FCC v. Consumers' Research, 606 U.S. ___ (2025)

NOTICE: This opinion is subject to formal revision before publication in the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, pio@supremecourt.gov, of any typographical or other formal errors.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 24–354 and 24–422

_________________

FEDERAL COMMUNICATIONS COMMISSION, et al., PETITIONERS

24–354v.

CONSUMERS’ RESEARCH, et al.

SCHOOLS, HEALTH & LIBRARIES BROADBAND COALITION, et al., PETITIONERS

24–422v.

CONSUMERS’ RESEARCH, et al.

on writs of certiorari to the united states court of appeals for the fifth circuit

[June 27, 2025]

Justice Kagan delivered the opinion of the Court.

Nearly a century ago, Congress charged the then-new Federal Communications Commission (FCC or Commission) with making communications services available, at affordable prices, to all Americans. That objective became known as “universal service.” Some decades on, near the turn of the 21st century, Congress reaffirmed its commitment to universal service while providing new and more detailed instructions to the FCC about how to achieve it. Under the amended statutory plan, the FCC would use required payments, called contributions, from telecommunications companies to subsidize basic communications services for consumers in certain underserved communities—particularly, rural and low-income areas. To carry out that mandate, the Commission established discrete subsidy programs for the consumers Congress had identified, set up a special fund to receive and disburse the companies’ payments, and enlisted a private corporation, called the Universal Service Administrative Company, to help manage that fund’s operations.

The question in this case is whether the universal-service scheme—more particularly, its contribution mechanism—violates the Constitution’s nondelegation doctrine, either because Congress has given away its power to the FCC or because the FCC has given away its power to a private company. We hold that no impermissible transfer of authority has occurred. Under our nondelegation precedents, Congress sufficiently guided and constrained the discretion that it lodged with the FCC to implement the universal-service contribution scheme. And the FCC, in its turn, has retained all decision-making authority within that sphere, relying on the Administrative Company only for non-binding advice. Nothing in those arrangements, either separately or together, violates the Constitution.

I

A

The Communications Act of 1934, ch. 652, 48Stat. 1064, established the FCC and empowered it to regulate communications services. In the Act’s very first provision, Congress instructed the FCC to pursue the goal now called universal service. The FCC, Congress stated, was “to make available, so far as possible, to all the people of the United States,” reliable communications services “at reasonable charges.” 47 U. S. C. §151.

The universal-service project arose from the concern that pure market mechanisms would leave some segments of the population without access to needed communications services. That is because providers of those services, also called carriers, can reap greater profits from some classes of customers than from others. Carriers, for example, make more money in urban areas than in rural ones because fixed costs in cities are spreadable over many more users. See S. Benjamin et al., Telecommunications Law and Policy §15.3.1, p. 763 (2d ed. 2006); In re Federal-State Joint Bd. on Universal Serv., 12 FCC Rcd. 8776, 8784 (1997) (Universal Service Order). Similarly, carriers may prefer to focus on business customers instead of residential or not-for-profit customers (like schools and libraries) because the former are willing to pay more for the same services. See id., at 8784. So carriers have incentives to neglect some kinds of customers in providing services or setting prices. See P. Huber, M. Kellogg, & J. Thorne, Federal Telecommunications Law §6.1.2, pp. 6–9, 6–10, and n. 26 (3d ed. Supp. 2022). The result, policymakers thought, would be severe inequities in access to communications systems, and a swiss-cheese-like communications network for the whole country. See ibid.; Universal Service Order, 12 FCC Rcd., at 8780–8781, 8783.

Under the original Act, the FCC addressed that concern—and promoted universal service—through a “patchwork quilt of implicit and explicit subsidies.” FCC, Report to Congress 5 (FCC 98–67, 1998). The earliest and dominant squares of that quilt were implicit subsidies, provided under the FCC’s authority to set “just and reasonable” rates. See Universal Service Order, 12 FCC Rcd., at 8784; 47 U. S. C. §201(b). Carriers then operated as regulated monopolies, with the FCC (and its state counterparts) superintending the rates they could charge different kinds of customers for different services. See Universal Service Order, 12 FCC Rcd., at 8784–8785; Huber, Federal Telecommunications Law §6.1.1.2, p. 6–7. And regulators used that authority over rates to shift costs: Long-distance rates were set enough above cost “to subsidize local rates, business rates to subsidize [non-business] rates, and urban rates to subsidize rural rates.” Id., p. 6–7. So, for example, a “large New York [City] brokerage firm making heavy use of long-distance service” would end up “subsidiz[ing] the local services of homeowners in Fishkill.” Ibid.; see Universal Service Order, 12 FCC Rcd., at 8784. Over time, though, towns like Fishkill benefited as well from explicit subsidy programs. The FCC, for example, began in the 1980s to levy assessments on long-distance carriers and disburse the proceeds to carriers serving high-cost communities. See id., at 8890–8892; 47 CFR §§36.601–36.641, 69.116(a) (1996). And the FCC began, through the so-called Lifeline program, to make payments directly reducing low-income consumers’ monthly phone bills. See Universal Service Order, 12 FCC Rcd., at 8957–8958; §§69.104(j)–(l), 69.117, 69.203(f )–(g).

In 1996, Congress overhauled the Act to “promote competition and reduce regulation” in the telecommunications sector. See Telecommunications Act, 110Stat. 56. As part of those reforms, Congress created a new framework for achieving universal service. The amended Act discarded the implicit subsidies embedded in ratemaking and substituted a plan for explicit transfer payments to ensure that basic communications services extend across the country. See §254; Universal Service Order, 12 FCC Rcd., at 8783–8784.

Section 254 of the amended statute requires every carrier providing interstate telecommunications services to “contribute,” in line with the statute and FCC rules, to a fund designed to “preserve and advance universal service.” §254(d). The FCC must use the money in that fund, now known as the Universal Service Fund, to pay for subsidy programs for designated populations and facilities needing improved access. See §§254(a), (e); Wisconsin Bell, Inc. v. United States ex rel. Heath, 604 U. S. ___, ___ (2025) (slip op., at 2). The statute, for example, continues the Lifeline program aiding low-income individuals. See §254(j). It directs the FCC to provide assistance to rural hospitals, as well as to schools and libraries. See §254(h)(1). And it instructs the FCC to expand communications access for consumers in “rural” and other “high cost areas.” §254(b)(3). The carrier contributions collected to support those programs, the Act further provides, must be “sufficient” to carry them out and so to “advance universal service.” §§254(d)–(e).

The statute also provides detailed guidance for identifying the specific communications services to which the statute’s beneficiaries should have access. On the one hand, the Act recognizes that those services, given the expected pace of technological change, are unlikely to stay static: Universal service, says the statute, is “an evolving level of telecommunications services that the Commission shall establish periodically” as it accounts for “advances in telecommunications and information technologies and services.” §254(c)(1). On the other hand, the Act specifies the relevant criteria in every period. In deciding which communications services the “definition” of universal service encompasses, the FCC “shall consider the extent to which” a service (1) is “essential to education, public health, or public safety”; (2) has, through market forces, “been subscribed to by a substantial majority of residential customers”; and (3) is in fact “being deployed in public telecommunications networks by telecommunications carriers.” §§254(c)(1)(A)–(C). So too, the Commission must evaluate whether a service can be made available at an “affordable rate[ ].” §254(b)(1). Congress thus struck a balance in establishing universal service’s metes and bounds—affording the FCC latitude to adapt to technological developments, but insisting that the FCC always look to whether services are essential, affordable, and widely used.

Echoing the provisions just described, Congress also listed six “principles” on which the FCC “shall base” all its universal-service policies. §254(b). First, “[q]uality services should be available at just, reasonable, and affordable rates.” §254(b)(1). Second, “all regions of the Nation” should have access to those services. §254(b)(2). Third, all consumers, “including low-income consumers and those in rural, insular, and high cost areas,” should have access to services that are “reasonably comparable” in quality and price to those in urban areas. §254(b)(3). Fourth, every carrier should make “an equitable and nondiscriminatory contribution” to the achievement of universal service. §254(b)(4). Fifth, the subsidies given to advance that goal should be “specific, predictable[,] and sufficient.” §254(b)(5). And sixth, “schools,” “libraries,” and “health care providers” should have access to services. §254(b)(6). That list concludes with a provision enabling the FCC to add “other principles” found both “consistent with” the Act and “necessary and appropriate for the protection of the public interest, convenience, and necessity.” §254(b)(7).

The Commission now operates, under the terms of the Act, four universal-service programs. Lifeline, which predated the Act, gives low-income consumers a markdown on their monthly phone bills, usually of $9.25 per month. See §254(j); 47 CFR §§54.400–54.424 (2024); supra, at 4. The High Cost program, which similarly built on a pre-1996 explicit subsidy program, furnishes carriers with funds enabling them to provide basic communications services in “rural, insular, and [other] high cost areas” at costs comparable to those charged in urban areas. §254(b)(3); §§54.302–54.322; supra, at 4. The E-Rate program subsidizes communications services for schools and libraries across the country, with greater discounts going to facilities in rural or low-income areas. See §§254(b)(6), (h)(1)(B); §§54.500–54.523; Wisconsin Bell, 604 U. S., at ___ (slip op., at 3). And finally, the Rural Health program supports rural hospitals and other health care facilities, enabling them to use telemedicine in caring for far-flung patients and ensuring that they receive the needed communications services at roughly the rates urban facilities pay. See §§254(b)(6), (h)(1)(A); §§54.600–54.633.

To calculate how much carriers must contribute to the Fund for those programs, the FCC has devised a formula, known as the “contribution factor.” 47 CFR §54.709(a). That factor is a fraction, expressed as a percentage, whose numerator is the Fund’s projected expenses for the upcoming quarter (the subsidy payments it will make plus overhead) and whose denominator is the total projected revenue of contributing carriers during that same period. See §54.709(a)(2). A carrier must pay into the Fund an amount equal to its own projected revenue multiplied by the contribution factor. See §54.709(a)(3). So, for example, if the FCC forecasts that it will need, in a given quarter, 25% of all carrier revenues to cover the costs of its universal-service programs, a carrier expecting to generate $100 million in revenue in that quarter will have to contribute $25 million. The carrier may then pass along to its customers the cost of its contributions. See §54.712(a). Every quarter, the FCC updates its expense and revenue projections, comes up with a new contribution factor, and announces it in a public notice. See §§54.709(a)(2)–(3). During the next 14 days, the FCC may revise the factor as it thinks proper. See §54.709(a)(3). If the FCC takes no action within that period, the factor is “deemed approved by the Commission” and goes into effect. Ibid. Carriers must then kick in to the Fund accordingly.

The FCC in 1998 appointed the Universal Service Administrative Company as the Fund’s “permanent Administrator.” §54.701(a). The Administrator, as we will call it, is a private, not-for-profit corporation owned by an association of carriers. See §54.5; Wisconsin Bell, 604 U. S., at ___ (slip op., at 2). It manages the day-to-day operations of the Fund, “bill[ing] and collect[ing] contributions from carriers” and “distribut[ing] the resulting pot of money, as FCC rules provide, to program beneficiaries.” Id., at ___ (slip op., at 2); see §54.702. More relevant here, the Administrator plays a role each quarter in producing the financial projections that end up determining the contribution factor. See §§54.709(a)(2)–(3). Specifically, the Administrator estimates the Fund’s expenses (again, the costs of the programs plus overhead) and adds up the estimates it receives from individual carriers about their revenues. See ibid. The Administrator then submits those figures, along with supporting documentation, to the Commission for approval and eventual use in calculating required contributions. See ibid.

B

In December 2021, the FCC proposed a contribution factor of 25.2% for the first quarter of 2022. Respondents—the non-profit organization Consumers’ Research, a carrier, and several consumers (collectively, Consumers’ Research)—filed comments (during the 14-day, post-notice period described above) requesting that the FCC instead set the contribution factor at 0%. In support of that submission, Consumers’ Research argued that the universal-service contribution scheme violates the Constitution’s nondelegation rule. The Commission took no action in response, so the 25.2% contribution factor went into effect.

Consumers’ Research then petitioned for review in the Court of Appeals for the Fifth Circuit. The en banc court granted the petition, replacing a panel decision to the contrary. See 109 F. 4th 743 (2024); 63 F. 4th 441 (2023). In the full Fifth Circuit’s view, the universal-service contribution mechanism is unconstitutional because of its so-called “double-layered delegation.” 109 F. 4th, at 782.

The court’s analysis proceeded by expressing “skeptic[ism]” about each of two aspects of the contribution scheme, while declining to rule on either one. Id., at 778. First, the court stated, Congress in Section 254 “may have delegated legislative power” to the FCC by giving it “the power to tax” carriers “without supplying an intelligible principle to guide [its] discretion.” Id., at 756. The court described Section 254’s limits on the Commission’s authority as “minimal,” “contentless,” and “amorphous.” Id., at 760, 761, 767. Nonetheless, the court decided not to decide whether Congress had impermissibly transferred authority to the Commission. Id., at 767. Second, the court continued, the Commission “may have impermissibly delegated the taxing power to private entities” by involving the Administrator in setting contribution amounts. Id., at 756. The court posited that the FCC had “de facto abdicate[d]” governmental responsibilities to the Administrator by giving it the “final say” on how much carriers pay into the Fund. Id., at 771. Again, however, the court demurred as to the bottom line, reserving judgment on whether the Administrator’s role in the contribution scheme violates the Constitution. See id., at 778.

The dispositive constitutional problem, the Fifth Circuit ultimately held, is “the combination of Congress’s sweeping delegation to FCC and FCC’s unauthorized subdelegation” to the Administrator. Ibid. (emphasis in original). Relying heavily on this Court’s decision in Free Enterprise Fund v. Public Company Accounting Oversight Bd., 561 U.S. 477 (2010), the court opined that “two or more things that are not independently unconstitutional can combine to violate the Constitution’s separation of powers.” 109 F. 4th, at 778 (emphasis in original). And that was true here for a pair of reasons. The “double-layered delegation,” the court thought, had “no foothold in history or tradition.” Id., at 782. And, the court went on, that delegation “undermine[s] democratic accountability” by obscuring whether Congress, the FCC, or the Administrator bears responsibility for the amount of contributions. Id., at 783–784.

We granted certiorari, 604 U. S. ___ (2024), and now reverse the decision below.[1] In this Court, Consumers’ Research separates what the Fifth Circuit combined. It contends (as it did below) that Congress’s delegation to the FCC violates the Constitution, and that the FCC’s delegation to the Administrator does so too. We reject each argument, and also reject the Fifth Circuit’s combination theory.

II

Article I of the Constitution provides that “[a]ll legislative Powers herein granted shall be vested in a Congress of the United States.” §1. Accompanying that assignment of power to Congress is a bar on its further delegation: Legislative power, we have held, belongs to the legislative branch, and to no other. See Whitman v. American Trucking Assns., Inc., 531 U.S. 457, 472 (2001). At the same time, we have recognized that Congress may “seek[ ] assistance” from its coordinate branches to secure the “effect intended by its acts of legislation.” J. W. Hampton, Jr., & Co. v. United States, 276 U.S. 394, 406 (1928). And in particular, Congress may “vest[ ] discretion” in executive agencies to implement and apply the laws it has enacted—for example, by deciding on “the details of [their] execution.” Ibid.; see Wayman v. Southard, 10 Wheat. 1, 46 (1825) (“[T]he maker of the law may commit something to the discretion of the other departments”); Whitman, 531 U. S., at 474–475 (A “degree of policy judgment” can “be left to those executing or applying the law”).

