As Far As Reasonably Practicable: Reimagining the Role of Congress in Agency Rulemaking

Note from the Editor: 

The Federalist Society takes no positions on particular legal and public policy matters. Any expressions of opinion are those of the author. We invite responses from our readers. To join the debate, please email us at [email protected]. 

Mr. Jayne’s article won First Place in the Article I Initiative Writing Contest, conducted by the Federalist Society’s Article I Initiative on the topic The Nondelegation Doctrine: Intelligible Principle or Unworkable Standard? Prof. Lillian BeVier, Hon. C. Boyden Gray, and Hon. Chris DeMuth were the esteemed judges for the Writing Contest. They completed a blind review of the submitted essays addressing the contest topic, and they selected this paper as the winner.

Abstract: 

This paper argues that the nondelegation doctrine is in need of resuscitation. It argues for adoption of a new “as far as reasonably practicable” standard, first articulated in the lesser-known case of Buttfield v. Stranahan, and for effectuating that standard with the application of statutory construction principles like the major questions doctrine to issues of nondelegation. The practical effects of this approach would be a judiciary more faithfully policing the constitutional separation of powers and spurring Congress to govern more responsibly. With the assistance of a revamped CBO, and informed by the examples of British Columbia and Idaho, Congress should take a greater role in generating regulations by establishing legislative impact accounting of proposed bills, institutionalizing the Congressional Review Act, and implementing the twin reforms of regulatory budgeting and retrospective review.

I. The Nondelegation Doctrine Needs to Be Resuscitated

The American experiment is predicated on the idea of a social contract, the notion that citizens are governed by consent that they can revoke and by representatives that they can hold accountable. Failure to faithfully enforce the nondelegation doctrine—the doctrine that the Constitution places limits on Congress’ authority to transfer its lawmaking powers to administrative agencies—deprives the citizen of both means of participating in government, because the regulator neither needs consent nor must give an account. Perhaps more importantly, threats to liberty abound when the power to define, enforce, and interpret the law accrue in one branch or department. That is why the most salient arguments against the current iteration of the nondelegation doctrine are constitutional, and why the separation of powers, far from being an anachronism, remains integral to “the system of government ordained by the Constitution.”[1]

Hence, the vesting clauses of the first three articles point to a tripartite framework with an exclusive role for each branch.[2] “All legislative Powers herein granted shall be vested in a Congress of the United States.”[3] The Necessary and Proper Clause implies a limit on the content of the laws that Congress can pass.[4] It is not enough for laws to be “convenient, or useful, or essential to another.”[5] The conjunction “and” implies that in addition to being necessary, they must be appropriate in allocating authority with respect to separation of powers principles (as well as consistent with federalism and individual liberty).[6] The Take Care Clause implies a reciprocal duty for the executive: that it must carry out the will of the legislature and not exercise its own prerogative.[7]

Aside from the constitutional perils, there are prudential reasons for revisiting the nondelegation doctrine as currently applied. Since 1935, general nonenforcement of the nondelegation doctrine under the intelligible principle standard has coincided with a shift in the locus of policymaking from Congress to government agencies. During the 2018 calendar year alone, Congress enacted 313 laws, but agencies issued 3,368 rules—a 1 to 11 ratio.[8] This shift seems in keeping with the ideal regime championed by modern administrative state architect James Landis, who saw the growth of the so-called Fourth Branch as both inevitable and desirable.[9] “The administrative process springs from the inadequacy of a simple tripartite form of government to deal with modern problems,” he wrote.[10]

But while agency officials often possess greater technical expertise than elected representatives, Article I establishes a finely wrought process to refine policy while maintaining its legitimacy as the product of representative government. This process brings together more than 500 senators and representatives, chosen from different constituencies, to shape the final outcome of what binds the public.[11] While this “sausage-making” often results in tradeoffs and compromises, it frequently ensures that multiple perspectives are considered and the worst proposals jettisoned from the resulting legislation. Agencies lack such a honing process. The resulting rules are often ill-conceived and ill-considered, popularly coined “red tape.” Regulators are rewarded for issuing new rules, rather than for effectively managing the interrelationship of an agency’s entire portfolio of existing rules.[12] Hence, they are rarely held responsible when their good intentions do not translate into good outcomes. Also troubling, agencies reach for outdated congressional delegations of power as a source of authority to pass rules that Congress never considered or would never support today.[13]

Regulations tend to accumulate, as they are added to, but seldom removed from, a growing stockpile of often duplicative, burdensome, or outdated rules. It would take someone three years, 108 days, four hours, and five minutes to read through the Code of Federal Regulations (CFR) at a rate of 250 words per minute for 40 hours a week.[14] Of course the task would be a Sisyphean one, as the Code is constantly in flux with the regular churn of the administrative state. Not only is it hard for Congress to police this ballooning code, it is difficult for the average citizen or small business to avoid running afoul of some arcane rule. When so many citizens become unwitting lawbreakers, institutional faith, trust, and respect suffer. The rule of law is compromised, threatening the legitimacy of the American experiment. As Madison cautioned, “It will be of little avail to the people, that the laws are made by men of their own choice, if the laws be so voluminous that they cannot be read, or so incoherent that they cannot be understood.”[15]

