Non Potest Delegari: How the Common-Law Principle of Agency Recasts the Nondelegation Doctrine

The nondelegation doctrine forbids Congress from delegating lawmaking authority to other government institutions. The Supreme Court has regularly emphasized the doctrine’s importance, but in practice, the Court’s rulings permit such wide delegations that many regard the doctrine to be dead. Yet the common law principle of agency provides a robust case to revive this doctrine.

Stemming from business interactions between private parties, the principle of agency holds that when a principal delegates a specific power to a specific agent possessing specific qualities, the agent cannot delegate that power to a third party. Likewise, Congress receives its power from the people, so it is an agent rather than a principal. Further, Congress is entrusted with a specific power, the authority to make laws coercively binding citizen action. Moreover, Congress possesses specific qualities tailored to this legislative task. Therefore, Congress may not subdelegate its delegated power to another institution. This framework unifies other arguments for the nondelegation doctrine and refutes objections to it.

This principle also speaks to the appropriate bright-line between acceptable and unacceptable delegation. Some potential bright-lines allow delegations based on their specificity, significance, or structural considerations, but this focus on ancillary factors ignores what makes law legislative: its coercive force. The principle of agency indicates that this power cannot be delegated, and this is true no matter how specific, significant, or structured it may be. Therefore, the principle of agency prompts a substantive standard based on the law’s content itself: Congress may not delegate the power to coercively bind citizen action. This leaves noncoercive authority (such as factfinding, nonbinding rulemaking, and enforcement choices) as acceptable delegations.

In 1815, a British court addressed a case about missing hats.[1] A woman had entrusted a shipment of millinery goods to the master of a ship so he could sell them for her once he arrived in the West Indies. But according to the shipmaster, he was unable to sell the hats there, so he entrusted them to someone else to sell them in Caraccas. Unfortunately, an earthquake struck Caraccas, the hats were destroyed, and the shipmaster could not give the woman proceeds from her goods. The matter ended up in court.[2]

Now, the man may have thought he was acting in the woman’s best interests by finding someone else who could sell the hats in a better market.[3] But he was the woman’s agent, and she had given authority over her property to him personally. As the court put it in this case, “[T]here being a special confidence reposed in the defendant with respect to the sale of the goods, he had no right to hand them over to another person, and to give them a new destination.”[4] Thus, when the shipmaster gave his delegated authority to a third party to accomplish the task for him, he betrayed the woman’s original charge to him. The court ruled against him.

As unfortunate as it was for the shipmaster, this decision simply applied a ubiquitous common-law rule about delegated authority.[5] Common-law judges in England and America recognized that an agent carried a personal trust from those who had sent him, so it was not his prerogative to delegate that trust away.[6] Their cases gave rise to a single fundamental principle about delegated authority: when a specific agent with specific personal qualities is invested with a specific power, then the agent cannot further delegate that power to anyone else.

At a minimum, this principle of agency ensures that women’s fashion can make it around the globe, but its implications reach much farther. It remarkably parallels the relationship between the people of a nation and their representative legislature, so it speaks directly to a controversial subject in American constitutional law: the nondelegation doctrine. This doctrine forbids Congress from delegating lawmaking authority to other governmental institutions, usually executive agencies.[7] Some view it as constitutionally obvious; others see it as a dead doctrine better left buried. And it has a curious history in American jurisprudence, seeming to be both indispensable and practically irrelevant to the nation’s highest court.[8] Notably, though, much of the conversation around the doctrine overlooks the fundamental political principles that the principle of agency would highlight.

The common-law principle of agency offers a powerful analogy that brings clarity to the political nondelegation debate. It prompts one to identify Congress as the agent of the people—an agent possessing specific qualities and therefore entrusted with specific powers. This framework shows that the nondelegation doctrine is a valid constitutional principle, and it implies a robust bright-line for what kinds of delegation are still acceptable. Accordingly, this argument proceeds in four parts. Part I explains the history of the nondelegation doctrine in American jurisprudence and its significance in contemporary legal discussion. Part II describes the principle of agency articulated in common-law sources. Then, Part III uses the principle of agency to show the nondelegation doctrine’s validity. Finally, Part IV uses the principle of agency to propose a substantive bright-line between acceptable and unacceptable delegation. All in all, this historic principle helps revive the doctrine in a much more robust form: Congress may not delegate the authority to issue rules coercively binding citizen action.

I. Nondelegation Refuses to Die

Although the nondelegation doctrine traces its roots to the first few decades of American jurisprudence, it is far from settled law and is still vigorously debated. In fact, the doctrine occupies a paradoxical place in the Supreme Court’s precedents; while affirmed in theory, it is almost inconsequential in practice.

From the nation’s first few decades to the 21st century, the Court has always seemed to agree with the doctrine in theory.[9] As Chief Justice John Marshall wrote in one of the nation’s first nondelegation cases, “It will not be contended that Congress can delegate to the courts or to any other tribunals powers which are strictly and exclusively legislative.”[10] Justice Antonin Scalia echoed him almost two centuries later, confirming that the Constitution “permits no delegation of those powers.”[11] Yet, because Congress cannot directly perform every necessary governmental task, the difference between appropriate and inappropriate delegation has been very difficult to nail down. Some early cases proposed standards to answer this question. The Supreme Court in The Cargo of the Brig Aurora v. United States[12] allowed Congress to delegate factfinding responsibilities to the executive and to pass laws contingent on his determinations.[13] A decade later, the Court in Wayman v. Southard[14] distinguished between “important subjects”[15] which the legislature must regulate directly and matters of “less interest”[16] for which the executive could “fill up the details.”[17] A century after that, in J.W. Hampton, Jr., & Co. v. United States,[18] the Court allowed the executive to determine tariff rates if Congress determined “an intelligible principle to which the person or body authorized to fix such rates is directed to conform.”[19] If the executive received such a principle to limit his actions in carrying out the statute, then he had not received a “forbidden delegation of legislative power.”[20] This “intelligible principle” metric has become the Court’s standard test.

Remarkably, however, only one statute has ever failed this test.[21] In Panama Refining Co. v. Ryan[22] and A.L.A. Schechter Poultry Corp. v. United States,[23] the Court ruled against a New Deal-era statute which had authorized the executive to regulate petroleum transportation[24] and set industry codes of “fair competition.”[25] The Court rejected these delegations because they did not provide sufficient criteria to guide the president’s actions.[26] But since then, the Court has allowed wide delegations to pass this same test. In Yakus v. United States,[27] the Court upheld a delegation of authority to an executive official to formulate “fair and equitable”[28] price controls. In National Broadcasting Co. v. United States,[29] the Court likewise approved a delegation of authority to the FCC to regulate broadcasting according to the “public interest, convenience, or necessity.”[30] This was intelligible enough for the Court, and thus it was acceptable for the executive to receive these broad rulemaking powers.[31]

By the 1970s, some scholars were characterizing nondelegation as a failed legal doctrine,[32] and as the intelligible principle test developed, it seemed virtually impossible for a statute to flunk. In both The Benzene Case[33] and Mistretta v. United States,[34] the Court upheld broad delegations which may not have had intelligible principles by themselves, but the Court determined that their broader statutory schemes or legislative histories gave sufficient principles to guide use of the delegated power.[35] In Whitman v. American Trucking Associations, Inc.,[36] the Court upheld a delegation to the EPA to set air quality standards with standards as loose as “requisite to protect the public health.”[37] The Court in Gundy v. United States[38] upheld a delegation to the attorney general to regulate the retroactive enforcement of a sex offender registration act, because the statute’s text, history, and purpose seemed to sufficiently specify the delegation.[39] Thus, given the Court’s repeated approval of delegations tasking the executive to pursue open-ended policy goals, the Court has curiously seemed to treat the doctrine as both indispensable and irrelevant. The idea behind it is affirmed as vitally important, but the focus on “unintelligible” delegations renders the doctrine virtually toothless—or, as some say, simply dead.[40] At the very least, it has been relegated to mere cameo appearances in the Court’s jurisprudence: technically present but never doing much.

Yet the doctrine has not died easily. Every decade since at least the 1970s has seen various judges or scholars arguing for its revival.[41] Some argue that nondelegation flows directly from the text of the Constitution,[42] others contend that delegation allows Congress to avoid proper political accountability,[43] still others claim that nondelegation is one part of a fundamental shift towards absolute power.[44] And despite the Court’s overall precedents, individual justices have occasionally expressed interest in reviving the doctrine.[45] There seems to be “something very fundamental . . . about the nondelegation doctrine that keeps resuscitating it when any rational observer would have issued a ‘code blue’ long ago.”[46] Indeed, the doctrine’s curious refusal to give up the ghost has earned it the title “Energizer Bunny of constitutional law: no matter how many times it gets broken, beaten, or buried, it just keeps on going and going.”[47]

In recent years, nondelegation has threatened even more vigorously to resurrect. Even though Gundy’s ruling upheld yet another broad delegation, Justice Neil Gorsuch’s scholarly dissent defended a set of guiding principles that were more substantive than the intelligible principle standard.[48] His dissent has sparked scholarly comment,[49] and at least four other Justices have seemed sympathetic to Gorsuch’s reasoning at various points.[50] Additionally, recent Supreme Court cases have begun to curtail the administrative state from multiple directions as they have reined in agency rulemaking authority,[51] the operations of administrative law courts,[52] and Chevron deference to agency interpretations of statutes.[53] In light of Gundy and other recent developments, one scholar observed in 2021 that “[w]e have been here before, but this time seems different.”[54]

In light of these circumstances, the hopes of nondelegation advocates rose in November 2024 when the Court granted certiorari in Federal Communications Commission v. Consumers’ Research.[55] This case disputed whether Congress violated the nondelegation doctrine by authorizing the FCC to determine monetary contributions for telecommunications carriers.[56] Yet any hopes of a reinvigorated doctrine were largely deferred by the Court’s decision in June 2025.[57] Both the majority and the dissent chose to evaluate the statute in question under the same long-standing intelligible principle test. On the one hand, Justice Elena Kagan’s opinion adopted that standard, and—unsurprisingly—the delegation passed.[58] In his dissent, however, Justice Gorsuch briefly referenced his own previous arguments for a more substantive delegation test.[59] Yet because the respondents had claimed that route was unnecessary, and had instead argued that the statute could not survive even the modern intelligible principle standard, he also decided to use that test—which, in his opinion, the statute failed.[60] Thus, both majority and dissent continued operating within the same long-standing framework rather than use this case to change the fundamental standard by which delegations are evaluated. Nondelegation advocates can only hope with Justice Gorsuch that the majority’s approach “will not stand the test of time,”[61] while they continue to promote “other guides, beyond the intelligible principle test, for assessing when Congress has impermissibly ceded legislative power.”[62] Until the Court weighs in on that question, this will continue to be a live debate.

Now, nondelegation is not merely an obscure dispute over legal doctrine. An academic debate is currently raging over the legitimacy of the administrative state as a whole, and the conflict is being waged on five main fronts. First is the question of nondelegation itself—whether Congress can delegate lawmaking power to other government agencies.[63] Second is the issue of administrative adjudication—whether administrative courts may carry out judicial proceedings outside of the judicial branch.[64] Third is the doctrine of judicial deference[65]—whether courts must defer to executive agencies’ interpretations of ambiguous statutes,[66] or even of agencies’ own regulations.[67] Fourth is the president’s ability to control his own executive employees—whether he may remove them or nullify their actions at will so that the entire executive branch remains politically accountable.[68] Fifth is a general argument about the separation of powers—whether the three core powers of government may be concentrated in the same branch when executive agencies craft binding rules, enforce them, and then adjudicate them.[69]

Delegation holds a central place in this multifaceted debate, for “[m]any of the features of the administrative state that anti-administrativists condemn . . . arguably follow simply from the phenomenon of delegation.”[70] Executive agencies exercise a combination of powers because Congress has delegated that authority to them. Administrative courts conduct judicial proceedings pursuant to congressional delegations. At least until the Supreme Court’s ruling in Loper Bright Enterprises v. Raimondo,[71] courts deferred to executive statutory interpretation because they assumed that Congress had delegated the authority to interpret.[72] And, in a sense, executive employees that are not answerable to the president are exercising a new kind of unanswerable power that Congress has delegated to them. Therefore, the nondelegation doctrine strikes at the heart of the modern administrative state. As Justice Elena Kagan put it in Gundy, if the contended delegation in that case were held unconstitutional, “then most of Government is unconstitutional.”[73] This debate is not merely a question of abstract legal theory, for it fundamentally affects how modern government works. So as this debate comes to a head, the stakes are high.

II. The Common-law Principle of Agency

Just as the nondelegation problem affects more than legal doctrine, the solution reaches deeper than legal doctrine as well.[74] Because delegation poses questions that touch the heart of representative government, a debate like this should not be “a technical slap-down, but . . . a dispute about principle.”[75] Beyond mere conclusions from case law that the doctrine is either dead or still alive,[76] a sound political philosophy is indispensable, for if one cannot understand what government fundamentally is, then one cannot understand what government should do. Fortunately, an old common law principle provides a solid philosophical framework to answer the delegation question. For centuries before the Constitution’s enactment, lawyers and judges had been assessing delegation within the private-law context of agency arrangements.[77]

The common law recognized that much human business must be carried out by persons acting on the authority of others.[78] Insurance agents could appoint clerks to sign their policies, landowners could hire agents to sell their properties, and shipowners could sell goods through their appointed supercargo. In such relationships, the one delegating authority was the “principal” and the one receiving the delegation was the “agent.”[79] The agent’s actions were considered to be taken by the principal himself.[80]

Because delegation was so integral to conducting business, the common law recognized that a principal may generally delegate any task that he could normally do himself.[81] But not so with an agent. An agent could not subdelegate (further delegate his charge to yet another agent), for it was assumed that his authority had been vested in him personally. As Justice Joseph Story explained,

[O]ne, who has a bare power or authority from another to do an act, must execute it himself, and cannot delegate his authority to another; for this being a trust or confidence reposed in him personally, it cannot be assigned to a stranger, whose ability and integrity might not be known to the principal, or who, if known, might not be selected by him for such a purpose.[82]

For example, if a landowner hired an agent to sell a piece of property, the agent could not further entrust it to a third party to sell it for him.[83] After all, the landowner had hired the agent because of the agent’s personal ability and integrity, but the landowner would not necessarily know the personal qualities of the third party. If the principal selected the agent “for her knowledge, skill, trustworthiness, or other personal qualities, he presumably gave the power to her, not anyone else,”[84] so the agent would undermine the purpose of the delegation by passing it further on. A Latin maxim summarized the principle succinctly: delegata potestas non potest delegari.[85]

Numerous common-law sources attested to this principle. Sir Edward Coke noted that a court of exchequer, having received a certain writ, “cannot make a commission to others concerning this matter, but ought to proceed legally themselves, because they have but delegatam potestatem, quae non potest delegari . . . .”[86] Matthew Bacon explained that an agent entrusted with a charge “cannot transfer it to another; for this being a trust and confidence reposed in the party, cannot be assigned to a stranger, whose ability and integrity were not so well thought of by him for whom the act was to be done . . . .”[87] Echoing this, Edward Sugden wrote that when “a power is given, whether over real or personal estate . . . if the power repose a personal trust and confidence in the donee of it, to exercise his own judgment and discretion, he cannot refer the power to the execution of another, for delegatus non potest delegare.”[88] According to the prominent agency-law theorist Samuel Livermore, because the power given to an agent is “a personal trust and confidence it is not in its nature transmissible, and if there be such a power to one person, to exercise his judgment and discretion, he cannot say, that the trust and confidence reposed in him shall be exercised at the discretion of another person.”[89] James Kent agreed:

The maxim is, that delegatus non potest delegare, and the agency is generally a personal trust and confidence which cannot be delegated; for the principal employs the agent from the opinion which he has of his personal skill and integrity, and the latter has no right to turn his principal over to another, of whom he knows nothing.[90]

In the 20th century, Floyd Mechem’s treatise on agency law reiterated that “powers which are conferred upon one in consideration of his personal qualities or characteristics, or as the result of special trust and confidence reposed in him, or which clearly contemplate the exercise of his personal knowledge, judgment or experience, should clearly be executed by him in person.”[91] These authors all expressed a ubiquitous principle of agency law: when a power was delegated to a specific agent on the basis of his specific personal qualities, it could not be further delegated.[92] A principal could generally delegate whatever he could do himself, but an agent could not subdelegate whatever he had been entrusted to do himself.

