Rethinking the Role of Agencies in Private Regulatory Enforcement

David Freeman Engstrom, Agencies as Litigation Gatekeepers, 123 Yale L. J. 616 (2013).

Over the past several decades, many scholars have weighed in on benefits and detriments of authorizing private parties to sue to enforce federal regulatory standards. They often take either of two opposing positions: Some argue that private enforcement is necessary to supplement underfunded and perhaps captured agency enforcement mechanisms; others contend that private enforcement undermines social welfare or even statutory goals by sacrificing officials’ prosecutorial discretion not to pursue cases that, while technically justified, would not further regulatory goals. Few scholars, however, have written about the trade-offs triggered by a choice between public and private enforcement.

In Agencies as Litigation Gatekeepers, David Engstrom views the issue as one of when and how agencies should control the use of private enforcement. He is not the first to write about vesting agencies with such gatekeeper functions. But, others who have written on the subject generally have done so within the context of a particular regulatory program or litigation regime. Agencies as Litigation Gatekeepers views the structure and control of private enforcement as a unique kind of regulatory problem that extends potentially to every regulatory program. Doing so allows the article to develop some theoretical insights into how private enforcement might be structured and how agencies might best further the use of private enforcement mechanisms.

Engstrom begins the article simply enough by reviewing the potential benefits and detriments of private enforcement. He then presents arguments why the blunt instruments of litigation reform are unlikely to mitigate the pitfalls of private enforcement without sacrificing many of its benefits. Based on these arguments, he concludes that “vesting administrative agencies with litigation ‘gatekeeper’ powers” is a preferable alternative. (P. 616.) This conclusion then frames the remainder of the article, in which Engstrom lays out a “taxonomy” of how agency litigation gatekeeping might be structured, and evaluates the potential benefits and problems of the various categories in his taxonomy.

Engstrom identifies five “dimensions” along which agency gatekeeping can be oriented. The first characterizes agency gatekeeping as “affirmative” versus “residual,” by which Engstrom means whether the agency will have explicit power to control private litigation, as opposed to mere exercise of residual powers such as intervention in a private enforcement action as an interested party or amicus. The second distinguishes between “retail” (i.e. case by case) and “wholesale” (across an entire program or class of private actions within a program) gatekeeping. The third asks whether the agency power is “binding” or “advisory,” the latter referring to an agency determination of the propriety of a private action that only has power to persuade the courts. The fourth addresses whether agency control is “passive” as opposed to “active,” essentially hinging on whether the agency can simply decree whether the private enforcement action may proceed without substituting itself actively in the action. Finally, the fifth dimension, which applies only to affirmative, binding gatekeeping, differentiates between agency “veto” and “license,” essentially based on the default ability of private entities to bring enforcement actions if the agency does not make an explicit determination.

To evaluate the different gatekeeper structures that result from various regions within Engstrom’s five gatekeeper dimensions, he reviews the functions the agency would need to perform to play its gatekeeper role in an ideal manner, and then discusses how some well-known imperfections in administrative processes—limits on institutional competence, regulatory capture, and undue political influence—might cause deviations from ideal gatekeeping. In doing so, Engstrom reveals an impressive knowledge of actual experiences in myriad contexts in which private regulatory enforcement plays a role. Engstrom’s taxonomy is too rich to allow me to even to summarize his entire analysis of how these imperfections might bear on litigation gatekeeping, but it is worth mentioning a few of his arguments and conclusions.

Engstrom finds wholesale gatekeeping less susceptible to regulatory pathologies than retail gatekeeping. He notes that wholesale gatekeeping is not particularly useful for enhancing the efficiency of private enforcement by winnowing out cases that are unlikely to prevail in court. But it is potentially valuable towards the ends of ensuring that private enforcement does not undermine fundamental statutory goals or compromise the public interest—ends to which, he contends, agency expertise and political accountability are well suited. He further concludes that capture, if it occurs at all, is not likely as great a problem for wholesale as opposed to retail gatekeeping. Engstrom does not discuss the problems of overly politicized wholesale gatekeeping, but given that such gatekeeping usually will be more salient and more likely to require the agency to resort to notice and comment rulemaking, transparency and accountability are likely to be greater for wholesale than retail gatekeeping. The downside of wholesale gatekeeping is that it is relatively indiscriminate in allowing an agency to differentiate meritorious claims from perverse ones (whatever metric one uses to evaluate merit). Overall, I read Agencies as Litigation Gatekeepers to strongly support wholesale gatekeeping in those contexts where the costs created by private actions are unlikely to vary with the particular factual circumstances of the case. Its view of the value of retail gatekeeping is much more guarded and dependent on the facts of any particular private enforcement suit.

The most significant value of Agencies as Litigation Gatekeepers, however, is not any precise prescriptions one might derive from its analysis. Rather, its value is in reconceptualizing the concerns raised by private enforcement as squarely falling within the ambit of matters that agencies might address. By its nature, the article is very general in its analysis, but it is certain to generate an interesting literature that applies, evaluates and expands on his taxonomy in particular regulatory contexts.


Viewing the Arbitrary and Capricious Test as a Set of Function-Specific Criteria

Louis J. Virelli III, Deconstructing Arbitrary and Capricious Review, 92 N.C.L. Rev. (forthcoming, 2014), available at SSRN.

The Administrative Procedure Act’s “arbitrary and capricious” standard has been a source of power for the courts, but also a source of bewilderment. It is a source of power because it provides courts with the authority to set aside agency action and, in particular, agency rulemaking, perhaps the most important and characteristic tool of regulatory governance. It is a source of bewilderment because its defining terms are enigmatic. Fairly early in its history, the D.C. Court of Appeals interpreted it as requiring courts to take a “hard look” at the agency’s action. Despite this formulation’s popularity, it has failed to dispel the mystery, first because it is excessively metaphorical, but even more seriously because it is deeply ambiguous. Does it mean that the court must take a hard look at the way the agency reached its decision (a procedural hard look), or rather that the court is insisting that the agency take a hard look at the evidence and arguments being presented to it (a substantive hard look)? The Supreme Court’s decision in Motor Vehicle Manufacturers Assoc. v. State Farm Ins. became the leading decision on the subject because it parsed the substantive hard look standard, providing at least some operationally defined criteria by which the agency’s application of the evidence can be assessed.

