Note from the Editor:
This article argues that, while judicial deference to agency decisions is often appropriate and constitutional, Auer deference gives agencies too much unreviewable discretion such that it violates the Constitution’s separation of powers.
The Federalist Society takes no positions on particular legal and public policy matters. Any expressions of opinion are those of the author. Whenever we publish an article that advocates for a particular position, as here, we offer links to other perspectives on the issue, including ones opposed to the position taken in the article. We also invite responses from our readers. To join the debate, please email us at firstname.lastname@example.org.
Together with the better-known Chevron deference rule, the doctrine articulated in Auer v. Robbins two decades ago—which makes reasonable administrative constructions of ambiguous administrative rules binding on courts in most circumstances—has become a focal point for concerns about the expanding administrative state. For good reason. Auer deference, even more than Chevron deference, enlarges administrative authority in ways at odds with basic constitutional structures and due process requirements.
I. Chevron: Deference from Lawful Delegation of Discretion
Chevron deference, named for the Supreme Court’s decision in Chevron U.S.A. v. National Resources Defense Council, Inc., requires federal courts to defer to reasonable agency decisions implementing an agency’s statutory mandate when the particular statutory instruction being implemented is ambiguous and not clearly at odds with the agency’s actions. Although often described (and sometimes applied by reviewing courts) as if courts were directed to defer to administrators’ interpretations of law, in essence Chevron—at least as originally constructed—tells courts to decide what laws mean, and only to defer to agencies’ decisions when courts determine that Congress did not speak to an issue.
In keeping with the terms of the Administrative Procedure Act (APA), Chevron also reads the law on judicial review as directing courts to assess some administrative decisions only for their reasonableness, not their correctness. When a court concludes that Congress did not specifically instruct an administrative agency on what to do—instead granting the agency discretion with respect to some aspect of its implementation of the law—courts then should check the agency’s exercise of discretion for its reasonableness and consistency with the limits of the law, not its consistency with judges’ view of better policy. In other words, courts decide what the law means, including the scope of discretion granted to administrators, while administrators make policy decisions when Congress gave them discretion under law. As amply documented, that was the original understanding and evident intent of the Chevron decision.
Chevron changed the law slightly by explicitly assuming that legislation’s ambiguity or silence on a given issue generally should be seen as granting the agency responsible for implementing that law discretion to take any action—to make any policy choice—as long as it is reasonable and not outside the scope of the law’s grant of discretion. Ambiguity, in other words, carried an implicit grant of authority for administrators to make decisions when implementing ambiguous directives—but only so far as they did not contradict the judicially determined limits of the law. At times, the Court has said this quite clearly, as it did, for example, in Smiley v. Citibank (South Dakota), N.A. Even before Chevron, the Court at times treated legislative authorization of agency action in broad, vague terms as evidence of an understanding that Congress was giving the relevant agency (or the President) authority to make policy within the scope of those terms. Chevron simply generalized that inference.
Understood this way, Chevron allows administrators to make different policy choices over time, recognizing distinct (and constitutionally appropriate) roles for the legislature, the executive branch, and the courts. This “original Chevron,” thus, is consistent with the scope of authority delegated to administrators by law; it also is consistent with the Constitution’s separation of powers among the three branches.
Chevron should be—and certainly has been—criticized for its unclear language and its often confused and confusing application by courts (“Chevron-in-practice”). There is ample reason to think that Chevron’s direction to reviewing courts should be abandoned in favor of the terms of the APA. But original Chevron should be applauded, at least, for rooting its understanding of the role of administrative authority in legally prescribed delegation.
Under original Chevron, as under the APA, deference follows delegation. Delegation is an essential prerequisite for deference—but not necessarily the whole game.
II. Auer Deference: Self-Delegation and Due Process Problems
Auer deference as framed by the Court’s decision—what could be termed “original Auer”—is markedly different in kind from original Chevron, although a version of Auer deference is not (if modified to limit the doctrine’s scope, make sure it is used only when there is a statutory commitment of deference, and tailored to minimize opportunities for unfair surprise in agency interpretations). Original Auer—strong deference to any agency interpretation of ambiguous agency rules, no matter the nature of the ambiguity or the means or timing for a later interpretation—was represented as similar to Chevron. But ambiguity in rules adopted by an agency cannot plausibly be evidence of a congressional commitment of authority to the agency. And administrative officials cannot confer additional discretionary authority on themselves. If judicial deference follows from legal delegation of discretionary authority to administrators—as in original Chevron—that delegation must be found in statutory or constitutional provisions, not in unclear agency rules. This connection between deference and delegation is the key to understanding what is wrong with Auer deference and why seemingly similar deference such as original Chevron nonetheless is sound. Before turning further to that issue, it is worth considering other objections to Auer deference.
