Adrian Vermeule, John H. Watson Professor of Law at Harvard University, is a bold, original, and often brilliant thinker—and at times his own worst enemy. Along with a cornucopia of trenchant insights on (administrative) law, his copious output contains cheap polemics; intemperate attacks on scholars and judges who peddle “libertarian” law; reckless flirtation with proto-fascist legal tropes; and wild theorizing backed by little more than ipse dixit.
That latter tendency in particular is on display in Law’s Abnegation: From Law’s Empire to the Administrative State. The book’s principal theme is the perennial tension between the lawful government of “classical” constitutionalism and the demands of a large-ish administrative state. The U.S. Supreme Court, on the eve of the New Deal, sought to find a workable formula. Crowell v. Benson (1932) gave administrators judicial “deference” on some (not all) questions of fact, though never on law. That formula, Vermeule seeks to demonstrate, proved unstable not merely on account of external events and exigencies but for compelling “internal” reasons embraced by judges and lawyers. Deference soon migrated to “mixed” questions of law and fact and from there to questions of law, eventually including the scope of an agency’s jurisdiction. That is where we are and, in Vermeule’s decided view, where we ought to be.
Much of this is substantially and importantly right; and for reasons I’ll explain, conservative or libertarian proponents of lawful, constitutional government would do themselves a big favor by taking Vermeule’s challenge seriously. Not that the author makes it easy: he manages to bury his forcefully argued point underneath a grand but hopelessly overwrought and misguided theory.
The Arc of Administrative Law
Ever since the 1930s, courts have increasingly deferred to administrative determinations of reasonableness. Vermeule’s central claim is that they have done so for compelling reasons that are “internal” to law itself. Law has not been “overcome”; it has “abnegated.” That somewhat odd term suggests a dereliction of duty, most of the time—but Vermeule intends nearly the opposite meaning.
“In area after area,” he writes,” lawyers and judges, working out the logical implications of their principles with a view to rational consistency, have come to the view that administrators should have broad leeway to set policy, to determine facts, to interpret ambiguous statutes, and even . . . to determine the boundaries of the administrators’ own jurisdiction.”
Vermeule dates the beginning of that process to Crowell v. Benson. In “a sweeping attempt to mediate the conflict between law and the administrative state in general terms,” the Supreme Court declared that it would defer to administrative determinations of fact so long as they were supported by what we now call “substantial evidence” (a highly deferential standard). Questions of law, as well as questions of “constitutional” and “jurisdictional” fact—meaning, respectively, whether the question at issue fell into Congress’ delegated jurisdiction or that of an executive branch agency—would remain subject to de novo judicial review.
Over time, this formula proved unstable. It came to be embodied, nominally at least, in the Administrative Procedure Act (APA), which instructs courts to decide “all relevant questions of law” while reviewing agencies’ factual determinations for “substantial evidence.” However, the trajectory of the case law tells a different story. Crowell’s “constitutional fact” category became effectively content-less with the vast expansion of Congress’ powers under the Commerce Clause, beginning in 1937 and completed in 1943. The delegation doctrine that underpinned Crowell and that was articulated more directly in A.L.A. Schechter Poultry Corp. v. United States (1935) was effectively discarded in 1944. And judicial deference soon expanded to “mixed” questions of law and fact. All this happened before the APA was written and enacted.
Whatever gravitational force the APA and its “all relevant questions of law” might have had failed to survive the rulemaking revolution of the 1970s and the Court’s eventual response in Chevron v. NRDC (1984), which famously instructed courts to defer to any “reasonable” agency interpretation of “ambiguous” statutory language. In its 2013 decision in City of Arlington v. FCC, the Court extended deference even to an agency’s determination of the limits of its own jurisdiction—a proposition that would have been unthinkable to the Crowell Court.
How and why did this happen? The formula of Crowell was ambiguous and perhaps “schizophrenic,” Vermeule observes. Compromises of that sort often erode because the lawyers who argue and decide cases value coherence and logical consistency. But why has the arc bent toward deference, rather than backward to legalism? Vermeule’s answer, presented in six of the better pages on American administrative law that I have read in quite some time, is marginalist thinking.