Admittedly, there were exceptions to the agency law nondelegation principle. An agent could delegate his power if his principal gave him express authority to do so, if there was a clear custom or tradition allowing subdelegation in a particular context, or if the transferred tasks were merely ministerial and did not require significant discretion to be placed in the third party.[93] Yet while agency nondelegation may not have been absolute, it was certainly presumptive.[94] In the absence of a recognized exception, delegated power could not be further delegated.

One other feature of agency law relevant to the nondelegation discussion is the classification based on the scope of the delegation. A principal could empower a “general agent” to perform any act required by the business at hand,[95] while a “special agent” was only empowered to perform a particular act for the principal.[96] Intermediate types of agents could perform a variety of actions within specific parameters.[97] Thus, in the common law, the breadth of an agent’s delegation affected what he could do and which of his actions were deemed valid. Regarding special agents, for example, “if the agent exceeds the special and limited authority conferred on him, the principal is not bound by his acts, but they become mere nullities, so far as he is concerned. . . .”[98] Therefore, in a principal-agency relationship, the details of the specific power delegated were also an important qualifier. Adding this to the broader rule, then, the common law principle of agency gives rise to a single fundamental premise about delegated power. When a specific power is delegated to a specific agent on the basis of his specific personal qualities, it cannot be further delegated.

This simple principle from private-law agency arrangements has profound implications for nondelegation in the political sphere. When applied to the political context, the precept prompts one to examine the specific powers, specific agents, and specific qualities of entire government institutions. In doing so, this centuries-old principle brings much-needed clarity to the murky nondelegation debate. In short, “the Supreme Court does not need to create a non-subdelegation doctrine. It needs only to rediscover it.”[99]

III. Defending the Nondelegation Doctrine

The nondelegation debate progresses at two different levels. Fundamentally, some disagree over whether the doctrine is constitutionally valid at all.[100] Yet even among the doctrine’s advocates, debate also persists over the line between legislative subdelegation to the executive and mere grants of appropriate discretion.[101] Any argument for nondelegation must address both questions, and the principle of agency does just that. As it highlights the specific powers, specific agents, and specific qualities of those agents within the political sphere, the principle of agency unifies arguments for nondelegation and helps refute objections raised against it.

A. Prominent Nondelegation Defenses

Scholars who argue for the nondelegation doctrine tend to rely on three considerations of political theory: constitutional text, separation of powers, and accountability for political leaders.[102] To begin with, a plain reading of the Constitution confirms the doctrine’s basic tenet. After all, legislative power is vested in Congress, executive power in the president, and judicial power in the courts.[103] So it would seem that the Constitution intends only one branch to legislate,[104] and nothing in the Constitution indicates otherwise.[105] What’s more, while the Constitution broadly vests “the executive Power” and “the judicial power” in the executive and judiciary, it specifically vests “[a]ll legislative powers herein granted” in Congress.[106] The word “all” indicates that Congress is the only federal entity that can exercise these legislative powers.[107] Notably, the document anticipates auxiliary executive departments and courts,[108] but it nowhere anticipates inferior lawmaking bodies.[109] This would imply that the Constitution only intends the amount of legislation that Congress can create directly. All this is significant because the Constitution created the federal government from scratch—it did not merely limit an already-functioning system.[110] This means the government does not fundamentally possess all power until the Constitution forbids it, but rather, the government fundamentally lacks any power unless the Constitution grants it.[111] Therefore, nondelegation is an a priori constitutional assumption, not some alien legal doctrine imposed on it.[112]

The separation of powers ideal supports the doctrine as well. Every government must fulfill three main roles: prescribing laws for action, applying them to particular cases, and carrying them out to full effect. But if a single institution fulfills multiple roles, it may craft laws with the intent to unfairly enforce or adjudicate them.[113] As James Madison warned, “The accumulation of all powers, legislative, executive, and judiciary, in the same hands . . . may justly be pronounced the very definition of tyranny.”[114] The Constitution recognizes this, so it establishes three distinct branches that may not trade these powers with each other. This way, the legislature must create laws that are clearly promulgated (to communicate its intentions to the citizenry and the executive) and fair to all affected (because it has no ability to enforce them arbitrarily). Yet when Congress delegates legislative power to the executive, laws are now made by the same institution that enforces them, which removes the separation of those two powers. Therefore, Congress should not distort the Constitution’s careful balance by delegating legislative power.[115]

Finally, the nondelegation doctrine keeps government accountable to the people. The Constitution itself recognizes how important accountability is; its frequent elections ensure that lawmakers are regularly “compelled to anticipate the moment . . . when their exercise of [power] is to be reviewed. . . .”[116] Of course, legislators must be accountable for the decisions they make, for only then can the public punish them for broken promises and poor judgment.[117] But when Congress delegates legislative authority, it no longer needs to shoulder responsibility for controversial decisions—it can simply tell an executive agency to decide the issue. Congress can claim the credit for a worthy goal (such as passing a broad “Clean Air Act”) but can then deflect blame onto the agencies that actually decide the details.[118] And even if voters do find out who really made the decisions, they are unable to democratically correct those unelected executive employees.[119] Since in a democratic government, controversial legislative decisions should be made by a politically accountable Congress and not a politically unaccountable bureaucracy, Congress should not delegate legislative power to executive employees. Ultimately, then, nondelegation scholars argue that the doctrine follows from both the Constitution’s text and from the separation-of-powers and political-accountability theories that undergird it.

B. The Principle of Agency Grounds the Case for Nondelegation

These are good reasons to believe that Congress should not delegate legislative authority. Yet there is a deeper level to the case for nondelegation, because the principle of agency emphasizes the underlying political reality from which constitutional text, separation of powers, and political accountability naturally follow. In fact, the American system assumes that the people have delegated specific power to a specific agent with specific qualities. This is seen through three main premises. First, the people, as principal, have vested power in Congress, as agent. Second, Congress possesses specific qualities tailored for its task. And third, Congress has received a specific power to carry out. Therefore, Congress cannot delegate legislative power to another institution. Moreover, the principle of agency unifies and grounds the other arguments for nondelegation in a way that they cannot do on their own.

1. Congress as an Agent

The private law context of agency arrangements may seem far removed from the political sphere, but the principles underlying it are just as applicable. Agency law provides a powerful analogy for the relationship between the people and the legislature of a nation. This is because political power originally arises from the people, and the legislature stands as their agent to exercise that power on their behalf. This assumption about the origin of political power is well-attested in Anglo-American political theory. As John Locke famously expressed it:

[T]he legislative cannot transfer the power of making laws to any other hands. For it being but a delegated power from the people, they who have it cannot pass it over to others. . . . [W]hen the people have said: We will submit to rules, and be governed by laws made by such men, and in such forms, nobody else can say other men shall make laws for them . . . .[120]

Thomas Jefferson observed similarly: “Our ancient laws expressly declare, that those who are but delegates themselves shall not delegate to others powers which require judgment and integrity in their exercise.”[121] The early 20th century political theorists Thomas Cooley and Charles Burdick agreed.[122] Contemporary voices argue this as well. As Justice Gorsuch explains in his Gundy dissent:

Our founding document begins by declaring that “We the People . . . ordain and establish this Constitution.” At the time, that was a radical claim, an assertion that sovereignty belongs not to a person or institution or class but to the whole of the people. From that premise, the Constitution proceeded to vest the authority to exercise different aspects of the people’s sovereign power in distinct entities.[123]

Gary Lawson puts an even finer point on it—he claims that the Constitution is similar to a power of attorney enabling one person to act on another’s behalf. Under this fiduciary instrument, the principal is clearly “We the People” and the agent is “Congress of the United States,”[124] so it is appropriate to apply the private-law principle of delegation to the U.S. political system.[125] Philip Hamburger would even rebrand the issue: the real question is not whether Congress may delegate, but whether it may sub-delegate the power it has already been delegated.[126] To suggest otherwise implies that power originally arises from the legislature instead of from the people.[127]

Moreover, American jurists have repeatedly applied this reasoning to court cases about legislative power. In Locke’s Appeal, the Supreme Court of Pennsylvania applied the principle of agency to the legislature:

That a power conferred upon an agent because of his fitness and the confidence reposed in him cannot be delegated by him to another, is a general and admitted rule. Legislatures stand in this relation to the people whom they represent. Hence it is a cardinal principle of representative government, that the legislature cannot delegate the power to make laws to any other body or authority.[128]

State courts made similar arguments in Rice v. Foster,[129] Parker v. Commonwealth,[130] People ex rel. Caldwell v. Reynolds,[131] and Thorne v. Cramer,[132] which all used agency language to describe the legislature and the people which have vested it with power. These courts followed the logical implications of the nature of American government. After all, in the private-law context, “the distinguishing features of the agent may briefly be said to be his representative character and his derivative authority”[133]—and this precisely describes the relationship between the American government and the American people. Legislators, executives, and judges do not possess original power; their power is delegated by the people who have established the government. Therefore, they are properly considered “agents,” and the principle of agency analogously applies to them, too.

2. Congress’s Specific Qualities

While Congress may not have “personal qualities” in the private-law sense, it certainly possesses distinctive institutional qualities in virtue of which it has received its power.[134] First, Congress possesses the prima facie requirement of every legislature: promulgating written legislation through a set process to govern future events, rather than determining issues case by case.[135] This written legislation must proceed through a bicameral vote and presentment before it can become binding law. This way, the nation’s laws will be less arbitrary,[136] more stable,[137] and more accessible to those bound by them.[138] Laws’ development will be open to public scrutiny, and their creators can be held accountable.[139] Second, individual legislators have certain qualifications. They come from various regions of the country, so that Congress will possess “knowledge of the affairs . . . of all the states.”[140] Their elections happen frequently, so that Congress maintains “an immediate dependence on, and an intimate sympathy with, the people.”[141] Third, Congress’s structure manifests deliberate calibration for legislative duties. It is a body composed of many individuals, which promotes “deliberation and circumspection.”[142] The number of representatives is large enough “to guard against too easy a combination for improper purposes” while still small enough “to avoid the confusion and intemperance of a multitude.”[143] Moreover, Congress’s two houses have different election schedules and different constituencies, so that both “a majority of the people, and then, . . . a majority of the states” must assent to legislation.[144] This further increases deliberation and helps ensure that small states still have a voice.[145] Fourth, Congress possesses one other distinctive quality—it exercises neither executive nor judicial powers. As the separation of powers principle implies, an institution that makes laws should not enforce or adjudicate them.[146] The executive even has a veto on legislative actions, so that “the less must be the danger of those errors which flow from want of due deliberation.”[147]

Notably, the other branches of government also have their own distinctive qualities ordered toward their own distinctive tasks,[148] and some of these are the exact opposite of the legislature’s. The framers refused to allow a single man to make law,[149] but the executive power rests on a single leader so that his operations retain their energy and personal responsibility.[150] The framers rejected unlimited terms for legislators,[151] but judges serve without limitation so as “to secure a steady, upright, and impartial administration of the laws.”[152] Evidently, the three branches are not different in name only—they are also different in nature.

In sum, Congress is a distinct, deliberative, representative body promulgating written law. Its distinct status, isolated from the other branches, preserves the separation of powers. Its deliberative qualities foster wise decisions.[153] Its representative qualities keep it responsive to the people.[154] Its set processes for promulgating written law preserve the liberty of the people from runaway government action.[155] Evidently, “[t]he constitutional gauntlet required to make law was a conscious feature of the charter’s longest article, not a bug.”[156] Congress’ specific qualities are ordered toward the intended end of wise, beneficial lawmaking—and it is precisely in virtue of these qualities that it has been delegated legislative power.[157]

3. Congress’s Specific Power

While Congress’s distinctive qualities are relatively straightforward, its specific power is harder to define precisely. Is legislative power merely the power to make any rules whatsoever? If so, then Congress would be responsible for all guidelines for every government employee as he goes about his daily work. Yet legislative power must have something to do with making rules, or it would not be called “legislative” in the first place. Some of America’s brightest political thinkers have admitted the difficulty of precisely defining and distinguishing this power.[158] Yet even though it is a difficult question, it is not an impossible one.[159] In fact, several founding-era and contemporary theorists center around a robust, substantive definition of this power.

Legislative power consists of two facets: essential legislative power and accompanying legislative power. The former describes, from the standpoint of political theory, the core activity that makes legislative power what it is by nature. The latter includes, from the details of a nation’s constitution, what else the legislature is assigned to do. Taken together, these two facets of legislative power are the “specific power” that the people have delegated to Congress.

Essential legislative power is the power to issue rules coercively binding citizen action.[160] Blackstone described law as “the rule of civil action,”[161] originating when all persons in a society “submit their own private wills”[162] to the supreme authority who establishes “general rules, for the perpetual information and direction of all persons in all points.”[163] Similarly, Locke characterized law as that by which the “people [are] bound,”[164] that which has “force over the subjects of that commonwealth.”[165] The legislator had “authority of making laws that shall be binding to the rest”[166] of the people. Likewise, Hamilton said that “[t]he essence of the legislative authority is to . . . prescribe rules for the regulation of the society . . . .”[167] (implying that they are “rules prescribed by the sovereign to the subject”[168]), and he described the legislature as the one that “prescribes the rules by which the duties and rights of every citizen are to be regulated.”[169] And Madison claimed that a law “by which personal liberty is invaded, property deprived of its value to the owner, and life itself indirectly exposed to danger”[170] is especially the province of the legislature. As Justice Gorsuch summarizes, the framers’ understanding of legislative power was “the power to adopt generally applicable rules of conduct governing future actions by private persons.”[171] Accordingly, law is what binds the actions of the citizens of a nation, and therefore the essential legislative power is the power to issue rules coercively binding citizen action.

Moreover, suppose the alternative was true instead. Suppose that the power to issue law coercively binding citizen action was not uniquely the province of the legislature, and that it could be the province of the executive or judiciary as well. If so, then laws obligating the people could originate in the branches least responsive to their consent,[172] and the legislature would become irrelevant for any laws operating within this two-branch system. This would be especially pernicious for criminal statutes, as the executive is already one of the parties to a criminal case, and it is a key constitutional principle that no citizen can be punished for a criminal act without the participation of all three branches of government.[173] So the separation of powers, due process, and the rule of law could be violated in many cases.[174] This is why “the natural dividing line between legislative and nonlegislative power was between rules that bound subjects and those that did not.”[175] Because of the practical harms of allowing this power to the other two branches, it makes the most sense to classify it as uniquely legislative—corroborating the definition drawn from Blackstone, Locke, Hamilton, and Madison.

