Cato Policy Analysis No. 312 August 11, 1998

Policy Analysis

Unlegislated Compulsion:
How Federal Agency Guidelines Threaten Your Liberty

by Robert A. Anthony

Robert A. Anthony is Foundation Professor of Law at George Mason University School of Law in Arlington, Virginia.

Executive Summary

The Constitution contemplates that our federal laws will be made by the people's elected representatives in Congress. But Congress, with the acquiescence of the Supreme Court, has delegated large chunks of lawmaking authority to the executive agencies. While agency-made law is a reality, it is legitimate only where the agency scrupulously observes the procedures Congress has specified for issuing regulations that are intended to have legally binding effect. Those procedures normally include publication of the proposed regulation, receipt of comments from the public, and publication of an explanation accompanying the final action. Regulations so issued are called "legislative rules."

All too often, though, agencies try to lay down binding rules without observing the procedures required by Congress. Those are called "nonlegislative rules." Agencies put forth memoranda, circulars, bulletins, guidance, manuals, press releases, policy statements, staff instructions, and similar informal documents. Even though those documents do not have legally binding effect, they have practical binding effect whenever the agencies use them to establish criteria that affect the rights and obligations of private persons. Except to the extent that such documents merely interpret existing law, it is improper to impose them with binding effect. Indeed, it is nothing short of autocratic or despotic for officials to take the unauthorized action of placing obligations on citizens without honoring the procedural requirements. The courts should strike down such actions, which have no place in our system of limited government under the rule of law.


Can a federal agency official decree that low areas like those on your land are "wetlands," which you cannot develop without a hard-to-get permit? Or declare that, if you try to see a particular doctor of your choice and pay her yourself, the doctor could be penalized under Medicare? Or establish a "model," based on worst-case assumptions, for predicting contamination from regulated waste?

The laying down of law like that should be done by the persons designated for that purpose by Article I of the Constitution, the people's representatives in Congress, the branch vested by the Constitution with "all legislative powers herein granted."

But Congress, abetted by the Supreme Court, has promoted the vast exercise of legislative authority by unelected agencies. If that sort of thing is to be tolerated, it is essential that procedures be implemented to contain the harm that agency officials can do. That is, some means by which federal agencies and officials attempt to lay down rules with binding effect are more undemocratic than others.

Rulemaking that complies with congressional authorizations and procedural mandates is least injurious to our democratic system. Rulemaking done otherwise can be autocracy and even tyranny. Indeed, when government officials announce new rules purporting to tell citizens what they can and cannot do--in a bulletin or a memorandum, for example--without acting within the authorizations and procedures specified by Congress, that is the stuff of dictatorship. A system in which officials freely imposed unlegislated or unauthorized rules through informal documents would be nothing short of despotism.

An indispensable element of a good government for America--one that affirms democracy and rejects dictatorship--is the foundational norm that officials cannot issue decrees without legislative authority.[1] That proposition lies near the heart of our freedoms. It marks a boundary between liberty and oppression. It is a vital element in our civil liberties.

The Courts: Our Bulwark against Agency Overreaching

It falls primarily to the courts, exercising their power to review agency actions, to ensure that the agencies do it right.

This paper will show how the courts can protect citizens against the agency practice of attempting to impose practically binding rules without observing the statutory procedural requirements for imposing binding rules. It also will show how the courts should limit the effects that agencies may legitimately give to documents that are not issued in accordance with those requirements.

Congress in the Administrative Procedure Act of 1946 established procedures that an agency must follow to promulgate binding rules and regulations on subjects within the area delegated to it by statute.[2] Those procedures, usually called "legislative rulemaking procedures," ordinarily require published notice and opportunity for public comment, followed by agency deliberation and explanation and publication of the final rule in the Federal Register. Sometimes an agency's authorizing statute specifies additional or different rulemaking procedures, but the APA pattern is the norm.[3]

If an agency follows the required procedures, and acts within the subject area in which it possesses statutory power to make rules, it is entitled to enforce the resulting rules, subject to certain judicial checks as described below. That is because the agency has acted under the lawmaking authority delegated to it by Congress and has observed the rulemaking procedures set down by Congress. It has legislated new rules.

Such lawmaking by agencies is distinctly a second best to legislation by the people's elected representatives in Congress. It bureaucratically rather than democratically imposes compulsion or mandatory standards on the people. And while those agency processes for developing rules involve some roughly democratic elements--including open notice, limited participation by interested members of the public, responsible agency deliberation and explanation, and official publication--they should be clearly seen for what they are: lawmaking by other than the people's lawmakers.

Congress has built up a habit of delegating large chunks of its constitutionally entrusted lawmaking power to federal agencies. The pernicious result is to vest legislative authority in the hands of unelected administrators, who often brandish enforcement authority and wield adjudicative power as well. The protection of the separation of powers--a centerpiece of our Constitution--is debilitated. Regrettably, the Supreme Court has not read the Constitution as placing any substantial curbs on this congressional bad habit. As a result, delegations have become pandemic across government, and the exercise of legislative powers so delegated fills hundreds of volumes in the Code of Federal Regulations.

Because legislative power has been so copiously transferred to federal agencies, people who would protect freedom must work to guard citizens against the anti-democratic tendencies that arise from delegation to unelected officials. We must ensure that the exercise of delegated legislative power is least offensive to our Constitution and laws.