In the paradigmatic administrative law case, a court is asked to examine—re-view, as we say—a decision that has been made by a duly authorized administrator. There are compelling reasons for assigning the initial decision to an agency or commission, starting with the fact that no serious person would want millions of workmen’s or Social Security claims to be determined by Article III courts, let alone Congress. What exactly do we gain, then, from having the decision made again, by independent courts? The Crowell Court, Vermeule argues convincingly, thought in those marginalist terms: If we (the court) review the entire decision de novo, there’s no point to having the commission in the first place.
Agencies, though, are specialists with tunnel vision, and they have an agenda. One could find value in the judges’ more generalist and impartial point of view. Against that, however, stands an array of arguments that point the other way: the sheer cost of running the system; the difficulty of drawing lines between fact and policy and law with any confidence; the monumental scale and complexity of modern-day administration; and (on the conventional understanding of Chevron) the fact that Congress actually wanted those sorts of decisions to be made by agencies.
Moreover, as Vermeule shows in two very powerfully argued chapters, marginalist support for judicial deference gains special force in the context of modern-day risk regulation, and the Supreme Court has spoken accordingly. So long as agencies proffer some colorable defense of their policy choices, courts have declined to intervene. Law has become marginal—both in the colloquial sense of not terribly important, and in the sense of operating as an outer boundary, but no more.
Had Vermeule stopped right there, Law’s Abnegation would be a very fine and important book. For good or ill, though—mostly ill—the author launches into a much more ambitious program, one that is instructive in its own way, but horridly misguided.
The Dworkinian Connection
Law’s Abnegation is about administrative law—but also about much more. In the subtitle (From Law’s Empire to the Administrative State) and in the opening and closing pages, the author rests on, yet also takes aim at, the late Ronald Dworkin’s grand theory that judges and lawyers, prompted by a concerns over legal “integrity,” will generate decisions that promise “fit” with the extant legal texts and precedents and the best (moral) “justification” they can think of. Over time, Dworkin theorized, the process will produce what he famously called “law’s empire.” Not so, says Vermeule. “Fit” and “justification” have produced, and should produce, abnegation. “The arc of the law”—not just administrative law, but law in general—“bends toward deference.” Why did Dworkin fail to see this? He never wrote a word about the administrative state and seemed to be unaware of its existence.
Stipulate that Vermeule is right about the arc of administrative law. It is very difficult to discern a pattern of more and more judicial deference on matters of voting rights, free speech, criminal procedure, or civil rights from gay marriage to guns. Never mind, though. Vermeule dismisses the rights stuff with a half-page shrug to the effect that it occupies only a “tiny subset of [public] issues.” He says the Supreme Court’s highly visible decisions on abortion, gay marriage, and the like hide the larger reality that “the Court has largely retreated from the public sphere, instead policing a subset of free-speech issues and personal liberties, and very occasionally policing government ‘takings’ of property rights.”
The Dworkinian’s answer to this half-hearted demurrer writes itself.
What you call a “tiny subset of issues,” young man, is the whole ballgame. Hercules and I never gave a rip about Administrative Law because that’s just about shuffling rents hither or yon: who cares? Instead, we committed ourselves to Taking Rights Seriously by fabricating the rights that the readers of the New York Review of Books always thought they had—and by prompting the Court, to that end, to lord it over the deplorables. By anyone’s lights, that enterprise has been a rousing success.
There is no point in quarreling over the relative size or importance of the parallel universes of rights imperialism and administrative law abnegation. In plain fact they have expanded in tandem, over the same timeframe. That observation calls for further thought, and perhaps a theory. Vermeule never even tries. That failure, and the obvious implausibility of the broader claims, cloud and undermine an entirely sensible and forceful argument about the trajectory of the administrative state and its law.
Why pick this fight?
