The Revival Debate

The Revival Debate

Courts have struck down a federal statute for delegating too much power on only two occasions, in Panama Refining and Schechter Poultry, and they far more often use the doctrine to read a statute narrowly than to invalidate it.¹ The test that governs is the one the earlier sections described: Congress may give executive officers and agencies broad rulemaking and regulatory authority, so long as it satisfies minimal requirements such as providing an intelligible principle. This section asks whether that low bar will last.

The bar is under pressure. Some Justices favor a formalist approach to the separation of powers, one that gives great weight to Article I’s vesting of legislative power in Congress and treats significant new policy as something only Congress may create. That approach sits uneasily with a test that a standard as open as “public interest” can satisfy.² The friction follows from what the test measures. An intelligible principle asks whether Congress named a policy and set limits the recipient must follow; it does not ask how weighty the delegated choices are. A statute can therefore pass while leaving an agency to settle questions of great economic and political consequence. A Justice who reads the vesting clause as reserving those questions to Congress finds the test too permissive, and a Justice who accepts that Congress cannot govern a complex society without broad directives finds it about right. The gap between broad delegation clauses and judicial pressure for narrower agency authority is likely to matter more in the future, and how Congress will respond remains to be seen.³

The four subsections below follow the debate in order. Individual Justices first urged a stricter doctrine; the 2019 decision in Gundy drew several of them together; one Justice proposed a replacement test; and the major questions doctrine meanwhile began to do similar work as a rule of statutory interpretation. That doctrine limits an agency’s reading of its statute when the agency claims power over a question of vast economic and political significance. Like nondelegation, it confines the reach of existing statutes without new legislation, and it is currently making a comeback.⁴ On an exam, state and apply the intelligible principle test first. Present a stricter approach as an argument that individual Justices have made and not as a holding, and treat the major questions doctrine as a separate route to narrowing an agency’s claimed power.

Earlier Calls for Revival

Every call for a stricter nondelegation doctrine starts from the words of Article I, § 1. Justice Thomas made the point in Whitman v. American Trucking Associations, Inc.: since 1928 the Court has treated the intelligible principle requirement as the only constitutional limit on congressional grants of power to agencies, yet the Constitution never mentions intelligible principles and says in much simpler terms that “All legislative Powers herein granted shall be vested in a Congress.”⁵ Three Justices have built on that text in three different ways: Justice Rehnquist in 1980, Justice Scalia in 1989, and Justice Thomas in 2001. Each wrote separately, and none spoke for the Court, so their positions are arguments in an open debate, and the intelligible principle test remains the law. They still repay study, because each names a different defect in a lenient test and so predicts a different way that a future majority might tighten it.

The first call came in Industrial Union Department v. American Petroleum Institute, the Benzene case. The Occupational Safety and Health Act told the Secretary of Labor to protect workers from toxic substances “to the extent feasible,” and the Secretary responded with a rule requiring expensive measures to minimize exposure to benzene, a chemical that causes cancer. Four Justices, led by Justice Stevens, set the rule aside because the Secretary had not made all the findings the statute required. Justice Rehnquist supplied the fifth vote on a different ground: the statute itself was an unconstitutional delegation. He called “to the extent feasible” a “legislative mirage” by which Congress avoided deciding when employers may run some risk of injury to workers given the high cost of protection, and he insisted that resolving fundamental, politically divisive policy questions is the essence of legislative authority and should not unnecessarily be left to an administrator unresponsive to the voters.⁶ In his view the Constitution forbids Congress to delegate major policy decisions to “politically unresponsive bureaucrats.”⁷

The charge rests on an institutional argument. The phrase “to the extent feasible” does not say how much risk to workers is tolerable or how much cost an employer must bear to reduce it, yet someone must answer those questions before a rule can issue. Under the statute, that someone is an official the voters did not elect. Legislators who voted for the Act can claim credit for protecting workers, while the agency absorbs the blame when the rule proves costly. Commentators who share Rehnquist’s concern make the incentive explicit: in the present climate, they say, Congress has too great a temptation to use open-ended delegations so that it can divert public blame from itself to the bureaucracy.⁸

Rehnquist was not alone. Interest in reviving the doctrine had grown through the 1970s, driven by the argument that Congress irresponsibly refused to resolve most policy disputes and delegated them to unelected decisionmakers, and by 1980 it appeared that a majority of the Court might act on that argument.⁹ The four-Justice plurality attributed to Congress a substantive standard that the statute did not contain, and it said that without that narrowing construction it too would have to conclude that the Act was an unconstitutional delegation.¹⁰ The plurality thus shared Rehnquist’s concern and chose a gentler remedy, narrowing the statute where he would have struck down the grant of power. Revival did not follow. The Court upheld five broad delegations in the years after Benzene and never fully explained why it apparently abandoned the effort.¹¹

