The Binary Executive
abstract. In recent years, the Supreme Court has begun to implement a “unitary” theory of the Executive, according to which the President alone holds the executive power. At the same time, the Court has greatly intensified its scrutiny of administrative policymaking, abandoning deference on questions of law and at times taking a steel-hard look at questions of policy and fact as well. These moves together create novel constitutional structures, internally contradictory jurisprudence, and unstable patterns of political rule. The unitary executive theory presumes that the President alone may wield executive power. But by wresting away agencies’ policymaking discretion, the Court itself exercises executive power. The Court is thus constructing the unitary executive with one hand and fragmenting it with the other. In this emerging regime, there are two chief executives: the President and the Court. The Executive is not unitary; it is binary. The binary executive is a constitutional anomaly that disturbs settled understandings, undermines the quality of government, and aggrandizes the Court at the expense of both the elected branches and its own legitimacy.
The old administrative law is ailing, and the new is not mature enough to take its place.1 Under the old administrative law of the twentieth century, courts deferred to agencies’ reasoned judgments,2 and agencies exercised regulatory power apart from the President’s policy preferences.3 The regulatory state thus remained relatively autonomous from both the judiciary and the White House. The new administrative law inverts this arrangement. In recent years, the Supreme Court has begun to implement a “unitary” theory of the Executive.4 According to this theory, the President alone holds the executive power. The Court has relied on this theory to restrict Congress’s ability to insulate administrative officials from presidential control.5 At the same time, the Court has greatly intensified its own scrutiny of administrative policymaking, abandoning deference on questions of law and at times taking a steel-hard look at questions of policy and fact as well.6 Whereas agencies were once fairly independent from both the President and the Court, they are now increasingly under the thumb of both.
These moves together create novel constitutional structures, internally contradictory jurisprudence, and unstable patterns of political rule. The unitary executive theory presumes that the President alone may exercise executive power. But by wresting away the policymaking discretion that Congress has delegated to executive agencies, the Court itself exercises executive power. Its recent rulings are best understood as “executive” insofar as they render particularized policy decisions without encoding general rules to govern the disposition of future cases. The Court is thus constructing the unitary executive with one hand and fragmenting it with the other. In this emerging regime, there are two chief executives: the President and the Court.7 The Executive is not unitary; it is binary. The binary executive is a constitutional anomaly that disturbs settled understandings, undermines the quality of governance, and aggrandizes the Court at the expense of both the elected branches and its own legitimacy.8
Though theoretically inconsistent, the unitary executive theory and aggressive judicial review nonetheless cohere in another, political respect. The binary executive squeezes the administrative state in a pincer movement. Under the Court’s unitary executive theory, the President may control how executive officers exercise their duties. Under the Court’s approach to judicial review, administrative expertise appears to count for little, at least on politically salient matters. The result is that agencies face pressure both from above, in the executive hierarchy, and laterally, from judicial review. The competencies that have historically justified administrative power, such as subject-matter specialization and expertise, organizational divisions of labor, internal checks, and deliberative processes lose their purchase in the face of these complementary constitutional pressures.9 In the emerging model, the executive power is either presidential or judicial. It is not administrative. The binary executive makes sense, therefore, not as a coherent legal theory, but as an effort to accomplish the “deconstruction of the administrative state.”10
In their hurry to undo and rework the pattern of American governance without a stable jurisprudential foundation, the Justices increasingly take institutional ownership over controversial matters of public policy. They make highly visible and consequential decisions without the electoral mandate, professional expertise, or public input that ordinarily accompany administrative action. At the same time, the Justices strip the executive branch of the expert and deliberative virtues that Congress instilled and the Court itself shaped over the course of the twentieth century. The binary executive is therefore less likely to render well-reasoned decisions and more likely to expose the Court to colorable charges of arbitrary, ideologically motivated adjudication. The “authority of the Court” may suffer from these political perils.11
It is no coincidence that this transition was already underway when Justice Breyer’s tenure on the Court came to an end. Breyer was a key architect of the traditional administrative law, offering a particularly nuanced and practical understanding of how Congress and administrative agencies operate.12 But he also laid an early foundation for the binary executive by arguing for enhanced judicial scrutiny of agency policy decisions that concern “major questions.”13 The Court has magnified Breyer’s approach to judicial review almost beyond recognition while rejecting his much more modest conception of presidential power.14 Breyer envisioned an important role for administrative independence and expertise under the mutually restrained supervision of the three constitutional branches. In the wake of his departure, the Court is entering uncharted waters of joint judicial and presidential supremacy.
Part I of this Essay identifies the fundamental features of the emerging regime of administrative law. Part II explains why it is apt to characterize this regime as a “binary” form of executive power. Part III highlights the risks the binary executive poses for constitutional norms, democratic accountability, and good government.
The administrative law we traditionally learn and teach in law school crystallized in the late 1970s and 1980s, as the Supreme Court responded to the transition from the New Deal order to the deregulatory regime of the late Carter and Reagan Administrations.15 This arrangement had two cornerstones. First, the President enjoyed a powerful supervisory role over administration but could not call all the shots. For example, when confronted in Morrison v. Olson16 with a constitutional challenge to the independent “special counsel” investigation into political corruption in the Reagan administration, the Court held the President’s power at bay. In his majority opinion, Chief Justice Rehnquist did not identify any conflict between the Constitution’s grant of executive power to the President and statutory restrictions on the removal of an inferior officer with limited prosecutorial jurisdiction. This approach was consistent with, and arguably deepened, the Court’s solicitude for legislative restraints on the President’s control of administration, which had been a hallmark of constitutional law since at least the 1930s.17
The second cornerstone of this arrangement was substantial (but not unlimited) administrative policymaking discretion. Reviewing a change in air-pollution-control policy at the Environmental Protection Agency (EPA), the Court held in Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc.18 that the judiciary must accept reasonable agency interpretations of statutory ambiguities. Chevron’s reasonableness requirement, and related principles of arbitrary-and-capricious review, had some teeth. The Court would scrutinize agencies’ justifications for policy changes to make sure they were internally consistent, guided by the relevant statutory factors, and cognizant of alternative approaches.19 But it also reined in lower courts’ more adventurous efforts to hone administrative procedures to promote public participation and fair decision-making20 and often took a lenient approach when reviewing agency policymaking.21
So much for the old way. The new approach moves in the opposite direction on both fronts. As to presidential power, the Court has adopted a unitary theory of the Executive. The modern incarnation of the unitary executive theory first appeared in the halls of the Reagan Department of Justice,22 but Justice Scalia’s dissent in Morrison introduced it into the case law.23 According to this controversial theory, the Constitution grants the President not “some of the executive power, but all of the executive power,”24 which means the President must be able to control all executive-branch officials.
