Corbin K. Barthold is Internet Policy Counsel at TechFreedom and previously served as Senior Litigation Counsel at Washington Legal Foundation.

Ed. Note: This WLF Legal Pulse post reflects the views of the author and not necessarily the views of Washington Legal Foundation or the author’s employer, TechFreedom. Mr. Barthold hopes that this post sparks a conversation. The WLF Legal Pulse invites, and will consider publishing, thoughtful responses that offer alternative views on how the Court should approach nondelegation, major questions, separation of powers, and the administrative state. Send queries or submissions to


“Five votes,” Justice William Brennan used to tell his law clerks. “Five votes can do anything around here.” No one took this cynical view of the Supreme Court more seriously than Brennan himself. He ceaselessly lobbied and cajoled his fellow justices in search of the fifth vote for a desired result. And his efforts paid off. Brennan’s clerks once lined the halls of the Court with placards commemorating the many landmark decisions the liberal justice wrote during his long career. Little did Brennan know, as he strolled through the halls admiring his many accomplishments, that the Court’s newest member had snuck up behind him. “My Lord, Bill,” exclaimed Antonin Scalia, “have you got a lot to answer for!”

The shoe is now on the other foot. The Court’s new conservative majority could, if it so chose, fully embrace Brennan’s Rule of Five, overturning and constricting liberal precedents, expanding preferred rights and doctrines, and otherwise pursuing a reactionary agenda as fast as useful cases can be brought before it.

The new majority is indeed likely to make some bold moves. But there is an obvious risk of going too far, too fast, in too many directions. Some of the justices in the new majority are alive to this fact. Many conservative politicians, lawyers, and activists may have no time for the literal “conservatism” of prudence and incrementalism. The same cannot necessarily be said of Chief Justice Roberts, Justice Kavanaugh, or Justice Barrett.

“Society cannot safely forget the limits of effective legal action,” Alexander Bickel cautioned, “and attempt to surrender to the Court the necessary work of politics.” Directed at Justice Brennan and the Warren Court (as, when Bickel wrote it, it was), that line meant one thing. Directed at the rising conservative majority, it means another—yet it remains true.

Nondelegation, Then and Now

The Court is poised to revisit, and is likely to revive, the constitutional bar on delegating legislative authority. The “nondelegation” rule is doubtless worthy of the justices’ attention. But it is also something that should be adjusted only with great care and caution.

The Constitution vests “all legislative Powers” in Congress. It follows, as Chief Justice Marshall observed, that Congress may not pass, or “delegate,” that power to others. What is “legislative” power? Surely it has something to do with the creation of binding rules that govern the behavior of the public. But even James Madison remarked the “obscurity which rei[g]ns in these subjects, and which puzzle the greatest adepts in political science.”

For more than a hundred years the question attracted little attention, because Congress rarely delegated much authority to other government bodies. Then, however, came the Progressive Era and the New Deal, and with them the creation of the administrative state. In response, the Court twice enforced nondelegation, holding in Schechter Poultry v. United States (1935) that Congress may not hand the President an open-ended power to write “codes of fair competition,” and in Panama Refining Co. v. Ryan (1935) that it may not grant him broad discretion to limit the interstate transportation of oil. Yet the justices soon retreated. Picking up a stray line from J. W. Hampton, Jr., & Co. v. United States (1928)—which upheld a law that empowered the President, upon finding certain facts, to impose tariffs—they began to allow any delegation that contains an “intelligible principle.” Under this test, Congress has successfully delegated to agencies the power to regulate in “the public interest,” or in a manner that is “fair and equitable,” or as necessary to protect “the public health.”

The “intelligible principle” test now faces the chopping block. Dissenting in Gundy v. United States (2019), Justice Gorsuch, joined by Chief Justice Roberts and Justice Thomas, lamented the Court’s “intelligible principle misadventure.” Justice Alito agreed to reconsider the Court’s position—a stance that has enabled agencies, in his words, “to adopt important rules pursuant to extraordinarily capacious standards”—in a later case. And Gundy was decided before Justices Kavanaugh and Barrett joined the Court. Kavanaugh recently wrote that “Justice Gorsuch’s thoughtful Gundy opinion raised important points,” and, as a law professor, Barrett described the intelligible principle test as “notoriously lax.”