Justice Scalia’s dissent in Mistretta v. United States rested on a different premise. He fully agreed with the Court that the Sentencing Reform Act could not be faulted for lacking intelligible, congressionally prescribed standards, so his objection did not turn on how precise the statute was.¹² His theory was that, strictly speaking, no delegation of legislative power is acceptable. The so-called excessive delegation cases, he wrote, really ask whether a grant of executive or judicial power is so general that it amounts to a delegation of legislative power.¹³

Some lawmaking is nonetheless lawful. Discretion, and with it a measure of lawmaking, is built into most executive or judicial action, and Congress decides through the specificity or generality of its commands how much of it a recipient receives.¹⁴ That is why courts may adopt rules of procedure and why the executive may specify what the “public interest, convenience or necessity” requires of radio and television licensees. In each case the lawmaking is ancillary, meaning incidental, to a power the recipient already holds: the judicial power to decide cases, or the executive power to grant and police licenses.¹⁵

The ancillary requirement answered a problem he thought courts could not solve head-on. The doctrine, he wrote, is “not an element readily enforceable by the courts,” because once it is conceded that no statute can be entirely precise, the debate over delegation becomes a debate over “a question of degree.”¹⁶ Precisely because courts cannot police degree, he argued, they must be rigorous in preserving the structural restriction that the power to make law cannot be exercised by anyone but Congress except in conjunction with a lawful exercise of executive or judicial power.¹⁷ The Sentencing Commission failed that restriction. Its lawmaking was divorced from execution and adjudication: it exercised no executive power of its own, did not answer to the President, and, not being a court, had no judicial power. Its power to make law was therefore “not ancillary but quite naked.”¹⁸ The situation would be no different in principle had Congress given the same sentencing power to the General Accounting Office or to members of its staff, and he treated the statute’s standards as irrelevant because they were, in his words, “plainly and simply, standards for further legislation.”¹⁹

The theory did not carry the Court. Twelve years after Mistretta, Justice Scalia himself wrote the majority opinion in Whitman, in which a majority hewed closely to its traditional approaches and the standards needed to uphold a federal statute remained broad.²⁰ His position therefore marks one edge of the critique: a stricter account of what counts as lawmaking, paired with doubt that judges can measure how much discretion is too much.

Justice Thomas’s concurrence in Whitman raised a third concern, the size of the delegated decision. Rehnquist had asked whether Congress dodged a hard choice, and Scalia had asked what job the lawmaking was attached to. Thomas joined the Court’s holding that the Clean Air Act directive was no less an intelligible principle than a host of others the Court had approved, yet he wrote separately because he was not convinced that the intelligible principle doctrine prevents all cessions of legislative power.²¹ He believed 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.’”²² On that view a statute could contain a perfectly clear standard and still fail, because a standard tells the agency what to pursue and says nothing about how weighty the decision is. He acknowledged that none of the parties had examined the constitutional text or asked the Court to reconsider its precedents, and he said that on a future day he would be willing to ask whether delegation jurisprudence had strayed too far from the Founders’ understanding of separation of powers.²³ The idea that some questions are too important for anyone but Congress to decide reappears as an element of the major questions doctrine, the subject of the last subsection.²⁴

Courts have seldom said why they declined these invitations, but the commentary suggests three explanations. The first answers Rehnquist’s complaint about unaccountable administrators. Broad delegations need not be undemocratic, because many shift policymaking discretion to appointees of a President who has an electoral base of his own, and a vague delegation can be defended as giving an incumbent administration the latitude to carry out the mandate the voters gave it.²⁵ The text offers a second response. Justice Stevens, concurring in Whitman, reasoned that Article I’s vesting of legislative power in Congress, like Article II’s vesting of executive power in the President, does not purport to limit the recipient’s authority to delegate, so a delegation that supplies a sufficiently intelligible principle is not inherently unconstitutional.²⁶

The second explanation is that no workable judicial test exists. Rehnquist’s formulation asks a judge to say which issues are fundamental and politically divisive, and Thomas’s asks a judge to say which decisions are significant enough to be legislative. Judges would have difficulty identifying delegations that involve fundamental issues or that stem from a congressional wish to avoid accountability, and a standard that subjective would put every new regulatory statute under serious threat, because a party burdened by an agency rule has every reason to argue, especially with hindsight, that Congress should have spoken in more detail.²⁷ The third, and perhaps most important, is that courts have found other ways to keep broad grants of power from becoming instruments of administrative oppression, namely the procedural requirements, judicial review, and narrowing constructions examined earlier in this chapter.²⁸

