What Seila Law Says About Chief Justice Roberts’ View of the Administrative State

In Seila Law LLC v. Consumer Financial Protection Board, the Supreme Court invalidated a statutory provision that protected the director of the Consumer Finance Protection Board (CFPB) from removal by the president except for “inefficiency, neglect of duty, or malfeasance in office.” Writing for the Court, Chief Justice John Roberts announced a new test for evaluating the constitutionality of “for cause” restrictions on presidential removal of high-level agency officials. Under this test, the Court asks whether the removal restriction applies to an official who is the head of a “single-head agency” or to the officials who collectively lead a “multimember expert agency,” prohibiting the former and permitting the latter. This test is remarkable both because it changes the law and because of how it changes the law: it lets the structure of the agency determine the degree of presidential control over its principal officers.

But Seila Law is remarkable in another way. To develop the new test, Chief Justice Roberts offered a vision of separation of powers that finally allows us to see fully how he views the administrative state. That vision helps to explain the real problem with the CFPB’s structure. It also helps to explain puzzling or shocking decisions that predate Seila Law. We can now understand what Roberts did a decade ago in Free Enterprise Fund v. Public Company Accounting Oversight Board, a removal decision that to this day seems to defy logic and put at risk important agency officials. We can also understand what he did a week before Seila Law in Department of Homeland Security v. Regents of the University of California (the DACA case), an administrative law decision that surprised most everyone and seemed to many like a betrayal of his conservative principles. Despite their differences, all these cases can be explained by the same concern for preventing abuse of power and protecting individual liberty in the administrative state, which Roberts traces to our constitutional structure.

In Seila Law, the Court confronted a for-cause removal restriction that was unlike those that it had seen in previous cases. The removal restriction applied to the sole head of an agency, the director of the CFPB, and the relevant precedent concerned removal restrictions that applied to the body of officials who led an agency, or to an individual inferior officer. The first question in the case (after reviewability of the case) was whether the prior decisions, Humphrey’s Executor and Morrison, controlled the analysis. In Humphrey’s Executor, the Court approved the constitutionality of the Federal Trade Commission (FTC), an agency led by a body whose members are removable by the president only for good cause, reasoning that the agency exercised “quasi-legislative” and “quasi-judicial” functions, not purely executive ones that would have precluded any removal restriction, as had been the case with the postmaster in the seminal removal decision Myers v. United States. In Morrison, the Court upheld a for-cause removal restriction on the independent counsel, an official who exercised predominantly executive functions, asking whether the removal restriction “impede[s] the President’s ability to perform his constitutional duty” and answering that question in the negative. Turning to the CFPB, the Court held that it did not fit either case because it was not led by a “group of principal officers,” like the FTC, and the director was not an “inferior officer,” like the independent counsel.

To explain more precisely why the prior cases did not apply and to determine what rule should, the Court—or really Chief Justice Roberts writing for the Court—offered a particular conception of the constitutional structure. Article II mentions two categories of officers: “principal officers” and “inferior officers.”In Roberts’ view, inferior officers like the independent counsel occupy their own box because by constitutional design, they possess limited power. He did not say much about the broader reasoning in Morrison and mostly cut to the chase: “we concluded that the removal protections did not unduly interfere with the functioning of the Executive Branch because ‘the independent counsel [was] an inferior officer under the Appointments Clause, with limited jurisdiction and tenure and lacking policymaking or significant administrative authority.’”

As to principal officers, Chief Justice Roberts said much more. He subdivided principal officers into two groups: those who are politically accountable through the president, and those who run agencies with certain “organizational features” that substitute for political accountability as a means of controlling agency policy. Although Article II does not subdivide principal officers in this way, in Roberts’ view, separation of powers does. The constitutional structure manifests two strategies for controlling power within the political branches—election and subdivision. On Roberts’ account, the Framers determined that the nation needed the energy and dispatch of a single Executive, which required the concentration of power in a single person “not bog[ged] . . . down with the ‘habitual feebleness and dilatoriness’ that comes with a ‘diversity of views and opinions.’” But the concentration of executive power in a single individual, even though separated from legislative power, raised the threat of fiat and entailed a commensurate check. The Framers made the president the most politically accountable official in the government—the only official elected by the entire nation.1

The legislative power was different. The political branch that makes law for the entire nation is prone to tyranny of the majority, and the Framers relied on subdivision into houses as a check.2 Subdivision was for the legislative power what political accountability was for the executive power. It prevented abuse of power and protected liberty in a way tailored to the type of threat the branch poses.

