The Early Years of Congress’s Anti-Removal Power

Abstract

Judges and scholars have long debated whether the Constitution provides the president with a power to remove executive officials. The Constitution, however, undoubtedly gives Congress tools to discourage the president’s use of such power. Perhaps most notably, the Appointments Clause makes it more difficult for the president to remove principal officers—even those whose views are out of the step with the president’s—because the president cannot know whether the Senate will consent to a preferred replacement. This is an example of what is dubbed Congress’s anti-removal power: Even if the president can remove, a motivated Congress can discourage the president’s use of that power. In ‘Congress’s Anti-Removal Power’, we used game theory to show why anti-removal tools are effective—viz., they increase the costs of presidential removal, resulting in less of it—and argued that such tools have been a longstanding feature of interbranch relations. This article focuses on the founding era to argue that Congress’s anti-removal power not only comports with the Constitution’s language, but is also a deliberate feature of the constitutional bargain. Not only did James Madison and Alexander Hamilton bless anti-removal tools, but early Congresses enacted statutes that discouraged removal. While the question of presidential removal attracted debate in the first Congress, the same does not appear to be true for these anti-removal features. The article thus concludes—in the spirit of dogs that do not bark—that Congress’s use of its anti-removal power finds support in both the Constitution’s text and founding-era thought and practice.

I. INTRODUCTION

One of the oldest constitutional debates in the United States concerns whether the president has constitutional authority to unilaterally remove executive officials. This debate divided the first Congress,¹ has been at the heart of foundational Supreme Court cases,² and still today generates much scholarly scrutiny. Now that the Supreme Court appears to be in the process of limiting Humphrey’s Executor (if not jettisoning it outright),³ it is also safe to say that this longstanding debate will soon become even more significant.

Alas, we do not have anything new to add to the debate over whether Article II of the Constitution gives the president a removal power. We do, however, have something to say about a related issue that has attracted much less attention. Whether or not the president has a removal power, the Constitution certainly gives Congress what we have dubbed an ‘anti-removal power’, ie the power to discourage presidents from exercising whatever removal authority they possess. In a recent article entitled ‘Congress’s Anti-Removal Power’, we used game theory to demonstrate how structural features of the Constitution—including but limited to the Appointments Clause and the power of each House of Congress to set its own rules—allow Congress to increase the president’s costs of removal (or decrease the benefits of removal, which amounts to the same thing), thereby often creating at least a measure of independence for agency officials.⁴ We also identified real-world examples of what appears to be successful uses of Congress’s anti-removal power. For example, it appears that no president has ever directly removed a comptroller of the currency, even though the only formal protection Congress provided the comptroller is a requirement that the president provide reasons for removal—a protection Congress provided precisely to discourage removal.

In ‘Congress’s Anti-Removal Power’, we also explained that once the underlying dynamic is understood, it is apparent that Congress has many options to strengthen the effectiveness of its anti-removal power. For example, Congress can increase the number of votes required for cloture (which would make it more difficult for the president who has removed an official to replace that official) or can constrain the power of acting officials (which would limit the ability of the president to evade the Senate’s confirmation process).⁵

In this article, we deepen our investigation into Congress’s anti-removal power by examining its operation at the founding and in the early years of Congress. In ‘Congress’s Anti-Removal Power’, we identified statements from Alexander Hamilton and James Madison that suggest that not only does the Constitution allow Congress to discourage removal, but also that Congress’s ability to do so is a deliberate feature of the Constitution. Here, we offer further support for that position by discussing several early statutes in which Congress made it more difficult for the president to use his removal authority, such as Congress’s decision in 1790 to create the Sinking Fund Commission—a five-person body that included the vice president and the chief justice as voting members. Although the president could in theory control the Commission’s direction because of his ability to remove the other three commissioners (all department heads), the Commission’s structure inherently tended to increase the president’s costs of removal. The Sinking Fund Commission is the starkest example of an early statute with features that discouraged removal, but it is not the only one.