Not all of the pressure has come from individual opinions. In Jarkesy v. SEC, the Fifth Circuit held in 2022 that a statute allowing the SEC to choose whether to enforce alleged securities fraud administratively or in court violated the nondelegation doctrine, because it lacked an intelligible principle to guide the choice.²⁹ The court reached that result under the existing test and did not adopt any of the three proposals above. The Supreme Court later affirmed on other grounds, so the Fifth Circuit’s nondelegation reasoning shows what lower courts may do with the doctrine and is not a holding of the Supreme Court.³⁰ Even so, the decision supports the view that the doctrine has taken on new life after a long period of disuse.³¹

When an exam stem describes a statute that leaves an agency to settle a divisive tradeoff under a phrase like “to the extent feasible,” begin with the intelligible principle test, because it remains the rule the Court applies. The three positions then supply arguments that a stronger answer can add and label as positions. Rehnquist’s diagnosis fits facts in which Congress avoided a hard choice, Scalia’s fits a recipient whose only function is writing rules, and Thomas’s fits a decision of great significance made under a clear standard. The next subsection takes up the 2019 case in which a plurality upheld a delegation while three Justices would have gone further.

Gundy v. United States

In Gundy v. United States (2019), a plurality began from the text of Article I, which vests “[a]ll legislative Powers herein granted” in Congress, and explained that the assignment of power carries with it a bar on further delegation.³⁶ It then restated the qualification developed earlier in this chapter: Congress may obtain the assistance of its coordinate branches, including by giving executive agencies substantial discretion, and a delegation is constitutional as long as Congress lays down an intelligible principle to which the delegate must conform.³⁷ The question was whether a provision of the Sex Offender Registration and Notification Act (SORNA) met that standard when it authorized the Attorney General to “specify the applicability” of the Act’s registration requirements to people convicted before the Act was passed.³⁸

Herman Gundy was one of the people the opinions call pre-Act offenders. The year before SORNA’s enactment he pleaded guilty under Maryland law to sexually assaulting a minor, and after his release from prison in 2012 he came to live in New York without registering there as a sex offender. He was convicted of failing to register, and he argued that Congress had unconstitutionally delegated legislative power when it authorized the Attorney General to specify how SORNA applied to pre-Act offenders. The district court and the Second Circuit rejected the claim, as had every other court to consider it, including eleven courts of appeals.³⁹

Congress enacted SORNA in 2006 because the state registration laws that preceded it varied widely and contained loopholes and deficiencies that had allowed over 100,000 sex offenders, about 20 percent of the total, to escape registration.⁴⁰ The Act states its purpose as protecting the public through a comprehensive national registration system, and it backs that system with criminal penalties of up to ten years for a registrant who knowingly fails to register after traveling in interstate commerce.⁴¹ The general rule in subsection (b) requires registration before an offender completes a prison sentence. Many pre-Act offenders had already been released and could not satisfy that rule, so subsection (d) gave the Attorney General authority to specify how the requirements apply to them and to prescribe rules for registering those unable to comply.⁴² The Attorney General used that authority in an interim rule issued in 2007 and a final rule issued in 2010, both applying the requirements in full to pre-Act offenders, and the rule has not changed since.⁴³

The plurality treated the case first as a problem of reading the statute. A nondelegation inquiry begins, and often nearly ends, with statutory interpretation, because the constitutional question is whether Congress supplied an intelligible principle to guide the delegate’s discretion, and a court cannot answer that until it knows what task the statute delegates and what instructions it gives.⁴⁴ The order matters in practice. Two judges can agree that a delegation needs an intelligible principle and still reach opposite results, because one reads the provision to leave the delegate a choice of policy while the other reads it to leave only the carrying out of a policy Congress already chose. Once the meaning is fixed, the plurality observed, the constitutional question may nearly answer itself.⁴⁵

Gundy’s reading of subsection (d) shows what a failing delegation would look like. He contended that the provision gave the Attorney General plenary power to decide whether pre-Act offenders must register at all and to change that policy for any reason at any time. The plurality accepted that, if the provision meant this, a nondelegation question would arise, and then it rejected the premise.⁴⁶ The Court had already interpreted § 20913(d) to require the Attorney General to apply SORNA to all pre-Act offenders as soon as feasible, and a fuller look at text, context, purpose, and history led to the same conclusion, so the Attorney General’s discretion extends only to considering and addressing feasibility issues.⁴⁷

