TABLE OF CONTENTS

This Essay concerns a constitutional puzzle, the puzzle of for-cause removal. For a century the Supreme Court has been attempting to answer a simple question: when is it constitutional for Congress to provide that an agency head or lower official can be removed only for cause?

The answer seems straightforward: almost always. The president’s responsibility under Article II is to see that the laws are faithfully executed. The standard forms of for-cause removal provide that an officer may be removed only for “inefficiency, neglect of duty, or malfeasance.” Conduct of this sort, of course, interferes with faithful execution of the law. Removal of an officer on these grounds, thus, is consistent with the president’s constitutional responsibility. Removal on other grounds, for example, that the officer has angered the president because his or her testimony before Congress embarrassed the president or the administration, would seem unconnected to the president’s executive authority under the Constitution. To be sure, the president might have that authority of unfettered removal by law. Congress can surely so provide under its authority to establish offices and specify the qualifications for appointment to them. And, since the so-called, much-discussed and hotly contested “Decision of 1789,” the president has been understood to have that authority as a default understanding whenever the president is the appointing authority and no necessary grounds for removal have been specified by statute.

But, of course, that is not the way the majority in Seila Law LLC v. Consumer Financial Protection Bureau put the question. For them the question was whether the Constitution gives the president the power to remove principal officers (those required by the Constitution to be appointed by the president and confirmed by the Senate) as a necessary feature of having been “vested” with “the executive power” by Article II. Relying on what Chief Justice John Roberts termed the “landmark decision” in Myers v. United States, the chief answered “yes.” Indeed, he interpreted Myers as establishing a general rule of unencumbered presidential removal authority for all executive officers. Cases permitting for-cause removal since Myers were distinguished as involving multimember commissions rather than single-headed agencies (Humphrey’s Executor v. United States) or as involving “inferior officers” (Morrison v. Olson). The more recent case of Free Enterprise Fund v. PCAOB was interpreted as reaffirming Myers and as refusing to extend the Humphrey’s exception to the novel situation of a for-cause protection (for members of the Public Company Oversight Accounting Board (PCAOB)) that could be invoked only by another commission (the Securities Exchange Commission (SEC)) whose members were also subject only to for-cause removal. For the majority, the Seila Law decision hinged on a simple extrapolation from historic precedents.

The dissenters agreed that precedent controlled. But for them Myers was no landmark. It was that rare case where Congress tried to give itself the power of removal over an executive officer, and the case had been so understood since Humphrey’s limited it to that situation a mere nine years after Myers was decided. Moreover, nothing in the Humphrey’s decision turned on the officer in question being a member of a multimember commission, nor was Morrison’s resolution of the removal question based importantly on the notion that Alexia Morrison, the independent counsel, was an inferior officer. Instead, the dissenters understood Morrison as permitting any removal limitation that did not impair the president’s ability to carry out his constitutional responsibility to see that the laws were faithfully executed. Free Enterprise Fund did indeed invalidate a double for-cause provision, but it used and reaffirmed Morrison’s “impediments test.” Nor did it question the legitimacy of the SEC’s for-cause removal protections, which were assumed to exist although nowhere stated in its legislation. In short, for the dissent, Humphrey’s, as reformulated by Morrison, is the landmark precedent—a precedent followed for nearly a century and firmly embedded in multiple congressional statutes that make a very substantial portion of high-level, federal officialdom removable only for cause. The cases finding removal provisions unconstitutional are the exceptions.

How could the majority and dissent see the Court’s precedents so differently? One part of the answer is that the Court’s precedents on for-cause removal are a jurisprudential train wreck. There is hardly a good constitutional argument to be found in the decisions taken as a whole. A second is that the majority and dissent begin with starkly different interpretations of Article II and the position of presidents in the constitutional system, and of the role the Court should play in overseeing how Congress structures administrative agencies. The dissenters affirm a view very similar to the one with which this Essay began. The majority’s view will be elaborated below when we return to a discussion of Seila Law.