To distinguish between the permissible and the impermissible in this sphere, we have long asked whether Congress has set out an “intelligible principle” to guide what it has given the agency to do. J. W. Hampton, 276 U. S., at 409. Under that test, “the degree of agency discretion that is acceptable varies according to the scope of the power congressionally conferred.” Whitman, 531 U. S., at 475. The “guidance” needed is greater, we have explained, when an agency action will “affect the entire national economy” than when it addresses a narrow, technical issue (e.g., the definition of “country [grain] elevators”). Ibid. But in examining a statute for the requisite intelligible principle, we have generally assessed whether Congress has made clear both “the general policy” that the agency must pursue and “the boundaries of [its] delegated authority.” American Power & Light Co. v. SEC, 329 U.S. 90, 105 (1946). And similarly, we have asked if Congress has provided sufficient standards to enable both “the courts and the public [to] ascertain whether the agency” has followed the law. OPP Cotton Mills, Inc. v. Administrator of Wage and Hour Div., Dept. of Labor, 312 U.S. 126, 144 (1941). If Congress has done so—as we have almost always found—then we will not disturb its grant of authority.

A

Although the intelligible-principle standard has focused our nondelegation doctrine for a century, Consumers’ Research and the dissent primarily argue that we must apply a different test here. Section 254, as just described, authorizes the Commission to raise revenue in the form of carrier “contribut[ions]” for universal-service programs. §254(d); see supra, at 4. Consumers’ Research views those required contributions as taxes. See Brief for Respondents 25–29; see also post, at 12–13, 29–30 (Gorsuch, J., dissenting). And it argues that tax statutes—and probably all revenue-raising statutes—have to satisfy a special nondelegation rule. For those statutes, Congress must set a “definite” or “objective limit” on how much money an agency can collect—a numeric cap, a fixed rate, or the equivalent. Brief for Respondents 32, 36; see id., at 33–35; see also post, at 14–16 (stating that a “tax rate” is likely required, but a “cap” may also suffice). Without such a limit, Consumers’ Research claims, no intelligible principle (however constraining) will do. See Brief for Respondents 46, 66. And all agree that Section 254 contains no determinate cap or formula. So, on Consumers’ Research’s telling, it effects an unconstitutional delegation.[2]

But this Court’s precedents foreclose that argument. Twice before, we have rejected a party’s request to create a special nondelegation rule for revenue-raising legislation. In J. W. Hampton, a taxpayer contended that when Congress is “exercis[ing] the power to levy taxes and fix customs duties,” it lacks the usual leeway to confer discretion on agencies. 276 U. S., at 409. “The [legal] authorities,” the Court responded, “make no such distinction.” Ibid. And then the Court set out the “intelligible principle” standard as the universal method for assessing delegations. Ibid. More than sixty years later, we reiterated that holding. In Skinner v. Mid-America Pipeline Co., 490 U.S. 212, 220 (1989), another litigant urged that “Congress’ taxing power” can be delegated only “with much stricter guidelines” than are normally used. Citing J. W. Hampton, we again—and unanimously—rejected that “two-tiered theory of nondelegation.” 490 U. S., at 220–221. “[N]othing” in the Constitution’s text or structure, Skinner explained, “distinguish[es] Congress’ power to tax from its other enumerated powers” “in terms of the scope and degree of discretionary authority that Congress may delegate to the Executive.” Ibid. Nor, the Court added, did history at all distinguish the two. See id., at 221 (“From its earliest days to the present, Congress, when enacting tax legislation,” has at times delegated “discretionary authority” to the Executive). So whether or not a tax is at issue—so say our cases—the usual nondelegation standard applies. And that standard is, again, trained on intelligible principles, not on numeric caps and “mathematical formula[s].” United States v. Rock Royal Co-operative, Inc., 307 U.S. 533, 577 (1939); see supra, at 11.[3]

The alternative test Consumers’ Research and the dissent propose also would throw a host of federal statutes into doubt. Relying on this Court’s nondelegation precedents, Congress has often enacted statutes empowering agencies to raise revenue without specifying a numeric cap or tax rate. See Reply Brief for Federal Petitioners 7–9 (listing nine “example[s]”). Indeed, such statutes are endemic in the sphere of financial regulation. The Federal Reserve Board, for instance, funds its operations by levying on Federal Reserve Banks “an assessment sufficient to pay its estimated expenses.” 12 U. S. C. §243. The Office of the Comptroller of the Currency (OCC) likewise “collect[s]” from OCC-chartered banks an “assessment, fee, or other charge” as the Office “determines is necessary or appropriate to carry out [its] responsibilities.” §16. And the Deposit Insurance Fund of the Federal Deposit Insurance Corporation (FDIC), which protects the savings of millions of Americans, is financed through charges on banks “which the Corporation may by regulation prescribe, after giving due consideration to the need to establish and maintain the [Fund’s] reserve ratio.” §1815(d)(1). In none of those (or many other) revenue-raising statutes does a number appear. So all would be on the constitutional chopping block under Consumers’ Research’s reasoning.

Consumers’ Research is conflicted about how to approach that problem, but sometimes tries to draw a line between taxes and fees. Our decision in National Cable Television Assn., Inc. v. United States, 415 U.S. 336 (1974) (NCTA), recognized that distinction, as Consumers’ Research notes. See Brief for Respondents 28. We there described “fees” as “bestow[ing]” a reciprocal “benefit on the [payor], not shared by other members of society.” NCTA, 415 U. S., at 341. By contrast, “taxes” are expected to “inure[ ] to the benefit” of the wider public. Id., at 343. In its brief, Consumers’ Research argues that its numeric-cap standard applies to both taxes and fees: As to either, Congress’s delegation to an agency must include an “objective upper limit[ ].” Brief for Respondents 37; see id., at 36–38. At argument, however, Consumers’ Research relied on the tax vs. fee distinction to get out from under the long list of statutes its position places in jeopardy. Most of those statutes, it argued, involve not taxes but fees, where the charge reflects simply “the value of the benefit to the” payor. Tr. of Oral Arg. 134; see id., at 148. So, Consumers’ Research suggested, if we label carrier contributions “taxes” and then make only taxes subject to the numeric-limit requirement, there would be minimal fallout from adopting its position. See ibid. The dissent joins Consumers’ Research in pressing that argument. See post, at 25–30.

But the problems with going down that road are substantial. First and as already shown, precedent forecloses it. Indeed, in rejecting a stricter test for delegations made “under Congress’ taxing power,” Skinner specifically noted that its position rendered irrelevant the question (which, we noted, “so exercised the District Court”) whether the charges there were “user fees” or a “form of taxation.” 490 U. S., at 223. Either way, the Court held, the delegation inquiry was just the same, and just the usual one. See ibid.[4] Second, Consumers’ Research offers no argument for why categorizing something as a fee rather than a tax should matter for delegation purposes. To the contrary, its brief suggests the difference should make no difference—that instead all revenue-raising measures should be treated the same. Brief for Respondents 36–38; Tr. of Oral Arg. 132–133 (repeating the point). The distinction is proposed only as an artificial method for limiting the effects of a holding in Consumers’ Research’s favor, should the court be too squeamish to go all the way.

And third, the distinction between taxes and fees, even if occasionally needed, is a morass—or as the Government (which levies both) puts it, “unbelievably murky in practice.” Tr. of Oral Arg. 52. A charge is a “fee,” according to NCTA, when it is for a “benefit” granted to the payor that is “not shared by other members of society.” 415 U. S., at 341. But articulating that test is a fair bit easier than applying it, because it is often hard to say whether a benefit is so “shared.” Consider the charge that banks pay to obtain FDIC Deposit Insurance. See supra, at 14. The banks benefit from that insurance (it helps them attract money and prevents bank runs). So maybe the charge looks like a fee? But then again, depositors also benefit (because the insurance protects their money) and so does the wider public (because everyone gains from having a stable banking system). So maybe the charge instead looks like a tax? Or take another example: the OCC’s assessments on banks. See ibid. Are they fees because a bank must pay them to hold an OCC-issued charter? Or are they taxes because the OCC uses the funds collected for regulatory programs benefiting the public? Consumers’ Research does not know. See Tr. of Oral Arg. 134 (describing the OCC statute as “kind of on the line,” “tough,” and “maybe . . . questionable”).[5]

Or finally and most relevantly: What category do carriers’ contributions belong in? Universal service is of course a public benefit. But carriers gain in tangible ways from having an all-inclusive network, and they often receive direct subsidies from the FCC’s universal-service programs. For those reasons, the carriers’ main trade associations view the contributions as fees. See Tr. of Oral Arg. 76, 80. The Government, by contrast, sees “genuine ambiguity” on the issue, but “assum[es]” they are taxes. Id., at 52–53. It is a good thing for the state of the law that we do not have to decide between the two, in this case or others raising a delegation challenge.[6]

And yet a greater problem inheres in the shared position of Consumers’ Research and the dissent: Whatever it applies to (just taxes or fees as well), its focus on numeric limits produces absurd results, divorced from any reasonable understanding of constitutional values. Under that view, a revenue-raising statute containing non-numeric, qualitative standards can never pass muster, no matter how much guidance those standards provide and how tight the constraints they impose. But a revenue-raising statute with a numeric limit will always pass muster, even if it effectively leaves an agency with boundless power. Consider a hypothetical raised at oral argument: Congress tells the FCC it can demand payments from carriers of any amount it wants up to $5 trillion. (The actual cost of universal service is, of course, a tiny fraction of that amount.) According to Consumers’ Research, that statute is permissible because . . . well, because Congress has set the $5 trillion figure. See id., at 124 (“[T]hen we would know that Congress itself has made that determination”); see id., at 123–127; Brief for Respondents 5, 63, 66; see also post, at 34–35. But so what? The purpose of the nondelegation doctrine is to enforce limits on the “degree of policy judgment that can be left to those executing or applying the law.” Mistretta v. United States, 488 U.S. 361, 416 (1989) (Scalia, J., dissenting). The anywhere-up-to-$5-trillion tax statute would not do that, whereas a statute with qualitative limits well might. In approving the former and precluding the latter, the Consumers’ Research approach does nothing to vindicate the nondelegation doctrine or, more broadly, the separation of powers.

B

We therefore return to the usual intelligible-principle test to decide whether the universal-service contribution scheme violates the Constitution’s nondelegation rule. The question is, again, whether Section 254 adequately guides the FCC in requiring contributions from carriers—whether it expresses the “general policy” the FCC must pursue in setting contribution amounts, as well as the “boundaries” it cannot cross. American Power & Light, 329 U. S., at 105. Here, that inquiry into the nature of the FCC’s discretion involves what turn out to be two closely related questions. First, how much money can the FCC raise through contributions? And second, on what things can it spend those funds? We consider each in turn, and find that Congress answered both. Congress, that is, imposed ascertainable and meaningful guideposts for the FCC to follow when carrying out its delegated function of collecting and spending contributions from carriers.

1

As Consumers’ Research notes, Section 254 imposes no quantitative but only qualitative limits on how much money the FCC can raise from carriers for universal service. There is not a number or a rate in sight. Instead, the statute directs the FCC to collect the amount that is “sufficient” to support the universal-service programs Congress has told it to implement. §§254(b)(5), (d), (e). That language replicates or resembles the statutory terms Congress has used in other revenue-raising statutes, as described above. See supra, at 14; see, e.g., 12 U. S. C. §243 (instructing the Federal Reserve Board to levy on banks “an assessment sufficient to pay its estimated expenses”).

Consumers’ Research argues that, even under our usual nondelegation test, the term “sufficient” does not do enough. That is because, in the Consumers’ Research view, it sets only “a floor—not a ceiling—on the FCC’s revenue-raising power.” Brief for Respondents 56. Or to put the point differently, Consumers’ Research thinks that the statute gives the FCC power, all on its own, to raise our hypothetical $5 trillion. See supra, at 18–19. And not unreasonably, it thinks that would pose a constitutional problem.

But in fact the word “sufficient” sets a floor and a ceiling alike. An amount of money is “sufficient” for a purpose if it is “[a]dequate” or “necessary” to achieve that purpose. Black’s Law Dictionary 1447 (7th ed. 1999). That means, of course, that the FCC cannot raise less than is adequate or necessary to finance the universal-service programs Congress wants. But it also means that the FCC cannot raise more than that amount. Were the FCC to raise, say, twice as much as needed, the revenue would not be “sufficient” but instead excessive. Cf. Whitman, 531 U. S., at 475–476 (similarly understanding the term “requisite” to mean “not lower or higher than is necessary”). Take another hypothetical from oral argument. If you told a friend to order a “sufficient” amount of food for five people and 500 boxes of pizza showed up at your house, you would not think he had followed instructions. See Tr. of Oral Arg. 135. So too with Congress and the Commission. Budgeting, to be sure, is not an exact science, so in one quarter the Commission may collect a bit more than it needs and in another a bit less. See 47 CFR §§54.709(b), (c) (telling the Administrator that if it collects “excess” or “inadequate” funds in a given quarter, it should compensate in the next one). But the Commission’s mandate is to raise what it takes to pay for universal-service programs; if the Commission raises much beyond, as if it raises much below, it violates the statute.

And the Commission has long viewed the statute in just that way. For many years, the Commission has construed the sufficient-funding directive to call for raising “an affordable and sustainable amount of support that is adequate, but no greater than necessary, to achieve the goals of the universal service program.” In re High-Cost Universal Serv. Support, 25 FCC Rcd. 4072, 4074 (2010); see Tr. of Oral Arg. 4 (Solicitor General explaining that Congress has authorized the FCC to collect “only what’s sufficient to achieve universal service, so no more than needed to support specified programs”). The Commission, in other words, sets the contribution factor to raise just enough money—a “sufficient” amount—to implement universal service as Congress directed.

2

To say that much, though, takes us only halfway, because it raises the question: Sufficient for what? If Section 254’s universal-service program is itself indeterminate—so that the FCC can turn it into anything the FCC wants—then the “sufficiency” ceiling will do no serious work. The FCC could operate—and collect contributions “sufficient” for—either the most barebones or the most extravagant program. But if Congress has given appropriate guidance about the nature and content of universal service, then that plus the “sufficiency” ceiling will defeat this challenge to the contribution system. For Congress will have provided intelligible principles to guide the FCC as it raises funds.

On this further, “for what” question, our nondelegation precedents provide context—showing what kinds of statutory schemes have passed, and what kinds have failed, the demand that Congress give adequate guidance. Those that have failed are fewer in number—in fact, only two—but offer object lessons about the amount of latitude Congress can confer. In one case, the statute empowered the President to bar the transport of petroleum products while “establish[ing] no criterion” and “declar[ing] no policy” for whether, when, or how he should do so. Panama Refining Co. v. Ryan, 293 U.S. 388, 415 (1935). The statute in the second case was even worse. It authorized the President to approve “codes of fair competition” for “the government of trade and industry throughout the country,” yet imposed “few restrictions” and “set[ ] up no standards” aside from a “statement of the general aims of rehabilitat[ing], correct[ing,] and expand[ing]” the economy. A. L. A. Schechter Poultry Corp. v. United States, 295 U.S. 495, 521–522, 541–542 (1935). The law thus gave the President “virtually unfettered” authority to govern the Nation’s trades and industries. Id., at 542.

