The Administrative Procedure Act (APA) is a broadly worded statute that has benefitted from case law to fill many of its gaps, ambiguities, and inconsistencies. But the case method directs judicial attention to slivers of APA inquiry that are required to resolve cases in as-applied challenges to rules and adjudications. There is another method of APA interpretation that has never been deployed in the statute’s seventy-seven year life—that of intentional collaboration between the Executive Branch and the Judiciary. Acting on their litigation and case management authorities as well as their unique power to persuade the Judiciary on questions of administrative procedure, the Attorney General can utilize APA rulemaking procedures to promulgate interpretations of key discrete provisions of the APA. The Attorney General does not need to do this through notice-and-comment rulemaking because the APA is not organic to the U.S. Department of Justice (DOJ). Nonlegislative rulemaking will suffice. Such guidance would facilitate judicial review of pure interpretive distillations of the statute. Through the iterative application of the simple remand rule, the goals of the APA would be advanced through interbranch dialogue with the Judiciary either upholding Attorney General interpretations of law or remanding them to DOJ for revision or abandonment. Certain provisions in the APA’s Judicial Review Chapter are most amenable to this process because they pertain to Executive Branch litigation conduct that is in the near-exclusive statutory domain of DOJ. Offering interpretations of key APA provisions through guidance would benefit society by increasing certainty by agency operators, promoting judicially consistent standards, accelerating the settled understanding of the statute, and increasing the political accountability of the Executive Branch.
Following the 1946 enactment of the Administrative Procedure Act (APA), federal agencies have abided by its procedures for legislative rulemaking of both notice-and-comment rulemaking and formal adjudication varieties, as well as its minimal strictures for nonlegislative rulemaking. As a consensus statute that was drafted to cover a broad diversity of federal agencies, the terms of the APA are broad and subject to gap-filling under common law principles. Congress has not established procedures to regulate all conduct of federal administrative agencies. Most notably, the statute neither meaningfully regulates the dominant form of agency action, informal adjudication, nor regulates agency investigations at all. Meanwhile, federal courts have filled gaps in the APA to answer core questions such as the very definition of a federal agency and agency action. To date, the Executive Branch—with the Attorney General as its chief legal officer—has not intentionally driven the development of the APA beyond litigating cases as they arise, issuing reports and views leading up to the APA’s enactment, publishing a manual following its enactment, and hosting a summit to promote the modernization of the APA in 2019.
Under the premise that legislative revision of the APA will not occur in the foreseeable future, this Article inquires whether the Executive Branch can offer a more collaborative and departmental approach with the Judiciary to promote the spirit of the APA: fair, transparent, and efficient government. By invoking its general regulating, policymaking, and unique litigation management authority, the U.S. Department of Justice (DOJ) could interpret portions of the statute that will, in turn, draw litigation challenging DOJ’s interpretations. Courts will either uphold the rule, which would fill gaps in the statute, or courts will likely remand for additional analysis and explanation as opposed to conclusively interpreting the APA, which would also fill gaps in the statute. By potentially issuing subregulatory rules that utilize the public participation element of notice-and-comment, this strategy would anticipate dissatisfied members of the regulated public suing to challenge DOJ’s APA regulations. Such a strategy will enable an iterative process of judicial review of DOJ’s work, with remands if necessary and affirmation if appropriate. The net result would be a more intentional and directed approach to resolving certain ambiguities that would advance the clarity of the APA. Specifically, DOJ could issue rules using 5 U.S.C. § 301 and 28 U.S.C. § 516 to implement the APA, particularly on litigation or evidentiary matters leading up to litigation as long as the rules are procedural in nature and not substantive. DOJ need not even consult with the public, however. So long as DOJ meets an exemption from the APA’s notice-and-comment requirements, the Attorney General could more quickly issue a guidance document or a letter to department heads. But assuming no remand is taken, an APA-interpreting rule will likely survive only if it is narrowly cabined on the Attorney General’s statutory powers of securing evidence and determining the conduct of litigation.
The process of regulating provisions of the APA under DOJ’s general litigation authority gives the Executive Branch skin in the game and credibility to persuade the federal courts. This manner of nonlegislative regulation would be a key vehicle to raise important questions about the APA’s contours to the Federal Judiciary in a way that the current process of incidental consideration as a part of deciding cases cannot easily provide.
Significant benefits would flow from this process. The regulated public will have access to more discrete procedures so that they can make better choices in areas subject to federal regulation. The federal government will benefit because the agencies represented by DOJ—as well as the geographically and topically dispersed units of DOJ—will have a unitary body of precise procedures by which to comport themselves. Federal courts will have the benefit of deliberate decisional material on the APA’s meaning in a way that a trial or appellate litigator cannot provide in the context of a single case. Adherents to the private rights role of the federal courts would encounter cases that squarely challenge pure interpretations of the APA; thus, those jurists would need to either affirm DOJ’s interpretation or set the interpretation aside under arbitrary and capricious review. Public rights adherents would have a convenient and classic mechanism to answer difficult questions. Congress would have a clearer picture of the policy outcomes of a more nuanced APA paradigm and consequently be better equipped to form consensus on how to reform or improve the statute. That body could then, in its prerogative, enshrine APA innovations in statute or preclude them.
A cost of this approach would be the utilization of scarce time and resources to draft the regulations and litigate in their defense. Another cost would be the potential political consequences of the Attorney General—as the signer of the nonlegislative instrument—actually describing what they think is the best approach for filling in gaps and areas of silence and curing inconsistencies in the APA. There is precedent for DOJ’s unique position to expound upon the nuances of the statute. That Department issued multiple, influential, pre-APA enactment reports in the 1940s and a manual for the APA post-enactment that is heavily deferred to by the Judiciary.
Only a handful of APA provisions are amenable to this kind of rulemaking. We identify two areas ripe for APA regulation that have not been comprehensively resolved by the U.S. Supreme Court. First, courts have wrestled with the scope of the administrative record to be reviewed in APA actions, which the APA simply states must be the “whole record.” The text is ambiguous and the subject of significant dispute. The evidence that courts consider when reviewing agency action is important and usually dispositive of the dispute, yet also subject to ambiguous and confusing common law. DOJ should define “whole record” as the body of evidence that courts must use when adjudicating their § 706 review function. Second, the APA empowers agencies and courts alike to postpone final rules, but that unusual authority has never been fleshed out by courts by virtue of its ephemeral use, i.e., typically opposing-party presidential transition periods. DOJ regulations could offer a comprehensive explication of that authority and help regularize disputes that often arise in an emergency posture.
To be clear, there are limits. We do not see all, or even most, of the APA as being amenable to regulation. Vague, impactful, or rarely litigated portions of the APA likely cannot be implemented via rulemaking. These include the APA’s definition of “agency,” subpoena subsections, and judicial review provisions related to agency discretion. These are much more distant from the litigation and evidence-gathering functions that give rise to DOJ regulation.
Our Article examines APA regulation by DOJ in particular due to its statutory ownership of federal housekeeping regulatory authority, but DOJ need not be the sole agency engaging in this work. Joint regulations could be of value. For example, administrative regulations on discrete actions taken by a given agency that consider the particulars of the authorizing statute and the nature of the substantive subject matter could be additive to the general administrative record regulation. Although we conclude that regulation of the APA is possible and beneficial, we recognize Congress’s primacy in this space and acknowledge that any new statutory pronouncements would take precedence over contrary guidance of the Attorney General.
Part I of this Article engages with the background of regulating the APA, with a significant emphasis on separation of powers principles and DOJ’s pre- and post-enactment influence on the APA. Part II follows by laying out how APA regulations can be achieved. Part III asks the question of whether the APA ought to be regulated and considers the benefits and costs of doing so. Part IV delves into the record rule, rule postponement, and investigations provisions of the APA and makes the case that they should be subject to regulation.
No federal agency has ever attempted to promulgate regulations to expound portions of the APA itself. Title 5 of the Code of Federal Regulations might be expected to contain any such regulations because the APA is codified in title 5 of the United States Code. However, that title contains only regulations concerning, for example, ethics and grants and agreements for agencies like the Office of Management and Budget under its subtitle headings. Even when the Executive Branch has considered creative approaches to APA practice, DOJ did not include regulation of the APA as a possibility.
