KAVANAUGH, Circuit Judge:

The United States is at war against al Qaeda, an international terrorist organization. Al Qaeda's stated goals are, among other things, to drive the United States from posts in the Middle East, to devastate the State of Israel, and to help establish radical Islamic control over the Greater Middle East. Al Qaeda uses terror to advance its broad objectives. Al Qaeda terrorists do not wear uniforms, and they target American civilians and members of the U.S. Military, as well as U.S. allies. After al Qaeda's attacks on the United States on September 11, 2001, Congress authorized the President to wage war against al Qaeda. That war continues.

In war, when the United States captures or takes custody of alien enemy combatants or their substantial supporters, it may detain them for the duration of hostilities. Moreover, the United States may try unlawful alien enemy combatants before military commissions for their war crimes. See Hamdi v. Rumsfeld, 542 U.S. 507, 518-24, 124 S.Ct. 2633, 159 L.Ed.2d 578 (2004); Ex parte Quirin, 317 U.S. 1, 26-45, 63 S.Ct. 2, 87 L.Ed. 3 (1942).

This case raises questions about the scope of the Executive's authority to prosecute war crimes under current federal statutes.

This particular dispute involves the military commission conviction of Salim Hamdan, an al Qaeda member who worked for Osama bin Laden. In 2001, Hamdan was captured in Afghanistan. He was later transferred to the U.S. Naval Base at Guantanamo Bay, Cuba.

Hamdan was not just detained at Guantanamo as an enemy combatant. He was also accused of being an unlawful enemy combatant and was tried and convicted by a military commission for "material support for terrorism," a war crime specified by the Military Commissions Act of 2006. See 10 U.S.C. § 950t(25); see also 10 U.S.C. § 950v(b)(25) (2006) (previous codification of same provision). Hamdan's conviction was based on actions he took from 1996 to 2001 — before enactment of the Military Commissions Act. At the time of Hamdan's conduct, the extant federal statute authorized and limited military commissions to try violations of the "law of war." 10 U.S.C. § 821.

As punishment for his war crime, Hamdan was sentenced by the military commission to 66 months' imprisonment, with credit for some time already served. Hamdan's sentence expired in 2008. Although the United States may have continued to detain Hamdan until the end of hostilities pursuant to its wartime detention authority, see Hamdi, 542 U.S. at 518-22, 124 S.Ct. 2633, Hamdan was transferred in late 2008 to Yemen and then released there. Even after his release, Hamdan has continued to appeal his U.S. war crimes conviction.

This appeal presents several issues. First, is the dispute moot because Hamdan has already served his sentence and been released from U.S. custody? Second, does the Executive have authority to prosecute Hamdan for material support for terrorism on the sole basis of the 2006 Military Commissions Act — which specifically lists material support for terrorism as a war crime triable by military commission — even though Hamdan's conduct occurred from 1996 to 2001, before enactment of that Act? Third, if not, did the pre-existing statute that authorized war-crimes military commissions at the time of Hamdan's conduct — a statute providing that military commissions may try violations of the "law of war," 10 U.S.C. § 821 — proscribe material support for terrorism as a war crime?

We conclude as follows:

First, despite Hamdan's release from custody, this case is not moot. This is a direct appeal of a conviction. The Supreme Court has long held that a defendant's direct appeal of a conviction is not mooted by the defendant's release from custody.

Second, consistent with Congress's stated intent and so as to avoid a serious Ex Post Facto Clause issue, we interpret the Military Commissions Act of 2006 not to authorize retroactive prosecution of crimes that were not prohibited as war crimes triable by military commission under U.S. law at the time the conduct occurred. Therefore, Hamdan's conviction may be affirmed only if the relevant statute that was on the books at the time of his conduct — 10 U.S.C. § 821 — encompassed material support for terrorism.

Third, when Hamdan committed the relevant conduct from 1996 to 2001, Section 821 of Title 10 provided that military commissions may try violations of the "law of war." The "law of war" cross-referenced in that statute is the international law of war. See Quirin, 317 U.S. at 27-30, 35-36, 63 S.Ct. 2. When Hamdan committed the conduct in question, the international law of war proscribed a variety of war crimes, including forms of terrorism. At that time, however, the international law of war did not proscribe material support for terrorism as a war crime. Indeed, the Executive Branch acknowledges that the international law of war did not — and still does not — identify material support for terrorism as a war crime. Therefore, the relevant statute at the time of Hamdan's conduct — 10 U.S.C. § 821 — did not proscribe material support for terrorism as a war crime.

Because we read the Military Commissions Act not to retroactively punish new crimes, and because material support for terrorism was not a pre-existing war crime under 10 U.S.C. § 821, Hamdan's conviction for material support for terrorism cannot stand. We reverse the judgment of the Court of Military Commission Review and direct that Hamdan's conviction for material support for terrorism be vacated.1

I

In 1996, Salim Hamdan traveled from his native Yemen to Pakistan and then to Afghanistan to participate in jihad. In Afghanistan, Hamdan attended an al Qaeda training camp. At the camp, Hamdan received weapons training, met Osama bin Laden, and listened to bin Laden's lectures.

Later in 1996, Hamdan became an al Qaeda driver. His duties included transporting personnel, supplies, and weapons between an al Qaeda guesthouse and al Qaeda's al Farouq training camp in Afghanistan. Eventually, Hamdan became Osama bin Laden's personal driver and bodyguard.