Objectors to Auer have given cogent reasons why courts should not grant deference to administrative interpretations merely because an agency’s rule is unclear. The most commonly voiced objections implicate, or directly invoke, due process concerns. One well-known objection—advanced notably by Professor (now Dean) John Manning—focuses on potential partiality, a corollary of permitting the body that writes rules to interpret them. The framers of the Constitution sought to avoid the potential for partiality inherent in this kind of arrangement by ensuring that legislative and judicial powers would reside with separate branches of the federal government. Another frequently voiced objection to Auer deference emphasizes the risk that an agency could revise its interpretation in ways that would unfairly surprise those who must comply with its rules—and might even choose less clear rule formulations for the purpose of providing leeway for different, not wholly foreseeable, applications. Judicial deference to agency rule interpretations reduces protections against these potential problems.
These objections provide a reasoned basis for skepticism about permitting administrators both to write rules and to interpret and apply rules. That skepticism, however, should not be the basis for barring any legal commitment of discretion to administrators to perform both functions and to receive deference for both. Neither objection, that is, explains why Congress should be disabled in all instances from granting administrators discretionary authority over rule interpretation, even in settings that do not carry serious risks of partiality or unfair surprise in administrative construction.
III. Administrative Discretion Explained: Legitimate Bases for Deference
There surely are legitimate reasons for granting discretion to administrators. Congress reasonably can conclude that certain decisions require confidential information, involve judgment calls on how to allocate agency resources, or will be better if informed by special expertise or experience. In such cases, delegations of decision-making authority are sensible and consistent with our constitutional structure.
Decisions necessarily based on information that cannot be widely disclosed—such as national security considerations relevant to selection, assignment, and retention of officials at the Central Intelligence Agency—properly can be assigned to the discretion of the relevant administrators. (These decisions have, in fact, been committed to CIA officials’ judgment by law.) The Supreme Court’s decision in Webster v. Doe, and especially the separate opinions by Justices Antonin Scalia (dissenting) and Sandra Day O’Connor (dissenting in part), recognizes the role of discretion in these judgments.
Discretionary authority also can be appropriate for determinations calling on judgments about the best use of agency resources or the best route to implement enforcement activities which require a balance of priorities and personnel and assessment of the effectiveness of alternative enforcement approaches. While prosecutorial discretion poses its own set of problems, the complex set of managerial and policy considerations relevant to prosecution decisions explains why courts, including the Supreme Court in Heckler v. Chaney, have deferred to administrators’ judgments on these matters.
Further, some determinations draw on technical, scientific, or experiential judgments that can be assigned to administrators’ discretion. Where in Washington, D.C., should a government building be located? Which building materials are best suited to a given structure in a particular climate? How should contracts for different quantities of and delivery schedules for concrete be compared to establish the comparability of the prices charged? What separation and siting of particular broadcast outlets best assures signal clarity and coverage? Such questions are most sensibly answered drawing on information or analysis more accessible to administrators than to reviewing judges. These sorts of questions frequently are presented in applying rules for decision as well as in framing regulations to guide future decisions.
Saying that some decisions sensibly can be committed to administrators’ discretion does not mean that any particular decision in fact has been committed to administrators’ discretion. Courts also should not assume simply because it would be reasonable to grant administrators discretion that Congress has provided statutory authority to exercise that discretion. But the fact that some decisions are sensibly committed to administrators’ discretion helps explain why sometimes deference to administrative rule implementation decisions might be required by law.
That was true in the case the Auer Court thought it was following, Bowles v. Seminole Rock & Sand Co. Seminole Rock was a dispute concerning application of rules for comparing prices in different contracts under the war-time Price Control Act of 1942—a dispute that should have been (and, judging from most of the language in the Court’s opinion, actually was) decided on grounds far narrower than the broad deference rule announced by the Court and later relied on as precedent for Auer. In fact, neither the Seminole Rock Court nor the Auer Court needed to invoke a rule of general deference to administrative interpretation of regulations, as both had other strong reasons for affirming the agency’s decision.