That absence is perhaps to be expected. The text of the APA today does not explicitly require any executive officer to establish regulations. This silence contrasts with other significant statutes that clearly make such pleas. The Immigration and Nationality Act, for example, states that the Secretary of Homeland Security “shall establish such regulations . . . as he deems necessary for carrying out his authority under the provisions of this chapter.” The law gives various cabinet officials—including the Attorney General, Secretary of Homeland Security, Secretary of State, and Secretary of Labor—freedom to establish regulations in more than a dozen other settings. Examples abound elsewhere in federal law: the Copyright Alternative in Small-Claims Enforcement Act of 2020, the Act to Prevent Pollution from Ships, and the Emergency Livestock Feed Assistance Act of 1988, to take a few at random. Some examples also arise in title 5, such as the Privacy Act and Freedom of Information Act (FOIA).
The original APA contained text that contemplated the issuance of regulations to further its aims. When it was enacted in 1946, the APA contained a public-disclosure component. Section 3 of the APA required agencies to publish adjudication orders and opinions or, “in accordance with published rule,” make them available to the public. The same section required agencies to provide public records to requestors “in accordance with published rule,” except for confidential matters. The Attorney General’s Manual on the APA clarified that this subsection encouraged agencies to “set forth in published rules the information or type of material which is confidential and that which is not.” It emphasized that the APA was not even requiring these rules to take the form of a regulation necessarily: “Regular publication of decisions in bound volumes or bulletins, as many agencies are now doing, will suffice.” Yet no agency appears to have ever published a rule under this authority. That result would not be surprising. One study has shown that of 390 delegating statutes surveyed between 1947 and 1987, no final rule was issued on those statutes in 230 cases, or 59% of the time.
Of the very few judicial opinions to discuss § 3, one appears to concern an agency’s failure to issue rules. The case arose out of the Federal Trade Commission’s (FTC’s) investigation into Macy’s upon belief that the department store was importing mislabeled wool products. In federal court, Macy’s challenged the FTC’s refusal to issue a subpoena to government employees involved in the negotiations over the importation of sweaters, arguing that the refusal violated § 3 because it was made by an unpublished, “secret” directive and not the “published rule” that the APA required. The Court ruled on other grounds, holding that Macy’s had filed the case prematurely and had to wait for a final agency order, at which point it had to petition for review in the court of appeals. The court opined, however, that it cannot be said that refusing a subpoena by an unpublished directive is clearly in defiance of the APA.
Section 3 of the APA lasted for twenty years until supplanted by FOIA. Section 3 “was plagued with vague phrases” that eventually led Congress to pass FOIA, a separate, strengthened public disclosure law. The next year, Congress passed a law recodifying FOIA into the APA at 5 U.S.C. § 552 but did not recodify the APA as a whole. Although that law was passed several years after the original APA, and FOIA is generally given distinct analytical consideration from the APA, the two share legislative ancestry.
The APA today lacks explicit implementation language. The statute focuses on the process for creating agency rules to apply their authorizing statutes, but not the possibility of implementing the APA itself. The bill’s drafters and relevant congressional committees apparently did not express views on whether the congressional imperatives and judicial review standards could or should be broken down into executive regulations. On the one hand, that silence might not be deliberate; the legislative history is equivocal. The road to the APA took about seventeen years, and the documents telling its story are “voluminous, scattered, and difficult to navigate without substantial pre-existing knowledge of the events that lead to the statute’s enactment.” On the other hand, “[t]he APA is not merely a statute: it is a superstatute.” According to the Senate Judiciary Committee, the APA was intended to be a “bill of rights for the hundreds of thousands of Americans whose affairs are controlled or regulated in one way or another by agencies of the Federal Government.” Such a high-order document naturally “contains many compromises and generalities and, no doubt, some ambiguities,” as the Supreme Court has observed. That is because, as an act following “years of deliberation and debate,” the APA “was a monumental compromise, and Congress necessarily left much of the Act vague.”
The next challenge is determining how, if at all, an agency could regulate the APA. As a threshold matter, the fact that the APA does not appear to have been regulated before does not, in our view, imperil future attempts to do so. Some 41% of congressional invitations to agencies to issue regulations go unanswered. There is precedent for the APA providing a rulemaking hook in the public-disclosure provisions of § 3 that has since been superseded by FOIA. Although § 3 was repealed and reconfigured into FOIA, suggesting Congress disapproved of continuing § 3 regulations, Congress continued to require regulations in FOIA. Section 3 was enacted contemporaneously with the remainder of the APA in 1946, which suggests a broader acceptance of regulation by its drafters.
There is also precedent for parts of the APA lying underused for years. Section 705 empowers an agency to stay an action it has taken pending judicial review and authorizes a court to stay an action pending review. That provision has been cited by 416 courts as of June 11, 2023, according to Westlaw, with 161 of those cites—about 38.7% of them—occurring just in the years 2018–2023.
Although the APA has not featured a rulemaking hook since § 3’s repeal in 1966, there are other vectors for regulation. The prime question is whether Congress left any statutory gaps in the APA or in other statutes that could be used to regulate the APA. It is in such gaps that sound rules ordinarily flourish in other contexts. After all, “rulemaking authority is not inherent in an administrative agency but is instead derivative of and dependent upon statutory authorization.” For an agency to regulate, it must have been gifted the authority by Congress to issue regulations of that sort. The grant might include a statutory gap or a statutory ambiguity, with explicit and implicit delegation. While limitations exist, such as the major questions doctrine’s requirement for a clear statement, it is generally commonplace for statutes to confer authority on agencies to make rules.
On that basis, we think that the intersection of two statutes provides a basis for the Attorney General to issue a narrow body of rules governing the Executive Branch’s treatment of the APA. Those provisions are 5 U.S.C. § 301, the general authority for an agency to self-regulate, and 28 U.S.C. § 516, which provides the Attorney General with authority to direct the conduct of federal litigation. Section 516 of title 28, which reserves the conduct of litigation and the securing of evidence to DOJ, ought to be paired with § 301 of title 5 because the latter, but not the former, expressly contemplates rulemaking. We examine each in turn.
The federal agency “Housekeeping Statute” gives each agency head the authority to make procedural rules governing the agency’s operations. It provides: “[T]he head of an Executive department or military department may prescribe regulations for the government of his department, the conduct of its employees, the distribution and performance of its business, and the custody, use, and preservation of its records, papers, and property.”
The Housekeeping Statute was enacted in 1789 “to help General Washington get his administration underway by spelling out the authority for executive officials to set up offices and file Government documents.” In 1979, the U.S. Supreme Court confirmed in Chrysler Corp. v. Brown that the Housekeeping Statute enables only procedural rules, not substantive rules. The Court analogized § 301 to the APA, which recognizes a class of rules exempt from notice-and-comment procedures—“rules of agency organization, procedure or practice”—as opposed to “substantive rules.” Similar to how the APA alleviates the hurdles an agency must surmount to enact “rules of agency organization, procedure or practice,” the Court held that § 301 was “simply a grant of authority to the agency to regulate its own affairs.”
The line between procedural and substantive rules can be murky in administrative law. One 1920s definition for procedural rules was “laws which have for their purpose merely to prescribe machinery and methods to be employed in enforcing these positive provisions.” The same author defined substantive laws as “laws which have for their purpose to determine the rights and duties of the individual and to regulate his conduct and relation with the government.”
Nevertheless, since Chrysler Corp., the federal courts of appeals have struck down several attempts by agencies “to find statutory authority for substantive regulation in the Housekeeping Statute.” For instance, § 301 “cannot be construed to establish authority in the executive departments to determine whether certain papers and records are privileged.” Nor could it be used to “issue binding regulations creating rights to work overtime which are enforceable by a federal court.” In so holding, the First Circuit noted that to view § 301 otherwise would be to recognize a “generalized grant of authority to all military and executive department heads . . . so broad and non-specific that we cannot reasonably conclude Congress intended this statute to serve as the sole basis for jurisdiction in the federal courts for this internal policy pronouncement.”
Some Housekeeping Statute regulations have survived challenges that they are not truly procedural in nature. Perhaps most famously in United States ex rel. Touhy v. Ragen, the U.S. Supreme Court held that agencies could use the Housekeeping Statute to issue regulations requiring agency preapproval before the release of its official records and information. As a later circuit court decision characterized Touhy, the regulations in that case “established which agency employees would produce documents in response to a subpoena duces tecum” and thus “dealt exclusively with internal administrative procedure.”
That said, § 301 has been used to justify innumerable modern rules. For example, DOJ “has interpreted § 301 to allow agencies not only to set rules for employee conduct while on the job, but also to regulate employee conduct outside the workplace that ‘may undermine the efficient operation of the Department or the effectiveness of employees in the performance of their duties.’” There may be tens of thousands of regulations relying on § 301, albeit often in conjunction with other statutory bases.