The Precedents

To the great benefit of the coherence of legal doctrine, and to the Republic itself, judicial consideration of the constitutional limits concerning legislative conditions on the removal of executive officers did not make its way to the Supreme Court until over a century after the ratification of the Constitution. The case was Myers v. United States, a suit for back pay by the estate of a postmaster of the second class who was removed from office in violation of the Tenure of Office Act. That statute, which has a colorful history, and which figured prominently in the impeachment trial of President Andrew Johnson, did not impose for-cause conditions on the removal of officers appointed by the president and confirmed by the Senate. It instead required that any such officer’s removal be approved in the Senate.

In opinions that stretched over 243 pages of the original U.S. Reports, ex-President Taft, then Chief Justice of the United States, for the majority, and Justices Brandeis and Holmes in dissent, exhaustively examined the debates in the first Congress, the ratification debates in the states and subsequent congressional discussions and practices. The upshot was a majority opinion that found that the Tenure of Office Act unconstitutionally interfered with the president’s power of removal by giving the Senate a role in the removal process that was not authorized by the text of the Constitution or institutionalized by subsequent practice. Of course, the president’s power of removal is nowhere specified in the Constitution either, and the Tenure of Office Act was a congressional practice that had been in effect since 1867. As is usual with such judicial excursions into past practices, the past is complicated, contradictory and difficult to recover. Its evidentiary significance is contested, and multiple judicial historians arrive at multiple meanings.

The Myers situation, involving as it did the Tenure of Office Act, is clearly distinguishable from the cases of most interest to the present inquiry, those raising the question of the constitutional status of legislative requirements that officers be removed only for cause. Dicta in the Myers majority opinion, however, questioned Congress’s power to qualify the president’s removal authority at all. But that dicta was disavowed ten years later in Humphrey’s Executor v. United States. Humphrey’s was another back pay case, this time by a removed member of the Federal Trade Commission (FTC), or rather his executor. William Humphrey’s estate complained that when removing Humphrey, President Roosevelt failed to claim that Humphrey’s conduct satisfied any of the grounds for removal that were specified in the FTC statute. Eschewing another foray into historical debates and institutional practices, the Court in Humphrey’s took an entirely different tack.

Surprisingly, according to the Humphrey’s majority opinion, the Federal Trade Commission did not exercise executive power and therefore did not run afoul of Myers’ broad dicta. I say “surprisingly” because if the FTC, established to carry out the functions specified in the Federal Trade Commission Act, is not an executive agency, what is it? On the majority’s view it was “quasi-legislative” and “quasi-judicial.” What does that mean? Apparently, because the FTC decides cases, it is quasi-judicial. Because it has the authority to adopt rules of some sort, it is quasi-legislative. And in the majority’s view the agency’s investigative powers are similar to the investigatory powers of a legislative committee. Hence those powers too are quasi-legislative. And on this view, the FTC’s officers are not executive officers like the postmaster whose estate brought suit in the Myers case.

This is, of course, constitutional nonsense. The Constitution recognizes three principal departments of government: the legislative, the executive and judicial. FTC commissioners are not judges and they are not legislators. Their task is to implement a federal statute. This is an executive function and those officers are executive officers.         

The Court struggled to distinguish the postmaster from the FTC commissioners by characterizing the former as engaged in “purely executive” functions. But, of course, what a purely executive function might be is as mysterious as what a quasi-judicial or quasi-legislative function might be. The Post Office, a freestanding executive agency, makes rules about what can be carried in the mails. Postmasters make judgments about whether things that are presented for carriage satisfy those rules. Why are these functions not quasi-legislative and quasi-judicial? The executive agency that makes no rules and decides no cases, at least informally, is difficult to imagine, much less identify.

