by Neil Gorsuch
So what would happen in a world without Chevron? If this goliath of modern administrative law were to fall? Surely Congress could and would continue to pass statutes for executive agencies to enforce. And just as surely agencies could and would continue to offer guidance on how they intend to enforce those statutes. The only difference would be that courts would then fulfill their duty to exercise their independent judgment about what the law is. Of course, courts could and would consult agency views and apply the agency’s interpretation when it accords with the best reading of a statute. But judicial review of the law’s meaning would limit the ability of an agency to alter and amend existing law. It would avoid the due process and equal protection problems of the kind documented in our decisions. It would promote reliance interests by allowing citizens to organize their affairs with some assurance that the rug will not be pulled from under them tomorrow, the next day, or after the next election. And an agency’s recourse for a judicial declaration of the law’s meaning that it dislikes would be precisely the recourse the Constitution prescribes—an appeal to higher judicial authority or a new law enacted consistent with bicameralism and presentment. We managed to live with the administrative state before Chevron. We could do it again. Put simply, it seems to me that in a world without Chevron very little would change—except perhaps the most important things.
Caring Hearts v. Burwell
Where the last case addressed the mixing of executive and judicial powers, the next excerpt discusses what can happen when the executive assumes legislative powers. Freed from the constitutional constraints associated with the legislative process (bicameralism and presentment) and endowed with vigor and energy, the executive branch agency at issue here had little to stop it from churning out so many new regulations—thousands upon thousands of new documents every year—that eventually even it couldn’t keep track of them all. A local company providing home health care to seniors found itself accused of violating rules that, as it turns out, weren’t even in existence when it rendered its services.
Executive agencies today are permitted not only to enforce legislation but to revise and reshape it through the exercise of so-called “delegated” legislative authority. The number of formal rules these agencies have issued thanks to their delegated legislative authority has grown so exuberantly it’s hard to keep up. The Code of Federal Regulations now clocks in at over 175,000 pages. And no one seems sure how many more hundreds of thousands (or maybe millions) of pages of less formal or “sub-regulatory” policy manuals, directives, and the like might be found floating around these days. For some, all this delegated legislative activity by the executive branch raises interesting questions about the separation of powers. For others, it raises troubling questions about due process and fair notice—questions like whether and how people can be fairly expected to keep pace with and conform their conduct to all this churning and changing “law.” But what if the problem is even worse than that? What happens if we reach the point where even these legislating agencies don’t know what their own “law” is?
That’s the problem we confront in this case. And perhaps it comes as little surprise that it arises in the Medicare context. Medicare is, to say the least, a complicated program. The Centers for Medicare & Medicaid Services (CMS) estimates that it issues literally thousands of new or revised guidance documents (not pages) every single year, guidance providers must follow exactingly if they wish to provide health care services to the elderly and disabled under Medicare’s umbrella. Currently, about 37,000 separate guidance documents can be found on CMS’s website—and even that doesn’t purport to be a complete inventory.
But how did CMS wind up confused about its own law? It began this way. Caring Hearts provides physical therapy and skilled nursing services to “homebound” Medicare patients. Of course, any Medicare provider may only charge the government for services that are “reasonable and necessary.” But Congress hasn’t exactly been clear about who qualifies as homebound or what services qualify as reasonable and necessary. So CMS has developed its own rules on both subjects—rules the agency has (repeatedly) revised and expanded over time. In a recent audit, CMS purported to find that Caring Hearts provided services to at least a handful of patients who didn’t qualify as “homebound” or for whom the services rendered weren’t “reasonable and necessary.” As a result, CMS ordered Caring Hearts to repay the government over $800,000.
The trouble is, in reaching its conclusions CMS applied the wrong law. The agency didn’t apply the regulations in force in 2008 when Caring Hearts provided the services in dispute. Instead, it applied considerably more onerous regulations the agency adopted only years later. Regulations that Caring Hearts couldn’t have known about at the time it provided its services. Regulations that even CMS concedes bore only prospective effect. And Caring Hearts can make out a pretty good case that its services were entirely consistent with the law as it was at the time they were rendered. So this isn’t (and never was) a case about willful Medicare fraud. Instead, it’s a case about an agency struggling to keep up with the furious pace of its own rulemaking.
This case lays bare a strange world where the government itself—the very “expert” agency responsible for promulgating the “law” no less—seems unable to keep pace with its own frenetic lawmaking. A world Madison worried about long ago, a world in which the laws are “so voluminous that they cannot be read” and constitutional norms of due process, fair notice, and even the separation of powers seem very much at stake. But whatever else one might say about our visit to this place, one thing seems to us certain: an agency decision that loses track of its own controlling regulations and applies the wrong rules in order to penalize private citizens can never stand.
