparently concedes that Appellant Mandan waived this claim before the District Court when he was asked whether he wished to pursue it. Yet, the dissent relies exclusively on cases involving forfeiture—not waiver—to argue that “exceptional circumstances” permit an appellate court to nevertheless consider the claim. The Supreme Court, though, has been clear: “Waiver is different from forfeiture.” United States v. Olano, 507 U.S. 725, 733, 113 S.Ct. 1770, 123 L.Ed.2d 508 (1993). While “[m]ere forfeiture ... does not extinguish an error,” waiver may. Id. (internal quotation marks omitted).
Of course, some errors cannot be waived—subject matter jurisdiction chief among them. In an attempt to shoehorn this case into that category, the dissent hints that a federal court may be without jurisdiction to approve a settlement agreement that requires the Executive to make an unappropriated expenditure. No authority is cited for that proposition and the legal signposts in this area instead point in the opposite direction. See Local No. 93 v. City of Cleveland, 478 U.S. 501, 523, 106 S.Ct. 3063, 92 L.Ed.2d 405 (1986) (“[T]he mere existence of an unexercised power to modify the obligations contained in a consent decree does not alter the fact that those obligations were created by agreement of the parties rather than imposed by the court.”); cf. Authority of the United States to Enter Settlements Limiting the Future Exercise of Executive Branch Discretion, 23 Op. O.L.C. 126, 128 (1999) (‘We do not believe ... that Article III bars federal courts from entering consent decrees that limit executive branch discretion whenever such decrees purport to provide broader relief than a court could have awarded pursuant to an ordinary injunction.”).
*1058Even if the “extraordinary circumstances” standard cited in the dissent were applicable, its terms are not met. The dissent claims that the “proper resolution is not in doubt” because we are presented with a “fully briefed, purely legal question.” See Dis. Op. at 1066-67. But, of course, the proper resolution of even fully briefed, purely legal questions can be doubtful. Circuit splits happen.
Moreover, the question presented here is not purely legal. Congress has appropriated funds for the Executive to settle “claims ... for defense of imminent litigation or suits against the United States ... [which] shall be settled and paid in a manner similar to judgments in like causes.” 28 U.S.C. § 2414. Concededly, the nonprofit organizations that will receive cy-pres distributions out of leftover settlement funds may not possess any claims against the United States. But there is no denying that the Settlement Agreement did in fact settle claims against the United States; namely, the claims of the members of the class. The question of whether this settlement was supported by a congressional appropriation turns on whether providing for cy-pres distribution of unclaimed settlement funds is “similar to judgments in like causes.” Id. That requires answering at least two questions that are not “purely legal”: What are “like causes” to this one? And how are judgments in such causes settled and paid? The record offers no answers to these questions, nor should it, because Appellant Mandan told the District Court that this issue was off the table. Not only has there been no fact-finding on these questions, there has been no adversarial presentation; these questions are not “fully briefed.” Nor can we say, without a proper factual record, that the proper resolution of the merits question is, in fact, “beyond any doubt,” Singleton v. Wulff, 428 U.S. 106, 121, 96 S.Ct. 2868, 49 L.Ed.2d 826 (1976).
These conclusions are not the product of “schadenfreude” or disrespect for the importance of separation of powers. Rather, they are mandated by a commitment to the proper role of the appellate court and the system of adversarial presentation. When properly presented in a case or controversy, courts vindicate the constitutional scheme of separation of powers. Otherwise, allegations of trespass onto Congress’ constitutional curtilage must be addressed by Congress—not the courts— and Congress has ample weaponry with which to defend its turf.