Skip to main content

I.  A Bracingly Simple Idea

It is a bracingly simple idea.

Article II, section 1 of the U.S. Constitution vests the executive power in “a president of the United States.” Those words do not seem ambiguous. Under the Constitution, the President, and no one else, has executive power. The executive is therefore “unitary.”1 It follows, as the night follows the day, that Congress lacks the power to carve up the executive—to say, for example, that the Secretary of Transportation is a free agent, immune from presidential control, or that the Secretary of Commerce can maintain their job unless the President is able to establish some kind of “cause” for removing them.2

On this view, the Supreme Court’s unambiguous embrace of the idea of the unitary executive in Myers v. United States3 was a golden moment in constitutional law, a ruling on which diverse people ought to be able to agree, and indeed one that they should enthusiastically embrace. And on this view, the Court’s messy, confusing, neologism-based, indefensible rejection of the unitary executive in Humphrey’s Executor v. Federal Trade Commission,4 upholding the independence of the Federal Trade Commission, was a dark stain, one of the lowest moments in the Court’s history and a prime candidate for inclusion in the “anticanon” of constitutional law. If that is so, the only serious question in the removal debate, for many decades, has been simple: Should Humphrey’s Executor be flatly overruled, or should it be confined as much as possible simply in deference to a longstanding precedent on which much of American government has been built?

A.  Minimalism and Maximalism

In Seila Law LLC v. Consumer Financial Protection Bureau,5 the Court invalidated the provision guaranteeing that the Director of the CFPB could not be removed by the President except for cause. But the Court did not overrule Humphrey’s Executor; it distinguished it, confining it to its facts. The resulting opinion, however, is deeply ambiguous, because it is not obvious what the legally relevant description of Humphrey’s Executor should now be taken to be.

On one reading, which we will call the “minimalist reading,” the Court’s opinion might be read to say: “We have made some mistakes in the past, and we may or may not overrule them, but if we can find any minimally plausible ground for distinguishing them, that is exactly what we will do.” On this view, the Court distinguished Humphrey’s Executor principally on the ground that the CFPB is headed by a single person, not a multimember commission, along the lines of then-Judge Kavanaugh’s earlier opinion for a D.C. Circuit panel.6 In simpler words: “Go forth, and sin no more.”

But in fact, there are significant strands of the Court’s opinion that seem far more ambitious; call these “the maximalist reading.” This reading arises because the Court repeatedly described the exception derived from Humphrey’s Executor as not extending to independent agencies that exercise significant executive power, as by rulemaking or enforcement in internal agency proceedings. In a crucial passage, the Court said that the baseline rule of Myers, granting the President at-will removal authority for all officers exercising executive power, is subject to “two exceptions—one for multimember expert agencies that do not wield substantial executive power, and one for inferior officers with limited duties and no policymaking or administrative authority.”7 The Court continued that these two exceptions “represent what up to now have been the outermost constitutional limits of permissible congressional restrictions on the President’s removal power”8 and refused to expand or add to them. Justice Thomas, who joined the majority opinion in relevant part, also wrote a separate opinion, joined by Justice Gorsuch, saying unambiguously that Humphrey’s Executor was wrong and that he would overrule it.9 In his words, “Humphrey’s Executor does not comport with the Constitution.”10

The maximalist reading, if pursued in future cases, would effect radical changes in administrative law and indeed the fabric of modern government. The main independent agencies with multiple heads wield broad rulemaking and enforcement powers; the Court’s ruling thus casts a legal cloud over the removal provisions for the commissioners and heads of the FTC, the FCC, the SEC, the NRC, the NLRB, and others. The constitutionality of those removal provisions would seem to depend on what, particularly, those agencies are authorized to do. Whether the maximalist reading is in fact pursued depends on many contingencies, but it is nonetheless significant that the Court read Humphrey’s Executor so narrowly that it might well be taken to have thrown the independence of most of the current independent agencies, and longstanding understandings of that decision, into grave doubt.

B.  Originalism and Dworkinism

There is also a major methodological ambiguity in Seila Law. Some of its defenders are likely to understand the ruling as a clear vindication of the Constitution itself, understood in terms of the original understanding of the text. Indeed, we predict that the decision will be taken as an originalist triumph and in two different ways: as an enthusiastic embrace of originalism as the proper method of constitutional interpretation and as an unquestionably proper use of originalism. Seila Law might even take its place with District of Columbia v. Heller11 as a defining example of originalism at work and as a vindication of that method.

We shall explore that possibility and raise some doubts about it. Taken in purely originalist terms, the decision might or might not be correct. For our purposes, the more fundamental point is that in our view, the Court’s opinion is not only, or not principally, an originalist one. The Court does not refer to the “original public meaning,” as many originalists do, and it does not work hard with the text and the history to show that in 1789, a widely shared understanding of the executive power would compel its conclusion. Indeed, there are major nonoriginalist strands to the opinion. To put things in slightly provocative terms, one might even call the majority’s opinion frankly Dworkinian, in the sense that it rests on an effort to put the existing fabric of law in the best constructive light by reference to considerations of political morality.12

In that respect, Seila Law can be illuminatingly understood as a form of constitutional common law13 and as responsive to emphatically contemporary concerns. The decision reflects anxiety about the powers of unaccountable bureaucrats freed from the constraining arm of the President (and hence We the People). Even while confining Humphrey’s Executor to its facts, it appeals throughout to high-level principles, such as “liberty” and “accountability,” to decide which of those facts are legally relevant. Is the number of agency heads, one versus many, relevant? The Court sometimes says it is not only relevant, but crucial; elsewhere, as we have described, the Court’s focus is on whether agencies wield executive power. The important thing is that on the terms of the Court’s own analysis, emphasizing those high-level principles—that is, which features of the history and caselaw are critical—is not simply read off from any previous precedent or from the original understanding. Rather, the Court arrives at its conclusions through high-level structural reasoning from what it sees as defining constitutional principles, and it expressly defends that structural reasoning as such.14

In other words, the conclusion and the analysis in Seila Law are rooted in large part in rich interpretations of abstract, contested principles of self-government and liberty. That is why the decision can be understood not only in originalist terms but also as a species of constitutional common law, or, more precisely, as a reflection of Dworkin’s notion of law as “integrity,” by which judges attempt both to “fit” existing legal materials and to “justify” them by making them the best that they can be.15 So understood, Seila Law is an exercise in fit, and an unusually creative one; it is also an exercise in justification, and an especially bold one.

In what follows, we explore these two ambiguities about the opinion, which involve respectively the scope of its holding and its methodology. Our exposition comes in four parts. Part II explores the idea of a unitary executive with reference to text and history and also with reference to changed circumstances. We attempt to show the ambiguities in the founding era that gave rise to reasonable, competing understandings of what was and was not settled. We also attempt to show that the emergence of the modern administrative state can be taken both to fortify and to undermine the argument for the idea of a strongly unitary executive. Part III discusses the Court’s disparate, inconclusive encounters with that idea. Part IV turns to Seila Law and its striking treatment of Humphrey’s Executor, which it simultaneously preserves (for now) and perhaps hollows out, thus endangering many contemporary independent agencies. Part V discusses implications and constitutional method, showing that the decision rests crucially on contested normatively laden views about the meaning of high-level principles. In the end, Seila Law is best seen as part of a much broader effort, in prominent circles, to constrain the operation of the regulatory state in general and of apparently unaccountable institutions in particular by referencing a distinctive understanding of constitutional principles.

II.  Two Kinds of Unitariness

In a sense, everyone agrees that the Constitution creates a “unitary executive.”16 There is one President, not an executive council, and the President is broadly in charge of the executive branch. But reasonable people strenuously disagree about what a unitary President entails.17 We begin by distinguishing two ways of thinking about the unitary executive and then we turn, respectively, to originalist and nonoriginalist disputes about which way is best. The result is a map with four possible positions.

A.  Strong and Weak

Some people believe in a strongly unitary presidency; others believe in a weakly unitary presidency. The former insist that at a minimum, the President has the constitutional authority to remove all noninferior policymaking officials who exercise executive power (and also to control their decisions).18 On this view, the executive power is the President’s alone, and any congressional effort to compromise that principle by limiting the President’s ability to fire executive branch officials is forbidden. All those who implement the law, including all those who exercise administrative authority, must be controlled by the President, at least in the sense of being at-will employees.19 The Court’s opinion in Seila Law seems to embrace this view, certainly as a matter of constitutional text and history.

