There is a basic, underlying logic to what Alexis de Tocqueville termed the “complex constitution” of the United States. The lower case “c” is important; Tocqueville used “constitution” in its broader signification, referring to state governments as well as the national government. In particular, he discussed how the national government was constituted to fit with the state governments, and vice versa.
While Tocqueville labeled the constitution of the United States “complex,” it’s actually simple to summarize. Publius put it this way in Federalist #14:
The [national] government is not to be charged with the whole power of making and administering laws. Its jurisdiction is limited to certain enumerated objects, which concern all the members of the republic, but which are not to be attained by the separate provisions of any. The subordinate governments, which can extend their care to all those other subjects which can be separately provided for, will retain their due authority and activity.
To be sure, while the broad outline is simple to trace, filling in the details requires a bit more attention. The trick that judges, legislators, and citizens have been wrestling with for over 200 years is drawing the boundary between policy areas that “can be separately provided for” by the states, and the “limited” set of objects that cannot be provided for by the separate actions of the states and so were delegated to the national government.
One recent approach seeking to explain the logic regarding which policy domains the federal Constitution leaves to the states and which are delegated to the national government is termed the “collective action Constitution.” This explanation locates the logic of the Constitution’s delegated powers in the Founders’ responses to interstate incentive structures that prevent states separately from being able to provide for the common good. Unfortunately, some of the scholars extolling the theory don’t seem to place much actual trust in their theory, and habitually deflect in favor of suboptimally expansive national authority when the theory would point to optimal resolution at the state level.
Justifying Delegations to the National Government
Publius argues in Federalist #14, quoted above, that power is retained by state governments on all matters which “can” be provided for separately by the states; power is delegated to the national government on matters which states cannot provide for separately.
This is the thread tying together the arguments for the grant of national power throughout The Federalist. Diffuse state action is accepted as the baseline; each grant of power to the national government is justified by a reason to depart from diffuse state policies. Publius reflects this mode of argument in Federalist #41 when he frames his discussion with the question, “Whether any part of the powers transferred to the general government be unnecessary or improper?”
Again and again the writers of The Federalist appealed to the basic logic of collective-action design. Madison, for example, observes in Federalist #10 that “the federal Constitution forms a happy combination in this respect; the great and aggregate interests being referred to the national, the local and particular to the State legislatures.” So, too, in Federalist #3, John Jay underscores the need for a national government “vested with sufficient powers for all general and national purposes.” He specifies an application in the next paper, discussing the need to nationalize defense policy given that separate state provisions of defense would be in “want of concert and unit of system.”
Later in The Federalist, Hamilton laments the “competitions of commerce” between the states and motivates the contracts clause by asserting an interstate justification. Madison similarly justifies national authority over patents and copyright because states recognized that they “cannot separately make effectual provision for either.”
Neil Siegel simply does not seem to believe his own analysis regarding the collective action Constitution.
At the Constitutional Convention, the assembly adopted Resolution VI of the Virginia Plan. The Resolution sketched the aspiration that Congress under the new Constitution would be empowered “to legislate in all cases to which the separate States are incompetent, or in which the harmony of the United States may be interrupted by the exercise of individual Legislation.”
Despite repeated appeals to the single principle, however, the question remained just where the boundary would be located that sorted a set of appropriate matters to the national government while reserving other matters to the states.
Chief Justice Marshall, writing for the Court in Gibbons v. Ogden in 1824, moved the needle on the question of what remained subject to state regulation (writing specifically about the Commerce Clause):
The genius and character of the whole government seem to be that its action is to be applied to all the external concerns of the nation, and to those internal concerns which affect the States generally, but not to those which are completely within a particular State, which do not affect other States, and with which it is not necessary to interfere for the purpose of executing some of the general powers of the government.
The last clause in the quotation, which is critical to Marshall’s argument, is typically neglected in commentary on the case. Significantly, Marshall identifies that the Constitution delegates authority to the national government based not only on the criterion that the matter affects states generally—that is, that the matter is not limited to a particular state but affects other states—but also provided that it is “necessary” for the national government to legislate in the stead of states.
The criterion of “necessity” need not be taken in a strict sense (no need to replicate Maryland’s argument in McCulloch v. Maryland). Nonetheless, Marshall’s criterion of “necessity” stands on its own as a separate criterion that needs to be met to justify national action; this is particularly true in discussing the collective action Constitution.
Note the upshot of Marshall’s analysis: There are some matters that do affect the states generally, yet continue to be reserved for state action because it is not necessary to assign the policy domain to the national government to achieve an optimal outcome.
To put the point in the language of the collective action Constitution: that separate state policies can impose externalities for other states is a necessary, but not a sufficient condition for nationalizing a policy domain.
State policies of all sorts frequently affect other states as a matter of course. But the existence of states’ policies that affect other states does not mean that the interaction of these state policies reflects pathological incentive structures that induce states to adopt suboptimal policies.
A concrete example might help. Consider the law at issue in United States v. Lopez, the 1995 case in which, for the first time since the Great Depression, the Supreme Court struck down a national law as beyond Congress’s Commerce Clause authority. Congress had adopted the Gun-Free School Zones Act of 1990, effectively nationalizing a policy domain that had previously been reserved to the states.
Consider the impact of diverse state laws regulating guns in school zones, and ask Marshall’s first question: Do the laws that states adopt regulating guns near school zones affect other states? The answer is obviously yes, with pertinent interstate effects being discussed at length in the case. Hence, if the question of whether states were “incompetent” to regulate guns in school zones depended solely on whether state choices “affected” other states, then the answer would be that the policy domain should be nationalized. (Recall that the majority’s response to these effects was to adopt an infelicitously formalistic economic/non-economic distinction.)
