• 31 August 2009

Chevron’s Mistake

Lisa Schultz Bressman - Vanderbilt Law School

Posted in , , , , , ,

In 1984, the Supreme Court issued an opinion that set forth a framework for addressing the problem of ambiguity in statutes that delegate authority to administrative agencies. That opinion, Chevron U.S.A. v. Natural Resources Defense Council,1 is the most famous case in administrative law. The Court stated that a reviewing court should first determine whether a statute is “clear” in relevant respects. A clear meaning is controlling on all those concerned.2 But if the statute is ambiguous in relevant respects, the court should defer to the agency’s interpretation as long as that interpretation is “reasonable.”3 Although the Court meant to crystallize if not revolutionize the approach for addressing ambiguity in regulatory statutes, it did not fully capture the process. Later decisions would offer further clarification and introduce further confusion. Some twenty-five years later, judges and scholars are still trying to sort out the proper method of statutory interpretation in the regulatory state.

In this Article, I attempt a step forward on that important project. I begin by offering a novel explanation for why Chevron falls short of offering the proper method of statutory interpretation. I argue that Chevron was never well calibrated to determine whether a statute is clear or ambiguous in relevant respects. The reason is that the decision fails to give courts adequate tools for determining whether Congress has delegated interpretive authority to the agency involved. How big is the problem? It is hard to measure, but there is a sense among professors who study this area that, Chevron notwithstanding, courts are too often in the business of invalidating agency interpretations.4 In addition, there is another part of the problem: courts too often uphold agency interpretations when they should instead invalidate them.

Where exactly does Chevron take a wrong turn? The decision pursues the correct goal—namely, to account for how Congress designs regulatory statutes. It recognizes that Congress may delegate authority to agencies to capitalize on agency expertise on complex issues or to obtain legislative consensus on contentious issues. And it recognizes that Congress may delegate expressly by granting regulatory authority in statutes or implicitly by leaving interpretive room in statutes. So, sensibly, Chevron asks courts to determine whether Congress has delegated such interpretive power to the agency involved.

Yet, despite its grasp of legislative realities, Chevron then reverts to a standard judicial conception of statutory interpretation. That (mis)conception has judges examining statutory text as if Congress meant to resolve rather than delegate an interpretive question. According to Chevron, courts should look for a “clear” meaning using the “traditional tools of statutory construction.”5 This process is not one-size-fits-all. That is, courts approach statutory text using different interpretive theories, including textualism, intentionalism, and purposivism. At the bottom of all those theories, however, lies a vision of legislative behavior that sees Congress as resolving concrete questions. For example, textualism sees statutory text as embodying a legislative deal, often peculiar when viewed in context but no less precise in its meaning. Intentionalism views statutory text as reflecting a collective legislative intent, albeit imperfectly expressing that intent. Purposivism regards statutory text as the manifestation of a general legislative goal, again as imperfectly expressing that goal. Despite their differences, all of these theories view Congress as striving for a relatively specific meaning. (In addition, textualism views statutory text and only statutory text as law.) When courts apply any of these theories, they are approaching regulatory statutes in much the same way as they approach any other statute: trying to recover the specific meaning of the words at issue.

To be sure, some courts apply Chevron mindful of a more agency-liberating spirit. They look for a “zone of indeterminacy” that permits them to uphold the agency interpretation if that interpretation is reasonable.6 The difficulty is that the doctrine does not ordain this type of judicial generosity. I claim that the doctrine itself should do better. As it stands, Chevron effectively primes courts to find a relatively specific meaning. Once courts find a “best” meaning, they must possess more than an average amount of judicial resistance to set aside that meaning in favor of the agency’s merely “reasonable” meaning, as Chevron optimistically counsels.7 Ardent textualists probably have a more difficult time abandoning their preferred meaning than others because they are especially committed to the pursuit of linguistic clarity.8 But no judge, textualist or otherwise, who takes seriously the search for linguistic clarity is immune.

