*This is the third post in a series on Cass Sunstein and Adrian Vermeule’s new book Law and Leviathan: Redeeming the Administrative State. For other posts in this series, click here.

In Law and Leviathan, an expertly written account of contemporary administrative law, Cass Sunstein and Adrian Vermeule show us how to fight a theory with a theory. The most fervid critics of the regulatory state, they charge, “venture dramatic, large-scale challenges” to administrative law, but wholly overlook the “concrete materials of actual [administrative] law.” (S&V 14.) When those building blocks are examined, the authors contend, the picture that emerges is a far cry from Stuart despotism and secret law. Rather, they demonstrate, contemporary administrative law doctrine is anchored to the rule-of-law ideals championed by Lon Fuller: precepts of fairness, notice, consistency, and transparency that infuse administrative law with “internal morality.” In practice, the authors are careful to acknowledge, the administrative state may not always live up to its noblest aspirations. But its heart is in the right place.

As a synthetic account of administrative law’s first-order structure, Law and Leviathan succeeds remarkably well. Why must agencies follow their own rules? Why must agencies state their justifications for action beforehand, not contrive them post hoc? Why is retroactive rulemaking so strongly disfavored? The answers to all of these puzzles (and many more, the authors show) are best discovered not by closely parsing the APA’s text or the Constitution, but instead by consulting a slim chapter of Fuller. When one examines the many rules governing how agencies may act, the rubric of internal morality can fairly be said “to unify a disparate array of judge-made doctrines, and perhaps even the field as a whole.” (S&V 42.)

If one moves up a plane, however, administrative law’s claim to internal morality becomes a bit harder to defend. Administrative law may function commendably to ensure that agencies act with internal morality, but the courts that enunciate those doctrines remain free to operate outside Fullerian guardrails. As I will explain below, several important doctrines that the Supreme Court applies to agency action are themselves vulnerable to critique on various Fullerian scores: (1) the doctrines are not transparent; (2) they are difficult to understand; (3) they are applied on an ad hoc basis; and (4) they create a disjoint between “the rules as announced” and “the rules as administered.” By applying these doctrines to agency action, the Court does indeed encourage agencies to honor Fullerian virtues—as the authors contend. The trouble is that in enunciating these doctrines, the Court itself often performs poorly on the metric of Fullerian morality. And because administrative law encompasses not only what agencies do, but also what courts do, a complete appraisal of administrative law’s claim to internal morality requires us to assess not only whether agencies must adhere to Fullerian ideals, but also whether the Court has held itself to Fullerian ideals.

To begin to fill in the latter piece, consider what it would take to say that the Court-developed doctrinal landscape of administrative law lived up to some of Fuller’s minimal requisites. In such a world, the tests for the constitutionality of delegation and for the judicial review of agency action would be transparent, in the sense that they would clearly communicate to Congress and to agencies which statutory delegations and regulations would be held legally valid. The rules governing judicial scrutiny of agency action would be understandable, not only to the most immediately “regulated parties” (Congress and the executive branch), but also more broadly to those who might seek the benefits or protections of such doctrines and to the lower courts that have to apply those doctrines. The doctrines of judicial review of agency action would not be ad hoc, but instead would be articulated in rule-like fashion and applied uniformly across like cases. And, last if not least, the rules governing judicial review in administrative law would work as advertised: the rules proclaimed to govern agency action would actually match up with the rules administered and applied by courts.

Does the doctrinal landscape of administrative law resemble that picture? In some ways, yes—but in other ways, not really. Let us take some recent examples from various areas of administrative law and see how they fare on these four Fullerian metrics.

Transparency: In Fuller’s rubric, a system of legal rules must have publicly available content, so that affected parties are “aware of the rules with which they must comply.” To meet that threshold in this context, Congress must be aware of the rules with which its delegations must comply in order for those delegations to be legally valid. And an agency must be aware of the rules with which its regulations must comport if they are to withstand legal challenges.

From a birds’-eye view, black-letter administrative law surely meets that test. Despite recent rumblings from five sitting justices,¹1.Gundy v. United States, 139 S. Ct. 2116, 2137-42 (2019) (Gorsuch, J., dissenting, joined by Roberts, C.J., and Thomas, C.J.); Gundy, 139 S. Ct. at 2131 (Alito, J., concurring in the judgment); Paul v. United States, 140 S. Ct. 342 (mem.) (2019) (Kavanaugh, J., respecting the denial of certiorari). the rule remains that delegations will be upheld so long as the statute contains an intelligible principle. And (bracketing requirements imposed by the APA on rulemaking) regulations will be upheld if they fall within the bounds of an agency’s delegated authority and constitutional limits.