There is, of course, a perfectly straightforward constitutional argument for distinguishing the Humphrey’s case from the Myers case. The Court might simply have said that Congress has the authority under the Constitution to create departments, determine their structures and functions and specify the qualifications for departmental officers, as well as the grounds upon which they may be removed. But, Congress may not under the Constitution appoint executive officers itself and, by implication, cannot insert itself into the removal process save by the constitutionally authorized process of impeachment and conviction. Beyond those principles the Constitution is silent.

The notion that the real problem with the Tenure of Office Act was the insertion of Congress into the removal process, other than by impeachment, was confirmed fifty years later in Bowsher v. SynarBut once again, the Court’s opinions invalidating the statute were analytically dubious and, in practical terms, a triumph of abstract reasoning over common sense. Without going into the details of this complex case, suffice it to say that a plurality of the court thought that the actions of the comptroller general at issue under the Gramm-Rudman-Hollings Act—in particular, the comptroller’s issuance of a report detailing the reductions in government expenditures that the president had to make to satisfy the “anti-deficiency” provisions of the Act—were unlawful because the statute providing for the comptroller general’s office permitted his removal for cause by a joint resolution of Congress. A plurality of the Court believed that this made the comptroller general an executive officer, and an executive officer’s tenure could not depend upon congressional action or upon a joint resolution subject to presidential veto. Other members of the Court reached the same result a different way. They deemed the comptroller general’s actions to be legislation without going through the constitutionally prescribed process of bicameralism and presentment.

But as Justice Byron White pointed out in his dissent, these analyses, in this context, conjured up visions of supposed medieval debates concerning the number of angels that could dance on the head of a pin. First of all, Justice White was doubtful that what the comptroller general did amounted either to execution of the law or to legislation. The complained of action was merely the issuance of a report. Nothing happened unless the president acted on it. Furthermore, even if the comptroller general’s report were regarded as execution of the law, the plaintiff’s theory of the case hinged on the claim that the removal provisions in the comptroller general statute made that officer subservient to Congress. But there was no evidence supporting this proposition. No comptroller general had ever been removed using these statutory provisions, or even threatened with removal. Moreover, in Justice White’s view, the statutory provisions added nothing to Congress’s existing constitutional authority to remove the comptroller general by impeachment.

And the plot thickens. Justice Blackmun, in dissent, made the obviously salient point that the Court was being asked to invalidate the automatic budget reduction provisions of one statute because of the removal provisions in another. But even conceding that the removal provisions in the General Accounting Office statute could not pass constitutional muster because of the executive functions also assigned to the comptroller general, Justice Blackmun could not see how that justified invalidating provisions in the budget reduction legislation. If the constitutional problem lay in the General Accounting Office statute, the removal powers provided in that statute could be declared to be unconstitutional should they ever be exercised.

To review the bidding, after three major opinions on the constitutional requirements for the exercise of removal authority with respect to federal officials, the Supreme Court had generated the following legal understandings: Congress could not insert itself into the removal process for executive officers—whoever those are—save through impeachment. And, Congress could condition the president’s removal power over federal officers (in yet unspecified ways) so long as those officer’s functions were not “purely executive”—whatever that means. Or maybe not. The approval of for-cause removal in Humphrey’s seems premised on the self-evidently silly proposition that the Federal Trade Commission is not an executive agency. Surely that is a wobbly precedent. In addition, there was no majority in Bowsher for the proposition that the comptroller general’s function under the Deficit Reduction Act was an executive function. And since the plurality opinion in that case agreed with the concurring justices that if it were a legislative function it would violate the bicameralism and presentment requirements for legislation, perhaps that’s what the case stands for. After all, the holding was that the Deficit Reduction Act was unconstitutional, not that the removal provisions in the statute creating the office of the comptroller general were. Up to this point this jurisprudence is in shambles. And it gets worse.