But if not the intelligible principle test, then what? In his Gundy dissent, Gorsuch offered little guidance. The main point, he insisted, is that “Congress makes the policy decisions,” while agencies may at most “fill up the details.” “Of course,” he was quick to add, “what qualifies as a detail can sometimes be difficult to discern.” That is an understatement. True, the policy-versus-details test has a strong pedigree: it comes from Chief Justice Marshall. But Marshall did not have to apply his test often; nor did he have to square it with the modern apparatus of government. He never had to fill up the details, so to speak.

Experts, Politics, and Pragmatism

Pundits rail against “Washington elites.” Politicians attack the “entrenched bureaucracy.” Lawyers speak derisively of the “headless fourth branch” of government. Broadly speaking, conservatives of all stripes agree: federal agencies should wield less power.

But what’s the plan? Even when shouted, contemptuous rhetoric and populist slogans will not, in fact, weaken the administrative state, let alone lead to more capable governance. The course of true reform does not run so smooth. Navigating it requires orchestration and hard work and tenacity, but most of all it demands retail-level competence. The only effective approach is to learn how agencies operate; articulately challenge their missteps and misdeeds in court; and form, articulate, and rally around detailed, plausible plans for making them leaner, more disciplined, more efficient, more accountable institutions. Under this approach, some populists will never be satisfied that the “swamp” has been sufficiently “drained.” But the simple truth is that it takes experts to combat experts.

At first blush, a sweeping revival of nondelegation might seem like an attractive shortcut—a simple, easy path to cutting the federal government down to size. Consider Gundy. At issue was a law that establishes sex-offender registration rules. That law delegates to the Attorney General the discretion to decide whether and how to apply those rules to offenders convicted before the statute’s enactment. As Justice Kagan recognized, writing for a plurality of the Court, this kind of discretion is a common feature of the United States Code. “If [this statute’s] delegation is unconstitutional,” she warned, “then most of Government is unconstitutional.” It is probably a bit much to assert that a different outcome in Gundy would have wiped out most of the federal government. To many on the right, though, making Kagan’s fear a reality, one way or another, is indeed the goal.

But is a nondelegation “quick fix” really a good idea? It is true that populist ultras will recoil at the idea that the justices should “only” place new “guardrails” around the exercise of administrative expertise. And it is also true that, as a matter of original constitutional structure, most government policy experts should be serving as advisors to the legislative branch, rather than as administrators in the executive branch and the independent agencies. Yet using nondelegation to abruptly gut the federal government would be sure to produce extreme unintended consequences. There is no telling how a judicial dismantling of the administrative state would affect the political system, the economy, or the judiciary itself. Put simply, making a beast of nondelegation would invite disaster. It would also make a mockery of the notion—a conservative notion, in the small-“c” sense—of the judiciary as the least dangerous branch.

It is far from clear, moreover, that conservatives would need to go so far to get what they want—or rather, as much of what they want as they can realistically obtain. When they attack the administrative state, conservatives tend to have bureaucratic political activism in mind. Think of the Department of Education dictating schools’ bathroom policies through the issuance of “Dear Colleague” letters. Think of the CDC using the pandemic to assert control over landlord-tenant law. Experts act as experts only when they apply special knowledge. Government administrators too often abuse their authority (indeed, risk discrediting themselves) by instead trying to exercise raw political power.

But much of what the administrative state does is not like that. When they stick to doing their jobs, administrators spend a lot of their time grappling with highly technical or scientific problems. At an oral argument a few terms ago, Justice Breyer noted that the FDA once had to decide whether “a particular compound should be treated as a single new active moiety, which consists of a previously approved moiety, joined by a non-ester covalent bond to a lysine group.” “Do you know how much I know about that?” he asked. (“Right, exactly,” he added, after the gallery laughed.) In a case about gene patenting, Justice Scalia once declined to join the portions of the majority opinion “going into fine details of molecular biology.” “I am unable to affirm those details on my own knowledge or even my own belief,” he explained. “Trust the experts” has become a phrase of opprobrium—and not without reason. It does not follow that judges or legislators are well equipped to tackle difficult questions of science, technology, engineering, or math.

