Deference Conservation—FOIA’s Lessons for a Chevron-less World

John C. Brinkerhoff Jr. & Daniel B. Listwa *


In SAS Institute Inc. v. Iancu, 1 138 S. Ct. 1348 (2018). the Supreme Court entered the next chapter in the long-winding debate over Chevron deference, which instructs courts to defer to an agency’s reasonable interpretation of its substantive statutes. 2 See Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 843 (1984). Writing for a five-member majority, Justice Gorsuch refused to affirm the doctrine, noting portentously that “whether Chevron should remain is a question we may leave for another day.” 3 Iancu, 138 S. Ct. at 1358. While defending the doctrine, Justice Breyer’s dissent declined a full-throated endorsement, instead classifying it as a simple “rule of thumb, guiding courts in an effort to respect that leeway which Congress intended the agencies to have.” 4 Id. at 1364 (Breyer, J., dissenting).

By teeing up a debate on the future of Chevron, Iancu calls attention to a fundamental question: What would it mean to do away with agency deference? Both Congress and the courts have debated deference as though it were amenable to easy intervention. Prior to Iancu, Justice Thomas and then-Judge Gorsuch both questioned Chevron’s constitutionality. 5 See Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring); Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1149 (10th Cir. 2016) (Gorsuch, J., concurring). After Iancu, Justice Kennedy called for the Court to “reconsider . . . the premises that underlie Chevron,” accusing the doctrine of producing “reflexive deference.” 6 Pereira v. Sessions, 138 S. Ct. 2105, 2120-21 (2018) (Kennedy, J., concurring). In Congress, one of the first pieces of legislation that the House passed following President Trump’s inauguration sought to eliminate Chevron deference by requiring that courts “decide de novo all relevant questions of law.” 7 The Act was passed within a broader regulatory reform bill. Regulatory Accountability Act of 2017, H.R. 5, 115th Cong. § 202. Such legislation has been (and almost certainly will continue to be) a priority of conservative activists. 8 Nearly identical bills were introduced in the previous Congress. See Separation of Powers Restoration Act of 2016, H.R. 4768, 114th Cong. § 2; Separation of Powers Restoration Act of 2016, S. 2724, 114th Cong. § 2. Given judicial rhetoric in Iancu and elsewhere, however, it appears that the Court might beat Congress to the punch.

The academic debate has, at times, examined the assumption that agency deference can be easily mandated away. Most notably, some commentators have taken the position that placing a de novo review provision in the Administrative Procedure Act (APA) would functionally do little to stop Chevron-type deference to statutory interpretations, 9 See, e.g., Nicholas R. Bednar & Kristin E. Hickman, Chevron‘s Inevitability, 85 Geo. Wash. L. Rev. 1392, 1408, 1460 (2017). building from William Eskridge and Lauren Baer’s findings that agencies often receive significant deference regardless of the formal regime applied. 10 See William N. Eskridge, Jr. & Lauren E. Baer, The Continuum of Deference: Supreme Court Treatment of Agency Statutory Interpretations from Chevron to Hamdan, 96 Geo. L.J. 1083, 1098, 1099 & tbl.1 (2008). Opponents of this position, such as Kent Barnett and Christopher Walker, counter with statistical evidence that Chevron has impacted decision-making within both the lower courts and the agencies themselves. 11 See Kent Barnett & Christopher J. Walker, Chevron in the Circuit Courts, 116 Mich. L. Rev. 1, 6 (2017) (addressing lower court use of Chevron); Christopher J. Walker, Inside Agency Statutory Interpretation, 67 Stan. L. Rev. 999, 1005 (2015) (addressing the role that Chevron plays in agencies’ internal processes).

This Essay offers a modest contribution to this debate. Rather than speculating from current usages of Chevron, it takes the novel approach of looking to another area of administrative law where courts already apply de novo review: Freedom of Information Act (FOIA) litigation. FOIA litigation reveals that those who argue that abrogating Chevron would have limited practical effect are right—but not for the reasons they think. In the FOIA context, rather than deferring to agencies on matters of statutory interpretation, courts have shifted deferential treatment to other aspects of litigation. The result is that the government wins 90% of the time, 12 Paul R. Verkuil, An Outcomes Analysis of Scope of Review Standards, 44 Wm. & Mary L. Rev. 679, 719, 730 (2002); accord 1 James T. O’Reilly, Federal Information Disclosure 227 (2015) (reaffirming the win rate). but the de novo standard is formally maintained.