United States v. Nichols
As a group, criminal defendants have never been much liked—and the persons involved here may be the least popular of all. But what would you think of a government that allows its chief prosecutor to write the crimes he gets to prosecute? That’s exactly what happened in this case, where Congress divested its legislative power and handed it to the attorney general, allowing him to write criminal laws for half a million people. In this Tenth Circuit dissent, I explore some of the consequences of this mixing of legislative and executive powers. After I joined the Supreme Court, these issues were to come before me again in Gundy v. United States (2019).
If the separation of powers means anything, it must mean that the prosecutor isn’t allowed to define the crimes he gets to enforce. Yet, that’s precisely the arrangement the Sex Offender Registration and Notification Act purports to allow in this case and a great many more like it. In the Act, Congress left it to the Attorney General to decide whether and on what terms sex offenders convicted before the date of SORNA’s enactment should be required to register their location or face another criminal conviction. So unusual is this delegation of legislative authority that to find an analogue you might have to look back to the time Congress asked the President to devise a code of “fair competition” for the poultry business—a delegation of legislative authority the Supreme Court unanimously rejected and Justice Cardozo called “unconfined and vagrant,” a “delegation running riot.” Even then you could be excused for thinking the delegation before us a good deal less cooped or caged than that one. After all, it doesn’t just grant some alphabet soup agency the power to write rules about the chicken trade. It invests in the nation’s chief prosecutor the authority to devise a criminal code governing a half-million people.
* * *
—
WHEN IT COMES TO sex offenders convicted after SORNA’s enactment, the statute is exquisitely detailed. It divides those persons into three tiers based on the seriousness of their offense. It specifies which sex offenses place offenders in which tiers. It requires tier I offenders to register their location for 15 years; tier II offenders to do so for 25 years; and tier III offenders to carry on registering for life. It explains what cond
itions merit reducing the registration period. On and on it goes for 22 pages.
But none of this automatically applies to Mr. Lester Nichols and others convicted of sex offenses before the Act’s passage. Instead, when it comes to past offenders, the Act says just this:
The Attorney General shall have the authority to specify the applicability of the requirements of this subchapter to sex offenders convicted before the enactment of this chapter…and to prescribe rules for registration of any such sex offender.
Yes, that’s it.
As the government acknowledges, this language leaves the Attorney General free to do nothing: the law “does not require the Attorney General to act within a certain time frame or by a date certain; it does not require him to act at all.” Alternatively, “[u]nder his delegated authority in Subsection (d), the Attorney General could” require all past offenders to register or “require some but not all to register.” Or, alternatively still, he could require those forced to register to “comply with some but not all of the registration requirements” applicable to future offenders in order to adapt the law as he thinks best for past offenders. After all, the statute grants the Attorney General authority to specify the applicability not of the Act as a whole, one way or another, but to specify the applicability of each of the various “requirements” contained within the Act. Even then, the Attorney General remains free to “change his mind at any given time or over the course of different administrations.” Given all this, it’s perhaps unsurprising how many circuits and commentators have observed that the degree of discretion invested in the Attorney General here is vast. It is so vast, in fact, that some (including the government itself) once suggested a narrower interpretation of the relevant statutory language would make more sense of the Act.
A majority of the Supreme Court, however, carefully considered and rejected any alternative reading and made plain that, as a matter of statutory interpretation, SORNA’s retroactive application hinges on the Attorney General. It went on to acknowledge that a statute investing so much authority in the Attorney General inevitably raises with it separation of powers questions. But, the Court said, it would leave those questions for another day. Justices Scalia and Ginsburg went further, expressing concern that the law “sail[s] close to the wind.” The day to decide the constitutional question the Court left open is now upon us. And, as it turns out, the statute doesn’t just sail close to the wind. It sails right into it.
* * *
—
ARTICLE I §1 PROVIDES that “[a]ll legislative powers herein granted shall be vested in a Congress of the United States.” Many times over and in cases stretching back to the founding, the Supreme Court has held that this language limits the ability of Congress to delegate its legislative power to the Executive. There’s ample evidence, too, that the framers of the Constitution thought the compartmentalization of legislative power not just a tool of good government or necessary to protect the authority of Congress from encroachment by the Executive but essential to the preservation of the people’s liberty. As Madison put it, “[n]o political truth is…stamped with the authority of more enlightened patrons of liberty” than the separation of powers because “[t]he accumulation of all powers, legislative, executive, and judiciary in the same hands…may justly be pronounced the very definition of tyranny.” By separating the lawmaking and law enforcement functions, the framers sought to thwart the ability of an individual or group to exercise arbitrary or absolute power. And by restricting lawmaking to one branch and forcing any legislation to endure bicameralism and presentment, the framers sought to make the task of lawmaking more arduous still. These structural impediments to lawmaking were no bugs in the system but the point of the design: a deliberate and jealous effort to preserve room for individual liberty.