Delegated power, it is vital to remember, is in all cases limited by the statutes conferring it. [4] An agency of the government cannot announce actions that legally bind citizens unless Congress by statute has delegated it the authority to do so. Nor may it bind citizens if it fails to observe procedures required by statute.

We must look to the courts to be the citizen's protectors against actions that reach beyond an agency's legitimate powers--either with regard to the subject matter covered or with regard to the procedures by which the agency has acted. And it must also be the courts that protect against actions that an agency attempts to apply with more binding force than it has the statutory power to impose. Through rigorous judicial review--available to combat improper agency actions--the courts can ameliorate the undemocratic doctrine that allows Congress to delegate legislative power to unelected agencies.

Manifestly, the citizen's protection relies heavily on having judges who will faithfully apply the law that limits the agencies, rather than judges who reflexively approve restrictive agency actions regardless of their basic legality in order to support results they find politically congenial. At bottom, we have no choice but to assume that our judges will be faithful, and therefore that the judicial system will in fact check and deter unauthorized agency encroachments on the people's freedoms.

Ways the Courts Test Agency Actions

The courts' touchstone for protecting the citizenry is a simple one: agencies may affect persons only to the extent authorized by the Constitution and acts of Congress. Especially must that be true where agencies purport to act upon the rights and obligations of citizens with binding legal effect or with binding practical effect.

Application of that simple precept is, however, complicated. It is complex because there are so many varieties of actions and documents by which agencies try to affect citizens' conduct. The determination of congressional authority in most cases involves a close analysis of the APA as applied in the setting of the agency action or document in question.

A valuable starting place is the realization that the bestowal of statutory authority over a general subject area does not empower the agency to take action regarding that subject in any way it chooses. The APA constrains the agency in three major ways, enforced by the courts through their review of the agency's actions. [5]

First, obviously, the agency must stay within those aspects of the subject area with which it is specifically authorized to deal by statute.[6] Each statute authorizing agency action explicitly or implicitly sets outer limits on the agency's authority.

Second, agency actions, to be valid, must satisfy further substantive conditions. In general terms, the agency action must be reasonable, in that its policy has sufficient factual support[7] and is not "arbitrary, capricious [or] an abuse of discretion."[8] Although the courts' review under those standards is enormously important in discouraging poorly supported agency issuances, it is not the main focus of our attention here.

Of principal importance for this paper is the third category of review standards, those governing the ways in which an agency acts procedurally.[9] It is an essential part of our liberty that agencies, when they impose obligations or standards on private persons, act only through procedures and in forms prescribed by Congress.

That may seem an elementary proposition, and indeed it is. Even though Congress established the normally required procedures in the APA, a tremendous thicket of confusion has arisen about the ways agencies act and about the effect their actions have on private persons and organizations. True, the law often is difficult to apply in specific cases. That makes it all the more important that there be great clarity about the governing procedural concepts, so that they may be applied with precision to particular cases and people can know where they stand.

Informal Agency Documents Purporting to Lay Down Rules

Agency regulations promulgated in accordance with legislative rulemaking procedural requirements will be upheld by the courts, provided that they are consistent with the relevant statutes and are not factually unsupported or otherwise unreasonable. Those are called "legislative rules" and are binding on the public.[10]

But what about "nonlegislative rules"--the primary focus of this paper--that myriad of other agency documents that are not promulgated through legislative rulemaking procedures?[11] May the agencies make the public adhere to "rules" set forth in guidances or memoranda or interpretations or manuals or bulletins or press releases or policy statements or Dear Colleague letters or enforcement guidelines or models or questions-and-answers or action levels or staff instructions or advisory opinions or other such informal documents?

Not only do those multifarious documents come in a variety of formats, they also can issue from an assortment of agency officials, at headquarters or agency field offices, and at levels high or low. One study listed 52 "levels of government," from secretary and deputy secretary down through assistant subdivision director and branch chief. [12] Informal agency documents can potentially be issued at any of those levels, and nobody is necessarily in charge. Only a tiny fraction of those documents is published in the Federal Register. The original northern spotted owl guidelines, for example, were merely photocopies of typewritten or word-processed pages, made available to interested persons by a regional office. [13]

In the usual situation, the mere act of issuing documents like those is not improper. The concern arises from the use that the agencies try to make of such documents, especially in imposing effects upon private parties. As we shall see, the ways in which agencies can legitimately use such documents are limited.

Can agencies treat such documents as binding in enforcement actions or in granting or denying approvals? That is, even though nonlegislative rules by definition do not have the force of law and therefore are not legally binding, can agencies give them practical binding effect by applying them regularly, as fixed criteria, in enforcement or approval proceedings? May an agency tell an applicant that to avoid enforcement action or to gain approval he or she must comply with a given nonlegislative guidance or memorandum or bulletin?

With one exception, the answer to those questions is, and should be, no. Although agencies often try to give binding effect to such unlegislated documents, they are not entitled to do so and the courts should stop them. The exception is for documents that interpret; as explained below, it is not procedurally improper for an agency to try to give them practical binding effect, even though the documents cannot legally bind the courts or the public.