Such a state of affairs weakens faith in political efficacy. Voters, rather than seeing their ballot-box choices reflected in policy, increasingly feel subject to the whims of faceless, unaccountable bureaucrats. “[T]he citizen confronting thousands of pages of regulations—promulgated by an agency directed by Congress to regulate, say, ‘in the public interest’—can perhaps be excused for thinking that it is the agency really doing the legislating.”[16]

This phenomenon of regulatory accumulation has important implications for the economy as well. Economists Michael Mandel and Diana Carew liken it to dropping pebbles in a stream.[17] One pebble or regulation is insignificant, but too many pebbles can dam a stream, and too many regulations can slow down an economy.[18] One study found that if regulations had been held constant at their 1980 levels, the economy would have been 25 percent larger in 2012 than it actually turned out to be, or $4 trillion larger, an average of $13,000 more in the pocket of every American.[19]

Congress will not fix the problem on its own. Its incentives are to pass general pronouncements of laudable goals but leave the tough tradeoffs to the executive branch, which it can then blame when implementation falls short of its ideal. In a study of four regulatory reform statutes that became law, Stuart Shapiro and Diana Moran found that all failed to reduce regulatory burdens.[20] In order to secure passage, they had to be watered down to the point of being mostly ineffectual, but they allowed policymakers to campaign on their adoption.[21] Even if the Regulatory Accountability Act[22] and the Regulations in Need of Scrutiny (REINS) Act[23] had received a floor vote in today’s politically fractious climate, they would likely have been gutted of any meaningful reform.

Courts have been reluctant to second-guess agencies. The Supreme Court has largely accepted the view of Landis: “Our jurisprudence has been driven by a practical understanding that in our increasingly complex society, replete with ever changing and more technical problems, Congress simply cannot do its job absent an ability to delegate power under broad general directives.”[24] Even Justice Antonin Scalia, exponent extraordinaire of the separation of powers, put it thus: “In short, we have ‘almost never felt qualified to second-guess Congress regarding the permissible degree of policy judgment that can be left to those executing or applying the law.’”[25] The idea that it would be impracticable and nonsensical to expect Congress to make all of the implementation decisions on its own informs a central rationale of the so-called “intelligible principle” standard as established in J.W. Hampton v. United States.[26] The intelligible principle standard has become the Court’s test for whether a given delegation is lawful. Congress can delegate quasi-legislative power to agencies or officials, so long as it gives them an intelligible principle to guide their discretion.[27] The practical effect of the standard is that courts have avoided placing any real limits on what Congress can assign to agencies.

But Alexander Hamilton rightly admonished that judges must do their duty as “faithful guardians of the Constitution.”[28] This means that courts must step into the breach. The Court recognizes that duty, and it has hence repeatedly reaffirmed the existence of a limit on congressional delegation and discoursed on the importance of such a limit.

Indeed, our nation’s foremost jurists have expressed concern about delegation. Chief Justice John Marshall is credited with first giving judicial expression—in Wayman v. Southard—to the doctrine that Congress cannot delegate “exclusively legislative” functions and must decide the “important subjects” if it assigns others to “fill up the details.”[29] Four years later, in Field v. Clark, the Court provided additional guidance when it defined a category of cases in which the nondelegation doctrine is not implicated: when Congress directs the executive to take certain actions upon a contingent event or the latter’s ascertainment of particular facts.[30] In Field, the Court upheld a grant of authority to the president to suspend congressionally prescribed tariff rates with countries he determined had imposed unequal and unreasonable duties on American shipping.[31] Still, the Court maintained, “That congress cannot delegate legislative power to the president is a principle universally recognized as vital to the integrity and maintenance of the system of government ordained by the constitution.”[32] 

In 1980, then-Justice William Rehnquist quoted this latter statement from Field in calling for resuscitation of the nondelegation doctrine.[33] In a concurrence in a case concerning a delegation of authority to the Labor Secretary to set the allowable level of benzene exposure in the workplace, he outlined the contours of a new standard consistent with Justice Marshall’s exposition in Wayman: “The most that may be asked under the separation-of-powers doctrine is that Congress lay down the general policy and standards that animate the law, leaving the agency to refine those standards, ‘fill in the blanks,’ or apply the standards to particular cases.”[34] He added, “It is the hard choices, and not the filling in of the blanks, which must be made by the elected representatives of the people.”[35]

Justice Clarence Thomas has carried the banner in the years since.[36] In a concurrence tracing the nondelegation doctrine’s rationale from Greek and Roman times, through English history, to J.W. Hampton,[37] he questioned the soundness of the intelligible principle standard before counseling a test more consistent with Justice Marshall’s criteria in Wayman, namely that Congress could not delegate “exclusively legislative” functions.[38] Justice Thomas quoted Professor David Schoenbrod at length for the proposition that what implicates the doctrine is not the degree or quantity of authority that is conferred, but its nature or quality.[39] Schoenbrod distinguishes between “rules statutes,” which define the parameters of allowable conduct, and “goals statutes,” which state only objectives; when Congress passes goals statutes and asks agencies to determine how to achieve those objectives, it impermissibly delegates legislative power to agencies.[40] In other words, the difference between Justice Marshall’s descriptions of “important subjects” and “fill[ing] up the details” is not about big picture versus nitty-gritty. It is about making law versus determining how to implement it. When it makes law, the government regulates private conduct; when it determines how to implement that law, it regulates itself.[41] Schoenbrod describes his “rules statute/goal statute distinction” as “fundamentally different” from the intelligible principle standard because it is rigidly formalistic in prohibiting all delegations of legislative power.[42] Justice Thomas seemed to endorse Schoenbrod’s test when he wrote that “[g]overnment may create generally applicable rules of private conduct only through the proper exercise of legislative power.”[43]