While the power to issue rules coercively binding citizen action is uniquely legislative, and no other branch should possess it, this does not rule out other powers from also being the province of the legislature.[176] After all, the Constitution entrusts numerous noncoercive powers exclusively to Congress, such as borrowing money, establishing post offices, and declaring war.[177] So by itself, the “coercive rule” standard does not reflect the entire text of the Constitution.[178] It is important to also recognize accompanying powers in a definition of legislative power. Among others, the Constitution exclusively grants Congress power to impose taxes, borrow money, appropriate public funds, regulate commerce, coin money, fix weights and measures, establish post offices and roads, issue patents, establish federal courts, declare war, maintain and regulate military forces, govern the nation’s new capital, and of course, make all laws necessary and proper for its assigned powers.[179] Some of these overlap with the principle of coercively binding citizen action; some do not. But in the American context, these are all considered legislative as well.[180]    

Therefore, in the American context, legislative power is the power to issue rules coercively binding citizen action, along with all other powers exclusively granted to Congress in the Constitution. This combination of subtext and text is faithful both to the document and to the principles undergirding it. And in terms of the principle of agency, this is the “specific power” which the people, as principal, have granted to Congress, as agent.

4. Applying the Principle of Agency

The principle of agency provides a powerful argument against subdelegation. Congress does not possess its own original power; it is the agent of the people. Moreover, it possesses specific qualities tailored for its task—it is a distinct, deliberative, representative body promulgating written law. Further, its delegated power has a concrete definition: it is the power to issue rules coercively binding citizen action, as well as any other powers vested in Congress by the Constitution. And “[f]rom a design standpoint alone, it is inconceivable that the framers would have taken such great care in whom they vested the people’s lawmaking power . . . but didn’t care a whit if this ‘most dangerous’ power was later re-delegated to the executive or unelected administrators . . . .”[181] Ultimately, because specific legislative power has been delegated to a specific legislative body possessing specific legislative qualities, this delegata potestas non potest delegari.

Alternatively, when legislative power is delegated to an institution with executive qualities, that institution will legislate like an executive. Granted, laws will be made more quickly and efficiently, for those are qualities that the executive is designed to have.[182] But the process will also lack many qualities that the legislature is designed to have. Laws may be more arbitrary or biased toward certain groups, and less stable and predictable. Laws may be unsympathetic to popular interests or regional concerns. Laws may lack the wisdom that comes from extended debate. And since laws will be made by the same institution that enforces them, one can expect more of the runaway government power this enables.[183] All in all, the people can expect laws marked less by “deliberation and wisdom, and best calculated . . . to secure their privileges and interests,” and more characterized by “energy . . . decision, activity, secrecy, and despatch.”[184] This should not be surprising, for the framers designed the executive to act this way. But there is a reason Hamilton praised the virtues of both “a single executive, and a numerous legislature.”[185] The framers did not design the executive to act this way with this power.[186]

The principle of agency does not merely concord with constitutional text, separation of powers, and political accountability. It constructs a broader political framework within which the other justifications find their place. This framework explains why the Constitution’s text says what it does—the Constitution vests “[a]ll legislative powers”[187] in the agent possessing the qualities tailored for the task. Separated power is one of those specific qualities that Congress is designed to exhibit. And political accountability is an intended goal of Congress’s specific structure. Further, the principle of agency prompts inquiry into matters that the other arguments do not necessarily reach, such as the nature of a legislature and of legislative power. Simple appeals to constitutional text may be legally sufficient but are philosophically unsatisfactory, because the Constitution is not merely words on a page. Its text reflects an underlying reality, a set of institutions with specific work assigned for each to do. Without that understanding, isolated provisions of the Constitution (notably, the Necessary and Proper Clause)[188] can be used to undermine the underlying reality that the Constitution was intended to establish. Further, bare appeals to separation of powers or political accountability do not identify the nature of the legislative power that is supposed to be separate or accountable. Opponents may contend that when the executive follows congressional directions to formulate rules, then he is merely “executing” those directions.[189] But a substantive definition of legislative power explains why that would still be legislative power, albeit in executive hands.[190] Opponents might also contend that legislative power includes a prerogative to delegate, and legislators are still are accountable to their constituents when they use it.[191] But the uniquely tailored qualities of each government branch explains why delegation between them cannot be their prerogative.[192] So the other nondelegation justifications must ground themselves in a concrete account of the nature of a legislature and of legislative power if they are to have teeth.[193] Ultimately, the principle of agency provides both an overarching context and an underlying justification for constitutional text, separation of powers, and political accountability.

Now, there is one obvious objection to the application of agency law in the nondelegation conversation: the common law’s exceptions to its own rule of agency. The common law does allow for subdelegation when the principal has expressly authorized it, when clear custom or tradition have allowed for it, or when merely discretionary or ministerial tasks were subdelegated.[194] However, these exceptions do not justify redelegation of legislative power, because they lack an analogy to the political context.

First, the express authorization exception does not apply to the American political context, because the people have not expressly authorized subdelegation in the Constitution. The only possible source of such an exception would be the Necessary and Proper Clause, which “does not even come close to constituting an express authorization for subdelegation.”[195] Second, appeal to custom does not conclusively support subdelegation, either. As Lawson explains,

[T]here could not possibly be a custom or tradition of subdelegation in a regime of separated powers such as that created by the Constitution when there was no established history of such separation-of-powers regimes from which to induce a custom or tradition. The American separation-of-powers structure, embodied in the Constitution and in the state constitutions that preceded it, was unique in human history, so references to other traditions are of limited value for understanding that structure.[196]

Third, though, the exception for discretionary and ministerial tasks does substantiate a much more formidable objection.[197] After all, there is a good case to be made that Chief Justice John Marshall reasoned along these lines when he distinguished between “important subjects” and matters of “less interest” in Wayman v. Southard.[198] However, attempts to define and draw a concrete line between non-discretionary and discretionary tasks are fraught with difficulty. Without a careful definition, allowing “ministerial tasks” to be subdelegated would allow for all manner of legislative delegations and leave the nondelegation doctrine toothless yet again.[199] Yet as Michael Rappaport observes, there is a way to interpret this ministerial exception such that the distinction remains clear: laws coercively binding citizen action are the “important subjects,” and noncoercive laws are “matters of less interest.”[200] Such a standard obviously cannot be gleaned from private-law cases, because private parties do not normally possess the power to draft laws coercively binding the actions of their fellow private parties. But it is a fitting standard for the political context, because drafting coercively binding laws is precisely the “important subject” that a legislature is created to handle. Therefore, the third common-law exception—allowing discretionary or ministerial duties to be subdelegated—harmonizes with the nondelegation doctrine if “ministerial” actions are those which do not coercively bind citizen action. And this is precisely what this article argues the nondelegation doctrine should stipulate.

C. Prominent Nondelegation Critiques

Yet not all agree that the nondelegation doctrine enjoys such resounding support. In fact, some adopt the same standards of constitutional text, separation of powers, and political accountability to argue against the doctrine.[201] But while their arguments raise important points, they fail to recognize the truths highlighted by the principle of agency.

Not all believe that the Constitution’s text forbids legislative subdelegation. Gillian Metzger argues exactly the opposite: the administrative state (including executive officials promulgating rules for society) is constitutionally necessary. First, she argues, “given the economic, social, scientific, and technological realities of our day,”[202] delegation is “mandatory in practice” if modern government is to handle its wide variety of responsibilities successfully.[203] Second, the Constitution requires that the executive “take Care that the Laws be faithfully executed.”[204] So if the administrative state is necessary to execute congressional legislation, then it is “constitutionally obligatory.”[205] This is a common sentiment in the conversation about delegation: our modern world is so complex that we simply need it. As the Supreme Court recognized in Mistretta v. United States, “[O]ur jurisprudence has been driven by a practical understanding that, in our increasingly complex society . . . Congress simply cannot do its job absent an ability to delegate power under broad general directives.”[206] For many, delegation of rulemaking authority is acceptable because government simply could not function without it. And for Metzger, this means the administrative state is constitutionally mandated.

Now, Metzger correctly observes that the nondelegation principle conflicts with the current scope of government operations.[207] Yet she assumes that current government activity dictates whether one accepts nondelegation, when it could very well go the other way around. Indeed, if the nondelegation doctrine is true, then it would be a positive case against expansive government operations. If Congress cannot delegate legislative power to other institutions, government operations must be limited to that for which Congress can directly legislate.[208] Notably, the Constitution anticipates “workload overload” for the executive and judicial branches. The individual president and the one Supreme Court cannot possibly handle all executive and judicial responsibilities, so the framers anticipated inferior executive departments and federal courts.[209] But the Constitution vests “[a]ll legislative powers” in a single, bicameral Congress.[210] This conspicuous absence of inferior legislative bodies suggests that the framers only intended the amount of legislation that Congress itself could enact.

Fundamentally, though, Metzger’s argument treats Congress like a principal. Principals can generally delegate whatever they can do themselves as the exigencies of the situation demand.[211] But agents may not, because they are entrusted with power personally. Because Congress is an agent vested with power because of specific qualities it alone possesses, it is not interchangeable with executive agencies.

Next, not all critics of nondelegation agree that legislative delegation violates the separation of powers. In fact, Eric Posner and Adrian Vermeule argue that statutory grants of rulemaking authority never transfer legislative power.[212] This is because, they claim, legislative power is “the authority to vote on federal statutes or to exercise other de jure powers of federal legislators,”[213] while executive power is “authority that the president exercises pursuant to a statutory grant.”[214] Therefore, when Congress passes a statute ordering the executive to issue rules, and the executive carries out this statute by promulgating such rules, it “isn’t lawmaking, but law execution.”[215] So when Posner and Vermeule agree that legislative power should not be delegated, they merely mean that executive officials cannot have the power to vote in Congress.[216]

Of course, if legislative power is “whatever congressmen do when voting” and executive power is “carrying out whatever Congress tells the executive to do,” then it would be practically impossible for the executive to ever exercise legislative power. But Posner and Vermeule’s critical mistake is to define governmental powers procedurally (according to how they are carried out) instead of substantively (according to the content of the actions themselves). Suppose their logic applied to the judicial branch, and suppose a judge issued a final judgment in a case that “the relevant legal rights and duties of the parties are whatever the President . . . determines those legal rights and duties to be.”[217] According to Posner and Vermeule, no one could accuse the judge of delegating judicial power, for he issued a ruling in a court (just as “voting in Congress” is the legislative power).[218] Moreover, no one could accuse the president of exercising judicial power, for he did not issue a ruling in court; he was merely executing the judge’s judgment.[219] Yet when judicial power is defined according to the content of the action itself (“adjudicating cases and controversies”), the judge clearly abdicated his responsibility and delegated it to another party to exercise instead. Similarly, suppose that Congress passed a law empowering the president to act as a judge over criminal trials. If the president followed these instructions and served as a judge, Posner and Vermeule would claim that he was merely “executing” that law. But regarding the content of his action, he would clearly be exercising judicial power.[220]

So legislative power is not merely the power to vote on legislation, but rather the power to issue rules coercively binding citizen action and to carry out any accompanying powers exclusively vested in Congress by the Constitution. And executive power is not merely the power to follow any congressional instructions; it “presupposes a proper legislative act to execute.”[221] Thus, legislative delegation does violate the separation of powers when it assigns the ability to issue coercive rules to another branch. Posner and Vermeule ignore the substantive powers granted only to Congress, and they overlook Congress’s unique legislative qualities tailored for those powers. But Congress is an agent, not a principal, so these cannot be ignored.

Finally, not all critics of nondelegation agree that the nondelegation doctrine would promote greater accountability. Some say that the doctrine would shift a massive amount of power to the judiciary without clear standards for its use. This is because no statute can be completely precise about every action the executive is to take.[222] Congress must expect the executive to use a certain amount of discretion in carrying out any law, so some amount of delegation is unavoidable. So it would seem, as Justice Scalia noted in his Mistretta dissent, that “the debate over unconstitutional delegation becomes a debate not over a point of principle, but over a question of degree.”[223] Because there is no easy metric for how much executive discretion is too much to still remain “executive,”[224] many scholars concede that the nondelegation doctrine is simply too indefinite to be a principle of constitutional law.[225] Because of its ad hoc nature, “judicial enforcement of the conventional doctrine would grant massive new authority to the federal judiciary, authority to second-guess legislative judgments about how much discretion is too much, without clear constitutional standards for answering that question.”[226]

While admirable for its desire to prevent judicial activism, this argument overreacts in the other direction: judicial abdication. First, courts can still reach relatively consistent conclusions even though hard cases sometimes arise. As Gaziano and Blevins observe,

That some statutory gap filling is inherently, implicitly, or expressly authorized is no ground to conclude that the courts could never develop a body of law that distinguishes the permissible from the impermissible. Unlike a legislature writing a comprehensive statute or code, courts need to decide only one application or problem at a time and can reconsider their prior distinctions relatively easily. It’s called jurisprudence.[227]

Second, this argument overlooks the cases which are not difficult at all. If any congressional statute delegates practically all policymaking responsibilities to the executive, then “no line drawing is necessary,”[228] and a court has no reason for hesitation.[229] In fact, it is the courts’ duty “to declare all acts contrary to the manifest tenor of the constitution void,”[230] so judicial abdication could be just as egregious as judicial activism. Finally, though, this objection only has teeth if its fundamental assumption is true—there is no robust standard. If there is an objective, principled distinction between appropriate and inappropriate delegation, then Scalia’s concern would no longer be salient. And so, still keeping the principle of agency in view, this discussion turns to the question of an appropriate bright-line.

IV. Defending a Nondelegation Bright-line

The nondelegation doctrine itself is controversial, but a workable bright-line is even more so. All must admit that some amount of executive discretion is needed to execute any law,[231] but there is no easy line between appropriate and inappropriate delegation.[232] This leads some of the doctrine’s sympathizers to hesitate to apply it,[233] while some of its opponents exploit this deficiency and reject the doctrine outright.[234] The Supreme Court itself seems entangled in a paradox: it formally recognizes that any delegation of legislative power violates the Constitution,[235] but “[t]he sticking point during the past 80 or more years has been how to properly enforce it.”[236] All the noble appeals to the Constitution and the separation of powers are overshadowed by one looming cloud—there seems no way to objectively apply this doctrine.

This is, admittedly, a “delicate and difficult inquiry.”[237] But it is not an impossible one. Scholars and judges throughout American history have repeatedly attempted delegation bright-lines, and these can be generally grouped into four categories: specificity, significance, structure, and substance.[238] Ultimately, the concepts highlighted by the principle of agency help determine which is the most sound.