At the same time, we have found intelligible principles in a host of statutes giving agencies significant discretion. So, for example, we upheld a provision enabling an agency to set air quality standards at levels “requisite to protect the public health.” Whitman, 531 U. S., at 472. We sustained a delegation to an agency to ensure that corporate structures did not “unfairly or inequitably distribute voting power” among security holders. American Power & Light, 329 U. S., at 104. And we affirmed authorizations to regulate in the “public interest” and to set “just and reasonable” rates, because we thought the discretion given was not unbridled. See, e.g., National Broadcasting Co. v. United States, 319 U.S. 190, 225–226 (1943); FPC v. Hope Natural Gas Co., 320 U.S. 591, 600 (1944); see supra, at 3. Of course, our cases did not examine those statutory phrases in isolation but instead looked to the broader statutory contexts, which informed their interpretation and supplied the content necessary to satisfy the intelligible-principle test. See, e.g., National Broadcasting, 319 U. S., at 226 (“It is a mistaken assumption” that the phrase “public interest” is “a mere general reference to public welfare without any standard to guide determinations”; rather, “[t]he purpose of the Act, the requirements it imposes, and the context of the provision in question show the contrary”); see also infra, at 29.

Section 254, for its part, provides the FCC with determinate standards for operating the universal-service program. The statute makes clear whom the program is intended to serve: those in rural and other high-cost areas (with a special nod to rural hospitals), low-income consumers, and schools and libraries. See §§254(b)(3), (6), (h)(1); see supra, at 4–6. And in provisions defining universal service and stating the program’s core “principles,” the statute provides specific criteria for which services those statutory beneficiaries should receive. §§254(b), (c)(1). In deciding whether a service falls within the program’s ambit, the FCC must consider whether the service has “been subscribed to by a substantial majority of residential customers.” §254(c)(1)(B). If that objective criterion is not met, the FCC generally may not subsidize the service. So too, the service must be one that can be made available to all consumers in all regions at “reasonable[ ] and affordable rates”—so more a basic than a budget-busting good. §§254(b)(1), (3).[7] And still more, the service must be “essential to education, public health, or public safety”—a necessity, not a luxury, in order to live in the world. §254(c)(1)(A). The conditions, each alone and together, have bite, creating a bounded program. Section 254 instructs the Commission to provide to an identified set of recipients a defined sort of benefit—widely used, generally affordable, and essential telecommunications services.

That limited conception of universal service is rooted in its history—except that the new statute, as compared with the old, holds the FCC to more specific requirements. As earlier explained, the 1934 Act charged the FCC with pursuing universal service—that is, with making available to all Americans telecommunications services “at reasonable charges.” §151. As then understood, that objective did not reach for the stars: Though quite important, it was also “relatively modest.” Benjamin, Telecommunications Law and Policy §18.2, at 863. The idea was to “reduc[e] the costs of basic telephone service and, in that way, increas[e] national subscribership.” Ibid. By the time of the 1996 amendments, though, new telecommunications technologies and services had emerged. And so Congress enabled the FCC, in carrying out universal-service programs, to do more. See id., at 863–864; see, e.g., §§254(b)(6), (h)(1). But still the statute’s policy was a circumscribed one: to provide to all (especially, the rural and poor) the services that most already had, if those services were both necessary and affordable. The key difference between the original statute and the amended one is a matter of means, and on that score, the latter provides far greater congressional guidance. Under the 1934 Act, the FCC gave implicit (and often obscure) subsidies under its general (i.e., “just and reasonable”) ratemaking authority. See supra, at 3; see also post, at 3 (explaining that in the pre-1996 regime, “[r]egulators manipulated rates” to fund universal service). Under the 1996 Act, the FCC gives explicit (and transparent) subsidies in accord with the detailed criteria described above. See supra, at 4, 23. So today, when the FCC carries out Congress’s century-old commitment to universal service, the statutory policy is clear and the statutory boundaries specific.

The proof is in the pudding: Each of the four programs the FCC now operates under Section 254 reflects Congress’s choices about universal service’s scope and content. The Lifeline program, which began under the original Act, advances a basic commitment. Now codified, it helps make phone service affordable to all Americans by providing a modest monthly subsidy. See §254(j); 47 CFR §§54.400–54.424; supra, at 4, 6. The High Cost program similarly implements a longstanding principle—to integrate rural communities into the Nation’s communications network. See §151; §§54.302–54.322; supra, at 3–4, 6. And it does so now in accordance with the statutory directive to ensure that rural and other high-cost areas have access to roughly the same needed services, at the same affordable prices, as urban areas do. §254(b)(3). The Commission’s other two programs, E-Rate and Rural Health, are of a piece. Specifically authorized in the amended Act, they underwrite services essential to education and healthcare, with a focus on underserved populations. §254(h)(1); §§54.500–54.523, 54.600–54.633; supra, at 6–7. Not one of those important but decidedly ordinary programs suggests an agency vested with unbridled discretion. Each provides communications services satisfying the Act’s listed criteria to the Act’s identified beneficiaries. And maybe surprisingly (given what we are used to when it comes to government programs), they have done so at roughly constant inflation-adjusted dollars. Compare USAC, 2000 Annual Report 5 (charging carriers $7.4 billion in 1999 when so adjusted), with USAC, 2024 Annual Report 18 (charging carriers $8.4 billion in 2024).[8] That is because in the amended Act, Congress made clear the parameters of the programs, and the FCC has operated within them.[9]

Consumers’ Research and the dissent tell a different story—that the Act gives the FCC boundless authority—but the provisions they point to show nothing of the kind. They first pluck out a few words to argue that the criteria for subsidizing services, described above, are not “real limit[s].” Brief for Respondents 55; supra, at 23. On affordability, both contend that because Section 254(b) states that services “should”—rather than “shall”—be made available at “reasonable[ ] and affordable rates,” there is in fact no such requirement. Brief for Respondents 47–48; see post, at 20–21. But that reading starts in the middle. The provision, starting from the start, says that the FCC “shall” base all universal-service policies on the principle (among others) that services “should be available” at “reasonable[ ] and affordable rates.” §254(b)(1). The mandatory “shall” requires the FCC to follow the affordability principle—which means providing all services at “reasonable[ ] and affordable rates.” And similarly as to whether a service is widely used and essential. Per Consumers’ Research and the dissent, Section 254(c)(1) says only that the FCC must “consider” those criteria and thus establishes no more than a “weak[ ] procedural requirement.” Brief for Respondents 54; see post, at 6–7, 19–22. But the list of criteria the FCC “shall” consider resides in the very “definition” of the “services” it can subsidize. §254(c)(1). The statute, read sensibly, does not tell the FCC to muse on those criteria before developing its own. Rather, it tells the FCC to “consider,” as to any given service, whether it satisfies the listed criteria (and therefore can be subsidized). And that is, indeed, how the Solicitor General, representing the FCC, understands the provision. See Tr. of Oral Arg. 10, 27–29, 177–178.

The dissent therefore fails in its related claim that the FCC can balance different universal-service criteria against each other—so, for example, fund a service because it is “essential to education” even though it has not been adopted “by a substantial majority” of customers. Post, at 6–7; see post, at 18–19; supra, at 23. Again, the Solicitor General has represented in this Court that each of the criteria has to be met. Tr. of Oral Arg. 10, 27–29, 177–178. In any event, and yet more important, we must “exercise [our] independent judgment in deciding” what power Congress has conferred. Loper Bright Enterprises v. Raimondo, 603 U.S. 369, 412 (2024). And for all the reasons given above, we view the statutory criteria—which, contra the dissent, post, at 18, define universal service—as separately mandatory. See §254(c)(1); supra, at 23, 26–27. Of course, the Commission may still have to strike balances in addressing those criteria, along with the statute’s other provisions. It may, for example, have to decide whether to make a service more affordable (by giving a larger subsidy) or instead extend it to a broader swathe of recipients. But that kind of discretion—balancing or no—does not raise a constitutional problem: A “degree of policy judgment,” as we have explained, “can be left to those executing or applying the law.” Whitman, 531 U. S., at 474–475.

We likewise see no constitutional issue in Section 254(c)(1)’s description of universal service as an “evolving level of telecommunications services that the Commission shall establish periodically” in light of “advances in telecommunications and information” services. According to Consumers’ Research and the dissent, that language enables the FCC to “redefine universal service” over time as it and only it “sees fit.” Brief for Respondents 8, 54; see post, at 18. But Congress’s statement that universal service should “evolve” is itself a direction—and a near-inevitable one, given the reality of technological change. If universal service did not evolve—if Congress had defined it as, say, a landline in every home (or, as the dissent would have it, “touch-tone [phone] service,” post, at 19)—the program would have long since become obsolete. The Act’s embrace of evolution—the permission it gives the FCC to subsidize different services now than 30 years ago—ensures that the universal-service program will be of enduring utility. But that conferral of discretion does not strip the statute of standards and constraints. The Commission still may fund only essential, widely used, and affordable services, for the benefit of only designated recipients. See supra, at 23. So Congress has ensured that the Commission will continue to carry out the same objectives according to the same criteria and principles, even as it has allowed adaptation to a changing technological landscape.

Finally, we do not view as Consumers’ Research does the provision in Section 254 enabling the FCC to articulate “[a]dditional principles,” beyond the six listed, to guide its universal-service programs. §254(b)(7); see supra, at 6. Recall that the added principles are ones the FCC “determine[s] are necessary and appropriate for the protection of the public interest, convenience, and necessity and are consistent with this chapter.” See supra, at 6. As Consumers’ Research sees it, the FCC can, through devising those principles, “rewrite its own authority.” Brief for Respondents 50 (capitalization altered); see post, at 18. But that is not so because, again, the added principles must be “consistent with” the rest of the statute. They cannot change any of the statute’s other principles, much less its conditions on what subsidies can go toward and who can receive them. The new principles can only operate, within those statutory parameters, to further channel the FCC’s discretion. So they are a way to superimpose self-restraint on congressional restraint, which is hardly improper. And the provision’s broadly framed reference to the “public interest” suggests nothing to the contrary. The public-interest requirement lies on top of the consistency requirement—connected with an “and,” not an “or”—and anyway is complementary to it. For we have long held that “the words ‘public interest’ in a regulatory statute” do not encompass “the general public welfare” but rather “take meaning from the purposes of the regulatory legislation.” NAACP v. FPC, 425 U.S. 662, 669 (1976); see New York Central Securities Corp. v. United States, 287 U.S. 12, 24–25 (1932); see supra, at 22–23. So the whole of the “[a]dditional principles” provision supplies a means of further implementing, rather than dispensing with, Congress’s instructions.

In a sense, each of the arguments Consumers’ Research and the dissent make about Section 254 suffers from the same flaw. At every turn, they read Section 254 extravagantly, the better to create a constitutional problem. As earlier seen, “sufficient” means to Consumers’ Research as much as the FCC wants, a floor without a ceiling. See supra, at 20. The statute’s mandatory conditions on funding services are instead mere suggestions, for the FCC to observe or not as it chooses. See supra, at 26–27. The phrase “evolving level” of service licenses the FCC to create a whole new program, unhindered by the statute’s existing standards and boundaries. See supra, at 27–28. And the possibility of “[a]dditional principles” coming from the FCC somehow subverts the limiting principles Congress put on the FCC, so that everything about universal service is up for grabs. See supra, at 28–29. All in all, the arguments do not show statutory construction at its best. Nor, relatedly, do they show proper respect for a coordinate branch of Government. Statutes (including regulatory statutes) should be read, if possible, to comport with the Constitution, not to contradict it. See, e.g., Ashwander v. TVA, 297 U.S. 288, 348 (1936) (Brandeis, J., concurring); Industrial Union Dept., AFL–CIO v. American Petroleum Institute, 448 U.S. 607, 646 (1980) (plurality opinion); West Virginia v. EPA, 597 U.S. 697, 722–723 (2022). That disposition nowhere appears in the efforts Consumers’ Research and the dissent make to force Section 254 past the Constitution’s breaking point.

Properly understood, the universal-service contribution scheme clears the nondelegation bar. The policy it expresses is clear and limiting. If, says the statute, a substantial majority of Americans has access to a communications service that is both affordable and essential to modern life, then other Americans should have access to that service too. And to make that happen, the statute continues, carriers should kick in the needed funds. At bottom, that is all the contribution scheme challenged here accomplishes. Through that statutory mechanism, the FCC raises sufficient funds (neither more nor less than needed) to bring to underserved Americans, mainly in rural and low-income communities, a bounded and commonplace set of communications services. The FCC no doubt exercises significant discretion in carrying out that charge. But it is discretion tethered to legislative judgments about the scope and content of the universal-service program. And so the main delegation at issue here, from Congress to the Commission, does not offend the Constitution.

III

The next question Consumers’ Research raises is whether a different delegation, now from the Commission to the Administrator (which, recall, is a private, not-for-profit corporation), independently flouts a constitutional command. Here, Consumers’ Research invokes what is commonly called the private nondelegation doctrine. See Brief for Respondents 74–75. In the leading case of Carter v. Carter Coal Co., 298 U.S. 238, 310–311 (1936), this Court struck down a statute authorizing certain coal producers to set maximum hours and minimum wages for the rest of the industry. We explained that the statute involved “delegation in its most obnoxious form” because it was made to “private persons whose interests” are often “adverse to the interests of others.” Id., at 311; see Schechter Poultry, 295 U. S., at 537. Consumers’ Research contends that the FCC has in like manner conferred governmental power on a private party, by (in its description) giving the Administrator carte blanche to set the contribution factor, which then determines what individual carriers pay into the Fund. See Brief for Respondents 3–4, 75; supra, at 7–8.

Carter Coal, though, has a counterpart case, addressing how Government agencies may rely on advice and assistance from private actors. In Sunshine Anthracite Coal Co. v. Adkins, 310 U.S. 381, 388 (1940), this Court considered a statute, enacted in response to Carter Coal, permitting boards of coal companies to propose minimum coal prices to a Government agency for “approv[al], disapprov[al], or modifi[cation].” That arrangement, we held, was “unquestionably valid.” 310 U. S., at 399. After all, we explained, the private boards “function[ed] subordinately to” the agency and were subject to its “authority and surveillance.” Ibid. As long as an agency thus retains decision-making power, it may enlist private parties to give it recommendations.

Here, the Administrator is broadly subordinate to the Commission. The FCC appoints the Administrator’s Board of Directors and approves its budget. See 47 CFR §§54.703(b)–(c), 54.715(c). The Administrator “may not make policy,” and must carry out all its tasks “consistent with” the FCC’s rules, “orders, written directives, and other instructions.” §54.702(c); Memorandum of Understanding Between the Federal Communications Commission and the Universal Service Administrative Company 2 (Oct. 17, 2024) (Memorandum of Understanding). And anyone aggrieved by an action of the Administrator may seek de novo review by the Commission. §§54.719–54.725. So in the relationship between the two, the Commission dominates.

And critically, that is as true in determining the contribution factor as in other matters: Although the Administrator plays an advisory role, the Commission alone has decision-making authority. Recall that each quarter’s contribution factor is a function of the carriers’ projected revenues and the Fund’s projected expenses. See supra, at 7. The Administrator makes the initial projections. On the revenue side, that means just doing arithmetic: The carriers submit their projections on FCC forms and the Administrator adds them up. See §§54.709(a)(2)–(3), 54.711(a). On the expense side, the Administrator’s estimates involve considerably greater effort—but still no policy-making. The FCC’s rules implementing the Act dictate the programs’ scope: For example, they set eligibility criteria for beneficiaries, provide formulas for calculating subsidies, and impose some funding caps. See, e.g., §§54.410, 54.507, 54.604–54.606.[10] Working within those rules, the Administrator estimates the programs’ cost. It then publicly reports those projections, along with supporting documents, to the Commission—on the revenue side, at least 30 days before a quarter starts, on the expense side, at least 60. See §54.709(a)(3). That gives the Commission a chance to review—and, if needed, to revise—the projections before approving final figures. See ibid. When the review process is complete, the Commission sets the contribution factor and posts it in a public notice. See ibid. The Commission then has 14 days to make additional changes before the factor is “deemed approved.” Ibid. So the Commission is, throughout, the final authority—just as the agency was in Sunshine Anthracite. The Administrator, following the FCC’s rules, makes recommendations. But the Commission decides whether or how to use them in setting the contribution factor.