By contrast, those who believe in a weakly unitary presidency insist that under the Necessary and Proper Clause, Congress has significant authority to limit the President’s authority of removal (and also supervision).20 They are likely to agree that with respect to some executive officers—the Secretary of Defense, the Secretary of State, the Attorney General—the President must have plenary removal authority. The President must have that authority where specific constitutional texts that make grants of power to the President, such as the Commander-in-Chief Clause, are implicated, and where tradition holds that core executive powers and prerogatives involving war, diplomacy, and foreign affairs are at issue. But they also believe that as a general matter, Congress has considerable room to structure the administrative state as it sees fit, especially where tradition suggests that agency independence is essential, as with respect to agencies that engage in financial regulation.21 Justice Kagan’s dissenting opinion in Seila Law embraces this position.22

Those who believe in a weakly unitary executive insist that Congress is able to immunize adjudicative officers from presidential control23; they add that some administrative functions might be exercised by people who are not subject to the President’s policy preferences.24 They believe that multiple authorities, including rulemaking and even prosecution, might be taken out of plenary presidential control so long as doing so does not prevent the President from exercising constitutionally specified functions, defined not broadly to mean control of all executive or administrative powers but more narrowly to include specific ones, such as the Commander-in-Chief power.25 Of course it is true that those who believe in a weakly unitary executive have to do considerable work to spell out what their position particularly entails. The general point is that under the Necessary and Proper Clause, Congress is permitted to carve out some important functions from presidential control.

Some participants in these debates speak in originalist terms; others do not.26 We can therefore identify four positions:

Strongly unitary(1)(2)
Weakly unitary(3)(4)

B.  Text and History

1.  Strongly unitary: the original meaning

Seeking to uncover the original public meaning of the founding document, those in Cell (1) begin with the text.27 Article II vests the executive power in “a President of the United States.”28 It also grants the President the power to “take Care that the Laws be faithfully executed.”29 On one view, these terms are exceedingly clear.30 They demonstrate that the President, and no one else, is in charge of execution of the laws. Invoking history, those in Cell (1) add that the contemporaneous debates show that the strongly unitary view reflects the original public meaning.31 In their view, those debates demonstrate that the framers and ratifiers sought to ensure that the executive branch would be accountable, coordinated, and energetic.32 The concentration of the relevant authorities in a single person was deemed necessary to achieve those goals. On this view, the original public meaning of the constitutional text compels Cell (1).

Cell (1) is often thought to have compelling structural justifications, as emphasized in the founding period. As Hamilton put it in The Federalist No. 70: “Decision, activity, secrecy, and dispatch will generally characterize the proceedings of one man in a much more eminent degree than the proceedings of any greater number.”33 In addition, “one of the weightiest objections to a plurality in the executive … is that it tends to conceal faults, and destroy responsibility.”34 A unitary executive is more clearly subject to the people and therefore well-suited to a self-governing nation. If Congress were authorized to divide the executive power—for example, by creating independent officials charged with implementing important aspects of federal law—all of the Constitution’s structural commitments, as specified by Hamilton, would be gravely undermined.

If there were any doubts (the Cell (1) view asserts), the Decision of 1789, as it is called, resolves them.35 In that year, an early Congress debated the President’s removal power and the unitariness of the executive at great length in the context of determining the legal relationship between early cabinet heads (Treasury, War, and Foreign Affairs) and the President.36 Various views were represented in that debate, but Congress ultimately concluded that, by Constitutional compulsion, those who execute the laws must be at-will employees of the President, at least if they work at sufficiently high levels.37 That conclusion, clearly emerging from the historical materials, is, on this view, authoritative with respect to the original public meaning of the Constitution.38

2.  Weakly unitary: the original public meaning

Those in Cell (3) respond either that the text is murky and inconclusive or that it rejects the strongly unitary view.39 In their view, the original understanding does not call for Cell (1). With respect to the text itself, Justice Holmes put it briskly but memorably: “The arguments drawn from the executive power of the President, and from his duty to appoint officers of the United States (when Congress does not vest the appointment elsewhere), to take care that the laws be faithfully executed, and to commission all officers of the United States, seem to me spider’s webs inadequate to control the dominant facts.”40 Justice Holmes did not spell that out, but on one view, the vesting of executive power in a President says essentially nothing about Congress’ capacity to insulate certain officials from presidential control. (It is a spider’s web.) The general idea that there is one President, and the general idea that the President has the executive power, need not be taken to resolve the specific question whether Congress can declare that some officials, executing the laws, are not his at-will employees.

If this proposition seems puzzling or provocative, those in Cell (3) add that tellingly, Hamilton himself, a strong believer in a unitary executive, specifically rejected Cell (1) in The Federalist and concluded that the removal power followed from the Appointments Clause. In his view, that meant that officials who were subject to advice and consent for their appointment could be made removable only with the consent of the Senate.41 In his words, “the consent of [the Senate] would be necessary to displace as well as to appoint.”42 Remarkably, he added this, in a passage that is much less well-known than it ought to be:

A change of the Chief Magistrate, therefore, would not occasion so violent or so general a revolution in the officers of the government as might be expected, if he were the sole disposer of offices. Where a man in any station had given satisfactory evidence of his fitness for it, a new President would be restrained from attempting a change in favor of a person more agreeable to him, by the apprehension that a discountenance of the Senate might frustrate the attempt, and bring some degree of discredit upon himself. Those who can best estimate the value of a steady administration, will be most disposed to prize a provision which connects the official existence of public men with the approbation or disapprobation of that body which, from the greater permanency of its own composition, will in all probability be less subject to inconstancy than any other member of the government.43

Thus far, the claim is not that Hamilton necessarily had the original public meaning right. It is only that the textual vesting of the executive power in the President, by itself, need not be taken to resolve the question at hand. Hamilton’s views would seem to be decisive on that particular question—and a strong point in favor of Holmes’s claim that the relevant words of the constitutional text are “spider’s webs inadequate to control the dominant facts.”44

The same conclusion holds for the Take Care Clause.45 One can wholeheartedly agree that the President is authorized and obliged to execute the laws faithfully while also insisting that Congress has the capacity to immunize some officials from the President’s plenary control. The duty of faithful execution need not entail the conclusion that the President can discharge law-implementing officials in the President’s discretion. On one view, the laws that limit their power to do that must themselves be faithfully executed.46 In any case, the Clause can be taken as a statement of a duty, phrased in the passive voice, not speaking to the particular question whether Congress can limit the President’s removal power. Hamilton’s views are relevant here as well. Hamilton had no problem with the Take Care Clause, but he also believed that Congress could condition removal of cabinet officials on the advice and consent of the Senate.

Consider in this regard the Opinions in Writing Clause: “The President … may require the Opinion, in writing, of the principal Officer in each of the executive Departments, upon any Subject relating to the Duties of their respective Offices.”47 On one view, this clause is hard textual evidence against the strongly unitary view, taken as an originalist matter. If the President is fully in control of the operations of all those who administer federal law, why would the framers and ratifiers deem it necessary to specify this particular power? The question might be taken to be rhetorical. Though the Opinions in Writing Clause raises many questions and can be understood in different ways, it takes some work to explain how it sits comfortably with the strongly unitary view of the presidency.48

As for post-ratification history, the Decision of 1789 contained a range of competing strands, with influential figures explicitly rejecting the strongly unitary view.49 The debates were complicated and messy—far more so than they might initially appear.50 Those outcomes, and the Decision itself, could be read in different ways. It would be possible to read them as rejecting the proposition that the strongly unitary view is constitutionally mandatory and instead as revealing a discretionary congressional choice, not compelled by the Constitution, to confer unrestricted removal authority over particular officials. After a careful study of the debates, for example, Jed Shugarman refers to “the legend of the Decision of 1789,” and concludes, “A majority of the first Congress opposed the powers cited by unitary theorists … On whether the president had exclusive removal power, the first Congress decisively answered no.”51

On another view, also based on a careful analysis of the debates, there was no clear Decision of 1789 that authoritatively settled the question whether Article II establishes a strongly or weakly unitary executive. On that view, “the outcome was not nearly as decisive as later actors would claim. The process by which the removal issue was settled produced only a murky, very tenuous precedent in favor of the president’s constitutional removal power.”52 Note too that any settlement, if it even counts as one, involved just three departments (State, War, and Treasury), whose particular functions might be thought to be essentially indistinguishable from the President’s own.53 If so, the Decision of 1789 might have no bearing on congressional efforts to immunize from presidential control the decisions of agencies with other sorts of authorities; consider the Federal Reserve Board, the Securities and Exchange Commission, the National Labor Relations Board, the Social Security Administration, and the Nuclear Regulatory Commission. Even if there was a Decision of 1789, it might not speak to the question whether Congress can immunize those agencies, and assorted others, from plenary presidential removal authority.