Another vehicle for regulating the APA comes not from title 5 of the U.S. Code, but from title 28, devoted to the Judiciary and DOJ. Section 516 of that title provides: “[E]xcept as otherwise authorized by law, the conduct of litigation in which the United States, an agency, or officer thereof is a party, or is interested, and securing evidence therefor, is reserved to officers of [DOJ], under the direction of the Attorney General.” Unlike § 301 of title 5, § 516 of title 28 does not expressly empower DOJ to issue regulations.
Section 516 aligns with the longstanding principle that the Attorney General maintains primary control over litigation on behalf of the United States. The Attorney General’s involvement throughout litigation, including at trial-level court proceedings, ensures a better handle on cases facing judicial review. The Attorney General’s power is so complete that they and their Assistant Attorneys General wield the power to force litigation positions on agencies.
Like the Housekeeping Statute, § 516 traces its roots back to the founding. The Judiciary Act of 1789 established the Office of the Attorney General. The Act vested in the Attorney General the absolute power to “prosecute and conduct all suits in the Supreme Court in which the United States shall be concerned” and to advise the President or heads of Departments regarding questions of law. That Act also vested the Attorney General with the plenary power to “conduct all suits” in which the United States may be a party. In the years following, the Supreme Court recognized the Attorney General’s authority to litigate on behalf of the United States, noting the authority might come from the Judiciary Act, Supreme Court decisions, or “the usage of the government.”
The Attorney General’s litigation authority was fortified by the 1870 Act, which established DOJ, headed by the Attorney General. In that law, the Attorney General’s role expanded to representing government agencies and departments. Particularly, § 14 of the 1870 Act directed that the Attorney General “shall . . . procure the proper evidence for . . . all suits and proceedings.” The 1870 Act furthered the goal of Congress to create a centralized mechanism for conducting and supervising government litigation. It created a mechanism for “a unity of decision, a unity in jurisprudence . . . in the executive law of the United States.”
The next step in this language’s evolution came in 1954. Section 361 of the Revised Statutes provided that the
officers of the Department of Justice, under the direction of the Attorney General, . . . shall, on behalf of the United States, procure the proper evidence for, and conduct, prosecute, or defend all suits and proceedings in the Supreme Court and in the Court of Claims, in which the United States, or any officer thereof, as such officer, is a party or may be interested.
Although the 1954 statute and its antecedents used the word “procure,” that was replaced with “secure” in 1966 when Congress enacted § 516 in its present form. Both the terms “procure” and “secure” suggest a statutory history of DOJ’s active and decisive role in gathering evidence to present during litigation.
Also in 1966, Congress enacted § 3106 of title 5, the corollary to § 516 of title 28. Whereas § 516 assigns DOJ primacy in handling litigation and the securing of evidence for the U.S. Government, § 3106 generally prohibits other agencies from assuming that role. It requires heads of executive departments to refer litigation matters, including the “securing of evidence therefor,” to DOJ except as otherwise authorized by law. To drive the point home, Congress emphasized that the Attorney General “shall supervise all litigation” in which the United States is a party and “all United States attorneys, assistant United States attorneys, and special attorneys appointed under [§] 543 of this title in the discharge of their respective duties.”
The U.S. Supreme Court and courts of appeals have primarily invoked § 516 to reiterate the well-established precedent that the Attorney General is charged with the “conduct of litigation” in which the United States or agencies are a party to, absent laws to the contrary. Section 516 creates an exception to the Attorney General’s presumptive litigation authority for non-DOJ litigating situations “otherwise authorized by law.” Similar language constrains non-DOJ agencies’ ability to argue in the Supreme Court: the authority to initiate appeal proceedings is cabined to the Attorney General or Solicitor General. But courts generally construe such language narrowly and only allow agencies to conduct independent litigation with explicit statutory authorization. Without valid statutory authorization, the Supreme Court has rejected an agency’s attempt to file a petition for a writ of certiorari in its own name, even though the agency possessed independent litigating authority at the district court level.
Grants of independent litigating authority are neither uniform nor comprehensive. As of 1978, at least thirty-one executive branch and independent agencies were authorized to conduct at least some of their own litigation, leading to an inconsistent statutory scheme in representing agencies in civil ligation. A spectrum lies between agencies like the FTC, which largely litigates its own cases with lesser remedial authority, and agencies like the Environmental Protection Agency (EPA), which have litigating authority extending to certain parts of their authorizing statutes. Some agencies’ litigating jurisdiction lies only in the district courts, like the Consumer Product Safety Commission, or only in the courts of appeals, like the Securities and Exchanges Commission.
Unlike the Housekeeping Statute, § 516 does not expressly authorize any agency to promulgate regulations. At best, it is silent on whether further rules are necessary to build out the regime where DOJ primarily handles the federal government’s litigation and evidence gathering. But if the text does not fully support using § 516 to make rules, then its historical use does.
The most common regulatory invocation of § 516 confirms DOJ’s sole right to represent the United States or agencies in suits. For example, DOJ has turned to § 516 as the fount for regulations concerning contract terms and conditions, patent and copyright infringement cases in federal procurement actions, certain Department of Agriculture claims, and individual-capacity suits against federal employees. Even when an agency enjoyed the authority to “determine the amount of, to settle, and to adjust any claims arising” under a particular program, the regulation made clear that the delegation at no time was intended to “diminish the authority of the Attorney General” under § 516.
Section 516 is not only referenced in regulations (as when regulations emphasize that they are not intended to impinge on § 516’s default arrangement), but also used as express authority for some regulations. The vast majority of these regulations are DOJ’s own regulations. They include regulations describing the minimum qualifications for annuity brokers to assist with structured settlements entered by the United States, regulations governing Young American Medals for youth engaging in acts of bravery or service, and the DOJ Antitrust Division’s business review procedure. Others are housekeeping matters governing litigation policy, like expressing DOJ’s “general overriding affirmative duty” to support open judicial proceedings. But § 516 has also been used by DOJ to “establish a uniform procedure for consular notification where nationals of foreign countries are arrested by officers of this Department on charges of criminal violations” to conform to the United States’ treaty obligations. The section is the foundation for “procedural guidance” to agencies responsible for enforcing Title VI of the Civil Rights Act of 1964. Section 516 is also the invoked authority for a regulation establishing how DOJ accepts concurrent federal criminal jurisdiction to prosecute offenses in tribal territory.
The question then becomes—What are the outer limits of § 516 regulation? Section 516 states that “the conduct of litigation . . . and securing evidence therefor, is reserved to officers of the Department of Justice, under the direction of the Attorney General.” Take both of those endowments in turn. The “conduct of litigation” reiterates a broad delegation to the Attorney General; to conduct means “[t]o manage; direct; lead; have direction; carry on; regulate; do business.” This facially broad definition could imply that DOJ could simply litigate an APA case without ever consulting with the agency whose action DOJ is defending. However, that power is tempered by DOJ’s historical restraint and recognition of its order in the Executive Branch. DOJ’s Office of Legal Counsel has previously cast this authority as a “supervisory” one, whereby DOJ “coordinate[s] the legal involvements of each ‘client’ agency with those of other ‘client’ agencies, as well as with the broader legal interests of the United States overall.” The Attorney General, under that opinion, “will accommodate the agency’s policy judgments to the greatest extent possible without compromising the law, or broader national policy considerations.” Notably, DOJ’s conduct is linked to “litigation,” not the antecedent agency action or any agency action following closure of the case. The Attorney General’s powers correlate to this one stage in the administrative state’s lifecycle.
This limited reading of § 516 aligns with Chief Justice John Marshall’s observation of the nondelegation doctrine. On the one hand are “important subjects” whose regulation occurs only by Congress, and on the other hand are matters of “less interest” for which the Executive Branch may “fill up the details.” DOJ recognizing that § 516 regulation should be limited to procedural matters directly linked with court filings and appearances respects any boundaries and reduces litigation risk in this uncharted territory.
We think that DOJ could issue rules using 5 U.S.C. § 301 and 28 U.S.C. § 516 to implement the APA, particularly on litigation or evidentiary matters in or leading up to litigation, so long as the rules are procedural in nature and not substantive. Those two statutes have individually supported countless regulations from agencies, particularly concerning internal agency procedure and DOJ litigating authority. Those fields converge, generally speaking, on an agency’s procedures for preparing for litigation and the litigation-induced evolution of roles played by the agency performing the action under review and DOJ spinning up to defend that action. However, we cannot say that every APA provision falls in that intersection. Although the APA, at a high level, is concerned with reasonably universal administrative procedure, its provisions cover a diverse set of administrative law topics.