The next contribution to the for-cause muddle seemingly drove a stake into the heart of the Myers opinion. Morrison v. Olson involved a challenge to the independent counsel provisions of the Ethics in Government Act of 1978, enacted in the wake of the Watergate scandal. That Act aimed to mitigate what Congress perceived to be the conflicts of interest inherent in entrusting the executive branch to conduct criminal investigations of the president, vice president, cabinet and certain high-level subcabinet officials, or high-level officials in the president’s election campaign.

The Act gave the attorney general 90 days to conduct a preliminary investigation upon receipt of “information sufficient to constitute grounds to investigate” any of these covered officials for a federal felony or serious misdemeanor. At the end of 90 days, the attorney general would report whether reasonable grounds to continue the investigation in fact existed to a special division of the U.S. Court of Appeals for the D.C. Circuit. If the attorney general determined that such grounds existed or failed to make a determination either way, the attorney general would then be obligated to apply to the special division for the appointment of an independent counsel. Upon receiving the application, the special division would appoint an independent counsel and issue an order defining the counsel’s prosecutorial jurisdiction.

With respect to all matters within the independent counsel’s jurisdiction, the Act granted her “full power and independent authority to exercise all investigative and prosecutorial functions and powers of the Department of Justice, the Attorney General, and any other officer or employee of the Department of Justice,” other than those relating to wiretap applications. The independent counsel was required, in general, to “comply with the written or other established policies of the Department of Justice respecting enforcement of the criminal laws.” 28 U.S.C. § 594(f). In all other respects, however, the attorney general was precluded from exercising any policy control over the independent counsel. This limitation was reflected in the very narrow grounds available to the attorney general for removing an independent counsel. The Act provided: “An independent counsel appointed under this chapter may be removed from office, other than by impeachment and conviction, only by the personal action of the Attorney General and only for good cause, physical or mental disability . . . or any other condition that substantially impairs the performance of such independent counsel’s duties.” 28 U.S.C. § 596(a)(1).

The respondent, Theodore Olson, attacked the statute on several grounds. We will put aside for the moment the claim that the independent counsel was improperly appointed because she was a superior officer who could be appointed, under the Constitution, only by the president with the advice and consent of the Senate. More pertinent to the present inquiry, Olson also claimed that the removal provisions of the statute improperly limited the president’s control over a purely executive officer. This claim was a straightforward interpretation of the state of the law following Myers and Humphrey’s.

“Wrong,” declared the Supreme Court majority. In pertinent part the majority opinion provided the following analysis of its prior jurisprudence:

Appellees contend that Humphrey’s Executor . . . [is] distinguishable from this case because [it] did not involve officials who performed a “core executive function.” They argue that our decision in Humphrey’s Executor rests on a distinction between “purely executive” officials and officials who exercise “quasi-legislative” and “quasi-judicial” powers. In their view, when a “purely executive” official is involved, the governing precedent is Myers, not Humphrey’s Executor. . . .

We undoubtedly did rely on the terms “quasi-legislative” and “quasi-judicial” to distinguish the officials involved in Humphrey’s Executor . . . from those in Myers, but our present considered view is that the determination of whether the Constitution allows Congress to impose a “good cause”-type restriction on the president’s power to remove an official cannot be made to turn on whether or not that official is classified as “purely executive.”

. . .

Considering for the moment the “good cause” removal provision in isolation from the other parts of the Act at issue in this case, we cannot say that the imposition of a “good cause” standard for removal by itself unduly trammels on executive authority. . . .

Nor do we think that the “good cause” removal provision at issue here impermissibly burdens the president’s power to control or supervise the independent counsel, as an executive official, in the execution of his or her duties under the Act. . . . [B]ecause 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.

The majority opinion in Morrison comes very close to establishing the principle argued for at the beginning of this Essay. While Justice Antonin Scalia’s somewhat hyperbolic dissent accuses the majority of treating a restriction, the for-cause provision, as an empowerment, his critique arguably is wide of the mark. At least implicitly the majority opinion views the president’s constitutional authority to oversee executive officers as limited to the authority to ensure “faithful execution” of the laws. Because it also views the “causes” embedded in the “good cause” removal requirement as encompassing any conduct that would interfere with faithful execution, there is no improper interference with the president’s constitutional responsibilities. On this view, the implicit view of the dissenters in Seila Law, for-cause removal restricts no power that the president legitimately holds.