In 2019, Justice Thomas and Chief Justice John Roberts joined Justice Neil Gorsuch in a sharp dissent on nondelegation grounds.[44] In a case considering the scope of the U.S. Attorney General’s authority to determine the applicability of a statute to offenders convicted before its enactment, Justice Gorsuch wrote a dissent offering what might be considered an alternative test to the one Justice Thomas has endorsed.[45] First, he said, Congress may delegate gap-filling duties (a reiteration of Wayman’s statement of the doctrine), but it must make the policy governing private conduct (a clarification of Wayman’s “exclusively legislative” duties and “important subjects”).[46] “Second, once Congress prescribes the rule governing private conduct, it may make the application of that rule depend on executive fact-finding” (a summation of the Field category of cases exempted under the nondelegation doctrine).[47] “Third, Congress may assign the executive and judicial branches certain non-legislative responsibilities.”[48] He further argued that the intelligible principle test has been misunderstood. An intelligible principle must “assign to the executive only the responsibility to make factual findings,” it must “set forth the facts that the executive must consider and the criteria against which to measure them,” “[a]nd most importantly,” Congress must make the policy judgments.[49]

A revival of the nondelegation doctrine now appears imminent. Recently, Justice Brett Kavanaugh cited the opinions of then-Justice Rehnquist and Justice Gorsuch discussed above.[50] He issued a statement respecting the denial of certiorari in a case because he said it raised an identical statutory interpretation issue that had already been decided in Gundy.[51] But he wrote separately to signal, like Justice Samuel Alito did in his Gundy concurrence,[52] that he would be open to revisiting the doctrine.[53] In summarizing then-Justice Rehnquist, he wrote that Congress must make the “major policy decisions with the president through the legislative process, and not through delegation to agencies”[54] Justice Kavanaugh referred to a “nondelegation doctrine for major questions” that could provide additional guidance for a new standard or test.[55]

This paper attempts to synthesize a new standard from the criteria offered by Justices Rehnquist, Thomas, Gorsuch, and Kavanaugh—one that could get five votes. It briefly considers what effect this new standard would have on lower courts and Congress. It then recommends that Congress implement significant institutional reforms to make a revived nondelegation doctrine workable.

II. A Revived Doctrine Needs a New Standard

The nondelegation doctrine needs resuscitation, but the intelligible principle standard is a dead letter. It should be discarded and replaced with a test that is more limiting and more readily administrable. An intelligible principle is a low bar, but Congress still manages to limbo right under it by passing vague generalities. The Court has upheld broad delegations with weak intelligible principles such as the FCC’s authority to grant broadcast licenses “if public convenience, interest, or necessity will be served thereby”;[56] the SEC’s authority to determine whether a holding company’s organization “does not unduly or unnecessarily complicate the structure, or unfairly or inequitably distribute voting power among security holders”;[57] and the wartime Office of Price Administration’s authority to fix “fair and equitable” commodity prices.”[58] The Court “consistently finds intelligible principles where less discerning readers find gibberish.”[59]

Since 1935, the Court has never struck down a statute for failing to articulate an intelligible principle. The test is difficult to enforce and administer. It offers meager guidance for courts, as it makes no distinction among the nature or degree of delegated authority.[60] Justice Thomas opined, “I believe that there are cases in which the principle is intelligible and yet the significance of the delegated decision is simply too great for the decision to be called anything other than ‘legislative.’”[61]

Commentators have argued that the intelligible principle test was never intended to be interpreted so broadly. Justice Thomas noted in another case that the intelligible principle test was formulated in a time when most of the delegations challenged before courts concerned conditional or contingent legislation.[62] These were laws in which Congress made the rules and the conditions under which the rules would be triggered or suspended, and then left to the executive only the duty of determining whether those conditions had taken effect.[63] Examples include delegations to the president to adjust tariff rates,[64] lift embargos,[65] and ban importation of inferior tea.[66] Many of these delegations concerned inherent Article II functions, warranting greater deference given the president’s role as the “sole organ” in foreign affairs.[67]

Justice Gorsuch has also questioned the status of J.W. Hampton as a seminal case:

No one at the time thought the phrase meant to effect some revolution in this Court’s understanding of the Constitution. . . . And when Chief Justice Taft wrote of an “intelligible principle,” it seems plain enough that he sought only to explain the operation of these traditional tests; he gave no hint of a wish to overrule or revise them.[68]

He went on to surmise that “the Court’s reference to an ‘intelligible principle’ was just another way to describe the traditional rule that Congress may leave the executive the responsibility to find facts and fill up details.”[69] Whatever its place in administrative law jurisprudence, the intelligible principle standard has failed to demarcate any limits on delegation or declare what the law is.

