A Concluding Post on Federal Agency Guidance and the Power to Bind, by Nicholas R. Parrillo
Over the last few weeks, Notice and Comment has hosted a symposium with contributions from fourteen scholars on the binding power of federal agency guidance, using as a focal point my article “Federal Agency Guidance and the Power to Bind,” recently published in the Yale Journal on Regulation. The contributions, many of which draw from (and link to) major prior publications by their respective authors, delve into several of the most important and difficult issues on this controversial subject. Collectively, they map the state of play in the discourse and debate. I believe they can (and hope they will) serve as an introduction to the subject for the interested scholar, student, or practitioner.
In this concluding post to the symposium, I seek — for the benefit of future readers — to draw together the fourteen contributions, grouping them around different themes and putting them in dialogue with each other. In doing so, I cannot fully summarize each contribution; hopefully I will capture at least one key point of each. Also, with only one or two exceptions, I will not attempt to respond to the contributions substantively. Given the depth and complexity of what the authors have said (and the prior research on which it often draws), a conscientious response to all would take another long article, not a blog post!
What Is Binding Impact? And Is It the Thing to Worry About?
In my article (summarized in this post), I sought to document the incentives regulated parties face to comply with guidance when it is operative. When these incentives are strong (as they often are), their impact can nonetheless be mitigated if officials keep their minds open to arguments why, in a given case, the guidance ought not to be operative — that is, if officials treat the document flexibly. But I also sought to document the pressures that officials feel not to be flexible — pressures from competing regulated parties, NGOs, the media, and Congress that often spring from rule-of-law values favoring consistency and predictability, ones that agencies would be remiss to simply ignore.
Some of the contributors see these values of consistency and predictability as having primacy, such that the widespread concern with guidance’s binding impact is essentially misplaced, or at least exaggerated. These scholars are elaborating general normative arguments to back up the more situational demands for consistency that my interviewees make and hear each day in the trenches. Peter Strauss makes this case across the board, extending his classic work on the subject. Jill Family makes the case in the specific area of immigration law, where rule-of-law concerns manifest somewhat differently than in the business-regulatory contexts on which I focused: concerns about economic competition among regulated parties seem not at issue, while “restrictions on fundamental life decisions” are at the fore, implicating the rule of law as a shield for individual rights and individual dignity.
If one is convinced that we should privilege consistency and predictability, then the courts’ predominant approach to guidance — striking it down simply because it binds — seems wrong. Consistent with Strauss, William Funk and Mark Seidenfeld argue that courts ought not review guidance for its binding impact, but for whether it is defensible in substance, e.g., for consistency with the relevant statutes and regulations. The concern about binding impact, says Seidenfeld, is satisfied by the principle that guidance may not have “independent force of law,” that is, the agency “must show that the violation of guidance in any particular circumstance permits the agency to act against the [regulated] entity in accordance with the agency’s authorizing statute and legislative rules.” Similar, Funk invokes the core reasoning of the D.C. Circuit’s ur-precedent Pacific Gas & Electric Co. v. FPC that an agency cannot rely upon a policy statement in an adjudication but “must be prepared to support the policy just as if the policy statement had never been issued.” According to Funk, the D.C. Circuit’s wrong turn came later, when it elevated some other fleeting language in Pacific Gas (about the need to be “tentative”) into a controlling principle. And Funk notes that the Supreme Court hasn’t endorsed the D.C. Circuit’s demand for tentativeness and could still reject it. Meanwhile, Michael Asimow argues for a congressional fix rather than a judicial one, but the model he offers — the Model State Administrative Procedure Act of 2010 — prohibits the agency from using guidance “to foreclose consideration of issues” in any administrative proceeding, which seems resonant with the core approach of Pacific Gas (and would obviate the need for a broader inquiry into de facto binding effect, an endeavor whose “futility” Asimow argues is proven by the stakeholder demands for consistency that I document).