Guidance could take a number of forms. The soundest route would be for DOJ to pass a regulation after notice-and-comment. That would forestall a challenge to how the rules were enacted and bolster the government’s case for any Skidmore deference. An open and upfront rulemaking process through notice-and-comment could also tee up the rule for judicial review (pending a plaintiff with standing), which might be beneficial: the courts would then be confronted with a pure legal question and be forced to rule on a number of the finer and more important points. That said, for the sake of expediency, DOJ is free to try regulating the APA through subregulatory rules, for example interpretive rules or policy papers.
DOJ is the right entity to promulgate any such rules. Section 301 of title 5 applies to all agency heads, enabling them to regulate in pursuit of housekeeping matters. But § 516 of title 28 grants the Attorney General the primary responsibility for conducting litigation and securing evidence for that litigation. DOJ is thus charged by § 516 and other relevant statutes to represent rules promulgated by administrative agencies. To let then-Assistant Attorney General Ted Olson summarize: “[T]he Attorney General is properly perceived by the people as the official charged with ensuring that the Executive Branch observes constitutional limits and that the laws enacted by Congress are faithfully executed.” Attorneys are assigned to a range of cases that vary in subject matter and administrative programs before the entire federal Judiciary. This exposure allows DOJ attorneys to understand how regulations are implemented and challenged under the APA and other statutes. DOJ is responsible for defending agency rules as counsel for the United States, but deference is afforded to agency’s views on policy matters. Because DOJ attorneys exercise independent judgment, the Department frequently provides advice to agencies on how to deal with impending litigation, but also how to avoid it. Anticipating potentially litigious issues associated with agencies’ policy objectives is one among the many vital roles of a DOJ attorney. However, DOJ’s litigation authority is not solely vested. But DOJ is indirectly democratically accountable, as it is staffed with principal Officers of the United States who may be removed at the conclusion of a Presidential term, if not earlier.
In the APA context, the Attorney General’s views carry unique weight. The Attorney General was heavily involved in drafting the APA by convening a Committee on Administrative Procedure. The Committee issued a report with two bills, one reflecting a liberal approach and one reflecting a conservative approach. The Attorney General also submitted a letter of support of the prevailing bill in 1945 along with an analysis of the bill. After the bill passed and the Attorney General’s preferred view of less restrictive constraints on agencies won out, he authored the Attorney General’s Manual on the Administrative Procedure Act. The manual laid out a sort of post hoc legislative history, though the accuracy of that history has been reasonably questioned as presenting “only one side of the debate” preceding the APA’s passage. Nevertheless, that manual “significantly influenced the way agencies and courts interpreted and applied the statute.” The Supreme Court has given the Attorney General’s Manual varying levels of deference, ranging from “some weight” to “persuasive.” Given this, the Attorney General’s regulation of the APA could carry persuasive power unique to their station. However, any new APA procedural rules likely will not constitute a sequel to the Attorney General’s Manual and share in the deference shown to it. In respecting the Attorney General’s Manual, the Supreme Court has emphasized that that treatise was a contemporaneous interpretation by DOJ. DOJ is free to partner with other agencies to issue joint rules, but at minimum, DOJ is best positioned to offer general administration of the cross-cutting requirements. DOJ has also shown some interest in recent years, having hosted a “Summit on Modernizing the Administrative Procedure Act” in 2019.
To clarify, there is a dearth of APA regulations, but not necessarily APA rules in general. When we discuss “regulating” the APA, we mean the full spectrum of informal rulemaking authority available to the Attorney General. The most durable and resource-intensive efforts along that spectrum are regulations that undergo notice-and-comment rulemaking and are published in the Federal Register. Such rules tend to be more durable than policy documents. It is not uncommon for DOJ to supersede previous policies, as Attorney General Jeff Sessions did en masse in 2017. The President may also order DOJ to consider doing so, as President Biden did in a 2022 executive order aimed to update criminal justice policies.
Subregulatory guidance, such as guidance documents or letters to department heads, is less sturdy but can be quite impactful. A key DOJ policy document is its comprehensive Justice Manual. To use other non-APA examples, Attorney General Eric Holder, in 2013, declared that he would no longer defend the Defense of Marriage Act because it was unconstitutional, having been passed less than fifteen years earlier in a bipartisan fashion and signed into law by Bill Clinton. Attorney General Holder also reformed how the U.S. government invokes the state secrets privilege through a memorandum that failed to even invoke his statutory authority but has nonetheless held firm since it was issued in 2009. Guidance such as interpretive rules or policy documents can still be challenged in an Article III venue. A regulated party who alleges that nonlegislative APA rulemaking has been applied unfairly to them can challenge the application to their case, and they still benefit from having a clear distillation of the government’s APA policy, the same as a rule that has undergone notice-and-comment. Which form of rulemaking may be best suited for the challenge identified in this Article will likely depend on the political environment and provision of the APA that DOJ seeks to regulate.
Recognizing that some agencies have independent litigating authority, we are not using DOJ here as a proxy for any government agency that has the authority to represent the United States in any defensive or affirmative APA lawsuit. The APA and the portions of it that we think the Attorney General could regulate are certainly not organic to DOJ. But § 516 and § 301 authorities that unlock the Attorney General’s litigation regulatory power are collectively organic to DOJ. Because § 301 and § 516 must be read with the APA, they empower DOJ to act. The APA gives way its general authority to more specific authority, each of which postdates the 1946 APA through their 1966 recodifications. DOJ can always issue advisory guidelines for agencies in its counseling function, which renders moot the question of whether independent litigating agencies could issue rules of the sort we discuss in this Article.
Before discussing whether an agency could make rules to implement or expound a given part of the APA, a natural first question is: why would an agency want to do so?
First, having the Executive Branch create universal rules to self-regulate conduct could complement the Judiciary’s occupancy of that function. Federal courts have taken a role in “impos[ing] rulemaking requirements that exceed the simple formula in the APA.” In any event, courts have forged ahead anyway, imposing obligations on agencies that arguably cross the line from judicial interpretation to judicial creativity. Examples include requiring agencies to create records in informal rulemaking, forecast what information it considered in drafting a proposed rule, disclose ex parte contacts in informal rulemaking, and include a robust, exhaustive explanation for its rulemakings. We have previously criticized courts’ eagerness to expand, without statutory support, administrative records beyond what agencies present. Blame could lie in congressional torpor, which emboldens courts to fill the void and conduct searching review of agency action. Relatedly, Congress’s decision not to flesh out certain dimensions of the APA at the outset may have contributed to this behavior. As a law professor, Antonin Scalia argued that the APA’s silence on informal adjudication procedures pushed federal courts to craft common law solutions to fill the gap. The silence may have been intentional because the APA was a consensus statute, but it exists, nonetheless.
Of course, an agency is always free to impose such requirements on itself. But a court cannot saddle agencies with these requirements under threat of the rule otherwise being set aside under the substantive arbitrary or capricious review standard. As the Supreme Court recognized in Vermont Yankee, there is “little doubt that Congress intended that the discretion of the agencies and not that of the courts be exercised in determining when extra procedural devices should be employed.” When courts do so, they violate not only the APA but also the spirit of administrative law.
The result, in the words of the Attorney General’s Committee on the APA, is diverting rulemaking energy away from regulations and toward subregulatory policies and directives: “Rules will either not be made or policy will be driven underground, as it were, and remain inarticulate or secret.” Kathryn Kovacs has identified the limited perspective of the Judiciary and unintended adverse consequences when courts impose agency procedures beyond the APA: “Courts are not well positioned to adjust the benefits and burdens of the regulatory state.” Court-made rules can ossify rulemaking, meaning that it increases the costs for agencies to initiate and follow through with rulemaking. Courts engaging in this practice raise separation of powers concerns, as they do not engage the public in their deliberations and are not electorally accountable. This can erode judicial legitimacy in the unique context of the APA. Finally, courts that circumvent Vermont Yankee discourage Congress from involving itself in the process (although it should be noted that, notwithstanding the spin-off of FOIA, Congress has largely left the APA alone since passing it in 1946). The upshot is that the President can be incentivized to undertake direct, unilateral action. Thus, regulating the APA could enhance the quality of government, especially in litigation, without offending the balance of separated powers.