Without a doubt, the framers’ concerns about the delegation of legislative power had a great deal to do with the criminal law. The framers worried that placing the power to legislate, prosecute, and jail in the hands of the Executive would invite the sort of tyranny they experienced at the hands of a whimsical king. Their endorsement of the separation of powers was predicated on the view that “[t]he inefficiency associated with [it] serves a valuable” liberty-preserving “function, and, in the context of criminal law, no other mechanism provides a substitute.”
Of course all this invites the question: how do you know an impermissible delegation of legislative authority when you see it? By its own telling, the Court has had a hard time devising a satisfying answer. But the difficulty of the inquiry doesn’t mean it isn’t worth the effort. After all, at stake here isn’t just the balance of power between the political branches who might be assumed capable of fighting it out among themselves. At stake is the principle that the scope of individual liberty may be reduced only according to the deliberately difficult processes prescribed by the Constitution, a principle that may not be fully vindicated without the intervention of the courts. And “[a]bdication of responsibility is not part of the constitutional design.”
Besides, putting the pieces together it turns out we do know a few things.
We know, for example, that Congress can leave “details” to the Executive. Congress can’t punt to the President the job of devising a competition code for the chicken industry. Such widely applicable rules governing private conduct must be enacted by the Legislature. But once Congress enacts a detailed statutory scheme on its own—once it says, for example, that margarine manufacturers must pay a tax and place a stamp on their packages showing the tax has been paid—Congress may leave to the President “details” like designing an appropriate tax stamp.
Of course, defining what qualifies as a detail is itself no detail. But whether or not something fairly denominated a detail is involved, we also know Congress may pass legislation the operation of which is conditioned on a factual finding by the President. So, for example, Congress may direct the President to lift a statutorily imposed trade embargo against Great Britain if he determines as a factual matter that it is no longer violating the United States’s neutrality. That’s clearly no trivial question the President may answer. But answer it he may so long as a clear legislative consequence follows from his factual finding.
While these are the most traditional delegation tests—is it a detail? do we have a clear legislative consequence hinging on a factual finding?—in more recent times the Court has gone further, allowing legislation to stand so long as it contains an “intelligible principle” to guide the exercise of Executive discretion. How intelligible the “intelligible principle” must be to pass muster is much debated. But we know, by way of example, that Congress may ask the EPA to set national air quality standards which are “requisite to protect the public health” subject to “an adequate margin of safety” because, as used in the statute, the term “requisite” demands a standard neither higher nor lower than necessary to meet the legislatively directed objective of protecting the public health with an adequate margin of safety.
Still, the Court has never expressly held that an intelligible principle alone suffices to save a putative delegation when the criminal law is involved. To be sure, the Court has applied the intelligible principle test to regulations that may be enforceable through criminal penalties. But the Court hasn’t endorsed the test in anything like the situation we face—legislation leaving it to the nation’s top prosecutor to specify whether and how a federal criminal law should be applied to a class of a half-million individuals. In fact, the Court has repeatedly and long suggested that in the criminal context Congress must provide more “meaningful[]” guidance than an “intelligible principle.”
Recently, the Supreme Court has suggested what a more “meaningful” standard might look like in the criminal context. Its discussion came in the course of a challenge to the Controlled Substances Act—legislation permitting the Attorney General to schedule various drugs as controlled substances, rendering their possession by una
uthorized persons illegal. The Court allowed the law to stand, but instead of applying the intelligible principle test alone it proceeded to stress the presence and importance of certain specific statutory features.
At least three “meaningful” limitations emerged: (1) Congress must set forth a clear and generally applicable rule that (2) hinges on a factual determination by the Executive and (3) the statute provides criteria the Executive must employ when making its finding.
* * *
—
WITH THAT MUCH GUIDANCE about delegation doctrine in hand, a few things come clear when we return to the statute before us. For one, it’s easy enough to see the similarities between our case and Schechter Poultry where the Court held Article I violated. Here as there Congress pointed to a problem that needed fixing and more or less told the Executive to go forth and figure it out. Meanwhile, it’s hard to see how ours might be likened to any of the cases turning away delegation challenges.
True, some might try to pass off the question of SORNA’s applicability to past offenders as a mere “detail.” But the statute before us leaves the Attorney General with “unfettered discretion to determine both how and whether SORNA [is] to be retroactively applied” to a half-million individuals under threat of criminal prosecution from his own deputies. And however far you want to bend the boundaries of what qualifies as a “detail,” it’s hard to see how that might qualify. Our case just isn’t anything like your grandfather’s tax stamp challenge.
