The Supreme Court says that Chevron has two steps: Is the statute ambiguous (Step One), and, if so, is the agency’s interpretation of the ambiguous provision a permissible one (Step Two)? Yet over the last three decades, the DC Circuit has inserted an intermediate step between Steps One and Two: Did the agency recognize that the statutory provision is ambiguous? If not, then the DC Circuit refuses to proceed to Chevron Step Two and remands the matter to the agency. This doctrine—which we dub “Chevron Step One-and-a-Half”—has led to dozens of agency losses in the DC Circuit and DC federal district court, but it has gone entirely unmentioned in administrative law casebooks and is rarely referenced in the academic literature. The few who have not ignored the doctrine have treated it with skepticism. Chief among those skeptics is now–Chief Justice John Roberts, who while a DC Circuit judge criticized his colleagues for applying the doctrine.

This Article presents a more sympathetic account of Chevron Step One-and-a-Half. After providing an overview of the Chevron Step One-and-a-Half doctrine, we offer several theories why Chevron Step One-and-a-Half cases continue to arise, even though agencies can avoid the doctrine by stating that they would hew to their view regardless of whether the relevant statutory provision is ambiguous. Some number of Chevron Step One-and-a-Half cases might be explained by the fact that agencies are ignorant of the doctrine or ambivalent about their own policies, but we suggest that there also may be strategic reasons why agency actors might maintain that a statute is unambiguous. For instance, agency lawyers with a preference for a particular reading (or with patrons who have such a preference) might seek to increase influence over policy by declaring that a statute can be interpreted only one way. Alternately, an agency might claim that a statute is unambiguous in order to reduce the probability that the White House’s Office of Information and Regulatory Affairs will second-guess the agency’s choice. In a similar manner, an agency might attempt to evade political accountability for an unpopular policy by claiming that the choice was compelled by Congress. Finally, an agency might maintain that a statute is unambiguous in order to “lock in” an interpretation so that future administrations cannot undo it. After identifying the potential causes of Chevron Step One-and-a-Half cases, we consider how courts ought to respond to the potential for strategic agency behavior. We suggest that, when viewed in this light, Chevron Step One-and-a-Half helps to uphold the theoretical justifications for Chevron deference. While Chevron Step One-and-a-Half remands also impose undeniable costs on administrative agencies, we argue that these costs ought to be evaluated against the considerable benefits that the doctrine potentially brings.


The Supreme Court’s decision in Chevron U.S.A. Inc v Natural Resources Defense Council, Inc1 has created a cottage industry in choreography. Justice John Paul Stevens’s opinion introduced the famous Chevron two-step.2 Thereafter, Professors Thomas Merrill and Kristin Hickman identified a Chevron “step zero”—“the inquiry that must be made in deciding whether courts should turn to the Chevron framework at all.”3 (Some subdivide Step Zero into two steps of its own, creating a four-step test.)4 A quarter century after the Chevron decision, Professors Matthew C. Stephenson and Adrian Vermeule declared that in fact “Chevron has only one step”: “whether the agency’s construction is permissible as a matter of statutory interpretation.”5 Some judges now view Chevron as “a three-step inquiry,”6 while others suggest that the number of steps probably does not matter much in practice.7 Little wonder, then, that others throw up their hands (or their feet?) and dismiss the entire step-defining exercise.8

One might infer from this choreographic confusion that we now have too many formulations of Chevron, with no need for another. One might also draw the conclusion that a doctrine designed to simplify judicial review of agency statutory interpretations has instead had the opposite effect. And yet one might infer from this Article’s title that the authors, rather than trying to streamline Chevron, are scheming to propose yet another step in the Chevron shuffle—or, more precisely, a half step. One might ask, quite fairly, whether the Chevron dance really needs another move. (One might also groan that we already have taken the dance metaphor too far.)9

Although we are mindful of Chevron fatigue, the fact remains that neither the one-step, the two-step, the three-step, nor the four-step formulation of Chevron captures an important doctrinal development that has occurred in the federal courts. In the classic Chevron two-step, the court asks at Step One “whether Congress has directly spoken to the precise question at issue”; if the answer is negative, then the court proceeds to Chevron Step Two and asks “whether the agency’s answer is based on a permissible construction of the statute.”10 The DC Circuit and a handful of other courts, however, now do something different. After deciding that the statute is ambiguous but before deciding whether the agency’s construction is permissible, these courts ask a separate question: whether the agency itself recognized that it was dealing with an ambiguous statute. In these courts, a misstep at this intermediate stage can be fatal to an agency’s cause: the court will remand—sometimes with vacatur—if the agency claimed that the statute is clear but the court concludes it is not. In other words, the agency will lose if it mistakenly says that the issue can be resolved at Chevron Step One while the court determines that it should be resolved at Chevron Step Two.

