Justice Scalia famously described the Lemon test as a “ghoul in a late-night horror movie that repeatedly sits up in its grave and shuffles abroad, after being repeatedly killed and buried . . .” For many, this aptly describes the nondelegation doctrine—a ghoul that haunts the administrative state. Almost a century has passed since the Supreme Court invoked the nondelegation doctrine to strike down a legislative delegation. But the nondelegation doctrine refuses to die. And perhaps, for good reason.

Under the nondelegation doctrine, some limit prohibits Congress from delegating its legislative powers. But in recent terms, the Court has maintained the near-century-long tradition of burying the nondelegation doctrine. It has done so by locating “intelligible principle[s]” that purportedly confine congressional delegations within constitutional bounds. In Gundy v. United States, the Court declined to revive the nondelegation doctrine to invalidate a delegation that allowed the Attorney General to “specify the applicability” of the Sex Offender Registration and Notification Act (SORNA) to citizens convicted before SORNA’s enactment. Justice Gorsuch dissented—joined by Chief Justice Roberts and Justice Thomas. And last term, the Court once again eschewed reviving the nondelegation doctrine. In FCC v. Consumers’ Research, the Court upheld a delegation to the FCC to determine how much telecommunications carriers must contribute to the Universal Service Fund to achieve “universal service.” Justice Gorsuch dissented again, joined by Justice Thomas and Justice Alito.  And most recently, several Justices questioned at oral argument in the tariff cases whether a delegation to the President to set tariffs would violate the nondelegation doctrine.

Two scenarios may explain nondelegation doctrine jurisprudence over the past century. Either Congress is exceptionally skilled at including an “intelligible principle” in statutes or the Supreme Court has not exactly enforced the doctrine.[1] In the latter—and I would argue more plausible—scenario, the nondelegation doctrine “exists” but remains largely toothless.[2] But it remains unclear whether some limit bars Congress from delegating its legislative power.

In a recent article, I argue that such a limit does exist: Congress cannot delegate authority to administrative agencies to unilaterally determine the actus reus of a criminal offense. In other words, Congress cannot equip agencies with carte blanche authority to create “administrative crimes.” Such delegations are inconsistent with the original meaning of “legislative Power[].”

A recent Ninth Circuit case may offer the Supreme Court another opportunity to revive the nondelegation doctrine. In United States v. Pheasant, the Bureau of Land Management (BLM) arrested and indicted a man in Nevada for an administrative crime—punishable by up to 12 months in prison. The crime? Operating an off-road vehicle on public lands “from a half-hour after sunset to a half-hour before sunrise without lighted headlights and taillights.”

That crime won’t be found in a federal statute. Section 1733(a) of the Federal Land Policy and Management Act (FLPMA) grants the Secretary of the Interior authority to issue regulations “with respect to the management, use, and protection of the public lands.” FLPMA then delegates carte blanche authority to the Secretary to criminalize acts on BLM land; anyone who “knowingly and willfully” violates a BLM rule or regulation may face a prison sentence of up to one year.

In Pheasant, the district court held that FLPMA violated the nondelegation doctrine because Section 303(a) granted the Secretary “unfettered legislative authority” to create administrative crimes. The Ninth Circuit reversed, holding that “[e]ven in the criminal context, the ‘intelligible principle’ test provides the controlling legal standard for evaluating non-delegation challenges.” And under the “intelligible principle” test, only an “exceedingly modest limitation” restrains Congress from delegating its legislative power. Unsurprisingly, the Ninth Circuit located a requisite intelligible principle in FLPMA: “The Secretary must develop a long-term management strategy to realize the land’s value in a sustainable way.” This purportedly limited BLM’s criminal lawmaking authority.

But on October 31, the nondelegation ghoul appeared once again. The Ninth Circuit denied rehearing en banc in Pheasant, but Judge Bumatay and Judge VanDyke offered thoughtful dissentals. Consistent with my article, Judge Bumatay explored how more than an “exceedingly modest limitation” restrains Congress from delegating its criminal lawmaking authority.

