Procedural & Doctrinal Pivots

Habeas and Mandamus: The Two Great Writs at the Edges of Custody

Federal practice Long-form analysis May 10, 2026

The writ of habeas corpus and the writ of mandamus are sometimes called sister writs. They share a courtroom, a vocabulary, and a common ancestor in the prerogative writs of the English King’s Bench. They are both, in the modern federal system, “extraordinary” remedies invoked after ordinary remedies have failed or are unavailable. And in the daily life of federal practice, lawyers sometimes plead them in the same complaint, side by side, as if they were interchangeable forms of the same plea: set my client free, or, in the alternative, make the government act.

They are not interchangeable. Habeas reaches custody; mandamus reaches duty. Habeas asks whether the petitioner’s body is being held in violation of law; mandamus asks whether an officer is refusing to perform a non-discretionary obligation owed to the plaintiff. The two writs occasionally do the same work — and the cases in which they do are often the most consequential immigration and post-conviction matters in the federal courts — but they do it from different sides of a doctrinal line that is older than the Republic.

This essay maps that line. It is intended as a single reference treatment for practitioners who file under both writs, for law students who have encountered them in separate chapters of separate casebooks, and for the noncitizen or federal prisoner who has been told, often by another prisoner, that “you should file mandamus on that.” The first half is doctrinal and historical. The second half is a practitioner-facing FAQ.

I. Common roots, divergent functions

Both writs descend from the family of prerogative writs issued by the Court of King’s Bench in the name of the sovereign — certiorari, habeas corpus, mandamus, prohibition, procedendo, and quo warranto. The English Parliament did not invent habeas in 1679; the Habeas Corpus Act of that year codified and disciplined a writ that had been in use for centuries, requiring jailers and other custodians to produce the body of the prisoner and to state the cause of detention. Mandamus, by the seventeenth century, had become the means by which the King’s Bench compelled inferior courts and royal officers to perform duties owed to the subject. The two writs were complementary instruments of the same superintending jurisdiction: one freed the body, the other moved the office.

The American Founding inherited both. The Suspension Clause of Article I, section 9 protected the privilege of the writ of habeas corpus from suspension absent rebellion or invasion. Section 14 of the Judiciary Act of 1789 conferred on the new federal courts the power to issue habeas and “all other writs not specifically provided for by statute” in aid of their respective jurisdictions — the direct ancestor of today’s All Writs Act, codified at 28 U.S.C. § 1651. Section 13 of the same Act gave the Supreme Court authority to issue mandamus to federal officers.

That section 13 grant produced the most famous mandamus case in American law. In Marbury v. Madison, 5 U.S. (1 Cranch) 137 (1803), Chief Justice Marshall held that William Marbury had a vested right to his commission, that mandamus was the appropriate remedy, and that the Supreme Court could not issue it — because Congress’s attempt to lodge mandamus jurisdiction in the Court’s original docket exceeded Article III. Marbury is the Court’s great judicial-review case, but it is also a mandamus case, and its doctrinal contribution to mandamus is durable: where there is a legal right, there is also a legal remedy, and that remedy reaches the executive when the executive is bound by law.

Four years later, the Court turned to the other writ. In Ex parte Bollman, 8 U.S. (4 Cranch) 75 (1807), Marshall held that the Supreme Court could issue habeas to inquire into the commitment of two of Aaron Burr’s associates by a lower federal court. Bollman established that federal habeas is a creature of statute, traceable to section 14 of the 1789 Act, and that the writ is available against detention by federal officers under federal authority. Together, Marbury and Bollman set the channels: mandamus runs to officers and courts; habeas runs to custodians.

The nineteenth century reinforced the channels. Kendall v. United States ex rel. Stokes, 37 U.S. (12 Pet.) 524 (1838), established that federal mandamus could compel the Postmaster General to perform a payment duty Congress had ordered — an early statement of the principle that executive officers are not above the law and that courts can compel performance of a strictly ministerial duty. Ex parte Yerger, 75 U.S. (8 Wall.) 85 (1868), reached the opposite kind of question and reached a related answer: even after Congress repealed the appellate-habeas provision of the 1867 Habeas Corpus Act, the Supreme Court retained appellate habeas jurisdiction under the residual 1789 framework. The lesson of Yerger — that Congress’s power to strip habeas jurisdiction has constitutional limits — reappears, transformed, in the immigration cases of the twenty-first century.