In addition, prominent legal historians have read both theory and practice during the founding era as an endorsement of the weakly unitary view, at least in some form.54 Some historians have agreed that first, some department heads (such as the Secretaries of State and Defense) must be subject to presidential control, so that the President can exercise their own constitutional authority but that others need not be, and second, that the Necessary and Proper Clause gives Congress considerable discretion to decide on the appropriate allocation of authority between the President and the administrative state.55 Those in Cell (3) can easily insist that their opponents in Cell (1) turn out to be living constitutionalists, motivated by emphatically present-day concerns and fears (about, say, accountability and liberty) while proclaiming a clear constitutional settlement that cannot, in fact, be traced to the original understanding and the founding generation.

C.  Nonoriginalist Approaches

For various reasons, many people are not originalists.56 They do not believe that the original public understanding is binding. They too commit to following the text, of course, but they insist that structural principles and inferences, institutional roles, changed circumstances, unanticipated problems, judicial precedents, longstanding practices, the views of Congress and the President, and new or emerging values are a legitimate part of constitutional interpretation, part of how the text’s meaning is best understood.57

In the context at hand, the rise of the modern administrative state, arguably an unanticipated development, might be taken to motivate nonoriginalist approaches to separation of powers questions.58 Nonoriginalist approaches can of course take diverse forms. In the context at hand, one form involves constitutional common law.59 Another form involves “translation”: efforts to understand the Constitution’s commitments in circumstances that the founding generation could not have anticipated.60 Yet another form, elaborated by Dworkin, suggests that constitutional interpretation requires judges to put the existing legal materials in the best constructive light.61

Cell (2) can be seen as an exercise in constitutional common law in translation62 or in Dworkinian law-as-integrity. The basic idea is that in a period in which the executive branch is wielding unanticipatedly expansive power that touches so many domains of domestic affairs, the founding commitments to accountability, dispatch, coordination, and energy call for strong unitariness, even if those commitments authorized weak unitariness two centuries earlier.63 It is one thing to say that in (say) 1800, Congress had the constitutional authority to immunize certain agencies and institutions, not so fundamental to national life, from plenary presidential control. It is quite another to say that Congress can carve out an assortment of crucial agencies affecting the economy in multiple ways, such as the Federal Communications Commission, the Federal Trade Commission, the Nuclear Regulatory Commission, and the CFPB, and let them do their work without control from the constitutionally specified executor of the laws. In the modern era, fidelity to constitutional commitments calls for insistence on presidential primacy, even if it did not quite do that in the founding era, when the administrative state was so much smaller and less central to people’s lives.

It is true that this approach to constitutional interpretation might seem looser and more speculative than originalism, supposing, quite controversially, that the latter yields straightforward answers. But because many areas of constitutional law have unmistakable common-law features, with new judgments emerging from the case-by-case process,64 we might want to insist that careful attention to the most central founding commitments, as opposed to particular founding-era understandings and practices, is an honorable way to engage in constitutional interpretation. And if we attend to those commitments, strong presidential oversight of the administrative state, including agencies now characterized as independent, might seem essential if we are to avoid devastating damage to the very principles invoked by Hamilton in defense of a unitary presidency.

The ironic conclusion, embraced by those in Cell (2), is that even if the founding generation did not believe that a strongly unitary presidency was necessary to promote their own deepest commitments, such a presidency is necessary now, given the sheer size and nature of the contemporary administrative state. Those in Cell (2) think that those in Cell (1) do not have the original understanding right. But for reasons of principle, they are happy to make common cause with them.

A firm rejoinder comes from Cell (4).65 Perhaps Cell (2) gets it exactly wrong. If fidelity is the goal, perhaps we should emphasize the risks of concentrated power and the importance of checks and balances—and the need to allow Congress to have some flexibility given the diversity of circumstances that give rise to new agencies. Perhaps the real concern, highlighted by changed circumstances, is the capacity of just one person, acting on their own, to move the government in their preferred direction. It should not be necessary to mention that that capacity is anathema to founding commitments; it raises the specter of monarchy.66 As James Landis put it in a famous translation-based defense of expert administrative tribunals:

The administrative process is, in essence, our generation’s answer to the inadequacy of the judicial and the legislative processes. It represents our effort to find an answer to those inadequacies by some other method than merely increasing executive power. If the doctrine of the separation of power implies division, it also implies balance, and balance calls for equality. The creation of administrative power may be the means for the preservation of that balance, so that paradoxically enough, though it may seem in theoretic violation of the doctrine of the separation of power, it may in matter of fact be the means for the preservation of the content of that doctrine.67

Or, as Abner Greene puts it in a sustained modern defense of Cell (4), “if we accept sweeping delegations of lawmaking power to the President, then to capture accurately the framers’ principles—principles that deserve our continuing adherence—we must also accept some (though not all) congressional efforts at regulating presidential lawmaking.”68

To compress what might be turned into a lengthy argument: If executive power is concentrated in the President, and if they are able to oversee everything now included in the immensely broad category of administration, we would see a kind of Madisonian nightmare.69 Given the risk of a discretion-wielding, immensely powerful set of administrative authorities concentrated in a single person, Congress should have the authority to insulate at least some such authorities from presidential control, simply in order to preserve diffusion of power.

The strongest example may well be the Federal Reserve Board: If Presidents could control it, they could promote their own short-term political interests by reducing interest rates to the detriment of the economy’s long-term health. In the year before an election, for example, the President might manipulate the Board’s decisions in such a way as to promote the President’s own prospects for reelection. A similar argument might apply to the Federal Communications Commission, whose independence might be deemed necessary to prevent a situation in which the President is punishing their political enemies and rewarding their political friends. After all, the FCC has the authority to give out and to renew licenses to radio and television stations. If the President is entitled to control its decisions, the President could manipulate the speech market in such a way as to promote their own political interests. These arguments could easily be generalized—for example, to the FTC, which has the authority to approve or disapprove mergers, and to a range of financial authorities whose decisions might bear on the short-term political prospects of the President.

To reach a conclusion on the underlying questions, it would be necessary to begin by deciding on the right approach to constitutional interpretation—that is, to choose between Cells (1) and (3) or Cells (2) and (4). Those who embrace originalism would have to undertake a careful investigation of history in order to choose between Cell (1) and Cell (3). If one rejects originalism, an assessment of how best to fit with founding commitments would determine the choice between Cell (2) and Cell (4). Our goal here is not to resolve those questions. At the same time, we agree that on originalist grounds, some executive officers—including the Secretary of State, the Secretary of Defense, the Secretary of the Treasury, and the Attorney General—must be at-will employees of the President. But we are not originalists, and we would give substantial weight both to judicial precedents and institutional practices over time, which means that we land in Cell (4), which can easily fit, broadly speaking, with judicial decisions since 1926.70

To be sure, general cells do not decide concrete cases,71 including Seila Law. Thus far, our more modest goal has been to offer a concrete sense of why many people believe that a strongly unitary executive is constitutionally mandatory and why many people reject that proposition.

III.  The Unitary Executive in the Supreme Court

Before Seila Law, the Court’s decisions, often adopting the weakly unitary position,72 are best understood as falling within Cell (4), though they could be defended by reference to Cell (3), and though Cells (1) and (2) make prominent appearances (and very much have come to the fore as a result of Seila Law). Here again, our goal is to give a brisk sense of the constitutional background rather than to offer a comprehensive treatment.

A.  The Early Era

Myers v. United States,73 the Court’s first sustained encounter with the issue, is essentially Cell (1), urging that the text and the original understanding call for broad presidential control over the executive branch, above all through plenary removal authority. Myers offers an elaborate discussion of text, structure, and history, including the Decision of 1789, and thus provides the foundation for much of contemporary analysis.74 For modern defenders of a strongly unitary executive, Myers is the shining and fixed star75 even though it was written by a former President, Chief Justice Taft, who might be expected to be particularly insistent on presidential prerogatives. It could even be said that for those who believe in a strongly unitary presidency, all of subsequent constitutional theory has been a series of footnotes to Myers.76

By contrast, the widely reviled opinion in Humphrey’s Executor,77 generally understood to validate the independent agency form, deals not at all with constitutional history and barely at all with constitutional text. Mounting a fundamental attack on the very idea of agency independence and asserting a strongly unitary view, President Franklin Delano Roosevelt asserted that the “inefficiency, neglect of duty, and malfeasance in office” standard, which restricted his power to discharge members of the Federal Trade Commission, was an unconstitutional restriction on his Article II authority to discharge them on whatever grounds he chose.78 Insofar as the Court unanimously rejected that argument, it made no effort to ground its analysis in something akin to Justice Holmes’s skepticism about the supposed clarity of the text, in constitutional structure, in changed circumstances, or in the ambiguities of the Decision of 1789. Instead, the Court emphasized the need to specify the relevant administrative functions. It held that Congress has the power to immunize members of independent agencies from plenary presidential removal authority insofar as such agencies exercise “quasi-judicial” functions (namely, adjudication) and “quasi-legislative” functions (namely, compiling reports for Congress).79