Roberts then imported these strategies to the removal context. In Roberts’ view, nothing in separation of powers prohibited Congress from making similar choices when designing agencies. An agency could have a head who was politically accountable, or an agency could have structural protections that substituted for political accountability as a means of preventing abuse of power. Thus, Congress has leeway to choose one form or the other depending on the particular needs of government. Roberts clarified through his analysis, however, that Congress was not permitted to relax the demands of the constitutional structure in either instance. In other words, when Congress creates an agency, it must ensure political accountability to a comparable degree or division of power to a comparable extent as the constitutional structure. Thus, Roberts stated that “individual executive officials may wield significant authority, but that authority remains subject to the ongoing supervision and control of the elected President.” When considering the director of the CFPB in particular, he clarified that “supervision and control” of individual agency heads means “meaningfully controlled (through the threat of removal).” This is where Congress went wrong in creating the CFPB: “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.”

Roberts scoffed at the dissent’s contention that the CFPB’s structure was permissible because modern presidents have mechanisms short of removal to control agency heads, like “oversight devices (e.g., centralized review of rulemaking or litigating positions), budgetary processes, personal outreach, and more.” Roberts characterized the dissent as “brainstorm[ing]” these mechanisms and the mechanisms themselves as “bureaucratic minutiae”—low-level trivialities, not constitution-level constraints. Perhaps political control over appropriations and appointment, which are in the Constitution, would have made it a closer case. He noted that the director is not beholden to either Congress or the president, as even most independent agencies are, for help navigating the appropriations process because funding was set in the statute. So too, the president may not have an opportunity to appoint a new director because the director’s term might extend past the president’s term in office. Roberts said that these features made lack of at-will removal “worse,” suggesting that control over appropriations and appointments are worthy forms of political accountability. But he was fairly clear that they could not substitute for at-will removal. The “insulation from removal by an accountable President is enough to render the agency’s structure unconstitutional.”

According to Roberts, if Congress wanted to limit presidential removal of the CFPB director in response to the need for expertise or impartiality, it could have designed the agency with structural protections in place of political accountability. He pointed to the FTC, the agency at issue in Humphrey’s Executor, as a model. As Roberts detailed, the FTC has certain “organizational features”: 

• The FTC is “[c]omposed of five members—no more than three from the same political party—the Board was designed to be ‘non-partisan’ and to ‘act with entire impartiality.’”
• “The FTC’s duties were ‘neither political nor executive,’ but instead called for ‘the trained judgment of a body of experts’ ‘informed by experience.’”
• “And the Commissioners’ staggered, seven-year terms enabled the agency to accumulate technical expertise and avoid a ‘complete change’ in leadership ‘at any one time.’”

The FTC is a multimember body, like Congress, where diverse perspectives would “‘promote deliberation and circumspection.’” Its duties required expertise and impartiality, and the appointment and removal of the commissioners (fixed and staggered terms) promoted the development of expertise over time and prevented political derailment. In Roberts’ view, these features, these “structural protections,” went the distance toward satisfying separation of powers. Of course, Congress had not taken this route in creating the CFPB. Even if the CFPB possessed some feature of an “expert” agency, it was not a multimember agency.

Reduced to a test, Roberts’ analysis comes down to this question: does the for-cause removal restriction apply to the head of a single-headed agency or to the officials who lead a multimember agency? For-cause removal restrictions are prohibited for the single heads and permissible for the multiple officials. This test puts an end to classifying principal officers by their functions (executive, quasi-legislative, quasi-judicial), which, as Roberts acknowledged, does not track how agencies are designed. In terms of simplicity, it confines ad hoc balancing of the official’s functions and the president’s constitutional duties to inferior officers, like the independent counsel. Instead, it makes agency structure the key to presidential removal.

At a deeper level, the new test provides Congress the leeway to choose political accountability or structural protections as a means of controlling agency power. Depending on the needs of government, Congress can choose one or the other. But it cannot choose neither, as it had done with the CFPB. The director was not subject to at-will presidential removal, and the CFPB was not a multimember agency. As a result, the CFPB’s structure was not just unusual in our governmental system, one that the Court had never seen before. It “clashe[d]” with the constitutional structure. The director was unleashed from any structure that could prevent abuse of power and protect individual liberty.

Seila Law makes a stunning change to removal law. And stepping back, the decision is striking in another way. Whether right or wrong, Roberts’ opinion provides a more complete and comprehensible view of how he sees the administrative state. After Seila Law, we can finally understand what Roberts was saying ten years ago in another significant removal decision, Free Enterprise Fund v. Public Company Accounting Oversight Board, and what effect that decision will likely have on other important agency officials. More surprising, we can also understand what he did just days before Seila Law in the highly charged DACA case, an administrative law decision, which reached a “liberal” political result.