We focus on these early statutes because it appears that unlike formal restrictions on the president’s removal power, statutes that merely discouraged removal by increasing the president’s political costs did not prompt serious constitutional debate. If Congress’s use of its anti-removal power conflicted with the founding generation’s understanding of the respective powers of the presidency and Congress, it seems likely that someone would have objected. Thus, in the spirit of the dog that does not bark, especially combined with affirmative statements from key framers, we conclude that founding-era practice also may support Congress’s use of its anti-removal power. This contention matters in terms of distributed power over removal of executive-appointed officials because it suggests that even without Humphrey’s Executor, Congress has other tools to create some measure of independence—a proposition of obvious contemporary significance but also one that enriches the growing historical literature on the origins and evolution of this longstanding separation-of-powers debate.

II.HOW CONGRESS’S ANTI-REMOVAL POWER WORKS

For more than 200 years, judges and scholars have been arguing about the president’s power over the executive branch. In ‘Congress’s Anti-Removal Power’, we contended that even if the president has an unfettered power to remove—that is to say, fire—executive branch officials, the Constitution also provides Congress with political tools to discourage the president’s use of that power.⁶ Here, we do not rehash the full framework. Instead, as background to this article’s discussion of founding-era materials, we briefly set forth the basics.

First, imagine a world in which the president could not only remove an official (say, the head of an agency), but could also costlessly replace that official with someone who shares and can implement the president’s preferences in every particular. In such a world, the agency should do precisely what the president wants. Otherwise, the president would remove the agency head and place someone else in her stead. Accordingly, even if the incumbent official would rather do something else, the president’s preferences would win out.

Of course, that is not the world we live in. In the real world, it can be quite costly for the president to remove someone. Removal may be controversial to voters, thus threatening the president’s reelection prospects or at least limiting the president’s ability to pursue other political objectives. Removal may also offend Congress—for example, as is often the case, if the removed official has a powerful sponsor in the Senate. Removal thus may require the president to provide an offsetting favor to the offended senator, and favors are not free. Removal also takes time. Even if it is certain that the Senate will confirm a replacement official, the process is not instantaneous; indeed, the Senate has many demands on its time. This means that the White House must ‘discount’ the value of having someone better (from the president’s perspective) in the office because the agency’s ability to function may be limited pending completion of the confirmation process. And where it is unclear whether the Senate will confirm a replacement more to the president’s liking, the president must also discount the value of that preferred replacement. Even if the ‘better’ person is confirmed, moreover, it may turn out that the president is disappointed with the person’s performance; when deciding whether to fire, the president must account for that risk that the devil he knows is better than who may come next.

A key insight of ‘Congress’s Anti-Removal Power’ is that these removal costs create a measure of decisional independence for incumbent officials. Because removal is costly, the president has an ‘indifference range’ in which the benefits of removal from the president’s perspective do not outweigh the costs of removal. In ‘Congress’s Anti-Removal Power’, we demonstrated this indifference range graphically,⁷ but the insight is intuitive. If the incumbent official and the president are on the same page except for a few small matters, the president may conclude that the person is ‘close enough’ to what the president wants that it would not be a good use of resources to remove that official. By contrast, if they are far apart on important issues, and the official refuses to yield—or who, unlike another potential officeholder, lacks the physical or mental ability to carry the president’s preferences out—the president will decide ‘that’s too much’ and remove that official. What falls within the president’s ‘close enough’ bucket versus the president’s ‘that’s too much’ bucket depends on how important the issue is to the president and how costly it would be to remove the incumbent official. Presidents, for example, presumably want the secretary of state to share the White House’s priorities in almost all respects and would conclude that nearly any realistic cost is worth paying to replace a secretary of state who does not do what the president wants. By contrast, the White House may be more willing to accept divergence for less important offices (from the president’s perspective), such as the head of a lower-profile agency, or for agency leaders who traditionally have been viewed as exercising a measure of decisional independence (eg regulators of elections, financial markets, or interstate and international communications).

