Prof. Thomas Merrill on "The Chevron Argument: Ambiguity Versus Delegation"

I'm delighted to pass along this item from Prof. Thomas Merrill (Columbia), whom readers might remember as a guest-blogger from a couple of years ago, and who is the author of The Chevron Doctrine: Its Rise, and Fall, and the Future of the Administrative State (2022):

Last week, the Supreme Court heard nearly four hours of argument about the Chevron doctrine—including whether it should be cast aside. Not surprisingly, much of the argument consisted of different conceptions of what Chevron means. The nub of the problem is to specify what must happen for a court to depart from ordinary statutory interpretation (Chevron's step one) and defer to a reasonable agency interpretation of the statute (Chevron's step two). The argument (in Loper Bright Enterprises, Inc. v. Raimondo and Relentless, Inc. v. Department of Commerce) largely proceeded on the assumption that a court must turn to step two if it concludes the statute is "ambiguous" or includes a "gap." At some points, however, the advocates and the Justices thought that the critical question was whether Congress has implicitly "delegated" interpretive authority to the agency.

If we attempt to solve the problem by parsing what was said in the 1984 Chevron case, we quickly encounter the difficulty that Justice Stevens said both things. He wrote, on the one hand, that the court should employ "traditional tools of statutory construction" to determine whether a statute is "silent or ambiguous with respect to the specific issue." In other words, ambiguity is the key.

But, on the other hand, he also wrote that if Congress makes "an express delegation of authority to the agency to elucidate a specific provision of the statute by regulation," the agency's interpretation will prevail unless it is "arbitrary, capricious, or manifestly contrary to the statute." He immediately added: "Sometimes the legislative delegation to an agency on a particular question is implicit rather than explicit," in which case also the court should accept "a reasonable interpretation made by the administrator of an agency." In other words, explicit or implicit delegation to the agency to interpret is the key.

Some readers of Chevron, including Justice Scalia, have reconciled these statements by concluding that an ambiguity or a gap in a statute just is an implicit delegation by Congress to the agency. He admitted that equating ambiguity with an implicit delegation is a "fiction[]," but he argued it was a salutary one.

Justice Stevens's opinion in Chevron did not explicitly link the two propositions in this fashion. Even if he did, the opinion in Chevron, which was regarded as a restatement of existing law when it was decided, should not be parsed as if were a constitutional provision.

In fact—for the point that I especially want to explore here—the "ambiguity" conception of Chevron and the "delegation" conception have very different implications for a number of difficult issues. This was abundantly revealed during the lengthy oral argument.

Perhaps the most far-reaching claim advanced by petitioners in the pending cases is that Chevron is unconstitutional because it violates a fundamental precept of Article III: that courts have an unfailing duty "to say what the law is" in cases that come before them (Marbury v. Madison). Chevron seems to say that courts should enforce the law only when it is unambiguous. Otherwise, agencies act as the primary interpreter, and courts must defer to the agency's view.

Solicitor General Elizabeth Prelogar and several of the Justices parried the constitutional claim with counterexamples, such as review under the traditional writ of mandamus, where courts would intervene only if the law was clear, or review of final convictions on habeas corpus, where Congress has limited relief to clear violations of the law (Antiterrorism and Effective Death Penalty Act of 1996). And, of course, Chevron itself acknowledged that when Congress expressly delegates authority to an agency to define a statutory term, review is deferential.

Overall, the delegation characterization seems more consistent with the broad range of history, which indicates that it is up to Congress to decide when judicial review is available and what the standard of review should be.

Sensing perhaps that the Court was not going to buy the constitutional argument, petitioners took a step back and insisted that Chevron is inconsistent with the Administrative Procedure Act. Section 706 of the APA instructs reviewing courts to "decide all relevant questions of law" and to set aside agency action "not in accordance with law." This seems to contemplate that courts must resolve all questions of law de novo, that is, in the exercise independent judgment about the meaning of the law. How, the petitioners repeatedly asked, can this be squared with a doctrine that requires courts to defer to an agency when the law is ambiguous?