The unitary executive transformed from a theory favored by a sole dissenter to one adopted by a majority of the Court in a series of twenty-first-century cases. First, in Free Enterprise Fund v. Public Co. Accounting Oversight Board,25 the Court held that Congress could not protect executive officers with two layers of for-cause removal protections. Then, in Seila Law LLC v. Consumer Financial Protection Bureau (CFPB),26 the Court held that Congress could not restrict the President’s ability to remove the sole Director of CFPB. Echoing Justice Scalia, Chief Justice Roberts based the Court’s holding that the President had the “prerogative to remove executive officials” on arguments about constitutional text, democratic accountability, and personal liberty.27 Roberts concluded that hierarchical control by an elected President both reinforces electoral control over administration and safeguards private liberties threatened by government action.28 This unitary logic continues to unfurl in other cases concerning the removal, as well as the appointment, of executive officers.29 A majority of the Justices seem keen to increase the President’s exclusive control over the executive branch.
The Court also grants great deference to the President on matters relating to national security. In Trump v. Hawaii,30 the Court rejected a constitutional challenge to President Trump’s “travel ban” from a number of majority Muslim countries, despite bigoted statements the President himself had made to justify earlier versions of the ban. In approving the ban, the Court emphasized the need to protect “the authority of the Presidency itself”31 and gave credence to the “worldwide review”32 that cabinet agencies had conducted to rationalize it.
Meanwhile, the Court has substantially tightened its review of administrative action in the domestic regulatory sphere. Until recently, Chevron was the north star for administrative law, providing a central, if ever-contested and evolving, principle for understanding the appropriate relationship between Congress, the Court, and the President.33 But the Court has not relied on Chevron to defer to agency action since 2016.34 Instead, the Court increasingly relies on the “major questions doctrine” to fill in its own interpretations of ambiguous statutory text.35 In Alabama Ass’n of Realtors v. Department of Health & Human Services,36 for example, the Court held that the Centers for Disease Control’s (CDC’s) eviction moratorium was likely unlawful—despite quite broad statutory text—because the “sheer scope of the CDC’s claimed authority . . . counsel[ed] against the Government’s interpretation.”37 Likewise, in National Federation of Independent Business (NFIB) v. Department of Labor (DOL), the Court upheld an injunction against the Occupational Safety and Health Administration’s (OSHA’s) vaccine-or-testing mandate because it was a “significant encroachment into the lives—and health—of a vast number of employees” without explicit congressional authorization.38 Finally, in West Virginia v. EPA,39 the Court held that EPA lacked authority to regulate the power grid as a whole under the Clean Air Act. The Court was incredulous that “Congress implicitly tasked [EPA], and it alone, with balancing the many vital considerations of national policy implicated in deciding how Americans will get their energy.”40 Such a vast exercise of authority “raise[d]” the majority’s “eyebrow.”41
The major questions doctrine is not the only example of the Court applying heightened scrutiny to agency action. In rejecting the Trump Administration’s effort to add a citizenship question to the census, the Court took the rare step of questioning the sincerity of the government’s proffered explanation.42 Rather than follow the more conventional approach of taking a hard look at the logical consistency of the government’s explanation, as Justice Breyer did in his partial concurrence,43 the majority concluded that the government’s explanation was pretextual.44 While the latter conclusion had some foundation in the case law45 (and the unusual facts and litigation history46), it also reflected the Court’s greater willingness to “probe the mental processes” of executive-branch decision-makers than in years past.47
The Court also took a very hard look at the Trump Administration’s effort to rescind the Deferred Action for Childhood Arrivals (DACA) policy. The Court emphasized that the Department of Homeland Security had failed to entertain less disruptive alternatives or consider the “serious reliance interests” involved.48 While the merits of this ruling were consistent with traditional principles of administrative law, the Court pushed those principles nearly to their breaking point by second-guessing an exercise of enforcement discretion—a matter usually left to executive judgment.49 For the second time in recent years,50 the Court intervened in the Secretary of Homeland Security’s exercise of their statutory authority to “establish national immigration enforcement policies and priorities.”51 More such cases are pending.52
To be clear, the point is not that the census and DACA cases were wrongly decided or inconsistent with the judicial role.53 Rather, the point is simply that these cases were highly nondeferential. As I’ll explain further in the next Part, this nondeferential approach is in deep tension with the Court’s view that the President alone exercises executive power.
The Supreme Court is making two moves at once: heightening presidential control of administration and intensifying judicial review of regulatory action. Together, these moves bisect the executive power, even as the Court adopts a “unitary” theory of executive power. By the executive power, I mean the power to carry out or implement the law.54 As we shall see, this simple definition raises problems on close inspection, as the judiciary also carries out the law in various ways. These boundary-drawing problems are manageable within a functionalist approach to the separation of powers, which recognizes that each branch may exercise a share of the powers exercised by the others. The Roberts Court, however, has adopted a formalist and unitary conception of executive power that precludes the judiciary from partaking in execution; at the same time, it has insisted on de novo review of agency policymaking. As a result, the Court is getting caught up in a maze of doctrinal and structural tensions, if not outright contradictions. While pumping up the presidency, the Justices are taking a share of executive power for themselves and acting collectively as the President’s cochief of the federal government. I will have more to say about why such practices are constitutionally unsound in the next Part. But first I will substantiate the charge.