A. The Specificity Proposal

Some propose a bright-line for delegations based on their specificity. Open-ended delegations are impermissible, but delegations that give the executive narrow instructions are permissible. This is a leading contender for the nondelegation bright-line, as one version—the “intelligible principle” test—has been the Supreme Court’s standard measure from the mid-1940s to the present.[239] In fact, this was the measure the Court adopted in its most recent nondelegation case, FCC v. Consumer’s Research.[240] The phrase itself dates back to the Court’s reasoning in J. W. Hampton Jr. & Co. v. United States, which held that “[i]f Congress shall lay down by legislative act an intelligible principle to which the person or body authorized to fix such rates is directed to conform, such legislative action is not a forbidden delegation of legislative power.”[241] Yet over the many decades since the phrase was first used in Hampton, the Court has interpreted “intelligible principle” to include almost any statement Congress makes about how an agency should carry out its powers[242]—sometimes, “no more than the articulation of a policy aim or relevant policy concerns.”[243] If any such content can be found in the statute, or even in its legislative history or larger statutory scheme,[244] then Congress is not deemed to have delegated legislative power.[245]

Unfortunately, this is a test that no law can seem to fail. Since the mid-1940s, no statute has ever fallen short of the standard, and “[t]he problem is not (just) that justices are easy graders; the test itself is flawed.”[246] Some compare the standard to a “worn-out elastic band” which continually expands with use.[247] The Hampton Court may not have intended it, but this principle has vindicated a delegation of broadcasting authority to the FCC as long as it was in keeping with “public interest, convenience, or necessity,”[248] as well as a delegation to the Occupational Safety and Health Administration to create workplace standards “reasonably necessary or appropriate to provide safe or healthful employment and places of employment.”[249] Under this vacuous standard, the Court acknowledges the nondelegation in theory but pays “lip service in practice.”[250]

Yet there is a deeper problem with all specificity-based bright-lines that ignore the nature of legislative power: they assume that a law’s scope is what makes it law. If this is true, then the executive can promulgate rules coercively binding citizen action, but as long as the executive does so within the confines of some limiting principle, then the executive is no longer legislating.[251] So specificity-based standards ignore the fact that a law can be very specific and still be a law. As Justice Thomas puts it:

[T]he Constitution does not speak of “intelligible principles.” Rather, it speaks in much simpler terms: “All legislative Powers herein granted shall be vested in a Congress.” . . . 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.”[252]

Law’s essence is not its scope; law’s essence is that it coercively binds private action.[253] And thus, even a specific delegation is still a subdelegation of legislative power if it binds citizens. Considering the principle of agency, then, specific delegations of rulemaking authority remain the wrong power for the executive branch, and even when delegated in small doses. Thus, a specificity bright-line does not satisfy the principle of agency.

B. The Significance Proposal

Others propose a bright-line for delegations based on their significance. While it may overlap with specificity, significance focuses on the importance of the matter that has been delegated. Certain matters are too consequential to delegate away to a nonlegislative body, while minor matters can afford to be transferred. This approach traces back to Chief Justice Marshall’s opinion in Wayman v. Southard: he distinguished between the “important subjects” which the legislature must handle and the subjects “of less interest” for which the executive may “fill up the details.”[254]

The major questions doctrine is one contemporary significance-based standard, articulated by Justice Brett Kavanaugh in his denial of certiorari in Paul v. United States[255] and referenced again in passing in his concurrence in Consumer’s Research.[256] According to this standard, Congress must “expressly and specifically decide the major policy question itself”[257] when the executive exercises authority over issues of “great economic and political importance.”[258] Descriptions of such important issues vary—perhaps they must be central to the statute, economically or politically significant, or fall under the executive agency’s expertise.[259] Regardless of the precise formulation, these factors all highlight the significance of the delegated power.[260]

But grounding a bright-line in significance is also unsatisfactory, because the definition of “major question” is inescapably situational.[261] It depends on the audience, as indicated by the FDA’s 100,000-page hearing record about the precise percentage threshold for peanut butter.[262] It also depends on the values selected as most important—certain issues (such as religious liberty) may not be financially weighty but may be extremely morally important.[263] Even insignificant choices can have significant effects—a small change in metropolitan air pollution thresholds can dramatically affect an entire city.[264] So sometimes, “in complex, technical areas, the devil really is in the details.”[265] Defining acceptable delegation as “that which is not too important” just kicks the can down the road—there is no all-encompassing definition of “important.”

There is also a deeper problem with significance-based bright-lines—they assume that law’s significance is what makes it legislative.[266] They presume that if the executive makes rules governing major questions, the executive is legislating, but if the executive’s rules govern minor questions, the executive is not legislating. This ignores the fact that a law may deal with either major or minor questions and still be law. After all, the Vesting Clause “[does not] distinguish between major and minor policy decisions; it bans all delegation of legislative power.”[267] Law’s essence is rulemaking that coercively binds private action,[268] so an insignificant delegation—if it coercively binds private action—is still delegation of legislative power albeit in a small dose. Thus, “[t]he test of permissible delegation should look not to what quantity of power a statute confers but to what kind . . . .”[269] And considering the principle of agency, even insignificant subdelegations of rulemaking authority remain the wrong power for the executive branch. Thus, significance standards also cannot satisfy the principle of agency.

C. The Structural Proposal

A third possible distinction turns on the structural context of the delegation. Structural safeguards could be external to the executive (such as congressional oversight of executive rulemaking) or internal (such as formal processes for executive rulemaking). Some could claim that unstructured delegations of power are impermissible, but delegations exercised within appropriate structural safeguards are acceptable. One example is the “constitutional administration” theory, which distinguishes administrative activity into its legislative, executive, and judicial components and empowers each constitutional branch of government to check the administrative activities corresponding to it (implying that such delegations are now acceptable).[270]

While structural bright-lines may make delegation more palatable,[271] they fail to satisfy the principle of agency on two counts. First, they do not honor the substantive nature of law. Their proponents suffer from a double bind—on the one hand, they could concede that the power delegated is still legislative, but that conflicts with the Constitution’s stipulation that “[a]ll legislative powers”[272] be vested in Congress. On the other hand, they must assume that the structure around a power renders it no longer legislative. The power itself (to issue rules coercively binding citizen action) could be the exact same either way; the mere context of how it was developed or exercised would make it legislative or not. But the nature of legislative power is not “rulemaking exercised under appropriate supervision,” nor is it “rulemaking promulgated through set deliberative processes.” These contextual factors are important, but they do not make law what it is.

Second, such bright-lines do not honor the full set of unique qualities Congress possesses as the agent of the people. No matter how structured it may be, delegation to an executive agency is still delegation to the wrong institution. That institution may mimic some of Congress’s deliberative qualities, but unless it becomes a carbon-copy of Congress, it will lack the full set of qualities that promote the separation of powers, wise decision-making, responsiveness to the people, and limited government action[273]—all of which are tailored to making law directly.[274] Detached congressional supervision does not accomplish the same end, for congressmen have the option to look the other way when politically convenient.[275] Congress possesses specific qualities so it can make law, not so it can (optionally) supervise other institutions lacking those qualities as they make the law instead.[276] No matter how structured or supervised a delegation of coercive rulemaking power may be, it is still a delegation of real legislative power to the wrong institution lacking the necessary qualities for lawmaking.

So these three bright-line proposals are all inconsistent with the principle of agency. Admittedly, legislative delegation would make sense if Congress were delegating its own original power, because principals generally have the authority to delegate whatever they may do themselves.[277] In this case, it would not matter much the identity of the agent Congress chose, and it would make practical sense to delegate only specifically delineated powers, to delegate mostly minor questions, or to delegate within structural safeguards.[278] But delegation of legislative power is really an issue of subdelegation,[279] because Congress does not possess its own original power. It is the agent of the people, possessing particular power vested in it because of its particular legislative qualities. This power cannot be delegated away. No amount of specificity, significance, or structure can turn a legislative power into executive power, or an executive body into a legislative one. Even if a delegation is very specific, very small, or very supervised, it is still the wrong power going to the wrong agent with the wrong particular qualities.

D. The Substance Proposal

So specificity, significance, and structure all are inconsistent with the principle of agency. But there is one other possibility: substance. Rather than assessing a delegation’s magnitude or institutional context, this standard examines the content of the delegation itself: power that is legislative in nature may not be subdelegated, while non-legislative power may. The twofold definition of legislative power implies a twofold bright-line:[280] if a delegation gives the executive the power to issue rules coercively binding citizen action, or transfers any other power that the Constitution vests in Congress alone, then the delegation is not permitted. Otherwise, the delegation is permitted.[281]

Unlike the alternatives, the substance-based standard reflects that which makes legislative power legislative—it coercively binds citizen action or is otherwise designated by a nation’s constitution as such. It also honors the previously mentioned independent arguments for the nondelegation doctrine. It respects the text of the U.S. Constitution, which vests “[a]ll legislative powers”[282] in Congress. It protects the separation of powers more than just in name, because it labels legislative power according to its content instead of merely the actor or procedure carrying it out. It preserves the heart of political accountability, by keeping coercive power firmly in the hands of those most accountable to the people. Finally, and most fundamentally, it honors the principle of agency that describes the entire American political landscape. A specific power must remain in the hands of the specially qualified agent to whom the people delegated it.

Yet such a categorical standard naturally prompts an important question—if no legislative power can be delegated, then what power can Congress appropriately delegate? There are at least four kinds of appropriate delegation,[283] none of which transfer Congress’s essential or accompanying legislative powers.[284]

1. Delegation of Factfinding

Congress may delegate to the executive the power to determine facts. As Justice Gorsuch put it in his Gundy dissent, “once Congress prescribes the rule governing private conduct, it may make the application of that rule depend on executive fact-finding.”[285] This principle has a well-established history in American jurisprudence.[286] It dates back at least to the case The Cargo of the Brig Aurora v. United States,[287] in which the Supreme Court upheld a trade prohibition that was to cease automatically after the president made a factual determination that Great Britain and France had stopped violating neutral commerce.[288] In cases like this, the executive is not creating a rule constraining private action; the constraining rule already exists. He is merely determining whether certain facts that trigger it are indeed the case. In fact, factfinding is logically implied by law execution. After all, every law that takes effect on a certain date requires the executive to at least check the calendar to determine whether that date has arrived.[289] Similarly, if Congress passes a law contingent on a state of affairs it clearly defines, then it is the role of the executive and judiciary to determine whether those states of affairs exist.[290]

While factfinding always requires discretion, it is qualitatively different than legislative discretion: it is not based on will.[291] The legislature uniquely possesses the “will” involved in government operations,[292] so it implements its discretion about what it wants society to look like. But the executive does not answer questions about what it wants society to look like; it merely carries the “force” involved in government operations.[293] Accordingly, its discretion is limited to determining facts—irrelevant of its own will in the matter—which determine how to apply that force.[294] So factfinding is not a legislative power, as it involves no will-based discretion, is an inescapable part of enforcing any law, and confers no new power on the executive. Congress may delegate to the executive the power to find facts upon which enforcement of the law depends.

2. Delegation of Nonbinding Rulemaking

Congress may delegate to the executive the authority to issue rules that do not coercively bind citizen action, as long as they are not explicitly vested in Congress by the Constitution. Such powers are not uniquely legislative and thus are not part of the unique power Congress has been granted in virtue of its specific qualities. Several broad kinds of rulemaking fall under this category.

First, the executive may issue rules directing the internal operations of the executive branch if such rules do not coercively bind citizen action. For example, Congress may delegate to the postal service the authority to determine daily schedules for post office workers, but not the authority to determine the circumstances under which a citizen can be prosecuted for mail fraud. This is because the executive, like any employer, must give his employees directives. When these directives are promulgated generally to all employees within a large government agency, they look much like law.[295] But at most, they can only be conditions of employment; if an employee breaks such regulations, the question would be “not whether the disobedient officer could be prosecuted, but merely whether he could be removed.”[296] So the power to delegate nonbinding rules for executive operations is not a transfer of legislative power.[297] Second, the executive may issue rules regarding foreign policy which do not coercively bind citizen action. The executive is not exercising legislative power when dealing with entities outside of the law’s domain. For example, the executive can issue rules governing enemy aliens residing in the country during wartime if they are outside of the obligation and the protection of the law.[298] Third, the executive may issue rules governing the military.[299] Military law has long been considered distinct from civilian law and is only enforceable in the military justice system, so it is not, properly speaking, a rule coercively binding citizen action.[300] Even though military law uses legislative delegation, “[t]he relationship of military officers to their subordinates is not the model for the relationship of executive officers to the people of the United States.”[301]

These are just three examples, among others, of appropriate executive rulemaking which does not coercively bind citizen action. And the distinction between rules that bind citizen action and those that do not is one of the most important distinctions relevant to the current discussion about the nondelegation doctrine of paramount importance. Advocates for delegation often confuse the two, arguing that nonbinding delegations are precedent for binding delegations.[302] But nonbinding delegations are not delegations of legislative power.

3. Delegation of Powers not Uniquely Legislative

Congress may delegate to the executive the authority to perform acts which are not uniquely vested in Congress, even if Congress could appropriately choose to perform them. Certain ancillary powers may be constitutionally appropriate for either Congress or another branch to exercise.[303] For example, as Chief Justice Marshall acknowledged in Wayman v. Southard, Congress has the ability to regulate federal courts, but it can also grant courts the authority to follow the forms of process for the states where each court sits.[304] As Marshall put it, “It will not be contended that Congress can delegate to the courts or to any other tribunals powers which are strictly and exclusively legislative. But Congress may certainly delegate to others powers which the legislature may rightfully exercise itself.”[305] Thus, “[t]he question is whether there are certain kinds of rulemakings that have to be done by Congress, even if there are many other kinds of rulemaking that can be done by either.”[306] In the American context, those that have to be done by Congress are the power to issue rules coercively binding citizen action and all other powers exclusively vested in Congress by the Constitution. Other delegations do not transfer legislative power if they merely clarify that another branch may exercise power it already possesses.[307]

4. Delegation of the Means of Enforcement

Finally, Congress may delegate to the executive the authority to determine the means of enforcing a law. If such a determination does not add any new coercive provisions, it is not an exercise of legislative power. It merely selects one possible means of enforcing an already-existing coercive provision. So Congress must determine “whether” something is to be done, and it may leave the executive to determine “how” that is to be done.[308]

But this must be carefully defined. A simple distinction between “whether” and “how” does not inherently limit any delegation. Suppose that Congress passed a law prohibiting “all transactions in interstate commerce that fail to promote goodness and niceness.”[309] Congress would indeed have determined “whether” goodness and niceness are to be promoted, leaving the president to determine “how” to promote them. Yet this law is so vague that “any attempt to implement this law would amount to creation of a new law.”[310] But one key question helps determine when law enforcement transforms into law creation: which institution is actually promulgating the coercive provision?

Legislative power is the power to issue rules coercively binding citizen action. Unpromulgated laws cannot bind citizens, because otherwise citizens could not know their duties under the law.[311] Therefore, laws must be announced publicly,[312] and the institution that issues the coercive measures publicly is essentially the lawmaking body. So in order for Congress’s actions to be “legislative,” the original statutes must promulgate the coercive measures sufficiently without later executive clarification. If statutes are too vague for this, then the real promulgated laws are the executive clarifications, and thus the executive acts as the legislative body.