To explain this phenomenon, we posit that courts engage in “deference conservation”—the idea that where Congress forecloses one mode of deference, alternative forms of deference become more pronounced. It is impossible to completely eliminate ambiguities from statutory interpretation. 13 Of course, judges have considerably different thresholds before finding such ambiguity. Compare Robert A. Katzmann, Response to Judge Kavanaugh’s Review of Judging Statutes, 129 Harv. L. Rev. F. 388, 398 (2016) (“[T]here will always be many cases where the statute can be viewed as ambiguous under any reading, where the statute is reasonably susceptible to more than one interpretation.”), with Raymond M. Kethledge, Ambiguities and Agency Cases: Reflections After (Almost) Ten Years on the Bench, 70 Vand. L. Rev. En Banc 315, 319-20 (2017) (explaining that “it should take a great deal more than a couple of competing dictionary definitions” to find ambiguity, while conceding that statutory ambiguities “happen, but they are pretty rare”). This Essay does not attempt to assess the appropriate threshold. How courts choose to resolve those ambiguities has policy consequences. In administrative law, judges have preferred to defer to the politically accountable executive. While Congress may be able to restrict deference by legislating a standard of review, pressures to defer remain—and, ultimately, a simple command for de novo statutory review will not prevent deference from manifesting somewhere in a court’s decision-making.

As its name suggests, the upshot of deference conservation is that existing agency win rates reflect deeper institutional concerns that will exert a gravitational force against changes to standards of review. In the case of FOIA litigation, the 90% win rate is commensurate with the near-total control over information that executive agencies enjoyed prior to FOIA’s enactment. 14 See John C. Brinkerhoff Jr., FOIA’s Common Law, 36 Yale J. on Reg. (forthcoming 2019) (manuscript at 21-22), Under the APA, commentators have pegged the government’s win rate at 60% to 70%. 15 David Zaring, Reasonable Agencies, 96 Va. L. Rev. 135, 169 (2010). Deference conservation contends that this outcome will remain roughly even in Chevron’s absence.

Part I introduces deference conservation through an analysis of judicial reactions to de novo review in FOIA litigation. Specifically, it finds that courts have shifted deference to agency assertions of fact and more subtly, adopted agency-friendly legal interpretations and procedural rules. Part II discusses what deference conservation would look like under a de novo APA system, identifying potential areas for future conservation and pointing to doctrinal areas in which it has already occurred in response to previous deference-constraining changes in law. Finally, Part III lays out the stakes of the principle.

I. FOIA’s Experience with Deference Conservation

FOIA litigation provides a clear example of Chevron-less administrative law—and thus, of deference conservation. A highly pro-plaintiff statute, FOIA mandates a “strong presumption in favor of disclosure,” 16 U.S. Dep’t of State v. Ray, 502 U.S. 164, 173 (1991). expressly provides for de novo review, 17 5 U.S.C. § 552(a)(4)(B) (2016). and demands that courts “narrowly construe[]” the statute’s nine exemptions, which are the “exclusive” avenues through which the government can withhold responsive information. 18 See Milner v. Dep’t of the Navy, 562 U.S. 562, 565 (2011) (first quoting FBI v. Abramson, 456 U.S. 615, 630 (1982); and then quoting EPA v. Mink, 410 U.S. 73, 79 (1973)). As commentators have concluded, FOIA’s provisions represent a desire to eliminate, in the FOIA context, deferential treatment of agency decision-making. 19 See Margaret B. Kwoka, Deferring to Secrecy, 54 B.C. L. Rev. 185, 196-99 (2013); David E. Pozen, Freedom of Information Beyond the Freedom of Information Act, 165 U. Pa. L. Rev. 1097, 1118 (2017); Nathan Slegers, Comment, De Novo Review Under the Freedom of Information Act: The Case Against Judicial Deference to Agency Decisions To Withhold Information, 43 San Diego L. Rev. 209, 216-18 (2006).