The probable answer is that Vermeule wants to borrow Ronald Dworkin’s authority for the proposition that “the law” has an “arc” that bends this way or that entirely for reasons internal to the law itself. With respect to administrative law, that claim is fairly described as suffering from fateful over-determination. More conventional, wholly plausible accounts have shown law’s “abnegation” to be a response to democratic demands; a public-choice-driven process of rent extraction; a reaction to such large-ish events as the Great Depression and World War II; or some combination of those and other economic and political factors, a few of them briefly suggested below.
Vermeule occasionally acknowledges such “external” factors in a half-sentence here or there. By the book’s final chapter, the acknowledgment nearly amounts to saying he didn’t really mean it: “It is not that, or not just that, large-scale forces of economics or politics have put the classical Constitution out of reach.” (Emphasis added.) It’s just that internal dynamics have also played a role, of a magnitude that the reader is left to guess. Still, Vermeule wants us to take law’s internal logic very seriously. His Introduction strongly suggests that law’s abnegation is entirely self-driven; and the insistent claim that law was not “overwhelmed” but “abnegated” gets repeated again and again, with a Tinkerbell-ish conviction that if you say it often enough and put it in italics, it will come true.
What exactly, though, does it mean that “the law” abnegates (or for that matter triumphs) by a logic all its own? No one this side of G.W.F. Hegel believes that “the Law” is an independent actor that works its will and way behind the backs of mortals. Certainly, Ronald Dworkin did not think so. His métier was sophistry, not metaphysics, and Law’s Empire (1986) had real-world princes: judges, and in particular “Herculean” judges who decide hard cases. Vermeule’s “abnegation” is likewise the product of real-world actors: the legal profession. “Judges and lawyers value logical consistency,” he writes. Thus, once administrative agencies showed up and the need for deference over some range was acknowledged, Crowell’s formula gave way to abnegation, for persuasive legal reasons.
The transposition of Dworkinian forms of legal justification from one context to another is a sleight of hand. Dworkin labored mightily to demonstrate that moral considerations of principle are intrinsic to law (for example, because such terms as “cruel and unusual punishment” call for moral judgment). Vermeule makes no attempt to explain why the “marginalist,” essentially pragmatic and functionalist concerns that push toward deference—considerations, mind you, of a kind that Professor Dworkin abhorred—would be “legal” and “internal” in that same sense. Without some such argument, though, the edifice collapses entirely into legal sociology: if practicing lawyers and judges embrace some argument, that makes it “legal.”
Sociology, it turns out, provides a good way of harmonizing the Dworkinian and Vermeulean universes and their divergent arcs. Dworkin’s theory hinged on judges who have to adjudicate “hard cases” by resort to principle where the law runs out or commands moral judgment. That program appeals not only to Justice “Hercules” Kennedy but also to the underpaid district or appellate judge who, amidst a flood of tedious drug and immigration cases, longs to do something meaningful in the rare cases that touch on concerns of an existential nature, most having to do with sexual intercourse.
Legal sociology cuts just the other way in the countless, boring AdLaw cases that turn on the true and correct understanding of “source” or “employee” or the “donning and doffing” of work clothes. In the latter cases, what do the judges know, and why should they care? The better part of wisdom, or at any rate the most economical strategy, is to declare the statute “ambiguous,” invoke “deference,” and in this way obviate the need for further thought in the case at hand while deterring future challenges to agency action.
In contrast to Vermeule’s “Dworkin didn’t get it” riff, this view of the matter makes perfect sense. Unable to maximize their returns in cash, judges maximize their reputational gains on the rights margin and their leisure on the AdLaw margin. That, though, is not Ronald Dworkin’s universe of “integrity”; it is Richard Posner’s view of the world. Its lodestar is not “integrity” but self-interest.