II. The modern statutory architecture

The modern federal system distributes the two writs across a small set of statutes. Habeas is governed by chapter 153 of Title 28: § 2241 (general habeas, the vehicle for federal prisoners challenging the execution of a sentence and for noncitizens challenging immigration detention), § 2254 (state-prisoner post-conviction), and § 2255 (federal-prisoner motion to vacate, channeled to the sentencing court and serving as the principal substitute for traditional habeas in collateral attacks on a federal judgment). The Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA) layered on a one-year statute of limitations, gatekeeping for second and successive petitions, and the Certificate of Appealability requirement of § 2253(c).

Mandamus, in its modern federal form, has two distinct sources. 28 U.S.C. § 1361, enacted in 1962, gives the district courts original jurisdiction of any action in the nature of mandamus to compel an officer or employee of the United States or any agency thereof to perform a duty owed to the plaintiff. That is the statute under which an applicant whose I-130 has languished sues USCIS, or under which a federal prisoner sues the Bureau of Prisons over good-time credits. The All Writs Act, 28 U.S.C. § 1651(a), supplies the appellate side: it authorizes federal courts to issue all writs “necessary or appropriate in aid of their respective jurisdictions,” and is the source of the modern petition for a writ of mandamus filed in a court of appeals against a district judge under Federal Rule of Appellate Procedure 21.

An administrative-law overlay matters in modern practice. Most federal mandamus actions in the immigration and benefits contexts are pleaded together with a claim under the Administrative Procedure Act. 5 U.S.C. § 706(1) directs reviewing courts to compel agency action unlawfully withheld or unreasonably delayed. When the government answers, the unreasonable-delay analysis runs through the six-factor test of Telecommunications Research & Action Center v. FCC, 750 F.2d 70, 79–80 (D.C. Cir. 1984) (the “TRAC factors”): a rule-of-reason for agency timing; statutory timetables; whether human health and welfare are at stake; the effect of expedition on competing agency priorities; the nature of the interests prejudiced; and the absence of any need to find impropriety in order to find unreasonable delay.

The two writs thus sit on opposite sides of a procedural and conceptual line. Habeas is a custody remedy, statutory in form but with a constitutional floor in the Suspension Clause. Mandamus is a duty remedy, statutory in form, with no comparable constitutional floor — Congress could, in theory, narrow § 1361 in a way it cannot narrow § 2241.

III. The custody line and the duty line

The custody requirement

Habeas under § 2241(c)(3) requires that the petitioner be in custody in violation of the Constitution or laws or treaties of the United States. The Supreme Court has read “custody” capaciously but not infinitely. In Jones v. Cunningham, 371 U.S. 236 (1963), the Court held that a parolee remains in custody because parole imposes conditions which significantly confine and restrain his freedom. A decade later, Hensley v. Municipal Court, 411 U.S. 345 (1973), extended custody to a defendant released on his own recognizance pending execution of sentence. Custody, after these cases, is not coterminous with physical confinement; it includes meaningful restraint short of bars and walls.

Custody is not, however, infinitely elastic. Maleng v. Cook, 490 U.S. 488 (1989) (per curiam), held that a sentence that has fully expired no longer satisfies the custody requirement, even where the prior conviction is being used to enhance a current sentence. Collateral consequences alone do not suffice. The corollary — that a noncitizen at liberty whose adjustment-of-status application has stalled is not in habeas-cognizable custody — is not academic. It is the principal reason mandamus, not habeas, is the only available federal remedy for tens of thousands of immigration-benefits claimants whose petitions sit unadjudicated.

The mandamus standard

The Supreme Court has consolidated the modern federal mandamus standard around three conditions, restated most authoritatively in Cheney v. United States District Court, 542 U.S. 367, 380–81 (2004): the petitioner must have no other adequate means to attain the relief he desires; the petitioner must show that [his] right to issuance of the writ is clear and indisputable; and the issuing court, in the exercise of its discretion, must be satisfied that the writ is appropriate under the circumstances.

“The remedy of mandamus is a drastic one, to be invoked only in extraordinary situations.”
— Allied Chemical Corp. v. Daiflon, Inc., 449 U.S. 33, 35 (1980) (per curiam).