There is a well-established, if fraught, division of labor in the United States’s constitutional system. Congress makes the law.55 The Executive implements and enforces the law, exercising discretion where a statute so permits.56 The judiciary interprets the law and the Constitution when an appropriate “case” or “controversy” comes before it.57 The courts may generally invalidate the exercise of administrative power when the Executive acts contrary to legal requirements, or without legal authority, to the detriment of a private party’s protected interests.58
The key distinction between the province of the judiciary and that of the Executive in this admittedly simplified rendering of the separation of powers is that the courts interpret the law, whereas the Executive exercises political discretion. This contrast goes all the way back to Marbury v. Madison.59 In explaining the scope of the writ of mandamus, the Marbury Court explained that “in cases in which the executive possesses a constitutional or a legal discretion, nothing can be more perfectly clear than that [executive officers’] acts are only politically examinable.”60 That is to say, the President could supervise officers in the exercise of their powers or remove officers who improvidently exercise their discretion, and the people could hold the President accountable for those choices. Or Congress might impeach officers for serious breaches of official duty. By contrast, “where a specific duty is assigned by law, and individual rights depend on the performance of that duty,” the writ would lie to compel the Executive to perform the obligation.61 The distinction between discretionary and obligatory “ministerial” duties laid a key foundation for nineteenth-century administrative law.62 During that era, review was “bipolar”—courts either applied de novo review to matters of administrative law and fact or treated them as committed to executive discretion.63
Twentieth-century administrative law softened but retained this distinction between law and discretion.64 Contests between the elected branches and the Supreme Court over the Interstate Commerce Commission’s policymaking powers yielded a settlement in which the Court would resolve questions of constitutional and statutory authority but could not, “under the guise of exerting judicial power, usurp merely administrative functions by setting aside a lawful administrative order upon [their] conception as to whether the administrative power ha[d] been wisely exercised.”65 In the New Deal era, the Court built on this approach and adopted deferential standards in reviewing questions of policy resolved during rulemaking or adjudicatory proceedings.66 The Administrative Procedure Act of 1946 (APA) subsequently enshrined agencies’ authority to make “policy” through rulemaking67 and stated that some matters could be “committed to agency discretion by law.”68 In the wake of the APA, the Court again affirmed that it could not substitute its view of wise policy for that of the relevant administrative agency.69 Notably, Chevron itself justified judicial deference on the ground that the President was accountable to the electorate for his “policy choices,” whereas unelected federal judges lacked such accountability.70
Judicial review of administrative action nonetheless routinely risks intrusion into executive discretion. Whenever a court is called upon to review general administrative policies, especially in the rulemaking context, it necessarily gets involved to some extent in the political domain.71 For example, when the Supreme Court says that a policy on seatbelts and airbags is unlawful because it is unjustified72 or that the rejection of a petition to regulate greenhouse gases has not been adequately explained,73 it is making (or unmaking) decisions that involve a substantial degree of political judgment. In such cases, the result is not clearly dictated by statutory text. Rather, it is based on a broader assessment that the agency was not being as thoughtful as Congress expected it to be. The Court, it seems, knows better.
The problem of distinguishing executive from judicial power is a theoretically deep one. While the Constitution assigns the “judicial Power”74 and the “executive Power”75 to different actors, that distinction is hardly self-interpreting or stable. John Locke did not separate the two powers at all, describing both the judicial and monarchical powers as “executive.”76 Montesquieu—whose theory of the separation of powers directly influenced the Framers—described the “executive power” as encompassing both a foreign-affairs power and the “power of judging.”77 The reason these thinkers conceived of the executive and judicial powers as one is that each power involves the application or implementation of law. While there is arguably a fundamental conceptual difference between the creation and the application of law,78 it is difficult to categorically distinguish the judicial task from executive or administrative ones. Under the Constitution, the judicial power only arises when there is a concrete and particularized “case” or “controversy.”79 But it has long been established that executive agencies may adjudicate some such disputes as well.80 Likewise, the judiciary frequently engages in an executive function when it determines how open-ended statutory schemes ought to operate by considering practicalities as well as text and principle.81
If the distinction between the executive and judicial powers is difficult to maintain at a conceptual level, and the courts have long engaged in some degree of policymaking when reviewing administrative action, what makes the current Court’s jurisprudence special? Why call this form of government the “binary executive,” while withholding that label from the Reagan-, New Deal-, or Progressive-era models?
There is difference in kind, rather than degree, between these earlier models and the Supreme Court’s current posture of review. As Bijal Shah observes, courts reviewing agency action engage in “judicial administration,” which is analogous to what Elena Kagan called “presidential administration.”82 The form of judicial administration practiced over the course of the twentieth century consisted of varying degrees of “oversight,” akin to the President’s indirect, supervisory role over agencies. With the decline of Chevron deference, however, the Court has moved from the role of “overseer” into the role of “decider,” akin to the role the President plays in the unitary theory of the Executive.83 The Court dictates how administrative discretion is to be exercised within the bounds of statutory authority.
