In Remedial Restraint in Administrative Law, Professor Nicholas Bagley argues that we should replace administrative law’s ordinary remand rule with a more restrained, context-specific standard of first assessing whether the parties challenging the action were actually prejudiced by agency error. He bases this argument in part on his belief that the states challenging the Obama Administration’s sweeping execu¬tive actions on immigration suffered no harm from the Department of Homeland Security’s failure to engage in notice-and-comment rule¬making. That is because, he argues, the states received notice through leaks to the media and had a chance to comment through their public complaints on cable news programs and elsewhere.
This Response agrees that administrative law should focus more on remedies. But of all the serious challenges facing the regulatory state today, the lack of this particular type of remedial restraint is not among them. On the contrary, the current rule-based ordinary remand rule plays an important role in preserving a proper separation of powers, in ensuring agencies exercise their congressionally delegated discretion in a nonarbitrary manner, and in facilitating a richer court–agency dia¬logue that allows courts to have a systemic effect on the administrative process. The benefits of the ordinary remand rule exceed any benefits of a more restrained, standard-like remedial approach. And the current rule avoids the costs of courts assessing, for instance, whether regu¬lation by press leakage or regulation by Twitter is an acceptable, harm¬less substitute for notice-and-comment rulemaking.


Last year, administrative law scholars and practitioners engaged in a fierce and wide-ranging debate regarding the Obama Administration’s landmark executive actions on immigration. 1 See Nicholas Bagley, Remedial Restraint in Administrative Law, 117 Colum. L. Rev. 253, 253–54, 270–71 (2017) [hereinafter Bagley, Remedial Restraint] (discussing issues and debates). These actions, under the Deferred Action for Parents of Americans and Lawful Permanent Residents (DAPA) guidance memorandum and its predecessor the Deferred Action for Childhood Arrivals (DACA) guidance memorandum, made available a discretionary form of relief from removal—“deferred action”—to an estimated four million of the more than eleven million noncitizens who were unlawfully present in the United States at the time. 2 Texas v. United States, 809 F.3d 134, 148 (5th Cir. 2015), aff’d, 136 S. Ct. 2271 (2016). Twenty-six states challenged certain parts of DAPA. 3 Id. at 146. The Fifth Circuit ultimately upheld the district court’s nationwide preliminary injunction that enjoined those parts of DAPA pending resolution of the case on the merits. 4 Id.; see also id. at 146–50 (detailing DAPA and this judicial challenge). The Supreme Court granted review of the decision. 5 United States v. Texas, 136 S. Ct. 906 (2016). After Justice Scalia’s death, however, the Fifth Circuit’s decision was affirmed without opinion by an equally divided Court. 6 United States v. Texas, 136 S. Ct. at 2272, reh’g denied, 137 S. Ct. 285 (2016).

This legal challenge presented a number of discrete questions con­cerning core administrative law principles: Do the states have standing under Article III of the U.S. Constitution to challenge these immigration enforcement policies in federal court? 7 See, e.g., Lujan v. Defs. of Wildlife, 504 U.S. 555, 561 (1992) (reiterating the three main elements for constitutional standing: injury in fact, causation, and judicial redressability). The parties also briefed and argued the additional statutory zone-of-interest standing question under the Administrative Procedure Act. See Texas v. United States, 809 F.3d at 162–63. Are such policies “committed to agency discretion by law” such that they are not subject to judicial review under the Administrative Procedure Act (APA)? 8 See, e.g., Heckler v. Chaney, 470 U.S. 821, 828 (1985) (“[B]efore any review . . . [under the APA], a party must first clear the hurdle of § 701(a) [requiring that] judicial review ‘applies, according to the provisions thereof, except to the extent that—(1) statutes preclude judicial review; or (2) agency action is committed to agency discretion by law.’” (quoting 5 U.S.C. § 701 (1982))). Turning to the mer­its, are these executive actions arbitrary and cap­ricious under the APA as contrary to the agency’s governing statute, the Immigration and Nationality Act? When granting review, the Supreme Court directed the parties to also address “[w]hether the [DAPA] Guidance violates the Take Care Clause of the Constitution, Art. II, § 3.” 9 United States v. Texas, 136 S. Ct. at 906. Finally, and most im­portantly for the purposes of this Response, should the challenged DAPA provisions have been subject to the APA’s notice-and-comment rulemaking requirements? The district court and the Fifth Circuit both ruled in the states’ favor on this question, finding that the states had established a substantial likelihood of success on the merits of this proce­dural claim. 10 Texas v. United States, 809 F.3d at 146. The district court did not rule on the states’ substantive claims, but the Fifth Circuit also concluded that the states had established a substantial likelihood of success on the merits of their substantive challenges that the DAPA provi­sions were arbitrary and capricious under the APA as exceeding the agency’s statutory authority under the Immigration and Nationality Act. See id. at 178–86; see also Texas v. United States, 86 F. Supp. 3d 591, 677 (S.D. Tex.), aff’d, 809 F.3d 134 (5th Cir. 2015), aff’d, 136 S. Ct. 2271 (2016).

As debates swirled around all of these questions in early 2016, Professor Nicholas Bagley advanced a provocative and somewhat contrar­ian position on the Yale Journal on Regulation’s blog regarding the notice-and-comment question:

It’s true that DAPA itself didn’t pass through the formal notice-and-comment process . . . . It doesn’t follow, however, that the administration never gave notice of DAPA or afforded the pub­lic an opportunity to comment. Quite to the contrary. Of all the possible defects in the deferred action program, lack of public input was not one of them. 11 Nicholas Bagley, No Harm, No Foul in Texas v. United States, Yale J. on Reg.: Notice & Comment (Jan. 19, 2016) [hereinafter Bagley, No Harm, No Foul], []; accord Bagley, Remedial Restraint, supra note 1, at 270–71 (repeating and further developing this argument). This Response takes at face value Bagley’s framing of the agency notice and states’ response, recognizing others may see the facts differently. See, e.g., Josh Blackman, Gridlock, 130 Harv. L. Rev. 241, 279–303 (2016); Josh Blackman, Immigration Inside the Law, 55 Washburn L.J. 31, 32–36 (2016).