Yet the Court has already articulated a suitable alternative standard in its line of nondelegation decisions, tucked away in the overlooked 1904 case of Buttfield v. Stranahan.[70] In Buttfield, the Court upheld a delegation of authority to the Secretary of the Treasury to ban importation of “impure” and “unwholesome” tea.[71] It held this case fell within the Field v. Clark fact-finding, contingent exception to the nondelegation doctrine because the statute at issue “fix[ed] a primary standard” for the Secretary to follow and gave that official the “mere executive duty to effectuate the legislative policy declared in the statute.”[72] In concluding a discussion of the petitioner’s nondelegation challenge, then-Justice Edward Douglass White added, “Congress legislated on the subject as far as was reasonably practicable.”[73] This could be read either as controlling precedent or as dicta. Either Congress is required to legislate as far as reasonably practical before it delegates any authority to the executive branch,[74] or, in this particular case, the Court made an additional observation that Congress had gone as far as it realistically could in designing the statutory scheme.

Regardless, this language suggests what could become a new standard or test by which to apply the nondelegation doctrine. A court faced with a challenge to a congressional delegation would determine whether Congress had legislated as far as reasonably practicable, leaving to agencies some gap-filling discretion that Congress would be unable to effectively exercise on its own. Administration of this standard could be assisted by several existing interpretive canons that are already frequently applied in nondelegation contexts, particularly the major questions doctrine.

An “as far as reasonably practicable” test would have several advantages over the existing intelligible principle standard. It would be more limiting and more administrable. As discussed below, when informed by a series of nondelegation canons, it would give reviewing courts better guidance than what the intelligible principle standard provides. Drawn from Buttfield, a progeny of Field, it avoids undoing a century of precedent. It is an incremental step in the right direction, staking a moderate position that is likely to garner at least five votes. For critics of a runaway administrative state, it would require Congress to stop passing the buck and pass legislation with greater specificity. On the other hand, for those worried that Congress is unable to discharge its duties in an increasingly complex society—a concern raised in Mistretta[75]—it acknowledges a gap-filling role for agencies in a modern technocracy. What is not “practicable” for Congress, whether because it is too detailed or too technical, can be assigned to subject matter experts in the branch tasked with enforcing the laws.

Two cases illustrate how the “as far as reasonably practicable” test could work. First, reviewing courts should be skeptical that Congress has legislated on a subject “as far as reasonably practicable” when an agency relies on a very old statute for a new grant of power or a novel interpretation of authority. For example, when Congress passed the Communications Act of 1934,[76] it established the FCC to regulate public use of the broadcast frequencies within the electromagnetic spectrum.[77] Congress could not have foreseen the development of satellite technology and dish receivers,[78] let alone cable television or the internet. Yet the FCC relied on the 1934 law to justify applying common-carrier regulations to internet service providers.[79] Under an “as far as reasonably practicable” standard, the FCC would be purporting to exercise an unlawful delegation of authority in doing so. Congress did not legislate far enough into this field—as far as reasonably practicable—for the FCC to promulgate its net neutrality rule.[80]

By contrast, Touby v. United States may be a case in which an as far as reasonably practicable standard would permit Congress to delegate gap-filling authority that is consistent with the nondelegation doctrine.[81] In Touby, the Court considered a provision of the Controlled Substances Act that allowed the Attorney General to temporarily add a controlled substance to a list of prohibited drugs if he determined it necessary to avoid threats to public safety.[82] To do so, he had to follow specified procedures and engage in fact-finding by evaluating a substance with reference to its history and current pattern of abuse; the scope, duration, and significance of its abuse; and what, if any, risk it posed to public health.[83] Here, Congress established the general policy and standards for the authority it was delegating and outlined the facts that needed to be ascertained before the Attorney General could add a drug to the list of prohibited substances. It would have been impracticable for Congress to withhold this authority, because new designer drugs were regularly hitting the streets before the normal drug scheduling process could make them illegal. While stopping short of endorsing the unanimous decision in Touby, Justice Gorsuch cited it in his Gundy dissent as a case pointing “in the direction of the right questions.”[84]

The phrase “as far as reasonably practicable” has been invoked in several areas of the law, such as the advisability of executing a search warrant in the daytime,[85] desegregation considerations in planning the construction of new schools,[86] and the standard of care in monitoring freight train wheels while in transit.[87] If not a universal term, it is a generally understood one. The concept is also adaptable enough to allow courts to apply it in different factual circumstances. The question is how to apply it to nondelegation, to make a standard or test flexible enough to apply to different facts and cases, but firm enough to limit judges’ discretion to principled decision-making. There is some tension in Chief Justice Marshall’s pronouncement of the doctrine, as his descriptions of “important subjects” and “fill up the details” could be read as delineating a matter of degree, while his term “exclusively legislative” appears to be a black-and-white, categorical definition.[88] Then-Justice Rehnquist channels Chief Justice Marshall when he emphasizes that Congress must “lay down the general policy” and make the “hard choices,” “leaving the agenc[ies]” only to “fill in the blanks.”[89] In summarizing then-Justice Rehnquist, Justice Kavanaugh wrote that Congress must make the “major policy decisions.”[90] An “as far as reasonably practicable” standard could determine the lawfulness of any delegation based on how much authority Congress hands over—whether it makes the “major policy decisions” or “general policy” when it designs the statutory scheme—so it expects more of Congress than under the Court’s current test and is therefore a step in the right direction.