In his Gundy dissent, Justice Gorsuch merely sketched an outline of what nondelegation should look like going forward. The justices’ goal, as they build on that outline, should be not to “burn it all down,” but rather to limit administrators’ ability to make nakedly political decisions under the guise of expertise.

Three Steps Toward a More Accountable Government

What, then, should the conservative justices do? Three things. First, they should carefully, selectively strike down a few delegations. Second, they should consider (re-)introducing a “process” element to the nondelegation analysis. And third, they should strengthen the limits on agencies’ ability to address major questions.

Step 1

Nondelegation is, first and foremost, a demand for precise language. Congress, however, long ago fell into the habit of issuing statutory demands that are irredeemably vague. Take Section 5 of the Federal Trade Commission Act, which empowers the FTC to ban “unfair methods of competition.” If this phrase is cabined by judicial precedents construing the Sherman Antitrust Act—as it historically has been by the courts—it is not so broad as to create a nondelegation problem. Yet the current Chair of the FTC, Lina Khan, wants to free the agency of this limiting construction. “Fairness” is in the eye of the beholder, and economic regulation is subject only to rational basis review. Little would stand between an FTC allowed to define “unfair” as it will, therefore, and the imposition on markets of what Edward Gibbon called the servile equality of despotism. A Congress obliged to wield “all legislative Powers” may not hand an agency that kind of discretionary authority.

Not every legislative use of the word “unfair” is an impermissible delegation. Nor should Congress be categorically barred from using other open-ended terms such as “feasible” or “reasonable.” But when one of these words is used to give an agency authority “as wide as the field of industrial regulation”—to borrow a phrase from Justice Cardozo’s concurrence in Schechter Poultry—that is “delegation running riot,” and “no such plenitude of power is susceptible of transfer.”

Justice Kagan has a point. The Court has neither the authority, nor the power, nor the institutional capacity to declare most of the government unconstitutional. The justices should be extraordinarily wary, therefore, of pressing the distinction between policy-making and detail-filling too far. Writing for the Court in Whitman v. American Trucking Associations (2001), Justice Scalia said that “the degree of agency discretion that is acceptable varies according to the scope of the power congressionally conferred.” A decision reviving nondelegation should confirm that this is so. Nondelegation should not come into play whenever Congress passes a question of policy to an agency.

Yet if nondelegation is to mean anything, it must bar Congress from effectively refusing to govern. The legislature may not delegate unlimited discretion over big questions of policy. It may not hand over unrestrained policy-making power “as wide as the field of industrial regulation.” In short, it may not simply tell an agency to “do justice!”—as it undoubtedly, on occasion, has done. The Court should identify the worst of these “do justice!” directives—Section 5 of the FTC Act, if construed too broadly, is one candidate—and strike them down.

Step 2

For the most part, Justice Gorsuch’s Gundy opinion treats delegation as binary. Can Congress instruct an agency to fill up the details? Yes. Can it let an agency make policy decisions? No. Still, Gorsuch acknowledged that Congress may lay down a rule and then “make the application of that rule depend on executive fact-finding.”

There is a perhaps related point, one that Gorsuch only hinted at. Recall that J.W. Hampton involved a tariff law. Why did the Court uphold that law? According to Gorsuch, because Congress “offered guidance on how to determine costs of production, listing several relevant factors and establishing a process for interested parties to submit evidence” (emphasis supplied). Indeed, the law allowed for what is now called notice and comment. Could it be, then, that whether a law is an improper delegation depends not only on the scope of the power the statute confers, but also on the amount of process the statute requires of the agency?

What Gorsuch left implied, several Supreme Court decisions say out loud. The Interstate Commerce Commission was constitutional, J.W. Hampton explained, in part because the Commission, before it could set “just and reasonable” shipping rates, had to “hear[] evidence and argument … from interested parties.” Writing for the Court in Wichita Railroad & Light Co. v. Public Utilities Commission (1922), Chief Justice Taft made a similar point. When handing “the regulatory police power” to an agency, he said, Congress, to avoid making “a pure delegation of legislative power, must enjoin upon [the agency] a certain course of procedure and certain rules of decision in the performance of its function.”