With respect to the APA’s notice requirement, Professor Bagley ass­erted that it was sufficient that the Obama Administration “leaked the proposal to the national media and held a Rose Garden press conference.” 12 Bagley, No Harm, No Foul, supra note 11. As for the APA’s public-comment requirement, it was suff­icient that state officials “objected vociferously” in public, with such ob­jections covered on CNN and Fox News. 13 Id. As for the requirement that the agency respond to significant comments, 14 See, e.g., Perez v. Mortg. Bankers Ass’n, 135 S. Ct. 1199, 1203 (2015) (explaining that under the APA “[a]n agency must consider and respond to significant comments received during the period for public comment”); Home Box Office, Inc. v. FCC, 567 F.2d 9, 35–36 (D.C. Cir. 1977) (“[O]pportunity to comment is meaningless unless the agency responds to significant points raised by the public.” (footnote omitted)). it was sufficient that “the [Justice Department’s] Office of Legal Counsel released a dense and closely reasoned opinion explaining why DAPA was a lawful exercise of the President’s enforce­ment discretion.” 15 Bagley, No Harm, No Foul, supra note 11. Even though the Obama Administration did not even attempt to engage in the procedural form­alities of notice-and-comment rulemaking, Bagley concluded that the states’ legal challenge on this procedural issue should fail because the APA instructs reviewing courts that “‘due account . . . be taken of the rule of prejudicial error.’” 16 Id. (quoting 5 U.S.C. § 706 (2012)); see also 5 U.S.C. § 706 (“In making the foregoing determinations, the court shall review the whole record or those parts of it cited by a party, and due account shall be taken of the rule of prejudicial error.”). In other words, “no harm, no foul.” 17 Bagley, No Harm, No Foul, supra note 11.

In Remedial Restraint in Administrative Law, Bagley builds on his blog post and develops a more comprehensive and nuanced argument in fa­vor of reinvigorating the harmless error standard for judicial review of federal agency action under the APA. 18 Bagley, Remedial Restraint, supra note 1, at 258–60. When confronted with an agency error, he argues, courts should discard the ordinary remand rule of generally vacating and remanding the agency action for further proceedings in favor of a more restrained, context-specific standard of first assessing whether the parties challenging the action were actually prejudiced by the error. 19 Id. at 263–65. The reason for this move, he further explains, is that the costs of the rule-based approach may outweigh the benefits. 20 Id. at 255. Bagley then chronicles how this standard should apply in a variety of ad­ministrative law contexts—from notice-and-comment failures to errors in agency adjudication. 21 See id. at 265–312.

As Bagley rightly observes, we lack a vigorous debate on questions of remedies in administrative law; this area deserves much more attention by scholars, practitioners, and courts (and perhaps even Congress). Remedial Restraint in Administrative Law is a welcome addition to our nascent conversation, 22 A number of scholars have focused recent attention on a variety of remedial issues in administrative law. On the ordinary remand rule and court–agency dialogue, see Emily Hammond Meazell, Deference and Dialogue in Administrative Law, 111 Colum. L. Rev. 1722 (2011); Christopher J. Walker, The Ordinary Remand Rule and the Judicial Toolbox for Agency Dialogue, 82 Geo. Wash. L. Rev. 1553 (2014) [hereinafter Walker, Ordinary Remand Rule]. On remand without vacatur in particular, see, e.g., Ronald M. Levin, “Vacation” at Sea: Judicial Remedies and Equitable Discretion in Administrative Law, 53 Duke L.J. 291, 384 (2003). On remedies in separation-of-powers structural chal­lenges, see Kent Barnett, To the Victor Goes the Toil—Remedies for Regulated Parties in Separation-of-Powers Litigation, 92 N.C. L. Rev. 481 (2014). And most recently on the nationwide injunction, see Samuel L. Bray, Multiple Chancellors: Reforming the National Injunction (UCLA Sch. of Law, Pub. Law & Legal Theory Research Paper Series Working Paper No. 16-54, 2016), (on file with the Columbia Law Review). one that should hopefully spark further conversa­tion. Similar to Professor Kathryn Watts’s review, 23 See Kathryn Watts, Rethinking Remedies, JOTWELL (Jan. 17, 2017), []. however, this is roughly the point at which my agreement with Remedial Restraint in Administrative Law ends. This Response does not attempt to present the comprehensive case against remedial restraint but instead focuses on two main points.

First, as discussed in Part I, one’s comfort with a more restrained, standard-like approach to the APA’s harmless error rule is likely corre­lated with one’s comfort with the constitutional and normative status of the modern administrative state. Although Bagley and I share similar foundational experiences in administrative law, 24 To name just a few similarities, we both started our careers clerking for federal judges who routinely review federal agency actions, and we worked together on the Justice Department’s Civil Appellate Staff, in which capacity we defended federal agencies in a variety of regulatory contexts. Moreover, we have both represented regulated entities in judicial challenges to federal agency action. we clearly do not share similar levels of comfort with the administrative state’s role in American governance, much less with the role of courts in supervising federal agency action. 25 Cf. Nicholas Bagley, The Puzzling Presumption of Reviewability, 127 Harv. L. Rev. 1285 (2014). For those of us who are less trusting of the federal bureaucracy, we are much less likely to find agency errors harmless—especially errors related to the structures and procedures that attempt to compensate for the regulatory state’s democratic deficits. The current rule-based approach of the ordinary remand rule better accounts for this distrust. And this rule-based approach is consistent with the text and structure of the APA’s appellate review model, especially as the model has evolved over the decades to address various separation-of-powers concerns.

Second, as discussed in Part II, Bagley’s call for remedial restraint is based in part on his intuitions about the costs and benefits of judicial re­view of agency action. 26 See, e.g., Bagley, Remedial Restraint, supra note 1, at 255 (“It’s not obvious, how­ever, that the benefits of a rule-bound approach outweigh its costs; indeed, there’s reason to fear that they don’t.”); id. at 257 (“Holding more agency errors harmless may not much affect agency incentives; to the extent it does, any uptick in agency misbehavior may not be sufficiently worrisome to warrant the reflexive invalidation of agency action.”). These intuitions seem to be shaped not only by an arresting level of trust in the modern administrative state, as noted above, but also by a striking amount of skepticism regarding judicial re­view’s ability to constrain arbitrary agency action. My own empirical work looking inside agency statutory interpretation and examining the court–agency dialogue on remand seems to suggest otherwise. 27 See, e.g., Christopher J. Walker, Inside Agency Statutory Interpretation, 67 Stan. L. Rev. 999 (2015) [hereinafter Walker, Inside Agency Interpretation]; Walker, Ordinary Remand Rule, supra note 22. For instance, it is reasonable to conclude that federal agencies regulate against the back­drop of judicial review. They are familiar with the judicial doctrines that govern administrative law, and they seem influenced by even varia­tion at the margins in judicial doctrine, such as the shift from Chevron to Skidmore deference. Moreover, the call for remedial restraint does not seem to appreciate some important costs of such a standard-like ap­proach, including a significant silencing of the court–agency dialogue and an accompanying impairment of the federal judiciary to play a more systemic role in protecting regulated individuals and entities from arbi­trary agency action. This systemic role is perhaps particularly important in the agency adjudication context, in which less sophisticated individu­als navigate the administrative process without representation and often lack the wherewithal to seek further review of erroneous agency actions.