The statement of purpose carries much of the argument. Congress announced that it was establishing a comprehensive national registration system, and a system cannot be comprehensive if the Attorney General could decline, for any reason or none, to cover pre-Act offenders, who for many years after enactment would be the great majority of offenders in the country.⁴⁸ The plurality called the mismatch between that purpose and Gundy’s reading stark, and it added that the legislative history shows the need to register pre-Act offenders was front and center in Congress’s thinking.⁴⁹ Every source points the same way: Congress decided that pre-Act offenders would register, and the Attorney General received the question of when and how.

With the meaning settled, the constitutional step was short. A delegation is permissible when Congress sets out an intelligible principle or, in a related formulation, when Congress makes clear the general policy the delegate must pursue and the boundaries of the delegate’s authority.⁵⁰ That standard has proved undemanding. The Court has almost never second-guessed how much policy judgment Congress leaves to those who apply the law, and it has found a delegation excessive only twice, in a single year, in Schechter Poultry and Panama Refining, each time because Congress had failed to articulate any policy or standard to confine discretion.⁵¹ It has repeatedly upheld broad grants, such as authority to regulate in the “public interest,” to set “fair and equitable” prices, and to fix “just and reasonable” rates.⁵² Measured that way, SORNA’s delegation conveyed Congress’s policy that the Attorney General require pre-Act offenders to register as soon as feasible, and the feasibility issues the Attorney General could address were administrative and, more specifically, transitional.⁵³ Registering a large population that is already out of prison raises real questions of timing and logistics, and the statute left those questions to the executive after Congress had settled that registration would occur. That division resembles the one the details formula drew earlier in this chapter between choosing a policy and working out how to carry it out.

The plurality closed with a warning: if SORNA’s delegation were unconstitutional, most of Government would be unconstitutional, because Congress depends on giving executive officials discretion to carry out its programs, and feasibility is among the judgments often left to them.⁵⁴ A rule that treated every direction to act “as soon as feasible” as lawmaking would force Congress to write the timetable for each program into the statute itself, and the Court has long recognized that Congress cannot do its job without the ability to delegate under broad general directives.⁵⁵

Justice Kagan’s opinion spoke for only a plurality, an opinion joined by more Justices than any other but fewer than a majority. Justices Ginsburg, Breyer, and Sotomayor joined her, and Justice Gorsuch dissented for three, joined by Chief Justice Roberts and Justice Thomas.⁵⁶ Justice Kavanaugh joined the Court after oral argument and took no part in the decision, so eight Justices decided the case.⁵⁷ The fifth vote to affirm came from Justice Alito, who concurred in the judgment, meaning he agreed with the result without joining the plurality’s reasoning. He wrote that if a majority of the Court were willing to reconsider the approach taken for the past 84 years, he would support that effort.⁵⁸ His was the decisive vote in a five-to-three affirmance, and he would not take that step in Gundy itself because no majority of the Court, as constituted for the case, supported it.⁵⁹

Because only four Justices joined the reasoning, describe Gundy as a judgment upholding the delegation, supported by a plurality applying the intelligible principle test, and not as a holding that the test is settled for good. The plurality needed Alito’s vote to prevail, and the dissent would have gone much further, as the next subsection explains.⁶⁰ The Court’s membership has changed since. Justice Barrett replaced Justice Ginsburg, one of the four in the plurality, and Justice Kavanaugh has expressed openness to the dissent’s position, so some observers now count at least five and perhaps six votes for reinvigorating the doctrine.⁶¹ Another count places Justices Kagan, Sotomayor, and Jackson as committed to the intelligible principle test, Justices Gorsuch, Roberts, and Thomas as preferring an alternative, and Justices Kavanaugh and Barrett as unclear.⁶² The Court has not returned to the issue since Gundy and has meanwhile strengthened the major questions doctrine, the subject of the last subsection.⁶³

