Symposium: Tampering with the structure of administrative law
on Jan 29, 2019 at 10:23 am
Adrian Vermeule is the Ralph S. Tyler, Jr., Professor of Constitutional Law at Harvard Law School.
In Kisor v. Wilkie, the Supreme Court has agreed to consider whether the doctrine of deference to agency interpretations of their own regulations should be discarded — or more precisely whether Auer v. Robbins and Bowles v. Seminole Rock & Sand Co. should be overruled, the first question in the certiorari petition and the only one the court granted.
In one sense, the stakes are lower than the petitioners and the critics of Auer deference acknowledge. Auer is cabined in many ways, especially by judicial review for clearly erroneous agency interpretations, and by agencies’ obligation to give reasons for their interpretations and for any change in policy that affects reliance interests (constraints recently emphasized by the Supreme Court). The constraints ensure that Auer deference only ever kicks in when agencies are interpreting law and making policy in reasonable ways.
In another sense, however, the stakes are high, because Auer deference hardly stands alone. It is not as though one can have the Auer doctrine, or not, without affecting any other aspects of administrative law, and without any follow-on consequences. Instead Auer is entangled with other basic principles and doctrines of administrative law, in ways that make it exceedingly difficult to isolate. The fabric of administrative law as a subject is unusually dense and interconnected to begin with, but this is especially true of Auer, which is closely intertwined with the fundamental default principle that agencies are presumed to have broad discretion to choose their own procedures and policymaking instruments free of judicial interference, except when statutes and binding regulations specify otherwise.
The upshot is that the Supreme Court should proceed with great caution in this area. Any blithe overruling of Auer would be imprudent in the extreme. It would, I suggest, tamper with the structure of administrative law, risking unpredictable consequences for other doctrines and in future cases. Radical critics of the administrative state would quickly attempt to use the decision in the lower courts to push the law well beyond where most of the justices are willing to go, inevitably causing disruption and instability.
I will emphasize three ways in which Auer deference is entangled with other basic principles — especially in the procedural posture at issue in Kisor, whose facts arguably present the strongest possible case for Auer deference. Because of these entanglements, overruling Auer would cause ramifying, unpredictable consequences throughout the structure of administrative law.
(1) Auer as a corollary of Chenery II. The petitioners in Kisor, at least at the certiorari stage, advance a concern that Auer deference allows and indeed encourages agencies to “circumvent” the restrictions and purposes inherent in notice-and-comment rulemaking. On this view, sometimes called the “self-delegation” argument, agencies have incentives to issue vague, ambiguous or conclusory rules and then clarify the rules through subsequent interpretation, thereby “delegating power to themselves.” The most ambitious version of the self-delegation argument suggests that this behavior not only undermines the notice-giving and deliberative functions of the ordinary rulemaking process, but also combines lawmaking and law-interpreting powers in the same set of hands, in violation of “constitutional principles” of the separation of powers.
Elsewhere, Cass Sunstein and I have argued that insofar as the self-delegation argument embodies a prediction about agency behavior said to result from particular incentives, it is largely evidence-free. And indeed, the latest serious research shows no evidence that agencies more often wrote vague regulations after Auer than before. Let us pause to absorb this: Much of the clamor against Auer has been premised on an empirical claim about agency behavior now shown to lack any discernible factual basis.
Here I will emphasize a different point, one I have also discussed at length elsewhere. There is no such thing as agency self-delegation, for agencies cannot actually expand the boundaries of their own legal authority. For an agency to enact a regulation, vague or specific, it must (barring unusual cases) have statutory authority to do so; when the agency exercises that authority, it is just a confusion to say that the agency has delegated new authority to itself. And courts always retain power to specify the clear boundaries of the agency’s authority. As Justice Antonin Scalia once put it in the context of deference to agency statutory interpretations under Chevron U.S.A. v. Natural Resources Defense Council, in words that apply with equal force in the Auer setting, “[t]he fox-in-the-henhouse syndrome is to be avoided not by establishing an arbitrary and undefinable category of agency decision-making that is accorded no deference, but by taking seriously, and applying rigorously, in all cases, statutory limits on agencies’ authority.” (Scalia was of course the author of Auer; his later doubts, I am arguing, were misguided on his own premises.)
