Libertarian Administrative Law, or Administrative Law?

John Harrison

82 U Chi L Rev Dialogue 134 (2015) [Essay PDF]

A Response to Cass R. Sunstein and Adrian Vermeule, Libertarian Administrative Law, 82 U Chi L Rev 393 (2015) [Article PDF]


This Essay is written with the advantage of hindsight, as the Supreme Court has now decided Perez v Mortgage Bankers Association1 (“Mortgage Bankers”) and Department of Transportation v Association of American Railroads.2   The DC Circuit’s decisions in those cases are leading instances of what Professors Cass Sunstein and Adrian Vermeule call “libertarian administrative law,” which “seeks to use administrative law to push and sometimes shove policy in libertarian directions, primarily through judge-made doctrines that lack solid support in the standard legal sources.”3 Sunstein and Vermeule argue that this trend “should be cabined by the Supreme Court or by the DC Circuit itself.”4

In both Mortgage Bankers and Association of American Railroads, the Supreme Court unanimously reversed the DC Circuit, but in neither did the Court read the riot act to the court of appeals. In Mortgage Bankers, three concurring justices took the trouble to grapple with the legal problem the lower court had tried to resolve, while ultimately rejecting its solution. That problem, which was created by the Supreme Court’s cases, is likely to soon receive the Court’s attention in an appropriate case. Association of American Railroads was decided on the narrowest of three possible grounds.5 The Court concluded that Amtrak is part of the government for purposes of the nondelegation doctrine, and so it did not have to consider that doctrine’s application to grants of authority to private people.6 Justice Samuel Alito, in concurrence, noted that the DC Circuit’s conclusion that Amtrak is private “understandably” relied on the statutory text.7 The Court did not take the opportunity to endorse the position that Sunstein and Vermeule attribute to its cases: grants of authority to private people are permissible as long as Congress provides an intelligible principle.8

Those cases are routine manifestations of the relationship between the Supreme Court and the DC Circuit, in which the lower court does much of the work of administrative law but with meaningful supervision. In my view, Sunstein and Vermeule have not shown that the Court confronts an emergency in that relationship because the DC Circuit has produced a body of libertarian administrative law that “lacks sufficient respect for existing law, including, emphatically, controlling precedents of the Supreme Court”;9 that is “without sufficient warrant in existing sources of law, including the decisions of the Supreme Court itself”; or that is “lawless."10 This Essay examines the DC Circuit cases that Sunstein and Vermeule discuss (except for Business Roundtable v Securities and Exchange Commission,11  which I am not qualified to assess) and argues that the DC Circuit has not produced a set of cases meeting that description. I argue that some of the authors’ characterizations of a decision as libertarian are inapt and that some of the cases they discuss are not administrative law. None of the administrative law decisions they discuss (again, with one possible exception) is a substantial departure from the Court’s precedents. The DC Circuit makes mistakes and judges are influenced by their policy views. In the old saw about the news business, those are “dog bites man” stories. These mistakes are unfortunate. But they are not out of the ordinary.

I.  Delegation

Professors Sunstein and Vermeule maintain that, in cases involving the nondelegation doctrine, the DC Circuit “has twice developed its own nondelegation doctrine, operating independently of the Supreme Court’s and in the face of that Court’s noticeable lack of enthusiasm for the doctrine.”12 They point to two cases, in both of which the Supreme Court reversed the DC Circuit. In the first case, American Trucking Associations, Inc v United States Environmental Protection Agency,13 the DC Circuit held that a provision of the Clean Air Act14 was an unconstitutional delegation of legislative power because it lacked the intelligible principle that the Court says must cabin grants of authority to executive agencies.15 As I will explain, in reversing that decision the Supreme Court did not question the lower court’s understanding of the doctrine. The second case is Association of American Railroads v Department of Transportation,16 in which the DC Circuit held that grants of regulatory authority to private people are per se unconstitutional.17 Finding that Amtrak is a private corporation, the court of appeals concluded that it could not be given regulatory authority.18 The DC Circuit’s holding that private persons may not be granted regulatory power rests on a quite plausible reading of the Court’s cases—a reading that Justice Alito endorsed in his concurring opinion in Association of American Railroads.19 The Court, however, found that Amtrak was part of the government, and so did not need to consider grants of power to private entities.