One might call this move “the Prill doctrine” in honor of Prill v National Labor Relations Board,11 the DC Circuit case from 1985 that is sometimes cited as the rule’s origin.12 Or perhaps one might call it “the Negusie doctrine” in honor of Negusie v Holder,13 a 2009 case in which the Supreme Court arguably applied the rule as well (though, as discussed below, we think the Negusie rule is of a slightly different nature).14 We choose to call it “Chevron Step One-and-a-Half,” because—well—that is what it is: a way station between Chevron Step One and Chevron Step Two. Whatever one calls it, however, we should recognize it for what it is and ask why it exists. This Article attempts to do just that.

While Chevron Step One-and-a-Half has up until now been a doctrine without a name, it is nonetheless a doctrine with skeptics. Perhaps chief among them is now–Chief Justice John Roberts. In a concurring opinion that he wrote while on the DC Circuit, then-Judge Roberts criticized his colleagues for invoking Prill and its progeny.15 To be sure, Roberts said that he had “no quarrel with the basic proposition . . . that when an agency erroneously concludes that a statutory interpretation is required by Congress, we should remand to give the agency an opportunity to interpret the statute in the first instance.”16 But he argued that the doctrine ought not apply unless there is “real and genuine doubt concerning what interpretation the agency would choose” if the agency were aware of the ambiguity.17 In all other cases, Roberts said, the doctrine we dub Chevron Step One-and-a-Half “outstrips its rationale”18 and “convert[s] judicial review of agency action into a ping-pong game.”19 Professor Nicholas Bagley, in turn, has urged that such “real and genuine doubt” will be vanishingly rare. “The very fact that an agency has read the statute in a particular way,” wrote Bagley, is itself strong “evidence” that the agency “prefers the interpretation it adopted to the one that it did not adopt.”20 And quite often that probative evidence will be backed up by a representation—in writing and signed by agency lawyers—stating that the agency would stay the course on remand even if the reviewing court were to conclude that the statute is susceptible of multiple meanings.21 In light of this reality, the chief justice and Bagley would have Chevron Step One-and-a-Half be invoked infrequently rather than become a standard part of the Chevron analysis.22

Indeed, regardless of whether one subscribes to the view of the chief justice and Bagley, one might wonder why Chevron Step One-and-a-Half is ever invoked. That is, why would an agency ever insist that a statute is unambiguous? From an agency’s perspective, disclaiming ambiguity seems like a self-inflicted wound, and an easily avoided self-inflicted wound at that. In fact, if agencies seek to enhance their own autonomy (as some common accounts of agency motivation maintain23), one might expect them always to argue that the statute is ambiguous so long as a nonfrivolous argument for ambiguity exists. After all, under Chevron, ambiguity acts as a grant of discretion,24 and discretion is power.25 Why would an agency disavow discretion it could credibly claim?

Yet Chevron Step One-and-a-Half remands have occurred in dozens of cases—including in high-profile matters.26 This means that agencies are denying (or at least declining to acknowledge) that the relevant statute is ambiguous, even when nonfrivolous arguments in favor of ambiguity are available (arguments that a court ultimately concludes are meritorious). So what is causing these agencies to stumble on the path to Step Two? We suggest several reasons why agencies might maintain that a statute is unambiguous even though such an assertion disadvantages the agency in litigation. The first, and perhaps most mundane, is agency ignorance: agency lawyers may simply be unaware of the Chevron Step One-and-a-Half doctrine. But while ignorance might be to blame the first time that an agency encounters the doctrine, the explanation is less plausible as time goes on, especially for agencies that have had actions remanded by the DC Circuit on Chevron Step One-and-a-Half grounds on multiple occasions.27 A second reason is agency ambivalence: The agency might believe that a particular result is statutorily compelled while also being unconvinced that it is an optimal policy. The agency might decide that if a court can find ambiguity when the agency could not, then the agency would be happy to consider adopting an alternative approach. We think that this too is a plausible explanation in some cases, though not in all.

We suggest here that even when agency actors are aware of the Chevron Step One-and-a-Half doctrine, and even when they have a firm preference for a particular interpretation, they sometimes may have strategic reasons to say that a statute is unambiguous despite the potentially negative litigation consequences. One such strategic motivation involves intra-agency politics: agency lawyers who prefer a particular outcome might claim that their preferred outcome is statutorily ordained, anticipating that nonlawyers within the agency will be ill-equipped to contest that claim. A second strategic motivation involves intra–executive branch politics: an agency might claim that a particular result is statutorily compelled so as to avoid having to convince the White House’s Office of Information and Regulatory Affairs (OIRA) that the agency’s preferred outcome is cost justified relative to feasible alternatives. A third strategic motivation involves interbranch politics: an agency might seek to shirk blame for an unpopular policy by claiming that responsibility lies with Congress. Fourth, an agency might maintain that a statute unambiguously points one way in an effort to prevent future administrations from choosing a different route.