Critical readings of Seminole Rock reveal that the Price Control Act did give substantial discretionary authority to the Office of Price Administration (OPA) and that the OPA’s reading of the rule had virtually every hallmark of a rule deserving deference. It was sensible, consistent with the text and with OPA practice, announced contemporaneously with the issuance of the rule, and publicized together with the rule. In fact, Seminole Rock presented the best possible case for deference: actual delegation of discretion in OPA’s implementation authority, no significant risk of partiality in how it interpreted and implemented the agency’s rules, and no risk of unfair surprise—a risk that was uniquely absent because of the simultaneous issuance and publication of the rule and its interpretation. In other words, the case that first articulated the rule adopted uncritically in Auer elided all of the principal objections to its application.
IV. Law-Making, Separation of Powers, and Delegation
Yet there is an additional consideration that was not discussed in either Auer or Seminole Rock: whether the delegation of authority itself is constitutional given the powers vested in each branch of government. The Constitution assigns separate powers to each of the three branches of the federal government, vesting legislative power in Congress, executive power in the President, and the judicial power in the Supreme Court and other courts created under Article III.
The terms of the vesting clauses are clear and instructive. While all three vesting clauses assign exclusive authority to one branch, unlike the unconditional assignments of executive and judicial power in Articles II and III, Article I vests in Congress only “all legislative Powers granted herein.” The limitation in that vesting clause emphasizes that the national government lacks the plenary powers of state governments and other inherently sovereign governments and also underscores the framers’ belief that the most fearsome and dangerous authority is the legislative power. James Madison, in Federalist No. 48, captured that sentiment in explaining the need for special constraints on the legislature, as historical observation revealed that the legislative branch was “everywhere extending the sphere of its activity and drawing all power into its impetuous vortex.”
Concern over legislative power explains why this power was subjected to special constraints. It can only be exercised through the agreement of majorities in both houses of Congress. The two houses of Congress are composed of representatives selected in different ways and at different times, to serve differently configured constituencies for different lengths of time, diffusing legislative power and guarding against ill-considered measures backed by temporary majorities moved by ephemeral passions. And the laws that survive the legislative gauntlet must be presented to the President for approval or veto. Madison, in Federalist No. 51, describes these precautions as a “remedy” for the fact that “[i]n republican government, the legislative authority necessarily predominates.”
Those who framed and ratified the Constitution took special pains to assure that the branches stayed within their assigned roles, seeking to “maintain in practice the necessary partition of power among the several departments.” They stressed that this most importantly included seeing that the carefully constructed limitations on the exercise of legislative powers would not be evaded. Congress cannot short-circuit the law-making process by lowering the vote needed to pass legislation, by allowing one house of Congress acting alone to pass laws, or by providing for law-making that bypasses presentment to the President. That is the understanding, for example, behind the Supreme Court’s decision in Immigration and Naturalization Service v. Chadha, which declared that Congress could not exercise a one-house veto of administrative decisions. The Court found that this veto procedure—which was defended as simply placing conditions on the exercise of lawfully granted administrative authority—amounted to making law without the constitutionally required procedures of bicameralism and presentment.
The same understanding applies to efforts to give another government official authority that is tantamount to law-making power, no matter what formal characterization is given to it. That is why the Supreme Court in Clinton v. New York found unconstitutional the Line Item Veto Act of 1996, which gave the President authority to veto three specific types of expenditure or tax benefit that had been conferred by a statute the President decided to sign into law. The Act effectively granted the President power to rewrite specific laws rather than accepting or rejecting them—and the Court appreciated that rewriting a law is no different than writing it in the first instance. The Court that decided the Clinton case understood that congressional passage of line-item veto authority was not a charitable act by which members of Congress ceded some of their power to the President. Instead, it was a means for advantaging certain interests and disadvantaging others—and, in exactly the same way, for advantaging some members of Congress and disadvantaging others. But more importantly, it was a means for evading constitutional constraints on law-making.
Together, the Chadha and Clinton cases stand for the proposition that, when Congress grants itself an exemption from ordinary law-making procedures, grants subordinate parts of the Congress law-making authority, or grants another official or entity parts of that authority, it is evading constitutionally required processes. However the evasion takes place, it is not permitted.
