Jason Bell v. J. Streeval

Opinion

PUBLISHED

UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 22-6189

JASON TYANN BELL, Petitioner – Appellant,

v.

J.C. STREEVAL, Warden of USP Lee, Respondent – Appellee.

------------------------------ PROFESSOR BRANDON HASBROUCK, Amicus Supporting Appellant.

Appeal from the United States District Court for the Western District of Virginia, at Roanoke. Elizabeth K. Dillon, Chief District Judge. (7:21-cv-00094-EKD-JCH)

Argued: September 26, 2024 Decided: August 6, 2025

Before THACKER, RICHARDSON, and BENJAMIN, Circuit Judges.

Affirmed by published opinion. Judge Richardson wrote the opinion of the Court with respect to Parts I and II(A)–(B), in which Judges Thacker and Benjamin joined. Judge Thacker wrote the opinion of the Court with respect to Part II(C), in which Judge Benjamin joined. Judge Richardson wrote an opinion concurring in the judgment with respect to Part II(C).

2

ARGUED: Dana Kagan McGinley, ARNOLD & PORTER KAYE SCHOLER LLP, Washington, D.C., for Appellant. Paul Theodore Crane, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee. ON BRIEF: Randy V. Cargill, Assistant Federal Public Defender, OFFICE OF THE FEDERAL PUBLIC DEFENDER, Roanoke, Virginia; R. Stanton Jones, Andrew T. Tutt, Washington, D.C., Kevin Cosgrove, Hafeez Khan, Hailey V. Sullivan, ARNOLD & PORTER KAYE SCHOLER LLP, San Francisco California, for Appellant. Kenneth A. Polite, Jr., Assistant Attorney General, Lisa H. Miller, Deputy Assistant Attorney General, Criminal Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; Adair Ford Boroughs, United States Attorney, Kathleen M. Stoughton, Assistant United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Columbia, South Carolina, for Appellee. Michael L. Rosenthal, Abigail P. Barnes, Joseph DuChane, Sameer Aggarwal, Austin S. Martin, COVINGTON & BURLING LLP, Washington, D.C., for Amicus Curiae.

3

RICHARDSON, Circuit Judge, writing for the Court in Parts I and II(A)–(B):

Jason Tywann Bell is asking for habeas relief under 28 U.S.C. § 2241. He argues that because his sentence was imposed under an unconstitutionally vague Guideline provision, he is entitled to resentencing. But to even advance this theory, he must satisfy the procedural requirements that Congress has imposed on when courts may grant postconviction relief. He does not.

Bell was convicted of, and sentenced for, a federal crime. So rather than seek habeas under § 2241, he was obligated to seek relief under § 2255. If he had done that, his motion would have been denied because this is not his first § 2255 motion and § 2255(h) permits second or successive motions only in limited circumstances that he does not satisfy.

Bell therefore argues that because § 2255 relief is unavailable, he should be permitted to seek habeas via § 2241 under the dictates of § 2255(e), the so-called “saving clause.” The saving clause permits someone—who must otherwise file under § 2255—to file under § 2241 instead if § 2255 is “inadequate or ineffective to test the legality of his detention.” But Bell’s failure to satisfy § 2255(h)’s limits on successive motions does not by itself make § 2255 inadequate or ineffective. Jones v. Hendrix, 599 U.S. 465, 482 (2023). So we lack jurisdiction over Bell’s § 2241 petition.

Bell also argues that if we interpret § 2255 to bar his requested habeas relief, then our interpretation will mean Congress unlawfully suspended the writ of habeas corpus when it enacted § 2255. See U.S. Const., Art. I, § 9, cl. 2. But precedent forecloses this

4

argument too. 1 We therefore affirm the district court’s order dismissing Bell’s petition for want of jurisdiction. I. Background A. Conviction And Postconviction Proceedings Bell pleaded guilty to two federal offenses: (1) attempted bank robbery, 18 U.S.C.

§ 2113(a), and (2) carrying and using a firearm during and in relation to a crime of violence, 18 U.S.C. § 924(c). When sentenced in October 2003, Bell faced mandatory Sentencing Guidelines since the Supreme Court had not yet made the Guidelines advisory in United States v. Booker. 543 U.S. 220, 245 (2005). Under those mandatory Guidelines, Bell qualified as a career offender based on his prior convictions for two “crime[s] of violence.” U.S.S.G § 4B1.1. This designation set his mandatory guidelines range at 262 to 327 months. Bell was sentenced to 274 months, plus a term of supervised release.

Bell has since moved several times for postconviction relief. His first § 2255 motion in 2004 raised a sentencing error and was denied. After a ten-year hiatus, Bell began filing motions again. The content of these motions is largely irrelevant—what does matter is that this is not Bell’s first postconviction motion.

B. Intervening Changes In Law—Booker, Johnson, Beckles, And Brown Bell now seeks habeas relief on the theory that the Guidelines’ career-offender provision, which mandatorily enhanced his sentence, is unconstitutionally vague. If he is

1

While this panel agrees that precedent forecloses Bell’s Suspension Clause claim, we disagree about which precedent does so. Compare § II(C), infra, with Concurring Op. at 42–43 (Richardson, J., concurring in the judgment).

5

correct on his petition’s merits, he was sentenced above his guidelines range by nearly 100 months. Understanding Bell’s merits theory—and why he has chosen Section 2241 as his procedural vehicle as opposed to Section 2255—requires walking through fifteen years of intervening Supreme Court and Fourth Circuit precedent.

To start, after Bell was sentenced in 2003, the Supreme Court made the Guidelines advisory rather than mandatory. The Court held in 2005 that judges could not constitutionally enhance a person’s sentence under the Guidelines by relying on facts not found by juries. Booker, 543 U.S. at 226–27. To remedy that constitutional problem, the Court severed and invalidated the portion of the Guidelines that made them mandatory. The resulting advisory guidelines did not raise that constitutional concern. Id. at 246.

A decade later, the Supreme Court limited the ways to identify career criminals based on their convictions. Under the Armed Career Criminal Act, a defendant is subject to more severe punishment for being a felon in possession of a firearm if he has three or more previous convictions for “violent felon[ies].” 18 U.S.C. § 924(e). Congress defined “violent felony” in the Act’s so-called “residual clause” to include any felony that “involves conduct that presents a serious potential risk of physical injury to another.” Id. § 924(e)(2)(B). But in 2015, the Supreme Court held that the residual clause was unconstitutionally vague because it provided no guidance on how “one go[es] about deciding” when the clause is satisfied. United States v. Johnson. 576 U.S. 591, 593 (2015).

The Supreme Court then combined Booker and Johnson in Beckles v. United States, 580 U.S. 256 (2017). Beckles addressed the career-offender provision in an older version of the Guidelines that enhanced criminal penalties for defendants with two prior felony

6

convictions for “crime[s] of violence.” Id. at 258–59 (citing U.S.S.G. § 4B1.2(a)(2) (Nov. 2006)). The Guidelines’ “crime of violence” definition, in its own residual clause, was “identically worded” to the unconstitutionally vague residual clause in Johnson; despite that, the Court upheld the Guidelines’ residual clause as constitutional because—unlike the Armed Career Criminal Act—the Guidelines are merely advisory. Id. at 267. While the Due Process Clause requires the public to have notice of substantive crimes and mandatory sentences, it does not require the same of discretionary sentencing guidance for judges. Id. at 262–63.

Beckles’s logic depended on the Guidelines’ advisory (i.e., non-mandatory) nature.

That left open a question—what about the people, like Bell, who were sentenced before Booker and thus under the mandatory Guidelines’ residual clause? There were two possible answers.

On the one hand, perhaps the Beckles court simply reserved the question whether the residual clause in § 4B1.2(a)(2) was unconstitutionally vague when the Guidelines were mandatory. This was Justice Sotomayor’s view. See Beckles, 580 U.S. at 281 n.4 (Sotomayor, J., concurring in the judgment). On the other hand, perhaps Beckles’s reasoning necessarily compelled the conclusion that a mandatory Guidelines residual clause would be unconstitutionally vague.

The difference between these two views matters for people like Bell who have previously filed a § 2255 motion. Second or successive § 2255 motions are ordinarily prohibited. But under one of two exceptions laid out in § 2255(h), such a motion is permitted when the Supreme Court establishes a “new rule of constitutional law, made

7

retroactive to cases on collateral review . . . that was previously unavailable.” See Dodd v. United States, 545 U.S. 353, 359 (2005) (quoting § 2255(h)); see also In re Graham, 61 F.4th 433, 442–43 (4th Cir. 2023). If the latter view—that Beckles compelled the conclusion that a mandatory Guidelines residual clause is unconstitutionally vague—is correct, then Beckles would have established a new rule of constitutional law, potentially opening the door for second or successive § 2255 motions. But this Court has rejected the latter interpretation, instead explicitly adopting Justice Sotomayor’s view. See United States v. Brown, 868 F.3d 297, 299 n.1 (4th Cir. 2017). Because the Supreme Court broke no new constitutional ground in Beckles, Bell cannot bring a second or successive § 2255 motion on his theory that the Guidelines’ residual clause is unconstitutionally vague. Even if Bell is right, 2 he cannot challenge his sentence by filing another § 2255 motion unless and until the Supreme Court creates a new retroactive constitutional rule. § 2255(h).