Besides textualism in the APA, judicial deference doctrine—such that it currently exists—offers another reason for courts to give agencies the opportunity to determine APA-compliant administrative procedures. At a minimum, it could pave the way for agencies to go around courts. Because no meta-rulemaking has apparently occurred under the APA, courts are unlikely to have had the opportunity to evaluate the ambiguousness of the APA statutory language. In any event, the Supreme Court has noted, if at a high level of generality, that the APA is perforated with compromise driven ambiguities. However, resorting to these judicial deference doctrines should be taken with great caution. The Supreme Court has been increasingly critical of administrative deference cases like Chevron. Without Chevron as an aid, agencies might not be able to regulate in the face of judge-made administrative rules but might be able to create new ones in the absence of conflict with the administrative common law. Some Justices have expressed readiness to do away with Brand X deference. The Court has already reconstructed the erstwhile Seminole Rock and Auer deference in Kisor v. Wilkie. Because Chevron and Brand X face a degree of existential threat, they would not be the ideal mechanism to interact with the Judiciary on administrative procedures. But at base, DOJ and the affected agency would still be able to argue for Skidmore deference. That veneer of deference could help DOJ—albeit in a way difficult to quantify—when DOJ grapples with the APA on an agency’s behalf in litigation through a duly promulgated rule.
Second, agency-led regulation has the potential to be more surgical, comprehensive, and flexible, even alongside the courts. The APA contains some ambiguities, and it is costly for agencies to operate in a standardless zone—left to repeat struggles without the benefit of considered, durable stability in how they apply the law. Specific problems with APA provisions that are not built out into rules are discussed below in Part IV. But one model is for agencies to prune and tuck APA regulations, including in conversation with the Judiciary.
When courts see APA procedural rules, they would likely look harder at the ordinary remand rule. Regulating the APA can provide a sharper framing of otherwise gray legal issues. This provides a basis for courts to opine in bona fide cases but then remand to the agency as necessary, rather than kicking the can down the road or incentivizing pocketed silos of district-level norms. Under our theory, a court might be more willing to remand the matter to the agency to decide anew with the benefit of DOJ assistance as opposed to keeping the matter for itself to decide. Thus, the ordinary remand rule is an important and useful mechanism for advancing the law in this area.
Third, the advantage of broadly applicable rules is that DOJ is best positioned to assess the best interests of the United States and best practices for serving those interests. The Attorney General has the statutory and historical role as the federal government’s top lawyer, backed up by their role in the constitutional order. Centralizing the litigation authority in DOJ aids the Attorney General in overseeing each “client” agency, especially in coordination of interagency representation. Congress has slightly complicated that conception by carving out exceptions to DOJ’s centralized litigation authority in the conferral of litigation authority to other agencies. DOJ’s Solicitor General of the United States often still has control over Supreme Court litigation.
But the attorney-client relationship between the Attorney General and agencies can come with difficult questions in terms of parsing policy judgments from legal judgments in litigation. With agency subject matter expertise, the former is generally deferred to agencies, while the latter is the province of the Attorney General. Given that agency decisions are increasingly scrutinized based under APA review, it is in the interest of such agencies to heed the advice of the Attorney General on how best to defend or avoid such claims in their cases.
To be sure, DOJ’s control over litigation has led to less than ideal working relationships between DOJ and agency lawyers. For example, DOJ may overrule agencies’ preferences for case management, including settling the matter. Agency lawyers have expertise in the subject area and an in-depth knowledge of the inner workings of the regulatory program. In other words, utilizing agency lawyers may present an advantage when engaging with issues on the merits. But if § 301 requires rules to be procedural—and § 516 does not suggest that it weakens that limitation—then the types of rules that can be created under those sections would seem to benefit from more procedural expertise. As litigators, DOJ has abundant procedural expertise. In the area of litigation strategy, it is in the United States’ interest for agency lawyers to abide by the best practices of litigating APA nonmerits issues determined by the Attorney General.
Facilitating a more cogent working relationship between DOJ and agency lawyers is paramount to the Executive Branch’s ability to defend agency decisions in litigation. Otherwise, the pattern will continue of agency lawyers advising the agencies on rulemaking but then, ex post, DOJ lawyers having to defend the agency lawyers’ advice in court. Consequentially, DOJ’s function of counseling agency decisions regarding matters of law is “general[ly] invisib[le].” With DOJ mainly serving its role as litigator rather than counselor, DOJ is left with a “fait accompli.”
One downside to having across-the-board regulations from DOJ is that if they are not carefully calibrated, they might not be attuned to the practical differences in the way that agencies act or are governed by their organic statutes. There is minimal historical support for such an approach. Touhy regulations, for one, are issued agency-by-agency. A court might also find delegation problems or major questions violations if DOJ issues rules that cover a wide variety of practice areas that extend beyond the management of conduct and/or the securing of evidence in litigation.
However, regulating portions of the APA is genuinely a more collaborative approach to filling in its gaps and silences. This is the purpose of interagency review for significant regulations under Executive Order 12,866 and its progeny. Agencies would have the ability to comment on and help shape any regulations that might affect their pre-litigation administrative procedures. By doing this, parties and the courts will have a refined and pure expression of legal interpretation on important questions that parties can, and likely will, freely challenge in the rulemaking or agency action capacity. Courts will, in turn, have a crystalized enunciation of the rule, which will result in better framing of the issues before the court. This is more efficient for the regulated public and the courts—and the government as well, which offers savings to the taxpayer.
Finally, there may be little downside to DOJ attempting to regulate the APA. Facing a § 301 or § 516 regulation, courts may hold that the APA does not permit implementing rules, that the wrong agency enacted the rule, or that an implementing rule is an unreasonable interpretation of the APA. Such regulatory efforts, even if unsuccessful for the agencies, may help them make Vermont Yankee arguments to future courts to avoid the creation of a new, conflicting judicial rule or the accretion of existing common law. It may also lay the groundwork for an agency to issue informal guidance within the agency or to other agencies to advise on best practices for complying with the APA.
Consider, for example, Department of Commerce v. New York, where the Supreme Court considered the Census Bureau’s inclusion of a citizenship question on the decennial census and whether APA challengers could properly supplement the administrative record. The Court held that a new tranche of documents could be added to the administrative record, building on an exception to the presumption of the record’s regularity. That exception, for a “strong showing of bad faith or improper behavior,” was first identified some fifty years prior in Citizens to Preserve Overton Park, Inc. v. Volpe but never used since. The Court found that standard met where the Secretary of Commerce claimed to need the citizenship question to enforce the Voting Rights Act, yet the challengers made a showing that the Voting Rights Act “played an insignificant role in the decisionmaking process.” Before a similar case arises in the future, DOJ could issue checklist rules as to what an agency must show before DOJ will defend the agency’s action and administrative record in federal court as good faith. Upon hearing that the agency satisfied DOJ’s guidance, a court may defer to that finding and be more inclined to reject application of the bad-faith exception and accept the record presented. It might not. But even in the latter case, the government is no worse off for having tried. Such cases, even when unsuccessful for the government, would still miss those instances in which DOJ guidance would incentivize agencies to better comply with the Department of Commerce standard and thereby pass judicial review. Such cases would also miss those instances in which DOJ finds the agency to fall well short and refuses to defend the agency action, and the matter does not reach the point where a court is considering whether the challengers have made a strong showing of bad faith or improper behavior. In all cases, DOJ should be the entity to consider and issue rules, be they nonlegislative notice-and-comment rules or less sturdy memoranda.
Focusing on whether § 301 and § 516 could justify regulation of particular parts of the APA, this Article first covers the contents of administrative records in 5 U.S.C. § 706 and then looks to pre-judicial review stays of enforcement in 5 U.S.C. § 705.
The first low-hanging fruit for Attorney General APA regulation is the contents of administrative records for informal adjudication and rulemaking. Agency action can be held “unlawful and set aside” if it is found to be “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” The reviewing court is compelled to “review the whole record” in determining the validity of agency action. The APA does not speak to (whole) record-generating procedures for informal adjudicative proceedings, which are effectively the overwhelming majority of proceedings under judicial review. This gap has led to instability, lack of predictability, and imprecision. The “whole record” has been judicially defined as materials considered or relied upon at the time the decision was made and carried out by the agency. But that definition is tenuous because it can vary depending on subject matter, context, and even inter- or intra-circuit differences. There are no general regulations for the “whole record” requirement.
We examine whether DOJ rules, whether regulatory or subregulatory, could assist agencies in performing their litigation-defense duties.