Thus understood, the actual decision was far more modest than it has long been taken to be;80 as we shall see, Seila Law corrects the historical record. Humphrey’s Executor was hardly a broad endorsement of the independent agency form as it is now understood. Because the Federal Trade Commission made policy only through adjudication (and lacked the authority to issue binding orders at that)81 and did not even try to exercise rulemaking authority until the 1970s,82 Humphreys Executor was very limited indeed. It does not hold that Congress may restrict the President’s ability to control the exercise of the authority to make rules. Its narrow holding, focused on the distinctive functions of the FTC in 1935, can in fact be defended on originalist grounds.83 Insofar as it allows Congress to immunize purely adjudicative officials from plenary presidential control, it did not even depart radically from Myers. In that case, the Court had stated:

Then there may be duties of a quasi-judicial character imposed on executive officers and members of executive tribunals whose decisions after hearing affect interests of individuals, the discharge of which the President cannot in a particular case properly influence or control. But even in such a case, he may consider the decision after its rendition as a reason for removing the officer, on the ground that the discretion regularly entrusted to that officer by statute has not been, on the whole, intelligently or wisely exercised. Otherwise, he does not discharge his own constitutional duty of seeing that the laws be faithfully executed.84

In this passage, the Myers Court recognized that adjudicatory authority is distinctive, in the sense that its exercise might be immunized from presidential influence or control. To be sure, Humphreys Executor goes further than Myers insofar as it rejects the idea that as a matter of constitutional right, the President may discharge adjudicatory officials simply on the ground that the President does not like their decisions. But that is not exactly a radical shift from Myers.

Insofar as the Court ruled that Congress can immunize from presidential control those who compile reports to inform legislation, the Article II objection is hardly at its strongest. After all, Congress can have its own staff; perhaps it can also say that certain officials, compiling those reports within an executive agency, are not at-will employees of the President.85 Recall Humphreys Executor has absolutely nothing to say about whether Congress has the authority to make rulemaking agencies independent of the President. As of 1935, then, and for the following decades as well, we might say that the Court held that the Constitution allows Congress to immunize adjudicative officers from at-will discharge by the President but did essentially nothing more than that to authorize Congress to exempt policymaking officials from plenary presidential control. This is so even if actual practice by independent agencies, including the Securities and Exchange Commission, the Federal Reserve Board, and the Federal Communications Commission, went far beyond what was necessarily authorized by Humphrey’s Executor.

B.  Modernity

Until Seila Law, the modern era had been largely defined by a continued endorsement of Cell (4), fortified by longstanding practice and Humphreys Executor as currently understood, butting up hard against Cell (1) and perhaps Cell (3), and on occasion Cell (2) as well. In Free Enterprise Fund v. Public Company Oversight Board,86 for example, the Court clearly accepted Humphreys Executor even as it also signaled a lack of enthusiasm for its decisions authorizing Congress to create independent agencies: “The parties do not ask us to reexamine any of these precedents, and we do not do so.”87 Its holding, invalidating two tiers of insulation from the President (an independent agency within an independent agency), was relatively narrow. But in reaching that holding, the Court also offered a degree of support for strong unitariness, not on originalist grounds but largely by reference to structural principles associated with the Take Care Clause:

The President cannot “take Care that the Laws be faithfully executed” if he cannot oversee the faithfulness of the officers who execute them. Here the President cannot remove an officer who enjoys more than one level of good-cause protection, even if the President determines that the officer is neglecting his duties or discharging them improperly. That judgment is instead committed to another officer, who may or may not agree with the President’s determination, and whom the President cannot remove simply because that officer disagrees with him. This contravenes the President’s “constitutional obligation to ensure the faithful execution of the laws.”88

Invoking text and history but also speaking broadly in terms of basic principle, the Court stated plainly: “It is his responsibility to take care that the laws be faithfully executed. The buck stops with the President, in Harry Truman’s famous phrase.”89 In words that could easily be invoked to challenge the very idea of independence, the Court added: “By granting the Board executive power without the Executive’s oversight, this Act subverts the President’s ability to ensure that the laws are faithfully executed—as well as the public’s ability to pass judgment on his efforts. The Act’s restrictions are incompatible with the Constitution’s separation of powers.”90 By emphasizing the need for the President to have the tools with which to take care that the laws be faithfully executed, the Court planted a seed that could grow over time. The central idea is that unitariness within the executive operates mainly through the particular obligation reflected in the Take Care Clause, which ensures that the President is ultimately responsible for faithful execution and which means that the President must have the tools to carry out that responsibility. That idea is very different from, or at least more specific than, any organizing principle that the Court had offered before.

Decided decades earlier, Morrison v. Olson91 is by contrast a clear Cell (4) case, allowing Congress to create an independent prosecutor who is empowered to investigate and prosecute wrongdoing on the part of high-level executive branch officials. In Morrison, the Court did not much engage constitutional text and history. Instead it spoke in pragmatic and functional terms, emphasizing that so long as the President could carry out their constitutionally assigned tasks, Congress could establish an office with a degree of independence from the President.92 The Court went out of its way to emphasize that the statutory restriction on the President’s removal power—”good cause”—left the President with a great deal of authority:

This is not a case in which the power to remove an executive official has been completely stripped from the President, thus providing no means for the President to ensure the “faithful execution” of the laws. Rather, because the independent counsel may be terminated for “good cause,” the Executive, through the Attorney General, retains ample authority to assure that the counsel is competently performing his or her statutory responsibilities in a manner that comports with the provisions of the Act. Although we need not decide in this case exactly what is encompassed within the term “good cause” under the Act, the legislative history of the removal provision also makes clear that the Attorney General may remove an independent counsel for “misconduct.”93

Here, then, is an emphasis on the centrality of the Take Care Clause, and an insistence that the “good cause” standard was compatible with it.

Even before Seila Law, it would have been fair to say that the view associated with Cell (1) enjoyed strong support on the current Court.94 It would have been fair to speculate that a majority believed that the independent agency form is constitutionally illegitimate and that Humphreys Executor was wrongly decided, at least as it is currently understood.95 For such justices, the only question is whether it should be overruled—which brings us directly to the case at hand.

IV.  Strongly Unitary

In some ways, the problem in Seila Law was straightforward. The Director of the CFPB was protected from presidential removal by the same statutory formula that was at issue in Humphrey’s Executor: inefficiency, neglect of duty, or malfeasance in office.96 The most obvious difference between the CFPB and the vast majority of independent agencies is that the former is headed by a single person whereas the latter are headed by multimember commissions.

A.  Options

In these circumstances, the Court had four principal options:

1.  It could have upheld the provision apparently guaranteeing independence to the CFPB on the grounds that the number of heads was constitutionally irrelevant and that Humphrey’s Executor essentially settled the constitutional issue.

2.  It could have interpreted the provision apparently guaranteeing independence to the CFPB so as to allow the President to have considerable authority over the bureau’s policymaking and concluded that, so interpreted, the provision is constitutionally unobjectionable. An interpretation that would allow the president that authority would have meant that the CFPB was not really independent at all—and hence could have weakened and perhaps met the constitutional objection.

3.  It could have overruled Humphrey’s Executor and thus struck down the independent agency form.

4.  It could have struck down the provision apparently guaranteeing independence to the CFPB on the ground that independence is permissible only for agencies headed by a multimember commission and not for agencies headed by a single person.

Each of these options could be accompanied, of course, by opinions with very different emphases. Option (2) is evidently attractive, and it could be written in numerous ways that allowed the President more or less in the way of policy control.97 The Court chose Option (4), but not with a minimalist opinion that strongly affirmed Humphrey’s Executor, and concluded, modestly and narrowly, that independence was unacceptable when the agency was headed by a single person. Instead, the Court offered the most forceful embrace of the strongly unitary view since Myers itself. In the process, the Court’s opinion seems to have gutted Humphrey’s Executor and in a sense to have confined it to its facts in a way that opens up a great deal of room for contemporary challenges to independent agencies.

