The Yale Law Journal


Equal Supreme Court Access for Military Personnel: An Overdue Reform

31 May 2021

abstract. Federal law currently provides for direct Supreme Court review of criminal convictions from almost all American jurisdictions, but not of most court-martial convictions. For them, an Article I court can veto access to the Supreme Court. This Essay argues for elimination of that veto.


One might think that every criminal defendant in the United States has a right to seek review by the Supreme Court. But Congress has largely blocked the path of a particular group of defendants: persons convicted at courts-martial. This is because, under the Military Justice Act of 1983 (the 1983 Act),1 the only court-martial convictions eligible for Supreme Court review are cases in which a service Court of Criminal Appeals has affirmed a death sentence; cases certified by one of the Judge Advocates General (JAG); extraordinary writ cases in which relief has been granted; and cases granted discretionary review by the United States Court of Appeals for the Armed Forces (CAAF), the highest court of the military justice system.2

In this Essay, we argue that the limitation on cases within CAAF’s discretionary jurisdiction, which account for the lion’s share of that court’s docket, is both unconstitutional and bad policy. First, by delegating to CAAF, an Article I court, the power to determine the Supreme Court’s jurisdiction over appeals, Congress violated the separation of powers. Second, in making a comparatively small class of cases nonreviewable for the ostensible purpose of reducing the Supreme Court’s caseload, Congress acted irrationally and violated Fifth Amendment equal protection. Third, by creating a category of unappealable cases that is essentially coterminous with the universe of military cases, and by conferring on CAAF a vague and effectively nonreviewable standard for granting review, Congress violated the Exceptions
Clause.3 Finally, by allowing the Supreme Court to exercise jurisdiction over cases certified for review by a JAG, but not cases in which an accused4 seeks review unless CAAF chooses to exercise its discretion to grant review, the system provides asymmetric access to justice in favor of the government.

In light of these defects, Congress should remove the jurisdictional limitations that currently deprive military personnel and anyone else subject to court-martial jurisdiction5 equal access to the Supreme Court. Whether or not a court might find the limitations unconstitutional (assuming the issue is litigable),6 Congress has an independent obligation to uphold constitutional norms.7

This Essay proceeds as follows. Part I describes the process of appellate review of courts-martial under the Uniform Code of Military Justice (UCMJ),8 as well as the significant change in Supreme Court jurisdiction that Congress enacted in 1983. Part II argues that a central feature of these changes is both unconstitutional and unfair on four grounds. The Essay concludes with an explanation of what should be done to remedy that defect and identifies other changes Congress might consider with respect to appellate review of court-martial convictions.

I. background: the military justice act of 1983

Enacted in 1950, the UCMJ prescribed for the first time a single disciplinary statute for all of the United States Armed Forces. Congress created the Court of Military Appeals to serve as a civilian court of last resort for those convicted at courts-martial for a broad array of civilian- and military-type offenses. Congress did not initially provide for direct appeal from the Court of Military Appeals to the Supreme Court. The only way a decision of the Court of Military Appeals could reach the Supreme Court was if the accused sought collateral review, such as habeas corpus,9 back-pay actions under the Tucker Act or the Little Tucker Act,10 or a federal-question civil action in district court.11 But in 1983, Congress expanded the Supreme Court’s certiorari jurisdiction to include direct appellate review of decisions of the Court of Military Appeals. That court, created under Article I of the Constitution,12 was renamed the “Court of Appeals for the Armed Forces” in 1994.13

Current law provides for direct appellate review of four categories of CAAF decisions by the Supreme Court: (1) cases in which a Court of Criminal Appeals has affirmed a death sentence, (2) cases certified for review by a JAG, (3) cases in which CAAF has granted discretionary review, and (4) any other case in which CAAF has granted relief, typically on petition for an extraordinary writ.14 Congress’s grant of jurisdiction is constitutional in the sense that Congress has the constitutional authority to extend the Supreme Court’s appellate jurisdiction to CAAF cases.15 But the Supreme Court’s jurisdiction over the enumerated categories is limited. CAAF’s docket is overwhelmingly discretionary, so most courts-martial are not afforded civilian appellate review within the military justice system. And CAAF’s decision to grant or deny a petition for review determines whether the Supreme Court can in turn exercise its certiorari jurisdiction. Article 67a,16 which Congress added to the UCMJ through the 1983 Act, expressly bars Supreme Court review of CAAF orders that deny petitions for review.17 Thus, only those Article 67(a)(3) cases in which CAAF grants discretionary review are eligible for Supreme Court review.18

