Member, Jenner & Block, Chicago. The author is grateful to the Faculty of Law and Administration of the University of Warsaw and to the Faculty of Laws, Queen Mary and Westfield College of the University of London, where much of the work on this essay was accomplished. The author is also grateful to the Frances Lewis Law Center of Washington and Lee University, which supported this project, and to the staff of the Wilbur C. Hall Law Library of Washington and Lee University, especially Sarah K. Wiant, John Doyle, and Judy Stinson. The author would also like to thank Jacob Corr, Mark H. Grunewald, Liz Heffernan, John W. Joyce, Leo Kanowitz, Jeffrey S. Lubbers, Brian C. Murchison, H. Jefferson Powell, John Schmelzer, Jeffrey T. Shaw, and Winnifred Fallers Sullivan, who provided helpful comments on earlier drafts, as well as Christopher Anderson, Angela Stewart and Meri Triades, who provided excellent research assistance. Additional helpful comments were provided at the Federal Administrative Law Judges Annual Conference, where a shorter version of this paper was presented on September 11, 2001.
"O body swayed to music, O brightening glance,
How can we know the dancer from the dance?"
In the conventional taxonomy of governmental actors and functions, administrators are said to "execute" or enforce the law. Like courts, however, administrators also interpret law. 1Otherwise, execution or enforcement of the law would not be possible. In some cases, the burden of interpretation may be relatively undemanding. Congress may have given relatively specific instructions as to what an administrator, applicant for benefits, or regulated industry is required to do in a set of carefully defined situations, leaving little room or need for administrative "interpretation." In many cases, however, the task of interpretation will be more substantial. Congress's instructions will be less precise, and their proper application to the class of circumstances at hand will be subject to contest, requiring recourse to relevant interpretive materials, analysis, and the exercise of judgment. Administrators will be required to interpret the law, and will do so in much the same way that courts do, albeit from a perspective that may be more practical and informed, as well as more obviously interested. At the far end of the spectrum, Congress may have legislated in such a way that its signals are so sketchy or contradictory as to make interpretation nearly impossible.
In addition, the work of administrators sometimes includes "lawmaking" or legislation.2 In some cases, Congress will have given that responsibility to administrators expressly, as when administrators are charged by statute with the responsibility for promulgating "legislative" or substantive rules-rules that are indistinguishable from statutes, have the force and effect of law, and directly create enforceable rights or duties.3 Often the responsibility for crafting legislative rules will entail the making of choices about policy issues, large or small. In this class of cases, the proper field of administrative action, as well as the degree of discretion to be accorded the administrator's choices and determinations, may be broad or narrow, as Congress presumably will (or could) have determined. In other cases, where Congress has not expressly delegated lawmaking authority to an administrator, the need for such administrative action may nonetheless be implicit in the structure and design of the relevant statutory scheme. In yet another class of cases, delegation of lawmaking authority will not be implicit in the statutory scheme, but administrative experience and practical application of the statute may have demonstrated the need for further elaboration or application of statutory norms and objectives. In these last two circumstances, the absence of congressional direction necessarily leaves unanswered important questions about the permissible scope and legal effect of administrative action. Among other things, the extent to which such action will be permitted at all, the permissible scope of the action, and the proper characterization of the action within those permeable but necessary categories of lawmaking and interpretation, may not always be clear. The differences in legal consequences, however, may be substantial.
Administrators, then, do not simply "enforce" the law; they also interpret law, make law, and decide on policy initiatives. Different legal consequences may flow from the characterization of administrative action as one or the other of these activities, and attempts to draw bright lines are often unconvincing. With these three points in mind, one can begin to appreciate the tension and complexity inherent in the practice of judicial review of administrative interpretations and applications of law. Which branch of government is to have the final word with respect to a particular category of administrative action, and, if that responsibility should happen to fall to the judicial branch, what degree of respect, if any, should be accorded to the views of administrators? In some cases, these questions will implicate basic principles relating to delegations of legislative power under a republican form of government. In all cases, the answers will depend on ascertaining various principles of deference to administrative action and their relationship to politics and the rule of law.
In 1984, the Supreme Court set forth an apparently simple test for resolving some of these questions. In Chevron U.S.A., Inc. v. Natural Resources Defense Council, Inc.,4 the Court created a now-familiar two-step test for determining whether particular administrative interpretations of law should be upheld. According to Chevron, a reviewing court, using traditional methods of statutory construction, must first determine whether Congress has "spoken directly to the precise question at issue."5 If the court concludes that Congress did have a specific intention with respect to the precise question at issue, Congress's intention controls, and the agency's interpretation will stand only if the court is satisfied that the interpretation reflects congressional intent. On the other hand, if the reviewing court determines that Congress did not have a specific intention with respect to the precise question at issue, the court must proceed to step two of the analysis. At step two, the court's role is limited to determining whether the administrator's interpretation of the statute is "reasonable."6 If the reviewing court determines that the administrator's interpretation is "a permissible interpretation of the statute," that interpretation must be given effect, even if the court, making a de novo determination based on traditional principles of statutory construction, might have chosen a different interpretation as being more plausible or efficacious. At least in theory, step two of the Chevron test creates a default rule favoring strong deference to agency interpretations of law.7 More generally, Chevron's two-step approach is thought to have replaced a pragmatic, contextual set of inquiries with a more systematic and disciplined mode of inquiry, and to have replaced an approach that made deference a matter of degree with an approach that makes deference an all-or-nothing matter.8
All this is commonplace. Nonetheless, students of administrative law soon learn that the apparent simplicity of the Chevron test is illusory at best, and that its application in practice is sometimes complex or uncertain. Just as Chevron quickly became one of the most frequently cited cases in American jurisprudence, it also quickly became a major focus for legal scholarship.9 Both trends continue today. Some scholars have regarded Chevron as a revolutionary break with the past; others have thought its impact limited. Some scholars have criticized Chevron on theoretical and doctrinal grounds; others have attempted in various ways to measure its impact empirically or statistically.10 The doctrinal relationship of the Chevron test to "arbitrary and capricious" review under the Administrative Procedure Act11 has been an abiding concern, and a number of commentators have made various suggestions for refining the Chevron framework or replacing it with another approach.12
The extent to which Chevron may have covered the field, and thereby rendered the older methods of analysis "an anachronism," as Justice Scalia has suggested,13 remains a central and hotly contested issue within the Court. At present, an overwhelming majority of the Court seems committed to the view that the Chevron test is not exclusive, and that less formal agency interpretations, such as those contained in opinion letters, policy statements, agency manuals, and enforcement guidelines should continue to be evaluated under Skidmore v. Swift & Company,14 which requires that the courts evaluate administrative interpretations by considering a number of factors, including the thoroughness of the agency's consideration and the strength of its reasoning.15 Nonetheless, some...