Solicitor General Prelogar responded that Section 706 should be read in light of pre-APA precedents that gave varying degrees of deference to agency interpretations. Looking at the text of Section 706, however, there is nothing to suggest that the APA was ratifying such decisions (the legislative history on this point is contested).

A much better argument is one grounded in implicit delegation. As Henry Monaghan observed before Chevron, if a court concludes that Congress has delegated authority to an agency to resolve an ambiguity or gap, then the court is simply determining, as a matter of independent judgment, that the relevant rule of decision is the one adopted by the agency.

The petitioners also made much of the fact that if there is no agency in the picture, a court will determine the best meaning of a statute, including the meaning of any ambiguities or gaps it may contain. How, they asked, can this settled understanding about the judicial role be set aside merely because an agency enters the picture? If courts can and must resolve ambiguities in statutes when there is no agency, surely they must do the same when an agency has offered its view of the matter.

SG Prelogar and the Justices sympathetic to Chevron tried to rebut this point by saying that agencies are likely to do a better job at resolving ambiguities and gaps, because of their expertise and familiarity with the statutory regime. As the petitioners pointed out, however, courts do not need Chevron to tap into the superior expertise of agencies. Courts can take advantage of agency insights under the Skidmore standard of review, which requires courts to consider agency interpretations from a variety of perspectives and endorse them if they are persuasive, but leaves the final judgment up to the court.

Once again, the implied-delegation theory offers a more principled justification for deferring to agency interpretations of statutes. If Congress has directed the agency to do the interpreting, then of course the courts should defer to the agency.

A further puzzle debated by the Justices concerned how disruptive it would be to overrule Chevron. The petitioners argued that none of the results reached in the thousands of cases that have been decided under Chevron would have to be relitigated, because Chevron is simply a method of deciding cases rather than a specific legal rule. As such, it is entitled to little weight as a matter of stare decisis.

Justice Barrett pointed out, however, that any prior decision upholding an agency interpretation at step two would be vulnerable to being relitigated if Chevron were overruled, because claimants could ask for a different outcome based on a court's determination of the best understanding of the statute. In effect, under the view that Chevron is triggered by ambiguities or gaps, overruling Chevron would open the door to massive relitigation of any case decided at step two.

Under the delegation conception, in contrast, the only thing settled by prior step-two decisions is that Congress has implicitly delegated authority to the agency to resolve the issue. This determination would be given effect as a matter of stare decisis, and relitigation of the determination would presumably be ruled out.

Finally, the petitioners made much of the Court's 2005 decision in National Cable & Telecommunications Association v. Brand X Internet Services, holding that agencies can override judicial decisions when the conditions for applying Chevron are met. This, they said, inverts the usual hierarchical relationship between agencies and reviewing courts, and promotes instability as one administration reverses the interpretations offered by the previous administration. These results would not be possible under Skidmore-type review, they pointed out, since all interpretations would be made by courts (giving appropriate consideration to agency views) and the judicial judgments reached would be final and binding as a matter of stare decisis. SG Prelogar endorsed Brand X, but she did not explain why it was correct.

The answer, once again, is that Brand X is required by the delegation conception of Chevron. If Congress has delegated interpretive authority to agencies, then agencies exercising that authority should be able to override contrary judicial interpretations. Judges may not like this, and the Court has strained to interpret its own precedents as step-one decisions to avoid this result (see T. Merrill, The Chevron Doctrine: Its Rise and Fall, and the Future of the Administrative State 148–58 (Harv. Univ. Press 2022) (discussing "Chevron and Conflicting Judicial Precedent")). To be sure, the petitioners' point about instability is well taken. But rather than replace Chevron with Skidmore, a better solution is to incorporate the traditional canon about the importance of longstanding and consistent agency interpretations into the "reasonableness" review that takes place under step two.