And now the central point: The costs to the president of removal are not static. The Constitution provides Congress with potent tools to increase the president’s removal costs—thereby creating a wider indifference range, and so greater space for independence from the White House. Consider the Appointments Clause.⁸ At least for principal officers, the president cannot simply remove an incumbent and replace the person with someone the president likes more—at least on not on a nontemporary basis. Instead, the Senate must confirm the president’s preferred replacement. The White House therefore must discount the benefits of a replacement both by the time it will take to get a new person confirmed and by the possibility that the Senate will reject the president’s nominee. By itself, this structural feature expands the president’s indifference range vis-à-vis an incumbent official. Even for inferior officers, moreover, the Appointments Clause allows Congress to vest appointment in the Senate; indeed, that is the constitutional default.⁹

The Constitution also provides that ‘[e]ach House may determine the rules of its proceedings’.¹⁰ One upshot of this cameral rules power is that the Senate can impose heightened cloture requirements for certain offices, which, in turn, cause the president to further discount the benefits of a remove-and-replace strategy. After all, the more votes necessary for confirmation, the greater the risk that a nomination will fail—thus requiring a steeper discount on the benefits of removal from the president’s perspective. Furthermore, because each House controls its own proceedings, if either House wants to, say, hold hearings following the removal of an official, it is free to do so. Perhaps Congress would only require hearings for certain officials for which it cares a lot about independence, and perhaps only if the president fails to articulate public good reasons for the firing. The critical point is that hearings should also increase the political costs of removal for the president, thus again widening the indifference range.

Congress, moreover, also has less direct ways to prevent removal. The Constitution, for example, provides Congress with powerful ‘anti-evasion’ tools to prevent the president from sidestepping the Senate’s confirmation process (which evasion, if allowed, should narrow the president’s indifference range). Indeed, if it wishes, Congress can generally prevent the president from making recess appointments, using acting or temporary agency leaders, or subdelegating policymaking to lower-level agency officials. Should Congress use them, these anti-evasion tools also expand a president’s indifference range between deciding whether to remove an official or keeping the person in place.¹¹ Although not the focus of this article, it must be underscored that these anti-evasion tools are critical to Congress’s broader anti-removal power, especially in modern times as presidents turn more and more to ‘actings’ to implement their regulatory policy agendas.¹² In recent years, scholars have suggested various reforms to the Federal Vacancies Reform Act and related statutes to weaken the president’s ability to bypass Congress when staffing agency leadership.¹³

In ‘Congress’s Anti-Removal Power’, we collected evidence suggesting that Congress’s ability to discourage removal can be potent. Presidents are often reluctant to remove certain officials, even when there is no significant formal barrier to removal. For example, although all the White House must do to remove the comptroller of the currency is provide a reason for doing so—the statute is clear that any reason will suffice, including, presumably, ‘I felt like it’—in practice presidents do not remove comptrollers.¹⁴ As Senator Samuel Clarke Pomeroy explained when Congress imposed this reason-giving requirement, the requirement’s force does not come because it is hard for a president to think of some possible reason, but because the president knows that the Senate believes some independence is important for the position and that ‘if the Senate [does] not approve of the reasons given by the President, they could refuse to confirm the successor’.¹⁵ Thus, as put by Senator Charles Buckalew (who proposed this requirement), a rational president will ‘not exercise [the removal] power unless he has good reasons for it’.¹⁶

Of course, Congress’s anti-removal power does not always prevent removal; sometimes the president (perhaps with good reason) will decide that the cost of removal is worth paying. And just because this anti-removal power exists does not mean every use of it is politically appropriate; the president’s power to remove executive branch officials is valuable for many reasons, including increasing public accountability for agency decisions. It is important to recognize, however, that even if statutory restrictions on removal are off the table, the Constitution gives Congress tools to discourage removal.