As noted, however, courts have effectively circumvented this mandate, creating a 90% government win rate—far higher than even under Chevron. While a comprehensive examination of the underlying reasons for this departure is beyond this Essay’s scope, two tactics warrant analysis: deference to agency affidavits and the courts’ tendency to adopt pro-government statutory and procedural interpretations. These strategies enable courts to defer to agencies without formally abrogating Congress’s imposition of de novo review.

A. Deference to Agency Affidavits

While courts acknowledge that they review FOIA decisions de novo, they have nonetheless established formal deference regimes for the affidavits that agencies submit to support summary judgment. Because discovery is nonexistent in FOIA litigation and nearly all cases are resolved by summary judgment, 20 See infra Part I.B. government affidavits are typically the only “factual” material in play. Thus, the government’s burden is the same for every case:

If an agency’s affidavit describes the justifications for withholding the information with specific detail, demonstrates that the information withheld logically falls within the claimed exemption, and is not contradicted by contrary evidence in the record or by evidence of the agency’s bad faith, then summary judgment is warranted on the basis of the affidavit alone. 21 ACLU v. U.S. Dep’t of Def., 628 F.3d 612, 619 (D.C. Cir. 2011).

As outlined below, courts have developed exemption-specific standards under which affidavits are assessed, which speak to the degree to which an agency must factually support its conclusory statements to meet its legal burden. Due to this deference, agencies are able to bend their factual assertions to meet their legal burdens. Analytically, this is the reverse of Chevron, where questions of law get special solicitude. But the result is the same—the agency wins.

The most consequential deference regimes exist in FOIA’s exemptions for national security and law enforcement issues. For the former, courts affirm agency decisions that “appear[] ‘logical’ or ‘plausible,’” 22 Wolf v. CIA, 473 F.3d 370, 374-75 (D.C. Cir. 2007) (quoting Gardels v. CIA, 689 F.2d 1100, 1105 (D.C. Cir. 1982)); accord Hamdan v. U.S. Dep’t of Justice, 797 F.3d 759, 774 (9th Cir. 2015); ACLU v. Dep’t of Justice, 681 F.3d 61, 69 (2d Cir. 2012). typically offering loaded statements such as that the judiciary is “in no position to second-guess” such decisions. 23 Stein v. Dep’t of Justice, 662 F.2d 1245, 1254 (7th Cir. 1981). For the latter, courts have deployed a variety of regimes depending on the issues in play. For example, in many circuits, law enforcement agencies need only provide “a colorable claim of a rational nexus” 24 See Pratt v. Webster, 673 F.2d 408, 420 (D.C. Cir. 1982) (quoting Malizia v. U.S. Dep’t of Justice, 519 F. Supp. 338, 347 (S.D.N.Y. 1981)), superseded on other grounds by statute, Freedom of Information Reform Act of 1986, § 1802(a), Pub. L. No. 99-570, 100 Stat. 3207, 3207-48 to -49 (codified at 5 U.S.C. § 552(b)(7) (2016)), as recognized in 100Reporters LLC v. U.S. Dep’t of Justice, 248 F. Supp. 3d 115, 160 (D.D.C. 2017). between information and a law enforcement function to meet FOIA’s requirement that the information be “compiled for law enforcement purposes.” 25 Pratt, 673 F.2d at 410 (quoting 5 U.S.C. § 552(b)(7) (1976 & Supp. IV 1980) (current version at 5 U.S.C. § 552(b)(7) (2016))). By accepting threadbare or conclusory statements, courts are, in effect, deferring to agency legal determinations by functionally lowering the bar for what it means to factually support a claim. 26 Examples of such conflations between law and fact to achieve deference exist outside of administrative law. See, e.g., Alex Carver, Note, Rethinking Conspiracy Jurisdiction in Light of Stream of Commerce and Effects-Based Jurisdictional Principles, 71 Vand. L. Rev. 1333, 1342 (2018) (observing similar deference to the legal question of whether conspiracy jurisdiction exists by accepting as true a plaintiff’s conclusory factual allegations that a conspiracy exists).