In contending otherwise, Consumers’ Research misunderstands the regulatory scheme. Its primary argument rests on the words “deemed approved” in the FCC’s regulations. Consumers’ Research takes that to mean that the Administrator’s projections can “take legal effect” just by the “deem[ing]” mechanism—that is, without receiving “formal FCC approval.” Brief for Respondents 80. But that account ignores everything that happens before the 14-day period following public notice. Prior to that time, the Commission reviews the Administrator’s projections, and either revises or approves them. Then, the Commission sets the contribution factor based on the vetted projections and issues it to the public. So the Administrator’s projections can have only the legal (or, indeed, practical) effect the Commission decides they should. Not the Administrator, but the Commission endorses final projections, converts them into a contribution factor, and formally promulgates them. At the end of all that action, the “deemed approved” provision just operates to shut off an additional two-week opportunity the Commission has to revise the published contribution factor—because something (including public comments like Consumers’ Research submitted) has caused it to change its mind. That provision does nothing to negate the Commission decision-making that has already taken place.

The alternative argument Consumers’ Research makes does not fare any better. Here, Consumers’ Research appears to concede that the FCC approves the projections going into the contribution factor; the problem instead is that the approval is too often automatic—simply “rubber-stamp[ing].” Id., at 82. But the relevant legal question is not how often the FCC revises the Administrator’s projections, just as in Sunshine Anthracite it was not how often the agency rejected the coal companies’ pricing advice. It is sufficient in such schemes that the private party’s recommendations (as is true here) cannot go into effect without an agency’s say-so, regardless of how freely given. See 310 U. S., at 399. This case suggests at least one reason why: It may not be clear what the ratio of approvals to rejections actually means. On the view of Consumers’ Research, the infrequency with which the Commission changes the Administrator’s publicly submitted projections shows that it simply is not paying attention. But an amicus brief submitted by former FCC Commissioners offers an alternative explanation—that the Administrator “informally shares its projections” with the Commission before it publicly submits them, so that much of the discussion between the two occurs behind the scenes. See Brief for Bipartisan Former Commissioners of the FCC as Amici Curiae 11; see also Memorandum of Understanding 7 (establishing that informal procedure). And yet a third account might suggest that the absence of frequent dispute reflects the limited role the Administrator performs in estimating the expenses of programs whose contours FCC regulations precisely define. See supra, at 32–33. The explanation, that is, would lie in the narrow scope of the assignment the FCC has given to the Administrator.

So the Commission’s transfer of accounting functions to the Administrator offers no reason for holding the universal-service contribution scheme invalid. In every way that matters to the constitutional inquiry, the Commission, not the Administrator, is in control.

IV

Consumers’ Research almost wholly ignores the basis of the decision below: that the “combination” of Congress’s grant of authority to the FCC and the FCC’s reliance on the Administrator for financial projections violates the Constitution, even if neither one does so alone. See 109 F. 4th, at 778 (emphasis in original). But because that theory accounts for the decision we are reviewing, we cannot close without addressing it briefly.

The Fifth Circuit, as noted earlier, founded its combination theory—that a constitutional non-violation plus a constitutional non-violation may equal a constitutional violation—on this Court’s decision in Free Enterprise Fund. See supra, at 9. There, we struck down a statute because it gave an executive officer two “layers of protection” from the President’s removal authority: The President was “restricted in his ability to remove a principal officer, who [was] in turn restricted in his ability to remove an inferior officer.” 561 U. S., at 483–484. Even granting that each layer of good-cause protection was alone permissible, we thought the combination was too much. The two together, more than either alone, insulated the officer from the President’s firing power, thus super-charging the officer’s “independence.” Id., at 496. That holding, in the Fifth Circuit’s view, gave rise to a “general principle”: “[T]wo constitutional parts do not necessarily add up to a constitutional whole.” 109 F. 4th, at 779. And the court thought that principle applied here. Even if Congress lawfully conferred discretion on the Commission and the Commission lawfully sought assistance from the Administrator, the combination was both “unprecedented” and “incompatible” with “democratic accountability.” Id., at 779, 783–784. So what the Fifth Circuit called “the universal service contribution mechanism’s double-layered delegation” had to go. Id., at 784.

But the court’s analogy and associated logic do not work. In Free Enterprise Fund, each of the two layers of for-cause protection limited the same thing—the President’s power to remove executive officers. And when combined, each compounded the other’s effect, so that the President was left with no real authority. Or otherwise said, the two layers of restrictions operated on a single axis with the one exacerbating (we thought exponentially) the other. But that reasoning has no bearing here. A law violates the traditional (or call it, for comparison’s sake, “public”) nondelegation doctrine when it authorizes an agency to legislate. And a law—whether a statute or, as here, a regulation—violates the private nondelegation doctrine when it allows non-governmental entities to govern. Those doctrines do not operate on the same axis (save if it is defined impossibly broadly). So a measure implicating (but not violating) one does not compound a measure implicating (but not violating) the other, in a way that pushes the combination over a constitutional line. “Two wrong claims do not make one that is right.” Pacific Bell Telephone Co. v. linkLine Communications, Inc., 555 U.S. 438, 457 (2009). If a regulatory scheme authorizes neither executive legislation nor private governance, it does not somehow authorize an unlawful amalgam. Contra the Fifth Circuit, a meritless public nondelegation challenge plus a meritless private nondelegation challenge cannot equal a meritorious “combination” claim.

And indeed Sunshine Anthracite as well as said so before. As earlier noted, that case involved a private nondelegation challenge—that a board of coal companies advising an agency played too great a role in setting industry prices. See 310 U. S., at 399; supra, at 31. In addition, the case involved a public nondelegation challenge—that even the agency could not set prices because Congress had failed to provide it with sufficient guidance. See 310 U. S., at 397–399. The Court discussed and rejected the one challenge; and then it discussed and rejected the other. See ibid. And then the Court stopped. It did not think some further “combination” analysis was required. That was because (1) an executive agency exercising only executive power, plus (2) a private entity exercising no government power (but merely giving advice) equals (3) a permissible constitutional arrangement.

V

When Congress amended the Communications Act in 1996, it provided the Commission with clear guidance on how to promote universal service using carrier contributions. Congress laid out the “general policy” to be achieved, the “principle[s]” and standards the FCC must use in pursuing that policy, and the “boundaries” the FCC may not cross. J. W. Hampton, 276 U. S., at 409; American Power & Light, 329 U. S., at 105. Our precedents do not require more. Nor do they prevent the Commission, in carrying out Congress’s policy, from obtaining the Administrator’s assistance in projecting revenues and expenses, so that carriers pay the needed amount. For nearly three decades, the work of Congress and the Commission in establishing universal-service programs has led to a more fully connected country. And it has done so while leaving fully intact the separation of powers integral to our Constitution.

We accordingly reverse the judgment of the Court of Appeals for the Fifth Circuit and remand for further proceedings consistent with this opinion.

It is so ordered.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 24–354 and 24–422

_________________

FEDERAL COMMUNICATIONS COMMISSION, et al., PETITIONERS

24–354v.

CONSUMERS’ RESEARCH, et al.

SCHOOLS, HEALTH & LIBRARIES BROADBAND COALITION, et al., PETITIONERS

24–422v.

CONSUMERS’ RESEARCH, et al.

on writs of certiorari to the united states court of appeals for the fifth circuit

[June 27, 2025]

Justice Kavanaugh, concurring.

This case presents a narrow but important nondelegation question: May Congress authorize the Federal Communications Commission to determine the monetary amount “sufficient” to fund certain telecommunications services, which in turn is the amount that telecommunications carriers must contribute to the Universal Service Fund? Applying the longstanding “intelligible principle” test set forth by this Court’s precedents, the Court today upholds that congressional delegation to the FCC. See Skinner v. Mid-America Pipeline Co., 490 U.S. 212 (1989); J. W. Hampton, Jr., & Co. v. United States, 276 U.S. 394 (1928).

I join the Court’s opinion and write separately to make two points. First, I will briefly outline what I understand to be the background and rationale behind the intelligible principle test that the Court has long used to assess congressional delegations of authority to the Executive Branch. Second, I will explain why congressional delegations to independent agencies—as distinct from delegations to the President and executive agencies—raise substantial questions under Article II of the Constitution.

I

A

From the start in 1789, Congress has delegated to the President the power to exercise discretion and policymaking authority when implementing legislation.[1] Those delegations have been a regular feature of American Government ever since.[2]

The Court has generally permitted such delegations. As to the text of the Constitution, the Court has rejected arguments that the President impermissibly wields legislative power when exercising discretion or policymaking authority delegated by Congress. Instead, the Court has reasoned that the President ordinarily exercises “executive Power” under Article II when implementing legislation—even if he employs discretion or policymaking authority when doing so and even if the Executive Branch issues legally binding regulations. See, e.g., Mistretta v. United States, 488 U.S. 361, 386, n. 14 (1989) (“[R]ulemaking power originates in the Legislative Branch and becomes an executive function only when delegated by the Legislature to the Executive Branch”); see also Loving v. United States, 517 U.S. 748, 777 (1996) (Scalia, J., concurring in part and concurring in judgment) (“What Congress does is to assign responsibilities to the Executive; and when the Executive undertakes those assigned responsibilities it acts, not as the ‘delegate’ of Congress, but as the agent of the People”); J. Manning, The Nondelegation Doctrine as a Canon of Avoidance, 2000 S. Ct. Rev. 223, 240, n. 90 (“The Court apparently believes that when a statute sets down an intelligible principle, the agency can be thought of as implementing legislative directions, rather than exercising legislative authority. . . . Under that view, the agency is engaged in law ‘execution,’ rather than receiving delegated legislative authority”).[3]

The history of congressional delegations and the Court’s understanding of Article II’s text correspond to what the Court has described as the practicalities of legislative and executive action. Congress delegates at least in part because it must adapt legislation to “complex conditions involving a host of details with which the national legislature cannot deal directly.” A. L. A. Schechter Poultry Corp. v. United States, 295 U.S. 495, 530 (1935); Panama Refining Co. v. Ryan, 293 U.S. 388, 421 (1935). And the Constitution “has never been regarded as denying to Congress the necessary resources of flexibility and practicality.” Schechter Poultry, 295 U. S., at 530; Panama Refining, 293 U. S., at 421. That flexibility enables Congress “to perform its function in laying down policies and establishing standards, while leaving to selected instrumentalities the making of subordinate rules within prescribed limits and the determination of facts to which the policy as declared by the legislature is to apply.” Schechter Poultry, 295 U. S., at 530; Panama Refining, 293 U. S., at 421. Even when the legislature might want to legislate more specifically in certain circumstances, a “certain degree of discretion, and thus of lawmaking, inheres in most executive . . . action.” Whitman v. American Trucking Assns., Inc., 531 U.S. 457, 475 (2001) (quotation marks omitted).

Although the Court has ruled that congressional delegations to the President are permissible as a matter of constitutional text and history, the Court has not said that “anything goes” with respect to those delegations. As Justice Gorsuch rightly says, Congress may not give “the President or an executive agency a blank check to legislate.” Post, at 12 (dissenting opinion). So “Members of Congress could not, even if they wished, vote all power to the President and adjourn sine die.” Mistretta, 488 U. S., at 415 (Scalia, J., dissenting). Congress likewise cannot merely assign the President to take over the legislative role as to a particular subject matter. See Schechter Poultry, 295 U. S., at 537–542; Panama Refining, 293 U. S., at 430. Rather, the Court has said, any congressional grant of authority must supply some guidance to the President—otherwise the President would no longer be exercising “executive Power” when implementing legislation.

But the question of where to draw that line can be difficult: At what point does a broad statutory delegation transform from (i) a permissible grant of discretion or policymaking authority for the President to exercise when implementing legislation into (ii) an impermissible delegation of legislative power? Justice Scalia phrased the issue this way: “Once it is conceded, as it must be, that no statute can be entirely precise, and that some judgments, even some judgments involving policy considerations, must be left to the officers executing the law . . . , the debate over unconstitutional delegation becomes a debate not over a point of principle but over a question of degree.” Mistretta, 488 U. S., at 415 (dissenting opinion).

To address that question of degree and ensure that the President is exercising executive power when implementing legislation, the Court in 1928 adopted the “intelligible principle” test. In its unanimous opinion in J. W. Hampton, Jr., & Co. v. United States, the Court speaking through Chief Justice (and former President) Taft stated: “If Congress shall lay down by legislative act an intelligible principle to which the person or body authorized to [act] is directed to conform, such legislative action is not a forbidden delegation of legislative power.” 276 U.S. 394, 409 (1928) (emphasis added). Rather, when implementing legislation that contains an intelligible principle, the President is exercising executive power. The inverse is also true: When Congress grants authority to the President without an intelligible principle to confine his action, Congress has impermissibly delegated legislative power, although the Court has found that to occur only “rarely.” Mistretta, 488 U. S., at 419 (Scalia, J., dissenting).

The intelligible principle test recognizes that “[a]t some point the responsibilities assigned can become so extensive and so unconstrained that Congress has in effect delegated its legislative power.” Loving, 517 U. S., at 777 (Scalia, J., concurring in part and concurring in judgment). But “until that point of excess is reached there exists . . . no delegation” of legislative power “at all.” Ibid.

For 97 years, the intelligible principle test set forth in J. W. Hampton has formed the foundation of the Court’s nondelegation doctrine. Under the test, as then-Justice Rehnquist succinctly framed it, Congress may “lay down the general policy and standards that animate the law, leaving the agency to refine those standards, ‘fill in the blanks,’ or apply the standards to particular cases.” Industrial Union Dept., AFL–CIO v. American Petroleum Institute, 448 U.S. 607, 675 (1980) (opinion concurring in judgment).

To be clear, the intelligible principle test is not toothless. But it does operate in a way that respects the President’s Article II authority to execute the laws—that is, to exercise discretion and policymaking authority within the limits set by Congress and without undue judicial interference. See, e.g., Whitman, 531 U. S., at 472–476; Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 635 (1952) (Jackson, J., concurring) (“When the President acts pursuant to an express or implied authorization of Congress, his authority is at its maximum, for it includes all that he possesses in his own right plus all that Congress can delegate”). Notably, the intelligible principle test was accepted and applied over the years by Justice Scalia, Chief Justice Rehnquist, and Chief Justice Taft—three jurists who, based on their Executive Branch experience and judicial philosophies, deeply appreciated the risks of undue judicial interference with the operations of the Presidency. See Whitman, 531 U. S., at 472–476 (Scalia, J., joined by Rehnquist, C. J., among others); Mistretta, 488 U. S., at 371–379 (majority opinion joined by Rehnquist, C. J.); id., at 415–416 (Scalia, J., dissenting); Industrial Union, 448 U. S., at 673–676, 685–686 (Rehnquist, J., concurring in judgment); J. W. Hampton, 276 U. S., at 409 (Taft, C. J.); cf. W. Kelley, Justice Scalia, the Nondelegation Doctrine, and Constitutional Argument, 92 Notre Dame L. Rev. 2107 (2017). The intelligible principle test has had staying power—perhaps because of the difficulty of agreeing on a workable and constitutionally principled alternative, or because it has been thought that a stricter test could diminish the President’s longstanding Article II authority to implement legislation.[4]

In any event, there of course can be difficult questions about how to apply the intelligible principle test to particular statutes. See Industrial Union, 448 U. S., at 646 (plurality opinion of Stevens, J.); id., at 685–686 (Rehnquist, J., concurring in judgment). But I agree with how the Court has applied the test in this case.