III.CONGRESS’S ANTI-REMOVAL POWER AT THE FOUNDING

As a matter of text and structure, the Constitution creates tools that Congress can use to discourage presidential removal. It does not follow, however, that Congress’s use of those tools for that purpose is a deliberate feature of the Constitution rather than an accidental one. After all, the power to coin money has allowed Congress to create a market for rare coins, but allowing collectors to trade pennies is surely not why the Coinage Clause exists. In other words, merely because Congress can use certain tools in the Constitution to discourage removal does not mean that the Constitution gives Congress those tools in order to discourage removal. This distinction matters because how one evaluates a particular use of a tool may depend on whether that use is an intended one. Here, we do not review the history of every clause that Congress can use to discourage removal. Instead, more modestly, we examine historical sources to see whether there is any support for the concept of an anti-removal power. We think there is.

As we explained in ‘Congress’s Anti-Removal Power’, both Hamilton and Madison endorsed congressional efforts to discourage removal. Hamilton explained in the Federalist that the Appointments Clause allows the Senate to prevent appointment of someone with ‘no other merit than that of coming from the same State to which he particularly belonged, or of being in some way or other personally allied to him, or of possessing the necessary insignificance and pliancy to render them the obsequious instruments of his pleasure’.¹⁷ Thus, the president should have ‘a strong motive to care in proposing’ nominees, else he be accused by the Senate of ‘a spirit of favoritism, or an unbecoming pursuit of popularity’.¹⁸ But Hamilton also explained that the Senate’s power to reject nominees would not only prevent the president from picking unworthy ones. Instead, ‘[w]here a man in any station had given satisfactory evidence of his fitness for it, a new President would be restrained from attempting a change in favor of a person more agreeable to him by the apprehension that a discountenance of the Senate might frustrate the attempt’.¹⁹ Although Hamilton did not use the modern jargon of game theory, his analysis seems to capture concepts like discounting benefits and indifference ranges.

Madison was the foremost defender of presidential removal during the Decision of 1789, but he also offered reasoning consistent with our analysis. In particular, he explained that there was no need to fear the president’s removal power because if the president should attempt to ‘displace from office a man whose merits require that he should be continued in it’, the president would ‘be impeachable by this House, before the Senate, for such an act of mal-administration; for I contend that the wanton removal of meritorious officers would subject him to impeachment and removal from his own high trust’.²⁰ Even short of impeachment, the removed person could form ‘combinations’ with others—no doubt including those in Congress—to harm the president politically.²¹ These risks ‘are considerations which will excite serious reflections beforehand in the mind of any man who may fill the Presidential chair’.²² Madison’s view that a tool as serious as impeachment could be used to discourage removal certainly suggests founding-era support for an anti-removal power.

Early practice offers additional support for Congress’s anti-removal power. Consider, for example, the Sinking Fund Commission. Proposed by Hamilton and enacted by the first Congress in 1790 to purchase securities, the Commission was comprised of ‘the President of the Senate [ie the vice president], the Chief Justice, the Secretary of State, the Secretary of the Treasury, and the Attorney General’.²³ In 1790, the vice president was John Adams, the chief justice was John Jay, the secretary of state was Thomas Jefferson, the Secretary of the Treasury was Hamilton, and the attorney general was Edmund Randolph—who all knew a thing or two about the Constitution. Both Madison and John Marshall also later served on the Commission. The Commission is relevant here because it could only make a purchase with the assent of three commissioners and the president.²⁴ Yet the president could not remove the vice president—at the time, a political opponent—or the Chief Justice. Scholars such as Christine Kexel Chabot argue that the Commission’s unusual structure shows that the Constitution does not give the president a removal power.²⁵ Others disagree.²⁶ In 2021, the Supreme Court addressed the issue in Collins v Yellen. Writing for the Court, Justice Alito reasoned that the Commission is not inconsistent with an inherent presidential removal power because the Commission did not ‘operate[] beyond the President’s control’, given that ‘three of those Commissioners were part of the President’s Cabinet and therefore removable at will’.²⁷

