The Supreme Court has a new mixed drink: it’s five parts formalism to four parts functionalism, with a dash of Humphrey’s Executor. This new drink was on full display in Seila Law LLC v. CFPB, in which five Justices concluded that Congress violated the separation of powers when it placed limits on the president’s power to fire the CFPB’s director. Those five Justices used what’s called the “formalist” approach, prevailing over the competing “functionalist” approach adopted by the four dissenting Justices. Throughout history, the Court has oscillated between formalism and functionalism, especially in cases involving the president’s power to remove public officials. For the latter half of the 20th century, functionalism appeared to be the dominant approach to removal-power cases—until two recent decisions from the Roberts court. Might formalism now be seeing a resurgence?
Formalism and Functionalism
We’ve all heard the phrase “separation of powers.” In civics class, we learned that the Framers of our Constitution designed a system of government in which the powers to make law, enforce law, and decide controversies about law would be “separated” among three branches of government: the Legislature (Congress), Executive (President), and Judiciary (Supreme Court), respectively.
Suppose Congress one day enacts a statute that tests this tripartite system of government. Perhaps the law allows someone in the Executive branch to write new laws, or gives someone in Congress the power to decide questions about a law’s meaning, or gives someone in the Judiciary the ability to prosecute violators of a law. Whatever the nature of the statute, how does a court go about determining whether it violates the separation of powers?
Two methods have emerged throughout history. They’re called “formalism” and “functionalism.” To be sure, neither method is set in stone; no court has ever said that formalism or functionalism requires looking at X, comparing it to Y, and then applying that to Z. Instead, legal scholars have mined the formalist and functionalist approaches from the Supreme Court’s separation-of-powers jurisprudence.
Formalism generally says that if a law attempts to share powers among two or more of the three distinct branches, the law is unconstitutional. It places the Constitution’s formal structure of our federal government above all else. Formalism underscores the distinct separation of the legislative, executive, and judicial branches and the powers they possess. Relying on the plain text of the Constitution, formalism is immediately skeptical of any effort to blend powers together in a single branch of government or stretch the same power over two different branches of government. So, given an Act of Congress that creates some sort of power-sharing arrangement, the mere fact that the statute ensures our government works “efficient[ly], convenient[ly], [or] useful[ly]” is not, by itself, sufficient to rescue the statute from unconstitutionality (INS v. Chadha (1983)). Formalism is thus more rigid and historical than its counterpart, taking a hard view of the “separation of powers” and the independence of the three branches.
Functionalism, on the other hand, is more flexible and forward-looking. A functionalist would permit a power-sharing law if it does not obstruct the proper functioning of a branch. It is guided by the principle of “checks and balances” and looks to the functions of each branch in an evolving society. Using the Constitution’s purpose (not necessarily its text) as the starting point, it asserts that the three branches have become more and more interdependent throughout our history and will continue to do so, and that a strict tripartite theory of government isn’t absolutely necessary. Thus, a power-sharing arrangement is constitutional if it maintains a proper balance of power between the branches according to the needs of our evolving society—even if such an arrangement doesn’t perfectly align with the Framers’ original understanding of the separation of powers.
Again, the Supreme Court has never laid out the tenets of either approach in so many words. Legal scholars and jurists continue to haggle over the minutiae of both doctrines, and some have gone so far as to call the entire debate protean in nature. (See the reading list at the end of this article.) But both doctrines have been used in one form or another in dozens of separation-of-powers cases.
That is especially so when it comes to the president’s removal power. Under Article II, the president has the power to appoint certain public officials, like heads of executive agencies. Incidental to this power to appoint officials is the power to remove them. Thus, if Congress passes a statute that limits the president’s removal power, the statute is often challenged as a violation of the separation of powers.
Since the early 1900s, the Supreme Court has decided eight cases concerning Congressional limits on the president’s removal power. Early on, formalism guided the Court’s reasoning. That changed in the 1980s under the Rehnquist Court, when functionalism controlled the outcome of two removal-power cases in as many years. A period of dormancy followed, during which several academics speculated that functionalism may have beat out formalism in the long run. However, in 2010 and 2020, the Roberts Court returned to the formalist approach. What follows is a history of the formalist versus functionalist debates and an analysis of the Roberts Court’s use of formalism.