Given the importance and ambiguity of the arbitrary and capricious test, it is hardly surprising that the scholarly literature on the subject is voluminous. One approach that commonly appears is the effort to articulate a single test or standard that would enable courts to determine whether an agency decision is arbitrary or capricious. In this innovative and insightful article, Louis Virelli adopts the opposite approach. His idea is to multiply the number of considerations that the arbitrary and capricious test includes, combining both substantive and procedural standards. The point of this proliferation is not to make judicial review more demanding; he agrees with the prevailing view that the agency is the principal decision maker in our system and is entitled to considerable deference from the judiciary. But he argues that the administrative decision-making process necessarily consists of various discrete, qualitatively different steps, and that the standard for arbitrary and capricious action should vary in accordance. Thus, the hard look doctrine should be viewed as “a collection of more targeted inquiries into specific aspects of agency action.”

The article divides the various stages in the agency’s decision making process into two main categories; first, the agency’s “modes of self-education and information gathering,” and second, the conclusions that the agency reaches on the basis of the information that it has obtained. It proposes that judicial review of each stage, which is labeled simply as first and second order review, should be distinguished. In essence, this means that the court will engage in both a procedural hard look at the way the agency reached its decision and a substantive hard look at the quality of its decision making. But Virelli is not simply suggesting that courts engage in each type of review, rather than choosing between them. Rather, he disaggregates each of these large and familiar categories into separate components, and it is these components, not the general categories, to which his task-specific tests apply. The first and second order, or procedural and substantive categories, which he readily concedes are “inexact,” serve largely as a means of organizing the more individualized criteria.

Within the first category, the specific functions are for the agency to define the scope of its research, to gather information within that ambit, to build a record and to give reasons for the conclusions it reaches on the basis of that record. In each area, it has a great deal of discretion, being the institution the legislature has chosen to implement its statute and being a repository of substantive expertise. But the agency might fail in carrying out each of its functions. It might define the scope of research too narrowly, omitting relevant and important issues. It might rely on poor quality information, with obvious methodological errors or inadequate quantities of data. It might fail to document whatever information it has obtained, and it might fail to explain how it used that information to arrive at its conclusions. Within the second category, the functions that agency is expected to perform are to consider relevant factors and to establish a rational connection between the data gathered and the conclusions reached. Here too, the agency has wide discretion, and here too, it can abuse that discretion. It might overlook issues that it needs to consider, either by statutory command or good decision-making practice, or it might engage in slovenly or defective analysis of the information it has gathered. After specifying the way that the arbitrary and capricious standard can be given meaning in the context of each of these procedural and substantive functions, Virelli goes on to elaborate on the advantages of his approach and, to his credit, a few of the disadvantages as well.

A number of observers have criticized legal scholarship in general, and administrative law scholarship in particular, for being “juriscentric,” that is, for focusing excessively on judicial decision making and ignoring or under-emphasizing the legislative and administrative realms where most of modern governance occurs. Virelli’s article is hardly exempt from that criticism; it is a piece about judicial review, not about, say, the way agencies resolve scientific uncertainties. But one of the many virtues of the article is that it takes a non-juriscentric approach to judicial review. Instead of focusing on the reviewing court’s decision-making process—what are its competencies, which sorts of information it needs, how should it articulate the standard that it is applying—the article focuses on the decision-making process in the agency. The standard of review, it suggests, should be determined by the task the agency is carrying out, and should be specific to that task. In other words, instead of deriving the content of the arbitrary and capricious test from judicial precedent, from the conceptual framework and long established standards of judicial review, the article attempt to derive this content from the administrative process itself. What is good practice, it inquires, at each stage in this process. To what standard would a conscientious agency hold itself if it were conducting its own internal review of its own decision making.

Is this article the golden key that will permanently unlock the mysteries of the arbitrary and capricious doctrine? Of course not. There is no such key because the problem is too large to yield to a single resolution. The question that judicial review of agency action confronts is the effectiveness and fairness of the basic means by which we, as a collectivity, govern ourselves. Modern government is regulatory in its essence. In seeking standards for judicial review of agency action, we are asking—albeit in new ways and in a contemporary setting—the age-old question: what is good government and what is bad government. This article, needless to say, has not settled that question, but it advances and clarifies the inquiry in a creative and insightful way, which is why I like it lots.


Judge Wald and Justice Scalia Dance the Chevron Two-Step

Gary Lawson & Stephen Kam, Making Law Out of Nothing at All: The Origins of the Chevron Doctrine, 65 Admin. L. Rev. 1 (2013), available at BePress.

If you are teaching administrative law this semester, you can look forward to a riveting discussion of Chevron. There have been volumes written on this topic, and here I plead guilty. But if you will indulge me for a moment, I’d like to recommend that you read one more article about Chevron.

Professor Gary Lawson and former student Stephen Kam collaborated to write Making Law Out of Nothing At All: The Origins of the Chevron Doctrine. Their mission is to explain why the Chevron decision is irrelevant to the Chevron doctrine. They write not to praise or criticize the case, but “to bury it.” Or, to put it another way, they explain why one cannot resolve the many questions relating to the Chevron doctrine by examining the Chevron decision. In their article, Lawson and Kam elucidate how the lower courts, particularly the D.C. Court of Appeals, transformed the unrevolutionary Chevron case into the revolutionary Chevron doctrine within just two short years.

The authors begin by explaining the state of the law pre-Chevron. Specifically, they contend that the world of administrative law deference was broken into two categories: (1) agency decisions involving pure questions of law; and (2) agency decisions involving the application of law to fact. Courts reviewed decisions involving pure questions of law de novo, while they reviewed decisions involving the application of law to fact with some deference, although the exact level of deference was unclear. Further, these standards were not absolute, but rather only rebuttable presumptions.