If that sociology seems a bit cynical, consider a complementary story: The arc of administrative law has bent toward deference for the simple and obvious reason that the New Dealers got to appoint their own judges and justices. It has continued to bend that way because the same cast of characters got to run the law schools and educate the next generations of lawyers in their miserable jurisprudence. The sustained force of this corpus juris impresses itself upon anyone who now teaches Federal Courts or Administrative Law (in its more Legal Process-ish version). To be sure, the enterprise does exhibit a certain insularity, if not internal logic: It screams “FDR” at every turn and has no Republicans in it. On this account, administrative law’s arc toward deference simply marks the triumph of a certain legal class. The “legal” arguments that propelled the law on that trajectory were the arguments that that class chose to believe. Whether these were any good is fairly debatable.
Abnegation, the Separation of Powers
Much of Law’s Abnegation, and especially the best parts of the book, tackle core questions of administrative law, such as agency procedure and the scope of judicial review. Vermeule fully recognizes, however, that those doctrines belong in a broader constitutional context. Here, as with administrative law more narrowly conceived, Vermeule mobilizes law’s “internal logic.” “The classical arguments” for constitutionalism and the separation of powers, he writes, “do not survive the abnegation of the law.”
From that starting point, he goes to town against Gary Lawson, Philip Hamburger, and their respective defenses of “classical” constitutionalism and critiques of the contemporary administrative state and its law. Their view that the classical constitution was “crushed by the swelling power of the state” is “wrong in the large and in the details,” Vermeule writes. Lawson and Hamburger “get the law itself wrong.” Constitutional institutions that the defenders of the old order understand as bulwarks of separated powers themselves created the administrative state, and so even Lawson must acknowledge its legitimacy. Likewise, Hamburger “fails to recognize that law’s abnegation . . . flows from law’s own internal logic.” Both of them “want to return to a baseline set of legal arrangements that proved unstable, because the lawyers and judges within those arrangements decided that abnegation had the better of the internal legal arguments.”
What are those “internal” arguments? Why, classical constitutional theory “is haunted by the prospect that some official, somewhere, might commit abuses.” Vermeule views this “abuse prevention” theory of the separation of powers as hopelessly outmoded, for three reasons. First, we have come to realize the excessive costs of strictly minimizing official abuse; the task is to optimize it. Second, we must worry not only about “public” but also “private” abuses of power. The administrative state accepts an increased potential for “public” abuse to enable officials to suppress “private” abuses that arise from the “self-interested behavior of economic actors wielding delegated state power under the rules of common law of property, tort, and contract, and under corporate law.” Third, the classical theory had nothing to say about bureaucracy; now, bureaucracy is everywhere. We have made those adjustments because the world—the “policy environment”—is moving so much faster than in 1787. Or, as Vermuele puts it, “Time is always of the essence, so institutions are forced to trade off the quality of policy against its timeliness.”
Our institutions responded to these factors by abnegating, and the law has done likewise. Vermeule illustrates the point with a discussion of two important administrative law doctrines: the so-called Chenery II principle, which says that agencies may freely choose their procedures (rulemaking or adjudication) unless Congress says otherwise; and so-called Auer deference, under which courts give agencies leeway to interpret their own regulations.
The AdLaw discussion is characteristically lucid and insightful. The broader analysis is wrongheaded and sloppy at every turn.
As an initial matter, it is preposterous to insist that classical theorists sought to “strictly minimize” public abuse. The way to accomplish that is anarchy, which wasn’t high on the Founders’ agenda. (The Federalist is a manifesto for “optimizing” abuse, in precisely Vermeule’s sense. Go read Hamilton on the need for an energetic and unitary executive.) His “private abuse” point boils down to the observation that we have to worry about corporate power—a notion that had occurred to jurists of the 19th century.
The idea that the classical theory had nothing to say about bureaucracy, apart from being empirically false, is upside down: whatever is new about modern bureaucracy is on Vermeule’s own account a consequence of abandoning the old way of thinking. It is disingenuous to mobilize the same thing as a contributing factor.