The standard’s pedigree is older. Kerr v. United States District Court, 426 U.S. 394, 403 (1976), required a showing that the right is clear and indisputable. Will v. United States, 389 U.S. 90, 95 (1967), characterized mandamus as the means to confine an inferior court to a lawful exercise of its prescribed jurisdiction or to compel it to exercise its authority when it is its duty to do so (internally quoting Roche v. Evaporated Milk Ass’n, 319 U.S. 21, 26 (1943)). And Ex parte Fahey, 332 U.S. 258, 259 (1947), called the writs drastic and extraordinary remedies reserved for really extraordinary causes.

Ministerial versus discretionary

The conceptual core of mandamus is the distinction between ministerial duty and discretionary judgment. As the Court put it in Heckler v. Ringer, 466 U.S. 602, 616 (1984), mandamus is intended to provide a remedy only if the defendant owes him a clear nondiscretionary duty. Mandamus can compel an officer to act; it cannot ordinarily compel the officer to act in a particular way, because where a statute or regulation commits the choice to the officer’s judgment, the right to a particular outcome is, by definition, not clear and indisputable.

This is why mandamus succeeds against pure delay (the duty to decide is non-discretionary; reasonableness of timing is reviewable) but almost always fails against substantive outcomes (denials, revocations, prosecution declinations). It is also why the everyday mandamus suit pleads § 1361 and APA § 706(1) in the alternative: the delay theory survives the discretion objection because reaching any decision is a ministerial duty, even if the content of the decision is not.

Channeling and exclusivity

Where Congress has fashioned a specific review scheme, mandamus is usually foreclosed because an “adequate alternative remedy” exists. Heckler v. Ringer, 466 U.S. at 616–17, applied that principle in the Social Security context. The immigration analog is 8 U.S.C. § 1252(b)(9) — the “zipper clause” — which we treat below. The mirror-image proposition, which we treat next, is that where Congress has stripped habeas, the Suspension Clause requires an “adequate substitute,” even if not the writ in name. Boumediene v. Bush, 553 U.S. 723, 771 (2008).

Where one writ reaches and the other does not

The two writs collide most usefully at three margins.

No custody, but unlawful agency inaction. A noncitizen at liberty whose application has stalled cannot file habeas. Mandamus, paired with the APA, is the proper vehicle.

In custody, but the harm is judicial delay rather than unlawful detention. When a habeas petition itself sits undecided for years, the petitioner cannot file another habeas (and may face AEDPA bars even if she could). The natural remedy is appellate mandamus to the court of appeals under the All Writs Act, asking the appellate court to direct the district court to rule.

In custody, and detention is unlawful. Habeas is the proper vehicle and mandamus is typically denied as duplicative — an “adequate alternative remedy” defeats the first Cheney prong.

IV. Mandamus to compel a habeas ruling

The most procedurally interesting intersection of the two writs is mandamus brought to compel a ruling on a long-pending habeas matter. The classic example is capital habeas. When a death-row petition sits under submission for years, the State (or, in different posture, the petitioner) may seek mandamus from the court of appeals.

The Supreme Court’s reference point is In re Blodgett, 502 U.S. 236 (1992) (per curiam), in which the State of Washington sought mandamus to compel a Ninth Circuit panel to decide a capital habeas appeal that had been under submission for more than two and a half years. The Court denied the petition on the strength of the panel’s explanation but warned, in a sentence that has become the operative standard, that any further postponements or extensions of time will be subject to the most rigorous scrutiny. Blodgett is the case for the proposition that appellate courts may, and on a sufficient showing must, intervene by mandamus where habeas adjudication has stalled.

AEDPA reinforced the principle in the special chapter applicable to opt-in capital states (28 U.S.C. ch. 154), which imposes statutory time limits on capital habeas review and provides for mandamus to enforce them. Outside that chapter, the framework is the ordinary one: an extraordinary remedy reserved for extraordinary delay, with the petitioner bearing the heavy Cheney burden.

A related question — whether a Certificate of Appealability denial can be reached by mandamus — was reframed by Hohn v. United States, 524 U.S. 236 (1998). Hohn held that a COA application is a “case” in the court of appeals subject to certiorari review under § 1254(1). Because certiorari therefore supplies an adequate alternative, mandamus to compel issuance or denial of a COA is now disfavored: the first Cheney prong cuts the other way.