Section 516, as previously discussed, reinforces that the authority to conduct litigation on behalf of the United States sits with the Attorney General. One enumerated facet of the Attorney General’s plenary power is “securing evidence” for the conduct of litigation. Although not explicitly under its § 516 evidence-securing powers, DOJ has offered guidance to agencies compiling an administrative record that will be the major subject of judicial review. In 1999, the Department’s Environment and Natural Resources Division, without attribution, issued a memorandum titled Guidance to Federal Agencies on Compiling the Administrative Record. The division released this guidance to agencies to create uniform measures in compiling the administrative record. The guidance encouraged agencies to think broadly about record scope, encouraging the exercise of “great care in compiling a complete administrative record” because “[i]t is worth the effort and may avoid unnecessary and/or unfortunate litigation issues later on.” One such admonition was to include all documents “used by or available to the decision-maker, even though the final decision-maker did not actually review or know about the documents and materials.” The ethos expressed in that guidance pervaded other parts of DOJ. In 2000, DOJ’s Executive Office for U.S. Attorneys platformed guidance that is generally consistent with the 1999 memorandum in the United States Attorneys’ Bulletin in 2000.
DOJ distanced itself from the memorandum just a few years later. A 2008 memorandum issued by that division’s Assistant Attorney General clarified that the 1999 memorandum did not dictate requirements to agencies. Rather, the division emphasized that agency record compilation is “an agency responsibility in the first instance and the Supreme Court has made clear that an agency has discretion in how to create the record to make and explain its decision.” The memorandum encouraged agencies to have “their own internal guidance,” but offered DOJ’s availability for “consult[ation]” in compiling the administrative record.
The Department later disowned the 1999 memorandum more forcefully, rescinding it in all but name. In 2017, in conjunction with briefing a Ninth Circuit case called In re United States, DOJ issued a new memorandum stating that the 1999 guidance “should be disregarded,” specifically noting that documents that were a part of the agency’s deliberative process should not be included in the record. The 2017 memorandum stated if agencies included deliberative materials, it would “chill free and frank agency deliberation . . . and may serve as a harmful precedent in other cases.” Subsequent DOJ policy documents on administrative law issues did not reference compiling an administrative record but reiterated the memorandum was discretionary and not “intended to be applied by a court.” The U.S. Attorney’s Office’s article from 2000 has apparently not been formally repudiated.
Although DOJ has recalibrated its advice to agencies on administrative record compilation, the Judicial Branch has not done the same. “Recognizing that the 1999 Guidance is not binding upon agencies,” one court found “that the Guidance nevertheless provides helpful insight into the types of documents and materials an agency should consider when assembling an administrative record.” This was notwithstanding the fact that DOJ had already distanced itself from that memorandum. The Ninth Circuit did the same, finding the memorandum “persuasive” even though it was “inexplicably rescinded the very same day” that the government filed the mandamus petition that initiated the appeal. Another court extensively cited the 1999 Guidance as authoritative, casting it as active guidance by DOJ to agencies as recently as 2021.
In lieu of the dated and somewhat clumsy directive from DOJ on compiling an administrative record, some agencies still derive influence from its basic tenets of the 1999 memorandum. Individual agencies have also moved beyond the DOJ memorandum, innovating within their own territories. The agencies with official policies on administrative records have sewn a patchwork of guidelines with varying levels of detail. The Department of Interior (DOI) issued a guidance document in 2006 to assist employees in compiling a decision file that could be used in forming an administrative record. It outlined that records should “contemporaneously document any decision, and if necessary, an administrative record (‘AR’) for judicial review.” The guidance stated there should be a designated employee, the “AR Coordinator,” to ensure to collection of proper documents. Such documents include: (1) those that were relied upon or considered by the agency, (2) documents available to the decisionmaker “at the time the decision was made . . . regardless of whether they were specifically reviewed by the decision-maker,” and (3) even documents the AR Coordinator believes are privileged. This nearly twenty-page guidance document appears to remain in effect.
Relatedly, the EPA issued a policy with the stated purpose of describing “EPA’s practices for compiling administrative records for use in litigation challenging EPA decisions.” The guidance encouraged gathering the record through the entire decisionmaking process but mentioned excluding inter- and intra-agency deliberative documents. Moreover, it laid out that the document was “consistent with” the DOJ recommendation in the 2008 memorandum that agencies develop guidance on the gathering and content of the administrative record.
The National Oceanic & Atmospheric Administration (NOAA) also recommends designating a coordinator like the DOI does. In developing its own guidance, the NOAA cited relevant case law and “informal guidance” provided by DOJ. However, it acknowledged that the 1999 DOJ guidance has been revised and does not represent a formal DOJ policy. Once an agency decision is officially challenged, NOAA requires a “Custodian” to be designated to compile and maintain all the documents that will make up the administrative record. It also recommended best practices for organizing and indexing the record. The document concludes with a helpful checklist that agency officials can reference in developing a plan to compile the necessary record.
Part 32 of the Internal Revenue Service Revenue Manual contours the procedural requirements for regulations projects. Even though the manual does not explicitly state that its focus is assembling an administrative record per se, it does guide employees on how to “compil[e] the file as soon as the regulation project is opened.”
These sample guidelines are not harmonized. DOI guidelines urge the record to “contain the complete ‘story’ of the agency decision-making process, [including] important substantive information that was presented to, relied on, or reasonably available to the decision-maker.” NOAA guidelines, in contrast, state that the administrative record “‘must include all documents that were directly or indirectly considered by the agency decision-maker’ and all documents that relate ‘to the action under consideration and inform[], or ha[ve] the potential to inform, the decision-maker.’” The difference is subtle: the EPA instructs its operators to collect documents “reasonably available” to the decisionmaker, while NOAA focuses on what documents were “directly or indirectly considered.” Agencies’ materials also differ on how to handle deliberative process documents.
Commentators have coalesced on the value of some form of guidance for administrative record compilation. The Administrative Conference of the United States (ACUS) has repeatedly weighed in on best practices for all agencies to compile APA records, subject to organic statutes. ACUS is an independent government agency established in 1964 by the Administrative Conference Act. Its mission is to study and improve the efficiency of administrative programs and procedures and to develop recommendations for action by proper authorities.
In 2013, ACUS issued recommendations regarding best practices in assembling, preserving, and certifying records for judicial review of informal rulemaking. ACUS recommended “that agencies develop a written policy for treatment of protected or privileged materials, including indexing, in public rulemaking dockets and in certification of the administrative record for judicial review, and that agencies make this policy publicly available.” Similarly, in 2022, representatives from the public and private sector published a recommendation through ACUS, the “Handbook on Compiling Administrative Records for Informal Rulemaking.” The handbook is a sixty-page guide that instructs agencies on optimal methods in compiling an administrative record. The 2013 ACUS committee, however, stated without explanation that the preparation of the record is “properly within the province of the agency.” The committee recommended that agencies publicize their record policies and provide them to DOJ if DOJ represents the agency in litigation.
That ACUS committee did not consider the approach we discuss here: DOJ informal rulemaking under § 301 or § 516 to govern the conduct of agencies. We conclude that DOJ could potentially issue rules on the subject of the APA record’s composition. DOJ could issue informal rules akin to formal, cross-government implementation of ACUS recommendations with which the Attorney General agrees. These rules could also be a superseding and binding update to previous instances of division-issued DOJ nonbinding guidance.
ACUS also recommended that agencies make rules concerning the administrative records for informal rulemaking (rulemaking on rulemaking). We recommend the creation of record-compilation rules for APA litigation concerning agencies’ informal adjudications instead. Because informal adjudication is the most quantitively prevalent body of agency action by a large margin, DOJ rules would reach a larger swath of conduct. DOJ rulemaking here would be facilitated by the fact that the APA generally imposes no positive procedures for informal adjudications, so agencies are left to pockets of case law and their own devices to comply with the APA’s strictures.
We conclude that DOJ could regulate informal adjudication records because 28 U.S.C. § 516’s “securing evidence” provision permits the Attorney General to issue persuasive guidance of what constitutes a complete administrative record, and such a rule is procedural, so the Attorney General could invoke 5 U.S.C. § 301. We have already discussed how § 301 textually, purposively, and historically permits agencies to issue their own rules of internal management, which roughly translates to procedural rules. The Attorney General’s § 516 authority is further contextualized by the notation that they may “secure evidence” for the “conduct of litigation.” The term “secure” means “to guaranty or make certain the . . . discharge of an obligation.” Congress, as previously discussed, revised the language in § 516 to enact the present language in 1966: modifying the Attorney General’s authority from being able to “procure” evidence to being able to “secure” it, and recognizing the Attorney General’s authority in “the conduct of litigation” as opposed to the former ability to “conduct, prosecute, or defend all suits and proceedings.” These distinctions suggest no meaningful difference for the purposes of this inquiry. Neither the Judiciary nor the scholarly literature have engaged with the “securing evidence therefor” provision of § 516. But read in harmony with “conduct of litigation,” “securing the evidence therefor” suggests a procedural role for DOJ when it comes to rulemaking.