B

I see no need in this case to try to spell out a definitive guide for applying the intelligible principle test, and it would probably not be possible to do so anyway. It is important, however, to emphasize three points.

First, as both the Court and Justice Gorsuch agree, under the intelligible principle test, “the degree of agency discretion that is acceptable varies according to the scope of the power congressionally conferred.” Whitman, 531 U. S., at 475; see ante, at 11; post, at 12 (dissenting opinion). Congressional delegations of authority to the President “must be judged ‘according to common sense and the inherent necessities of the governmental co-ordination.’ ” Industrial Union, 448 U. S., at 675 (Rehnquist, J., concurring in judgment) (quoting J. W. Hampton, 276 U. S., at 406).

Second, many of the broader structural concerns about expansive delegations have been substantially mitigated by this Court’s recent case law in related areas—in particular (i) the Court’s rejection of so-called Chevron deference and (ii) the Court’s application of the major questions canon of statutory interpretation. Cf. Paul v. United States, 589 U. S. ___ (2019) (statement of Kavanaugh, J., respecting denial of certiorari).

To elaborate: Although the nondelegation doctrine’s intelligible principle test has historically not packed much punch in constricting Congress’s authority to delegate, the President generally must act within the confines set by Congress when he implements legislation. So the President’s actions when implementing legislation are constrained—namely, by the scope of Congress’s authorization and by any restrictions set forth in that statutory text. See Loper Bright Enterprises v. Raimondo, 603 U.S. 369, 394–396, 404 (2024).

On top of that, when interpreting a statute and determining the limits of the statutory text, courts presume that Congress, in the domestic sphere, has not delegated authority to the President to issue major rules—that is, rules of great political and economic significance—unless Congress clearly says as much. See West Virginia v. EPA, 597 U.S. 697, 721–724 (2022). Courts “presume that Congress intends to make major policy decisions itself, not leave those decisions to agencies.” Id., at 723 (quotation marks omitted). That major questions canon reflects both background separation of powers understandings and the commonsense interpretive maxim that Congress does not usually “hide elephants in mouseholes” when granting authority to the President. Whitman, 531 U. S., at 468; see, e.g., Biden v. Nebraska, 600 U.S. 477, 501–506 (2023); FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120, 159–160 (2000); Industrial Union, 448 U. S., at 645 (plurality opinion); ICC v. Cincinnati, N. O. & T. P. R. Co., 167 U.S. 479, 494–495, 509 (1897).

Third, in the national security and foreign policy realms, the nondelegation doctrine (whatever its scope with respect to domestic legislation) appropriately has played an even more limited role in light of the President’s constitutional responsibilities and independent Article II authority. See Loving, 517 U. S., at 772–773; Youngstown, 343 U. S., at 636, n. 2 (Jackson, J., concurring); Zemel v. Rusk, 381 U.S. 1, 17–18 (1965); United States v. Curtiss-Wright Export Corp., 299 U.S. 304, 319–322 (1936); Marshall Field & Co. v. Clark, 143 U.S. 649, 691 (1892). In “the area of foreign affairs, Congress ‘must often accord to the President a degree of discretion and freedom from statutory restriction which would not be admissible were domestic affairs alone involved.’ ” Industrial Union, 448 U. S., at 684 (Rehnquist, J., concurring in judgment) (quoting Curtiss-Wright, 299 U. S., at 320).

In addition, the major questions canon has not been applied by this Court in the national security or foreign policy contexts, because the canon does not reflect ordinary congressional intent in those areas. On the contrary, the usual understanding is that Congress intends to give the President substantial authority and flexibility to protect America and the American people—and that Congress specifies limits on the President when it wants to restrict Presidential power in those national security and foreign policy domains. See Youngstown, 343 U. S., at 635–638 (Jackson, J., concurring); see also Hamdi v. Rumsfeld, 542 U.S. 507, 519 (2004) (plurality opinion); Dames & Moore v. Regan, 453 U.S. 654, 678–679 (1981); Zemel, 381 U. S., at 8–9; Al–Bihani v. Obama, 619 F.3d 1, 38–41, 48–52 (CADC 2010) (Kavanaugh, J., concurring in denial of rehearing en banc); C. Bradley & J. Goldsmith, Foreign Affairs, Nondelegation, and the Major Questions Doctrine, 172 U. Pa. L. Rev. 1743, 1789–1801 (2024). The canon does not translate to those contexts because of the nature of Presidential decisionmaking in response to ever-changing national security threats and diplomatic challenges. Moreover, in those areas, the President possesses at least some independent constitutional power to act even without congressional authorization—that is, in Youngstown category 2.[5]

II

Congressional delegations to independent agencies, as distinct from delegations to the President and executive agencies, raise substantial Article II issues.

Critiques of broad congressional delegations sometimes focus on officials described as “unaccountable bureaucrats.” But that label does not squarely fit delegations to executive agencies. In those circumstances, the President and his subordinate executive officials maintain control over the executive actions undertaken pursuant to a delegation. And the President is elected by and accountable to all the American people. See Myers v. United States, 272 U.S. 52, 123 (1926).

Rather, the problems with delegations to “unaccountable” officials primarily arise from delegations to independent agencies. Independent agencies are headed by officers who are not removable at will by the President and who thus operate largely independent of Presidential supervision and direction. Those independent agency heads are not elected by the people and are not accountable to the people for their policy decisions. Unlike executive agencies supervised and directed by the President, independent agencies sit uncomfortably at the outer periphery of the Executive Branch. Although this Court has thus far allowed such agencies in certain circumstances, they belong to what has been aptly labeled a “headless Fourth Branch.” Freytag v. Commissioner, 501 U.S. 868, 921 (1991) (Scalia, J., concurring in part and concurring in judgment) (quotation marks omitted); see Humphrey’s Executor v. United States, 295 U.S. 602, 628–629 (1935); see also In re Aiken Cty., 645 F.3d 428, 439–446 (CADC 2011) (Kavanaugh, J., concurring).

This case involves a congressional delegation of authority to the FCC. The FCC has commonly been viewed as an independent agency headed by five Commissioners. But at oral argument in this case, the Government correctly pointed out that the FCC formally is not an independent agency because “the FCC does not have statutory for-cause removal protections”—in other words, no statutory text restricts the President’s authority to remove FCC Commissioners at will. Tr. of Oral Arg. 54. And as the Government indicated, this Court’s usual practice, given the text and structure of Article II, is not to infer for-cause removal protections from statutory silence. See Kennedy v. Braidwood Management, Inc., 606 U. S. ___, ___ –___ (2025) (slip op., at 19–20); Shurtleff v. United States, 189 U.S. 311, 314–315 (1903). For those reasons, I tend to agree with the Government that the FCC, in light of the statutory text, should not be considered an independent agency.

If the FCC were an independent agency, however, then a serious Article II delegation problem would arise, in my view. When Congress delegates authority to the President or an executive agency, the exercise of that delegated authority is controlled by the President who was elected by and is accountable to the people. See Myers, 272 U. S., at 123, 163–164. But when Congress delegates authority to an independent agency, no democratically elected official is accountable. Whom do the people blame and hold responsible for a bad decision or policy adopted by an independent agency? Such a system of disembodied independent agencies with enormous power over the American people and American economy operates in substantial tension with the principle of democratic accountability incorporated into the Constitution’s text and structure, as well as historical practice and foundational Article II precedents. “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 Bd., 561 U.S. 477, 499 (2010); see Myers, 272 U. S., at 163–164; see also Morrison v. Olson, 487 U.S. 654, 724–727 (1988) (Scalia, J., dissenting).

There are at least two possible solutions to the problem caused by congressional delegations of authority to independent agencies. One is to overrule (or significantly narrow) Humphrey’s Executor so that the heads of all or most independent agencies are removable at will by the President, and thus supervised and directed by the President. A second option would be to apply a more stringent version of the nondelegation doctrine to delegations to independent agencies. For example, to take one possibility, independent agencies might need to first submit proposed rules to Congress for approval in the legislative process before the rules can take effect.

I will not prolong the point here. Congressional delegations of policymaking authority to independent agencies raise significant Article II issues. In an appropriate case, this Court should address that problem.

*  *  *

As the Court explains, Congress has delegated authority to the FCC with respect to the Universal Service Fund in accordance with the longstanding intelligible principle test. If the FCC were an independent agency, however, the question would be more difficult. Because that issue is not presented in this case, I join the Court’s opinion in full.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 24–354 and 24–422

_________________

FEDERAL COMMUNICATIONS COMMISSION, et al., PETITIONERS

24–354v.

CONSUMERS’ RESEARCH, et al.

SCHOOLS, HEALTH & LIBRARIES BROADBAND COALITION, et al., PETITIONERS

24–422v.

CONSUMERS’ RESEARCH, et al.

on writs of certiorari to the united states court of appeals for the fifth circuit

[June 27, 2025]

Justice Jackson, concurring.

Respondents in this case have challenged the Federal Communications Commission’s universal-service program under both the traditional nondelegation doctrine and the private nondelegation doctrine. The Court properly rejects both challenges today, and I join the Court’s opinion in full. I write separately to express my skepticism that the private nondelegation doctrine—which purports to bar the Government from delegating authority to private actors—is a viable and independent doctrine in the first place. Nothing in the text of the Constitution appears to support a per se rule barring private delegations. And recent scholarship highlights a similar lack of support for the doctrine in our history and precedents. See, e.g., A. Volokh, The Myth of the Federal Private Nondelegation Doctrine, 99 Notre Dame L. Rev. 203 (2023).

In today’s case, none of the parties addressed these concerns, and the Court had no reason to consider them sua sponte because respondents’ private nondelegation claim failed on its own terms. But we should tread carefully before entertaining challenges under this theory in the future. “When the Constitution’s text does not provide a limit to a coordinate branch’s power, we should not lightly assume that Article III implicitly directs the Judiciary to find one.” Consumer Financial Protection Bureau v. Community Financial Services Assn. of America, Ltd., 601 U.S. 416, 446 (2024) (Jackson, J., concurring).

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 24–354 and 24–422

_________________

FEDERAL COMMUNICATIONS COMMISSION, et al., PETITIONERS

24–354v.

CONSUMERS’ RESEARCH, et al.

SCHOOLS, HEALTH & LIBRARIES BROADBAND COALITION, et al., PETITIONERS

24–422v.

CONSUMERS’ RESEARCH, et al.

on writs of certiorari to the united states court of appeals for the fifth circuit

[June 27, 2025]

Justice Gorsuch, with whom Justice Thomas and Justice Alito join, dissenting.

Within the federal government, Congress “alone has access to the pockets of the people.” The Federalist No. 48, p. 334 (J. Cooke ed. 1961) (J. Madison). The Constitution affords only our elected representatives the power to decide which taxes the government can collect and at what rates. See Art. I, §8, cl. 1. Throughout the Nation’s history, Congress has almost invariably respected this assignment. As this Court observed some decades ago, it would represent “a sharp break with our traditions” for Congress to abdicate its responsibilities and “besto[w] on a federal agency the taxing power.” National Cable Television Assn., Inc. v. United States, 415 U.S. 336, 341 (1974).

Today, the Court departs from these time-honored rules. When it comes to “universal service” taxes, the Court concludes, an executive agency may decide for itself what rates to apply and how much to collect. In upholding that arrangement, the Court defies the Constitution’s command that Congress “may not transfer to another branch ‘powers which are strictly and exclusively legislative.’ ” Gundy v. United States, 588 U.S. 128, 135 (2019) (plurality opinion) (quoting Wayman v. Southard, 10 Wheat. 1, 42–43 (1825)).

Still, things could be worse. Because today’s misadventure “sits unmoored from surrounding law,” I have reason to hope its approach will not stand the test of time. Loper Bright Enterprises v. Raimondo, 603 U.S. 369, 425 (2024) (Gorsuch, J., concurring) (internal quotation marks omitted). And even as the Court swallows a delegation beyond anything yet seen in the U. S. Reports, it also signals, unmistakably, that there are some abdications of congressional authority, including in the very statute before us, that the present majority isn’t prepared to stomach.

I

If you look closely at your phone bill, you will likely notice a charge for “universal service.” Perhaps you have wondered what that is and why you are paying for it. As it turns out, in 47 U. S. C. §254, Congress has authorized the Federal Communications Commission (FCC) to subsidize a number of disparate programs under the umbrella of “universal service.” The FCC selects which programs to pursue and how much they should cost. To fund them, the agency taxes telecommunications companies at a rate it controls. By regulation, those companies are then free to pass the charges along to consumers like you. This case involves a challenge to that scheme. To appreciate the questions it poses for us, some background helps.

A

The phrase “universal service” has carried different meanings at different times. Originally, it referred to “a telephone network that covers all of a country.” M. Mueller, Universal Service: Competition, Interconnection, and Monopoly in the Making of the American Telephone System 1 (1997). And it meant one network in particular: the Bell System owned by the American Telephone and Telegraph Company (AT&T). AT&T’s president coined the slogan in 1907—“One System, One Policy, Universal Service”—to boost Bell’s nascent monopoly. Id., at 4, 96. “Universal,” as AT&T used it, focused less on telephone service for all than on making sure AT&T provided all the service. And the slogan proved apt: By the 1920s, the Bell System, fighting “under the banner of universal service,” had conquered the U. S. telephone market. Id., at 146.

Over time, “the term ‘universal service’ took on a new meaning.” P. Huber, M. Kellogg, & J. Thorne, Federal Telecommunications Law §6.1.1.2 (3d ed. Supp. 2022) (Huber). For much of the 20th century, it referred to a policy aimed at making landline local phone service “available to all consumers at a reasonable cost.” Ibid. Even so, AT&T’s monopoly remained at the heart of it all. As with other monopolistic public utilities, federal and state governments regulated the rates the Bell System could charge. See Verizon Communications Inc. v. FCC, 535 U.S. 467, 477 (2002). And, for decades, that was the key to universal service: Regulators manipulated rates to expand Americans’ access to telephones. See Huber §6.1.1.2. So, for example, “[l]ong-distance rates were used to subsidize local rates, business rates to subsidize residential rates, and urban rates to subsidize rural rates.” Ibid.

That system of implicit subsidies worked as long as the same family of companies served all telephone customers. See Verizon Communications, 535 U. S., at 480–481. But the scheme began to falter in the 1970s and 1980s, as new long-distance carriers entered the picture, and an antitrust consent decree spun off AT&T’s long-distance business into a separate company, with newly independent “Baby Bells” now providing local service. See NYNEX Corp. v. Discon, Inc., 525 U.S. 128, 130–131 (1998); Bell Atlantic Corp. v. Twombly, 550 U.S. 544, 549 (2007). At that point, regulators could no longer depend on the Bell System to subsidize local rates by inflating long-distance rates. See Huber §6.2.1.2.

Still, parts of the old universal-service regime hung on. Because the Baby Bells continued to enjoy regional monopolies over local phone service, regulators could still rely on them to provide some implicit subsidies, charging higher rates to some customers while offering below-cost service to others. See id., §6.2.1. The FCC pitched in, too, by requiring long-distance carriers to subsidize local providers, and by establishing a “Lifeline” program to help low-income households afford local phone service. See Rural Telephone Coalition v. FCC, 838 F.2d 1307, 1311–1312 (CADC 1988); Huber §6.2.2.3; ante, at 4.