The Supreme Court has said that Congress may give regulatory authority to federal agencies only if it gives them an intelligible principle to guide their choices.20  In American Trucking Asso­ciations, Judge Stephen Williams identified an unusual problem: Some airborne substances pose health risks at any concentration above zero.21 There is no threshold below which they are safe. With no threshold, the DC Circuit reasoned, the EPA lacked “any determinate criterion for drawing lines.”22

The Supreme Court reversed, with Justice Antonin Scalia writing the majority opinion.23 In response to the DC Circuit’s concern that the agency was left to make an arbitrary choice, the Court pointed to decisions in which it had upheld vague grants of authority, and denied that an intelligible principle must be determinate.24 The Court did not say that a wholly indeterminate principle could be intelligible; it instead appeared to read the Clean Air Act as merely vague, not empty of content as the court of appeals thought.25 The Court thus was rejecting the DC Circuit’s interpretation of the statute, not its understanding of the nondelegation principle. American Trucking Associations indicates that a statute genuinely providing no criteria at all would be unconstitutional.26

American Trucking Associations also bears on Association of American Railroads. Scalia made clear that, strictly speaking, legislative power is nondelegable.27 An intelligible principle does not make a delegation of legislative power permissible. Rather, it keeps a grant of decisionmaking authority from being a delegation of legislative power and hence impermissible.28  An intelligible principle thus is a necessary condition for such a grant to be constitutional.29  Whether it is a sufficient condition is another question, and whether private people are subject to the same principles as government actors is very much another question.

In Association of American Railroads, Alito said that a grant of regulatory authority to a private person is unconstitutional, full stop: “As to the merits of this arbitration provision, I agree with the parties: If the arbitrator can be a private person, this law is unconstitutional. Even the United States accepts that Congress ‘cannot delegate regulatory authority to a private entity.’”30 The intelligibility of the statutory principles the arbitrator is to apply apparently did not matter to Alito. Scalia implied the same conclusion in his dissent in Mistretta v United States:31

Precisely because the scope of delegation is largely uncontrollable by the courts, we must be particularly rigorous in preserving the Constitution’s structural restrictions that deter excessive delegation. The major one, it seems to me, is that the power to make law cannot be exercised by anyone other than Congress, except in conjunction with the lawful exercise of executive or judicial power.32

Private people may not exercise executive or judicial power.

Alito’s stance is consistent with the reasoning in Currin v Wallace,33 which though now well aged is the Court’s most recent substantial analysis of grants of authority to both government agencies and private people considered under the rubric of nondelegation.34 Congress had given the Secretary of Agriculture authority to regulate tobacco markets, including authority over grading quality and condition.35 The statute also provided that regulation was conditioned on a two-thirds favorable vote among tobacco growers.36 Chief Justice Charles Evan Hughes, writing for the Court, found that Congress had given enough policy guidance that there was “no unfettered discretion lodged with the administrative officer.”37 He had a different response concerning the involvement of private parties: there was no delegation to the tobacco growers because “Congress ha[d] merely placed a restriction upon its own regulation by withholding its operation as to a given market ‘unless two-thirds of the growers voting favor[ed] it.’”38

Hughes’s reasoning suggests that he thought that intelligible principles were not relevant to private parties. If he had believed that an intelligible principle directed to the Secretary were enough to sustain the entire arrangement, he probably would not have explained the growers’ role on a different ground, and would have relied only on the intelligible principle given the Secretary. If he had believed that the growers themselves had to be subject to an intelligible principle, he would have decided the case differently because the growers themselves were free to vote as they chose. Hughes’s way of analyzing the issues makes sense if he believed that private parties may be involved in exercises of government power to only a limited extent, and that beyond those limits decisions must be made by government agencies—which in turn must be subject to an intelligible principle.

The Supreme Court has never held that grants of authority to a private person are permissible as long as the private person is given an intelligible principle to apply.39 Instead, its cases support the conclusion that grants of regulatory authority to private people acting on their own are unconstitutional per se, just as true delegations of legislative power are.

II.  Commercial Speech

Professors Sunstein and Vermeule describe two DC Circuit First Amendment commercial speech cases as “administrative law” cases. They classify those cases as administrative law cases partly on the ground that intermediate scrutiny is much like arbitrariness review.40 Intermediate scrutiny under the First Amendment balances competing interests: restrictions on commercial speech must be substantially related to an important government interest.41 That principle follows if such restrictions are costly and the courts must ensure that the costs are justified. The premise that restrictions on commercial speech are costly distinguishes intermediate scrutiny from arbitrariness review under the Administrative Procedure Act.42  (APA). In a First Amendment case, once the regulated party has shown that the law at issue does indeed limit commercial speech, it has established that the law has an undesirable consequence and the question becomes whether that consequence is counterbalanced by an advantage. In contrast, under the APA, the courts ask whether the agency’s reasoning was sound.43 Costs are not assumed to be significant as they are under the First Amendment. The two inquiries are quite different from one another, and it is plausible to think that intermediate scrutiny is more demanding in some rough sense.44

III.  Interpretative Rules

Professors Sunstein and Vermeule criticize the DC Circuit’s Paralyzed Veterans doctrine, which the Supreme Court has since rejected in Mortgage Bankers.45 In Paralyzed Veterans of America v D.C. Arena LP,46 the DC Circuit said that if an agency wants to change its interpretation of a rule having the force and effect of law, it must do so through the notice-and-comment process.zed

47 Subsequent DC Circuit cases, including Mortgage Bankers Association v Harris,48 relied and elaborated on that principle.49 That doctrine was not libertarian. The outcome in Paralyzed Veterans itself favored regulation,50 and in a world with much regulation, a rule that inhibits regulatory change has no particular valence. The Paralyzed Veterans opinion was joined by Judge Harry Edwards, who is no libertarian and is free with his strictures when he thinks it appropriate.51 Subsequent panels, whether composed of friends or opponents of regulation, were bound by that case. They may have moved at small margins for ideological reasons, but in applying that principle they were following—not departing from—the standard legal materials.

