Prev Top Next

II. An Economic Theory of Federalism

Federalism describes the arrangement under which a group of equally-sovereign states combine to form a union in which they cede some sovereignty to a central government but retain some sovereignty, too.1 To use an economic analogy, the formation of a federation is like the merger of independent firms. But the analogy is not precise. Some economic mergers are of limited scope and duration—e.g., mergers made for the purpose of pursuing a joint advertising campaign for a few months—while others are complete—e.g., mergers that result in the extinction of the independent firms and the creation of an entirely new business entity. Political mergers are usually between these extremes: the members of the federation do not disappear but retain some measure of sovereignty and the arrangement is meant to last for an indefinite (usually, long) period.2

Presumably, when independent countries first form a federation, they perceive the benefits of the arrangement to exceed the costs. And, presumably, they establish the particular institutions of the federation (the representation rules, jurisdictional and issue competence, voting rules, and the like) so as to maximize the excess of benefits over costs. However, the affairs of the union and of the constituent states may change over time so that either the costs of remaining united will temporarily exceed the benefits or the benefits of union will only be realizable under a different institutional structure. The ideally-designed federal system will anticipate these periods, distinguish between those episodes that call for dissolution of the federation and those that call for restructuring, and provide sufficiently flexible institutions for use in adjusting optimally to the changes.

My purpose in this section is to outline an economic theory of federalism. 3 My central focus will be on the appropriate division of governmental responsibility between the constituent states and the federal government. And I hope also to show that as the factors that explain this appropriate division change, the assignment of government responsibilities ought to change as well. There may be periods in which the central government in the federation should be weak and its duties narrowly defined and other periods in which the central government should be strong and its powers wide. This suggests that, even in an ideally-defined federation, the relationship between the central government and the constituent state governments cannot be immutably fixed. The appropriate division of responsibility should, instead, vary between the center and its constituents.4 Precisely these same real factors should also explain differences between federations.

First, I examine the anticipated benefits of forming a federal union as seen from the perspective of the member states. Then I look at the costs of the federation, with particular attention to the public-choice aspects of a federal system. Next I use the economic theory to explain the dynamics of federalism—i.e., the ways in which changes over time in the costs and benefits of being federated can cause changes in the appropriate assignment of governmental responsibilities as between the different layers of government. I conclude with an account of two recent proposals for restructuring federal-state relations in the United States.

A. The benefits of a federal structure of government

Any theory of federalism should explain two central things: first, why constituent states would willingly cede some degree of sovereignty to a superior government and retain some independent sovereignty for themselves, and, second, which governmental responsibilities ought to be assigned to the central government and which to the member-state governments.

Addressing the first of these two central issues, we begin with the presumption that, when independent countries form a union, they do so because they perceive the benefits of union as exceeding the costs. This section seeks to specify those benefits.5 We shall examine the anticipated costs in the next section.

Economic analysis suggests four benefits to the member states from forming a federal union:

(1) minimizing costs and maximizing benefits that cross jurisdictional borders;

(2) realizing economies of scale in the provision of governmental services;

(3) minimizing the monopolistic distortions of government; and

(4) implementing optimal redistributive policies.

The first economic factor that might lead independent states to consolidate has to do with cross-jurisdictional costs and benefits. For example, suppose that production in Country A generates pollution that imposes uncompensated costs on the inhabitants of Country B.6 Clearly the polluters are beyond the reach of the regulators in Country B, and, from the point of view of the inhabitants of Country B, there will be too much pollution from Country A.

Alternatively, suppose that employment opportunities are slim in Country B but good in Country A and that there is labor mobility between the two countries. If there is a publicly-funded job-training program in Country B, the beneficiaries of that program would include the employers and other inhabitants of Country A. The taxpayers of Country B would be paying more than they need to pay for job-training, while the taxpayers and employers in Country A would be paying less. Following the usual economic logic, the governments of both countries would have incentives not to provide the appropriate amount of job-training.