These contributions raise the interesting question of what it practically means to keep an agency from relying on guidance as an independent basis for decision in particular proceedings. In particular, there’s the question of whether the agency’s obligation to build its case from the ground up in particular proceedings provides sufficient protection for interested parties other than the regulated entities subject to those proceedings. Richard Pierce, recalling his direct involvement in Pacific Gas, describes how GM — a third party dealing with regulated pipelines whose allocations of gas were covered by the contested policy statement — failed on a facial challenge but then prevailed by intervening in every adjudication where the agency sought to apply the guidance to a particular pipeline. Of course this is the exception that proves the rule: guidance normally will not face an opponent as formidable as GM circa 1974. As scholars like Nina Mendelson and David Franklin have emphasized, confining review of guidance to individual applications may do little to protect parties other than regulated entities including regulatory beneficiaries, and accordingly skeptics of the binding-effect test that is the current basis for facial challenges have suggested the need to make substantive challenges available to such beneficiaries somehow (as Strauss says in his contribution, if an NGO has “credibly established a present threat to members’ health”).
Given the legitimate demands for consistency that agencies face from regulated parties and other stakeholders, I argue that if an agency is to avoid rigidly following a document that hasn’t gone through notice and comment, its best solution is to engage in principled flexibility: “for each departure the agency makes [from guidance], it gives a written explanation that is accessible to other agency officials and to the public, with the understanding that the exception then becomes generally applicable to like cases prospectively” (p. 181). I recognize that the provision of explanations eats up agency resources, meaning that this solution cannot be implemented everywhere all the time, and I suggest priorities for when to expend the time and money necessary for it.
Three contributors accept my approach in principle while interrogating where and how it can be implemented. At the outset, there is the question of whether principled flexibility (or any flexibility, for that matter) is an appropriate goal for all types of guidance. My study focused on whatever guidance interviewees considered to be legally nonbinding. This definitely included guidance about the agency’s use its discretion (policy statements). And it included guidance about the agency’s interpretation of law (interpretive rules) insofar as interviewees thought such guidance should be nonbinding, though not all interviewees (or courts) agree on whether it should. ACUS, in adopting Recommendation 2017-5 (which includes principled flexibility, among other things), confined it to policy statements. In his contribution, Ron Levin gives an update on an ACUS project now underway to determine the Recommendation’s applicability to interpretive rules. In a study conducted with Blake Emerson, he finds that agency officials generally do think flexibility is appropriate in the use of interpretive guidance, though interestingly some think it shouldn’t be provided in individual applications but instead through proceedings for wholesale revision of the guidance (an approach that I would say takes principled flexibility to its logical extreme).
Kristin Hickman asks whether and why agencies will be motivated to engage in principled flexibility, which she thinks a “sensible solution to the guidance problem.” She recognizes that some agencies may have reasons of their own to try it. But “a little judicial review goes a long way,” and she suggests that reviewing courts could spur agencies to take this approach, particularly by accepting the Supreme Court’s recent signals to lower the finality hurdle to review of guidance and by understanding principled flexibility as an application of the familiar APA mandate that agency action be reasoned and not arbitrary.
Sean Griffith points out that my discussion of principled flexibility focuses on the supply side but not the demand side. I take as given that regulated parties will ask for departures from guidance, and I analyze how officials do and should react to those requests. Griffith asks about the demand side: will regulated parties make requests? He explains how government enforcement settlements against firms have pressed them to hire increasing numbers of compliance officers, whose divided loyalty to shareholders and regulators may cause them not to “fully inform their firms of . . . the potential to obtain an exemption [from guidance] if they fear that doing so might lead to less compliance” or would threaten their own “base of authority within the firm.” To work, principled flexibility may require reforms not only within the agency but also within the firm.
Processes for the Formulation and Issuance of Guidance
It’s possible that, even if firms demand departures from guidance, agencies will usually find it hard to engage in principled flexibility simply due to resource constraints. If that is true, then rigidity in the application of much guidance should perhaps be taken as given, and the second-best solution may be to provide more opportunity for regulated parties to participate in the agency’s formulation of guidance when it is issued, at a wholesale level. Indeed, participation at the time of issuance may be the only practical way to protect the interests of third parties (such regulatory beneficiaries), who, unlike Pierce’s GM, will not usually have the resources to challenge guidance’s application in individual proceedings.
