Federalism as a System of Administration

Author: Michael C. Dorf

Title of Author: Professor of Law, Columbia University

Lecture at Peking University Law School

Ni hao ma and greetings from the United States. Let me begin by saying how honored I am to be here today and to have this opportunity to share with you some of my thoughts about federalism in the United States as well as elsewhere, and to invite you through your questions to help me learn more about your system.

My topic today is “Federalism as a System of Administration.” Federalism, as you know, is a word used to describe a system of government in which there is both a central government with special responsibilities and state or provincial governments, that have their own distinct responsibilities, and which have, some though not necessarily all, elements of sovereignty. Federal systems today are diverse. The United States is the oldest continuously functioning federal system of government, but Australia, Canada, Germany, India, Mexico, and Switzerland, to name just a few, are countries that also have robust systems of federalism. Perhaps the most interesting experiment in federalism in recent years, has been the European Union. Unlike the other federal systems, however, the European Union is not a sovereign state. Rather, it is made up of fully sovereign states, some of which are themselves federal systems.

What do all of these diverse federal systems have in common? In each one, there is, as I said, a central government and then state or provincial governments, or as they’re called in Germany, “l?nder.” In addition, most constitutional systems that employ federalism attempt to separate competencies between the federal governments and the state governments. However, there are also important differences. Let me give an example comparing the United States and Canada, two countries that have similar populations and similar histories, but somewhat different federal regimes.

In the United Sates, the power to regulate commerce, in particular interstate commerce, is a competency of the national government, while the power to address regulate most forms of violent crime such as, murder and theft, is a competence of the state governments. By contrast, in Canada, the power to prohibit and punish crimes is a competency of the national government, whereas the power to regulate most commercial matters is a state competency. This difference has important consequences when it comes to judicial efforts to enforce federalism. In the United States, as the power of the national government has grown, the courts have tended to read the delegation of authority to the national government to regulate interstate commerce very broadly. That has enabled the federal government, with its power to regulate interstate commerce, to regulate all sorts of activity that in earlier eras would have been thought to be within the unique competence of the states. In Canada, there has been similar pressure to expand the domain of federal law, but there, the pressure has fallen within the power to proscribe crimes, so that the Canadian legislature at the national level has used its power to regulate crime as the source, or you might say, the pretext, for regulating what were previously thought to be matters within the provincial competencies.

The common thread in both the United States and Canada is that in each country the courts and the national legislature have viewed the written division of competencies between the national and regional or state governments pragmatically. They have not paid undue attention to the exact wording of the constitutional limits, instead allowing the assertion of broad power by the respective national governments when it seemed appropriate for accomplishing important government objectives.

Let me give another example of the variety of arrangements consistent with the idea of federalism. In Germany, the grundgesetz or constitution sets out two mechanisms for the enforcement and administration of federal law. First, Article 83 states that in general, the l?nder are responsible for executing federal law. In doing so, the l?nder act as agents of the federal government. There is also a second mechanism for the enforcement of federal law in Germany. With respect to some particular powers—such as the foreign service and national defense—the federal government directly executes federal law. Although the powers carried out directly by the federal government are important, the general rule is administration by the l?nder.

Contrast the German approach with that of the United States. The American Constitution does not expressly say whether laws enacted by Congress, our national legislature, are to be implemented by the federal government or by the state governments. One might think, therefore, that the choice between direct federal implementation and implementation by the states is a matter committed to the discretion of Congress. That is what Congress thought when it passed the Brady Act in 1993. The Act mandated criminal background checks for purchasers of handguns. Although the background checks would eventually be performed through a national computer database, initially the system required that state officials perform the background checks. Officials from the states of Arizona and Montana objected to the requirement, and in a 5-4 decision in 1997, the United States Supreme Court ruled that the Brady Act was unconstitutional because the federal Congress lacked the authority to require state officials to implement federal law.

