Rather than hiding regulatory elephants in mouse holes, Congress can extend the reach of the administrative state only through clear statements.
Professors Cass Sunstein and Adrian Vermeule of the Harvard Law School are perhaps today’s two most distinguished defenders of the modern administrative state. They are not wooly-eyed thinkers who pretend that the administrative state is perfect in all of its particulars. As their book title suggests, however, they do believe the current rules do an adequate job of using law to restrain Hobbes’s Leviathan, primarily by reconciling two imperatives: maintaining sufficient flexibility to allow administrative agencies to discharge their delegated functions in the modern, post-New Deal state, and simultaneously offering sufficient protections for individual rights and privileges against serious government abuse.
The first element of their defense claims that the standard safeguards contained in Lon L. Fuller’s most famous book, The Morality of Law (1969), are for the most part observed in modern administrative law. Their second, and more impassioned, plea is that it is wholly proper for the modern state to allow administrative agencies substantial discretion in interpreting the law, a power granted by the two hardy standbys, Chevron v. Natural Resources Defense Council (1984) and Auer v. Robbins (1997).
Our authors take on, albeit only indirectly, conservative critics of the current synthesis, such as Professors Philip Hamburger and Gary Lawson, as well as Justice Neil Gorsuch, whom they collectively dub as the “New Coke” (as opposed to the more successful Coke Classic?)—that is, as the intellectual heirs of Sir Edward Coke, the preeminent early 17th century jurist. However, Sunstein and Vermeule remain confident that key Supreme Court justices, most notably Chief Justice John Roberts, care enough about continuity and stability in the legal order to resist this right-wing onslaught. For my part, I have renewed the attack on the modern administrative state in my recent book, The Dubious Morality of Modern Administrative Law, whose thesis is not addressed by Sunstein and Vermeule, but whose pages cover many of the cases and issues that I discuss here.
My objections to the Sunstein/Vermeule position can be summarized in three propositions. First, it is incorrect to argue that Fuller’s procedural principles offer a sufficient defense of the administrative state. Second, Sunstein and Vermeule mistakenly treat administrative law as a self-contained set of procedural rules that operate independently of the substantive rules they are designed to enforce. In most cases, therefore, our authors regrettably never detail the underlying factual record, nor do they quote the relevant statutory or judicial language. Together, these two significant omissions create a dangerous disconnect between the law in action and the law on the books. Third, as a consequence of these omissions, Law and Leviathan insists that rescuing Chevron and Auer from their many critics need address the doctrinal niceties only. In so doing they commit the far greater sin of ignoring the blatant and avoidable abuses of the administrative state in area after area, whether it be in environmental, labor, civil rights, security, patent, or water law.
The Undeveloped Morality of Administrative Law
Fuller sets the stage by way of his now-famous example of a hapless, if well-intentioned, sovereign named Rex who misfires on a variety of procedural protections that are everywhere necessary for a just legal system. These include protections against vices that make it difficult for firms to conduct their businesses confident that they will remain on the right side of the law—ad hoc decisions, insufficient notice of legal obligations, retroactive legislation, inconsistent rules, impossible demands, frequent rule changes, and a want of congruence between the rules as announced and the rules as administered. Notably, Fuller was only talking about legislation. Accordingly, he did not mention the common law safeguards appropriate for adjudication before both administrative or judicial bodies, most critically the right to a trial before a neutral arbiter and an opportunity to present a full defense. Nor does his account of the unitary Rex open the discussion to pressing issues of separation of powers. Each of these points must be included in any systematic account of administrative law.
Sunstein and Vermeule optimistically claim that by and large the constraints built into the administrative state respect Fuller’s commands, and thus bolster the “internal” morality of administrative law. But their point breaks down almost immediately because the modern administrative state gives little or no protection against retroactive legislation and related abuses. Likewise, the modern deference to administrative agencies allows for dramatic flip-flops in administrative law whenever political control shifts from one party to the other. To give an example, the innocent phrase “waters of the United States” was once interpreted to mean navigable waters on which boats or logs could float. But after a collusive 1975 settlement between the Secretary of the Army and the Natural Resources Defense Council, the Army Corps of Engineers issued a new definition which covered not only navigable and nonnavigable waters, but also lands adjacent to navigable waters. Ten years later, a deferential Supreme Court, relying on Chevron, sustained those regulations. So vindicated, the Corps imposed a fine of $37,500 per day, later doubled, on a landowner for filling in dirt on a dry lot separated by several built-on lots from any navigable river. While that excess was ultimately rebuffed in the Supreme Court on ad hoc grounds, similar extravagant claims are still being made. On these developments, Sunstein and Vermeule remain silent.