In Free Enterprise, Roberts, writing for the Court, invalidated a removal restriction on members of the Public Company Accountability and Oversight Board (PCAOB). The PCAOB’s members are inferior officers appointed by the Securities and Exchange Commission (SEC), whose members are removable for cause by the president. In Seila Law, Roberts cited and quoted his opinion in Free Enterprise quite a bit. But when he wrote that opinion ten years before, the logic of it was clearer in his head than on paper. The question was whether the dual-layer removal restriction unduly interfered with the president’s constitutional duties, and Roberts said yes—the PCAOB is too remote from the president for him to meaningfully control its members. Although the Court has upheld, and the president can tolerate, one layer of removal, two layers is one too many. Justice Stephen Breyer, in dissent, was perplexed. To paraphrase his argument, the removal restriction on the members of the PCAOB did not take a bigger bite out of the president than the removal restriction on the SEC already did. The problem is that the president cannot fire the members of the SEC over a disagreement, including a disagreement whether or not to fire a member of the PCAOB for cause (say, because the member of the PCAOB took a bribe). Even if the SEC could remove the members of the PCAOB at will, the president is stuck with the SEC’s decision whether to fire the members of the PCAOB. Where is the logic?

It turns out that Justice Breyer was right as to level of presidential control, the political accountability check, but, in light of Seila Law, Roberts can be understood as focused on a different problem: the structural check. Roberts described the dual for-cause removal structure as a “diffusion of power,” a “diffusion of accountability,” and a “dispersion of responsibility.” The logic goes something like the following: The SEC has the standard multimember agency structure, in which structural protections substitute for political accountability (i.e., at-will presidential removal) as a check on agency power. But the PCAOB is not a standard multimember agency. It is a multimember agency led by a group of inferior officers, which creates confusion. In order to be inferior officers for appointment purposes, the members of the PCAOB must be subordinate to their boss, the SEC. If they are subordinate, then the multimember structure of the PCAOB may not be doing what it does for the SEC—providing a structural check on the agency’s power in place of accountability. You need one or the other—either meaningful accountability (i.e., at-will removal by someone) or meaningful multimember structure (i.e., the members truly check each other and their boss does not overrule them). With the PCAOB’s structure, you have an unacceptable mix of both or at least obfuscation of control. Instead of a division of power, you have a “diffusion of power,” a “diffusion of accountability,” and a “dispersion of responsibility.”

This understanding is helpful in another way. Justice Breyer was also concerned in Free Enterprise that the majority’s decision jeopardized a long list of agency officials protected by dual for cause structures, most notably administrative law judges (ALJs), who are removable for cause by the Merit Systems Protection Board, whose members are removable for cause by the president. He raised this concern again in Lucia v. SEC, which concerned whether SEC ALJs are inferior officers for purposes of the Appointments Clause or mere employees of the government. Justice Breyer has less reason to worry after Seila Law, and with the benefit of the Court’s decision in Lucia. In Lucia, the Court held that SEC ALJs are inferior officers rather than mere employees of the government, and therefore their appointment by SEC staff violated the Appointments Clause. To decide the issue, the Court considered the features of SEC ALJs, observing that they have “extensive powers—the ‘authority to do all things necessary and appropriate to discharge his or her duties’ and ensure a ‘fair and orderly’ adversarial proceeding . . . ‘includ[ing], but [ ] not limited to,’ supervising discovery; issuing, revoking, or modifying subpoenas; deciding motions; ruling on the admissibility of evidence; administering oaths; hearing and examining witnesses; generally “[r]egulating the course of” the proceeding and the “conduct of the parties and their counsel”; and imposing sanctions for ‘[c]ontemptuous conduct’ or violations of procedural requirements.” As the Court summarized, “an SEC ALJ exercises authority ‘comparable to’ that of a federal district judge conducting a bench trial.” The Court found that these features render SEC ALJs inferior officers rather than mere employees.

But the same features that make SEC ALJs inferior officers also address Justice Breyer’s concern about the constitutionality of their dual for-cause removal structure. For some agencies, the structural check is division of power, a multimember expert form. But an agency’s structural check, to be meaningful, must correspond to the agency’s particular duties. Given an SEC ALJ’s functions, we would expect the structural protections to be comparable to those of federal trial judges. Notwithstanding the removal layers, the SEC ALJ’s structure does not cause a “diffusion of power,” like the PCAOB’s does. It is just the way the constitutional structure would expect all ALJs to look and operate. ALJs and many of the agencies that worried Justice Breyer in Free Enterprise now seem safer.