In both examples, Countries A and B can reduce the inefficiencies by somehow internalizing these externalities. One method of doing so would be through a Coasean bargain. Thus, in the pollution example, Country B could pay Country A's polluters not to pollute, and, in the job-training example, employers in Country A could pay the government in Country B to provide training programs. To pursue the economic analogy I introduced earlier, the countries could enter into a contractual arrangement with respect to, say, pollution or job-training programs, or both. This contract would be a limited merger for a specific activity and, probably, for a limited duration. Indeed, there are many such limited inter-governmental contracts (regional compacts in the United States with regard to lotteries and hazardous-waste disposal are examples, as are the early compacts of the European Steel and Coal Community).

These bargaining solutions to the cross-jurisdictional externality problem may not be practicable. The costs of negotiating and enforcing such bargains may be so high as to make them impossible. Moreover, inter-governmental contracts do not amount to federal unions. What, then, is the connection between cross-jurisdictional externality problems and the formation of a federation? One possibility is that the externality-generating activities are so pervasive (because of, say, geography and custom) in both directions that it would be simpler for the countries to consolidate than to have a series of contractual arrangements dealing with each external cost or benefit. Thus, we might tentatively conclude that where cross-jurisdictional externality problems are numerous, the consolidation of the two governments may be the optimal means of resolving the inefficiencies. A corollary of this conclusion would be that inter-governmental cooperation through quasi-contracts would be an optimal means of dealing with specific cross-jurisdictional externalities.

Thus far, I have been focusing on the formation of a federal union. But we may invoke these same considerations about cross-jurisdictional externalities to explain how governmental responsibilities ought to be assigned once a federal union has been established. Thus, if the costs and benefits of an action, whether public or private, stray across jurisdictional lines, then the highest level of government that can fully internalize the costs and benefits of the action ought to take responsibility. Only then will a governmental authority have the appropriate incentives to regulate the externality-generating activity.7

As a final example of this point, consider the regulation of interstate commerce. The temptations of state governments to favor their own citizens over foreign interests are strong; as a result, state governments tend to confer benefits on their own citizens and to impose costs on citizens of other states. To internalize these costs and benefits to a trading area, responsibility for interstate commerce should lie with the national government and not with lower-level governments.8

A second economic factor that might lead independent states to consolidate has to do with the realization of economies of scale in the provision of governmental services. Suppose that there is some governmental function—say, tax collection—for which there are economies of scale, in the sense that the average cost of collecting taxes falls as the number of taxpayers becomes greater. If these economies would be realized for a population that is greater than that of any one country, then it may make sense for the function to be performed by a consolidated government of the independent countries. As was the case with cross-jurisdictional externalities, the countries do not necessarily have to form a federal union in order to realize these economies of scale: they could maintain their independence and agree to form a joint taxation authority.9

As we did when we considered cross-jurisdictional externalities, we need here to distinguish between the situation in which independent states realize economies of scale in the provision of governmental services through a series of inter-governmental agreements from one in which they can only do so through consolidation in a federation. And, as before, the distinction must lie in the fact that when the number of governmental services for which economies of scale is large, consolidation is superior to mere inter-governmental cooperation. We may also posit that inter-governmental cooperation through Coasean bargains would be the superior means of dealing with economies of scales in governmental services when there are only a few such services for which the average cost of provision continually falls.

As before, we need to distinguish between the formation of a federation and the assignment of responsibilities once the federation has been created. Once a federation is in place, efficiency argues for assigning particular governmental responsibilities to that level of government by which it can be most efficiently performed—whether the member-state or the federal-government level. When the members of a federation negotiate the assignment of a governmental responsibility, they will have an incentive to take economies of scale (as well as other factors) into account.