B.  Not Minimalist

Toward the very beginning, the Court’s opinion, by Chief Justice Roberts, signaled the sheer magnitude of the issue in the case: “Under our Constitution, the ‘executive Power’—all of it—is ‘vested in a President,’ who must ‘take Care that the Laws be faithfully executed.”98 Just a few paragraphs thereafter, the Court declared categorically, “The President’s power to remove—and thus supervise—those who would wield executive power on his behalf follows from the text of Article II, was settled by the First Congress, and was confirmed in the landmark decision Myers v. United States.”99 As the Court explained, “[L]esser officers must remain accountable to the President, whose authority they wield.”100 Emphasizing what it saw as the Decision of 1789, the Court explicitly and fully endorsed the analysis in Myers; it put a kind of halo around both the holding and the analysis. By contrast, it did not have a favorable word to say about Humphrey’s Executor.

With Myers as the defining case, the Court said that it had recognized only two exceptions to the strongly unitary executive. The first is “expert agencies led by a group of principal officers removable by the President only for good cause”; the second is “certain inferior officers with narrowly defined duties.”101 The Court said that it would not “extend these precedents to … an independent agency that wields significant executive power and is run by a single individual.”102 Pointedly, it said that it “need not and do[es] not revisit our prior decisions”103—which is not the most enthusiastic endorsement.

At this point, it would have been possible for the Court to avoid the constitutional question and to embrace Option (2). The Court might have said that the statutory standard allows the President the kind of policymaking control that Article II requires. That conclusion would have had large implications for the President’s relationship to a host of independent agencies and the Court refused to offer it. On this count, at least, the Court’s analysis is a ringing endorsement of agency independence—as a statutory matter. More specifically:

Humphrey’s Executor implicitly rejected an interpretation that would leave the President free to remove an officer based on disagreements about agency policy. In addition, while both amicus and the House of Representatives invite us to adopt whatever construction would cure the constitutional problem, they have not advanced any workable standard derived from the statutory language. Amicus suggests that the proper standard might permit removals based on general policy disagreements, but not specific ones; the House suggests that the permissible bases for removal might vary depending on the context and the Presidential power involved. They do not attempt to root either of those standards in the statutory text. Further, although nearly identical language governs the removal of some two-dozen multimember independent agencies, amicus suggests that the standard should vary from agency to agency, morphing as necessary to avoid constitutional doubt. We decline to embrace such an uncertain and elastic approach to the text.104

The Court added, “Without a proffered interpretation that is rooted in the statutory text and structure, and would avoid the constitutional violation we have identified, we take Congress at its word that it meant to impose a meaningful restriction on the President’s removal authority.”105 These words may or may not mean that the President entirely lacks policymaking control over independent agencies. But they certainly mean that the statutory provision that grants independence is a “meaningful restriction”— which means, in the Court’s view, that the constitutional issue could not be avoided.

In an important sense, however, the Court did revisit Humphrey’s Executor, if only by interpretation. The Court took that decision in accordance with its historical context rather than with modern-day understandings of the authorities of the FTC (including rulemaking). In 1935, it would have been plausible to say that the agency did very little that would qualify as narrowly “executive,” and insofar as it engaged in quasi-adjudicative and quasi-legislative actions, its powers were sharply limited (and quasi!). As the Seila Law Court had it, all that was central to the Court’s ruling.106 On this count, Seila Law is exceedingly important; even in the modern era, the Court had never interrogated or narrowed Humphrey’s Executor in this way.

The Court emphasized that the FTC had several features: it consisted of five members; it had bipartisan membership; and its members had staggered, seven-year terms.107 And the Court did not stop there. It added that the agency’s functions were quasi-judicial and quasi-legislative and in distinctive ways.108 They were quasi-judicial insofar as the FTC acted as an adjudicator.109 They were quasi-legislative insofar as the FTC undertook investigations and compiled reports for Congress.110 Hence the Humphrey’s Executor exception was strictly limited to “multimember expert agencies that do not wield substantial executive power.”111 If taken at face value, this reading of Humphrey’s Executor, while historically accurate, is astonishing, because the major multimember independent agencies assuredly do wield “substantial executive power,” however defined, and because the passage seems to raise the possibility that their independence is unconstitutional to that extent. Pointedly, the Court declined to embrace the view that Humphrey’s Executor allows Congress to make agencies independent because and when they engage in rulemaking.112

The second exception, stemming above all from Morrison, merely said that Congress can protect an inferior officer “lacking policymaking or significant administrative authority” with a for-cause provision.113 So understood, that exception seems quite narrow. (It is also puzzling; the power to initiate criminal prosecutions might well be taken to qualify as “significant administrative authority.” But we bracket that point.)

By contrast, the Director of the CFPB “is hardly a mere legislative or judicial aid.”114 On the contrary, the Director “possesses the authority to promulgate binding rules fleshing out 19 federal statutes, including a broad prohibition on unfair and deceptive practices in a major segment of the U.S. economy.”115 Nor does the Director merely submit “recommended dispositions to an Article III court.”116 It is worth emphasizing this point; unlike the FTC, “the Director may unilaterally issue final decisions awarding legal and equitable relief in administrative adjudications.”117 In addition, the Director’s “enforcement authority includes the power to seek daunting monetary penalties against private parties on behalf of the United States in federal court—a quintessentially executive power not considered in Humphrey’s Executor.”118

In refusing to extend its prior rulings to this new situation, the Court did not speak in originalist terms. Instead, it worked by reference to historical practice and to high-level principles. Thus, the Court emphasized that the CFPB’s structure “is almost wholly unprecedented.”119 Single-headed independent agencies have been very rare. The most prominent example is the Social Security Administration, whose independence “is comparatively recent and controversial.” Further, the Social Security Administration “lacks the authority to bring enforcement actions against private parties” and is largely an adjudicative agency.120

Even more fundamentally, the Court held that the CFPB’s structure was incompatible with the constitutional design, understood in the large, which “scrupulously avoids concentrating power in the hands of any single individual,” except in the case of the President, “the most democratic and politically accountable official in Government … elected by the entire nation.”121 By contrast, the “single-Director structure contravenes this carefully calibrated system by vesting significant governmental power in the hands of a single individual accountable to no one.”122 The result is to threaten, at once, the core structural principles of self-government and liberty.

Importantly, and in a bow to Cell (4), the Court showed that it was not oblivious to the rise of “‘a vast and varied federal bureaucracy.’”123 In light of that size and variety, the Court said that it had a sharpened “duty to ensure that the Executive Branch is overseen by a President accountable to the people.”124 This is not quite a point about translation, but it is not so far from it, and it can be readily seen as Dworkinian or a form of constitutional common law. The basic idea is that the structural argument for presidential control is strengthened, rather than weakened, by the existence of the modern administrative state.

V.  New Doubts, In Brief

It is clear that the CFPB is permitted to continue in operation, now as an executive agency.125 A noteworthy implication is that it can now be made subject to the review process that is overseen by the Office of Information and Regulatory Affairs, which includes a complex process of interagency review and various analytical requirements, including cost-benefit analysis.126 For better or for worse, it is permissible to require CFPB regulations to be submitted to OIRA for application of the standard process. It is therefore subject to the control of the Executive Office of the President and ultimately to the judgments of the President personally.

But some fundamental questions are now suddenly open. We have seen that the Court adopted an extremely narrow reading of Humphrey’s Executor, emphasizing that the FTC (1) did not issue binding rules; (2) merely submitted recommended dispositions to federal courts; and (3) did not wield substantial executive power, which would be wielded if, for example, the FTC were authorized to seek monetary penalties against private parties.127 Many contemporary independent agencies, including the FTC, exercise much broader powers than that. If the narrow reading of Humphrey’s Executor is correct, then many of those agencies are in serious constitutional trouble. And indeed, the Court went out of its way to invite that conclusion. In an important footnote, it said that “we take the decision on its own terms, not through a gloss added by a later Court in dicta.”128 It added that “what matters is the set of powers the Court considered as the basis for its decision, not any latent powers that the agency may have had not alluded to by the Court.”129 The Court need not have said all of this, or any of it. It could easily have relied solely on the single fact that the CFPB was a single-headed agency. It could have been silent on the precise authorities of the FTC as the Court understood them in 1935 and thus left the independent status of the multimember commissions unquestioned.

In this light, it would be possible to mount a serious challenge to many of the current independent agencies, including the Securities and Exchange Commission, the Federal Communications Commission, the Federal Reserve Board, the Nuclear Regulatory Commission, and even the Federal Trade Commission in its current form. The work of each of these agencies—and all other independent agencies—must now be scrutinized to see if they (1) issue binding rules, (2) issue final decisions, in adjudications, ordering relief, (3) seek monetary penalties against private parties, and (4) display other features of concern to the Seila Law majority, such as independent budgetary authority. The relationship among these factors is unclear. Perhaps the presence of all four is enough to require an agency to be subject to presidential removal at will; perhaps the presence of even one is enough. It is easily imaginable, at least in principle, that hardly any independent agencies would survive.