To allow binding rules to be issued in informal documents would remove the incentive for agencies to observe the legislative procedural requirements laid down by Congress. That is true even though the informal documents would have only a practical binding effect rather than the force of law. An agency could accomplish just as stringent and effective a regime of regulation through guidances and memoranda as it could by using the more laborious notice-and-comment procedures set by the APA. The agency would avoid the challenge of allowing public participation and building a worthy rulemaking record that those procedures entail, and the accompanying delay. Regulation would become not only far less disciplined but vastly easier. The temptation to overregulate, and to regulate without balanced information and consideration, would be hard for agencies to resist. The adverse impact upon the liberties of citizens and upon the economy would be incalculable.

Again, in our system of limited government, citizens should not be affected by the federal government beyond the extent authorized by the Constitution and acts of Congress. Freedom from being bound by unlegislated documents issued by unelected agency officials is a basic American liberty.

To understand how the courts should protect that liberty, we must now turn to the complex work of classifying agency actions and documents in accordance with the APA's taxonomy and identifying the standards by which the courts should judge agency actions and documents.

Some Central Legal Distinctions Concerning Agency Rules

As stated above, legislative rules are those promulgated in accordance with statutory requirements for establishing rules carrying the force of law. They have legal binding effect.

Rules not so promulgated are nonlegislative rules. This latter category includes documents of any degree of informality that purport to have general applicability and that implement, interpret, or prescribe law or policy.[14]

Within the realm of nonlegislative rules, those that actually interpret existing statutes or regulations are called interpretive rules.[15] Nonlegislative rules that do not interpret are policy statements, regardless of how they are titled or characterized by the agency.

A central concern of this paper is that agencies often try to give such nonlegislative rules practical binding effect by using them to establish fixed criteria that the agency will consistently enforce.

The APA contains an exemption from its notice-and-comment legislative rulemaking requirements for interpretive rules and policy statements. [16] But this does not mean that documents that come under the exemption can be used by agencies to bind the public. To the contrary, as will be explained, policy statements may not be used to bind, either legally or as a practical matter. Interpretive rules cannot be used to bind legally and can be used to bind practically to only a limited extent.

Policy statements that attempt to impose practical binding effect should be invalidated by the courts on procedural grounds, for failure to observe required legislative rulemaking procedures. Interpretive rules that attempt to impose practical binding effect should not be invalidated on procedural grounds, but should be invalidated whenever the courts disagree with the interpretations they contain.

Those propositions result from the courts' application of the APA in their review of agency actions.

Policy Statements (Documents That Do Not Interpret)

If an agency issues a (noninterpretive) guidance, memorandum, manual, bulletin, or like document with the intent of binding the public, or with the effect of binding the public, the courts will strike it down for nonobservance of legislative rulemaking requirements. Such court actions are fairly numerous, [17] but undoubtedly they deal with only a small part of the manifestly widespread agency misuse of those documents. Here are just a few of those cases:

Through a manual rather than a legislative rule, the Bureau of Indian Affairs denied general assistance to full-blooded unassimilated Indians who lived near the reservation but not on it. The Supreme Court held that the secretary's failure to treat that eligibility requirement as a "legislative-type rule, render[ed] it ineffective." [18]

An owner sought to fill portions of real property for building development. Under the Clean Water Act, dredged or fill material is a pollutant, which may be placed in "navigable waters" only by permit from the Corps of Engineers. The Corps asserted jurisdiction over the land in question on the basis of a memorandum issued by its deputy director. The memorandum declared that isolated waters usable as habitat for migratory birds were "navigable waters." Conceivably that document could have been treated as an interpretation of the term "navigable waters," albeit a stupefyingly strained one. But the court treated it as noninterpretive and, finding that it was intended to be binding, set it aside for failure to observe APA rulemaking procedures.[19]

A demonstrator in Lafayette Park in front of the White House was criminally prosecuted for violating "conditions" restricting storage of property in the park. Because the document setting forth the conditions was not part of the properly promulgated Park Service regulations, the appeals court held it to be "null and void" and reversed the conviction. [20]

The administrator of the Occupational Safety and Health Administration issued a nonlegislative document that would charge employers with discrimination unless they paid wages to union representatives who accompanied OSHA personnel conducting inspections of the employers' premises. The appeals court scathingly denounced that "high-handed rulemaking" as "offensive to our basic notions of democratic government." [21]

Assuredly it was that. Examples could readily be multiplied. Rather often, agencies try to use informal documents to create practically binding new law on the cheap, gambling that affected persons will not take them to court. An applicant for a permit or for a benefit or other approval might well deem it impolitic to sue an agency that possesses the power to deny its pending application or a future one. Thus, the agency may be able to issue an effectively binding document and get away with it. Nevertheless, where the agency has not followed the procedures required by Congress, it has no more authority to impose its will than would a bullying stranger.

There is a proper and indeed often valuable role for the agency policy statement: as an expression of the agency's tentative position that does not purport to be binding. By setting forth its current thinking about the position it expects to take in the future, the agency provides useful guidance to staff and the affected public. But the indispensable condition is that the agency must maintain an "open mind," [22] so that affected persons, before the policy is applied to them, can have a meaningful opportunity to challenge the policy and make a case for an alternative or modified position or for abandonment of the position already tentatively stated. The agency should keep itself prepared to fundamentally rethink and revise its position, not merely to consider exceptions or waivers. Where the agency does those things, its policy statement may lawfully be issued without observance of APA notice-and-comment procedures. Such a policy statement comes within the exemption for "general statements of policy" in the APA's rulemaking provision, while a policy statement that the agency treats as binding does not.