Lawson and Kam digress briefly to describe and distinguish legal deference from epistemological deference. Legal deference is deference owed to someone simply because of the identity of the entity making the decision; think of parents or a board of directors or juries. Chevron deference is a form of legal deference. In contrast, epistemological deference is deference earned from experience and wisdom; think of a mentor or a consultant to a corporation or an expert witness. Skidmore deference is a form of epistemological deference. These categories, which the authors created, may help you understand and explain the difference in how the doctrines should be applied and why.

The pre-Chevron deference world was not a simple one; hence, a clearer two-step deference framework would have been appealing to the lower court judges, especially those on the D. C. Circuit who routinely faced these issues. Along comes Chevron, a case that no one expected to alter this deference framework; as we know, Justice Stevens, Chevron’s author, intended to do no more than restate existing law. You don’t need me to explain either the decision or the doctrine, but what happened after Chevron was decided merits explanation, as it illustrates the development of the Chevron doctrine. First, in General Motors Corp. v. Ruckelshaus, 742 F.2d 1561 (D.C. Cir. 1984), Judge Patricia Wald, a friend of administrative law, ignored the pre-Chevron categorization step and framed Chevron as a “scope-of-review doctrine.” Next came the first clear delineation of Chevron as a two-step process in Reting v. Pension Benefit Guaranty Corp., 744 F.2d 133 (D.C. Cir. 1984). It might come as no surprise that Judge Wald authored that opinion as well. Then, the D.C. Circuit issued two more Chevron opinions authored by, yes you guessed it, Judge Wald. In Railway Labor Executives’ Ass’n v. United States Railroad Retirement Board, 749 F2d. 856 (D.C. Cir. 1984), Judge Wald articulated and applied the two step-analysis she had delineated in Reting; however, in Pennsylvania Public Utility Commission v. United States, 749 F.2d 841 (D.C. Cir. 1984), she did not refer to Chevron despite resolving what she called “substantial issues of statutory interpretation.” Id. at 849. As Lawson and Kam note, “By the end of 1984, the D.C. Circuit … was applying the Chevron two-step episodically at best. Even the judge who birthed the Chevron doctrine was not applying it consistently.” (P. 50.)

But the schizophrenia did not linger. By Chevron’s first anniversary, Judge Wald’s Chevron two-step had firmly taken root in the circuits, although it had not yet migrated to the Supreme Court. However, in 1986, three judges from the D.C. Circuit authored law review articles identifying the Chevron decision as revolutionary. One of these judges, Antonin Scalia, ascended to the Supreme Court on September 26, 1986. Justice Scalia brought with him expertise in administrative law, as well as an understanding of Chevron’s significance: “The more Chevron mandates deference, the more power flows from the judiciary to the executive.” (P. 61.)

INS v. Cardoza-Fonseca , 480 U.S. 421 (1987), set the stage for battle. Trying to return the analysis to pre-Chevron categorization, Justice Stevens explicitly and pointedly stated, “The question [in the case]… is a pure question of statutory construction for the courts to decide.”  Id. at 446. Justice Scalia astonishingly responded that “the Court badly misinterprets Chevron.” Id. at 455 (Scalia, J., concurring). Presumably, Justice Scalia was not suggesting that Justice Stevens misinterpreted the opinion he had authored just three years prior; more likely Justice Scalia meant that Chevron had taken on a life of its own regardless of Justice Stevens’ intent when writing the opinion. As we know, Justice Scalia won the battle and today, the Chevron doctrine bears little resemblance to the decision that birthed it.

My brief summary does not do justice to the depth and complexity of Lawson and Kam’s analysis and the variety of cases they cover, including some from other judges such as Ken Starr and Stephen Breyer. What, if anything, does this article add to the Chevron discussion? I have always found the Chevron two-step to be unintuitive: why aren’t judges deciding pure questions of law, perhaps with agencies acting as expert advisors? Having used Lawson’s administrative law text, I was already familiar with the pre-Chevron categorization framework, but I was unaware of the large role Judge Wald played in forming the Chevron doctrine and why she might have done so. Also, from reading the article, I better understand the Scalia/Stevens INS v. Cardoza-Fonseca battle. Perhaps most importantly, the piece is written with great humor and wit (the headings are all pop culture references, for example) and relatively few footnotes, making it an easy read. And while the authors explain well why the Chevron case and the Chevron doctrine have little relationship to each other, the article will likely not, as they had hoped, bury the former.


What Does It Feel Like To Have OIRA Review Your Rule?

Lisa Heinzerling, Inside EPA: A Former Insider’s Reflections on the Relationship Between the Obama EPA and the Obama White House, Pace Envtl. L. Rev. (forthcoming), available at SSRN.

Ever wondered what it is like—really like—to be an agency official confronting review by the Office of Information and Regulatory Affairs (OIRA) of your agency’s rule? Readers of JOTWELL’s administrative law blog are disproportionately likely to be part of the small group that wonders about such things, and this post has some very good news for them.

Surely, the best way to find out what it is really like to run a rule through OIRA would be to become an insider, serving as a high-ranking official at a major rulemaking agency. Most of us will never have that option. Fortunately for outsiders, a leading administrative law scholar, Professor Lisa Heinzerling of Georgetown University Law Center, did. She left academia for two years to serve as Senior Climate Policy Counsel to EPA Administrator Lisa Jackson from January to July 2009 and then as Associate Administrator of the Office of Policy from July 2009 to December 2010. Now back in the academic fold, she has written a fascinating account of the way that centralized White House review has affected agency rulemaking during the Obama administration.

Inside EPA begins by concisely telling the story of the evolution of centralized executive review from the Nixon through the Obama administrations. A theme of running through this history is how OIRA’s informal, practical power can exceed the limitations that appear on paper in the governing executive orders. For instance, responding to criticisms of President Reagan’s EO 12,291, President Clinton imposed a variety of seemingly strict transparency and deadline requirements on OIRA in EO 12,866. There is less here than meets the eye, however, because OIRA has “taken the position that only when a regulatory action is sent to OIRA through official channels—which include a computer system used for the purpose of facilitating the transfer of rules between the agencies and OIRA—do the transparency requirements of EO 12,866 kick in.”  (P. 9.) This stance leaves OIRA free to intervene earlier in the rulemaking process without leaving fingerprints. (P. 9.) It also gives a rationale for OIRA’s infamous refusal during the Bush II administration to upload EPA’s draft endangerment finding on greenhouse gases. After all, “[i]f OIRA does not upload the package, … it is as if it was never sent to OIRA; no clock begins ticking, and the package does not does not appear on OIRA’s website listing rules under review.” (P. 10.) EO 12,866 might look like it addresses problems of OIRA delays and transparency. It doesn’t.