Marshall’s analysis does not end here; it considers a second criterion: Is it necessary for the national government to regulate guns near school zones? The collective-action Constitution version of Marshall’s question is this: Do states face pathological interstate incentive structures that systematically induce states to underregulate guns near school zones?
That is, granting that state choices in the policy domain will naturally affect other states, does, say, Texas’s choice of regulating guns in school zones change the incentives other states face to regulate the matter as they see fit? And there the natural answer is just as obviously “no.” States may differ regarding the level of regulation they choose, but those choices do not create pathological interstate incentives that induce or pressure other states to change the level of regulation they choose.
In this area, different state policies simply reflect different state policy preferences and do not interact in such a way as to undermine the policies that other states adopt.
To illustrate the difference between interstate incentive structures in which state policies merely affect other states and interstate incentive structures that undermine other state policies, consider the game illustrated below. It shows a simplified and highly stylized version of a prisoner’s dilemma game between states.
Prisoner’s dilemmas exist in numerous policy areas, but for concreteness, let’s consider the policy domain of child labor. Each state has the choice of either banning child labor or permitting child labor. The voters of both states prefer to ban child labor rather than to permit it. The problem, however, is that if only one state bans child labor while the other state permits it, mobile capital will move to the non-banning state, causing significant economic contraction in the banning state and economic growth in the non-banning state.
As a result of the pathological incentive structure that exists between the states, and despite unanimous support for banning child labor, without the ability of coordinate policy jointly, the states nonetheless both permit child labor in equilibrium; both states choose to implement the outcome that they regard as suboptimal. (Note that both states prefer the outcome in which both states ban child labor.) As a result of the pathological interstate incentive structure between the states, separated state policy choices result in states being “incompetent” to regulate child labor.
Not only do the actions of one state affect the other state, in order to regulate child labor, it is “necessary” for the national government to implement the regulation. Separate state actions cannot realize the outcome. (Or, at least, the nature of the interstate game creates pressures for states that deter them from banning child labor, though they unanimously regard a universal ban as optimal.)
The Cost of Promiscuously Nationalizing Policy Domains
It is important to underscore that, just as it is costly not to nationalize a policy domain in which states face a pathological incentive structure preventing them from separately adopting optimal policies, so too it is costly to nationalize policy domains in which states do not face pathological incentive structures. That is, the general welfare of the nation is reduced by nationalizing policy domains that should not be nationalized.
Hamilton noted the optimality of diffused state action (in the absence of pathological interstate incentives) in Federalist #32. He wrote of “the utility and necessity of local administrations for local purposes.” That is, he recognized the positive value of reserving appropriately state matters to state governments. Nationalizing these policy matters would represent a net loss of “utility” (in Hamilton’s phrase).
The general welfare of the nation is reduced by nationalizing policy domains that should not be nationalized.
Just as there is a loss to the general welfare when the nation fails to nationalize policy matters where states cannot provide for the general welfare, there is also a loss when the nation nationalizes policy matters in which separate state action does provide for the general welfare.
The illustration below depicts a simple game in which uncoordinated state activity maximizes national welfare. In the game, voters in two states have different preferences over which policy they prefer. Each state adopts the different policy its voters prefer. (State 1 implements Policy A, State 2 implements Policy B, the payoff across the two states is (2,2).)
If the policy decision these states make were to be nationalized, only one of the two policy options would be adopted and implemented uniformly across both states. Irrespective of which policy was adopted, the voters of one state would be left worse off than they were with uncoordinated policy implementation.
When we talk about policy diffusion and federalism, scholars emphasize the possibility of informational advantages of policy implementation at lower levels rather than at higher levels. Those may exist, but presumably less in the informationally-rich policy environment of the modern era relative to earlier periods. Often overlooked in the focus on informational advantages, however, is that voters in different states may simply have different preferences over policy content. These preferences may reflect not just preferences over substance, but different preferences between state populations over risk, different preferences over the relative weighting of law and liberty, different preferences for spending on public goods relative to private goods, and more.
These differences in preferences make hash out of the simple aggregation-of-preference rationale that, for example, Neil Siegel adds to his otherwise careful discussion of how the Constitution was designed to solve interstate collective action problems. He consequently would have the national government improperly usurp state policies when those policies do not reflect collective action problems.
State policies that he would nationalize that are not justified by the actual collective action part of his analysis are Covid-19 policies (and future pandemic policies), infrastructure investment, responses to opioid addiction, and gun violence (despite providing a tepid defense of the Supreme Court’s Lopez decision as consistent with his theory).
Simply consider pandemic policies. As I’ve pointed out before, states can take draconian action at their borders to prevent disease from crossing in from neighboring states. The national government can assist with policing borders if states have difficulty doing so. Nonetheless, it is far from obvious that differing state preferences over liberty and lockups, over continuing school and economic activity versus the ordering of the confinement of massive proportions of state populations, and more, were optimal or necessary.
In this policy domain and others, Siegel simply does not seem to believe his own analysis regarding the collective action Constitution. The addition of aggregating preferences across states makes his version of the collective action Constitution all but indistinguishable from a national government that faces few limits to usurping state policies.
The irony is that these examples are at variance with much of Siegel’s exposition of the basic collective action Constitution. That is, there is an inherent logic to the US small-c constitution, to how national power and state power fit together and complement each other. The nation’s welfare is promoted when we recognize the Constitution’s collective action logic, and when judges enforce it.