Policy statements that set forth tentative positions, and that therefore do not need to be legislatively promulgated, can in appropriate circumstances be reviewed by the courts. The courts often will not regard such tentative documents as being "ripe for review." [23] But, when they do, they certainly do not have to approve the agency position. With policy statements that were not promulgated legislatively, the courts should apply a fairly stringent standard to test the agency position. In addition to their normal inquiries about factual support and whether the action was arbitrary or capricious or an abuse of discretion, the courts ought to consider the "underlying wisdom" of nonlegislatively issued policy statements, as advocated by Judge MacKinnon in the 1974 seminal case.[24] More specifically, the inquiry should examine agency policy for the reasonableness of its underlying assumptions, of its goals, and of the way in which the policy serves those goals.[25] That articulation of review standards is consistent with the relatively few cases on the subject.

Most frequently, however, agency policy statements are not treated as tentative. Agencies try to use them to tell people what they must or can or cannot do, and that is truly "offensive to our basic notions of democratic government."[26] An agency should not issue a policy document that it treats as binding unless it does so through the legislative rulemaking process. Otherwise the courts can be expected to overturn it. To be distinguished is the situation in which the document (or a portion thereof) interprets existing legislation, to which we now direct our attention.

Interpretive Rules (Nonlegislative Documents That Interpret)

Agencies and officials can set forth their interpretations of existing statutes or regulations in a variety of documents, legislative and nonlegislative.

If an interpretation is promulgated in a rule or regulation through legislative rulemaking procedures, then the rule containing that interpretation, like other legislative rules, can have legal binding effect. But more frequently agencies issue their interpretations in informal documents like interpretive memoranda, bulletins, guidelines, circulars, and the like. Those nonlegislatively promulgated interpretations are called interpretive rules.

Since they are not issued through legislative rulemaking procedures, interpretive rules cannot have legal binding effect. The constitutional basis for legal effect simply does not exist where an agency has failed to observe the procedures established by Congress for agencies to follow in making law.

And interpretive rules can have practical binding effect only to a limited extent. Agencies may legitimately attempt to apply their informally issued interpretations, but the courts are not bound by them. [27] The courts may and should form their own independent interpretations and should overturn agency interpretations with which they do not agree. Thus, until the courts have accepted a nonlegislative interpretation, the agency's efforts to enforce it are on shaky ground.

Undoubtedly, though, an agency often can induce affected private parties to comply, or intimidate them into compliance, during the interim before the courts have acted. Some persons who are affected may be in a position where they simply have to live with the agency interpretation. To that extent, a nonlegislative agency interpretation can have practical binding effect.

Friends of unrestrained governmental power have argued for an enormous broadening of that practical binding effect, which would result in agencies dictating citizens' conduct through informal memoranda, guidances, and the like.

Stretching the famous Chevron case, agencies and others have argued that the courts must accept all "reasonable" agency interpretations, even where the interpretations are issued informally rather than through proper legislative rulemaking procedures. [28]

The Chevron doctrine holds that a reviewing court must accept an agency's interpretation of a statute it administers, unless the interpretation is contrary to statute or is unreasonable. That doctrine gains its potency from the fact that, since the courts must accept agencies' reasonable interpretations, those interpretations have the effect of prescribing the law that the courts will apply. They thus bind the courts and the public.

That doctrine applies to interpretations that have been embodied in legislative rules. [29] Where the agency has exercised congressionally delegated rulemaking power in an assigned subject area, and in doing so has followed the procedures required by Congress, the agency's interpretation (if reasonable) must prevail in court, even where the court acting independently would have arrived at a different interpretation. Although interpretation of federal statutes is one of the federal courts' preeminent functions, under Chevron a court is not free to interpret the statute for itself but must accept the position duly legislated by the agency. The court "defers" to the agency in those circumstances for a structurally fundamental reason: It is the intent of Congress that propositions (including interpretations) properly laid down by legislative rulemaking procedures have the force of law, and therefore must be applied by the courts. Thus, when an agency reaches its interpretation by using its delegated lawmaking powers, Congress intends that the agency interpretation shall be the law for the court to apply. The exercise of congressionally conferred authority has displaced the normal duty of the courts to interpret statutes. The court merely reviews such an agency interpretation for reasonableness, which is not a very exacting standard of review and is no more searching than what the courts apply generally to other legislative rules. [30]

Should the Chevron doctrine of accepting "reasonable" agency interpretations be extended to cover interpretations set forth in informal nonlegislative formats like manuals and guidances and memoranda? The answer must be an emphatic no.

The alternative would allow something approaching agency dictatorship. Agency officials could issue interpretations in the form of memoranda and other informal documents--which do not receive the public input and the senior-level deliberation that legislatively issued rules must get--and the courts would have to accept them if "reasonable." Since the courts would be bound to accept them, those informal interpretations would have binding effect on the public as well. The result would be that, as a practical matter, staff members and regional office officials and other functionaries could declare the law, simply by issuing informal interpretations. The agencies could dispense with legislative rulemaking for all of their interpretations and just use informal documents.

Under a regime in which the courts were obliged to accept informal agency interpretations, the public would lose the protection of the courts' exercising their traditional role of independently interpreting the statutes and regulations in question. The courts would be largely neutered by limiting their intervention to cases where the agency interpretation can somehow be found "unreasonable." Independent review, by contrast, guards against overzealous or self-interested interpretations by the agencies, which, after all, are often in a position adverse to affected private parties.