In August 1996, Osama bin Laden publicly declared war on the United States. That declaration came after various al Qaeda terrorist attacks, including the 1993 bombing of the World Trade Center. In 1998, bin Laden issued a fatwa calling for the indiscriminate killing of Americans, including American civilians. Hamdan was fully aware of bin Laden's public statements targeting the United States.

In August 1998, al Qaeda operatives bombed U.S. Embassies in Kenya and Tanzania, killing 257 people, including 12 Americans. Hamdan was generally aware that such an attack was planned. Around the time of the attack, Hamdan assisted Osama bin Laden in evacuating from Kandahar and moving around Afghanistan.

Later in August 1998, asserting the President's Article II power of self-defense, President Clinton ordered the U.S. Military to bomb targets in Afghanistan in an attempt to kill bin Laden. Bin Laden narrowly avoided being killed in that military action.

In October 2000, at the direction of bin Laden and senior al Qaeda leaders, al Qaeda bombed the U.S.S. Cole off the coast of Yemen, killing 17 Americans and injuring many others. Around that time, Hamdan returned to Afghanistan from Yemen.

In August 2001, Hamdan drove bin Laden to various planning meetings in Afghanistan. Several days before September 11, 2001, bin Laden told Hamdan that they had to evacuate their compound because of an impending operation. Hamdan drove bin Laden to Kabul. They later moved to a series of locations around Afghanistan.

On September 11, 2001, al Qaeda attacked the United States, killing thousands of civilians and causing massive long-term damage to the American economy and way of life.

In the days following the attacks of September 11, 2001, Congress passed and President George W. Bush signed the Authorization for Use of Military Force. That law authorized the President

Pub. L. No. 107-40, 115 Stat. 224 (2001).

Consistent with the 2001 Authorization for Use of Military Force, President Bush directed the use of force to kill or capture and detain al Qaeda operatives, and where appropriate to try unlawful al Qaeda combatants who had committed war crimes. On October 7, 2001, as part of the overall operation, President Bush ordered U.S. troops into Afghanistan to wage war against al Qaeda there, as well as against the Taliban government that was in control of Afghanistan and had been supporting and harboring al Qaeda.

On November 13, 2001, the President issued an executive order establishing military commissions to try al Qaeda members and aiders and abettors who had committed war crimes as defined under the "laws of war" or other "applicable laws." Military Order of Nov. 13, 2001, 66 Fed. Reg. 57,833; 57,833-34. The executive order did not purport to rely solely on the President's constitutional authority; rather, it cited two separate statutes as congressional authorization for the President to employ military commissions: the 2001 Authorization for Use of Military Force and 10 U.S.C. § 821, the long-standing statute that authorized military commissions to try violations of the "law of war."

In November 2001, Hamdan was captured in Afghanistan while driving toward Kandahar. The car he was driving contained two anti-aircraft missiles. Also in the car was an al Qaeda-issued document that authorized the bearer to carry a weapon in Afghanistan. Hamdan's captors turned him over to U.S. authorities. He was later transferred to Guantanamo Bay, Cuba, and the U.S. Military detained him there as an enemy combatant.

At Guantanamo, Hamdan not only was detained as an enemy combatant but also was eventually charged with one count of conspiracy and was to be tried before a military commission as an unlawful enemy combatant who had committed war crimes.2 Hamdan raised various legal objections to the prosecution, and the case ultimately wound its way to the Supreme Court. The Supreme Court held that the military commission rules then in place contravened statutory limits because the rules did not comply in certain respects with statutory restrictions contained in 10 U.S.C. § 836. See Hamdan v. Rumsfeld, 548 U.S. 557, 613-35, 126 S.Ct. 2749, 165 L.Ed.2d 723 (2006). The Court split 4-3 on and thus did not decide a separate issue: whether conspiracy was a cognizable charge in a military commission under the "law of war" for purposes of 10 U.S.C. § 821. Compare Hamdan, 548 U.S. at 595-612, 126 S.Ct. 2749 (Stevens, J., plurality opinion) (conspiracy is not a law of war crime), with id. at 697-706, 126 S.Ct. 2749 (Thomas, J., dissenting) (conspiracy is a law of war crime). (Justice Kennedy did not address that issue; Chief Justice Roberts did not participate in the case.)

In the Hamdan case, several Justices specifically invited Congress to clarify the scope of the President's statutory authority to use military commissions to try unlawful alien enemy combatants for war crimes. See Hamdan, 548 U.S. at 636, 126 S.Ct. 2749 (Breyer, J., concurring); id. at 636-37, 126 S.Ct. 2749 (Kennedy, J., concurring).

In the wake of the Supreme Court's decision in Hamdan, Congress enacted a new military commissions statute. See Military Commissions Act of 2006, Pub. L. No. 109-366, 120 Stat. 2600. Of particular relevance here, Congress expanded military commissions beyond prosecuting violations of the generic "law of war," spying, and aiding the enemy, which were the crimes listed by statute at the time. See 10 U.S.C. §§ 821, 904, 906. Of most importance here, Congress alleviated some of the uncertainty highlighted in Hamdan about the phrase "law of war" in 10 U.S.C. § 821 by listing a large number of specific war crimes that could be charged by military commission, including conspiracy and material support for terrorism. See § 3(a), 120 Stat. at 2630. (In 2009, Congress enacted a new Military Commissions Act; that law did not make changes relevant to this case. See Pub. L. No. 111-84, 123 Stat. 2574.)