But the standard could also determine whether a delegation is lawful based on the kind of authority that is handed over, as the term “exclusively legislative” implies. In quoting Professor Schoenbrod and seemingly endorsing his “rules statute/goals statute distinction,” Justice Thomas seems to favor a strict categorical approach whereby any legislative power—including the ability to make authoritative interpretations of laws—left to agencies is an unlawful delegation. Though as Adam White notes, by joining Justice Gorsuch’s Gundy dissent, Justice Thomas may be signaling that he is amenable to a more modest approach.[91] Justice Gorsuch parallels Chief Justice Marshall and then-Justice Rehnquist when he writes that Congress “may always authorize executive branch officials to fill in even a large number of details,”[92] but he appears to gesture at a categorical approach too. He does not speak of “general policy” or “major policy” but only of “policy.”[93] He writes that Congress must make the policy judgments and leave to the executive “only the responsibility to make factual findings.”[94] Such a rule predates J.W. Hampton and its introduction of the intelligible principle standard. Adopting it could signal a return to the Field approach, a general policy of nondelegation with a categorical exception for executive fact-finding. If viewed in this light, the “as far as reasonably practicable” standard is merely a faithful application of Field’s fact-finding nondelegation exception. Recall that the Buttfield Court explicitly applied Field’s fact-finding principle in its decision.[95] With Field in view, “as far as reasonably practicable” could be applied as prohibiting the delegation of any legislative power. An “as far as reasonably practicable” standard would press Congress to settle the primary policy questions and define clear standards to cabin agencies’ discretion—certainly more than the extant intelligible principle standard does.

Whether it turns on the amount of delegated authority or on a categorical classification, this standard would likely rely on other established doctrines to make it work. It can be informed by what essentially is the current nondelegation standard or test: a series of statutory construction canons. Practically, the existing nondelegation doctrine does not so much limit the laws that Congress can pass as it limits the way agencies can construe statutes.[96] An “as far as reasonably practicable” test would be an umbrella standard encompassing a series of nondelegation canons, the most important of which would be the major questions doctrine. Justice Kavanaugh invoked this doctrine in his recent statement inviting the Court to revisit the nondelegation doctrine, and he appears inclined to see it incorporated within any new standard.[97]

The term “Major Questions Doctrine” comes from an article by Justice Stephen Breyer in which he discusses the degree of deference that courts should give to how agencies interpret their governing statutes and make their rules.[98] Justice Breyer wrote that courts should assume Congress has considered and decided the major questions in a statute and should therefore accord agencies less deference on major questions than on “interstitial matters.”[99] The Court cited Justice Breyer’s article when it held the FDA lacked the authority to regulate tobacco products.[100] It has invoked the major questions doctrine or its rationale on several occasions since.[101]

Justice Gorsuch described the major questions doctrine in Gundy as a sort of workaround to the nondelegation doctrine and its intelligible principle standard: “Although it is nominally a canon of statutory construction, we apply the major questions doctrine in service of the constitutional rule that Congress may not divest itself of its legislative power by transferring that power to an executive agency.”[102] As Adam Gustafson has proposed, “Although the major questions doctrine began as an exception to Chevron deference, it can operate more broadly as a nondelegation canon of statutory construction.”[103] Echoing other administrative law experts, Gustafson suggests using the Executive Order 12866[104] definition of a “significant regulatory action”—agency actions that would have an annual impact on the economy of $100 million or more—as an administrable standard for determining when a major question is presented.[105] This $100 million threshold is used elsewhere, such as in the Regulatory Flexibility Act.[106] In other words, if an agency proposes a regulation that is “major” enough that the Office of Information and Regulatory Affairs (OIRA) is tasked with reviewing it, then it should be able to show statutory language evincing congressional authorization for such a rule. Applying an “as far as reasonably practicable” standard suggests that a policy judgment with a $100 million price tag requires Congress to exercise a requisite degree of decision-making. Charging an agency with a mere intelligible principle should not suffice.

Other nondelegation canons could include the doctrine of avoidance, whereby a court would construe an ambiguous statute narrowly to avoid raising separation of powers problems; lesser deference for agencies’ novel uses of older, more established terms in statutes; greater deference for interpretations of broader, more general terms; and more leeway for delegations of highly technical decision-making.[107] Though imperfect, an “as far as reasonably practicable” standard, buttressed by continued application of the major questions doctrine, is an incremental step in the right direction.

The practical effect of a revived nondelegation doctrine, guided by an “as far as reasonably practicable” standard and applied through a series of nondelegation canons, would be a more hands-on judiciary, a more responsible Congress, and a more fettered administrative state. These changes would not happen overnight but would gradually take effect as institutional incentives were realigned. Applying the new approach, courts would put Congress and agencies on notice that the days of broad delegations are over. Agencies would see many of their lawmaking efforts frustrated and turn to Congress for clear direction. A Congress forced to take more responsibility for the everyday requirements and restrictions that bind its citizens would be more accountable to the public. And as the public became aware of this growing accountability, it might spur Congress to become even more involved in agency rulemaking.