This Response concludes with a warning about introducing such dra­matic change to administrative law without considering its effects on the rest of the modern administrative state. The ordinary remand rule does not operate in isolation, but it is just one part of a calibrated system for judicial review of federal agency action that attempts to strike the proper balance between facilitating the exercise of congressionally delegated agency discretion and reinforcing agency procedures and structures that ensure such agency discretion is not exercised in an arbitrary or capri­cious manner.

I. Bureaucracy and Distrust

To understand the proper role of the APA’s harmless error standard in our current approach to judicial review of federal agency action, it is helpful to frame administrative law’s larger, decades-long struggle to bal­ance agency discretion with judicial review to constrain arbitrary agency action. The Administrative Procedure Act of 1946 attempted to strike such a balance and in so doing embraced an appellate model of judicial review. 28 Bagley, Remedial Restraint, supra note 1, at 258–60. See generally Thomas W. Merrill, Article III, Agency Adjudication, and the Origins of the Appellate Review Model of Administrative Law, 111 Colum. L. Rev. 939 (2011). Under this model, courts review agency actions similarly to how appellate courts review trial court decisions in civil litigation. 29 See Walker, Ordinary Remand Rule, supra note 22, at 1554–56 (discussing administrative law’s appellate review model in greater detail). The appel­late review model in the civil litigation context is based on the record from the prior proceeding, and the reviewing court does not engage in inde­pendent fact-finding. Likewise, the standard of review reflects the com­parative expertise of the various institutions, with more or less def­erential review depending on whether the issue is more factual or le­gal, respec­tively. It is thus not surprising that the APA incorporates a judi­cial review principle that “due account shall be taken of the rule of prejudicial error.” 30 5 U.S.C. § 706 (2012). As the Attorney General’s 1947 manual on the APA asserts, this prejudicial error rule means that “errors which have no substantial bearing on the ultimate rights of the parties will be disregarded.” U.S. Dep’t of Justice, Attorney General’s Manual on the Administrative Procedure Act 110 (1947). A similar “harmless error” rule was included in the original 1937 Federal Rules of Civil Procedure in that errors at trial are not grounds for relief “unless refusal to take such action appears to the court inconsistent with substantial justice.” 31 Fed. R. Civ. P. 61 (1937) (repealed 2007). In 2007, Rule 61 was stylistically modi­fied to instruct that “the court must disregard all errors and defects that do not affect any party’s substantial rights.” Fed. R. Civ. P. 61 (as amended Apr. 30, 2007, effective Dec. 1, 2007).

Unlike the intrabranch relationship between appellate and trial courts, however, the relationship between courts and agencies implicates separation-of-powers concerns among all three branches of government. On the Article I front, “[t]he presumption that the reviewing court has superior competence to answer questions of law is rebutted by the fact that Congress often delegates law-elaboration authority first and foremost to the agency.” 32 Walker, Ordinary Remand Rule, supra note 22, at 1555 (citing, inter alia, Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 983 (2005)). On the Article II front, most federal agencies are squarely located within the Executive Branch, with so-called independent agencies more geographically ambiguous. As such, those federal agencies have certain law-execution discretion—separate from the congressionally delegated authority—under Article II. 33 See Christopher J. Walker, Avoiding Normative Canons in the Review of Administrative Interpretations of Law: A Brand X Doctrine of Constitutional Avoidance, 64 Admin. L. Rev. 139, 173–82 (2012) (exploring these dual separation-of-powers concerns in the context of Chevron deference and modern constitutional avoidance). To be sure, some scholars have advanced a more robust Article II theory. See, e.g., Jack Goldsmith & John F. Manning, The President’s Completion Power, 115 Yale L.J. 2280, 2282 (2006) (arguing that the President has constitutional authority to fill in the holes of any statutory scheme, subject to congressional override). Administrative law’s appellate model of judicial review has thus evolved to incorporate a number of agency-deference doctrines that reflect these separation-of-powers values. Chevron deference comes immediately to mind. 34 Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 842–44 (1984) (instructing courts to defer to reasonable agency interpretations of ambiguous provisions in statutes the agency administers); see also Merrill, supra note 28, at 999 (noting “in response to the deregulation movement, the model was sufficiently elastic to permit a further modification in the appropriate division of authority in resolving questions of law, most prominently with the Chevron decision in 1984”). As Professor Thomas Merrill has explained, one reason why administrative law’s approach to judicial review has persevered is that “[t]he appellate review model has . . . proven to be flexible at the macro level.” 35 Merrill, supra note 28, at 998.

A contrasting set of separation-of-powers concerns is also at play. The eras of federal lawmaking by common law and by statute have ceded to an era of federal lawmaking by regulation. 36 See, e.g., Dakota S. Rudesill, Christopher J. Walker & Daniel P. Tokaji, A Program in Legislation, 65 J. Legal Educ. 70, 72 (2015) (“In 2013 alone, federal agencies filled nearly 80,000 pages of the Federal Register with adopted rules, proposed rules, and notices. By contrast, the 113th Congress [filled] . . . a total of 1,750 pages in the Statutes at Large.” (footnote omitted)). As Chief Justice Roberts remarked, “The administrative state ‘wields vast power and touches al­most every aspect of daily life’” such that “[t]he Framers could hardly have envisioned today’s ‘vast and varied federal bureaucracy’ and the authority administrative agencies now hold over our economic, social, and political activities.” 37 City of Arlington v. FCC, 133 S. Ct. 1863, 1878 (2013) (Roberts, C.J., dissenting) (quoting Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 499 (2010)). To be sure, the Supreme Court has not over­ruled the nondelegation doctrine, which instructs that Congress cannot delegate legislative power to federal agencies but, instead, must delegate lawmaking power only based on an “intelligible principle.” 38 J.W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 409 (1928). Yet the doc­trine currently imposes no real constraints; the Court has yet to find any delegation to be unconstitutional since the APA’s enactment. 39 See Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 474–75 (2001) (discussing nondelegation doctrine precedent); see also Dep’t of Transp. v. Ass’n of Am. R.Rs., 135 S. Ct. 1225, 1246 (2015) (Thomas, J., concurring in the judgment) (“We have held that the Constitution categorically forbids Congress to delegate its legislative power to any other body, but it has become increasingly clear to me that the test we have applied to distin­guish legislative from executive power largely abdicates our duty to enforce that prohibi­tion.” (citing Whitman, 531 U.S. at 472)). When one combines the Court’s lack of an effective doctrine to patrol excessive congressional delegations of lawmaking authority to federal agencies with “the rise and rise of the administrative state,” 40 Gary Lawson, The Rise and Rise of the Administrative State, 107 Harv. L. Rev. 1231 (1994). it should come as no surprise that many have a deep distrust of the American bureaucracy and a fear of arbitrary exercises of agency lawmaking discretion. As Professor Aaron Nielson and I have argued, a central principle of administrative law is (or at least should be) that discretion can be dangerous. 41 Aaron L. Nielson & Christopher J. Walker, Strategic Immunity, 66 Emory L.J. 55, 57 (2016) (“The danger is that although discretion can be and, indeed, usually is used for the public’s benefit, it can also serve self-interested ends.”).