Thus, in Germany the federal government can enlist the l?nder in enforcement of federal law, but in the United States, the federal government cannot. Nonetheless, there is a mechanism by which the American Congress can obtain state cooperation in the enforcement of federal law. It is quite common in the United States for the national Congress to pass a law and then to give money to the states to carry out that law. Examples include our system of unemployment insurance, our system of welfare and job training, and such regulatory matters as providing clean air and clean water. In each of these contexts and others, the federal government has passed legislation and made money available to the states to carry out that legislation, provided the states take the money.

（In Germany, except for the constitutional requirement for the l?nders to execute federal laws, are there any incentives such as money for the l?nders to do so?）

In principle, the American states can decline the federal money, but in practice they have very strong incentives to accept it. Federal money comes, in the first instance, from taxation of activity throughout the country. For example, as a citizen of the state of New York, I pay taxes both to New York State and to the national government. If New York State decides not to participate in the federal system of unemployment insurance then if I become unemployed, New York State foregoes the money from the federal government attributable to the taxes I paid, and so there is a powerful incentive for the states to take the money from the federal government to implement national policy. However, because the states retain a formal right to refuse the federal funds, the courts allow Congress to pay the states to implement federal law, even though they do not allow Congress simply to order the states to implement federal law.

In practice, therefore, the American system is not all that different from the German system. Using its power to spend money, Congress can usually obtain state enforcement of federal law. Still, there will be circumstances in which states will decline federal funds and even if they do not, requiring Congress to attach implementation requirements to federal grants limits the flexibility of Congress in setting budget priorities. Why, then, did the U.S. Supreme Court read the Constitution as prohibiting Congress from issuing a simple directive to the states?

Part of the answer is historical. The five Justices of the Supreme Court who voted to prohibit what they called the “commandeering” of the state executive branch thought that when the Constitution was adopted in 1789, it was generally understood that the federal government would act only directly on citizens, not indirectly through states. That reading of history is controversial, however, and furthermore, it is not entirely clear that the ambiguous language of the Constitution should be read to reflect eighteenth century views, when the world has changed considerably since then.

The functional, modern reason why the majority in the Brady Act case prohibited the commandeering of the state executive is that these Justices were trying to preserve elements of state sovereignty. Therefore, we might ask, which system better preserves the prerogatives of the state governments—the German system or the American system? On the one hand, the American system is more respectful of the autonomy of the states because it says the states cannot be ordered to do something they do not want to do. On the other hand, the German system has the functional effect of giving the states a greater role, because in the German system, when there is a new national program, there is not necessarily a new national bureaucracy that draws power away from the states. Instead, each new national program adopted under one of the provisions of the German constitution requiring l?nder administration actually increases the power of the l?nder, and for that reason, four of the nine judges in the Brady Act case dissented. They said that the majority of the court thought it was helping the states by giving them a new power to resist orders from the national government, but in fact, this rule would harm the states because it would give the national government no choice but to create a national bureaucracy to administer each new program.

You may at this point be wondering why one would care about preserving autonomy for the states. Why is it important to respect the state of New York, the province of Ontario, or the land of Bavaria? These are after all artificial entities, not individuals with feelings. I want to suggest that there are two kinds of reasons for caring about state autonomy.

One kind of reason is historical. In many federal systems, we can trace the origin of the national government to an earlier time when the constituent states were independent sovereigns. Germany, prior to unification in the nineteenth century, is an example. Some people say that the United States is also an example, although this is a controversial proposition, because prior to the American Revolution of 1776, all of what later became the states were colonies of Great Britain. In order to say that the states became sovereign, one has to assert that between the Declaration of Independence of 1776 and the adoption of the United States Constitution in 1789, the States exercised sovereignty. But this a matter of some delicacy because the States that declared independence from Great Britain in 1776 did so collectively assembled in Congress, and so there is a reasonably strong argument that New York, Pennsylvania, Virginia and the other colonies never possessed full sovereignty.

