Suspect Spheres, Not Enumerated Powers: A Guide for Leaving the Lamppost

Richard Primus* and Roderick M. Hills, Jr.** May, 2021

Despite longstanding orthodoxy, the Constitution’s enumeration of congressional powers does virtually nothing to limit federal lawmaking. That’s not because of some bizarrely persistent judicial failure to read the Constitution correctly. It’s because the enumeration of congressional powers is not a well-designed technology for limiting federal legislation. Rather than trying to make the enumeration do work that it will not do, decisionmakers should find better ways of thinking about what lawmaking should be done locally rather than nationally. This Article suggests such a rubric, one that asks not whether Congress has permission to do a certain thing but whether a certain kind of lawmaking is more prone to pathology at the national or the state level. That inquiry could identify “suspect spheres”: areas of policymaking where federal law calls for more justification than elsewhere. Federal legislation within suspect spheres would not necessarily be subject to judicial invalidation, but the judgment that legislation falls within a suspect sphere could underwrite softer forms of judicial resistance to nationalization. We illustrate the suspect-spheres model with a principle of federalism we call the corporate nondelegation doctrine, by which federal delegations of power to private corporations are to be treated skeptically. Early on, that principle animated Madison’s opposition to the Bank of the United States and much of the Jacksonian approach to federalism. It later underwrote the Supreme Court’s decision in Schechter Poultry. In the current century, the idea that the corporate nondelegation doctrine defines a suspect sphere helps explain otherwise puzzling judicial behaviors in federalism cases, including the presumption against preemption and the resistance to the individual mandate of the Affordable Care Act. By illustrating the possibility of a suspect-sphere approach, we suggest a tool that might be useful at a time of destructively polarized national politics, when rubrics for allocating some polarizing issue spaces to state-level decisionmakers might help lower the national temperature.

Introduction

Officially, the Constitution’s enumeration of congressional powers marks the boundary between two domains. Within the domain of its enumerated powers, Congress may legislate, subject to limits like those in the Bill of Rights. Outside that domain, legislative power belongs exclusively to the states.¹ The official account further imagines significant space on each side of the line and, more pointedly, that the zone of exclusive state jurisdiction is bigger than the zone where Congress may act. In James Madison’s oft-quoted words, “[t]he powers delegated by the proposed Constitution to the federal government are few and defined. Those which are to remain in the State governments are numerous and indefinite.”²

This official picture bears little resemblance to the way American federalism actually works. In reality, Congress is not much limited by its enumerated powers, and national lawmaking is normal across a very broad swath of policymaking space. Moreover, the enumeration’s failure to confine Congress to a smaller domain does not reflect some bizarre judicial inability to understand and enforce the Constitution. It is consistent with constitutional text, history, and structure, and it accords with longstanding practice. As a historical matter, it is probably a mistake to think that the guiding purpose behind the drafting of Article I, Section 8 was the limitation of federal legislative power.³ And as a conceptual matter, the Constitution’s enumeration of congressional powers is not well suited for marking limits beyond which the national government cannot go. So it should not be surprising that courts have been unable to use the enumeration as the basis for any practically successful set of rules for limiting national power. “Enumerationism,” to use Professor David Schwartz’s term, is a failed ideology of federalism, one that cannot explain either the content or the limits of Congress’s powers.⁴

One might adjust to this reality by abandoning federalism-based limits on Congress, except for those that are specifically established in the Constitution’s text.⁵ But alternative adjustments are also conceivable. Instead of trying to identify what should be left to the states by making negative inferences from the list of enumerated congressional powers, perhaps constitutional decisionmakers⁶ could think directly about what sorts of national lawmaking should be treated as suspect.

Many readers will be skittish about going down that road. Identifying domains in which Congress should not legislate—even above and beyond those that the Constitution textually proscribes—calls to mind the “traditional government functions” test from National League of Cities.⁷ And National League of Cities was a failure.⁸ In recent decades, analysts have accordingly been skeptical that courts have the institutional capacity to identify and enforce affirmative federalism-based limits beyond those spelled out in the Constitution’s text.⁹

To a significant extent, we share that skepticism. We also recognize, however, that the fall of National League of Cities has coexisted with continued assertions by constitutional decisionmakers that some sorts of congressional lawmaking not proscribed by the Constitution’s text are disfavored on federalism grounds. Sometimes, as in the commandeering cases, the concern is that federal law must not interfere unduly with the operations of state governments.¹⁰ At other times, the concern has been that Congress might inappropriately federalize specific policy domains.¹¹

We do not believe that all these concerns have been well-founded. Federal law is present in pretty much every policy domain—in our view, unproblematically so. Our aim in pointing out that constitutional practice robustly features arguments like these is not to endorse the view that Congress must stay its hand in any particular policy areas. Instead, the present point is about the kinds of arguments constitutional decisionmakers have typically made when advocating (categorical or presumptive) limits on federal legislation. Despite the official account, those arguments have often proceeded by identifying areas where Congress should not legislate, rather than by looking at Congress’s enumerated powers and asking what is left out. For the most part, decisionmakers have long behaved—and properly so—as if Congress has the power to legislate in general, except where its legislation would bump up against some limiting principle.

To express that thought with a set of modern terms, the federalism-based limits under which Congress labors have mostly been external limits (that is, affirmative prohibitions) rather than internal limits (that is, limits inhering in the grant of powers to Congress).¹² Sometimes these external limits are enforced directly, as when a court strikes down federal law because it trenches on some unwritten state prerogative. Anticommandeering decisions like Printz v. United States provide examples.¹³ Sometimes these limits are enforced in softer ways, as with clear-statement rules used to construe federal statutes narrowly. A recent example is Bond v. United States.¹⁴ And the source of federalism-based resistance to Congress in cases like Printz and Bond is not to be found in anything about the Constitution’s enumeration of congressional powers. To map the zones of resistance, one needs to treat federalism-based limits on national lawmaking as affirmative principles whose sources mostly lie outside that enumeration. Indeed, they mostly lie outside the constitutional text entirely.

This Article has two aims. The first is to show the impracticality of generating sensible limits on national lawmaking by reference to the enumerated powers of Congress. The second is to sketch a potential alternative. Rather than defining a zone of exclusive state jurisdiction consisting of whatever residuum Congress’s enumerated powers do not reach, we recommend identifying “suspect spheres” in which national lawmaking should be presumptively disfavored. Those spheres need not be fixed and categorical across time: they might shift as changing conditions alter the domains in which skepticism toward national lawmaking is reasonable. The challenge, therefore, is not to identify a static list of suspect spheres but to articulate the tools that decisionmakers should use to identify legislation as suspect and then to apply the appropriate suspicion.

To be successful, such an approach must meet three criteria. First, rather than being deduced from the text of the enumeration, the content of the suspect spheres must be products of constitutional construction, rooted in political and legal experience.¹⁵ Second, the approach must not ask courts to erect and police static limits on Congress, as National League of Cities tried to do. We imagine that most of the constraining work within this model will be done extrajudicially, by agencies and by Congress itself.¹⁶ And when cases do come to court, we believe in judicial deference to Congress’s judgment that federal law is warranted, even within suspect spheres. But perhaps courts can help prevent Congress (and the national executive) from reaching such judgments lightly. For that purpose, we envision courts using “soft” constraints like antipreemption and clear-statement canons, as well as some limits on Chevron¹⁷ and State Farm¹⁸ deference to executive rulemaking. In the most extreme case, where federal lawmakers regulated in a suspect sphere and were utterly unable to produce a plausible explanation for doing so, the stronger medicine of judicial review might be justified.¹⁹ But such cases should be rare, and even if they occurred, the relevant federal actions would be held invalid only because insufficient reasons were given, not because legislation of a certain kind was off-limits.

Third and finally, a successful suspect-spheres approach would have to deliver some substantively valuable benefit of federalism. Of the various benefits that well-implemented federalism can deliver, one that is especially needed right now is reducing the stakes of national politics.²⁰ Bitter and accelerating conflict across the partisan divide is doing significant damage to the constitutional system.²¹ Limitations on the scope of national lawmaking might reduce the temperature of that conflict—if those limitations were to offer something important to people affiliated with both major parties. Limits that appeal to only one side—say, by forbidding Congress to nationalize health care or by preempting local gun control laws without delivering something of comparable value to the other side of the aisle—would just exacerbate the conflict. Limits appealing only to one side would also be unstable, because the other side would have no incentive to respect that limit when it holds the levers of national power (including the judicial levers). But if federal lawmaking could be limited by principles with some amount of transpartisan appeal, those limits might be able to sustain constitutional coalitions—that is, coalitions of citizens in both parties strong enough to resist each party’s temptation to renege on the compromise when it is in power. In the best-case scenario, building such coalitions might be part of building a healthier constitutional culture more broadly.

Readers might reasonably wonder whether any principles that could be used to limit national lawmaking actually meet all of these criteria. To illustrate the possibility, we offer an example we call the corporate nondelegation doctrine. Ever since the 1790s, Americans have worried about delegations of federal power to private corporations.²² To be sure, people have worried about corporate power at every level of government, local as well as national.²³ But corporate influence over national lawmaking has been particularly worrisome: partly because of fears that the national government is distinctively susceptible to capture by networks of financial insiders, partly because the national government has more capacity to do damage if captured. This resistance to delegating national power to private corporations has often united political coalitions capable of sustaining constitutional principles. We illustrate that history by reference to episodes from the eighteenth, nineteenth, and twentieth centuries, from the struggles over the Bank of the United States to the death of the National Industrial Recovery Act.²⁴

We also suggest that the corporate nondelegation doctrine helps explain otherwise-perplexing features of twenty-first century federalism jurisprudence. Our leading examples are the principles governing statutory preemption²⁵ and the conflict over the individual mandate of the Affordable Care Act.²⁶ In these domains, courts (and others) have been animated by intuitions that federal law is operating in domains where it should not. But as long as constitutional lawyers look to the Constitution’s enumeration of congressional powers as the guide to where federal law is and is not proper, the real bases of those intuitions will remain obscure, and clearsighted decisionmaking will be difficult. It would accordingly be wise to abandon the idea that the Constitution’s enumeration of congressional powers provides a basis—let alone an authoritative basis—for dividing policy space between a domain in which national lawmaking is appropriate and a domain in which it is not.

Freeing ourselves of that myth does not mean abandoning federalism. Federalism is enormously valuable, and it can and should be maintained. But there is little point in preserving the idea of a limiting enumeration if in fact the enumeration of congressional powers does little or no practical work for federalism. And it doesn’t. The actual forces sustaining federalism have long been a combination of external limits like the ones we explore and process limits built into the structure of governmental decisionmaking.²⁷ Because enumeration has not done much work, federalism would continue more or less unchanged if we recognized the idea of a limiting enumeration as the myth that it is. Except, of course, for one respect in which that recognition would improve matters: setting aside the myth would let analysts and decisionmakers focus more forthrightly on the actual reasons for federalism-based limits on national lawmaking.

In Part I of this Article, we explain why the enumeration of congressional powers does not supply useful limits on national lawmaking. In Part II, we describe the alternative of external limits in the form of suspect spheres. In Part III, we use the corporate nondelegation doctrine to illustrate this alternative, tracing its role in constitutional federalism from the First Congress to the New Deal. Finally, in Part IV, we show that the corporate nondelegation doctrine is alive and consequential in the twenty-first century. We suggest that better decisions could be made if we recognized it and things like it for what they are: nontextual external limits on federal lawmaking, untethered to anything in the Constitution’s enumeration of congressional powers.

There’s an old joke in which a police officer on night patrol comes across a drunk man crouching under a lamppost. The officer asks the drunk man what he is doing, and the drunk man explains that he lost his wallet and is searching for it. Intending to be helpful, the officer crouches alongside the drunk man and joins in the search for the wallet. After a minute or two of looking without success, the officer asks the drunk man, “Where were you when you lost your wallet?” The drunk man points to the other side of the street. The officer, perplexed, asks the obvious next question: “So why are you looking here?” “Well,” the drunk man says, “there’s no light over there. Here, there’s a lamppost.”

Searching for sensible limits to national lawmaking by looking at the Constitution’s enumeration of congressional powers is a bit like looking for the drunk man’s wallet under the lamppost. Something about the environment—the light, or the textual list of powers—makes it seem like an auspicious place to look. But the thing being sought is not there to be found. To be clear, there is no guarantee of success even if we look somewhere else. The drunk man is not wrong to think that the other side of the street is dark. But we know for certain what will happen if we look in the wrong place. The only rational alternatives are to look in another place or give up trying. One way or the other, it is time for constitutional federalism to leave the lamppost.

I. How Internal Limits Fail to Identify Reserved State Powers

Consider two ways to identify limits on Congress’s power to legislate. First, one might try to reason from the Constitution’s textual enumeration of congressional powers. Working from the premise that “[t]he enumeration of powers is also a limitation of powers,”²⁸ one could reason that Congress can do nothing except what the Constitution affirmatively specifies. According to this expressio unius–style argument, the Constitution’s affirmative specification of many congressional powers implies that every other type of legislative power belongs to some other legislator.²⁹ The Tenth Amendment reinforces this perspective by providing that whatever powers the federal government does not enjoy are reserved either “to the States . . . or to the people.”³⁰ In keeping with a modern terminological convention, we call limits rooted in the enumeration of powers internal limits, because they are derived from—internal to—the grant of power itself.³¹

By contrast, one might instead specify limits on Congress by focusing on affirmative reasons why Congress should not enact certain kinds of legislation. Congress may not abridge the freedom of speech,³² take private property without just compensation,³³ or impose cruel and unusual punishments.³⁴ Laws transgressing those limits are unconstitutional even if they fit within Congress’s enumerated powers. Think of a regulation of commerce that takes private property or a law punishing counterfeiters by boiling them in oil. The reasons why Congress may not enact such laws are rooted in considerations located outside of—external to—the Constitution’s enumeration of congressional powers. We call these external limits.

The external limits in the examples just given—no taking of private property without just compensation, and no cruel and unusual punishments—are rooted in considerations of individual rights. But external limits can also be rooted in considerations of federalism. At the Founding, the Establishment Clause and the Second Amendment were both federalism-based external limits: irrespective of Congress’s enumerated powers, the Establishment Clause blocked a type of legislation that might endanger the peace of the Union if done centrally rather than locally, and the Second Amendment blocked a type of legislation that the central government might use to arrogate power at the expense of the states. Other federalism-based external limits on Congress include the prohibitions against Congress’s unilaterally redrawing state boundaries,³⁵ dictating the location of a state’s capital,³⁶ or taxing a state’s own tax revenue.³⁷

We argue below that judicial enforcement of federalism through internal limits is doomed to failure. Neither the text nor the spirit of the enumeration in Article I provides workable materials for any sensible theory of federalism. A rational approach to constitutional federalism must recognize this fact.

A. Internal Versus External Limits

The Venn diagrams below use the distinction between internal and external limits to depict three conceptions of the relationship between state and federal legislative power.

Figure 1

Figure 1 depicts a conventional (albeit contested) “internal limit” conception of congressional power. This conception imagines that the enumerated powers of Congress define a small space within the universe of potential legislation. It then imagines an extensive residuum of exclusive state power occupying the remainder of that universe. The striped oval represents areas that are exclusively federal, like the power to declare war. The gray oval represents areas of concurrent federal and state power, where states can legislate until and unless Congress legislates to the contrary. But even when combined, Congress’s zones of exclusive and concurrent jurisdiction form only a moderately sized island in a vast sea. That image reflects Madison’s famous statement—really, his lament—that federal powers “are few and defined,” while those of “the State governments are numerous and indefinite.”³⁸ State law is the presumptively normal form of law; federal law is the exception that must be justified.