Eventually, however, Congress decided that universal service had to be “ripped apart and rebuilt afresh.” Huber §2.10. In the Telecommunications Act of 1996, 110Stat. 56, Congress “fundamentally restructure[d]” the local telephone market. AT&T Corp. v. Iowa Utilities Bd., 525 U.S. 366, 371 (1999). No more, Congress declared, should the Baby Bells enjoy regional monopolies over local phone service; now, they must face competition, too. See ibid. To achieve that objective, Congress required the Baby Bells to share their networks with new entrants seeking to offer landline local phone services. See Twombly, 550 U. S., at 549.[1] But Congress also recognized that its new approach would deal “a fatal blow” to “the preexisting system of universal service,” for there would no longer be monopolies whose rates regulators could adjust to subsidize some customers at the expense of others. R. Krotoszynski, Reconsidering the Nondelegation Doctrine: Universal Service, the Power To Tax, and the Ratification Doctrine, 80 Ind. L. J. 239, 282 (2005).

So Congress had to reimagine “universal service” again. In §254 of the Telecommunications Act, Congress used the term “universal service” for the first time and invested it with a new meaning. See Huber §6.3. Gone was Bell’s old idea that universal service meant a single network of wires covering the country. Gone, too, was the idea that the Bell monopoly should subsidize basic telephone service by inflating other customers’ rates. Repurposing the slogan of universal service once more, Congress told the FCC to decide for itself what the concept meant and to fund programs consistent with its understanding. See §254(c)(1). And to pay for those programs, Congress authorized the agency to tax a broad base of interstate “telecommunications carrier[s]” and “provider[s].” §254(d).

B

1

To understand how the scheme works, start with the programs the FCC may fund. Section 254 describes “universal service” as “an evolving level of telecommunications services” that the agency must both “preserve” and “advance.” §§254(b)(5), (c)(1). To determine which specific services to fund and at what level, §254(c)(1) directs the FCC to “consider” four factors. Those factors look to “the extent to which” a service (A) is “essential to education, public health, or public safety,” (B) has “been subscribed to by a substantial majority of residential customers,” (C) is “being deployed . . . by telecommunications carriers,” and (D) is “consistent with the public interest, convenience, and necessity.” §§254(c)(1)(A)–(D).

On top of those four factors, the statute supplies six further “principles” in §254(b). So, for instance, the agency must “base” its funding decisions on the principles that “[q]uality services should be available at just, reasonable, and affordable rates,” §254(b)(1), and that “[a]ccess to advanced telecommunications and information services should be provided in all regions of the Nation,” §254(b)(2). In addition, the FCC may adopt other new “principles” that it “determine[s]” to be “necessary and appropriate.” §254(b)(7). To date, the FCC has exercised that authority twice. One new principle requires “competitive neutrality” among providers and technologies,[2] and the other encourages “support for advanced services” including “broadband networks.”[3]

From this mash of four factors and six (now eight) principles, the FCC must discern which programs it wishes to fund and to what degree. And it falls to the FCC to “ ‘balance’ ” these “factors” and “ ‘principles’ ” “ ‘against one another when they conflict.’ ” Reply Brief for Federal Petitioners 11–12 (quoting Qwest Corp. v. FCC, 258 F.3d 1191, 1200 (CA10 2001)). So, for instance, if the FCC finds that a particular service is “essential to education,” §254(c)(1)(A), but not “subscribed to by a substantial majority of residential customers,” §254(c)(1)(B), the agency must pick which part of the statute prevails. As the FCC has long put it: “[A]ll four criteria enumerated in section 254(c)(1) must be considered, but not each necessarily met.” In re Federal-State Joint Bd. on Universal Serv., 12 FCC Rcd. 8776, 8809 (1997).

Still, that is not quite the end of it. At least when it comes to schools, libraries, and healthcare providers, two additional provisions—§254(c)(3) and §254(h)(2)—permit the agency to pay for “advanced” and “additional” services that go “above the baseline of what’s been considered universal service.” Tr. of Oral Arg. 42, 47. Consistent with these provisions, the FCC has funded programs without regard to whether they satisfy the four factors outlined in §254(c)(1). See 12 FCC Rcd., at 9008–9011.

Over time, the services the agency has funded have evolved considerably. So, for example, in 1996 the FCC debated whether to subsidize “touch-tone service,” not just old rotary phones. 61 Fed. Reg. 10503 (1996). (The answer: Yes. 12 FCC Rcd., at 8809.) By 2011, the FCC “comprehensively reform[ed] and modernize[d]” its universal-service goals to include expanding access to internet services nationwide. In re Connect America Fund, 26 FCC Rcd. 17663, 17667 (2011). More recently, the agency has announced that the Universal Service Fund will help put Wi-Fi on school buses. In re Modernizing the E-Rate Program for Schools and Libraries, FCC No. 23–84 (2023) (declaratory ruling).

2

Once the FCC decides which programs to support, it must figure out how to pay for them. On that score, §254(d) offers this instruction: “Every telecommunications carrier that provides interstate telecommunications services shall contribute, on an equitable and nondiscriminatory basis, to the specific, predictable, and sufficient mechanisms established by the Commission to preserve and advance universal service.” §254(d); see §254(b)(4). In addition to those “mandatory” contributions from common carriers, the statute also grants the FCC “permissive” authority to compel contributions from “[a]ny other provider of interstate telecommunications,” including noncommon carriers, “if the public interest so requires.” §254(d); 12 FCC Rcd., at 9178. Essentially, the agency must figure out whom to tax and how much.

Taking up the question whom to tax, the agency has said that every telecommunications carrier must “contribute” a share of its revenue from interstate and international telecommunications services (think long-distance calls). 47 CFR §54.706 (2024). But over time, the FCC has also expanded the roster of companies who must contribute, so that it now includes providers of prepaid calling cards and internet-based calling. See 71 Fed. Reg. 38781, 43667 (2006). Currently, the FCC does not tax carriers’ broadband revenues (think internet). But some have suggested that, too, should change. See In re Report on the Future of the Universal Serv. Fund, 37 FCC Rcd. 10041, 10088–10094 (2022).

After deciding whom to tax, the agency must determine how much to collect from each carrier. For that, the FCC relies on the Universal Service Administrative Company, a Delaware not-for-profit corporation. Congress has not expressly authorized the FCC to outsource its responsibilities under §254. But in 1997, the FCC directed an association of carriers to create the Administrative Company, and the agency has assumed the task of defining that company’s structure and role. See Brief for Federal Petitioners 4; 47 CFR §§54.703, 54.705; ante, at 7–8. Among other things, FCC regulations ensure that a supermajority of the Administrative Company’s board consists of directors who represent industry insiders (like carriers) and groups that benefit financially from universal-service programs (like libraries and schools). §54.703(b)(1).

How does the Administrative Company help calculate the tax each carrier must pay? Each quarter, the company estimates the upcoming expenses of the FCC’s universal-service programs. §54.709(a)(3). Once the FCC approves that figure, the company next estimates carriers’ total revenues from interstate telecommunications, based on their self-reported figures. This is known as the “contribution base.” Ibid. Finally, the FCC calculates the ratio of projected expenses to the contribution base, which yields the “contribution factor,” or the percentage of its revenue each carrier must pay. §54.709(a)(2); ante, at 7.

As the scope of the FCC’s programs has expanded, so have the taxes the agency collects to fund them. In 1998, universal-service disbursements totaled about $2.29 billion. Universal Service Administrative Co., 1999 Annual Report 2. In 2024, that figure swelled to about $8.59 billion—nearly double, adjusted for inflation. Universal Service Administrative Co., 2024 Annual Report 4. To pay for that increase, the “contribution factor” (or tax rate) has risen, too. In 1998, carriers paid less than 4% of their revenue from interstate and international telecommunications. 63 Fed. Reg. 35931 (1998). Today, that figure is nearly 37%. FCC, Public Notice, DA 25–223 (Mar. 13, 2025).[4]

One might wonder why the Administrative Company, dominated as it is by industry insiders, has allowed universal-service contributions to grow so dramatically. FCC regulations supply at least a partial explanation: “Federal universal service contribution costs may be recovered . . . through a line item on a customer’s bill.” 47 CFR §54.712(a). So, in the end, it is consumers who pay for the agency’s universal-service programs.

II

A

In 2022, a carrier, a non-profit group, and several consumers (collectively, respondents) challenged the present universal-service scheme. Under the Constitution, they observed, Congress must set the federal government’s tax policies. And, they argued, §254 offends that rule because it allows the FCC and the Administrative Company to decide how much tax to collect and at what rate. The Fifth Circuit largely agreed with respondents’ submissions. See 109 F. 4th 743 (2024) (en banc). The FCC, an association of carriers, and others (collectively, petitioners) then sought our review.

As the dispute comes to us, it presents three questions. First, did Congress violate the Constitution by delegating to the FCC the power to tax? Second, did the FCC violate the Constitution by subdelegating some of its authority to the private Administrative Company? And third, even if neither of those features independently offends the Constitution, does their combination? As I see it, this case begins and ends with the first question. Section 254 impermissibly delegates Congress’s taxing power to the FCC, and knowing that is enough to know the Fifth Circuit’s judgment should be affirmed.[5]

Even when it comes to that first question, there is much we need not address. Elsewhere, I have urged the Court to reconsider its approach to assessing legislative delegations in light of the Constitution’s original meaning and historic practice. See Gundy, 588 U. S., at 149 (dissenting opinion). But respondents tell us we need not do so here. Instead, they argue, §254’s delegation of authority to the FCC cannot survive even the most forgiving standard this Court has devised for analyzing delegations: The modern version of the “intelligible principle” test. See Brief for Respondents 65–66. So I will focus on that test, how it applies here, and why §254 fails it.[6]

The Court and I approach our task from common ground. As the Court acknowledges, the Constitution vests “[a]ll” federal legislative power in Congress. Art. I, §1; see ante, at 10. Necessarily, that assignment means “no other” branch of government may exercise legislative power. Ante, at 10. To enforce that rule, this Court has developed the “intelligible principle” test. Ante, at 11 (internal quotation marks omitted); cf. J. W. Hampton, Jr., & Co. v. United States, 276 U.S. 394 (1928). All agree, too, that the test must do something to stop Congress from giving the President or an executive agency a blank check to legislate. See ibid.

On top of all that, the Court and I agree that the intelligible principle test is not one size fits all. Ante, at 11. Instead, “contex[t]” matters. Ante, at 22. Among other things, that means that the “ ‘degree of agency discretion that is acceptable’ ” depends on “ ‘the scope of the power congressionally conferred.’ ” Ante, at 11 (quoting Whitman v. American Trucking Assns., Inc., 531 U.S. 457, 475 (2001)). So, for instance, Congress might permissibly give an agency wide leeway in designing a tax stamp. See In re Kollock, 165 U.S. 526, 537 (1897). But Congress must give far more detailed instructions if it wants an agency to regulate an entire industry. See A. L. A. Schechter Poultry Corp. v. United States, 295 U.S. 495, 541–542 (1935).

B

From that common ground, however, my path and the Court’s begin to diverge. I would start by examining the nature of the power Congress assigned to the FCC. Under §254, the FCC may compel carriers to “contribute” money to support what everyone agrees is a government program. See Wisconsin Bell, Inc. v. United States ex rel. Heath, 604 U. S. ___, ___ (2025) (slip op., at 2). That is a quintessential tax—a “compulsory contribution to the support of government.” 17 Oxford English Dictionary 677 (2d ed. 1989) (defining “tax”). Several FCC commissioners, including at least two FCC chairmen, have seen it the same way, referring to universal-service “contributions” as “taxes.”[7] So have academics and other informed commentators.[8] Before us, as well, the FCC has said that it is “willing” to have this Court treat universal-service contributions “as a tax.” Tr. of Oral Arg. 53. Really, if this compulsory contribution is not a tax, “[w]hat else would you call it?” Bondi v. VanDerStok, 604 U. S. ___, ___ (2025) (slip op., at 19).

Taxation ranks among the government’s greatest powers. Indeed, it is arguably the federal government’s “most important . . . authorit[y].” The Federalist No. 33, p. 205 (A. Hamilton). As this Court has put it, the “power to tax is the one great power upon which the whole national fabric is based.” Nicol v. Ames, 173 U.S. 509, 515 (1899); see also McCulloch v. Maryland, 4 Wheat. 316, 431 (1819) (“[T]he power to tax involves the power to destroy”). Reflecting as much, the Constitution provides that all legislation “for raising Revenue” must “originate in the House of Representatives,” the only popularly elected chamber at the time of the Constitution’s adoption. Art. I, §7, cl. 1; see Amdt. 17. As the framers saw it, “the Chamber that is more accountable to the people should have the primary role in raising revenue.” United States v. Munoz-Flores, 495 U.S. 385, 395 (1990).

That context matters. To survive the intelligible principle test, a delegation involving such a significant power must supply more significant limits on an agency’s discretion than when Congress confers some lesser authority. That is not to say some “different and stricter” test applies when Congress delegates the power to tax. See Skinner v. Mid-America Pipeline Co., 490 U.S. 212, 222–223 (1989). It is instead to recognize that what qualifies as an intelligible principle depends on “context” and “the nature of the particular constitutional powers” at issue. Lichter v. United States, 334 U.S. 742, 778 (1948).

What exactly does the intelligible principle test require in this context? Surely, history must count for something. And it supplies at least one clear standard. As far as I can tell, and as far as petitioners have informed us, this Court has never approved legislation allowing an executive agency to tax domestically unless Congress itself has prescribed the tax rate. See, e.g., Michigan Central R. Co. v. Powers, 201 U.S. 245, 297 (1906) (suggesting that “a direct legislative determination of the rate” avoids “abdication of the legislative function”); 1 T. Cooley & C. Nichols, Law of Taxation 194 (4th ed. 1924) (Cooley & Nichols) (“The nondelegable powers . . . include . . . the fixing of the rate of taxation”); J. Hines & K. Logue, Delegating Tax, 114 Mich. L. Rev. 235, 239 (2015) (Hines & Logue) (“[D]elegating some control over income tax rates . . . would be unprecedented in U. S. history”); cf. Tr. of Oral Arg. 39–40, 57, 78.[9]

Applying that insight here poses petitioners with a serious problem. “[A]ll agree” that Congress has not set the rate at which the FCC may exact contributions. Ante, at 12. Instead, §254 delegates to the FCC the power to determine which services to fund and thus the amount of money to collect. See Part I–B–1, supra. To secure those funds, the agency uses the Administrative Company to set a “contribution factor”—i.e., the percentage of its revenues a carrier must pay. See Part I–B–2, supra. That’s a tax rate. And, remember, that rate has grown from less than 4% of carriers’ applicable revenues in 1998 to nearly 37% today—all based on the agency’s say-so and without any change to the statute. Ibid. Nothing in the U. S. Reports suggests that an executive agency may exercise that kind of power over taxation.

To be sure, petitioners identify an exception to the historic rule that only Congress may set tax rates. Sometimes, they point out, Congress has declined to supply a rate and instead opted to cap the total sum the Executive may collect. See Brief for Petitioner SHLB Coalition et al. 38–39; Brief for Respondents 33–36; Reply Brief for Federal Petitioners 6–7; see Veazie Bank v. Fenno, 8 Wall. 533, 542 (1869). But none of this solves petitioners’ problem. For one thing, petitioners identify no example of a lump-sum delegation outside of direct taxes on property—a unique context where the Constitution’s apportionment requirement makes it practically impossible for Congress to set taxes by rate.[10] For another, and even setting aside that difficulty, the lump-sum exception still would not save §254. As the Court acknowledges, the statute before us “contains no determinate cap.” Ante, at 12. Instead, the FCC gets to decide for itself how much to collect—and, over time, has exercised that authority to double that amount. See Part I–B, supra.