Such independent review is plainly required by the language of the APA[31] and by pre-Chevron Supreme Court precedent. The long-standing doctrine of the Skidmore case should govern. Under it, a reviewing court should give respectful consideration to a nonlegislative agency interpretation but then should reach its own independent interpretation, overturning any agency interpretation with which it does not agree.[32]

That is no more than the citizen deserves. Independent judicial interpretation is a fundamental of our constitutional scheme.[33] It can be displaced where the agency's interpretation represents the exercise of congressionally delegated power to declare law through legislative rulemaking. Where the agency acts informally, however, it cannot claim that Congress has authorized it, rather than the courts, to make the controlling interpretation. The agency is not fulfilling a delegated legislative function, and the courts should not be deferential.

The requirement of independent review for informal agency interpretations ought to be reaffirmed by the Supreme Court,[34] to ensure continuation of the courts' indispensable role as the bulwark against agency overreaching.

Agencies' Interpretations of Their Own Regulations

In one important field, regrettably, the courts have abdicated their role as the citizens' protectors against agency overreaching. That abdication, which has received little attention, ought to be deeply troubling to those who are concerned about unlegislated agency compulsion.

The core of the problem is the Supreme Court's approach to judicial review of interpretations by agencies of their own rules and regulations.

Let us take a fact pattern from a recent case: [35] Where regulations requiring Medicare providers to make "reasonable collection efforts" are interpreted in an agency manual to require the use of collection agencies in specified circumstances, how should a reviewing court act toward the agency's interpretation?

Should the court act impartially, independently forming its own interpretation of the regulation, the way courts do every day when they interpret statutes? Or should the reviewing court act deferentially, upholding the agency position if it is reasonable, as the courts do with an agency interpretation embodied in a legislative rule?

Neither of those, it seems. The court should act abjectly. In a series of opinions dating from before the APA, the Supreme Court has repeatedly pronounced a review standard that imposes on the agencies almost no restraint whatsoever. In 1994, for example, it stated:

[T]he agency's interpretation must be given "'controlling weight unless it is plainly erroneous or inconsistent with the regulation.'" . . . [W]e must defer to the Secretary's interpretation unless an "alternative reading is compelled by the regulation's plain language or by other indications of the Secretary's intent at the time of the regulation's promulgation." [36]

There are grave flaws in that wretched formula, which is one of the most brazenly anti-democratic pronouncements ever to issue from the Supreme Court. As a practical matter, it gives agencies and even individual lower level officials the power to tell people what they can and cannot do through casual documents, provided only that the positions can be claimed to interpret the agency's regulations.

Most pervasively hurtful is the Court's failure to instruct the lower courts to apply this slovenly formula only to genuine interpretations. Recall that a document that does not interpret existing law is a policy statement, which the agency must issue through legislative rulemaking if it is going to treat it as binding. But if the document can be called an "interpretation," an agency not only can claim exemption from legislative rulemaking requirements but also can claim "controlling weight" such that the courts "must defer" to it.

The threshold issue thus always ought to be whether a document is authentically interpretive. The Supreme Court has said nothing that explicitly or clearly requires that agency documents be examined to ensure that they are interpretive. As a quite natural result, the courts are likely simply to accept assertions that the documents are interpretive and consider only whether they are plainly erroneous or inconsistent with the regulations. Even where they do ask whether the document in question actually interprets the regulation, the courts usually speak in conclusory terms, without rigorous analysis of whether there is a logical connection between the substance of a regulation and the document purporting to interpret it.[37] Consequently, in practice the courts might well enforce agency documents that could not be fairly said to interpret regulations.

If agency positions are not "plainly erroneous or inconsistent with the regulation," they are routinely accorded the nearly bulletproof "controlling weight" called for by the Supreme Court's review formula. That gives such agency documents a prodigious binding effect, since affected private parties have almost no realistic chance of getting judicial relief.

A bizarre result is that the binding effect of informal documents purporting to interpret regulations can be even stronger than the effect of formal legislative rules, in the following way. An interpretation of a statute set forth in a legislative rule must be accepted by the reviewing court if found to be reasonable.[38] And other legislatively promulgated regulations, which do not interpret statutes, must also satisfy a reasonableness standard on judicial review. [39] But once an informal document is (correctly or incorrectly) treated as an interpretation of a regulation, it gets controlling weight except in the unlikely circumstance of being found plainly erroneous or inconsistent with the regulation. Recent formulations by the Supreme Court omit any reasonableness requirement for interpretations of regulations.[40]

The supine posture of the courts has created a dangerous opportunity for agencies to impose unlegislated compulsion upon citizens. An agency can impose a practical binding effect on the public whenever it sets forth a new requirement in a memorandum or other informal document that it declares to be an interpretation of an existing regulation. Affected private parties will have to live with the document unless they believe that they can persuade a court that it is not really an interpretation, or that it is plainly erroneous or inconsistent with the regulation. It is probably even harder to overturn an informal document on those grounds than it is to overturn a procedurally proper legislative rule, promulgated in accordance with the statutory authorities and procedures established by Congress and having the force of law.[41]

Obviously, this is upside down. Not only does it give binding force to informal documents that Congress has not authorized, but it gives them greater force than formal documents that Congress has authorized. In sum, the courts' toothless review standard enables the agencies to dictate to the public, with near-complete binding force, without authority from Congress.