Recognizing he is procedurally barred by § 2255(h), Bell advanced that same merits theory before the district court via a different procedural vehicle—§ 2241. The district court determined that Bell could not proceed under § 2241 and dismissed his habeas petition. Bell then appealed. During the pendency of his appeal, he finished serving his prison sentence and was released.

2

Several sister circuits have agreed with Bell that a mandatory-Guidelines residual clause is unconstitutionally vague under Johnson. See, e.g., Shea v. United States, 976 F.3d 63, 81–82 (1st Cir. 2020); Cross v. United States, 892 F.3d 288, 291, 294 (7th Cir. 2018). Those cases involved federal prisoners bringing their first § 2255 motion, however, and therefore did not involve the second-or-successive procedural hurdle Bell faces.

8

II. Discussion Bell’s appeal presents three questions. First, as a threshold matter, has Bell’s release from prison stripped us of jurisdiction over this case? Second, was Bell entitled to file under § 2241, instead of § 2255, due to the arguable unconstitutionality of his sentence? And third, would denying Bell access to relief under § 2241 violate the Constitution’s Suspension Clause?

The answer to all three questions is no. We possessed jurisdiction at the time Bell filed his habeas petition and still do. But Supreme Court precedent forecloses his ability to seek habeas relief under § 2241. And denying him the § 2241 procedural vehicle in this case does not violate the Suspension Clause. Accordingly, we affirm the district court and dismiss Bell’s petition for want of jurisdiction. 3 A. We Have Jurisdiction Over Bell’s Appeal We begin with the threshold question. While this appeal was pending, Bell was released from prison. This raises two potential jurisdiction issues.

First, because § 2241 is limited to people “in custody” by its plain text, “[a]

reasonable reader might think we lack jurisdiction over a habeas petition when the petitioner” is “released from custody.” Plymail v. Mirandy, 8 F.4th 308, 314 (4th Cir. 2021). But precedent says that reader would be wrong. It’s true courts lack the “jurisdiction to entertain [an] application” for habeas under § 2241(c) if the application was

We review the availability of the saving clause and whether § 2255 is inadequate

3

de novo. Farkas v. Butner, 972 F.3d 548, 553 (4th Cir. 2020). We also review the applicability of the Suspension Clause de novo. Cf. United States v. Schnittker, 807 F.3d 77, 81 (4th Cir. 2015).

9

not made by a person “in custody.” But this jurisdictional check for custody is performed “at the time an action is filed,” not later. Plymail, 8 F.4th at 314. That Bell has been released from prison does not change the fact that he filed his § 2241 application for habeas while he was in prison—unquestionably while he was “in custody.” Thus, his petition is not barred by the requirement that a habeas petitioner be “in custody” under § 2241(c).

Second, “a logical assumption is that the release of a prisoner from [prison] renders a habeas petition moot,” stripping us of jurisdiction under Article III’s case-or-controversy requirement. Plymail, 8 F.4th at 315. But this assumption, too, is wrong. “A case becomes moot only when it is impossible for a court to grant any effectual relief whatever to the prevailing party.” Knox v. Serv. Emps. Int'l Union, Loc. 1000, 567 U.S. 298, 307 (2012) (cleaned up). And we can still provide effectual relief to Bell.

Bell is currently serving three years of supervised release. If Bell is right that he was unconstitutionally sentenced to 100 months more than he should have been—that is, if he were to prevail—then a decision from this Court would lead to his resentencing. And at the resentencing, the court could take his excess prison time into account and adjust the length or conditions of his term of supervised release. Cf. United States v. Johnson, 529 U.S. 53, 60 (2000) (permitting a trial court to “modify” or “terminate” an individual’s term of supervised release under 18 U.S.C. § 3582(e)(2) in the event an individual is incarcerated for too long). Therefore, because it is not “impossible” for us to “grant

10

effectual relief,” Knox, 567 U.S. at 307, his claim is not moot. 4 We turn now to the next question: whether he can use the § 2241 procedural vehicle.

B. Bell’s Habeas Petition Is Not Cognizable Under § 2241 The district court determined below that Bell could not bring a habeas petition under § 2241 because he failed to satisfy the § 2255(e) saving clause. On appeal, Bell asks us to reverse. We cannot do so. A federal prisoner may not seek postconviction relief through § 2241 unless § 2255 is “inadequate or ineffective” to test his detention’s legality. § 2255(e). Bell insists he meets that standard because he cannot satisfy the standard to file a second or successive § 2255 motion. But the Supreme Court has rejected that argument. Under Jones v. Hendrix, the mere failure to satisfy § 2255(h)’s procedural limitations does not make § 2255 inadequate or ineffective. He therefore cannot avail himself of § 2241. 5

4

In some cases, a petitioner’s case is not mooted even after being “unconditionally released” from custody—that is, he may continue to challenge the legality of his conviction even after any term of supervised release. The Supreme Court has explained that a criminal conviction can bring collateral consequences that linger beyond custody, such as the loss of voting rights and the privilege of serving on a jury. Carafas v. LaVallee, 391 U.S. 234, 237–38 (1968). And the Supreme Court has further instructed us to “presume[] that collateral consequences exist.” Plymail, 8 F.4th at 315 (citing Sibron v. New York, 392 U.S. 40, 55 (1968)). So the “mere ‘possibility of consequences collateral to the imposition of sentence is sufficiently substantial to justify our dealing with the merits.’” Id. (quoting Pollard v. United States, 352 U.S. 354, 358 (1968)). The Carafas line of cases, however, does not apply here. Bell is challenging the length of his sentence, not the validity of his underlying conviction. And so even if he wins, any potential collateral consequences flowing from his conviction will remain the same.

5

When Bell originally filed his § 2241 petition in the district court, the Supreme Court had not yet decided Hendrix. Consequently, the district court applied this circuit’s pre-Hendrix saving clause jurisprudence to Bell’s petition. But our prior jurisprudence is no longer good law after Hendrix, and we do not apply it on appeal. We explain this conclusion in more depth below.

11

1. The structure of the various “habeas” rules.

To understand why Bell cannot use § 2241 requires a background on how various federal postconviction statutes interact with one another.

“Habeas” colloquially refers to a whole family of common-law writs, pre-American Revolution English statutes, federal statutes, and state statutes. The most important common-law habeas writ is habeas corpus ad subjiciendum (roughly translated, “you shall have the body to submit”). This is “a writ antecedent to statute” that has “its root deep into the genius of our common law.” Farkas, 972 F.3d at 553 (quoting Rasul v. Bush, 542 U.S. 466, 473 (2004)). The writ of habeas corpus ad subjiciendum allows a person to challenge “the causes[] and extent” of their detention, thus permitting “a court to examine the grounds” for the prisoner’s confinement. Id. (citing 3 William Blackstone, Commentaries on the Laws of England 133 (1st ed. 1768)). For its service in protecting liberty, it is famously called “the Great Writ.” Id.; Stone v. Powell, 428 U.S. 465, 474 n.6 (1976) (citation omitted). 6 As its importance waxed in England, the writ was codified in various forms by statutes, Farkas, 972 F.3d at 553, though the common-law forms remained too,

6

Though the habeas corpus ad subjiciendum earned the lofty title, this is a relatively recent development—and historians are right to step carefully around anachronism. Various forms of habeas corpus appeared in England as early as the Middle Ages. Often, different forms of the writ were used by different courts, but used similar language, served similar functions, and influenced each other’s development. Alongside the ad subjiciendum was another, almost equally important form of the writ: the corpus cum causa. The writs’ functions were similar, and both evolved in parallel until the ad subjiciendum became dominant. Paul D. Halliday, Habeas Corpus: From England to Empire 17–18, 61–62 (2010).

12

see Paul D. Halliday & G. Edward White, The Suspension Clause: English Text, Imperial Contexts, and American Implications, 94 Va. L. Rev. 575, 631 (2008).

The developed Great Writ came to “receive[] explicit recognition in the Constitution” via the Suspension Clause. Farkas, 972 F.3d at 553 (quoting Rasul, 542 U.S. at 474); see also Brandon L. Garrett & Lee Kovarsky, Federal Habeas Corpus: Executive Detention and Post-Conviction Litigation 22 (2d ed. 2024). We discuss the contours of the Suspension Clause in depth later, but in general, “judges in all camps tend to view the Suspension Clause’s reference to the ‘privilege of the writ of habeas corpus’ as a textual reference to the writ as it existed in England, before the American revolution.” Garrett & Kovarsky, supra, at 15.