Note three consequences for reasoning about legislation that follow from this image. First, the conception of national legislation as exceptional focuses analysis on defining and justifying the island of federal power: the question to be answered is “What, if anything, justifies Congress’s choice to pass a law?”³⁹ Second, because it treats the vast black sea as an undifferentiated whole, this conception has nothing to say about whether local control is more important or more desirable in some areas within the states’ jurisdiction than in others. Third, and by the same token, this conception has nothing to say about whether some potential legislation lying within the gray zone of concurrent power might still be better suited for local rather than national policymaking, such that Congress should think twice about acting even though its action would fall within its enumerated powers.

Figure 1 bears little resemblance to reality. There is no meaningful sense in which federal laws are a small island in a sea of reserved state powers, nor has there been at any point in our lifetimes. Federal law is as normal a part of our regulatory landscape as state law. All aspects of your job—wages and hours,⁴⁰ working conditions,⁴¹ hiring and promotion criteria,⁴² collective bargaining⁴³—are regulated by federal law. So is the air you breathe,⁴⁴ the car you drive,⁴⁵ the food you eat,⁴⁶ the water you drink and clean with,⁴⁷ and the energy that powers your home, office, farm, or factory.⁴⁸ And it is uncontroversial that Congress has authority to preempt a vast swath of the state law that still remains in force, setting aside (for example) state laws regarding torts, contracts, property, and crime.⁴⁹ State law remains immensely important, but not because it operates in a space that Congress cannot reach.

Figure 2

The sphere of exclusive federal power (striped) could still be imagined as a small subset of all possible legislative power. But the zone of concurrent federal and state authority occupies the vast majority of the space. The residuum of exclusive state jurisdiction is less like an encompassing sea and more like a few discrete ponds on the margins of the law. Laws penalizing the mere possession of firearms that have never traveled in interstate commerce, for instance, would fall into one of those corner patches of black.⁵⁰ But laws regulating the possession of firearms that have moved in interstate commerce—a large majority of all guns—do not.⁵¹

To be sure, state law can still regulate firearms that move in interstate commerce. States regulate not only in their exclusive-power zones but also in the large (gray) zone where their power is concurrent with that of Congress, so long as the specific legislation they enact is not preempted. There is a lot of important state law out there. Our earlier list of areas regulated by federal statute—your job, your car, your food, the very air around you—is also a list of things regulated by state law. Most of the social world is subject to both state and national legislation, until and unless Congress makes itself the exclusive authority through preemption.

The systems that Figures 1 and 2 depict are different—obviously and radically so. But they share something important. Although exclusive state jurisdiction is large in Figure 1 and small in Figure 2, both diagrams present the allocation of powers between national and local legislatures exclusively by reference to what Congress is allowed to do. In neither diagram does any part of the analysis ask whether some area of law is particularly important for states to control. Nor—and this is not the same thing—does any part of the analysis ask whether some sorts of lawmaking are likely to go badly if done at the national level. As a result, these ways of thinking make no space for the possibility that Congress may have the enumerated power to enact a piece of legislation that, for federalism reasons, it should not.

Consider: Congress has an enumerated power to lay and collect taxes.⁵² A federal law requiring states to pay 100 percent of the money that they collect as tax revenue to the federal treasury would be, straightforwardly, a law for the laying and collecting of taxes. It would also be a terrible idea for federalism reasons: it would destroy a great deal of the states’ ability to govern. We have little trouble saying that such a federal tax law would be unconstitutional. The reason for that unconstitutionality, however, cannot be discerned by examining the enumerated powers of Congress. Instead, the reason turns on a need of the states, or, put differently, on the destructiveness of a certain kind of national lawmaking, even though that lawmaking falls within the textually specified powers of Congress. Such considerations are invisible in an account that takes the enumeration of congressional powers as the basis for allocating legislative powers among state and national legislatures. On any such account—whether it looks like Figure 1 or Figure 2—the zones in which state jurisdiction cannot be overridden by federal law have no internal coherence. They are random snippets left over when the set of congressional powers is subtracted, like the oddly shaped and practically useless bits of cloth left over after a tailor cuts out pieces for a suit.

Contrast this internal-limit approach to Congress’s powers with the external-limit approach visualized in Figure 3 below:

Figure 3

In Figure 3, everything within the gray square, including the interior of the two circles, is within the enumerated powers of Congress. But within that overall space, a particular zone is carved out as belonging presumptively to the states on the basis of considerations that override Congress’s enumerated powers. Thus, the black circle representing presumptive state jurisdiction is not defined merely as a space in which Congress’s enumerated powers have run out. A federal law might be an exercise of Congress’s enumerated powers and yet fall within the black zone, rendering the law unconstitutional on federalism grounds. So the black space is not a residuum. It is an affirmative reserve.

In this third model, which we think is both more sensible and closer to reality than Figures 1 and 2, the zone of presumptive state jurisdiction is not determined by reference to the enumerated powers of Congress. It is defined by factors independent of—external to—the enumeration, such as the importance of a state’s being able to make certain kinds of laws or the danger of letting certain kinds of laws be made at the national level. Sometimes these two conditions are opposite sides of the same coin. At other times, one is at work rather than the other.

To explore the value of thinking in the way that Figure 3 suggests, consider the Supreme Court’s anticommandeering doctrines. In its first anticommandeering decision, New York v. United States, the Court rested its analysis on a historical claim about the Founders’ having deliberately rejected a system in which Congress operated directly on the states and a structural claim about democratic accountability.⁵³ In the second anticommandeering decision, Printz v. United States, the Court added an argument about traditional practice (i.e., that Congress through American history has not commandeered state officials and that Congress’s self-restraint on this point has signaled its awareness of a constitutional limitation)⁵⁴ and an argument that letting Congress conscript the resources and personnel of all fifty state governments would instantly make the federal government much more powerful than it would otherwise be.⁵⁵ Reasonable people disagree about the strength of these rationales. For present purposes, what matters is that all these considerations are matters of history and structure. None of them is about the text of the enumerated powers of Congress, nor even about the purpose or the spirit of the enumeration. Indeed, New York’s and Printz’s arguments about the Founders or political accountability or traditional practice would be just as strong (or weak) if the Constitution expressly gave Congress a general police power. So on the understanding that considerations like these are the justifications for the anticommandeering doctrines, those doctrines are external limits on the powers of Congress, rooted in considerations lying outside the Constitution’s enumeration of congressional powers. They occupy the black circle in Figure 3, defined not by the edge-of-the-picture triangles in Figure 2 but rather by principles independent of the shape of Congress’s powers.

B. The Limits of Textual Inference

Constitutional interpreters commonly try to use the textual enumeration of congressional powers as a basis for reasoning about what sorts of legislation should be left to the states. Considered more carefully, however, the constitutional text cannot do the work that these interpretive moves require. In the end, the texts of the clauses specifying congressional powers cannot provide adequate guidance for identifying types of legislation that should (conclusively or presumptively) be within the domain of states.

1. “The Enumeration Presupposes Something Not Enumerated”

There are various inspirations for using the enumeration to identify types of legislation that Congress should not pass: the Tenth Amendment, the expressio unius canon, and (a misunderstanding of) Chief Justice Marshall’s dictum that “[t]he enumeration presupposes something not enumerated.”⁵⁶ On this view, the fact that the Constitution affirmatively lists a bunch of specific congressional powers shows that Congress enjoys less than general legislative jurisdiction.⁵⁷

This way of thinking has important limits. For one, it is not true that the Constitution’s enumeration of powers and the rule stated by the Tenth Amendment necessarily preclude Congress from regulating any particular thing. Sometimes a set of enumerated powers turns out, in practice, to cover all the ground that a grant of general jurisdiction would.⁵⁸ A legislature with seven enumerated powers would have the power to do anything that a legislature of general jurisdiction had the power to do if the seven enumerated powers were the power to make law applicable on Sundays, the power to make law applicable on Mondays, and so forth. Whether any particular set of enumerated powers amounts to something less than a power of general jurisdiction is a question that can only be answered at retail, by looking at the specific set of powers granted and asking whether they leave some part of the social world uncovered. Maybe, as applied to the circumstances of the social world in 2020, the enumerated powers of Congress are sufficient warrant for anything that Congress might choose to legislate, subject to the blocking force of external limits. Or maybe not. The Supreme Court has given the latter answer.⁵⁹ But for reasons one of us has explained at length elsewhere, it is a mistake to think that there must be things left over.⁶⁰ What David Schwartz has called the “mustbesomething rule”⁶¹—that is, the idea that the enumeration of powers signals that there must be something congressional power does not reach—is simply a fallacy.

But even if it were true that the enumerated powers of Congress must leave something unreachable by federal legislation, that interpretive principle would not help in identifying which domains of legislation should be off limits. The idea that the enumerated powers must leave something beyond Congress’s reach is the animating principle of Lopez,⁶² and in the twenty-six years since Lopez, the application of this principle has led only to the sort of exclusive state jurisdiction pictured in Figure 2, in which Congress is denied power over disconnected and marginal bits of the social world. Congress cannot regulate the possession of firearms that have not moved in interstate commerce,⁶³ and Congress cannot create a private cause of action for victims of gender-motivated violence as a remedy for purely intrastate crimes.⁶⁴ Those limitations matter in the lives of the people affected by those specific rules. But those limitations play no functional role in a system of helpful federalism, unless “federalism” just means “blocking exercises of national power in a couple of disconnected places.” Put differently, if all that the enumeration of powers can say about what sorts of legislation should be left to states is said in Lopez and Morrison—or even Lopez, Morrison, and three or four other cases like them—then the enumeration of powers has little of value to say on the topic.

2. Inferences from Absences

Interpreters who would like to identify specific domains of reserved state power by imagining the photographic negative of Congress’s enumerated powers might press their case, however, by reasoning from apparently specific omissions from Congress’s arsenal, rather than from the fact of the enumeration overall. The intuition is easy to grasp. If we tell our students that class will meet on Monday, Wednesday, and Friday, they will reasonably understand us to mean that class will not meet on Tuesday and Thursday (to say nothing of Saturday and Sunday). The same logic can be applied to constitutional clauses. For example, Congress has the power to regulate commerce “with foreign Nations, and among the several States, and with the Indian Tribes.”⁶⁵ Chief Justice Marshall famously wrote that the specification of three kinds of commerce implied the existence of a fourth—purely intrastate commerce—to which the power given in this clause does not extend.⁶⁶

But an interpreter who treated this interpretive move as generally valid would go badly astray. Article I, Section 8, Clause 5 gives Congress the power to coin money.⁶⁷ The next clause gives Congress the power “[t]o provide for the Punishment of counterfeiting the Securities and current Coin of the United States.”⁶⁸ If the affirmative specification of particular congressional powers implies the absence of other powers of the same kind that the Constitution does not affirmatively specify, then the affirmative specification of a congressional power to criminalize counterfeiting would imply that Congress lacks the power to criminalize tax evasion. After all, the Constitution’s express grant to Congress of a power to protect the money it coins by punishing counterfeiters stands in clear contrast to the Constitution’s omission of a congressional power to impose punishments in aid of its power to collect taxes. On this interpretation, Congress would have the power to lay and collect taxes but no power to provide for the punishment of those who do not pay their taxes.

This is nonsense: no reasonable person doubts that Congress has the power to punish tax evaders (and mail thieves and patent infringers and so forth). That authority is implicit, either as incident to Congress’s taxing power or as part of the authority specified by the Necessary and Proper Clause. To be sure, a close reader of Section 8 would notice the express specification of a power to punish counterfeiters and the absence of any parallel power to punish tax evaders. But the inference that Congress lacks those other powers of punishment is based on a false premise: that the existence of Clause 6 is a necessary condition for Congress’s having the power to punish counterfeiters. Even if there were no Clause 6, Congress would have the power to punish counterfeiters, just as Congress has the power to punish tax evaders.⁶⁹ Yes, this line of reasoning means that Clause 6 is superfluous, and many constitutional interpreters resist the idea that the Constitution contains surplusage.⁷⁰ But it does, just as most documents drafted by large committees do. Section 8 may be a competently drafted collection of clauses, but it is not a finely wrought work of art.⁷¹ Nor, a fortiori, is the full set of the forty-three constitutional clauses scattered throughout the Constitution conferring power on Congress, only eighteen of which are in Section 8.

In short: some clauses enumerating congressional powers specify powers (in this example, to criminalize counterfeiting) that Congress would have even if those powers were not enumerated. It is thus not the case that the enumeration’s failure to include a given power means that any particular authority is denied. This is true even where the power not enumerated is comparable to, and comparably important with, a power that the Constitution does bother to enumerate. And once it is clear that the enumeration of certain powers does not reliably imply the absence of similar powers not enumerated, it is hard to sustain the idea that the limits of congressional power can be inferred by thinking about what the enumeration of powers seems to leave out.⁷²

3. The Extratextualism of “Necessary and Proper”

On one possible framing, the foregoing problem with inferring the absence of congressional powers from silences in the Constitution’s power-conferring clauses arises from the Necessary and Proper Clause. That clause makes it impossible to construe the boundaries of the full set of congressional powers without recourse to extratextual considerations. Even if all of Congress’s other powers had clear internal limits—a heroic assumption, given the many interpretive questions that might arise about those other powers—the Necessary and Proper Clause would mean that no analysis of congressional power could fix the limit of what Congress can do based on the texts of the power-conferring clauses alone, without the need for substantive judgment about the propriety of congressional action. After all, the Necessary and Proper Clause means that any conclusion that some action lies beyond Congress’s ken requires a judgment that that action is not necessary and proper. And judgments about necessity and propriety are substantive.⁷³

Some writers have been tempted by the thought that modern decisionmakers can avoid making substantive judgments about necessity and propriety by reading the Necessary and Proper Clause to embody a set of eighteenth-century understandings. The idea is that a jurisprudence of original meanings would allow them to implement the judgments that the Founders built into the Necessary and Proper Clause, rather than having to make judgments of their own.⁷⁴ But any attempt to define the scope of the Necessary and Proper Clause by reference to original meanings must cope with the fact that from the very beginning, Americans with different substantive views about national power made different judgments about what the Necessary and Proper Clause authorized.⁷⁵ For reasons one of us has explained at length, a choice to interpret the Clause in accordance with the views of some members of the Founding generation rather than others would be irreducibly normative.⁷⁶ One could make such a choice, but then one’s own choices, not the Founding generation’s, would define the scope of congressional power. In short, the phrase “necessary and proper” will not mechanically draw a boundary between the zone of congressional power and the zone of regulation reserved exclusively to state legislatures—not even if one reads that phrase as the Founding generation read it.

The situation with the Tenth Amendment is similar. Its text suggests the existence of spheres of decisionmaking where Congress should not intervene. But like the Necessary and Proper Clause, the Tenth Amendment does not identify any particular sphere as falling within that category. It neither names nor even suggests criteria for identifying the fields of regulation that (either conclusively or presumptively) should be handled by local decisionmakers. Those fields are instead defined by reference to whatever powers have not been delegated to the United States. So when the scope of Congress’s sixty-plus powers turns out to be pretty broad, the residuum becomes vanishingly small, and the Tenth Amendment’s text offers no resources for identifying where it is important to push back.