After laying out the various reasons why Chevron Step One-and-a-Half cases might continue to arise, we consider whether these reasons justify the doctrine’s existence. Despite the hostility that the doctrine faces from some quarters, we suggest that Step One-and-a-Half produces potentially important benefits. In particular, this half step advances the values that motivate (and help justify) Chevron in the first place. If agencies are entrusted with discretionary power on the grounds that they are more accountable than courts, then judicial review should encourage agencies to take responsibility for their decisions. Chevron Step One-and-a-Half can serve to encourage this accountability in an administrable way. It can also help to ensure that deference is reserved for cases in which agencies employ the expertise that they (at least ostensibly) have and that courts do not have. To be sure, the doctrine imposes costs as well, in the form of judicial remands and further litigation. Neither the benefits nor the costs are easily quantifiable, and we cannot confidently say whether the net welfare effects of the doctrine are positive or negative. What we can say, though, is that the potential benefits of Chevron Step One-and-a-Half have been overlooked so far. This Article seeks to bring those potential benefits to the fore.

This Article concludes by considering the choices courts must make when applying the Chevron Step One-and-a-Half doctrine. For example, should the doctrine be triggered whenever an agency fails to acknowledge that a statutory provision might be ambiguous—or only when the agency affirmatively states that the statutory provision is clear? Our proffered justifications for the doctrine might counsel in favor of the former, but the DC Circuit tends to lean toward the latter view. Likewise, should the application of Chevron Step One-and-a-Half result in a remand to the agency—or should it also result in vacatur of the agency’s rule? Remand without vacatur reduces Chevron Step One-and-a-Half’s costs, and yet those “costs” are arguably the doctrine’s virtues—the costs of Chevron Step One-and-a-Half potentially deter agencies from hiding their cards.

We realize, of course, that Chevron Step One-and-a-Half further complicates the Chevron analysis. The simpler version of Chevron set out in administrative law casebooks, however, does not descriptively reflect what is happening in the nation’s “administrative law court.”28 We hope that by naming the Chevron Step One-and-a-Half doctrine, explaining how it might be justified, delineating its contours, and describing its applications, we can cut through some of the complexity.

* * *

This Article proceeds as follows. Part I introduces Chevron Step One-and-a-Half with an illustrative example. Part II describes the origins of this doctrine and explains why Step One-and-a-Half is consistent with, but nonetheless distinct from, other administrative law doctrines like hard-look review and the bar on post hoc rationalizations. Part III then addresses a key puzzle presented by Step One-and-a-Half: Why do these cases continue to arise? The answer, we submit, is central to why Chevron Step One-and-Half can be a beneficial doctrine. Part IV then tentatively defends Step One-and-a-Half from its critics by demonstrating the connection between the doctrine and Chevron’s accountability and expertise justifications; whatever one thinks of the doctrine, it is a mistake to focus exclusively on its costs while ignoring its benefits. Finally, Part V considers some of the difficult choices that courts must make when applying Chevron Step One-and-a-Half in concrete cases.

  • 1. 467 US 837 (1984).
  • 2. See id at 842–43.
  • 3. Thomas W. Merrill and Kristin E. Hickman, Chevron’s Domain, 89 Georgetown L J 833, 836 (2001). See also Cass R. Sunstein, Chevron Step Zero, 92 Va L Rev 187, 191 n 19 (2006) (crediting Merrill and Hickman for the article’s title).
  • 4. See, for example, William S. Jordan III, Judicial Review of Informal Statutory Interpretations: The Answer Is Chevron Step Two, Not Christensen or Mead, 54 Admin L Rev 719, 725 (2002) (“[T]he Court erected a new four step test to replace what we once knew as the Chevron two step.”).
  • 5. Matthew C. Stephenson and Adrian Vermeule, Chevron Has Only One Step, 95 Va L Rev 597, 598 n 4, 599 (2009). But see Kenneth A. Bamberger and Peter L. Strauss, Chevron’s Two Steps, 95 Va L Rev 611, 624–25 (2009) (rejecting Stephenson and Vermeule’s argument); Richard M. Re, Should Chevron Have Two Steps?, 89 Ind L J 605, 608 (2014) (same).
  • 6. See Alaska Wilderness League v Jewell, 788 F3d 1212, 1217 (9th Cir 2015) (“[U]nder Chevron . . . we engage in a three-step inquiry when reviewing an agency’s interpretation of a statute.”); Restrepo v Attorney General of United States, 617 F3d 787, 792 (3d Cir 2010) (“When confronted with a potential Chevron application, we administer a three-step analysis.”).
  • 7. See, for example, Carter v Welles–Bowen Realty, Inc, 736 F3d 722, 731 (6th Cir 2013) (Sutton concurring) (arguing that “[i]f you believe that Chevron has two steps, you would” reach a result one way, and “[i]f you believe that Chevron has only one step,” you would reach the same result another way).
  • 8. See, for example, Brett M. Kavanaugh, Book Review, Fixing Statutory Interpretation, 129 Harv L Rev 2118, 2154 (2016):

    In short, the problem with certain applications of Chevron, as I see it, is that the doctrine is so indeterminate—and thus can be antithetical to the neutral, impartial rule of law—because of the initial clarity versus ambiguity decision. . . . [W]e need to consider eliminating that inquiry as the threshold trigger.