So far, so good. Yet, the opinion enunciates a vague standard for determining whether statutory provisions limiting the president’s removal authority are consistent with the Constitution. Whether such a provision “unduly trammels” the president’s executive authority surely lies in the eyes of the beholder. Moreover, the majority fails to ground its conclusions in a clear statement that where administrative power is derived from legislation, rather than the president’s independent powers under the Constitution, the president’s directive authority is limited to ensuring faithful execution. And, finally, because of the Court’s suggestion that there may be some, as yet unspecified, group of executive officers for whom even for-cause removal would be improper, the door was left open for further developments that might muddy the waters, particularly in the hands of justices more sympathetic to Justice Scalia’s commitment to the Article II vesting clause as a fount of unitary authority to direct administration.  

Free Enterprise Fund v. PCAOB stirred up the silt and perhaps cemented Myers’s claim to be a true zombie precedent. There was no telling when it might spring back to life, as it did in Seila Law. As previously noted, Free Enterprise Fund invalidated the for-cause removal provision protecting the members of the Public Company Accounting Oversight Board because the removal power over those officers was lodged in the Securities and Exchange Commission whose members were themselves, by stipulation of the parties, removable only for cause. In explaining its decision, the majority had this to say: 

Since 1789, the Constitution has been understood to empower the President to keep these officers accountable—by removing them from office, if necessary. See generally Myers v. United States . . . . This Court has determined, however, that this authority is not without limit. In Humphrey’s Executor v. United States, . . . we held that Congress can, under certain circumstances, create independent agencies run by principal officers appointed by the President, whom the President may not remove at will but only for good cause. Likewise in . . . Morrison v. Olson, . . . the Court sustained similar restrictions on the power of principal executive officers—themselves responsible to the President—to remove their own inferiors. The parties do not ask us to reexamine any of these precedents, and we do not do so.

This was, indeed, a striking statement of the state of the law. Note that Myers is cited for the general principle that presidents are empowered under the Constitution to remove officers to keep them accountable. There is no mention of the fact that Myers did not involve a for-cause removal provision but, instead, a requirement for senatorial acquiescence in the removal. Humphrey’s is described as holding only that Congress can “under certain circumstances” create independent agencies susceptible to removal only for cause. And Morrison is described as sustaining for-cause removal for inferior officers in situations where the removing authority is a principal officer removable at will. Bowsher is ignored, perhaps because the Court does not know what to make of it, or perhaps because the majority wants to downplay the common feature of that case and Myers: the insertion of Congress into the removal process.

The majority then continues:

We are asked, however, to consider a new situation not yet encountered by the Court. The question is whether these separate layers of protection may be combined. May the President be restricted in his ability to remove a principal officer, who is in turn restricted in his ability to remove an inferior officer, even though that inferior officer determines the policy and enforces the laws of the United States? 

We hold that such multilevel protection from removal is contrary to Article II’s vesting of the executive power in the President. 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.”

This is yet another striking statement. Presumably the constitutional infirmity here is that the president cannot remove SEC commissioners who failed to remove PCAOB members who, in the president’s view, are not faithfully executing the law. But there is no explanation of why the president’s power is so restricted. SEC commissioners who fail to remove faithless PCAOB members are themselves failing to faithfully execute the law. Such commissioners should be removable notwithstanding their judicially assumed, for-cause removal protections. Failure to see this, or even to address the issue, suggests something else: the majority sees the president’s removal power as encompassing something more than the responsibility to see that the laws are faithfully executed.