Formalism and Functionalism in Practice
Formalism’s entrance on the judicial stage was in Myers v. United States (1926). An 1876 Act of Congress stated that the Senate must approve the president’s firing of any postmaster. Postmasters were inferior officers. They were appointed by the president and confirmed by the Senate, and they had purely executive duties. Chief Justice (and former president) William Howard Taft held that the Act violated the separation of powers. Using the formalist approach, Taft asserted that the power to remove federal officials belongs to the Executive branch alone, not the Legislative branch. For Congress to “draw to itself, or to either [chamber] of it, the power to remove or the right to participate in the exercise of that power . . . would be . . . to infringe the constitutional principle of the separation of governmental powers,” Taft wrote. He therefore struck down Congress’ limitation on the president’s power to fire postmasters. His opinion was demonstrably formalist, invalidating Congress’ attempt to arrogate to itself a power that belongs to the Executive branch. In doing so, his opinion laid out quite a broad view of the president’s removal power.
That breadth was short-lived. Nine years later, the Court cabined the president’s removal power in Humphrey’s Executor v. United States (1935). As a preliminary matter, it should be noted that Humphrey’s Executor does not fit snugly in either the formalist or the functionalist camp. Justice Sutherland’s majority opinion at times appears to use principles from both approaches. Plus, legal scholars have disagreed about Humphrey’s Executor’s approach, with some calling it formalist and others calling it functionalist. Compare in the reading list Manning, at 1952 n.55 (calling it functionalist); Strauss, at 489 n.4 (calling it formalist); Krotoszynski, at 950 n.43 (functionalist); Brown, at 1540 n.120 (formalist); Swire, at 1766 (functionalist). (More broadly, legal scholars have disagreed about the soundness of Humphrey’s Executor in general.)
At issue was the 1914 Free Trade Commission (FTC) Act, which stated that the president can remove an FTC commissioner only for “inefficiency, neglect of duty, or malfeasance in office.” Put differently, the president could fire FTC commissioners only for cause, not at will.
Unlike in Myers, the Court upheld the limitation. Justice Sutherland, writing for the majority, distinguished the facts of Myers from those of Humphrey’s Executor. As opposed to the postmaster in Myers, Sutherland said, the FTC did not have purely executive functions and its commissioners were not mere agents of the Executive branch. Instead, Congress created the FTC as an administrative agency and gave it “quasi-legislative” and “quasi-judicial” powers. Sutherland then held that Congress had not intruded on the president’s removal power because the FTC was not itself part of the Executive branch. Because the FTC was, in Sutherland’s mind, more an arm of the legislative and judicial branches, Congress had the authority to prescribe limits on the removal of the FTC’s commissioners. In other words, Sutherland’s opinion suggests that the Constitution does not give the president the power to remove officers whose duties are not purely executive. To conclude otherwise would permit the president to undermine Congress’ power to establish protections for (quasi-)legislative and judicial actors. (Again, subsequent commentators have derided several dicta in Humphrey’s Executor. Some have gone so far as to call for its overruling, including current Justices Clarence Thomas and Neil Gorsuch.)
In 1986, the Court decided Bowsher v. Synar, and it is as good a case as any to compare the two doctrines. At issue was the Gramm-Rudman-Hollings Act of 1985, which empowered the Comptroller General to make federal budget cuts (an executive power) without presidential oversight. The Act also stated that the Comptroller could be removed only by Congress via joint resolution or by the president for cause.
In an opinion written by Chief Justice Burger, the Court struck down the removal restrictions. Burger’s formalism was loud and clear. Quoting the paradigmatic formalist case, INS v. Chadha (1983), Burger began by reiterating that “‘[the] Constitution sought to divide the delegated powers of the new Federal Government into three defined categories, Legislative, Executive, and Judicial.'” The Constitution thus “does not contemplate an active role for Congress in the supervision of officers charged with the execution of the laws it enacts.” But here, Congress sought to have just such a role. The Act not only gave the Comptroller considerable power to execute laws concerning the federal budget; it then made the Comptroller General “answerable only to Congress,” thereby giving Congress the power to supervise the execution of federal budget laws. “The structure of the Constitution does not permit Congress to execute the laws,” Burger wrote, and “it follows that Congress cannot grant to an officer under its control what it does not possess.”