Courts and litigants also use the word “habeas” to describe the federal statutes that govern postconviction review by federal courts. 7 See 28 U.S.C. §§ 2241, 2243, 2244, 2253, 2254, and 2255. Section 2241 “is considered the general grant of habeas authority.” 8 Garrett & Kovarsky, supra, at 45. Sections 2244, 2253, and 2254 govern federal court review of state custody, providing jurisdiction to do so while simultaneously imposing procedural limitations. Review of federal custody, meanwhile, is largely governed by § 2255. While § 2255 is an independent statutory privilege and not technically “habeas”

7

Postconviction review is also often referred to as “collateral review,” as it is outside the original proceeding.

8

The first federal habeas statute was § 14 of the 1789 Judiciary Act. 1 Stat. 73, § 14. This grant of authority was split in two. See Garrett & Kovarsky, supra, at 45. The statutory authority to issue the Great Writ—writs of habeas corpus ad subjiciendum—was codified in § 2241. Id. There are also other habeas writs, none of which are relevant here. The statutory authority to issue them is the All Writs Act, 28 U.S.C. § 1651(a). Id.

13

in the sense of the common-law writ inherited by the Colonies, it governs almost all federal postconviction review.

In 1996, Congress worked a sea change in federal statutory postconviction law. See Antiterrorism and Effective Death Penalty Act of 1996, Pub. L. No. 104-132, 110 Stat. 1214 (“AEDPA”). AEDPA is “Congress’ latest pronouncement on the writ of habeas corpus.” Farkas, 972 F.3d at 554. Among other things, AEDPA changed and recodified several habeas statutes, sharply limiting the scope of postconviction relief in federal court. See, e.g., Taylor A.R. Meehan, Postconviction Remedies, Retroactivity, and Montgomery v. Louisiana’s Other New Rule, 88 Mo. L. Rev. 1077, 1100 (2024); Brandon L. Garrett & Kaitlin Phillips, AEDPA Repeal, 107 Cornell L. Rev. 1739, 1756 (2022).

With this background, we can turn to Bell’s claim.

2. Bell cannot use the saving clause as interpreted by Hendrix.

As noted above, most federal detainees are required to pursue postconviction relief under § 2255—not § 2241. Section 2255 is not an optional alternative to § 2241—“the general rule [is] that convicted federal prisoners must proceed under § 2255.” Farkas, 972 F.3d at 555 (emphasis added). There is, however, one exception to the general rule: the so-called saving clause. Id. (citing § 2255(e)). The saving clause provides that if § 2255 is “inadequate or ineffective” for a person “to test the legality of his detention,” he may proceed directly under § 2241. § 2255(e). After much discussion in the courts of appeals,

14

the Supreme Court recently explained when § 2255(e) is satisfied. See Hendrix, 599 U.S. at 474–75.

In Hendrix, the Supreme Court held that the petitioner could not bring a second or successive 9 motion under § 2255(h) because his claim of innocence was based on a new statutory ruling, rather than a new constitutional ruling. The plain text of § 2255(h) explains this result: Second or successive motions based on changes in law must allege “a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court, that was previously unavailable.” § 2255(h) (emphasis added). 10 Claims based on changes in statutory law will not do. Hendrix, 599 U.S. at 476–77. Section 2255(h) describes “two—and only two—conditions in which a second or

9

What counts as a “second or successive” motion is deceptively self-explanatory.

Often it just means any § 2255 motion that comes after a person has filed their first. Makes sense, as second comes after first. But not always—“the phrase second or successive” “is a term of art” that “does not necessarily refer to all habeas filings made second or successively in time.” Rivers v. Guerrero, 145 S. Ct. 1634, 1643 (2025) (quotations omitted). Some types of literally successive motions have been deemed not “second or successive” in the relevant sense. See, e.g., Slack v. McDaniel, 529 U.S. 473, 478 (2000) (motions filed after an initial motion was dismissed without adjudication on the merits for failure to exhaust state remedies); Panetti v. Quarterman, 551 U.S. 930, 947 (2007) (claims brought in successive motions under Ford v. Wainwright, 477 U.S. 399 (1986), when those claims become ripe).

10

Second or successive § 2255 motions are also permissible if the movant presents “newly discovered evidence that, if proven and viewed in light of the evidence as a whole, would be sufficient to establish by clear and convincing evidence that no reasonable factfinder would have found the movant guilty of the offense.” § 2255(h)(1). But this provision is irrelevant to Bell, both because he does not present new evidence and because he is challenging his sentence, not his conviction.

15

successive § 2255 motion may proceed,” and a change in statutory law is not one of them. Id. at 477. 11 Because § 2255 was procedurally unavailable, Jones argued that he should be able to use § 2255’s exception, the saving clause, to route his claim through § 2241. In essence, he argued that if a legally innocent person cannot test the legality of his detention under § 2255 because of § 2255(h), then § 2255 is “inadequate or ineffective.” While not without intuitive appeal, the argument was rejected by the Supreme Court.

In addressing Jones’ claim, the Supreme Court explained § 2255(e)’s scope. The Court first explained that the “sole purpose” of § 2255 is to provide administrative convenience. Id. at 473. A federal prisoner can be sentenced in one judicial district but imprisoned in another. In such cases, the former district court would contain the records relevant to the prisoner’s habeas petition, yet the prisoner would have to file his § 2241 habeas petition against the prison warden in the latter court. Id. at 474. The geographic split from this traditional method of habeas filing resulted in “serious administrative

11

Setting aside those procedural limits in § 2255(h), the underlying claim in Hendrix was plainly meritorious. The petitioner was convicted under § 922(g)(1) before Rehaif v. United States, 588 U.S. 225 (2019). Rehaif held that a defendant’s knowledge of his status that disqualifies him from owning a firearm is an element of a § 922(g) conviction. See Hendrix, 599 U.S. at 470. Rehaif overruled the Eighth Circuit precedent that Jones’s convicting court had applied and that the Eighth Circuit had applied on direct appeal. Id. So the prosecution in Jones’s case had not proven an essential element of his 922(g) count, making him legally innocent. But there is no legal innocence exception to § 2255(h)’s strictures: “Congress has chosen finality over error correction.” Id. at 480.

16

problems.” Id. at 473–74 (quoting United States v. Hayman, 342 U.S. 205, 212 (1952)). 12 To remedy these problems, Congress enacted § 2255, which requires a prisoner to file in “the court which imposed [his] sentence,” § 2255(a).

The saving clause’s narrow scope reflects its rationale. Resort to § 2241 via the saving clause and filing in the judicial district of imprisonment is permitted only in “the unusual circumstances in which it is impossible or impracticable for a prisoner to seek relief from the sentencing court.” Hendrix, 599 U.S. at 474–75. And impossibility and impracticability are not defined with reference to the prisoner’s ability to succeed; rather, they are defined with reference to the prisoner’s ability to seek relief from the sentencing court. Id. 13 Nor could it reading AEDPA as a cohesive whole. Congress in AEDPA established clear procedural restrictions on postconviction relief by motion. And as the Supreme Court emphasized repeatedly in Hendrix, any reading of § 2255(e) that shunts procedurally barred claims into § 2241 would turn the saving clause into an “end-run around” those procedural restrictions, making AEDPA “curiously self-defeating.” Id. at 477, 479. The Court thus forbade us from making AEDPA “internally inconsistent” by

12

In addition to the evidentiary problems this posed, this concentrated habeas petitions in the handful of judicial districts areas where prisons were located, forcing those few district courts to process “an inordinate number of habeas corpus actions.” Hayman, 342 U.S. at 213–14.

13

For example, a prisoner would lack access to his sentencing court if the sentencing court were dissolved (i.e., it literally “no longer exists.”). Hendrix, 599 U.S. at 474 (citing Witham v. United States, 355 F.3d 501, 504–05 (6th Cir. 2004)).

17

applying the saving clause to claims that may be meritorious but are procedurally barred. Id. at 479 (quotation omitted).

Under Hendrix, then, Bell’s case is straightforward—it is essentially the same as Jones’s was in Hendrix itself. Bell’s petition, like Jones’s petition, would have been procedurally barred had it been brought as a second or successive § 2255 motion because it does not fall into either of the two exceptions laid out in § 2255(h). So Bell, like Jones, brought his claim through § 2241, arguing that he could do so because of the saving clause. But it remains possible and practicable for Bell to file his § 2255 motion with his sentencing court. Hendrix, 599 U.S. at 474. So Bell, like Jones, cannot use the saving clause. That Bell, like Jones, may have a meritorious claim of legal innocence stymied by § 2255(h)’s procedural restrictions does not make § 2255 “inadequate or ineffective.” “The inability of a prisoner . . . to satisfy [§ 2255(h)] does not mean that he can bring his claim in a habeas petition under the saving clause. It means that he cannot bring it at all.” Id. at 480; see also Slusser v. Vereen, 36 F.4th 590, 592 (4th Cir. 2022).

Unable to prevail under Hendrix, Bell attempts to escape its reasoning in two ways.