If the administrative record is a procedural issue, then it is plausibly within the Attorney General’s supervising authority. In effect, if the conduct of litigation encompassed nonmerit issues, compiling or offering guidance on the administrative record can fall in that sphere of department guidance.
The process of compiling and preparing an administrative record for production is arguably a nonmerits issue. Courts have suggested that errors in the lodging of the complete administrative record with the record are procedural in nature. More broadly, the legislative history of the APA supports this reading. While there were many proposed bills before the passage of the APA in 1946, the core tenet of the legislation was to “settle and regulate the field of Federal administrative law and procedure.”
There are limitations on any informal rules that DOJ could issue. The Senate Judiciary Committee, concerning the bill that became 5 U.S.C. § 706, explained that judicial review of “the whole record” meant that “courts may not look only to the case presented by one party, since other evidence may weaken or even indisputably destroy that case.” That does not mean that the court can or must look beyond just the agency’s one-party account of the record. The Attorney General’s Manual comments that “the phrase ‘whole record’ was not intended to require reviewing courts to weigh the evidence and make independent findings of fact; rather, it means that in determining whether agency action is supported by substantial evidence, the reviewing court should consider all of the evidence and not merely the evidence favoring one side.” Somewhat relatedly, the APA laid out rules for administrative hearings. There are no formal rules of evidence for such proceedings, but the Senate Judiciary Committee remarked that the procedures “must be the same as those prevailing in courts of law or equity in nonadministrative cases.” That appears to be an evidence-securing gap that DOJ could fill. For example, the Attorney General could decide that administrative record guidance mirrors rules of evidence or civil procedure that are used in court proceedings. It would likely cross the line into substance territory if DOJ issued a rule that a court should give particular evidence any particular weight.
Every agency action is unique, and the contents of the record depend on what the decisionmaker considered in rendering the agency action in that moment. One-size-fits-all regulation might seem at odds with that reality. DOJ litigators, meanwhile, tend to be generalists. While they often develop deep subject matter expertise by virtue of litigating the same types of cases day in and day out, and working with the same agencies, they often will not have the subject matter expertise of agency lawyers. That said, APA regulations can provide useful parameters for record compilation without binding agencies’ hands in every instance, especially because the agencies will remain involved in APA litigation. The agencies could also be involved in any DOJ rulemaking process; they would presumably submit detailed comments, either formally or through bureaucratic back channels. There is also a spectrum of specificity for agency informal adjudications: the general rules that DOJ would promulgate would still be more specific than the enunciations of a majority of the courts. Agencies, in turn, could and should then issue their own more specific guidance.
Section 516 is sufficient for DOJ to pass a procedural rule governing how agencies compile administrative records and then submit them to DOJ litigators for use in federal litigation. The statute expressly references DOJ’s ordinarily exclusive authority to secure evidence. Section 301, the Housekeeping Statute, permits the Attorney General to implement that authority by expressly giving DOJ rulemaking authority for the “performance of its business,” which includes the “conduct of litigation” under § 516. Supplementing DOJ’s litigator role with a counseling component would allow DOJ to best defend agencies upon judicial review involving the nonmerit issue of record composition. Counseling would also help insulate any DOJ guidance from judicial review, although it may not receive the Skidmore deference due to regulations. However, as previously discussed, courts have given such deference against DOJ over its protests.
As far as the content of the rule, the guidance for agencies must be more nuanced than providing a high-level parroting definition of an administrative record because agencies can already access that information by perusing appellate opinions in the jurisdiction where they intend to act. The DOJ rules should contain some level of detail. The benefit for regulated parties is that agencies could apply the DOJ standard to their discrete actions based on the particularities of their work. For an agency such as U.S. Citizenship and Immigration Services, the DOJ guidance could be sufficiently nuanced as to distinguish between particular immigration benefit types, such as immigrant and nonimmigrant benefits. That kind of guidance would be valuable not only for the regulated public but also for the government as a whole because administrative records will be more consistent and disciplined.
A DOJ rule could specify items that an agency should presumptively consider to be part of the record: a launch-point checklist similar to what the NOAA has used. A rule could establish what internal affirmations or certifications the agency must make to the litigators to assure the litigators that they are receiving the whole administrative record and presenting it to the court or the opposing party in good faith. A rule could also designate the identity or position title of the record custodian. The individual could be the agency’s chief information officer, an operational leader, or an ad hoc position of indeterminate rank, similar to the DOI’s AR Coordinator role. Although the APA does not require an administrative record to be certified, it has become common practice for agencies to do so, and some courts require it, in apparent contravention of Vermont Yankee. This has led to litigation over the certification, for example, when a certification is properly worded. The rule could specify which official within the agency must certify the record, the circumstances in which that authority may be delegated (perhaps no further down than the Senior Executive Service), and the parameters for the certification’s content.
The rule could have some limitations. Any effort should pertain only to the administrative records for informal adjudications. The APA itself suggests the inputs for other types of agency action. For example, the administrative record for formal adjudications must include the full hearing record, while informal rulemaking records include notices of rulemaking, transcripts of oral presentations, and committee reports. Only with informal adjudication does the APA provide no clues, save for the perfunctory notice of decision that is required by 5 U.S.C. § 555(e). The absence of statutory procedures makes informal APA adjudication a higher value endeavor for rulemaking.
A court could conceivably sustain a challenge to DOJ’s ability to use the Housekeeping Statute and § 516 to promulgate rules applicable to other agencies, perhaps on the basis that the delegation of authority in those statutes does not permit cross-government rulemaking.
Consider a court hearing a case arising in the following manner. A plaintiff aggrieved by an agency action sues the agency. The agency produces an administrative record. Plaintiff challenges the adequacy and completeness of the administrative record. The agency responds that the record comports with the new DOJ rules on record compilation. A court decides that the DOJ rules are unlawful themselves under the APA, notwithstanding any judicial deference. Therefore, the defending agency cannot rely on those rules, at least in part. If its reliance was outcome-determinative, the agency may have to expand what it produces as a record by producing more documents that the plaintiff sought.
If a court accepts such a challenge, DOJ might still be able to issue rules governing the compilation of administrative records by the Department’s own non-litigating components: the Bureau of Alcohol, Tobacco, Firearms, and Explosives; the Federal Bureau of Investigation; the Federal Bureau of Prisons; the Drug Enforcement Administration; and the Executive Office for Immigration Review, to name a few. Each of those bureaus and offices is subject to APA challenges and must put forward an administrative record. Even guidance that is applicable just to them would be beneficial.
Another provision of the APA that is ripe for Attorney General regulation is the first sentence of 5 U.S.C. § 705. The two-sentence section entitled “Relief Pending Review” has stated the following since 1966:
When an agency finds that justice so requires, it may postpone the effective date of action taken by it, pending judicial review. On such conditions as may be required and to the extent necessary to prevent irreparable injury, the reviewing court, including the court to which a case may be taken on appeal from or on application for certiorari or other writ to a reviewing court, may issue all necessary and appropriate process to postpone the effective date of an agency action or to preserve status or rights pending conclusion of the review proceedings.
As with the “whole record” provision in 5 U.S.C. § 706, there do not appear to be any regulations for § 705, the statutory text is ambiguous, and the courts diverge on its interpretation.
According to the Attorney General’s Manual, the first sentence of § 705 is “a restatement of existing law.” The legislative history of the APA corroborates this reading in that the first sentence grants administrative authority to issue a stay of its actions. The authority in this sentence “may be used only to suspend a rule before it is effective, not after it is effective but before the compliance deadline.” The significant question for the first sentence is when an agency can or should find that “justice so requires” the postponement of agency action when it is facing judicial review. The term “justice” usually entails an agency providing some basis or justification for postponing the effective date. Reasons that agencies have offered include the desire to “administer[] a nationwide program in a uniform fashion,” avoiding the confusion and disruption that could be created if the agency implemented a decision that was subsequently overturned, and allowing the agency time to “mount an appropriate defense of the rule.” “[T]his provision requires agencies to make some showing that a suspension is necessary to enable judicial review over the original rule to proceed in a ‘just’ manner.” A failure to do so can result in a court rejecting the postponement.
Analogous words, “[w]hen an agency finds that justice so requires,” are found in the Federal Rules of Civil Procedure. For instance, Rule 15(a)(2) states that federal district courts “should freely give leave [to amend pleadings] when justice so requires” if a party fails to do so as per Rule 15(a)(1). This is a liberal standard with a low threshold. Courts should grant pleadings amendments unless there would be undue prejudice, undue delay, or representative of affirmative bad faith involved. A court’s denial of such a motion to amend without any apparent justification would be seen as an abuse of discretion. A similar standard applies to relief from a final judgment or order. The court must act on “just terms.”