In some sense, the Court has already recognized the deficiencies of Chevron. It has been fiddling with Chevron in a way consistent with my criticisms. In United States v. Mead Corp.,9 for example, the Court returned to the question of whether Congress has delegated interpretive authority to the agency involved, this time by focusing specifically on the relationship between congressional delegation and administrative procedures. The Court reasoned that Congress would not intend for an agency to possess binding interpretive authority unless it authorized (and the agency utilized) procedures that are appropriate for issuing binding interpretations.10

In my view, Mead does not remedy the problem because, like Chevron, it is right in theory but not in practice. Specifically, it does not give courts an adequate metric for evaluating questionable administrative procedures. Mead instructs courts to consider whether the procedures at issue promote fairness and deliberation, which is a standard judicial conception of the values that procedures serve—and a fuzzy one at best. I do not deny that procedures, if sufficiently formal, serve these rule-of-law values. My point is that courts, if interested in tracking congressional delegation, should be equally interested in how procedures fit into the strategic legislative picture. Why are procedures important to Congress when delegating interpretive authority to an agency? If courts focused on answering this question, then Mead might truly focus on the connection between delegation and procedures. Furthermore, courts might have more guidance in applying the decision. As Mead stands, courts are seriously confused about what it means and how it applies.11

Rather than fiddling, I argue that the time has come to replace the standard judicial theories and conceptions with the factors that matter to Congress when designing regulatory statutes. I admit that identifying those factors is no easy task. Yet the decisions themselves suggest a compass. Both Chevron and Mead reflect impulses that can be traced to political science literature, and in particular, positive political theory (PPT). This literature does not reveal how Congress designs particular statutes. Rather, it contains hypotheses about legislative behavior that might furnish assumptions about legislative behavior in the statutory interpretation context. Courts already rely on assumptions about legislative behavior when they apply Chevron and Mead.  In my view, PPT suggests more accurate ones and therefore can produce better interpretation of statutes.

PPT scholars offer two hypotheses that are of particular interest in fashioning a new approach to statutory interpretation. First, PPT scholars assert that Congress is more likely to delegate if the issue is complex or an obstacle to a law’s passage.12 Congress lacks the time and competence to handle many issues and will be blamed for any burdens imposed specifically by statute. Congress also faces obstacles to consensus that are alleviated by choosing ambiguous language. Of note, legal scholars have also underscored the potential for Congress to use “deliberate ambiguity” when writing statutes.13 Second, PPT scholars assert that Congress is more likely to delegate if it can monitor subsequent agency interpretations through administrative procedures or has reasonable assurances that agency preferences are likely to roughly track legislative preferences.14 Notwithstanding its appeal, delegation comes with a price. Congress must monitor subsequent agency action to ensure that it roughly tracks legislative preferences. Administrative procedures are an ingenious form of legislative monitoring because they shift the costs of monitoring to constituents. Rather than having to discover information about agency action itself, Congress gives constituents a right to participate in the process so that those constituents may learn about agency action as it unfolds and alert Congress to intervene on their behalf.

I suggest that courts should examine the sorts of considerations that PPT theorists and legal scholars have identified as indications of congressional delegation. Assuming Congress delegates to agencies for reasons related to the complexity or contentiousness of the issue addressed by a statute, courts should consider the complexity or contentiousness of the issue. Assuming Congress delegates to agencies under conditions that minimize principle-agent concerns, courts should consider the adequacy of administrative procedures for this purpose or should consider other evidence of likely legislative-administrative preference alignment. To isolate these factors, courts should examine the face of the statute as well as its legal or social context. The stronger the case for interpretive delegation, the more courts should hesitate to read the statutory text as “clear” on the theory that Congress intended such clarity.