The appellate review model for administrative law has evolved to ad­dress some of these concerns as well. For instance, in the 1970s the model adapted to embrace “hard look” review under the APA’s arbitrary-and-capricious standard to discourage arbitrary agency action. Hard look review, as the Supreme Court has instructed, requires reasoned deci­sionmaking in that:

[A]n agency rule would be arbitrary and capricious if the agency has relied on factors which Congress has not intended it to consider, entirely failed to consider an important aspect of the problem, offered an explanation for its decision that runs counter to the evidence before the agency, or is so implausible that it could not be ascribed to a difference in view or the prod­uct of agency expertise. 42 Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins., 463 U.S. 29, 43 (1983).

As Professor Merrill has observed, this hard-look “approach required no fundamental alteration in the appellate review model. Courts simply lay­ered a more aggressive monitoring of the quality of agency reasoning on top of the standard review of the factual record from the original model.” 43 Merrill, supra note 28, at 999.

The administrative law principles established by the Supreme Court in the 1940s in the Chenery decisions similarly reformed the appellate re­view model in light of separation-of-powers concerns. First, the Court departed from “the settled rule” in the civil litigation context that a trial court’s decision “must be affirmed if the result is correct ‘although the lower court relied upon a wrong ground or gave a wrong reason.’” 44 SEC v. Chenery Corp. (Chenery I), 318 U.S. 80, 88 (1943) (quoting Helvering v. Gowran, 302 U.S. 238, 245 (1937)). In­stead, “[t]he grounds upon which an administrative order must be judged are those upon which the record discloses that its action was based.” 45 Id. at 87. Or, as the Chenery II Court rearticulated the rule shortly after the APA’s enactment, “A reviewing court, in dealing with a determina­tion or judgment which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds in­voked by the agency.” 46 SEC v. Chenery Corp. (Chenery II), 332 U.S. 194, 196 (1947).

Separation-of-powers concerns, as Professor Kevin Stack has convinc­ingly argued, motivate this first Chenery principle. In particular, this principle “promotes core values of the nondelegation doctrine in ways that supplement the enforcement of the intelligible principle require­ment.” 47 Kevin M. Stack, The Constitutional Foundations of Chenery, 116 Yale L.J. 952, 958 (2007). It does so by increasing political accountability of the agency’s action, by helping to prevent arbitrary and capricious agency action, and by “provid[ing] assurance that accountable agency decision-makers, not merely courts and agency lawyers, have embraced the grounds for the agency’s actions, and that the agency decision-makers have exercised their judgment on the issue in the first instance.” 48 Id. at 958–59. Professor Jon Michaels’s important work on “administrative separation of powers”—the call to view “administrative law through the lens of a secondary, sub­constitutional separation of powers that triangulates administrative power among politically appointed agency leaders, an independent civil service, and a vibrant civil society” 49 Jon D. Michaels, An Enduring, Evolving Separation of Powers, 115 Colum. L. Rev. 515, 520 (2015); cf. Miriam Seifter, Complementary Separation of Powers, 91 N.Y.U. L. Rev. Online 186, 199 (2016), [] (“Civil society, acting alone, cannot replicate Congress along relevant dimensions, but it can complement Congress and further ideals of inclusion and public engagement. It may be the energetic involvement of all of the players . . . that does the work of checking and balancing in the modern administrative state.”). For an in-depth discussion of Professor Michaels’s work on administrative separation of powers, see Jon D. Michaels, Separation of Powers All the Way Forward: The Theory and Practice of Constitutional, Administrative, and Privatized Government (forthcoming 2017) [hereinafter Michaels, Separation of Powers] (on file with the Columbia Law Review). —further underscores the need for this Chenery principle and related doctrines. As Professor Michaels argues, when reviewing agency action, courts should “consider whether a fair, inclusive administrative process has been short-circuited, disabled, or unduly inter­fered with in a manner that precludes or limits meaningful participation by all three administrative rivals.” 50 Michaels, Separation of Powers, supra note 49 (manuscript at 230). It is worth noting that, like the heading of this Part, Professor Michaels analogizes administrative separation of powers to John Hart Ely’s Democracy and Distrust: “When thinking about republican government in an administrative rather than legislative arena, reinforcing representative democracy becomes reinforcing rivalrous administra­tion.” Id. (manuscript at 229–30); see also id. (manuscript at 229) (“We can piggyback on the powerful but not uncontroversial Carolene Products/Ely approach and retool it for use in the administrative arena.” (citing United States v. Carolene Prods. Co., 304 U.S. 144, 152 n.4 (1938); John Hart Ely, Democracy and Distrust 102–03 (1980))).

The second Chenery principle—the ordinary remand rule—is yet an­other modification to the appellate review model. As the Chenery I Court announced, if a court concludes that an agency’s decision is erroneous, the general rule is to remand to the agency to consider the issue anew—as opposed to the court deciding the issue itself. 51 SEC v. Chenery Corp. (Chenery I), 318 U.S. 80, 95 (1943) (remanding the matter to the agency because the “administrative order cannot be upheld unless the grounds upon which the agency acted in exercising its powers were those upon which its action can be sustained”); accord Chenery II, 332 U.S. at 196. “[T]he guiding princi­ple,” as the Court reiterated several years after the Chenery cases, “is that the func­tion of the reviewing court ends when an error of law is laid bare. At that point the matter once more goes to the [agency] for reconsideration.” 52 Fed. Power Comm’n v. Idaho Power Co., 344 U.S. 17, 20 (1952). The Court has rearticulated this ordinary remand rule over the years in a variety of administrative law contexts. 53 See, e.g., Negusie v. Holder, 555 U.S. 511, 517 (2009) (“When the BIA has not spoken on ‘a matter that statutes place primarily in agency hands,’ our ordinary rule is to remand to ‘giv[e] the BIA the opportunity to address the matter in the first instance in light of its own expertise.’” (quoting INS v. Orlando Ventura, 537 U.S. 12, 16–17 (2002) (per curiam))). As I have explored elsewhere, the ordinary remand rule applies not only to ques­tions of fact—as is the case in the civil litigation context—but also to ques­tions of the application of law to fact, policy judgments, and even certain questions of law. 54 See Walker, Ordinary Remand Rule, supra note 22, at 1561–79 (tracing the evolu­tion of the ordinary remand rule). It is the ordinary rule, “except in rare circum­stances.” 55 Orlando Ventura, 537 U.S. at 16 (“‘[T]he proper course, except in rare circum­stances, is to remand to the agency for additional investigation or explanation.’” (quoting Fla. Power & Light Co. v. Lorion, 470 U.S. 729, 744 (1985))). Those rare exceptions, I have argued, should be limited to when there are “inadequate or errone­ous subsidiary findings” 56 Henry J. Friendly, Chenery Revisited: Reflections on Reversal and Remand of Administrative Orders, 1969 Duke L.J. 199, 224. that the court is fairly confident do not affect the outcome, when the agency lacks the authority to decide the issue, or when Congress has provided for a trial de novo of the issue. 57 The first two exceptions are discussed in Walker, Ordinary Remand Rule, supra note 22, at 1566–68, 1620, and the third in Stephanie Hoffer & Christopher J. Walker, The Death of Tax Court Exceptionalism:
“[T]here is another exception of particular relevance here: when APA § 706(2)(F) applies and ‘the facts are subject to trial de novo by the reviewing court.’ . . . Logically, then, if the reviewing court is empowered to conduct a trial de novo, the court is not required to remand (though it retains discretion to do so) because de novo review allows the court to take the unusual step of substituting its judgment for that of the agency.”
99 Minn. L. Rev. 221, 267 (2014) [hereinafter Hoffer & Walker, Death of Tax Court Exceptionalism] (quoting 5 U.S.C. § 706(2)(F) (2012)) (footnote omitted).