First, he attempts to distinguish the case by arguing that Hendrix involved a change in statutory law, while his involves a change in constitutional law. That is indeed a distinction—but one without a difference. Bell’s problem is that, like Jones, he fails to satisfy § 2255(h). To be sure, Jones failed because § 2255(h) makes no exception for new statutory law, while Bell fails because § 2255(h)(2)’s exception for new constitutional law does not apply unless the Supreme Court has said that there is new (and retroactive)

18

constitutional law. But the reason someone fails to satisfy § 2255(h) does not change the saving clause’s availability. The result for both is the same.

In fact, our answer is clearer than the answer was in Hendrix itself. The Supreme Court interpreted § 2255(h)’s enumerated list of two exceptions to give rise to the “negative inference” that Congress intended to provide only those two exceptions, foreclosing the possibility of an implied third exception for new statutory claims smuggled into the saving clause. Id. at 477. That negative inference is even stronger for Bell because one of those two enumerated exceptions already addresses changes in constitutional law and “speaks exactingly to the circumstances that permit a constitutional claim.” Farkas, 972 F.3d at 559. 14 We will not assume that Congress meant to implicitly permit kinds of constitutional claims when Congress has addressed other kinds explicitly. 15

14

There was some suggestion that the several related but not dispositive changes in constitutional law that followed Bell’s sentencing are “unusual” and therefore Bell satisfies § 2255(e)’s strictures. Cf. Hendrix, 599 U.S. at 478. But as we have explained, “unusual” refers to the inability to seek relief, not the inability to obtain it. Moreover, there is nothing “unusual” about denying relief to habeas petitioners who would have benefitted from changes in law that happen after they are sentenced. See Teague v. Lane, 489 U.S. 288 (1989) (judge made limits on when retroactive changes in law apply on collateral review); Greene v. Fisher, 565 U.S. 34, 38 (recognizing that collateral relief for state prisoners under § 2254(d) is limited to mistakes based on the law that existed at the time of the state’s last adjudication on the merits, rather than law made after).

15

Bell could have raised his theory in his first § 2255 motion. Had he done so, § 2255(h)’s bar on certain constitutional claims would not have applied. Whether fair or not, Congress requires a prisoner to raise available theories in the first instance and within a set amount of time. See §§ 2255(a), (f). If he chooses not to, his options moving forward will be more limited. But in any event, § 2255 was available and adequate to test his current theory in his first petition, had he chosen to argue it.

19

Bell’s second contention is that a decision against him would render the saving clause vestigial. If Hendrix forecloses all statutory claims, and we foreclose constitutional claims (that do not fit into §§ 2255(a), (h)(2)), the saving clause would never apply—or so the argument goes. It is true that, were the saving clause to be rendered “mere surplusage,” that would offend our presumption “that the legislature intended each portion [of a statute] to be given full effect.” Navy Fed. Credit Union v. LTD Fin. Servs., LP, 972 F.3d 344, 359 (4th Cir. 2020) (quoting Fontenot v. Taser Int’l, Inc., 736 F.3d 318, 327 (4th Cir. 2013)).

But this second argument simply misunderstands Hendrix. Far from foreclosing statutory claims, Hendrix makes clear that both constitutional and statutory claims can be routed through the saving clause into § 2241. Hendrix just clarifies that this happens only in the “unusual circumstances in which it is impossible or impracticable for a prisoner to seek relief from the sentencing court.” 599 U.S. at 474–75. Bell’s argument thus proceeds on a false premise.

So Bell’s claim under the saving clause is foreclosed by Hendrix for the same reasons given in Hendrix itself.

3. Hendrix abrogated this Court’s contrary precedent.

That was all the law that applies today. Bell, however, appeals a decision based on pre-Hendrix decisions from this Court that established when petitioners could bring sentencing challenges under § 2241 via the saving clause. See Bell v. Streeval, No. 7:21- CV-00094, 2022 WL 329229, at *6–7 (W.D. Va. Feb. 3, 2022). We do not apply that law. Instead, we resolve Bell’s appeal without applying those cases because we recognize that, after Hendrix, they are no longer good law. See Henderson v. United States, 568 U.S. 266,

20

271 (2013) (“The general rule is that an appellate court must apply the law in effect at the time it renders its decision.” (cleaned up)).

Before Hendrix, a trilogy of Fourth Circuit cases interpreted the saving clause. First came In re Jones, 226 F.3d 328, 333–34 (4th Cir. 2000), overruled by Hendrix, 599 U.S. at 477. In re Jones had essentially identical facts to Hendrix: A prisoner was barred from bringing a second or successive motion under § 2255(h) for a change in statutory law and so used the saving clause to bring a habeas petition under § 2241 instead. To evaluate his petition, this Court created a three-part test to determine when the saving clause was available. Id. That test reasoned that “§ 2255 is inadequate and ineffective to test the legality of a prisoner’s conviction when,” among other requirements, “the prisoner cannot satisfy the gatekeeping provisions of § 2255 because the new rule is not one of constitutional law.” 16 Id (emphasis added).

Second came United States v. Wheeler, 886 F.3d 415 (4th Cir. 2018). Extending In re Jones from conviction challenges to sentencing challenges, this Court in Wheeler laid out a similar multipart test that allowed a prisoner to bring a § 2241 habeas petition when,

16

More specifically, In re Jones laid out a three-part test: “Accordingly, we conclude that § 2255 is inadequate and ineffective to test the legality of a conviction when: (1) at the time of conviction, settled law of this circuit or the Supreme Court established the legality of the conviction; (2) subsequent to the prisoner's direct appeal and first § 2255 motion, the substantive law changed such that the conduct of which the prisoner was convicted is deemed not to be criminal; and (3) the prisoner cannot satisfy the gatekeeping provisions of § 2255 because the new rule is not one of constitutional law.” 226 F.3d at 333–34.

21

among other requirements, “the prisoner is unable to meet the gatekeeping provisions of § 2255(h)(2) for second or successive motions.” 17 Id.

The last case in this trilogy is Young v. Antonelli, 982 F.3d 914 (4th Cir. 2020).

Young brought a Section 2241 habeas petition arguing—like Bell in this case—that a new Supreme Court case interpreting language in 21 U.S.C. § 841 applied to equivalent language in a Guidelines sentencing enhancement. Id. at 915–16 (citing Burrage v. United States, 571 U.S. 204, 206 (2014)). But Young was a federal detainee who had previously filed a § 2255 motion and thus was barred from using § 2241 unless he satisfied § 2255(e). We acknowledged that Burrage was a change in statutory law that the Supreme Court had not yet applied to the Sentencing Guidelines, much less holding them to be retroactive. This meant if Young had filed a § 2255 motion, it would have been barred by § 2255(h). Id. at 918–19. Nevertheless we, as an inferior court, found that Burrage applied to the Sentencing Guidelines; and we, as an inferior court, held them to be retroactive. Id. at 919– 20. Proceeding on the premise that those were permissible moves, we found that Young satisfied the test we had laid out in Wheeler and so could use the saving clause to bring a § 2241 habeas petition to challenge his Guidelines enhancement. Id.

17

Again, more specifically, Wheeler laid out a four-part test: “[W]e conclude that § 2255 is inadequate and ineffective to test the legality of a sentence when: (1) at the time of sentencing, settled law of this circuit or the Supreme Court established the legality of the sentence; (2) subsequent to the prisoner’s direct appeal and first § 2255 motion, the aforementioned settled substantive law changed and was deemed to apply retroactively on collateral review; (3) the prisoner is unable to meet the gatekeeping provisions of § 2255(h)(2) for second or successive motions; and (4) due to this retroactive change, the sentence now presents an error sufficiently grave to be deemed a fundamental defect.” 886 F.3d at 429 (citing In re Jones, 226 F.3d at 333–34).

22

This trilogy of cases forms a tower, with Young built on Wheeler and Wheeler built on In re Jones. But in Hendrix the Supreme Court explicitly overruled In re Jones by name. Hendrix, 599 U.S. at 477. And when the Supreme Court demolished that foundation, the Wheeler and Young floors crumbled with it. 18 Hendrix explained that § 2255 is only inadequate or ineffective when access to the sentencing court is “impossible or impracticable.” See Hendrix, 599 U.S. at 477–78. The Hendrix test for when the saving clause is available, then, is unrelated to whether the prisoner is procedurally barred by § 2255—and so supplants the test created in In re Jones, expanded in Wheeler, and expanded again in Young.