Discretionary review cases account for the overwhelming majority of CAAF’s caseload. Unlike other federal or state appellate courts,19 CAAF enjoys veto power over access to the Supreme Court.20 From time to time, litigants have sought certiorari despite the 1983 Act’s limiting provisions. But these appeals are routinely denied.21 As we argue below, CAAF’s veto power offends various constitutional provisions and principles and is inequitable.

II. unconstitutional limitations

A. Separation of Powers

The 1983 Act’s jurisdictional limitations violate the separation of powers. In the recent case of United States v. Ortiz, the Supreme Court held that direct appellate review of CAAF decisions by the Supreme Court comports with Article III, even though both the military justice system and CAAF are situated within the executive branch.22

There is, however, an additional dimension to the military justice system’s placement in one of the political branches. Congress has authorized both CAAF and the four JAGs—the chief uniformed executive-branch lawyers—to substantially control the Supreme Court’s military justice docket. As mentioned above, CAAF, an Article I court located in the executive branch,23 can block Supreme Court review by denying a petition for grant of review under Article 67(a)(3). Similarly, a JAG may prevent review by refusing to “certify” a case to CAAF.24 Although the former circuit courts could control access to the Supreme Court by refusing to issue a certificate,25 they at least were part of the judicial branch. The same is true of the gatekeeping arrangements at issue in Hohn v. United States26 and Felker v. Turpin.27 Authorizing executive-branch officials like CAAF judges and JAGs to insulate specific cases from direct appellate review by the Supreme Court is thus unprecedented. If the arrangement ever came before the Supreme Court, at least three current Justices would likely object on the ground that Congress had conferred “the Government’s ‘judicial Power’ on entities outside Article III.”28

“[T]he Constitution prohibits one branch from encroaching on the central prerogatives of another.”29 For nearly a century (that is, since the Judges’ Bill of
30), one of the Supreme Court’s “central prerogatives” has been to select the cases it wishes to decide.31 The 1983 Act’s jurisdictional limitations therefore improperly place a “central prerogative” of the judicial branch in the hands of executive-branch officials.

B. Equal Protection

Because the 1983 Act’s jurisdictional limitations deny military personnel equal protection of the law, they also violate due process.32

There is no constitutional right to appellate review,33 and until 1879 federal criminal convictions were not subject to appellate review at all.34 Review of such convictions by writ of certiorari dates to the Evarts Act,35 while appellate review of courts-martial by multimember boards dates to 1920, when Congress enacted Article of War 50½.36

Congress has broad authority under the Exceptions and Regulations Clause, and no duty to expand the Supreme Court’s certiorari jurisdiction to provide direct appellate review for courts-martial. Indeed, until the 1983 Act, there was no certiorari jurisdiction over courts-martial.37 But like any federal legislation, the 1983 extension of jurisdiction must comport with the Constitution. Put differently, Congress cannot restrict the scope of its expansions of the Supreme Court’s jurisdiction in ways that violate the equal-protection component of Fifth Amendment due process.

The 1983 Act’s limitations reflected a fear that the Supreme Court might be burdened by military certiorari petitions. A House Report noted: “In view of current concerns about the Supreme Court’s docket, the legislation has been drafted in a manner that will limit the number of cases subject to direct Court review.”38

It is difficult to accept that justification given the absence of a comparable concern over any other category of litigants seeking direct appellate review.39 For example, it seems unreasonable to afford every federal or state misdemeanant untrammeled access to the Supreme Court, while denying that same access to military personnel who have been convicted of serious offenses—offenses that may be familiar civilian crimes and may have no service connection40—leading to prison terms in the double digits or even life without parole.

In the fiscal year preceding enactment of the 1983 Act, the Court of Military Appeals denied 2,421 petitions for grant of review, dismissed another 46, and either granted or remanded 310.41 Thus, as many as 2,467 additional cases might have been eligible for certiorari but for the limitations imposed by the 1983 Act.