And the notion that the “rate of change” in the “policy environment” compels abnegation is made of the silly “functionalism” which says that flies are black so that you can spot and swat them on a white wall. It is at least equally respectable to argue that, especially when the rate of social and economic change accelerates, you need a stable legal framework—shelter from the storm, not a mutable government that responds to every stakeholder’s whim and so, paradoxically, ends up making law for the few, not the many. The question of whether one or the other response is more promising is part empirics, part normative, part intuition. You can’t just assert the answer.
Beyond these objections lie more fundamental ones. To repeat, certain demands of modern-day government do make it difficult to defend the classical conception in all its rigor. But what we have here is an abject failure to think on the margin, by a scholar who purports to embrace that mode of thinking. Surely there must be some point at which the tradeoffs are no longer worth making—where the potential for official abuse overwhelms the gains from official discretion; where an obsession with rooting out “private” abuse undermines efficient government and efficient markets; and where an administrative apparatus that responds eagerly and promptly to every press release of this or that policy “community” produces nothing but rents, confusion, and demoralization.
In short, if some efficient frontier marks the tradeoffs between law and the opposing demands of the administrative state, the questions that remain are: where on that curve we are right now; whether we want to move left and up or right and down; and whether we are anywhere near the curve or miles below it.
None of these questions perturb Vermeule. He simply asserts that the policy mix is “optimal” and that “abuses of power” are “the unavoidable by-product of a package solution that is increasingly desirable overall, as the rate of policy change increases.” Why should we believe this? Because he said so. Suck it up.
Surely, though, it does not take a rock-ribbed constitutionalist to wonder whether a tidal wave of unorthodox lawmaking and rulemaking has bought us anything worth having; whether “functional” adjustments to rapid “policy change” may have produced massive dysfunction; or whether effectively unbounded agency discretion is a tradeoff or rather spells the joint maximization of public and private abuse. Questions of this sort are subjects of a rich public and scholarly debate. They lie wholly beyond Vermeule’s purview.
Taking Rights a Bit More Seriously
What’s there in all this for a constitutionalist to take seriously? If Vermeule’s central point about administrative law’s arc is substantially right (as it is), proponents of the rule of law and “classical” constitutionalism must at some point identify, articulate, and defend a durable synthesis between uncompromising formalism (originalism, constitutionalism) and the administrative state.
Acceptance of that challenge will not come easily. To date, the rule-of-law camp has largely contented itself with a mix of high-level constitutional argument (Lawson’s formalism or Hamburger’s deep constitutionalism) and reformist proposals to redress a wide array of abusive practices and misguided doctrines. (Pass the Regulations from the Executive in Need of Scrutiny Act! End Auer deference!) Along the way, we celebrate with mostly unwarranted enthusiasm and wild confirmation bias the occasional judicial decision or opinion that seems to herald the second coming of constitutional government.
This posture is understandable. Constitutional fundamentalism has undeniable attractions, and it is in many ways easier to defend than a more faint-hearted position. Reformist calls for somewhat less judicial deference, a somewhat harder look, and somewhat less delegation have obvious appeal because we have so obviously passed the point of diminishing marginal returns on Vermeule’s optimal-abuse frontier: practically any move in reverse is worth making.
In the end, though, an intellectual program that goes no further will produce disillusionment. We will not demolish the entire administrative state, period. And a pragmatic program for somewhat more legislative control and judicial review will prove inherently unstable, both politically (it may not survive the election of a Republican President and Congress) and in a deeper, Vermeulean sense. The demand for more law, like the demand for deference, operates on the margin. Once the real costs of legalism come into view, marginalism will again push in the opposite direction. Any Goldilocks formula is arbitrary and will crumble in the ordinary course of legal business.
The search for a more stable arrangement would have to start with the proposition that classical constitutionalism, rightly understood, allows a fairly wide berth for the administrative state—provided one keeps the underlying precepts and intuitions in mind. To illustrate the point, it helps to start where Vermeule starts, with Crowell v. Benson.