V. Federal prisoners: why mandamus almost never works

For federal prisoners attacking their convictions or sentences, § 2255 is the ordinary vehicle, filed in the sentencing court. The savings clause at § 2255(e) preserves access to traditional § 2241 habeas only when the § 2255 remedy is “inadequate or ineffective to test the legality of [the prisoner’s] detention.”

The Supreme Court’s recent encounter with the savings clause, Jones v. Hendrix, 599 U.S. 465 (2023), narrowed it considerably. Justice Thomas, for a 6-3 majority, held that the savings clause does not allow a federal prisoner to evade AEDPA’s bar on second or successive § 2255 motions by relying on an intervening statutory-interpretation decision. Justices Sotomayor, Kagan, and Jackson dissented sharply. After Jones, the savings clause is a narrow channel.

Mandamus filings from federal prisoners fail in nearly every imaginable posture for a single reason: § 2255 (or, where it does not reach, the residual § 2241 channel) is presumed to be an adequate alternative remedy under the first Cheney prong. The narrow exceptions in practice are mandamus to compel a ministerial Bureau of Prisons act (good-time computation, sentence-credit application) where the duty is non-discretionary; and the Blodgett-style mandamus to compel a district court to rule on a long-pending § 2255 motion. Even those are rare.

What this means for counsel is that mandamus is almost never the first move for a federal prisoner. The right move is the § 2255 motion, the § 2255(e) savings-clause § 2241 (where it survives Jones), or, where the harm is institutional rather than judicial, an APA-style claim or a Bivens action. Mandamus is the last instrument on the bench, and it is the right instrument only when no other is at hand.

VI. The immigration hybrid — delay, detention, and the zipper clause

Nowhere are habeas and mandamus pleaded together as often as in immigration practice. The reason is structural: the modern immigration system runs on adjudications and detentions, and the two writs map onto the two sides of that system — mandamus for delayed adjudications, habeas for unlawful detentions.

Mandamus and the APA: delay litigation

USCIS adjudicates millions of applications a year. When those adjudications stall — a five-year-old I-130, a four-year-old I-485, a three-year-old I-589 — the noncitizen has no custody on which to base habeas, and the petition-for-review framework reaches only final orders of removal. The remaining route is a complaint in district court invoking 28 U.S.C. § 1361 (mandamus to compel an officer of the United States to perform a duty owed the plaintiff), 5 U.S.C. § 706(1) (compel agency action unlawfully withheld or unreasonably delayed), and 28 U.S.C. § 1331 (federal-question jurisdiction). The government’s standard answer is that the duty owed is to adjudicate, not to grant, and on that point the government is correct — the relief a court can order is a decision, not a particular decision. The unreasonableness inquiry runs through the TRAC factors. In practice, complaints filed under this regime resolve more often than not by stipulation or remand, after the agency moves the file to adjudication.

For naturalization applicants, an even stronger statutory hook exists. 8 U.S.C. § 1447(b) provides that, if USCIS fails to act on a naturalization application within 120 days after the examination, the applicant may sue in district court, and the court “has jurisdiction over the matter and may either determine the matter or remand the matter, with appropriate instructions, to the Service.” That second clause is the rare power to take an immigration adjudication out of the agency entirely. It is technically not mandamus — it is original jurisdiction over the underlying adjudication — but it functions as a more potent cousin in the same family of compel-the-government remedies.

Habeas in immigration detention

The other half of the docket is detention. Section 2241 has been the principal vehicle by which noncitizens challenge:

• Indefinite post-removal-period detention — Zadvydas v. Davis, 533 U.S. 678 (2001), held that 8 U.S.C. § 1231(a)(6) does not permit indefinite detention and adopted a presumptively reasonable six-month detention period beyond the ninety-day removal window, after which the detainee may seek release on a showing that there is no significant likelihood of removal in the reasonably foreseeable future.
• Pre-removal mandatory detention — Demore v. Kim, 538 U.S. 510 (2003), held § 1226(c) detention of removable criminal aliens during proceedings constitutional, while leaving open the prolonged-detention question that has driven a generation of as-applied litigation.
• Bond and procedural protections — Jennings v. Rodriguez, 583 U.S. 281 (2018), rejected the statutory reading that would have required periodic bond hearings under §§ 1225(b), 1226(a), and 1226(c), but left the constitutional question for remand. Post-Jennings habeas litigation has continued to develop the as-applied due-process framework in the lower courts.