We suspect that a majority of the Court would not be prepared to be so aggressive. If it were, the Court would wreak havoc on institutions that have been central to the U.S. government for many decades. But for the first time in decades, the constitutional status of the independent agencies has become insecure.

VI.  The Living Constitution

Seila Law claims to be an originalist and textualist opinion—and in some respects, it is. But the Court did not spend a great deal of time on the original public meaning of Article II; it did not engage the competing historical accounts. As is not infrequently the case with originalism, at critical points, the analysis turns to contestable interpretations of abstract principles to decide what the Constitution means. As far as express text is concerned, we have explored the view that the Vesting Clause and the Take Care Clause grant the President unrestricted removal power. But the Constitution says essentially nothing specific about removal as such, and it offers only a few sideways glances at what would today be considered administrative officers. But even if one is inclined to read the Vesting Clause and the Take Care Clause to create a strongly unitary executive, to extrapolate from Article II very specific and reticulated rules—for example, that multimember independent agencies are constitutionally tolerable while single-member independent agencies are intolerable, or that two levels of for-cause removal are intolerable even if one level is tolerable—requires additional premises. The Court finds such premises in “structural inferences” that yield high-level abstract “principles” such as “liberty,” “accountability” and “separation of powers.” As Justice Kagan wrote in dissent, acutely on this methodological point130:

It is bad enough to “extrapolat[e]” from the “general constitutional language” of Article II’s Vesting Clause an unrestricted removal power constraining Congress’s ability to legislate under the Necessary and Proper Clause. It is still worse to extrapolate from the Constitution’s general structure (division of powers) and implicit values (liberty) a limit on Congress’s express power to create administrative bodies. And more: to extrapolate from such sources a distinction as prosaic as that between the SEC and the CFPB—i.e., between a multi-headed and single-headed agency. That is, to adapt a phrase (or two) from our precedent, “more than” the emanations of “the text will bear.” By using abstract separation-of-powers arguments for such purposes, the Court “appropriate[s]” the “power delegated to Congress by the Necessary and Proper Clause” to compose the government.

Our point here is not that Justice Kagan is necessarily right or the majority necessarily wrong. It is that the majority is engaged in a process of reasoning from abstract principles of constitutional-political morality—what the Court calls “first principles”131—and adopting contested conceptions of those principles in ways that can reasonably be called Dworkinian. If this is originalism, one wonders what is not originalism.132 It is not surprising, therefore, that the dissent discusses at some length Dean John Manning’s argument133 that the Constitution contains no general, abstract principles such as “the separation of powers.” Nor is it surprising that the Court squarely rejected that argument, stating that although “there [is not a] a ‘separation of powers clause’ or a ‘federalism clause’ [in the express constitutional text, t]hese foundational doctrines are … evident from the Constitution’s vesting of certain powers in certain bodies.”134 For the Court, the text’s distribution of certain powers to certain bodies are in fact evidence of true, underlying, subsistent constitutional principles that are themselves part of “the law.” If this is not a form of structural analysis, and Dworkinian, it isn’t clear what would be.

Consider the Court’s most critical legal move, seemingly confining Humphrey’s Executor to its facts and then declaring that new “exceptions” to presidential removal will not be allowed. This turns out to be a normatively complicated exercise, one resolvable only by arguments about the best interpretation of abstract principles of political morality, such as “liberty” and “accountability.” The problem is that which facts are legally relevant cannot be read off from Humphrey’s Executor, let alone from the vesting of the “executive power” or the Decision of 1789, in any simple way. Humphrey’s Executor happened to involve an agency whose name begins with an “F”; is that legally relevant? Why not? Or to take a somewhat less random fact: Humphrey’s Executor happened to involve an agency that was concerned with monopoly and fair competition. Is that relevant? Certainly the fact that the agency’s name began with F, and that the FTC dealt with monopoly and fair competition, formed no part of the Court’s rationale in 1935. Likewise, Humphrey’s Executor also happened to involve an agency headed by multiple members, but the difference between multiple heads and a single head was also not a part of the Court’s rationale in 1935.135 If that feature is to make a difference, unlike the myriad of other features one might discern in “the facts,” there will have to be a normative argument to that effect.

The Court certainly does offer such arguments; it is not true that the Court simply declares, by fiat, that Humphrey’s Executor is different. It suggests that in a multimember agency, the members check each other in some way, and that this produces a kind of mutual accountability, an additional safeguard for constitutional liberty136:

The CFPB’s single-Director structure contravenes this carefully calibrated system by vesting significant governmental power in the hands of a single individual accountable to no one.… [T]he Director may unilaterally, without meaningful supervision, issue final regulations, oversee adjudications, set enforcement priorities, initiate prosecutions, and determine what penalties to impose on private parties. With no colleagues to persuade, and no boss or electorate looking over her shoulder, the Director may dictate and enforce policy for a vital segment of the economy affecting millions of Americans.… [T]he agency’s single-Director structure means the President will not have the opportunity to appoint any other leaders— such as a chair or fellow members of a Commission or Board—who can serve as a check on the Director’s authority and help bring the agency in line with the President’s preferred policies.

The point of these arguments is not that, somewhere in the text or original understanding of Article II, a distinction is drawn between single-member-headed agencies and multiple-member-headed agencies. It is that the Court is attributing to the “structure” of the whole Constitution broad principles and then arguing for a particular conception of those principles. Justice Kagan characterized the Court’s conception of “liberty,” for example, as one of “anti-power-concentration.”137 The Court’s particular conception is not some straightforward exercise in originalism; it is closer to political philosophy. It is contestable and premised on a thick, normative view of constitutional liberty.

On the merits, it is hardly obvious, of course, that the Court in fact has the right conceptions of these contested principles. It is hardly obvious that the Court was right to say that a single-headed independent agency is worse, from the constitutional point of view, than an independent agency headed by a multimember commission. Justice Kagan offered a forceful response, based on different conceptions of the same principles138:

If a removal provision violates the separation of powers, it is because the measure so deprives the President of control over an official as to impede his own constitutional functions. But with or without a for-cause removal provision, the President has at least as much control over an individual as over a commission—and possibly more. That means the constitutional concern is, if anything, ameliorated when the agency has a single head.

Ultimately, however, the merits are not the point. The key thing is that the Court’s “originalism” ends up deeply engaged in arguments about competing conceptions of abstract principles of political morality.

In this light, Seila Law is best understood not as a vindication of the original understanding, and hence as a Cell (1) case, but in terms of Cell (2), as a response to contemporary fears and concerns, expressed in terms of political morality. Many observers, and several of the justices themselves, have expressed grave concerns about the rise of a powerful administrative apparatus that often exercises broad discretion.139 Those concerns have been manifested in multiple domains, including heightened interest in the nondelegation doctrine140 and intense skepticism about judicial deference to agency interpretations of law.141 Serious concerns about the administrative state seem to define the modern era.142

On one view, the rise of the modern administrative state is a kind of constitutional barnacle, raising real threats to core constitutional values. That view could be founded in Cell (2), not Cell (4). The idea of a strongly unitary executive might seem a necessary corrective.143 If administrative agencies are going to have a great deal of discretion and wield awesome power, a minimal requirement might be that they must act under the constraining arm of an elected President. Perhaps that is an essential way of maintaining continuity with core constitutional commitments under unanticipated circumstances. On this view, Seila Law can be seen as the culmination (thus far) of a sustained and emphatically modern effort to insist on the strongly unitary executive as a response to the problem of administrative discretion.144 Indeed, it can be taken as a close sibling to District of Columbia v. Heller,145 also written in originalist terms but plausibly taken as a response to contemporary concerns.

Here is another way to put the point. We have noted that in his work on legal reasoning, Dworkin urged that judges have a duty to fit the existing legal materials and also to justify them, in the sense of putting them in the best constructive light.146 Seila Law can easily be seen as a case for which the existing legal materials were indeterminate. As we have seen, plausible opinions could take multiple forms. In choosing the approach it did, the majority claimed to speak in originalist terms, but it quickly turned to judgments, grounded in abstract principles, about what would make the constitutional order the best that it could be. On that view, the Court’s emphasis on accountability and liberty, and its sharpened duty under contemporary circumstances, were not throwaway lines. They were essential.

VII.  Conclusion

For many decades, there has been a sharp dispute between those who believe in a strongly unitary presidency, in accordance with the idea that the President must have unrestricted removal power over high-level officials entrusted with implementation of federal law, and those who believe in a weakly unitary presidency, in accordance with the view that Congress may, under the Necessary and Proper Clause, restrict the President’s removal power so long as the restriction does not prevent the President from carrying out the President’s constitutionally specified functions.