So far, so good, on the metric of transparency. On closer inspection, however, both doctrines are far more muddled. A case in point is Gundy v. United States, in which the four-justice plurality (with a fifth vote reluctantly supplied by Justice Alito’s concurrence) deflected a nondelegation challenge by reading into SORNA a standard to guide the Attorney General: that convicted sex offenders should be registered “as soon as feasible.”²2.Gundy, 139 S. Ct. at 2125. The Gundy plurality’s approach, as the authors suggest, would entail that courts “should interpret statutes in such a way as to avoid giving blank checks to administrative agencies” and that courts should instead read “seemingly open-ended statutes in a way that supplies an intelligible principle.” (S&V 123-24.) That is all very well, at least insofar as it’s a standard that lower courts can conscientiously attempt to apply to future nondelegation challenges to “open-ended statutes.” But what about Congress—the entity that must draft statutes with this standard looming over it? It seems doubtful that members of Congress are aware of when courts will view statutes as providing “blank checks”; as the 4-3 split in Gundy itself showed, whether a statutory check is blank is in the eye of the beholder. And if a court should happen to decide that a statutory check is in fact blank, it is still less likely that Congress could know what court-supplied intelligible principle may lurk in the offing. The reading given to the statute in Gundy, for example, was not one that could have been predicted readily—not even by close readers of SORNA nondelegation caselaw. Of the eleven courts of appeals to opine on a nondelegation challenge to SORNA, not a single one understood the statute to contain the limiting principle (“as soon as feasible”) that four justices ultimately discerned in it.³3.United States v. Nichols, 775 F.3d 1225, 1231-1232 (10th Cir. 2014), rev’d on other grounds, 136 S. Ct. 1113 (2016); United States v. Richardson, 754 F.3d 1143, 1145-1146 (9th Cir. 2014) (per curiam); United States v. Cooper, 750 F.3d 263, 266-272 (3d Cir. 2014); United States v. Goodwin, 717 F.3d 511, 516-517 (7th Cir. 2013); United States v. Kuehl, 706 F.3d 917, 918-920 (8th Cir. 2013); United States v. Sampsell, 541 Fed. Appx. 258, 259 (4th Cir. 2013) (per curiam); United States v. Parks, 698 F.3d 1, 7-8 (1st Cir. 2012); United States v. Felts, 674 F.3d 599, 606 (6th Cir. 2012); United States v. Guzman, 591 F.3d 83, 91-93 (2d Cir. 2010); United States v. Whaley, 577 F.3d 254, 262-264 (5th Cir. 2009); United States v. Ambert, 561 F.3d 1202, 1212-1214 (11th Cir. 2009).

The obvious problem Gundy illustrates is that nondelegation doctrine is only nominally defunct. The nondelegation doctrine survives, albeit disguised in the form of statutory interpretation. That incarnation of nondelegation doctrine may advance separation of powers values, but it does so at the cost of the clarity lauded by Fuller. Post-Gundy, it has become harder to say that Congress (or the public) can know what Congress’s past delegations will be taken to mean, and also harder to state with confidence how broadly Congress can write a future delegation without having that statute be read as a blank check. To a Fullerian focused on law’s transparency, there’s not much to commend about this state of affairs.

Turn now to agencies and whether it can fairly be said that agencies are on notice of when their regulations will pass legal muster. Here, the major questions exception rears its befuddling head. If an agency’s regulation addresses a matter of “deep ‘economic and political significance,’” then the agency’s interpretation of ambiguities in the statute will not receive deference by the Court—which must then evaluate the meaning of the statute de novo. In its most recent application in King v. Burwell, the major questions exception seemed to merge with the nondelegation doctrine: though the Court refused to hold that Congress had delegated to the agency the authority to expend billions in tax credits, the Court was nonetheless willing to reach that same result on the Court’s own de novo reading of the statute. The upshot today is that in some uncertain class of cases, agencies cannot rely on judicial deference to sustain their regulations and must instead persuade a court that Congress has by statute commanded the result the agency wishes to achieve. The King Court’s renewed embrace of the major questions exception, with its overtones of nondelegation, has thrown a shadow across a slew of regulatory actions, both during the Trump Administration and already in the Biden Administration. Yet it remains unclear—which is to say, it remains not transparent—when the major questions exception applies. As with the nondelegation doctrine, this problem of transparency in the very heartland of administrative law is a Fullerian deficit.