The impracticality of deriving robust limits on Congress’s powers from the limits of the Necessary and Proper Clause is well illustrated by the newest commandeering decision, Murphy v. NCAA.⁷⁷ Unlike earlier Supreme Court opinions forthrightly acknowledging that the anticommandeering doctrines do not simply follow from the content of the enumerated powers,⁷⁸ Justice Alito’s opinion in Murphy tried to justify the anticommandeering principle strictly in terms of the internal limits of Congress’s enumerated powers. According to the Murphy Court, the reason why commandeering is unconstitutional is neither more nor less than this: Congress is authorized to do only those things described in its enumerated powers, and nothing in the enumerated powers authorizes Congress to commandeer state officials.⁷⁹

This internal-limits presentation of a doctrine whose real grounding lies in external limits is more than a little unsatisfying. It is of course true that the Constitution enumerates no congressional power to commandeer state decisionmakers. But neither does the Constitution enumerate a power to charter corporations or ration the cultivation of wheat. Those other powers are implied by the Necessary and Proper Clause as reasonable ways of executing projects valid under other congressional powers. Chartering a corporation (say, the United States Postal Service) might be a reasonable step toward executing Congress’s postal power, and rationing the cultivation of wheat might be a reasonable way to regulate national and international commerce in grain. Why, then, is the power to commandeer state officials not also a reasonable way to carry out Congress’s acknowledged powers? For example, the law in Printz directed state officials to help regulate retail sales of a consumer good usually shipped in interstate commerce.⁸⁰ Why isn’t such a law necessary and proper for the regulation of commerce? Nothing in Murphy explains the Court’s skepticism about inferring a commandeering power with the same ease as, say, a corporation-chartering power. A high court can announce by fiat that requiring state officials to implement federal statutes is not “proper.” But such a judgment is not justified by the meaning of the word “proper.” It rests on considerations that lie outside Article I’s text—considerations that Murphy ignores as it seeks to derive the anticommandeering rule from the textual enumeration of powers.

4. The Limits of Purpose-Based Reasoning

When constitutional text alone cannot deliver, interpreters sometimes wisely read that text in light of broader constitutional purposes. In this vein, some theorists have thought that if we can determine the purpose for which the Constitution enumerates congressional powers, we can also establish the content, and therefore the limits, of the textually conferred powers more clearly.⁸¹ But this strategy faces a basic obstacle: limiting federal power might not be the predominant purpose of Article I’s enumeration. Deriving some sort of limit on Congress from the spirit of that enumeration, therefore, is like making soup from a rusty nail in the folk tale—an act of self-deception in which the “nail” (i.e., the overall purpose of Article I) adds nothing but a pretext for supplying ingredients having nothing much to do with the text’s apparent purpose.⁸²

By way of illustration, consider Robert Cooter and Neil Siegel’s important argument that the purpose of Article I’s enumeration of congressional powers is to enable Congress to solve collective action problems among the states.⁸³ Cooter and Siegel would use that account of Article I’s purpose to resolve ambiguities in constitutional text, justifying federal regulation where such collective action problems exist while barring federal law elsewhere.⁸⁴ But even if this approach can succeed in explaining why the Constitution should be read to confer certain powers on Congress, it cannot without more explain why the Constitution should not be read to confer other powers on Congress. That the Constitution aims to empower Congress to solve collective action problems provides no reason to restrict congressional power to such scenarios. Maybe Congress should be understood to have the power to solve collective action problems and also to act in some other situations.

Cooter and Siegel accordingly argue for a background principle by which there should be no national regulation except where the collective-action rationale recommends it. Their “internalization principle,” which resembles the European idea of “subsidiarity,”⁸⁵ holds that all governmental power should presumptively be exercised by the smallest possible political unit, because “local residents possess better information than nonresidents” and “have stronger incentives than nonresidents to monitor the officials responsible for creating and maintaining [local public goods].”⁸⁶ The internalization principle is thus less an inference from the text of the Constitution than an attempt to give rational content to a photographic negative of the idea that Congress is empowered to solve collective action problems. The thought process runs like this: the Constitution seeks to empower Congress to solve collective action problems; therefore, the Constitution does not empower Congress to act in other scenarios; therefore, some principle unites those other scenarios.

It is difficult, however, to derive Cooter and Siegel’s “internalization principle” from Article I’s actual enumeration. As they recognize, Article I, Section 8 reflects the shortcomings rather than the virtues of decentralization.⁸⁷ Article I is accordingly not useful for inferring when the deference ordinarily required by the Necessary and Proper Clause should be suspended. Such an inference requires some concept of national collective action problems—that is, of ways in which Congress is liable to fail to aggregate citizens’ preferences fairly. From the beginning of the Republic, prominent constitutional theorists offered up such theories of congressional failure, pointing to (for instance) the collective action problems confronted by a heterogenous, continental-scale population in monitoring and controlling their elected representatives.⁸⁸ Whatever one might think of such theories’ relevance to constitutional meaning, one cannot derive such theories from the spirit or purpose of Article I, simply because Article I was directed at an entirely different—indeed, opposite—problem.

To justify constitutional limits on Congress, one needs a theory explaining when and to what extent Congress has perverse incentives to overcentralize, as well as a theory of how the answers to that question map on to scenarios where local decisionmaking is especially subject to pathologies of its own.⁸⁹ Just as the answers to those questions will not emerge from the specific texts of the Constitution’s clauses vesting power in Congress, they will not emerge from an account of the purpose for which the Constitution as a whole confers power on Congress. After all, the answers to these questions are not about the purpose for which Congress holds power. They are about the typical pathologies of national decisionmaking relative to the pathologies of local decisionmaking. To map that terrain, we need tools that lie outside not only the text but also outside the purpose of Article I’s enumeration.

5. This Enumeration of Powers Is No Way to Limit a Legislature

The constitutions of some federal systems respond to this problem with a straightforward solution. Rather than leaving the domains of subnational decisionmakers to be inferred as the things omitted from a list of enumerated national powers, they affirmatively specify the topics that should be (conclusively or presumptively) reserved to subnational decisionmakers. For example, the Canadian Constitution expressly specifies that the provincial governments presumptively enjoy the power to legislate on matters including education,⁹⁰ property,⁹¹ and the provinces’ own civil services.⁹²

These affirmative specifications do not eliminate all uncertainty about the legislative jurisdictions of the national and local legislatures in Canada. The powers specified as belonging to each legislature must still be interpreted, and sometimes a piece of regulation might seem to come within grant of power to both the national and a subnational government.⁹³ But the written constitution does supply a basis for identifying the local legislatures as the preferred bodies for certain kinds of regulation. So when national power makes law in one of these spaces, albeit in the apparent exercise of some authority granted to the national legislature, there is a basis in constitutional text for asking whether national power is being read too broadly.

Why doesn’t the United States Constitution use the same strategy? In part, the Framers assumed a relatively small federal footprint,⁹⁴ so as a general matter they did not invest much effort in affirmatively marking the domain of the states. (The qualifier “as a general matter” is important. In the matter of slavery, where some delegates at the Convention were particularly invested in protecting local decisionmaking, the Constitution was written with affirmative limits on Congress, such as the Nonimportation and Export Clauses of Article I, Section 9.⁹⁵) As it happens, the federal footprint has expanded, in ways enabled by the enumerated powers even if not foreseen by the Founders. The Constitution’s failure to specify an area of presumptive subnational authority has accordingly left us in a situation where the text provides few resources for identifying external limits.

It would be a mistake, however, to think that this phenomenon is solely the result of a vast and unforeseen expansion of federal governance unfolding over many generations. It is also a function of the Constitution’s having been written predominantly for the purpose of empowering the national government rather than limiting it.⁹⁶ Obviously, constitutional law must do both of those things. But the Framers went to Philadelphia to solve the problem of insufficient central power, so it should not be surprising that the document they drafted concentrates on the measures that would solve that problem—that is, on conferring rather than limiting national power. To the extent that Article I, Section 8 sounds in federalism, therefore, it makes sense to think of it as having been crafted much more for the purpose of empowering the national government than for the purpose of limiting it.

To many constitutional lawyers, the qualifier “to the extent that Article I, Section 8 sounds in federalism” might seem unnecessary. What was Section 8 written to do, a normal intuition runs, if not to distinguish the zone of national power from that of state power? But that intuition misses something important about Section 8, which is that Section 8 sounds not just in federalism but also in the separation of federal powers. Section 8 lists powers not of the national government as an undifferentiated entity but of Congress. Along one dimension, that means Congress rather than the states. But along another dimension—and a pretty important dimension at that—it means Congress rather than the president or the judiciary. To the extent that Section 8 was drawn to address the separation of national powers, it might not be a helpful guide on questions of federalism at all.

This aspect of Section 8 is easy to see not just because the content of Section 8 has done work in important separation-of-powers controversies⁹⁷ but also by reading Section 8 against the background of two important texts in whose shadow the Philadelphia Convention worked. The first was the Articles of Confederation, which contained its own enumeration of powers delegated to Congress. A fair amount of Article I, Section 8 echoes Article IX of the Articles of Confederation, which empowered the Congress to borrow money on the credit of the United States, to regulate trade with Indian tribes, to coin money and regulate its value, to fix standards for weights and measures, to establish post offices, to declare war, to grant letters of marque and reprisal, to make rules concerning captures on land and on water, to provide for the trial of piracies and felonies committed on the high seas, to build and equip a navy, and to make rules for the government and regulation of the land and naval forces.⁹⁸ It seems reasonable to assume that the Constitution’s drafters had the Articles handy when they did their work, and it is also easy to understand why the Constitution’s drafters might have thought it important to repeat the express mention of powers explicitly specified in the Articles of Confederation. First, omitting specific mention of such powers might invite the inference that a power conferred by the Articles was not conferred by the Constitution. Second, even if future interpreters of the Constitution correctly concluded that those powers were meant to be vested in the national government even though not specifically mentioned, they might then divide on the question of where in the national government those powers were lodged. Under the Articles, all powers vested in the United States were held by a single institution: the United States in Congress assembled. But under the Constitution, “Congress” became just one of three branches of government—and it could not automatically be assumed that the powers held by the United States in Congress assembled under the Articles would be powers of Congress in the new three-branch system. Some of the powers that the Articles had given to the United States in Congress assembled—like the power to send and receive ambassadors and the power to direct military operations—would under the Constitution be allocated to the president.⁹⁹ To omit express instructions about who in the new government could exercise which powers specifically mentioned in the Articles would therefore beg some obvious questions. The express allocation of powers enumerated in the Articles—whether to Congress or to another institution—may have seemed like an easy solution to that problem. And to the extent that Section 8 was a solution to that problem, it was the solution to a problem sounding in the separation of powers rather than in federalism.

The second piece of context heightens the importance of separation-of-powers questions to the composition of Section 8. In his canonical Commentaries on the Laws of England, William Blackstone provided a list of powers that belonged to the King, rather than Parliament, under the British constitution. That list included the powers to regulate commerce, naturalize aliens, coin money, regulate weights and measures, establish courts, declare war, issues letters of marque and reprisal, and raise and regulate armies and navies.¹⁰⁰ Several of those powers were expressly given to the United States under the Articles of Confederation. Other powers that Blackstone described as belonging to the King, like the power to naturalize aliens and the power to create (nonadmiralty) courts, were not delegated to the United States under the Articles but would presumably (or at least plausibly) exist somewhere in the government that the Constitution created. And unless the Constitution specified otherwise, Blackstone’s association of those powers with the King would have grounded an argument for treating them as vested in the president. Those powers, the intuition would run, are powers appropriate for a government’s chief executive officer, as Blackstone’s presentation indicated.¹⁰¹ So if the Framers wanted those powers to be exercised by Congress rather than the president, they would be well advised to say so expressly. And they did.¹⁰²

Situating Article I, Section 8 against these two basic pieces of background should emphasize problems with treating that Section as if it reflected, as an integrated whole, the Framers’ views about what sorts of powers states should be able to exercise without national interference. A fair amount of Section 8 might exist simply because particular powers appeared on previous lists—lists that were not written to instantiate anything like the federal arrangements envisioned by the Framers. One of those lists (from the Articles of Confederation) reflected a federal system that the Framers were rejecting, and the other (Blackstone’s) was not about federalism at all. Seen in that light, trying to use Section 8 as a guide to the limits of national legislative jurisdiction seems like a case of pressing a text into service for a purpose that did not guide its creation. It should not be surprising if the text turns out not to work so well for that imposed purpose.

We wish to be clear about the limits of our present claim. We do not know exactly how the principal drafters or the Constitutional Convention as a whole thought about the origins or purposes of Article I, Section 8. As far as we can tell, nobody does.¹⁰³ The surviving historical records that might answer the question are spotty and often unreliable, and we have attempted no painstaking construction of the drafting history. Maybe the drafters, consciously expecting interpreters to read their work against the Articles and Blackstone, expressly allocated many specific powers to Congress so as to avoid unwanted but foreseeable inferences that those powers were lodged elsewhere. Or maybe in some more inchoate way the earlier lists made certain powers salient, and the drafters mentioned them mostly because they were, from experience, powers of the kind that get mentioned in documents like these. One way or another, it is hard to read the Articles of Confederation and Blackstone’s list of Crown powers without thinking that Section 8 was drafted in their shadow. And to whatever extent dynamics like these drove the drafting of Section 8, that Section was crafted with attention to something other than the question of what powers should be exercised by states rather than by the new national government. Indeed, to whatever extent Section 8 was crafted to prevent future decisionmakers from taking Blackstone as a guide to presidential power, it was not crafted as a technology of federalism at all. It was crafted with the assumption that things left off the list might belong to the president.

To be sure, some important parts of Section 8 are there as a result of the Convention’s deliberate judgment about what powers should be exercised centrally in the reconstituted federal system, not (or not only) because of a concern about the division of power between the president and Congress. The power to tax and the power to regulate commerce are two important examples: letting the general government wield those powers was a central part of the Convention’s project.¹⁰⁴ And in a few respects, Section 8 is indeed naturally read as drafted for the purpose of limiting national power. The requirement that duties, imposts, and excises be uniform throughout the United States is an example,¹⁰⁵ as is the proviso that Congress can secure exclusive rights in writings and discoveries only “for limited [t]imes.”¹⁰⁶ Even those limitations, however, might not reflect the idea that some sorts of regulation should be done locally rather than nationally. (Congressional copyrights and patents are not time-limited so that state governments can issue permanent copyrights and patents;¹⁰⁷ congressional appropriations for armies are not limited to two years so that state governments can make military appropriations for the longer term.¹⁰⁸) To the extent that Section 8 is about federalism rather than the separation of national powers, it seems to have been written with primary attention to the functions of the national government rather than those of the states. Identifying the sorts of decisionmaking that are better done locally was not the activity in which the drafters of Section 8 were engaged. So it should not be surprising that their work product is not a good guide for answering that question.

C. Conclusion: Toward External Limits

In light of all of the problems with using the Constitution’s clauses enumerating congressional powers as a guide to identifying a boundary between state and national legislative domains, one might simply give up the attempt. Perhaps we should reconcile ourselves to a system in which Congress can regulate where it pleases and state law governs in those spaces where Congress tolerates it. Accepting that world would not mean giving up federalism. For a variety of reasons, many of them matters of structure and process, states are likely to be important decisionmakers on a host of issues as far into the future as one could possibly see. But the enumeration of Congress’s powers seems to play little if any practical role in maintaining the federal system.

We think that constitutional decisionmakers should give up the attempt to allocate power between Congress and the states by reference to the Constitution’s enumeration of congressional powers. But we do not think that constitutional decisionmakers should give up the attempt to allocate power between Congress and the states. Instead, constitutional law should approach questions about that allocation with better tools. To use a set of categories introduced earlier, some of those tools should be external limits, not internal ones. That is, they should be affirmative specifications of kinds of legislation that Congress should not enact, or at least kinds of legislation that Congress should not enact lightly.

II. The Possibility of Suspect Spheres

External limits on federal power are nothing new. Unlike the suspect spheres that we urge, however, most prior schemes of external limits on federal power marked categorical limits that Congress was flatly prohibited from overstepping. Consider, for example, the statement in United States v. Butler that Congress could not regulate agriculture,¹⁰⁹ or the attempt to prevent the commerce power from reaching “labor” and “manufacturing.”¹¹⁰ Such limits reflected an attitude of “dual federalism,”¹¹¹ in which judicially enforced constitutional doctrines “divide up the world into spheres of state and federal primacy.”¹¹²

We agree with (and have contributed to) the scholarship pronouncing dual federalism unworkable.¹¹³ The project of having courts police categorical boundaries among regulatory domains like “commerce” and “labor” and “manufacture” was incoherent, as was the attempt to establish a judicially enforceable boundary between “traditional governmental functions”¹¹⁴ and other things that ordinary governments do. To put the point doctrinally, if our idea were entirely successful, National League of Cities would remain as dead in the future as it is today.