To be sure, the majority never says as much. But it also never explains why this double for-cause-removal insulation matters. It certainly did not matter to the prevailing parties. Because the Court deemed the PCAOB’s tenure provisions to be severable from its organic statute, the Court’s holding portends no actual change in the PCAOB’s authorities or operations. The plaintiffs in this case, whose purpose was to evade the Board’s regulatory authority, may wonder what they have won. But more importantly, as Justice Breyer explains in his dissent, the majority decision, while perhaps a victory for unitary executive theory, has no practical effect on the president’s authority over the PCAOB other than, oddly enough, to give the SEC the power to remove a board member when the president would not have done so.

Invalidating the PCAOB’s tenure provisions now allows the SEC—even over presidential objections—to fire a PCAOB member on policy grounds, even if that member neither abuses his or her authority, nor violates or fails to enforce the law. To see why this is so, consider this thought experiment: Imagine first that the president and the SEC agree, either that a PCAOB member should or should not be removed. In this case neither agency’s for-cause removal protections matter. Now imagine that the SEC declines to discharge a PCAOB member on mere policy grounds when the president wishes it would. He would have no recourse against the SEC unless the SEC’s exercise of discretion was itself so unjustifiable as to create good cause for dismissal of one or more SEC commissioners. But here the removal of the Board’s for-cause protections are not relevant. The president’s problem is the SEC’s insulation from discretionary removal. Finally, imagine that the SEC exercises its new at will removal power in a situation where the president would have preferred to keep the PCAOB member in office. Again, unless the SEC’s exercise of discretion itself amounts to inefficiency, neglect of office, or malfeasance—which the majority takes to be the relevant standard for SEC removals—the president can do nothing. In this latter situation—where the president would have wanted to protect the PCAOB member from removal—it is actually the invalidation of the PCAOB tenure provisions that enables the SEC to take a position contrary to the preference of the White House. 

So the Free Enterprise Fund wins but gets no relief from PCAOB regulation. And the president’s removal authority is affirmed by a remedy that adds nothing to the president’s actual removal power. Justice Breyer’s invocation of angels and pins in Bowsher was not idle chatter.

Seila Law v. CFPB

Seila Law continues the tradition of taking questionable cases and deciding them on questionable grounds. First, the only two circuit courts that had considered this question in Seila Law prior to the grant of certiorari ruled against the petitioner’s position (as did a third circuit on the day Seila Law was argued before the Supreme Court). There was no circuit court split that would normally form the basis for the acceptance of a petition for certiorari. Second, these circuit court decisions were not close. The only circuit court judge to have written an opinion agreeing with the petitioner is now-Justice Brett Kavanaugh. Third, instead of defending the statute, the Justice Department and the CFPB both agreed with the petitioner. The Justice Department does sometimes decline to defend statutes that it believes are unconstitutional. But for an agency to take the position that its own statute is unconstitutional is a true rarity. In order to maintain adversity, the Supreme Court, therefore, had to appoint Paul Clement, a former solicitor general, to defend the statute against Seila Law’s separation-of-powers claim.

The peculiarities go deeper yet. The case was brought by a solo practitioner who was resisting a demand for documentary evidence concerning the violation of several statutes that the CFPB has jurisdiction to enforce. The connection between the for-cause removal provision and the petitioner’s injury, that is, the requirement that he respond to an evidentiary demand, was thus tenuous from the beginning. It disappeared when the then-director of the CFPB resigned and was replaced by an acting director. Pursuant to an opinion by the Justice Department’s Office of Legal Counsel, the acting director took the position that he was removable without cause. And, when a new director was appointed and confirmed, she took the same position. Because both the acting director and the new director both also continued on multiple occasions to seek the information that the petitioner sought to withhold, the CFPB’s evidentiary request was based on the authority of officials who took the position that they were removable without cause. The Supreme Court thus agreed to decide a case in which the claimed illegality no longer had any possible link to the petitioner’s injury. Alexander Bickel must have turned many times in his grave. Myers dicta has come back to life, and the passive virtues are dead.