As Hendrix explained, and as we reiterated above, when Congress enumerated “two—and only two—conditions in which a second or successive § 2255 motion may proceed,” it intended to bar all other types of second or successive motions. Id. at 477. That means that second or successive motions premised on changes in this Court’s precedent cannot proceed through the saving clause, contrary to our pronouncements in the trilogy. See In re Jones, 226 F.3d at 333–34 (requiring changes in the “settled law of this circuit or the Supreme Court” (emphasis added)); Wheeler, 886 F.3d at 429 (same); Young, 982 F.3d at 918 (same). Only second or successive motions premised on “a new rule of constitutional law, made retroactive to cases on collateral review by the Supreme Court”

Other courts have recognized this. See, e.g., Scheetz v. Ciolli, No. 22-1300, 2023

18

WL 6366049, at *3 (10th Cir. Sept. 29, 2023); Barrie v. United States, No. 1:22-CV-203, 2023 WL 5044977, at *3 (E.D. Va. Aug. 8, 2023); Johnson v. Dunbar, No. 5:22-3677, 2023 WL 5211052, at *3 (D.S.C. July 12, 2023); Murphy v. United States, No. 1:21-00053, 2023 WL 9103097, at *5 (S.D.W. Va. Sept. 13, 2023).

23

will do. § 2255(h)(2) (emphasis added). So although Hendrix does not mention Wheeler or Young by name, the reasoning in those cases has been fundamentally rejected. 19 Along with Jones, they are no longer good law. See Rose v. PSA Airlines, Inc., 80 F.4th 488, 504 (4th Cir. 2023) (“Where prior decisions in our Circuit use reasoning inconsistent with Supreme Court authority, we are not bound to follow them.”) (internal quotations omitted).

At bottom, Hendrix controls this case, and Bell can find no recourse in our pre-

Hendrix precedent.

C. Bell’s Inability To Obtain Relief Does Not Violate The Suspension Clause

THACKER, Circuit Judge, writing for the Court in Part II(C):

Last, we consider Bell’s argument that denying him access to relief under § 2241 would violate the Constitution’s Suspension Clause. This argument is resolved by our decision In re Vial, 115 F.3d 1192 (4th Cir. 1997) (en banc), which held that “the limitations imposed on second and successive § 2255 motions by [§ 2255(h)] do not constitute a suspension of the writ.” Id. at 1198.

As discussed supra, § 2241 relief is unavailable to Bell because the § 2255(e)

“saving clause does not authorize [] an end-run around [§ 2255(h)].” Jones v. Hendrix, 599 U.S. 465, 477 (2023). This means that Bell’s petition for habeas relief is subject to and defaulted by “the limitations imposed on second and successive § 2255 motions by [§ 2255(h)].” In re Vial, 115 F.3d at 1198. As we have recognized, this barrier to relief

19

To be clear, the merits question about Burrage addressed in Young was not presented in this case. This Court takes no position on that part of Young’s analysis.

24

does not implicate the Suspension Clause. Id. Indeed, as Appellee pointed out at oral argument, we “need go no further than that” to dispose of Bell’s Suspension Clause argument. Oral Argument at 15:35–15:49, Bell v. Streeval, No. 22-6189 (4th Cir. Sept. 26, 2024), https://www.ca4.uscourts.gov/OAarchive/mp3/22-6189-20240926.mp3 (“This court in In re Vial . . . held that the limitations in 2255(h) do not violate the Suspension Clause. So under that holding, this panel need go no further than that.”).

Our concurring colleague opts for a different approach. According to the concurrence, in Hendrix the Supreme Court established a bright line rule that “the Suspension Clause only protects claims that would have been cognizable in habeas at our country’s Founding.” Post at 29 (discussing Hendrix, 599 U.S. at 482–83 (2023) (dismissing petitioner’s constitutional avoidance Suspension Clause argument on the ground that “it would extend the writ of habeas corpus far beyond its scope when the Constitution was drafted and ratified” (internal quotations omitted))). And, according to the concurrence’s historical analysis, Founding era habeas relief was confined to claims alleging a jurisdictional defect in a prisoner’s underlying judgment. Id. at 30 (“[A] prisoner at the Founding had no privilege vindicable by habeas to be freed from custody imposed as a criminal sentence unless some jurisdictional defect afflicted his judgment.”). Thus, according to the concurrence, since Bell’s constitutional claim in his § 2241 petition does not allege a jurisdictional defect in his underlying judgment, it is not the type of claim that is within the scope of the protection enshrined by the Suspension Clause. The concurrence posits, therefore, that denying Bell relief through recourse to the saving clause does not amount to a suspension of the writ.

25

We do not read Hendrix to establish such a bright line rule on the scope of the protection enshrined by the Suspension Clause. In Hendrix, the Court considered the Suspension Clause in the context of the petitioner’s constitutional doubt canon argument, which asserted that denying the petitioner the chance to raise his Rehaif * claim in a § 2241 petition would “raise[] serious constitutional questions.” Hendrix, 599 U.S. at 482. The petitioner in Hendrix argued, notwithstanding the merits of the lower circuit’s rule that “the Suspension Clause refers to [the] specific legal instrument that existed [in 1789][,]” the petitioner’s incarceration still “pose[d] constitutional concerns.” Brief for Petitioner at 36, Jones v. Hendrix, 599 U.S. 465 (2023) (No. 21-857). From this footing, the Hendrix Court held that the petitioner’s Suspension Clause argument failed because it “would extend the writ of habeas corpus far beyond its scope ‘when the Constitution was drafted and ratified.’” Hendrix, 599 U.S. at 482–83 (quoting Dep’t of Homeland Sec. v. Thuraissigiam, 591 U.S. 103, 107 (2020) (quoting Boumediene v. Bush, 553 U.S. 723, 746 (2008))).

Facially, this assertion in Hendrix would appear to support the proposition espoused by the concurrence that the Suspension Clause is confined to claims cognizable in habeas at the Founding. Post at 29 (“[T]he Suspension Clause only protects claims that would have been cognizable in habeas at our country’s Founding.”). The authorities that the Hendrix Court relied on in support of its own assertion, however, undermine the

*

Rehaif v. United States, 588 U.S. 225, 237 (2019) (holding that a defendant’s knowledge “that he knew he belonged to the relevant category of persons barred from possessing a firearm” is an element the Government must prove in a prosecution for violation of 18 U.S.C. § 922(g)).

26

concurrence’s interpretation of it. Specifically, in Thuraissigiam, the Court relied on a concession from the parties to avoid answering “whether the scope of the writ as it existed in 1789 defines the boundary of the constitutional protection [of the Suspension Clause].” Thuraissigiam, 591 U.S. at 116 n.12. The Court consequently analyzed whether the Suspension Clause was limited to the habeas right as it existed at the Founding, without deciding whether that represented the full scope of the right protected by the Suspension Clause. Likewise, Hendrix relied on Boumediene, which expressly cautioned, “at the absolute minimum the [Suspension] Clause protects the writ as it existed when the Constitution was drafted and ratified.” Boumediene, 553 U.S. at 746 (citation and internal quotation marks omitted).

We do not take the Hendrix Court’s reliance on these authorities to be inadvertent.

Rather, we interpret the Court to be deliberately citing them, to mimic the analysis of Thuraissigiam based on the Hendrix petitioner’s concession that his incarceration violated the Suspension Clause even if the Clause is confined to the scope of habeas relief available at the Founding. That is, Hendrix did not resolve the question about the scope of the Suspension Clause en toto. It merely decided the question presented to it by the petitioner, about whether his incarceration “pose[d] constitutional concerns” even conceding the lower circuit’s interpretation of the Suspension Clause. This reading explains the Court’s reliance on authorities expressly disavowing adjudication of the scope of the Suspension Clause protection, Thuraissigiam, 591 U.S. at 116, and authorities that describe Founding era habeas relief as the floor rather than the ceiling of the Suspension Clause protection, Boumediene, 553 U.S. at 746. Thus, we are not “uncertain[]” as to the holding of Hendrix.

27

Post at 41. We simply apply a more thorough analysis of the holding of Hendrix based on the reasoning supplied in the Hendrix opinion itself.

This interpretation of the majority decision in Hendrix is replicated in Justice Jackson’s dissenting opinion. Justice Jackson interprets the Hendrix majority as “admit[ting] that, at a minimum, the Suspension Clause protects the right of habeas corpus as it existed at the time of the founding.” Hendrix, 599 U.S. at 528 (Jackson J., dissenting) (citing Hendrix, 599 U.S. at 482–83). Moreover, Justice Jackson couches the Hendrix majority’s read of the Suspension Clause as a “suggestion” -- the kind of ambiguous terminology that is dissonant to the legal pronouncement the concurrence perceives. Id. at 528 n.24 (Jackson J., dissenting) (“I reject the majority’s suggestion that the Suspension Clause protects only the scope of the great writ as it existed in the founding era.”) (first emphasis supplied) (second emphasis in original). Notably, Justice Jackson’s point went unrebutted by the Hendrix majority.

Certainly, “[i]t goes without saying that the majority opinion, not the gloss that the [dissent] seeks to place thereon, is controlling.” Dababnah v. Keller-Burnside, 208 F.3d 467, 471 n.3 (4th Cir. 2000). But in this specific context, where the Hendrix majority relied on authorities that specifically reserved the ultimate question on the scope of the Suspension Clause, the dissent’s description is not “gloss.” Rather, it is a plausible interpretation of the majority’s decision, considering the proper context. One we agree with, pursuant to our own independent consideration of the Hendrix majority decision.