But the actual number of eligible cases would have been substantially smaller than 2,467 because the Court of Military Appeals had long accepted Article 67(a)(3) petitions that did not identify any errors,42 even though the UCMJ has always required a showing of “good cause.”43 Because petitions for certiorari require good cause (and then some),44 many Article 67(a)(3) petitions would have been dead on arrival anyway. Indeed, CAAF’s Clerk of Court has estimated that twenty percent or less of Article 67(a)(3) petitions are submitted “on the merits”—that is, without identifying any particular errors.45 Startling as that figure is, it has been even higher in the past.46

Even assuming Congress’s caseload rationale could pass muster in principle, changed circumstances can render unlawful what was once lawful.47 And circumstances have indeed changed. CAAF’s last three annual reports chronicle a dramatic decline from 1983. In fiscal years 2018, 2019, and 2020, CAAF denied or dismissed only 328, 369, and 336 Article 67(a)(3) petitions, respectively. On average, this represents an eighty-six percent decline from 1983. The data show that any incremental burden on the Supreme Court would be much lower than when the 1983 Act was passed.

To the extent that the enactment of jurisdictional limitations reflected a prediction about how many certiorari petitions would be filed once the law took effect, experience again teaches that Congress’s fears were unfounded. According to the Pentagon’s Military Justice Review Group (headed by a respected former Chief Judge of CAAF), “Even in those cases eligible for Supreme Court review, petitions to the Supreme Court have been filed in only a fraction of the cases.”48 “On average over the past five years, fewer than a dozen petitions per year have been filed with the Supreme Court for review under Article [67a] according to data compiled by the U.S. Court of Appeals for the Armed Forces.”49 In 2010, the Congressional Budget Office predicted that a bill then under consideration “would make several hundred servicemembers eligible to file petitions each year, but that only a small portion of those individuals would pursue review by the Supreme Court (based on the experience of individuals whose cases currently qualify for Supreme Court review).”50

Even without adjusting the data for (i) the roughly twenty percent of cases that state no issues,51 (ii) the dismissals involving time-barred petitions or petitions plainly outside CAAF’s jurisdiction, and (iii) the fact that many Article 67(a)(3) petitioners who could seek certiorari do not do so, invalidating the 1983 Act’s jurisdictional limitations would make no discernible difference in the Supreme Court’s certiorari caseload, which fell to 5,411 cases in the October 2019 Term.52 Even if the changes Congress made to the first tier of appellate review in the Military Justice Act of 201653 lead to an uptick in the number of Article 67(a)(3) petitions, the rationale for the limitations, as weak as it was in 1983, will still have been seriously eroded.

The caseload-reduction rationale appears even more irrational when considered alongside other jurisdictional schemes granting litigants certiorari access. When Congress passed the 1983 Act, a handful of minor jurisdictions were still outside the reach of a writ of certiorari from the Supreme Court.54 Most of those jurisdictions are now subject to the Court’s direct appellate review.55 Except for individuals prosecuted in American Samoa, the Wake Island Court, and Native American tribal courts, every criminal defendant in the United States now has the right to apply for a writ of certiorari.56 This includes individuals convicted in the district courts, territorial courts, the local courts of the District of Columbia, state courts, and even the military commissions being tried at Guantanamo Bay, Cuba. It seems incomprehensible that Congress has afforded convicted enemy belligerents57 freer access to the Supreme Court than American military personnel. Given that general and special courts-martial are criminal proceedings, render judgments of conviction, and have the power to impose prison terms up to life without parole,58 preventing military personnel from accessing the Supreme Court is irrational on its face.

Finally, the limitations’ capricious effect is manifest not only on the face of the 1983 Act, but also in CAAF’s exercise of its authority. CAAF regularly grants review (and clears the path for potential certiorari petitions) for the sole purpose of correcting typographical or similar minor errors in proceedings below.59

C. Exceptions Clause

Article III, Section 2, of the Constitution provides that “the supreme Court shall have appellate Jurisdiction, both as to Law and Fact, with such Exceptions, and under such Regulations as the Congress shall make.”60 The 1983 Act’s jurisdictional limitations create an Exceptions Clause “exception” that is unconstitutional for three reasons.