The heart of Inside EPA is a section entitled, “The Common Law of 13,563.” In it, Professor Heinzerling explains how, to her mind, centralized executive review has actually worked during the Obama administration. She divides her account of the “common law” into five subsections, each asks and attempts to answer a different question. The questions follow—as well as some highlights from her answers.

Who Decides? Professor Heinzerling reports that, from her perspective, “it was often hard to tell who exactly was in charge of making the ultimate decision on an important regulatory matter.” (P. 14.) She notes that Professor Cass Sunstein, former administrator of OIRA and administrative-law superheavyweight, has described OIRA as an “information aggregator” that collects from a variety of sources across the White House and the broader executive branch. On this view, resistance to a rule could come from many different sources—it could be the Chief of Staff; it could be another agency head. Sitting at EPA, Professor Heinzerling couldn’t tell—which, on a moment’s reflection, is pretty remarkable. (P. 15.)  Her general sense, however, was that OIRA wielded much more clout than the mild-mannered “information aggregator” description suggests.  Along these lines, she cites Sunstein’s remark in his recent book, Simpler: The Future of Government, that he had the power to throw “highly touted rules, beloved by regulators, onto the shit list.” (P. 16, citing Simpler at 6.) “Shit,” of course, might lie in the eye of the beholder.

What Is Reviewed? OIRA reviews whatever it wants to review. (P. 18.)

Why Do Rules Fail? It can be hard to tell, but Professor Heinzerling identifies several possibilities. A rule might fail because it creates more costs than benefits on a formal, monetized cost-benefit analysis. (P. 22.) For those who think that excessive reliance on this type of analysis is a very bad idea—and Professor Heinzerling has long been a strong critic—this possibility is naturally very troubling. Another possibility is that OIRA determines that the rule is flat-out wrong “on the merits.” (P. 24, citing Simpler at 27.) But exactly who is in a better position to determine the “merits” of an EPA rule? EPA or OIRA?

When Does Review End (and Begin)? Forget the deadlines of EO 12,866. Agencies can, ahem, “ask” for indefinite extensions of OIRA’s deadlines. And OIRA asks agencies to ask for them—and the agencies, knowing what is good for them, say “yes.” (P. 27.)

What Are We Told? “OIRA follows, and allows the agencies to follow, almost none of the disclosure requirements of EO 12,866.” (P. 29.)

Professor Heinzerling closes with some pointedly normative observations and a simple, partial prescription for improvement. As applied, the regime of EO 12,866 and EO 13,563 raises serious problems of transparency, misallocation of power, and accountability. OIRA could alleviate many of these problems by following the letter and spirit of the EOs themselves. (P. 35.) Left unsaid, perhaps as an exercise for the reader, is the problem of why OIRA does not do so—or why somebody with real power over this agency does not insist that it do so.

(P.S. We now have Professor Heinzerling’s account of OIRA review from an EPA perspective, and we have Professor Sunstein’s account of OIRA review in works such as Simpler. The former is a noted critic of OIRA-style cost-benefit review; the latter is one of its greatest proponents. Maybe next time, as an experiment, Professor Heinzerling could run OIRA and Professor Sunstein could serve at EPA.)


Soft Institutional Design

Margo Schlanger, Offices of Goodness: Influence Without Authority in Federal Agencies, U. Mich. Pub. L. Res. Paper No. 353 (September 9, 2013), available at SSRN.

Margo Schlanger is a law professor at Michigan well-known for her work on prisons, structural reform litigation, and civil liberties, but not (yet) on administrative law as such. Perhaps for precisely that reason, she has given us here a novel, plausible and important account of a new species of administrative institution, one that administrative lawyers have heretofore failed to describe in general terms. A “new” species not in the sense that the species is new to the world, of course, but in the sense that it is newly identified by theory. Field zoologists discover species or traits of species that complicate or overturn established theoretical taxonomies; W.H. Caldwell famously proved that the platypus is a mammal that nonetheless lays eggs (“monotremes oviparous, ovum meroblastic”—so ran the immortal telegram). Likewise, field research on institutional design in the wild often does more for the progress of knowledge than a dozen nth-decimal refinements on whiteboard models of administrative interaction.

The novel institutional form here is the “Office of Goodness,” an office embedded within a larger agency and tasked with promoting or enforcing an extrinsic value that is orthogonal to the agency’s mission, or even one that constrains the agency’s mission. Schlanger headed the Office of Civil Rights and Civil Liberties embedded within the Department of Homeland Security from 2010 to 2012, and she draws upon her personal experiences with the effort to temper the imperatives of security by a measure of attention to liberty and security. But there are no war stories here, only informed illustrations of the larger theme. And Schlanger identifies similar offices from elsewhere in the government.

The striking feature of Offices of Goodness is that they typically lack operational legal authority over the host agency’s decisions, in any hard sense. Although Offices of Goodness may have power to delay or clear decisions, their officers are generally empowered to consult, to advise, or to report to outside third parties, but not to dictate or countermand operational decisions. Even so, Schlanger convincingly documents through case studies that Offices of Goodness can really wield “influence without authority.” Just as we now understand that there is such a thing as “soft law,” which accomplishes its ends through non-coercive mechanisms of signaling, coordination and information, so too there is such a thing as soft institutional design that gives power to agents indirectly. Rather than granting them rights and powers of legal control, soft institutional design works through mechanisms of participation in group decision-making, agenda-setting, and the ability of actors committed to Goodness to pull fire alarms that outside actors will hear. Agencies that do not want the fire alarm to go off will then anticipate or accommodate objections based on Goodness.