To be sure, were the Supreme Court to decide that the for-cause removal provision in the CFPB statute is unconstitutional and that that provision could not be severed from the remainder of the statute, then the petitioner would be entitled to the relief that he was seeking. And when granting certiorari, the Court requested that all parties brief the severability question, even though the petitioner, for obvious reasons, had not raised that question in its petition for certiorari. But, of course, the Supreme Court viewed the removal provision as severable from the remainder of the statute—as it had in every other case in which a removal provision had been invalidated. Once again the petitioner won and lost simultaneously.            

Ruling for the petitioner on the merits was also peculiar because of the nature of the petitioner’s argument. In summary, the petitioner argued that the for-cause removal provision relating to the director of the CFPB was unconstitutional because it applies to a single director rather than to a collegial body such as the FTC, which was the relevant agency in Humphrey’s. This is the classic distinction without a difference. Indeed, the differences that the petitioner’s argument highlights seem to have the matter exactly backwards. According to petitioner’s theory, protecting a single person from removal by a for-cause requirement interferes more with the president’s authority to see that the laws are faithfully executed than applying that same protection to a body composed of multiple persons. This, by itself, makes no sense.

The process of removing one FTC commissioner for cause is exactly the same as the process for removing one CFPB director for cause. And, if the president believes that the FTC is misbehaving, bringing the Commission into line may well require multiple removals, which is surely more difficult than achieving the same result by removing a single CFPB director. Moreover, the FTC is less subject to presidential control than the CFPB. The president may not, for example, fill up the Commission with commissioners of the president’s own party. The CFPB can have at its head a presidential loyalist devoted to the president’s program, as it currently does. Nor is the FTC, by general agreement, subject to presidential oversight via Office of Management and Budget (OMB) review of its demands for information from private parties or of its major rulemaking proposals. The CFPB is.

That the CFPB head has a five-year term may be problematic because it is possible that some one-term presidents would not have an opportunity to choose a CFPB head. But that would seem to bear only indirectly on the president’s removal authority. And the staggered terms of FTC commissioners means that at least some of them may not leave office during a one-term president’s time in office. If Humphrey’s is still law, which the majority affirms, the petitioner should have lost.

The majority avoids this result by raising other accountability issues with respect to the CFPB. The majority opinion argues first that the CFPB is a less accountable body than the FTC because it has a guaranteed source of funds that does not require annual congressional appropriations. What this has to do with the president’s authority to oversee the executive branch is a mystery. Indeed, what this has to do with accountability to Congress is a mystery. Appropriations are provided by statute. The CFPB’s guaranteed source of funds is provided by statute. If Congress wants to rein in the CFPB, it can do so by amending that statute to make the CFPB subject to annual appropriations legislation, or by threatening to do so. Nor is the guaranteed source of funding for the CFPB exceptional. Most major financial regulators at the federal level, the Federal Reserve, the Federal Deposit Insurance Program, and the Comptroller of the Currency are all self-funding, and the heads of all of those agencies are protected by for-cause removal provisions. The same is, of course, true of the Social Security Administration, which is headed by a single commissioner removable only for cause. None of these arrangements have ever been thought to raise constitutional difficulties.

There are many more questionable claims about the history of the structure of federal administrative agencies and the history of for-cause removal in the majority opinion. But, in the end the argument comes down to the simple proposition that administrators protected by for-cause removal provisions destroy the link between citizens and their government and create dangerously powerful and unaccountable administrators. This position gets at first principles and is worth discussing in its own functional terms.

First, consider the question of accountability. Accountability is a complex concept that demands highly contextual analysis. It involves the answer to a series of questions that address who is accountable to whom, about what, through what processes, based on what criteria and with what consequences. Administrative agencies are established by law and are accountable to whoever appoints and removes them, whoever has authority to oversee their activities, whoever has the authority to provide them with resources and whoever has authority to determine the legal validity of their actions. The CFPB is, first, accountable to the president, who appoints its director and who, under the statute as written, could remove that director for inefficiency, neglect of duty or malfeasance. In addition, as mentioned above, the president oversees all significant rulemaking by the agency through the OMB in the Executive Office of the President (EOP).