Justice Byron White, in dissent, called Burger’s opinion “distressingly formalistic.” It is in White’s dissent that we see functionalism laid bare. White argued that the simplistic, age-old tripartite division of power no longer holds in 1986. He writes:
“In an earlier day, in which simpler notions of the role of government in society prevailed, it was perhaps plausible to insist that all ‘executive’ officers be subject to an unqualified Presidential removal power. . . . [B]ut with the advent and triumph of the administrative state and the accompanying multiplication of the tasks undertaken by the Federal Government, the Court has been virtually compelled to recognize that Congress may reasonably deem it ‘necessary and proper’ to vest some among the broad new array of governmental functions in officers who are free from the partisanship that may be expected of agents wholly dependent upon the President.”
White was thus concerned about just how substantial had been the growth of the administrative state, the complexities surrounding the federal budget, and the duties assigned to officers tasked with enforcing Congress’ budget laws. In this modern form of government, White asserted, the old-fashioned formalist approach is inadequate, for it cannot account for the changes in the federal government’s structure and size that had taken place. Then came the functionalist dicta: “The role of this Court,” White argued, “should be limited to determining whether the Act so alters the balance of authority among the branches of government as to pose a genuine threat to the basic division between the lawmaking power and the power to execute the law.” Burger’s formalism, of course, would stop White at “alters the balance of authority among the branches of government.” A formalist need go no further to conclude that an Act of Congress that so alters that balance violates the separation of powers as envisioned by the Framers.
White could not have known in 1986 that the functionalism underlying his dissent would become the Court’s dominant approach in future removal-power cases. And yet that’s exactly what happened. Just two years after Bowsher (during which time William Rehnquist succeeded Warren Burger as Chief Justice), the Court decided Morrsion v. Olson (1988). At issue was the 1978 Ethics in Government Act, passed in the wake of the Watergate scandal. The Act established the “independent counsel,” a court-appointed independent attorney who was tasked with investigating potential crimes committed by high-ranking federal officials (an executive function). The Act also stated that the Attorney General could remove any independent counsel only for “good cause.”
Imagine, for a moment, how a formalist would approach this case. Formalism holds that a permissible Congressional limit on the president’s power to remove a public official hinges on whether Congress has given that official the power to execute the law. If the answer is “yes”—like in Myers and Bowsher—then the Congressional limit violates the separation of powers. If the answer is “no”—like in Humphrey’s Executor—then the Congressional limit might be valid. Here, the independent counsel has purely executive duties, namely investigating and prosecuting potential violations of federal law. Therefore, Congress’ “good-cause” protection for the independent counsel violates the separation of powers. And indeed, Justice Antonin Scalia adopted this line of reasoning hook, line, and sinker. He wrote in Morrison that “the line of permissible restriction upon removal of principal officers lies at the point at which the powers exercised by those officers are no longer purely executive.”
Yet Justice Scalia was the lone dissenter. Five Justices who relied on formalism to strike down the removal restriction in Bowsher then switched sides and relied on functionalism to uphold the removal restriction in Morrison just two years later.
Chief Justice Rehnquist, writing for the majority, held that the Congressional removal limits are constitutional. Rehnquist jettisoned the rigid formalist approach, declaring that “the determination of whether the Constitution allows Congress to impose a . . . 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.'” Instead, the “real question is whether the removal restrictions are of such a nature that they impede the President’s ability to perform his constitutional duty, and the functions of the officials in question must be analyzed in that light” (emphasis added).
Thus, if faced with a statute that gives removal protections to a wholly executive officer but does not impede the proper functioning of the Executive Branch, the functionalist would uphold it (whereas the formalist would strike it down). And that is exactly the case here, thought Rehnquist. Although the independent counsel undoubtedly performs executive acts, a Congressional limit on the president’s power to fire the independent counsel does not “sufficiently deprive the President of control over the independent counsel to interfere impermissibly with his constitutional obligation to ensure the faithful execution of the laws,” he wrote.