I acknowledge that a focus on “delegating” factors has limitations. First, it might not constitute much of a test because courts might find a delegation more often than not. Many issues in regulatory statutes are complex, and many agencies have relatively formal procedures for resolving interpretive questions. Accordingly, it is fair to describe my approach as a presumption of judicial deference to agency interpretations. My claim is that such a presumption is the right one. By the same token, I am not advocating judicial abdication. Courts should not defer to agency interpretations when circumstances suggest that Congress did not mean to delegate.

Second, courts might find evidence of a delegation but not necessarily to the agency. When Congress seeks to avoid contentious issues, for example, it could delegate to courts rather than an agency. In such circumstances, a finding of delegation is still significant. Even if courts retain interpretive authority for themselves, they might exercise that authority differently knowing that Congress did not specify a meaning. They would be justified in eschewing the traditional theories of statutory interpretation and considering other factors, including the agency’s views of wise policy. This approach offers a modern understanding for a famous old administrative law case, Skidmore v. Swift & Co.15

I am not alone in seeing real advantages in attempting to build delegating factors into statutory interpretation. Looking at two recent Supreme Court decisions, I see certain Justices as moving toward this approach. In Zuni Public School Dist. No. 89 v. Department of Education,16 Justice Breyer set aside Chevron’s question of meaning to focus on the sort of factors that I identify here. He returned to the statutory text but only as confirmation of a delegation, whereas Justice Scalia invalidated it on plain meaning grounds. In Gonzales v. Oregon,17 Justice Kennedy purported to apply a standard interpretive principle, but he did so in consideration of the sorts of factors that I emphasize here. Interestingly, he found no delegation based on the factors. He rejected the agency’s interpretation, whereas Justice Scalia offered a plain meaning argument to uphold it. These cases demonstrate not only that my approach can work, but that it can work both ways—to uphold or invalidate an agency interpretation.

Why should courts adopt the approach that I advocate? First, I believe that the law should seek to minimize fictions. Chevron does not track congressional delegation, although it purports to do so. My approach minimizes the fiction about congressional delegation by basing the doctrine on better assumptions about the legislative design of regulatory statutes. Second, I argue that the law should incorporate these assumptions for normative reasons. The problem with decisions like Chevron and Mead is not their focus on congressional delegation, but their overly judicial approach to ascertaining the existence of delegation. The focus on delegation is sensible. It continues an old tradition of promoting fidelity to congressional will, not as to specific statutory meanings but as to the involvement of agencies in choosing such meanings. In addition, it promotes other critical values. The reasons for which Congress delegates generally correlate to the reasons for which deference is normatively desirable: agency expertise and political accountability. On the flip side, Congress is not always inclined to delegate, and agencies are not always inclined to use formalized procedures. Precluding agency interpretive authority under these circumstances respects congressional will and promotes rule-of-law values.

In the Article, I address institutional and normative objections to my approach. The institutional objection concerns the difficulty of applying this test. As I previously mentioned, lower courts are already struggling with the doctrine in this area. It might be that they cannot handle one more complexity. I argue that my approach is not necessarily worse and may be better than current law, though I leave open the possibility of a truly streamlined rule: a presumption of judicial deference to agency interpretations stronger than Chevron has ever required.

The normative objection concerns the role of courts in policing broad delegations through statutory interpretation. Courts have enlisted canons of construction and even theories of interpretation to ensure that Congress delegates no more broadly than is normatively desirable.18 This well-established approach is seemingly precluded by my own. In fact, I show that it can coexist with my approach, applying in select cases when courts determine that nondelegation values are paramount. But before asking courts to determine which approach—the prodelegation or antidelegation approach—is appropriate in particular cases, I highlight a different alternative. In my prior work, I have argued that it is possible and promising for courts to police broad delegations under the arbitrary and capricious test of administrative law rather than as a matter of statutory interpretation. The idea is something of a compromise between prodelegation and antidelegation values or, more specifically, an insistence that policymaking exhibit a modicum of restraint and rationality at some point in the process. If Congress chooses to delegate broadly, courts should ensure that agencies exercise their delegated authority in a disciplined manner.