After passage of the 2006 Military Commissions Act, Hamdan was charged anew before a U.S. military commission on one charge of conspiracy and one charge, containing eight specifications, of material support for terrorism.

At his military commission trial, Hamdan was acquitted of conspiracy but convicted of five specifications of material support for terrorism. In August 2008, he was sentenced to 66 months' confinement and credited for having already served most of that time.

When his sentence ended later in 2008, the war against al Qaeda had not ended. Therefore, the United States may have continued to detain Hamdan as an enemy combatant. See Hamdan, 548 U.S. at 635, 126 S.Ct. 2749; Hamdi v. Rumsfeld, 542 U.S. 507, 518-24, 124 S.Ct. 2633, 159 L.Ed.2d 578 (2004). But in November 2008, Hamdan was transferred by the U.S. Military to Yemen, and he was then released on or about January 8, 2009, in Yemen.

After his release, Hamdan nonetheless continued to appeal his U.S. military commission conviction. On appeal to the en banc Court of Military Commission Review, Hamdan argued (i) that Congress lacked authority under Article I of the Constitution to make material support for terrorism a war crime triable by military commission; (ii) that in any event, the 2006 Military Commissions Act, which listed material support for terrorism as a war crime, could not be retroactively applied to him because his conduct occurred from 1996 to 2001; and (iii) that the statute in effect at the time of his alleged conduct — 10 U.S.C. § 821, which limited military commissions to violations of the "law of war" — did not authorize prosecution of material support for terrorism as a war crime. In 2011, the Court of Military Commission Review affirmed the conviction. See United States v. Hamdan, 801 F.Supp.2d 1247 (C.M.C.R.2011) (en banc).

By statute, Hamdan has an automatic right of appeal to this Court. See 10 U.S.C. § 950g.

II

We must first address the issue of mootness — that is, whether this appeal is moot because Hamdan has been released from U.S. custody. Although the parties agree that the appeal is not moot, mootness is a jurisdictional question that we must independently consider. See United States v. Juvenile Male, ___ U.S. ___, 131 S.Ct. 2860, 2864-65, 180 L.Ed.2d 811 (2011); Sibron v. New York, 392 U.S. 40, 50 n. 8, 88 S.Ct. 1889, 20 L.Ed.2d 917 (1968).

This case is a direct appeal of a military commission conviction. In the criminal context, a direct appeal of a criminal conviction is not mooted by a defendant's release from custody. See Sibron, 392 U.S. 40, 88 S.Ct. 1889. The Supreme Court has so ruled in part because of the collateral legal consequences of a conviction — namely, the possibility that the defendant could commit or be tried for a new offense, the punishment for which could take account of a past conviction. Those collateral consequences are of course present in virtually all criminal cases (other than, for example, when the defendant has died after the conviction and thus obviously cannot commit a new offense). The same collateral consequences are present in military commission conviction cases. See, e.g., MANUAL FOR MILITARY COMMISSIONS, Rules 1001(a)(2), 1001(b)(1)(A) (2012) (in military commission sentencing, the prosecution may "introduce evidence of military or civilian convictions, foreign or domestic, of the accused" as an aggravating factor); 18 U.S.C. § 3553(a)(1) (sentencing courts shall take into account "the history and characteristics of the defendant").3 Applying the relevant Supreme Court precedent, we therefore conclude that a direct appeal of a military commission conviction is likewise not mooted by the defendant's release.

To be sure, that principle generally does not apply to the habeas context where a detainee is challenging the basis for executive detention. Such a habeas case is sometimes moot after the detainee's release. See Spencer v. Kemna, 523 U.S. 1, 8-14, 118 S.Ct. 978, 140 L.Ed.2d 43 (1998); Gul v. Obama, 652 F.3d 12, 17 (D.C.Cir. 2011). In our recent habeas decision in Gul, where a former Guantanamo detainee objected to a military detention determination after his release, this Court dismissed the case as moot.

But Hamdan is not just a military detainee; he has been convicted of a war crime by military commission. Therefore, our recent decision in Gul does not control here. Rather, this case is controlled by the principle that a direct appeal of a conviction is not mooted by the defendant's release from custody.

This case is not moot.

III

Under a law now codified at 10 U.S.C. § 821, Congress has long authorized the Executive to use military commissions to try war crimes committed by the enemy. See Ex parte Quirin, 317 U.S. 1, 63 S.Ct. 2, 87 L.Ed. 3 (1942). That statute authorizes military commissions to try violations of the "law of war" — a term, as we explain below, that has long been understood to mean the international law of war. See Hamdan v. Rumsfeld, 548 U.S. 557, 603, 610, 126 S.Ct. 2749, 165 L.Ed.2d 723 (2006) (plurality); id. at 641, 126 S.Ct. 2749 (Kennedy, J., concurring); Quirin, 317 U.S. at 27-30, 35-36, 63 S.Ct. 2. Two other longstanding statutes separately authorize military commission prosecutions for spying and aiding the enemy. See 10 U.S.C. §§ 904, 906.4

After the Supreme Court's 2006 decision in Hamdan, Congress enacted a new military commissions statute that, among other things, clarified the scope of the Executive's authority to try war crimes. See Military Commissions Act of 2006, Pub. L. No. 109-366, 120 Stat. 2600. Of particular relevance here, Congress expanded military commissions beyond trying violations of the generic "law of war," spying, and aiding the enemy. Congress instead also listed a large number of specific war crimes that could be tried by military commission, including conspiracy and material support for terrorism. See id. § 3(a), 120 Stat. at 2630 (now codified at 10 U.S.C. § 950t).