The Suspension Clause forms a constitutional floor under all of this. INS v. St. Cyr, 533 U.S. 289, 305 (2001), avoided what it called “a serious Suspension Clause issue” by reading AEDPA and IIRIRA not to eliminate § 2241 review of pure questions of law in the removal context. Boumediene v. Bush, 553 U.S. 723 (2008), held that the Suspension Clause reaches Guantánamo and that the Detainee Treatment Act’s review procedures were not an “adequate substitute” for habeas. The lesson of these cases is that habeas, even when nominally restricted by statute, retains a constitutional minimum that reaches at least to the inquiry whether the petitioner is in custody for an unlawful reason.

The zipper clause and what it forecloses

The doctrinal counterweight on the immigration side is 8 U.S.C. § 1252. Subsection (a)(5) channels review of final orders of removal exclusively to the courts of appeals on a petition for review. Subsection (b)(9) — the “zipper clause” — consolidates all questions of law and fact . . . arising from any action taken or proceeding brought to remove an alien into that same petition-for-review process. The Court in Jennings v. Rodriguez read § 1252(b)(9) narrowly enough to leave detention claims outside its reach, on the ground that detention is collateral to removal rather than “arising from” it. But the courts of appeals have continued to wrestle with how broadly the zipper clause sweeps, and 2025–26 has seen renewed litigation about whether First Amendment–targeted ICE detentions can be reached by habeas at all or must be channeled into the petition for review.

DHS v. Thuraissigiam, 591 U.S. 103 (2020), drew a third line. The Court held, 7-2, that an expedited-removal asylum-seeker who sought additional procedural review through § 2241 fell outside the scope of the writ as it existed when the Constitution was adopted. Because Thuraissigiam sought an opportunity to apply for asylum rather than release from custody, his claim was not within the historical scope of habeas. Thuraissigiam is the most consequential habeas-narrowing immigration case of the last decade. It does not foreclose habeas for release from custody, and it does not foreclose mandamus or APA actions for delay. It does, however, push some kinds of immigration claims out of habeas and into other channels.

The hybrid complaint

Modern immigration-detention practice often pleads habeas under § 2241, mandamus under § 1361, APA delay under § 706, and declaratory relief under 28 U.S.C. § 2201 in a single complaint. The custody question controls habeas; the duty question controls mandamus; the unreasonable-delay question controls APA. A court that finds no custody will dismiss the habeas count but may still grant mandamus relief on a non-discretionary duty unreasonably withheld. A court that finds custody and unlawful detention may grant habeas relief and treat the alternative claims as moot. The discipline the practitioner must keep is to plead the writs separately and explicitly — conflation invites motions to dismiss for misjoinder of theories.

VII. State court echoes

State courts retain their own original-jurisdiction mandamus traditions, often more robust than the federal version. The state “writ of mandate” (California, in different terminology) and analogous proceedings in other states are routinely used to compel parole boards to grant hearings, departments of corrections to apply credits, and trial courts to rule on motions that have languished. State courts also retain their own habeas traditions, distinct from federal review under § 2254. The line between the two is the same line we have already drawn: substantive challenges to a sentence belong in habeas; ministerial duties belong in mandamus. State systems have generally been less willing than federal courts to allow habeas as a vehicle for collateral attacks once direct review has run, which gives mandamus a slightly larger niche on the state side.

The federal practitioner should note one practical consequence: a state habeas petitioner who has exhausted state remedies and is moving to federal § 2254 cannot use mandamus to leapfrog over an unfavorable state-court ruling, both because mandamus reaches only federal officers under § 1361 and because federal courts will not direct state courts by writ of any kind absent extraordinary justification.

VIII. Practitioner FAQ

The following is intended for lawyers and informed pro se litigants. It is not legal advice. The cases discussed in the essay above are the load-bearing authorities for each answer.

The two writs share a name in legal vocabulary — the extraordinary writ — and a place in the constitutional architecture. They do different work. Habeas is the writ of bodies and walls; mandamus is the writ of duties and offices. The federal practitioner who keeps the line clear, and who pleads each writ to the harm it is built to address, will not waste motion on either.