In Mortgage Bankers, the Court held that interpretative rules need not be adopted with notice and comment.52 The Court assumed that the rule at issue in that case was interpretative.53 The APA is clear on that point, and Vermont Yankee Nuclear Power Corp v National Resources Defense Council, Inc54 (“Vermont Yankee”) underlines the principle that courts may not add to the procedures required by statute.55

On that score, Mortgage Bankers was obviously correct. However, three concurring justices raised a difficulty: under Bowles v Seminole Rock & Sand Co56 (“Seminole Rock”) and Auer v Robbins,57 agency rules or statements that purport to interpret regulations are given strong deference and therefore can have a practical effect similar to an amendment to an earlier regulation.58 Paralyzed Veterans itself responded to this concern, pointing out that the APA requires that when a regulation has to be adopted through notice and comment, an amendment to that regulation must also be adopted through notice and comment.59 The Court in Mortgage Bankers denied not the soundness of that principle but rather its relevance, noting that interpretative rules are not amendments to the rules they gloss.60

The DC Circuit in the Paralyzed Veterans cases may have been reaching for a solution that is obscured by nonstatutory categories often used by judges and commentators. At the end of Mortgage Bankers, the Court addressed Mortgage Bankers’ argument that the rule at issue was actually legislative and not interpretative. The Court found that argument had been waived before it.61 The argument was also unhappily formulated because it rested on the common but doubtful assumption that every regulation is either legislative or interpretative.62 The APA refers to interpretative rules but does not mention legislative rules.63 

Courts and commentators developed the concept of legislative rules to label those agency actions that, like legislation, change legal relations.64 The concept was developed in contrast with that of interpretative rules, which do not have that effect.65  Concepts developed in dealing with particular problems are subject to a common error: two categories that are known to be mutually exclusive may incorrectly be assumed to be collectively exhaustive. If a court knows that a case falls into one of two nonoverlapping categories, the question whether there is some other category does not arise. When that happens, it is easy to assume, or to speak as if one assumes, that the two categories are the only possibilities.

So it may be with legislative and interpretative rules: although the concepts do not overlap, they also may not exhaust the possible kinds of rules. A rule that has the form of an interpretation and that does not change legal relations might nevertheless not be interpretative under the APA if its effects on outcomes are strong enough. Seminole Rock and Auer give quite strong effect to rules that take the form of interpretation.66 Rules with that effect might not be interpretative, even though they are not legislative. If they are not interpretative, they may be adopted only with notice and comment.67 The APA does not provide that only legislative rules must be adopted with that procedure; it requires that all rules that are not interpretative (and that do not fall into one of the other exceptions) must be adopted with notice and comment.68

The drafters of the APA may have expected that any rule that purports to be only an interpretation would be interpretative. They may also have assumed that any rule that purports to be only an interpretation would not have the legal consequences that agency interpretations are given under Auer. Those expectations cannot both be satisfied as long as Auer is controlling precedent.

IV.  Standing

Professors Sunstein and Vermeule describe a number of DC Circuit decisions regarding standing, most of which I have not read, that find standing for regulated parties and a lack of standing for beneficiaries of regulation.69 The authors “do not contend that these decisions are implausible or that a majority of the Supreme Court would disagree with all or most of them,” but they do maintain that “it is reasonable to say that almost all of them could have gone the other way.”70 As Sunstein and Vermeule describe them, those cases are libertarian and fall under the category of administrative law in a broad sense (though some rest on Article III and not the APA), but they are not cases in which the DC Circuit has departed from well-established Supreme Court precedent.

If Sunstein and Vermeule are correct about the standing cases they discuss, those cases are par for the course for courts of appeals. Judges are often influenced by their policy views in close cases, and in questioning Sunstein and Vermeule’s main conclusions, I do not doubt that many decisions by the DC Circuit have been influenced by judges’ policy judgments. As I understand Sunstein and Vermeule, they claim that DC Circuit libertarians have sometimes decided lawlessly or almost lawlessly for policy reasons, not just that they have adopted reasonable interpretations that accord with their own policy positiona.71 As the authors describe them, the standing cases discussed support the latter conclusion, not the former.