Put differently, it is unclear what legal baseline could support the complaint that Auer allows agencies to “circumvent” the notice-and-comment process. If agencies decide to exercise their authority other than through rulemaking, and otherwise possess legal discretion to do so, there is no “circumvention” at all. They are not “delegating” new authority to themselves, but rather exercising authority they already possess. The ordinary case is that agencies have legal authority to proceed either through rulemaking or through adjudication and interpretation, or to use some mix of these — such as a regulation that supplies some content at Time 1 and adjudication or interpretation that supplies further content at Time 2.
The Supreme Court has repeatedly affirmed and reaffirmed, most famously in Securities and Exchange Commission v. Chenery Corp. (Chenery II) and National Labor Relations Board v. Bell Aerospace Co., that agencies have broad discretion to choose which of these instruments to use, and when to use them. Crucially, this includes the discretion to decide how specifically to regulate at any given time. One basic rationale for this principle is that agencies often have excellent reasons not to use rulemaking to settle all issues at the first opportunity. Leaving policy indeterminate, unspecific or open-ended for some period enables agencies to acquire more information, learn by experience and avoid premature freezing of law and policy.
Auer is best understood as a corollary of this bedrock principle. Auer deference allows agencies — always only within the bounds of reasonable interpretation, and subject to judicial review for clear misreading — to postpone an exercise of interpretive authority from a rulemaking at Time 1 to a later adjudication or other form of interpretation, such as a guidance, at Time 2. Auer thus recognizes in law all the good reasons agencies may have to (sometimes) leave regulations initially under-specified, depending upon circumstances, especially the desire to proceed cautiously through a series of interpretations, so as not to prematurely freeze the law through general binding rules. Auer deference does not affect the ultimate scope of the agency’s power, but rather the timing of the exercise of that power — precisely what core principles of administrative law commit firmly to agency discretion. Overruling Auer, then, would tamper with one of the core structural principles that allocate procedural discretion between agencies and courts.
Petitioners in cases like Kisor typically want to focus not on (formal) adjudication but on interpretation through guidances, which they take to be the best case for their critique of Auer. It is thus extremely awkward for the petitioners in Kisor that, in their own case, the agency interpretation was instituted by means of a relatively formal adjudication, not a guidance. This means that the principle of agency procedural discretion over the choice between rulemaking and adjudication is at maximum force.
For the Supreme Court to overrule Auer in this setting would raise grave questions about the status of Bell Aerospace and Chenery II. It is not at all difficult to imagine that lower courts, wishing to go much further than the court’s majority, would extrapolate the supposed principles of the decision to attempt to undermine Chenery II as well (After all, Justice Robert Jackson’s sulphurous dissent in that case is a sort of ancestor of the most extreme criticisms of the administrative state in our day.). Overruling Auer, particularly in the circumstances of Kisor, would invite instability and adventurism with respect to agency discretion over procedure.
(2) Auer and the combination of functions. Return to the argument that Auer allows agencies to combine lawmaking and law-interpreting functions in a single set of hands. I have already indicated why I think Auer does no such thing. Consider for a moment, however, what havoc an overruling of Auer on these grounds would threaten to cause. It is not as though the Supreme Court has no caselaw on the combination of functions in agencies. There is extensive caselaw arising from direct constitutional challenges to agency structure (usually brought under the rubric of procedural due process, the claim being that the combination of functions creates impermissible bias in the adjudicator). And the court has repeatedly said there is no general constitutional problem about the combination in a single agency of lawmaking, law-enforcement and adjudication. Auer’s critics have never successfully explained how their separation-of-powers criticism is consistent with this basic principle. In reality many of the radical critics make no secret of wanting to undermine that principle as well, implying that, for example, the structure of the Securities and Exchange Commission, which combines lawmaking, law-enforcement and adjudicative functions, is constitutionally indefensible. Yet even the justices skeptical of Auer, except perhaps for Justice Clarence Thomas, show no signs of wanting to go so far. Overruling Auer on this ground would destabilize a long-settled principle supported by a large supermajority of the justices.
