Administrative Crimes: A Qualified Defense

On his way out, President Trump sought to “protect Americans from overcriminalization” by trying to limit the criminal enforcement of regulatory offenses. Hostility to administrative crimes is growing at the Supreme Court too, in part as an outgrowth of concern about the administrative state generally, but also because the very notion that someone can be prosecuted for an offense defined by an agency and not Congress seems abhorrent to Justice Gorsuch and others. Such prosecutions regularly occur, however, not just for the SEC Rule 10b-5 violations most familiar to readers, but for hoarding personal protective equipment during the COVID-19 pandemic, violating Iranian sanctions, and a host of other offenses defined by agencies pursuant to congressional delegation. That Chevron deference to agency interpretation of the relevant legislation is often urged makes matters even worse for critics, but, at least so far, the main problem is the very idea that an institution other than Congress can restrict “liberty.”

In a forthcoming essay in the Yale Journal on Regulation, I undermine this critique of administrative crimes by showing how Congress has also delegated power to define federal crimes to other partners and that the well-accepted justifications for doing so are just as valid (or problematic) as they are for granting that power to regulatory agencies.  I then argue that the elimination of administrative crimes would not only do little to change the mix of defendants actually prosecuted but also considerably reduce the specificity and clarity of federal criminal law.

The formal critique of criminal offenses defined by agencies is straightforward and powerful. The very nature of criminal law in a liberal democracy requires that the moral condemnation of conduct come from the community, speaking though its representatives. Canons like the “rule of lenity” ensure that legislators, not courts or prosecutors, will be the sole source of and fully accountable for criminalization.  Moreover, such separation of powers promotes not merely accountability but liberty, with the requirements of bicameralism and presentment providing salutary checks on the supply of criminal law.

Notwithstanding its power, however, this critique simply does not reflect the reality of federal criminal law.  Even as the U.S. Supreme Court regularly denies the existence of “common law crimes,” the judicial hand in defining federal offenses is sustained and often quite heavy.  Certainly the scope of mail or wire fraud owes far more to courts than to Congress.  Defenses are almost completely judicial creations, and so, originally at least, was corporate criminal liability.  The Supreme Court’s claims that it sets mens rea requirements only as a matter of statutory interpretation are wholly disingenuous. Criminal civil rights prosecutors are regularly based on judicial interpretations of the Due Process Clause and other constitutional provisions.[1] And of course there are criminal antitrust prosecutions.

Perhaps the insulation of courts from the political process and their adherence to precedent-based reasoning makes their accountability of less concern than that of agencies.  But the transparency of the agency rulemaking process and the susceptibility of judicial lawmaking to agenda control and case selection by prosecutors means that rule-based offenses will be far more attentive to due process values like notice and ex ante specificity than crimes largely defined by case law.

While congressional delegations of crime-definition authority to federal courts are generally implicit, masked by a fiction of “adherence to legislative intent,” those to state law are explicit and sustained, as when state law predicates become the bases for federal offenses or when, under the Assimilative Crimes Act, state penal law fills gaps in the criminal law applicable to federal enclaves.  Foreign law too provides the basis for a number of federal criminal offenses, not just the Lacey Act – which, because of its broad focus on foreign fish and wildlife violations that might not even be criminal, has attracted critical attention – but money laundering provisions that target funds obtained in the course of violations of another country’s corruption statutes.  And the scope of the federal piracy statute is set by the “law of nations.”

To be sure, the risk that these diverse congressional delegations will impose strange non-federal norms on federal criminal defendants is critically limited by the absolute gatekeeping authority of federal prosecutors.  No federal charges will be brought absent a federal executive decision to do so. But that gatekeeping extends to administrative crimes as well. Indeed, while a separation of powers formalist might condemn all instances in which Congress is not the dominant definer of federal criminal law, a functionalist more open to the virtues of collaborative lawmaking, the deficiencies of the federal legislative process, and the pathologies of unbridled prosecutorial discretion might value existing institutional dynamics.  Just consider the Rule 10b-5 offense – authorized by Congress, unleashed by the SEC, developed by courts, and controlled (in its criminal form) by prosecutors.  That is certainly not an optimal arrangement if the gold standard is thoughtful ex ante legislative specificity by a democratically accountable legislature.  But those familiar with Congress’ actual work might value the alternative legislative arrangements that have developed – hence the concern about the prosecutorial license permitted by the Second Circuit’s interpretation of 18 USC § 1348 in United States v. Blaszczak,[2] which allows prosecutors to avoid the “personal benefit” requirement developed in the 10b-5 insider trading cases.

Eliminating administrative crimes would not likely change who gets prosecuted. Already when it comes to pursuing environmental crimes, prosecutors turn to Title 18 in almost 45 percent of cases, charging standard conspiracy, false statement, obstruction, and fraud statutes. And we would see a massive uptick in charges of conspiracies to defraud the government of information, since so many of the regulatory violations that are currently pursued criminally will inevitably entail some degree of deception. Doing away with explicit administrative crimes would thus vastly reduce the notice and legislative specificity of the charged offenses, with the prosecutorial landscape dominated by sweeping statutes that, because the federal interests implicated leave no room for constraining federalism canons, the Supreme Court refuses to curtail.

Even as we appreciate the structural and procedural protections that flow from the current regime of agency rule promulgation and prosecutorial discretion, we might still worry about the risk of overcriminalization were an agency to focus exclusively on its own regulatory concerns and enforcement realities and not on the prospect that a rule might become a basis for a criminal prosecution when violated “willfully.”  One proposal addressing this risk would require that agencies “list and make generally available in full text all regulations that carry potential criminal penalties, and perhaps that Congress then be required to ratify any such regulation before it can provide the basis for a criminal prosecution.”[3] In the same vein, the Trump executive order requires, among other things, (a) that all future regulations explicitly indicate whether violation of any provision therein can itself be a basis for criminal liability, and (b) that future regulations “explicitly state a mens rea requirement for each such provision or identify the provision as a strict liability offense.”[4]

Marginal transparency efforts can’t hurt.  But if one looks not for formal attention to abstract separation of power principles but real care for “liberty” in the form of well-crafted and specific legislation, administrative crimes, however flawed, look a lot better than any realistic alternative.


[1] For an exploration of the extensive judicial role in federal crime definition, see Daniel Richman, Kate Stith & William Stuntz, Defining Federal Crimes (2d ed. 2018), available free in e-book form by emailing and asking for an access code to “Casebook Connect.”

[2] 947 F.3d 19 (2d Cir. 2019).  The Supreme Court recently vacated and remanded this case for “consideration in light of Kelly v. United States, [140 S. Ct. 1565 (2020)].”  2021 WL 78043 (mem.) (Jan. 11, 2021). But as Kelly does not implicate the circuit’s “personal benefit” analysis, the latter may well survive, at least in the near term.

[3] Michael B. Mukasey & Paul J. Larkin, The Perils of Overcriminalization 2, Heritage Found.: Legal Memo at 5 (Feb. 12, 2015.

[4] Executive Order 13980, Protecting Americans From Overcriminalization Through Regulatory Reform (Jan. 18, 2021), 86 Fed. Reg., No. 13, at 6817.

This post comes to us from Professor Daniel Richman at Columbia Law School. It is based on his recent article, “Defining Crime, Delegating Authority – How Different are Administrative Crimes?,” available here.