III. Congress Should Prepare for a More Active Role

Congress should prepare to take a more active role in generating regulations at all stages of the rulemaking process. As the Senate does for treaties and appointments, the whole Congress should do for significant regulations: advise and consent. It should vote—at a minimum—on all significant regulations.

In order to do so effectively, Congress must hire considerably more staff. It should begin anticipating, tracking, and analyzing regulations at the bill-drafting stage and conduct independent, ongoing analyses of agency regulatory actions. Congress should limit agency rulemaking by implementing a regulatory budget, and it should establish a process for periodically reviewing the CFR for rules that reflect excessive delegation, do not justify their costs, or are otherwise unlawful or imprudent.

A. Additional Congressional Staff

First, Congress would need to authorize and appropriate funding for more staff. In order to legislate on subjects “as far as reasonably practicable,” it needs expanded resources to further develop policies at the drafting stage, to conduct more effective oversight of the executive by more closely scrutinizing its proposed and finalized rules, and to conduct retrospective review of existing regulations. Naturally, this might entail larger committee staffs with additional subject-matter experts. But perhaps more importantly, Congress would need an institutional counterweight to the administration that it oversees, a rival to the Office of Management and Budget (OMB) and OIRA.[108] The natural place to house such an entity would be the Congressional Budget Office (CBO). CBO’s principal role is to forecast the effects of budget, tax, and spending policy.[109] It estimates the revenue and costs of proposed bills.[110] A revamped CBO could help Congress reassert its constitutional prerogative by providing reports on, estimates of, and recommendations about regulations. It is possible that either the General Accountability Office (GAO) or Congressional Research Service could perform a similar function, as all three operate under strict rules of nonpartisanship and objectivity.[111] This paper proposes CBO, but regardless of which entity is used, it is clear Congress needs more personnel. From 1975 to 2015, CBO, GAO, and the Congressional Research Service have seen their combined staffs shrink by 45 percent.[112]

B. Legislative Impact Accounting

Congress should be involved in generating regulations even before passing the enabling statutes that empower agencies to promulgate new rules. Scholars Jason Fichtner, Patrick McLaughlin, and Adam Michel propose that CBO be tasked with estimating a bill’s regulatory impact along with its effect on the federal budget.[113] Their system of “Legislative Impact Accounting” calls for scoring and tracking of regulations beginning with new bills.[114] An independent office like CBO would forecast the impact of proposed legislation, not just on the budget, but on the economy as a whole, estimating the likely regulatory effects on things like direct compliance costs, employment rates, technological disruptions, and future innovation.[115] They note, “The European Commission provides impact assessments on all legislation by the European Parliament.”[116]

In addition to economic forecasting, CBO could be charged with reviewing legislation to spot potential legal and constitutional delegation issues, providing a more holistic assessment of a bill’s legal consequences that goes beyond the focus of the individual members and committee staff who are its chief authors. A legal office within CBO or a similar entity could provide additional expertise, paying particular attention to circumstances in which Congress has not legislated as far as reasonably practicable or would need to decide major questions. Fichtner, McLaughlin, and Michel characterize their legislative impact accounting proposal as a continual feedback loop that conveys to Congress information about regulations and, by extension, their authorizing legislation.[117] It would begin with an assessment of proposed legislation prior to voting and continue with analysis of agencies’ regulatory actions.[118] Congress could then make better informed decisions about how to respond to agency behavior, particularly at budget time.[119] Such a feedback loop makes sense given the two additional recommendations discussed below.

C. Greater Oversight and Regular Use of the Congressional Review Act

Secondly, CBO should review the significant regulatory actions that OIRA includes in its semiannual Unified Agenda of Regulatory and Deregulatory Actions and prepare detailed analyses of them for members of Congress. Rather than relinquishing the responsibility for regulatory analysis to OIRA, CBO could be double-checking the executive branch’s work and providing regular advice to lawmakers as agencies carry out their legislative mandates. The good news is that Congress has an important tool at its disposal: the Congressional Review Act (CRA).[120] Passed in 1996, the CRA laid dormant for many years; it was used only once in 2001 to strike an unpopular ergonomics rule.[121] Yet in the wake of the 2016 election, President Trump signed 15 joint resolutions of disapproval passed by Congress under the CRA to nullify regulations issued by agencies in the final year of the Obama Administration.[122] Congress should institutionalize it as part of regular order.

The CRA requires a rule-issuing agency to submit a report and copy of the rule to GAO and both houses of Congress, which is then forwarded to the chairman and ranking member of the committees that have jurisdiction over the rule.[123] The report’s submission starts a 60-day clock during which Congress may initiate filibuster-proof, fast-track procedures to schedule a vote on whether to strike the rule.[124] If both chambers vote to strike it, they can submit a joint resolution to the president; if he signs it, or if Congress overrides his veto, the rule is quashed. In addition, if the resolution succeeds, the agency is forbidden from issuing a rule that is “substantially the same” unless Congress later takes action to empower it to do so.[125] This latter point is critical, because it can serve to deter agencies from passing so-called “midnight rules” in the final year of an outgoing administration. Such a tactic can backfire, as an incoming Congress can not only void the rule, but also prevent the agency from passing it or a substantially similar one in the future.