In this Part, we offer a way of thinking about certain kinds of policymaking not as exclusively reserved to the states but as suspect when engaged in at the national level. We begin, in Section II.A, by explaining the value of trying to construct such suspect spheres. Our focus is the possibility that well-chosen limits on national policymaking could reduce the stakes of national politics. At a time when polarization in national politics threatens to convulse every part of the constitutional system, devices for lowering the temperature are well worth considering.

In Section II.B, we ask readers to imagine suspect spheres as partly analogous to constitutional doctrine in the realms of due process and equal protection, where courts take harder looks at governmental action deemed “suspect.” Due process doctrine, for instance, requires governments to justify displacing parents’ judgment about what serves the best interests of their children because the sphere of child-rearing is presumptively controlled by parents. This does not mean that governments do not regularly displace parental judgments about children’s welfare. It just means that, when they do so, courts will demand more and better reasons. By analogy, a suspect-sphere approach to federalism would demand that Congress be especially plain and provide stronger reasons when regulating in areas where local decisionmaking is presumptively preferable. The intuition animating this approach is that local and national decisionmaking processes are vulnerable to different (if overlapping) sorts of pathologies. Where national legislation is likely to be more vulnerable to pathology, national legislation is to that extent suspect. But “suspect” is not the same as “forbidden.” And we stress that the analogy is only partial: the role we imagine for courts is weaker for federalism’s suspect spheres than it is in the realm of individual rights.

In Section II.C, we differentiate our approach from some famously unworkable attempts to enforce federalism through external limits at earlier times in history. As noted above, we have no intention of reviving National League of Cities, nor the more general idea of dual federalism, and our approach does not imagine the existence of static domains appropriate for local control (like “agriculture”¹¹⁵ or “family law”¹¹⁶). Nor do we treat state law as normal and federal law as an exception to be contained. Instead, our approach sees both Congress and the state legislatures as normal decisionmaking institutions, each with tendencies toward pathological decisionmaking and neither inherently more trustworthy than the other.¹¹⁷ And crucially, though our conception seeks to raise the cost that Congress must pay to make policy in suspect spheres, it need not suggest a warrant for the judiciary to stop Congress from going places where Congress is determined to go.

Finally, in Section II.D, we address the method by which constitutional decisionmakers could identify federalism’s suspect spheres. That identification would be a matter of constitutional construction rather than of textual interpretation.¹¹⁸ Decisionmaking engaging in that construction would reason from history, structure, and ideas traditionally held within the American constitutional system. We freely acknowledge that the constructive reasoning we envision will not offer determinate right answers. There will be disagreements, and decisions will require the exercise of judgment. We see no way of reasoning about federalism that avoids those challenges.

A. Why Bother? Decentralization as Temperature Control

In times of highly polarized politics, federalism might have special value in mitigating partisan rancor. By allowing people in different regions of a country to adopt different policies on issues where there is intense but reasonable disagreement, federal regimes reduce the stakes of national politics.¹¹⁹ Indeed, they do so in more than one way. First, assuming some correspondence between local public opinion and local law, they permit many people who would be unhappy with controversial policy decisions made nationally to live under rules that they prefer. Second, the subnational jurisdictions could sustain a party out of political power at the national level by demonstrating the feasibility or even desirability of its program.¹²⁰ On this account, federal regimes operate as political insurance policies. In return for reducing each party’s capacity to advance all of its policy priorities nationwide when it is the party in power at the national level, decentralization protects each party against maximum loss when it is the party out of power. The costs of losing any particular round of national elections are reduced. And each party then has less incentive to treat every aspect of national politics as a fight to the death.¹²¹

It is not hard to see why federal regimes offering this benefit can be politically fragile. The party in power must have some credible assurance that its forbearance will be reciprocated when the wheel of fortune turns and the party in power and party out of power trade places. Such credible commitment could be difficult to provide. A party currently out of power might promise to exercise self-restraint in the future in return for the party in power’s self-restraint in the present, but such promises are difficult to enforce. The failure of an incumbent president or congressional majority to exploit its moment of power might accordingly seem more like the concession of a chump than like mature self-restraint. And incumbent parties in power face considerable pressure from their electoral bases to make the most of lawmaking power while they hold it.

This problem is especially intense where political parties are highly polarized and ideologically programmatic. In a system of nonprogrammatic parties, members of different parties have overlapping ideologies that give legislators incentives to delegate policymaking to subnational institutions, thereby ducking controversial decisions that might alienate substantial numbers of their own party.¹²² Recall here how New Deal Democrats delegated the administration of social welfare programs to state and local officials in an effort to assure the party’s Southern wing that national officials would leave issues of race and gender to local control.¹²³ Likewise, the McCarran-Ferguson Act’s turning insurance regulation over to the states in 1945 has proven durable in part because insurers and state insurance commissioners are not programmatically supported by one party and opposed by the other.¹²⁴ By contrast, where national parties are tightly unified by rival and nonoverlapping ideologies, the party in power has more incentive to impose its will on the party out of power at the subnational level. Elections at the subnational level, in effect, become second-order elections on national issues, with state and local legislators running on the record of the national parties—abortion, immigration, war in Afghanistan—rather than on their performance regulating their particular jurisdictions.¹²⁵ Where rank-and-file party members are impatient for comprehensive policy change, party leaders who supported federalism when their party was out of power predictably abandon such after winning a national election.¹²⁶ In short, it is hard for the rival parties to reach the sort of federalist settlement that protects each side from total defeat, even though the system overall might be well served if federalism could play that role.

We acknowledge and largely share the post-Garcia view that courts are poorly suited for policing the boundaries of federalism. But the difficulty that parties in and out of power face in making credible commitments to each other points to an important corollary: the fact that courts are problematic facilitators of federalism does not mean that Congress is optimally equipped for the job, either. So a sophisticated understanding of each institution’s strengths and weaknesses might be consistent with a limited judicial role in this project. Temperature-reducing federalism is something from which everyone benefits, but it is difficult to achieve without a little constitutional help. A few judicial speed bumps—for instance, plain-statement rules requiring Congress to be especially clear when drafting statutes that gratuitously nationalize issues with high cultural or partisan salience—can help each party in Congress achieve temperature-reducing insurance that they both want but cannot easily achieve unaided.¹²⁷ In this sense, the judiciary in a suspect-spheres model provides Congress with something more like a helping hand than an oppositional constraint.¹²⁸

Again, we do not imagine external limits that place any specific domains of policymaking completely off limits to Congress: our system has been down that road before. But we do imagine an approach that would raise the costs of the national government’s preempting subnational policy within suspect spheres. To whatever degree it succeeded, such an arrangement would give each political party some basis for confidence that its self-restraint when it is the party in power would be reciprocated when it is the party out of power.

Critical to this idea of using external limits to ground a modus vivendi between the party out of power and the party in power is that the limits chosen actually provide benefits on both sides of the partisan aisle. A theory of suspect spheres that allowed federal law to preempt policies favored in Blue States (say, medical-marijuana laws) but not policies favored in Red States (for instance, open- or concealed-carry laws) would have little traction as a bipartisan truce. The motivation for the party in power to accede to limits on its power is that it will benefit from those same limits when it becomes the party out of power, so only a system of external limits promising bipartisan benefits can serve the desired purpose. The concept of “suspect spheres” must accordingly be general enough to provide benefits to both parties.

B. Suspect National Legislation: A Partial Analogy to Substantive Due Process

The idea that some congressional legislation could be categorized as falling into “suspect spheres” on federalism grounds relies on a rough analogy to constitutional doctrine in the realms of due process and equal protection, where courts take harder looks at governmental action deemed “suspect.” In all these contexts, background principles of electoral accountability suggest a default principle of significant judicial deference to legislative decisionmaking. In cases involving the Necessary and Proper Clause as well as cases involving equal protection and due process, that default is generally called rational basis review.

In the due process and equal protection contexts, that deference is officially reduced when courts examine governmental decisions that raise suspicions about pathologies in the relevant decisionmaking processes (like racial prejudice) or especially onerous burdens (for instance, on private liberties deemed “fundamental”).¹²⁹ Formally, analysis under the Necessary and Proper Clause does not use different tiers of scrutiny in this manner. But judicial doctrine, in practice, plainly raises the level of scrutiny in particular contexts without much explanation, sometimes by pegging the degree of suspicion to the particular enumerated power being executed (the “congruence and proportionality” standard in cases arising under Section 5 of the Fourteenth Amendment being an easy example).¹³⁰ As Aziz Huq has argued, this approach to degrees of suspicion does not make much sense.¹³¹ It does not follow, however, that tiers of scrutiny pegged to factors unconnected to Article I powers are equally senseless. Indeed, a system of varying tiers of scrutiny for federal legislation is a de facto part of current federalism doctrine, inasmuch as the courts give narrow readings to federal statutes deemed to “upset the usual constitutional balance of federal and state powers.”¹³²

Gregory v. Ashcroft illustrates that this “usual balance,” whatever it might be, is not defined by Article I. In reading the Age Discrimination in Employment Act (ADEA)¹³³ narrowly in order to uphold Missouri’s mandatory retirement age for state judges, Justice O’Connor’s majority opinion admitted that the statutory language did not naturally lend itself to the narrow reading that she chose.¹³⁴ Because the literal reading of the ADEA was deemed to “interfere[] with this decision of the people of Missouri, defining their constitutional officers,”¹³⁵ the Gregory majority held that the ADEA should not be read to cover judges unless that coverage would “be plain to anyone reading the Act.”¹³⁶ Article I says nothing whatsoever about whether Congress can interfere with state definitions of its constitutional officers. Gregory’s principle that congressional legislation of that sort is to be narrowly construed is, in effect, a statement that such legislation falls into a suspect sphere, and that sphere is grounded in considerations independent of the Constitution’s enumeration of congressional powers.

The Court’s jurisprudence suggests that such suspect-sphere analysis can guide the judiciary’s approach to federal regulations of private persons as well as state officials. Consider Bond v. United States.¹³⁷ The defendant in Bond tried to inflict a skin rash on her husband’s lover with an arsenic-based compound.¹³⁸ Though far from the sort of atrocity the drafters of international chemical weapons conventions typically contemplate, that conduct straightforwardly satisfied the prohibition on “chemical weapons” described in the Chemical Weapons Convention Implementation Act of 1998, which prohibits any knowing “use” of any “toxic chemical,” except as specified by the statute.¹³⁹ The Supreme Court nonetheless vacated her conviction, relying on Gregory v. Ashcroft’s principle that statutes ought to be read against “the background principle that Congress does not normally intrude upon the police power of the States.”¹⁴⁰

The Court in Gregory and Bond did not merely apply a canon of construction that counsels narrow readings of federal statutes whose meaning is otherwise in equipoise. It put a robust thumb on the scale against federal law over some range of topics.¹⁴¹ Indeed, given the statutory text at issue in each case, and the degree to which each decision seems determined to resist the ordinary meaning of that text, the metaphor of the thumb might be too meager: Gregory and Bond seem more like several fingers, if not an entire fist. But in each case, it is plausible to think that Congress might not have wanted the outcome that a less federalism-conscious construction of the statute would have supported, so it is not unreasonable to think that the Court’s resistance to federalization in those cases would give way if Congress came back with a pellucid indication of an intention to federalize. In sum, the Court’s construction of the statutes in Gregory and Bond tempered the reach of federal law without denying Congress’s authority to make policy if Congress really wants to do so in the future.

Reasonable people can differ about whether Gregory and Bond were rightly decided. But regardless of the merits of those two particular decisions, “plain-statement rules” of the kind the Court deployed in those cases show how heightened scrutiny can limit power without walling topics categorically out of Congress’s jurisdiction. In the presence of a serious federalism concern, the judiciary can interpret statutes somewhat obstructively. But the qualifier “somewhat” is important. Congress remains able to make policy, even in these precise areas, if it proffers a suitably strong justification or manifests its intention with sufficient clarity.

In this sense, the Gregory–Bond canon has a function analogous to the triggers for heightened scrutiny under the doctrine of substantive due process. Consider, by analogy, the doctrine that state laws regulating parental decisions regarding the education and care of children should be subject to heightened scrutiny.¹⁴² Nothing in this doctrine forecloses extensive state regulation of parenting decisions. State law indeed regulates everything from vaccination to school attendance without offending the Fourteenth Amendment.¹⁴³ Nevertheless, state decisions that second-guess parenting decisions without reason are subject to searching judicial examination that other types of state law do not receive.¹⁴⁴ There is no due process category of “parenting” that is off-limits to the law, but state laws that regulate parental decisionmaking are in a suspect sphere where legislation requires more justification than usual and legislatures receive less than ordinary deference.¹⁴⁵ Suspect spheres in federalism works the same way in decisions like Gregory and Bond.

Clear-statement rules of the kind used in Gregory and Bond are not the only devices by which courts might raise the costs of federal lawmaking in suspect spheres. Depending on the context, courts might reduce the deference they show to the legal interpretations,¹⁴⁶ factual findings,¹⁴⁷ or guidance documents¹⁴⁸ issued by federal agencies. Whatever form reduced judicial deference takes, however, it is important that suspicion not be fatal and that Congress have the last word, so long as its determinations are backed by sufficiently plausible reasons. The suspect spheres defined by the Gregory–Bond canon are thus weaker than the due process and equal protection limits, which require some sort of weighty interest to justify intrusions into parental authority. If a court holds a statute invalid on the grounds that it violates due process or equal protection, Congress cannot overcome the court’s judgment simply by saying, “Actually, we really want to do that.”¹⁴⁹ By contrast, the suspect-sphere external limits suggested by the Gregory–Bond canon are soft. The courts are putting up resistance, but they are not policing static and inviolable boundaries.

C. Suspect Spheres Versus Dual Federalism

The “soft” nature of external limits within a suspect-sphere approach is one way that the suspect-spheres model would not reproduce the well-known problems of the system called “dual federalism.” To repeat, dual federalism is the theory that state and federal governments should swim in their own lanes, regulating distinct, nonoverlapping topics separated by a rigid constitutional barrier.¹⁵⁰ That image of American federalism has rightly been criticized as both descriptively fanciful and conceptually unworkable.¹⁵¹ The conceptual problem is that it makes little sense to imagine that categorical terms like “education,” “criminal law,” “family law,” “commerce,” and “taxation” define mutually exclusive regulatory realms that can be allocated separately to state or federal control. The social world does not come in separate packages corresponding to categories like these. Attempts to carve the social world into such categories with legal rules predictably invite lawyerly efforts to manipulate the labels in ways that win cases without making substantive regulatory sense.¹⁵² The descriptive problem is essentially a manifestation of the conceptual problem: dual federalism imagines a world in which state and federal governments pursue separate goals with distinct resources—in Gerken’s phrase, “engaged in the governance equivalent of parallel play”¹⁵³—but the actual world looks nothing like that. In reality, state and federal officials “govern shoulder-to-shoulder in a tight regulatory space,”¹⁵⁴ cooperating or fighting with each other, in each of many shared policy domains, to accomplish overlapping purposes. This intertwining of federal and state governance is longstanding, pervasive, and ordinary.

Our suspect-sphere approach avoids these problems. It does not imagine that the areas where federal policy should not lightly displace state policy correspond to static categories like “education” and “family law.” Instead, it seeks to identify reasons why local control is sometimes preferable and then to put a thumb (or a few fingers) on the local side of the scale when those reasons apply. Unlike the dual-federalism model, the suspect-spheres model embraces the reality of mixed federal and state governance in regulatory sphere after regulatory sphere.¹⁵⁵ It identifies areas where federal policy should not lightly displace state policy, but it never imagines substantive policy realms in which Congress simply may not act.