A third economic factor to be considered by states contemplating the formation of a federation has to do with the minimization of monopolistic distortions generated by the act of governing. Every government has monopolistic elements, and one of the economically-interesting aspects of a federal system is that it minimizes the monopolistic distortions attributable to government. Federalism does this by replacing a single government with multiple governments—i.e., by introducing competition among governing units.10 How does this consideration help to explain the formation of a federation? Perhaps independent states recognize the distortions of monopolistic government and the benefits of competition and affirmatively choose a federation so as to minimize those distortions and maximize those benefits. Additionally, once a federal structure is in place, this economic factor can help to explain to which level of government should be assigned certain governmental responsibilities. For instance, if the social benefits of government can be just as efficiently secured from placing a governmental responsibility at any level of the federal system, then that responsibility should be assigned to the level at which the monopolistic distortions will be minimized. Generally speaking, and again subject to the assumption that the governmental service can be equally-efficiently supplied by any level, the monopoly distortions at the federal level are worse than those at the member-state level. If a state abuses its monopoly position by, say, imposing a confiscatory tax, some of those adversely affected can migrate to other states with less burdensome taxes. (And, of course, the knowledge that inhabitants may leave in response to bad policy will serve to mitigate the monopolistic impulses of local governments.) By contrast, this corrective migration is not nearly as likely in response to a bad policy at the federal level. People may be willing to move from Illinois to Wyoming to avoid what they consider to be the depredations of the Illinois government, but it is unlikely that they would be willing to move to another country in response to a similarly burdensome policy from Washington, D.C.

There is also a further (and different) positive aspect to placing authority at the member-state level. It is not just that competition among the states (combined with free migration) will minimize the possibilities of monopoly distortion. In addition, there are all the positive benefits of competition. In the United States, the states constitute fifty different experiments designed to attract taxpayers and jobs.11 As they frame policies to accomplish that attraction, the states are promoting national economic growth and expansion. They are also searching for the most efficient solutions to common problems, such as tort reform, welfare reform, and health care reform.12

The final economic consideration in the formation of a federation and in the efficient division of authority between member states and the federal government regards optimal redistributive policies. As we saw above, states (even those not affiliated in a federation) compete among themselves to attract taxpayers and jobs. But although the immigration of investment and jobs is desirable, the immigration of welfare recipients from other states is a drain on the receiving community's resources.13 Generous welfare benefits paid for by high taxes will drive out the wealthy and bring in the poor.14 For that reason, independent states who care deeply about redistribution and who otherwise compete for resources may choose to affiliate in a federation. Moreover, once the federation is established, there is an economic reason for giving responsibility for redistributive policies to the federal government, rather than leaving it to the member-state governments.15

B. The costs of a federation

Let us assume that a group of independent states perceive that the benefits of federation are substantial. They will not federate unless they are convinced that the costs of doing so are less than the benefits. What costs can those states anticipate bearing if they form a federation? We may classify these costs as fixed or variable. The fixed costs are those of negotiating and drafting the original agreement to form a federal union. Although these may be considerable, they are, nonetheless, fixed costs and, therefore, should not affect the on-going operations of the resulting federal system. The variable costs are those that depend on the scope of the federation's activities, and of these costs the largest is that of coordinating the relations between the sovereign member states and the federal government.

There are several reasons for believing these costs to be particularly important. First, in the founding document of the federation the division of authority between the federal government and those of the member states will probably not be exact but rather specified according to general standards. There will be uncertainty about what is and is not allowed to the various levels of government, and this uncertainty will cause mistakes by both levels of government. In some instances a government will overreach its authority, and in others it will refrain from exercising authority when it might legitimately do so. Either mistake will impose costs on the federation. Second, there must be some ongoing means of resolving these disputes about legitimate authority. For example, in the United States the Supreme Court serves as the ultimate arbiter of federal-state disputes; in other federal systems, the political process may occupy this role.16 The costs of this dispute resolution can be large.

