Opinion for the Court filed by Circuit Judge KAVANAUGH, with whom Senior Circuit Judge RANDOLPH joins, and with whom Circuit Judge HENDERSON .joins as to Parts I, IV, and V.
Concurring opinion filed by Senior Circuit Judge RANDOLPH.
Opinion concurring in part and dissenting in part filed by Circuit Judge HENDERSON.
KAVANAUGH, Circuit Judge:INTRODUCTION AND SUMMARY
This is a case about executive power and individual liberty. The U.S. Government’s executive power to enforce federal law against private citizens—for example,- to bring criminal prosecutions and civil én-forcement actions—is essential to societal order and progress, but simultaneously a grave threat to individual liberty.
The Framers understood that threat to individual liberty. When designing the executive power, the Framers first separated the executive power from the legislative and judicial powers. “The declared purpose of separating and dividing the powers of government, of course, was to ‘diffus[e] power the better to secure liberty.’ ” Bowsher v. Synar, 478 U.S. 714, 721, 106 S.Ct. 3181, 92 L.Ed.2d 583 (1986) (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 635, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (Jackson, J., concurring)). To ensure accountability .for the exercise of executive power, and help safeguard liberty, the Framers then lodged full responsibility for the executive power in the President of the United States, who is elected by and accountable to the people. The text of Article II provides quite simply: “The executive Power shall be vested in a President of the United States of America.” U.S. Const, art. II, § 1. And Article II assigns the President alone the authority and responsibility to “take Care that the Laws be faithfully executed.” Id. § 3. As Justice Scalia . explained: “The purpose of the separation and equilibration of powers in general, and of the unitary Executive in particular, was not merely to assure effective government but to preserve individual freedom.” Morrison v. Olson, 487 U.S. 654, 727, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988) (Scalia, J., dissenting).
Of course, the President executes the laws with the assistance of subordinate executive officers who are appointed by the President, often with the advice and consent of the Senate. To carry out the executive power and be accountable for the exercise of that power, the President must be able to control subordinate officers in executive agencies. In its landmark decision in Myers v. United States, 272 U.S. 52, 47 S.Ct. 21, 71 L.Ed. 160 (1926), authored by Chief Justice and former President Taft, the Supreme Court therefore recognized the President’s Article II authority to supervise, direct,-and remove at will subordinate officers in the Executive Branch, .
In 1935, however, the Supreme Court carved out an exception to Myers and Article II by permitting Congress to create independent agencies that exercise executive power. See Humphrey’s Executor v. United States, 295 U.S. 602, 55 S.Ct. 869, 79 L.Ed. 1611 (1935). An agency is considered “independent” when the agency heads are removable by the President only for cause, not at will, and therefore are not *6supervised or directed by the President. Examples of independent agencies include well-known bodies such as the Federal Communications Commission, the Securities and Exchange Commission, the Federal Trade Commission, the National Labor Relations Board, and the Federal Energy Regulatory Commission. Those and other established independent agencies exercise executive power by bringing enforcement actions against private citizens and by issuing legally binding rules that implement statutes enacted by Congress.
The independent agencies collectively constitute, in effect, a headless fourth branch of the U.S. Government. They exercise enormous power over the economic and social life of the United States. Because of their massive power and the absence of Presidential supervision and direction, independent agencies pose a significant threat to individual liberty and to the constitutional system of separation of powers and checks and balances.
To help mitigate the risk to individual liberty, the independent agencies, although not checked by the President, have historically been headed by multiple commissioners, directors, or board members who act as checks on one another. Each independent agency has traditionally been established, in the Supreme Court’s words, as a “body of experts appointed by law and informed by experience.” Humphrey’s Executor, 295 U.S. at 624, 55 S.Ct. 869 (internal quotation marks omitted). The multi-member structure reduces the risk of arbitrary decisionmaking and abuse of power, and thereby helps protect individual liberty.
In other words, to help preserve individual liberty under Article II, the heads of executive agencies are accountable to and checked by the President, and the heads of independent agencies, although not accountable to or checked by the President, are at least accountable to and checked by their fellow commissioners or board members. No head of either an executive agency or an independent agency operates unilaterally without any check on his or her authority. Therefore, no independent agency exercising substantial executive authority has ever been headed by a single person.
Until now.
In the Dodd-Frank Act of 2010, Congress established a new independent agency, the Consumer Financial Protection Bureau. As proposed by then-Professor and now-Senator Elizabeth Warren, the CFPB was to be another traditional, multi-mem-ber independent agency. See Elizabeth Warren, Unsafe at Any Rate: If It’s Good Enough for Microwaves, It’s Good Enough for Mortgages. Why We Need a Financial Product Safety Commission, Democracy, Summer 2007, at 8, 16-18. The initial Executive Branch proposal in 2009 likewise envisioned a traditional, multi-member independent agency. See Department of the TReasuey, Financial RegulatoRy RefoRM: a New Foundation: Rebuilding Financial SuPERVisiON and Regulation 58 (2009). The House-passed bill sponsored by Congressman Barney Frank and championed by Speaker Nancy Pelosi also contemplated a traditional, multi-member independent agency. See H.R. 4178, 111th Cong. § 4103 (as passed by House, Dec. 11, 2009).
But Congress ultimately departed from the Warren and Administration proposals, and from the House bill. Congress established the CFPB as an independent agency headed not by a multi-member commission but rather by a single Director.
Because the CFPB is an independent agency headed by a single Director and not by a multi-member commission, the Director of the CFPB possesses more uni*7lateral authority—that is, authority to take action on one’s own, subject to no check— than any single commissioner- or board member in any other independent agency in the U.S. Government. Indeed, as we will explain, the Director enjoys more unilateral authority than any other officer in any of the three branches of the U.S. Government, other than the President.
At the same time, the Director of the CFPB possesses enormous power over American business, American consumers, and the overall U.S. economy. The Director unilaterally enforces 19 federal consumer protection statutes, covering everything from home finance to student loans to credit cards to banking practices. The Director alone decides what rules to issue; how to enforce, when to enforce, and against whom to enforce the law; and what sanctions and penalties to impose on violators of the law. (To be sure, judicial review serves as a constraint on illegal actions, but not on discretionary decisions within legal boundaries; therefore, subsequent judicial review of individual agency decisions has never been regarded as sufficient to excuse a structural separation of powers violation.)
That combination of power that is massive in scope, concentrated in a single person, and unaccountable to the President triggers the important constitutional question at issue in this case.
The petitioner here, PHH, is a mortgage lender and was the subject of a CFPB enforcement action that resulted in a $109 million order against it. In seeking to vacate the order, PHH argues that the CFPB’s status as an independent agency headed by a single Director violates Article II of the Constitution.
The question before us is whether we may extend the Supreme Court’s Humphrey’s Executor precedent to cover this novel, single-Director agency structure for an independent agency. To analyze that issue, we follow the history-focused approach long applied by the Supreme Court in separation of powers cases where, as here, the constitutional text alone does not resolve the matter.
Two recent Supreme Court decisions exemplify that historical analysis. In its 2010 decision in Free Enterprise Fund v. Public Company Accounting Oversight Board, the Supreme Court held that the new Accounting Oversight Board at issue in that case—with two levels rather than one level of for-cause protection insulating the independent agency heads from the President—exceeded the bounds on traditional independent agencies and thus violated Article II. 561 U.S. 477, 514, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010). In so ruling, the Court emphasized, among other things, the novelty of the Board’s structure: “Perhaps the most telling indication of the severe constitutional problem with the PCAOB is the lack of historical precedent for this entity.” Id. at 505, 130 S.Ct. 3138 (internal quotation marks omitted). In its 2014 decision in NLRB v. Noel Canning, the Supreme Court held that recess appointments in Senate recesses of fewer than 10 days were presumptively unconstitutional under Article II. — U.S.-, 134 S.Ct. 2550, 2567, 189 L.Ed.2d 538 (2014). Why 10 days? The Court explained: “Long settled and established practice is a consideration of great weight in a proper interpretation of constitutional provisions regulating the relationship between Congress and the President.” Id. at 2559 (internal quotation marks and alteration omitted). And the historical practice of Presidents and Senates had established a de facto 10-day line so that recess appointments in recesses of fewer than 10 days were impermissible. See id. at 2567.
As those two cases illustrate, history and tradition are critical factors in sepa*8ration of powers cases where the constitutional text does not otherwise resolve the matter, As Justice Breyer wrote for the Court in Noel Canning, that bedrock principle—namely, that the “longstanding practice of the government can inform our determination of what the law is”—is “neither new nor controversial.” Id. at 2560 (internal quotation marks and citation omitted) (quoting McCulloch v. Maryland, 17 U.S. 4 Wheat. 316, 401, 4 L.Ed. 579 (1819) and Marburg v. Madison, 5 U.S. 1 Cranch 137, 177, 2 L.Ed. 60 (1803)).
In this case, the single-Director structure of the CFPB represents a gross departure from settled historical practice. Never before has an independent agency exercising substantial executive authority been headed by just one person.
The CFPB’s concentration of enormous executive power in a single, unaccountable, unchecked Director not only departs from settled historical practice, but also poses a far greater risk of arbitrary decisionmak-ing and abuse of power, and a far greater threat to individual liberty, than does a multi-member independent agency. The overarching constitutional concern with independent agencies is that the agencies are unchecked by the President, the official who is accountable to the people and who is responsible under Article II for the exercise of executive power. Recognizing the broad and unaccountable power wielded by independent agencies, Congresses and Presidents of both political parties have therefore long endeavored to keep independent agencies in check through other statutory means. In particular, to check independent agencies, Congress has traditionally required multi-member bodies at the helm of every independent agency. In lieu of Presidential control, the multi-member structure of independent agencies acts as a critical substitute check on the excesses of any individual independent agency head—a check that helps to prevent arbitrary decisionmaking and thereby to protect individual liberty.
This new agency, the CFPB, lacks that critical check and structural constitutional protection, yet wields vast power over the U.S. economy. So “this wolf comes as a wolf.” Morrison v. Olson, 487 U.S. at 699, 108 S.Ct. 2597 (Scalia, J,, dissenting).
In light of the consistent historical practice under which independent agencies have been headed by multiple commissioners or board members, and in light of the threat to individual liberty posed by a single-Director independent agency, we conclude that Humphrey’s Executor cannot be stretched to cover this novel agency structure. We therefore hold that the CFPB is unconstitutionally structured.
What is the remedy for that constitutional flaw? PHH contends that the constitutional flaw means that we must shut down the entire CFPB (if not invalidate the entire Dodd-Frank Act) until Congress, if it chooses, passes new legislation fixing the constitutional flaw. But Supreme Court precedent dictates a narrower remedy. To remedy the constitutional flaw, we follow the Supreme Court’s precedents, including Free Enterprise Fund, and simply sever the statute’s unconstitutional for-cause provision from the remainder of the statute. Here, that targeted remedy will not affect the ongoing operations of the CFPB. With the for-cause provision severed, the President now will have the power to remove the Director at will, and to supervise and direct the Director. The CFPB therefore will continue to operate and to perform its many duties, but will do so as an executive agency akin to other executive agencies headed by a single person, such as the Department of Justice and the Department of the Treasury. Those executive agencies have traditionally been headed by a single person precisely be*9cause the agency head operates within the Executive Branch chain of command under the supervision and direction of the President. The President is a check on and accountable for the actions of those executive agencies, and the President now will be a check on and accountable for the actions of the CFPB as well.
Because the CFPB as remedied will continue operating, we must also address the statutory issues raised by PHH in its challenge to the $109 million order against it.1 PHH raises three main statutory arguments.
First, PHH argues' that the CFPB incorrectly interpreted Section 8 of the Real Estate Settlement Procedures Act to bar so-called captive reinsurance arrangements involving mortgage lenders such as PHH and their affiliated reinsurers. In a captive reinsurance arrangement, a mortgage lender (such as PHH) refers borrowers to a mortgage insurer. In return, the mortgage insurer buys reinsurance from a mortgage reinsurer affiliated with (or owned by) the referring mortgage lender. We agree with PHH that Section 8 of the Act allows captive reinsurance arrangements so- long, as the amount paid by the mortgage insurer for the reinsurance does not exceed the reasonable market value of the reinsurance..
Second, PHH claims that, in any event, the CFPB departed from the consistent prior interpretations issued by the Department of. Housing and Urban Development, and that the CFPB then retroactively applied its new interpretation, of the Act agaihst PHH, thereby violating PHH’s due process rights. We again agree with,PHH: The CFPB’s order violated bedrock, principles of due process.
Third, in light of our ruling on the constitutional and -statutory issues, the CFPB on remand still will have an opportunity to demonstrate that the relevant mortgage insurers in fact paid more than reasonable market value to the PHH-affiliated rein-surer for reinsurance, thereby making disguised payments for referrals in contravention of Section 8. PHH claims, however, that much of the alleged misconduct oc*10curred outside of the three-year statute of limitations and therefore may not be the subject of a CFPB enforcement action. The CFPB responds that, under Dodd-Frank, there is no statute of limitations for any CFPB administrative actions to enforce any consumer protection law. In the alternative, the CFPB contends that there is no statute of limitations for administrative actions to enforce Section 8 of the Real Estate Settlement Procedures Act. We disagree with the CFPB on both points. First of all, the Dodd-Frank Act incorporates the statutes of limitations in the underlying statutes enforced by the CFPB in administrative proceedings. And under the Real Estate Settlement Procedures Act, a three-year statute of limitations applies to all CFPB enforcement actions to enforce Section 8, whether brought in court or administratively.
In sum, we grant PHH’s petition for review, vacate the CFPB’s order against PHH, and remand for further proceedings consistent with this opinion. On remand, the CFPB may determine among other things whether, within the applicable three-year statute of limitations, the relevant mortgage insurers paid more than reasonable market value to the PHH-affiliated reinsurer.
In so ruling, we underscore the important but limited real-world implications of our decision. As before, the CFPB will continue to operate and perform its many critical responsibilities, albeit under the ultimate supervision and direction of the President. Section 8 will continue to mean what it has traditionally meant: that captive reinsurance agreements are permissible so long as the mortgage insurer pays no more than reasonable market value for the reinsurance. And the three-year statute of limitations that has traditionally applied to agency actions to enforce Section 8 will continue to apply.
With apologies for the length of this opinion, we now turn to our detailed explanation and analysis of these important issues.
I
PHH is a large home mortgage lender. When PHH and other lenders provide mortgage loans to homebuyers, they require certain homebuyers to obtain mortgage insurance. Mortgage insurance protects lenders by covering part of the lenders’ losses if homebuyers default on their mortgages. Homebuyers pay monthly premiums to the mortgage insurer for the insurance.
In turn, mortgage insurers may obtain mortgage reinsurance. In the same way that mortgage insurance protects lenders, mortgage reinsurance protects mortgage insurers. Reinsurers assume some of the risk of insuring the mortgage. In exchange, mortgage insurers pay a fee (usually a portion of the homebuyers’ monthly insurance premiums) to the reinsurers.
In 1994, PHH established a wholly owned subsidiary known as Atrium Insurance Corporation. Atrium provided reinsurance to the mortgage insurers that insured mortgages generated by PHH. In return, PHH often referred borrowers to mortgage insurers that used Atrium’s reinsurance services. That is known as a “captive reinsurance” arrangement, which was not uncommon in the industry at the time. According to PHH, the mortgage insurers did not pay more than reasonable market value to Atrium for the reinsurance.
Originally passed by Congress and signed by President Ford in 1974, the Real Estate Settlement Procedures Act is a broad statute governing real estate transactions. One of its stated purposes was “the elimination of kickbacks or referral fees that tend to increase unnecessarily *11the costs of certain settlement services.” 12 U.S.C. § 2601(b)(2).
To achieve that objective, Section 8(a) of the Act, which is'titled “Prohibition against kickbacks and unearned fees,” provides: “No person shall give and no person shall accept any fee, kickback, or thing of value pursuant to any agreement or understanding, oral or otherwise, that business incident to or a part of a real estate settlement service involving a federally related mortgage loan shall be referred to any person.” Id. § 2607(a). In plain English, Section 8(a) prohibits, as relevant here, paying for a referral—for example, a mortgage insurer’s paying a lender for the lender’s referral of homebuying customers to that mortgage insurer.
Standing aloné, Section 8(a) perhaps might have been construed by government enforcement agencies to cast doubt on a mortgage lender’s referrals of customers to mortgage insurers who in turn purchased reinsurance from a reinsurer affiliated with the lender. But another provision of the Real Estate Settlement Procedures Act, Section 8(c), carved out a series of expansive exceptions, qualifications, and safe harbors related to Section 8(a). Of relevance here, Section 8(c) provides: “Nothing in this section shall be construed as prohibiting ...' (2) the payment to any person of a bona fide salary or compensation or other payment for goods or facilities actually furnished or for services actually performed....” Id. § 2607(c).
Before the creation of the CFPB in 2010, the Department of Housing and Urban Development, known as HUD, interpreted Section 8(c) to establish a safe harbor allowing bona fide transactions between a lender and a mortgage insurer (or between a' mortgage insurer and a lender-affiliated reinsurer), so long as the mortgage insurer did not pay the lender for a referral. HUD therefore interpreted Section 8(c) to allow captive reinsurance arrangements so long as the mortgage insurer paid no more than reasonable market value for the reinsurance. If the mortgage insurer paid more than reasonable market value for the reinsurance, then a presumption would arise that the excess payment was indeed a disguised payment for the referral, which is impermissible under Section 8(a). HUD repeatedly reaffirmed that interpretation, and the mortgage lending industry relied on it.
When Congress created, the CFPB in 2010, Congress provided that the CFPB would take over enforcement of Section 8 from HUD. By regulation, the CFPB carried forward HUD’s rules, policy statements, and guidance, subject of course to any future change by the CFPB.
Therefore, under Section 8(c), as authoritatively interpreted by the Federal Government, PHH as a mortgage lender could refer customers to mortgage insurers who obtained reinsurance from Atrium—so long as the mortgage insurers paid Atrium no more than reasonable market value for the reinsurance.
Or so PHH thought. In 2014, notwithstanding Section 8(c) and HUD’s longstanding interpretation, the CFPB initiated an administrative enforcement action against PHH. The CFPB alleged that PHH’s captive reinsurance arrangement with the mortgage insurers violated Section 8.
Under the CFPB’s newly minted interpretation, Section 8 prohibits most referrals made by lenders to mortgage insurers in exchange for the insurer’s purchasing reinsurance from a lender-affiliated rein-surer. The CFPB said that Section 8 bars such a captive reinsurance arrangement even when the mortgage insurer pays no more than reasonable market value to the reinsurer for the reinsurance.
*12In its order in this case, the CFPB thus discarded HUD’s longstanding interpretation of Section 8 and, for the first time, pronounced its new interpretation. And then the CFPB applied its new interpretation of Section 8 retroactively against PHH, notwithstanding PHH’s reliance on HUD’s prior interpretation. The CFPB sanctioned PHH for previous actions that PHH had taken in reliance on HUD’s prior interpretation, even though PHH’s conduct had occurred before the CFPB’s new interpretation of Section 8. The CFPB ordered PHH to pay $109 million in disgorgement and enjoined PHH from entering into future captive reinsurance arrangements.
PHH petitioned this Court for review. A motions panel of this Court (Judges Henderson, Millett, and Wilkins) previously granted PHH’s motion for a stay of the CFPB’s order pending resolution of the merits in this case.