However that dispute is resolved, we observe that, at a minimum, the Commission’s structure surely made it at least somewhat difficult for the president to control it in the real world. Imagine that, say, the attorney general agreed with the vice president and chief justice that the Commission should not make a purchase that the president favored. What could the president do? Obviously, he could remove the attorney general and urge the Senate to confirm a new one, thus again ensuring that at least three commissioners shared the president’s view. By itself, however, requiring the president to toss out the attorney general over a securities issue—not even a question of law enforcement—would have been costly from the president’s perspective, especially if the president believed that the attorney general was doing a fine job in his main position. But the costs would not end there. The president would also know that if the president were to remove the attorney general, the attorney general could (and presumably would) create a political mess. For one thing, the attorney general could accurately tell the public that he was following the lead of the chief justice, thus cloaking his decision in the judiciary’s prestige. For another, the vice president—who, by definition, would be backed by formidable national constituency—would be there to publicize the president’s rejection of the views of his own attorney general and the chief justice. Of course, if a dispute between the president and the attorney general were significant enough, the president presumably would remove him. But, for relatively small matters, it is difficult to imagine that any rational president would pay the cost. It just wouldn’t be worth it.

To be sure, we are unaware of such a scenario playing out—though, notably, the Commission did find itself deadlocked when chief justice Jay was unable to attend one of its meetings.²⁸ But the risk of such a scenario would have been apparent to the First Congress after even a moment’s consideration. Indeed, this potential is plain on the statute’s face. Yet instead of prompting anything like the weighty constitutional debate that occurred just one year earlier in the Decision of 1789, the historical record does not report anyone raising a constitutional objection.²⁹ This is not because the Commission was unimportant; to the contrary, Jefferson had a ‘vehement dislike for the sinking fund’, yet he ‘took his seat on the Commission without any recorded objection to its structure’.³⁰ Although this example is not dispositive, it suggests that the founders were not opposed to efforts by Congress to increase the president’s removal costs, especially where the president retained formal removal authority.

Congress did something similar with the Mint Commission. The Mint Act of 1792 required inspection by numerous officials in the executive branch (the secretary of state, attorney general, treasury secretary, and comptroller of the treasury), but also, again, the chief justice.³¹ Here, presidential control was obviously stronger than with the Sinking Fund Commission because the president had more of ‘his’ people involved, and the vice president was not included. But once more, the chief justice was involved, which presumably would have increased the president’s political costs of removal in cases where the removed individual could credibly claim to be on the Chief Justice’s side. In modern administrative law, scholars speculate that lawyers engaged in intra-agency debates may disguise policy arguments as legal conclusions in order to dissuade opposition.³² The effect would be much greater where the official whose views differed from the president’s could invoke the chief justice as support.

The Mint and Sinking Fund Commissions were unusual in that they included the chief justice (and vice president, for the Sinking Fund Commission)—and not just traditional executive branch officials. But Congress in the early years also created more traditional multi-member agency structures, such as a patent authority headed by the secretary of state, the attorney general, and the secretary of war.³³ As Chabot explains, ‘[m]ultimember boards or commissions were commonplace in other domestic statutes passed by the First Congress’.³⁴

It is important to underscore that structuring an agency as a multi-member commission can be part of Congress’s anti-removal power. Justice Kagan made a similar observation in her dissent in Seila Law LLC v CFPB: ‘A multimember structure reduces accountability to the President because it’s harder for him to oversee, to influence—or to remove, if necessary—a group of five or more commissioners than a single director. Indeed, that is why Congress so often resorts to hydra-headed agencies’.³⁵ Congress’s decision to create multi-member commissions in the nation’s early years at least suggests that the founding generation recognized it had an anti-removal toolkit.