Deference conservation is perhaps predictable in such high stakes areas. But, importantly, deference to affidavits extends to other exemptions. For example, courts grant agencies “considerable” 27 Pfeiffer v. CIA, 721 F. Supp. 337, 339-40 (D.D.C. 1989). or “particular” 28 Island Film, S.A. v. Dep’t of the Treasury, 869 F. Supp. 2d 123, 134-35 (D.D.C. 2012). deference when they invoke the deliberative process privilege, which requires agencies to prove that disclosure would “discourage candid discussion within the agency.” 29 Access Reports v. Dep’t of Justice, 926 F.2d 1192, 1195 (D.C. Cir. 1991) (quoting Dudman Commc’ns Corp. v. Dep’t of the Air Force, 815 F.2d 1565, 1568 (D.C. Cir. 1987)). Additionally, courts “generally” defer to agencies under the confidential commercial information exemption due to the executive’s presumed expertise in the subject area. 30 Ctr. for Auto Safety v. U.S. Dep’t of Treasury, 133 F. Supp. 3d 109, 120 (D.D.C. 2015) (quoting United Techs. Corp. v. U.S. Dep’t of Def., 601 F. 3d 557, 563 (D.C. Cir. 2010)). The rationale underlying this deference arose in the “reverse-FOIA” context, see CNA Fin. Corp. v. Donovan, 830 F.2d 1132, 1155-56 (D.C. Cir. 1987), which occurs when a third party sues to reverse an agency’s decision to disclose information. The Center for Auto Safety Court imported this deference to FOIA litigation by directly quoting a reverse FOIA case.

That deference exists in these areas is particularly significant. While factual deference in national security and law enforcement has been noted outside of the FOIA context, 31 See Robert M. Chesney, National Security Fact Deference, 95 Va. L. Rev. 1361, 1380 (2009). no such solicitude is associated with these more mundane matters of fact-finding in areas other than FOIA litigation. On the other hand, statutory deference is commonly afforded in a closely related set of matters—agency procedure. Courts have recognized that Chevron deference is broadly appropriate for an agency’s internal procedural matters, a determination motivated at least partially by the idea that an agency is better positioned to understand a decision’s effects on its functioning. 32 See, e.g., Chem. Waste Mgmt. v. EPA, 873 F.2d 1477, 1478-83 (D.C. Cir. 1989) (applying Chevron deference to give the agency significant discretion in choosing between formal and informal adjudication); see also Vt. Yankee Nuclear Power Corp. v. Nat. Res. Def. Council, 435 U.S. 519, 544 (1978) (quoting Fed. Power Comm’n v. Transcon. Gas Pipe Line Corp., 423 U.S. 326, 333 (1976) (per curiam)) (reasserting the “very basic tenet of administrative law that agencies should be free to fashion their own rules of procedure” on the grounds that agencies should have “discretion in deciding how, in light of internal organization considerations, it may best proceed”). This same concern would seem to justify deference regarding matters of deliberative process, illustrating a form of factual deference that addresses a concern that would otherwise be resolved through statutory deference.

All of these factual deference regimes act as effective substitutes for statutory deference by accepting as sufficient what would be—in other circumstances—insufficient factual showings. Functionally, then, courts defer to agencies in a sub-rosa manner on the legal question of what set of facts is sufficient to substantiate the relevant legal standard, circumventing Congress’s imposition of de novo review.

B. Putting a Thumb on the Scale

This Subpart outlines another, broader group of methods by which courts undercut de novo review: the tendency to interpret FOIA’s substance and procedures in a manner that systematically favors the government. These interpretations run counter to explicit statutory directives to limit agency withholdings to “specifically stated” exemptions 33 5 U.S.C. § 552(d) (2016). and to place the burden of proof on the government. 34 Id. § 552(a)(4)(B). And yet, the interpretations are maintained by courts that are unwilling to withdraw from the executive branch the deference they have determined is warranted.

Looking to substance, numerous commentators have concluded that FOIA case law is riddled with pro-government precedent. 35 See, e.g., David E. Pozen, Note, The Mosaic Theory, National Security, and the Freedom of Information Act, 115 Yale L.J. 628, 631 (2005). Providing an exception that proves the rule, the Supreme Court recently savaged a forty-year-old, previously unquestioned interpretation—one of FOIA’s most consequential pro-government doctrines—noting that “the only way to arrive at [the interpretation] is by taking a red pen to the statute—‘cutting out some’ words and ‘pasting in others’ until little of the actual provision remains.” 36 Milner v. Dep’t of the Navy, 562 U.S. 562, 573 (2011) (quoting Elliott v. U.S. Dep’t of Agric., 596 F.3d 842, 845 (D.C. Cir. 2010)). These sorts of interpretations—common and largely unchallenged in FOIA case law—clearly do not represent narrow constructions in favor of disclosure. 37 See Brinkerhoff, supra note 14 (manuscript at 11-21) (detailing other examples).