Courts have already been addressing this situation in the criminal context through an analogous principle:[313] the void-for-vagueness standard.[314] According to this test, if the original law enacted by elected legislators does not provide fair notice to the parties affected, then the law is too vague.[315] Initially, a “fair notice” requirement can apply to the executive enforcers. A delegation must provide a rule for executive action,[316] one which makes it possible to determine whether the executive has disobeyed it. Merely instructing agencies to act “reasonably” or “in the public interest” does not do this.[317] Yet primarily, “fair notice” applies to the citizens themselves. Congress must pass laws crafted such that citizens can tell whether they are violating the original statute without later executive or judicial narrowing of its terms. As one void-for-vagueness case put it, laws must “give the person of ordinary intelligence a reasonable opportunity to know what is prohibited, so that he may act accordingly.”[318] But not even a person of extraordinary intelligence could necessarily know whether he is violating a statute requiring his actions to be done “reasonably” or “in the public interest;” he must wait for executive regulations to clarify what those mean. And if the original congressional statute cannot serve as grounds for convicting a citizen in court, then Congress did not promulgate the law. If the executive clarifies the statute such that it can be used to convict a citizen, then the clearly promulgated coercive provision actually comes from the executive.[319] In this light, an “enforcement choice” clarifying a law promoting “goodness and niceness” would instead be a legislative decision.

Therefore, Congress may delegate enforcement choices to the executive as long as the original congressional law clearly promulgates the coercive provisions and provides fair notice to the executive actors and citizens the law affects.[320] Such delegations do not transfer the legislative authority to issue rules coercively binding citizen action, so they are consistent with the substantive bright-line for nondelegation.[321]

All in all, then, the substantive standard honors Congress’s constitutional status as the federal government’s only legislative body and its political status as the agent of the people. It allows delegations of nonlegislative responsibilities such as factfinding, nonbinding rulemaking, shared nonlegislative powers, and enforcement choices. But because the people have vested in Congress alone the power to issue rules coercively binding citizen action, all those coercive rules—and any other exclusively legislative powers vested by the Constitution—must come from Congress itself. Other bright-line attempts fail to reflect the principle of agency by focusing on ancillary aspects of delegation and overlooking the nature of legislative power and the specific qualities of the legislature. Yet because Congress is the agent of the people, vested with specific legislative power in virtue of its specific legislative qualities, its specific power cannot be subdelegated away. The line between appropriate and inappropriate delegation must be a substantive line between legislative and nonlegislative power.

V. Conclusion

The principle of agency does more than keep hats safe on trips around the world—it prompts reexamination of the political foundations of the American system. In doing so, it recasts much of the nondelegation conversation by requiring clear descriptions of the specific powers and specific qualities assigned to Congress as the people’s agent. It unifies common justifications for the doctrine by placing them within an overarching political framework. It refutes objections which overlook the substantive nature of legislative power and the specific legislative qualities Congress possesses. It also implies a robust, objective bright-line for courts to use, while showing that other proposals overlook the true nature of legislative power. Without a robust principle of agency, conversation about nondelegation tends to assume that Congress is a principal dispensing its own original authority. This assumption easily devolves into appeals to necessity, redefinitions of legislative power, or bright-line proposals that are too loose or subjective to resist inappropriate subdelegation. Ultimately, the principle of agency grounds the nondelegation conversation in fundamental political theory, not just technical legal details.

Nondelegation is deeply rooted in the structure of the American system. Perhaps this explains why the doctrine has never truly died. It has persisted through the American legal conversation despite courts watering it down and academics dismissing it altogether; “[i]ndeed, it seems someone catches a glimpse of this elusive phoenix stirring in the ashes every 20 years.”[322] Despite all the concerns about “our increasingly complex society”[323] seeming to demand legislative delegation, the “civics-book model of legislators legislating, executives executing, and judges judging has enormous intuitive—and legitimating—power.”[324] But without substantive accounts of Congress’s qualities and powers, the civics-book model will only go so far. Hopefully, a principled revival of the doctrine will do more than just stir the ashes.

Now a further question looms. Nondelegation is a hard pill to swallow not primarily for theoretical reasons, but for practical ones; it is incompatible with how modern government operates. Without legislative delegation, Congress must limit the scope of government lawmaking to that which it can directly legislate,[325] which calls into question the legitimacy of major aspects of the administrative state.[326] So if executive agencies should not issue rules coercively binding citizen action, then is “most of Government”[327] indeed unconstitutional? Should courts begin hacking away at the administrative state?

This argument for recasting the nondelegation doctrine is primarily one of theory, not of policy. If the argument from the principle of agency is true, then much of modern government is indeed unconstitutional; how courts should implement that truth is a separate question.[328] Yet regardless of implementation, nondelegation advocates must always stand by the logical conclusion of the doctrine they promote. Unfortunately, some advocates for a stronger nondelegation doctrine are quick to clarify that their proposals would not invalidate the modern administrative state.[329] This is curious, however, because the modern administrative state stands on the lax versions of the doctrine that have held sway for the last several decades. If the lax versions are misguided, it stands to reason that the modern administrative state is, as well. And nondelegation’s opponents notice the intellectual inconsistency when its advocates refuse to stomach this conclusion.[330] So nondelegation advocates should openly admit that the doctrine is logically inconsistent with the modern administrative state because Congress may not delegate to the executive the power to draft rules coercively binding citizen action. It is time for judges and legislators to remember that delegata potestas non potest delegari and to hold legislative delegations to be void. Power cannot be exercised contrary to the conditions upon which it was given, so attempts to do so must be ruled illegitimate. As Hamilton once expressed it:

There is no position which depends on clearer principles, than that every act of a delegated authority, contrary to the tenor of the commission under which it is exercised, is void. No legislative act therefore contrary to the constitution can be valid. To deny this would be to affirm that the deputy is greater than his principal; that the servant is above his master; that the representatives of the people are superior to the people themselves . . . .[331]

All in all, legislative subdelegation violates the principal-agent relationship between the American people and the American government. No servant should become greater than his master.

[1]  Catlin v. Bell, 4 Camp. 183 (1815).

[2]  See id. at 183-84.

[3]  See Gary Lawson, A Private-Law Framework for Subdelegation, in The Administrative State Before the Supreme Court 123, 127 (Peter J. Wallison & John Yoo eds., 2022).

[4]  Catlin, 4 Camp. at 184.

[5]  See id.

[6]  See infra Section II.

[7]  Technically, this is the “public nondelegation doctrine” because it prohibits Congress from delegating lawmaking authority to other government institutions, as opposed to the “private nondelegation doctrine,” which prohibits Congress from delegating lawmaking authority to private parties. Jacob D. Charles & A. H. Miller, Violence and Nondelegation, 135 Harv. L. Rev. F. 463, 463 (2022), https://harvardlawreview.org/forum/vol-135/violence-and-nondelegation [https://perma.cc/GT3Y-F3GT]. This thesis only addresses the public nondelegation doctrine, although many of the principles behind public nondelegation would apply equally—if not more so—to private nondelegation. If it is inappropriate to delegate lawmaking power to unelected government employees, then delegating this power to those who perform no role in government would seem even more egregious. Note also that this article uses the terms “laws/lawmaking” and “rules/rulemaking” interchangeably, unless specifically noted (e.g., “nonbinding executive rulemaking” is not lawmaking). In this article, both terms refer to provisions that coercively bind citizen action. See infra Section III.B.3.

[8]  See infra Section I.

[9]  See Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 42 (1825); Todd Gaziano & Ethan Blevins, The Nondelegation Test Hiding in Plain Sight: The Void-for-Vagueness Standard Gets the Job Done, in The Administrative State Before the Supreme Court 45, 47 (Peter J. Wallison & John Yoo eds., 2022) (“Every current member of the Supreme Court has agreed that the legislative nondelegation rule is crucial to the Constitution’s separation of powers.”).

[10]  Wayman, 23 U.S. (10 Wheat) at 42.

[11]  Whitman v. American Trucking Associations, Inc., 531 U.S. 457, 472 (2001).

[12]  11 U.S. (7 Cranch) 382 (1813).

[13]  Id. at 382.

[14]  Wayman, 23 U.S. (10 Wheat.) at 1.

[15]  Id. at 43.

[16]  Id.

[17]  Id.

[18]  276 U.S. 394 (1928).

[19]  Id. at 409.

[20]  Id.

[21]  Kristin E. Hickman, Nondelegation as Constitutional Symbolism, 89 Geo. Wash. L. Rev. 1079, 1089 (2021).

[22]  293 U.S. 388 (1935).

[23]  295 U.S. 495 (1935).

[24]  Pan. Refin., 293 U. S. at 406

[25]  Schechter, 295 U.S. at 521-23.

[26]  Pan. Refin., 293 U.S. at 414-15; Schechter, 295 U.S. at 531.

[27]  321 U.S. 414 (1944).

[28]  Id. at 420, 422, 426-27.

[29]  319 U.S. 190 (1943).

[30]  Id. at 216.

[31]  See Hickman, supra note 21, at 1090.

[32]  Id. at 1091.

[33]  Indus. Union Dep’t, AFL-CIO v. Am. Petroleum Inst. (The Benzene Case), 448 U.S. 607 (1980).

[34]  488 U.S. 361 (1989).

[35]  The Benzene Case, 448 U.S. at 641; Mistretta, 488 U.S. at 363-370. As Gary Lawson comments about Mistretta, “It was painfully obvious that the Court in Mistretta was trying to take the nondelegation doctrine off of the constitutional agenda for the foreseeable future.” Gary Lawson, Delegation and Original Meaning, 88 Va. L. Rev. 327, 329-30 (2002).

[36]  531 U.S. 457 (2001).

[37]  Id. at 465.

[38]  Gundy v. United States, 139 S. Ct. 2116 (2019).

[39]  Id. at 2124-30.

[40]  See, e.g., Hickman, supra note 21, at 1081. See also Eric A. Posner and Adrian Vermeule, Interring the Nondelegation Doctrine, 69 U. Chic. L. Rev. 1721, 1722-23 (2002).

[41]  See Hickman, supra note 21, at 1081 (referring to various scholars who have argued for the doctrine’s revival); Lawson, supra note 35, at 331 (referring to various scholars who have argued for the doctrine’s revival).

[42]  See, e.g., Lawson, supra note 35.

[43]  See, e.g., David Schoenbrod, Power Without Responsibility: How Congress Abuses the People through Delegation (1995).

[44]  See Philip Hamburger, Is Administrative Law Unlawful? 377-402 (2014).

[45]  See Indus. Union Dep’t, AFL-CIO v. Am. Petroleum Inst. (The Benzene Case), 448 U.S. 607, 673-75 (1980) (Rehnquist, J., concurring in the judgment); Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 486-87 (2001) (Thomas, J., concurring).

[46]  Lawson, supra note 35, at 332.

[47]  Id. at 330.

[48]  Gundy v. United States, 139 S. Ct. 2116, 2136-37 (2019) (Gorsuch, J., dissenting).

[49]  See, e.g., Hickman, supra note 21.

[50]  Justice Clarence Thomas and Chief Justice John Roberts joined Justice Gorsuch’s opinion, while Justice Alito, concurring in the judgment, said he would be willing to reconsider the Court’s approach to nondelegation. See Gundy, 139 S. Ct. at 2131 (Alito, J., concurring in the judgment). Although Justice Kavanaugh did not participate in Gundy, he noted in a denial of certiorari during the same year that “Justice Gorsuch’s thoughtful Gundy opinion raised important points that may warrant further consideration in future cases.” Ronald W. Paul v. United States, 140 S. Ct. 342, 342 (2019) (Kavanaugh, J., respecting the denial of certiorari).

[51]  See West Virginia v. EPA, No. 20-1530 (U.S. Jun. 30, 2022) (reining in the EPA’s delegated rulemaking authority because it affected major questions without clear congressional authorization).

[52]  See Jarkesy v. Sec. & Exch. Comm’n, No. 22-859 (U.S. Jun. 27, 2024) (holding that common lawsuits seeking punitive remedies must be tried by Article III courts, instead of Congress delegating such cases to administrative courts).

[53]  See Loper Bright Enters. v. Raimondo, 144 S. Ct. 2244 (2024) (overturning forty years of precedent directing courts to defer to reasonable executive agency interpretations of statutes, under the theory that when Congress passed an ambiguous statute, it was delegating interpretive authority to the implementing agency). Cf. Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984), rev’d, 144 S. Ct. 2244 (2024).

[54]  Hickman, supra note 21, at 1080.

[55]  No. 24-354 (cert. granted Nov. 22, 2024).

[56]  See Brief for Petitioner at I, Fed.Commc’ns. Comm’n v. Consumers’ Rsch., No. 24-354 (U.S. Jan. 8, 2025).

[57]  Fed. Commc’n Comm’n v. Consumers’ Rsch., No. 24-354 (U.S. Jun. 27, 2025).

[58]  Id. at 19.

[59]  Id. at 11 (Gorsuch, J., dissenting) (citing Gundy v. United States, 139 S. Ct. 2116, 2130 (2019)).

[60]  Id. at 11 (Gorsuch, J., dissenting).

[61]  Id. at 2 (Gorsuch, J., dissenting).

[62]  Id. at 37 (Gorsuch, J., dissenting).

[63]  See, e.g., Gary Lawson, The Rise and Rise of the Administrative State, 107 Harv. L. Rev. 1231, 1237-41 (1994).

[64]  See, e.g., id. at 1246-48.

[65]  Id. at 1247-48.

[66]  See Chevron U.S.A. Inc. v. Natural Res. Def. Council, Inc., 467 U.S. 837 (1984), rev’d, 144 S. Ct. 2244 (2024). However, because Chevron was overruled in June of 2024, this aspect of the debate may no longer be salient.

[67]  See Auer v. Robbins, 519 U.S. 452 (1997).

[68]  See, e.g., Lawson, supra note 63, at 1241-46.

[69]  See, e.g., id. at 1248-49. In his article, Lawson also critiques the fundamental problem that the administrative state has sprung up to address: the expansive powers that the federal government has assumed for itself in violation of the Constitution’s specifically enumerated powers. This is indeed a root cause of many of the smaller critiques, for those more detailed criticisms focus on specific measures that the federal government has adopted to manage the massive range of powers that it has assumed.

[70]  Gillian E. Metzger, 1930s Redux: The Administrative State Under Siege, 131 Harv. L. Rev. 1, 92 (2017).

[71]  Loper Bright Enters. v. Raimondo, 144 S. Ct. 2244 (2024).

[72]  See Metzger, supra note 70, at 94.

[73]  Gundy v. United States, 139 S. Ct. 2116, 2130 (2019).

[74]  See Hamburger, supra note 44, at 379.

[75]  Id.

[76]  See id.

[77]  See Lawson, supra note 3, at 127.

[78]  See Joseph Story, Commentaries on the Law of Agency, as a Branch of Commercial and Maritime Jurisprudence § 2 (6th ed. 1863).

[79]  Id. at § 3.

[80]  Id. at § 2.

[81]  Id. See also id. at § 6 (“In general it may be stated (as has been already intimated), as a rule of the common law, that, whenever a person has a power, as owner, or in his own right, to do a thing, he may do it by an agent.”).

[82]  Id. at § 13.

[83]  Id.

[84]  Hamburger, supra note 44, at 386.

[85]  Story, supra note 78, at § 13. The Latin translates as “delegated power cannot be [further] delegated.”

[86]  1 Edward Coke, The Second Part of the Institutes of the Laws of England 597 (1797). The Latin translates as “[they have] delegated power, which cannot be [further] delegated.”

[87]  1 Matthew Bacon, A New Abridgement of the Law 320 (Henry Gwillim, ed., 5th ed. 1798). Several editions of Bacon’s treatise emphasized this principle well into the 19th century. See Lawson, supra note 3, at 131-32.