Schechter Poultry provides the clearest example of the principle. By the time that case was decided, Congress had passed the FTC Act, Section 5 of which is as we know quite broad. Why was Section 5 valid, while the law at issue in Schechter Poultry was an invalid delegation? One reason, the Court concluded, was that under Section 5 the FTC must issue a “formal complaint,” supply “notice and hearing,” make “findings of fact supported by adequate evidence,” and submit to “judicial review.” The law at issue in Schechter Poultry, the National Industrial Recovery Act, “dispense[d] with this administrative procedure.”

Here again Chair Khan’s understanding of Section 5 is instructive. Not only does Khan think the FTC may define “unfair,” as it appears in Section 5, however it wants; she believes that the FTC may dispense with case-by-case procedures—the ones mentioned in Schechter Poultry—and instead use Section 5 to issue binding rules. If the amount of process a statute requires informs whether the statute is a permissible assignment of power, the constitutionality of Section 5 could hinge on the validity of Khan’s reading of the statute. In other words, Section 5 might stand if it demands process and case-by-case adjudication, but fall, as an unlawful delegation, if it does not.

This distinction makes some sense. “Formal administrative procedure,” observed United States v. Mead Corp. (2000), tends “to foster the fairness and deliberation that should underlie” an “administrative action” with “the effect of law.” When that procedure limits the agency to case-by-case adjudication, moreover, it forces the agency to take an incremental, common-law approach, subject to thorough judicial oversight. We want experts to act as experts. We want them to stick where possible to addressing difficult technical matters, rather than subjects where, as Aristotle put it, “certainty is impossible and opinions are divided.” Finding more ways to make agencies act slowly, deliberately, and case-by-case would seem like progress. An agency that must take its time, cross its t’s and dot its i’s, and stick to narrow pronouncements will, one hopes, have a harder time making easy pronouncements about politics.

May Congress expand an agency’s policy-making power if it adds the right procedural strings to that power’s exercise? Which procedural strings might permit how much more power? These are questions the Court should take seriously.

Step 3

Under Chevron v. NRDC (1984), a court defers to an agency’s reasonable reading of an ambiguous statute the agency is charged with administering. When the agency’s reading touches on a major question of policy, however, this “Chevron deference” is withheld, and the court determines the correct reading of the statute for itself.

The “major questions” rule does not stop Congress from granting broad authority to agencies; it simply forces Congress to state clearly when it is doing so. The benefits of the rule are several. For one thing, Congress can still give agencies ample room to deal with technical matters—to regulate in areas that involve new active moieties or the fine details of molecular biology. For another, agencies are discouraged from trying to sneak outside their areas of expertise; from using ambiguous statutes to drift into political matters better left to legislators. For yet another, the legislative branch must at least take responsibility (i.e., write a clear announcement in a statute) for failing to take responsibility (i.e., passing authority to an agency). Although Congress remains able to duck certain policy choices, Congress must admit when it is doing so.

The major questions rule has the further virtue of being widely accepted at the Supreme Court. Soon after Chevron was decided, then-Judge Breyer proposed that considering the importance of a question is a natural step in determining whether Congress wanted the question resolved by an agency. After all, “Congress is more likely to have focused upon, and answered, major questions, while leaving interstitial matters to answer themselves in the course of the statute’s daily administration.” In the decades since now-Justice Breyer made that observation about “major questions,” the Court has applied the major questions rule in a half-dozen cases, in opinions joined by justices across the ideological spectrum.

Yet as the Congressional Research Service notes, “the Court has applied the ‘major questions doctrine’ in a somewhat ad hoc manner.” It is unclear how, precisely, the rule fits within the Chevron framework. Confusion remains even over what, exactly, makes a question a “major” one.