These considerations apply even to an ideal federal system, one in which the member states have devised an ideal arrangement for realizing their joint goals and continue to act in good faith. In a federation that is less than ideal, the variable costs of resolving problems are heightened. Here are two considerations relevant to this issue. First, suppose that political constituencies are so powerful at the member-state level that they distort policies away from socially-desirable goals and toward their own ends. Those political interest groups may be less powerful at the federal level, and, therefore, it may be less distorting to assign governmental responsibility at the federal level, even though other considerations would seem to argue for assignment to the member-state governments.17 Second, recognize that interstate competition may not always be desirable. There may be instances in which competition among the constituent states leads not to experimentation that produces the best outcome but rather to a "race to the bottom." That is, there might be so much competition among the member states that the result is a minimal set of governmental activities that is less than socially optimal. Central rules precluding or regulating interstate competition may prevent this race to the bottom. It is difficult to distinguish ex ante those circumstances in which interstate competition is good from those in which it is bad. Nonetheless, this difficulty is an example of the practical difficulties that may beset even a reasonably well-crafted federation.18

C. A public-choice theory of federal-state relations

What constraints does politics impose on the optimal design of a federation? To see the impact of this consideration on the issues of federalism, we must turn to public choice theory. That theory explains political behavior (of constituents, bureaucrats, and politicians) using the same rational choice models that economists use to model economic behavior. Public choice theory sees statutes, for example, as the result of forces of supply and demand in the political marketplace. Some interest groups demand favorable legislation, and some politicians supply it.19

How might the insights of public choice theory apply to the issues of the design of a federation? The relevant literature could be read to suggest that, taking the case of the United States as an example, the federal government would almost always exert its power, under the Supremacy Clause,20 to preempt local law (either by regulating or failing to regulate) so as to receive the political and other benefits of providing services for constituents. This prediction notwithstanding, in the United States there is a division of responsibility between the federal and state governments in which, in certain areas, the federal government willingly abstains from regulation, leaving the field to those state governments, when, in fact, the federal government could exert its authority. The actual division of regulatory responsibility is not one that always obeys the predictions of the economic theory of federalism outlined in the previous section. Rather, in the United States at least, the federal government sometimes takes a strong interest in regulating things that seem purely local. For example, the federal government issues charters to banks that are not going to do business across state lines. At other times, the federal government leaves regulatory authority to the states in areas in which one might say that there was a strong national interest. An example is Delaware's preeminence in United States corporate law. Because the economic theory of federalism (as developed thus far) cannot easily explain these phenomena, we need an alternative theory.

To begin, consider that the interest groups who demand regulation prefer, in general, federal regulation to state regulations. This is because it is generally cheaper to obtain passage of one federal bill rather than to obtain passage of 50 separate ones originating in the states. In addition, even if a regulatory bill is passed at the state level, support must still be obtained from the federal level to assure that there will be no conflicting national regulation. In contrast with a state regulation, a competent federal regulation clearly preempts the field. A third reason that demanders of regulation prefer the federal solution is that federal law is generally considered to be of higher quality than state law.21 Lastly, federal law is more difficult to avoid than is state law (because of the greater costs of migrating out of the country by comparison to the costs of migrating to another state). Parties adversely affected by regulation may escape state regulations but cannot escape federal law. Why should the demanders of regulation prefer that state of affairs? Sometimes the purpose of regulation is to "raise rivals' costs" by requiring them to provide a higher-quality output or to do so in a more expensive fashion than they otherwise would, and the inescapability of federal regulation often fits into such a strategy. If that tactic is successful, the rivals may be forced from the market. For all these reasons, interest groups generally prefer federal to state regulation.

Nonetheless, federal regulators sometimes defer to state regulators. Why? Clearly the answer must be that in some cases the preemptive suppliers of regulation (the Congress) gain more from deference than from satisfying the demands of the political interest groups. In all such cases it must also be true that those who demand regulation prefer that regulation be done at the state level. Professor Macey identifies three sets of conditions under which federal deference to the states is likely to occur:

(1) where "interest groups have made an expropriable investment in a particular set of local regulations" and the value of that investment would be dissipated by federal regulation22;

(2) where "the political-support-maximizing solution for a particular regulatory issue differs markedly from jurisdiction to jurisdiction"23; and

(3) when Congress can escape subsequent adverse political reaction by delegating a controversial set of decisions to the states.24