The historical question is interesting for constitutional scholars in the United States, and also a question that has surprisingly important consequences for the outcome of concrete cases. For example, in 1995 our Supreme Court ruled that a state could not place limits on the number of terms a member of Congress elected from that state could serve in the national government. In this case, which was another 5 to 4 split, the majority said that that the Constitution was a creation of the people of the United States as a whole rather than a creation of the people of each individual state acting as States. Therefore, a state could not limit the terms of office of a member of the national legislature.

Although the term limits case appeared to reject the idea that the American states were fully sovereign entities before the Constitution was adopted, other recent cases come closer to accepting that view, and the issue will continue to be debated in the courts and the academic journals. In the meantime, perhaps the clearest example of this first kind of reason for caring about state autonomy that I can give you comes not from the United States, but from contemporary Europe.

Whatever the status of the American States in the late eighteenth century, nobody doubts today that Italy, Spain, and the other members of the European Union are all sovereign states. When people in contemporary Europe worry about the power of the European Union they are worried that they will lose their national identity. Citizens of France want to remain French even as they want to partake of the common economic, social, and political benefits that arise out of being part of a unified Europe. Most of the debate in Europe about the future of the European Union has been a debate about the nature of federalism in Europe. And in that debate, it is clear that what people worry about is the loss of sovereignty. Here I want to offer an historical perspective.

Today it would be ridiculous for somebody in the United States who lived in, let’s say, the state of Massachusetts to complain that some program of the national government is harming his identity as a citizen of Massachusetts. And that is because Massachusetts has been part of the United States for over two hundred years. By contrast, it would hardly be ridiculous for a citizen of Denmark or Ireland to worry about the loss of national identity that will come from power-moving from Denmark and Ireland to the European Union. The question for people who care about national identity in Europe is whether it is possible to write legal documents, whether they are treaties or a constitution for Europe, that will be able to preserve the prerogatives of state sovereignty even as they accelerate economic forces that are likely to blur cultural distinctions. My hypothesis is that if the European Union succeeds in creating a continent-wide economic and political system, national identity is likely to diminish. And at that point, perhaps one or two or more generations away, the people of the new Europe will have to decide whether they want to retain legal structures that maintain national distinctions if those distinctions are no longer felt by the people.

I said a few moments ago that there are two sorts of reasons for caring about state autonomy and I have just now been discussing the first: Respect for the sovereign entities that joined the federal system. In circumstances in which the member states continue to maintain a strong sense of sovereignty—as in the European Union today—such respect will be a matter of accepting popular opinion. The people do not want to give up their national identities. After the passage of time, once the people no longer feel strong attachments to the member states, respect for the sovereign entities that joined the federal system becomes a matter of honoring a past bargain. That is essentially the situation of the United States today. We honor state autonomy out of respect for the entities that the states once were. It is a little bit like keeping a promise to a dead ancestor.

Accordingly, within a country like the United States today, respect for state sovereignty is non-instrumental. Respect for state sovereignty need not be defended in terms of the contemporary goals it serves.

Some of the Justices of the U.S. Supreme Court think that non-instrumental claims for state sovereignty are a sufficient reason to limit the powers of the national government in favor of the states. Other judges and scholars—including myself—are less comfortable with this justification for modern doctrine. We respect the wishes of the dead either because we believe that is the way to honor their spirits or because we believe it will give comfort and security to the living. Neither justification applies to states. I am not a theologian, but know of no claim that states have immortal souls. Nor does it make much sense to say that state autonomy must be preserved today to give comfort to other states contemplating joining the federal union. If there are such states, they can enter or not on whatever terms are now established.

For me, therefore, the case for protecting state autonomy through federalism must be an instrumental one. It must be that how federalism advances worthwhile contemporary projects. Not surprisingly, therefore, I would associate the instrumental virtues of federalism with the advantages that arise out of decentralized administration. Instrumental federalism, however, is not identical to decentralization. For example, France is not a federal system but nonetheless includes some elements of decentralization in its administration of laws and policies. The French national government frequently assigns powers and tasks to local government. By contrast, in a system of instrumental federalism, there are actually formal legal barriers to the central government asserting some powers.