In this vein, Jamie Conrad draws upon my research to argue “that agencies should be required to provide notice and an opportunity for comment on drafts of sufficiently momentous guidance documents.” He notes my finding that existing examples of “notice and comment lite” on guidance don’t usually involve nearly the same process burden as legislative rulemaking, largely because judicial review and OMB review are much lighter to nonexistent. Of course, as Conrad acknowledges, the “most difficult issue to resolve would be how to formulate the threshold below which notice and comment would not be required.” In my ACUS report, I was reluctant to recommend anything but a very high threshold. That was partly because of my concern about ossifying the guidance-making process, based to a large degree on my observations of OSHA, though Conrad provides some very interesting evidence that OSHA’s experience is sui generis and that its avoidance of process for issuing guidance has recently been counter-productive even from its own perspective.
That said, Conrad’s proposal would increase the cost of adopting guidance at least somewhat, and as Stuart Shapiro points out, there are already other factors in play that are also increasing those costs, namely OMB’s new treatment of guidance under the Congressional Review Act and the Supreme Court’s possible rollback of Auer deference. Shapiro cautions that as guidance’s costs increase, agencies will substitute other methods of policymaking, and while these may sometimes be notice-and-comment rules, they can just as easily be less-transparent means like case-by-case adjudication or strategic dissemination of information.
Expanding Our Understanding of Guidance’s Power
Evaluating my assessment of the reasons why regulated parties follow guidance, Ming Hsu Chen points out that I see these mainly through an instrumental lens, for example, I discuss firms’ need to maintain a good relationship with the agency as an investment in trust and goodwill that can be “cashed in” when the firm inevitably fails to comply on some issue. Chen argues that a full understanding of guidance’s impact must also consider “value-driven, non-instrumental incentives for cooperation” with agencies. Relatedly, Sam Halabi emphasizes the heterogeneity of guidance and in particular asks whether the category should be understood to include “scholarly publications” by agency officials that provide “academic information” independent of the official power of the authors. Reading these contributions, I’m reminded of Daniel Carpenter’s book on the FDA, in which he argues that an agency’s power consists not merely of enforcement authority but also includes “conceptual power,” that is, “the ability to shape the content and structure of human cognition itself,” something FDA has done to how people generally think about drug development (pp. 15-17).
I agree that persons and firms who conform their behavior to guidance do so partly for reasons that aren’t reducible to self-interested instrumental thinking. My choice of a relatively instrumental interpretation (which, as Chen notes, is mutually constituted with my choice of which agencies to study) had to do with the concept of “coercion,” which has long loomed over this subject. People endlessly complain about guidance being coercive, but often in a frustratingly vague manner, and I therefore wanted to trace allegedly coercive interactions down to their micro-foundations, in as concrete a way as possible. My focus on instrumental considerations (and on agencies in which those considerations tend to be prominent) was a product of my interest in interactions that I thought people would tend to label “coercive.” Although coercion is a notoriously contested concept, it seems to me that it’s easiest to say behavior is coerced when the thinking behind it is instrumental: “I don’t want X to happen to me, and to avoid it, I will follow this guidance.” By contrast, I think there’s less of a tendency to say behavior is coerced (at least in our political culture) when it arises partly from the actor’s own values or the actor’s own factual perceptions, as when the actor says: “I believe Y is right, and that doing Z will further Y, so I will do Z,” in a scenario where the agency’s guidance informs the actor’s understanding that Y is right, or that Z as a factual matter furthers Y. But while I considered an instrumental focus to be most relevant to interactions conventionally labeled as coercive — and “coercion” to be the controversy demanding most immediate attention — there is no question that guidance’s less instrumental and more value-informing or perception-shaping properties deserve more attention in the future of this debate.
I’m delighted that this group of scholars has convened here for this exchange. It has been great to be part of it. Thanks are owed to several people: Richard Murphy, who selected and invited the contributors; Chris Walker, who originally hatched the idea for the symposium; and Susan Wang and the editors of JREG, who made it all happen.
Nicholas R. Parrillo is a professor of law at Yale and a public member of the Administrative Conference of the United States.
This post is the final installment of a symposium on federal agency guidance. The rest of the posts in this symposium can be viewed here.