A subtler deference-conferring tactic is the judiciary’s use of procedural levers. Courts have utilized neutral-sounding procedural rules to favor government exemption claims. Margaret Kwoka has collectively labeled these procedures “unspoken deference.” 38 Kwoka, supra note 19, at 221. Among other procedural rules, Kwoka identified three relevant areas of deference: (1) the general ban on discovery in FOIA cases; (2) the near-inescapable demand that FOIA cases be settled at the summary judgment stage, which prevents challenging agency affidavits’ credibility; and (3) the widespread practice of granting agencies multiple summary judgment attempts should they fail. 39 See id. at 224-25, 228-29, 231-32. Other areas of “unspoken deference” also exist. For example, courts frequently refuse to give weight to the testimony of plaintiffs’ witnesses, reasoning that even when the witnesses are “eminent and well informed,” they are not “entitled to the deference accorded to those who have [a] statutory duty.” 40 Berman v. CIA, 378 F. Supp. 2d 1209, 1218 (E.D. Cal. 2005); see also Rush v. Dep’t of State, 748 F. Supp. 1548, 1549, 1551, 1554 (S.D. Fla. 1990) (rejecting an affidavit written by the ambassador who created the records at issue); Wash. Post Co. v. Dep’t of Def., No. 84-2949, 1987 U.S. Dist. LEXIS 16108, at *19-20 (D.D.C. Feb. 25, 1987) (rejecting an affidavit written by a U.S. Senator who read the information as a member of the Foreign Relations Committee).

Both the substantive and procedural measures can be understood as the judiciary putting a pro-government thumb on the scale—in direct contravention of FOIA’s mandate. In so doing, courts can achieve the deference they find institutionally mandated.

II. The APA’s Susceptibility to Deference Conservation

Judicial responses to FOIA’s congressionally mandated de novo review provides a central insight for current Chevron debates: Courts can and most likely will blunt the impact of a statute or holding that mandates, without more, de novo review of agency statutory interpretations. The APA is perhaps even more susceptible to this outcome than FOIA. Commentators have long observed the common-law-like approach courts take to the APA, where large swaths of doctrine are drawn not from statutory interpretation but institutional, “pragmatic and normative concerns . . . developed incrementally through precedent.” 41 Gillian E. Metzger, Embracing Administrative Common Law, 80 Geo. Wash. L. Rev. 1293, 1295 (2012). Under this view, the APA functions as a “restatement” of preexisting equity doctrines that are still subject to judicial control. See Steven J. Lindsay, Note, Timing Judicial Review of Agency Interpretations in Chevron‘s Shadow, 127 Yale L.J. 2448, 2478 & n.162 (2018). This methodological approach clearly would allow courts to formally drop Chevron deference, while amending doctrine elsewhere to counterbalance this shift.

Perhaps the most susceptible area for deference conservation can be found in the judiciary’s tendency to strategically conflate issues of law and fact. In NLRB v. Hearst Publications, Inc., an important pre-APA case, the Court justified its use of heightened deference by characterizing what was largely a question of law—whether “newsboys” were employees or independent contractors—as a question of fact. 42 322 U.S. 111, 124-31 (1944), overruled in part by Nationwide Mut. Ins. Co. v. Darden, 503 U.S. 318 (1992). Jerry Mashaw describes this characterization as but one of the “different means” with which “[t]he Supreme Court has experimented . . . for allocating policy choice to agencies or to courts.” 43 Jerry L. Mashaw, Rethinking Judicial Review of Administrative Action: A Nineteenth Century Perspective, 32 Cardozo L. Rev. 2241, 2243 (2011).