The lesson of Chadha and Clinton similarly undergirds the delegation doctrine (also referred to as the “non-delegation doctrine”). Although the essence of the doctrine was articulated much earlier, its classic formulation was given in Justice John Marshall Harlan’s opinion for the Court in Field v. Clark: “That Congress cannot delegate legislative power to the President is a principle universally recognized as vital to the integrity and maintenance of the system of government ordained by the Constitution.” The separate opinion of Justice Lucius Lamar, for himself and Chief Justice Melville Fuller, similarly declared: “That no part of this legislative power can be delegated by Congress to any other department of the government, executive or judicial, is . . . universally recognized as a principle essential to the integrity and maintenance of the system of government ordained by the Constitution.”
Perhaps the best explanation for the delegation doctrine, however, is contained in Chief Justice John Marshall’s opinion for the Court in Wayman v. Southard: “It will not be contended that Congress can delegate to the Courts, or to any other tribunals, powers which are strictly and exclusively legislative. But Congress may certainly delegate to others, powers which the legislature may rightfully exercise itself.” Powers that fall into the first category—powers that Congress must exercise itself and cannot delegate to other officials—involve making rules on matters of such importance that they “must be entirely regulated by the legislature itself.” The second category is comprised of subjects “of less interest,” where Congress properly may make “general provisions” and leave it to others to “fill up the details.”
Unfortunately, although Wayman captures the approach taken for the nation’s first (almost) 150 years, the Court, in J.W. Hampton, Jr., & Co. v. United States, gave a different explanation of what the Constitution requires. Chief Justice (and former Chief Executive) William Howard Taft, writing for the Court, stated: “Congress has found it frequently necessary to use officers of the executive branch within defined limits, to secure the exact effect intended by its acts of legislation, by vesting discretion in such officers to make public regulations interpreting a statute and directing the details of its execution . . . .” Distinguishing what is constitutionally permitted from what is forbidden, Taft wrote: “If Congress shall lay down by legislative act an intelligible principle to which the person or body authorized to [implement the law] . . . is directed to conform, such legislative action is not a forbidden delegation of legislative power.”
Following Hampton, the Court has only twice—never after 1935—found that legislation did not contain “an intelligible principle.” Even assignments that merely instructed agencies to act as “the public interest, convenience, and necessity” require or to set prices that are “generally fair and equitable” have passed the test, along with a lot of other vague, multi-faceted, amorphous directives. The result has been a virtual abandonment of serious attention to the way legislative commitments of authority to administrative officials fit (or do not fit) the Constitution’s divisions of governmental powers.
V. Auer Deference’s Larger Delegation Problem
The loss of a serious, direct judicial brake on legislative grants of power to administrators has permitted the enormous expansion of government regulation of the economy and of many aspects of health, safety, and personal behavior (retirement savings, family and child-raising decisions, and much more). Much of this regulatory structure covers subjects that long had been thought beyond the ambit of federal power.
In addition to long, detailed, and often internally inconsistent statutes, huge portions of the network of regulatory controls has come from officials who were not elected to Congress (or, indeed, elected at all) but who have been deputized to implement the laws, both through specific applications of law in particular settings and through adoption of more general, law-like rules. The Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, for example, which covers more than 2,000 pages, has given rise to more than 22,000 pages of implementing rules. Today’s Code of Federal Regulations contains roughly 200,000 pages of agency-generated rules—approximately nine to ten times as many pages as the congressionally-passed laws collected in the United States Code—that are enforced through threat of criminal punishment, civil penalties, denial of valuable privileges, loss of benefits, and damaging publicity.
This does not mean that there is no judicial control over administrators’ exercise of law-making authority. Courts have interpreted particular statutes as inconsistent with specific assertions of agency authority. The Supreme Court, for example, in Food and Drug Administration v. Brown & Williamson Tobacco Corp., rejected the FDA’s assertion of power to regulate tobacco sales under the Food, Drug and Cosmetic Act. The FDA had suddenly discovered this power almost 60 years after the Act’s adoption and despite the repeated failure of efforts to gain congressional approval for tobacco regulation. Courts are unlikely to defer to agency claims of authority that would effect major changes in the law without clear statutory basis, especially when the claims are inconsistent with long-standing interpretations of the law. The interpretive canon behind this inclination was pithily captured in Justice Scalia’s observation that “Congress . . . does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions—it does not, one might say, hide elephants in mouseholes.”