Others have worried, not unreasonably, that trying to differentiate between policy areas where federalization is and is not suspect is hopeless. In Ernest Young’s formulation, the attempt to defend federalism by treating congressional action more skeptically in some regulatory fields than in others “reintroduces the same confusion that led to the demise of dual federalism in the first place.”¹⁵⁶ To avoid that problem, Young urges courts to adopt a general presumption against preemption—that is, a skeptical posture toward any federal law that conflicts with state law. That solution would indeed prevent courts from having to make contestable judgment calls, and as such it has an undeniable appeal. But it also has significant weaknesses. For one thing, if Young were right that courts should abandon all hope of differentiating among fields of regulation substantively, it is not clear why the solution should be a wall-to-wall presumption against preemption rather than a wall-to-wall presumption in favor of preemption. That too would eliminate the need for sorting, and it might be more consistent with the Constitution’s central purpose of enabling national legislation.¹⁵⁷ Moreover, as Young candidly acknowledges, the Supreme Court has never endorsed such an indiscriminate speed bump to all federal preemption.¹⁵⁸ Instead, it has found preemption more readily in commercial transactions, admiralty, and banking than in criminal law, family law, tort law, and land use law.¹⁵⁹

Just as it does not make sense to treat all laws as equally suspect under the Fourteenth Amendment, it does not make sense to treat with equal suspicion all federal legislation that displaces any state law. The better course of action in either case is to tailor presumptions to the relevant risks. Used well, suspect spheres can do in federalism cases an important part of the work that heightened scrutiny does in individual rights cases: increase the burden of justification for types of law that have a greater risk of proceeding from pathological reasons or imposing indefensible burdens. And if the courts avoid treating policy fields as statically divided between governments, and similarly avoid the idea that they have the ultimate authority to police that division, a suspect-spheres approach will not reproduce the problems that make dual federalism nonviable.

D. Suspect Spheres as Constitutional Construction

The next hard question, of course, is how constitutional decisionmakers could identify the suspect spheres. In what instances, exactly, should Congress be reluctant to legislate, and courts be willing to blunt and delay national action, in order to deliver the decentralization that might moderate the stakes of national politics? The text of the Constitution will not answer that question. Instead, the shape of the suspect spheres must be a matter of constitutional construction,¹⁶⁰ one that draws upon sources from within what scholars sometimes call the “small-c constitution.”¹⁶¹ Those sources include historical experience for two reasons. First, experience furnishes useful knowledge—indeed, knowledge not available to the Founders—about how American federalism actually works. Second, patterns of historical practice can legitimize constitutional arrangements with the venerability bestowed by time, both with and without the assistance of formal precedential decisionmaking, judicial or otherwise.¹⁶²

We are keenly aware of the many ways in which “looking to history” for constitutional authority can go wrong. Many attempts to ground constitutional law in historical experience are shallow, peremptory, or downright fictional. Even good accounts of history are regularly challengeable with other good accounts. Put generally, “history” is not a force that directs determinate constitutional principles; it is a source of raw material that can be used to support constitutional arguments.¹⁶³ But to make a point just a little polemically, the same could be said about the text of the written Constitution, or at least about most of the clauses of the written Constitution that are at issue in cases that make their way to court. The potential for drawing multiple and conflicting inferences from the historical record is, accordingly, not a sufficient reason to abandon history as a source of constitutional authority. It is simply a reason to try to use the resource intelligently, with due attention to what it is being used to show in any given instance and whether it can in fact show that thing. And sometimes, historical information is rather useful for sound constitutional decisionmaking. For example, historical knowledge sometimes suggests a popular consensus about constitutional meaning,¹⁶⁴ a political settlement that courts ought to accept for the sake of stability,¹⁶⁵ or a natural experiment showing the practical consequences of a political system.¹⁶⁶

Our arguments about federalism and history are intended to demonstrate the potential for establishing criteria that can help decisionmakers identify areas of regulation where federal lawmaking should be treated with some suspicion. We investigate that potential in part by noticing themes in the past behavior and historical intuitions of American decisionmakers trying to make federalism work. We do not claim that ours is the only possible way of reading the historical record, and we do not claim that facts about history compel our recommendations for constitutional law. We hope to show, however, that a suspect-spheres approach would not be a radical break with American constitutionalism as practiced to this point. On the contrary, it would have thematic antecedents in the constitutional past. In sum, it would be a strategy consistent with insights that can be gleaned from one responsible reading of American constitutional history—both the history of how federalism has actually operated and the history of various attempts, more and less successful, to make it work.

III. Illustration: The Corporate Nondelegation Doctrine

The external limits that we propose must meet a set of demanding standards. They must pick out types of federal law as presumptively suspect without preventing intermingled federal and state policymaking. They must find some support in American constitutional practice, both to suggest their functionality and to help establish their legitimacy as a tool of constitutional reasoning. And they must support a bipartisan truce aimed at lowering the stakes of national political conflict, one that politicians on both sides would favor but cannot achieve without some mechanism for credible constitutional commitment.

One might reasonably ask whether such a doctrine is practically possible. In what follows, we seek to illustrate the possibility of a soft-limit, suspect-sphere approach with a case study drawn from the United States’s constitutional traditions regarding delegations of governmental power to private corporations. After considering this attempt, readers can draw their own conclusions about the viability of such an approach.

The potential suspect sphere we describe is animated by a principle that we call the corporate nondelegation doctrine. The essence of that doctrine is suspicion toward federal laws that delegate broad immunities or powers to private corporations. In this Part, we illustrate this principle’s deep constitutional history. It played an important role in James Madison’s attack on the First Bank of the United States, in Andrew Jackson’s attack on the Second Bank, and in the Supreme Court’s invalidation of the National Industrial Recovery Act (NIRA) of 1934. As that set of examples suggests, the corporate nondelegation doctrine functioned both as a legal conception animating judicial action (as with NIRA) and as a principle of constitutional politics (as with Madison and Jackson). At the end of this Part, we will be in a position to notice some ways in which the corporate nondelegation doctrine helps make sense of modern federalism cases.

The idea that delegations of power to private corporations are especially suspect was a dominant theme of antebellum American constitutional law.¹⁶⁷ But it had no foundation in the Constitution’s text, whether in the enumeration of congressional powers or anywhere else. The text of the Constitution is silent about Congress’s power to confer powers on corporations. That silence is not an accident. At the Convention, the delegates considered enumerating a specific power to charter corporations but decided against it, largely because they worried that such an express power would provoke opposition during the ratifying process.¹⁶⁸ (The delegates did not think that their declining to enumerate such a power would prevent Congress from chartering corporations; they just thought it would make the issue less salient.) Then, in the state conventions, some prominent Anti-Federalists tried and failed to attach amendments prohibiting federal aid to corporations.¹⁶⁹ So the adopted constitutional text remained silent on the issue, with each side’s having failed to make that text conform to its preferences.

Subsequently, politicians opposed to federal aid to corporations often made their arguments in constitutional terms. Those arguments were sometimes presented as if they flowed from the text of Article I’s enumeration of congressional powers.¹⁷⁰ But in truth, the argument against the constitutionality of aid to corporations did not turn on any specific constitutional text. It turned on the idea that there were special risks to allowing large private corporations to manipulate Congress. In other words, the principle that these politicians argued for is best understood as an external limit on Congress—a prohibition arising from a reason Congress should not act—and not an internal limit resulting from anything about the scope of the enumerated powers.

A. Madison and the First Bank

Consider Madison’s 1791 stand against the bill to charter the Bank of the United States as an example of an argument rooted in historical sources external to the text of Article I. Madison’s argument enlisted a new canon rooted in an extratextual consideration—a presumption that federal laws conferring benefits on private corporations were especially suspect as incidental means for executing enumerated powers. According to Madison, the proposed power to charter a bank was not a necessary and proper means because it was “a great and important power” rather than an incidental one.¹⁷¹ The measure of its “importance” was rooted in Madison’s extratextual assessment that such a bank charter threatened citizen equality and democratic equality by enriching and empowering a set of already-wealthy stockholders,¹⁷² committing the government to deal only with a single corporation,¹⁷³ and inviting the corruption of officeholders and the undue influence of the financial class.

Madison’s conclusion that bank charters fell within a suspect sphere of federal powers rested on two points critical for understanding how suspect spheres work. First, Madison rooted the canon entirely in considerations of constitutional purpose external to the text of the enumeration. Second, Madison derived these external considerations from Anglo-American culture and history rather than from the text of the Constitution.

Consider, first, how Madison’s suspicion of congressionally conferred corporate charters rests on reasoning external to anything in Article I. As Jonathan Gienapp has powerfully argued, Madison’s argument contained several important interpretative innovations.¹⁷⁴ One such innovation, recently highlighted by William Baude, is the presumption against inferring that Congress’s incidental powers included any “important” powers, on the ground that a reasonable drafter would not leave important matters to implication.¹⁷⁵ That “important” powers might be excluded from the scope of incidental powers was a plausible claim. But the “importance” of any claimed incidental power had nothing to do with inferences from constitutional text. Instead, Madison inferred that chartering corporations was an “important” power because it posed risks to important extratextual values like political equality.¹⁷⁶ Put more generally, although it makes sense to presume that “important” matters should not be lightly inferred as “necessary and proper” for the execution of expressly enumerated powers, one cannot derive the list of “important” powers not specified in the Constitution from reading the text of Article I, and Madison did not try.

How, then, did Madison support his inference that the power to charter a bank was an “important” power? He relied on cultural considerations external to the text of Article I, resonating with an Anglo-American tradition of suspicion toward corporate power already of long standing in 1791. A generation before Madison was born, it was powerfully manifested in the pamphlet attacks on the corruption of the British government by the South Sea Company in the 1720s.¹⁷⁷ At least one set of those pamphlets—Cato’s Letters, by John Trenchard and Thomas Gordon¹⁷⁸—was widely read in the American colonies and helped make the denunciation of corporate finance—“stock-jobbing,” in 18th century jargon¹⁷⁹—a prominent theme of eighteenth-century American political discourse.¹⁸⁰ These tropes were staples of American polemics against the British government after 1773, when Parliament enacted the Tea Act to bail out the East India Company. Like the South Sea Company, the East India Company had loaned millions of pounds to the government and in return secured major economic concessions—including a monopoly on the right to sell tea in North America.¹⁸¹ And as was to become important in the Bank debate, Madison and other leading Americans had come to think that the tendency to be corrupted by financial insiders was not equal in all governments. The centralized governments of large empires were especially vulnerable.¹⁸²

Madison’s characterization of a congressional power to create and patronize “an incorporated monopoly bank”¹⁸³ as particularly dangerous, and therefore as not to be read as implied in the Constitution, drew on this deep well of rhetoric and ideology. It was that tradition, rather than the text of Article I or any implicit purpose of subsidiarity embedded in that text, that allowed Madison to distinguish the Bank of the United States from postal monopolies¹⁸⁴ and other less frightening fruits of Congress’s early legislation. And for former Anti-Federalists, Madison’s embrace of suspicion towards private corporations with special connections to the government was common sense. Throughout the ratification debates, Anti-Federalists had urged that continental-scale legislatures like the proposed Congress were at risk of being captured by financiers, merchants, and other elites—“natural aristocrats,” in their phrase—dominant in Eastern Seaboard cities.¹⁸⁵ Behind this fear of capture lay the reality that urban elites enjoyed organizational advantages that would matter in politics. Communication among merchants, lawyers, and financiers across a continent was facilitated by networks created by trade, universities, and newspapers, most of which were controlled by Federalists.¹⁸⁶ If national politics became a competition to influence the allocation of wealth, farmers in the hinterland who enjoyed none of those organizational assets would be at a severe disadvantage.¹⁸⁷

Madison himself was a recent convert to the Anti-Federalists’ way of looking at this aspect of political mobilization in geographically large polities. In 1787, in Federalist 10, Madison had famously praised the great extent of the United States as a positive feature that would prevent factional legislation.¹⁸⁸ He changed his mind within four years.¹⁸⁹ In his 1791 essay “Consolidation,” Madison warned that the diversity of interests in the “extended Republic” that he formerly praised as a solution to the problem of faction might instead lead to paralysis in Congress, because “neither the voice nor the sense of ten or twenty millions of people, spread through so many latitudes as are comprehended within the United States, could ever be combined or called into effect.”¹⁹⁰ Rather than a healthy legislature that could break and control the violence of faction, the result might be an ineffective legislature and a corresponding rise of the policymaking power of the executive branch. When Hamilton as Secretary of the Treasury succeeded in getting Congress to charter the Bank and then used open-market operations to manage credit, Madison and others of like mind saw a realization of Anti-Federalist fears that the federal government’s scale made it prone to capture by financiers.

One might argue that Madison’s reliance on a traditional hostility toward aid for corporations sounds in partisan politics rather than constitutional construction. That would be too quick. Madison explicitly framed his argument in substantively constitutional concerns: he formally distinguished between his “remarks on the merits of the bill” and his argument against “the authority of Congress to pass it,”¹⁹¹ thus making the standard constitutional advocate’s move of saying, “This is a bad bill, but apart from that, it’s unconstitutional, so you must oppose it even if you like its policy.” Moreover, Madison and his allies presented their opposition to the Bank as flowing from—indeed, required by—the Republic’s fundamental nature. To recognize a congressional power to charter the Bank, Madison warned, would transgress the Constitution’s “essential characteristic” as the plan of a limited government.¹⁹² In other words, Madison was making an argument about both the structure of American government and its ethos—and arguments about the structure and ethos of American government are, in an important sense, substantively constitutional arguments, whether or not they are arguments about the meaning of words in the written Constitution.¹⁹³

Moreover, Madison’s framing his argument as constitutional was not merely a rhetorical maneuver aimed at persuading his fellow congressmen to vote against the Bank. It was made necessary by the imperatives of building a coalition capable of protecting the fundamental interests with which he was concerned. If citizens were to resist Hamilton and his Eastern financiers, they would have to bridge divisions among rival interest groups, including Madison’s own landed gentry and the growing population of the West. Madison articulated this challenge directly in “A Candid State of Parties,” a 1792 essay that again showed a marked change from the views Madison had expressed as Publius. “The anti republican party,” Madison warned, would try to prevent the formation of a unified opposition “by reviving exploded parties, and taking advantage of all prejudices, local, political, and occupational, that may prevent or disturb a general coalition of sentiments.”¹⁹⁴ (A more prosaic formulation might simply have said that the financiers might hold on to power by buying western support with infrastructure and easy credit.) To counteract this effort, the gentry needed to “bury[] all antecedent questions” and “banish[] every other distinction than that between enemies and friends to republican government.”¹⁹⁵Framing the case against the Bank as a fundamental issue of constitutional principle helped.¹⁹⁶

That Madison characterized his argument as constitutional did not, by itself, confer constitutional status on either his reading of the Necessary and Proper Clause or his underlying suspicion of corporate power. Congress voted against him, after all, and the Bank’s charter became law.¹⁹⁷ But Hamilton’s victory in 1791 did not extinguish Madison’s substantive constitutional concern. Eventually, suspicion of the national government’s involvement with corporations and monopolies became a central component of the Democratic Party’s platform. In that form, the corporate nondelegation doctrine shaped the dominant reading of the Constitution prior to the Civil War.

B. Jackson and Anticorporate Constitutionalism

Like Madison’s argument against the First Bank, Andrew Jackson’s argument against the Second Bank of the United States owed little to the text of Article I, Section 8. It was instead an argument for an external limit, rooted in the fear that federal aid to private corporations would favor privileged insiders. And unlike Madison’s proposed limit, Jackson’s had legs: it became the Democratic Party’s consensus position until the Civil War.