Following Morrison was Mistretta v. United States (1989), in which Justice Harry Blackmun again relied on the functionalist approach to uphold for-cause removal protections for members of the U.S. Sentencing Commission. Blackmun expressly rejected the argument that the removal restrictions would frustrate the Judiciary’s ability to perform its constitutionally required functions. “[W]e see no risk,” Blackmum wrote, “that the . . . Act’s removal provision will prevent the Judicial Branch from performing its constitutionally assigned function of fairly adjudicating cases and controversies.”
Justice Scalia was again alone in dissent. He lamented: “Today’s decision . . . treat[s] the Constitution as though it were no more than a generalized prescription that the functions of the Branches should not be commingled too much—how much is too much to be determined, case-by-case, by this Court.” Mistretta thus presents the formalist-functionalist dichotomy in high relief. Was the Congressional limitation on one of the president’s powers constitutional? Blackmun, the functionalist, said yes; the limitation would not stymie the functioning of the Executive branch. Scalia, the formalist, said no; the Constitution’s separation of powers does not allow for an ad hoc, how-much-is-too-much approach.
From 1989 to 2010, the Court did not decide any removal-power cases. During that time, John Roberts succeeded William Rehnquist as Chief Justice (and several Justices entered and exited too). Since Roberts’ confirmation in 2005, the Court has decided just two removal-power cases: Free Enterprise Fund v. Pub. Co. Accounting Oversight Bd. (2010) and Seila Law, LLC v. Consumer Financial Protection Bureau (2020). Both were decided 5:4 on ideological lines. In both cases, the five Justices in the majority relied on the formalist approach, and the four in dissent relied on the functionalist approach. Chief Justice Roberts authored both majority opinions.
In Free Enterprise Fund, the Court faced an “unusual” two-layered removal-restriction scheme. The Public Company Accounting Oversight Board was a division of the Securities and Exchange Commission (SEC). Its members were appointed by the SEC director and could be removed only for good cause by the SEC (whose commissioners, in turn, were also insulated with good-cause protections). The statute did not expressly give the president the power to remove the Board’s members.
Roberts held that the scheme violates the separation of powers. A double layer of good-cause protection “is contrary to Article II’s vesting of the executive power in the President,” he argued. Much like the situation in Bowsher, the Board’s members were not directly accountable to the president. Rather, they answered to the SEC. This setup “transforms” the Board’s independence, Roberts wrote, since “[n]either the President, nor anyone directly responsible to him, nor even an officer whose conduct he may review only for good cause, has full control over the Board.” The president was therefore “stripped” of his “ability to . . . hold his subordinates accountable for their conduct.”
To make matters worse, the Board’s members were executive officers and thus unelected officials. Their dual-layered removal protections left them unaccountable not only to the president but to the American people as well. Thus, Roberts emphasized, the president “can neither ensure that the laws are faithfully executed, nor be held responsible for a Board member’s breach of faith.” By giving the Board such thick protection, Congress had “subvert[ed]” the power of the Executive Branch and undermined the “public’s ability to pass judgment on [its] efforts.” Therefore, Roberts concluded that the removal-protection scheme violates the separation of powers.
In hindsight, there are hints of functionalism in Roberts’ reasoning. A quick glance might leave the reader with the impression that the crux of Roberts’ argument was that the dual-layered removal-protection scheme inhibited the president’s ability to exercise the duties of his office—and thus obstructed the proper functioning of the Executive Branch. Such a reading may not be too far off the mark.
However, the language and tenor of his opinion is decidedly formalist. He cites the “landmark” case of Myers while simultaneously giving only a brief gloss of Humphrey’s Executor and distinguishing the holding of Morrison. He adopts a strong reading of Article II’s Vesting Clause and then concludes that the statute at issue had encroached on the president’s power under the Vesting Clause (and, by extension, the separation of powers).
Finally, Roberts rejected a purebred functionalist approach proffered by the Board and Justice Breyer in dissent. Breyer called the Board “the kind of practical accommodation between the Legislature and the Executive that should be permitted in a workable government.” But the Board’s practicality means little if it is masked from oversight by an elected official, Roberts countered. Quoting Bowsher, “the ‘fact that a given law or procedure is efficient, convenient, and useful in facilitating functions of government, standing alone, will not save it if it is contrary to the Constitution,’ for ‘[c]onvenience and efficiency are not the primary objectives—or the hallmarks—of democratic government.'”