VI. Discretion Consistent with Constitutional Authority
Legislative commitments of power to administrators, however, include both constitutional delegations of authority—even expansive ones—and delegations that exceed the constitutional safeguards provided by bicameralism and presentment. Look first at the positive side of the ledger. Consider, for example, the authority given to the Central Intelligence Agency for certain national security matters. Among other things, the CIA’s Director is instructed to “protect intelligence sources and methods from unauthorized disclosure.” The Supreme Court has upheld exercises of that authority, deferring to various decisions by the Agency. It upheld the CIA’s use of contracts requiring employees to protect secret information and to secure permission before publishing material that might disclose such information. It also deferred to the Agency’s refusal of requests for release of information under the Freedom of Information Act, such as identities of sources, that the Agency deemed would put national security at risk.
The Webster case involved another matter in which statutory authority grants the CIA discretion. The case addressed the CIA’s decision to dismiss an employee after the Office of Security and CIA Director determined that his continued employment posed a potential threat to national security. The National Security Act of 1947 specifically grants discretion over these matters, declaring: “Notwithstanding . . . the provisions of any other law, the Director of Central Intelligence may, in his discretion, terminate the employment of any officer or employee of the Agency whenever he shall deem such termination necessary or advisable in the interests of the United States.” The law makes clear that the determination is committed to agency discretion and that the courts should defer to the exercise of that discretion.
As Justice Scalia explained in that case, several considerations reinforce the basis for deferring by removing the question from the scope of judicial authority to review altogether. First, the issue on which review was sought was expressly committed to the administrator’s discretion. Second, the law containing the directly relevant provision gave “extraordinary deference” to the Director. And, third, “the area to which the text pertains is one of predominant executive authority and of traditional judicial abstention.” Scalia concluded that “it is difficult to conceive of a statutory scheme that more clearly reflects . . . ‘commit[ment] to agency discretion by law’ . . . .” The third consideration was central to the arguments of both Justice Scalia and Justice O’Connor in Webster. They both emphasized that it is not merely the nature of the statutory commitment but the consistency of that commitment with constitutional assignments of authority among the branches. Justice O’Connor, focusing on the concept behind the third reason for deference given by Justice Scalia, declared:
The functions performed by the Central Intelligence Agency and the Director of Central Intelligence lie at the core of “the very delicate, plenary and exclusive power of the President as the sole organ of the federal government in the field of international relations.” United States v. Curtiss–Wright Export Corp., 299 U.S. 304, 320 (1936). The authority of the Director of Central Intelligence to control access to sensitive national security information by discharging employees deemed to be untrustworthy flows primarily from this constitutional power of the President, and Congress may surely provide that the inferior courts are not used to infringe on the President’s constitutional authority.
In the same vein, Justice Scalia emphasized that there are “certain issues and certain areas that [are] beyond the range of judicial review,” including those so intimately bound to matters within the constitutional domain of the executive branch that insulation from review reflects “a traditional respect for the functions of the other branches.”
Determinations that are committed to agency discretion by law or are of such a nature—given constitutional assignments of executive authority and the type of discretionary judgments necessary to effectuate them—that they are incompatible with judicial review are matters on which courts should defer. This deference properly encompasses not only the act of crafting regulations but also of implementing administrative authority through decisions in individual cases, through interpretation of agency rules, or through combinations of all of these activities. The sort of deference addressed in Auer is appropriate here—not because an agency that writes a rule should be able to interpret the rule, but because any agency action strictly within the executive’s domain, especially when buttressed by statutory confirmation of the understanding that this is distinctly the executive’s role, is properly within the agency’s discretion. When authority has been constitutionally delegated, courts should accept the exercise of discretion attached to that delegation and defer to it.