Both positions can claim some support from the original understanding of relevant clauses; both can also claim to keep faith with constitutional commitments in light of dramatically changed circumstances, above all the rise of the modern administrative state. In Seila Law, the Court wholeheartedly accepted the strongly unitary position, in an opinion that appeared to accept Humphrey’s Executor but that read the case so narrowly that it left a great deal of room for constitutional challenges to many independent regulatory commissions in their present form. The Court’s analysis purports to be rooted in the original understanding of the constitutional text, and its conclusion can certainly claim support from that understanding. But that conclusion might also be understood and defended as a response to contemporary fears and concerns about a powerful, discretion-wielding administrative state—and as reflective of a judgment that a necessary response to those concerns is a firm insistence on firm presidential control.


Cass R. Sunstein is the Robert Walmsley University Professor, Harvard University. Adrian Vermeule is the Ralph S. Tyler, Jr. Professor of Constitutional Law, Harvard Law School.

Author’s note: We are grateful to David Strauss for excellent comments and to Eli Nachmany for exceptional research assistance and valuable comments.

1 See Steven Calabresi and Christopher Yoo, The Unitary Executive (2008).

2 We are speaking here only of the removal power, not the directive power, that is, the power to issue orders to those who implement the law.

3 272 U.S. 52 (1926). We believe that the phrase “unambiguous embrace” is accurate, but as we shall see, the Court did not conclude that all those who operate within the executive branch are unequivocally subject to the President’s will.

4 295 U.S. 602 (1935).

5 Seila Law v. Consumer Fin. Protection Bureau, 140 S. Ct. 2183 (2020).

6 PHH Corp. v. Consumer Fin. Protection Bureau, 839 F.3d 1 (D.C. Cir. 2016), vacated and remanded on rehearing en banc, 881 F.3d 75 (D.C. Cir. 2018).

7 Seila Law, 140 S. Ct. at 2199–2200.

8 Id. at 2200 (internal quotation marks and citation omitted).

9 See Seila Law, 100 S. Ct. at 219–21 (Thomas, J., concurring in part and dissenting in part).

10 Id. at 2216.

11 554 U.S. 570 (2008).

12 See Ronald Dworkin, Law’s Empire (1986).

13 See David A. Strauss, The Living Constitution (2010).

14 See Seila Law, 140 S. Ct. at 2202–04.

15 See Dworkin, supra note 12.

16 See Calabresi and Yoo, supra note 1; Lawrence Lessig & Cass R. Sunstein, The President and the Administration, 94 Colum. L. Rev. 1 (1994).

17 For different views, see, e.g., Elena Kagan, Presidential Administration, 114 Harv L. Rev. 2245 (2001); Steven Calabresi & Saikrishna Prakash, The President’s Power to Execute the Laws, 104 Yale L.J. 541 (1994).

18 See Calabresi & Prakash, supra note 17. We are bracketing here the definition of “inferior officers” and the precise relationship between them and the President. As noted, an important issue, not explored here. is whether the President is able to control them by issuing directions they are obligated to obey, even if they are not his at-will employees.

19 As discussed below, those who believe in a strongly unitary President might accept some limits on this principle; they might not believe in the directive power. On that issue, see Cass R. Sunstein and Adrian Vermeule, Presidential Review: The President’s Statutory Power Over Independent Agencies, Geo. L. J. (forthcoming 2021); Robert V. Percival, Who’s In Charge? Does the President Have Directive Authority Over Agency Regulatory Decisions?, 79 Fordham L. Rev. 2487 (2011).

20 See Gerhard Casper, An Essay in Separation of Powers: Some Early Versions and Practices, 30 Wm. & Mary L. Rev. 211 (1989); Lessig & Sunstein, supra note 16.

21 See id.

22 See Seila Law, 140 S. Ct. 2224–45 (Kagan, J., concurring in the judgment with respect to severability and dissenting in part).

23 Wiener v. United States, 357 U.S. 349 (1958).

24 Morrison v. Olson, 487 U.S. 654 (1988).

25 Id. at 691.

26 Some influential and informative work on this topic distinguishes between “formal” and “functional” approaches, where the former term refers to constitutional text and the latter term refers to constitutional goals and purposes. See Peter L. Strauss, Formal and Functional Approaches to Separation of Powers Questions: A Foolish Inconsistency, 72 Cornell L. Rev. 488 (1987). It is not clear that this distinction easily maps onto recent decisions, discussed below.

27 See Calabresi & Prakash, supra note 17, at 551.

28 U.S. Const. art. II, § 1.

29 U.S. Const. art. II. § 3.

30 See Calabresi & Prakash, supra note 17, at 576–77.

31 See id. at 617.

32 See Joseph Postell, Bureaucracy in America 49–57 (2017).

33 The Federalist No. 70 (Alexander Hamilton).

34 Id.

35 See Myers v. United States, 272 U.S. 52 (1926); Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 492 (2010); Saikrishna Prakash, New Light on the Decision of 1789, 91 Cornell L. Rev. 1021 (2006).

36 See id. at 1026.

37 See Geoffrey P. Miller, Independent Agencies, 1986 S. Ct. Rev. 41, 47.

38 See Prakash, supra note 35, at 1026 (“In passing three acts in 1789 that assumed the President enjoyed a preexisting removal power, majorities in the House and Senate affirmed the executive power theory on three separate occasions. Members of Congress understood that votes in favor of the acts were votes favoring the executive power theory. Following these votes, members of Congress and newspaper accounts repeatedly described the final acts as endorsing the theory that the Constitution granted the president a removal power”).

39 See Casper, supra note 20; Jerry Mashaw, Creating the Administrative Constitution (2012); Peter Shane, The Originalist Myth of the Unitary Executive, 19 U. Pa. J. Const. L. 322 (2016).

40 Myers, 272 U.S. at 177 (Holmes, J., dissenting).

41 The Federalist No. 77 (Alexander Hamilton).

42 Id.

43 Id.

44 Myers, 272 U.S. at 177 (Holmes, J., dissenting).

45 See Andrew Kent et al., Faithful Execution and Article II, 132 Harv. L. Rev. 2111 (2019).

46 See id.

47 U.S. Const. art. II, § 2.

48 See Akhil Reed Amar, Some Opinions on the Opinions Clause, 82 Va. L. Rev. 647 (1996); see also Peter L. Strauss, Overseer, or “The Decider?”: The President in Administrative Law, 75 Geo. Wash. L. Rev. 696 (2007).

49 This point is acknowledged in Prakash, supra note 35, at 1024–25, 1039 even though he finds the general conclusion authoritative. For a rejoinder, see Postell, supra note 32, at 84–89.

50 See Jed Shugarman, The First Congress Rejected Unitary Presidentialism: The Indecisions of 1789: Strategic Ambiguity and the Imaginary Unitary Executive (June 22, 2020) (unpublished manuscript),; Jed Shugarman, The Decisions of 1789 Were Non-Unitary: Removal by Judiciary and the Imaginary Unitary Executive (June 22, 2020) (unpublished manuscript),

51 Id. at 3, 1.

52 See Postell, supra note 32, at 84. Note, however, that Postell adds that “the practice of removals largely followed the Decision of 1789 during the first several decades of American history.” Id. at 89.

53 For discussion on some of the complexities with the example of the Treasury Department, see Lessig & Sunstein, supra note 16.

54 See Mashaw, supra note 39; Casper, supra note 20.

55 Casper, at 235.

56 See Strauss, supra note 13; Stephen Breyer, Active Liberty (2006); Dworkin, supra note 12; John Hart Ely, Democracy and Distrust (1980).

57 See Breyer, supra note 56; NLRB v. Noel Canning, 573 U.S. 513 (2014).

58 On the relevance of changing circumstances in general, see Strauss, supra note 13; Lawrence Lessig, Fidelity and Constraint (2019).

59 See Strauss, supra note 13.

60 See Lessig, supra note 58.

61 See Dworkin, supra note 12.

62 See Lessig & Sunstein, supra note 16, at 88.

63 See id. at 101.

64 See Strauss, supra note 13.

65 See Abner S. Greene, Checks and Balances in an Era of Presidential Lawmaking, 61 U. Chi. L. Rev. 123 (1994).

66 On that specter, see Cass R. Sunstein, Impeachment: A Citizen’s Guide (2017).

67 James Landis, The Administrative Process 46 (1925). For discussion, see Adrian Vermeule, Bureaucracy and Distrust: Landis, Jaffe and Kagan on the Administrative State, 130 Harv. L. Rev. 2463 (2017).