When Congress delegates the authority to unilaterally determine the actus reus of a federal criminal offense, it violates the nondelegation doctrine. The original meaning of legislative power requires Congress to include a criminal penalty and an actus reus in order to “enact” criminal law. A delegation to create an entire criminal code regulating the private conduct of citizens is antithetical to the original meaning of legislative power. But this is precisely what Congress—and the Court—has allowed to proliferate over the past century. The Ninth Circuit’s treatment of Pheasant is no different.

Since United States v. Grimaud in 1911, the Court has allowed Congress to “enact” criminal law by merely establishing a criminal punishment for a statutorily undefined crime. In Grimaud, a sheepherder appealed a criminal conviction for failing to obtain a sheep grazing permit in violation of a United States Forest Service regulation. The Court determined that Congress—not the agency—criminalized grazing sheep without a permit because it enacted the specific punishment for violating the agency’s regulations. Under Grimaud, Congress need not identify an actus reus in a criminal statute to “enact” criminal law.

Grimaud rests on a fatal flaw. The Court confused a necessary condition for a sufficient one. It assumed that Congress “enacts” criminal law solely by setting a criminal punishment. That is wrong.

First, as an original matter, a “crime” consisted of at least two necessary conditions: a criminal act (actus reus) and a penalty. Blackstone, for instance, defined a “crime or misdemeanor” as “an act committed or omitted, in violation of a public law either forbidding or commanding it.” And James Madison emphasized that “on criminal subjects, it is proper, that details should leave as little as possible to the discretion of those who are to apply and to execute the law.” These “details” included the actus reus of a criminal offense. And because Congress had the exclusive authority to enact federal crimes, it could not delegate the authority to unilaterally define an actus reus to the Executive Branch. As Judge Bumatay recounts, the First Congress enacted criminal statutes with great specificity, clearly identifying the conduct that would permit the government to deprive citizens of their liberty, or even life. That is not to say Congress never granted discretion to the Executive Branch in applying criminal statutes. Congress occasionally granted the Executive Branch limited discretion to define the objects of the conduct Congress prohibited.

Second, early precedent confirms that Congress only “enacts” criminal law through including a penalty and a criminal act. The Marshall Court emphasized that Congress “must first make an act a crime” and “affix a punishment to it.” Chief Justice Marshall understood legislative power to require Congress to “define a crime, and ordain its punishment.” True, these early precedents involved challenges to the authority of the federal judiciary to create common law crimes. But the Marshall Court clearly understood “legislative Power[]” to impliedly require Congress to include a penalty and a criminal act when enacting criminal law. Third, it was not until over a century after the Founding that Congress began to delegate authority to unilaterally determine the actus reus of a criminal offense. Together, this historical evidence of the Constitution’s original meaning suggests that Congress must include more than an “exceedingly modest limitation” when it delegates its criminal lawmaking authority.

Consider recent precedent in light of Judge Bumatay’s dissental. In Gundy, Justice Gorsuch’s dissent emphasized that SORNA “endow[ed] the nation’s chief prosecutor with the power to write his own criminal code governing the lives of a half-million citizens.” But SORNA did not equip the Attorney General with carte blanche authority to create new federal offenses regulating the private conduct of convicted sex offenders. It did not, for instance, equip an agency with rulemaking authority to create any federal offenses insofar as the offenses were consistent with the “public interest” or some other capacious “intelligible principle.” Congress created SORNA’s requirements—not the Attorney General. Congress “only” delegated authority to decide whether and how SORNA’s specified requirements would apply to half a million citizens. SORNA’s delegation is indeed problematic. But it pales in comparison to the delegation at issue in Pheasant.