First, an exception must be an exception, rather than a general rule. “An ‘exception’ implies a minor deviation from a surviving norm; it is a nibble, not a bite.”61 The 1983 Act’s limitations are no legislative nibble. For CAAF’s most recent term, discretionary review was denied in over eighty-nine percent of the cases that were filed under Article 67(a)(3).62

Second, the class of cases that the 1983 Act excludes is effectively undefined. Article 67(a)(3) only requires review for “good cause,” a phrase that does little to guide or limit the exercise of CAAF’s discretion.63 As a result, the jurisdictional limitations improperly delegate legislative authority.64 Indeed, “good cause” is a capacious concept, which vests significant discretion in officials exercising their pertinent powers.

Third, good-cause determinations are ordinarily subject to judicial or appellate
65 With Article 67a(a), however, Congress has insulated CAAF’s denials of review—which are by definition findings that good cause does not exist—from the only court that can properly review those denials: the Supreme Court. This arrangement creates yet another constitutional problem by vesting the power to determine the metes and bounds of an Exceptions Clause exception in a court that is subordinate to the Supreme Court. This violates the One Supreme Court Clause.66 “[T]he Constitution provides for one Supreme Court, quite as clearly as it provides for one President.”67

Unless CAAF grants discretionary review, a case that falls within Article 67(a)(3) is outside the Supreme Court’s appellate jurisdiction. CAAF thus controls a part of the Supreme Court’s docket. This arrangement improperly makes it pro tanto coequal with the Supreme Court.68

To be sure, a lower court could attempt to influence the likelihood that the Supreme Court would grant certiorari by a variety of means, such as disposing of a case by a per curiam or unpublished opinion or by calling attention to various prudential factors that militate against a grant of certiorari. But it is a different matter entirely for a lower court to possess explicit, formal power to insulate a case from direct appellate review by the Supreme Court. This arrangement is in special tension with our constitutional framework when the court enjoying that extraordinary power is an Article I court and the officials exercising it are in the executive branch.69

Remarkably, since 2010, CAAF has gone so far as to require petitioners whose cases it previously remanded and who are back before it for discretionary review to “specify” the “issue or issues on which certiorari review would be sought, whether related to the remand or to the original decision.”70 It has then proceeded to deny appellate review, even when petitioners who intended to seek certiorari have complied with its rules.71 Plainly, CAAF exercises its own judgment as to whether such petitioners have an issue that warrants review by the Supreme Court. This is gatekeeping with a vengeance.72 It is also circumstantial evidence that when a case comes before CAAF for discretionary review the first time around (that is, before a remand), CAAF may consider whether it is certworthy, since there is no obvious basis for applying a different standard of “good cause” before and after a remand.

Whether CAAF is a good or bad judge of certworthiness is immaterial given the One Supreme Court Clause.73 The only officials who can permissibly decide whether a petition meets the Supreme Court’s standards for certiorari are the Justices.74

D. Due Process

The 1983 Act’s jurisdictional limitations favor the prosecution. This is a result of the JAGs’ certification power. Under Article 67(a)(2), each JAG has the power to send a case to CAAF.75 Cases decided by CAAF on certificate for review are within the Supreme Court’s certiorari jurisdiction.76 The JAGs rarely exercise their power to certify for the benefit of an accused.77 All but one of the twelve certificates the JAGs filed in CAAF’s last three terms were for the benefit of the prosecution.78 This is not surprising since, as Judge Baker observed in United States v. Arness, the JAGs “are not independent and impartial judicial entities,” but rather “represent the government” and “are closely aligned with the government.”79 In practical terms, therefore, the certification provision affords the prosecution an appeal as of right, in contrast to the discretionary review hurdle a petitioner must overcome under the standard review provision. This produces a major asymmetry in the parties’ access to certiorari

To our knowledge, in no other jurisdiction can the prosecution appeal a final judgment of conviction as of right, while the defendant must show good cause in order to obtain discretionary review. That American appellate justice speaks with one voice on this structural question is evidence that a solitary exception (here, the military justice system) offends due process.81 This systemic asymmetry is aggravated by the fact that a JAG may cross-certify an issue for review in a case in which the accused has filed a petition for a (discretionary) grant of review in that same case. The former’s issue must be addressed by CAAF (unless, for example, it is moot or academic),82 but the latter’s need not be. Nothing inherent in the military justice system, the crimes it punishes, the interests it vindicates, or the history of the discriminatory provisions here at issue justifies this glaring disparity.83