Schlanger is alert to the standing risk that Offices of Goodness will be co-opted or neutered by mission-oriented administrators or by principals in Congress or the White House, who may want political credit for promoting Goodness but don’t want Goodness offices with real teeth. Those risks are real, and she is right to be worried about them (assuming, of course, that one shares whatever conception of Goodness underpins whatever Office is at issue in a given case. Nothing in Schlanger’s institutional analysis depends one iota on any particular substantive conception of Goodness). Yet reading her case studies, one is struck not so much by the fragility of Goodness as by its power—by the sheer surprising power of awarding Goodness a seat at the table, wherever that table may happen to be.

Recent work in psychology has underscored the scarcity of cognitive resources and the resulting limits of attention at the individual level. All decisions and choices are made subject to limited time, limited information, and the limited processing capability of the brain. At the group level, there are equally sharp limits to institutional attention, emphasized by Herbert Simon and many others. Those limits imply that values extrinsic to the agency’s ordinary mission, or even constraints on the agency’s mission, may often be slighted not out of malice or bad faith, but just because no one thinks about them. Merely putting a representative of Goodness in the room and at the table to raise the relevant issues, to steer the group’s attention in certain directions, may produce outsize effects even if those representatives have no coercive power whatsoever. Neglect is one of the parents of bad government, and Offices of Goodness may ameliorate neglect even if they are helpless in the face of willful and deliberate abuse.

Schlanger has given us one of very few treatments of soft institutional design — perhaps the very first, depending whether one thinks that earlier papers on privacy offices and inspectors-general properly belong in the same category, or instead raise somewhat distinct issues. In any event, those papers did not abstract from the particulars. Schlanger has thus gone well beyond the earlier efforts by generalizing her analysis to the level of species-identification. The result is a memorable contribution to the institutional design of the administrative state.


What Is the Real Effect of OIRA Application of Cost Benefit Analysis?

Michael A. Livermore, Cost-Benefit Analysis and Agency Independence, 81 U. Chi. L. Rev. (forthcoming, 2014), available at SSRN.

The Office of Information and Regulatory Affairs (OIRA) in the Office of Management and Budget (OMB) has been applying cost-benefit analysis (CBA) to major rules issued by exec­­­utive branch agencies for over thirty years. The practice has long been controversial among legal academics but the debates have taken place against a generally-agreed set of beliefs about the effects of OIRA application of CBA—it increases the power of the President to control policy making by the bureaucracy. Michael Livermore’s meticulously researched and well argued article challenges this enduring belief.

Livermore argues that the relationship between OIRA application of CBA to major rules and presidential influence over policy making by the bureaucracy is far more complicated than the standard account can capture. He contends that the practice has actually increased agency autonomy by providing agencies a means of protecting themselves from presidential control. In Livermore’s view, the practice of OIRA application of CBA to agency rules provides agencies with a “safe harbor” they can access by dominating the process for developing the methodology government uses to apply CBA.

Livermore’s research and analysis focuses almost entirely on the relationship between OIRA and the Environmental Protection Agency (EPA)—the agency that has issued by far the largest number of rules that have been subjected to OIRA review through application of CBA. Livermore documents in detail the ways in which EPA has determined the methodology that OIRA uses to apply CBA. EPA has hired large numbers of staff economists and consulting economists who write studies, books, and articles that dominate the field of CBA.

EPA can, and has, hired far more economists than OIRA, and EPA has consulting contracts with the most prolific and most respected economists who write in the field of CBA. As a result, OIRA has no choice but to adopt as its own the methodologies that EPA supports. Livermore documents the history of every important component of CBA methodology to show how each component that OIRA now applies had its origin in studies conducted by economists under contract with EPA.

The history of two of the most important components of CBA methodology illustrates Livermore’s point well. The most important single input to any CBA of a potential life-saving rule is the estimate of the value of the statistical lives saved by the rule. The methods OIRA uses for this purpose were derived entirely from EPA studies. More recently, estimates of the social cost of carbon dioxide emissions have occupied center stage in the heated debates about whether to continue to use coal as our primary source of electricity. Livermore describes the ways in which EPA has used its position as the dominant source of expertise on regulatory economics to persuade OIRA to adopt EPA’s estimate of the social cost of carbon to evaluate the costs and benefits of the rules EPA is proposing to reduce dramatically use of coal to generate electricity in the United States.

As Livermore recognizes, his research raises many important questions that he is unable to answer. To what extent are his findings applicable to agencies other than EPA? If the traditional simplistic explanation for the decisions of each of the last five Presidents to instruct OIRA to apply CBA to agency rules is unsupportable, in what ways does OIRA application of CBA help Presidents influence policy making? Perhaps each President believed that agencies should use CBA to make major policy decisions and that instructing OIRA to apply CBA to agency rules would further that goal by forcing agencies to retain the services of large numbers of economists as a means of insulating their decisions from presidential control? Perhaps each President created a decision making process that favors EPA because each was more comfortable with the views of his EPA Administrator than his OIRA Administrator? Scholars need to test each of these and scores of other hypotheses that are suggested by Livermore’s article. Livermore has forced us to think more deeply about the inherently complicated relationship between OIRA application of CBA and presidential influence over rulemaking by executive branch agencies.


Remedying Structural Separation-of-Powers Violations

Kent H. Barnett, To the Victor Goes the Toil—Remedies for Regulated Parties in Separation-of-Powers Litigation, 92 N.C.L. Rev. (forthcoming, 2014), available at SSRN.

This coming Term, the U.S. Supreme Court is set to decide National Labor Relations Board v. Noel Canning, a case involving the constitutionality of the President using his recess appointment power to fill various vacancies on the National Labor Relations Board (NLRB). Unless the Court ducks the issues presented in the case, Noel Canning promises to become yet another important case in a string of recent decisions involving structural challenges to federal administrative agencies—challenges that have sought to limit agencies’ power based upon the Appointments Clause, the President’s recess appointment power, the President’s general Article II powers, and the judiciary’s Article III powers. For example, in 2010 in Free Enterprise Fund v. Public Company Accounting Oversight Board the Court held that the dual for-cause restrictions placed on removal of members of the Public Company Accounting Oversight Board (PCAOB) violated separation-of-powers principles. Similarly, in 2011 in Stern v. Marshall the Court held that a non-Article III bankruptcy court could not constitutionally enter a final judgment on a state-law tortious interference counterclaim.