In modern jurisprudence, courts have already effectively collapsed the two separate APA-mandated standards of factual review—“arbitrary and capricious” and “substantial evidence”—such that the differences between them are “largely semantic.” 44 See Ass’n of Data Processing Serv. Orgs., Inc. v. Bd. of Governors of the Fed. Reserve Sys., 745 F.2d 677, 684 (D.C. Cir. 1984) (quoting Aircraft Owners & Pilots Ass’n v. FAA, 600 F.2d 965, 971 n.28 (D.C. Cir. 1979)). In the related comparison between the “substantial evidence” and “clearly erroneous” standards for review of fact, the Court has conceded that it could not find a single case in which the standards would produce different outcomes, chalking this result up to the “intangible factors” that encompass each instance of factual review. 45 See Dickinson v. Zurko, 527 U.S. 150, 162-63 (1999). This malleable area of review is ripe for the sort of counterbalancing we anticipate would occur were Chevron to be set aside. Courts could compensate for the more searching review demanded on legal conclusions by subtly shifting what factual assumptions they consider sufficiently reasonable to survive review.

Furthermore, deference conservation is not merely theoretical in the APA context. Courts have already deployed it in certain procedural areas. An apt example can be seen in the “remand without vacatur” doctrine, where courts remand a rule for deficiencies without vacating it. Courts’ use of this doctrine was once almost nonexistent. 46 See Richard J. Pierce, Jr., Seven Ways to Deossify Agency Rulemaking, 47 Admin. L. Rev. 59, 75 (1995). However, this reticence dissipated following the Court’s holding in Bowen v. Georgetown University Hospital that agencies generally could not adopt a rule that applies retroactively. 47 488 U.S. 204, 208 (1988). Citing directly to the dire consequences of Georgetown, the D.C. Circuit formulated a test for using remand without vacatur and countenanced its use. 48 Allied-Signal, Inc. v. U.S. Nuclear Regulatory Comm’n, 988 F.2d 146, 150-51 (D.C. Cir. 1993). A decisive majority of circuit courts subsequently deployed the doctrine with greater frequency, 49 See Stephanie J. Tatham, Admin. Conference of the U.S., The Unusual Remedy of Remand Without Vacatur 27 (2014) (finding eight circuits that have deployed the doctrine). effectively mitigating the practical consequences of Georgetown without touching its actual holding.

There is nothing special about deference that would shield it from these effects. Chevron itself is an administrative common law creation. The Chevron Court did not cite the APA a single time, reasoning instead from policy and pre-APA precedent. 50 See John F. Duffy, Administrative Common Law in Judicial Review, 77 Tex. L. Rev. 113, 192 (1998). As an empirical matter, lower courts already resist the Court’s attempts to limit Chevron’s application. Barnett and Walker note that courts “largely ignore” the Court’s signals to not apply Chevron to “sensitive matters.” 51 Barnett & Walker, supra note 11, at 72. And when considering the balancing tests that determine if Chevron applies, they rarely invoke certain factors that would increase their discretion. 52 See id. Thus, as Judge Kethledge contends, where statutes often present “a dense undergrowth of sections and subsections and subsections,” an agency’s interpretation “offer[s] a path already cleared.” 53 Kethledge, supra note 13, at 324. Unsurprisingly, even if the correct answer remains hidden, courts often prefer to “proceed[] straightaway down the cleared path.” 54 Id.

Administrative common law gives courts weapons to resist congressional attempts to overturn Chevron and force courts to cut through the undergrowth—characterizing questions of law as questions of fact, closing off review through procedural or jurisdictional determinations, and preserving erroneous agency decisions. The experience with FOIA suggests that such alternative forms of deference would become more pronounced if courts could not officially defer to agencies on matters of statutory construction.

III. Implications of Deference Conservation

Ambiguities cannot be entirely removed from statutory interpretation. 55 See supra note 13; see also The Federalist No. 37, at 229 (James Madison) (Clinton Rossiter ed., 1961) (“[N]ew laws, though penned with the greatest technical skill and passed on the fullest and most mature deliberation, are considered as more or less obscure and equivocal . . . .”). While judges have historically deployed aides, such as interpretive canons, to ascertain meaning in the face of ambiguity, 56 See, e.g., John C. Brinkerhoff Jr., Ropes of Sand: State Antitrust States Bound by Their Original Scope, 34 Yale J. on Reg. 353, 380 n.168 (2017) (noting the use of common law holdings to interpret ambiguous provisions in statutes). deference regimes have long been their preference in the administrative law context. This preference is strong enough that the judiciary openly admits to subjugating its best judgment on legal interpretation—ostensibly its constitutional wheelhouse—to agency conclusions. 57 See Antonin Scalia, Judicial Deference to Administrative Interpretations of Law, 1989 Duke L.J. 511, 521 (noting that while the frequency with which this subjugation occurs depends on the willingness of judges to find ambiguity, Chevron still calls for judges to occasionally set aside their best interpretive judgments).

