The central concern of administrative law is how to control agency discretion. Agencies are handed enormous authority, and administrative law consists primarily – indeed, almost exclusively – of a set of doctrines designed to inform, curb, or enable other actors to oversee discretionary agency actions. Administrative law is preoccupied with establishing procedures to prevent agency abuse and designing oversight by non-agency players – the President, Congress, private stakeholders, and, most obviously, the judiciary. All the core doctrines of administrative law are generally understood as implementing basic decisions regarding institutional choice: who does what? How should power be divided up amongst these institutions?
In Allocating Powers Within Agencies, Elizabeth Magill and Adrian Vermeule convincingly argue that in operation core administrative law doctrines are not only about institutional choice but also institutional design. That is, they do not merely allocate authority between agencies and other actors; they also have important consequences for who does what within agencies, for how the institution is designed. The paper elegantly reviews a number of familiar doctrines and explains the impact they have on how power is allocated within agencies. The article’s title seems to portend a discussion of how to go about constructing agency organizational charts. In fact, the article is not about such conscious allocation of responsibility at all. Rather, it explains how doctrines established without consideration for their impact on internal agency operations do in fact significantly affect how power is allocated within agencies. It thus makes an explicit and implicit plea that these impacts be thought through rather than incidental and haphazard.
The important, inarguable, but often overlooked starting point is that, to adapt Kenneth Shepsle’s famous article title about Congress, “an agency is a they, not an it.” Judicial opinions and much scholarship tend to speak of “the agency,” as if it were a unitary entity. But, as with the executive branch as a whole, “unitariness,” whatever its normative appeal, is hard to come by in the real world, where critical decisions result from the involvement of political and career appointees, of high-ranking and low-ranking staff, of lawyers, scientists, economists, technical experts, and public affairs departments. There is an existing literature on conflicts and collaborations between individuals wearing different hats within agencies. But Magill and Vermeule are (almost) the first to draw the link between administrative law doctrines that are always seen as involving institutional choice. (I say almost, because there is at least one previous example of such an effort, duly credited by the authors. That is Don Elliott’s article explaining how Chevron has made the lawyers at the Environmental Protection Agency relatively less important and the policy and technical people relatively more so. E. Donald Elliott, Chevron Matters: How the Chevron Doctrine Redefined the Roles of Congress, Courts, and Agencies in Environmental Law, 16 Vill. Envtl. L.J. 1 (2005).)
So what are some examples? One set of doctrines concerns, of course, judicial review. For example, take the Chenery principle that agency action can be upheld only on the basis of the rationale offered by the agency at the time it took the challenged action – no post hoc rationalizations. The Chenery principle actually has little bearing on the allocation of authority among government institutions; it comes into play in settings when the agency has authority to act, but will be allowed to do so only if it goes about it in the “right way,” which means getting the explanation right before the agency finds itself in court. “So understood,” write Magill and Vermeule, “Chenery’s crucial effect is to reallocate power horizontally within agencies, away from the lawyers who rationalize policy after the fact and to agency personnel who formulate policy before the fact.”
Chenery is not alone. Chevron operates in similar fashion; that was Don Elliott’s point. This in turn means that the stakes in “Chevron step zero,” the yes/no decision controlled by United States v. Mead Co., 533 U.S. 218 (2001), are not merely about allocating authority between courts and agencies (the institutional choice question), but about allocating authority between different professionals within agencies (the institutional design question). Hard look review has more complicated tendencies, in some settings tending to lead agencies to emphasize the scientific character of their analysis, in others empowering lawyers at the expense of scientists and other policy experts. Magill and Vermeule’s analysis is not limited to doctrines of judicial review; they also consider doctrines concerning structure and process, such as executive-created agencies, OIRA review, and separation of functions.
The foregoing concerned horizontal allocation of functions – i.e., the question of which professionals will dominate agency decisionmaking at any given level of the agency hierarchy. Magill and Vermeule offer a similar analysis regarding vertical allocation. For example, an ongoing debate over Mead concerns when it is exactly that an agency will be deemed to have acted with the force of law, as required to trigger Chevron. One school of thought, embraced, though in somewhat different fashion, by Justice Scalia and by David Barron and Elena Kagan, holds that deference is appropriate when the agency interpretation is sufficiently authoritative, or comes from a sufficiently high-ranking official. While this approach has not been embraced by a Supreme Court majority, it would be a stark example of how a doctrine generally conceived as involving allocating authority between institutions (here, agencies and courts), would have a direct impact on the allocation of authority within agencies. The authors also explore consequences for the level at which agency decisions are made of the principle that agencies are bound by their own rules, ALJ independence, and other doctrines.
The distinction between “institutional design” and “institutional choice” is somewhat slippery in this setting; it raises a classic level of generality issue here. If the institution is “the government of the United States,” then we have only problems of institutional design (though federalism would then raise problems of institutional choice); if we label offices within agencies “institutions,” then we have only problems of institutional choice. But that terminological point is hardly central. Magill and Vermeule’s essential argument — that doctrines generally understood to be about the power of agencies within the government have important consequences for how power is allocated within agencies – is important and novel. While the authors’ observations are preliminary and not empirically grounded, they trigger a new way of thinking about old problems, and point to an important and almost completely neglected field of inquiry.