[88]  Edward Sugden, A Practical Treatise on Powers 144-45 (1808). The Latin translates as “one to whom [power] has been delegated cannot [further] delegate.”

[89]  1 Samuel Livermore, A Treatise on the Law of Principal and Agent; and of Sales by Auction 54 (1818).

[90]  2 James Kent, Commentaries on American Law 495 (1827).

[91]  1 Floyd Mechem, A Treatise on the Law of Agency § 124 (2d ed. 1914).

[92]  See Lawson, supra note 3, at 133 (“I am not aware of any 18th- or 19th-century source that contradicts these authors on the basic agency law of subdelegation. . . . Founding-era private-law lawyers understood very well the problem of subdelegation of authority.”) (citation omitted).

[93]  Id. at 134-136. But see infra Section III.B.4 for why these exceptions do not apply to the political nondelegation principle.

[94]  See id. at 133.

[95]  Story, supra note 78, at § 17.

[96]  Id.

[97]  Id. at § 18.

[98]  Id. at § 126.

[99]  Lawson, supra note 3, at 128.

[100]  See, e.g., Metzger, supra note 70, at 72, 89–90; Posner & Vermeule, supra note 40, at 1723.

[101]  Hickman, supra note 21, at 1086-88. Unless indicated otherwise by the context, this article uses the terms “(legislative) delegation” and “(legislative) subdelegation” interchangeably to refer to the same phenomenon: the legislative branch delegating legislative duties to a non-legislative branch. “Delegation” is the commonly used term in the literature; “subdelegation” is a more precise description of what is happening in such cases.

[102]  Admittedly, concerns about political theory are only one aspect of the debate. Lively academic disputes swirl over the doctrine’s historical grounding. See, e.g., Ilan Wurman, Nondelegation at the Founding, 130 Yale L.J. 1490 (2021); Keith E. Whittington & Jason Iuliano, The Myth of the Nondelegation Doctrine, 165 U. Pa. L. Rev. 379 (2017). These debates are necessary contributions to the conversation, but—while not dismissing the relevance of historical analysis—the aim of this article is to analyze the doctrine’s grounding in political theory.

[103]  U.S. Const. art. I, § 1; id. art. II, § 1; id. art. III, § 1.

[104]  See Whitman v. Am. Trucking Ass’ns, 531 US 457, 472 (2001).

[105]  Some may claim that the Necessary and Proper Clause—which empowers to Congress to “make all Laws which shall be necessary and proper for carrying into Execution” the constitutionally-granted powers—justifies delegation if it is “necessary and proper” to handle modern government responsibilities. U.S. Const. art. I, § 8. However, as Gary Lawson observes, actions justified by the Necessary and Proper Clause must both be “necessary” and “proper.” See Lawson, supra note 35, at 347. An array of historical legal sources show that “proper” means “within one’s own jurisdiction,” so any “proper” action must accord with how power is already allocated among the branches of the federal government. Id. at 347-48. Thus, delegating legislative power to a non-legislative branch would not be “proper.”

[106]  U.S. Const. art. I, § 1; id. art. II, § 1; id. art. III, § 1.

[107]  Hamburger, supra note 44, at 387.

[108]  See U.S. Const., art. II, § 2; id. art III, § 1.

[109]  Hamburger, supra note 44, at 387.

[110]  Lawson, supra note 35, at 336.

[111]  See id. at 337.

[112]  Hamburger, supra note 44, at 379.

[113]  See 1 Baron de Montesquieu, The Spirit of Laws 216 (1750).

[114]  The Federalist No. 47 at 249 (James Madison) (George W. Carey & James McClellan eds., 2001). This was so important to Madison that he admitted that—if the Constitution violated it—“no further arguments would be necessary to inspire a universal reprobation of the system.” Id.

[115]  See Wurman, supra note 102, at 1526.

[116]  The Federalist No. 57, at 297 (James Madison) (George W. Carey & James McClellan eds., 2001).

[117]  See generally Jonathan Adler, A “Step Zero” for Delegations, in The Administrative State Before the Supreme Court 161, 165 (Peter J. Wallison & John Yoo eds. 2022); David Schoenbrod, A Judicially-Manageable Test to Restore Accountability, in The Administrative State Before the Supreme Court 346, 346 (Peter J. Wallison & John Yoo eds. 2022); Cf. U.S. Const. art. I, § 5.

[118]  See Adler, supra note 117, at 164; see also Schoenbrod, supra note 117, at 350.

[119]  It is important for nondelegation advocates not to overstate the case here. If the argument stopped at mere claims of popular accountability, the doctrine would instead be undermined, for to maximize popular accountability, Congress would have to delegate its decision-making back to national popular votes for every issue. This indicates that the doctrine cannot merely rest on a single, isolated political principle such as popular accountability. Rather, its true grounding is in the nature of a legislative body itself—a nature which carefully balances principles such as popular accountability against opposing principles such as selective membership. See The Federalist No. 10, at 46 (James Madison) (George W. Carey & James McClellan eds., 2001). If legislative power is delegated to any institution too far on either side of this balance, the nondelegation doctrine steps in to intervene. Thus, by forcing one to examine all of the specific qualities of the legislature, the principle of agency grounds the doctrine much more robustly than isolated political principles do.

[120]  John Locke, The Second Treatise of Government, in John Locke: Political Writings 261, § 141 (David Wootton, ed., Penguin Books 1993) (1690). Admittedly, some scholars have argued that Locke’s statement does not apply to the nondelegation debate, because they claim Locke distinguishes between the terms “transfer” (permanent alienation of power, which is impermissible) and “delegate” (temporary cession of power that can be resumed, which is permissible). See Julian Davis Mortenson & Nicholas Bagley, Delegation at the Founding, 121 Colum. L. Rev. 277, 307-08 (2021). However, Ilan Wurman provides a cogent historical counterargument, showing that founding-era sources used the terms interchangeably and that Mortenson and Bagley’s distinction between the two would render the term “transfer” meaningless. See Wurman, supra note 102, at 1518-1522.

[121]  Thomas Jefferson, Notes on the State of Virginia 132 (Lilly & Wait 1832).

[122]  Thomas Cooley argued that the legislature cannot “substitute the judgment, wisdom, and patriotism of any other body for those to which alone the people have seen fit to confide this sovereign trust.” 1 Thomas Cooley, A Treatise on the Constitutional Limitations 224 (8th ed. 1927). Charles Burdick drew attention to the “fundamental principle of American constitutional law that the legislative branch of the government cannot delegate its essential legislative function to any other agency.” Charles K. Burdick, The Law of the American Constitution: Its Origin and Development § 60 (1922).

[123]  Gundy v. United States, 139 S. Ct. 2116, 2133 (2019) (Gorsuch, J., dissenting) (alteration in original).

[124]  Lawson, supra note 3, at 131.

[125]  See id. at 127 (“There are good reasons to think that the Constitution draws on private-law background norms for much of its meaning, and the subdelegation problem is an excellent candidate for elaboration in private-law terms.”).

[126]  See Hamburger, supra note 44, at 377.

[127]  Id. at 385.

[128]  Commonwealth ex rel. McClain v. Locke (Locke’s Appeal), 72 Pa. 491, 494 (1873).

[129]  4 Del. 479, 489 (Del. 1847).

[130]  6 Pa. 507, 515 (1847) (“Among the primal axioms of jurisprudence, political and municipal, is to be found the principle that an agent, unless expressly empowered, cannot transfer his delegated authority to another, more especially when it rests in a confidence, partaking the nature of a trust, and requiring for its due discharge, understanding, knowledge, and rectitude. The maxim is, delegata potestas non potest delegari.”).

[131]  10 Ill. (5 Gilm.) 1, 11 (1848) (“To the general assembly have the people delegated the legislative powers of the government, only limited and controlled by the federal and State constitutions, and it is insisted that these powers can not be delegated to any body of men or any portion of the people, upon the principle that delegated powers can not be delegated.”).

[132]  15 Barb. 112, 116 (N.Y. Gen. Term 1851) (“[A legislator] cannot delegate to others the trust which has been expressly confided to him, by reason of his supposed knowledge and sound judgment. Delegata potestas, non potest delegati [sic], is a settled maxim of the common law, in full force at the present day; and never more applicable than to the case of a legislator . . . .”).

[133]  Mechem, supra note 91, at § 26.

[134]  See Hamburger, supra note 44, at 386.

[135]  Cf. Locke, supra note 120, at § 136.

[136]  See id. at § 136.

[137]  See Gundy v. United States, 139 S. Ct. 2116, 2135 (2019) (Gorsuch, J., dissenting).

[138]  Cf. Locke, supra note 120, at § 136.

[139]  See Gundy, 139 S. Ct. at 2135 (Gorsuch, J., dissenting).

[140]  The Federalist No. 53, at 279 (James Madison) (George W. Carey & James McClellan eds. 2001).

[141]  The Federalist No. 52, at 273 (James Madison) (George W. Carey & James McClellan eds. 2001).

[142]  The Federalist No. 70, at 365 (Alexander Hamilton) (George W. Carey & James McClellan eds. 2001).

[143]  The Federalist No. 55, at 288 (James Madison) (George W. Carey & James McClellan eds. 2001).

[144]  The Federalist No. 62, at 321 (James Madison) (George W. Carey & James McClellan eds. 2001). Granted, the constituency of a modern senator is no longer his state legislature as it was when The Federalist was written. Still, a modern senator must look to his whole state’s interests more than a representative from a particular district within the state, so he does still represent his whole state. In that sense, even today’s congressional legislation requires assent from both the people and the states.

[145]  See Larry Alexander & Saikrishna Prakash, Reports of the Nondelegation Doctrine’s Death Are Greatly Exaggerated, 70 U. Chic. L. Rev. 1297, 1301 (2003); see also Gundy v. United States, 139 S. Ct. 2116, 2134 (2019) (Gorsuch, J., dissenting).

[146]  See The Federalist No. 47, supra note 114, at 249 (James Madison).

[147]  The Federalist No. 73, at 381 (Alexander Hamilton) (George W. Carey & James McClellan eds. 2001).

[148]  See Wurman, supra note 102, at 1525.

[149]  The Federalist No. 70, supra note 142, at 363 (Alexander Hamilton).

[150]  Id. at 363, 366-67.

[151]  The Federalist No. 52, supra note 141, at 273 (James Madison).

[152]  The Federalist No. 78, at 402 (Alexander Hamilton) (George W. Carey & James McClellan eds. 2001); see also The Federalist No. 79, at 408 (Alexander Hamilton) (George W. Carey & James McClellan eds. 2001).

[153]  See Gundy v. United States, 139 S. Ct. 2116, 2134 (2019) (Gorsuch, J., dissenting).

[154]  See Adler, supra note 117, at 164.

[155]  See Gundy, 139 S. Ct. at 2134 (Gorsuch, J., dissenting).

[156]  Gaziano & Blevins, supra note 9, at 49; see also Gundy, 139 S. Ct. at 2134 (Gorsuch, J., dissenting).

[157]  See Hamburger, supra note 44, at 386.

[158]  See The Federalist No. 37, at 182-83 (James Madison) (George W. Carey & James McClellan eds. 2001); Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 46 (1825).

[159]  Madison still indicates his belief in a distinct legislative power. See The Federalist No. 48, at 256 (James Madison) (George W. Carey & James McClellan eds. 2001). John Marshall similarly acknowledges that legislative power has a distinct nature. See Wayman, 23 U.S. at 42 (“It will not be contended that Congress can delegate to the Courts, or to any other tribunals powers which are strictly and exclusively legislative.”).

[160]  The qualifier “coercive” is important, as it refers to laws or rules for which there are legal penalties for disobedience. As Philip Hamburger puts it, legislative acts “purport to bind subjects not merely in the sense of reaching a settled decision about them, but in the deeper sense of legally obliging, constraining, or interfering with them.” Hamburger, supra note 44, at 2. Certainly, many kinds of government action can affect citizen action by altering the environment in which citizens live. However, they only coercively bind citizen action if they impose legal penalties for noncompliance. Under this interpretation, then, imposing a tax or passing a criminal statute do coercively bind citizen action, while conducting factfinding, issuing regulations governing internal government operations, and distributing government benefits do not coercively bind citizen action.

[161]  1 William Blackstone, Commentaries *52, see also id. *44.

[162]  Id. at *52.

[163]  Id. at *53.

[164]  Locke, supra note 120, at § 141.

[165]  Id. at § 9.

[166]  Id. at § 212.

[167]  The Federalist No. 75, at 388 (Alexander Hamilton) (George W. Carey & James McClellan eds. 2001).

[168]  Id.

[169]  The Federalist No. 78, supra note 152, at 402 (Alexander Hamilton).

[170]  James Madison, The Report of 1800, in 17 The Papers of James Madison 303, 324 (David B. Mattern, J.C.A. Stagg, Jeanne K. Cross & Susan Holbrook Perdue eds., 1991), https://founders.archives.gov/documents/Madison/01-17-02-0202 [https://perma.cc/7YEG-QE4K].

[171]  Gundy v. United States, 139 S. Ct. 2116, 2133 (2019) (Gorsuch, J., dissenting); see also Dep’t of Transp. v. Ass’n of Am. Railroads, 135 S. Ct. 1225, 1242-44 (2014) (Thomas, J., concurring in the judgment) (marshalling an array of historical sources to show that it is particularly the legislature’s role to formulate “generally applicable rules of private conduct”).

[172]  See Hamburger, supra note 44, at 84.

[173]  Mark Chenoweth and Richard Samp, Reinvigorating Nondelegation with Core Legislative Power, in The Administrative State Before the Supreme Court 81, 102 (Peter J. Wallison & John Yoo eds. 2022).

[174]  See id.

[175]  Hamburger, supra note 44, at 84. One line of questioning helps clarify the import of this definition: “How far can a non-legislator proceed on a certain course of action before we would say he is making law?” Certainly, if an FDA official writes a letter to a friend advising against a certain drug, that would not be law. Nor if the official issues a public proclamation expressing his disapproval of the drug. Nor if the official issues a rule to his employees to prioritize inspections of this particular drug. However, if the official issues a rule legally prohibiting pharmacists from selling the drug, he is doing something qualitatively different from before. He is, in Hamilton’s words, “prescrib[ing] the rules by which the duties and rights of every citizen are to be regulated”—at least, for every citizen involved in the manufacture, sale, and consumption of the drug. The Federalist No. 78, supra note 152, at 402 (Alexander Hamilton). Thus, there must be something uniquely legislative about the power to issue rules coercively binding citizen action.

[176]  Hamilton himself described legislative power in a manifold way, recognizing both the essential power and an example of an accompanying power. See The Federalist No. 78, supra note 152, at 402 (Alexander Hamilton).

[177]  U.S. Const., art. I, § 8.

[178]  Among contemporary theorists, Philip Hamburger relies exceptionally heavily on the binding-subjects standard to define legislative power. See Hamburger, supra note 44, at 85 (“When executive rules purport to bind subjects, they create a regime of legislation outside the law and the lawmaking institutions and processes established by the Constitution. In contrast, executive acts that do not alter the binding duties of subjects typically are not legislative and therefore are not usually part of the problem examined here.”). However, even he concedes that the binding-subjects standard cannot be all-encompassing. See id. at 84-85.