These two Chenery principles, taken together, provide important con­text for the scope of the APA’s harmless error charge that “due account shall be taken of the rule of prejudicial error.” 58 5 U.S.C. § 706. In other words, it is not at all clear, especially in light of these separation-of-powers concerns, that, “[a]s a matter of what the APA says, . . . harmless error review should be as central to administrative law as it is to conventional litigation.” 59 Bagley, Remedial Restraint, supra note 1, at 259. To be sure, Bagley notes a number of these separation-of-powers concerns, see id. at 260–62, and how they have limited the APA’s harmless error rule. He argues that these concerns have resulted in “the rule of prejudicial error ha[ving] gone missing.” Id. at 262. This Part, however, explains that the harmless error standard has not gone missing; it just plays a much more limited role in light of these constitutional concerns than it might in the context of civil litigation.

Bagley is certainly correct, though, that the APA’s harmless error rule has developed as a “rule-like approach to administrative remedies” as opposed to a more context-based standard that includes a robust investigation into the prejudice caused by the agency error. 60 Id. at 257. In light of the Chenery principles and related adaptations to administrative law’s appellate review model, the inquiry into whether an agency error is harmless has largely been limited to an agency’s “inadequate or erron­eous subsidiary find­ings,” 61 Friendly, supra note 56, at 224. and even then only when the reviewing court is confident that those minor errors would not affect the outcome. 62 See Walker, Ordinary Remand Rule, supra note 22, at 1566–68. Remand is, as the Supreme Court has often repeated, the general rule, “except in rare circumstances.” 63 INS v. Orlando Ventura, 537 U.S. 12, 16 (2002) (per curiam) (quoting Fla. Power & Light Co. v. Lorion, 470 U.S. 729, 744 (1985)); accord Negusie v. Holder, 555 U.S. 511, 523 (2009); Gonzales v. Thomas, 547 U.S. 183, 186 (2006) (per curiam).

Such an approach to harmless error, although perhaps less searching than in the civil litigation context, is fully consistent with the text and structure of the APA. In light of the separation-of-powers con­cerns that motivate judicial review of agency action, it turns out that most agency errors will be deemed harmful, especially errors with respect to agency processes and structures. Some errors may well rise to the level of structural error—to borrow from the criminal law context—such that no showing of prejudice is required. 64 For examples of structural errors in the criminal law context, see Sullivan v. Louisiana, 508 U.S. 275, 281 (1993) (“Denial of the right to a jury verdict of guilt beyond a reasonable doubt is certainly a [structural] error [when harmless error analysis does not apply], the jury guarantee being a ‘basic protectio[n]’ whose precise effects are unmeasurable, but without which a criminal trial cannot reliably serve its function.” (cita­tion omitted)); Waller v. Georgia, 467 U.S. 39, 49 (1984) (holding “the defendant should not be required to prove specific prejudice in order to obtain relief for a violation of the public-trial guarantee”); Tumey v. Ohio, 273 U.S. 510, 535 (1927) (“No matter what the evidence was against him, he had the right to have an impartial judge.”); United States v. Kimbrel, 532 F.3d 461, 469 (6th Cir. 2008) (“‘Structural’ errors [such as the Batson error here], by contrast, which affect the ‘entire conduct of the trial from beginning to end,’ are not subject to harmless-error review.”(quoting Arizona v. Fulminante, 499 U.S. 279, 309 (1991))). For other errors not considered to be structural enough to exempt them entirely from the harmless error in­quiry, concerns about due process, bureaucratic legitimacy, or adminis­trative separation of powers may persuade courts to adopt a “strong” pre­sumption of prejudice. 65 Kristin E. Hickman & Mark Thomson, Open Minds and Harmless Errors: Judicial Review of Postpromulgation Notice and Comment, 101 Cornell L. Rev. 261, 311 (2016) (arguing “a strong presumption against the validity of postpromulgation notice and com­ment best respects the balance between an express statutory command for prepromulga­tion notice and comment and a particularized harmless error rule”).

Many of the agency errors Bagley highlights as meriting a more searching prejudice inquiry can be considered structural errors—or at least errors affecting substantial procedural rights so as to require at least a strong presumption of prejudice. On the rulemaking front, if the agency is required to utilize notice-and-comment rulemaking but fails to do so (such as was at issue in United States v. Texas), that goes to the heart of the administrative process. 66 See Bagley, Remedial Restraint, supra note 1, at 269–74 (discussing examples of errors in rulemaking). It is not about an error at trial, to borrow from the civil and criminal contexts, but a lack of a trial entirely. At the very least “[t]he entire conduct of the [agency proceeding] from be­ginning to end is obviously affected” by the lack of notice-and-comment rulemaking. 67 Fulminante, 499 U.S. at 309–10 (1991) (explaining judicial bias and lack of representation for a criminal defendant are structural errors). Judicial inquiry into whether the parties were harmed when they received notice by press leakage and were able to voice their con­cerns on cable television news shows has no place in the context of struc­tural error. 68 It is important to underscore that individuals challenging agency action, as dis­cussed in the Introduction, must still establish that they were injured in fact by the agency’s actions for the purposes of standing under Article III and that their interests are within the zone of interests for purposes of judicial review under the APA.

This seems to be Professor Watts’s main critique of the argument:

This kind of reasoning, in my mind, threatens to eviscerate Section 553 of the APA, allowing informal dialogue between an agency and interested parties to substitute for Section 553’s carefully defined procedures. Effectively, it would allow the LA Times, Fox News, CNN and other media channels to displace the Federal Register as the place where interested parties must look to find—and to learn how and when to comment on—proposed agency rules. That is not consistent with the APA. Nor would it help to bolster the public’s perception of the legitimacy of agency decisionmaking. 69 Watts, supra note 23.