II
In challenging the enforcement action against it, PHH raises a fundamental constitutional objection to the entire proceeding. According to PHH, the CFPB’s structure violates Article II of the Constitution because the CFPB operates as an independent agency headed by a single Director. PHH argues that, to comply with Article II, either (i) the agency’s Director must be removable at will by the President, meaning that the CFPB would operate as a traditional executive agency; or (ii) if structured as an independent agency, the agency must be structured as a multi-member commission. We agree.
A
We begin by describing the background of independent agencies in general and the CFPB in particular.
As the Supreme Court has explained, our Constitution “was adopted to enable the people to govern themselves, through their elected leaders,” and the Constitution “requires that a President chosen by the entire Nation oversee the execution of the laws.” Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 499, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010). Under the text of Article II, the President alone is responsible for exercising the executive power. The first 15 words of Article II of the Constitution provide: “The executive Power shall be vested in a President of the United States of America.” U.S. Const. art. II, § 1. And Article II assigns the President alone the authority and responsibility to “take Care that the Laws be faithfully executed.” Id. § 3. Article II makes “emphatically clear from start to finish” that “the president would be personally responsible for his branch.” Akhil Reed AmaR, AmeRICa’s Constitution: A Biography" 197 (2005); see also Neomi Rao, Removal: Necessary and Sufficient for Presidential Control, 65 Ala. L. Rev. 1205, 1215 (2014) (“The text and structure of Article II provide the President with the power to control subordinates within the executive branch.”).
To exercise the executive power, the President must have the assistance of subordinates. See Free Enterprise Fund, 561 U.S. at 483, 130 S.Ct. 3138. The Framers therefore provided for the appointment of executive officers and the creation of executive departments to assist the President “in discharging the duties of his trust.” Id. (internal quotation marks omitted); see U.S. Const. art. II, § 2.
In order to maintain control over the exercise of executive power and take care that the laws are faithfully executed, the President must be able to supervise and direct those subordinate executive officers. *13See Free Enterprise Fund, 561 U.S. at 498-502, 130 S.Ct. 3138. As James Madison- stated during the First Congress, “if any power whatsoever -is in its nature Executive, it is the power of appointing, overseeing, and controlling those who execute the laws.” 1 Annals of Congress 463 (Madison) (1789) (Joseph Gales ed., 1834).
To supervise and direct executive offi-cérs, the President must be able to remove those officers at will. See generally Myers v. United States, 272 U.S. 52, 47 S.Ct. 21, 71 L.Ed. 160 (1926). Otherwise, a subordinate could ignore the Presidént’s supervision and direction without fear, and the President could do nothing about it. See Bowsher v. Synar, 478 U.S. 714, 726, 106 S.Ct. 3181, 92 L.Ed.2d 583 (1986) (“Once an officer is appointed, it is only the authority that can remove him, and not the authority that appointed him, that he must fear and, in the performance of his functions, obey”) (internal quotation marks omitted). The Article II chain of command depends on the President’s removal power. As James Madison explained: “If the President should possess alone the power of removal from office, those who are • employed in the execution of the law will be in their proper situation, and the chain of dependence be preserved; the lowest officers, the middle grade, and the highest, will depend, as they ought, on the President, and the President on the community.” 1 Annals of Congress 499 (Madison). The Supreme Court recently summarized the Article II chain of command this way: “The Constitution that makes the President accountable to the people for executing the laws also gives him the power to do so. That power includes, as a general matter, the authority to remove those who assist him in carrying out his duties. Without such power, the President could not be held fully accountable for discharging his own responsibilities; the buck would' stop somewhere else.” Free Enterprise Fund, 561 U.S. at 513-14, 130 S.Ct. 3138.
In the late 1800s and the early 1900s, as part of the Progressive Movement and an emerging belief in expert, apolitical, and scientific answers to certain public policy questions, Congress began creating new expert agencies that were independent of the President but that exercised executive power. The heads of those independent agencies were removable by the President only for cause, not at will, and were neither supervised nor directed by the President. Some early examples included the Interstate Commerce Commission (1887) and the Federal Trade Commission (1914). Importantly, the independent agencies were all multi-member bodies: They were designed as non-partisan expert bodies that would neutrally and impartially issue rules, bring law enforcement actions, and resolve disputes in their respective jurisdictions.
In a 1926 decision written by Chief Justice and former President Taft, the Supreme Court ruled that, under Article II, the President must be able to supervise, direct, and remove at will certain executive officers. The Court stated: “[W]hen the grant of the executive power is enforced by the express mandate to take care that the laws be faithfully executed, it emphasizes the necessity for including within the executive power as conferred the exclusive power of removal,” Myers, 272 U.S. at 122, 47 S.Ct. 21.
A few years later, based on his reading of Article II and the Court’s 1926 decision in Myers, President Franklin Roosevelt vigorously contested the idea that Congress could create independent agencies and thereby prevent the President from controlling the executive power vested in those independent agencies. President Roosevelt did not object to the existence of the agencies; rather, he objected to the *14President’s lack of control over these agencies, which after all were exercising important executive power.
The issue came to a head in President Roosevelt’s dispute with William E. Humphrey, a commissioner of the Federal Trade Commission. Commissioner Humphrey was a Republican holdover from the Hoover Administration who, in President Roosevelt’s view, was too sympathetic to big business and hostile to the Roosevelt Administration’s regulatory agenda. Asserting his Authority under Article II, President Roosevelt fired Commissioner Humphrey. Humphrey contested his removal, arguing that he was protected against firing by the statute’s for-cause removal provision, and further arguing that Congress possessed authority to create such independent agencies without violating Article ll. The case reached the Supreme Court in 1935.
At its core, the case raised the question whether Article II permitted Congress to create independent agencies whose heads were not removable at will and would operate free of the President’s supervision and direction. Representing President Roosevelt, the Solicitor General argued that the case was straightforward and controlled by the text and history of Article II and the Court’s 1926 decision in Myers. But notwithstanding Article II and the decision in Myers, the Supreme Court upheld the constitutionality of independent agencies—a decision that so incensed President Roosevelt that it helped trigger his ill-fated court reorganization plan in 1937. See Humphrey’s Executor v. United States, 295 U.S. 602, 624, 631-32, 55 S.Ct. 869, 79 L.Ed. 1611 (1935). In allowing independent agencies, the Humphrey’s Executor Court found it significant that the Federal Trade Commission was intended “to be non-partisan,” to “act with entire impartiality,” and “to exercise the trained judgment of a body of experts appointed by law and informed by experience.” Id. at 624, 55 S.Ct. 869 (internal quotation marks omitted). Those characteristics, among others, led the Court to conclude that Congress could create an independent agency “wholly disconnected from the executive department.” Id. at 630, 55 S.Ct. 869. According to the Court, Congress could therefore limit the President’s power to remove the commissioners of the Federal Trade Commission and, by extension, Congress could limit the President’s power to remove the commissioners and board directors of similar independent agencies. Id. at 628-30, 55 S.Ct. 869.2
In the wake of the 1935 Humphrey’s Executor decision, independent agencies have continued to play an enormous role in the U.S. Government. The independent agencies possess massive authority over vast swaths of American economic and social life.
Importantly, however, the independent agencies have traditionally operated—and continue to operate—as multi-member “bod[ies] of experts appointed by law and informed by experience.” Id. at 624, 55 *15S.Ct. 869 (internal quotation marks omitted).3 ' '
The independent agency at issue here, the CFPB, aróse out of an idea originally proposed by then-Professor and now-Senator Elizabeth Warren. In 2007, concerned about balkanized and inconsistent federal law enforcement of consumer protection statutes, Professor Warren advocated that Congress create a new independent agency, which she called a Financial Product Safety Commission. This new agency would centralize and unify federal law enforcement to protect consumers. See Elizabeth Warren, Unsafe at Any Rate: If It’s Good Enough for Microwaves, It’s Good Enough for Mortgages. Why We Need a Financial Product Safety Commission, Democracy, Summer 2007, at 8,16-18.
The agency proposed by Professor Warren was to operate as a traditional multi-member independent agency. The subsequent Executive Branch proposal for such a new agency likewise contemplated a mul-ti-member structure. See Department op THE TREASURY, FINANCIAL REGULATORY REFORM: A New Foundation: Rebuilding Financial Supervision and Regulation 58 (2009). The originally passed House bill sponsored by Congressman Barney Frank and supported by Speaker Nancy Pelosi also would have created a traditional multi-member independent agency. See H.R. 4178, 111th Cong. § 4103 (as passed by House, Dec. 11, 2009).
But Congress ultimately strayed from the Warren and Executive Branch proposals, and from the House bill, as well as from historical practice, by creating an independent agency with only a single Director. See Dodd-Frank Wall Street Reform and Consumer' Protection Act, § 1011, 12 U.S.C. § 5491. Congress made the Director of the CFPB removable only for cause—that is, for “inefficiency, neglect of duty, or malfeasance in office”—during the Director’s fixed five-year term. See 12 U.S.C. § 5491(c)(3); Humphrey’s Executor, 295 U.S. at 620, 55 S.Ct. 869. Under the statute, the President therefore may not supervise, direct, or remove at will the Director. As a result, this statute means that a Director appointed by a President may continue to serve in office even if the President later wants to remove the Director based on policy disagreement, for example. This statute also means that a Director may even continue' to serve under a new President (at least until the Director’s statutory five-year tenure has elapsed), even though the new President might strongly disagree with the Director about policy issues or the overall direction of the agency.
At the same time, Congress granted the CFPB broad authority to enforce U.S. consumer protection laws. Under the Dodd-Frank Act, the CFPB possesses the power to “prescribe rules or issue orders or guidelines pursuant to” 19 distinct consumer protection laws. 12 U.S.C. § 5581(a)(1)(A); see also id. § 5481(14). That power was previously exercised by seven different government agencies. See id. § 5581(b) (transferring to the CFPB “[a]ll consumer financial protection functions” previously exercised by the Board of Governors of the Federal Reserve, the *16Comptroller of the Currency, the Office of Thrift Supervision, the Federal Deposit Insurance Corporation, the National Credit Union Administration, and select functions of the Department of Housing and Urban Development and the Federal Trade Commission). The CFPB may pursue actions to enforce the consumer protection laws in federal court, as well as in administrative actions before administrative law judges, and may issue subpoenas requesting documents , or testimony in connection with those enforcement actions. See id. §§ 5562-5564. The CFPB has the power to impose a wide range of legal and equitable relief, including restitution, disgorgement, money damages, injunctions, and civil monetary penalties. Id. § 5565(a)(2). And all of this massive power is lodged in one person—the Director— who is not supervised, directed, or checked by the President or by other directors.
Because the Director alone heads the agency without Presidential supervision, and in light of the CFPB’s broad authority over the U.S. economy, the Director enjoys significantly more unilateral power than any single member of any other independent agency. By “unilateral power,” we mean power that is not checked by the President or by other colleagues. Indeed, other than the President, the Director of the CFPB is the single most powerful official in the entire United States Government, at least when measured in terms of unilateral power. That is not an overstatement. What about the Speaker of the House, you might ask? The Speaker can pass legislation only if 218 Members agree. The Senate Majority Leader? The Leader needs 60 Senators to invoke cloture, and needs a majority of Senators (usually 51 Senators or 50 plus the Vice President) to approve a law or nomination. The Chief Justice? The Chief Justice must obtain four other Justices’ votes for his or her position to prevail. The Chair of the Federal Reserve? The Chair needs the approval of a majority of the Federal Reserve Board. The Secretary of Defense? The Secretary is supervised and directed by the President. On any decision, the Secretary must do as the President says. So too with the Secretary of State, and the Secretary of the Treasury, and the Attorney General.
To be sure, the Dodd-Frank Act requires the Director to establish and consult with a “Consumer Advisory Board.” See id. § 5494. But the advisory board is just that: advisory. Nothing requires the Director to heed the Board’s advice. Without the formal authority to prevent unilateral action by the Director, the Advisory Board does not come close to equating to the check provided by the multi-member structure of traditional independent commissions.
The Act also, in theory, allows a super-majority of the Financial Stability Oversight Council to veto certain regulations of the Director. See id. § 5513. But by statute, the veto power may be used only to prevent regulations (not to prevent enforcement actions or adjudications); only when two-thirds of the Council members agree; and only when a regulation puts “the safety and soundness of the-United States banking system or the stability of the financial system of the United States at risk,” a standard unlikely to be met in practice in most cases. Id. § 5513(c)(3)(B)(ii); see S. Rep. No. 111-176, at 166 (“The Committee notes that there was no evidence provided during its hearings that consumer protection regulation would put safety and soundness at risk.”); see also Todd Zywicki, The Consumer Financial Protection Bureau: Savior or Menace?, 81 Geo. Wash. L. Rev. 856, 875 (2013) (“[Substantive checks on the CFPB can be triggered ... only under the extreme circumstance of a severe *17threat to the safety and soundness of the American financial system. It is likely that this extreme test will rarely be satisfied in practice.”); Recent Legislation, Dodd-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. 2123, 2129 (2011) (“[T]he high standard for vetoing regulations ... will be difficult to establish.”). The veto power could not have been used in this case to override the Director’s determination regarding Section 8, for example. As with the consultation requirement, the Act’s limited veto provision falls far short of making the CFPB the equivalent of a multi-member independent agency.
Finally, the Act technically makes the CFPB part of the Federal Reserve for certain administrative purposes. See, e.g,, 12 U.S.C. § 5491(a); see also id. § 5493. But that is irrelevant to the present analysis because the Federal Reserve may not supervise, direct, or remove the Director.
In short, when measured in terms of unilateral power, the Director of the CFPB is the single most powerful official in the entire U.S. Government, other than the President. Indeed, within his jurisdiction, the Director of the CFPB can be considered even more powerful than the President. It is the Director’s view of consumer protection law that prevails over all others. In essence, the Director is the President of Consumer Finance. The concentration of massive, unchecked power in a single Director marks a departure from settled historical practice and makes the CFPB unique among traditional independent agencies, as we will now explain.
B
As a single-Director independent agency exercising substantial executive authority, the CFPB is the first of its kind and a historical anomaly, Until this point in U.S. history, independent agencies exercising substantial executive authority have all been multi-member commissions or boards. A sample list includes:
• Interstate Commerce Commission (1887)
• Federal Reserve Board (1913)
• Federal Trade Commission (1914)'
• U.S. International Trade Commission (1916)
• Federal Deposit Insurance Corporation (1933)
• Federal Communications Commission (1934)
• National Mediation Board (1934)
• Securities and Exchange Commission (1934)
• National Labor Relations Board (1935)
• Federal Maritime Commission (1961)
• National Transportation Safety Board (1967)
• National Credit Union Administration (1970)
• Occupational Safety and Health Review Commission (1970)
• Postal Regulatory Commission (1970)
• Consumer Product Safety Commission (1972)
• Nuclear Regulatory Commission (1974)
• Federal Energy Regulatory Commission (1977)
• Federal Mine Safety and Health Review Commission (1977)
• Federal Labor Relations- Authority (1978)
• Merit Systems Protection Board (1978)
• Defense Nuclear Facilities Safety Board (1988)
• National Indian Gaming Commission (1988)
*18• Chemical Safety and Hazard Investigation Board (1990)
• Surface Transportation Board (1995)
• Independent Payment Advisory Board (2010).4
Have there been any independent agencies headed by a single person? Prior to oral argument, in an effort to be comprehensive, the Court issued an order asking the CFPB for all historical or current examples it could find of independent agencies headed by a single person removable only for cause. The CFPB found only three examples: the Social Security Administration, the Office of Special Counsel, and the Federal Housing Finance Agency. Tr. of Oral Arg. at 19. But none of the three examples has deep historical roots. Indeed, the Federal Housing Finance Agency was created only in 2008, about the same time as the CFPB. The other two are likewise relatively recent. And those other two have been constitutionally contested by the Executive Branch, and they do not exercise the core Article II executive power of bringing law enforcement actions or imposing fines and penalties against private citizens for violation of statutes or agency rules. For those reasons, as we will explain, the three examples are different in kind from the CFPB and other independent agencies such as the FCC, the SEC, and FERC. Those examples therefore do not count for much when weighed against the deeply rooted historical practice demonstrating that independent agencies are multi-member agencies. To borrow the words of Justice Breyer in Noel Canning, as compared to the settled historical practice, “we regard these few scattered examples as anomalies.” NLRB v. Noel Canning, — U.S.-, 134 S.Ct. 2550, 2567, 189 L.Ed.2d 538 (2014); see also Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 505-06, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010).
First, the CFPB cited and primarily relied on the example of the Social Security Administration, which is an independent agency headed by a single Social Security Commissioner. See 42 U.S.C. §§ 901(a), 902(a). But the current structure of the agency is relatively recent. The Social Security Administration long existed first as a multi-member independent agency and then as a single-Director executive agency within various executive departments, most recently the Department of Health and Human Services. Only in 1994 did Congress change the Social Security Administration to a single-Director indepen*19dent agency. Importantly, when the agency structure was altered in 1994, President Clinton issued a signing statement expressing his view that the change in the agency’s structure was constitutionally problematic. See President William J. Clinton, Statement on Signing the Social Security Independence and Program Improvements Act of 1994, 2 Pub. Papers 1471, 1472 (Aug. 15, 1994). The status of that agency’s structure therefore is constitutionally contested. In those circumstances, the historical precedent counts for little because it is not settled. Cf. Noel Canning, 134 S.Ct. at 2563-64, 2567 (discounting prior example of appointments during “fictitious” inter-session recess because of Senate Committee’s strong opposition to those appointments); INS v. Chadha, 462 U.S. 919, 942 n.13, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983) (discounting prior statutory legislative veto provisions because Presidents had objected to those provisions). If anything, when considered against the “settled practice,” the Social Security example only highlights the “ano-mal[y]” of an independent agency headed by a single person. Noel Canning, 134 S.Ct. at 2567.
Moreover, the Social Security Administration is not a precedent for the CFPB because the Social Security Commissioner does not possess unilateral authority to bring law enforcement actions against private citizens, which is the core of the executive power and the primary threat to individual liberty posed by executive power. See Morrison v. Olson, 487 U.S. 654, 706, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988) (Scalia, J., dissenting). The Social Security Administration does not have unilateral power to impose fines or penalties on private citizens in Social Security benefits cases. Instead, the bulk of the Social Security Administration’s authority involves supervision of the adjudication of private claims for benefits. Although the agency does possess limited power to seek civil sanctions against those who file improper claims, the Commissioner may initiate such a proceeding “only as authorized by the Attorney General”—who is an executive officer accountable to the President. 42 U.S.C. § 1320a-8(b).
Second, the CFPB also cited the example of the Office of Special Counsel, an independent agency headed by a single Special Counsel. The Office has a narrow jurisdiction and mainly enforces certain personnel rules against government employers and employees, such as the prohibition against improper political activity by government employees. Like the Social Security Administration, the Office of Special Counsel lacks deep historical roots. Its single-Director structure was established in 1978. Also like the Social Security Administration, the constitutionality of the Special Counsel has been contested since its creation. Under President Carter, the Department of Justice opined that the Special Counsel “must be removable at will by the President” and expressed opposition to a for-cause removal restriction for the Special Counsel. Memorandum Opinion for the General Counsel, Civil Service Commission, 2 Op. O.L.C. 120,120 (1978). When Congress passed subsequent legislation regarding the Office of Special Counsel, President Reagan vetoed the Ml due to “serious constitutional concerns” about the Office’s status as an independent agency. See President Ronald Reagan, Memorandum of Disapproval on a Bill Concerning Whistleblower Protection, 2 Pub. Papers 1391, 1392 (Oct. 26, 1988). The history of the Office of Special Counsel confirms what one Special Counsel himself has acknowledged: the agency is “a controversial anomaly in the federal system.” K. William O’Connor, Foreword to ShIGEKI J. SUGIYAMA, PROTECTING the INTEGRITY OF THE MERIT SYSTEM: A LEGISLATIVE History of the Merit System Princi*20PLES, PROHIBITED PERSONNEL PRACTICES AND the Office of the Special Counsel, at v (1985). The status of the agency remains constitutionally contested and does not supply a persuasive, historical precedent for the CFPB’s structure. Cf. Noel Canning, 134 S.Ct. at 2563-64, 2567; Chadha, 462 U.S. at 942 n.13, 103 S.Ct. 2764.