The interpretive step is also where the plurality is most exposed. A different account of the same statute describes Congress as unsure whether to require registration of pre-Act offenders and as leaving that decision to the Attorney General without any standard for making it.⁶⁴ That description mirrors the plurality’s. On the plurality’s reading Congress decided that pre-Act offenders register and left the Attorney General the timing; on the other reading Congress left the decision itself open. Because the constitutional step turns entirely on which reading is correct, a disagreement about the text becomes a disagreement about whether the delegation is valid. The plurality described its reading as the one that text, context, purpose, and history make clear, which distinguishes it from the narrowing canon discussed earlier, where a court adopts a limiting reading specifically to avoid a constitutional doubt.⁶⁵ The dissent rejected the plurality’s approach to the intelligible principle standard, which it considered misunderstood and wrongly used, and it treated imposing SORNA’s requirements on about 500,000 pre-Act offenders as the kind of question that should count as major.⁶⁶

Steps two and three carry the weight, because the result depended on which reading controlled. A challenger who wants a court to reach the constitutional question must persuade it that the statute leaves the policy choice open, and a defender wins by showing, as the plurality did with the statement of purpose, that Congress made the choice and left only the details.

Gorsuch's Alternative Test

In his Gundy dissent, Justice Gorsuch, joined by Chief Justice Roberts and Justice Thomas, urged the Court to abandon the modern intelligible principle test and to ask instead the questions that the framers’ idea of legislative power supports.⁷⁰ The framers, he argued, understood legislative power to mean the power to adopt generally applicable rules of conduct governing the future actions of private persons.⁷¹ On that definition the constitutional system breaks down if Congress can announce vague aspirations and leave others to write the legislation that realizes them, because the people vested the power to prescribe rules limiting their liberties in Congress alone.⁷² Because the opinion is a dissent, the framework below is a proposal and not the law. The Court still applies the rule taught earlier in this chapter: Congress has broad discretion to delegate rulemaking and regulatory authority to executive officers and agencies so long as it satisfies minimal requirements such as supplying an intelligible principle.

The dissent begins with the history of the phrase it wants to retire. Chief Justice Taft used the words “intelligible principle” in J.W. Hampton, Jr., & Co. v. United States in 1928, and the dissent reads the remark as just another way of describing the traditional rule that Congress may leave the executive the responsibility to find facts and fill up details.⁷³ Two facts support that reading. Taft wrote the phrase seven years before Schechter Poultry and Panama Refining, and it did nothing to stop those challenges from succeeding by lopsided votes. The tariff statute in Hampton itself directed the President to investigate the relative costs of production at home and abroad and to impose duties that would equalize them, so Congress arguably made every policy decision and left the President only fact-finding and details.⁷⁴ The phrase then sat more or less unused until the late 1940s, when lawyers began arguing that it had silently displaced all earlier teaching, and the dissent likens that development to treating an isolated line from an opinion as if it were a statute.⁷⁵ In the dissent’s judgment the resulting standard has no basis in the original meaning of the Constitution, in history, or even in the decision from which the phrase was taken, and it has been used to permit delegations that would fail under any other account.⁷⁶

The two approaches ask different questions, and that difference explains why the history matters to the dissent. The intelligible principle test asks whether Congress named a policy and set limits, which broad value words can satisfy. The dissent asks who actually made the policy choices. It describes the separation of powers as a procedural guarantee that requires Congress to assemble a social consensus before choosing the nation’s course on policy questions.⁷⁷ Each of the three principles below asks whether Congress did that work itself or left the executive to do it, and a court can answer each from the text of the statute.

The first principle is that, so long as Congress makes the policy decisions when regulating private conduct, it may authorize another branch to fill up the details.⁷⁸ The Court drew the underlying line early, between “the delegation of power to make the law, which necessarily involves a discretion as to what it shall be,” and the conferral of “authority or discretion as to its execution, to be exercised under and in pursuance of the law.” The Vesting Clause bars Congress from handing off its legislative powers, including to executive agencies, but it leaves Congress free to task others with the executive responsibility of “fill[ing] up the details” of the laws it creates.⁷⁹ Chief Justice Marshall’s opinion in Wayman v. Southard distinguished important subjects that the legislature must regulate entirely itself from subjects of less interest, where a general provision and a power to fill up details will do, and the dissent follows the same theme through the tax stamps for margarine packages in In re Kollock and the rules for public forests that the Secretary of Agriculture was authorized to adopt.⁸⁰ What unites these cases is that Congress must set standards “sufficiently definite and precise to enable Congress, the courts, and the public to ascertain” whether its guidance has been followed.⁸¹ The principle tolerates delegation of particulars because a particular, such as the design of a stamp, follows from a policy Congress has already chosen, and a court can compare the delegate’s rule against that policy. This is the details formula taught earlier in this chapter.