What is more, that state of affairs creates a powerful incentive for agencies to adopt vague regulations, as a basis upon which the agency can later establish new policy without observing legislative rulemaking procedures. With a fuzzy regulation in place, the agency can then put forth informal documents that state concrete new positions, in the guise of "interpreting" the regulation. Of course, the courts ought to apply the threshold test of whether the new document truly interprets the regulation; if the regulation is vacuous and devoid of concrete content, it presents nothing for a genuine interpretation to grab hold of, and therefore the agency's new position does not interpret anything and would have to be established by legislative rulemaking if the agency means to impose it in a binding way. But the courts do not regularly or thoroughly exercise the threshold test. In those circumstances, an agency may readily beable to avoid legislative rulemaking procedures, where it should have used them, and still get away with it. In the view of four dissenters, the Supreme Court recently let an agency do just that. [42]

A further curse of the Supreme Court's readiness to accept informal documents interpreting regulations is that it enables the agencies to tilt the playing field in their own favor and against private parties. In many situations there is a fair question as to what the regulations in question mean.[43] Agencies are often vigorous partisans. They will issue interpretations with the object of strengthening their regulatory hand, or gaining an advantage over private adversaries in enforcement litigation, or limiting their distribution of largesse. Such cases make obvious the injustice of allowing an agency to be the judge of its own cause by giving its interpretive positions "controlling weight."

The proper standard of review should be independent interpretation of the regulation by the court, under the doctrine of the Skidmore case. [44] Under that standard, the court gives consideration to the agency's position but is not obliged to accept it, even if it can be deemed reasonable.

By issuing a regulation, the agency has set forth a legislative act, for all to see. All should be able to determine the fair interpretations of it, just as they may do with an act of Congress. And all should be equally able to argue for the interpretations they thus derive, without being trumped in court by the agency under the one-sided review formula of the Supreme Court. The Court has disregarded the plain command of APA 706, which requires the reviewing court to "decide all relevant questions of law . . . and determine the meaning or applicability of the terms of an agency action."

Were they to engage in truly independent review, the courts would once again become the citizens' protectors against the agencies' self-serving or politically motivated "interpretations." Casual or informal agency interpretations would no longer impose on citizens the powerful practical binding effect that now flows from the Supreme Court's review formula. No more could the agencies act with virtual impunity in subjecting citizens to interpretations that, as independently assessed by the courts, are incorrect. Nor could the agencies costlessly assert new policy positions by claiming that they are interpretations when in fact they are no such thing.

The Supreme Court has extended extraordinary shelter to agency documents that purport to interpret agency regulations. That nurturing of what are often high-handed and unauthorized agency actions is profoundly undemocratic. And it affronts the separation of powers, by allowing executive agencies to impose their will without either congressional authority or meaningful judicial restraint.

The courts' indulgence of unlegislated compulsion of private persons is an anachronism. It is the legacy of a statist mindset that comes from the pre-APA era. The courts of earlier times, coping with depression and war and redistribution, exerted every effort to affirm dubiously authorized agency assertions of power. Such a judicial posture should have no place in today's system of limited government under the rule of law.


1. "The legislative power of the United States is vested in the Congress, and the exercise of quasi-legislative authority by governmental departments and agencies must be rooted in a grant of such power by the Congress and subject to the limitations which that body imposes." Chrysler Corp. v. Brown, 441 U.S. 281, 302 (1979).

2. Administrative Procedure Act, 5 U.S.C. 553 (1946).

3. Examples of statutes that require more extensive rulemaking procedures than those of the APA include 15 U.S.C. 57a (Federal Trade Commission) and 20 U.S.C. 1232 (Department of Education). On the other hand, rules involving military or foreign affairs functions and those involving agency management or public property, loans, grants, benefits, or contracts are exempt from the rulemaking requirements of APA 553(a), though they must nevertheless fulfill other requirements to be valid legislative rules. See Robert A. Anthony, "Interpretive Rules, Policy Statements, Guidances, Manuals and the Like--Should Federal Agencies Use Them to Bind the Public?" Duke Law Journal 41 (1992): 1311. Exceptions to the required use of legislative rulemaking procedures are also allowed for agency housekeeping rules ("rules of agency organization, procedure or practice") and where the agency for stated good cause finds that such procedures are "impracticable, unnecessary or contrary to the public interest." APA 553(b)(3)(A) and APA 553(b)(B).

4. "It is axiomatic that an administrative agency's power to promulgate legislative regulations is limited to the authority delegated by Congress." Bowen v. Georgetown University Hospital, 488 U.S. 204, 208 (1988).

5. Those constraints are imposed by the APA's vitally important provision that spells out the courts' "scope of review"--that is, the standards by which the reviewing courts will "hold unlawful and set aside agency actions, findings, and conclusions." APA, 706.

6. Agency actions are to be set aside if they are "in excess of statutory jurisdiction, authority, or limitations, or short of statutory right." APA 706(2)(C). They are also to be set aside if they are "otherwise not in accordance with law." APA, 706(2)(A).

7. Agency actions (including some rules) that are required by statute to be made on the basis of a formal record, involving a trial-type administrative hearing, must be "supported by substantial evidence." APA, 706(2)(E). Most of the actions discussed in this paper are rules, and most rules do not have to be made on a formal record with a trial-type hearing. Nevertheless, the courts have held that they must be supported by the same quantum of evidence as actions governed by the "substantial evidence" rule. Association of Data Processing Service Orgs. v. Board of Governors of the Federal Reserve System, 745 F.2d 677 (1984).