Hamdan argues that Congress lacked authority under Article I of the Constitution — namely, the Define and Punish Clause — to define material support for terrorism as a war crime subject to trial by a U.S. military commission.5 Hamdan maintains that Congress's authority under the Define and Punish Clause is limited to proscribing offenses that are already illegal under international law. And Hamdan contends that material support for terrorism is not a recognized international-law war crime. The Government responds that Hamdan's focus on the Define and Punish Clause alone is misplaced. According to the Government, the Declare War Clause and other war clauses in Article I, as supplemented by the Necessary and Proper Clause, independently authorize Congress to establish military commissions to try an enemy's war crimes. And the Government further contends that Congress's broad authority under the Declare War Clause is not constrained by the evolving and often difficult to discern standards of international law. Therefore, as the Government sees it, Congress has authority to make material support for terrorism a war crime triable by military commission.

We do not decide that antecedent question. Even assuming arguendo that Congress had authority under its various Article I war powers to establish material support for terrorism as a war crime in the Military Commissions Act of 2006,6 we conclude that the Act did not authorize retroactive prosecution for conduct that was committed before the Act's enactment and was not prohibited by U.S. law at the time the conduct occurred. Here, Hamdan's conduct occurred from 1996 to 2001 — before enactment of the Military Commissions Act. And as we will explain, the federal statute in effect at the time of Hamdan's conduct — 10 U.S.C. § 821 — did not authorize prosecution for material support for terrorism.

A

As is clear from the text of the Military Commissions Act of 2006, Congress was quite concerned about the ex post facto implications of retroactively prosecuting someone under the Act for conduct committed before its enactment. Congress tried to deal with any ex post facto problem by declaring in the text of the statute that "[t]he provisions of this subchapter codify offenses that have traditionally been triable by military commissions. This chapter does not establish new crimes that did not exist before its enactment, but rather codifies those crimes for trial by military commission." § 3(a), 120 Stat. at 2624. The Act continued: "Because the provisions of this subchapter (including provisions that incorporate definitions in other provisions of law) are declarative of existing law, they do not preclude trial for crimes that occurred before the date of the enactment of this chapter." Id.

As Congress well understood when it appended this unusual statement to the statute, the U.S. Constitution bars Congress from enacting punitive ex post facto laws. See U.S. CONST. art. I, § 9, cl. 3 ("No Bill of Attainder or ex post facto Law shall be passed."). Among other things, the Ex Post Facto Clause bars laws that retroactively punish conduct that was not previously prohibited, or that retroactively increase punishment for already prohibited conduct. See Collins v. Youngblood, 497 U.S. 37, 42, 110 S.Ct. 2715, 111 L.Ed.2d 30 (1990); Calder v. Bull, 3 U.S. 386, 3 Dall. 386, 1 L.Ed. 648 (1798) (opinion of Chase, J.). The Ex Post Facto Clause thus prevents Congress and the Executive from retroactively applying a federal criminal statute to conduct committed before the statute was enacted.

As Congress itself recognized in the statutory text, retroactive prosecution by military commission could similarly raise serious constitutional issues, at the very least. As stated in the statutory text, however, Congress believed that the Act codified no new crimes and thus posed no ex post facto problem. As we explain below, Congress's premise was incorrect. The statute does codify some new war crimes, including material support for terrorism. The question for ex post facto purposes is this: If Congress had known that the Act was codifying some new crimes, would Congress have wanted the new crimes to be enforced retroactively? To begin with, the statutory text reveals a tight causal link between (i) Congress's belief that the statute codified only crimes under pre-existing law and (ii) Congress's statement that the statute could therefore apply to conduct before enactment. That causal link suggests that Congress would not have wanted new crimes to be applied retroactively. The Executive Branch agrees with that interpretation of the statute, stating: "Congress incorporated ex post facto principles into the terms of the MCA itself." Brief for the United States at 66. At a minimum, we know that the statutory text does not contemplate or address the possibility of retroactively applying new crimes, leaving us with at least something of an ambiguity. And courts interpret ambiguous statutes to avoid serious questions of unconstitutionality. See Rapanos v. United States, 547 U.S. 715, 738, 126 S.Ct. 2208, 165 L.Ed.2d 159 (2006) (plurality opinion of Scalia, J.) (constitutional avoidance where statute "raises difficult questions" of constitutionality); Cherokee Nation of Oklahoma v. Leavitt, 543 U.S. 631, 646, 125 S.Ct. 1172, 161 L.Ed.2d 66 (2005) (avoiding an interpretation that "may violate the Constitution"). To avoid the prospect of an Ex Post Facto Clause violation here, we interpret the Military Commissions Act of 2006 so that it does not authorize retroactive prosecution for conduct committed before enactment of that Act unless the conduct was already prohibited under existing U.S. law as a war crime triable by military commission. In this case, therefore, Hamdan's conviction stands or falls on whether his conduct was prohibited by the pre-existing statute, 10 U.S.C. § 821, at the time he committed the conduct.7