V.  Commitment to Agency Discretion by Law

In Cook v Food & Drug Administration,72 the DC Circuit concluded that the FDA was under a statutory duty to refuse admission to the country of a drug that the agency had concluded was misbranded and unapproved.73  In Sierra Club v Jackson,74 the DC Circuit found that the EPA was not under a statutory duty to take steps, including the commencement of litigation, to prevent the construction of three air pollution–emitting facilities in Kentucky.75  Professors Sunstein and Vermeule characterize both of those decisions as libertarian and subject them to sarcasm that requires French.76 In Cook, they say, the DC Circuit decided “more or less to ignore the instructions of the Supreme Court by means of irrelevant distinctions.”77 Cook is correct, and Sunstein and Vermeule have not established that it is libertarian; I think it is not.78 Sierra Club is, in my view, wrong, but it is not a serious departure from established legal principles.

Cook involved the Federal Food, Drug, and Cosmetic Act79 (FDCA), which governs the importation of drugs into the United States.80 Under the FDCA, if the FDA concludes that a drug offered for import is misbranded or is an unapproved new drug, “then such article shall be refused admission” to the United States.81 In 2009, the FDA detained two shipments of thiopental, a drug used in lethal injections, because the agency had grounds to believe that thiopental was misbranded or unapproved within the meaning of the statute.82 When state officials told the FDA that the drugs were used for lethal injection, the shipments were released into the country.83  In 2011, the FDA issued a policy statement stating that it would exercise its enforcement discretion to allow admission of thiopental, although the agency agreed that thiopental is a misbranded and unapproved new drug under the FDCA.84

According to Sunstein and Vermeule, Cook was a serious departure from Heckler v Chaney,85 an important case about agency discretion. The Court in Chaney distinguished the facts before it from those in Dunlop v Bachowski,86 which had concluded that an agency had a judicially enforceable duty to commence a lawsuit on behalf of a private person.87 The statute in Bachowski had an if-then structure ending in a directive to the agency.88 The Labor-Management Reporting and Disclosure Act of 195989 (LMRDA) provided that the Secretary of Labor “shall investigate such complaint and, if he finds probable cause to believe that a violation . . . has occurred . . . he shall . . . bring a civil action.”90 Then-Justice William Rehnquist explained in Chaney that ordinarily the decision whether to bring enforcement litigation is committed to agency discretion by law—and hence is not subject to judicial review—because there is no law for a reviewing court to apply.91 Most of the time, Congress does not provide criteria that the agency must follow in allocating its enforcement resources.92 But if Congress does supply such criteria, the presumption of commitment to agency discretion may be overcome. That presumption was overcome in Bachowski because Congress gave guidance: if the Secretary finds probable cause of a violation, he must bring a civil action.93

The provision governing the criminal sanctions at issue in Chaney itself, by contrast, did not tell the FDA when to recommend that the Attorney General prosecute.94 Nor did the Court find any other indication that Congress meant to do so. Instead, the Court noted that “[t]he section on criminal sanctions states baldly that any person who violates the [FDCA]’s substantive prohibitions ‘shall be imprisoned . . . or fined.’”95 Further, the Court explained, “Respondents argue that this statement mandates criminal prosecution of every violator of the Act but they adduce no indication in case law or legislative history that such was Congress’ intention in using this language, which is commonly found in the criminal provisions of Title 18 of the United States Code.”96 The Court was therefore “unwilling to attribute such a sweeping meaning to this language, particularly since the Act charges the Secretary only with recommending prosecution; any criminal prosecutions must be instituted by the Attorney General.”97

The provision at issue in Cook, although part of the FDCA, is like that in Bachowski and not Chaney.98 Indeed, the case for a judicially enforceable duty was stronger in Cook than it was in Bachowski. The two statutes are the same in that they have if-then structures, with the first clause describing the conditions under which the agency is to act.99 Under the FDCA, if the Secretary finds that a drug is misbranded or unapproved, he must refuse it admission to the country.100 The statute provides the criteria by which to determine whether a drug is misbranded or unapproved, and in Cook the FDA agreed that those criteria were met.101 Cook is easier than Bachowski because the subsequent clause in Cook did not direct the initiation of enforcement proceedings. As the Court noted in Chaney, litigation is often costly and the number of cases an agency can bring typically exceeds its resources.102 Deciding that a drug should be refused admission to the country costs very little once the FDA has determined that it is misbranded or unapproved. (Actually identifying and intercepting packages that contain such articles may be costly, but deciding that their importation is forbidden is not.)103 Because refusal of admission is relatively inexpensive, it is easy to believe that the FDA is required to refuse admission to every single drug it determines is misbranded or unapproved.