You can think of instrumental federalism as a pre-commitment strategy. Citizens designing a system of federalism where there previously may or may not have not been independent sovereign states deliberately carve up some aspects of national sovereignty and give them irrevocably to the state and local units because they worry that without that kind of firm decentralization, people in the central government with good intentions will take too much power for themselves and do a poor job of running the government.

I should say that I myself do not have a strong view about whether instrumental federalism or decentralization of the sort we see in France is generally a better system of government, but I do want to suggest that in a country like the United States today the arguments for strong protection of state autonomy should be arguments for instrumental federalism rather than non-instrumentalism, because any claim to real sovereignty in the American states is no longer plausible at this point in our history.

I shall shortly return to the instrumental role played by federalism in the United States, but first, let me ask: What about China? Here, I must begin with humility. I know very little about Chinese administration, and there is a saying in English that a little knowledge is a dangerous thing. Aware of that danger, I want to offer a few thoughts about how a country like China, that is to say, a very large country with both a large urban population and a large rural population, might think about federalism.

Let me put aside any question of non-instrumental federalism. From what I know of Chinese history over the course of the last several thousand years, various rulers have governed smaller and larger portions of China, but today no one would claim, for example, that Guangdong or Sichuan is a sovereign entity. These provinces, whatever their history over the course of thousands of years are now considered irrevocably part of China. I will assume that either the Chinese provinces never were fully sovereign entities or if they were, they long ago ceded that sovereignty to the national government in Beijing.

If I were asked for advice about applying federalism in a country with the characteristics of China, what would I say? Champions of federalism draw up long lists of the virtues of federalism compared to completely centralized governments. I think I can boil down all of these virtues to one. The central virtue of instrumental federalism is humility. In a large country like China or the United States it is impossible for any one person or any one organization, no matter how benevolent, competent, and wise, to know everything that needs to be known to address practical problems in all of their complexity.

Let me give an example from the United States. Take the question of air and water pollution. A number of years ago, petroleum refiners in the United States began adding certain chemicals to gasoline to make the gasoline burn more cleanly, producing fewer toxins in the air. It turned out, however, that the chemical that made the gasoline burn more cleanly and thus improved the quality of the air, also sometimes leaked into the water where it became a toxin itself. As a result, there was a movement to ban this particular additive from gasoline. Although I studied physics as an undergraduate, I am not an expert in the biology and chemistry at issue, and so I cannot tell you whether the aggregate benefits from cleaner air outweigh the aggregate costs from dirtier water that arise from this additive, which is known as MBTE, the letters designating the chemical. I can say this however: It is not obvious that the right way to address the health effects of MBTE is through a national rule either categorically prohibiting its use in gasoline or categorically permitting it. Some parts of the country have very good air quality, but vulnerable water. Other parts of the country have very good water quality but vulnerable air, and some parts of the country have either good air and good water or vulnerable air and vulnerable water.

A lawmaker infused with the humble spirit of instrumental federalism might well think that the best balance between clean air and clean water will vary from place to place. And once the lawmaker reaches that conclusion, it is not obvious that the central authority ought to make the decision about how to strike the balance rather than permitting that decision to be made at the state or local level. There is a danger of course that one can go too far with the spirit of instrumental federalism, because every problem that confronts mankind has local dimensions. And that is why the version of instrumental federalism that I favor does not simply decentralize all power but asks in each case whether there is a useful role for the central government in setting broad policy.

For example, in the case of clean air and clean water, the federal government might say: “Here is the minimum standard for air and here is the minimum standard for clean water. We leave it to each state to determine for itself whether it is necessary to prohibit the use of particular gasoline additives in order to achieve the minimum standard for clean water.”