The second principle is that, once Congress prescribes the rule governing private conduct, it may make the application of that rule depend on executive fact-finding.⁸² The dissent’s model is the embargo statute in Cargo of the Brig Aurora v. United States, where Congress wrote the embargo and made it apply to one belligerent if the President found that the other had stopped interfering with American trade.⁸³ The finding was consequential, yet the policy was still Congress’s, because Congress fixed the result for each answer the President might give. This principle differs from the first in what the executive produces. Under the first, the delegate writes subordinate rules that carry out the policy; under the second, Congress writes the whole rule and the executive reports whether the condition for its operation exists.

The third principle is that Congress may assign the executive and judicial branches certain non-legislative responsibilities.⁸⁴ The Constitution vests all federal legislative power in Congress, but that power sometimes overlaps with authority the Constitution separately vests in another branch, so a statute conferring wide discretion may raise no separation-of-powers problem when the discretion is exercised over matters already within the scope of executive power.⁸⁵ The dissent suggests that the foreign-affairs embargo in Brig Aurora may be an example, and that the decision letting the President prescribe aggravating factors for court-martial death sentences may rest in part on the President’s commander-in-chief authority.⁸⁶ It adds that Wayman might be explained the same way for the judiciary, because courts have Article III power to regulate their own practice even without a statute.⁸⁷ This principle differs in kind from the first two. Those require Congress to make the policy choice; this one permits broad discretion because the recipient is not receiving a legislative power at all but exercising one the Constitution already gives it. The code-approval power in Schechter Poultry marks the limit: nobody thought that power inhered in the executive, and the President approved a lengthy code of fair competition written by a group of possibly self-serving poultry butchers.⁸⁸

The dissent turns these principles into working questions through Touby v. United States, which upheld the Attorney General’s power to schedule a drug temporarily after considering the drug’s history and pattern of abuse, the scope, duration, and significance of the abuse, and the risk to public health. The dissent reads the Court’s stress on those constraints as treating the statute as sound because it assigned an essentially fact-finding responsibility, and it says that, whether or not one agrees with that reading, Touby began to point back toward the right questions.⁸⁹ The questions are whether the statute assigns the executive only the responsibility to make factual findings, whether it sets out the facts the executive must consider and the criteria for measuring them, and, most importantly, whether Congress rather than the executive made the policy judgments.⁹⁰

The third step does the real work, and the first two supply its evidence. A statute can mention facts and criteria and still leave the policy choice open if the criteria do not determine the outcome, which is why an answer cannot stop at noticing that the statute contains some standard.

The dissent’s own application to SORNA shows the questions at work. The government had admitted in an earlier case that the provision left the Attorney General free to impose all of the Act’s requirements, some of them, or none of them on roughly 500,000 pre-Act offenders, to choose which of those offenders to cover, and to change course at any time or across administrations.⁹¹ The dissent found no single policy decision about pre-Act offenders on which Congress even tried to speak, and no case in which the Court had upheld executive authority over such matters as mere details.⁹² The embargo in Brig Aurora was just as consequential as this registration regime, so the difference lies in who decided whether the restriction applied. In Brig Aurora Congress decided and the President found a fact; under the dissent’s reading of SORNA, the Attorney General decided.

The dissent also compared the provision to the 1935 decisions. If Congress could not give the President discretion to ban or allow interstate shipment of petroleum, it reasoned, then Congress could not give the Attorney General discretion to apply or not apply SORNA’s requirements and to change his mind at any time.⁹³ The third principle offered no rescue, because SORNA gave the Attorney General power to prescribe the rules by which the duties and rights of citizens are determined, a quintessentially legislative power with no overlapping executive authority such as foreign affairs.⁹⁴ The dissent answered the plurality’s as-soon-as-feasible reading by noting that “feasible” could mean technological, economic, administrative, or even political feasibility, an evasive standard that would let the agency make important policy choices while frustrating meaningful judicial review, and that the government offered the reading only after the constitutional challenge arrived.⁹⁵

The dissent’s account of the danger shows the purposes behind its questions. Congress could not reach consensus on pre-Act offenders and passed the problem to the Attorney General, who, freed from assembling a broad supermajority, applied the statute retroactively to a politically unpopular minority, adopted different positions across administrations, and sacrificed fair notice.⁹⁶ Congress meanwhile claimed credit for comprehensively addressing the problem and left the Attorney General to sort it out, so the officials who could be blamed were not the ones who had made the choice.⁹⁷ The dissent insists that its test does not doom the administrative state. Congress may still authorize executive officials to fill in even a large number of details, to find facts that trigger a generally applicable rule, and to exercise non-legislative powers, and it may commission agencies or experts to study a problem and recommend legislative language.⁹⁸ On that view the separation of powers forecloses no substantive outcome and requires only that Congress do the deciding.