Which brings us, at last, to Seila Law v. CFPB. A product of the 2010 Dodd-Frank Act, the CFPB is an independent agency tasked with regulating consumer financial products. It wields considerable executive power: eighteen federal financial statutes fall under its jurisdiction; it can investigate any “unfair” or “deceptive” practice in the consumer-finance sector by, inter alia, serving subpoenas and prosecuting civil suits in federal court; and it may initiate administrative proceedings to force compliance with those eighteen consumer-finance laws. Crucial here, the CFPB is headed by a single director whom the president can remove only for “inefficiency, neglect of duty, or malfeasance in office” (i.e., good cause).
Roberts—joined by Justices Thomas, Alito, Gorsuch, and Kavanaugh—struck down the removal protections for the CFPB’s director, holding that they violate the separation of powers. Roberts begins by reiterating Myers’ holding that Article II gives the president broad—nearly “unrestrictable”—removal powers, and Congressional attempts to restrict that removal power are presumptively invalid. He seems compelled to label Humphrey’s Executor and Morrison as the two “exceptions” to that general rule, but he then distinguishes both cases from the one at bar. Unlike the FTC in Humphrey’s Executor, the CFPB is (1) headed by a single individual who is not nonpartisan and (2) not a mere legislative or judicial aid; it “acts as a mini legislature, prosecutor, and court, responsible for creating substantive rules for a wide swath of [financial] industries, prosecuting violations, and levying knee-buckling penalties against private citizens,” Roberts writes. And unlike the independent counsel in Morrison, the CFPB’s director is not an inferior officer. The director instead wields considerable policymaking and administrative power.
After he notes the “historical anomaly” that is the CFPB’s structure, Roberts’ formalism steps up to the plate. It has shades of his argument in Free Enterprise Fund. Central to his argument is the fact that the Framers designed the Executive Branch specifically so that a single individual wielding immense power is accountable to every American. To Roberts, it follows that those who exercise the president’s power on his behalf—executive officials—must be accountable to the president (so that, by extension, they remain accountable to the American people). Thus, any executive officer who is not accountable to the president flouts the Framers’ design. The CFPB’s director fails Roberts’ accountability test because of the director’s removal protections.
Roberts begins with a survey of the Framers’ writings. When the Framers were structuring our fledgling country, they recognized that, in the long run, “structural protections against abuse of power were critical to preserving liberty,” Roberts explains (quoting Bowsher). So, first they divided sovereignty between state governments and the federal government. Then they divided the powers of the federal government among three coordinate branches, one of which in turn they further split into two chambers.
The Executive Branch, however, they left undivided. On the one hand, they knew that the unique responsibility of enforcing every law passed by Congress required “[d]ecision, activity, secrecy, and dispatch” (Alexander Hamilton, The Federalist No. 70). On the other hand, they were aware of the immense power that would accompany the leader of the Executive Branch. Their solution? A single person—the president—would be the nation’s chief executive, and he would be accountable to every American citizen (as he is the only public official elected by the entire nation, save for the vice president).
The Framers’ strategy was thus “straightforward,” Roberts says: “divide power everywhere except for the Presidency, and render the President directly accountable to the people through regular elections.” True, the executive officers whom the president appoints to help him carry out his duties will “still wield significant authority.” But fret not: “that authority remains subject to the ongoing supervision and control of the elected President,” Roberts argues. With such oversight, “’the chain of dependence [is] preserved,’ so that ‘the lowest officers, the middle grade, and the highest’ all ‘depend, as they ought, on the President, and the President on the community’” (quoting James Madison in the First Congress).
Let’s return to the CFPB—and heavens, do problems emerge. Its single-director structure “vest[s] significant governmental power in the hands of a single individual” who is “neither elected by the people nor meaningfully controlled . . . by someone who is,” Roberts asserts. What is his justification for the latter attribute? The good-cause removal protections. Just like he argued in Free Enterprise Fund, Roberts says the CFPB’s director is not directly accountable to the president because the president has no power to fire the director without cause. Implicit in Roberts’ argument is the assumption that accountability hinges on at-will removal power. As long as the CFPB director hasn’t been “inefficien[t],” hasn’t “neglected [his] duty,” and hasn’t committed “malfeasance in office,” Congress has prohibited the president from firing the director even if the director is not enforcing consumer-finance laws according to the manner in which the President (and, by extension, the American people) thinks is proper. Thus, an unelected official has the power to wield the sword of the Executive Branch with “accountab[ility] to no one.” This scheme “contravenes the carefully calibrated system” the Framers designed, Roberts concludes.