The provision concerning criminal enforcement at issue in Chaney was not an if-then imperative directing an agency to take certain actions under certain conditions because it was not an imperative directed to an agency. As Rehnquist pointed out, that provision stated that whoever committed certain acts should be fined or imprisoned.104 Read as an imperative, that statement is directed to the courts that impose criminal penalties. Executive agencies bring prosecutions, but unlike the LMRDA, the FDCA does not command the initiation of legal proceedings.105 Rehnquist explained that the FDCA’s criminal sanctions provision reads like several provisions in Title 18, but did not elaborate on that point.106 He may have meant (and in any event, it is the case) that a directive that courts punish those who commit a crime is not an order to the executive to prosecute everyone it thinks has done so. Instead, the explicit imperative is addressed to the judiciary, and there is no implicit imperative directed toward anyone else. This implication is a natural inference from the absence of criteria and from the executive’s limited litigation resources, but it is an inference that defeats a claim of implicit meaning—not one that undermines explicit meaning.107 Chaney does not refuse to read “shall” to mean “shall”; it refuses to read “shall” addressed to one branch of government as a command to another.

Sierra Club seems to me to have been wrongly decided. The statute at issue had an if-then structure with an imperative directed to the agency.108 The DC Circuit concluded that the command to take such measures as were necessary left the agency so much latitude in deciding what was necessary that there was “no law to apply.”109 Courts regularly review agency decisions about the appropriate way to meet some statutory goal, so judgments of necessity are reviewable in general.110 Insofar as the DC Circuit took Chaney to reduce the mandatory force of “shall” in statutes directing the initiation of legal proceedings, it misinterpreted that case. But right or wrong, Sierra Club was a decision within the range of reasonable disagreement, not a manifestation of systematic libertarian lawlessness.

VI.  The Supreme Court and the Lower Courts

This Essay has engaged with Professors Sunstein and Vermeule in considerable detail because their argument itself is quite detailed. I think their central claim is not well taken, but I may be wrong. Thus, in this final Part I will use one of our points of disagreement to say something about the relationship between the Supreme Court and the lower courts, especially the DC Circuit.

Sunstein and Vermeule read Chaney to stand for a broad principle about enforcement discretion.111 I think it stands for a substantially narrower principle that is not about enforcement in general but rather about the initiation of proceedings. Putting a drug on a list falls into the broader category but not the narrower. Students of the common law may think that whether to read a precedent broadly or narrowly is really just a question of the rhetoric used in explaining a result reached on other grounds. That may be true when the Supreme Court deals with its own earlier cases; but if it is also true about lower courts dealing with the Supreme Court’s precedents, then the hierarchical structure of the federal judiciary does not function as normally understood. The Supreme Court reviews only a tiny fraction of court of appeals decisions;112 its influence operates through precedents, and operates only if lower courts really follow those precedents and do not just claim to do so. If there is to be meaningful vertical stare decisis, questions about the meaning of precedents, such as the question of the scope of cases like Chaney, must be real and important.

One way for the Supreme Court to increase lower court compliance with its decisions is for the Court to articulate the case-reading principles that lower courts should use.113 Had it reviewed and reversed Cook, for example, the Court might have explained why Chaney should be read broadly. Simply saying that it applies to enforcement in general would not address the methodological question, but methodology is crucial because of the Court’s limited capacity. By building a body of principles about the precedential scope of its own cases, the Court could increase the binding force of those cases, as the norms of case reading would provide additional guidance and hence constraint for the lower courts.

A program of elaborating the doctrine of vertical stare decisis itself would be very useful to all levels of the federal judiciary. Of course, it is possible that in the process of trying to do that, the justices would discover that they do not agree enough to generate a body of coherent, transsubstantive norms about the lower courts’ correct approach to the Supreme Court’s precedents. That discovery would stymie the program I have suggested, but it would also show that the lower courts are not departing from the Supreme Court’s norms when they decide how to read a case.


           James Madison Distinguished Professor and Joseph C. Carter Jr Research Professor, University of Virginia School of Law

      1     2015 WL 998535 (US).

      2     2015 WL 998536 (US).

      3     Cass R. Sunstein and Adrian Vermeule, Libertarian Administrative Law, 82 U Chi L Rev 393, 398 (2015).

      4     Id at 402.

      5     See Part I.

      6     Association of American Railroads, 2015 WL 998536 at *8–9.

      7     Id at *9 (Alito concurring). The statute provides that Amtrak is “not a department, agency, or instrumentality of the United States Government.” 49 USC § 24301(a)(3).

      8     According to Sunstein and Vermeule, the DC Circuit’s decision in Association of American Railroads was “an opportunity to bring home the message that the DC Circuit has repeatedly failed to hear: at least outside very extreme circumstances, invalidation on nondelegation grounds is not permissible in contemporary administrative law.” Sunstein and Vermeule, 82 U Chi L Rev at 423 (cited in note 3). The Solicitor General did not ask the Court to say that grants of authority to private people are permissible provided that Congress gives an intelligible principle. The Government had sought certiorari because of the importance of the statute, and not so that the Court could clarify that the same lax standard applies to grants of authority to both private people and government agencies. See Petition for a Writ of Certiorari, Department of Transportation v Association of American Railroads, Docket No 13-1080, *11–12 (US filed Mar 10, 2014) (available on Westlaw at 2014 WL 953507).