B

Before enactment of the Military Commissions Act in 2006, U.S. military commissions could prosecute war crimes under 10 U.S.C. § 821 for violations of the "law of war." The Government suggests that at the time of Hamdan's conduct from 1996 to 2001, material support for terrorism violated the "law of war" referenced in 10 U.S.C. § 821. It is true that in the text of the Military Commissions Act of 2006, Congress declared its belief that material support for terrorism was a pre-existing crime under the law of war and thus under 10 U.S.C. § 821. See § 3a, 120 Stat. at 2624. But exercising our independent review, as we must when considering the ex post facto implications of a new law, see Calder v. Bull, 3 U.S. 386, 3 Dall. 386, 1 L.Ed. 648 (1798) (opinion of Chase, J.); Marbury v. Madison, 5 U.S. 137, 1 Cranch 137, 2 L.Ed. 60 (1803), we conclude otherwise. Material support for terrorism was not a war crime under the law of war referenced in 10 U.S.C. § 821 at the time of Hamdan's conduct.

Analysis of this issue begins by determining what body of law is encompassed by the term "law of war" in 10 U.S.C. § 821. The Supreme Court's precedents tell us: The "law of war" referenced in 10 U.S.C. § 821 is the international law of war.8 See Hamdan, 548 U.S. at 603, 126 S.Ct. 2749 (plurality) (act is law of war offense when "universal agreement and practice both in this country and internationally" recognize it as such) (internal quotation marks omitted); id. at 610, 126 S.Ct. 2749 (analyzing international sources to determine whether conspiracy was "recognized violation of the law of war"); id. at 641, 126 S.Ct. 2749 (Kennedy, J., concurring) ("the law of war" referenced in 10 U.S.C. § 821 "derives from rules and precepts of the law of nations" and is "the body of international law governing armed conflict") (internal quotation marks omitted); Quirin, 317 U.S. at 29, 63 S.Ct. 2 ("law of war" referenced in 10 U.S.C. § 821 is a "branch of international law"); id. at 27-28, 63 S.Ct. 2 (The "law of war" is "that part of the law of nations which prescribes, for the conduct of war, the status, rights and duties of enemy nations as well as of enemy individuals."); see also Instructions for the Government of Armies of the United States in the Field (Lieber Code), General Orders No. 100, arts. 27 & 40 (Apr. 24, 1863) (describing the law of war as a "branch" of the "law of nations"); O.L.C. Memorandum from Patrick F. Philbin to Alberto R. Gonzales 5 (Nov. 6, 2001) ("laws of war" are "considered a part of the `Law of Nations'"); id. at 29, 63 S.Ct. 2 ("the term `law of war' used in 10 U.S.C. § 821 refers to the same body of international law now usually referred to as the `laws of armed conflict'").9

We turn, then, to the question whether material support for terrorism is an international-law war crime.

It is true that international law establishes at least some forms of terrorism, including the intentional targeting of civilian populations, as war crimes. See, e.g., Rome Statute of the International Criminal Court art. 8(2)(b), July 17, 1998, 2187 U.N.T.S. 90; Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Geneva IV), art. 33, Aug. 12, 1949, 6 U.S.T. 3516, 75 U.N.T.S. 287; COMMISSION OF RESPONSIBILITIES, CONFERENCE OF PARIS 1919, VIOLATION OF THE LAWS AND CUSTOMS OF WAR 17 (Clarendon Press 1919) (the Allied Nations condemned Germany for "the execution of a system of terrorism" after World War I).

But the issue here is whether material support for terrorism is an international-law war crime. The answer is no. International law leaves it to individual nations to proscribe material support for terrorism under their domestic laws if they so choose. There is no international-law proscription of material support for terrorism.

To begin with, there are no relevant international treaties that make material support for terrorism a recognized international-law war crime. Neither the Hague Convention nor the Geneva Conventions — the sources that are "the major treaties on the law of war" — acknowledge material support for terrorism as a war crime. See Hamdan, 548 U.S. at 604, 126 S.Ct. 2749 (plurality); Geneva Convention Relative to the Protection of Civilian Persons in Time of War (Geneva IV), Aug. 12, 1949, 6 U.S.T. 3516, 75 U.N.T.S. 287; Hague Convention (IV) Respecting the Laws and Customs of War on Land and Its Annex, Oct. 18, 1907, 36 Stat. 2277.

Nor does customary international law otherwise make material support for terrorism a war crime. Customary international law is a kind of common law; it is the body of international legal principles said to reflect the consistent and settled practice of nations. See RESTATEMENT (THIRD) OF FOREIGN RELATIONS LAW OF THE UNITED STATES § 102(2) (1987) ("Customary international law results from a general and consistent practice of states followed by them from a sense of legal obligation"). It is often difficult to determine what constitutes customary international law, who defines customary international law, and how firmly established a norm has to be to qualify as a customary international law norm. Cf. Sosa v. Alvarez-Machain, 542 U.S. 692, 124 S.Ct. 2739, 159 L.Ed.2d 718 (2004).10

But here, the content of customary international law is quite evident. Material support for terrorism was not a recognized violation of the international law of war as of 2001 (or even today, for that matter). As we have noted, the Geneva Conventions and the Hague Convention do not prohibit material support for terrorism. The 1998 Rome Statute of the International Criminal Court, which catalogues an extensive list of international war crimes, makes no mention of material support for terrorism. See Rome Statute of the International Criminal Court, July 17, 1998, 2187 U.N.T.S. 90. Nor does the Statute of the International Tribunal for the Former Yugoslavia, the Statute of the International Tribunal for Rwanda, or the Statute of the Special Court for Sierra Leone. See Statute of the International Tribunal for the Former Yugoslavia, adopted by S.C. Res. 827, U.N. Doc. S/RES/827 (1993), reprinted in 32 I.L.M. 1159, 1192; Statute of the International Tribunal for Rwanda, adopted by S.C. Res. 955, U.N. Doc. S/RES/955 (1994), reprinted in 33 I.L.M. 1598, 1602 (includes terrorism itself as a crime); Statute of the Special Court for Sierra Leone art. 3(d), Jan. 16, 2002, 2178 U.N.T.S. 138 (same). Nor have any international tribunals exercising common-law-type power determined that material support for terrorism is an international-law war crime.