There is both a structural and a public legitimacy point here. As Professor Watts further explains, “Procedural fastidiousness, in my mind, plays a very important role in bolstering public perceptions of agency legitimacy and attending to agencies’ democracy deficit.” 70 Id. See generally Louis L. Jaffe, Judicial Control of Administrative Action 320 (1965) (“The availability of judicial review is the necessary condition, psychologically if not logically, of a system of administrative power which purports to be legitimate, if not valid.”). If Bagley’s context-specific standard were adopted, a reviewing court may well have to assess in the next case whether rulemaking by Twitter is a proper, harmless substitute for notice-and-comment rulemaking.

Bagley’s other rulemaking examples do not fare much better. He argues that a context-specific harmlessness standard should apply to “logical outgrowth” challenges, in which a regulated entity challenges a final rule for lack of fair notice because the final rule is not a logical out­growth of the proposed rule. 71 See Bagley, Remedial Restraint, supra note 1, at 274–83 (discussing examples). Again, notice is a critical procedural right embedded in the APA’s notice-and-comment rulemaking process. It should be structural error if the public is not given fair notice of the final rule, or at the very least this procedural right should be considered so fundamental so as to impose a strong presumption of prejudice. This is particularly important because the substantive standard for a logical out­growth challenge already encompasses a prejudice inquiry: whether the public had fair notice. 72 See, e.g., Long Island Care at Home, Ltd. v. Coke, 551 U.S. 158, 174 (2007) (“The object [of the logical outgrowth rule], in short, is one of fair notice.”).

The same is true of challenges to legislative rules that are published without a notice-and-comment period and then subjected to a postprom­ulgation comment period. 73 See Bagley, Remedial Restraint, supra note 1, at 283–92 (discussing examples). Public comment on a proposed rule is a core procedural right, subject to limited statutory “good cause” exceptions. 74 Administrative Procedure Act 5 U.S.C. § 553(b) (2012) (“Except when notice or hearing is required [], this subsection does not apply . . . when the agency for good cause finds (and incorporates the finding and a brief statement of reasons therefor in the rules issued) that notice and public procedure thereon are impracticable, unnecessary, or contrary to the public interest.”). In their seminal treatment of this issue, Mark Thomson and Professor Kristin Hickman argue that “[t]he potential consequences of ignoring notice and comment, combined with the agency’s ability to prevent such a defect in the first place, makes it fair to put the burden of proving harmlessness on the agency.” 75 Hickman & Thomson, supra note 65, at 313. There are compelling arguments that this widespread postpromulgation notice practice, absent a proper good cause justification, is a structural error such that prejudice should not be questioned.

Bagley is on firmer ground in arguing that some inadequacies in agency explanations are harmless. After all, the Supreme Court’s articula­tion of hard-look review expressly recognizes that not all inadequate explanations merit remand; instead, courts should “uphold a decision of less than ideal clarity if the agency’s path may reasonably be discerned.” 76 Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins., 463 U.S. 29, 43 (1983) (internal quotation marks omitted) (quoting Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 286 (1974)). This approach is consistent with Judge Henry Friendly’s exception to the ordi­nary remand rule for “inadequate or erroneous subsidiary findings.” 77 Friendly, supra note 56, at 224; accord Walker, Ordinary Remand Rule, supra note 22, at 1566–68. Of course, Bagley is not just talking about agency explanations that are “of less than ideal clarity.” Instead, his main explanatory error is when an agency erroneously declares that a statute is unambiguous. 78 See Bagley, Remedial Restraint, supra note 1, at 296–301. Under D.C. Circuit precedent in Prill v. NLRB, the court will remand such errors for the agency to consider anew as opposed to the court deciding the interpretation issue itself in the agency’s favor. 79 Prill v. NLRB, 755 F.2d 941 (D.C. Cir. 1985). The Supreme Court seems to have endorsed a variant of this doctrine in Negusie v. Holder, in which it remanded an inter­pretation question to the agency because the agency had erroneously believed prior judicial precedent foreclosed any statutory ambiguity. 555 U.S. 511, 514 (2009); see also Walker, Ordinary Remand Rule, supra note 22, at 1578–79 (discussing Negusie and its separation-of-powers foundation).

There are certainly agency costs to the Prill remand doctrine. As Professors Daniel Hemel and Aaron Nielson have recently explained, however, there are also a number of important benefits that arguably out­weigh these costs, including imposing incentives on agencies not to engage in strategic behavior. 80 Daniel J. Hemel & Aaron L. Nielson, Chevron Step One-and-a-Half, 84 U. Chi. L. Rev. (forthcoming 2017) (manuscript at 38–49), (on file with the Columbia Law Review) (developing their case for the benefits of this doctrine). Professor Michaels’s theory of admin­istrative separation of powers, discussed above, underscores another set of costs and benefits surrounding the Prill remand doctrine: Such remand ensures that the agency’s statutory interpretation is not just the product of the agency litigators and political appointees but that it results from wide­spread participation of all three administrative rivals—political appoin­tees, civil service, and civil society. 81 See supra notes 49–50 and accompanying text.

In sum, Bagley is no doubt correct that the current rule-based approach to remand in administrative law leads to some additional costs, including a heightened risk of false positives—cases in which relatively harmless agency errors are remanded to the agency, resulting in addi­tional delay or even the agency’s abandonment of the regulatory effort. He has certainly flagged a few of the most egregious examples of agency errors that seemed to impose little, if any, harm on the regulated entities seeking judicial review. But for those concerned about bureaucracy and distrust, we are much more troubled about false negatives—cases in which there are harmful agency errors that nevertheless are ignored because the court erroneously finds no prejudice. Bagley’s more restrained, context-based standard would certainly increase the rate of false negatives. That there is even a debate about the harmfulness of rule­making by press leakage should prove this point. In a federal system now dominated by lawmaking by regulation and by courts charged with pro­tecting against arbitrary exercises of agency discretion, calls for more remedial restraint in administrative law focus on eliminating the wrong type of errors.

II. The Costs of Remedial Restraint

For those who are not convinced that the modern administrative state poses sufficient dangers to retain the current rule-based approach to judicial remedies for agency errors, it is worth considering some of the other costs of shifting to a more restrained, standard-like approach. For ease of organization, this Part divides such costs between rulemaking and adjudication, though the costs undoubtedly overlap both categories of agency action.

A. For Agency Rulemaking

As detailed in Part IV of Remedial Restraint in Administrative Law, Bagley’s call for remedial restraint is based, at least in part, on his skepti­cism about the effect of judicial review on agency behavior. Simply put, he believes that, on the margins, “holding more agency errors harmless might not much affect agency incentives.” 82 Bagley, Remedial Restraint, supra note 1, at 313–14.