Seila Law also helps to explain Roberts’ thinking in the DACA case decided a week before. That case was a challenge to the Trump administration’s decision to rescind two Obama-era immigrant relief programs—Deferred Action for Childhood Arrivals (DACA), which allowed unauthorized aliens who entered the country as children a two-year forbearance on removal, and Deferred Action for Parents of Americans and Lawful Permanent Residents (DAPA), which allowed parents who are unauthorized aliens to remain with their children who were U.S. citizens or lawful permanent residents for the same forbearance period. Roberts, writing for the majority—himself and Justices Ginsburg, Breyer, Sotomayor, and Kagan—held that the rescission was reviewable and remanded it to the Department of Homeland Security (DHS) as lacking an adequate explanation and as arbitrary and capricious under the Administrative Procedure Act (APA).

Roberts’ vote, let alone his opinion, was unexpected and seemed to many like a stick in the eye of the president. But it reflects the very same view of the administrative state—stated most generally, that agency power must be meaningfully checked to prevent abuse of power and protect individual liberty—that he would use just one week later to strengthen the hand of the president in a constitutional context. In the DACA case, Roberts reaffirmed long-standing principles of administrative law, despite avowed presidential control of DHS policy. With the hindsight of Seila Law, it is not surprising that Roberts enforced “foundational principle[s] of administrative law,” as he referred to them. These principles have long been understood to prevent abuse of power and protect individual liberty. They have also been applied specifically to address concerns about presidential control—that presidential control sometimes produces arbitrary agency action rather than checking it. In other words, they ensure that control of agency power, built in their structure, is meaningful in operation. As such, they fit perfectly into Roberts’ view of the administrative state.

Roberts was direct about his position on administrative law. For example, he defended the requirement that an agency provide a contemporaneous explanation for its decision (which Justice Brett Kavanaugh called “an idle and useless formality”), as “promot[ing] ‘agency accountability’ by ensuring that parties and the public can respond fully and in a timely manner to an agency’s exercise of authority,” “instill[ing] confidence that the reasons given are not simply ‘convenient litigating position[s],’” and facilitating “‘the orderly functioning of the process of [judicial] review.’” He also fully embraced Motor Vehicle Manufacturers Ass’n v. State Farm Mutual Automobile Insurance Co., the decision famous for elaborating the arbitrary and capricious test of the APA, and said that the Trump administration had committed precisely the same error as the Reagan administration by letting presidential preferences get in the way of considered judgment: “[The administration’s] reasoning repeated the error we identified in one of our leading modern administrative law cases, Motor Vehicle Manufacturers Association of the United States, Inc. v. State Farm Mutual Automobile Insurance Co.” Whatever constitutional prerogative the president has to control agency officials, he must follow the rules of administrative law like everyone else.

In the DACA case, Roberts showed a commitment to administrative law that is consistent with how he approaches removal of agency officials under constitutional law—as requiring a meaningful check against abuse of power to protect individual liberty. This insight into Robert’s thinking may not deflect criticism of his decision in the DACA case, just as his decision in Seila Law is subject to criticism. It does suggest, however, that Roberts is not nearly as inconsistent or unpredictable in these cases as it may seem to both sides. In addition, it reveals that administrative law is more secure in his hands than we might have thought.

Seila Law helps to show that Chief Justice Roberts has a particular view of the administrative state. To be clear, he has no love for the administrative state. In a prior case, he indicated that he would vote to revive the nondelegation doctrine under Article I to invalidate broad regulatory statutes and constrict agency authority. Roberts joined Justice Neil Gorsuch’s dissent making this point. It is difficult to know what becomes of removal law or administrative law if he gets the chance. But as long as this administrative state lives, we now have a better understanding of where he stands: agencies must have meaningful checks, in their design as a matter of constitutional law, and in their actual operation as a matter of administrative law, to prevent abuse of power and protect individual liberty.

• 1See Seila Law, 140 S. Ct. at 2203 (quoting Federalist No. 70 (A. Hamilton)) (“To justify and check that authority—unique in our constitutional structure—the Framers made the President the most democratic and politically accountable official in Government.”)).
• 2See Seila Law, 140 S. Ct. at 2203 (quoting Federalist No. 70) (“The Framers viewed the legislative power as a special threat to individual liberty, so they divided that power to ensure that ‘differences of opinion’ and the ‘jarring of parties’ would ‘promote deliberation and circumspection’ and ‘check excesses in the majority.’”)).