The CFPB is also accountable to Congress, which can amend its statute at any time and investigate any and all of its activities either through congressional hearings or investigation by the General Accountability Office (GAO). And, of course, the director must be approved by the Senate when appointed. The CFPB is also responsible for complying with all general congressionally enacted requirements that structure the ways in which all administrative agencies must conduct their business.

Finally, the CFPB is accountable to law through judicial review. And that review makes the CFPB incidentally accountable to any person or entity affected by its regulatory activities. CFPB rulemaking must, subject to judicial invalidation, satisfy the rulemaking requirements of the Administrative Procedure Act (APA). Procedurally, the APA requires that agencies permit comments and argument by any interested person which, if pertinent, must be responded to by the agency. Substantively, the statute demands that any rules adopted be reasonable and based on the record developed in the rulemaking proceeding. And, to the extent that the CFPB seeks enforcement against any party, that party has rights to trial-type hearings and judicial review.

The idea that the inability of the president to remove the head of the CFPB for no reason having to do with the faithful execution of the agency’s responsibilities makes the agency dangerously unaccountable is simply preposterous. The notion that the structure of the CFPB undermines American democratic governance is similarly simpleminded.

If American democracy were structured such that elections produced presidents who had constitutional authority to create offices, unilaterally appoint all high-level government officials and direct all their activities in accordance with the president’s preferences, limiting the presidential power of removal would be an assault on democracy. There are constitutional systems that come very close to this model. As a Peruvian friend once commented to me, “We have a democracy; we elect our dictators.” That this is not the American system is so blindingly obvious that I will not here rehearse the many ways in which it misunderstands American government as constituted in our founding document and practiced on the ground.

If the Supreme Court in Seila Law v. CFPB had held merely that an agency structured like the CFPB—that is, having a single director, with a term of office longer than a single presidential term—must include the power of the president to remove the director without stating any cause for removal, the holding would have limited effect. But, by invoking Myers as the fountain of for-cause removal jurisprudence and by treating all cases that have deviated from Myers as exceptions to a general constitutional principle that the president must have unfettered removal authority over every principal executive officer in the federal government, the Court goes further. It makes all for-cause removal provisions constitutionally suspect. That suspicion is not limited to principal officers. Mr. Myers, a postmaster of the second class who brought the now “landmark” case, was not a principal officer.

Moreover, the majority’s vision of how the government is organized is arresting. According to that vision, Article II gives all executive power to the president. All other officers of the federal government are the president’s agents. Every power they exercise is the president’s power and therefore must be made accountable to the president as if the president were exercising the power himself. The statement of the president’s duty to see that the laws are faithfully executed is not a description of the purpose for presuming a constitutionally based presidential removal authority, as the dissent would have it, but is instead an amplification of the absolute executive authority that the majority finds in the vesting clause.

Unlike the dissenters’ vision of American constitutional government, which understands Congress to have broad discretion to structure administrative agencies and bureaus to meet the diverse and ever shifting requirements of public policy, Congress is nowhere to be found in the majority’s analysis. The necessary and proper clause of Article I, which charges Congress with making all laws necessary and proper for carrying into execution its enumerated powers, and all other powers lodged in the government or any officer or department, is not mentioned. That most congressional statutes delegate power, not to the president, but to particular officers of the government is ignored. The protection of a single officer from removal without cause is understood to be a threat to both liberty and democracy. But, somehow, giving absolute power to a single person to direct the actions of every principal officer in the government (perhaps with some exceptions for multimember commissions) on pain of discretionary removal is said to be necessary to the protection of those same values. If that vision prevails, we are perhaps on the way to joining my Peruvian friend in the way that we describe our government.