Having interviewed and surveyed hundreds of agency officials over the last few years, I do not share Bagley’s deep skepticism about the effects of changing the harmless error rule on agency incentives, espe­cially considering the role of thousands of civil service lawyers who work throughout the federal regulatory state. For instance, in 2013 I adminis­tered a 195-question survey of federal agency rule drafters that covered a variety of topics related to agency statutory interpretation and rule draft­ing. 83 The full findings are reported in Walker, Inside Agency Interpretation, supra note 27. See also Christopher J. Walker, Chevron Inside the Regulatory State: An Empirical Assessment, 83 Fordham L. Rev. 703 (2014) [hereinafter Walker, Chevron Inside the Regulatory State] (further exploring findings related to administrative law’s deference doctrines); Christopher J. Walker, Inside Regulatory Interpretation: A Research Note, 114 Mich. L. Rev. First Impressions 61 (2015) (exploring findings related to regulatory inter­pretation). The survey was modeled on Professors Lisa Bressman and Abbe Gluck’s pioneering study on congressional drafting. See Lisa Schultz Bressman & Abbe R. Gluck, Statutory Interpretation from the Inside—An Empirical Study of Congressional Drafting, Delegation, and the Canons: Part II, 66 Stan. L. Rev. 725 (2014); Abbe R. Gluck & Lisa Schultz Bressman, Statutory Interpretation from the Inside—An Empirical Study of Congressional Drafting, Delegation, and the Canons: Part I, 65 Stan. L. Rev. 901 (2013). The survey was administered at seven executive departments (Agriculture, Commerce, Energy, Homeland Security, Health and Human Services, Housing and Urban Development, and Transportation) and two independent agencies (Federal Communications Commission and Federal Reserve). In total, 128 agency rule drafters replied, resulting in a thirty-one percent response rate. Although confidentiality concerns imposed significant methodological limitations on the survey—including anonym­ity as to the individual respondent and the respondent’s re­spective agency—the survey responses shed considerable light on how judicial review affects agency behavior in the rulemaking context. 84 For more on the study methodology and its accompanying limitations, see Walker, Inside Agency Interpretation, supra note 27, at 1013–18.

First and foremost, nearly nine in ten rule drafters surveyed strongly agreed (46%) or agreed (41%)—and another 11% somewhat agreed—that “[w]hen drafting rules and interpreting statutes, agency drafters such as yourself think about subsequent judicial review.” 85 Walker, Chevron Inside the Regulatory State, supra note 83, at 722 (quoting the survey question). Moreover, these rule drafters know their judicial deference doctrines. Among all twenty-two interpretive tools tested in the survey, Chevron deference was the most known by name (94%) and most reported as playing a role in rule drafting (90%). 86 Walker, Inside Agency Interpretation, supra note 27, at 1061–63. The next most known tools were: the ordinary meaning canon (92%), Skidmore deference (81%), and the presumption against preemption of state law (78%). 87 Id. at 1019 fig.1. The tools most reported after Chevron deference as playing a role in rule drafting were: the whole act rule (presumption of consistent usage throughout statute) (89%), the ordinary meaning canon (87%), the Mead doctrine (by concept) (80%), noscitur a sociis (associated-words canon) (79%), and legislative history (76%). 88 Id. at 1020 fig.2.

When most of the rule drafters surveyed indicated they think about judicial review and “use” administrative law doctrines when regulating, they seem to suggest that judicial review affects agency behavior. For instance, the rule drafters surveyed understood quite well how different deference doctrines affect agency win rates in courts: About four in five strongly agreed (38%) or agreed (45%)—and another 17% somewhat agreed—that “[i]f Chevron deference (as opposed to Skidmore deference or no deference) applies to an agency’s interpretation of an ambiguous statute it administers, the agency is more likely to prevail in court.” 89 Walker, Chevron Inside the Regulatory State, supra note 83, at 723 (quoting the survey question). Recent empirical work in the circuit courts seems to confirm the agency rule drafters’ impressions that the deference doctrines indeed affect agency win rates. See Kent Barnett & Christopher J. Walker, Chevron in the Circuit Courts, 115 Mich. L. Rev. (forthcoming 2017) (manuscript at 6), (on file with the Columbia Law Review) (finding, inter alia, nearly a twenty-five percentage point differ­ence in agency win rates when the circuit courts applied Chevron deference than when they refused to apply it). Indeed, two in five rule drafters surveyed agreed (31%–33%) or strongly agreed (7%–10%)—and another two in five somewhat agreed (40%–45%)—that a federal agency is more aggressive in its interpretive efforts if it is confident that Chevron deference (as opposed to Skidmore defer­ence or de novo review) applies. 90 Walker, Chevron Inside the Regulatory State, supra note 83, at 722–24 & fig.3. These findings are presented in percentage ranges because the survey explored this issue with two questions that were worded in slightly different ways. See id. at 723–24.

These findings, especially considering the study’s methodological limitations, only scratch the surface of the complex relationship between courts and agencies. Much more work needs to be done. But the findings are consistent with my own qualitative experience in interviewing agency officials for this study and others. 91 These findings are also consistent with Connor Raso’s empirical work on how agencies avoid complying with statutory rulemaking procedures. In reviewing judicial deci­sions regarding agency rulemakings from 1995 through 2012, Raso finds that agencies are much more likely to avoid statutory rulemaking requirements—more than 90% of the time—when there is a low threat of a lawsuit. Connor Raso, Agency Avoidance of Rulemaking Procedures, 67 Admin. L. Rev. 65, 68–69 (2015). Raso thus concludes that judicial review is a necessary though “inconsistent and highly imperfect enforcement mechanism.” Id. at 127. Agency officials, especially agency law­yers, pay close attention to judicial developments in administrative law and think about judicial doctrines when regulating. Since they are attentive to shifts at the margins in judicial review doctrines, such as the shift from the rule-based Chevron doctrine to the standard-like Skidmore doctrine, I suspect that federal agencies would similarly respond to a shift from a rule-based remand rule to a context-specific standard. To be sure, not every judicial message is received clearly, nor is every judicial com­mand followed perfectly. Agencies no doubt engage in strategic behavior to preserve their regulatory actions from judicial override. 92 Indeed, agency lawyers have developed a playbook for successfully navigating judicial review of agency action. See Christopher J. Walker, How to Win the Deference Lottery, 91 Tex. L. Rev. 73, 77–87 (2012) (detailing the playbook). But deep skepticism about the ability of judicial review to check arbitrary agency behavior seems overstated.