III

Having failed to identify a single example where this Court has approved a tax delegation like this one, petitioners and the Court propose a workaround. Yes, they concede, §254 “imposes no quantitative . . . limits on how much money the FCC can raise.” Ante, at 19. But, they contend, Congress has provided guidance that amounts to a “qualitative” cap. Ibid.; see, e.g., Brief for Federal Petitioners 29–30. To support that claim, petitioners and the Court emphasize that §254(d) requires the FCC to collect taxes “sufficient . . . to preserve and advance universal service.” The word “sufficient,” they submit, “imposes an obligation” on the FCC “to ensure that the [Universal Service] Fund is large enough, but not too large, to achieve the statutory goals” Congress supplied. Brief for Petitioner SHLB Coalition et al. 23. And, they continue, Congress has laid out those goals in a “detailed” way that “cabin[s] the FCC’s exercise of delegated authority.” Id., at i.

Even taken on its own terms, I find that response unpersuasive. For argument’s sake, assume that, instead of fixing the rate at which the FCC can tax, Congress may permissibly impose a numerical limit on receipts. Assume, too, that “qualitative” instructions may sometimes provide guidance functionally equivalent to a numerical limit. Even then, §254’s “qualitative” instructions hardly measure up. Really, the numbers speak for themselves. Recall that in 1998, the FCC disbursed about $2.29 billion. Part I–B–2, supra. A quarter century later, that figure hit about $8.59 billion. Ibid. Even adjusting those figures for inflation, the FCC has nearly doubled the amount of tax it collects. Ibid. Far from supplying “qualitative” directions akin to a numerical cap, §254 supplies little more than a blank check.

Truth be told, the Court does not find its own response entirely persuasive, either. It upholds the ability of the FCC to tax and spend for universal-service programs under §254(c)(1). See ante, at 21–26. But, recall, §§254(c)(3) and (h)(2) also allow the FCC to support “advanced” and “additional” services without regard to §254(c)(1)’s factors. See Part I–B–1, supra. And in a remarkable footnote, the Court scruples to say those two provisions sufficiently constrain the agency’s taxing power. See ante, at 26, n. 9. All the Court is willing to defend, in other words, is whatever level of taxation suffices to support universal service as defined under §254(c)(1), and no more. Consider each of these moves in turn.

A

Start with the parts of §254 the Court does defend—the programs the FCC supports pursuant to §254(c)(1). Together with §254(d), that provision merely tells the FCC to tax carriers in an amount “sufficient” to “preserve and advance universal service.” To my eyes, it’s hard to see how that direction might be fairly analogized to a numerical cap. The statute, remember, does not say what “universal service” is, and the phrase bears no established meaning. To be sure, I have a good sense of what “universal service” meant when the Bell System used it to protect its monopoly. See Part I–A, supra. I have an idea, too, of what the phrase came to mean later, as a shorthand for the practice of regulating rates so most Americans could afford basic phone service. See ibid. But, as we have seen, the 1996 Act “ripped apart” the old notion of universal service. Huber §2.10. Going forward, Congress declared, the meaning of universal service would “evolv[e].” §254(c)(1). In what directions? Congress left that for the FCC to work out.

Of course, the statute proceeds to offer some direction to the agency about what qualifies as “universal service,” and thus how much it can tax and spend. But, even viewed charitably, that guidance can hardly be described as the functional equivalent of a numerical cap. Just recall what the statute actually says. It instructs the FCC to discern the “evolving” meaning of “universal service” from the primordial soup of four factors found in §254(c)(1), as supplemented by six principles discussed in §§254(b)(1)–(6), as well as whatever further principles (two and counting) the agency chooses to devise under §254(b)(7). See Part I–B–1, supra. Many of these factors and principles address competing goods, and any effort to weigh them all can yield no more certainty than asking “ ‘whether a particular line is longer than a particular rock is heavy.’ ” National Pork Producers Council v. Ross, 598 U.S. 356, 381 (2023) (plurality opinion). Recognizing as much, the FCC acknowledges that it falls to the agency to decide how best to proceed (and therefore how much to tax) when the statute’s abundance of factors and principles “conflict.” Brief for Federal Petitioners 31 (internal quotation marks omitted). In no way can this “preface of generalities as to permissible aims” be fairly compared to a firm cap. Schechter Poultry, 295 U. S., at 537.

Experience proves the point. In 1996, recall, the hot debate was whether to subsidize “touch-tone service.” 61 Fed. Reg. 10503. But in 2011, the FCC steered the Universal Service Fund away from simply making basic telephone service available and toward expanding access to “broadband, both fixed and mobile.” 26 FCC Rcd., at 17670. By 2019, the FCC’s Connect America Fund was spending nearly $5 billion annually on high-speed internet services for “rural or remote areas that are costly to serve.” GAO, A. Von Ah, Telecommunications, FCC Should Enhance Performance Goals and Measures for Its Program To Support Broadband Service in High-Cost Areas 1 (GAO–21–24, Oct. 2020). And why stop there? Nothing, it seems, would prevent the FCC from choosing to tax and spend to provide a mobile satellite internet device (like Starlink) to everyone who owns a business, home, or hunting cabin in rural America. Cf. In re Application for Review of Starlink Servs., LLC, 38 FCC Rcd. 12201, 12205–12206 (2023) (revoking a previous $885 million award to Starlink).

Searching for some way (any way) to support the notion that §254(c)(1) contains a “qualitative” cap on how much the FCC can tax and spend, the Court eventually resorts to rewriting the statute. Now, it says, the FCC can fund a service only if it meets all of subsection (c)(1)’s four factors. Ante, at 23, 27. So, the Court stresses, subsection (c)(1)(B) asks whether a particular service “has ‘been subscribed to by a substantial majority of residential customers.’ ” Ibid. As a result, the Court reasons, the FCC cannot fund any service unless most Americans already have it. Ibid. Likewise, the Court asserts, §254(b)’s instruction that the FCC “shall base” its funding decisions on the principles discussed there means that the FCC “must” satisfy all of those principles before funding any program. See ante, at 23, 26.

It’s a nice theory. But it bears no resemblance to the law Congress adopted. By its terms, §254(c)(1) requires the FCC only to “consider the extent to which” each factor applies. The statute nowhere says each (or any) of those criteria “has to be met” before the agency can fund a particular service. Ante, at 27. In fact, the FCC has long and consistently understood the statute to mean the opposite: “[A]ll four criteria enumerated in section 254(c)(1) must be considered, but not each necessarily met.” 12 FCC Rcd., at 8809 (emphasis added).[11]

The same goes for subsection (b). It says that the FCC “shall base” funding decisions on various “principles.” But each of those principles is framed as a “should,” not a “must.” So, for example, subsection (b)(1) provides that the FCC “should” make “[q]uality services . . . . available at . . . affordable rates.” And subsection (b)(2) says that the FCC “should” ensure “[a]ccess to advanced . . . services” is “provided in all regions of the Nation.” In this context, “[t]he term ‘should’ indicates a recommended course of action, but does not itself imply the obligation associated with ‘shall.’ ” Qwest Corp., 258 F. 3d, at 1200. Reflecting that understanding, the FCC has long read §254(b) as requiring it to consult all of that provision’s various principles, but not as “impos[ing] inflexible requirements” on universal-service programs. Connect America Fund Phase II Auction, 33 FCC Rcd. 1428, 1468, n. 229 (2018). Instead, each provision supplies “ ‘only a principle, not a statutory command,’ ” which the agency may “ ‘ignore’ ” in service of other principles found in the statute. Rural Digital Opportunity Fund Phase I Auction, 35 FCC Rcd. 6077, 6119, n. 262 (2020) (quoting Alenco Communications, Inc. v. FCC, 201 F.3d 608, 621 (CA5 2000)); see also In re Lifeline & Link Up Reform & Modernization, 27 FCC Rcd. 6656, 6759, n. 636 (2012); Rural Cellular Assn. v. FCC, 588 F.3d 1095, 1103 (CADC 2009).[12]

In its zeal to save subsection (c)(1) programs (why else ignore what the statute actually says?), the Court throws all that aside and seizes the drafting pen. So the meaning of “universal service” evolves once more. Only now, it is the Court’s creation through and through. And if that were not bad enough, the Court’s late-night rewrite hardly helps its cause. Even as revised, the statute still falls well short of imposing anything like a numerical cap on how much the FCC can tax and spend. Suppose tomorrow the agency decides to ensure “every American [has] a cell phone and a cell phone plan.” Tr. of Oral Arg. 62–63. Could anyone complain that “a substantial majority of residential customers” do not use cell phones? §254(c)(1)(B). I doubt it.

Not only does the Court’s new statute fail to deliver on its only assignment, it promises to backfire, too. Rather than preserve the status quo, as the Court so clearly desires, its revisions threaten to render existing programs illegal—all while leaving the FCC (and program beneficiaries) guessing about the implications for future initiatives. Take an example. Back in 2017, the FCC launched a multibillion-dollar effort to promote “broadband service in unserved high-cost areas.” In re Connect America Fund, 32 FCC Rcd. 1624 (2017). Among other things, the program subsidizes certain high-speed services that, the FCC has acknowledged, are not yet embraced by “a substantial majority of residential customers.” Id., at 1631 (internal quotation marks omitted). If we had simply confessed the obvious—that this statute is unconstitutional—Congress could have responded easily with the simple addition of a rate or, perhaps, a cap. But now? Now, the FCC can fund a program only if it satisfies all subsection (b) principles and all subsection (c) factors—a novel requirement that calls existing programs into question and promises profound implications for future ones as well. Far from avoiding any short-term disruption, the Court’s new statute promises plenty of chaos of its own.[13]

B

So much for the Court’s renovation of §254(c)(1). Now consider the two provisions the Court cannot bring itself to defend. Sections 254(c)(3) and (h)(2), recall, permit the FCC to fund “additional” and “advanced” services for “schools, libraries, and health care providers.” See Part I–B–1, supra. Even the Court is unwilling to say that these provisions impose a “qualitative” cap—and understandably so. When it comes to deciding what programs to fund under §254(c)(3) and §254(h)(2), the FCC is unconstrained by any of the subsection (c)(1) factors the Court rewrites and leans on so heavily today. See In re Federal-State Joint Bd. on Universal Serv., 12 FCC Rcd., at 9008–9011; Tr. of Oral Arg. 42, 47. Subsection (c)(3) makes that explicit: The FCC may designate services for support under that provision “[i]n addition to the services included in the definition of universal service under paragraph (1).”

Here, too, experience illustrates just how uncapped the FCC’s §254(c)(3) and §254(h)(2) programs really are. For some years, the FCC has relied on those provisions to fund internet access at schools and libraries. 89 Fed. Reg. 67304–67305 (2024). But, in 2024, the FCC announced that it would also begin funding “Wi-Fi hotspots and services to be used off-premises by students, school staff, and library patrons.” Id., at 67304, 67318. As far as the FCC sees it, subsections (c)(3) and (h)(2) might allow it to collect enough taxes to supply take-home hotspots to anyone with a library card. See Tr. of Oral Arg. 43–44. Or maybe even take-home Starlink devices for library patrons nationwide, to help shrink the “digital divide between [Americans] with access to broadband at home and those without.” In re Addressing the Homework Gap Through the E-Rate Program, FCC No. 24–76, p. 12 (2024).

Rather than address the constitutionality of the FCC’s power to tax and spend for §254(c)(3) and §254(h)(2) programs, the Court dodges the question. Buried in a footnote midway through its opinion, it offers this: “We have no occasion to address any nondelegation issues raised by Sections 254(c)(3) and (h)(2),” the Court says, because while respondents challenged “the contribution scheme generally,” they did not name those provisions “in particular.” Ante, at 26, n. 9.

It is a perplexing maneuver, and I suspect the parties will find it quite the surprise. Not one of them suggested cleaving off portions of the statute in this way. Still, the Court’s late-breaking move is, in one sense, to its credit. Though it is unwilling to say aloud that any part of §254 fails the intelligible principle test, neither can the Court bring itself to bless such a lavish delegation of taxing authority. As a result, respondents remain free on remand, or in a future proceeding, to renew their attack on the constitutionality of whatever contributions the FCC demands for its subsection (c)(3) and (h)(2) programs. And that in itself is a notable development: Today marks the first time in a long time that the Court has confronted a statutory delegation and found no way to save it.

IV

Return, now, to the portion of §254 the Court is willing to defend—the programs the FCC supports pursuant to subsection (c)(1). No one disputes that part of the statute lacks a tax rate or a numerical cap. And rewritten or not, that provision lacks anything approaching a “qualitative” cap as well. Faced with these difficulties, in the end petitioners and the Court resort to changing the conversation. Now, they say, look to other fields. The Court has allowed agencies “to raise revenue” through fees “without specifying a numeric cap or . . . rate.” Ante, at 14. The Court has allowed agencies to set “ ‘just and reasonable’ rates” and to regulate in the “ ‘public interest.’ ” Ante, at 22. And, petitioners and the Court reason, if those delegations are permissible, this one must be, too. Ante, at 22–23; Brief for Federal Petitioners 11. Failing all else, petitioners and the Court add that they see no practical value in requiring Congress to speak more clearly about the scope of permissible universal-service taxes. Ante, at 18–19; Tr. of Oral Arg. 8–9.

But the Court’s comparisons disregard its own insight that context matters in applying the intelligible principle test. Ante, at 11, 22. Nor, in my view, is it any answer to say that legislation supplying a rate or real cap might still leave the FCC with some measure of discretion. Though the Constitution does not require Congress to make every decision, there are some choices that belong to Congress alone—including setting a tax’s rate or, at least, capping receipts.

A

Start with the Court’s assertion that “Congress has often enacted statutes empowering agencies to raise revenue without specifying a numeric cap or tax rate.” Ante, at 14. To sustain that point, the Court relies on a list of nine examples offered by the government. Ibid. (citing Reply Brief for Federal Petitioners 7–9). The private petitioners highlight the same provisions, which, it seems, provide “the best precedents” for §254. Tr. of Oral Arg. 82–84; see id., at 8, 31, 37–38.

Those provisions all have something in common: Each describes a fee. And that makes them poor benchmarks for a tax delegation. Fees, by definition, are payments made in “compensation for a service provided to, or alternatively compensation for a cost imposed by, the person charged the fee.” Mueller v. Raemisch, 740 F.3d 1128, 1133 (CA7 2014); accord, Pace v. Burgess, 92 U.S. 372, 375–376 (1876). For that reason, fees carry a built-in intelligible principle: The government cannot collect more money than it needs to offset a real-world cost or benefit. See National Cable Television Assn., 415 U. S., at 341–342.

Consider one of the government’s examples, the Federal Deposit Insurance Corporation (FDIC). See ante, at 16. The FDIC finances its Deposit Insurance Fund through “[i]nsurance fees” paid by FDIC-insured banks. 12 U. S. C. §1815(d). By statute, the fee each bank pays must be “keyed to the risk of the bank’s insolvency.” M. Ricks, Money as Infrastructure, 2018 Colum. Bus. L. Rev. 757, 803 (citing §1817(b)). In other words, the fees offset a cost the banks impose on the FDIC—namely, their risk of failure, which the FDIC’s insurance must underwrite. That need to compensate for a particular cost provides an intelligible principle cabining the FDIC’s discretion and limiting the fees it can charge to what the agency needs to cover insurance payouts. See id., at 803–804.

A similar principle explains the fee-setting authority of federal courts. See Reply Brief for Federal Petitioners 8. By statute, courts of appeals may charge “fees” in amounts “prescribed from time to time by the Judicial Conference of the United States.” 28 U. S. C. §1913. As the Judicial Conference has recognized, that provision authorizes courts to “charg[e] for services provided,” such as docketing an appeal or admitting an attorney to practice. Report of the Proceedings of the Judicial Conference of the United States 12, 14 (Sept. 16, 2008). So here, too, fee levels reflect a cost imposed or a benefit received by the payor.