Paul Larkin argues that the CRA could apply to far more than regulations passed through notice-and-comment rulemaking.[126] A recent GAO opinion[127] suggests the CRA’s reach could extend to guidance documents, policy statements, and other sub-regulatory items.[128] If so, Congress would be able to vote on a wide swath of agency activity beyond notice-and-comment regulation. Additionally, because the 60-day clock on a rule does not start until Congress receives its report from the agency,[129] Larkin suggests that potentially thousands of rules that were never properly submitted to Congress could be reviewed under the CRA today.[130] This would allow Congress to begin retrospective review of regulations that have been in effect for some time. A reconstituted CBO could assist members in deciding what to prioritize.

While the CRA’s reach is limited because it must comport with Article I requirements of bicameralism and presentment, and most presidents would veto challenges to their own agencies’ regulations, the CRA holds promise for at least some improvement. Its regular use might even discipline agencies to pass better regulations.

D. Retrospective Review

Finally, Congress needs to periodically audit the federal corpus of regulation through a process of retrospective review. At first blush, this might not sound reasonably practicable. Given that there were 63,645 pages in the Federal Register and 185,434 pages in the CFR at the end of 2018,[131] such a review would require additional resources beyond what Congress currently commits to its oversight of government agencies. But Congress can draw encouragement from what the Trump administration, British Columbia, and Idaho have done to reexamine old rules that are on the books.

The Trump administration’s approach to executive branch rulemaking demonstrates how regulatory budgets and retrospective review work together. Pursuant to Executive Orders 13771[132] and 13777,[133] agencies are currently scrutinizing regulations as part of a regulatory budget and retrospective review. There is no reason Congress cannot bring these functions in house. Briefly, a regulatory budget is a cap on agency rulemaking.[134] It is an attempt to limit the total amount of regulation by placing the cost of regulation on the regulator.[135] Rather than merely making new rules, agencies must be “rule managers,” regulating within fixed limits such that each regulation entails a tradeoff. In order to issue a new regulation, an agency must make room for it within the amount of allowable regulation, often by rescinding an existing regulation.[136] The amount of regulation allowed within a given budget can be measured in different ways, and Executive Order 13771 uses two metrics within each executive branch agency: a cap on the total cost of an agency’s regulatory burden to the economy, and a 2-for-1 requirement that each proposed rule be offset by identification of two existing rules for elimination.[137] Aside from limiting the amount of new regulation, a regulatory budget provides the incentive for agencies to conduct retrospective review of their existing regulations: In order to make new rules, they must find offsets by digging through their stockpile to select rules they are willing to part with. Executive Order 13777 establishes the contours of this retrospective review by requiring the designation of “regulatory reform officers” to lead “regulatory reform task forces” in their implementation of the Executive Order 13771 regulatory budget. Under this charge, agency lawyers and economists are auditing the rules on their books, and according to several observers, they have helped slow the growth of regulation.[138]

Though an improvement, these executive orders will be effective only so long as a president chooses to keep them in place. Congress should make them permanent by institutionalizing regulatory budgeting and retrospective review, and by putting CBO in charge of monitoring compliance. There are at least three ways it could do this. First, Congress could simply pass the substance of Executive Orders 13771 and 13777 into law. It could then, perhaps through the appropriations process, annually set caps for the regulatory costs that agencies may impose on the economy and establish limits on the number of new rules or regulatory actions by conditioning issuance of new regulations on the rescission of old ones. Under this approach, OMB and OIRA would still be in the driver’s seat of setting and ensuring compliance with the regulatory budget for individual agencies. CBO would oversee OMB and OIRA and advise members of Congress on remedial actions.

A variation of this framework would put CBO in charge of recommending individual rules to Congress for removal, rather than deferring to OMB and OIRA. CBO could refer rules to committees with jurisdiction over their subject matter, the committees could make recommendations, and Congress could vote on whether to keep them. Given the size of the code and time pressure of scheduling votes on thousands of regulations, CBO could review sections on a staggered, multi-year schedule, reviewing the entire code perhaps once every decade. Here, Congress need not adopt the CRA provision that prevents an agency from reissuing rules that have been voted down by Congress.[139] Such a provision might raise the stakes of any retrospective review beyond what is helpful to encourage removal of old rules and secure the president’s signature.

A third option would be to use an independent, impartial commission to review regulations. It could be modeled on the Base Realignment and Closure (BRAC) Commission established by Congress to determine, in an apolitical manner, which bases to close.[140] The BRAC Commission was initially established as part of a post-Cold War drawdown to shrink the defense budget.[141] Congress squabbled over the issue, as members sought to keep open the bases in their districts that were sources of jobs and boons to their local economies.[142] Rather than incur their constituents’ wrath for voting on closures, they agreed to let the Commission decide. The Commission’s experts recommended closures that made sense from a cost-savings standpoint. The only way Congress could stop a closure was to pass a joint resolution of disapproval. These procedures allowed members in districts with pending base closures to save face by publicly opposing the closure and voting for a joint resolution of disapproval, because it was unlikely enough similarly situated members would muster enough votes to thwart the Commission’s recommendations.[143] That is exactly what happened, and the Commission was a success.[144] Like the second option presented above, a regulatory review commission would take the authority to decide on individual rules out of the hands of the agencies. But unlike the second option, it would place them not in the hands of Members of Congress, but in an independent body, insulated from special interests, political incentives, and institutional pressures.[145] It is important to note, however, that unlike the first two options, this third method would not address the underlying constitutional issue of requiring Congress to make the major legislative or policy decisions. It is merely a practical means for removing regulations that are already on the books, many of which were issued pursuant to excessive delegations in the first place.