The concurrence asserts that our application of In re Vial is incorrect, since the decision is “incompatible” with Hendrix. Post at 41. But that conclusion only holds water

28

if one accepts the concurrence’s interpretation of Hendrix with respect to the Suspension Clause as a starting premise. And the concurrence fails to account for how its interpretation of Hendrix coheres with the context of that decision or the authorities the Court applied. Instead, the concurrence adopts a perspective that was not proffered by any party to this appeal and is at odds with Justice Jackson’s own unrebutted interpretation of her colleagues’ ruling. This footing does not provide adequate ground to overrule our own en banc precedent. Thus, we decline to take that step in this appeal. We instead adopt the more modest disposition proposed by Appellee and hold that denying Bell habeas relief pursuant to § 2255(h) does not amount to a suspension of the writ.

* * *

There is no constitutional right for every worldly wrong, no judicial solution for every perceived problem. The scope of the habeas statutes reflects Congress’ balance between the benefits of further review and the harms of endless litigation. In Bell’s case, Congress has chosen finality over error correction. We are bound to follow that choice. Section 2255 may be unavailable to Bell, but it is nevertheless adequate and effective to test the legality of his claim. So he cannot bring a habeas petition under § 2241. Bell’s further arguments are foreclosed by precedent. Regardless of the merits of his claim, the district court must be AFFIRMED.

29

RICHARDSON, Circuit Judge, concurring in the judgment with respect to Part II(C):

I agree that precedent forecloses Bell’s Suspension Clause claim but diverge from my good colleagues on which precedent does so. The Court in Part II(C) retreats to one of our prior decisions—a decision whose reasoning I find incompatible with recent Supreme Court precedent. On my view, like every other issue in this case, Bell’s Suspension Clause claim turns on Hendrix.

The Suspension Clause reads: “The Privilege of the Writ of Habeas Corpus shall not be suspended, unless when in Cases of Rebellion or Invasion the public Safety may require it.” U.S. Const. art. 1, § 9, cl.2. Bell contends that if he cannot use the saving clause, then he cannot get any relief at all under AEDPA’s scheme, and so Congress will have suspended habeas corpus in violation of the Suspension Clause. But the Suspension Clause only protects claims that would have been cognizable in habeas at our country’s Founding. Hendrix, 599 U.S. at 482–83 (“This Suspension Clause argument fails because it would extend the writ of habeas corpus far beyond its scope when the Constitution was drafted and ratified.” (cleaned up)). Bell’s claim extends beyond the ambit of late eighteenth-century habeas because he seeks to challenge the judgment of a court of competent jurisdiction on its merits. I would thus find that the Suspension Clause cannot help him.

The Suspension Clause is often invoked in—yet rarely relevant to the disposition of—modern-day habeas cases. This confusion arises because in modern usage, “habeas” has two meanings: a colloquial sense and another, distinct, constitutional sense. As

30

explained above, in colloquial usage, “habeas” often refers to the various statutory provisions governing postconviction review today. But when the Constitution speaks of the “Writ of Habeas Corpus,” it refers to the common-law tradition this country inherited from England. See Farkas, 972 F.3d at 553. It is this tradition—not any subsequent statutory enactments—that “receive[d] explicit recognition in the Constitution.” Id. (citing Rasul, 542 U.S. at 474) (alteration in original). While today’s statutory postconviction review regime permits claims that reach beyond the Founding-era scope of the writ, that is a policy choice Congress has made—not a constitutional command.

The only question relevant for Bell’s constitutional claim, then, is whether he would have had access to habeas at the Founding. “At the founding, a sentence after conviction ‘by a court of competent jurisdiction’ was ‘in itself sufficient cause’ for a prisoner’s continued detention.” Hendrix, 599 U.S. at 483 (quoting Brown v. Davenport, 596 U.S. 118, 129 (2022)). 1 In other words, “the black-letter principle of the common law [is] that the writ was simply not available at all to one convicted of [a] crime by a court of competent jurisdiction.” Paul M. Bator, Finality in Criminal Law and Federal Habeas Corpus for State Prisoners, 76 Harv. L. Rev. 441, 466 (1963). Put one more way, a prisoner at the Founding had no privilege vindicable by habeas to be freed from custody imposed as a criminal sentence unless some jurisdictional defect afflicted his judgment.

1

See also Dallin H. Oaks, Legal History in the High Court—Habeas Corpus, 64 Mich. L. Rev. 451, 468 (1966) (“[A] court disposing of a habeas corpus petition could not go behind the conviction for any purpose other than to verify the formal jurisdiction of the committing court.”).

31

To understand why common-law habeas—and thus the Suspension Clause—was unconcerned with the merits of jurisdictionally sound judgments, it is helpful to consider what habeas corpus was originally for. Since long before the Founding, habeas was a remedy for lawless confinement. See, e.g., Edward Coke, The Second Part of the Institutes of the Lawes of England 55 (London, Fischer & Young 1642); 3 William Blackstone, Commentaries on the Laws of England *137. And lawless confinement could take many forms, not just incarceration. For this reason, habeas had many applications, extending so far as to even provide recourse to “a wife confined by her husband.” Halliday, supra, at 43–44; see also id. at 35–38 (habeas as remedy for impressment into military service). Its sweep and versatility rightly earned it the moniker, “palladium of Liberty.” Id. at 8.

For all its many uses, one stood out: “Habeas is at its core a remedy for unlawful executive detention.” Dep’t of Homeland Sec. v. Thuraissigiam, 591 U.S. 103, 119 (2020) (quotation omitted). In pre-Founding England, detention by local executive officers who claimed no authorization but their titles was a common problem. The writ developed as a potent weapon against such detention. Halliday, supra, at 30. 2 This “classical function”— warding off inadequately authorized “detention by the executive or the military,” see Bator, supra, at 475—endures in America today as “an important judicial check on the

2

Reaching this point took centuries and bloodshed. Until the seventeenth century, habeas was at best an unreliable remedy for executive detention. See, e.g., Darnel’s Case (The Case of the Five Knights) (1627) 3 How. St. Tr. 1 (K.B.) (denying writ because confinement was at King’s command). Only by statute and then war was this use of habeas fully secured. See Habeas Corpus Act 1640, 16 Car. 1 c. 10 (expanding the writ to reach detentions by the King and Privy Council); Habeas Corpus Act 1679, 31 Car. 2 c. 2 (entrenching developments in habeas jurisdiction).

32

Executive’s discretion in the realm of detentions.” Hamdi v. Rumsfeld, 542 U.S. 507, 536– 38 (2004); see also Boumediene v. Bush, 553 U.S. 723, 739–45 (2008). Like preventing impressment and spouses confining one another, preventing summary executive detention falls within habeas’s core historical function because it aims at the same ill: confinement without law.

The problem with using habeas for postconviction review is that imprisonment as a criminal sentence does have the authorization of law. That law, of course, is a court’s judgment. As the Supreme Court held in Ex parte Watkins and reaffirmed three terms ago in Davenport, “a judgment of conviction” is “in itself sufficient cause” for detention. Davenport, 596 U.S. at 129 (quoting Ex parte Watkins, 28 U.S. 193, 202 (1830)). This concept is a familiar one. Historically, although a judgment may have been wrong, it was nevertheless final because of res judicata: “A judgment, in its nature, concludes the subject on which it is rendered, and pronounces the law of the case. The judgment of a court of record whose jurisdiction is final, is . . . conclusive on all the world.” Watkins, 28 U.S. at 202–03; see Davenport, 596 U.S. at 129 (relying on Watkins for this point). 3 But as with res judicata generally, there is “an important exception” to this rule: A judgment was not immune to collateral attack if the court rendering it lacked jurisdiction. Davenport, 596 U.S. at 130. For habeas purposes, a jurisdictionally defective “judgment [was] an absolute nullity” and could thus be set aside. Watkins, 28 U.S. at 203. By contrast, merits errors

3

The statutory scope of postconviction relief today no longer maps cleanly onto the traditional rules of preclusion. But that reflects Congress’s choice to offer more postconviction process than is constitutionally required. The Suspension Clause imposes only a floor, not a ceiling, on what Congress can do.

33

“constituted no ground for relief” and accordingly lay outside the concern of the Suspension Clause. Davenport, 596 U.S. at 129 (cleaned up).

Of course, criminal defendants could still seek error-correction on the merits. But that error-correction relied other procedural vehicles, not habeas. Similar to today, common-law habeas existed alongside writs of error, appeals, and certiorari: All of these guarded against erroneous judgments, but they were used in different ways. See, e.g., Sir John Baker, An Introduction to English Legal History 145–48, 156–60 (5th ed. 2019). “Even in its widest application . . . [habeas] [did] not enable an appeal on the merits of a decision to imprison. Its function [was] to question the lawfulness, not the inherent correctness, of an imprisonment.” Id. at 158. Early American courts followed this historical distinction. They required a “writ of error or other direct remedy”—not habeas, which is a collateral remedy—to reverse the “judgment of a court of competent jurisdiction.” Ex parte Toney, 11 Mo. 661, 662 (Mo. 1848) (emphasis added); Watkins, 28 U.S. at 202–03. As far as habeas was concerned, a judgment from a court of competent jurisdiction ended the matter.