To be sure, there are differences between the Court’s practice of judicial review and the President’s control over administration. The Court generally exercises appellate jurisdiction over challenges to administrative action, whereas the President need not wait for an appeal to come to their desk.84 But the differences are not as great as they may initially seem. The White House, like the Court, conducts “review” of administrative action through the regulatory-review process at the Office of Information and Regulatory Affairs (OIRA).85 The Court, for its part, frequently reviews challenges brought by industry representatives, public-interest groups, and states, rather than individual parties.86 Review in these cases sounds more in public policy than the defense of individual rights, given that these parties usually sue to vindicate collective political interests. The Court may not conduct cost-benefit analysis the way OIRA does, nor does it respond directly to electoral incentives the way the President does. But this difference merely underscores that the Court is performing traditionally executive tasks without the benefit of the tools and the salutary incentives the Executive has. This is the fundamental flaw in the binary executive, to be explored further in the next Part.
For now, consider the inherently political nature of the Court’s task when it reviews broad grants of statutory authority. Recall the theory underlying Chevron deference: where a regulatory statute is ambiguous, Congress intends the administering agency to assume “policy-making responsibilities” and exercise discretion in interpreting and applying its commands.87 The relevant agency could, in the performance of those duties, “rely upon the incumbent administration’s view of wise policy to inform its judgments.”88 But when the Court acts as decider, it “fill[s]” in those statutory “gap[s]” itself89 and thus displaces the Executive’s view of wise policy with its own. The Court presents its policy determinations as the ineluctable consequence of a clear legal requirement. Presenting such policy choices as mere interpretation obfuscates the exercise of executive power.90
The Court’s approach is particularly novel because it combines this aggressive posture of review with its plenary theory of presidential power. Twentieth-century administrative law did not treat the President as the controller of all administrative action. So, when the Court exercised policymaking powers in reviewing agency action, it did not necessarily infringe upon the President’s power. Rather, it could be seen as operating in a distinct “administrative” sphere subject to the control of each of the three constitutional branches. By contrast, the unitary view the Court has recently embraced posits that the President alone has the power to control agency policymaking. As a result, when the Court engages in de novo interpretation of ambiguous regulatory statutes, it directly competes with the President for the authority to decide matters of policy. With both the unitary executive theory and de novo review ascendant, there is no longer any independent administrative space to mediate conflict between presidential and judicial prerogatives. The White House and the Court instead face off as constitutional principals governing the same province of power, each purporting to determine how law works on the ground.
The binary executive incorporates a formalist viewpoint that heightens the conflict between plenary presidential and judicial control over administration. Over the course of the twentieth century, the view that the President and the President alone had mandatory constitutional control over the administration of law was considered eccentric.91 The predominant understandings, by contrast, recognized multiple points of departure between presidential power and the administration of law.92 For some constitutional theorists, executive power was not synonymous with administrative power.93 For others, the true fault line across all three branches was between “politics” and “administration.”94 And for the Court itself, “quasi judicial” and “quasi legislative” powers were not subject to presidential control, even when exercised by administrative agencies.95 Constitutional jurisprudence, more broadly, did not conceive of the separation-of-powers scheme in exclusively formalist terms, where there were only three kinds of power exercised by three separate constitutional actors. Rather, administrative and structural constitutional law were at least as functionalist as they were formalist and frequently acknowledged the overlap and coordination among the three branches.96
If one maintains a functionalist approach to the separation of powers, there is nothing particularly problematic about the conceptual ambiguity and practical contestation at the boundaries of executive and judicial power.97 The functionalist will acknowledge that courts and agencies have overlapping responsibilities. The courts engage in a “dialogue” with agencies and the other constitutional branches over the common task of governance.98 The goal in this arrangement is to ensure that the three branches “share the reins of control” over agencies99 without displacing one another’s complementary powers, competencies, and judgments. The Court has historically managed the risk of judicial aggrandizement by evaluating an agency’s reasons for its action, the cogency of its explanation, and its compliance with procedural requirements, rather than by directly interrogating the substance of the agency’s decision.100 While this approach does not eliminate judicial policymaking, it creates a workable division of labor between agencies and courts. For instance, the hard-look analysis the Court conducted in reviewing the rescission of DACA would make sense from a functionalist perspective, even if it approaches the boundary of the judicial role. Since there is nothing inherently problematic about the Court sharing in the exercise of policymaking power, the Court may probe the record of the agency’s decision-making in great depth where serious questions of public policy are at issue.
That pragmatic mixture of judicial and administrative functions is unavailable under the formalist theory of the unitary executive. The formalist asserts that executing is categorically different from judging and that each must be done independently of the other.101 Judicial review of agency policymaking then becomes problematic to the extent that it requires the courts to engage in policymaking themselves.102 It is theoretically possible to maintain the formalist division of labor if judicial review is restricted to determining whether an agency action is “plainly or palpably inconsistent with the law.”103 If the agency acts manifestly contrary to a statute’s terms to the detriment of a private party, judicial review of that policy is not only permissible but necessary to ensure that the executive branch does not act beyond its powers.
But the Supreme Court currently exercises a form of judicial review that is orders of magnitude more intense than that highly deferential approach. Consider the vaccine-or-testing-mandate case, NFIB v. DOL.104 The Court elected to resolve that case, as well as the eviction-moratorium case, through its orders, or “shadow,” docket—a once-rare but increasingly frequent means of intervening in important issues of national policy.105 This comparatively abbreviated and opaque form of adjudication, issued in unexplained or cursorily explained orders, aggravates the risk that the Court’s “review” of a major policy decision constitutes its own discretionary political decision.