However, deference conservation suggests that Chevron and its compatriots are more than simple canons or tiebreakers. Rather, courts have infused them with implicit, quasi-constitutional force rooted in institutional competencies. Indeed, the same institutional insecurities that courts have used to justify their defiance of de novo review in FOIA cases were also present in Chevron itself, where the Court reasoned that “[j]udges are not experts in the field, and are not part of either political branch of the Government.” 58 Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 865 (1984).

But this conclusion raises additional considerations on both sides of the Chevron issue. First, it challenges Chevron’s supporters to recognize the heightened stakes of the debate, in which Chevron is not the only variable in the broader scheme of a court’s procedural and substantive levers. With the possible exception of Justice Scalia’s originalist arguments, 59 See United States v. Mead Corp., 533 U.S. 218, 241-242 (2001) (Scalia, J., dissenting) (“[Chevron] was in accord with the origins of federal-court judicial review . . . principally exercised through the prerogative writ of mandamus.”). But see Aditya Bamzai, The Origins of Judicial Deference to Executive Interpretation, 126 Yale L.J. 908, 913-14 (2017) (challenging this conclusion). prominent justifications of Chevron largely focus on non-constitutional arguments, such as Justice Breyer’s “rule of thumb” defense 60 See SAS Inst. Inc. v. Iancu, 138 S. Ct. 1348, 1364 (2018) (Breyer, J., dissenting). or the framing of Chevron as a form of judicial abstention. 61 See David M. Hasen, The Ambiguous Basis of Judicial Deference to Administrative Rules, 17 Yale J. on Reg. 327, 360 (2000). These arguments would presumably be moot if Congress or the Supreme Court explicitly overruled Chevron. However, judicial circumvention of FOIA’s commands demonstrates that deference would likely persist. A meaningful defense of deference must go beyond discussions of statutory interpretation.

Instead, deference conservation suggests that Chevron’s defenders should explore more explicitly Chevron’s constitutional dimensions, both in the institutional concerns underlying courts’ decisions and the rejoinders that argue Chevron is unconstitutional. These separation of powers arguments—as well as delegation arguments more generally 62 Bednar and Hickman offer an example of such a point, arguing that those who object to Chevron deference are really concerned with the degree to which Congress has delegated open-ended policymaking to agencies. See Bednar & Hickman, supra note 9, at 1453-56. —address concerns broader and more fundamental than those that typify the debates on Chevron, which are focused on the intricacies of statutory interpretation. Because these arguments are not limited to the statutory interpretation context, they possess the inherent breadth to reach questions of factual and procedural deference, permitting analysis that confronts deference conservation directly. Such a discussion would cut more to the heart of the issue than shadowboxing over statutory ambiguities.

Second, deference conservation challenges the doctrine’s detractors, suggesting that Chevron should not be treated as the sole (or even primary) cause behind any alleged loss of judicial independence. To be sure, this distinction might not affect the bottom line of formalist critiques that are based on the APA’s mandate that courts “shall decide all relevant questions of law” 63 E.g., Bamzai, supra note 59, at 985-90 (quoting 5 U.S.C. § 706 (2016)) (noting the tension between the APA’s text and Chevron). or the view that Article III requires courts to “say what the law is.” 64 E.g., Michigan v. EPA, 135 S. Ct. 2699, 2712 (2015) (Thomas, J., concurring) (quoting Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177 (1803)). Such arguments are concerned with statutory deference, which would technically be ameliorated by deference conservation’s shift of deference into non-statutory areas of litigation. However, for opposition that is rooted, either partially or fully, in reducing administrative discretion or promoting liberty, 65 See, e.g., City of Arlington v. FCC, 569 U.S. 290, 314-15 (2013) (Roberts, C.J., dissenting) (offering “the danger posed by the growing power of the administrative state” as a reason to limit Chevron). our conclusion suggests the focus is better placed on broader institutional dynamics rather than isolated deference regimes. Deference conservation ensures that the baselines in administrative discretion would be protected even with changes to Chevron.