Even though significant attention has been given to the constitutional merits of these and other recent cases, exceedingly little attention has been given by litigants, the courts and scholars to a subsidiary question lurking in the background of the cases: What should the proper remedy be when separation-of-powers violations are found to exist in the structures of federal administrative agencies? Professor Kent Barnett, an assistant professor at the University of Georgia School of Law, quite perceptively identifies this little-noticed question and begins to try to answer it in a forthcoming article titled To the Victor Goes the Toil—Remedies for Regulated Parties in Separation-of-Powers Litigation, which is soon to be published in the North Carolina Law Review. Given that the Noel Canning case is looming on the Court’s docket and various other structural challenges have been brought challenging the newly-formed Consumer Financial Protection Bureau (CFPB), Professor Barnett’s article is extremely timely. Indeed, it is a “must read” for courts and litigants involved in structural separation-of-powers cases as well as constitutional and administrative law scholars.

At its heart, Professor Barnett’s main claim is that when faced with remedying structural separation-of-powers violations, courts often order ineffectual remedies—such as simply severing the structural defect from the agency’s organic act but otherwise leaving the agency fully operative—that fail to compensate regulated parties for harm suffered. Nor do the sorts of remedies often ordered by courts generally prevent future harm, achieve deterrence or incentivize regulated parties to seek redress for structural separation-of-powers violations, according to Professor Barnett. Hence, in structural separation-of-powers cases, Professor Barnett believes that courts often fail to order remedies that further three key remedial values: (1) providing compensation; (2) incentivizing the pursuit of redress; and (3) deterring violations.

To support his basic claim about the ineffectiveness of the kinds of remedies often ordered by courts, Professor Barnett walks through the varying remedies ordered by the federal courts in a number of important structural separation-of-powers cases, including Buckley v.Valeo, Free Enterprise Fund v. Public Company Accounting Oversight Board, Stern v. Marshall, and Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board. For example, he describes how the Court in Free Enterprise Fund—after holding unconstitutional the double for-cause restrictions on removal of members of the PCAOB—did not prevent the PCAOB from taking action under the Sarbanes-Oxley Act. Instead, as Professor Barnett notes, the Court simply severed the offending provision from the statute, rendering the PCAOB’s members “removable at will and thus sufficiently subject to presidential oversight to cure the structural defect.” The Court then remanded the matter to the PCAOB for further proceedings, leaving the board fully operative. According to Professor Barnett, this meant that “[u]ltimately, the prevailing litigant incurred significant costs only to end up in the same place it began, except this time before a potentially resentful board.” It also meant that “future litigants—especially those that repeatedly interact with an agency—might think it wise to ignore other structural defects, no matter how obvious or serious.”

After making his central claim that the kinds of remedies that courts often order in structural separation-of-powers cases like Free Enterprise Fund are ineffectual, Professor Barnett turns to propose some possible solutions and responses. Specifically, one very simple and very practical suggestion he makes to improve remedies is that parties should better brief the issue of remedies for the courts. This common-sense suggestion is one that parties should take to heart and that judges would likely welcome. After all, Professor Barnett notes that internal Supreme Court documents indicate that then-Justice Rehnquist lamented the lack of briefing on remedial issues in Buckley v. Valeo involving the Federal Election Commission.

Two other possible means of improving remedies that Professor Barnett mentions include: (1) having courts order more potent injunctive relief, such as more frequently requiring Congress to reconfigure the agency in a constitutionally sound manner as the Court did in Buckley v. Valeo; and (2) having Congress legislate remedies by statute by, for instance, providing scheduled, monetary damages for structural harms. However, both these proposals, as Professor Barnett himself acknowledges, might pose their own problems and are likely to be controversial. Hence, Professor Barnett raises two additional potential responses: (1) more judicial candor about the judiciary’s inability to render effective structural remedies for regulated parties and about the aspirational nature of structural limitations; and (2) a reassessment of the interest regulated parties have in pursuing structural violations.

In the end, all of the possible responses that Professor Barnett suggests are really just sketches—or tentative ideas—relating to proposed solutions. But the sketches are nonetheless quite useful because, at a minimum, they should help to kick start a broader conversation among scholars, litigators, legislators and judges about how to best deal with remedies in structural separation-of-powers cases. This is likely the greatest contribution of Professor Barnett’s article; it should help to bring more attention to an important topic that has to date largely been ignored and that is very timely given ongoing litigation involving the NLRB and CFPB.

If Professor Barnett’s article does indeed succeed in kick starting additional discussion about the adequacy of remedies for structural violations, care should be taken to view his article as merely the very useful opening of conversation on the topic—not the end. This is because, as Professor Barnett himself acknowledges, the article somewhat narrowly assesses the suitability of structural remedies through the lens of a regulated party’s interests. Other competing interests and values, like the importance of avoiding regulatory chaos and the proper role of the courts, are given some limited attention, but these interests seem to be treated by Professor Barnett as secondary to the interests of regulated parties. Professor Barnett does not hide this fact but rather quite frankly acknowledges that his goal in the article is assessing how best to protect and vindicate the interests of regulated parties. Hence, future conversations about the adequacy of remedies for structural violations—which Professor Barnett’s article hopefully will help to get going—would benefit from other more diverse vantage points being brought to bear, including the vantage point of various governmental interests.


Does Congress Really Mean To Delegate Interpretative Authority To Agencies?

This coming year marks Chevron’s 30th anniversary. Westlaw reports that Chevron has been cited in over 66,000 sources, including in nearly 13,000 articles. Despite the ink already spilled, until now no one had empirically investigated the core assumption underpinning Chevron deference—that Congress actually intends to delegate interpretative authority to federal agencies when it leaves ambiguities in statutes the agencies administer.

Professors Abbe R. Gluck and Lisa Schultz Bressman recently sought answers to this question (and many others). They interviewed 137 congressional counsels, asking them 171 questions about the canons, legislative history, and administrative law doctrines. Their findings will appear in a two-part series in the Stanford Law Review.  The 125-page part one was just published, along with a methods appendix, and there is a lot to like about this piece.