Given that Madison and his allies treated their suspicion of private corporations as a matter of constitutional dimension, one might have expected something like the corporate nondelegation doctrine to triumph once the Democratic-Republicans became the dominant party. But matters were not so straightforward. Once the Democratic-Republicans took control of the national government, they seemed to abandon the principle. Confronted with the practical problems of providing credit and infrastructure for a continental-scale nation-state, they gradually accepted not only the idea that the national government would sponsor a private bank with monopoly privileges but also the related idea that the national government should subscribe to the stock of private canal- and road-building companies.¹⁹⁸ In 1816, President Madison signed the Second Bank of the United States into law, ambiguously relying on a newly minted¹⁹⁹ argument that Congress’s Article I, Section 8 power to coin money authorized the Bank or, alternatively, that the question was settled by political precedents established between 1791 and 1811.²⁰⁰

But just as Representative Madison’s legislative defeat in 1791 did not purge constitutional discourse of the corporate nondelegation idea, neither did President Madison’s decision to set it aside in 1816. The idea persisted, and a generation later President Andrew Jackson made it a centerpiece of the Democratic Party’s platform and eventually the basis for the federal government’s monetary and infrastructure policy. In those forms, the corporate nondelegation doctrine provided a framework for federalism—one rooted in ideology, tradition, and the structure of national politics, though not in the text of the Constitution.

Jackson strenuously opposed federal aid to private corporations, which he characterized as special privileges for influential insiders. By manipulating the national government to subscribe to road- and canal-building companies, Jackson charged, financiers were using “artful expedients to shift upon the Government the losses of unsuccessful private speculation”²⁰¹ and pushing Congress to exercise “an authority unknown to the Constitution and beyond the supervision of our constituents.”^{²⁰²Andrew Jackson, Second Annual Message to Congress (Dec. 6, 1830), https://www.presidency.ucsb.edu/documents/second-annual-message-3 [https://perma.cc‌/G7WC-KZ68]. Jackson invoked this ban on federal subscriptions of private corporations’ stock to justify his pocket veto of an appropriation to purchase stock in the Louisville & Portland Canal Company. See 7 Reg. Deb. app. at v (1831).} His objection to the Second Bank was similar. The use of a single private bank as the federal government’s fiscal agent, Jackson maintained, was a contrivance to grant special privileges for the further enrichment of a financial insider class. Decrying “artificial distinctions, . . . titles, gratuities, and exclusive privileges” that “make the rich richer and the potent more powerful,” Jackson argued that the privileges conferred by the Second Bank’s charter, from owning real estate and holding federal revenues to immunity from state taxes, were all too extensive to delegate to a private corporation.²⁰³

Some modern readers might misread Jackson here as offering a general objection to Congress’s delegating important powers rather than exercising those powers itself. But in fact, the problem Jackson saw was more particular. He had no objection, after all, to Congress’s delegating enormously broad power to the legislatures of the federal territories.²⁰⁴ The objection was to a private body’s exercising delegated governmental powers. And in keeping with that particular worry, Jackson specifically situated his concern within the same tradition of anticorporate suspicion that animated the Anti-Federalists. In his famously candid remarks to Nicholas Biddle, the president of the Second Bank of the United States, Jackson explained that he had been afraid of banks “ever since I read the history of the South Sea bubble.”²⁰⁵

We do not know in what text Jackson had read about the South Sea Bubble. But whatever it was, it was not Article I, Section 8. Jackson’s worries, though eminently constitutional in the sense described above, were not naturally expressed by anything in the Constitution’s enumeration of congressional powers. They were worries about the likely pathologies of national lawmaking, best addressed through external limits—that is, affirmative prohibitions on things like granting corporate charters—rather than by negative implication from the powers affirmatively given to Congress.

Jackson’s constitutional argument, in sum, followed exactly Madison’s argument that chartering a private corporation like the Bank of the United States was too “great and important” a power to be lightly inferred because such delegations were incompatible with general ideas of democratic equality implicit in the Constitution. This sort of inference owed nothing much to any text in Article I or, for that matter, in any other article of the Constitution.²⁰⁶ Instead, Jackson grounded his argument in his understanding of Anglo-American culture and history dating back to the South Sea Bubble.

One might reasonably ask why Jackson should use this extratextual concern to limit Congress’s powers in particular, rather than state and federal governmental power more generally. Jackson’s own antipathy for financial corporations was certainly ecumenical enough to support that expansive form of the doctrine: as Jackson famously told Biddle, “I do not dislike your Bank any more than all banks.”²⁰⁷ Not surprisingly, Jackson indeed called for a more total separation between government and finance, urging that state governments’ use of private banks as fiscal agents to regulate the supply of currency violated Article I, Section 10, which prohibits states from issuing “bills of credit” or paper money.²⁰⁸ More generally, the Democratic Party’s “hard money” radicals urged a strict constitutional principle barring all aid to private corporations, under the slogan “separation of bank and state[!]”²⁰⁹

But that extreme position could not hold together Jackson’s political coalition, which was an uneasy alliance between westerners hungry for infrastructure and credit, urban radicals suspicious of inflation and banks, and southerners suspicious of the federal government.²¹⁰ Important Democratic Party leaders like David Henshaw were also up-and-coming financiers who resented the Bank of the United States’s “aristocratic” dominance but who also did not want to push state and federal government entirely out of the banking sector.²¹¹ The more limited principle that the federal government should stay away from subsidizing private banks could, by contrast, deliver something important to both urban radicals (limiting revenue to private banks) as well as rising leaders like Henshaw (less competition from a federal “monopoly”) while preserving western farmers’ access to credit.

In other words, the Jacksonian understanding of federalism was a coalition-supporting compromise. This compromise eventually became the Democratic Party’s strictly enforced constitutional dogma with the Independent Treasury System, under which the federal government would hold its own revenue in the form of coin stored in governmentally owned “subtreasuries,” on the constitutional theory that private corporations could not be given power over public functions like control of the currency supply.²¹²

The antebellum Democratic Party’s anticorporate federalism, in sum, intertwined constitutional principle and practical politics. Indeed, the principle helped make the politics possible. Each constituency within the Jacksonian coalition needed assurance that the others would accept an arrangement that foreclosed some policies that each one favored, thus making everyone in the coalition better off as long as everyone was willing to compromise somewhere. Without that assurance, some Democrats would have been tempted to defect to the Whigs to pursue their top policy preferences. (Some did.)²¹³ By articulating the Party’s position in terms of constitutional principle, and by rigorously ostracizing those who defied party “regularity” as traitors to the Constitution, the party leadership was able to signal a stable commitment to hold the coalition together.

The Jacksonian Democrats’ version of anticorporate federalism, like Madison’s, owed little or nothing to the text of Article I. It was a theory of suspect spheres, not one of enumerated powers. If the text of Congress’s enumerated powers were the true measure of what Congress could constitutionally do, it would have been child’s play to justify federal laws doing all sorts of things to which the Democrats were constitutionally opposed. Federal power over the currency supply was obviously useful for interstate commerce and federal borrowing. Federal charters and subsidies for rail corporations, turnpikes, and canal companies would obviously facilitate interstate commerce and the delivery of the mails. To Democrats, however, these links to Congress’s enumerated powers were not enough to justify federal aid to corporations, because recognizing a power to confer such aid carried the dangers of anti-republican corruption. The Democrats could not and did not infer those dangers from anything in the letter or spirit of the list of powers in Article I. That list, after all, reflects the benefits, not the dangers, of national power. Instead, they read the Necessary and Proper Clause through the lens of suspicion they constructed from popular belief and tradition.

C. Schechter Poultry and the Obvious Answer

After the Civil War, the corporate nondelegation doctrine was contested by banks and railroads seeking to exercise delegated federal regulatory power. These banks and railroads failed in that quest: the corporate nondelegation doctrine prevailed. The reasons why the corporate nondelegation doctrine prevailed had nothing to do with the enumeration of powers in Article I, Section 8. Nonetheless, the repeated refusal of Congress to delegate regulatory power to banking, railroad, and manufacturing companies created a settled expectation of suspicion towards such delegations, one so powerful that the U.S. Supreme Court eventually took it to be obvious that such delegations were constitutionally forbidden.

The Republicans who took control of the federal government after 1860 did not share the Democrats’ anticorporate constitutionalism. They did not restore the old system of vesting monetary policy in a single, privately owned Bank,²¹⁴ but they were perfectly willing to let Congress give direct aid to private corporations. Federally chartered railroad companies were prominent examples.²¹⁵ Moreover, the Republican-controlled judiciary construed the National Bank Acts to have broad preemptive force, thus significantly expanding nationally chartered banks’ immunities from state regulation.²¹⁶ Throughout the late nineteenth and early twentieth centuries, important Republican constituencies persistently lobbied for a return to Hamilton’s vision of an economy managed by leaders of private corporations, on the theory that only those managers had the necessary expertise. Private financiers decried the incompetence of the Treasury Department²¹⁷ and called for the money supply to be regulated instead by an association of private bankers.²¹⁸ Railroad companies complained that bureaucrats did not understand rate making²¹⁹ and called for the federal government to let agreements among the railroads themselves set the rules for “pooling” (that is, cooperation among private lines) as well as the rates for transportation.²²⁰ And in the wake of their experience with cartelization by the War Industries Board during World War I, the leaders of nationwide firms called for the relaxation of antitrust laws in favor of “self-government in business,”²²¹ by which the corporations themselves would act to prevent “destructive competition” that could undermine high wages and product quality.²²²

Given the persistence of these efforts by powerful members of the Republican coalition, it is noteworthy that all these campaigns to have the federal government turn economic management over to private industry failed spectacularly. The plan to delegate monetary policy to a bevy of private banks was laughed out of Congress.²²³ Instead, the Federal Reserve Act placed monetary policy under a board controlled by Senate-confirmed presidential appointees.²²⁴ Rather than acceding to railroad managers’ pleas for self-governing pools, Congress vested the power to regulate mergers and rates in the Interstate Commerce Commission.²²⁵ These refusals to vest federal authority in private corporations were not articulated as resting on any constitutional prohibition. But the idea that private businesses should not be trusted to exercise regulatory power was strong enough to command bipartisan majorities, even in the face of the determined contrary efforts of powerful interest groups. In that respect, the idea functioned like a constitutional principle, albeit not one enforced by the courts or associated with specific words in the Constitution.

What often converts a long-accepted principle enforced by policymakers’ self-restraint rather than by courts into a rule that courts will enforce as a constitutional prohibition is the breach of that principle by some innovative policymaker. In the case of the principle disfavoring delegations of federal power to private corporations, the innovation-provoking judicial reaction was part of the New Deal’s National Industrial Recovery Act (NIRA). That statute, which was intended to help pull the economy out of the Great Depression, vested the regulation of wages, hours, working conditions, production standards, and prices in private trade associations loosely supervised by the president.²²⁶

A broad coalition of stakeholders initially supported this reform. Industrialists hoped to stabilize prices.²²⁷ Trade unions hoped for the protection of collective bargaining.²²⁸ Anti-market intellectuals dreamed of an economy in which the needs of consumers, labor, and industry might be balanced by expert planners.²²⁹ Before long, however, critics characterized the NIRA as empowering industry cartels that squeezed smaller businesses while keeping wages stagnant and prices high.²³⁰ Taking stock two years after the law’s passage, the nonpartisan Brookings Institution called the Act “an exercise in counterproductivity” that “on the whole, retarded recovery.”²³¹ And in that environment, the Supreme Court in A.L.A. Schechter Poultry Corp. v. United States ruled unanimously that the NIRA unconstitutionally delegated federal legislative power.²³²

Because the president had to approve the “codes of fair competition” that the NIRA’s trade associations formulated, the delegation of power held unconstitutional in Schechter was formally a delegation to the president, not a delegation to private corporations.²³³ But the federal government lacked the manpower to review the trade associations’ proposed codes, so presidential approval was perfunctory.²³⁴ As a practical matter, the codes were privately drafted regulation with the force of federal law. Indeed, the Roosevelt Administration tried to defend the propriety of the codes precisely on the ground that they were not governmental impositions but “rules of competition deemed fair for each industry by representative members of that industry—by the persons most vitally concerned and most familiar with its problems.”²³⁵

The Court was having none of that. On the contrary, it regarded the idea that private entities should engage in public regulation—long pressed by Republican business interests—as completely beyond the pale:

[W]ould it be seriously contended that Congress could delegate its legislative authority to trade or industrial associations or groups so as to empower them to enact the laws they deem to be wise and beneficent for the rehabilitation and expansion of their trade or industries? Could trade or industrial associations or groups be constituted legislative bodies for that purpose because such associations or groups are familiar with the problems of their enterprises? And, could an effort of that sort be made valid by such a preface of generalities as to permissible aims as we find in [the NIRA]? The answer is obvious. Such a delegation of legislative power is unknown to our law and is utterly inconsistent with the constitutional prerogatives and duties of Congress.²³⁶

A court that asks a constitutional question and intones that “The answer is obvious” demonstrates the highest level of confidence in its judgment—the kind that might attend a decision striking down a law whose text began with “This is a Bill of Attainder” or “The free exercise of religion is hereby prohibited.” But Schechter’s nondelegation doctrine has no clear textual home. One could read the doctrine into the Vesting Clause of Article I, which provides that legislative powers are vested in Congress, but that reading is easily contestable. As Adrian Vermeule and Eric Posner have argued, anyone who makes rules on the authority of a federal law authorizing rulemaking can be understood as acting pursuant to an exercise of legislative power by Congress, rather than as exercising legislative powers granted to Congress by the Constitution.²³⁷ In the early Republic, the Constitution was clearly read to permit Congress to authorize private entities to make public policy: look no farther than the Bank of the United States.²³⁸ Why then was it “obvious” that the NIRA’s codes of fair competition were unconstitutional?

The answer lies not in the text of Article I but in political battles of the eighteenth and nineteenth centuries. Madison had formulated, and Jackson had vindicated, the principle that the federal government should not give regulatory responsibilities to private corporations. That principle held firm during an era of another political party’s dominance after the Civil War. Attempts to resurrect the model of government through private business expertise were not just beaten back but repudiated in moralistic terms; the old Jacksonian rhetoric against special privileges was refracted through Republican treatise writers like Thomas Cooley, Christopher Tiedeman, and John Dillon.²³⁹ In the era of Schechter Poultry, that rhetoric united progressive justices like Louis Brandeis with libertarian conservative justices like James McReynolds who rose to prominence as a government lawyer prosecuting antitrust laws against monopolies. Like other broadly held and deeply felt principles about the limits of government, opposition to federal delegations of power to private corporations became entrenched as constitutional law. And one good sign of how broadly and deeply the principle was held was, of course, that a unanimous Supreme Court could pronounce it “obvious,” and say that “[s]uch a delegation of legislative power is unknown to our law,” even in the absence of clear constitutional text or a line of precedential decisions directing that conclusion.

As was true in Jackson’s time, one of the doctrine’s features enabling it to be shared by such a broad coalition was its construction as a principle about federalism rather than about government in general. At the state and local levels, governments then and now enlist any number of private organizations to administer government programs. It might be the Salvation Army or the SPCA. If the corporate nondelegation doctrine were applied to state and local governments, all that infrastructure would be jeopardized. State and local governments would need either to cut back radically on their public services or to build up robust independent bureaucracies. It is not an accident that at the federal level, the formalization of a constitutional nondelegation doctrine coincided with the emergence of the robust administrative state, staffed by full-time public employees who could carry out policies that might otherwise have to be run through private implementers. By restricting the nondelegation doctrine to the federal government, the Court gave force to a broadly held view about federal power while allowing the states to structure the interaction between public and private power in many different ways.