Three cautions keep the proposal in perspective. First, it has not been adopted. Only three Justices joined the dissent, the Court has not returned to the issue since Gundy, and the views of Justices Kavanaugh and Barrett remain unclear.⁹⁹ Second, critics question its historical foundations, pointing out that the opinion offers many general founding-era statements about liberty and the separation of powers but little specific evidence that the framers would have disapproved of the intelligible principle test itself, much less that they expected courts to police the size and powers of the administrative state.¹⁰⁰ Third, the framework is hard to apply to a statute like the Clean Air Act provision upheld in Whitman. Setting national air quality standards is a highly consequential decision that is not easily described as a detail, and much of what the EPA does could be called fact-finding about health risks, yet the agency must also decide how ample the margin of safety should be. The statute limits the agency to a single factor, health risks, and arguably makes the crucial policy judgment itself, that those risks be avoided without regard to cost, so how the test would apply remains unclear.¹⁰¹

The Court has meanwhile continued to police broad grants under another name, the major questions doctrine, which Justice Gorsuch describes as nominally a canon of statutory construction applied in service of the constitutional rule that Congress may not divest itself of legislative power by transferring it to an executive agency.¹⁰² The last subsection takes up that doctrine. On an exam, state the intelligible principle test as the governing rule, present these three principles as the dissent’s proposed alternative, and then run the statute through the questions above, noting where the answer would change.

Major Questions as a Substitute

In West Virginia v. EPA (2022), the Court explained that in certain extraordinary cases both separation of powers principles and a practical understanding of legislative intent make a court reluctant to read into ambiguous statutory text the delegation an agency claims to find there.¹⁰⁶ In those cases the agency needs “something more than a merely plausible textual basis” and “must point to ‘clear congressional authorization’ for the power it claims.”¹⁰⁷ That requirement is the major questions doctrine, a rule of statutory interpretation under which an agency asserting highly consequential power must show that Congress clearly gave it that power. The doctrine works on the meaning of the statute, so it lets a court stop an agency without declaring any delegation unconstitutional, and it leaves Congress free to grant the power later by saying so plainly.

The majority built the rule from earlier decisions in which an agency had pointed to real statutory words and still claimed too much. In Utility Air, the agency’s reading of “air pollutant” would have given it permitting authority over millions of small sources, such as hotels and office buildings, that had never been subject to such requirements. In Gonzales v. Oregon, the Attorney General claimed power to rescind the license of any physician who prescribed a controlled substance for assisted suicide, even in a State where that was legal. In NFIB v. OSHA, the agency imposed a mandate under which 84 million Americans had to obtain a COVID-19 vaccine or undergo weekly testing at their own expense.¹⁰⁸ All of these assertions had a colorable textual basis, yet the Court reasoned that “common sense as to the manner in which Congress [would have been] likely to delegate” such power made it very unlikely that Congress had done so.¹⁰⁹

The reason for demanding more than plausible text lies in how Congress writes when it means to hand over something large. Extraordinary grants of regulatory authority are rarely accomplished through “modest words,” “vague terms,” or “subtle device[s].”¹¹⁰ A Congress that wants an agency to remake an industry has plain language available and can be expected to use it. When the claimed power instead rests on an ambiguous or ancillary clause, the likelier explanation is that Congress never granted it, and the Court presumes that Congress intends to make major policy decisions itself and not leave them to agencies.¹¹¹ A clear-statement rule therefore forces a choice on Congress: it must decide the major question itself or delegate it expressly. That keeps the policy decision with the branch Article I assigns to make it, and no statute has to be struck down to get there. The cost is that the trigger, a “major” question, has no precise definition, as the discussion of factors below shows.¹¹²

West Virginia itself supplies both poles. The Clean Power Plan aimed to mitigate climate change by encouraging utilities to shift from coal toward cleaner forms of electricity generation, and EPA relied on § 111(d) of the Clean Air Act for its authority.¹¹³ The Court recognized that EPA may regulate carbon dioxide emissions from the utility sector and that the statutory language was broad, but it concluded that the claimed power was too consequential to infer that Congress meant to grant it.¹¹⁴ The agency’s reading required an innovative construction of § 111(d), and it rested on a very different kind of policy judgment, namely that it would be “best” if coal made up a much smaller share of national electricity generation.¹¹⁵ Under the agency’s earlier view of the same provision, EPA’s role had been limited to ensuring the efficient pollution performance of each individual regulated source, and a source already operating at that level left the agency nothing more to do.¹¹⁶ That source-by-source use of § 111 is the kind of ordinary regulatory power the doctrine leaves alone, while the new theory treated the provision as authority to reshape which fuels generate the nation’s electricity. The same logic later disposed of a loan forgiveness program. In Biden v. Nebraska (2023), the Secretary of Education’s plan to forgive over $430 billion in federal student loans for 20 million borrowers was a decision of such magnitude and consequence that it rested with Congress itself, or with an agency acting under a clear delegation that the majority could not find in the relevant statutes.¹¹⁷