That risk materialized in NFIB v. DOL. The statute in question states that the Secretary of Labor
shall provide . . . for an emergency temporary standard . . . if he determines (A) that employees are exposed to grave danger from exposure to substances or agents determined to be toxic or physically harmful or from new hazards, and (B) that such emergency standard is necessary to protect employees from such danger.106
The Court did not contest the sufficiency of the agency’s factual finding that COVID-19 was both physically harmful and a new hazard, nor that the mandate was necessary to protect employees from danger. Rather, the Court made the peculiar claim that COVID was “not an occupational hazard” because it also exists outside of the workplace.107 In response to the obvious objection that OSHA regulates many risks that exist beyond the workplace—like fire or asbestos—the Court responded that a vaccination “cannot be undone at the end of the workday.”108 But the Court did not explain why that fact was relevant to the statutory scheme. In the absence of careful reasoning, the Court’s interference in executive policy was an exercise of will, governed not by legal rules but by broad, subjective values about the appropriate scale and use of political power. The Court did not, in legislative fashion, lay down a general rule. Instead, like a President, the Court deemed a particular exercise of power inappropriate.
In West Virginia v. EPA,109 the Court again waded into a political dispute using questionable methods of statutory interpretation. The regulation at issue, designed to reduce greenhouse-gas emissions, had never been implemented and, according to the government, never would be.110 Therefore, the Court’s decision was, as Justice Kagan noted in dissent, an “advisory opinion” untethered from a concrete controversy.111 Though the statute granted EPA authority to set the “best system of emission reduction,”112 the Court found that this delegation could not encompass the authority to regulate the power grid as a whole. Without a clear anchor in the statutory text or established principles of construction, the Court determined that the power EPA claimed was simply too great to be inferred from the statute’s broad delegation of authority. “This Court could not wait,” Kagan observed, “to constrain EPA’s efforts to address climate change.”113
One might reply that these conclusions do flow from legal requirements—in particular, the “major questions doctrine” invoked by the majority.114 However, the version of the major questions doctrine the Court deployed in these cases is both legally faulty and highly laden with discretionary political judgment. As a result, reliance on the major questions doctrine is not legal interpretation at all, but rather an exercise of raw political power. This exercise of power is best characterized as executive because the cases failed to lay down a general and administrable rule to govern future cases or private conduct. Rather, in its haste to decide, the Court merely resolved the policy questions at issue by resorting to broad political values. Such an exercise of power is most analogous not to legislation, but rather to executive directives.
Prior to the COVID-19 cases,115 the major questions doctrine usually operated as an exception to Chevron deference: the Court would not defer to agencies’ resolution of questions that involved matters of “deep economic and political significance.”116 That principle helped courts determine whether Congress meant to give agencies—rather than the judiciary—the power to interpret an ambiguous provision. Such an interpretive principle may or not be advisable, but it is broadly consistent with the judicial role. It instructs courts to interpret certain statutes for themselves rather than deferring to agencies. There is a world of difference between that narrower major questions doctrine and the new one, which simply denies agencies the authority to exercise significant power without explicit congressional authorization.
Today’s stronger version of the major questions doctrine cuts the interpretive task short. A consistent feature of the recent major-questions cases is abbreviated, if not cursory, interpretive analysis. Once the Court determines that an exercise of administrative power is big, it simply determines that the agency does not have authority to act. Given the ambiguity of concepts like “major,” “deep,” and “vast,” as well as the lack of clarity about how “express” statutory authorization must be, the doctrine creates great leeway for courts to make subjective judgments concerning the scope and content of administrative power.
If the major questions doctrine were grounded in actual or constructive evidence of legislative intent, as Chief Justice Roberts claimed in West Virginia,117 it would be an exercise of the Court’s ordinary interpretive power. But it is not.118 The text and legislative history of the APA, for instance, show that legislators expected agencies to resolve major questions.119 Even the conservative opponents of the New Deal who urged the adoption of a stringent set of administrative procedures, such as Roscoe Pound, did not object to agencies exercising “important quasi judicial or quasi legislative action” per se.120 Rather, they objected to the procedural informality and thus unconstrained discretion open to administrative officers. Congress’s solution was to provide a set of transsubstantive procedural requirements and expectations to enhance public accountability.121 When agencies decided “[m]atters of great import” through rulemaking, Congress expected them to incorporate fulsome public participation, above and beyond the opportunity for public comment.122 Likewise, “important issues of law or policy” could be adjudicated when agency heads reviewed impartial hearing examiners’ initial decisions.123 Either form of decision would be subject to judicial review under deferential standards.124 The APA, in other words, contemplates that agencies may decide important political questions. There would have been little reason for an APA in the first place if Congress did not expect agencies to make major decisions. Congress designed the Act to inform and control this routine exercise of administrative power.
It might be argued that times have changed. Recent empirical studies indicate that most congressional staffers today do not believe Congress intends to allow agencies to resolve major policy questions by enacting ambiguous provisions.125 But these results have limited external validity with regard to the stronger major questions doctrine, especially as applied to old statutes. Rather than assuming that Congress did not intend for courts to defer to agencies on such questions, the stronger major questions doctrine assumes that major administrative powers do not exist if not expressly granted. Moreover, contemporary staffers’ awareness of major-questions precedents also tells us little about legislators’ intent in enacting much older statutes like the Clean Air Act, the Occupational Safety and Health Act, and the Public Health Service Act, at times when the major-questions precedents did not exist. To the contrary, it is facially implausible to assume that legislators during the expansive regulatory era between the New Deal and Civil Rights Revolution in the 1970s could not reasonably have intended for agencies to exercise major powers in pursuit of statutory objectives. Broad grants of regulatory power—constrained by judicial review of agency reason-giving and progressively broadening public-interest standing—were par for the course.126
For this reason, it is difficult to argue that the major questions doctrine somehow respects or protects the power of the legislature. The doctrine instead constrains the legislative power by requiring perfect clarity when an agency’s exercise of power might plausibly be construed as major. It does not fit the mold of other extratextual principles, which merely help to elucidate and circumscribe ambiguous text based on factually substantiated assumptions about legislative objectives. For instance, courts will sometimes interpret text by reference to the “mischief” that Congress set out to address when it enacted a statute.127 This can help to clarify what sorts of conduct a statute allows or prescribes. Unlike the mischief rule, the major questions doctrine does not rest on well-grounded claims about the nature of linguistic or legal communication.128 While it may be common sense to assume that Congress enacts statutes to fix problems, it is not common sense to assume that Congress grants all significant powers expressly. Such an assumption might actually prevent Congress from addressing vast, pressing, and shifting social problems by requiring an unrealistic degree of precision about the best means to do so. As noted, implicit delegation of major powers to agencies has been a routine feature of legislation, and one explicitly contemplated by legislators, going back at least as far as the APA—the “subconstitution” for the American administrative state.129
Proponents of the major questions doctrine may argue that it is not based on legislative intent; rather, it is a substantive canon of construction meant to avoid nondelegation difficulties. However, as Daniel Deacon and Leah Litman recently pointed out, the major questions doctrine, as currently formulated, does not actually safeguard the constitutional value of nondelegation.130 As Chief Justice Roberts said in forbidding EPA to regulate the electric grid, “a decision of such magnitude and consequence rests with Congress itself, or an agency acting pursuant to a clear delegation from that representative body.”131 If delegation of a major decision to an agency is not a problem so long as the delegation is “clear,” then the major questions doctrine cannot cure any nondelegation issue. At most, the doctrine increases the cost of enacting expansive regulatory legislation, requiring express statutory language for a major delegation.