To see an example of this theory of federal deference,25 consider the provision of corporate law by state legislatures. There are two different views about the efficiency of federal versus state regulation of corporations. Those in favor of state regulation argue that competition among the states produces efficient corporate law that maximizes the value of firms and of shareholder wealth.26 Those in favor of a federal law (which would, in the United States system, probably preempt state law) argue that competition among the states for corporate-chartering revenues will lead to a "race to the bottom"—i.e., to a set of minimally-acceptable corporate laws that will tend to transfer wealth from shareholders to managers.27

In point of fact, the competition among the states for corporate charters is over. Delaware is the winner. More than 40 percent of the companies listed on the New York Stock Exchange are chartered in Delaware. Of those firms that decide to reincorporate for whatever reason, 82 percent of them do so in Delaware. Why? It is not the case that Delaware's corporate code is vastly superior to or more lenient than those of other states. In fact, it is indistinguishable from those of many other states. Delaware's predominance in corporate law is due to another, unexpected factor—its small size. Delaware obtains about 16 percent of its state revenues from franchise taxes on corporate chartering. This is such a high percentage of Delaware's budget that it constitutes a credible bond of the state's promise not to alter its corporate law precipitously.28 The highly-specialized bar in Delaware has particular competence in corporate law, and the judiciary in the state is also well acquainted with corporate law. There will be no or few surprises in litigation over corporate-law matters in that state.29

The interest groups in the state of Delaware—the bar, Delaware corporations, investment bankers, the state, and others—have a very strong stake in maintaining this regulatory regime as it is. In Macey's terms, they do not want the value of "their specific capital investments ... destroyed [,as they might be,] if the federal government enacted a pervasive system of federal corporate law that preempted the field."30

Contrast this explanation—i.e., the public-choice explanation—for the division of federal and state responsibilities with that of the economic theory of the previous sections. That latter theory of federalism suggests that state regulation would dominate where there are no externalities. But, as we have seen, state regulation—at least in the case of corporate-chartering laws—sometimes predominates even where there are externalities. "The [economic theory of federalism mistakenly] assumes that the full costs and benefits of a particular legal regime to the public is what motivates the local decisionmaker, while the [public-choice] model ... focuses on the fact that interest groups have a strong incentive to press for local solutions to their regulatory problems."31

D. The economics of comparative federalism

Having seen the economic factors that motivate the formation of a federation and the division of governmental responsibility between the central and member-states levels, we are now in a position to combine our economic theories of federalism and of comparative law to look at differences in federations. It might be helpful to begin by characterizing a traditional view of these differences and then contrasting that view with the one that comes from the economic theories. One certainly recognizes that among the federations in the world today—e.g., the United States, Australia, Switzerland, the European Union, Canada, Argentina, and the Commonwealth of Independent States—there are vast differences. I believe that most commentators would attribute those differences exclusively to historical and political-ideological factors. Moreover, I suspect that most commentators would assert that there is no general theory of federalism (or of comparative law) that can reconcile these differences. Rather, they would probably assert that each federation is a special case, understandable only on its own particular historical terms.

By contrast, I make two claims, one for the economic theory of federalism and one for the economic theory of comparative law. First, I claim that the economic theory of federalism identifies the factors that explain the general characteristics of any federation, including their particular divisions of governmental responsibility. Second, I claim that the economic theory of comparative law explains the different rules and institutions among the federations by appealing to differences in social costs and benefits, resource endowments, notions of fairness, and historical, social, and political constraints on what is legally acceptable.

Perhaps I can forcefully show the distinction between these two ways of looking at comparative federalism by considering current trends in federal-state relations in the United States. There is apparently a trend underway toward a weaker federal government and stronger state governments.32 Most observers would attribute this trend to a change in the prevailing ideology of voters: voters are more conservative than in the recent past and generally favor a relatively smaller federal government and relatively stronger state and local governments. By contrast, I would assert that, if a trend toward weaker central government exists, that trend is a response to underlying real factors that alter the costs and benefits of assigning governmental responsibilities in the federation. If this assertion is correct, then we ought to be able to perceive real economic forces as the cause, independent of changes in which political party controls the national or state governments. We shall look at evidence to this effect shortly.