Take one more example from the government’s brief. See Reply Brief for Federal Petitioners 8–9. Congress has authorized the Animal and Plant Health Inspection Service to charge fees “sufficient” “to cover the cost of providing agricultural . . . inspection” for certain “commercial aircraft.” 21 U. S. C. §136a(a)(1)(A). That provision might sound broad because it lacks a numerical limit. In context, though, it leaves little for the agency to do except arithmetic: To set the fee, it “divid[es] the total costs of inspecting Commercial Aircraft by the total number of Commercial Aircraft.” Air Transp. Assn. of Am., Inc. v. United States Dept. of Agriculture, 37 F. 4th 667, 675 (CADC 2022).[14]

Petitioners and the Court would set all that to the side. In their telling, this Court’s decision in Skinner “rendered irrelevant” the question whether a particular exaction involves a tax or a fee. Ante, at 15 (citing 490 U. S., at 223). Not so. Skinner, to be sure, indicated that the intelligible principle test applies to tax delegations. Id., at 223. And under that test, Skinner presented an easy case. The statute at issue there told the government (numerically) how much money it could raise. Id., at 219–220 (“the ceiling on aggregate fees . . . is set at 105 percent of the aggregate appropriations made by Congress for that fiscal year”). In light of that and other instructions, the Court held, the statute provided an intelligible principle regardless of whether it imposed a tax or a fee. Id., at 220–223. That is all the decision held. It certainly did not invite courts to ignore relevant context or disregard the basic distinction between fees (which incorporate an intelligible principle by nature) and taxes (which do not).[15]

Perhaps sensing as much, the Court ultimately retreats from precedent into pragmatism. Trying to distinguish taxes from fees, it contends, would risk plunging the Court into “a morass.” Ante, at 16. But the job is hardly some feat fit for Hercules alone. Courts must, and do, routinely distinguish between taxes and fees in many contexts. See, e.g., United States v. United States Shoe Corp., 523 U.S. 360, 366–370 (1998) (distinguishing between fees and taxes for purposes of the Export Clause, U. S. Const., Art. I, §9, cl. 5).

Besides, any morass here is of the Court’s own making. In its view, the tax/fee distinction depends on whether the charge “is for a ‘benefit’ granted to the payor that is ‘not shared by other members of society.’ ” Ante, at 16 (quoting National Cable Television Assn., 415 U. S., at 341). But that formulation is not the usual one. In defining fees, the typical question is whether the fee compensates for a benefit or cost that sets the payor apart. See, e.g., United States v. Sperry Corp., 493 U.S. 52, 60–61 (1989); Massachusetts v. United States, 435 U.S. 444, 462–463, and n. 19 (1978); Pace, 92 U. S., at 375–376; Mueller, 740 F. 3d, at 1133; Hines & Logue 257, n. 107; GAO, S. Irving, Federal User Fees: A Design Guide 4, n. 4 (GAO–08–386SP, May 2008); 1 Cooley & Nichols 97–98, 109–110.[16] Think back to the FDIC. The public undoubtedly benefits from the fees banks pay for deposit insurance; we all enjoy a safer banking system. But banks impose a cost on the FDIC (their risk of failure) that distinguishes them from the general public. Insurance fees compensate for that special cost. That is why they are fees, not taxes.

Once we understand fees correctly, the category plainly includes all the government’s examples and excludes §254 contributions. When carriers pay into the Universal Service Fund, they do not gain any special benefit, such as permission “to practice law or medicine or construct a house or run a broadcast station.” National Cable Television Assn., 415 U. S., at 340. (A different provision, 47 U. S. C. §158, authorizes “application fees.”) Nor do §254 contributions offset some regulatory cost that carriers impose on the FCC or on society. (That falls to “regulatory fees” the FCC collects under §159.) Instead, §254 takes money from some (carriers) and gives it to others (libraries, schools, and the like). See In re Incomnet, Inc., 463 F.3d 1064, 1066–1067 (CA9 2006). In short: Section 254 creates a classic tax-and-spend scheme, not a fee. Even the FCC does not dispute the point. Tr. of Oral Arg. 53; see Part II–A, supra.[17]

B

Beyond fees, petitioners and the Court offer a second reason to ignore the fact that §254(c)(1) programs can be funded through taxes without a rate or fixed cap. After all, they observe, this Court has rejected nondelegation attacks on “authorizations to regulate in the ‘public interest’ and to set ‘just and reasonable’ rates.” Ante, at 22 (quoting National Broadcasting Co. v. United States, 319 U.S. 190, 225–226 (1943), and FPC v. Hope Natural Gas Co., 320 U.S. 591, 600 (1944)). And, set next to those vague standards, petitioners and the Court suggest, §254 offers “clear and limiting” guidance. Ante, at 30.

This argument neglects the Court’s own admonition that the intelligible principle test is context dependent. See ante, at 11, 22. It begins by asking “what instructions” the statute in question provides to constrain an agency’s discretion. Gundy, 588 U. S., at 136 (plurality opinion). Answering that question is a matter of statutory interpretation. Ibid. And “[i]t is a fundamental canon of statutory construction that the words of a statute must be read in their context,” West Virginia v. EPA, 597 U.S. 697, 721 (2022) (internal quotation marks omitted), because language used in one setting may carry a meaning it does not have in others, see Yates v. United States, 574 U.S. 528, 537 (2015) (plurality opinion). So even if a particular statutory term evokes “well-known and generally acceptable standards” in one domain, that does not mean the same term will necessarily supply similar guidance when used in other “uncharted fields.” Fahey v. Mallonee, 332 U.S. 245, 250 (1947); see also Schechter Poultry, 295 U. S., at 534 (new context may give terms “a much broader range and a new significance”).

This Court has recognized this point many times, including when it comes to phrases like “public interest.” So, for example, the Court has held that phrase may contain enough “concrete” meaning to survive the intelligible principle test in the “context” of broadcast licensing, where the government has to allocate a limited spectrum of publicly owned airwaves. National Broadcasting Co., 319 U. S., at 216, 226 (internal quotation marks omitted); see M. McConnell, The President Who Would Not Be King 334 (2020) (McConnell). But, the Court has also found, the same “public interest” criterion offers inadequate guidance to the FCC when it comes to raising revenue. National Cable Television Assn., 415 U. S., at 341.

The same goes for the phrase “ ‘just and reasonable.’ ” Ante, at 22. This Court has sometimes found that phrase satisfies the intelligible principle test when it comes to setting rates for regulated monopolies like public utilities—a context where it incorporates “concepts with a long history at common law.” H. Friendly, The Federal Administrative Agencies: The Need for Better Definition of Standards, 75 Harv. L. Rev. 863, 873, n. 53 (1962); see, e.g., ICC v. Cincinnati, N. O. & T. P. R. Co., 167 U.S. 479, 501 (1897).[18] Accordingly, in that sphere, the “traditional regulatory notion of the ‘just and reasonable’ rate” may mean something. Verizon, 535 U. S., at 481. But outside that sphere, the same phrase may amount to little more than an instruction to go forth and do good. See Schechter Poultry, 295 U. S., at 539–540; G. Lawson, Delegation and Original Meaning, 88 Va. L. Rev. 327, 386 (2002).

Here’s the point: Just because a phrase carries a well-understood historic meaning in one context does not mean the same phrase “ ‘in the abstract’ ” will suffice in every other setting to satisfy the intelligible principle test. Reply Brief for Federal Petitioners 3. So the fact that petitioners and the Court can point to past decisions approving the use of a broad phrase in a different domain proves nothing. Instead, it falls to petitioners and the Court to show that the statutory terms presently before us, properly understood in their particular context, do in fact provide significant constraints on the FCC as it exercises a significant power.

That is a burden petitioners and the Court have not carried and cannot carry. When §254(c)(1) speaks of “universal service,” it does not invoke “ ‘concepts with a long history at common law.’ ” Supra, at 31. To the contrary, everyone agrees that §254 “ripped apart” the old understanding of “universal service” and cleared the ground for a new one. Huber §2.10; see Part I–A, supra; Brief for Federal Petitioners 3; Brief for Petitioner SHLB Coalition et al. 3. Worse, even if one term or another in §254(c)(1) or §§254(b)(1)–(7) might claim some meaning-giving ancestry, the statute’s mix of four factors and six (or more) principles lacks any such pedigree. Whether those factors and principles need only be considered, as the statute says, or whether they must always be met, as the Court now suggests, the FCC truly must blaze its own trail.

Notice, too, where trails like this lead. If context could be cast aside, and a phrase like “just and reasonable” might suffice in every season, nothing would stop Congress from granting agencies limitless legislative power. Congress might delegate to the Secretary of Education the authority to set a “just and reasonable” tax on university endowments in order to fund universal education—defined, of course, according to four factors and six (or more) incommensurable principles. Congress might instruct the Secretary of Health and Human Services to impose a “just and reasonable” tax on pharmaceutical sales in order to subsidize “universal health coverage,” defined as an “evolving level” of care that should be available “at just, reasonable, and affordable rates” to patients “in all regions of the Nation.” Or Congress might let the Treasury Department set whatever “just and reasonable” income tax rates were needed to trim the national debt to a level “consistent with the public interest, convenience, and necessity.” If these possibilities strike you as unhinged, it is because you appreciate that the permissible scope of delegation must depend on context and the nature and scope of the power conferred.[19]

C

Running out of precedents to work with, the Court suggests, finally, that it would be “absurd” to ask Congress to provide more guidance than it has. Ante, at 18. The argument runs this way. Respondents suggest, and I agree, that Congress could readily cure §254’s nondelegation problem. All it needs to do is set a tax rate or (perhaps) specify a numerical cap or its equivalent. The Court does not dispute that a legislatively defined tax rate would amount to meaningful guidance. But, the Court contends, some hypothetical caps could be pointless. What if Congress capped receipts at $5 trillion? Ibid. To suggest that such a law would provide an “intelligible principle,” while insisting that §§254(b) and (c)(1) do not, strikes the Court as mindless formalism. Ibid.

Up to a point, I agree. It may well be that a rate is usually required to give a domestic tax law an intelligible principle outside the context of direct taxes. See Part II–A, supra. Imagine, for instance, that Congress told the IRS to collect $50 trillion of income tax from the American people and left it at that. Few, I suspect, would suggest that instruction supplies sufficient guidance. But even assuming that a cap alone might be permissible when it comes to §254, the Court too readily dismisses its constitutional value.

Forcing Congress to supply some cap, any cap, would advance the nondelegation doctrine’s purpose of ensuring “that the lines of accountability [remain] clear.” Gundy, 588 U. S., at 155 (Gorsuch, J., dissenting). If Congress adopted the Court’s ludicrously hypothetical $5 trillion universal-service tax, the American people would at least know whom to thank when the corresponding charges showed up on their phone bills. Even if the FCC chose to collect only a fraction of that amount, every Member of Congress would have to explain to his constituents where he stood on a potential $5 trillion tax. Far more realistically, of course, Congress would never contemplate such a silly law, precisely to avoid those awkward conversations (and the electoral consequences that could follow).

And that’s exactly the point. The framers divided power among legislative, executive, and judicial branches not out of a desire for formal tidiness, but to ensure ours would indeed be a Nation ruled by “We the People.” See id., at 152. By vesting executive power in a single President, the framers hoped to ensure vigorous enforcement of the laws. See United States v. Arthrex, Inc., 594 U.S. 1, 28 (2021) (Gorsuch, J., concurring in part and dissenting in part). And by vesting judicial power in life-tenured judges, they hoped to ensure laws would be applied fairly by those insulated from political pressure. See SEC v. Jarkesy, 603 U.S. 109, 147–148 (2024) (Gorsuch, J., concurring). But, as the framers saw it, the power to make the laws that govern our society belongs to elected representatives more accountable to the people in whose name they act. See Gundy, 588 U. S., at 154–157 (Gorsuch, J., dissenting). And nowhere did the framers see that principle as applying with greater force than in the field of taxation. See Part II–A, supra.

In so many other arenas, this Court vigorously polices the Constitution’s allocation of power. We have refused to tolerate congressional intrusions on powers reserved to the President. See Gundy, 588 U. S., at 168 (Gorsuch, J., dissenting). We have prohibited the Executive from encroaching on power vested in Congress. Ibid. We have found unconstitutional, too, legislation seeking to confer judicial power on the other branches. Ibid.

Yet there is one exception. When Congress has willingly surrendered its power to the Executive Branch, this Court’s responses can only be described as feeble. Always, to be sure, the Court dutifully recites the creed that “[l]egislative power . . . belongs to the legislative branch, and to no other.” Ante, at 10. Too often, though, these professions amount (at most) to faith without works, and the results are not hard to see. Today, the “vast majority” of the rules that govern our society are not made by Congress, but by Presidents or agencies they struggle to superintend. J. Adler & C. Walker, Delegation and Time, 105 Iowa L. Rev. 1931, 1975 (2020). Those rules reflect not the public deliberations of elected representatives, but the concerns of small cadres of elites. And as those cadres turn over from administration to administration, the rules revolve, too, inflicting whiplash on those who must live under them. If there is any consistency over time, it may be because Presidents and their deputies do not always call the shots: Lower level officials, unknown to the public and sometimes even to the White House, now make many of the rules we live by. See N. Rao, The Hedgehog & the Fox in Administrative Law, 150 Daedalus 220, 228–229 (Summer 2021).

To its credit, the Court has sometimes mitigated its failure to police legislative delegations by deploying other tools, like the major questions doctrine and de novo review of statutory terms, to ensure the Executive “act[s] within the confines set by Congress.” Ante, at 8 (Kavanaugh, J., concurring); see Gundy, 588 U. S., at 166–168 (Gorsuch, J., dissenting). But every doctrine has its limits. What happens when Congress, weary of the hard business of legislating and facing strong incentives to pass the buck, cedes its lawmaking power, clearly and unmistakably, to an executive that craves it? See id., at 156. No canon of construction can bar the way. Then, our anemic approach to legislative delegations leaves the Court with a choice. It can permit the delegation to stand and move us all one step further from being citizens in a self-governing republic and one step closer to being subjects of quadrennial kings and long-tenured bureaucrats. Or the Court can, as it does today, usurp legislative power, rewrite the statute, and dictate its own terms for Congress’s surrender. Either way, we wind up in much the same place, only now with judges, rather than Presidents or bureaucrats, making our laws.

There is another way. The Constitution promises that our elected representatives in Congress, and they alone, will make the laws that bind us. To honor that commitment, historical practice and our cases suggest other guides, beyond the intelligible principle test, for assessing when Congress has impermissibly ceded legislative power, as I have pointed out before. See id., at 157–162. As I have observed, too, when Chief Justice Taft first used the phrase “intelligible principle,” he did not aim to displace those traditional guides, only to summarize them. Id., at 162 (citing J. W. Hampton, 276 U. S., at 409); see also n. 15, supra. Someday, soon, we should find our way back. By employing the modern, enfeebled form of the intelligible principle test, we do the Constitution no favors. And by approving a delegation of Congress’s taxing power unprecedented in this Court’s history, we risk making matters worse yet.[20]

Still, there is room for some optimism. The Court today cannot bring itself to say that §§254(c)(3) and (h)(2) survive even its milquetoast version of the intelligible principle test. The Court also refuses to sustain §254(c)(1) as enacted, feeling obliged instead to rewrite that provision before upholding it. I can imagine worse outcomes than those small steps toward home. But we can and should do better. When it comes to other aspects of the separation of powers, we have found manageable ways to honor the Constitution’s design. This one requires no less of us.

Respectfully, I dissent.