A multi-headed agency leadership structure is far from the only example from the founding era of Congress’s use of its anti-removal power. In a recent article, Chabot has identified other early statutes that made presidential control of executive action more difficult.³⁶ Beyond mimicking features of the Sinking Fund Commission, the first Congress also enacted succession provisions that would allow the secretaries of foreign affairs, treasury, and war (the original executive departments) to appoint inferior officers who would take ‘charge and custody of all records, books, and papers appertaining’ if the president removed the respective secretary.³⁷ Although this would not prevent removal, it presumably sometimes could make removal somewhat more difficult. As Chabot observes, ‘Having the Secretary’s hand-picked assistant retain Department records would seem to limit a Machiavellian scheme to oust a law-abiding Secretary . . . as removal would automatically transfer Department records to a second officer who was understood to be loyal to the Secretary’.³⁸ These succession provisions thus may also be examples of Congress exercising its anti-removal power by making it marginally more costly for the president to evade the Senate confirmation process.

Countersigning requirements are another example of Congress making presidential control more difficult. For tariffs and related collections efforts (a key source of federal funds) at larger ports, for example, Congress in 1789 declared, among other things, that naval officers were required ‘to “countersign all permits and clearances granted by the collector”’, thereby making it ‘impossible for a single collector to carry out key revenue functions unless the naval officer agreed’.³⁹ Such ‘shared responsibility’, Chabot explains, ‘not only limited recording errors, but it also curbed temptation for officers to accept bribes in exchange for dishonest recording’.⁴⁰ Similarly, the Treasury Act of 1789 required four officers to approve any fund disbursement.⁴¹ According to Chabot, Congress’s decision to embed such ‘nonunitary structures’ into the executive branch suggests that the founding generation rejected the unitary executive theory.⁴² We are not sure that conclusion follows—at least if it means that the president does not have removal power under Article II.

For our purposes, however, again, whatever one concludes about the president’s removal power, such countersigning requirements surely at least discouraged removal. In the collections context, for example, if the president were unhappy, he would have to consider the costs of removing two government officials. Moreover, the Collections Act allowed the collector to appoint a deputy who could act in place of a disabled, sick, or otherwise absent collector,⁴³ which may also have applied if the president removed the collector. Thus, the president would either have been required to remove and replace two people, or at least remove one of them and threaten the other with removal—which would be a credible threat, but also one that adds to the president’s transaction costs. And even then, the collector may have already appointed a deputy collector who would temporarily fulfill the collector’s duty until a successor was appointed. Either way, removal would be more difficult, thus expanding the president’s indifference range. For similar reasons, in the Treasury context, the fact that multiple individuals were involved in disbursement of funds made it more difficult—as a practical matter—for the president to act.

We do not claim that these examples definitively prove that the founding generation intended that Congress should be able to discourage presidential removal. It is noteworthy, however, that whereas formal restrictions on presidential removal prompted fervent constitutional debate, it does not appear that the same can be said for these other statutes. That fact, combined with Hamilton’s and Madison’s statements, suggests not only that the constitutional text and structure support Congress’s anti-removal power, but also that it is a deliberate feature of the Constitution.

IV. CONCLUSION

Basic economics teaches that because Congress can make removal more costly for the president, we should expect less removal. There is also good reason to think Congress’s power to discourage removal is a deliberate feature of the Constitution. Not only was an anti-removal power endorsed by Hamilton and Madison, but early congressional practice suggests that discouraging removal—in contrast with formally restricting removal—did not trigger constitutional alarms. This history is worth remembering, especially as scholars debate the nature and scope of the president’s removal power.

For helpful comments, thanks are due to Christine Kexel Chabot, Shay Elbaum, Jenn Mascott, Jed Shugerman, and participants at the Stanford University Histories of Presidential Power Conference.

Footnotes