      9     Sunstein and Vermeule, 82 U Chi L Rev at 400 (cited in note 3).

     10    Id at 401.

     11    647 F3d 1144 (DC Cir 2011).

     12    Sunstein and Vermeule, 82 U Chi L Rev at 417 (cited in note 3).

     13     175 F3d 1027 (DC Cir 1999).

     14     Pub L No 88-206, 77 Stat 392 (1963), codified at 42 USC § 7401 et seq.

     15     American Trucking Associations, 175 F3d at 1034.

     16     721 F3d 666 (DC Cir 2013).

     17     Id at 670.

     18     Id at 668.

     19     Association of American Railroads, 2015 WL 998536 at *12 (Alito concurring).

     20     See Whitman v American Trucking Associations, Inc, 531 US 457, 472 (2001).

     21     American Trucking Associations, 175 F3d at 1034.

     22     Id.

     23     American Trucking Associations, 531 US at 462, 476.

     24     Id at 474–76.

     25     Id at 474–75.

     26     See id at 475–76. Scalia presaged a latitudinarian view of the nondelegation principle in one of his prior dissenting opinions. See Mistretta v United States, 488 US 361, 415 (1989) (Scalia dissenting) (“But while the doctrine of unconstitutional delegation is unquestionably a fundamental element of our constitutional system, it is not an element readily enforceable by the courts.”). Sunstein and Vermeule say that the Court in American Trucking Associations “largely relied on its own precedents, pointedly quoting its statement” that it does not second-guess Congress’s decisions regarding the permissible extent of delegation, and the authors cite a passage in that case that in turn quotes Scalia’s dissenting opinion in Mistretta. Sunstein and Vermeule, 82 U Chi L Rev at 418–19 & n 109 (cited in note 3), citing American Trucking Associations, 531 US at 474–75, quoting Mistretta, 488 US at 416 (Scalia dissenting).

     27     American Trucking Associations, 531 US at 472 (noting that the text of the Vesting Clause “permits no delegation” of legislative power).

     28     See id.

     29     See id.

     30     Association of American Railroads, 2015 WL 998536 at *12 (Alito concurring), quoting Association of American Railroads, 721 F3d at 670.

     31     488 US 361 (1989).

     32     Mistretta, 488 US at 416–17 (Scalia dissenting).

     33     306 US 1 (1939).

     34     Id at 15–18.

     35     Id at 5–6, citing Act of August 23, 1935, 49 Stat 731, codified as amended at 7 USC § 511 et seq.

      36     7 USC § 511d.

      37     Currin, 306 US at 17.

      38     Id at 15.

      39     Such a doctrine could not be justified on the ground that government and private power are indistinguishable, because the legal authority created by genuinely private rights does not need to be governed by an intelligible principle. For example, holders of federal patents decide whether to assign or license their patents for their own reasons, not reasons given to them by statute. The requirement that grants of regulatory authority to private people have an intelligible principle thus calls for a distinction between those grants and truly private rights.

     40     See Sunstein and Vermeule, 82 U Chi L Rev at 424–25 (cited in note 3).

     41     See Central Hudson Gas & Electric Corp v Public Service Commission of New York, 447 US 557, 562–64 (1980).

     42     Pub L No 79-404, 60 Stat 237 (1946), codified as amended in various sections of Title 5.

     43     See Ronald M. Levin, et al, A Blackletter Statement of Federal Administrative Law, 54 Admin L Rev 1, 42–43 (2002).

     44     By limiting my argument to those cases’ status as administrative law, I do not mean to imply that they are incorrect under the Supreme Court’s commercial speech doctrine; I have not undertaken to resolve that question in this Essay. The Court’s commercial speech cases have been criticized as quite intrusive into regulation, so it is certainly possible that the DC Circuit decisions that Sunstein and Vermeule discuss are both intrusive and proper applications of precedent. See, for example, Nat Stern, In Defense of the Imprecise Definition of Commercial Speech, 58 Md L Rev 55, 72–75 (1999) (compiling a list of such criticisms).

     45     Mortgage Bankers, 2015 WL 998535 at *3.

     46     117 F3d 579 (DC Cir 1997).

     47     Id at 586.

     48     720 F3d 966 (DC Cir 2013).

     49     Id at 969 (discussing the elements of the Paralyzed Veterans analysis). See also Alaska Professional Hunters Association, Inc v Federal Aviation Administration, 177 F3d 1030, 1033–36 (DC Cir 1999) (applying Paralyzed Veterans to find invalid Federal Aviation Administration regulations).

     50     See Paralyzed Veterans, 117 F3d at 588 (finding that the agency interpretation at issue was “not sufficiently distinct to the regulation to require notice and comment”).