Roberts offers a few other reasons why the CFPB’s structure makes the director’s removal protections problematic. For example, the director serves a term of five years, which means some Presidents may never have the chance to appoint a director while in office. “[A]n unlucky President might get elected on a consumer-protection platform and enter office only to find herself saddled with a holdover Director from a competing political party who is dead set against that agenda,” Roberts writes (emphasis in original).
In addition, the CFPB isn’t even subject to budgetary accountability. It doesn’t receive its funds from Congress via the federal budget process; it receives its funds from the Federal Reserve (which is similarly exempt from the normal appropriations process), giving the CFPB great “financial freedom.” This makes it all the more likely that the agency “will ‘slip from the Executive’s control, and thus from that of the people,’” Roberts notes (quoting Free Enterprise Fund).
Finally, Roberts rejects functionalist arguments in Justice Kagan’s dissent. Kagan contends that Congress has “discretion . . . to structure administrative agencies as the times demand, so long as the President retains the ability to carry out his constitutional duties.” But Roberts counters that such a standard is “amorphous” and “provides no real limiting principle.” “[T]he urge to meet new technological and societal problems with novel governmental structures must be tempered by constitutional restraints that are not known—and were not chosen—for their efficiency or flexibility.”
Two data points barely a trend make. But in two similar removal-power cases, Chief Justice Roberts has steered the Court down the formalist road. And with the same driver at the wheel, Roberts has had the chance to set out and clarify his removal-power jurisprudence and, perhaps, set the Court up for a long trip down the formalist path. To Roberts, the president alone heads the Executive Branch. His immense power must be checked by accountability to the people. If he delegates executive power to unelected officials, those officials must in turn remain accountable to the people. If the president’s power to remove those officials is restricted, those officials are insulated from the normal accountability process. Thus, such Congressional limitations on the president’s removal power violate the separation of powers because they leave the power of the Executive Branch in the hands of unelected, unaccountable officials.
A third, fourth, and fifth data point from the Roberts court would better test my hypothesis. We will get a third data point next term in Collins v. Mnuchin, which concerns the structure of the Federal Housing Finance Agency (FHFA). Strikingly similar to the CFPB, the FHFA is exempt from the normal Congressional appropriations process and is headed by a single director who may be removed by the president only for cause. Ceteris paribus, Roberts’ opinions in Free Enterprise Fund and Seila Law bode ill for the FHFA.
- William N. Eskridge, Jr., Relationships between Formalism and Functionalism in Separation of Powers Cases, 22 Harv. J. L. & Pub. Pol’y 21 (1998).
- John F. Manning, Separation of Powers As Ordinary Interpretation, 124 Harv. L. Rev. 1939 (2011).
- Peter L. Strauss, Formal and Functional Approaches to Separation-of-Powers Questions—A Foolish Inconsistency?, 72 Cornell L. Rev. 488 (1987).
- Rebecca L. Brown, Separated Powers and Ordered Liberty, 139 U. Penn. L. Rev. 1513, 1522–31 (1991).
- Ronald J. Krotoszynski, Jr., Partisan Balance Requirements in the Age of New Formalism, 90 Notre Dame L. Rev. 941, 950–61 (2015).
- Thomas W. Merrill, The Constitutional Principle of Separation of Powers, 1991 S. Ct. Rev. 225, 225–35 (1991).
- Gillian E. Metzger, The Roberts Court and Administrative Law, 2019 S. Ct. Rev. 1, 39–48 (2019).
- Samuel W. Cooper, NOTE: Considering ‘Power’ in Separation of Powers, 46 Stan. L. Rev. 362, 367–85 (1994).
- Peter P. Swire, NOTE: Incorporation of Independent Agencies into the Executive Branch, 94 Yale L.J. 1766 (1985).