Differences in federations should also be attributable, at least in part, to differences in economic factors (as well as to differences in notions of fairness and in acceptability, as defined by historical, social, and political constraints). Focusing only on the economic factors, consider how differences in the extent of external costs and benefits might explain differences in federal systems. Recall that the economic theory suggested that cross-jurisdictional external costs and benefits would create an incentive for independent states to affiliate and that the more pervasive those externalities were, the stronger would be the case for a more complete affiliation. Moreover, we saw that, in an existing federal system, member-state governments should handle local externalities (including local public goods), while the central government should handle cross-jurisdictional externalities (and public goods). Suppose that there are two federations, one of which is geographically compact, densely populated, and economically developed and the other of which is spread over a very large territory, sparsely populated, and economically under-developed. In the first federation, pollution is much more likely to be a significant external cost that is beyond the ability of member-state governments to regulate optimally. Instead, the central government will have major responsibilities for environmental regulation. The second federation may not yet have significant pollution problems. Local regulation is likely to be sufficient for the optimal social control of what little pollution there is, and the central government's role in environmental regulation would, therefore, be small. As a result, the first federation is likely to have, all other things equal, a stronger central government, while the second federation is likely to have a weaker central government. This difference is attributable to economic factors, not to different cultural or ideological norms.

As another example, consider the impact on different federations of differences in the efficient scale of production of governmental services. To take taxation as an example, population densities, technology, and taxation possibilities may make the appropriate kind of tax and the locus of the taxation power very different in federations. One federal system may find that the optimal scale of taxation is to rely principally on indirect taxes at the local level while another may find the optimal scale to be direct taxes at the central level (with redistribution to the member states).33

Finally, what about the public-choice considerations raised in the previous section? Might they account for significant differences in federal systems? Recall from the previous section that the public-choice theory of federalism predicted that the starting point in dividing regulatory authority between the central and member-state governments might be that the central government would regulate everything and that member-state governments would regulate only if the center was willing to defer; the center would defer to the member states only if this deference furthered the interests of politically-influential constituents and of federal politicians. If we follow the logic of public choice theory with respect to differences in federations, we could predict that differences in federations could, in part, be attributed to differences between the countries in the strength and political influence of their constituent groups. If interest groups are primarily local—i.e., they do not cross jurisdictional lines, then they will be content with local regulation. Member-state governments will be relatively stronger, and central governments will be relatively weaker. However, if the concerns of interest groups cross jurisdictional borders, then they will prefer central regulation. Member-state governments will be relatively weaker, and central governments will be relatively stronger.