Moreover, the Office of Special Counsel is not a precedent for the CFPB because the Office of Special Counsel is primarily responsible for enforcing personnel laws against government agencies and government employees. Unlike the CFPB, the Office of Special Counsel does not have authority to enforce laws against private citizens, and does not have power to impose fines and penalties on private citizens.5
Third, the CFPB cited Congress’s 2008 creation of a single head of the new Federal Housing Finance Agency. See Housing and Economic Recovery Act of 2008, Pub. L. No. 110-289, § 1101, 122 Stat. 2654, 2662 (codified at 12 U.S.C. §§ 4511-4512). That agency is a contemporary of the CFPB and merely raises the same question we confront here. A body created only in 2008 obviously does not constitute a historical precedent for the CFPB,
Although not a regulatory agency precedent and not an example cited by the CFPB as precedent for its single-Director structure (for good reason), there is at least one other modern example of an independent entity headed by one person. It is the now-defunct independent counsel law that was upheld in Morrison v. Olson, 487 U.S. 654, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988). But that decision did not expressly consider whether an independent agency could be headed by a single director. The independent counsel, moreover, had only a limited jurisdiction for particular defined investigations. Id. at 671-72, 108 S.Ct. 2597. In addition, the independent counsel experiment ended with nearly universal consensus that the experiment had been a mistake and that Justice Scalia had been right back in 1988 to view the independent counsel system as an unconstitutional departure from historical practice and a serious threat to individual liberty. See id. at 699, 108 S.Ct. 2597 (Scalia, J., dissenting) (“this wolf comes as a wolf’); see also Stanford Lawyer, Spring 2015, at 4 (quoting Justice Kagan’s statement that Justice Scalia’s dissent in Morrison is “one of the greatest dissents ever written and every year it gets better”). The independent counsel experience, if anything, strongly counsels caution with respect to single-Director independent agencies.6
So that’s all the CFPB has, and that’s not much. As Justice Breyer stated when *21facing a similar (actually, a more robust) historical record in Noel Canning, the few examples offered by the CFPB are “anomalies.” 134 S.Ct. at 2567. Or as the Court put it in Free Enterprise Fund when confronting a novel structure, a “handful of isolated” examples does not count for much when assessed against an otherwise settled historical practice. 561 U.S. at 505, 130 S.Ct. 3138. To be sure, in “all the laws enacted since 1789, it is always possible that Congress” created some other independent agencies like the CFPB “that exercise[ ] traditional executive functions” but are headed by single Directors. Free Enterprise Fund v. Public Company Accounting Oversight Board, 537 F.3d 667, 699 n.8 (D.C. Cir. 2008) (Kavanaugh, J,, dissenting); see also Noel Canning, 134 S.Ct. at 2567 (“There may be others of which we are unaware.”). But “the research of the parties and the Court has not found such a needle in the haystack.” Free Enterprise Fund, 537 F.3d at 699 n.8 (Kavanaugh, J., dissenting). “Even if such an example were uncovered,” there is no question that this kind of single-Director independent agency “has been rare at best.” Id.
The bottom line is that there is no settled historical, practice of independent agencies headed by single Directors who possess the substantial executive authority that the Director of the CFPB enjoys. The CFPB is exceptional in our constitutional structure and unprecedented in our constitutional history. See Who’s Watching the Watchmen ? Oversight of the Consumer Financial Protection Bureau: Hearing Before the Subcomm. on TARP, Financial Services and Bailouts of Public and Private Programs of the H. Comm, on Oversight and Government Reform, 112th Cong. 77 (2011) (statement of Andrew Pincus) (“Dodd-Frank sets up for the Bureau an unprecedented structure that consolidates more power in the director than in the head of any other agency that regulates private individuals and entities.”); Recent Legislation, Dodd-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. 2123, 2130 (2011) (“[T]he CFPB’s design is troubling because of its unprecedented nature.”); Note, Independence, Congressional Weakness, and the Importance of Appointment: The Impact of Combining Budgetary Autonomy with Removal.Protection, 125 Harv. L. Rev. 1822, 1824 n.15 (2012) (CFPB’s lack of a multi-member board is “atypical for independent agencies and will amplify the Director’s independence”); Todd Zywicki, The Consumer Financial Protection Bureau: Savior or Menace? 81 Geo. Wash. L. Rev. 856, 899 (2013) (“[T]he agency structure Congress chose for the CFPB—a single-director structure, devoid of accountability, and with vast, ill-defined powers— appears to be unique in recent American history.”).7
C
The CFPB’s departure from historical practice matters. A long line of Supreme Court precedent tells us that history and tradition are important guides in separation of powers cases that, like this one, are not resolved by the constitutional text alone. As Justice Breyer wrote for the *22Supreme Court in Noel Canning, the “longstanding practice of the government can inform our determination of what the law is.” NLRB v. Noel Canning, — U.S. -, 134 S.Ct. 2550, 2560, 189 L.Ed.2d 538 (2014) (internal quotation marks and citation omitted). Justice Breyer quoted James Madison’s statement that it was “foreseen at the birth of the Constitution, that difficulties and differences of opinion might occasionally arise in expounding terms & phrases necessarily used in such a charter ... and that it might require a regular course of practice to liquidate & settle the meaning of some of them.” Id. (internal quotation marks omitted). Justice Breyer explained, moreover, that the Court “has treated practice as an important interpretive factor even when the nature or longevity of that practice is subject to dispute, and even when that practice began after the founding era.” Id.
All of this, Justice Breyer stated, is “neither new nor controversial.” Id. Consider the following:
• “In separation-of-powers cases this Court has often put significant weight upon historical practice.” Zivotofsky v. Kerry, — U.S.-, 135 S.Ct. 2076, 2091, 192 L.Ed.2d 83 (2015) (internal quotation marks omitted) (quoting Noel Canning, 134 S.Ct. at 2559).
• “We therefore conclude, in light of historical practice, that a recess of more than 3 days but less than 10 days is presumptively too short to fall within the Clause.” Noel Canning, 134 S.Ct. at 2567.
• “Perhaps the most telling indication of the severe constitutional problem with the [agency] is the lack of historical precedent for this entity.” Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 505, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (internal quotation marks omitted).
• “[W]hen we face difficult questions of the Constitution’s structural requirements, longstanding customs and practices can make a difference.” Commonwealth of Puerto Rico v. Sanchez Valle, — U.S. -, 136 S.Ct. 1863, 1884, 195 L.Ed.2d 179 (2016) (Breyer, J., dissenting).
• “[Traditional ways of conducting government give meaning to the Constitution.” Mistretta v. United States, 488 U.S. 361, 401, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989) (internal quotation marks and alteration omitted) (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 610, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (Frankfurter, J., concurring)).
• “Deeply embedded traditional ways of conducting government cannot supplant the Constitution or legislation, but they give meaning to the words of a text or supply them.” Youngstown, 343 U.S. at 610, 72 S.Ct. 863 (Frankfurter, J., concurring).
• “A legislative practice such as we have here, evidenced not by only occasional instances, but marked by the movement of a steady stream for a century and a half of time, goes a long way in the direction of proving the presence of unassailable ground for the constitutionality of the practice, to be found in the origin and history of the power involved, or in its nature, or in both combined.” United States v. Curtiss-Wright Export Corp., 299 U.S. 304, 327-28, 57 S.Ct. 216, 81 L.Ed. 255 (1936).
• “Long settled and established practice is a consideration of great weight in a proper interpretation of constitutional provisions of this character.” The *23Pocket Veto Case, 279 U.S. 655, 689, 49 S.Ct. 463, 73 L.Ed. 894 (1929).
• “Such long practice under the pardoning power and acquiescence in it strongly sustains the construction it is based on.” Ex parte Grossman, 267 U.S. 87, 118-19, 45 S.Ct. 332, 69 L.Ed. 527 (1925).
• “[A] page of history is worth a volume of logic.” New York Trust Co. v. Eisner, 256 U.S. 345, 349, 41 S.Ct. 506, 65 L.Ed. 963 (1921).
• “[I]n determining the meaning of a statute or the existence of a power, weight shall be given to the usage itself—even when the validity of the practice is the subject of investigation.” United States v. Midwest Oil Co., 236 U.S. 459, 473, 35 S.Ct. 309, 59 L.Ed. 673 (1915).
• “[WJhere there is ambiguity or doubt [in the words of the Constitution], or where two views may well be entertained, contemporaneous and subsequent practical construction are enti-tied to the greatest weight.” McPherson v. Blacker, 146 U.S. 1, 27, 13 S.Ct. 3, 36 L.Ed. 869 (1892).
• “[A] doubtful question, one on which human reason may pause, and the human judgment be suspended, in the decision of which the great principles of liberty are not concerned, but the respective powers of those who are equally the representatives of the people, are to be adjusted; if not put at rest by the practice of the government, ought to receive a considerable impression from that practice.” McCulloch v. Maryland, 17 U.S. 4 Wheat. 316, 401, 4 L.Ed. 579 (1819).
Stated simply, in separation of powers cases not resolved by the constitutional text alone, historical practice matters a great deal in defining the constitutional limits on the Executive and Legislative Branches.8 The Supreme Court’s recent decisions in Noel Canning and Free Enterprise Fund illustrate how *24the Court considers historical practice in this context.9
In Noel Canning, the Supreme Court speaking through Justice Breyer stressed the importance of history when assessing the constitutionality of a novel practice—in that case, Presidential recess appointments in Senate recesses of fewer than 10 days. The Court said: “Long settled and established practice is a consideration of great weight in a proper interpretation of constitutional provisions regulating the relationship between Congress and the President.” Noel Canning, 134 S.Ct. at 2559 (internal quotation marks and alteration omitted). Based on that history, the Supreme Court ruled that a Senate recess of “less than 10 days is presumptively too short” for constitutional purposes. Id. at 2567. Importantly, the text of the Constitution did not draw any such 10-day line. But the historical practice between the President and the Senate had settled on a 10-day line.
In ruling out recess appointments in recesses of fewer than 10 days, the Noel Canning Court stated that it had “not found a single example of a recess appointment made during an intra-session recess that was shorter than 10 days.” Id. at 2566. The Court explained that the “lack of examples suggests that the recess-appointment power is not needed in that context.” Id. Although the Court did find “a few historical examples of recess appointments made during inter-session recesses shorter than 10 days,” the Court stated: “But when considered against 200 years of settled practice, we regard these few scattered examples as anomalies.” Id. at 2567.
According to the Court, therefore, allowing recess appointments in Senate recesses of fewer than 10 days would depart from the settled historical practice and alter the relative powers of the President and Senate over appointments. So, too, disallowing recess appointments in Senate recesses of 10 or more days would depart from settled historical practice. In Noel Canning, the Supreme Court therefore converted that historical 10-day practice into a constitutional rule.10
The Supreme Court engaged in the same kind of history-based analysis in Free Enterprise Fund. Independent agency heads are ordinarily removable for *25cause by the President. In that case, however, the new Accounting Oversight Board’s members were removable only for cause by the Commissioners of the SEC, and the. SEC Commissioners in turn were understood to be removable only for cause by the President. In other words, there were two levels of for-cause removal between the President and the Accounting Oversight Board.
The Supreme Court drew a line between one level of for-cause removal, which was the structure of traditional independent agencies, and two levels of for-cause removal, the novel structure of the new Accounting Oversight Board. See Free Enterprise Fund, 561 U.S. at 484, 130 S.Ct. 3138. The Court ruled that the latter was unconstitutional. The Court drew that line in part because historical practice had settled on one level of for-cause removal for a President to remove the head of an independent agency. There were at most “only a handful of isolated” precedents for the new Board. Id. at 505, 130 S.Ct. 3138. The vast majority of the extant independent agencies had only one level of for-cause removal. And as the Court noted, there was a meaningful difference between one level of for-cause removal and two levels of for-cause removal in terms of an agency’s insulation from Presidential control. See id. at 495-96, 130 S.Ct. 3138. Therefore, the Court invalidated the structure of the new Board.11
Those two cases well illustrate the broader jurisprudential principle long applied by the Supreme Court: In separation of powers cases not resolved by the constitutional text alone, historical practice matters.
D
The CFPB marks a major departure from the settled historical practice requiring multi-member bodies at the helm of independent agencies. Because this case is not resolved solely by the constitutional text, at least as the text was interpreted in Humphrey’s Executor, the CFPB’s departure from historical practice matters to the analysis. And the departure from historical practice matters even more in this instance because this departure from historical practice threatens individual liberty. The historical practice of structuring independent agencies as multi-member commissions or boards is the historical practice for a reason: It reflects a deep and abiding concern for safeguarding the individual liberty protected by the Constitution.
“The Framers recognized that, in the long term, structural protections against abuse of power were critical to preserving liberty.” Bowsher v. Synar, 478 U.S. 714, 730, 106 S.Ct. 3181, 92 L.Ed.2d 583 (1986); see also id. at 721, 106 S.Ct. 3181 (“The declared purpose of separating and dividing the powers of government, of course, was to ‘diffus[e] power the better to secure liberty.’”) (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 679, 635, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (Jackson, J., concurring)). When describing Article II, Justice Scalia put the point this way: “The purpose of the separation and equilibration of powers in general, and of the unitary Executive in particular, was not merely to assure effective government but to preserve individual freedom.” Morrison v. Olson, 487 U.S, 654, 727, 108 S.Ct. 2597, *26101 L.Ed.2d 569 (1988) (Scalia, J., dissenting).
The basic constitutional concern with independent agencies is that the agencies are unchecked by the President, the official who is accountable to the people and who is made responsible by Article II for the exercise of executive power. Recognizing the broad and unaccountable power wielded by independent agencies, Congress has traditionally required multi-member bodies at the helm of independent agencies. In the absence of Presidential control, the multi-member structure of independent agencies acts as a critical substitute check on the excesses of any individual independent agency head—a check that helps to prevent arbitrary decision-making and abuse of power, and thereby to protect individual liberty.
But this new agency, the CFPB, lacks that critical check and structural constitutional protection. And the lack of the traditional safeguard threatens the individual liberty protected by the Constitution’s separation of powers.
How do multi-member independent agencies fare better than single-Director independent agencies in protecting individual liberty? As compared to single-Director independent agencies, multi-mem-ber independent agencies help prevent arbitrary decisionmaking and abuses of power, and thereby help protect individual liberty, because they do not concentrate power in the hands of one individual. The point is simple but profound. In a multi-member independent agency, no single commissioner or board member possesses authority to do much of anything. Before the agency can infringe your liberty in some way—for example, initiating an enforcement action against you or issuing a rule that affects your liberty or property—a majority of commissioners must agree. That in turn makes it harder for the agency to infringe your liberty. As the current Chair of the Federal Trade Commission has explained, it takes “a consensus decision of at least a majority of commissioners to authorize, or forbear from, action.” Edith Ramirez, The FTC: A Framework for Promoting Competition and Protecting Consumers, 83 Geo. Wash. L. Rev. 2049, 2053 (2015). In a multi-member agency, even though each individual commissioner is not accountable to or checked by the President, each commissioner is at least still accountable to his or her fellow commissioners and needs the assent of a majority of commissioners to take significant action.
In addition, unlike single-Director independent agencies, multi-member independent agencies “can foster more deliberative decision making.” Kirti Datla & Richard L. Revesz, Deconstructing Independent Agencies (and Executive Agencies), 98 Cornell L. Rev. 769, 794 (2013). Relatedly, multi-member independent agencies benefit from diverse perspectives and different points of view among the commissioners and board members. The multiple voices and perspectives make it more likely that the costs and downsides of proposed decisions will be more fully ventilated. See Marshall J. Breger & Gary J. Edles, Established by Practice: The Theory and Operation of Independent Federal Agencies, 52 Admin. L. Rev. 1111, 1113 (2000) (independent agencies “are also multi-member organizations, a fact that tends toward accommodation of diverse or extreme views through the compromise inherent in the process of collegial decisionmaking”); Jacob E. Gersen, Administrative Law Goes to Wall Street: The New Administrative Process, 65 Admin. L. Rev. 689, 696 (2013) (“[A] multimember board allows for a representation of divergent inter*27ests in a way that a single decisionmaker simply cannot.”); Glen 0. Robinson, On Reorganizing the Independent Regulatory Agencies, 57 Va. L. Rev. 947, 963 (1971) (“It is not bipartisanship as such that is important; it is rather the safeguards and balanced viewpoint that can be provided by plural membership.”); cf. ■Harry T. Edwards, The Effects of Collegiality on Judicial Decision Making, 151 U. Pa. L. Rev. 1639, 1645 (2003) (“[C]ol-legiality plays an important part in mitigating the role of partisan polities and personal ideology by allowing judges of differing perspectives and philosophies to communicate with, listen to, and ultimately influence one another in constructive and law-abiding ways.”). '
In short, the deliberative process and multiple viewpoints in a multi-member independent agency can help ensure that an agency does not wrongly bring an enforcement action or adopt rules that unduly infringe individual liberty.
As compared to a single-Director structure, a multi-member independent agency also helps to avoid arbitrary decisionmak-ing and to protect individual liberty because the multi-member structure—and its inherent requirement for compromise and consensus—will tend to lead to decisions that are not as extreme, idiosyncratic, or otherwise off the rails. Cf Stephen M. Bainbridge, Why a Board? Group Deci-sionmaking in Corporate Governance, 55 Vand. L. Rev. 1, 12-19 (2002) (summarizing experimental evidence finding group decisionmaking to be superior to individual decisionmaking). A multi-member independent agency can only go as far as the middle vote is willing to go. Conversely, under a single-Director structure, an agency’s policy goals “will be subject to the whims and idiosyncratic views of a single individual.” Joshua D. Wright, The Antitrust/Consumer Protection Paradox: Two Policies at War with Each Other, 121 Yale L.J. 2216, 2260 (2012) (internal quotation marks omitted); cf Recent Legislation, Doddr-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. 2123, 2128 (2011) (multi-member commission structure “reduces the variance of policy and improves accuracy through aggregation”); Michael B. Rappa-port, Essay, Replacing Independent Counsels with Congressional Investigations, 148 U. Pa. L. Rev. 1595, 1601 n.17 (2000) (“[Ijndependent agencies tend to be headed by multimember commissions, which function to prevent aberrant actions_”).
Relatedly, as compared to a single-Director independent agency, a multi-mem-ber independent agency provides the added benefit of “a built-in monitoring system for interests on both sides because that type of body is more likely to produce a dissent if the agency goes too far in one direction.” Rachel E. Barkow, Insulating Agencies: Avoiding Capture Through Institutional Design, 89 Tex. L. Rev. 15, 41 (2010). A dissent, in turn, can serve “as a ‘fire alarm’ that alerts Congress and the public at large that the agency’s decision might merit closer scrutiny.” Id.-, see also Dodct-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. at 2128 (the “presence of dissenters” in agency proceedings “provides new information and forces the proponent to articulate a coherent rationale, thus acting as a constraining force”).