(3) Auer and stare decisis. Critics of Auer, and the petitioners in Kisor, are quick to argue that precedents on the allocation of interpretive authority between agencies and courts should receive less stare decisis force than do substantive judicial interpretations of statutes or regulations. They had better argue so; after all, the foundational case for Auer, Bowles v. Seminole Rock, was decided in 1945, and the doctrine is embodied in literally thousands of decisions and relied upon daily by courts and agencies. Although there is an interesting argument that Bowles originally rested on different presuppositions than Auer, so that Auer misunderstood Bowles, that argument is nearly irrelevant today. Countless exercises of agency rulemaking, congressional lawmaking and judicial decision-making in the ensuing decades have proceeded against the background of a settled practice of deference to agency interpretations of their own rules. Under ordinary stare decisis standards, if they apply, the argument for overruling Auer and Bowles would be a nonstarter. Therefore, the critics need an argument, any argument, that those standards don’t apply.
But this is largely uncharted territory, which implies uncertainty and risk for the Supreme Court. Any such exception to stare decisis would be ill-defined and highly susceptible to manipulation and expansion in later cases, with unpredictable consequences in a large and indefinite range of other areas. Surely the stare decisis force of Chevron would be equally affected, for similar reasons; and although a few of the justices most critical of Auer would also be happy to overrule Chevron, others almost certainly disagree. But Chevron is just the beginning of the knock-on consequences. What about the foundational precedents on agency decision-making procedure, such as SEC v. Chenery Corp. (Chenery I) (agency decisions stand or fall on the reasons agencies themselves gave in the administrative proceedings) and Accardi v. Shaughnessy (agencies must follow their own rules) — are they vulnerable on the same grounds? After all, the statutory basis for these fundamental principles is obscure at best. Are all these now to be deemed to have lessened stare decisis effect? If so, the prospect of a revolution looms. Note that, in the two cases I just mentioned, overturning precedent would actually expand the discretion of agencies, so even the justices who want to shackle the administrative state ought to be wary of destabilizing the law. Cooking up a new approach to precedent yields a toxic brew that can be harmful even to its creators.
In the three settings I have mentioned, overruling Auer creates palpable risks. One cautious disposition — probably the path of wisdom — would be to reaffirm Auer on the facts of the case actually before the Supreme Court, in which the agency interprets its own regulation in a (relatively formal) adjudicative proceeding, and to leave for another day questions about the reach and applicability of Auer in cases in which the agency interprets its regulation (solely) through guidances — the context that most concerns Auer’s critics. This implies that Kisor presents the strongest context for reaffirming Auer and is an unpromising case for overruling the doctrine altogether. If Kisor was, all along, a poor vehicle for raising the large question whether Auer should be overruled — well, the U.S. solicitor general did point that out.
* * *
Past cases linked to in this post:
Accardi v. Shaughnessy, 347 U.S. 260 (1954)
Auer v. Robbins, 519 U.S. 452 (1997)
Bowles v. Seminole Rock Co., 325 U.S. 410 (1945)
Chevron U.S.A. v. Natural Res. Def. Council, 467 U.S. 837 (1984)
National Labor Relations Board v. Bell Aerospace Co., 416 U.S. 267 (1974)
Securities Comm’n v. Chenery Corp., 318 U.S. 80 (1943)
Securities Comm’n v. Chenery Corp., 332 U.S. 194 (1947)