In sum, the defenders of Chevron had multiple reasons to endorse the delegation reading rather than the ambiguity reading. The ability of Congress to delegate interpretive authority to different entities defeats the Article III argument, explains why Chevron is consistent with the APA, distinguishes cases of agency review from cases where there is no agency in the picture, bolsters the case for preserving Chevron as a matter of stare decisis, and rationalizes Brand X. Unfortunately, the defenders of Chevron, by switching back and forth between the ambiguity conception and the delegation conception, made the logic of the doctrine harder to discern.

How would the Chevron doctrine change if the Court affirmed that it is based on delegation rather than ambiguity? One potential change is that courts would have to engage in a more wide-ranging inquiry at step one. It would not be possible to zero in on a single word or gap, declare it ambiguous, and endorse the agency interpretation. Sometimes an implicit delegation would be obvious, as when Congress directs an agency to set "reasonable" rates or "appropriate" environmental standards. The petitioners conceded that Congress can delegate authority to make "policy" in these ways, but the delegations also entail authority to resolve a variety of legal questions (such as whether a reasonable rate includes a risk-adjusted rate of return or whether appropriate standards require a consideration of costs).

In other cases, step one would require a more far-reaching examination of whether Congress's use of terms of art, like "navigable waters," imports a delegation of interpretive authority (Sackett v. EPA, 2023), or whether the context in which terms appear, such as a delegation of authority to "modify" tariff-filing requirements, imports authority to deregulate an industry (MCI v. AT&T, 1994). In still other cases, a sequence of legislative enactments would suggest that Congress has or has not delegated authority to interpret broad terms in particular ways (FDA v. Brown & Williamson Tobacco Corp., 2000). The Supreme Court has been moving strongly in the direction of this kind of wide-ranging consideration of context in statutory interpretation cases in any event, so this clarification of step one would not be that remarkable.

Would changing the focus from ambiguity to delegation create more uncertainty about whether it is appropriate to move to step two? Perhaps. But as Justice Kavanaugh has insisted in extrajudicial writing and as he stressed again in the argument, the ambiguity trigger has proved to be a great source of uncertainty itself. He pointed out that lower court judges have very different notions about what constitutes "ambiguity." Some find ambiguity all the time, and others never find it. And the petitioners could point to the two decisions under review: the D.C. Circuit (in Loper Bright) found the statute ambiguous and deferred to the agency under step two, but the First Circuit (in Relentless) concluded that the statute is clear and supported the agency interpretation.

In determining whether an agency has been delegated authority to interpret, Prelogar pointed out that in United States v. Mead Corp. (2001), the Court limited Chevron to cases where Congress has delegated authority to the agency to act with the force of law. An agency cannot act with the force of law unless it has been delegated regulatory authority over the question at issue; so answering the Mead question gets a court a fair way toward answering whether the agency has been delegated authority to interpret.

Another implication is that the Court would probably want to beef up step two. As recognized in Chevron and as acknowledged by the petitioners, exercises of delegated authority by agencies are subject to arbitrary-and-capricious review. Usually this is more searching than a perfunctory conclusion that the agency's interpretation is minimally rational. Two candidates for a beefed-up step two would be (1) to require courts to consider the consistency with which the agency has maintained its interpretation and the reasons it has given for any change and (2) to give agencies additional points for rendering interpretations in formats in which the public has been given notice and an opportunity to make its views known to the agency. Prelogar seemed to endorse these modifications in a sympathetic back-and forth with Justice Kagan near the end of the marathon argument.

A deeper reason why the delegation model of the Chevron doctrine may encounter resistance is the campaign of the conservative Justices to limit the power of Congress to delegate authority to administrative agencies. Justices who seem to allude fondly to returning to the pre-1935 conception of permissible delegation or even who create the "major questions doctrine" as a kind of reverse-Chevron are unlikely to embrace a conception of Chevron as grounded in a delegation to the agency of interpretive authority. Overruling Chevron and making Skidmore the universal standard for judicial review of agency interpretations of law would be more consistent with this campaign, as Michael Dorf has explained in a series of blog posts. But writing an opinion reaching this result is made difficult by the logic of delegation, which is very much embedded in Chevron jurisprudence—and provides the superior rationale for the doctrine.