B. For Agency Adjudication

Another cost of remedial restraint—and the accompanying decline in judicial remands based on agency errors—concerns the court–agency dialogue. As Professor Emily Hammond has explored in the rulemaking context, and I have explored in the agency adjudication context, the ordinary remand rule facilitates an important dialogue between courts and agencies. 93 See Meazell, supra note 22, at 1739–72; Walker, Ordinary Remand Rule, supra note 22, at 1585–99. This dialogue brings judicial expertise, aided by input from the parties to the lawsuit and other friends of the court, to the administrative process on remand. Indeed, courts have developed a num­ber of tools to enhance their dialogue with federal agencies on remand without upsetting the separation-of-powers balance between courts and agencies. For instance, in cases in which courts are skeptical of the agen­cy getting it right on remand, concerned about undue delay, or wor­ried about the petitioner getting lost on remand, some circuits require the agency to provide notice of its final determination, retain panel juris­diction over the matter, or set deadlines for an agency response to the remand. Others suggest that administrative judges be replaced on remand, certify issues for decision on remand, or set forth hypothetical answers in dicta or concurring opinions. 94 See Walker, Ordinary Remand Rule, supra note 22, at 1590–99, 1607–14 & tbl.2. Similar tools can be used to mitigate the delay or other agency costs on remand that seem to motivate Bagley’s call for remedial restraint.

This court–agency dialogue can assist the agency in developing bet­ter procedures and arriving at a better substantive outcome in the case remanded. But the dialogue serves an even more important objective, especially in the agency adjudication context, of allowing the federal judiciary to play a more systemic role in protecting regulated individuals and entities from arbitrary agency action. Agency adjudication in a vari­ety of regulatory contexts has been plagued with inconsistency in that similarly situated individuals are not treated the same. Caseloads can be crushing on agency adjudicators, training and competency can vary greatly across adjudicators, internal agency oversight via the administra­tive appeals process may be ineffective, and the regulated individuals often have varying levels of expertise to successfully navigate the adminis­trative process. 95 See, e.g., Hoffer & Walker, Death of Tax Court Exceptionalism, supra note 57, at 268–89 (discussing disparities in tax adjudication context); Christopher J. Walker, Referral, Remand, and Dialogue in Administrative Law, 101 Iowa L. Rev. Online 84, 94–95 (2016) [hereinafter Walker, Referral] (same for immigration adjudication).

Remand and court–agency dialogue play a particularly important systemic role in agency adjudication contexts in which many individuals do not have legal representation and often lack the wherewithal to seek further review of erroneous agency actions. I have previously made this point in the context of immigration (and tax) adjudication:

[H]aving a systemic effect is particularly important for immigra­tion adjudication and in other agency adjudication contexts where less-sophisticated individuals navigate the agency process, oftentimes without legal representation. In those circumstances, it is much more likely that individuals will not seek judicial review of erroneous agency decisions—either because they lack the sophistication to navigate the judicial process or have other­wise procedurally defaulted meritorious claims in the admini­strative process. Only by remanding and forcing the agency to correct systemic errors can the court help these individuals who fail to seek judicial review. 96 Walker, Referral, supra note 95, at 93; accord Hoffer & Walker, Death of Tax Court Exceptionalism, supra note 57, at 279–81 (making a similar argument in the tax adjudication context).

To be sure, under Bagley’s more restrained remand standard, many cases would still be remanded, and through those remanded cases, the court–agency dialogue would continue. But, as discussed in Part I, there would certainly be fewer remands and more false negatives—cases in which the agency error was erroneously deemed harmless. Perhaps more impor­tantly, the increased focus on the harms of the parties before the court could shift attention away from the systemic harms caused by the agency’s insufficient processes, erroneous substantive positions, or other­wise arbitrary and capricious actions. The rule-based ordinary remand rule and accompanying court–agency dialogue help address these sys­temic problems by leading to improved consistency and quality of deter­minations not just in cases that eventually reach the courts but, more importantly, in the (much greater number of) cases that are never appealed.


The modern administrative state poses many dangers to democracy and to the American form of government. Most federal lawmaking these days occurs by regulation. Such regulation is subject to very deferential judicial review. Although Congress continues to exert some control over federal agencies via oversight and appropriations, substantive legislative efforts to direct regulatory lawmaking have become virtually nonexistent in recent years. (And when Congress does legislate, the federal agencies that are affected by the legislation play a substantial, often confidential, role in drafting it. 97 See Christopher J. Walker, Legislating in the Shadows, 165 U. Pa. L. Rev. (forthcoming 2017) (manuscript at 3), (on file with the Columbia Law Review). ) The charitable view of current lawmaking by regula­tion is that federal agencies have had to get more creative with their stale statutory mandates to address changed circumstances while still respecting the bounds of such congressional delegations. The more cyni­cal view is that federal agencies believe they “can’t afford to wait for Congress” to provide statutory authority and permission to address the problem, instead opting to “go[] ahead and mov[e] ahead without them” via regulation. 98 Josh Blackman, The Constitutionality of DAPA Part II: Faithfully Executing the Law, 19 Tex. Rev. L. & Pol. 213, 269 (2015) (emphasis in original) (internal quotation marks omitted) (quoting President Barack Obama regarding the executive actions on immigration at issue in United States v. Texas).

Of all these challenges facing administrative law today, the lack of remedial restraint (at least as Bagley conceives of it) is not among them. On the contrary, the rule-based ordinary remand rule plays an important role in preserving a proper separation of powers, in ensuring agencies exercise their congressionally delegated discretion in a nonarbitrary manner, and in facilitating a richer court–agency dialogue that helps courts have a systemic effect on the administrative process. The benefits of the ordinary remand rule far exceed any benefits of a more restrained standard-like remedial approach. And this rule-based approach avoids the costs of courts assessing, for instance, whether regulation by press leakage or regulation by Twitter should be an acceptable, harmless substitute for notice-and-comment rulemaking.

If Bagley’s call for remedial restraint does gain traction, however, it is important to assess how his proposal would affect the rest of admin­istrative law. As Nielson has recently noted in a related context, Seminole Rock (or Auer) deference “does not exist in a vacuum but rather is part of an interconnected network of administrative law doctrines. When one part of the network is changed, that change reverberates across administrative law. Prudence suggests that [the] Supreme Court should understand those interconnected consequences before changing important doctrines.” 99 Aaron L. Nielson, Beyond Seminole Rock, 105 Geo. L.J. (forthcoming 2017) (manuscript at 7), (on file with the Columbia Law Review). Administrative law already has a number of doc­trines in place to address Bagley’s concerns about the agency delay and inefficiency caused by judicial review.

United States v. Texas, discussed in the Introduction, is illustrative. Article III’s standing requirement calls for the challengers to demon­strate an injury in fact (fairly traceable to the agency’s action and likely to be redressed by the judicial relief sought). The APA requires that the challengers’ interests fall within the zone of interests of the statute. The APA exempts from judicial review agency actions committed to agency discretion by law. And judicial review under the APA and related judicial deference doctrines is highly deferential—much more so than appellate review in the civil litigation context.

Accordingly, introducing a much more searching harmless error standard could risk upsetting this calibrated system for judicial review that strives to balance congressionally delegated agency discretion with judicial review to prevent such agency discretion from being exercised in an arbitrary or capricious manner. The current calibration is not perfect, but it is much better than the more remedially restrained system for which Bagley advocates.