Moreover, multi-member independent agencies are better structured than single-Director independent agencies to guard against “capture” of—that is, undue influence over—the independent agencies -by regulated entities or interest groups, for example. As then-Professor Elizabeth Warren noted in her original proposal for a multi-member consumer protection agency: ‘With every agency, the fear of regula*28tory capture is ever-present.” Elizabeth Warren, Unsafe at Any Rate: If It’s Good Enough for Microwaves, It’s Good Enough for Mortgages. Why We Need a Financial Product Safety Commission, Democracy, Summer 2007, at 8, 18. Capture can infringe individual liberty because capture can prevent a neutral, impartial agency assessment of what rules to issue and what enforcement actions to undertake. In a multi-member agency, however, the capturing parties “must capture a majority of the membership rather than just one individual.” Lisa Schultz Bressman & Robert B. Thompson, The Future of Agency Independence, 63 Vand. L. Rev. 599, 611 (2010); see also RobeRt E. Cushman, The Independent REGULATORY COMMISSIONS 153 (Octagon Books 1972) (1941) (noting, in reference to Federal Reserve Act of 1913, that it “seemed easier to protect a board from political control than to protect a single appointed official”); Barkow, Insulating Agencies, 89 Tex. L. Rev. at 38 (“[Ojnly one person at the apex can also mean that the agency is more easily captured.”); Robinson, On Reorganizing the Independent Regulatory Agencies, 57 Va. L. Rev. at 962 (“[T]he single administrator may be more vulnerable” to interest group pressures “because he- provides a sharper focus for the concentration of special interest power and influence.”).
Importantly, all of those features and benefits of multi-member independent agencies are not merely accidental or coincidental byproducts. Those points were in the minds of the Members of Congress who helped launch independent agencies. For example, Senator Newlands, the sponsor of the legislation creating the Federal Trade -Commission, emphasized the need for a commission rather than a single Director: “If only powers of investigation and publicity are given[,] a single-headed organization, like the Bureau of Corporations, might be the best for the work; but if judgment and discretion are to be exercised, or if we have in contemplation the exercise of any corrective power hereafter, or if the broad ends above outlined are to be attained, it seems to me that a commission is required.” 51 Cong. Rec. 11,092 (1914). In his leading study of independent commissions, Robert Cushman, former staff member of President Franklin Roosevelt’s Committee on Administrative Management, analyzed the creation of the Federal Trade Commission and explained: “The two ideas, a commission and independence for the commission, were inextricably bound together. At no point was it proposed that a commission ought to be set up unless it be independent or that an independent officer should be created rather than a commission.” Cushman, The Independent Regulatory Commissions, at 188; see also The President’s Committee on Administrative Management, Report of the Committee with Studies of Administrative Management in the Federal Government 216 (1937) (noting “popular belief that important rule-making functions ought to be performed by a group rather than by a single officer, by a commission rather than by a department head” as one reason “for the establishment of independent regulatory agencies”).
Examining the consistent historical practice here, we can see, moreover, that the consistent historical practice reflects the deep values of the Constitution. The Constitution as a whole embodies the bedrock principle that dividing , power among multiple entities and persons helps protect individual liberty. The Framers created a federal system with the national power divided among three branches. The Framers “viewed the principle of separation of powers as the absolutely central guarantee of a just Government.” Morrison v. Olson, 487 U.S. at 697, 108 S.Ct. 2597 (Scalia, J., dissenting).
*29And to protect liberty, the same kind of checks and balances principle also influenced how the Framers allocated power within the three national branches. For example, the Framers divided the Legislative Branch into two houses, each with multiple members. No one person operates as the Legislator-in-Chief. Rather, 535 Members ■ of Congress do so, divided among two Houses. Likewise, the Framers established “one supreme Court” composed of multiple “Judges” rather than a single judge. No one person operates as the lone Justice of the Supreme Court. Rather, the Court consists of one Chief Justice and several Associate Justices, all of whom have equal votes on cases. “Even a cursory examination of the Constitution reveals the influence of Montesquieu’s thesis that checks and balances were the foundation of a structure of, government that would protect liberty.” Bowsher, 478 U.S. at 722, 106 S.Ct. 3181.
Of course, the one exception to the pon-stitution’s division of power among multiple parties within the branches is the President, who is the lone head of the entire Executive Branch. But the President is the exception that proves the rule. For starters, the Framers were concerned that dividing the executive power among multiple individuals would render the Executive Branch too weak as compared to the more formidable Legislative Branch. See The Fedeealist No. 48, at 309-10 (James Madison) (Clinton Rossiter ed., 1961) (“[I]t is against the enterprising ambition” of the Legislative Branch “that the people ought to indulge all their jealousy and exhaust all their precautions. The legislative department derives a superiority in our governments .... ”). The Framers sought “energy in the executive.” The Federalist No. 70, at 424 (Alexander .Hamilton).
At the same time, the Framers certainly recognized the risk that a single President could lead to tyranny or arbitrary decision-making. To mitigate the risk to liberty from a single President, the Framers ensured that the President had “a due dependence on the people.” Id. The President is nationally elected by the people. In choosing the President, “the whole Nation has a part, making him the focus of public hopes and expectations.” Youngstown, 343 U.S. at 653, 72 S.Ct. 863 (Jackson, J., concurring). Presidential candidates are put through the wringer precisely because of the power they may someday wield. In other words, the Framers concentrated executive power in a single President on the condition that the President would be nationally elected and nationally accountable.
The President is therefore the exception to the ordinary constitutional practice of dividing power among multiple entities and persons. Apart from the President, the Constitution reflects the basic commonsense principle that multi-merfiber bodies—the House, the Senate, the Supreme Court—do better than single-member bodies in avoiding arbitrary decisionmaking and abuses of power, and thereby protecting individual liberty. That background constitutional principle further supports the conclusion here that a single-Director independent agency lies outside the norm and poses a risk to individual liberty. After all, the Director of the CFPB is not elected by the people and is of course not remotely comparable to the President in terms of accountability to the people.
Having identified the ways in which mul-ti-member independent agencies surpass single-Director independent agencies in protecting liberty, we must acknowledge that multi-member. independent agencies do not always meet that potential. For example, some members of multi-member independent agencies may occasionally move in lockstep, thereby diminishing the benefits of multi-member bodies. It can be *30harder to find five highly qualified commissioners than just one highly qualified commissioner. Moreover, multi-member bodies are often not as efficient as single-headed agencies and'can be beset by contentious relations among the members. See Breger & Edles, Established by Practice, 52 Admin. L. Rev. at 1181 (“even a single member” can throw a wrench into the works); Datla & Revesz, Deconstructing Independent Agencies, 98 Cornell L. Rev. at 794 (“The downside that accompanies increased deliberation is the slowness inherent in group action.”) (internal quotation marks omitted). That said, “[cjonvenience and efficiency are not the primary- objectives—or the hallmarks—of democratic government.” Bowsher, 478 U.S. at 736, 106 S.Ct. 3181 (internal quotation marks omitted). Indeed, so as to avoid falling back into the kind of tyranny that they had declared independence from, the Framers often made, trade-offs against efficiency in the interest of enhancing liberty.
In any event, notwithstanding some failings and downsides, multi-member independent agencies are superior to single-Director independent agencies in preventing arbitrary decisionmaking and abuse of power, and thereby protecting individual liberty.
For that reason and others, both before and after Humphrey’s Executor, Congress has structured independent agencies as multi-member agencies. Indeed, the multi-member agency form has become “synonymous with independence.” Breger & Edles, Established by Practice, 52 Admin. L. Rev. at 1137. As Justice Breyer noted in Free Enterprise Fund: “Agency independence is a function of several different factors ... including] .A its composition as a multimember bipartisan board Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 547 (2010) (Breyer, J., dissenting). Likewise, Professor Barkow has explained that “multimember design” is one of the “[t]ra-ditional [ljodestars” of agency independence. Barkow, Insulating Agencies, 89 Tex. L. Rev. at 26; see also Peter L. StRauss, An Introduction to Administrative Justice in the United States 15 (1989) (defining “independent regulatory commission[s]” as “governmental agencies headed by multi-member boards acting collegially on the regulatory matters within their jurisdiction”) (internal quotation marks omitted); Bressman & Thompson, The Future of Agency Independence, 63 Vand. L. Rev. at 610 (independent agencies, unlike Executive Branch agencies, are “generally run by multi-member commissions or boards”); Dodd-Frank Act Creates the Consumer Financial Protection Bureau, 124 Harv. L. Rev. at 2128 (“Most independent agencies have multimember boards.... ”); Paperwork Reduction Act of 1980, Pub. L. No. 96-511, 94 Stat. 2812, 2814 (defining “independent regulatory agency” by reference to 17 multi-member agencies) (internal quotation marks omitted).
E
To sum up so far: In order to preserve individual liberty and ensure accountability, Article II of the Constitution assigns the executive power to the President. The President operates with the assistance of subordinates, but the President acts as a critical check on those subordinates. That check provides accountability and protects against arbitrary decisionmaking by executive agencies, thereby helping to safeguard individual liberty. Article II has been interpreted by the Supreme Court to allow independent agencies in certain circumstances. Independent agencies lack the ordinary constitutional checks and balances that come from Presidential supervision and direction. But to ensure some check against arbitrary decisionmaking and to *31help preserve individual liberty, independent agencies have traditionally been structured as multi-member bodies where the commissioners or board members can check one another. The check from other commissioners or board members substitutes for the check by the President. As an independent agency with just a single Director, the CFPB represents a sharp break from historical practice, lacks the critical internal check on arbitrary deci-sionmaking, and poses a far greater threat to individual liberty than does a multi-member independent agency. All of that raises grave constitutional doubts about the CFPB’s single-Director structure.12
Before rendering a final conclusion on the CFPB’s constitutionality as currently structured, however, we must address several other arguments.
First, in considering precedents for the single-Director structure of the CFPB, one might wonder about all of the executive departments and agencies headed by a single person. Why don’t they provide a precedent for the CFPB? Consider for example the Department of Justice, the Department of the Treasury, the Department of State, the Department of Defense, and the EPA, all headed by a single person.
As should be clear by now, the distinction, of course, is that those departments and agencies are executive agencies. They operate within the Executive Branch chain of command under the supervision and direction of the President, and those agency heads are removable at will by the President. The President is a check on those agencies. Those agencies are accountable to the President. The President in turn is accountable to the people of the United States for the exercise of executive power in the executive agencies. So a single person at the helm of an executive agency is perfectly constitutional.13
By contrast, independent agencies are unaccountable to the President and pose a greater threat to individual liberty because they operate free of the President’s supervision and direction. Therefore, they traditionally have been headed by multiple members who check one another. An independent agency operates as “a body of experts appointed by law and informed by experience.” Humphrey’s Executor v. United States, 295 U.S. 602, 624, 55 S.Ct. 869, 79 L.Ed. 1611 (1935) (internal quotation marks omitted).
Second, some may say that Congress’s creation of the single-Director structure is unlikely to give Congress any greater influence over the CFPB than Congress possesses over a multi-member independent agency. That is perhaps true, although perhaps not. Either way, however, the Supreme Court has emphasized that congressional aggrandizement is not a necessary feature of a separation of powers violation in this context. The Court squarely said as much in Free Enterprise Fund. See Free Enterprise Fund v. Public Company Accounting Oversight Board, *32561 U.S. 477, 500, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (“Even when a branch does not arrogate power to itself, therefore, it must not impair another in the performance of its constitutional duties.”) (internal quotation marks omitted). And to take an obvious example of the point, if Congress enacted legislation converting the Department of Justice into an independent agency, there would be no formal congressional aggrandizement. But there is little doubt that such legislation would violate Article II. See Morrison v. Olson, 487 U.S. 654, 695, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988) (Congress may not impair the President in performance of constitutionally assigned functions). Congressional aggrandizement is not a necessary condition for an Article II violation in this context.
Relatedly, one might think that a single head of an independent agency might actually be more responsive to the President than multiple heads of an independent agency are, thereby reducing the risk of arbitrary decisionmaking and mitigating the Article II concern with a novel single-Director independent agency. But there is no meaningful difference in responsiveness and accountability to the President. Whether headed by one, three, or five members, an independent agency is not supervised or directed by the President, and its heads are not removable at will by the President. With independent agencies, the President is limited in essence to indirect cajoling. Cf. Elena Kagan, Presidential Administration, 114 Harv. L. Rev. 2245, 2323 (2001) (“[A] for-cause removal provision would buy little substantive independence if the President, though unable to fire an official, could command or, if necessary, supplant his every decision.”). 14 As Justice Scalia once memorably noted, an attempt by the President to direct (or threaten to remove) the head of an independent agency with respect to a particular substantive decision is statutori*33ly impermissible and likely to trigger “an impeachment motion in Congress.” Tr. of Oral Arg. at 60, Free Enterprise Fund, 561 U.S. 477, 130 S.Ct. 3138, 177 L.Ed.2d 706. That is true whether there are one, three, or five heads of the independent agency. The' independent- status of an independent agency erects a high barrier between the President and the independent agency, regardless of how many people head the independent agency on the other side of the barrier. So a structure with a single independent agency head entails no meaningful benefit over a multi-member independent agency in terms of Presidential control over the independent agency.
Although the single-Director structure does not necessarily give more control to the President over an independent agency, one might say from the other direction that the structure at least does not diminish the President’s power beyond the diminishment already caused by Humphrey’s Executor, and thus should not form the basis of an Article ÍI violation. In other words, some might say that Humphrey’s Executor already greatly reduced Presidential power, and this novel- structure is merely a variation on Humphrey’s Executor rather than a further diminishment of Presidential power. To begin with, that may not be true. A President may be stuck for his or her entire four-year term with a single Director appointed by a prior President with different views. Generally, the members of multi-member agencies serve staggered terms, and the President will at least have an opportunity to appoint some new commissioners over the course of his or her first term.
In any event, although it is true that Article II violations often involve diminishment of Presidential power, neither Humphrey’s Executor nor any later case gave Congress a free pass, without any boundaries, to create independent agencies that depart from history and threaten individual liberty. Humphrey’s Executor does not mean that anything goes. See Free Enterprise Fund, 561 U.S. at 514, 130 S.Ct. 3138. In that respect, keep in mind (in case we have not mentioned it enough already) that the Constitution’s separation of powers is not solely or even primarily concerned with preserving the powers of the branches. The separation of powers is- primarily designed to protect individual liberty. See Stern v. Marshall, 564 U.S. 462, 483, 131 S.Ct. 2594, 180 L.Ed.2d 475 (2011) (“Yet the dynamic between and among the branches is not the only object of the Constitution’s concern. The structural principles secured by the separation of powers protect the individual as well.”) (internal quotation marks omitted) (quoting Bond v. United States, 564 U.S. 211, 222, 131 S.Ct. 2355, 180 L.Ed.2d 269 (2011)); Bowsher v. Synar, 478 U.S. 714, 721, 106 S.Ct. 3181, 92 L.Ed.2d 583 (1986) (“The declared purpose of separating and dividing the powers of government, of course, was to ‘diffus[e] power the better to secure liberty.’ ”) (quoting Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 635, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (Jackson, J., concurring)); Clinton v. City of New York, 524 U.S. 417, 450, 118 S.Ct. 2091, 141 L.Ed.2d 393 (1998) (Kennedy, J., concurring) (“Liberty is always at stake when one or more of the branches seek to transgress the separation of powers.”). As with the broader separation of powers, moreover, a key purpose of Article II is to preserve individual liberty. See Morrison v. Olson, 487 U.S. at 727, 108 S.Ct. 2597 (Scalia, J., dissenting) (“The purpose of the separation and equilibration of powers in gendral, and of the unitary Executive in particular, was not merely to assure effective government but to preserve individual freedom.”),
So the single-Director independent agency—-which, as we have explained, is a *34structure that departs from settled historical practice and threatens individual liberty far more than a multi-member independent agency does—poses a constitutional problem even if it does not occasion any additional diminishment of Presidential power beyond the significant diminishment already caused by Humphrey’s Executor itself.15
Third, in considering the constitutionality of the CFPB’s structure, some might speak of the CFPB as a one-off congressional experiment and say we should let it go as a matter of judicial restraint. But even apart from the fundamental point that our job as judges is to enforce the law, not abdicate to the political branches, cf. Boumediene v. Bush, 553 U.S. 723, 765-66, 128 S.Ct. 2229, 171 L.Ed.2d 41 (2008), we cannot think of this as a one-off case because we could not cabin the consequences in any principled manner if we were to uphold the CFPB’s single-Director structure. As the Supreme Court has warned: “Slight encroachments create new boundaries from which legions of power can seek new territory to capture.” Stern, 564 U.S. at 503, 131 S.Ct. 2594 (internal quotation marks omitted). Justice Frankfurter captured it well in his opinion in Youngstown: “The accretion of dangerous power does not come in a day. It does come, however slowly, from the generative force of unchecked disregard of the restrictions that fence in even the most disinterested assertion of authority.” 343 U.S. at 594, 72 S.Ct. 863 (Frankfurter, J., concurring). That fairly describes what a ruling upholding the CFPB’s single-Director structure would mean. As the CFPB acknowledged at oral argument, a ruling in its favor would necessarily allow all extant independent agencies to be headed by one person. Tr. of Oral Arg. at 18-19. We would be green-lighting Congress to make other heads of independent agencies a single Director rather than a multi-member commission. A single-Director SEC, with the power to unilaterally impose $500 million penalties? A single-Director FCC, *35with the power to unilaterally require “net neutrality”? A single-Director NLRB, with the power to unilaterally supervise employer-employee relations nationwide? That’s what we would be ushering in with a ruling upholding the CFPB’s single-Director structure.
At a more general level, however, some might think that judges should simply defer to the elected branches’ design of the administrative state. But that hands-off attitude would flout a long, long line of Supreme Court precedent. Agreement by the two elected branches at a particular moment or period in time has never been a ground for the courts to simply defer regardless of whether the legislation violates the Constitution’s separation of powers. Far from it. See Free Enterprise Fund, 561 U.S. at 497, 508, 130 S.Ct. 3138 (invalidating structure of Public Company Accounting Oversight Board); Boumediene, 553 U.S. at 765-66, 792, 128 S.Ct. 2229 (invalidating provision of Military Commissions Act); Clinton, 524 U.S. at 448-19, 118 S.Ct. 2091 (invalidating Line Item Veto Act); Metropolitan Washington Airports Authority v. Citizens for the Abatement of Aircraft Noise, Inc., 501 U.S. 252, 266-69, 111 S.Ct. 2298, 115 L.Ed.2d 236 (1991) (invalidating structure of Metropolitan Washington Airports Authority Board of Review); Bowsher, 478 U.S. at 733-34, 106 S.Ct. 3181 (invalidating Comptroller General’s powers under “reporting provisions” of Balanced Budget and Emergency Deficit Control Act); INS v. Chadha, 462 U.S. 919, 942 n.13, 957, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983) (invalidating legislative veto provision of Immigration and Nationality Act); Buckley v. Valeo, 424 U.S. 1, 134-35, 140, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) (invalidating structure of Federal Election Commission); Myers v. United States, 272 U.S. 52, 47 S.Ct. 21, 71 L.Ed. 160 (1926) (invalidating provision requiring Senate consent to President’s removal of executive officer). In that same vein, even though a particular President might accept a novel practice that violates Article II, “the separation of powers does not depend on the views of individual Presidents, nor on whether the encroached-upon branch approves the encroachment.” Free Enterprise Fund, 561 U.S. at 497, 130 S.Ct. 3138 (internal quotation marks and citation omitted). A President cannot “choose to bind his successors by diminishing their powers.” Id.