In sum, Chevron’s mistake is in thinking about statutory ambiguity too much as a problem of words rather than as a manifestation of politics. Congress certainly uses words in statutes to make policy. But Congress also uses words in statutes to avoid setting specific details, often delegating that responsibility to the agency involved. Chevron comes into play precisely at that moment—when Congress transfers authority to an agency not in so many words. Yet Chevron directs reviewing courts to consider what words Congress provided as the best indication of implicit delegation. Moreover, it allows courts to use interpretive theories that are designed to find a relatively specific meaning. Instead courts should examine other signals of congressional delegation before they draw conclusions about the significance of words. Like any other approach to statutory interpretation, mine is not without its own difficulties. In the end, I maintain that it offers a better method of interpretation in this context. Courts protect important normative values—indeed, the values for which Chevron and Mead themselves purport to stand—when they attempt to regard regulatory statutes in their realistic legislative context.dingbat

 

Acknowledgments:

Copyright © 2009 Duke Law Journal.

Lisa Schultz Bressman is Professor of Law at Vanderbilt University Law School.

This Legal Workshop Editorial is based on the following Article:   Lisa Schultz Bressman, Chevron’s Mistake, 58 DUKE L.J. 548 (2009).

  1. Chevron, U.S.A, Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984).
  2. See id. at 842.
  3. See id.
  4. See, e.g., Thomas J. Miles & Cass R. Sunstein, Do Judges Make Regulatory Policy? An Empirical Investigation of Chevron, 73 U. CHI. L. REV. 823 (2006).
  5. Chevron, 467 U.S. at 842 & n.9.
  6. Peter Strauss, One Hundred Fifty Cases per Year: Some Implications of the Supreme Court’s Limited Resources for Judicial Review of Agency Action, 87 COLUM. L. REV. 1093, 1124 (1987).
  7. See Jacob A. Gersen & Adrian Vermeule, Chevron as a Voting Rule, 116 YALE L.J. 676, 960-61 (2007).
  8. See Thomas W. Merrill, Textualism and the Future of the Chevron Doctrine, 72 WASH. U. L.Q. 351, 372 (1994).
  9. United States v. Mead Corp., 533 U.S. 218 (2001).
  10. Id. at 231-33.
  11. See Lisa S. Bressman, How Mead Has Muddled Judicial Review of Agency Action, 58 VAND. L. REV. 1443, 1469-74 (2005).
  12. See, e.g., David Epstein & Sharyn O’Halloran, The Nondelegation Doctrine and the Separation of Powers: A Political Science Approach, 20 CARDOZO L. REV. 947, 967 (1998).
  13. See, e.g., Joseph A. Grundfest & A.C. Pritchard, Statutes with Multiple Personality Disorders: The Value of Ambiguity in Statutory Design and Interpretation, 54 STAN. L. REV. 627 (2002); Victoria F. Nourse & Jane S. Schacter, The Politics of Legislative Drafting: A Congressional Case Study, 77 N.Y.U. L. REV. 575 (2002).
  14. See, e.g., Mathew D. McCubbins, Roger G. Noll & Barry Weingast, Administrative Procedures as Instruments of Political Control, 3 J.L. ECON. & ORG. 243, 246 (1987).
  15. Skidmore v. Swift & Co., 323 U.S. 134 (1944).
  16. Zuni Pub. Sch. Dist. No. 89 v. Dep’t of Educ., 127 S. Ct. 1534 (2007).
  17. Gonzales v. Oregon, 126 S. Ct. 904 (2006).
  18. See, e.g., John F. Manning, Textualism as a Nondelegation Doctrine, 97 COLUM. L. REV. 673, 702-25 (1997); Cass R. Sunstein, Nondelegation Canons, 67 U. CHI. L. REV. 315, 338 (2000).

Post a Comment (all fields are required)

You must be logged in to post a comment.