The Court’s choice to limit and direct how congressional power is put to work may or may not be wise as a matter of policy. But it certainly does not flow from the Court’s responsibility to “say what the law is” in appropriate cases or controversies.132 Rather, it flows from value-based decisions about how much and what sorts of governmental power ought to be exercised. Where property and personal liberty may be broadly impacted by a governmental policy, the Court at times declines to enforce the law against affected parties. That kind of enforcement discretion traditionally lies with the Executive, not with the Court.133
The Justices relying on the major questions doctrine in recent cases would surely reject this characterization. In his concurrence in NFIB, for example, Justice Gorsuch said, “The only question is whether an administrative agency in Washington, may mandate the vaccination or regular testing of 84 million people. Or whether . . . that work belongs to state and local governments and the people’s elected representatives in Congress.”134 On this account, the Court was simply regulating the proper constitutional relationship between Congress and the Executive, and between the federal government and the states. It was not deciding, Gorsuch insisted, but instead determining “[w]ho decides.”135
This defense is revealing. Note that Justice Gorsuch does not acknowledge that the President is also one of “the people’s elected representatives.” Nor does he acknowledge that he and his colleagues have elsewhere treated the President as the sole possessor of the executive power. According to the Justices’ own unitary executive theory, it was not an “administrative agency in Washington,” but rather the President, who imposed the vaccine-or-testing mandate. But acknowledging the President’s electoral mandate and constitutional authority over these matters would complicate Gorsuch’s effort to paint the vaccine-or-testing mandate as an exercise of extraconstitutional power—one that licensed judicial intervention in an intensely political decision.
NFIB presents a series of decisions by three constitutional actors: Congress’s decision to create OSHA and give it emergency powers relating to workplace safety; the President’s decision to use that authority to issue a vaccine-or-testing mandate; and, finally, the Court’s decision to countermand the President’s order, relying on an unsound doctrine of its own recent invention. Perhaps, then, Justice Gorsuch felt compelled to clarify that “[t]his Court is not a public health authority”136 because it was acting like one. OSHA imposed a vaccine mandate; the Court stopped it. If the question is “[w]ho decides,” then the answer in this case was the Court.
One might argue that I am making too much of a handful of recent cases and one particularly objectionable doctrine. To be sure, it is not as though the Court invariably relies on extratextual standards to second-guess and invalidate executive-branch decisions: just this term, the Court rejected a challenge to the Biden Administration’s revocation of the Trump Administration’s remain-in-Mexico policy,137 as well as a challenge to statutory interpretations offered by the Centers for Medicaid and Medicare Services.138 But even in these cases, the Court has turned away from the deference regime that existed in traditional administrative law—one that would be more consistent with the unitary executive theory. Each of these decisions involved ambiguous text and policymaking discretion, which, under Chevron, would have warranted deference.139 Instead, the Court reviewed each interpretive question de novo. Because the traditional tools of construction do not include a judicial prerogative to resolve questions of policy left open by the statute and delegated to an administrative agency,140 these decisions reallocate executive power from the relevant administrative agency to the Court itself.
More recently, the Court left the Fifth Circuit’s vacatur of the Biden Administration’s immigration-enforcement guidance in place pending review on the merits.141 The lower court’s vacatur relied on a narrow conception of immigration-enforcement discretion and an expansive gloss of the court’s jurisdiction to conclude that the guidance was unlawful.142 Immigration enforcement has long rested squarely within the bounds of executive power: it implicates foreign-policy decisions over which the President has constitutionally inherent authority, involves the allocation of scarce enforcement resources, and flows from a complex maze of congressional statutes that expressly grant enforcement discretion to executive officials.143 By leaving the lower court’s vacatur in place before ruling on the merits, the Supreme Court kept the judiciary’s hands on the tiller of executive power.
The Court has thus shown a clear propensity to tighten the screws on executive action, even as it proclaims that all executive power is held in the hands of the President. Its new brand of administrative law may augment both the President’s powers and its own. But, as the next Part explains, this binary-executive model leaves the Court vulnerable to charges of ideologically motivated decision-making and is likely to undermine the quality of constitutional government.
The binary executive is constitutionally fraught. It combines maximalist theories of presidential and judicial power that work at cross-purposes. It also undermines political accountability while threatening the expert bases of administrative decision-making. Furthermore, it implicates the Supreme Court in politically salient and controversial decisions, thus threatening its already-troubled legitimacy.