So the natural question is: Was Bell sentenced pursuant to a judgment from a court of competent jurisdiction? Bell says no. According to Bell, a court lacks “jurisdiction” when it commits constitutional error, such as in his case when it (allegedly) imposes a sentence under an unconstitutional mandatory sentencing guideline. Oral Arg. at 9:30– 11:00. In asserting this view, Bell has company; the dissent in Hendrix and the dissent in

34

Davenport argued the same. See Hendrix, 599 U.S. at 528–29 (Jackson, J., dissenting); Davenport, 596 U.S. at 146–50 (Kagan, J., dissenting). 4 To Bell’s credit, the question he raises is not a simple one. “[T]he line between mere errors and jurisdictional defects was not always a ‘luminous beacon’ and [has] evolved over time.” Davenport, 596 U.S. at 129 (quoting Bator, supra, at 470); see also Danforth v. Minnesota, 552 U.S. 264, 271–72 (2008) (tracking the “expan[sion]” of the concept of jurisdiction). “Jurisdiction,” after all, “is a word of many, too many, meanings.” Steel Co. v. Citizens for a Better Env’t, 523 U.S. 83, 90 (1998) (quotation omitted). But though the boundary where jurisdictional defect ends and mere error begins may be blurry, the basic concept of jurisdiction that prevailed at the Founding is clear: Jurisdiction is the power to enter judgment in a case. See Rose v. Himely, 8 U.S. 241, 269 (1808), overruled on other grounds by Hudson v. Guestier, 10 U.S. 281 (1810); see also Grignon’s Lessee v. Astor, 43 U.S. 319, 338 (1844); Fauntleroy v. Lum, 210 U.S. 230, 234–35 (1908); Ryan C. Williams, Jurisdiction as Power, 89 U. Chi. L. Rev. 1719, 1729–30 (2022); William Baude, The Judgment Power, 96 Geo. L.J. 1807, 1850 (2008).

Still, this basic definition does not foreclose Bell’s view. One could reasonably believe that, in one sense of the word “power,” courts lack the power to make constitutional mistakes (like rely on unconstitutional statutes or misunderstand the scope of constitutional rights). If that were so, such merits errors might indeed be jurisdictional. See Hendrix,

4

Several academics also disagree with the Supreme Court majority. See, e.g., Jonathan R. Siegel, Habeas, History, and Hermeneutics, 64 Ariz. L. Rev. 505, 527 (2022).

35

599 U.S. at 529 n.25 (Jackson, J., dissenting) (citing Ex parte Siebold, 100 U.S. 371, 376– 77 (1880)); Davenport, 596 U.S. at 146–49 (Kagan, J., dissenting).

That this conception of power is best supported by dissents is telling. The majority of the Supreme Court has ruled out the possibility that merits errors, like the one Bell cites, deprive a court of its power to act: “[A] habeas court could ‘examin[e] only the power and authority of the court to act, not the correctness of its conclusions.’” Davenport, 596 U.S. at 129 (quotation omitted). Treating these two concepts as one—even when the merits error is constitutional—elides Davenport’s careful distinction, “trivializ[ing] . . . the writ” in a way “inconsistent with the presumption of finality that traditionally attached to criminal convictions.” Id. at 130 (quoting Brown v. Allen, 344 U.S. 443, 536, 543 (1953) (Jackson, J., concurring in the judgment)); see also Hendrix, 599 U.S. at 485–87 & n.7 (rejecting the argument that merits errors are jurisdictional and describing Ex parte Siebold as “seemingly abandoned”). 5 It would also contradict longstanding precedent. The modern Court has explained that the boundaries of Davenport’s jurisdictional-defect rule can be

5

In addition to American cases with a confused sense of jurisdiction, there are English cases that seemed to allow merits review using habeas. Many of these can be explained by the fact that writs of habeas corpus were not always used alone. For instance, something akin to direct review of convictions could be had in King’s Bench by writ of certiorari—not to be confused with the Supreme Court’s writ by the same name, which no longer bears a meaningful connection to the English writ. See Frank J. Goodnow, The Writ of Certiorari, 6 Pol. Sci. Q. 493, 501–05 (1891). The King’s Bench’s writ of certiorari was often aided by a ministerial form of habeas corpus necessary to physically produce the prisoner, whose presence was mandated by the longstanding norm that someone’s liberty should not be adjudicated in absentia. See 2 Matthew Hale, Historia Placitorum Coronæ 401 (London, Nutt & Gosling 1736); see also id. at 211. The ambit of habeas corpus when combined with the certiorari writ in England should therefore not be confused with the ambit of habeas alone, which was much more limited after a court had entered judgment. See Halliday, supra, at 118–19.

36

found in the older cases on which it relies. See Hendrix, 599 U.S. at 483–86 & n.8; Davenport, 596 U.S. at 128–130 & nn.1–2. And that precedent belies the notion that a court is powerless to mistake or misapply the law. So long as a court has the power to enter judgment, errors do not render that judgment a nullity.

The jurisdictional defect concept is not, however, a “habeas petitioner always loses”

rule; Bell’s problem is simply that he does not allege a true jurisdictional defect. Under the precedents the modern Court has stressed, three jurisdictional conditions are relevant to habeas—none of which covers constitutional merits errors. See Davenport, 596 U.S. at 128–30 & nn.1–2. To have jurisdiction in the sense relevant on habeas, the convicting court must have “had jurisdiction of the party, and of the offense for which he was tried, and has not exceeded its powers in the sentence which it pronounced.” The Ku Klux Cases, 110 U.S. 651, 653–54 (1884). If those three conditions were satisfied, the habeas court could “inquire no further.” Id.

The Supreme Court laid out these conditions in slightly varying terms across several early American habeas cases. See, e.g., Ex parte Parks, 93 U.S. 18, 21–23 (1876) (requiring the convicting court to have “jurisdiction of the person,” “of the offence,” and to have “[done] no act beyond the powers conferred upon it”); Ex parte Reed, 100 U.S. 13, 23 (1879) (similar); see also William S. Church, A Treatise of the Writ of Habeas Corpus § 362 (1886) (stating that habeas is unavailable when the convicting court possesses “jurisdiction of the person, of the subject-matter, and to render the particular judgment assailed”). Despite a temporary “shift” toward a broader conception of habeas relief

37

toward the tail end of the 20th century, the Supreme Court has recently confirmed the correctness of those early American cases. Davenport, 596 U.S. at 128–130 & nn.1–2.

The first required condition—that the court have jurisdiction over the person—

roughly tracks our modern notion of personal jurisdiction. The idea is that the convicting court must have been able to exert power over that specific individual. See, e.g., Ex parte Bigelow, 113 U.S. 328, 329–30 (1885) (confirming that the convicting court “had jurisdiction of the prisoner, who was properly brought before the court”). The historical baseline for this jurisdiction was territorial: “Where a party is within a territory, he may justly be subjected to its process, and bound personally by the judgment pronounced.” Picquet v. Swan, 19 F. Cas. 609, 612 (C.C. D. Mass. 1828) (Story, Cir. J.). For this reason, a sovereign generally had the power to criminalize and punish conduct occurring within its borders. See Joseph Story, Commentaries on the Conflict of Laws § 539 (6th ed. 1865); see also Emma Kaufman, Territoriality in American Criminal Law, 121 Mich. L. Rev. 353, 361–65 (2022). But this rule of thumb was neither exclusive nor exhaustive. As Story also acknowledged, “nations generally assert a claim to regulate the . . . acts of their own citizens, wherever they may be domiciled.” Story, supra, § 540; see also Kaufman, supra, at 375–96. And although a sovereign might lay claim to the persons within its territory, seldom will a single court have jurisdiction over every person within its territory—take, for instance, a court-martial without jurisdiction over civilians. See, e.g., Wise v. Withers, 7 U.S. 331, 337 (1830); Reid v. Covert, 354 U.S. 1, 19–21 (1957); see also Robert Leider,

38

Retiring Military Jurisdiction over Military Retirees, 68 Vill. L. Rev. 751, 777 (2023). 6 While questions of this sort may present hard cases, Bell has given me no reason to believe his convicting court lacked jurisdiction over him.