Understandability: As the foregoing discussion has presaged, the doctrines of judicial review of agency action are not easy to understand nowadays. And the unfortunate fact is that they seem to be becoming less, not more, understandable. Chevron began its career as a relatively simple two-step. These days, it is more of a minuet. We have long been accustomed—thanks to Sunstein—to speaking of a Chevron Step Zero, but nowadays scholars and courts (with justification) point to a Chevron Step One-And-A-Half, a Chevron Step Three,⁴4.CASA de Maryland, Inc. v. Trump, 971 F.3d 220, 252 (4th Cir.), reh’g en banc granted, 981 F.3d 311 (4th Cir. 2020). a Chevron four-step—even a Chevron “Staircase” that interposes six additional interstitial steps between Steps One and Two. Similarly, Kisor v. Wilkie ratified the “Chevronization of Auer” by turning “a simple, straightforward shortcut for quickly and predictably resolving potentially thorny cases” into a morass of “multifarious exceptions and qualifications” that—at its inception—boasts no fewer than five steps. The authors contend that the “reticulated framework[s]” of Kisor and Chevron deference allow agencies and courts to achieve a “stable if uneasy equilibrium” between “competing views.” (S&V 131.) To those critics of Chevron who premise their claims on constitutional or statutory originalism, of course, the whole Chevron framework appears illegitimate. But even if our metric for evaluating Chevron’s legitimacy is not originalism (of whatever stripe) but Fullerian ideals, the issue doesn’t vanish entirely. It’s just that on the latter metric, it’s the “reticulat[ion],” rather than the “framework,” that poses the problem.

Ad Hoc Doctrine: It would go a league too far to say that administrative law doctrine suffers from Fuller’s first failure—the “failure to achieve rules at all, so that every issue must be decided on an ad hoc basis.”⁵5.Lon Fuller, The Morality of Law (1963), at 39. Administrative law is obviously awash with rules and doctrines, and certainly not every issue in administrative law is decided ad hoc. Yet there are still decisions that it’s hard to read without fearing that they may bear, in invisible ink, a Bush v. Gore style “good for this ride only” stamp. The census case is perhaps the most notable recent candidate for this dubious laurel.⁶6.Dep’t of Commerce v. New York, 139 S. Ct. 2551 (2019). The Court remanded, holding that the Commerce Department’s asserted justification for including a citizenship question on the census questionnaire was pretextual. As Sunstein and Vermeule explain, “the Court’s pretext analysis was notably limited,” but “indisputably indicated that in ‘unusual circumstances’ a reviewing court might declare an agency’s sole stated reason pretextual, in light of the whole course of the agency’s behavior, and thus remand for a further explanation or a change of course by the agency.” (S&V 140.) Now, as a rule for guiding agency behavior—as a holding that demands that agencies candidly state the justifications for their actions—this outcome is aligned with the Fullerian commitment to “congruence between the rules as announced and their actual administration.” (S&V 140 (quoting Fuller)). Viewed from a broader angle, however, the case appears in a different light. The Court’s terse endorsement of the occasional availability of pretext review—review that will be applied only in “unusual circumstances” depending on a court’s perception that an agency’s reasons are “contrived” in light of the “whole course” of the agency’s behavior—was itself ad hoc. And that decision will in turn produce more ad hoc decisions in the years to come: lower courts will now sometimes find official pretext to doom agency action, and sometimes not, and regardless will make that finding without the guidance of rule-like criteria from the Court and instead on an all-things-considered basis. Looked at from that perspective, it’s harder to conclude that—on net—the census case promoted administrative law’s fit with Fullerian precepts (though the case’s outcome may be defended for other reasons).

Rules as Announced and Rules as Administered: The occasional ad hoc decision, or even line of ad hoc decisions, is one thing, but what about administrative law more broadly? Does administrative law display a mismatch, a lack of congruence, between the rules as announced and the rules as administered? Here, the answer is straightforward—it does. Indeed, the authors themselves point to that answer: as they acknowledge, there is a “degree of irony” in their position, in that they defend principles of administrative law’s morality that are “nowhere expressly laid down in the APA.” (S&V 9.) That text is “worthy of great respect,” they note, but “the text does not exhaust the law, which also contains Fullerian principles that promote reasoned administrative lawmaking and that are hallowed by long usage, the explicit endorsement of the courts, and the tacit acquiescence of Congress.” (S&V 97.) A plank of the book’s argument is that (descriptively) the APA’s text is not the alpha and omega of administrative law, and that (normatively) this is a good thing.