     51     See, for example, Halbig v Burwell, 758 F3d 390, 414 (DC Cir 2014) (Edwards dissenting):

The majority opinion ignores the obvious ambiguity in the statute and claims to rest on plain meaning where there is none to be found. In so doing, the majority misapplies the applicable standard of review, refuses to give deference to the IRS’s and HHS’s permissible constructions of the ACA, and issues a judgment that portends disastrous consequences.

     52     Mortgage Bankers, 2015 WL 998535 at *3.

     53     Id at *10 (noting that because the parties litigated the regulation as an interpretative rule, the Supreme Court would not now reclassify it as legislative).

     54     435 US 519 (1978).

     55     Id at 524 (“Agencies are free to grant additional procedural rights in the exercise of their discretion, but reviewing courts are generally not free to impose them if the agencies have not chosen to grant them.”).

     56     325 US 410 (1945).

     57     519 US 452 (1997).

     58     See Seminole Rock, 325 US at 414 (noting that an administrative interpretation of a regulation is “of controlling weight unless it is plainly erroneous or inconsistent with the regulation”); Auer, 519 US at 461, quoting Seminole Rock, 325 US at 414 (noting that the Secretary of Labor’s interpretation of his own regulations is “controlling unless plainly erroneous or inconsistent with the regulation”) (quotation marks omitted). Justice Alito, concurring in Mortgage Bankers, did not dismiss those concerns but noted that the Paralyzed Veterans doctrine was not “a viable cure for these problems.” Mortgage Bankers, 2015 WL 998535 at *10 (Alito concurring). Justice Scalia, concurring in the judgment, stated that the DC Circuit’s approach was “a courageous (indeed, brazen) attempt to limit the mischief” caused by Seminole Rock and Auer, but ultimately found it “unlawful.” Id at *12 (Scalia concurring). Justice Clarence Thomas thought the lower court’s doctrine “inconsistent with the Administrative Procedure Act” and concluded that it “must be rejected.” Id at *13 (Thomas concurring). He was concerned that Seminole Rock and Auer might be unconstitutional because they effect “a transfer of the judicial power to an executive agency.” Id (Thomas concurring).

     59     Paralyzed Veterans, 117 F3d at 586–87.

     60     Mortgage Bankers, 2015 WL 998535 at *6.

     61     Id at *10. The Court did not decide whether the argument had also been waived below. Id (“[E]ven assuming MBA did not waive the argument below, it has done so in this Court.”).

     62     See id at *3–4 (discussing these two types of rules as distinguished by the APA).

     63     The APA does refer to substantive rules, which include interpretative rules. 5 USC § 553(d). Nonsubstantive rules apparently include some rules of “agency organization, procedure, or practice,” which need not be promulgated with notice and comment. 5 USC § 553(b).

     64     See Mortgage Bankers, 2015 WL 998535 at *3 (noting that legislative rules have the force and effect of law).

     65     See id at *4 (noting that interpretative rules “do not have the force and effect of law”) (quotation marks omitted).

     66     See Seminole Rock, 325 US at 414 (stating that an administrative interpretation of a regulation is “of controlling weight unless it is plainly erroneous or inconsistent with the regulation”).

     67     See Mortgage Bankers, 2015 WL 998535 at *3 (observing that “the notice-and-comment requirement ‘does not apply’ to ‘interpretative rules, general statements of policy, or rules of agency organization, procedure, or practice’”).

     68     5 USC § 553(b).

     69     See Sunstein and Vermeule, 82 U Chi L Rev at 455 & n 295 (cited in note 3) (collecting cases).

     70     Id at 456.

     71     See id at 401 (“While most of the decisions that we discuss cannot quite be described as lawless, some can, and as a whole they go beyond the boundaries of appropriate interpretation of the law as it now stands.”).

     72     733 F3d 1 (DC Cir 2013).

     73     Id at 10.

     74     648 F3d 848 (DC Cir 2011).

     75     Id at 855–56.

     76     Sunstein and Vermeule, 82 U Chi L Rev at 462–63 (cited in note 3) (“[A] consistently libertarian judge would be inclined to . . . conclude that the presumption of reviewability is not overcome in the case of environmental enforcement, even though it had been in CookEt voilaSierra Club, decided in 2011.”).