Congress could improve on the Trump administration’s model in important ways. Currently, the Executive Order 13771 budget may only apply to about 8 percent of federal regulations.[146] James Broughel and Laura Jones advise broadening the scope of the budget—beyond the small number of “significant regulations” that currently count in the cost-caps and 2-for-1 offsets—to include counts of regulatory restrictions or requirements in the CFR.[147] One way to measure these restrictions or requirements is to comb through the CFR for terms like “shall,” “must,” “may not,” “prohibited,” and “required.”[148] In some instances, these terms signify agency behavior, but in others, they define applicable rules of private conduct. Using a measure like regulatory restrictions or requirements captures more regulatory activity within the budget and allows regulations to be considered for adoption or rescission that have not undergone the prudent, but often complicated and time-consuming, process of cost-benefit analysis.

Lawmakers would also do well to consider the case study of British Columbia, as it demonstrates how regulatory budgets can be effective with simple measurements. This westernmost Canadian province, with a well-diversified economy and a population comparable to that of Louisiana, undertook a remarkable turnaround at the turn of the century.[149] In the 1990s, it was in last place in Canada for growth and employment.[150] A survey of mining companies in British Columbia scored the province last out of 31 jurisdictions.[151] In 2001, under new leadership, the province set a goal of reducing its “regulatory requirements” by a third in three years.[152] Broughel and Jones note that the measurement and definition used—“regulatory requirement”—was key to its success: “British Columbia’s two-for-one policy applied broadly to most requirements found in the province’s regulations, legislation, forms, and interpretive policies. The [U.S.] policy, by contrast, requires only that a relatively small number of legally ‘significant’ rules be offset.”[153] By 2004, the province had exceeded its retrospective review target, reducing regulatory requirements by 37 percent.[154] It institutionalized those reforms, and by 2015, it had cut 43 percent of its regulatory requirements.[155] The Canadian federal government took note and adopted a 1-for-1 regulatory budget.[156] However structured, retrospective review and regulatory budgeting can help lawmakers rein in the excesses of the administrative state.

Finally, Congress should start adding sunset provisions, or expiration dates, to the majority of its future statutes. Requiring reauthorization of statutes can “induce Congress to revisit, reassess, and recalibrate existing programs” to ensure they reflect up-to-date information and considered evaluation of agency behavior.[157] These statutes should also include sunsets on all regulations issued in pursuance of their expired authorizing legislation, as it should be unlawful for agencies to continue issuing rules without a current grant of authority in effect.

This reform would shift the burden of proof, so to speak, from Congress to the agencies. In conducting a retrospective review, Congress or CBO bears responsibility for identifying the regulations and defending its decision to vote disapproval or order their rescission. Sunset provisions shift the burden to agencies, especially when they must undertake the rulemaking process anew and are “subjected to public scrutiny, cost-benefit analysis and perhaps even court challenges.”[158] Far from a new idea, sunset provisions predate the republic and have been proposed by Thomas Jefferson, Alexander Hamilton, and William O. Douglas.[159]

While the prospect of an expiring regulatory code before a gridlocked Congress may give some pause, Idaho has demonstrated how sunset provisions can work without causing dire consequences. In January 2019, Governor Brad Little signed an executive order requiring regulators to identify two rules for repeal for every new one proposed.[160] Little was establishing a regulatory budget for the Gem State that mirrored Executive Order 13771. Yet another impetus for reform came from an odd quirk of the state’s government: the Idaho Legislature must reauthorize the entire regulatory code each year.[161] After a rancorous legislative session ended in April, lawmakers left town without reauthorizing the code.[162] The impending expiration provided the Little Administration with a rare opportunity to create a regulatory code from scratch that could be presented to lawmakers at the start of the next session in 2020.[163] By the end of the year, Little claimed to have cut 30,936 restrictions from the 72,000 that were on the books prior to expiration, which would make Idaho the least regulated state in the nation.[164]

Legislative impact accounting, regulatory budgets, retrospective review, and sunset provisions have traditionally been the recommendations of economists to improve the quality of regulations and minimize their tradeoffs. Moving forward, Congress can deploy them to reassert its authority and to prevent and correct lawless delegations to agencies.

IV. Conclusion

It has been said that the nondelegation doctrine had one good year back in 1935.[165] But the doctrine is far from dead. It just needs to be resuscitated. Time will tell whether the doctrine will get another good year, but recent developments leave room for optimism. And if the Court is a lagging indicator, then perhaps Congress, encouraged by reforms in states like Idaho, will move first to take the initiative and reestablish itself in the regulatory process. That would go a long way toward restoring its rightful place in the constitutional order.