68 Greene, supra note 65, at 124.

69 See The Federalist No. 51 (James Madison).

70 The reference is to Myers v. United States, 272 U.S. 52 (1926), taken up below. Many years ago, one of the present authors (Sunstein) argued in favor of Cell (2). See Lessig & Sunstein, supra note 16. He remains there, with the proviso that he would not overrule Humphrey’s Executor and that the statutory standard—inefficiency, neglect of duty, malfeasance—can be interpreted to give the President the requisite authority. See Cass R. Sunstein & Adrian Vermeule, Presidential Review, 109 Geo. L.J. 637 (2021). But candor compels an acknowledgement that he now believes that the historical evidence on behalf of the strongly unitary executive is somewhat stronger than he once thought and that the structural argument, emphasizing changed circumstances, on behalf of the strongly unitary argument is somewhat weaker than he once thought.

71 We are playing, of course, on Justice Holmes’s famous formulation: “General propositions do not decide concrete cases.” Lochner v. New York, 198 U.S. 45, 76 (1905) (Holmes, J., dissenting).

72 See Humphrey’s Executor v. United States, 295 U.S. 602 (1935); Morrison v. Olson, 487 U.S. 654 (1988); Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477 (2010).

73 272 U.S. 52 (1926).

74 See generally id.

75 See Calabresi & Yoo, supra note 1, at 18.

76 Cf. Alfred North Whitehead, Process and Reality 39 (1978) (“The safest general characterization of the European philosophical tradition is that it consists of a series of footnotes to Plato.”).

77 295 U.S. 602 (1935). On the reviling, see, e.g., Miller, supra note 37, at 93. In stating that the opinion in Humphrey’s Executor is widely reviled, we do not mean to suggest that Myers is without its critics, certainly of its sheer breadth. See, e.g., Casper, supra note 20. And despite the reviling, it is important to emphasize that Humphrey’s Executor has had a constitutive effect on the structure of the administrative state, arguably more so than Myers. We are grateful to Daphna Renan for pressing this point.

78 295 U.S. at 618.

79 Id. at 629.

80 Peter L. Strauss, The Place of Agencies in Government: Separation of Powers and the Fourth Branch, 84 Colum. L. Rev. 573, 615 (1984) (noting that President Roosevelt “had given Commissioner Humphrey no particular directive; he had asked no advice that Humphrey then refused to give; he did not, perceiving insubordination, direct [Humphrey] to leave” and therefore the Court did not resolve the question “whether the President could give the FTC Commissioners binding directives … or what might be the consequences of any failure of theirs to honor them”).

81 See pages 29–30 infra.

82 See Nat’l Petroleum Refiners Ass’n v. Fed. Trade Comm’n, 482 F.2d 672 (D.C. Cir 1973).

83 See Mashaw, supra note 39.

84 272 U.S. at 135.

85 The reason for the word “perhaps” is that a court could say (1) Congress can have its own staff while also saying (2) Congress cannot immunize executive branch officials, compiling reports for it, from the plenary control of the President.

86 130 S. Ct. 3138 (2010).

87 Id. at 3174. A valuable treatment is Peter L. Strauss, On the Difficulties Of Generalization – PCAOB in the Footsteps of Myers, Humphrey’s Executor, Morrison and Freytag, 32 Cardozo L. Rev. 2255 (2011).

88 130 S. Ct. at 3174.

89 Id. at 3152.

90 Id. at 3155.

91 487 U.S. 654 (1988).

92 Id. at 691.

93 Id. at 692.

94 See In re Aiken County, 645 F.3d 428, 440 (D.C. Cir. 2011) (Kavanaugh, J., concurring in the judgment). Consider in particular this passage:

Reading only the text of Article II, one would assume that the Nuclear Regulatory Commission would report to the President, not the President to the Nuclear Regulatory Commission. If two agencies in the Executive Branch were not on the same page (as may happen in this case if the Nuclear Regulatory Commission rejects the Department of Energy’s withdrawal application), the President presumably would have the authority to resolve that disagreement. If an agency were departing from the President’s preferred course (as the Nuclear Regulatory Commission may do), the President presumably would have the authority to prevent that. And if an agency were taking too long to make a critical legal or policy decision (as appears to be the case with the Nuclear Regulatory Commission), the President presumably would have the authority to fix that as well.

95 Recall the Court’s pointed statement in Free Enterprise Fund: “The parties do not ask us to reexamine any of these precedents, and we do not do so.” Free Enter. Fund v. Pub. Co. Acct. Bd., 561 U.S. 477, 483 (2010).

96 See Humphrey’s Executor, 295 U.S. at 623.

97 See Sunstein & Vermeule, supra note 19.

98 Seila Law, 140 S. Ct. at 2191.

99 Id. at 2191–92 (citation omitted).

100 Id. at 2197.

101 Id. at 2192 (emphasis in original).

102 Id.

103 Id.

104 Id. at 2206 (citations omitted).

105 Id. at 2207.

106 See id. at 2198.

107 See id. at 2198–99

108 See id.

109 See id.

110 See id.

111 Id. at 2199–2200. The Court added this qualification in a footnote:

The Court’s conclusion that the FTC did not exercise executive power has not withstood the test of time. As we observed in Morrison v. Olson, 487 U. S. 654 (1988), “[I]t is hard to dispute that the powers of the FTC at the time of Humphrey’s Executor would at the present time be considered ‘executive,’ at least to some degree.” Id., at 690, n. 28. See also Arlington v. FCC, 569 U. S. 290, 305, n. 4 (2013) (even though the activities of administrative agencies “take ‘legislative’ and ‘judicial’ forms,” “they are exercises of—indeed, under our constitutional structure they must be exercises of—the ‘executive Power’” (quoting Art. II, §1, cl. 1)).

Id. at 2198 n.2. The relationship between this paragraph and the Court’s main analysis is puzzling. One way to square the two is to say that even if the FTC’s adjudicative-type functions and legislative-type functions “must be” executive, they are not the kinds of functions that must be subject to plenary presidential control.

112 See id. at 2200.

113 Id. at 2199 (quoting Morrison, 487 U.S. at 691).

114 Id. at 2200.

115 Id.

116 Id.

117 Id.

118 Id. (footnote omitted).

119 Id. at 2201.

120 Id. at 2202.

121 Id. at 2203.

122 Id.

123 Id. at 2207 (quoting Free Enterprise Fund, 561 U.S. at 499).

124 Id. at 2207.

125 This much follows from the Court’s analysis of the severability question. That analysis warrants a detailed discussion, but it would take us beyond our topic here.

126 See Cass R. Sunstein, The Office of Information and Regulatory Affairs: Myths and Realities, 126 Harv. L. Rev. 1838 (2013).

127 See Seila Law, 140 S. Ct. at 2198.

128 Id. at 2200 n.4.

129 Id.

130 Seila Law, 140 S. Ct. at 2243–44 (Kagan, J., concurring in the judgment with respect to severability and dissenting in part) (internal citations omitted).

131 Seila Law, 140 S. Ct. at 2206.

132 We are bracketing, and not necessarily rejecting, the proposition that the strongly unitary understanding, in its purest form, can be justified on originalist grounds; see above for competing views on that proposition. Our claim here is the Court’s analysis, and the distinctions that the Court deems critical, cannot be defended simply by reference to originalism.

133 See John Manning, Separation of Powers as Ordinary Interpretation, 124 Harv. L. Rev. 1939 (2011).

134 Seila Law, 140 S. Ct. at 2205.

135 It is true that the Court referred on several occasions to the fact that the FTC was a multimember commission—by way of description, not legal argument. Humphrey’s Executor did in fact make the point about the seven-year terms of the several commissioners as “necessary to the effective and fair administration of the law” in the sentence just before discussing the insulation from complete change in the leadership. See Humphrey’s Ex’r v. United States, 295 U.S. 602, 624 (1935). But that was no part of its holding or rationale, and it would be a stretch even to consider the references to be dicta.

136 Seila Law, 140 S. Ct. at 2203–04.

137 Id. at 2243 (Kagan, J., concurring in the judgment with respect to severability and dissenting in part).

138 Id. at 2255.

139 For a summary, see Cass R. Sunstein & Adrian Vermeule, Law and Leviathan (2020).

140 See, e.g., Gundy v. United States, 139 S. Ct. 2116, 2131–48 (2019) (Gorsuch, J., dissenting).

141 See, e.g., Kisor v. Wilkie, 139 S. Ct. 2400, 2425–48 (2019) (Gorsuch, J., dissenting).

142 See generally Philip Hamburger, Is Administrative Law Unlawful? (2014).

143 See Mistretta v. United States, 488 U.S. 361, 422–27 (1989) (Scalia, J., dissenting).

144 See Miller, supra note 37.

145 554 U.S. 570 (2008).

146 See Dworkin, supra note 12.