The second required condition—that the court have jurisdiction over the offense—

roughly tracks our modern notion of subject matter jurisdiction. Though the convicting court may have power over someone generally, it can convict for a crime “only when the offense charged is within the class of offenses placed by the law under its jurisdiction.” Bowen v. Johnston, 306 U.S. 19, 24 (1939); see also Ex parte Bigelow, 113 U.S. at 329– 30 (similar). When a sovereign allocates jurisdiction among its courts, it need not give every court power over all offenses, just as it need not give every court power over every person. Some courts may adjudicate only petty offenses. See, e.g., Callan v. Wilson, 127 U.S. 540, 556 (1888) (vacating felony conviction by D.C. police court because, among other reasons, its jurisdiction, “as defined by existing statutes, does not extend to the trial of infamous crimes or offenses punishable by imprisonment in the penitentiary”); Miller v. Snyder, 6 Ind. 1, 3 (1854) (similar). Others lack criminal jurisdiction altogether. See, e.g., Wedmore v. Indiana, 122 N.E.2d 1, 2–3 (Ind. 1954) (probate court conviction for assault

6

Questions of personal jurisdiction were more acutely relevant in English practice.

In the centuries leading up to the Founding, England had many courts with extremely limited personal jurisdiction. When they acted beyond that jurisdiction, their judgments were deemed “coram non judice.” The Marshalsea, 77 Eng. Rep. at 1038.

39

and battery). 7 While questions of this sort may present hard cases, Bell’s federal crime plainly fell within his convicting court’s subject matter jurisdiction.

As some of these examples suggest, jurisdiction over person or offense might fail not because proceedings happen in the wrong court but because the wrong sovereign tries to handle the case. In those cases, it’s hard to cut a clear distinction between personal and subject matter jurisdiction because both are subsumed within this fundamental problem. Take the example of crimes committed within a national park. See Bowen, 306 U.S. at 28– 30 (exclusive federal-court jurisdiction even though the park lies within a state). Or crimes committed by members of some Indian tribes. See Ex parte Mayfield, 141 U.S. 107, 115– 16 (1891) (exclusive tribal-court jurisdiction under treaty). And usually the federal courts lack power to adjudicate state common-law offenses. See Ex parte Coy, 127 U.S. 731, 757–58 (1888). 8 Likewise, the federal courts have exclusive jurisdiction over “crime[s] against the laws of the United States only.” Ex parte Bridges, 4 F. Cas. 98, 105 (C.C. N.D.

7

In England, this principle was of surpassing importance given its complex, multilayered system of courts. See, e.g., Sir John Baker, Sources of English Legal History: Public Law to 1750 at 332 (2024) (habeas to ecclesiastical courts); id. at 339 (habeas to administrative bodies); William F. Duker, The English Origins of the Writ of Habeas Corpus: A Peculiar Path to Fame, 53 N.Y.U. L. Rev. 983, 1023–25 (1978) (habeas to courts of admiralty).

8

This is not to suggest, however, that criminal jurisdiction over the same conduct is always exclusive. See, e.g., Heath v. Alabama, 474 U.S. 82, 88 (1985) (“When a defendant in a single act violates the peace and dignity of two sovereigns by breaking the laws of each, he has committed two distinct offences . . . for each of which he is justly punishable.” (quotations omitted)).

40

Ga. 1875) (Bradley, Cir., J.). 9 In cases like these, habeas would presumably be available for a prisoner convicted by a court belonging to the wrong sovereign. But again, Bell alleges no such defects, and none are apparent.

Most of the time, when a court has jurisdiction of the person and of the offense, its judgment will be jurisdictionally sound. But the third condition requires that a court have jurisdiction to impose the punishment it doles out. The Ku Klux Cases, 110 U.S. at 653– 54. Errors in assessing those punishments are generally merits errors, not jurisdictional ones. See In re Eckart, 166 U.S. 481, 482–83 (1897) (“explaining that “an error” in determining what “punishment” is “authoriz[ed]” for “the offense charged . . . does not present a jurisdictional defect”); see also Davenport, 596 U.S. at 129 n.1. But some such errors, we have been told, may be jurisdictional. See In re Mills, 135 U.S. 263, 270 (1890) (finding a sentence contrary to clear statutory command beyond the jurisdiction of the court); Ex parte Snow, 120 U.S. 274, 285 (1887) (granting habeas when the trial court sentenced a man for three crimes although the indictment charged only one); Ex parte Lange, 85 U.S. 163, 176 (1874) (granting habeas where the court had already meted out a sentence following conviction and then purported to impose a second one). 10 Weaving

9

Justice Bradley went on to explain that it would be a “manifest incongruity for one sovereignty to punish a person for an offense committed against the laws of another sovereignty.” Id. And in Bridges, there could be no doubt how the principle applied, because Congress had expressly given jurisdiction to the federal courts, “exclusive of the state courts, of all crimes and offenses cognizable under its authority.” Id.

The Supreme Court has repeatedly taken care to distinguish the theory animating

10

Snow and Lange cases from one that would authorize factual or legal innocence claims on habeas. See Hendrix, 599 U.S. at 485 n.7 (“[I]f a court has jurisdiction of the case the writ (Continued)

41

these cases, and others, together is a difficult task. But I see no need to thread the needle in this case, as Bell has not alleged anything resembling a claim that the court lacked jurisdiction to impose his 274-month sentence. So I would leave these problems for another day, comfortable with the conclusion that Bell’s judgment was jurisdictionally sound.

True jurisdictional defects like these, however, are thankfully rare. Mostly, courts act within their jurisdiction. Here, Bell’s allegations amount to an ordinary merits error, not a jurisdictional defect. See Oral Arg. at 9:30–11:00. And for that reason I would find the Suspension Clause is irrelevant to his claim. See Hendrix, 599 U.S. at 482–87.

My good friends take a different approach, urging that Hendrix is irrelevant to Bell’s Suspension Clause claim. On their view, uncertainty about Hendrix’s holding suggests we can just apply our pre-Hendrix decision In re Vial, which held that § 2255(h) categorically does not violate the Suspension Clause. 115 F.3d 1192 (4th Cir. 1997) (en banc). I disagree. First, while Hendrix presents hard questions, whether the protections afforded by the Suspension Clause are tied to the scope of habeas at the Founding is not one of them. Second, In re Vial’s categorical holding does not permit ignoring Hendrix’s Suspension Clause analysis, as In re Vial’s reasoning is incompatible with Hendrix.

of habeas corpus cannot be employed to re-try the issues, whether of law, constitutional or otherwise, or of fact.” (alteration and emphasis original) (quoting Glasglow v. Moyer, 225 U.S. 420, 429 (1912)). What is important in Snow and Lange is that the defendant was not in fact convicted—not whether the conviction turned on accurate facts or whether those facts met the offense’s elements or even whether the offense satisfied constitutional rules. Though the distinction is subtle, the Court continues to insist that it is important. See, e.g., Davenport, 596 U.S. at 129 & n.1; see also Hendrix, 599 U.S. at 485–86 & n.8.

42

As the Court acknowledges in Part II(C), the opinion in Hendrix “facially” supports the proposition that “the Suspension Clause is confined to claims cognizable in habeas at the Founding.” Ante, at 25; see Hendrix, 599 U.S. at 483 (relying on the scope of the writ at the Founding to announce the court-of-competent-jurisdiction rule); id. at 487 (making clear that the “Suspension Clause does not constitutionalize” post-Founding developments). I do not, contrary to the Court, think that an inferior court may disregard what Hendrix says based on a belief that the Supreme Court failed to adequately justify its holding. Ante, at 25–27. Nor do I think that the characterizations of the Hendrix majority given by Hendrix’s dissenting voices change the majority’s holding. Id. at 27–28.

This Suspension Clause holding in Hendrix is inconsistent with our Court’s reasoning in In re Vial, so that prior decision can no longer be applied. See Rose, 80 F.4th at 504. In re Vial held that § 2255(h)’s second or successive bar for federal detainees does not violate the Suspension Clause because § 2254’s second or successive bar (located in § 2244(b)) for state detainees does not violate the Suspension Clause. 115 F.3d at 1197– 98. But that reasoning cannot be squared with Hendrix itself. At the Founding, federal habeas did not extend to state detainees—this only changed upon the passage of the Habeas Corpus Act of 1867. Consequently, the constitutionality of § 2254 cannot bear on the constitutionality of § 2255(h) after Hendrix.

On my view we ought to give Bell the benefit of modern caselaw, and since In re Vial habeas jurisprudence has gone through a sea change. Bell alleges that his court lacked competent jurisdiction, and Hendrix tells us that a prisoner in executive detention, or a prisoner convicted by a court which lacked competent jurisdiction, might have Suspension

43

Clause protection even if he has filed successive petitions. That’s because his judgment is a nullity, and a claim of that sort was cognizable in habeas at the Founding. See Ex parte Watkins, 28 U.S. at 203. In other words, if Bell is right that his convicting court had a true jurisdictional defect, he would have a claim under Hendrix that § 2255(h) cannot constitutionally be applied against him. In re Vial would improperly preclude any such challenge under Hendrix’s framework. So it cannot stand. Rather than applying In re Vial—which did not apply anything resembling the test Hendrix required for Suspension Clause inquiries—I think Bell is entitled to the benefit of the Supreme Court’s current law, even if his effort here fails.