In addition to setting minimum standards, the central government can play an extremely useful role in facilitating action by state governments by collecting information about successful and unsuccessful state practices. In other words, the federal government can act as a clearinghouse that gives technical support to the states. And in fact, in the United States, in just those areas where state cooperation in federal programs has been achieved by grants of money from the federal government to the states, we see the federal government providing such technical assistance. For example, the federal Environmental Protection Agency assists state Departments of Environmental Protection in their efforts to monitor air and water quality.

Of course not every question is best addressed through this combination of the national government setting minimum standards and state and local government carrying them out. There are some instances in which a uniform national policy is essential. Take, for example, the question of labeling. In the United States, there are a variety of laws requiring that certain kinds of products be labeled. For example, cigarettes must bear a warning telling smokers the health risks associated with tobacco. Those warnings are required by federal law, but the same federal law that requires warnings also forbids the states from requiring additional warnings. The reason for that prohibition is to protect the integrity of the national market. One of the benefits of living in a large country is the ability to take advantage of goods that can be cheaply manufactured because each product run is very large. However, if each of the fifty states and the thousands of local governments could require additional, distinct warnings on cigarettes or other products offered for sale in those jurisdictions, then the makers and packagers of those products would face added costs in customizing their products to distinct markets. Therefore, in most areas in which we see labeling requirements in the United States the national government both requires the labeling and prohibits additional labeling requirements by state and local governments.

Interestingly, even though businesses usually prefer less regulation to more regulation and less restrictive regulation to more restrictive regulation, they are often eager to have federal regulation even if it is somewhat more restrictive than the state equivalent would be, provided that the effect of the federal regulation is to prohibit additional state and local regulation. One can see similar dynamics, of course, in global markets. The World Trade Organization, for example, prohibits protectionist practices but also has a general policy favoring unified markets on the theory that this will lower the cost of customizing products to national markets.

I have discussed two examples of how we might think about instrumental federalism. In the environmental case, I suggested that there is a need for some federal regulation and that is because: Number one, the national government has a greater capacity to gather information about health risks than any individual or locality; and Number two, environmental concerns typically have a national, not to mention a global, dimension to them. Again, in the United States, we have been faced with the problem for many years now of a phenomenon called acid rain. As a result of the burning of fossil fuels in the middle part of the country, the rain that falls on the eastern part of the country contains pollutants that are harmful to water quality, and so, any effective environmental policy regarding the burning of fossil fuels is going to have to combine both national and state or local regulation.

The second example I gave, that of product labeling, was one in which I suggested that having a uniform regulatory regime fits best with the policy of a single market.

Finally, we might imagine that there are domains in which there is no useful role for the national government. But here I must confess that I find it hard to come up with a single good example. In the United States, there are many subject areas that are chiefly matters of state law. For example, rules governing contracts, marriage and divorce, real property, and street crime are all principally matters of state law, but they are not exclusively matters of state law because each of these areas can have a dimension that affects the nation as a whole. Bank robbery, for example, is a federal offense because banks do business across state lines, and violence in a bank can threaten the national economy. There is no general federal family law, but numerous federal laws rely upon state definitions of marriage in setting eligibility requirements for federal benefits. Rules governing real property, I said, are primarily state rules, but the use of property can have an important impact on the natural environment and, as I suggested a moment ago, environmental problems arising from individual or corporate use of private property do not respect state or, for that matter, national boundaries.

It is interesting to note that the people who wrote the United States Constitution in 1787 appear to have been aware of the problem I have just identified, that is, the difficulty of identifying specific domains that ought to be exclusively within the power of the state governments. They wrote out a list of powers given to the national government, and they also specified that whatever was not on the list was reserved to the states, but our courts have long recognized that in the course of exercising one of its listed powers, the federal government will—if only accidentally—often exercise overlapping jurisdiction with the states.