Finally, the JAGs’ power to certify a case to CAAF and thereby render it eligible for certiorari suffers even more severely from a delegation defect than CAAF’s power to deny a petition for grant of review. This is so because the UCMJ has never prescribed a standard for certification. Congress has not “la[id] down by legislative act an intelligible principle” with which the JAGs are “directed to conform.”84 Not even “good cause,” as vague as it is,85 is required, and the Manual for Courts-Martial and service regulations, unlike CAAF’s Rule 21(b)(5), do not even attempt to shed light on the matter.86

Beyond requiring that the issue certified be one of law, the JAGs’ discretion under Article 67(a)(2) is completely unfettered. All they need do is “appropriate[ly] notif[y]” one another and the Staff Judge Advocate to the Commandant of the Marine Corps.87

The Court of Military Appeals long ago rejected an equal-protection challenge to Article 67(a)(2), reasoning that the asymmetry was warranted in the interest of fostering “certainty in, and uniformity of, interpretation of the Uniform Code in each armed force, as well as for all the armed forces.”88 Whether or not that explanation sufficed with respect to that subset of the parties’ lopsided access to CAAF before Congress expanded the scope of the Supreme Court’s certiorari jurisdiction in 1983, it falls far short of justifying the disparate access that has existed since then.


Congress acted too cautiously when it expanded the Supreme Court’s certiorari jurisdiction in 1983. Whether or not the result was fair to military personnel at the time (we do not believe it was), it is now beyond dispute that the 1983 Act’s limitations are unfair, shockingly discriminatory, and serve no substantial federal interest. If anything, they undermine efforts to promote morale within the military by treating military personnel as second-class citizens. After nearly forty years of shortchanging military personnel, those limitations should be repealed.

Congress should enact legislation that provides military personnel with equal access to the Supreme Court. This was recommended over a decade ago by the blue-ribbon nongovernmental Commission on Military Justice.89 More recently, a 2018 workshop conducted at Yale Law School updated the 2006 UN Draft Principles Governing the Administration of Justice Through Military Tribunals,90 and included a specific recommendation that “[a]ccess to the State’s highest court with jurisdiction over criminal cases should be available to persons convicted by military courts on an equal footing with persons convicted by civilian courts.”91 Congressional action placing military personnel on an equal footing with other criminal defendants with respect to Supreme Court access is thus long overdue. Achieving equal footing requires repealing the second sentence of Article 67a(a) and amending Article 67(a)(3). The only potential complication is the political one that corrective legislation would have to pass through both the armed services and judiciary committees—hardly an insuperable obstacle.

To be sure, there are other questions that Congress could usefully address in connection with the appellate review of courts-martial. For example, it could conclude, in an era of austerity, that CAAF’s caseload has so diminished over time that its docket should be transferred lock, stock, and barrel to the U.S. Court of Appeals for the District of Columbia Circuit.92 Barring that, it would make sense to streamline CAAF’s discretionary jurisdiction by permitting an appeal as of right by either side from all final decisions of the four Courts of Criminal Appeals.93 Under such a system, CAAF could still dispose of many cases summarily (and without oral argument), just as the Article III courts of appeals may do.94 This change would not only end its improper role as a gatekeeper—an outcome we strongly endorse—but would also remove the need for two cycles of briefing, first at the petition stage and then again after review has been granted.95

Changes like those would serve the interest of efficiency and judicial economy, but have little to do with either fairness or the demands of the Constitution. Whether or not Congress takes these other steps, it should delay no further in fixing the fundamental flaw examined in this Essay.

Eugene R. Fidell is a Senior Research Scholar, Yale Law School; Adjunct Professor, NYU Law School; of counsel, Feldesman Tucker Leifer Fidell LLP, Washington, D.C. Brenner M. Fissell is an Associate Professor of Law, Hofstra University. Philip D. Cave is President, National Institute of Military Justice; Commander, Judge Advocate General’s Corps, U.S. Navy (Ret); Partner, Cave & Freeburg LLP, Washington, D.C. The authors are grateful to Max J. Goldberg, Yale Law School Class of 2022, for research assistance, and to Bill Eskridge, Dick Fallon, Eric Freedman, Ed Hartnett, Carlos Vasquez, and Steve Vladeck for help as we thought through these issues.