The Court has not given a clear test for deciding which questions are major. In West Virginia it cited several factors: the potentially immense scope of the authority under the agency’s reading, the obscurity of the statutory source, the lack of precedent for the regulation, Congress’s refusal to enact similar legislation, and the agency’s lack of expertise in the area.¹¹⁸ Novelty did real work in the earlier decisions as well, since the Court found it telling that OSHA, in its half century of existence, had never used its authority to impose such a remarkable measure.¹¹⁹ Because the trigger is a matter of degree, an exam answer should argue the factors on the facts and should not announce the label and stop.

The doctrine’s link to nondelegation lies in what it accomplishes. The Court had long enforced the nondelegation doctrine through statutory interpretation, giving narrow constructions to delegations that might otherwise be unconstitutional, and in 1980 a plurality found it unreasonable to assume that Congress gave an agency unprecedented powers in the absence of a clear mandate.¹²⁰ The major questions doctrine applies that narrowing technique to claims of vast scope. Justice Gorsuch describes it as nominally a canon of statutory construction applied in service of the constitutional rule that Congress may not divest itself of legislative power by transferring it to an executive agency.¹²¹ It functions like nondelegation because it substantially limits the reach of existing statutes and prevents their expansion in certain contexts without new legislation.¹²²

The doctrine also asks something the intelligible principle test does not. Under the test taught earlier, Congress has broad discretion to delegate so long as it meets minimal requirements such as supplying an intelligible principle. A statute can pass that test and still fail to authorize a particular power, because in a major case the agency must show more than an intelligible principle: it must show clear congressional authorization.¹²³ An examinee who finds a standard in the statute has therefore answered only the first question.

The opinions in West Virginia disagree about where the doctrine comes from, and the disagreement shows what it is a presumption about. The majority rests it on two grounds at once, separation of powers and a practical judgment about what Congress is likely to have meant, and it presents the doctrine as a canon of interpretation.¹²⁴ Justice Gorsuch’s concurrence supplies a constitutional home. He wrote that, much as constitutional rules about retroactive legislation and sovereign immunity have corollary clear-statement rules, “Article I’s Vesting Clause has its own: the major questions doctrine.”¹²⁵ He traced a version of the rule to 1897, when the Court refused to presume that Congress had given the Interstate Commerce Commission power over railroad rates “from any doubtful and uncertain language.”¹²⁶ In his account the Court applies the rule, like other clear-statement rules, to ensure that the government does “not inadvertently cross constitutional lines,” and the lines at stake include self-government, fair notice, federalism, and the separation of powers.¹²⁷

Justice Kagan’s dissent rejects the premise that delegation is a constitutional concern. Relying on scholarship, it states that the founding era was not concerned about delegation, that the first Congress gave sweeping authority to the Executive Branch over matters such as territorial administration, Indian affairs, and the national debt, and that barely anyone objected on delegation grounds.¹²⁸ The dissent adds practical reasons that explain why Congress delegates. Members of Congress often lack the expertise to regulate sensibly, and they cannot anticipate changing circumstances, so a rational Congress tells EPA to select “the best system of emission reduction” and lets the agency adapt that choice over time.¹²⁹ The three opinions answer different questions. The majority treats the doctrine as a presumption about legislative intent that Congress can overcome by speaking clearly, the concurrence treats it as a safeguard that Article I itself demands, and the dissent finds neither history nor practicality behind it.

The choice among these accounts matters for what the Court is likely to do next. The majority’s opinion implies, though it does not expressly hold, that Congress could have delegated the decision to the agency if it had done so clearly, and some observers conclude that the Court is therefore less likely to reach the nondelegation question, since it can block a broad delegation by this route.¹³⁰ The intelligible principle test remains the governing constitutional standard until the Court says otherwise, and the major questions doctrine operates beside it as a way of reading statutes. A full treatment of the doctrine, including how courts decide what counts as major, belongs to the administrative-law reading.