In this case, moreover, it bears mention that Congress’s choice of a single-Director CFPB was not an especially considered legislative decision. There are no committee reports, nor substantial legislative history, delving into the benefits of single-Director independent agencies versus mul-ti-member independent agencies. The CFPB has identified no congressional hearings studying the question. Congress apparently stumbled into this single-Director structure as a compromise or landing point between the original Warren multi-member independent agency proposal and a traditional executive agency headed by a single person.
Fourth, one might argue that the CFPB’s decisions are checked by the courts, so we should not worry about the single-Director structure. But .much of what an agency does—determining what rules to issue within a broad statutory authorization and when, how, and against whom to bring enforcement actions to enforce the law—occurs in the twilight of judicially unreviewable discretion. Those discretionary actions have a critical impact on individual liberty. And courts do not review or only deferentially review such exercises of agency discretion. See Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 844-45, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984); *36Motor Vehicle Manufacturers Association of U.S., Inc. v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 41-43, 103 S.Ct. 2856, 77 L.Ed.2d 443 (1983); Heckler v. Chaney, 470 U.S. 821, 831-33, 105 S.Ct 1649, 84 L.Ed.2d 714 (1985). Therefore, the probability of judicial review of some agency action has never excused or mitigated an otherwise extant Article II problem in the structure of the agency. See, e.g., Free Enterprise Fund, 561 U.S. 477, 130 S.Ct. 3138, 177 L.Ed.2d 706; Buckley, 424 U.S. 1, 96 S.Ct. 612, 46 L.Ed.2d 659.
From another direction, one might argue that the CFPB is checked by Congress through Congress’s oversight and ultimate control over appropriations. To begin with, Congress does not have the power to direct the Director or to remove the Director at will. Congress cannot supervise or direct the Director regarding what rules to issue, what enforcement actions to bring (or decline to bring), or how to resolve adjudications. More to the point, by further impairing the President’s control over the Executive Branch, day-to-day congressional control over an executive or independent agency generally would exacerbate, rather than mitigate, any Article II problem with the structure of the agency. To satisfy Article II, the check on an agency must come from the President or from other internal Executive Branch or agency checks, not from Congress. The bottom line, as the Supreme Court said in Bowsher, is that the “separated powers of our Government cannot be permitted to turn on judicial assessment of whether an officer exercising executive power is on good terfes with Congress.” 478 U.S. at 730, 106 S.Ct. 3181. “The Framers did not rest our liberties on such bureaucratic minutiae.” Free Enterprise Fund, 561 U.S. at 500, 130 S.Ct. 3138.16
In sum, the CFPB departs from settled historical practice regarding the structure of independent agencies. And that departure''makes a significant difference for the individual liberty protected by the Constitution’s separation of powers. Applying the Supreme Court’s separation of powers precedents, we therefore conclude that the CFPB is unconstitutionally structured because it is an independent agency headed by a single Director,17
*37III
Having concluded that the CFPB is unconstitutionally structured because' it is an independent agency headed by a single Director, we must decide on the appropriate remedy. When the constitutional problem involves a provision of a statute, the legal term for that question is severability. In light of this one specific constitutional flaw in the Dodd-Frank Act, must we strike down that whole Act? Or must we strike down at least those statutory provisions creating the CFPB and defining the CFPB’s duties and authorities? Or do we just narrowly strike down and sever the one for-cause removal provision that is the source of the constitutional problem?
Not surprisingly, PHH wants us, at a minimum, to strike down the CFPB and prevent its continued operation, if not strike down the entire Dodd-Frank Act. But Supreme Court precedent on sever-ability demands a narrower remedy for the CFPB’s constitutional flaw.
. “Generally . speaking, when confronting a constitutional flaw in a statute, we try to limit the solution to the problem, severing any problematic portions while leaving the remainder intact.” Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 508, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (internal quotation marks omitted). The “normal rule is that partial, rather than facial, invalidation is the required course.” Id. (internal quotation marks omitted). That is true so long as we conclude that (i) Congress would have preferred the law with the offending provision severed over no law at all; and (ii) the law with the offending provision severed would remain “fully operative as a law,” Id. at 509, 130 S.Ct, 3138 (internal quotation marks omitted).
First, in considering Congress’s intent with respect to severability, courts must decide—or often speculate, truth be told—whether Congress would “have preferred what is left of its statute to no statute at all.” Ayotte v. Planned Parenthood of Northern New England, 546 U.S. 320, 330, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006); see also Alaska Airlines, Inc. v. Brock, 480 U.S. 678, 685, 107 S.Ct. 1476, 94 L.Ed.2d 661 (1987) (“[T]he unconstitutional provision must be severed unless the statute created in its absence is legislation that Congress would not have enacted.”). Importantly, courts need not speculate and can presume that Congress wanted to retain the constitutional remainder of the statute when “Congress has explicitly provided for severance by including a severability clause in the statute.” Alaska Airlines, 480 U.S. at 686, 107 S.Ct. 1476; see also id. (“[T]he inclusion of such a clause creates a presumption that Congress did not intend the validity of the statute in question to depend on the validi*38ty of the constitutionally offensive provision”).
In this case, as was the case in Free Enterprise Fund, “nothing in the statute’s text or historical context makes it evident that Congress, faced with the limitations imposed by the Constitution, would have preferred no” CFPB at all (or no Dodd-Frank Act at all) to a CFPB whose Director is removable at will. 561 U.S. at 509, 130 S.Ct. 3138 (internal quotation marks omitted). Indeed, the Dodd-Frank Act itself all but answers the question of presumed congressional intent through its express severability clause, which instructs: “If any provision” of the Act “is held to be unconstitutional, the remainder of’ the Act “shall not be affected thereby.” 12 U.S.C. § 5302. It will- be the rare case when a court may ignore a .severability provision set forth in the text of the relevant statute. See Alaska Airlines, 480 U.S. at 686, 107 S.Ct. 1476. We have no reason or basis to tilt at that windmill in this case.
Second, we also must look at “the balance of the legislation” to assess whether the statute is capable “of functioning” without the offending provisions “in a manner consistent with the intent of Congress.” Id. at 684-85, 107 S.Ct. 1476 (emphasis omitted); see also United States v. Booker, 543 U.S. 220, 227, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005) (“[T]wo provisions ... must be invalidated in order to allow the statute- to operate in a manner consistent with congressional intent.”). That prong of the severability analysis in essence turns on whether the truncated statute is “fully operative as a law.” Free Enterprise Fund, 561 U.S. at 509, 130 S.Ct. 3138 (internal quotation marks omitted). To take just one example, in Mar-bury v. Madison, the Court concluded that Section 13 of the Judiciary Act of 1789 was unconstitutional in part. 5 U.S. 1 Cranch 137, 148, 179-80, 2 L.Ed. 60 (1803). But the Court did not disturb the remainder of the Judiciary Act. Id. at 179-80.
Here, the Dodd-Frank Act and its CFPB-related provisions will remain “fully operative as a law” without the for-cause removal restriction. Free Enterprise Fund, 561 U.S. at 509, 130 S.Ct. 3138 (internal quotation marks omitted). Operating without the for-cause removal provision and under the supervision and direction of the President, the CFPB may still “regulate the offering and provision of consumer financial products or services under the Federal consumer financial laws,” 12 U.S.C. § 5491(a), much as the Accounting Oversight Board has continued fulfilling its regulatory mission in the wake of the Supreme Court’s decision in Free Enterprise Fund.18 Moreover, the CFPB’s operation as an executive agency will not in any way prevent the overall Dodd-Frank Act from remaining operative as a law.
To be sure, one might ask whether, instead of severing the for-cause removal provision, we should rewrite and add to the Dodd-Frank Act by restructuring the CFPB as a multi-member independent *39agency. But doing so would require us to create a variety of new offices, designate one of the offices as Chair, and specify various administrative details of the reconstituted agency. All of that “editorial freedom” would take us far beyond our judicial capacity. Free Enterprise Fund, 561 U.S. at 510, 130 S.Ct. 3138. In addition, that approach would thwart the ongoing operations of the CFPB unless and until the President nominated and the Senate confirmed new members, potentially shutting the agency down for months if not years. No Supreme Court case in comparable circumstances has adopted such an approach. We may not do so here. Of course, if Congress prefers to restructure the CFPB as a multi-member independent agency rather than as a single-Director executive agency, Congress may enact new legislation that creates a Bureau headed by multiple members instead of a single Director. Cf. id. (“Congress of course remains free to pursue any of these options going forward.”).
In similar circumstances, the Supreme Court in Free Enterprise Fund severed the unconstitutional for-cause provision but did not otherwise disturb the Sar-banes-Oxley Act or the operation of the new Accounting' Oversight Board created by that Act; See id. at 508-10, 130 S.Ct. 3138. Similarly, in a recent cáse involving the Copyright Royalty Board, we severed the for-cause provision that rendered that Board unconstitutional, but did not otherwise disturb the copyright laws or the operation of the Copyright Royalty Board. See Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board, 684 F.3d 1332, 1340-41 (D.C. Cir. 2012). We do the same here.
In light of Congress’s clear textual expression of its intent regarding severability, and because the Dodd-Frank Act and the CFPB may function without the CFPB’s for-cause removal provision, we remedy the constitutional violation here by severing the for-cause removal provision from the statute. As a result, the CFPB now will operate as an executive agency. The President of the United States now has the power to supervise and direct the Director of the CFPB, and may remove the Director at will at any time.19
ÍV
Because our constitutional ruling will not halt the CFPB’s ongoing operations or the CFPB’s ability to uphold the $109 million order against PHH, we must also consider PHH’s statutory objections to the CFPB enforcement action in this case.
In its enforcement action against PHH, the CFPB alleged that PHH violated Section 8 of the Real Estate Settlement Pro*40cedures Act. Passed by Congress and signed by President Ford in 1974, the Act dramatically reformed the real estate industry. One of the textually stated purposes of the Act was “the elimination of kickbacks or referral fees that tend to increase unnecessarily the costs of certain settlement services.” 12 U.S.C. § 2601(b)(2).
To further that purpose, Section 8(a) of the Act bans payments for referrals in the real estate settlement process. Section 8(a) provides: “No person shall give and no person shall accept any fee, kickback, or thing of value pursuant to any agreement or understanding, oral or otherwise, that business incident to or a part of a real estate'settlement service involving a federally related mortgage loan shall be referred to any person.” Id. § 2607(a).20
Importantly for this case, however, Section 8(c) contains a series of qualifications, exceptions, and safe harbors. Of relevance here, Section 8(c) carves out a safe harbor against overly broad interpretations of Section 8(a): “Nothing in this section shall be construed as prohibiting ... (2) the payment to. any person of a bona fide salary or compensation or other payment for goods or facilities actually furnished or for services actually performed.” Id. § 2607(c)(2).
In 1996, PHH, a mortgage lender, began participating in so-called captive reinsurance agreements. PHH would refer borrowers' to certain mortgage insurers. Those mortgage insurers, in turn, would purchase mortgage reinsurance from Atrium, a wholly owned subsidiary of PHH. According to PHH, this was not a problem under Section 8 because the mortgage insurers would pay no more than reasonable market value to Atrium for the reinsurance they purchased. PHH argues that the mortgage insurers were thus paying reasonable market value for reinsurance from Atrium, as allowed by the statute’s safe harbor, and were not paying anything for the referrals made by PHH, which would have been unlawful.21
Many other mortgage lenders did the same thing as PHH. They did so in part because the U.S. -Department of Housing and Urban Development, known as HUD, the federal government agency responsible for enforcing this real estate law, repeatedly said (beginning in 1997) that captive reinsurance arrangements were permissible under Section 8 so long as the mortgage insurer paid no more than reasonable- market value for the reinsurance.
In this action against PHH, however, the CFPB changed course and, for the first time, interpreted Section 8 to prohibit captive reinsurance agreements even if the mortgage insurers pay no more than reasonable market value to the reinsurers. The CFPB then retroactively applied that new interpretation against PHH based on conduct that PHH engaged in before the CFPB issued its new interpretation.
*41PHH advances two alternative and independent arguments on the statutory issue. First, PHH argues that the CFPB misinterpreted Section 8(c). Second, in the alternative, PHH argues that the CFPB violated bedrock due process principles by retroactively applying its new interpretation of the statute against PHH. We agree with PHH on both points.
A
The basic statutory question in this case.is not a close call. The text of Section 8(c) permits captive reinsurance arrangements where mortgage insurers pay no more than reasonable market value for the reinsurance. Section 8(c) .contains a broad range of exceptions, qualifications, and safe harbors to Section 8(a). As relevant here, Section 8(c) creates a safe harbor, stating: “Nothing” in Section 8 “shall be construed as prohibiting” the “payment to any person of a bona fide. salary or compensation or other payment for goods or facilities actually furnished or for services actually performed.” See 12 U.S.C. § 2607(c)(2). Nothing means nothing.
Section 8(a) prohibits, in this context, payment by a mortgage insurer to a lender for the lender’s referral of a customer to the mortgage insurer. But Section 8(a) and 8(c) do not prohibit bona fide payments by the mortgage insurer to the lender for other services that the lender (or the lender’s subsidiary or affiliate) actually provides to the mortgage insurer.
How do we determine whether the mortgage insurer’s payment to the lender was a bona fide payment for the reinsurance rather than a disguised payment for the lender’s referral of a customer to the insurer? As HUD had long explained,, the answer is commonsensical: If the payment to the lender-affiliated reinsurer is more than the reasonable market value of the reinsurance, then we may .presume that the excess payment above reasonable market value was not a bona fide payment for the reinsurance but was -a- disguised payment for a referral. Otherwise, there is no basis to treat payment of reasonable market value for the reinsurance as a prohibited payment for the referral—assuming, of course, that the reinsurance was actually provided. In other words, in the text and context.of this statute, a.bona-fide payment means a payment of reasonable market valued.22
To be sure, one might say that the mortgage insurer—although paying reasonablé market value for the reinsurance—would have preferred not to purchase reinsurance at all or to purchase it from a different reinsurer. In that sense, the lender’s actions create a kind of tying arrangement in which the lender says to the mortgage insurer: ‘ We will refer customers to you, but only if you purchase another service from our affiliated reinsurer, albeit at reasonable market value. But the statute does not proscribe that kind of arrangement. As relevant here, Section 8(a) proscribes payments for referrals. Period. It does not proscribe other transactions between the lender and mortgage insurer. Nor does it proscribe a tying arrangement, so long as the only payments exchanged are bona fide payments for services arid not payments for referrals.
The CFPB says, however, that the mortgage insurer’s payriient for the reinsurance is not “bona fide” if it was part of a tying *42arrangement. That makes little sense. Tying arrangements are ubiquitous in the U.S. economy. To be sure, tying arrangements are outlawed in certain circumstances, but they were not outlawed by Section 8 in the circumstances at issue here.23 A payment for a service pursuant to a tying arrangement does not make the payment any less bona fide, so long as the payment for the service reflects reasonable market value. A bona fide payment means a payment of reasonable market value.
Recognizing, however, that an aggressive government enforcement agency or court might interpret other transactions between businesses in the real estate market as connected to, conditioned on, or tied to referrals, and might try to sweep such transactions within the scope of Section 8(a)’s prohibition, Congress explicitly made clear in Section 8(c) that those other transactions were lawful so long as reasonable market value was paid and the services were actually performed. In other words, Section 8(c) specifically bars the aggressive interpretation of Section 8(a) advanced by the CFPB in this case. Section 8(c) was designed to provide certainty to businesses in the mortgage lending process. The CFPB’s interpretation flouts that statutory goal and upends the entire system of unpaid referrals that has been part of the market for real estate settlement services.
Our interpretation of the text accords with the longstanding interpretation of the Department of Housing and Urban Development. For decades, HUD explained to mortgage lenders that captive reinsurance arrangements where reasonable market value was paid were entirely permissible under Section 8. Indeed, HUD adopted a rule, Regulation X, under which captive reinsurance arrangements were permitted so long as the insurer paid reasonable market value for the reinsurance. See 24 C.F.R. § 3500.14(g) (2011); see also 24 C.F.R. § 3500.14(e)-(f) (1992). That regulation remains in place as a CFPB regulation. See 12 C.F.R. § 1024.14 (2016). Yet in its decision here and its argument to this Court, the CFPB has not adhered to the regulation. On the contrary, the CFPB now says the opposite of what HUD’s prior interpretations and Regulation X all say. In the next section, we will consider the due process implications of the CFPB’s retroactive application of its about-face. For now, we simply note that the CFPB’s interpretation flouts not only the text of the statute but also decades of carefully and repeatedly considered official government interpretations.
Our interpretation of the text also accords with the statute’s multiple purposes, as revealed by the text. One goal of the statute was to eliminate payments for referrals because “referral fees ... tend to increase unnecessarily the costs of certain settlement services.” 12 U.S.C. § 2601(b)(2). Another purpose of the statute, as the text shows, was to allow market participants to refer customers to other *43service providers, albeit without demanding or receiving payment for the referral. Id. § 2607(a). After all, such referrals often enhance the efficiency of the homebuy-ing process. Another purpose was to assure market participants that they could engage in transactions—other than payments for referrals—so long as reasonable payments were made for services actually performed. Id. § 2607(c); see also Glover v. Standard Federal Bank, 283 F.3d 953, 964 (8th Cir. 2002); Geraci v. Homestreet Bank, 347 F.3d 749, 751 (9th Cir. 2003). If payments for services actually performed reflect the reasonable market value of the services, as they must to fall within Section 8(c), then they square with the Act’s various purposes.
Our interpretation of the text also aligns with how key Members of Congress intended Sections 8(a) and 8(c) to work together. When the Real Estate Settlement Procedures Act was reported out of the Senate Committee on Banking, Housing and Urban Affairs in 1974, the accompanying committee report stated: “Reasonable payments in return for services actually performed or goods actually furnished are not intended to be prohibited.” S. Rep. No. 93-866, at 6 (1974). Note the Senate Committee’s use of the word “reasonable.” Here, the CFPB has argued that the phrase “bona fide payment” in the statute somehow means something different from “reasonable payment.” CFPB Br. 29 & n.18. But the Senate Committee, following the commonsense meaning, expressly equated the two terms. Contrary to the CFPB’s strained interpretation, the committee report indicates that those Members of Congress intended Sections 8(a) and 8(c) to mean what they say and to say what.they mean: Payments for referrals are proscribed, but payments for other services actually performed are permitted, so long as the payments reflect reasonable market value.
In seeking to defend its interpretation, the CFPB argues that its interpretation of the Real Estate Settlement Procedures Act is entitled to Chevron deference: But Chevron instructs us at step one to' first employ all of the traditional tools of statutory interpretation, as we have done. See Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837, 843 n.9, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984). After we employ those tools, only if an ambiguity remains do we defer to the agency, if its interpretation is at least reasonable. Here, we conclude at Chevron step one that the statute permits captive reinsurance arrangements. Indeed, Section 8(c) eliminates any potential ambiguity" that might have "existed if all we had were Section 8(a) alone. Section 8(c) clearly permits captive reinsurance arrangements so long as the mortgage insurer pays reasonable market value for reinsurance actually provided. So the CFPB’s interpretation fails at Chevron step one. Cf. Kingdomware Technologies, Inc. v. United States, — U.S. - , 136 S.Ct. 1969, 1979, 195 L.Ed.2d 334 (2016); FERC v. Electric Power Supply Association, — U.S. -, 136 S.Ct. 760, 773 n.5, 193 L.Ed.2d 661 (2016); Adams Fruit Co. v. Barrett, 494 U.S. 638, 642, 110 S.Ct. 1384, 108 L.Ed.2d 585 (1990); Loving v. IRS, 742 F.3d 1013, 1021-22 (D.C. Cir. 2014). For those same reasons, if we reached Chevron step two, we would conclude that the CFPB’s interpretation is not a reasonable interpretation of the statute in light of the statute’s text, history, context, and purposes.