VII. Discretion at Odds with Constitutional Authority
In contrast, some assignments of discretionary authority to administrators plainly are in tension with—if not wholly in violation of—the constitutional division of roles for the different branches. As Chief Justice Marshall said in Wayman almost 200 years ago, the legislative power vested in Congress encompasses decisions on all “important subjects, which must be entirely regulated by the legislature itself . . . .” Executive branch officers may make rules for less important matters; and judges, in the course of deciding cases, may articulate rationales for their decisions that guide future decisions. But the truly important choices—those that are most politically-freighted, that are most seriously contested, and that have the most significant impact on society—cannot be left to other officials any more than they could be made by the legislature without bicameral agreement and presentment to the President. That especially applies to choices involving “rules for the regulation of the society,” which Alexander Hamilton termed the “essence of the legislative authority.”
The point applies even more broadly. Not only is Congress charged with making big decisions about the ordering of American life, it must decide important aspects of how those big decisions will be implemented—those aspects of making law can be just as important. But the less important details of implementation surely can be assigned to others. If Congress provides benefits for veterans or for citizens with serious, work-limiting disabilities, it does not need to decide every benefit claim or prescribe every detail for how to evaluate medical evidence or what weight to give evidence submitted by treating physicians relative to evidence from other authorities. But it cannot simply declare that veterans or disabled Americans should be taken care of and leave the when, what, and how to administrators. Similarly, if congressional majorities conclude that they should prevent fraudulent or misleading communications with consumers from harming the economy, Congress cannot simply grant an administrative agency broad authority “to prevent ‘unfair, deceptive, or abusive [consumer financial] acts or services’” and let the agency decide what acts or services it considers to be bad and what to do about it.
In the absence of a serious judicial doctrine limiting the scope of delegation, laws have effected large-scale transfers to administrators of authority to make rules over critically important and politically salient issues. So, for example, the Consumer Financial Protection Bureau was given authority to regulate anyone who offers or provides “a consumer financial product or service” in the broad, ambiguous terms quoted above. Worse, some agencies have asserted authority over matters doubtfully within their delegated powers. The FDA’s claimed authority over tobacco sales, rejected in the Brown & Williamson case, is one example. The Federal Communications Commission’s assertion of authority to regulate the pricing practices of internet service providers is another. Deferring to administrators’ decisions on such matters exacerbates the problem of judicial unwillingness to insist that important choices for regulation of private conduct be made by Congress through constitutionally-mandated processes. This point has been made by many scholars and jurists objecting to Chevron-in-practice: when agencies make critical policy decisions on important matters, they exercise legislative authority.
The problem is far greater, however, when deference is extended to rule interpretations in the way Auer requires. Auer expands the range of agency decisions to which courts should defer from first-level actions that directly implement constitutionally-questionable grants of authority based on statutes to second-level actions that implement agency rules. While Chevron provides for deference to an agency’s initial policy choices made in framing rules—primarily through “notice-and-comment” rulemaking processes designed to elicit relevant information, to allow expansive public participation in the rule-framing process, and to provide some degree of advance warning on how the agency will act—Auer requires deference to follow-on choices made in an array of rule interpretations and applications, generally using quite different processes that do not contain the features of notice-and-comment.
If administrative law-making is problematic, allowing administrators to remake the law repeatedly—to revise the meaning of agency-made law through new interpretations of admittedly unclear agency rules—should be doubly problematic. Consider, for example, the Department of Education’s change in interpretation of a Title IX regulation respecting segregation of school bathroom and locker room facilities by sex. The Department used a private letter ruling to dramatically alter an unambiguous rule some four decades after its adoption. The decision of a Fourth Circuit panel to defer to that interpretation certainly was a questionable extension of Auer, but it illustrates the potential range of decisions that Auer (or Auer-like) deference can entail—across time, over different vehicles for announcing changes in administrative position, and across different views of appropriate assertions of government power.
Deference to second-level agency decision-making enlarges the set of agency determinations subject to presumptive authority and vastly enlarges the set of such determinations claiming policy-based deference that are likely to be made without the sorts of procedures generally deemed best suited to informing both the administrative decision-makers and the public. The result is not only that law is made by the wrong officials without the processes constitutionally required for making law by Congress; it is made without even using the processes that agencies are supposed to utilize in writing substantive rules.