FLPMA equips the Secretary of the Interior “with the power to write his own criminal code governing the lives” of any citizen that happens to set foot on BLM land. Or—as Judge Bumatay put it—Congress effectively said “do as the minions of the Secretary of the Interior tell you—or go to jail.”

BLM’s authority is broader than many may suspect. BLM manages over 245 million acres of federal land. The BLM thus serves as the legislature, governor, and police force on 1 out of every 10 acres of land in the United States. And it’s difficult to imagine a crime that could not relate to a “long-term management strategy to realize the land’s value in a sustainable way.”

The BLM does not lack an imagination; it has criminalized a wide range of activities, including watching “noise producing devices” unreasonably loudly, repositioning a table at a campground, adopting wild horses, and placing antennas on rocks. A citizen may face up to 12 months in prison for violating any of these administrative crimes.

But Pheasant only scratches the surface of the labyrinth of administrative crimes. Congress often proclaims that a violation of “any rule or regulation” promulgated by a particular agency may constitute a federal offense. These delegations authorize agencies to create administrative crimes regulating a wide range of private conduct. Congress “enacts” criminal law solely by establishing a punishment—usually a Class A or Class B misdemeanor—while the agency independently determines the criminal act(s).

Pheasant may present the Supreme Court with a unique opportunity to revive the nondelegation doctrine without destabilizing the administrative state. After all, it is difficult to argue that federal prosecutors need more federal crimes to select from. Nor would Congress need to enact criminal legislation regulating the minutiae of activities on federal land to achieve its policy goals.[3] Congress could itself criminalize particular acts while delegating discretion to define the “precise objects” of the prohibited conduct. Congress may, for example, prohibit driving of vehicles that endanger wildlife on BLM land and delegate more limited discretion to BLM to determine precisely when and how particular vehicles would be prohibited. Or Congress could equip federal agencies with more robust civil enforcement mechanisms that don’t result in citizens unwittingly pleading guilty to federal offenses as they do under BLM’s current scheme.[4] Either solution would foster more, not less, congressional deliberation. For the past century, Congress has circumvented bicameralism and gifted administrative agencies blank checks to create new federal crimes.

The Constitution, as originally understood, prohibits Congress from freely handing over its criminal lawmaking reins. Grimaud’s premise that Congress must only specify a penalty to “enact” criminal law is wrong; Congress must also identify an actus reus. Grimaud has sanctioned a system where only an “exceedingly modest limitation” restrains Congress from delegating criminal lawmaking authority. But the Constitution demands more. If the Supreme Court revisits the nondelegation graveyard, Judge Bumatay’s Halloween dissental may haunt administrative crimes for years to come.

Nicolas Elliott-Smith is a Law Clerk at Sullivan & Cromwell LLP. His views do not necessarily represent the views of his employer.

---

[1] The bar is quite low. The Court has located an “intelligible principle” in the phrases “the public interest,” “fair and equitable,” and “requisite to protect the public health.” Whitman v. Am. Trucking Associations, 531 U.S. 457, 474 (2001) (summarizing nondelegation precedents).

[2] This does not imply that the Court has not enforced nondelegation principles through other means. One possible explanation for the Court’s unwillingness to revive the nondelegation doctrine is that the major questions doctrine has enabled the Court to avoid reaching the constitutional issue—the nondelegation doctrine—through statutory construction.

[3] The Assimilative Crimes Act also serves as a backup—incorporating state law that is applicable to conduct occurring on federal lands when the act is not punishable by a federal statute. See 18 U.S.C. § 13.

[4] The majority of BLM’s administrative crimes are processed as “petty offenses” without any judge, prosecutor, or defense lawyer reviewing the case. They are often resolved when the defendant pays the Central Violations Bureau a fine, known as a “forfeiture of collateral,” which has the effect of pleading guilty to the offense. See Brief of Amici Curiae Ninth Circuit Federal Public and Community Defender Offices in Support of Defendant-Appellee, United States v. Pheasant, No. 23-991 (9th Cir. June 13, 2025).