Commentators on international law have similarly explained that material support for terrorism is not an international-law war crime. See, e.g., ANDREA BIANCHI & YASMIN NAQVI, INTERNATIONAL HUMANITARIAN LAW AND TERRORISM 244 (2011) ("there is little evidence" that a proscription of "material support for terrorism" is "considered to be part of the laws and customs of war"). Nor is the offense of material support for terrorism listed in the JAG handbook on the law of war. See U.S. ARMY JAG, LAW OF WAR HANDBOOK (Maj. Keith E. Puls ed., 2005); see also Jennifer K. Elsea, The Military Commissions Act of 2006: Analysis of Procedural Rules and Comparison with Previous DOD Rules and the Uniform Code of Military Justice 12 (CRS, updated Sept. 27, 2007) ("defining as a war crime the `material support for terrorism' does not appear to be supported by historical precedent") (footnote omitted).

In short, neither the major conventions on the law of war nor prominent modern international tribunals nor leading international-law experts have identified material support for terrorism as a war crime. Perhaps most telling, before this case, no person has ever been tried by an international-law war crimes tribunal for material support for terrorism.

Not surprisingly, therefore, even the U.S. Government concedes in this case that material support for terrorism is not a recognized international-law war crime. No treaty that the Government has cited or that we are aware of identifies material support for terrorism as a war crime. And the Government further admits: The "offense of providing material support to terrorism, like spying and aiding the enemy, has not attained international recognition at this time as a violation of customary international law." Brief for the United States at 48; see also id. at 55-56 (same).

To be sure, there is a strong argument that aiding and abetting a recognized international-law war crime such as terrorism is itself an international-law war crime. And there are other similar war crimes. But Hamdan was not charged with aiding and abetting terrorism or some other similar war crime. He was charged with material support for terrorism. And as the Government acknowledges, aiding and abetting terrorism prohibits different conduct, imposes different mens rea requirements, and entails different causation standards than material support for terrorism. If the Government wanted to charge Hamdan with aiding and abetting terrorism or some other war crime that was sufficiently rooted in the international law of war (and thus covered by 10 U.S.C. § 821) at the time of Hamdan's conduct, it should have done so.

The Government latches on to a few isolated precedents from the Civil War era to prop up its assertion that material support for terrorism was a pre-existing war crime as of 2001 for purposes of 10 U.S.C. § 821. There are several independent reasons that those cases fail to support the Government's argument. First, the Civil War cases did not involve any charges of material support for terrorism. Instead, several cases involve guerillas who were punished for taking up "arms" as "insurgents" — that is, for direct attacks rather than material support. See, e.g., G.O. No. 15, HQ, Dep't of the Mississippi (Apr. 3, 1862), 1 OR ser. II, at 472-76. Others were convicted of "joining, aiding and assisting a band of robbers and bandits" — in other words, what we would likely call aiding and abetting, not material support. G.O. No. 19, HQ, Dep't of the Mississippi (Apr. 24, 1862), 1 OR ser. II, at 478. In short, those precedents are at best murky guidance here. Cf. Hamdan, 548 U.S. at 602, 126 S.Ct. 2749 (plurality) (requiring "plain and unambiguous" precedent). Second, those Civil War commissions were in part military tribunals governing certain territory — which are a separate form of military commission subject to a separate branch of law, and not the kind of law-of-war military commission at issue here. As others have suggested, their precedential value is therefore limited. See Hamdan, 548 U.S. at 596 n. 27, 126 S.Ct. 2749; id. at 608, 126 S.Ct. 2749 (plurality) (The "military commissions convened during the Civil War functioned at once as martial law or military government tribunals and as law-of-war commissions. Accordingly, they regularly tried war crimes and ordinary crimes together.") (citation omitted). Third, and perhaps most to the point, those cases do not establish that material support for terrorism was a war crime recognized under international law as of 1996 to 2001 when Hamdan committed his conduct, which is the relevant inquiry under 10 U.S.C. § 821. The Government contends that those Civil War precedents illuminate what it calls the "U.S. common law of war" — not the international law of war. But the statutory constraint here imposed by 10 U.S.C. § 821 is the international law of war. As the Government told the Supreme Court in Quirin, "This `common law of war' is a centuries-old body of largely unwritten rules and principles of international law which governs the behavior of both soldiers and civilians during time of war." Brief for the United States at 29, in Quirin, 317 U.S. 1, 63 S.Ct. 2 (emphasis added) (citation omitted). To be sure, U.S. precedents may inform the content of international law. But those Civil War precedents fail to establish material support for terrorism as a war crime under the international law of war as of 1996 to 2001. And even the Government admits that material support for terrorism was not an international-law war crime as of 1996 to 2001.