8. APA, 706(2)(A).

9. The courts shall set aside agency actions found to be "without observance of procedure required by law." APA, 706(2)(D).

10. This statement is true irrespective of whether the rule interprets existing legislation or not. A rule that interprets and explains existing legislation is tested by the standard of Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984), under which a rule interpreting a statutory provision that is ambiguous must be accepted by the courts if it is reasonable. Ibid. at 842-44. A rule that does not interpret or explain existing law (but rather creates new law) is tested by the "arbitrary and capricious" standard as articulated most prominently in Motor Vehicle Mfrs. Ass'n v. State Farm Mutual Auto. Ins. Co., 463 U.S. 29, 40-44 (1983), which is also regarded as a "reasonableness" test. In both situations, of course, compliance with statute and factual support are also required.

11. We are here concerned with agency actions that have some degree of general lawmaking effect on citizens, even though they have not been promulgated by legislative rulemaking procedures. Those actions or documents are nevertheless embraced within the APA's definition of a "rule": "[T]he whole or a part of an agency statement of general or particular applicability and future effect designed to implement, interpret or prescribe law or policy . . . ." APA, 551 (4). That definition includes guidances, bulletins, and other nonlegislative documents as well as properly promulgated legislative rules. The APA definition goes on to include agency statements "describing the organization, procedure, or practice requirements of an agency . . . ," but this paper is not concerned with such adjectival rules; it is concerned only with substantive rules affecting the primary conduct of private persons--that is, rules affecting private rights, duties, and obligations.

12. See Paul Light, "How Thick Is Government?" American Enterprise, November-December 1994, pp. 59-60. In addition to legislative rules and nonlegislative rules, agency actions can include individualized rulings called adjudications, in which the agency passes upon the rights, obligations, or eligibilities of a private party or a group of identifiable persons. See APA, 551(6-9). Typically, the agency either is deciding an enforcement action against the private party or is passing upon an application for an approval or benefit. The agency's reasoning or conclusion embodied in the decision can create a precedent that will be considered or applied in later cases, and thus can have effects like those of a rule. This paper does not further address lawmaking by adjudicatory decision. The fact that those styles of generating agency law or policy are governed by strikingly different procedural requirements (strict for rules but loose for adjudications) is a regrettable anomaly, flowing from distinctions drawn in the APA and confirmed by Supreme Court decisions holding that agencies may choose to generate their doctrines through case-by-case adjudication rather than by rulemaking. See NLRB v. Bell Aerospace Co., 416 U.S. 267 (1974).

13. U.S. Fish and Wildlife Service, Region 1, "Procedures Leading to Endangered Act Compliance for the Northern Spotted Owl," Portland, Ore., July 1990.

14. The APA definition of a "rule," APA 551(4), is not limited to legislative rules.

15. To interpret is to derive a proposition from existing law whose meaning compels or logically justifies the interpretation. To be susceptible to interpretation, the existing law must have some tangible meaning (rather than empty or vague language like "fair and equitable" or "just and reasonable") that can be interpreted. There must be a logical connection between the substantive content of the existing statute (or legislative rule) and the substantive content of the document that purports to interpret it, such that the latter flows fairly from the former.

16. APA, 553(b)(A).

17. See cases summarized in Anthony, "Interpretive Rules, Policy Statements, Guidances, Manuals and the Like," pp. 1355-59.

18. Morton v. Ruiz, 415 U.S. 199, 236 (1974).

19. Tabb Lakes, Ltd. v. United States, 715 F. Supp. 726 (E.D. Va. 1988), aff'd without opinion, 885 F.2d 866 (4th Cir. 1989).

20. United States v. Picciotto, 875 F.2d 345 (D.C. Cir. 1989).

21. Chamber of Commerce of the United States v. OSHA, 636 F.2d 464, 470 (D.C. Cir. 1980).

22. See McLouth Steel Products Corp. v. Thomas, 838 F.2d 1317, 1325 (D.C. Cir. 1988).

23. See cases cited in Robert A. Anthony and David A. Codevilla, "Pro-Ossification: A Harder Look at Agency Policy Statements," Wake Forest Law Review 31 (1996): 667, 672-73 n. 21.

24. See Pacific Gas and Electric Co. v. FPC, 506 F.2d 33, 40 (D.C. Cir. 1974).

25. See Anthony and Codevilla, pp. 688-92.

26. Chamber of Commerce of the United States, at 470.

27. It is not procedurally improper for an agency to issue an interpretive rule and attempt to give it binding effect. By contrast, it is procedurally improper for an agency to issue a policy statement to which it gives binding effect. An agency may assert its interpretation in a nonlegislative document like a guidance or an interpretive bulletin--and may act upon that interpretation, for example, by using it to deny an application or as the theory under which it brings an enforcement action--up to the point that a court rejects it as an incorrect interpretation. It is not procedurally improper for the agency thus to try to enforce a nonlegislative interpretative document. That does not mean, however, that the interpretation is legally binding, either on the courts or on the public. Where an interpretation has been set forth in an informal nonlegislative format, rather than in a legislative rule, the courts are not required to accord it any deference beyond a respectful consideration; they are free to derive their own interpretations of the law or regulation that is being interpreted. Skidmore v. Swift & Co., 323 U.S. 134 (1943); see Robert A. Anthony, "Which Agency Interpretations Should Bind Citizens and the Courts?" Yale Journal on Regulation 7 (1990): 1.