One further point: our discussion of Schechter Poultry as an instance of the corporate nondelegation doctrine at work should not be mistaken for a recommendation that courts make that doctrine into a hard external limit to be enforced with the power of judicial review. If it were, it might be a sickly recommendation given how infrequently courts have struck down federal laws as invalid delegations under Schechter.²⁴⁰ As explained earlier, the suspect-spheres approach is a guide for nonjudicial policymakers; within the judiciary, it is a possible source for soft limits like narrowing constructions. Schechter, indeed, has been the basis for a nondelegation canon of statutory construction.²⁴¹

Whatever the doctrinal merits of Schechter, our purpose here is neither to endorse nor to criticize its holding but rather to add to the stock of examples demonstrating the work that the corporate nondelegation doctrine has done for constitutional decisionmakers at earlier times in American history. Schechter is especially useful for illustrating how a constitutional principle once exclusively enforced by the political branches in party platforms and presidential vetoes can migrate into judicially enforced doctrine, after the course of history has made it seem “obvious.” In the next Part, we suggest that the work it has done is not confined to earlier times.

IV. Traces of the Corporate Nondelegation Doctrine in the Roberts Court

The corporate nondelegation doctrine illustrates how an external limit on Congress’s powers can emerge from history and be put to use as the basis for a constitutional truce that advances some of the benefits of federalism. In this Part, we suggest that attention to the corporate nondelegation doctrine can help explain and even refine some aspects of modern federalism doctrine. As two examples, we discuss the antipreemption canon that helps decide cases like Altria Group, Inc. v. Good²⁴² and the issue of the ACA’s individual mandate in National Federation of Independent Businesses v. Sebelius.²⁴³

A. The Presumption Against Preemption: Altria Group v. Good

The Supreme Court in Altria Group rejected a cigarette manufacturer’s argument that a state-law fraud claim against it was preempted by federal law.²⁴⁴ The plaintiffs in that case argued that Altria’s advertising of certain cigarettes as “light” or “low tar” was deceptive and therefore in violation of Maine’s Unfair Trade Practices Act (MUTPA), which prohibits deceptive advertising.²⁴⁵ Altria—the parent company of Phillip Morris USA—pointed to section 5(b) of the federal Cigarette Labelling Act, which provides that “[n]o requirement or prohibition based on smoking and health shall be imposed under State law with respect to the advertising or promotion of any cigarettes the packages of which are labeled in conformity with the provisions of this chapter.”²⁴⁶ But according to the Supreme Court, section 5(b) did not protect Altria because MUTPA was not a state law “based on smoking and health.”²⁴⁷ Its concern was truth in advertising, which is a different thing. And although Congress clearly intended the federal regime for health-based regulation of cigarette packaging to be exclusive, it did not follow that Congress also intended to prevent states from imposing their generally applicable laws against fraud.²⁴⁸

The Court’s analysis was contestable. In dissent, four justices argued that because the alleged fraud was the advertising of cigarettes as less unhealthy than they actually were, the plaintiffs’ claim was very much “based on smoking and health.”²⁴⁹ If the cigarettes had been manufactured in Mexico but advertised as made in the United States, the dissenters suggested, then the advertising would be fraudulent and an antifraud claim would not be based on health.²⁵⁰ But the essence of the plaintiffs’ actual claim was that they had been induced to smoke in a way that was worse for their health than they believed, and Section 5(b) establishes the federal government as the sole regulator of health-based interests in the area of cigarette packaging.²⁵¹

How should a court adjudicate between these two possible interpretations? The majority in Altria Group answered that question by invoking a presumption against federal preemption. Quoting Rice v. Santa Fe Elevator Corp.,²⁵² a 1947 decision whose significance we will further explore below, the majority framed its analysis as follows:

When addressing questions of express or implied pre-emption, we begin our analysis “with the assumption that the historic police powers of the States [are] not to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress.” That assumption applies with particular force when Congress has legislated in a field traditionally occupied by the States.²⁵³

Within that framework, it is easy to reach the majority’s conclusion. Perhaps the dissent’s view is reasonable, the logic runs. But it is not clear, let alone “clear and manifest,” that anything in Congress’s purposes is served by prohibiting states from imposing “a general rule that creates a duty not to deceive.”²⁵⁴ All things considered, the Rice presumption against preemption might not have put as many fingers on the scale in Altria Group as presumptions against federalization put on the scale in Gregory and Bond. But the Rice presumption does seem to do important work in Altria Group. Certainly the dissenters thought it did: writing for four justices, Justice Thomas made an attack on the Rice presumption central to the dissenting argument.²⁵⁵ In Thomas’s view, no such presumption is warranted. Where the question is express preemption, the dissent insisted, there is no reason to slant the reading of a statute.²⁵⁶ Some scholars have taken a similar view: Caleb Nelson argues that the “notwithstanding” language of Article VI’s Supremacy Clause means that there can be no presumption against federal preemption.²⁵⁷

What if anything justifies the Rice canon? To explore that question, it helps to think about the Court’s further statement that the presumption against preemption is especially strong “when Congress has legislated in a field traditionally occupied by the States.”²⁵⁸ For reasons discussed earlier, it is not at all straightforward to distinguish legislation in fields “traditionally occupied by the States” from other kinds of legislation. After all, both the states and the federal government have long made law about all sorts of things.²⁵⁹

There is, however, a different way of thinking about the “tradition” justifying Altria’s antipreemption canon. At least insofar as firms assert a “frustration of purpose” preemption defense, the antipreemption canon can be understood as a corollary of the corporate nondelegation doctrine.²⁶⁰ Unlike “impossibility preemption,”²⁶¹ frustration-of-purpose preemption defines a “ceiling” of precautions above which the firm is free to go if it so chooses. The usual justification is that some federal statute contains an unwritten purpose to create a nationwide market for goods or services free from state regulations that might be protectionist or parochial.²⁶² In this context, however, “market” is just a neutral-sounding word for the firm itself. In practice, frustration-of-purpose preemption operates as a delegation of power to private firms: those firms that meet a federal minimum standard are empowered to decide whether to go beyond that federal standard, free from interference by state policymakers.²⁶³ Moreover, the federal delegation is not merely a power to decide whether to exceed some federally defined minimum level of care in the present. It also delegates to regulated firms considerable practical power to affect whether the federal standards will be updated over time, because understaffed federal agencies sometimes lack the personnel to collect and review data about adverse effects arising from the use of the firms’ products after those products have been approved and marketed.²⁶⁴

Preemption’s delegation of power to private firms naturally invites the question of oversight. Are federal officials up to the task of monitoring those firms to ensure that federal standards of care are not only obeyed but updated? The answer is not always encouraging. Sometimes the federal agency in charge of overseeing firm compliance is, if not itself a fox guarding the henhouse, nevertheless somewhat fox friendly. Federal agencies are not always run by apolitical experts: their chiefs often serve at the pleasure of presidents who are sympathetic to the appeals of regulated firms. Banks, pharmaceutical companies, automakers, and other firms put a lot of pressure on federal agencies to preempt state laws simply to liberate those firms from state oversight, without necessarily substituting federal standards or enforcement.²⁶⁵ Public-interest activists can fight back, of course: citizen groups can sometimes outmobilize industry groups.²⁶⁶ But if federal agencies preempt state laws, state politicians cannot respond to citizen pressure. And Congress might be too mired in gridlock or agenda overload to fill the oversight gap that preemption has created.²⁶⁷

The antipreemption canon can thus be understood as protection for public interests likely to be slighted by de facto delegations to private firms that are too loosely supervised by overtaxed federal agencies. In the form advocated by Professor Catherine Sharkey, for instance, the canon forces federal agencies to expressly identify the risks regulated by preempted state law and explain why federal law provides an effective substitute.²⁶⁸ On this account, the justifications for protecting a state role in policymaking remain robust even when the states operate in a field where Congress has undoubted lawmaking power, because the purpose of the states is not to regulate the residuum where there is allegedly no federal interest but to remedy dysfunctions of the federal lawmaking process.²⁶⁹ The key question in preemption cases accordingly ought to be this one: has some pathology of federal governance—capture by a private interest group or information overload or gridlock—caused the scheme of federal regulation to ignore an important policy concern that state law has addressed? And no court will arrive at that appropriate question if it tries to reason about preemption from the text of Congress’s enumerated powers. After all, the text of Article I, Section 8 is a product of reasoning about what state lawmaking will not suffice for, not about what problems might attend lawmaking at the national level.²⁷⁰

In short, the text of Congress’s enumerated powers cannot explain the antipreemption canon, but the corporate nondelegation doctrine can. If one understands the states’ regulatory domain as whatever is not within the powers expressly delegated to Congress, the antipreemption canon seems senseless, because federal preemption cases generally arise squarely within Congress’s enumerated commerce power. The questions concern the preemptive force of federal statutes regulating commerce in nationally marketed products like cars, pharmaceuticals, and cigarettes.²⁷¹ So it is perhaps not surprising that the Supreme Court has never provided any coherent explanation for the antipreemption canon beyond a vague respect for an undefined brand of “federalism.” And in the absence of a more coherent explanation, it is perhaps also not surprising that a plurality of the Court has encouraged lower courts not to hesitate to find state laws preempted.²⁷²

But once one abandons the idea that the domains appropriate for state regulation can be identified by reasoning from what Congress’s enumerated powers leave out, one can make headway towards justifying the antipreemption canon in terms of those interests that the federal government is likely to neglect. Indeed, the Court already seems implicitly motivated in its preemption decisions by unspoken factors having nothing to do with Article I—for instance, in its apparent reluctance to preempt state tort law standards.²⁷³In keeping with the analysis suggested here, that reluctance might be rooted in the fact that tort standards tend to protect interests that federal statutes often slight or ignore, including the important interests in compensation and deterrence. Even when federal statutes do provide a minimum standard of care for firms doing business in interstate commerce, they rarely provide for the compensation of persons harmed by the violation of those standards. In contrast, state tort law places compensation and deterrence in the foreground.

The reality that those interests are more likely to be vindicated by state law than by federal law ought to be part of preemption analysis. But if courts try to assess questions about the interaction between state and federal law by reference to the enumeration of congressional powers, those functions of state law will be left out. The most compelling justification for the antipreemption canon will then remain invisible.

B. Mandatory Purchases: NFIB v. Sebelius

In National Federation of Independent Businesses v. Sebelius (NFIB),²⁷⁴ five justices wrote that the Commerce Clause and the Necessary and Proper Clause did not give Congress the power to enact the minimum coverage provisions—better known as the “individual mandate”—of the Affordable Care Act (ACA).²⁷⁵ The argument they presented on this point was all about internal limits. Framed by the major premise that Congress must be limited by its enumerated powers and the minor premise that a reading of the Commerce and Necessary and Proper Clauses broad enough to authorize the individual mandate could authorize any legislation not affirmatively prohibited, these five justices adopted narrower readings of those clauses.²⁷⁶ In particular, they endorsed the view that the commerce power permits Congress to regulate only people who are already active in the relevant commercial markets.²⁷⁷

We think that something about the individual mandate was indeed constitutionally problematic. But the five justices who adopted an activity–inactivity distinction as an internal limit on the commerce power went badly wrong. The constitutional difficulties that the individual mandate authentically raised had nothing to do with the supposed internal limits of the commerce power. They were problems of external limits. As relevant here, the individual mandate fell within the suspect sphere defined by the corporate nondelegation doctrine because it compelled citizens to do business with large private corporations.

That problem would not have justified a court’s invalidating the mandate. As noted earlier, the suspect-spheres approach generates soft external limits rather than hard ones. But awareness that the corporate nondelegation doctrine is part of American constitutional culture can help illuminate the struggle over the ACA’s individual mandate, rendering what is otherwise a bizarre argument more comprehensible. In what follows, we first discuss the weaknesses of NFIB’s internal-limit approach. We then explain how a focus on corporate delegation would have been more illuminating.

1. Under the Lamppost

For ease of exposition, our discussion of the internal-limit objections to the individual mandate will focus on Chief Justice Roberts’s presentation of the argument. Roberts’s analysis began by asking whether the mandate could be an exercise of Congress’s power “to regulate commerce,” as specified in Section 8.²⁷⁸ According to Roberts, the mandate was not a regulation of existing commerce, because the people it coerced were not engaging in relevant commerce at all. Instead, it was a law forcing people to engage in commerce, thus creating commerce to be regulated.²⁷⁹ And in Roberts’s interpretation of Section 8, Congress’s power to regulate commerce cannot include a power to create commerce.²⁸⁰ So the Commerce Clause could not justify the mandate.

Now the problem. Suppose that Roberts was right—though the point is contestable—that the mandate required the creation of commerce where previously there was none. Why, exactly, can a power to regulate commerce not include the power to create it? Roberts’s principal reason was textual. If the power to regulate includes a power to create, Roberts noted, some of the powers enumerated in Section 8 would become superfluous. In particular, the enumerated power to “coin Money” would be superfluous in light of the enumerated power to “regulate the Value” of money, and the powers to “raise and support Armies” and “provide and maintain a Navy” would be superfluous in light of the enumerated power to “Regulat[e] . . . the land and naval Forces.”²⁸¹

That argument founders on a point explained in Part I: any plausible reading of the Constitution’s text makes some of the enumerated powers superfluous. There should be nothing shocking about that point. It is completely normal for long documents drafted by committees to have some superfluous text. Article I, Section 8 clearly does. For example, if Section 8 did not affirmatively specify a congressional power to provide for the punishment of people who counterfeit federal securities, Congress would clearly still have that authority, just as Congress has the power to provide for the punishment of people who evade their taxes or steal the mail, even though no such punishment power is expressly specified.²⁸² In other words, an entire clause of Section 8—the Counterfeiters Clause—is superfluous. So the fact that reading a power “to regulate commerce” as encompassing the power to require commerce would make text in Section 8 superfluous is not a terribly strong reason to preclude that reading.

Roberts next engaged the possibility that even if the Commerce Clause did not authorize Congress to enact the individual mandate, the Necessary and Proper Clause might.²⁸³ The government’s theory, after all, was that pushing more insureds into the pool would make it financially viable for the insurance companies to cover the costs of the additional coverages they were required to provide under the ACA.²⁸⁴ Those new requirements, like covering people with preexisting conditions, were straightforwardly regulations of entities already involved in commerce. So the mandate, which made those requirements viable, could be understood as necessary and proper for carrying the ACA’s valid commerce regulations into execution.

In Roberts’s view, however, the Necessary and Proper Clause did not authorize the mandate.²⁸⁵ Roberts tried to cling tightly to the terms of Article I’s enumeration of congressional powers in two ways. First, he repeated his argument about the semantics of the phrase “regulate commerce,” observing that the mandate to buy insurance would “create the necessary predicate to the exercise of an enumerated power” rather than regulate commerce already in existence.²⁸⁶ But it is not clear why that should matter. Even if the commerce power standing alone does not authorize the regulation of people who are not engaged in the relevant commerce, it does not follow that the Necessary and Proper Clause cannot. The limitation of congressional authority under the Commerce Clause to preexisting commerce rests on the fact that that Clause is about commerce. The Necessary and Proper Clause is not. And surely the Necessary and Proper Clause authorizes Congress to do things in aid of its commerce legislation that Congress could not do on the basis of the Commerce Clause alone. That’s the point of having the Necessary and Proper Clause.

Accordingly, Roberts’s second argument relied on McCulloch v. Maryland’s suggestion that the Necessary and Proper Clause does not license the exercise of “great substantive and independent power[s]” beyond those specifically enumerated.²⁸⁷ McCulloch’s concept of “great” powers that cannot be lightly inferred, however, is just an echo of Madison’s “great and important powers” claim.²⁸⁸ As Madison understood, a prohibition on inferring “great and important” powers from express powers cannot be operationalized without some explanation of why a challenged power is “great and important.” Madison’s explanation, as we noted in Section III.A, rested entirely on considerations external to Article I: Madison denounced the proposed Bank as a monopolistic threat to private equality, drawing on a deep suspicion of governmental aid to private corporations dating back to the South Sea Bubble.²⁸⁹ In contrast, Roberts contended that “great” powers cannot lightly be inferred but did not offer much to explain why the power to require individuals to purchase insurance is “great.” Instead, Roberts simply repeated his claim that Congress may not create “the necessary predicate to the exercise of an enumerated power” by mandating private purchases.²⁹⁰ That this argument adds nothing to Roberts’s objections to grounding the mandate in the commerce power can be seen by quoting the passage in full:

The individual mandate . . . vests Congress with the extraordinary ability to create the necessary predicate to the exercise of an enumerated power.