     77     Id at 458.

     78     Cook required an agency to limit the rights of property owners. Sunstein and Vermeule argue that, “[i]n substantive terms, the decision is classically libertarian; opposition to the death penalty is a cause on which many libertarians of left and right converge.” Id at 462. The fact that a position is held by many libertarians does not make it libertarian, and Sunstein and Vermeule do not say that any of the judges who decided Cook are themselves opposed to capital punishment. For example, the late Professor Robert Nozick believed that “some deserve to die, to be killed, in punishment for their actions.” Robert Nozick, Philosophical Explanations 377 (Harvard 1981). Nozick explained that he had alternated in support for and opposition to the institution of capital punishment because he was not sure whether it showed adequate respect for the value of the guilty person, even though the guilty person deserves to die. Id at 378. It is also possible to believe that some wrongdoers deserve to be killed and that killing them is consistent with respecting their value, but to nevertheless oppose capital punishment in practice on the ground that the government is too likely to make mistakes in applying it. The belief that governments make mistakes, though widely held among libertarians, is hardly a distinctively libertarian position. That belief certainly could lead a nonlibertarian who approves the death penalty in principle to oppose it in practice. One thus might say that opposition to the death penalty is a classically government-skeptical position, because opposition to it is one on which those of either libertarian or nonlibertarian persuasion who believe that governments make mistakes converge.

     79     Pub L No 75-717, 52 Stat 1040 (1938), codified as amended at 21 USC § 301 et seq.

     80     Cook, 733 F3d at 3.

     81     21 USC § 381(a)(4).

     82     See Cook, 733 F3d at 4.

     83     See id.

     84     See id.

     85     470 US 821 (1985).

     86     421 US 560 (1975).

     87     Id at 568.

     88     See id at 560 n 2.

     89     Pub L No 86-257, 73 Stat 519, codified as amended at 29 USC § 401 et seq.

     90     Chaney, 470 US at 833, quoting 29 USC § 482 (ellipses in original).

     91     Chaney, 470 US at 832.

     92     See Cass R. Sunstein, Reviewing Agency Inaction after Heckler v. Chaney, 52 U Chi L Rev 653, 670 (1985).

     93     See Chaney at 832–36, citing Bachowski, 421 US 560.

     94     See Chaney, 470 US at 835, quoting 21 USC § 372.

     95     Chaney, 470 US at 835 (ellipses in original).

     96     Id, citing 18 USC § 471 (counterfeiting), 18 USC § 1001 (false statements to government officials), and 18 USC § 1341 (mail fraud).

     97     Chaney, 470 US at 835.

     98     Cook involved 21 USC § 381, a provision of the FDCA. Cook, 733 F3d at 3. Chaney involved 21 USC §§ 352 and 355, also provisions of the FDCA. Chaney, 470 US at 823–24. Bachowski dealt with 29 USC § 482, a provision of the LMRDA. Bachowski, 421 US at 566, 569.

     99     Compare 21 USC § 381(a) (“If it appears from the examination of such samples or otherwise that . . . such article is adulterated, misbranded or [an unapproved new drug] . . . then such article shall be refused admission.”), with 29 USC § 482 (“The Secretary shall investigate such complaint and, if he finds probable cause to believe that a violation . . . has occurred . . . he shall . . . bring a civil action.”).

     100   21 USC § 381(a).

     101   Cook, 733 F3d at 11.

     102   Chaney, 470 US at 831–32.

     103   The FDA receives notifications from the Customs Service of the entry of articles regulated by the FDA and electronically screens those entry data against the criteria for lawful entry. See Cook, 733 F3d at 3–4. Adding a drug to the screening list is an inexpensive step.

     104   Chaney, 470 US at 835.

     105   See id.

     106   Id, citing 18 USC § 471, 18 USC § 1001, and 18 USC § 1341.

     107   The argument for an implicit command to prosecute is that the provision says that whoever commits the offense shall be fined or imprisoned, that offenders can be punished by the courts only if they are first prosecuted by the executive, and that the executive must therefore prosecute everyone who commits the offense. That conclusion is implausible, primarily because the executive does not have the capacity to identify or to prosecute every offender. A simpler solution is to say that provisions like this one have a modifier implied by the context of an imperative addressed to the courts: every offender who is identified as such by the judicial process (that is, everyone who is convicted) shall be fined or imprisoned.

     108   See 42 USC § 7477.

     109   Sierra Club, 648 F3d at 856 (quotation marks omitted).

     110   See Levin, et al, 54 Admin L Rev at 42–43 (cited in note 43).

     111   See Sunstein and Vermeule, 82 U Chi L Rev at 460–61 (cited in note 3).

     112   In most years, the Supreme Court grants certiorari on less than 4 percent of petitions filed. See Richard J. Lazarus, Advocacy Matters before and within the Supreme Court: Transforming the Court by Transforming the Bar, 96 Georgetown L J 1487, 1493 (2008).

     113   The question of reading precedents may well be especially important to the DC Circuit. In some areas, especially criminal procedure, the Supreme Court decides enough cases on sufficiently different sets of facts that lower courts can often rely on factual analogies. Many other decisions by the Court resolve very specific legal questions—often ones of statutory interpretation—that then apply straightforwardly in many subsequent lower court cases. But when reviewing a decision by the DC Circuit, the Court often addresses a general question of administrative law that it will not revisit for years or even decades. Principles of case reading are especially useful to a court that must apply a small body of precedents addressing general issues.