Certainly, there are strong reasons to defend the Court’s interpretive approach to the APA’s text, and even the evident inconsistencies within that approach. To put it mildly, the APA was not drafted during a heyday of textualism. And the Court’s eschewal of textualist methodology when it comes to certain provisions of the APA (in particular, Section 706) is in line with a long-running tradition of employing non-textualist and non-originalist approaches to kindred domains involving procedural, jurisdictional, and remedial statutes and rules. Yet the fact that the law as announced in the APA’s text differs so markedly from the law as administered by the courts is itself a not-insignificant issue, at least if we are evaluating administrative law doctrine solely through the lenses of Fuller’s avuncular spectacles.

* * *

As these briefly sketched illustrations suggest, though the argument from Fuller laid out in Law and Leviathan has powerful explanatory force and profound justificatory appeal, its project of “redeeming” the administrative state is, in one sense, a bounded one. The safeguards of internal morality that the book distills and defends are safeguards that mainly hem in agencies taking agency action and that in consequence (and to a lesser degree) may influence Congress as it drafts statutes. What remains unbroached by this account is the question of how the Court’s administrative lawmaking fares on the various metrics of Fullerian internal morality. Unlike agencies—which must attend to Fullerian guardrails or else suffer the consequences—nobody can make the Court follow Fullerian ideals; the extent to which the Court achieves internal morality in its decision-making is essentially up to it. Perhaps predictably, then, when one searches for internal morality in the Court’s articulation of administrative law doctrine, the imprint of Fullerian values grows much fainter.

For the authors’ project, the Court’s failure to hold itself more faithfully to Fullerian ideals in its own handiwork has two related, though distinct, implications. The first is its downstream effect on agencies. The more difficult it becomes for agencies, lower courts, and private parties to know and understand what first-order rules apply to agency action, the more haphazard will be the effects of those first-order rules—which will have the knock-on consequence of undermining the guardrails of internal morality that Sunstein and Vermeule rightly celebrate. The second is that the Court’s inattentiveness to Fullerian values attenuates administrative law’s claim to overall morality. Administrative law extends beyond agencies to the Court, if only because such an important portion of the work that agencies do eventually winds up before the Court. If administrative law as a whole is to be said to possess internal morality, then internal morality ought to inhere not only in how agencies regulate and decide, but also in how courts regulate and decide. A full bookkeeping of administrative law’s internal morality, and hence its claims to legitimacy, must therefore account for both types of internal morality, as well as the dynamic relationship between them.

In sum, even as Sunstein and Vermeule have given us reason to praise administrative law’s internal morality in many domains, they have also given us a basis on which to critique its internal morality in an additional important domain: within the Court’s own handiwork, a constitutive aspect of administrative law, the values of internal morality could be more fully (Fullerly?) realized. To be sure, it may set too high a bar to insist that administrative law as a whole be imbued not just with internal morality at the level of agency action, but also internal morality at the level of the Court’s decisional caselaw. As Vermeule has elsewhere taught us, and the authors here also note (S&V 97), comporting with principles of (internal) legality is itself a value that has to be traded off against other values. Among these crosscutting demands are welfarist considerations, the need for some play in the joints in cases with hard facts, the need for a system that will accommodate change over time, and an understandable inclination to continually tinker with and re-adjust long-standing legal frameworks rather than suddenly jettison them. The Court, like an administrative agency defending a complicated but comprehensible regulation, might reasonably respond that in striking a trade-off between these competing demands, its decisional caselaw handily surpasses the floor of the “morality of duty” even if it falls well short of reaching the firmament of the “morality of aspiration.” Yet the authors’ eloquent defense of Fullerian values—safeguards that they hail as representing the very minimal constitutive elements of legality!—should prompt us all to consider seriously whether that trade-off is being struck correctly by the Court. For a century or more, the Court has exerted steady pressure on agencies to keep them within the guardrails of internal morality as they go about doing their work of regulating and adjudicating. Perhaps it would not go amiss for the Court to take some of its own medicine, as it goes about enunciating administrative law doctrine.

Mila Sohoni is a Professor of Law at the University of San Diego School of Law.