In short, material support for terrorism was not an international-law war crime under 10 U.S.C. § 821 at the time Hamdan engaged in the relevant conduct.

* * *

Because we read the Military Commissions Act not to sanction retroactive punishment for new crimes, and because material support for terrorism was not a pre-existing war crime under 10 U.S.C. § 821, Hamdan's conviction for material support for terrorism cannot stand. We reverse the decision of the Court of Military Commission Review and direct that Hamdan's conviction for material support for terrorism be vacated.

So ordered.

GINSBURG, Senior Circuit Judge, concurring:

I join the decision of the Court but, with respect to its holding Mr. Hamdan's appeal of his criminal conviction by military commission is not moot, I do so only because precedent so dictates. I write separately to explain the unfortunate state of that precedent, which requires us to review the conviction of a man long since transferred to Yemen who, even if his conviction were overturned and he were to hear of it, would not be affected in any way by the result. Because today's decision has no actual consequence for Mr. Hamdan, his case is moot in fact, though, curiously, not in law.

The Supreme Court "presumes" the appeal of a criminal conviction is not moot unless "it is shown that there is no possibility that any collateral legal consequences will be imposed on the basis of the challenged conviction." Sibron v. New York, 392 U.S. 40, 57, 88 S.Ct. 1889, 20 L.Ed.2d 917 (1968); see United States v. Juvenile Male, ___ U.S. ___, 131 S.Ct. 2860, 2864, 180 L.Ed.2d 811 (2011) (per curiam) ("When the defendant challenges his underlying conviction, this Court's cases have long presumed the existence of collateral consequences"). The Government concedes, as Hamdan's counsel contends, that it cannot show there is "no possibility" Hamdan's conviction will have a collateral legal consequence for him. The parties' mutual desire to have the court decide this case on its merits is of no moment, however; an Article III court has an "independent obligation to be sure [it] ha[s] jurisdiction," High Plains Wireless, LP v. FCC, 276 F.3d 599, 605 (D.C.Cir.2002), which here requires us to determine whether the case has become moot on appeal.

A criminal conviction may and often does have consequences beyond the penalties imposed in the sentence. In Sibron, the Court held the defendant's appeal of his conviction was not moot, even though his sentence had expired during the pendency of the appeal, because that conviction, left undisturbed, could increase his sentence if he were later to be convicted of another crime. 392 U.S. at 56, 88 S.Ct. 1889; accord United States v. Morgan, 346 U.S. 502, 512-13, 74 S.Ct. 247, 98 L.Ed. 248 (1954) ("Although the term has been served, the results of the conviction may persist. Subsequent convictions may carry heavier penalties"). Similarly, in Carafas v. LaVallee, 391 U.S. 234, 88 S.Ct. 1556, 20 L.Ed.2d 554 (1968), the Court concluded that a continuing civil disability stemming from a criminal conviction, such as a bar to voting in state elections or to serving as a juror, also keeps a criminal appeal from becoming moot. Even an adverse immigration consequence, including a bar on re-entering United States, may suffice to keep a case alive and hence to preserve appellate jurisdiction. See, e.g., United States v. Hamdi, 432 F.3d 115, 121 (2d Cir.2005).

Although, in considering a challenge to a criminal conviction,* "the Court [has] abandoned all inquiry into the actual existence of specific collateral consequences and in effect presumed that they exist[]," Sibron, 392 U.S. at 55, 88 S.Ct. 1889, the presumption is rebutted if the alleged collateral consequences are foreclosed as a matter of law. In Perez v. Greiner, 296 F.3d 123 (2d Cir.2002), the Second Circuit held moot the direct appeal of a criminal conviction on the ground there was "no material possibility that [the defendant] will suffer collateral consequences on the basis of the challenged conviction," id. at 125. The defendant in that case had been deported when his appeal was heard and was ineligible to reenter the country because of an earlier conviction. With the defendant "permanently barred from this country on a wholly separate ground, the currently challenged ... conviction [could] have no meaningful effect on his admissibility and hence [could not] serve as a possible collateral consequence." Id. at 126.

Hamdan and the Government each point to a collateral consequence of Hamdan's conviction. Hamdan claims his conviction for material support of terrorism makes him subject to permanent mandatory exclusion from the United States. See 8 U.S.C. § 1182(a)(3)(B)(i)(I), (V). For its part, the Government claims Hamdan's conviction may expose him to an enhanced sentence if in the future he commits a new offense and is tried therefore in a civilian or military court of the United States. See DEP'T OF DEFENSE, MANUAL FOR MILITARY COMMISSIONS, Rule 1001(b)(1)(A) (2010) ("The trial counsel may introduce [in a sentencing proceeding] evidence of [prior] military or civilian convictions, foreign or domestic, of the accused"); 18 U.S.C. § 3553(a)(1) (sentencing court shall consider "the history and characteristics of the defendant"). The adverse collateral consequence raised by Hamdan is foreclosed as a matter of law. The adverse collateral consequence posed by the Government is so far-fetched that the application of the Sibron presumption in this case risks making our opinion merely advisory.