To be sure, the classical administrative law principles in place before the New Deal avoided such massive abuses. As Aditya Bamzai has shown in a 2017 article (never discussed by Sunstein and Vermeule), the calling card of nineteenth-century administrative law was that any settled practice by an administrative agency was entitled to judicial deference. That deference arose because custom combined expertise and consistency, developed by public officials in charge of dealing with the routine business of the time—personnel matters of military officers and civil servants, patent and land grants, and tariffs. Unlike our modern legal framework, it was the unprincipled deviation from the rule that courts rejected, not the uniform practice. Both Chevron and Auer rejected the wisdom of that incremental approach by allowing a huge flip under the Clean Air Act and the Fair Labor Standards Act, respectively, each on partisan grounds. Indeed, through Justice Thomas no less, National Cable & Telecommunications Ass’n v. Brand X Internet Services (2005) held that Chevron deference applies even when agencies make inconsistent determinations over time, to which Sunstein and Vermeule merely comment that “Brand X’s approach may simply be at odds with Fullerian intuitions about inconsistency over time,” full stop.
The key issue of retroactive changes in legal rules tells the same sad tale. Fuller dismissed these changes as a “monstrosity” when practiced by Rex. If anything, the delegation of Rex’s legislative authority to an administrative power makes a retroactive shift in rules even more dangerous, given the widespread tendency of delegated bodies to take positions more extreme than those of the enacting legislature. Take, for instance, the expansive regulations explicating the definition of “harm” under the 1973 Endangered Species Act. With one stroke of the pen, the federal government was allowed to commandeer vast tracks of land for habitat protection without having to pay a dime in compensation at all. This danger has only been magnified since the 1990s through the widespread use of administrative guidances under the Food and Drug law and the Civil Rights laws to circumvent the need to subject proposed rules to notice and comment proceedings. An example of such abuse is found in the Obama administration’s “Dear Colleague Letter” of 2011, which stripped the accused of key protections on burden of proof and cross examination in sexual harassment cases.
Sunstein and Vermeule pass by these examples in silence. They do address, if unsatisfactorily, Gundy v. the United States (2019), where they defend the congressional delegation of discretion to the Attorney General (AG) as to the application of an onerous statutory registration requirement to sexual offenders convicted prior to the passage of the statute. Tracking the opinion of Justice Elena Kagan, they reason that the statute required the AG to apply the registration requirement to the “maximum extent feasible,” even though the AG never made any careful analysis of this key up/down issue that the Congress could have decided for itself. Thankfully, Sunstein and Vermeule are out of the mainstream on this issue, as a unanimous right/left coalition of amicus curiae briefs insisted that in effect, such an interpretation as applied to criminal cases deviated from Fuller’s notion of justice.
This cavalier attitude in criminal cases has led to serious violations of basic constitutional protections in other regulatory proceedings, which Fuller also deplored. Sunstein and Vermeule claim that it is “the traditional and mainstream understanding in American public law that when agencies—acting within a statutory authority—make rules, interpret rules, and adjudicate violations, they exercise executive power, not legislative or judicial power.” To be sure, it is often hard to disentangle legislative from executive decisions given the inherently difficult contours of the delegation doctrine. But it is utterly inexcusable to throw judicial powers into that mix when specialized Article I courts can be set up with limited terms in office, preserving the key element of judicial independence.
The Abnegation of Due Process
Without that separation, abuses come thick and fast. One administrative horror story is Lucia v. Securities and Exchange Commission (2018), which considered the constitutionality of the SEC appointment of administrative law judges to try serious cases—unsurprisingly, the SEC ran up an impressive string of victories using its home court advantage. In a Kafkaesque adjudication replete with procedural and substantive shortcuts taken by the administrative law judge, Ray Lucia, a financial planner and investment advisor, was banished from his profession for the flimsiest of reasons. Justice Kagan gave Lucia a new trial, having concluded the designation of the administrative law judge did not meet the requirements of the Appointments Clause. But that glitch was easily remedied: the SEC issued proper letters of appointment, leading to yet another trial which barred Lucia from the securities industry. (Lucia was permitted the right to reapply to the SEC—good luck with that—and settled from sheer exhaustion, financial and emotional.) What was needed from the Supreme Court was a clear declaration that such an intra-agency trial, where the SEC acts as both prosecutor and judge, are an ugly affront to the guarantees of procedural due process.