The policy and ethics of captive reinsurance arrangements no doubt can be debated, as can the policy and ethics of the wide variety of similar tying and referral arrangements that are ubiquitous in the American economy. But the initial question before us (and that was before the CFPB) *44is not one of policy or ethics. The question is one of law. Under Section 8(a) and Section 8(c), the relevant questions are whether the payment from the. mortgage insurer to the lender-affiliated reinsurer is bona fid,e—that is, commensurate with the reasonable market value of the reinsurance— and whether the services were actually performed. If so, then the payment is permissible, as HUD had long stated.
The CFPB obviously believes that captive reinsurance arrangements are harmful and should be illegal. But the decision whether to adopt a new prohibition on captive reinsurance arrangements is for Congress and the President when exercising the legislative authority. It is not a decision for the CFPB to make unilaterally. See King v. Burwell, — U.S. -, 135 S.Ct. 2480, 2496, 192 L.Ed.2d 483 (2015) (“In a democracy, the power to make the law rests with those chosen by the people.”). .
We hold that Sections 8(a) and 8(c) of the Real Estate .Settlement Procedures Act allow captive reinsurance arrangements so long as the mortgage insurance companies pay no more than reasonable market value to the reinsurers for services actually provided. On remand, the CFPB may determine whether the relevant mortgage insurers in fact paid more than reasonable market value to the reinsurer Atrium for the reinsurance.24
B
Even if the CFPB’s interpretation of Section 8 were permissible, it nonetheless represented a complete about-face from the Federal Government’s longstanding prior interpretation of Section 8. Agency change is not a fatal flaw in and of itself, so long as the change is reasonably explained and so long as the new interpretation is consistent with the statute. See FCC v. Fox Television Stations, Inc., 556 U.S. 502, 514-16, 129 S.Ct. 1800, 173 L.Ed.2d 738 (2009). But change becomes a problem—a fatal one—when the Government decides to turn around and retroactively apply that new interpretation to proscribe conduct that occurred before the new interpretation was issued. Therefore, even if the CFPB’s new interpretation were consistent with the statute (which it is not), the CFPB violated due process by retroactively applying that new interpretation to PHH’s conduct that occurred before the date of the CFPB’s new interpretation.
Before the creation of the CFPB in 2010, the Department of Housing and Urban Development administered the Real Estate Settlement Procedures Act. In 1997, HUD sent a letter to a mortgage company. The mortgage company had requested that HUD clarify the application of Section 8 of the Real Estate Settlement Procedures Act to captive reinsurance arrangements.
In the letter, HUD analyzed the relationship between Sections 8(a) and 8(c). HUD said that “Subsection 8(c) of RESPA sets forth various exemptions from these prohibitions.” Letter from Nicolas P. Ret-sinas, Assistant Secretary for Housing, Department of Housing and Urban Development, to Countrywide Funding Corporation 3 (Aug: 6, 1997) (J.A. 251-58). HUD further stated that its “view of captive reinsurance” was that “the arrangements are permissible” if “the payments to the reinsurer: (1) are for reinsurance services ‘actually furnished or for services performed’ and (2) are bona fide compensation that does not exceed the value of such services.” Id. (J.A. 253).
*45The 1997 HUD letter was widely disseminated and relied on in the industry. In 2004, a title association again asked HUD about the legality of captive reinsurance programs under the Real Estate Settlement Procedures Act. HUD restated the position it had taken in 1997 with respect to captive mortgage reinsurance. As it had in 1997, HUD wrote that captive reinsurance agreements are permissible if. the payments made to the reinsurer (1) are “for reinsurance services actually furnished or for services performed” and (2) are “bona fide compensation that does not exceed the value of such services.” Letter from John P. Kennedy, Associate General Counsel for Finance and Regulatory Compliance, Department of Housing, and Urban Development, to American Land Title Association 1 (Aug. 12, 2004) (J.A. 259).
In accord with -those letters, - HUD’s Real Estate Settlement Procedures: Act regulations were set forth in Regulation X, which was first issued in 1976 and updated, as relevant here, in 1992, See 24:C.F.R. §§ 3500.01-3500.14 (1977); see also 24 C.F.R. § 3500.14 (1993).. As it initially read, Regulation X stated: “The payment and receipt of a thing of value that bears a reasonable relationship to the value: of the goods or services received by the person or company making- the payment is not prohibited by RESPA section 8. To the extent the thing of value is in excess of the reasonable value of the goods provided or services performed, the excess is not for services actually rendered and may be considered a kickback or referral fee proscribed by RESPA section 8.” 24 C.F.R. § 3500.14(e) (1977). Regulation X was slightly reworded in 1992: “If the payment of a thing of value bears no reasonable relationship to the market value of the goods or services provided, then the excess is not for services or goods actually performed or provided.” . 24 C.F.R. § 3500.14(g)(3) (1993). Regulation X described “bona fide” payments for services actually performed as payments that “Section 8 of RESPA permits.” Id. § 3500.14(g)(1). After the CFPB inherited HUD’s enforcement and rulemaking authority under the Act, the CFPB itself codified HUD’s Regulation X provisions governing Section 8. See 12 C.F.R. § 1024.14(g) (2012).
In our Court, the CFPB acknowledges that, at the time of PHH’s conduct, Regulation X stated “that, if a payment bears no reasonable relationship to the value of the services provided, then the excess may be a payment for a referral.” CFPB Br. 31 n.23. But the CFPB argues that “this does not mean that, if the payment does bear a reasonable relationship to the value of the sendees provided, then those payments are never for referrals.” Id. The CFPB’s interpretation is a facially nonsensical reading of Regulation X. As Regulation X made clear, if an insurer makes a payment at reasonable market value for services actually providéd, that payment is not a payment for a referral.
HUD’s consistent and repeated interpretation of Section 8 was widely known and relied on in the mortgage lending industry. It was reflected in the leading treatise on the Act. See James H. PanNabecker & David Stemler, The Respa Manual: A Complete Guide to the Real Estate Settlement Procedures Act § 8.04[6][a] (2013). And courts had acknowledged and approved HUD’s interpretation. See, e.g., Glover v. Standard Federal Bank, 283 F.3d 953, 964 (8th Cir. 2002) (the “permissive language of Section 8(c) ... clearly states that reasonable payments for goods, facilities or services actually furnished are not prohibited by RESPA, even when done in connection with the referral of a particular loan to a particular lender”); cf. Carter v. Welles-Bowen Realty, Inc., 736 F.3d 722, 728 (6th Cir.2013) (Section 8(c)(2) is a *46“safe harbor[ ]” from Section 8(a)’s “ban on referral fees”); Geraci v. Homestreet Bank, 847 F.3d 749, 751 (9th Cir. 2003) (describing Section 8(c)(2) as a “safe harbor” and noting that HUD, when evaluating whether payments from mortgage lenders to mortgage brokers fall within Section 8(c), considers whether the payments for services “are reasonably related to the value of the ... services that were actually performed”).
At the time PHH engaged in its captive reinsurance arrangements, everyone knew the deal: Captive reinsurance arrangements were lawful under Section 8 so long as the mortgage insurer paid no more than reasonable market value to the reinsurer for reinsurance actually furnished.
In 2015, however, the CFPB decided that captive reinsurance agreements were prohibited by Section 8. The CFPB then applied its new interpretation of Section 8 retroactively against PHH, ruling against PHH based on conduct that had occurred as far back as 2008. The retroactive application of the CFPB’s new interpretation violated the Due Process Clause.
The-Due Process Clause limits the extent to which the Government may retroactively alter the legal consequences of an entity’s or person’s past conduct. That anti-retroactivity principle “is deeply rooted in our jurisprudence, and embodies a legal doctrine centuries older than our Republic.” Landgraf v. USI Film Products, 511 U.S. 244, 265, 114 S.Ct. 1483, 128 L.Ed.2d 229 (1994); see also Eastern Enterprises v. Apfel, 524 U.S. 498, 547, 118 S.Ct. 2131, 141 L.Ed.2d 451 (1998) (Kennedy, J., concurring in the judgment and dissenting in part) (“[F]or centuries our law has harbored a singular distrust of retroactive statutes.”).
Retroactivity—in particular, a new agency interpretation that is retroactively applied to proscribe past conduct— contravenes the bedrock due process principle that the people should have fair notice of what conduct is prohibited. As the Supreme Court has emphasized, “individuals should have an opportunity to know what the law is and to conform their conduct accordingly.” Landgraf, 511 U.S. at 265, 114 S.Ct. 1483. Due process therefore requires agencies to “provide regulated parties fair warning of the conduct a regulation prohibits or requires.” Christopher v. SmithKline Beecham Corp., — U.S. -, 132 S.Ct. 2156, 2167, 183 L.Ed.2d 153 (2012) (internal quotation marks and alteration omitted); see also FCC v. Fox Television Stations, Inc., — U.S. -, 132 S.Ct. 2307, 2317, 183 L.Ed.2d 234 (2012) (“A fundamental principle in our legal system is that laws which regulate persons or entities must give fair notice of conduct that is forbidden or required.”); cf. United States v. Pennsylvania Industrial Chemical Corp., 411 U.S. 655, 674, 93 S.Ct. 1804, 36 L.Ed.2d 567 (1973) (“Thus, to the extent that the regulations deprived PICCO of fair warning as to what conduct the Government intended to make criminal, we think there can be no doubt that traditional notions of fairness inherent in our system of criminal justice prevent the Government from proceeding with the prosecution.”); Cox v. Louisiana, 379 U.S. 559, 571, 85 S.Ct. 476, 13 L.Ed.2d 487 (1965) (“[U]nder all the circumstances of this case, after the public officials acted as they did, to sustain appellant’s later conviction for demonstrating where they told him he could would be to sanction an indefensible sort of entrapment by the State— convicting a citizen for exercising a privilege which the State had clearly told him was available to him. The Due Process Clause does not permit convictions to be obtained under such circumstances.”) (internal quotation marks and citation omitted); Bouie v. City of Columbia, 378 U.S. *47347, 350-51, 84 S.Ct. 1697, 12 L.Ed.2d 894 (1964) (Due Process Clause violated when state punished defendants “for conduct that was not criminal at the time they committed it” because the “underlying principle” of fair warning dictates that “no man shall be held criminally responsible for conduct which he could not reasonably understand to be proscribed”) (internal quotation marks omitted); Raley v. Ohio, 360 U.S. 423, 438-39, 79 S.Ct. 1257, 3 L.Ed.2d 1344 (1959) (“There was active misleading. The State Supreme Court dismissed the statements of the Commission as legally erroneous, but the fact remains that at the inquiry they were the voice of the State most presently speaking to the appellants. We cannot hold that the Due Process Clause permits convictions to be obtained under such circumstances.”) (internal citation omitted).25
In SmithKline, for example, the Supreme Court refused to defer to the Department of Labor’s changed interpretation of a regulation because the regulated industry “had little reason to suspect that its longstanding practice” violated the law. 132 S.Ct. at 2167. Neither the relevant statute nor any regulations provided clear notice of the Department of Labor’s new interpretation. “Even more important,” the Court said, was that “despite the industry’s decades-long practice,” the “DOL never initiated any enforcement actions” or “otherwise suggested that it thought the industry was acting unlawfully.” Id. at 2168.
In SmithKline, in a sentence that all but decides the case before us, the Supreme Court further stated: An “agency should not change an interpretation in an adjudicative proceeding where doing so would impose new liability on individuals for past actions which were taken in good-faith reliance on agency pronouncements.” Id. at 2167 (internal quotation marks and alterations omitted) (quoting NLRB v. Bell Aerospace Co., 416 U.S. 267, 295, 94 S.Ct. 1757, 40 L.Ed.2d 134 (1974)). The Court elaborated: “It is one thing to expect regulated parties to conform their conduct to an agency’s interpretations once the agency announces them; it is quite another to require regulated parties to divine the agency’s interpretations in advance or else be held liable when the agency announces its interpretations for the first time in an enforcément proceeding and demands deference.” Id. at 2168. Because automatically accepting the Department of Labor’s new interpretation “would result in precisely the kind of unfair surprise against which our cases have long warned,” the Supreme Court refused to defer to the Department of Labor’s retroactive application of a changed interpretation of its own regulations. Id. at 2167 (internal quotation marks omitted).
*48All of those fundamental anti-retroactivity principles are Rule of Law 101. And all of those fundamental anti-retroactivity principles fit this case precisely. PHH did not have fair notice of the CFPB’s interpretation of Section 8 at the time PHH engaged in the conduct at issue here. PHH participated in captive reinsurance arrangements in justifiable reliance on the interpretation stated by HUD in 1997 and restated in 2004. The CFPB therefore violated due process by retroactively applying its changed interpretation to PHH’s past conduct and requiring PHH to pay $109 million for that conduct.
The CFPB retorts that there is a presumption in favor of retroactive application of agencies’ interpretations of ambiguous statutes. CFPB Br. 42-43. But here, the CFPB was changing the Government’s longstanding interpretation of that statute and then applying that changed interpretation retroactively. The CFPB’s decision was a reversal of position—an “abrupt departure” from a consistent, longstanding position. Clark-Cowlitz Joint Operating Agency v. FEBC, 826 F.2d 1074, 1081 (D.C. Cir. 1987) (en banc) (internal quotation marks omitted). The Due Process Clause does not allow retroactive application of such a change.
The CFPB responds that nothing, including the 1997 letter, gave regulated entities such as PHH a reason to rely on HUD’s interpretation. CFPB Br. 44-45. But in the 1997 letter, the Presidentially appointed and Senate-confirmed Assistant Secretary of HUD stated: “I trust that this guidance will assist you to conduct your business in accordance with RE SPA.” Letter from Nicolas P. Retsinas, Assistant Secretary for Housing, Department of Housing and Urban Development, to Countrywide Funding Corporation 8 (Aug. 6, 1997) (J.A. 258). We therefore find this particular CFPB argument deeply unsettling in a Nation built on the Rule of Law. When a government agency officially and expressly tells you that you are legally allowed to do something, but later tells you “just kidding” and enforces the law retroactively against you and sanctions you for actions you took in reliance on the government’s assurances, that amounts to a serious due process violation. The rule of law constrains the governors as well as the governed.
The CFPB protests that the HUD pronouncements were not reflected in a binding HUD rule. To begin with, that is wrong. As discussed, Regulation X reflected HUD’s longstanding interpretation that Section 8(c) allowed payments of reasonable market value for services actually performed. See 12 C.F.R. § 1024.14 (2012) (CFPB codification of Regulation X Section 8 provisions); 24 C.F.R. § 3500.14 (2011) (HUD Regulation X Section 8 provisions). In any event, the CFPB is confusing (i) the administrative law issue of whether an- agency rule is sufficiently authoritative to obtain Chevron deference or to constitute a norm of proscribed conduct that the agency may enforce and (ii) the due process issue of whether an agency statement pronouncing the legality of certain conduct was sufficiently official for citizens to rely on it as the citizens arranged their conduct. To trigger the latter due process protection, an agency pronouncement about the legality of proposed private conduct need not have been set forth in a rule preceded by notice and comment rulemaking, or the like. Here, the agency guidance was provided by top HUD officials and was given repeatedly. Although we do not imply that those two conditions are necessary to justify citizens’ reliance for purposes of the Due Process Clause, they are. surély sufficient. Here, the regulated industry reasonably relied on those agency pronouncements.
*49Put aside all the legalese for a moment. Imagine that a police officer tells a pedestrian that the pedestrian can lawfully cross the street at a certain place. The pedestrian carefully and precisely follows the officer’s direction. After the pedestrian arrives at the other side of the street, however, the officer hands the pedestrian a $1,000 jaywalking ticket. No one would seriously contend that the officer had acted fairly or in a manner consistent with basic due process in that situation. See Cox v. Louisiana, 379 U.S. 559, 571, 85 S.Ct. 476, 13 L.Ed.2d 487 (1965). Yet that’s precisely this case. Here, the CFPB is arguing that it has the authority to order PHH' to pay $109 million even though PHH acted in reliance upon numerous government pronouncements authorizing precisely the conduct in which PHH engaged.
The Due Process Clause does not countenance the CFPB’s gamesmanship. As Justice Kennedy eloquently explained in a related scenario: “If retroactive laws change the legal consequences of transactions long closed, the change can destroy the reasonable certainty and security which are the very objects of property ownership. Groups targeted by retroactive laws, were they to be denied all protection, would have a justified fear that a government once formed to protect expectations now can - destroy them. Both stability of investment and confidence in the constitutional system, then, are secured by due process restrictions against severe retroactive legislation.” Eastern Enterprises, 524 U.S. at 548-49, 118 S.Ct. 2131 (Kennedy, J., concurring in the judgment and dissenting in part); see also General Electric Co. v. EPA, 53 F.3d 1324, 1328-29 (D.C. Cir. 1995) (“In the absence of notice—for example, where the regulation 1 is not sufficiently clear to warn a party about what is expected of it—an agency may not deprive a party of property by imposing civil or criminal liability.”); Satellite Broadcasting Co. v. FCC, 824 F.2d 1, 3-4 (D.C. Cir. 1987) (“Traditional concepts of due process ... preclude an agency from penalizing a private party for violating a rule without first providing adequate notice of the substance of the rule— Otherwise the practice of administrative law would come to resemble ‘Russian Roulette.’).
In sum, even if the CFPB’s new interpretation of Section 8 were a permissible interpretation of the statute, which it is not, the CFPB’s. interpretation could not constitutionally be applied retroactively to PHH’s conduct that occurred before that new interpretation.26 On remand; to reiterate, the CFPB may determine whether the relevant mortgage insurers paid more than reasonable market value to the reinsurer Atrium, which is what the statute proscribes and what HUD’s longstanding pronouncements provided.27
*50V
In order to hold PHH liable, the CFPB must therefore show that the relevant mortgage insurers paid more than reasonable market value to Atrium for the reinsurance. On remand, the CFPB may attempt to make that showing, assuming that any relevant conduct by PHH occurred within the applicable statute of limitations period. That in turn brings us to the statute of limitations issue. PHH contends that most of its relevant activity occurred outside of the three-year statute of limitations applicable in this case.
“Statutes of limitations are intended to promote justice by preventing surprises through the revival of claims that have been allowed to slumber until evidence has been lost, memories have faded, and witnesses have disappeared.” Gabelli v. SEC, — U.S. -, 133 S.Ct. 1216, 1221, 185 L.Ed.2d 297 (2013) (internal quotation marks omitted). Statutes of limitations also “provide security and stability to human affairs” by affording “certainty” about “a defendant’s potential liabilities.” Id. (internal quotation marks omitted).
The general working presumption in federal civil and criminal cases is that a federal civil cause of action or criminal offense must have some statute of limitations and must not allow suits to be brought forever and ever after the acts in question. See 28 U.S.C § 2462; 18 U.S.C. § 3282. As Chief Justice Marshall stated, allowing parties to sue “at any distance of time” would be “ utterly repugnant to the genius of our laws. In a country where not even treason can be prosecuted after a lapse of three years, it could scarcely be supposed that an individual would remain forever liable to a pecuniary forfeiture.” Adams v. Woods, 6 U.S. 2 Cranch 336, 342, 2 L.Ed. 297 (1805).
The Dodd-Frank Act authorizes the CFPB to “conduct hearings and adjudication proceedings” to enforce the Real Estate Settlement Procedures Act. 12 U.S.C. § 5563(a). The Real Estate Settlement Procedures Act, in turn, provides that the CFPB may “bring an action to enjoin violations” of Section 8. Id. § 2607(d)(4). As it now reads, the Real Estate Settlement Procedures Act also provides that “actions” brought by various government agencies, including the CFPB, to enforce Section 8 “may be brought within 3 years from the date of the occurrence of the violation.” Id. § 2614.