The Court’s theory of executive power and its administrative-law jurisprudence are dissonant. If the President has exclusive control over law’s administration, then the Court should not be able to countermand policy judgments that are left to officers under the President’s control. The unitary executive theory would have no legal consequence if executive officers had no political discretion: in that event, all executive decisions would be ministerial and subordinate executive officers’ actions would be dictated by statutory requirements. It is only because the careful execution of law frequently involves discretion that the President might have space to direct how subordinate officers do their jobs. Yet the Court’s maximalist theory of review strips the President of authority over many discretionary decisions that the unitary theory purports to vest in him. The result is that the executive power is divided between the President and the judiciary.
There is a clear loser in this arrangement: the administrative state. Both unitary presidential control and de novo judicial review of policymaking dismiss or at least sideline the relevance of procedures and practices that have long been thought to legitimate administrative action: professional expertise, public participation in rulemaking, and reason-giving.144 When the President simply controls agency output and has the authority to dismiss agency staff with policymaking responsibilities, there is little room for agencies to make decisions informed by their staffs’ subject-matter expertise and the concerns and values of affected constituencies. When the Justices simply resolve any significant policy question according to their own reading of ambiguous statutory terms, rather than probing the agency’s justification in depth, the distinctive blend of legal, political, and pragmatic reasoning that characterizes agency statutory interpretation loses its purchase.145 Instead, regulatory output is determined by the competing and arbitrary influence of presidential and judicial dictates.
The result is constitutionally problematic for some of the same reasons advanced by proponents of the unitary executive theory. Alexander Hamilton’s original argument for executive “unity” rests in large part on the value of clear lines of political accountability running from the lowest-ranking executive officers up to a single chief.146 He argued that, where there are multiple chief executives,
[i]t often becomes impossible, amidst mutual accusations, to determine on whom the blame or the punishment of a pernicious measure, or series of pernicious measures ought really to fall. It is shifted from one to another with so much dexterity, and under such plausible appearances, that public opinion is left in suspense about the real author.147
The Court itself recently relied on a similar argument to expand the President’s removal power.148 But the same problem of confused authorship arises when the Court participates in the exercise of executive power. When the Court frequently thwarts the executive branch’s major policy initiatives, it becomes more difficult for the public to accurately attribute responsibility to the President for policies enacted (or not enacted) during their administration.
Another Hamiltonian problem with the binary executive is that it simply makes policymaking more conflictual and less efficient, especially under urgent conditions. Hamilton argued that where two officers are “clothed with equal dignity and authority,” there is “always danger of difference of opinion,” which may cause “animosity” and “impede or frustrate the most important measures of government, in the most critical emergencies of the state.”149 There is no doubt that, during the “critical emergencies of the state” posed by COVID-19 and climate change, we have had more than one cook in the kitchen. The government cannot act with “[d]ecision, activity, secrecy, and dispatch”150 when the Court is prepared to thwart government action via emergency relief granted on the shadow docket. Worse, as Hamilton suggested, conflict over matters of policy merely multiplies opportunities to foment the extraordinary political animosity and polarization that already exist in the American polity. This problem is especially worrisome given that the judicial pole of the binary executive is made up of life-tenured Justices who often disagree vociferously—and publicly—about the correct course of action. One could hardly imagine a less auspicious context for energetic and accountable policy implementation.
It is worth noting the particular target of Hamilton’s criticism: a set of proposals for an executive council of state, which had been considered but rejected during the Constitutional Convention.151 One model for the council would have placed the Chief Justice of the Supreme Court at the head of an executive body responsible for advising the President.152 Others would have given a council of revision, including members of the judiciary, the power to veto national legislation.153 The principle that the Supreme Court may not issue advisory opinions is grounded in part in the Convention’s rejection of such proposals.154
The Court’s recent jurisprudence arrogates to the judiciary this executory role, which the Framers considered and ultimately rejected. In West Virginia v. EPA,155 when the Court passed judgment on a regulatory scheme that had never been and would never be put into effect, it took on a role analogous to that of a council of state, providing the Executive with an ex ante ruling on whether an exercise of authority might or might not be lawful. That function is core to administrative courts in other countries, such as the French Conseil d’État, which do participate actively in the exercise of executive power.156 But it is foreign to the constitutional tradition of the United States.
One need not adopt Hamilton’s unitary theory of the Executive to recognize that this diffusion of power and responsibility between the President and the Court is problematic. There are strong grounds to believe that the Constitution does not require, as theorists of the unitary executive posit, that the President alone exercises the executive power, without any space for dissensus and plurality within the executive branch. As Peter M. Shane and Peter L. Strauss have argued, the Opinion Clause implies the existence of multiple, discrete duties within the executive branch, as well as shared responsibility among the President and the principal officers.157 Likewise, I have argued that the “Departments” contemplated by the Constitution and established by Congress constrain the exercise of executive power with stable regulations, administrative practices, and reason-giving requirements.158 Such pluralistic conceptions of executive power compensate for a diminution in direct political accountability with gains in expertise, reasoned explanation, and public participation. Agencies have the benefit of organizational structures and professional staff capable of sifting for relevant information, collecting public comments, and bringing scientific expertise and political values to bear.159
The binary executive offers no such compensation. The Court’s approach to judicial review has generally short-circuited its own traditional processes of reasoned decision-making as well as the Executive’s. With the notable exception of the DACA case, Department of Homeland Security v. Regents of the University of California,160 the Court’s recent administrative-law opinions do not probe the soundness of an agency’s explanation; instead, the Justices often refuse to consider the agency’s policy justification. In the major-questions cases, the Court relies on extratextual principles to cabin executive discretion and avoid thoroughly analyzing legislative text, structure, and history. Once the Court has determined that a question is major and finds the statute lacks perfect clarity, it does not probe the record to assess whether the agency decision was well supported, as it would in an ordinary arbitrary-and-capricious-review case. It thus avoids the hard work of hard-look review. This approach results in less fulsome analysis. In other cases, such as Biden v. Texas,161 the Court has transformed questions of immigration-enforcement policy into matters of pure legal interpretation without any deference to the administering agencies. Such oblique interventions sound in political judgment but do not provide other officials or the public with policy-based explanations. Discretion masquerades as law.