Similar abuses took place in Oil States v. Greene’s Energy (2018), where in a post-grant patent review proceeding, the America Invents Act empowered the head of the Patent Trial and Appeal Board to displace Article III courts, appoint whatever judges from a long list he wanted to hear a case, and then to further stack the panel with more judges, even himself, if he did not like the initial outcome. The Due Process issues cry out to be addressed, yet Sunstein and Vermeule remain silent.
Indeed, the most indefensible omission in this entire book is the large number of breakdowns in the administrative law system that genuinely matter to the health and welfare of a firm. At one point, Sunstein and Vermeule ask what all the fighting is really about: “Is constitutional liberty really at risk if an agency is allowed to interpret the phrase ‘subject to’ or the word ‘diagnosis,’ within the bounds of textual meaning? ‘Bound books’? ‘Dairies’? Is liberty less at risk if, in the face of ambiguity, courts, composed of generalist judges, interpret terms on their own?”
These remarks refer to two well-known Supreme Court cases, United States v. Mead Corporation (2001), an old-style tariff case, and Kisor v. Wilkie (2019), a complex disability benefits determination case, and Sunstein and Vermeule’s rhetorical challenges deserve an emphatic response: Yes! Sure, the public rightly does not much care about the precise definition of “books” or “dairies,” or even arcane questions about veteran’s benefits. But it does care very much about the precise question in Auer, which was whether police sergeants and lieutenants fall into the class of “executive, administrative, and professional” employees who are exempt from the overtime regulations of the 1938 Fair Labor Standards Act. Look at any standard definition of these terms, and the case for recognizing the exemption is a no-brainer. Allowing political actors like then-Secretary of Labor Robert Reich to make mincemeat of the law shows why generalist judges, skilled in the art of statutory interpretation, should have the final say in interpreting statutory words. All of the endless wrinkles on judicial deference to administrative decisions quickly disappear if courts follow the basic architecture of the APA, which under Section 706 assigns to the courts the task of de novo interpretation of all key text, whether clear or ambiguous.
At this point, the remainder of the APA scheme falls into place. Treat the administrative agency as a rough substitute for a district court, whose decisions on simple questions of fact should be respected unless they are clearly erroneous, but whose judgments on mixed questions of fact and law (e.g., did certain actions constitute medical negligence?) are subject, roughly speaking, to intermediate scrutiny. Sunstein and Vermeule make spirited efforts to undermine this tripartite division of judicial oversight. But it is just as serviceable in administrative law, as it has been for centuries in ordinary civil litigation.
Moreover, once that conventional scheme of judicial review is disregarded, modern administrative law goes further astray by using a double-standard for grants and denials of permits. The so-called “hard look” doctrine announced in Motor Vehicle Manufacturers Association v. State Farm (1983)—again ignored by Sunstein and Vermeule—lets federal courts perform searching reviews of administrative permit approvals, while simultaneously requiring administrative denials of permits only to withstand rational basis review.
This divergence has created a genuine present-day crisis in the construction of new airports, bridges, roads, nuclear plants, and pipelines. Under the critical 1971 D.C. Circuit Court of Appeals decision Calvert Cliffs’ Coordinating Committee v. U.S. Atomic Energy Commission, the National Environmental Policy Act of 1969 (NEPA) was read to let any environmental group challenge any administrative permit in court. The result has been the end of new nuclear power plants in the United States, as well as a consistent effort to strangle fossil fuels by launching endless attacks on such major projects as the Keystone XL Pipeline and the Dakota Access Pipeline. In January 2020, the Trump administration urged a major update in these regulations, which I strongly supported, but which has been fiercely opposed by environmental groups. Sadly, the current NEPA regime is often counterproductive, all too often stopping projects that would reduce environmental risks by replacing dangerous, older facilities with safer new ones. Yet once again, Sunstein and Vermeule devote not a single word of Law and Leviathan to this contentious set of contemporary issues. Only dairies and daybooks indeed!
The case for major administrative law reform rests on the need to respond a growing crisis on the ground, and so long as a studied indifference to the true state of administrative law persists, the rot will continue: Leviathan will overcome Law, big time.