The CFPB says that no statute of limitations applies to its case against PHH. CFPB Br. 38. The CFPB advances two primary arguments. First, the CFPB contends it is broadly authorized to bring enforcement actions under the Dodd-Frank Act, and the CFPB says that the Dodd-Frank Act contains no statute of limitations on CFPB enforcement actions brought in an administrative proceeding, as opposed to in court. Notably, that broad argument would apply to all 19 of the consumer protection statutes that the CFPB enforces, and would mean that no statute of limitations applies to CFPB administrative actions enforcing any of those statutes.
*51Second, if the Dodd-Frank Act does not override the statutes of limitations in all of the underlying statutes enforced by the CFPB, meaning that the CFPB must abide by the statutes of limitations in the underlying statutes, the CFPB contends that the statute at issue here—the Real Estate Settlement Procedures Act—imposes a three-year statute of limitations only on those enforcement actions that the CFPB brings in court. According to the CFPB, the Real Estate Settlement Procedures Act does not impose any statute of limitations for those enforcement actions that the CFPB brings in administrative proceedings.
Neither of the CFPB’s arguments is correct.
First, the CFPB argues that we should ignore any statute of limitations contained in the Real Estate Settlement Procedures Act. Instead, the CFPB claims that we should look to the general enforcement provisions of the Dodd-Frank Act because those Dodd-Frank provisions, according to the CFPB, trump the statutes of limitations in the underlying statutes enforced by the CFPB.
Under the Dodd-Frank Act, the CFPB may bring an enforcement action either in an administrative action or in court. See 12 U.S.C. §§ 5563-5564. According to the CFPB, that choice matters, for statute of limitations purposes. The CFPB says that the Dodd-Frank “provision that authorizes court actions includes a statute of limitations,” but the “provision authorizing administrative enforcement does not.” CFPB Br. 38 (emphasis added). Because the CFPB challenged PHH’s conduct through an administrative action rather than in court, the CFPB concludes that there is no applicable statute of limitations.
Importantly, the CFPB’s Dodd-Frank-based argument—if accepted here—would apply not only to actions to enforce Section 8 of the Real Estate Settlement Procedures Act. The CFPB’s argument that it is nob bound by any statute of limitations in administrative proceedings would extend to all 19 of the consumer protection laws that Congress empowered the CFPB to enforce. Cf. Integrity Advance, LLC, 2015-CFPB-0029, Doc. No. 33, CFPB Opposition to Motion to Dismiss, at 12 (arguing no statute of limitations applies to CFPB administrative action to enforce the Truth in Lending Act and the Electronic Fund Transfer Act).
The CFPB’s argument misreads the enforcement provisions of the Dodd-Frank Act. Section 5563 authorizes the CFPB “to conduct hearings and adjudication proceedings ... in order to ensure or enforce compliance with” 19 federal consumer protection laws, in addition to other rules, regulations, and orders. 12 U.S.C. § 5563(a). But Congress limited the enforcement power granted in Section 5563. The CFPB may enforce those federal laws “unless such Federal law specifically limits the Bureau-from conducting a hearing or adjudication proceeding.” Id. § 5563(a)(2) (emphasis, added). Obviously, one such “limit” is a statute of limitations. By its terms, then, Section 5563 ties the CFPB’s administrative adjudications to the statutes of limitations of the various federal consumer protection laws it is charged with enforcing.28 The Dodd-Frank Act therefore makes clear that in its enforcement action against PHH, the CFPB was bound *52by any statute of limitations located in the Real Estate Settlement Procedures Act.
Second, as to the Real Estate Settlement Procedures Act itself, the CFPB argues that the three-year limitations period in Section 2614 of that Act applies only to CFPB actions to enforce Section 8 in court, not to CFPB administrative actions to enforce Section 8 before the agency. We again disagree. Section 2614 supplies the appropriate statute of limitations period not only for CFPB actions to enforce Section , 8 that are brought in court, but also for CFPB actions to enforce Section 8 that are brought administratively.29
The first part of Section 2614 specifies a general one-year statute of limitations for any “action pursuant to” Section 8 “brought in the United States district court or in any other court of competent jurisdiction.” Id, § 2614.
The second part of Section 2614 supplies a longer, three-year statute of limitations for “actions” to enforce Section 8 “brought by the Bureau, the Secretary, the Attorney General of any State, or the insurance commissioner of any State.” Id. In this second part of Section 2614, the term “actions” is not limited to actions brought in court. Section 2614 does not specify a jurisdiction or forum for actions by the Bureau, the Secretary, the Attorney General of any State, or the insurance commissioner of any State. Section 2614 simply requires that those actions be brought within a three-year limitations period.
On its face, the statute of limitations for actions under Section- 8 is therefore straightforward: Private plaintiffs can bring actions under Section 8 only in court. Private plaintiffs cannot bring administrative actions. For those private-party suits, a one-year statute of limitations applies. The relevant government enforcement agencies—including the CFPB—may bring actions to enforce Section 8 in courts or in administrative proceedings. For those cases, a three-year statute of limitations applies.
In response, the CFPB claims that the term “actions” in Section 2614 refers only to court actions, not to administrative actions. The CFPB argues that Congress uses the word “proceedings” rather than “actions” when it wants to refer to administrative actions. That is flatly wrong. Indeed, the Dodd-Frank Act itself, which amended Section 2614 to its current form, directly contradicts the CFPB’s assertion about the meaning of the term “action.” The Dodd-Frank Act repeatedly uses the term “action” to encompass court actions and administrative proceedings. See, e.g., id. § 5497(d)(1) (“If the Bureau obtains a civil penalty against any person in any judicial or administrative action under Federal consumer financial laws..,,”); id. § 5537(b)(1) (establishing grant program for States “to hire staff to identify, investigate, and prosecute (through civil, administrative, or criminal enforcement actions) cases involving misleading or fraudulent marketing”); id. § 5538(b)(6) (“Whenever a civil action or an administrative action has *53been instituted by or on behalf of the Bureau_”); id. § 5565(c) (subsection entitled “Civil money penalty in court and administrative actions”). The same can be said for various provisions scattered throughout the U.S. Code. See, e.g., 7 U.S.C. § 2279d (“Such liability shall apply to any administrative action brought before October 21,1998, but only if the action is brought within the applicable statute of limitations.... ”); 15 U.S.C. § 78u-6(a)(l) (“The term ‘covered judicial or administrative action’ means any judicial or administrative action brought by the Commission under the securities laws that results in monetary sanctions ■ exceeding $1,000,000.”); 42 U.S.C. § 9628(b)(p(B) (“The President may bring an administrative or judicial enforcement action under this chapter_”); 49-U.S.C. § 60120(a)(1) (“The maximum amount of civil penalties for administrative enforcement actions under section 60122 shall not apply to enforcement actions under this section.”).
The CFPB also cites BP America Production Co. v. Burton, 549 U.S. 84, 127 S.Ct. 638, 166 L.Ed.2d 494 (2006). There, the Supreme Court ruled that 28 U.S.C. § 2415(a)—a civil statute of limitations provision for “every action for money damages” brought by the Government—encompassed only court actions, and not agency enforcement actions. BP America, 549 U.S. at 89, 101, 127 S.Ct. 638 (internal quotation marks and emphasis omitted). To arrive at that conclusion, the Court looked to a wide array of textual and structural clues in that statutory scheme. For example, the Court noted that the “key terms in th[e] provision—‘action’ and ‘complaint’'—are ordinarily used in connection with judicial, not administrative, proceedings.” Id. at 91, 127 S.Ct. 638. That conclusion was reinforced by Congress’s use of the word “action” as part of the term “action for money damages,” which is “generally used to mean pecuniary compensation or indemnity, which may be recovered in the courts.” Id. at 91-92, 127 S.Ct. 638 (internal quotation marks omitted). The Supreme Court also noted Congress’s use of the term “right of action” in the same provision, which is defined as the “right to bring suit; a legal right to maintain an action, with suit meaning any proceeding ... in a court of justice.” Id. at 91, 127 S.Ct. 638 (internal quotation marks omitted) (quoting Black’s Law Dictionary 1488, 1603 (4th ed. 1951)).
At the very most, BP America articulated a presumption that the term “action” means court proceedings. But it is at most a presumption. BP America certainly never said that the term “actions” always means actions in court. Far from it. Indeed, Supreme Court cases interpret the term “actions” to encompass administrative actions. See West v. Gibson, 527 U.S. 212, 220-21, 119 S.Ct. 1906, 144 L.Ed.2d 196 (1999); Pennsylvania v. Delaware Valley Citizens’ Council for Clean Air, 478 U.S. 546, 557-60, 106 S.Ct. 3088, 92 L.Ed.2d 439 (1986).
The question of whether the term “actions” in a particular statute encompasses administrative actions thus turns on the overall text, context, purpose, and history of the statute. Here, the textual and contextual clues convincingly demonstrate that administrative actions are covered. Unlike in BP America, the key part of Section 2614—which refers to “actions” brought by the CFPB—speaks of an “action” generically and is not limited to an “action for money damages.” Section 2614 also lacks other “key terms” like' “complaint” or “right of action” that were present in the statute at issue in BP America.
The broader purpose and history of the Dodd-Frank Act strongly reinforce the conclusion that the CFPB is bound by a three-year statute of limitations in its ad*54ministrative actions to enforce Section 8. Before 2010, HUD could not bring administrative enforcement actions to enforce Section 8. HUD could sue only in court. The CFPB acknowledges that a three-year statute of limitations applied to all of those HUD actions to enforce Section 8. When passing the Dodd-Frank Act in 2010, Congress empowered the CFPB (taking over for HUD) to enforce Section 8 not just in courts, but also in administrative actions. Importantly, the CFPB has complete discretion to institute enforcement actions in courts or through administrative actions. See 12 U.S.C. §§ 5563-5564. And the CFPB can obtain administratively all of the remedies that it could obtain in court. Id. § 5565(a)(2). The CFPB’s theory is that Congress—for some unstated reason—did not carry forward the three-year statute ■ of limitations for CFPB 1 administrative actions to enforce Section 8. Under the CFPB’s theory, the agency therefore can always circumvent the three-year statute of limitations simply by bringing the enforcement action administratively rather than in court. But Congress did not suggest that by transferring authority from HUD to the CFPB, it intended to relax the longstanding three-year statute of limitations.
Moreover, “Congress ‘does not, one might say, hide elephants in mouseholes.’ ” Puerto Rico v. Franklin California Tax-Free Trust, — U.S. -, 136 S.Ct. 1938, 1947, 195 L.Ed.2d 298 (2016) (quoting Whitman v. American Trucking Associations, Inc., 531 U.S. 457, 468, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001)). If by means of the Dodd-Frank Act, “Congress intended to alter” the fundamental details of the statutes of limitations for enforcement of this critical consumer protection law, “we would expect the text of the amended” statute “to say so.” Id. (internal quotation marks omitted). In other words, we would expect Congress to actually say that there is no statute of limitations for CFPB administrative actions to enforce Section 8, especially given that the CFPB has full discretion to pursue administrative actions instead of court proceedings and can obtain all of the same remedies through administrative actions that it can obtain in court. But the text of Dodd-Frank says no such thing. Nor, moreover, has the CFPB cited any legislative history that says anything like that.
Of course, there is good reason Congress did not say that the CFPB need not comply with any statutes of limitations when enforcing the Real Estate Settlement Procedures Act administratively. That would be absurd. Why would Congress allow the CFPB to bring administrative actions for an indefinite period, years or even decades after the fact? Why would Congress create such a nonsensical dichotomy between CFPB court actions and CFPB administrative actions? The CFPB has articulated no .remotely plausible reason why Congress would have done so. See Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 575, 102 S.Ct. 3245, 73 L.Ed.2d 973 (1982) (“absurd results are to be avoided” where “alternative interpretations consistent with the legislative purpose are available”). The CFPB’s interpretation is especially alarming because the agency can seek civil penalties in these administrative actions. 12 U.S.C. § 5565(a)(2). But the Supreme Court has emphatically stressed the importance of statutes of limitations in civil penalty provisions. As the Supreme Court stated in Gabelli: “Chief Justice Marshall used particularly forceful language in emphasizing the importance of time limits on penalty actions, stating that it “would be utterly repugnant to the genius of our laws’ if actions for penalties could ‘be brought at any distance of time.’ ” 133 S.Ct. at 1223 (quoting Adams, 6 U.S. at 342); see also 3M Co. v. Browner, *5517 F.3d 1453, 1457 (D.C. Cir. 1994) (“Justice Story, sitting as a circuit justice in a civil penalty case, made the same point as Chief Justice Marshall: ‘it would be utterly repugnant to the genius of our laws, to allow such prosecutions a perpetuity of existence.’”) (quoting United States v. Mayo, 26 F.Cas. 1230, 1231, (No. 15754) (C.C.D. Mass. 1813)).
The absurdity of the CFPB’s position is illustrated by its response to a hypothetical question about the CFPB’s bringing an administrative enforcement action 100 years after the allegedly unlawful conduct. Presented with that question, the CFPB referenced its prosecutorial discretion. But “trust us” is ordinarily not good enough. Cf. McDonnell v. United States, — U.S. -, 136 S.Ct. 2355, 2372-73, 195 L.Ed.2d 639 (2016) (declining to construe a statute “on the assumption that the Government will use it responsibly”) (internal quotation marks omitted). The CFPB also suggested that the equitable defense of laches might apply to such a case, and that “a court would look askance at a proceeding” initiated 100 years after the challenged conduct occurred. CFPB Br. 38 n.28. We need not wait for an enforcement action 100 years after the fact. This Court looks askance now at the idea that the CFPB is free to pursue an administrative enforcement action for an indefinite period of time after the relevant conduct took place. A much more logical, predictable interpretation of the agency’s authority is that the three-year limitations period in Section 2614 applies equally to CFPB court actions and CFPB administrative actions. And most importantly for our purposes, that is what the relevant statutes actually say.30 * * *
We grant PHH’s petition for review, vacate the CFPB’s order, and remand for further proceedings consistent with this opinion. On remand, the CFPB may determine, among other things, whether, consistent with the applicable three-year statute of limitations, the relevant mortgage insurers paid more than reasonable market value to Atrium.
So ordered.
. If PHH fully prevailed on its constitutional argument, including with respect to severability, the CFPB could not continue operating unless and until Congress enacted new legislation. As a result, we could not and would not remand to the CFPB for any further proceedings in this case. By contrast, even if PHH fully prevails on the statutory issues, we still will have to remand to the CFPB for the agency to conduct the proceeding in accordance with the appropriate statutory requirements, under which PHH may still be liable for certain alleged wrongdoing. In other words, PHH’s constitutional and severability argument, if accepted, would afford it full relief from any CFPB enforcement action and thus would afford it broader relief than would its statutory arguments. For that reason, we have no choice but to address the constitutional issue first. The constitutional issue cannot be' avoided in any principled way. We therefore respectfully but firmly disagree with Judge Henderson’s suggestion in.her separate opinion that the constitutional issue can be avoided. In our view, failing to decide the constitutional issue here would be impermissible judicial abdication, not judicial restraint.
Moreover, apart from that necessity in this case, when a litigant raises a fundamental constitutional challenge to the very structure or existence of an agency enforcing the law against it, the courts ordinarily address that issue promptly, at least so long as jurisdictional requirements such as standing are met. See, e.g., Free Enterprise Fund, 561 U.S. at 490-91, 130 S.Ct. 3138; Morrison v. Olson, 487 U.S. at 669-70, 108 S.Ct. 2597; Buckley v. Valeo, 424 U.S. 1, 12, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976). That was the approach we took in both Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board, 684 F.3d 1332, 1334, 1336-37 (D.C. Cir. 2012), and Raymond J. Lucia Cos. v. SEC, 832 F.3d 277, 283-84 (D.C. Cir.2016). It can be irresponsible for a court to unduly delay ruling on such a fundamental and ultimately unavoidable structural challenge, given the systemic ramifications of such an issue.
. To cabin the effects of Humphrey’s Executor on the Presidency, some have proposed reading the standard for-cause removal restrictions in the statutes creating independent agencies to allow for Presidential removal of independent agency heads based on policy differences. But Humphrey’s Executor itself rejected that interpretation. As the Supreme Court recently explained, Humphrey’s Executor refuted the idea that "simple disagreement” with an agency head's "policies or priorities could constitute 'good cause’ for its removal.” Free Enterprise Fund, 561 U.S. at 502, 130 S.Ct. 3138. The correct reading of the "for-cause” restrictions, the Court stated in Free Enterprise Fund, is that they "mean what they say” and preclude removal except in cases of inefficiency, neglect of duty, or malfeasance in office. Id.
. The independent agencies have,been designed, moreover, to avoid "the suspicion of partisan direction.” Humphrey's Executor, 295 U.S. at 625, 55 S.Ct. 869. The independent agency heads are appointed by the President with the advice and consent of the Senate (or appointed for a temporary period by the President alone in appropriate Senate recesses). By statute, certain independent agencies must include members of both major political parties. See, e.g., 15 U.S.C. § 41 (Federal Trade Commission); 15 U.S.C. § 78d(a) (Securities and Exchange Commission); 15 U.S.C, § 2053(c) (Consumer Product Safety Commission); 42 U.S.C. § 7171(b)(1) (Federal Energy Regulatory Commission).
. In general, an agency without a for-cause removal statute is an executive agency, not an independent agency, because the President can supervise, direct, and remove at will the heads of those agencies. That said, in the period from Myers (1926) to Humphrey's Executor (1935), Congress created several multi-member agencies that did not include for-cause provisions, apparently because Congress believed that Myers had outlawed making agencies independent. Those agencies included the FCC and the SEC. After Humphrey's Executor, those multi-member agencies were nonetheless treated as independent agencies. Cf. Free Enterprise Fund v. Public Company Accounting Oversight Board, 561 U.S. 477, 487, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (deciding case on assumption that SEC is an independent agency); Wiener v. United States, 357 U.S. 349, 352-54, 78 S.Ct. 1275, 2 L.Ed.2d 1377 (1958). But because those agencies' statutes do not contain express for-cause provisions, some suggest that those agencies should be treated as executive agencies. See Kirti- Datla & Richard L. Revesz, Deconstructing Independent Agencies (and Executive Agencies), 98 Cornell L. Rev. 769, 834-35 (2013); Note, The SEC Is Not an Independent Agency, 126 Harv. L. Rev. 781, 801 (2013). We need not tackle that question in this case and do not imply an answer one way or the other about the executive or independent status of the multi-member agencies without express for-cause removal provisions.
. Because the Social Security Administration and the Office of Special Counsel do not exercise the core executive power of bringing law enforcement actions and because they have narrow jurisdiction, a holding invalidating the single-Director structure of the CFPB would not necessarily invalidate the single-Director structure of the Social Security Administration and the Office of Special Counsel. That said, if , those two agencies are unconstitutionally structured, the remedy would presumably be the same remedy as in Free Enterprise Fund-, severing the for-cause provision so that the agencies would continue to fully operate, albeit as traditional executive agencies rather than independent agencies, Cf. infra pp.37-39. We do not address those questions here.
. Some have suggested that the CFPB Director is similar to the Comptroller of the Currency. But unlike the Director, the Comptroller is not independent. The Comptroller is removable at will by the President. See 12 U.S.C. § 2 (“The Comptroller of the Currency shall be appointed by the President, by and with the advice and consent of the Senate, and shall hold his office for a term of five years unless sooner removed by the President, upon reasons to be communicated by him to the Senate.”).