28. See Brief of the United States as Amicus Curiae Supporting Petitioners, City of Chicago v. Environmental Defense Fund, 114 S. Ct. 1588 (1994) (no. 92-1639) (arguing that a memorandum must be given Chevron deference because it was "reasonable"). See also Russell L. Weaver, "Chevron, Martin, Anthony and Format Requirements," Kansas Law Review 40 (1992): 598 ("Deference should not depend on whether an agency has exercised lawmaking authority").

29. It probably also applies to interpretations set forth in formal agency adjudicative decisions.

30. See Anthony, "Which Agency Interpretations Should Bind Citizens and the Courts?" pp. 15-25.

31. Section 706 of the APA provides that "the reviewing court shall decide all relevant questions of law, interpret constitutional and statutory provisions, and determine the meaning or applicability of the terms of an agency action."

32. See also Batterton v. Francis, 432 U.S. 416, 425 n. 9 (1977).

33. "It is emphatically the province and duty of the judicial department to say what the law is." Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803).

34. 34. See Anthony, "Which Agency Interpretations Should Bind Citizens and the Courts?" The Supreme Court sidestepped the issue in City of Chicago v. Environmental Defense Fund at 1594 n. 5.

35. University Health Services, Inc. v. Health & Human Services, 120 F.3d 1145 (11th Cir. Aug. 28, 1997).

36. Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 512 (1994) (quoting Martin v. OSHRC, 499 U.S. 144, 150-51 (1991), Bowles v. Seminole Rock & Sand Co., 325 U.S. 410, 414 (1945), and Gardebring v. Jenkins, 485 U.S. 415, 430 (1988)). The key elements derive from the pre-APA Seminole Rock case, in which the Court stated: "[T]he ultimate criterion is the administrative interpretation, which becomes of controlling weight unless it is plainly erroneous or inconsistent with the regulation." Seminole Rock at 414. Another often-cited case is Udall v. Tallman, 380 U.S. 1 (1965), which quoted the language just cited from Seminole Rock and added: "When the construction of an administrative regulation rather than a statute is in issue, deference is even more clearly in order. . . . If, therefore, the Secretary's interpretation is not unreasonable, if the language of the orders bears his construction, we must [uphold it]." Udall at 16-18. The Supreme Court has recently repeated its test in an even more deferential form: The Secretary's interpretation of his own regulations "is, under our jurisprudence, controlling unless 'plainly erroneous or inconsistent with the regulation.'" Auer v. Robbins, 117 S. Ct. 905, 911 (1997) (quoting Robertson v. Methow Valley Citizens Council, 490 U.S. 332, 359 (1989) and Bowles v. Seminole Rock, 325 U.S. 410, 414 (1945)). That version does not even include a qualification for alternative readings compelled by the regulation's plain language or by other indicators of the secretary's intent, as stated in the quotation in the text above, nor does it include a reasonableness requirement.

37. A capital example is a recent case in which the Supreme Court seemed to think it sufficient that the manual was not inconsistent with the regulation. Shalala v. Guernsey Memorial Hospital, 115 S. Ct. 1232, 1239 (1995), discussed in Robert A. Anthony, "The Supreme Court and the APA: Sometimes They Just Don't Get It," Administrative Law Journal of American University 10 (1996): 7. Far from finding a logical link, the Court recognized that "the only question unaddressed by the otherwise comprehensive regulations" was the subject of the manual provision at issue. Shalala v. Guernsey at 1238 (emphasis added). A lower court did look for but could not find a logical link between interpretation and regulation in Hoctor v. U.S. Dep't of Agriculture, 82 F.3d 165 (7th Cir. 1996) (assertedly "interpretive" memorandum, requiring eight-foot fence around enclosure of dangerous animals, could not be derived or excogitated from a regulation requiring a structure appropriate to protect and contain the animals).

38. Chevron U.S.A., Inc. at 842-44.

39. Motor Vehicles Manufacturers Ass'n v. State Farm Mutual Automobile Ins. Co., 463 U.S. 29 (1983); and Citizens to Protect Overton Park v. Volpe, 401 U.S. 402 (1971).

40. See Thomas Jefferson University at 512 and Auer at 911, both quoted previously. But see Paralyzed Veterans of America v. D.C. Arena, L.P., 117 F.3d 585, 584 (D.C. Cir. 1997) ("we very much doubt that we would defer to an unreasonable agency interpretation of an ambiguous regulation") (emphasis in original).

41. See Richard J. Pierce, "Seven Ways to Deossify Agency Rulemaking," Administrative Law Review 47 (1995): 83.

42. Thomas Jefferson Univ. at 524-25 (Thomas, J., dissenting).

43. See, for example, the regulations involved in Shell Oil Co. v. Babbitt, 125 F.3d 172 (3d Cir. 1997).

44. See Anthony, "The Supreme Court and the APA," pp. 4-12; and John F. Manning, "Constitutional Structure and Judicial Deference to Agency Interpretations of Agency Rules," Columbia Law Review 96 (1996): 612.

Full Text of Policy Analysis No. 312 (HTML)

1998 The Cato Institute
Please send comments to webmaster