The adverse immigration consequence alleged by Hamdan is impossible as a matter of law because Hamdan is already subject to mandatory exclusion from the United States regardless whether his conviction stands. The Immigration and Naturalization Act (INA) provides: "Any alien who ... has engaged in a terrorist activity ... is inadmissible." 8 U.S.C. § 1182(a)(3)(B)(i)(I). The Government provided overwhelming evidence, none of which Hamdan bothers to dispute, demonstrating that he engaged in "terrorist activity" within the meaning of the INA, including the provision of material support for terrorism.* The INA makes Hamdan inadmissible not for his conviction, which we reverse,* but rather for having knowingly supported terrorist activities, a historical fact we cannot reverse. Cf. Fletcher v. Peck, 10 U.S. (6 Cranch) 87, 135, 3 L.Ed. 162 (1810) ("The past cannot be recalled by the most absolute power"). Nor is it conceivable that an immigration court, which applies a standard of proof much more favorable to the Government than does a military commission, compare 8 U.S.C. § 1229a(c)(2) ("In the proceeding [for deciding whether an alien is admissible,] the alien has the burden of establishing ... the alien is clearly and beyond doubt entitled to be admitted"), with 10 U.S.C. § 949l(4) (in a military commission "the burden of proof to establish the guilt of the accused beyond a reasonable doubt is upon the United States"), would find Hamdan in fact had not provided any of the five types of material support for which he was convicted and therefore may be admitted to the United States.* Because Hamdan is already barred from entering the United States due to his past involvement in terrorism, his current conviction has no incremental effect upon his admissibility and hence the immigration consequence he proffers cannot serve as a basis for our jurisdiction. Cf. Gul, 652 F.3d at 19-20 (where detention at Guantanamo Bay, rather than designation as enemy combatant, is the ground for inadmissibility, immigration consequence of challenge to designation "too speculative to sustain the exercise of our jurisdiction").

The only other collateral consequence alleged is the Government's preposterously hypothetical prospect of an enhanced sentence if Hamdan is in the future convicted in the United States for committing another crime. The Supreme Court held in Sibron that the hypothetical future sentencing enhancement is sufficient to support the presumption that the conviction being appealed will have an adverse collateral consequence and hence to keep the appeal from being moot. Sibron, 392 U.S. at 56, 88 S.Ct. 1889. Subsequent cases, however, cast doubt upon the continuing validity of Sibron as applied to this case. In Spencer v. Kemna the Court declined to extend the Sibron presumption to the appeal of a parole revocation because any collateral consequence in a future sentencing "was contingent upon [the defendant again] violating the law, getting caught, and being convicted." 523 U.S. 1, 15, 118 S.Ct. 978, 140 L.Ed.2d 43 (1998); see also Lane v. Williams, 455 U.S. 624, 633 n. 13, 102 S.Ct. 1322, 71 L.Ed.2d 508 (1982) ("The parole violations that remain a part of respondents' records cannot affect a subsequent parole determination unless respondents again violate state law, are returned to prison, and become eligible for parole. Respondents themselves are able — and indeed required by law — to prevent such a possibility from occurring").

That is, although in Sibron a conviction was presumed to have adverse consequences for the defendant in a future hypothetical sentencing, in Spencer the hypothetical sentencing consequences of a defendant's parole revocation were held insufficient to keep his case from being moot because such consequences are speculative and depend upon future unlawful conduct by the defendant. Both holdings were categorical; they did not depend at all upon the particular defendant's probability of recidivating. Therefore, the defendants' future crimes, apprehension, and conviction were equally speculative in both cases. It is entirely unclear, therefore, how the hypothetical sentencing consequences of a parole revocation could be too speculative to support a finding of collateral consequences, while the hypothetical sentencing consequences of a conviction could be concrete and certain enough to support the presumption of collateral consequences, and hence Article III jurisdiction, in all criminal appeals.

Nonetheless, "[i]f a precedent of [the Supreme] Court has direct application in a case, yet appears to rest on reasons rejected in some other line of decisions, the Court of Appeals should follow the case which directly controls, leaving to [the Supreme] Court the prerogative of overruling its own decisions." Rodriguez de Quijas v. Shearson/Am. Express, Inc., 490 U.S. 477, 484, 109 S.Ct. 1917, 104 L.Ed.2d 526 (1989). Accordingly, because Hamdan's case is a direct appeal of his criminal conviction rather than review of a parole revocation as in Spencer, the Court is bound to hold the Sibron presumption applies and therefore the hypothetical future sentencing consequences of Hamdan's conviction are sufficient to keep his appeal from being moot.

Finally, I note that although this is an "appeal of a criminal conviction," we have strayed far from the familiar territory of Sibron and its progeny, which deemed sentencing consequences the antidote to mootness. The criminal conviction in each of those cases was entered in a regularly constituted civilian court, and the criminal defendant served time in a domestic prison before being released into the sovereign territory of the United States. As such, upon his release the defendant was subject to the criminal laws of the United States and of the State in which he was located. Recidivism being common in the United States, it is unfortunately reasonable to suppose such a defendant may again be convicted for violating a state or federal law.* By contrast, Hamdan is presumably in Yemen and is certainly not in the United States; so far as the record shows, he has never entered the United States nor expressed any desire to do so; and he is barred, both legally and physically, from entering the United States. As a result, the only possible future sentencing consequence of his conviction by a military commission would be through extraterritorial application of our criminal law to a federal crime yet to be committed, or through a successive prosecution in a military commission for a future violation of the law of war. It is, to say the least, far more speculative that Hamdan may find himself again being sentenced in the United States than it is an domestic criminal may recidivate and find himself before a domestic criminal court.