In extending substantial deference to administrative fact-finders (while reserving questions of law and of constitutional and jurisdictional fact to Article III courts), the Crowell Court did not make an ad hoc, ill-fated concession to the bare necessities of “the administrative state.” Or at least, it did not think it did so. It had a time-honored model in mind: the jury. Juries, too, find facts but not law. They receive judicial instructions on the bounds of their mandate and deference on their determinations unless these are obviously crazy. Thus, just as juries function as “adjuncts” to ordinary courts, so do administrative fact-finders (and for that matter bankruptcy judges and magistrates).
While that view may sound absurd in an age when the federal courts function as adjuncts, partners, and enforcement agents of federal agencies, it was not entirely absurd at the time and in context. Crowell involved adjudication, not rulemaking; and it involved an underlying dispute between two private parties, rather than a dispute between the government and a regulated entity. Pursue the analogy, then, and a Vermeulean thought: Why did marginalist thinking (“What can judges really contribute?”) not also infect the ordinary operation of juries? What does a judge have over the parties’ peers? Why is jury nullification a crackpot cause and deference to administrators the rule?
The most plausible answer has to do with basic intuitions about law and rights. In a civil jury case, one party wins and the other loses; in a criminal case, the defendant will go to jail or get off. At bottom, we think, the answer must be governed by law, not whim or discretion; and the decisionmaker cannot also be the prosecutor. In Marbury v. Madison’s fine phrase, courts cannot “sport away the vested rights of others”; and the court must actually be a court.
Perhaps the arc of administrative law bent because that intuition gradually (actually, rather rapidly) gave way, for reasons good and bad. Foremost, what happens when the dispute is not (as in Crowell) between private parties but between the government and a private claimant? Classical constitutionalism had a straightforward answer: An action against a government official is just like a private, common law action (for trespass, assumpsit, and the like). The difference is that the official has a defense that is not available to private defendants: authority conferred by statute. The question then becomes whether the official’s act was authorized or ultra vires. In this universe, courts do not re-view anything at all: they make a decision in the first instance.
Here’s the catch. The private law set-up, including its ultra vires inquiry, puts the reasonableness of official action entirely beyond judicial purview—not because judges think on the margin but because they think in formalist terms: The law belongs to us, and considerations of convenience belong to the executive. That all-or-nothing model has an internal, constitutionally grounded logic. But it is exceedingly hard to sustain in the face of bureaucracies with vast discretionary powers to do what is reasonable in the bureaucrats’ minds. In that universe, “No reasonableness review” looks like a prescription for rampant abuse. Control administrative discretion, though: instead of making a decision, courts will be re-viewing, on a merits appeal, a decision already made by a duly authorized official. Lo, behold the Crowell conundrum—now you are playing in Vermeule’s universe. “Given the existence of administrative agencies,” he writes, “the judicial role is not decisionmaking, but review.” The entire abnegation story rests on this “appellate review” model.
That is entirely legit. The appellate review model is in fact the foundation of American administrative law. While its adoption was hardly unavoidable, it may be too deeply entrenched to permit any foundational challenge. Still, contemplation of the point suggests that the rule-of-law camp might benefit from reorienting itself from abstract debates over judicial review, administrative reasonableness, and delegation to more concrete questions of private right. To wit, can this official do this to you? The marginalist case for abnegation looks overwhelming in Chevron’s original domain of pre-enforcement challenges to rulemakings. It looks deeply suspect when agencies railroad individuals and sport away their rights.
Recent, important contributions to the literature build on that insight, and a few authors have been sufficiently bold to sketch a program for an administrative state that conforms to constitutional, rights-based precepts and intuitions. Any such program would be the work of decades; but that is the challenge. Law’s Abnegation brings it into focus.
Reading Vermeule can be an exasperating experience. He floats arguments he knows to be over-the-top, while refusing to engage rule-of-law arguments that he knows are there for the taking. Perhaps, though, he just wants someone to make those arguments. If that is his project, more power to him.
 The most ambitious contribution is Robert R. Gasaway and Ashley C. Parrish, “Administrative Law in Flux: An Opportunity for Constitutional Reassessment,” George Mason Law Review 24 (forthcoming in 2017). More on it in a future post.