Still, the previous two observations should be hedged in one important way: Deference conservation does not suggest that a change in review standards would be meaningless. 66 We do not address the debate on whether Congress can legislatively change precedents relating to the judiciary’s interpretive process. See Daniel B. Listwa, Comment, Uncovering the Codifier’s Canon: How Codification Informs Interpretation, 127 Yale. L.J. 464, 471-72 (2017) (arguing that legislated interpretive rules should be relevant to judges). The abrogation of Chevron would almost certainly reshape litigation involving administrative agencies. While mapping out the exact contours of such an alternative interpretive regime is both beyond the scope of this Essay and secondary to our point, some observations follow easily from our conclusion.

To begin, because deference conservation is admittedly not absolute, the mere act of abrogating Chevron might send a signal to lower courts, emboldening them to be less friendly to government litigants at the margins. 67 Cf. Daniel B. Listwa & Charles Seidell, Note, Penalties in Equity: Disgorgement After Kokesh v. SEC, 35 Yale. J. on Reg. 667, 699 (2018) (discussing Supreme Court signaling to another governmental body, the Securities and Exchange Commission, how they expect that body to utilize discretion). Similarly, agencies anticipating such a shift might themselves become less aggressive in their statutory interpretations. 68 Conceivably such an anticipatory shift in agencies’ legal positions could result in the percentage of wins remaining high, while still effecting a substantive change in the relative flexibility of agencies in interpreting the law.

Further, funneling deference into other areas of litigation—as we argue has occurred in the FOIA context—would have practical effects. As Nicholas Bednar and Kristin Hickman observe, although a formal repudiation of the precedent would likely lead most judges to “stop applying Chevron, at least in name,” functionally similar deference to agency statutory interpretations may continue, either under the umbrella of Skidmore or in some other form. 69 Bednar & Hickman, supra note 9, at 1457-59. This is a limited form of deference channeling—shifting from one statutory deference regime to another. However, as our examination of FOIA litigation has shown, deference can also be channeled out of statutory interpretation and into other parts of litigation, potentially with important consequences.

For example, if a court were to construe a legally-imbued issue as a question of fact and thereby defer to the agency’s finding, the ultimate precedential value would be much more limited than if the court were to uphold the agency’s decision as a matter of statutory interpretation. Alternatively, if courts were to counterbalance more searching review through a more exacting standing doctrine, the effect would be to stunt precedential development in administrative law altogether. Instead of merely deferring to agencies, courts could shield them from legal challenge entirely by erecting higher standing requirements. 70 Standing doctrine has long been viewed as flexible, with many arguing that standing determinations are influenced by judicial assessment of the underlying merits issues. See Nancy C. Staudt, Modeling Standing, 79 N.Y.U. L. Rev. 612, 614 n.14 (2004) (listing articles that have made this argument). This flexibility makes it susceptible to deference conservation. The result would be fewer decisions on the merits. These subtler effects of deference conservation may also help explain the evidence that Chevron does exert a meaningful force in lower court review of administrative law. Recognizing deference conservation, in other words, broadens the conversation about deference to agencies, presenting new paths forward for our understanding of the administrative state.


The absence of deference on matters of statutory interpretation does not imply the absence of the institutional pressures for judges to defer to agencies. FOIA litigation illustrates that where Chevron is unavailable, courts will construct other means to defer, such as through facts and procedures. Tinkering with only formal deference regimes—whether by narrowing or even overruling Chevron—misses the forest for the trees in the larger separation-of-powers debate. Only by grappling with the implications of deference conservation can our understanding of the administrative state confront head-on the challenges imposed by the nation’s constitutional structure.

John C. Brinkerhoff Jr. received his J.D. from Yale Law School in 2018. Daniel B. Listwa will receive his J.D. from Yale Law School in 2019. The Authors would like to thank Professors Bill Eskridge, Jerry Mashaw, and Chris Walker for their suggestions and insights; the editors of the Stanford Law Review Online, especially Raphael Ginsburg, Andra Lim, Jane Kessner, Liz Klein, and Emma Eastwood-Paticchio, for their thoughtful assistance; Leigh Terry and Sarah Kellner for their patience and support; and Charlie Seidell, Sam Adkisson, and Brian Lipshutz for providing excellent feedback on early drafts. All errors and omissions are our own.