    See also Jack M. Beermann, End the Failed Chevron Experiment Now: How Chevron Has Failed and Why It Can and Should Be Overruled, 42 Conn L Rev 779, 834 n 218 (2010) (“The number of steps in Chevron in any given case may turn out to depend on who writes the Court’s opinion.”).

  • 9. But others have taken it further. See, for example, Northpoint Technology, Ltd v Federal Communications Commission, 412 F3d 145, 151 (DC Cir 2005) (Henderson) (“[U]nder the Chevron two-step, we stop the music at step one if the Congress has directly spoken to the precise question at issue. . . . But if the statute is silent or ambiguous, we dance on . . . [to] step two.”) (quotation marks omitted); Braintree Electric Light Department v Federal Energy Regulatory Commission, 667 F3d 1284, 1288 (DC Cir 2012) (Garland) (“[T]he Chevron two-step is a dance for the court.”).
  • 10. Chevron, 467 US at 842–43.
  • 11. 755 F2d 941 (DC Cir 1985).
  • 12. See id at 942. See also Nicholas Bagley, Remedial Restraint in Administrative Law, 117 Colum L Rev 253, 296–301 (2017) (discussing “Prill cases,” “Prill claims,” and “Prill violation[s]”).
  • 13. 555 US 511 (2009).
  • 14. See notes 130–41 and accompanying text.
  • 15. See PDK Laboratories Inc v United States Drug Enforcement Administration, 362 F3d 786, 808–09 (DC Cir 2004) (Roberts concurring in part and concurring in the judgment).
  • 16. Id at 808 (Roberts concurring in part and concurring in the judgment).
  • 17. Id (Roberts concurring in part and concurring in the judgment).
  • 18. Id at 809 (Roberts concurring in part and concurring in the judgment).
  • 19. PDK Laboratories, 362 F3d at 809 (Roberts concurring in part and concurring in the judgment), quoting Time, Inc v United States Postal Service, 667 F2d 329, 335 (2d Cir 1981).
  • 20. Bagley, 117 Colum L Rev at 297 (cited in note 12).
  • 21. See id at 297–98.
  • 22. See PDK Laboratories, 362 F3d at 809 (Roberts concurring in part and concurring in the judgment); Bagley, 117 Colum L Rev at 301 (cited in note 12).
  • 23. See, for example, John C. Coffee Jr, The Future as History: The Prospects for Global Convergence in Corporate Governance and Its Implications, 93 Nw U L Rev 641, 702 (1999) (“[T]he occasions are rare on which a regulatory agency has voluntarily ceded control without some compelling need that required such a surrender. The usual assumptions of political science are that public agencies act to maximize their powers, just as private firms seek to maximize revenues or profits.”). See also Talk America, Inc v Michigan Bell Telephone Co, 564 US 50, 69 (2011) (Scalia concurring) (“By contrast, deferring to an agency’s interpretation of its own rule encourages the agency to enact vague rules which give it the power, in future adjudications, to do what it pleases.”).
  • 24. See Peter L. Strauss, “Deference” Is Too Confusing—Let’s Call Them “Chevron Space” and “Skidmore Weight”, 112 Colum L Rev 1143, 1145 (2012) (explaining that Chevron “empower[s]” agencies “to act in a manner that creates legal obligations”).
  • 25. See, for example, Burlington Truck Lines, Inc v United States, 371 US 156, 167 (1962).
  • 26. See, for example, Teva Pharmaceuticals USA, Inc v Food & Drug Administration, 441 F3d 1, 4–5 (DC Cir 2006) (declaring in a case involving a major cholesterol drug that “[w]e therefore generally remand for an agency to make the first interpretation of an ambiguous statutory term when it has failed to do so previously”).
  • 27. The Federal Communications Commission, the Department of Health and Human Services, the Department of Transportation, and the National Labor Relations Board are all repeat losers before the DC Circuit in Chevron Step One-and-a-Half cases. See Online Appendix.
  • 28. The Contribution of the D.C. Circuit to Administrative Law, 40 Admin L Rev 507, 509 (1988) (quoting remarks by Judge Patricia Wald).