The major questions rule is before the Court this term, in West Virginia v. EPA. The case involves a previously obscure provision of the Clean Air Act—Section 111(d). Enacted in 1970, the provision was invoked by the EPA a handful of times to require the use, at existing power plants, of more environmentally friendly equipment or practices. In 2013, however, after Congress failed to pass climate-change legislation, President Obama directed the EPA to implement a cap-and-trade program by itself. The agency duly “discovered” the necessary authority in Section 111(d). Although Section 111(d) empowers the agency to establish “standards of performance” for power plants, the EPA’s Clean Power Plan used the provision to phase out many facilities altogether. The EPA’s reading of the statute would enable the agency to fundamentally restructure the nation’s energy industry. (The case arrives at the Supreme Court as a challenge to the Trump administration’s repeal of the Obama-era rule.)

Several of the many parties in West Virginia want the Court to clarify and bolster the major questions rule. As these parties explain, the Court should take the following steps:

  • Declare that the major questions analysis precedes the Chevron analysis. Only in the absence of a major question—or in the presence of a clear legislative grant of power over one—does a court go on to consider whether a statute is clear (“Chevron step one”), and, if not, whether the agency’s reading of the statute is reasonable (“Chevron step two”—the point at which the agency may claim deference). In the parlance of administrative law professors, a court looks for a major question at “Chevron step zero.”
  • Explain that there are several ways a question could qualify as “major.” Does the regulation at issue have a large economic impact? Does it address a politically contentious topic? Does it infringe on authority traditionally held by the states? If so, apply the major questions rule. Further, some regulations vastly expand agency power. Some abruptly discover agency power in unexpected places (recall the Court’s oft-used quip about elephants in mouseholes). Some look suspiciously like bills that failed in Congress. These are additional clues that the major questions rule applies.
  • Explain how Congress, when assigning authority over a major question, is to “speak clearly.” This might mean simply confirming that the statute must support the agency action without any need for Chevron. Or it might mean ruling that, when it wants to grant an agency authority over a major question, Congress usually must say so expressly. A presumption might exist, in other words, that an agency may not use open-ended statutory terms, such as “reasonable,” to justify major actions. To prevail in West Virginia, for instance, the parties defending the Clean Power Plan would need to show that Congress authorized the EPA specifically to impose a cap-and-trade program, shut down certain power plants, and restructure the energy industry.

As that last point suggests, a case can be made that the major questions rule should be decoupled from Chevron altogether. This might merely streamline the analysis. Or it might produce a “specific authorization” requirement along the lines just sketched.

The Court may have already taken a step in this direction, in a pair of decisions addressing emergency pandemic measures. One of the decisions, issued last year, reimposes a stay of the CDC’s pandemic eviction moratorium. The other, handed down this month, blocks enforcement of OSHA’s workplace vaccine mandate. Each decision declares that Congress must “speak clearly” when authorizing an agency to resolve a major question—yet neither decision mentions Chevron. In a concurring opinion in the vaccine mandate case, Justice Gorsuch, joined by Justices Thomas and Alito, took the next step, strongly suggesting that an agency should need specific authorization (“you may impose a vaccine mandate,” say, rather than “you may impose emergency measures”) to resolve a major question.

“When one legal doctrine becomes unavailable to do its intended work,” Justice Gorsuch wrote in his Gundy dissent, “the hydraulic pressures of our constitutional system sometimes shift the responsibility to different doctrines.” He then cited the major questions rule as a sign that the Court “still regularly rein[s] in Congress’s efforts to delegate legislative power.” If this was not intended as praise of the major questions rule, it should have been. The Court should use West Virginia as a launching pad for the construction of a clear, robust major questions protocol.

Taming the Administrative State, Responsibly

The Court should rein in the “fourth branch” of government. But how? The conservative legal movement has an enviable problem: it has reached the point where the goal is at hand, but the devil is in the details.

The Court should revive nondelegation—but modestly. And it should ensure that Congress can still empower agencies to solve hard problems, so long as they do so subject to rigorous procedures backed by judicial review and in accord with clear legislative mandates. With a few shrewd, deliberate rulings, the Court could spur more focused agency action, increase both congressional and administrative accountability, and encourage better legislation, all without damaging itself or society.