Because this is Jotwell’s Administrative Law Section, I’ll focus on the administrative law findings (Part IV). But for those (like me) who also teach legislation, the findings on Congress’s use of various interpretative tools are well worth a read. As set forth in a table perfect for classroom use, congressional drafters did not know many of our favorite Latin canons (expressio unius, noscitur a sociis, ejudsem generis) by name but actually used them in practice. Other court-created tools were well known but ultimately rejected. For instance, more than 50% said dictionaries were never or rarely used to determine what terms to use in statutes; as one drafter put it, “no one uses a freaking dictionary.” (I remain curious about the distinct and unasked question whether congressional drafters think courts should consult dictionaries to help discern the meaning of statutory terms.)

Gluck and Bressman asked 45 questions about administrative law. Chevron, it turns out, was the most known, by name, of any canon in the study. In other words, the overwhelming majority (82%) of drafters are aware of—and thus legislate against—this background principle. Of those interviewed, 58% said Chevron plays a role when drafting; 31% indicated in comments that statutory ambiguity results in judicial deference to agency interpretations; and 29% reported that Chevron forces them to think about how precisely to draft. Nine in ten (91%) stated that one reason for statutory ambiguity is to delegate decision-making to agencies, with lack of time (92%), complexity of issue (93%), and need for consensus (99%) being other predominant reasons. Gluck and Bressman conclude that Chevron deference is not the reason drafters leave ambiguities; instead, the reasons mirror those articulated in Chevron—delegation of decision-making, implementation, expertise, etc. Notwithstanding, Chevron forces agencies to think more about delegation and precision in drafting.

Mead is a different story. Unlike Chevron’s 82% name recognition, only 28% knew Mead by name.  Yet, by concept, Gluck and Bressman conclude that Mead was a “big winner,” in that 88% indicated that the signal emphasized in Mead—authorization of notice-and-comment rulemaking—is always or often relevant to whether drafters intend for an agency to have interpretative authority. In other words, contrary to Justice Scalia’s Mead dissent, Congress’s authorization of agency rulemaking or formal adjudication may be one of the stronger signals of its intent to delegate interpretative authority to agencies. (To be sure, Justice Scalia has made clear—most recently this Term in City of Arlington v. FCC—that procedural formality is a sufficient, though not necessary, signal for deference.)

Speaking of City of Arlington, Gluck and Bressman asked drafters if they delegate major policy questions to agencies. Confirming Justice Scalia’s intuition that Congress does not “hide elephants in mouseholes,” more than 60% said Congress does not intend to delegate major policy questions.  Gluck and Bressman provide another classroom-ready chart to illustrate the types of ambiguities drafters intend agencies to fill—with implementation details (99%) and areas of agency expertise (93%) leading the way and major questions garnering less than 40%. The survey did not ask whether drafters intend to delegate by ambiguity authority for agencies to determine the scope of their own statutory jurisdiction—the question City of Arlington answered in the affirmative. But based on their responses regarding major questions, it is probably sensible to infer that congressional drafters would take issue with the majority’s conclusion in City of Arlington.

Gluck and Bressman uncover many other insights, including how Congress deals with multiple federal agencies, federal/state agency interaction, and longstanding agency interpretations. But I’ll conclude with legislative history. As I have explored elsewhere, courts and scholars have long debated which (and when) interpretative tools should be used under Chevron. For instance, many judges use legislative history at Chevron step one to determine the meaning of ambiguous terms. This study’s findings suggest alternative uses. In particular, 94% of drafters indicated that one purpose of legislative history is to shape the way agencies interpret statutes, and 21% described it as a mechanism for agency oversight. Gluck and Bressman suggest that these findings support the use of legislative history “as a relevant signal of delegation” (perhaps at step zero or one). That’s not an unreasonable conclusion. If a main purpose of legislative history is to shape agency interpretation (beyond shedding light on ambiguous terms), however, there may be a stronger argument that it is best used at Chevron step two to ascertain the reasonableness of an agency’s interpretation.

In all events, to borrow a line from Justice Scalia, “[i]t is indeed a wonderful new world that the [Gluck and Bressman study] creates, one full of promise for administrative-law professors in need of tenure articles and, of course, for litigators.” I, for one, am following their lead and surveying agency rule drafters to explore their understanding and use of various interpretative tools and administrative law doctrines in the rulemaking process. I look forward to reading part two of their study when published later this year.


Assessing Agency Legitimacy

David Markell and Emily Hammond Meazell, Administrative Proxies for Judicial Review: Building Legitimacy from the Inside-Out, 37 Harv. Envtl. L. Rev. (forthcoming 2013), available at SSRN.

The great question underlying American administrative law is that of agency legitimacy. Administrative agencies, whose heads don’t answer to the voters and whose decisions for the most part are not subject to effective popular checks, have dubious democratic bona fides. Where do they get off, then, mandating rules of conduct and imposing punishments backed up by the coercive power of the state? A crucial part of the answer, in American administrative-law thinking, has rested on the institution of judicial review: We can trust agencies to exercise their delegated authority, the classic argument runs, and we can treat that authority as legitimate, because we can rely on courts to take action if the agencies step out of line.

But as administrative-law scholars well know, the judicial-review focus has limitations. David Markell and Emily Hammond Meazell, in their paper Administrative Proxies for Judicial Review: Building Legitimacy from the Inside-Out, note that few administrative decisions ever go before a court. Judicial review of most agency decisions is neither cost-effective nor practical; review of others is precluded by law. This, the authors urge at the beginning of their paper, is one of the “great paradoxes of administrative law,” raising the question, “What else is there to legitimize unreviewable agency action?”

Markell and Meazell respond by studying the universe of 58 citizen petitions, filed over the past quarter-century, asking that EPA withdraw a state’s authorization to administer a federal environmental-law program. There’s reason to believe that these petitions would be ineffective, and that the agency’s response typically would be desultory. For one thing, it would be impractical—almost unthinkable—for the agency ever to grant such a petition: if EPA ever actually withdrew state authority to administer a regulatory program, it would have to take over administration of the program itself. The agency doesn’t have the resources to do that sort of thing. For another, a decision by EPA to ignore such a petition, or to investigate it only half-heartedly before moving on to other concerns, wouldn’t be subject to judicial review. Heckler v. Chaney provides secure protection against judicial oversight in these cases, and leaves the agency free to do what it will.