Examining the relationship between statutorily-directed deference and constitutional-structural principles clarifies the essential objection to Auer and the limits of that objection. When Congress by law confers discretionary authority that does not exceed its constitutional power to delegate functions to an administrator, courts should respect that assignment of authority unless it violates other specific constitutional commands. A different rule should apply when delegations are at most only arguably consistent with the Constitution. When that is the case, it means that the delegations probably are not consistent with the Constitution, even if they comply with the delegation doctrine as it has been interpreted by courts over the past eight decades to accommodate the modern administrative state. In this setting, deference—especially the sort of serially expanding deference Auer embraces to cover successive levels of administrative determination—exacerbates the problems with delegation.
A reinvigorated delegation doctrine would solve the major Auer problem directly, and elimination of Auer-like deference would be clearly preferable to retaining the doctrine in its current form. Short of that, demanding that the statutory basis for deference is clearly articulated—that Congress plainly convey authority for administrators to exercise discretion at the second level of administrative rule implementation as well as the first level of more direct statutory implementation—would provide a modest first step in cabining problems associated with constitutionally questionable delegations of law-making authority. Those who embrace the rule of law, whether advocates or opponents of the modern administrative state, should support that step.
 519 U.S. 452 (1997).
 476 U.S. 837 (1984).
 While the APA provides the general rules for federal agency processes and for judicial review of agency action, the terms for review in Chevron were governed instead by the Clean Air Act. The relevant provisions in that law, however, mirror the review provisions of the APA.
 Gary Lawson & Stephen Kam, Making Law Out of Nothing at All: The Origins of the Chevron Doctrine, 65 Admin. L. Rev. 1 (2013); Thomas W. Merrill, Justice Stevens and the Chevron Puzzle, 106 Nw. U. L. Rev. 551 (2012).
 517 U.S. 735, 740–41 (1995).
 486 U.S. 592 (1988).
 470 U.S. 821 (1985).
 325 U.S. 410 (1945).
 This section draws on Ronald A. Cass, Delegation Reconsidered: A Delegation Doctrine for the Modern Administrative State, 40 Harv. J.L. & Pub. Pol’y 148 (2016).
 U.S. Const. art. I, § 1 (emphasis added).
 Federalist No. 51 (James Madison).
 462 U.S. 919 (1983).
 524 U.S. 417 (1998).
 143 U.S. 649, 692 (1892) (emphasis added).
 Field, 143 U.S. at 697 (Lamar, J., concurring in judgment) (emphasis added).
 23 U.S. (10 Wheat.) 1, 42–43 (1825).
 Id. (emphasis added).
 276 U.S. 394 (1928).
 Hampton, 276 U.S. at 406.
 Id. at 409 (emphasis added).
 529 U.S. 120 (2000).
 Whitman v. American Trucking Associations, Inc., 531 U.S. 457, 468 (2001).
 50 U.S.C. §403(d)(3).
 Snepp v. United States, 444 U.S. 507 (1980).
 Central Intelligence Agency v. Sims, 471 U.S. 159 (1985).
 50 U.S.C. §403(c).
 Webster, 486 U.S. at 615–16 (Scalia, J., dissenting).
 Id. at 616.
 Id. at 605–06 (O’Connor, J., concurring in part and dissenting in part).
 Id. at 608–09 (Scalia, J., dissenting).
 Wayman, 23 U.S. (10 Wheat.) at 43.
 Federalist No. 75 (Alexander Hamilton).
 See Recent Legislation: Administrative Law—Agency Design—Dodd-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. 2123, 2125 (2011) (quoting the Dodd-Frank Wall Street Reform and Consumer Protection Act of 2010, §1031(b), 124 Stat. at 2006).
 While some older laws used similarly broad phrasing (think of the Sherman Antitrust Act of 1890, 26 Stat. 209, outlawing “every contract . . . or conspiracy in restraint of trade or commerce among the several states”), those laws tended to be (or, certainly, could and should have been) cabined by the historical usage of the terms, the terms’ common law roots, and the intentionality requirements traditionally required for the courts to conclude that someone had violated such laws.
 See United States Telecom Ass’n v. Federal Communications Comm’n, no. 15-1063 (D.C. Cir. Jun. 14, 2016); Verizon Communications, Inc. v. Federal Communications Comm’n, 740 F.3d 623 (D.C. Cir. 2014); Comcast Corp. v. Federal Communications Comm’n, 600 F.3d 642 (D.C. Cir. 2010).
 See G.G. v. Gloucester County School Bd., 822 F.3d 709, 737 (4th Cir. 2015) (Niemeyer, J., concurring in part and dissenting in part).