Putting aside the matter of how one figures out the areas best left to exclusive state and local jurisdiction, let me return to the central question I posed for myself. Let us assume that the art of governing a large country will require that some decisions be made exclusively at the national level, that other decisions will be best be made exclusively at the state or local level (although it will be difficult to specify by any sweeping categorical criteria what those decisions are), and that some third category of decisions, like our gasoline additive question, will best be made through cooperation between the central government in setting standards and the state and local governments in carrying them out. Is there any reason to think that a system of instrumental federalism is better than a system of unitary government with occasional one-time decentralizations or utilization of branch offices for dealing with relatively local problems?

To this question I want to give a tentative answer of “yes.” The reason why one might favor federalism to unitary government relying sometimes on branch offices, is the natural human tendency to overestimate one’s own capacities relative to those of others. If the national government gets to decide whether a particular problem is the sort that is best solved by the national government, by the state or local governments, or by some combination, there is a risk that the national government will tend to favor its own decision making. Constitutional and other legal mechanisms that guarantee a role for the states in the enactment and or the implementation of national legislation are a hedge against the tendency of institutions to aggrandize their own power.

Notice that the form of the argument that I have just put for federalism as opposed to unitary government with occasional use of branch offices is one that is similar in form to the argument generally advanced for the separation of powers. It is sometimes said that the reason to separate functions of government into an executive branch, a legislative branch, and a judicial branch is to prevent all power from falling into the same hands. That argument as it is usually made in the United States says that it is worth sacrificing the efficiency that comes from a parliamentary system of government in which executive and legislative powers are combined, for the liberty that comes from preventing potentially oppressive laws from being adopted or enforced. In other words, there is an association in the American literature between limiting power and a kind of libertarianism that comes at the expense of government accomplishing the goals that serve the people collectively.

That is not the argument that I am making for federalism. My argument is that efficiency in the sense of achieving the most practical solutions to the problems confronting society is actually furthered by limiting the power of a central government to prefer its own action over that of the states. To put the point finally in one additional way, I have suggested not a libertarian argument for limiting government power, but a pragmatic one. I tend to agree with the parliamentary critics of American separation of powers that our system of government makes it a little too difficult for the government to act. At the same time, however, federalism conceived as a system of administration utilizes limits on the national government for the purpose of empowering the government to solve problems even if perhaps paradoxically it empowers through limitations.

Let me conclude by noting that I have not said anything at all about how a system of government that aims at instrumental federalism ought to enforce the federalism provisions of its constitution and other legal documents. In the United States, constitutional scholars have long debated whether the courts should play an active role in policing the federalism provisions of the Constitution. From the mid-1930s until the early 1990s, our Supreme Court took the view that it was up to Congress to apply the federalism limits and that this was acceptable because Congress was selected through a mechanism that gives representation to the states. Our national Congress consists of two houses, a lower house from geographical districts of roughly equal population, and an upper house that gives each state two senators regardless of population. In addition, any fundamental change in our Constitution requires ratification by three fourths of the state legislatures, and finally, in the selection of our president citizens vote not as a country as a whole, but on a state by state basis. These structural safeguards, the Supreme Court used to argue, are sufficient to ensure that the national government will respect the limits on its power in favor of the states.

Over the last decade or so, however, the Supreme Court has taken a more active role in policing the bounds of federalism and has struck down a number of acts of Congress on the ground that they exceeded the powers granted to Congress in the Constitution. Those decisions have been the subject of much discussion and debate, and I have been a participant in that debate, but it strikes me that viewed from a broad perspective, the question of whether courts are going to enforce federalism limits is less important than the question of whether to have federalism at all. I have suggested today that countries like the United States and China, not to mention Australia, Mexico, and some other states, can use federalism as a system of administration even if there is no basis for treating the state units as sovereign or even quasi-sovereign entities. As I said, the principal reason for doing so is a matter of humility, and in that spirit, having already taken up too much of your time, I should humbly end by thanking you for your attention and inviting your questions and comments.