Extending these rationales to an agency’s postponement of a final rule under § 705, “when justice so requires,” would be a low standard that favors agency decisionmaking. In short, “when justice so requires” is likely a reviewable standard, but if an agency postpones a final rule with justification, it would be difficult for a litigant to refute such agency action in the absence of bad faith, substantial prejudice, or undue delay. Still, a DOJ rule expounding the justice standard could help bring order to a relatively untread area of law, given the rapid acceleration of § 705 cases discussed earlier. This would be particularly helpful to new administrations to execute policy changes in the first months of presidential transmissions.
The Supreme Court has only cited 5 U.S.C. § 705 in three cases. In Abbott Laboratories v. Gardner, the Court held that if the agency believes issuance of a stay would “significantly impede enforcement [of the final agency action] or [would] harm the public interest,” it does not need to delay enforcement of a regulation and can oppose any motion for a judicial stay. If an agency opposes a judicial stay, it would “scarcely . . . be doubted that a court would refuse to postpone the effective date of an agency action if the Government could show . . . that delay would be detrimental to the public health or safety.” Therefore, if the government opposes the judicial stay, the burden is on the opposing party to show the stay will not be harmful to the public interest.
Section 705 is rarely addressed by the courts of appeals. A First Circuit case referenced § 705, but it was not directly implicated in the proceedings. A Third Circuit case clarified that judicial review cannot commence until the agency has acted on a petition for rulemaking. An agency acting does not include examining requests and seeking additional information from petitioners. Rather, such actions do not make a claim ripe for judicial review until determinations on such requests are final.
The second sentence of § 705 confers upon the reviewing court discretionary functions to stay agency action “to the extent necessary to prevent irreparable injury.” Section 705 grants powers to a court, not an agency, as in the first sentence. The stay power is also more limited. The Attorney General’s Manual suggested that the reviewing court is not given plenary power, for example, the “power to order interim payment of grants or benefits the denial of which is the subject of review.” Attorney General Tom C. Clark further qualified an irreparable injury as the “historical condition of equity jurisdiction,” and stated that the “[m]ere maintenance of the status quo for the convenience of parties pending judicial review of agency action will not be adequate ground for the exercise of this stay power.” This distinction is relevant and supported by congressional intent because when the bill that eventually became the APA was introduced to the Senate, it read: “to the extent necessary to . . . afford an opportunity for judicial review of any question of law or prevent irreparable injury.” Attorney General Clark noted that “upon such conditions as may be required” is a “balance [of] equities” for the reviewing court to weigh. Factors to weigh such considerations include whether postponing agency action will adversely affect parties, even those not present, and parties seeking postponement may need to, “furnish security to protect such other persons from loss resulting from postponement.” In short, the scope for a judicial stay of agency action is limited and tethered to equitable grounds, such as preventing an irreparable injury. The Attorney General noted the general procedural provisions governing preliminary injunctions, restraining orders, issuances of interlocutory injunctions, and temporary stays would likely be applicable in these court proceedings. But to highlight the difference between the first and second sentences, the latter is implicated when a challenger to agency action claims an “irreparable injury” and usually seeks preliminary injunctive relief. At that point, the agency likely has not invoked the first sentence and postponed the final rule. By nature, this relief is not affirmative, but rather temporary. However, some statutes, such as the Magnuson-Stevens Act, explicitly preclude a reviewing court from granting preliminary injunctive relief in an action challenging a promulgated regulation. This effectively precludes Sentence 2 judicial action.
DOJ could issue a rule implementing the first sentence of § 705. Title 28 U.S.C. § 516 would likely provide support so long as the rule pertains to the conduct of litigation. Sentence 1 postponements necessarily pertain to the conduct of litigation and possibly the antecedent stage of securing evidence because an agency cannot enact a postponement unless the agency action is “pending judicial review.”
The regulation could establish standards for when an agency may postpone a rule or adjudication before its effective date. It could require the agency to, for example, submit a statement to DOJ identifying the validity of its reasoning, the consistency between the postponement and the reasons previously given for undertaking the agency action, and when, if ever, the agency expects to resume the action. The purpose of this information would be to place the regulated public in the best position to discern agency behavior and for the agency to have the best tools with which to prevail.
Alternatively, DOJ could establish standards for when it will make an argument before the courts. This kind of rule would be deeply supported under the conduct-of-litigation statute (§ 516) or the Housekeeping Statute (§ 301). A useful analogue might be the state secrets guidance and issuance procedures that the Attorney General laid out early in the Obama Administration. Then-Attorney General Eric Holder took a substantive standard—the invocation of a significant privilege which results in the covered information being completely excluded from the litigation—and imposed limits on its use, applicable to all agencies party to or interested in the litigation. The new policy required agencies to make a showing to DOJ that the legal standard for invoking the privilege was met. The agencies must narrowly tailor their requests, perhaps more restrictively than the outer limits of the privilege. The Attorney General imposed substantive limits—a refusal to defend an invocation of the privilege to “restrain competition” or to “conceal . . . administrative error.” Finally, the policy imposed significant process requirements, including the creation of a state secrets review committee and the requirement for personal Attorney General approval and periodic reporting to Congress. The memorandum does not even state its implementing authority, but § 301 and § 516 are plausible contenders if the policy were ever challenged in court as ultra vires. Similar to the state secrets policy, so too could the Attorney General impose limits on how and when agencies may undertake § 705 postponements. Analogous procedural and reporting requirements would seem well within DOJ’s § 301 and § 516 wheelhouse, even putting aside the substantive limitations on when the Attorney General will defend a state secrets privilege invocation.
There are currently no regulations implementing the APA. We think that is not by accident. We have identified two APA provisions as being amenable to rulemaking, but we believe that many other provisions of the APA are not amenable to Attorney General regulation.
A few subsections might, however, tempt DOJ’s policy apparatus or the Office of Management and Budget. The Executive Branch might consider establishing discrete procedures for the issuance of administrative subpoenas and warrants under APA § 555(c) and (d). There are certainly good policy reasons for regulating this section. This would provide the first meaningful regulation of agency investigations, which would protect the civil liberties of investigative targets and reduce the risk of judicial correction of administrative investigative excesses by stabilizing the subject matter within intentionally designed, reasonable constraints. Sections 701(a)(1) and (2) make clear that agency actions are immune to judicial review when “statutes preclude judicial review” or the action “is committed to agency discretion by law.” A regulation that expounds the scope of preclusion or agency discretion would be very valuable to the arm of the government defending most agencies in litigation. So too, DOJ might wonder whether it could regulate the somewhat hazy categories of military or foreign affairs functions, which are exempt from notice-and-comment rulemaking under § 553(a)(1).
These examples likely cross the line. For one, regulations of these APA provisions are more substantive in character than the provisions we have discussed unless they are carefully drawn to be procedurally focused. For another, these kinds of regulations do not squarely arise in the conduct of litigation unless the Attorney General expressly focuses on litigating positions DOJ attorneys are authorized to take. The record rule in § 706 is about evidence in litigation, which squarely ties to the Attorney General’s evidence-securing power in 28 U.S.C. § 516. The stay mechanisms of § 705 can only occur immediately before or during litigation, which is also within the Attorney General’s purview. These jurisdictional and notice-and-comment provisions lie out of reach for DOJ but help demonstrate the limiting principle that we discern from the APA’s intersection with the Housekeeping Statute and § 516.
The APA is steadily marching toward the century mark of its original enactment. It is not likely to be amended any time soon. The statute remains the best mechanism to date to achieve statutory optimization and promote public policy in a more honest and accountable government. Congress’s renewed attention in this realm would be beneficial.
Absent legislation, APA rulemaking in a few limited areas is a valuable additional tool of APA elucidation on top of the case method. It provides a more active and efficient model of establishing litigation procedure notice, consistency, legitimacy, and sophistication. DOJ is uniquely situated to enact these rules. Although notice-and-comment regulation may be possible and defensible in an Article III court, DOJ may find it optimal to issue subregulatory guidance to other agencies or articulate the conditions in which it will not defend an administrative record or contest a pre-enforcement stay. The costs are low relative to the benefits, but the sine qua non is having the political will to act and the prioritization to do so.
The authors thank Azniv Nalbandian and Anthony Licata for their research support, as well as Erin Blondel, Sam Bray, Greg Dickinson, Alan Morrison, and Dick Pierce for their helpful comments.