This is in no way an authority that is ‘narrow in scope,’ Comstock, supra, at 148, or ‘incidental’ to the exercise of the commerce power, McCulloch, supra, at 418. Rather, such a conception of the Necessary and Proper Clause would work a substantial expansion of federal authority. No longer would Congress be limited to regulating under the Commerce Clause those who by some preexisting activity bring themselves within the sphere of federal regulation. Instead, Congress could reach beyond the natural limit of its authority and draw within its regulatory scope those who otherwise would be outside of it.²⁹¹

The power to require uninsured people to buy insurance, in other words, is a too “substantial expansion of federal authority” because it is a power to require uninsured people to buy insurance.²⁹² Nothing in this passage goes beyond Roberts’s argument that the commerce power authorizes only “regulating . . . those who by some preexisting activity bring themselves within the sphere of federal regulation.”

One more observation is in point here. In addition to its other weaknesses, the internal limit on congressional power on which NFIB insisted can contribute little to the maintenance of a healthy federal system. The action–inaction distinction maps nothing about the different pathologies to which national and local decisionmaking are prone. In other words, there is no general reason to think that Congress, but not state and local governments, will act corruptly or abusively in some way that policing the action–inaction line would curtail. Indeed, that line probably cannot prevent Congress from doing anything, given the ease with which “prohibited” regulations of inaction can be redrafted to satisfy constitutional concerns. In NFIB itself, the individual mandate was sustained as an exercise of a different enumerated power: the power to lay and collect taxes.²⁹³ Even the joint dissenters, who maintained that the individual mandate as enacted was not justified as a tax, agreed that Congress could impose an ACA-style mandate to buy insurance through the taxing power if it wanted to.²⁹⁴ More generally, any Congress aware of NFIB’s action–inaction rule could avoid that limit with trivially simple drafting precautions, merely by expressing a mandate as a condition on commercial activity. “No person shall be permitted to buy bread or milk or gasoline, unless that person either carries health insurance or pays a monetary assessment” is a regulation of people already involved in commerce. So even if the Court had invalidated the ACA on the strength of the action–inaction distinction, that internal limit might have been a limit good for attacking national legislation on one occasion and one occasion only.

2. Crossing the Street

The weakness of the internal-limit arguments about action and inaction in NFIB raises this question: why did so many constitutional lawyers, including eventually five justices of the Supreme Court, come to find those arguments persuasive? Part of the explanation surely lies with many people’s passionate opposition to the ACA. Across the political spectrum, thoughtful people acting in good faith often overestimate the merits of legal arguments the results of which they like.²⁹⁵ For many opponents of the ACA, the action–inaction argument offered a ready vehicle for articulating strong opposition to an unwanted law, an opposing-party president, or both. That said, we do not think that the impulse to articulate constitutional opposition to the ACA was entirely attributable to policy distaste for the ACA or partisan antipathy toward President Obama. There were also other reasons why constitutional skepticism toward the individual mandate resonated.

Some of those reasons were rooted in an inchoate but powerful sense that the mandate was at odds with the American constitutional tradition. The justices who ultimately endorsed the Commerce Clause objection to the ACA wrote that the individual mandate was suspect because it was novel.²⁹⁶ Whether the ACA was in fact novel is a contestable proposition.²⁹⁷ But the justices’ sense that it was novel and that its novelty was troubling bespoke their intuition, right or wrong, that the ACA ran against the accumulated wisdom of the past.

We do not think that the accumulated wisdom of constitutional history, or anything else, would have justified a decision to strike down the ACA. We do think, however, that well-socialized constitutional lawyers could reasonably feel that something about the individual mandate was in tension with American constitutional tradition. That tension had nothing to do with a distinction between action and inaction. Instead, we suggest that a significant but unrecognized source of constitutional unease about the individual mandate was the traditional idea that we are calling the corporate nondelegation doctrine.

The individual mandate was not simply a command to do something. Nor even a command to buy something. It was a command to buy something from large private corporations. As we explained in Part III, concern about the dangers involved when the federal government puts its muscle directly behind private corporations has done important work in American constitutional history. To many people, that aspect of the ACA felt wrong, even if it was not easy to articulate the reason in textual or doctrinal terms.

Consider how Judge Jeffrey Sutton wrote about the mandate when a case challenging the ACA was before the Sixth Circuit. At the end of a long and perceptive analysis, Judge Sutton concluded that the mandate and the ACA were constitutional.²⁹⁸ But he also gave memorable expression to the intuition that something about the mandate seemed constitutionally off. In a part of his opinion distinguishing between the ACA’s individual mandate and other affirmative commands that the federal government issues to private citizens, such as the duty to serve as a juror or to register for the draft, Judge Sutton wrote:

To argue that Congress’s power to enlist individuals to defend the country’s borders proves that it may enlist individuals to improve the availability of medical care gives analogy a bad name. There is a difference between drafting a citizen to join the military and forcing him to respond to a price quote from Aetna.²⁹⁹

Judge Sutton did not explore the substance of that difference. He assumed, justifiably, that readers would share his intuition without its having to be explained—and perhaps even without quite being able to articulate its basis. But the basis for that intuition need not be mysterious, so long as one is willing to locate constitutional intuitions in historical attitudes not codified in the written Constitution.

The intuition Judge Sutton articulated is rooted in a traditional distaste for clothing large private corporations with the power of federal law. That distaste goes back to the nineteenth century: note the resemblance between the individual mandate and the “taxless financing” of corporations that Democratic-Republicans and Jacksonian Democrats despised.³⁰⁰ The essence of such taxless finance was that the government conferred special privileges on a private corporation in exchange for that corporation’s providing some sort of public service. In the case of transportation infrastructure, the government provided private corporations with exclusive rights to tolls or public land, and the corporations built bridges, canals, and railroads for the public. In the case of the Bank of the United States, Congress gave a private corporation the lucrative privilege of holding the government’s revenue and acting as its fiscal agent, and the corporation facilitated the provision of credit and currency to the private economy. These schemes had considerable attractions. They allowed the government to produce the policy benefits it sought without having to raise funds through taxation. But they also made federal politics a forum where corporate interests could seek sweetheart deals, thus provoking both fury and constitutional ideology from Andrew Jackson and his followers.³⁰¹

The ACA’s individual mandate was a form of taxless financing. If Congress had the necessary political will, it could eliminate the private health insurance sector and insure all Americans directly, in a single-payer system, using money raised with taxation. Even the joint dissenters in NFIB agreed that such a system would be constitutional.³⁰² But instead, Congress arranged for private corporations to provide the needed insurance. Those corporations had to get something in return, something bestowed by federal law: increased revenue from citizens now forced to carry private insurance. And in states where only a small number of insurance companies sold policies on the ACA-defined insurance exchange, those corporations gained both a guaranteed revenue stream and some degree of monopoly power.³⁰³ The historical analogies are not hard to spot.

In this light, consider the significance of Congress’s choice not to include a “public option” in the ACA. During its deliberations, Congress considered having the federal government itself act as one of the insurers in the marketplace. Citizens could then choose to buy insurance from a private corporation or, if they preferred, to exercise the “public option” and buy directly from the government.³⁰⁴ Like the Independent Treasury System championed by hard-money Jacksonian Democrats,³⁰⁵ the public option would have weakened the delegation of power to private corporations, thus allaying suspicions that private corporations would exploit their special privilege for undeserved financial gain. But proposals to create a public option were blocked by a filibuster threat from Senator Joseph Lieberman, who coincidentally or not represented the state in which many major insurance companies were headquartered at the time.³⁰⁶ One might plausibly analogize Lieberman’s role to the pro-banking “soft money” Democrats who prevailed on Jackson to adopt a “pet bank” system in which certain favored state-chartered banks would receive the right to hold federal revenue.³⁰⁷ And with the public option gone, the ACA presented itself to previously uninsured citizens in the form of large private corporations holding out their hands for payment, on the authority of the federal government.

For people who would like to understand why so many Americans were primed to conclude that the individual mandate was unconstitutional, the tradition of corporate nondelegation is significantly more illuminating than the action–inaction distinction on offer in NFIB. Unlike the idea that the commerce power has a built-in distinction between action and inaction, which was newly minted by (very smart) anti-ACA lawyers,³⁰⁸ the corporate nondelegation idea is authentically a long-standing feature of American constitutionalism. Rather than resting on a dodgy inference about superfluous text and some forced readings of case law, it goes to the practical substance of what might be problematic about the individual mandate. What’s more, the fact that the corporate nondelegation doctrine is an idea about federalism also helps explain one of the puzzles about attitudes toward the ACA outside the courts: how so many people who believed the ACA’s individual mandate abominable could have a different intuition about the state-level individual-mandate system created by Massachusetts.³⁰⁹ One could say, as critics often did, that the difference is simply that states have general police power and the federal government doesn’t.³¹⁰ But we suspect that the vehemence with which some people who accepted the Massachusetts mandate reviled the federal one calls for some additional explanation. To be sure, we do not say that the corporate nondelegation doctrine completely explains the phenomenon. But that doctrine has always been based on fears about the kinds of corruption to which Congress is distinctly vulnerable.

The corporate nondelegation doctrine is also better able than the action–inaction distinction to deliver the benefits of federalism, including the fostering of a constitutional truce between a party in power and a party out of power. The action–inaction distinction marshaled in NFIB seemed custom-tailored to eliminate one law opposed by one party. It offered something important to the people on one side of the partisan divide and nothing to the other. Indeed, given the ease with which that formal distinction can be overcome by a legislature that knows it needs to comply with that rule, the action–inaction distinction did not offer either party any plausible assurance that the same rule would protect their interests in the future. (An alternative-universe Republican Congress that wanted to make all Americans purchase firearms could enact a law prohibiting the purchase of goods in interstate commerce, except for people who own firearms; an alternative-universe Democratic Congress could use the same strategy to make all Americans purchase KN95 face masks during a pandemic.) In contrast, the idea that federal delegations to private corporations are suspect has resonance for people across party lines. In the past, it appealed to Jacksonian Democrats and McKinley Republicans, to Brandeis and to McReynolds. In the twenty-first century, it offers something to the proto–Tea Party Republicans who protested bank bailouts in 2008 and something to the left-leaning Democrats who call for an antioligarchical constitution.³¹¹

Again, our point is not that the Court should have held the individual mandate unconstitutional on the strength of the corporate nondelegation doctrine. The mandate’s corporate delegation aspect might make it suspect, but as we have explained, the “suspect spheres” approach does not purport to create hard external limits on congressional power, to be enforced by courts in the exercise of judicial review. In nonjudicial policymaking, it offers a resource to decisionmakers trying to craft law that makes American federalism the best that it can be. Congressional attention to the corporate delegation problem while the ACA was pending would have counseled support for the public option, or for a single-payer system, or for comprehensive rate regulation preventing insurers’ decisions from becoming a law unto the consumer in jurisdictions where competition was thin, or for something else that avoided creating a federal law compelling ordinary citizens to supply revenue for large private insurers—and exacerbating political opposition to what was already a bitterly contested proposal. In the courts, the corporate nondelegation doctrine should not support hard external limits, but it can inspire softer ones, like clear-statement rules or caution in finding state laws preempted. But there was no valid room for such measures on the question of the ACA’s individual mandate. Congress deliberated extensively and wrote a law that clearly established the minimum coverage requirement. Under those conditions, and in the absence of hard constitutional limits from some other source, courts should get out of the way. So we agree again with Judge Sutton, who, despite his implicit criticism of laws requiring people to respond to price quotes from Aetna, ultimately concluded that the powers of Congress were sufficient for enacting the ACA.

The adoption of the action–inaction idea by five justices in NFIB is thus doubly unfortunate. First, by rewarding the development of an internal-limit strategy ostensibly based on the text of Section 8, it gave credence to the notion that the enumerated powers of Congress are an important feature of American federalism. That notion is misleading. We would be better off letting it go. Second, by vindicating the idea that whatever constitutional unease people felt about the individual mandate was attributable to its falling on the wrong side of a metaphysical divide between activity and inactivity, it deflected attention from the possibility that something more substantive was at work. If we paid more attention to that substantive thing, we might make better decisions in the future—both in the specific context of health insurance policy and in the general project of federalism.

Conclusion

Some bad habits die hard. The practice of looking to the Constitution’s enumeration of congressional powers as if that enumeration said something important about American federalism is deeply rooted, and many constitutional lawyers would find it difficult to give it up. It is not hard to see why. Textual arguments can give both the decisionmaker and the audience the sense of “real law” in the form of a positive enactment that can be traced back to an authoritative moment of lawmaking. Where the text in question is Article I, the appeal is all the greater, both because the relevant lawmaking moment has the exalted status of the constitutional founding and because American constitutional culture has long treated the idea of a limiting enumeration as a sacred creed. But under careful examination, the idea of the limiting enumeration turns out to be largely illusory. In reality, American federalism is and has long been mostly a federalism of process and a federalism of nontextual external limits.

To be sure, constitutional decisionmakers determined to make the textual enumeration of congressional powers into a device imposing meaningful limits on Congress could find ways of doing so. Excellent constitutional lawyers are often wonderfully creative readers of the Constitution’s text. But the limiting interpretations of Congress’s powers that courts might read into that text are unlikely to recover any principles of federalism authentically placed there by the Framers or to instantiate principles of federalism that would be sensible in light of the actual workings of modern American government.³¹²

Instead, interpreters who purport to discover sensible rules of federalism by making negative inferences from the enumeration of congressional powers are more likely, in perfectly good faith, to present something that is not reasonably identified as a principle of constitutional federalism (for example, a distinction between activity and inactivity within the regulation of commerce) as if it actually were. To put the point a little polemically, that is a bit like picking up a pebble from under the lamppost, handing it to the drunk man, and saying “Here is your wallet.” It purports to get the job done by finding the thing sought in the place where people were looking. But the thing found will not do the work that it is supposed to do.

Rather than making misdirected arguments about the internal limits of textually enumerated powers, we recommend that constitutional decisionmakers ask a better set of questions. What are the distinctive pathologies of congressional lawmaking relative to local lawmaking? Is it possible to answer that question in ways that support principled limits that would reduce, rather than exacerbate, the tension between the party in power and the party out of power? If so, can public decisionmakers act on those principles in ways appropriate to their institutional roles?

Our example of the corporate nondelegation doctrine illustrates how such a principle might function. We leave it to readers to judge whether the example helps make the suspect-spheres approach plausible. If it does, scholars of federalism should try to identify other such principles and to work out the institutional arrangements for enforcing them. And if it does not, the implication cannot be a return to the seductive comfort of enumerated powers. It must instead be a commitment to think of better ways of identifying external limits—or else to abandon entirely the idea of substantive federalism-based constitutional limits on national lawmaking. However hard it is to search the other side of the street, nothing useful will be found under the lamppost.

* Theodore J. St. Antoine Collegiate Professor of Law, The University of Michigan Law School.

** William T. Comfort, III Professor of Law, New York University School of Law. The authors thank Daryl Levinson, Julian Mortenson, David Schwartz, Anna Searle, Audrey Springer-Wilson, Adam Wallstein, Tyler Washington, Kenji Yoshino, and the participants in workshops at New York University School of Law, the University of Michigan Law School, and the University of Pennsylvania Carey Law School.