Despite this concession, Hamburger tends to focus exclusively on that “core” legislative power at the risk of neglecting those accompanying legislative powers. Such a narrow description of legislative power risks promoting a standard that does not accurately reflect the text of the Constitution—whatever one’s ideas of the document’s philosophical substructure, one’s theories must still accord with the document’s actual text. See Lawson, supra note 35, at 345. Thus, it is important to maintain that congressional power consists of both binding power and other accompanying powers granted to it in the Constitution, both of which are equally “legislative” in the American context, and neither of which can be delegated. As Chenoweth and Samp conclude, a court’s nondelegation analysis “should not be limited to cases involving controls on private conduct. The Vesting Clause prohibits all delegation of legislative power.” See Chenoweth & Samp, supra note 173, at 98 (emphasis in original).

[179]  U.S. Const. art. I, §§ 8-9.

[180]  See Alexander & Prakash, supra note 145, at 1308. But three notes of clarification are necessary. First, the charge to “include whatever the Constitution says” does not exclude powers that are unstated but obviously implied by the text. For example, in addition to taxing, spending, and drafting criminal statutes, Chenoweth and Samp identify two other powers in their list of “core legislative powers” which cannot be delegated: resolving policy debates and checking the executive branch. See Chenoweth & Samp, supra note 173, at 97. Second, given how necessary it is to incorporate the Constitution’s details, one might wonder why the definition of legislative power does not simply reduce to “the powers granted to Congress in the Constitution.” If this were so, there would be no way to identify whether actions not mentioned in the Constitution were legislative in nature. Executive agencies could promulgate rules restricting access to the internet, for example, and there would be no grounds to brand this as executive lawmaking, because the Constitution does not mention the internet and so it could not be considered a legislative power. Therefore, since there are more conceivable legislative powers than those mentioned in the Constitution, it is necessary to have a definition of legislative power that identifies when another branch of government is carrying out those as well. Third, the argument presented in this article does not condone as constitutional every possible means of coercively binding private action. For example, if Congress passed a law forbidding the founding of any new businesses in the nation after January 1, 2027, that would indeed coercively bind citizen action—and thus would be legislative—but it would not be constitutional. Thus, this argument by no means approves every conceivable action Congress could take to further either the essential or accompanying powers. It merely claims that, whether constitutional or not, they are appropriately called “legislative.”

[181]  Gaziano & Blevins, supra note 9, at 49.

[182]  See The Federalist No. 70, supra note 142, at 363 (Alexander Hamilton).

[183]  Admittedly, it is possible to modify executive agency procedures such that they possess “copies” of legislature features. For example, many federal regulations are made through a notice-and-comment process, which mimics some features of deliberation. Yet in order to completely copy all of the safeguards the framers deemed necessary to put in the Constitution, one would have to create a legislative body that is simply a copy of Congress. Anything short of that implies that at least some of the framers’ safeguards were unnecessary to include in the Constitution. See generally Gaziano & Blevins, supra note 9, at 61-62.

[184]  The Federalist No. 70, supra note 142, at 363 (Alexander Hamilton).

[185]  Id.

[186]  Cf. The Federalist No. 62, supra note 144, at 321 (James Madison) (“[T]he facility and excess of law-making seem to be the diseases to which our governments are most liable . . . .”).

[187]  U.S. Const., art I. § 1.

[188]  See supra note 105.

[189]  See Posner & Vermeule, supra note 40. See also infra Section III.C for a refutation of their claim.

[190]  See Lawson, supra note 3, at 128-29.

[191]  See id. at 129.

[192]  See id.

[193]  See id.

[194]  See id. at 134-35.

[195]  Id. at 135; see also Lawson, supra note 35, at 347-48.

[196]  Lawson, supra note 3, at 134 (citation omitted).

[197]  See id. at 135-36.

[198]  23 U.S. (10 Wheat.) 1, 43 (1825).

[199]  See Lawson, supra note 3, at 143. Lawson, although a strong advocate for nondelegation himself, presents a prime example of this difficulty. Drawing from private-law agency cases, he claims that matters which require “general skill and knowledge” should not be delegated, while matters which require “localized skill or knowledge” would be matters of less interest. Id. at 142. Admittedly, Lawson is here attempting to offer a perspective that would be considered reasonable given the realities of existing doctrine, rather than offering a definitive statement of interpretative truth. Yet some may construe it as the latter, so it is important to understand why limiting delegations to “ministerial matters of less interest” requiring “localized knowledge” is not a robust limiting principle. This distinction is grounded in the circumstances that a law covers, not in the nature of law itself. In other words, it does not consider whether the delegated power coercively binds private action, but rather whether it requires localized knowledge to implement it (whether coercive or not).

This is a problem when congressional power expands, because then Lawson’s proposed standard would justify increases in delegation. If Congress passes an unprecedently intrusive law requiring unprecedented amounts of localized knowledge to implement, then this standard would authorize unprecedented amounts of legislative delegation to cover all of those “ministerial” responsibilities, regardless of whether that placed unprecedented amounts of coercive legislative power into executive hands. Thus, this argument is only a conditional argument for nondelegation. (Contrastingly, a nondelegation argument from the nature of legislative power can prevent this—Congress would be limited to only passing legislation whose coercive provisions were limited enough that the legislature could draft them all itself.)

Furthermore, Lawson’s proposed standard of “limited knowledge” remarkably resembles the argument that justifies administrative law in the first place: Congress has limited capacity and cannot possibly handle all legislative responsibilities in our complex modern age. Thus, pro-delegation advocates could borrow Lawson’s idea and simply claim that the current administrative state requires so much localized knowledge that delegation is appropriate in all its many forms. (Of course, if Congress is indeed intended to draft all coercive laws, then the number of government operations requiring coercive laws should be limited to what Congress can itself enact.) Therefore, a “localized knowledge” framing of “matters of less interest” does not stand firm as a robust limiting principle.

[200]  See Michael Rappaport, A Two-Tiered and Categorical Approach to the Nondelegation Doctrine, in The Administrative State Before the Supreme Court 195, 207 (Peter J. Wallison & John Yoo eds. 2022) (“Marshall [in Wayman v. Southard] distinguishes between ‘important subjects which must be entirely regulated by the legislature’ and those of ‘less interest,’ which need not be. But nothing in Marshall’s opinion is clearly inconsistent with reading it as indicating that all laws that involve matters governed by the strict tier [i.e., affecting the private rights of life, liberty, and property] are important subjects that must be entirely legislated by Congress.”) (emphasis in original).

[201]  These are among the most prominent arguments from legal and political theory. Again, arguments about legal history and founding-era practice constitute a vigorous parallel debate, one which is also an important facet of this discussion. See supra note 102.

[202]  Metzger, supra note 70, at 7.

[203]  Id. at 91.

[204]  U.S. Const., art. II, § 3.

[205]  Metzger, supra note 70, at 72, 89-90.

[206]  Mistretta v. United States, 488 U.S. 361, 372 (1989).

[207]  Even so, there is reason to doubt this assumption in all cases—after all, Congress directly enacts the extensive Internal Revenue Code. See Gaziano & Blevins, supra note 9, at 83-84.

[208]  Metzger observes that many opponents of the administrative state are strikingly reluctant to argue against delegations that seem necessary for the modern world—which, in her opinion, require the administrative state. See Metzger, supra note 70, at 7. However, the inconsistencies of the administrative state’s opponents do not prove that they are wrong, but could rather suggest that they should be more consistent.

[209]  See U.S. Const. art. II, § 2, id. art III, § 1; see also Hamburger, supra note 44, at 387.

[210]  U.S. Const., art. I, § 1; see also Hamburger, supra note 44, at 387.

[211]  See Story, supra note 78, at § 2.

[212]  Posner & Vermeule, supra note 40, at 1723.

[213]  Id.

[214]  Id. at 1725.

[215]  Id. at 1725-26.

[216]  In addition to the problems detailed in this section, Posner and Vermeule’s position leads to a paradox: according to them, the Constitution bars Congress from giving the Secretary of Commerce one vote out of 435 votes in the House of Representatives. However, they imply that the Constitution allows the Secretary of Commerce to promulgate rules binding the entire nation on his own—which is the same effect as having a majority (218 votes) in the House of Representatives. It is unclear why the Constitution would scrupulously forbid one and yet willingly allow the other. See Alexander & Prakash, supra note 145, at 1300.

[217]  Id. at 1302.

[218]  Id. at 1302-3.

[219]  Id. at 1303.

[220]  See id. at 1318.

[221]  Wurman, supra note 102, at 1537.

[222]  See Mistretta v. United States, 488 U.S. 361, 415 (1989) (Scalia, J., dissenting).

[223]  Id.

[224]  See Cass Sunstein, Nondelegation Canons, 67 U. Chic. L. Rev. 315, 326-27 (2000).

[225]  See Lawson, supra note 35, at 354-55 (referencing scholars making the bright-line objection).

[226]  Sunstein, supra note 224, at 327. Interestingly, this argument has prompted some conservative judges to join their liberal colleagues in rejecting a robust nondelegation doctrine—the former because of their faith in the competence of administrative experts, and the latter because of their resistance to judicial activism based on a judge’s personal preferences (which such a vague doctrine would seem to invite). See Gaziano & Blevins, supra note 9, at 54.

[227]  See Gaziano & Blevins, supra note 9, at 52.

[228]  Chenoweth & Samp, supra note 173, at 88.

[229]  See id.

[230]  The Federalist No. 78, supra note 152, at 403.

[231]  Cf. Richard Epstein, The Dubious Morality of Modern Administrative Law 34 (2020).

[232]  See Chenoweth & Samp, supra note 173, at 86.

[233]  See, e.g., Mistretta v. United States, 488 U.S. 361, 415 (1989) (Scalia, J. dissenting) (claiming simultaneously that “It is difficult to imagine a principle more essential to democratic government than that upon which the doctrine of unconstitutional delegation is founded,” but also that,

while the doctrine of unconstitutional delegation is unquestionably a fundamental element of our constitutional system, it is not an element readily enforceable by the courts. Once it is conceded, as it must be, that no statute can be entirely precise, and that some judgments, even some judgments involving policy considerations, must be left to the officers executing the law and to the judges applying it, the debate over unconstitutional delegation becomes a debate not over a point of principle, but over a question of degree.

[234]  See, e.g., Metzger, supra note 70, at 88; Posner & Vermeule, supra note 40, at 1728-29; Peter H. Schuck, Delegation and Democracy: Comments on David Schoenbrod, 20 Cardozo L. Rev. 775, 791 (1998-1999).

[235]  Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 472 (2001).

[236]  See Gaziano & Blevins, supra note 9, at 47.

[237]  Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 46 (1825) (observing that “[t]he difference between the departments undoubtedly is that the legislature makes, the executive executes, and the judiciary construes the law; but the maker of the law may commit something to the discretion of the other departments, and the precise boundary of this power is a subject of delicate and difficult inquiry, into which a court will not enter unnecessarily.”).

[238]  These four categories do not include every possible bright-line, but most major approaches fall into one or another of them. One notable outlier is the self-admittedly circular bright-line proposed by Gary Lawson in 2002. See Lawson, supra note 35. Examining Chief Justice Marshall’s distinction between important subjects and matters of less interest, Lawson argues that this boils down to a single rule: “Congress must make whatever decisions are important enough to the statutory scheme in question so that Congress must make them.” Id. at 361. Now, if circularity is the best the doctrine has to offer, then its advocates should do as Lawson does and present the idea as credibly as possible. But if circularity is indeed the best the doctrine has to offer, then both liberal and conservative judges have good reason to reject it outright. Circular measures should only be a matter of last resort when there are no other standards that would communicate something substantive to the judiciary other than “you decide.” Thankfully, it is possible to develop a noncircular substantive standard, one which may not clearly slice through all ambiguous situations, but which certainly draws more of an objective line than laying aside the project and admitting circularity. See infra Section IV.D. Notably, Lawson seems to have backed away from this circular standard in recent years. In a book published in 2022, he pivots to championing the principle of agency within the nondelegation conversation. See Lawson, supra note 3, at 124-25 (summarizing his former view), and id. at 126-52 (attempting to deduce a substantive, noncircular doctrine from the nature of the Constitution as a fiduciary document).

[239]  See Chenoweth & Samp, supra note 173, at 81. Other specificity-based standards have been proposed as well; see, for example, Martin Redish’s “political commitment” principle. He argues that since legislators must be accountable to their constituents, their statutes must be specific enough to communicate their position on the issue at hand. See Martin H. Redish, The Constitution as Political Structure 136-37 (1995). This is similar to the proposal offered in this thesis because it relies on the idea of communication (cf. “promulgation”) of substantive content to the citizens. See infra Section IV.D. Yet it does not quite identify the core of the issue. It focuses on whether the legislators’ stances are promulgated, rather than whether the law’s binding provisions are promulgated. Opting for the latter approach will achieve the former automatically and is closer attuned to what law actually is. In a government of laws and not of men, citizens are bound by the law’s binding provisions, not by their legislators’ intentions.

[240]  Fed.Commc’ns Comm’n v. Consumers’ Rsch., No. 24-354, slip op. at 19 (U.S. Jun. 27, 2025).

[241]  J. W. Hampton, Jr. & Co. v. United States, 276 U.S. 394, 409 (1928).

[242]  Chenoweth & Samp, supra note 173, at 81.

[243]  Adler, supra note 117, at 171.

[244]  Hickman, supra note 21, at 1091

[245]  Chenoweth & Samp, supra note 173, at 81. Admittedly, the current scope of the intelligible principle test reaches far beyond what Taft likely intended in 1928.

[246]  Id.

[247]  Id. at 89. Those who are well-versed in the Court’s nondelegation cases may find Chenoweth and Samp’s rehearsal of the Court’s standard rationale humorously familiar.

Indeed, the Court’s discussion of the intelligible-principle test often consists of little more than reciting rote assertions: that the Court has not invalidated a statute under the nondelegation doctrine since 1935; that since then it has upheld, “without deviation, Congress’s ability to delegate power under broad standards”; and that the challenged statute provides at least as much guidance regarding how it is to be implemented as did statutes previously deemed to provide the requisite intelligible principles. Id. (citation omitted).

[248]  Nat’l. Broad. Co. v. United States, 319 U.S. 190, 216 (1943). See also Hickman, supra note 21, at 1090.

[249]  Indus. Union Dep’t, AFL-CIO v. Am. Petroleum Inst. (The Benzene Case), 448 U.S. 607, 642-46 (1980).

[250]  Chenoweth & Samp, supra note 173, at 81.

[251]  Of course, this assumes the charitable version of the intelligible principle test: “If a statute supplies such a principle, the test deems Congress not to have delegated legislative power.” Id. (emphasis in original). By claiming that delegations no longer transfer legislative power if they are specific enough, this makes out specificity as a defining feature of what it is to be a law. Now, this criticism could be dodged by claiming that power still remains legislative power when it is delegated under a specific intelligible principle. In this case, though, the advocate would have to explain why legislative power is being delegated under a Constitution that explicitly reserves “all legislative power” to the “Congress of the United States.” U.S. Const. art. I, § 1, see also Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 472 (2001).

[252]  Whitman, 531 U.S. at 487 (Thomas, J., dissenting) (citation omitted).

[253]  See supra Section III.B.3.

[254]  Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 43 (1825).