This is not to say that prior courts were innocent of making political considerations, or to invite a return to a fictional past where law once reigned supreme. The Court has long participated in the exercise of executive power to some degree. To reiterate, the binary executive is distinctive because of its novel combination of unilateral presidential power with de novo review of executive policymaking, under the banner of formalism and the strict separation of powers. In this precarious posture, the Court exhibits a false consciousness, insisting on the purely legal character of conclusions that are plainly laden with policy preferences. An approach that embraced some combination of functionalism, a plural theory of the Executive, or deferential review might effectively manage the ever-present tension between law and politics. The binary executive cannot. As a result, the Court’s mode of pretextual legal explanation merely stands in the way of forthrightly justifying the policy judgments that lie just beneath the surface.
The problem is related to the judicial style of reasoning. As David E. Pozen and Adam M. Samaha have argued, there are certain “anti-modalities” of constitutional reasoning that the Court is not permitted to trade in—in particular, express policy judgment.162 That is equally true in administrative law, where the Court is not supposed to make a substantive judgment about the wisdom of agencies’ decisions. The Court comes closest to acknowledging its political commitments when it references the generalized menace that federal regulations pose to liberty, property, and popular sovereignty, without also acknowledging the benefits of such regulations.163 But constitutional proprieties seem to preclude the Court from elucidating its political theory in greater detail. This prevents the Court from thinking clearly about the consequences of its decisions and the public from assessing whether the Court has exercised its discretion wisely. Judgments become both less sound and less transparent, decreasing public welfare and accountability.164
One can see the Court straining underneath the constraints of legal nicety in a way that reflects the Justices’ assertion of political power. As the Court wades ever deeper into political decision-making, it begins to take on more of the trappings of American presidentialism. Since President Wilson, the American presidency has been what Jeffrey K. Tulis calls a “rhetorical presidency.”165 Presidents use speeches to shape public opinion, mobilize constituencies for policy change, and gather support for legislative proposals. Today we are increasingly treated to a rhetorical Court, both on and off the bench.166 As Jon D. Michaels observes, Justices and judges increasingly act as participants within political disputes—”ballers”—rather than as neutral umpires.167 In public speeches, the Justices sometimes make their political preferences known, speaking out in favor of political perspectives and particular litigants who may come before the Court again.168 In judicial opinions concerning administrative power, the Justices make value-based arguments against the administrative state, emphasizing vague, underspecified threats it poses to individual freedom and popular sovereignty.169 In his concurrence in West Virginia v. EPA, Justice Gorsuch even went so far as to associate the administrative state with racial bigotry.170 These claims do not make sense under a formalist conception of legal reasoning. They register, instead, as partisan appeals and indictments. Justices speak not only to one another but to the broader public in attempting to shape the basic commitments and policy goals of the federal government.
One problem with our rhetorical Court is that it is a much less democratically accountable body than the President or Congress. As Alexander M. Bickel recognized and Antonin Scalia affirmed, the Court is in many respects a “counter-majoritarian” institution, designed to check democratic political power.171 When the Court makes bad policy judgments, it is therefore extremely difficult for the political process to police and correct its errors. It is not possible to simply elect different judges. Laws must be changed, perhaps even judges added, or else the advice-and-consent process must await the Justices’ retirements or deaths. This is part of the reason judges ought to exercise the “passive virtues” when they review matters touching upon the political process.172 The current Court is lacking in these virtues, eager to get into the game and exercise its share of political power. There were sound prudential reasons, for instance, to decline to hear the challenge to the Clean Power Plan, since the plan was not in effect and would never go into effect.173 By wading into that “political thicket,”174 the Court took some ownership over the government’s future capacity to address a global environmental crisis.
The issue of ownership raises a final difficulty for the binary executive. The more the Court enters controversial political waters, the more likely it is to alienate and even anger significant portions of the American public. A judiciary that lacks the political authorization, doctrinal resources, and institutional competence to exercise executive power risks losing credibility when its makes and unmakes major policy decisions. The Court surely exercises more power over the short term when it takes a portion of the executive power for itself. But in the process, it may cast its own authority into doubt.
The new administrative law is still young. It might, in time, mature into a stable and coherent set of doctrines. Perhaps the Justices have simply been in a hurry to transform the law governing the executive branch and will work out the kinks later. But there is reason to be skeptical that law will work itself pure in this way. The binary executive is inherently destabilizing. It allocates governmental power haphazardly between two potentially conflicting poles—the Court and the presidency. This approach heightens the risks of fracture within public law and our broader constitutional structure. It opens the door for episodic judicial breaches into executive departments, explicable not in terms of doctrine or principle but rather naked political preference.
Such forays are a mistake. The administrative institutions created and tended to by generations of legislators, presidents, and justices serve practical human purposes. In addition to providing concrete services and protections to citizens and taxpayers, the administrative state itself provides a fulcrum on which to balance the values expressed in current politics and those embodied in older legal norms.175 At their best, agencies bring political and economic forces into conversation with legal and scientific disciplines, yielding decisions that are at once motivated by policy judgment and sensitive to normative constraint. If the Court prevents agencies from mediating such tensions between law and politics, the Court will have to take on more of that work for itself. Neither it, nor the public, nor the constitutional order are likely to benefit from such an extraordinary accrual of power in a few unelected hands.
Professor of Law, UCLA School of Law. Thanks to Mark Lemley, Leah Litman, and Bijal Shah for their comments and Sarah Walker and the staff of the Yale Law Journal for their editorial input.