. The historical practice is further illustrated by the quorum provisions that are applicable to independent agencies. Those quorum provisions reinforce the settled understanding that independent agencies are to have multiple members. Cf. New Process Steel, L.P. v. NLRB, 560 U.S. 674, 130 S.Ct. 2635, 177 L.Ed.2d 162' (2010); Márshall J. Breger & Gary J. Edles, Established by Practice: The Theory and Operation of Independent Federal Agencies, 52 Admin. L. Rev. 1111, 1182 & app, (2000) (summarizing independent agency quorum requirements).
. The Supreme Court has heavily relied on historical practice as a guide not just in separation of powers cases, but also in federalism cases. In several federalism cases in the last 25 years, the Court has invalidated novel congressional statutes that alter the traditional federal-state balance. See New York v. United States, 505 U.S. 144, 177, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992) ("The take title provision appears to be unique. No other federal statute has been cited which offers a state government no option other than that of implementing legislation enacted by Congress.”); Printz v. United States, 521 U.S. 898, 905, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997) ("[I]f, as petitioners contend, earlier Congresses avoided use of this highly attractive power, we would have reason to believe that the power was thought not to exist.”); Alden v. Maine, 527 U.S. 706, 744, 119 S.Ct. 2240, 144 L.Ed.2d 636 (1999) ("Not only were statutes purporting to authorize private suits against noncon-senting States in state courts not enacted by early Congresses; statutes purporting to authorize such suits in any forum are all but absent from our historical experience. ... The provisions of the FLSA at issue here, which were enacted in the aftermath of Parden, are among the first statutory enactments purporting in express terms to subject nonconsenting States to private suits.”); United States v. Windsor, - U.S. -, 133 S.Ct. 2675, 2692, 186 L.Ed.2d 808 (2013) ("DOMA, because of its reach and extent, departs from this history and tradition of reliance on state law to define marriage.”); cf. National Federation of Independent Business v. Sebelius, — U.S. -, 132 S.Ct. 2566, 2586, 183 L.Ed.2d 450 (2012) (binding opinion of Roberts, C.J.) ("But Congress has never attempted to rely on that power to compel individuals not engaged in commerce to purchase an unwanted product.”); id. at 2649 (joint dissent of Scalia, Kennedy, Thomas, and Alito, JJ.) ("[T]he relevant histoiy is not that Congress has achieved wide and wonderful results through the proper exercise of its assigned powers in the past, but that it has never before used the Commerce Clause to compel entry into commerce.”).
. Of course, if the constitutional text is sufficiently clear, then the existence of any historical practice departing from that text is not persuasive, See, e.g., INS v. Chadha, 462 U.S. 919, 944-46, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983); Powell v. McCormack, 395 U.S. 486, 546-47, 89 S.Ct. 1944, 23 L.Ed.2d 491 (1969). Here, the question concerns the scope of Humphrey’s Executor—which, depending on one’s perspective, requires either an analysis of a court-created exception to Article II or an analysis of ambiguous constitutional text in Articles I and II. Either way, in resolving those kinds of separation of powers questions, history and tradition play a critical role. See Noel Canning, 134 S.Ct. at 2559-60; Free Enterprise Fund, 561 U.S. at 505-06, 130 S.Ct. 3138.
. Justice Scalia concurred in the judgment for four Justices in Noel Canning, arguing as relevant here that the text of the Constitution rendered intra-session recess appointments unconstitutional even in Senate recesses of 10 or more days, But Justice Scalia did not disagree with the Court’s claim that historical practice often matters in separation of powers cases, which is the relevant point for our purposes. See Noel Canning, 134 S.Ct. at 2594 (Scalia, J„ concurring in the judgment) ("Of course, where a governmental practice has been open, widespread, and unchallenged since the early days of the Republic, the practice should guide our interpretation of an ambiguous constitutional provision.”). Rather, Justice Scalia stated that the constitutional text in that case was sufficiently clear and dispositive that resort to historical practice was unnecessary and unwarranted. See id. at 2592; see generally John F, Manning, Separation of Powers as Ordinary Interpretation, 124 Harv, L. Rev, 1939 (2011).
. Justice Breyer dissented for four Justices in Free Enterprise Fund. But importantly, he dissented not because he disagreed with the Court's point that historical practice matters, but.rather primarily because he did not see a meaningful difference—in practical, analytical, or constitutional terms—between one and two levels of for-cause removal. See Free Enterprise Fund, 561 U.S. at 525-26, 130 S.Ct. 3138 (Breyer, J., dissenting).
. In identifying and cataloging the problems with a single-Director independent agency, we do not in any way question the integrity of the current Director, a man of substantial accomplishment and of longstanding and dedicated devotion to public service and the public good. Cf. Morrison v. Olson, 487 U.S. 654, 731, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988) (Scalia, J., dissenting) (similarly describing the Special Division judges and independent counsel at issue in that case). But the constitutionality of an agency structure "must be adjudged on the basis of what it permits to happen.” Id.
. Congress may of course establish executive agencies that are headed by multiple individuals (although it rarely does so), -but each member must be removable at will by the President for the agency to maintain its status as an executive agency.
. The for-cause removal restrictions attached to independent agencies ordinarily prohibit removal except in cases of inefficiency, neglect of duty, or malfeasance. Those restrictions have significant impact both in law and in practice. See Free Enterprise Fund, 561 U.S. at 502, 130 S.Ct. 3138 (for-cause restrictions “mean what they say”); Freytag v. Commissioner of Internal Revenue, 501 U.S. 868, 916, 111 S.Ct. 2631, 115 L.Ed.2d 764 (1991) (Scalia, J., concurring in part and concurring in the judgment) ("independent regulatory agencies such as the Federal Trade Commission and the Securities and Exchange Commission” are "specifically designed not to have the quality ... of being subject to the exercise of political oversight and sharing the President’s accountability to the people”) (internal quotation marks and alteration omitted); Mistretta v. United States, 488 U.S. 361, 411, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989) (for-cause provisions are “specifically crafted to prevent the President from exercising coercive influence over independent agencies”) (internal quotation marks omitted). Humphrey’s Executor and Wiener v. United States show, for example, that for-cause removal requirements prohibit dismissal by the President due to lack of trust in the administrator, see Humphrey’s Executor, 295 U.S. at 625-26, 55 S.Ct. 869, differences in policy outlook, id, or the mere desire to install administrators of the President’s choosing, Wiener, 357 U.S. 349, 356, 78 S.Ct. 1275, 2 L.Ed.2d 1377 (1958). In Morrison v. Olson, the Court therefore took it as a given that "the degree of control exercised by the Executive Branch over an independent counsel is clearly diminished in relation to that exercised over other prosecutors, such as the United States Attorneys, who are appointed by the President and subject to termination at will.” 487 U.S. at 696 n.34, 108 S.Ct. 2597; see also Buckley v. Valeo, 424 U.S. 1, 133, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976) ("The Court in {Humphrey’s Executor] carefully emphasized that ... the members of such agencies were to be independent of the Executive in their day-today operations...,”); Humphrey's Executor, 295 U.S. at 628, 55 S.Ct. 869 (independent agencies “cannot in any proper sense be characterized as an arm or an eye of the executive”),
. In its brief, PHH has expressly preserved the argument that Humphrey’s Executor should be overruled. The reasoning of Humphrey’s Executor of course was inconsistent with the reasoning in the Court's prior decision in Myers. See Humphrey's Executor, 295 U.S. at 626, 55 S.Ct. 869 ("In so far as” the expressions in Myers are "out of harmony with the views here set forth, these expressions are disapproved.”). The Humphrey’s Executor decision subsequently has received significant criticism. See Geoffrey P. Miller, Independent Agencies, 1986 Sup. Ct. Rev. 41, 93 ("Humphrey's Executor, as commentators have noted, is one of the more egregious opinions to be found on pages of the United States Supreme Court Reports.”); Peter L, Strauss, The Place of Agencies in Government-Separation of Powers and the Fourth Branch, 84 Colum. L. Rev. 573, 611-12 (1984) ("Remarkably, the Court did not pause to examine how a purpose to create a body ‘subject only to the people of the United States’—that is, apparently, beyond control of the constitutionally defined branches of government— could itself be sustained under the Constitution.”). Moreover, the reasoning of Humphrey’s Executor is in tension with some of the reasoning of the Supreme Court’s recent decision in Free Enterprise Fund. See In re Aiken County, 645 F.3d 428, 444-46 (D.C. Cir. 2011) (Kavanaugh, J., concurring); Neo-mi Rao, Removal: Necessary and Sufficient for Presidential Control, 65 Ala. L. Rev. 1205, 1208 (2014). Of course, overruling Humphrey's Executor would not mean the end of the agencies that are now independent. The agencies would simply transform into executive agencies supervised and directed by the President. So the question is not the existence of the agencies; the question is the President's control over the agencies and the resulting accountability of those agencies to the people. In any event, as a lower court, we of course must follow Supreme Court precedent. It is not our job to decide whether to overrule Humphrey's Executor. But it is emphatically our job to make sure that Humphrey’s Executor is applied in a manner consistent with settled historical practice and the Constitution’s protection of individual liberty.
. On top of the Director's unilateral power to issue rules and take enforcement actions to enforce 19 separate consumer protection statutes, the CFPB is not subject to the ordinary annual appropriations process. Instead, the Dodd-Frank Act requires the Board of Governors of the Federal Reserve to transfer "from the combined earnings of the Federal Reserve System” the amount "determined by the Director,” not to exceed 12 percent of the "total operating expenses of the Federal Reserve System.” 12 U.S.C. § 5497(a)(l)-(2). As those who have labored in Washington well understand, the appropriations process brings at least some measure of oversight by Congress. According to PHH, the CFPB's exemption from that process enhances the concern in this case about the massive power lodged in a single, unaccountable Director. That said, the single Director would constitute a constitutional problem even if the CFPB were subject to the usual appropriations process. The CFPB’s exemption from the ordinary appropriations process is at most just "extra icing on” , an unconstitutional "cake already frosted.” Yates v. United States, — U.S.-, 135 S.Ct. 1074, 1093, 191 L.Ed.2d 64 (2015) (Kagan, J., dissenting). In any event, Congress can always alter the CFPB's funding in any appropriations cycle (or at any other time). Section 5497 is not an entrenched statute shielded from'future congressional alteration, nor could it be, See, e.g., Manigault v. Springs, 199 U.S. 473, 487, 26 S.Ct. 127, 50 L.Ed. 274 (1905).
. Nothing in our opinion casts any doubt on traditional structures under which Congress may establish a process for designating the Chair of an independent board or independent commission, and for assigning the Chair various additional administrative responsibilities. Those responsibilities are distinct from substantive authority. A Chair may not unilat*37erally issue a rule, unilaterally bring an enforcement action, or unilaterally decide an adjudication. See Marshall J. Breger & Gary J. Edles, Established by Practice: The Theory and Operation of Independent Federal Agencies, 52 Admin. L. Rev. 1111, 1166-67 (2000) ("As our survey of some thirty federal multi-member agencies suggests, all of the reorganization statutes and their progeny fundamentally assign substantive authority to the agency as :a whole and administrative authority to the chairman.”). We note, moreover, that many Chairs traditionally are removable at will by the President from their position as Chair, albeit not from the commission. See Rachel E. Barkow, Insulating Agencies: Avoiding Capture Through Institutional Design, 89 Tex. L. Rev. 15, 38 & n.124 (2010).
Nor does our decision cast any doubt on the independent status of administrative law judges who are protected by for-cause provisions. Those judges conduct only adjudications (of a sort) and are not covered or affected in any way by our decision here.
. The Dodd-Frank Act contains a five-year tenure provision for the Director, see 12 U.S.C. § 5491(c)(1), akin to the similar 10-year tenure provision for the Director of the FBI and the 5-year tenure provision for the Commissioner of the IRS. See Crime Control Act of 1976, § 203, reprinted in 28 U.S.C. § 532 note (FBI Director "may not serve more than one ten-year term"); 26 U.S.C. § 7803(a)(1)(B) (term of the IRS Commissioner "shall be a 5-year term"). But under Supreme Court precedent, such tenure provisions do not prevent the President from removing at will a Director at any time during the Director's tenure. See Parsons v. United States, 167 U.S. 324, 343, 17 S.Ct. 880, 42 L.Ed. 185 (1897). Therefore, we need not invalidate and sever the tenure provision. If such a provision did impair the President's ability to remove the Director at will, then it too would be unconstitutional, and it would be invalidated and severed.
. We need not here consider the legal ramifications of our decision for past CFPB rules or for past agency enforcement actions. We note, however, that this is not an uncommon situation. For example, in just the last few years, the NLRB, the Public Company Accounting Oversight Board, and the Copyright Royalty. Board have all been on the receiving end of successful constitutional and statutory challenges.to. their structure and legality. See NLRB v. Noel Canning, —— U.S. -, 134 S.Ct. 2550, 189 L.Ed.2d 538 (2014); New Process Steel, L.P. v. NLRB, 560 U.S. 674, 130 S.Ct. 2635, 177 L.Ed.2d 162 (2010); Free Enterprise Fund, 561 U.S. 477, 130 S.Ct. 3138, 177 L.Ed.2d 706; Intercollegiate Broadcasting System, Inc., 684 F.3d 1332. Without major tumult, the agencies and courts have subsequently worked through the resulting issues regarding the legality of past rules and of past or current enforcement actions. See, e.g., Noel Canning v. NLRB, 823 F.3d 76, 78-80 (D.C. Cir. 2016); Intercollegiate Broadcasting System, Inc. v. Copyright Royalty Board, 796 F.3d 111, 118-19 (D.C. Cir. 2015). Because, as we will explain in the next section, the CFPB’s enforcement action against PHH in this case must be vacated in any event, we need not consider any such issues at this time.
. Section 8 of the Act is codified at 12 U.S.C. § 2607. For consistency, we refer to Section 8 rather than Section 2607.
. It is worth noting that Sections 8(a) and 8(c), as relevant here, do not speak directly to transactions between mortgage lenders and homebuyers. Instead, those two provisions speak to the transactions between the mortgage lender and mortgage insurer. The sections prohibit one specific kind of activity in that market: payment to the lender by the mortgage insurer for the lender's referral of a customer to the mortgage insurer.
Although not required by Section 8(c)(2), PHH nonetheless typically provided its borrowers with a disclosure. The disclosure said that if a borrower selected a mortgage insurer with which PHH had a referral arrangement, the insurer would pay a reinsurance fee to-Atrium, which was affiliated with PHH. See J.A. 332.
. When we use the phrase “reasonable market value” in this opinion, we use that phrase as shorthand for a payment that bears a reasonable relationship to the market value of the services performed or products provided, as HUD has long explained it. We do not opine on what constituted reasonable market value for the reinsurance at issue in this case. That factual question is not before us.
. Tying arrangements are rarely prohibited in the American economy, unless the party doing the tying has market power. Otherwise, tying arrangements can be beneficial to consumers and the economy by enhancing efficiencies and lowering costs. As the Supreme Court has stated, "Many tying arrangements ... are fully consistent with a free, competitive market.” Illinois Tool Works, Inc. v. Independent Ink, Inc., 547 U.S. 28, 45, 126 S.Ct. 1281, 164 L.Ed.2d 26 (2006); see also National Fuel Gas Supply Corp. v. FERC, 468 F.3d 831, 840 (D.C. Cir. 2006). In this context, moreover, the Real Estate Settlement Procedures Act allows vertical integration of lenders and other settlement service providers under its affiliated business provisions. If such vertical integration is allowed, it would not make much sense to conclude that similar vertical contractual relationships are proscribed.
. If a mortgage insurer did pay more than reasonable market value for reinsurance, the disgorgement remedy is the amount that was paid above reasonable market value.
. In tiie criminal context, Article I's two Ex Post Facto Clauses bar retroactive criminal statutes. That principle is so fundamental to the protection of individual liberty that the Framers included it in the original Constitution, and made it applicable against both the National and State governments. See U.S. Const, art. I, § 9, cl. 3; id. art. I, § 10, cl. 1. The Framers well understood that a free society could not function if retroactive punishment were tolerated. See id.) see also Landgraf v. USI Film Products, 511 U.S. 244, 266-67, 114 S.Ct. 1483, 128 L.Ed.2d 229 (1994); The Federalist No. 84, at 511-12 (Alexander Hamilton) (Clinton Rossiter ed., 1961) ("[T]he subjecting of men to punishment for things which, when they were done, were breaches of no law, and the practice of arbitrary imprisonments, have been, in all ages, the favorite and most formidable instruments of tyranny.”); cf. George Orwell, 1984, at 40 (1949) (“Day by day and almost minute by minute the past was brought up to date. ... [N]or was any item of news, or any expression of opinion, which conflicted with the needs of the moment, ever allowed to remain on record.”),
. To be clear, Section IV-A and Section IVB of tills opinion represent alternative holdings on the question of whether the CFPB permissibly determined that PHH violated Section 8. As alternative holdings, both holdings constitute binding precedent of the Court. See Association of Battery Recyclers, Inc. v. EPA, 716 F.3d 667, 673 (D.C. Cir.2013).
. Proving that the,mortgage insurer paid more than reasonable market value—and thus made a disguised payment for the referral—is an element of the Section 8 offense that the CFPB has the burden of proving by a preponderance of the evidence. See 12 C.F.R. § 1081.303(a) (2016); see also Director, Office of Workers' Compensation Programs, Department of Labor v. Greenwich Collieries, 512 U.S. 267, 271, 276, 114 S.Ct. 2251, 129 L.Ed.2d 221 (1994) (APA's usé of "burden of proof” in 5 U.S.C. § 556 places both burden of persuasion and burden of production on proponent of order); 12 U.S.C. § 5563(a) (CFPB is authorized to conduct adjudication proceedings "in the manner prescribed by *50chapter 5 of title 5,” which includes Administrative Procedure Act burden of proof requirements in 5 U.S.C. § 556). The CFPB characterizes this issue as an affirmative defense. That is wrong. If there were express payments in exchange for referrals in this case, and PHH was trying to argue that the payments nonetheless were justified under some exception, that might potentially fit within the affirmative defense box. But here, there were no such express payments in exchange for referrals. It is the CFPB's burden to prove that the payments for reinsurance were more than reasonable market value and were disguised payments for referrals.
. Similarly, for actions the CFPB brings in court under any of the 18 pre-existing consumer protection statutes, the CFPB may only “commence, defend, or intervene in the action in accordance with the requirements of that provision of law, as applicable.” 12 U.S.C. § 5564(g)(2)(B).
. In full, Section 2614 provides: "Any action pursuant to the provisions of section 2605, 2607, or 2608 of this title may be brought in the United States district court or in any other court of competent jurisdiction, for the district in which the property involved is located, or where the violation is alleged to have occurred, within 3 years in the case of a violation of section 2605 of this title and 1 year in the case of a violation of section 2607 or 2608 of this title from the date of the occurrence of the violation, except that actions brought by the Bureau, the Secretary, the Attorney General of any State, or the insurance commissioner of any State may be brought within 3 years from the date of the occurrence of the violation.” 12 U.S.C. § 2614. Note that the referenced Section 2607 of Title 12 is Section 8 of the Real Estate Settlement Procedures Act.
. We do not here decide whether each alleged above-reasonable-market value payment from the mortgage insurer to the reinsurer triggers a new three-year statute of limitations for that payment. We leave that question for the CFPB on remand and any future court proceedings.
, Accordingly, I concur in Parts I, IV and V of the majority opinion.