[255]  Ronald W. Paul v. United States, 140 S. Ct. 342, 342 (2019) (Kavanaugh, J., respecting the denial of certiorari).

[256]  Fed. Commc’n Comm’n v. Consumers’ Rsch., No. 24-354, slip op. at 9 (U.S. Jun. 27, 2025) (Kavanaugh, J., concurring).

[257]  Paul, 140 S. Ct. at 342.

[258]  Id. at 342. Admittedly, this was one of two options Justice Kavanaugh presented; the other was that Congress may “expressly and specifically delegate to the agency the authority both to decide the major policy question and to regulate and enforce.” However, he seems to immediately doubt this second category. Further, the way Justice Kavanaugh incorporates the major questions idea into his argument indicates he is sympathetic to it. See Hickman, supra note 21, at 1125.

[259]  See Hickman, supra note 21, at 1124.

[260]  There are other methods of determining whether a power could be “major enough” to prevent its delegation. David Schoenbrod once suggested that the courts replace the intelligible principle standard and adopt a significance standard with a very precise definition: “significant regulatory action.” According to Executive Order 12,866, this would be action that has an “annual effect on the economy of $100 million or more.” If any regulation increases or decreases costs by that amount, it would be considered significant and trigger the nondelegation doctrine. See David Schoenbrod, Consent of the Governed: A Constitutional Norm that the Court Should Substantially Enforce, 43 Harv. J. L. Pub. Pol’y 213, 258-65 (2020). But this standard singles out only one interpretation of significance—economic significance—leaving other truly significant factors such as morality and social consequences by the wayside. Further, it does not solve the fundamental problem: the importance of the regulated subject is not what makes legislative action legislative.

[261]  See Hickman, supra note 21, at 1125-26.

[262]  See id. at 1118-20.

[263]  See Chenoweth & Samp, supra note 173, at 96.

[264]  See Adler, supra note 117, at 173.

[265]  Id.

[266]  See supra note 251.

[267]  Chenoweth & Samp, supra note 173, at 96.

[268]  See supra Section III.B.3.

[269]  David Schoenbrod, The Delegation Doctrine: Could the Court Give it Substance?, 83 Mich. L. Rev. 1223, 1227 (1985).

[270]  See Ilan Wurman, Constitutional Administration, 69 Stan. L. Rev. 359, 361-365 (2017).

[271]  See Chenoweth & Samp, supra note 173, at 82.

[272]  U.S. Const, art. I, § 1.

[273]  See Gaziano & Blevins, supra note 9, at 62.

[274]  See supra Section III.B.2.

[275]  See Schoenbrod, supra note 117, at 350.

[276]  Some may also claim that Congress theoretically reserves the right to take back a delegated power if the executive is misusing it. Yet, any attempt to reclaim it would not be under Congress’s complete control, as it would require the approval of the president. See Chenoweth & Samp, supra note 173, at 82. And the executive would rarely have incentive to give up such power. See id. at 94. Thus, while this justification may hold if Congress were delegating to those over whom it has complete control—such as its committee staff—it does not hold for delegation to the executive. See Wurman, supra note 102 at 1521.

[277]  See Story, supra note 78, at § 2.

[278]  Indeed, delegations that are specific, small, or supervised are often less egregious than they would otherwise be. But they are not the core reason why delegation is wrong—that has to do with the nature of the specific power and the specific qualities of the institutions at play.

[279]  See Hamburger, supra note 44, at 377.

[280]  See supra Section III.B.3.

[281]  See supra note 160.

[282]  U.S. Const. art. I, § 1.

[283]  Note that these four kinds of acceptable delegation may overlap with each other in specific delegations—for example, a single delegation may involve both factfinding and nonbinding internal government operations.

[284]  In addition to these four, there is also one a priori authority that the executive may exercise: the authority to interpret statutes pursuant to its executive responsibilities. This is hardly a “delegation,” but rather a logically required aspect of enforcing any law—the enforcer must determine what the law means. (Judicial review does not make it any less necessary for the executive to make its own determinations before enforcing a law.) And this power to interpret does not require a transfer of legislative power, either. First, it is a logically necessary aspect of enforcement, so it cannot be considered a uniquely legislative task, and accordingly it has historically been considered an executive or judicial responsibility. See Rappaport, supra note 200, at 204. Second, legislative activity is essentially a question of “will.” See The Federalist No. 78, supra note 152, at 402 (Alexander Hamilton). Yet statutory interpretation is a fact-based, rather than a value-based exercise. Its goal is to determine the content of a law as it is, not as it should be. Therefore, not even the minimal value-based judgments required in statutory interpretation are examples of executive policymaking, but they rather attempt to discover the facts of what the actual legislature intended. The question is primarily not what the executive’s will is in the matter, but what the legislature’s will was in the matter. See Rappaport, supra note 200, at 204, 205, 208.

Now, there is one situation in which the executive authority to interpret would allow delegation of real legislative power to the executive: when Congress passes an open-ended statute so vague that, while interpreting it, the executive essentially crafts the substance of the coercive rules. That is inappropriate, because such a law would violate the void-for-vagueness principle. See infra Section IV.D.4.

[285]  Gundy v. United States, 139 S. Ct. 2116, 2136 (2019) (Gorsuch, J., dissenting).

[286]  See Rappaport, supra note 200, at 204-5.

[287]  11 U.S. (7 Cranch) 382 (1813).

[288]  See Lawson, supra note 35, at 363-64.

[289]  See id. at 364.

[290]  See id. Indeed, it would be difficult to conceive of factfinding as a legislative role at all. Congress would have to repeatedly pass individual laws determining that states of affairs (specified in other laws) now exist. Furthermore, this factfinding role is implied in the very idea of enforcement. Every police officer walking his beat is assigned a core task: identify states of affairs around him that violate the law and then take appropriate action. Likewise, every executive agency is tasked to identify the states of affairs that their relevant statutes address and proceed according to their instructions.

[291]  See Hamburger, supra note 44, at 107.

[292]  See The Federalist No. 78, supra note 152, at 402.

[293]  See id.

[294]  This is not to say that the executive will not wrap its own will into such determinations (what political actors do in practice often does not align with the theories behind the institutions in which they serve). Rather, factfinding is not the kind of question that demands an answer based on the executive’s will.

[295]  Philip Hamburger indicates that this was the way that printed regulations for internal executive operations developed during the early years of the American republic. See Hamburger, supra note 44, at 87.

[296]  See id. Hamburger clarifies that “[o]f course, the officer could be prosecuted if he took a bribe or otherwise violated general prohibitions defined and imposed by law, but he was not accountable at law for mere disobedience.” See id. at 87-88. Hamburger indicates that this was also in keeping with early American practice.

[297]  Cf. 1 William Blackstone, Commentaries *49 (acknowledging the difference between external laws binding citizens and internal choices the government makes about its own operations as it goes about enforcing them).

By the sovereign power, as was before observed, is meant the making of laws, for wherever that power resides, all others must conform to and be directed by it, whatever appearance the outward form and administration of the government may put on. For it is at any time in the option of the legislature to alter that form and administration by a new edict or rule, and to put the execution of the laws into whatever hands it pleases; by constituting one, or a few, or many executive magistrates: and all the other powers of the state must obey the legislative power in the discharge of their several functions, or else the constitution is at an end.

[298]  See Hamburger, supra note 44, at 88 (indicating that this was also in keeping with early American practice).

[299]  See Gaziano & Blevins, supra note 9, at 52-53.

[300]  See Hamburger, supra note 44, at 394-95.

[301]  Id. at 395.

[302]  See id. at 85. This article does not attempt to conclusively weigh in on the debate over the nondelegation doctrine’s historical credentials, but by way of passing reference, the distinction between binding and nonbinding rules is a powerful argument against some pro-delegation arguments from historical precedent. See Wurman, supra note 102, at 1538 (observing that “[n]one of the early statutes that Mortenson and Bagley, Chabot, and Parrillo discuss disproves” the nondelegation rule under the conception that “any rule governing private conduct or altering private rights is ‘legislative.’”); Hamburger, supra note 44, at 83. See also Wurman, supra note 102, at 1554 (arguing that the First Congress’s practice was consistent with this standard).

[303]  This is why this article limits “legislative power” to “the power to issue rules coercively binding citizen action, and to exercise all other powers granted exclusively to Congress in the Constitution.”

[304]  See Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 42-43 (1825).

[305]  Id.

[306]  See Wurman, supra note 102, at 1534 (emphasis in original).

[307]  See Gundy v. United States, 139 S. Ct. 2116, 2137 (2019) (Gorsuch, J., dissenting); Gaziano & Blevins, supra note 9, at 52.

[308]  Telephone Interview with Molly Nixon, Separation of Powers Attorney, Pacific Legal Foundation (December 12, 2024).

[309]  Lawson, supra note 35, at 340.

[310]  Id.

[311]  See Thomas Aquinas, Summa Theologica I-II, Q.90, art. 4 (trans. Fathers of the English Dominican Province, Encyclopedia Brittanica 1923).

[A] law is imposed on others by way of a rule and measure. Now a rule or measure is imposed by being applied to those who are to be ruled and measured by it. Therefore, in order that a law obtain the binding force which is proper to a law, it must be applied to the men who have to be ruled by it. Such application is made by its being notified to them by promulgation. Therefore promulgation is necessary for the law to obtain its force.

See also Locke, supra note 120, at § 137.

For all the power the government has being only for the good of the society, as it ought not to be arbitrary and at pleasure, so it ought to be exercised by established and promulgated laws, that both the people may know their duty, and be safe and secure within the limits of the law, and the rulers too kept within their due bounds, and not to be tempted by the power they have in their hands to employ it to such purposes, and by such measures, as they would not have known, and own not willingly.

[312]  See 1 William Blackstone, Commentaries *45-46 (providing a vivid historical image for the menace that unpromulgated laws pose to society).

[313]  See Gaziano & Blevins, supra note 9, at 55.

[314]  Cf. Gundy v. United States, 139 S. Ct. 2116, 2142 (2019) (Gorsuch, J., dissenting) (explaining how the void-for-vagueness standard has been used as a substitute for the nondelegation doctrine to enforce the separation of powers). Void-for-vagueness is not just a substitute, however. As Gaziano and Blevins show, it is a standard that actually gives the nondelegation doctrine teeth. See Gaziano & Blevins, supra note 9, at 45-70.

[315]  See Gaziano & Blevins, supra note 9, at 45-70. Gaziano and Blevins provide a compelling argument for the void-for-vagueness test in the nondelegation context. Yet there is one small difference between their argument and the argument set forth here. Gaziano and Blevins describe void-for-vagueness through two main tenets: first, that laws must “be clear enough to provide fair notice,” and second, that laws must “be enacted by elected legislators to ensure democratic legitimacy.” See id. at 45. But they mostly emphasize the second aspect—laws must be passed by elected legislators so that they are not arbitrary. See id. at 45, 46, 56. They are reacting to an overemphasis on the fair notice proposal, so their emphasis makes sense. See id. at 64. However, this article equally highlights the “fair notice” tenet by meshing it with the “elected legislature” tenet into a single principle: the original law as enacted by the elected legislators must be clear enough to provide fair notice to the parties affected. Thus, it requires a twofold question about any law. First, can the enforcers and the public tell whether they are following or violating it? Second, can they tell this by looking only at the original congressional statute as passed by the legislators, without executive or judicial constructions narrowing it further? This maintains the “fair notice” tenet as a key question alongside the “elected legislature” tenet.

[316]  See Chenoweth & Samp, supra note 173, at 90.

[317]  See id.

[318]  Grayned v. City of Rockford, 408 U.S. 104, 108 (1972).

[319]  The same logic applies to the courts. Courts cannot save otherwise unconstitutional delegations of legislative power through “saving constructions” that impose limits on how the law can be applied. While courts are supposed to interpret laws constitutionally if possible, “[i]n the vagueness context . . . courts have sometimes read limits into a statute that don’t exist, slipping from interpretation into invention.” See Gaziano & Blevins, supra note 9, at 63. Thus, in this situation, the final form of the promulgated coercive rule binding citizen action would come from the judiciary, not the legislature. Attempts to get around the nondelegation doctrine by relying on “canons of delegation” are vulnerable to this criticism. See, e.g., Sunstein, supra note 224 at 315-17.

[320]  Not only does this prohibition flow from the essence of law as a coercive rule, but it is an important and practical due process consideration. There is little difference between executive power exercised under a broad delegation of power and simple arbitrary discretion. See Whittington & Iuliano, supra note 102, at 390.

[321]  Critics may lodge the same criticism against this standard that was lodged against the “intelligible principle” or “important questions” bright-lines: isn’t this vague as well? Doesn’t this leave the determination about the “reasonable person” up to individual judges? Isn’t this bright-line just as subjective as the others? These shortcomings do not undermine the void-for-vagueness standard in the same way that they undermine the other standards. Indeed, sometimes it may be difficult to determine whether a particular law is specific enough to communicate to the executive or the private citizen what they are to do. But this is the kind of ambiguity demanded by the ambiguity of the real world, not by congressional abdication of policymaking responsibility. The contrast between the amorphous “is it intelligible?” and the concrete “is it sufficient to serve as grounds for a conviction in court?” should make this clear. Unlike the intelligible principle standard, the void-for-vagueness standard is a test that delegations can actually fail. See Gaziano and Blevins, supra note 9, at 56 (“While the Supreme Court has not invalidated a single statute under the nondelegation doctrine itself since 1935, it has regularly applied the void-for-vagueness doctrine to strike down improper delegations”). Gaziano and Blevins catalogue a list of cases showing how, since at least 1921, the Court has used this standard to successfully strike down vague laws. See id. at 56-60; see also id. at 54-55.

[322]  Adler, supra note 117, at 161.

[323]  Mistretta v. United States, 488 U.S. 361, 372 (1989).

[324]  Lawson, supra note 35, at 332.

[325]  See supra Section III.C.

[326]  See, e.g., Lawson, supra note 3, at 1237-48 (critiquing legislative delegation, administrative adjudication, judicial deference, unaccountable executive employees, and violations of the separation of powers). Many of these only occur because Congress can delegate power to the administrative state. See Metzger, supra note 70, at 92, 94. And while the nondelegation doctrine specifically prohibits delegating legislative power to non-legislative branches of government, if the doctrine is true then the same rationale behind it immediately calls into question the delegation of judicial powers to non-judicial branches of government, etc.

[327]  Gundy v. United States, 139 S. Ct. 2116, 2130 (2019).

[328]  See, e.g., Lawson, supra note 3, at 147-151 (arguing that reliance interests also have a role to play in constitutional decision-making, so there is a case for treating past violations of nondelegation more leniently than future violations).

[329]  See, e.g., Lawson, supra note 3, at 143 (notifying his readers that “if one is looking for limiting principles to shield key elements of the modern regulatory state from a reinvigorated subdelegation doctrine, this [his particular proposal] is a possible avenue to explore.”).

[330]  Cf. Metzger, supra note 70, at 95 (“Few anti-administrativists are willing to seriously challenge delegation, and judicial anti-administrativism in particular has a notably rhetorical air, seemingly unwilling to follow through on the radical implications of its constitutional complaints.”); id. at 36.

[331]  The Federalist No. 78, supra note 152, at 403.

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 welcome responses to the views presented here. To join the debate, please email us at [email protected].