But that’s not what the authors found. Their study finds that interested parties have filed petitions at a steady pace of about three per year for the past 16 years; the typical petition is attorney-written and legally sophisticated, and includes a local environmental nonprofit among its signatories. EPA during the study period “virtually never simply denied a petition or ignored it.” Rather, it responded by launching an informal investigation into the concerns the petition raised. With respect to more than 80% of those concerns, EPA generated some document or set of documents acknowledging the concern and describing the outcome it reached. With respect to about 70%, it explained its decision in light of the relevant legal standards and the data available to it. And with respect to just over half of the concerns raised by the petitions, EPA’s process (informal investigation, followed by negotiations with the state) resulted in a formal commitment by the state to make measurable responsive changes in its program administration. The authors identify all of these (responsiveness, reason-giving, and results) as key internal metrics for agency legitimacy.

There are limits to Markell and Meazell’s data, and the authors acknowledge them. They code, for example, whether the agency negotiated an agreement with the state; they were unable to code whether it was a good agreement. The findings are nonetheless important.

What has caused EPA to internalize legitimacy principles as well as it has, engaging in public, reasoned decision-making coupled with reason-giving even in the absence of statutory constraint or judicial review? And why do withdrawal petitions work at all, given that actual withdrawal isn’t a credible threat? The authors suggest tentative answers.  They note EPA’s agency culture, its network of regional offices (so that petitions are filed against a backdrop of relationships with local actors), and EPA’s need to support its political legitimacy in the eyes of local stakeholders. The petition mechanism puts local environmental actors at the table with governments; it gives EPA political cover, and targets of opportunity, in its dealings with state actors; and it gives actors inside state government political cover in their dealings with each other, so that state-government actors seeking change can point to the EPA investigation as forcing it. (The withdrawal proceeding thus has a lot in common with the AALS re-accreditation process, including the opportunity for the accreditor to exert useful pressure even though ultimate denial is implausible.)

Markell and Meazell’s paper is better at raising questions than at providing answers, but two things make it a thought-provoking and admirable one. One is the set of questions it addresses: What makes agency action legitimate? How can we work our way through to new thinking about agency legitimacy? What makes some agency processes work better—or less well—than others? The second is its eagerness to connect theory with practice. Rather than articulating a theoretical framework and stopping there, the authors use their framework to structure their examination of actual agency process, to see how well the data fits the theory. This is good work. We need more of it.


Public Agencies Going Private

Jon D. Michaels, Privatization’s Progeny, 101 Georgetown Law Journal (forthcoming 2013), available at SSRN.

Administrative law loves binaries. There are executive agencies and independent regulatory commissions. There are federal bureaucratic organizations and state entities. There are public agencies and private firms. In recent years, scholars have examined organizations from different sides of a “boundary” working together – for instance, executive agencies and private firms in the new governance literature and separate federal and state agencies in federalism work. There is little discussion, however, on organizations essentially operating at or near the boundary.

Jon Michaels’s previous work has focused on national security agencies turning outward from the public sphere, specifically using government contracting, for classic government tasks. In his forthcoming piece, Privatization’s Progeny, Michaels turns inward, exploring how privatization is reshaping a wider class of public institutions themselves, through “the marketization of bureaucracy” and “government by bounty.”

Michaels starts his article with a concise, straightforward summary of privatization’s economic and political benefits. Other scholars tend to emphasize the economic benefits; therefore, Michaels’s attention to how privatization can foster control over the administrative state demonstrates some novelty in well-worn discussions. He then takes a clever turn from the thought provoking privatization literature (including his past work): instead of showing how the national security apparatus reaches out to the private sector (through government contracting, for example), he looks at how the broader administrative state is transforming inward as well as shifting the terms of contracting itself.

Michaels concentrates on changes to the civil service and the growth in bounty initiatives in government contracting. For the former, Michaels examines “unprecedented revisions to civil servants’ (1) collective bargaining rights, (2) wages and benefits, (3) promotion protocols, and (4) job security.” For the latter, Michaels details how social-impact bonds have the “potential to outperform traditional contracting vis-à-vis efficiency, cost-savings, and political payout.” Others have written about the undermining of collective bargaining rights and government prizes, but not together. (Nicholas Parillo’s historical work makes some interesting connections between Michaels’s two case studies; Parillo examines “how American lawmakers remade governance by shifting public officers’ monetary compensation away from profit-seeking arrangements—such as fees-for-service and bounties—and toward fixed salaries.”).

The heart of the article rests in the implications for the administrative state. Michaels is ambitious here. Although he may not convince all readers of his normative conclusions, he succeeds in getting them to think seriously about important issues. As for the marketization of bureaucracy, Michaels largely bemoans the changes to the civil service. His account of the postal service is compelling (as is his analogy of educators teaching to the test). As for government by bounty, Michaels also worries about the ramifications. His analysis of the Food and Drug Administration’s priority-review voucher program raises interesting issues, though firms may not act in the way he predicts.

In both the marketization of the bureaucracy and government by bounty discussions, the consequences for administrative law are nicely sketched. For instance, Michaels has an interesting take on judicial review of questions of law in light of the marketization of the administrative state. Michaels argues that if marketization produces “greater efficiencies, budgetary savings, and more complete unitary control over the administrative state,” “we might finally see a sharper division between reversal rates under Skidmore and Chevron.” As Michaels explains, “[t]his would be so because marketization ought to heighten courts’ Chevron deference, and lessen Skidmore deference.” Perhaps because Chevron (and Skidmore) is such trampled-on ground in administrative law, I enjoyed the discussion of the government-contractor defense and state action (in some sense opposite sides of the same issue) even more.

There are very clever and striking insights in this article. I am already wrestling with its implications for a project I have about bureaucracy on the boundary. Administrative law scholars would be wrong to think of Michaels as only a national security scholar. He is a national security scholar, but he is also an administrative law scholar. In this forthcoming piece, as with his other work, Michaels shows that he can make significant contributions to our understanding of meaningful issues in administrative law.