Prev Top Next

1 There are also examples in which the nation-state precedes the federation. For example, both Belgium and Germany were nation-states that created governmental sub-entities that then shared power with the central government. 2 Secession or the orderly dissolution of the federation would be analogous to the break-up of a firm into constituent parts. 3 Law and economics has been one of the most important and productive innovations in legal scholarship of the twentieth century. Yet its contributions to the issues of constitutional law, including federalism, are relatively modest. Many of the early efforts to look at federal-state relationships from an economic perspective focused on tax matters but did not take a broader view. See, e.g., Gordon Tullock, "Federalism: Problems of Scale," 6 Pub. Choice 19 (1969); Wallace Oates, Fiscal Federalism (1972); and Susan Rose-Ackerman, "Does Federalism Matter?" 89 J. Pol. Econ. 152 (1981). An exception is Richard A. Posner, Economic Analysis of Law 635-36 (4th ed. 1992) and Richard A. Posner, The Federal Courts: Crisis and Reform ch. 6 (1985). Nonetheless, Judge Posner's analysis is as much about the optimal size of the nation-state as it is about the division of governmental responsibilities among different layers in a federation. I elaborate this distinction in this section of the paper. 4 There is no reason to posit that the division of authority should change monotonically; that is, in a well-crafted federal system, no level of government should amass power and responsibility over time. Nonetheless, some scholars have argued that there was a tendency for power in the United States to flow to the national government. "The erosion of local autonomy may well have been inevitable, given the constitutional structure. Whatever the founders' intentions, the rules they wrote are skewed in favor of national power. In cases of conflict between state and federal law, federal law wins. If there is disagreement over constitutional rules, a department of the federal government, the courts, serves as umpire. And the principal structural protection for federalism, the direct representation of state legislatures in the Senate, was eliminated by the seventeenth amendment." Michael W. McConnell, "Federalism: Evaluating the Founders' Design," 54 U. Chi. L. Rev. 1481, 1488 (1987). 5 The thought experiment on which I am focusing imagines equal, independent states to be contemplating federal union. I shall not consider here the important question of how an existing union decides whether or not to add a new member and, if so, on what terms. 6 If the polluter's actions impose external costs only within the jurisdiction of a single country, then that country has the appropriate incentives to regulate the polluter's behavior. The benefits of so doing will accrue to the citizens of that country. But if some of the costs spill across state lines, then this diminishes the governmental incentive to regulate the polluter because some of the benefits will accrue outside the jurisdiction. 7 For the argument that these interstate externalities are neither a necessary nor a sufficient condition for federal control, see Perry Shapiro and Jeff Petchey, "The Welfare Economics of Environmental Regulatory Authority: Two Parables on State vs. Federal Control" (Working Paper, Australian National University, Center for Studies in Federalism, 1995). 8 This is the economic justification for the Commerce Clause in the United States Constitution and for the dormant Commerce Clause, both of which I discuss in Section III. For an economically-minded discussion of the Commerce Clause, see Richard Epstein, "The Proper Scope of the Commerce Power," 73 Va. L. Rev. 1387 (1987). Note that independent countries need not form a federation in order to take advantage of the cooperative regulation of trade. They could do so through Coasean bargaining, leading to a limited contractual arrangement. Indeed, the World Trade Organization is (as was its predecessor, GATT) precisely a multi-lateral contract that falls well short of federal union and establishes rules on inter-country commerce and an equivalent dormant Commerce Clause restriction on the members of the WTO. Arguably, the European Union, prior to the 1980s, was simply a limited contractual arrangement among independent nations rather than a federation. 9 Or they could agree to hire a private firm to collect all their taxes. 10 Note that this factor is the only one thus far considered that specifically addresses the novelty of a federal government. All the other factors heretofore considered address the optimal size of the nation-state as much as they address the optimality of a federal structure. 11 Justice Brandeis wrote in dissent in New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932): "It is one of the happy incidents of the federal system that a single courageous state may, if its citizens choose, serve as a laboratory, and try novel social and economic experiments without risk to the rest of the country." Professor Susan Rose-Ackerman has argued that this may be an overly-optimistic view of the role of constituent states in a federation. (Susan Rose-Ackerman, "Risk Taking and Reelection: Does Federalism Promote Innovation?" 9 J. Legal Stud. 593 (1980). See also Susan Rose-Ackerman, "Does Federalism Matter? Political Choice in A Federal Republic," 89 J. Pol. Econ. 152 (1981).) She examines the incentives for individual politicians to develop innovative policies in three different governmental structures: a highly centralized government, a decentralized multiple-government setting, and a federal government that is between these extremes. In neither of the extremes does she find much of an incentive for a politician intent on being reelected to adopt innovative policies. Professor Rose-Ackerman says that a two-tiered system like that in a federation improves matters but not much. In a federal system politicians may choose to innovate in order to run for federal office or to be appointed to a federal office. Additionally, she notes, in a federation low-level elected office is more attractive than in a more simple governmental structure. The winner of a local election can try for a higher office. Thus, there is more competition for lower-level offices and the returns to risk-taking innovation are generally greater. One last improvement that federalism accomplishes is that politicians who aspire to higher office have an incentive to adopt policies that may generate benefits across jurisdictions. Clearly this incentive is not as strong in the alternative governmental systems considered. But Professor Rose-Ackerman finds that these effects still may not produce as much innovation as would be deemed socially optimal. To get closer to that result, the central government in a federation may wish, she suggests, to give grants to local governments in order to subsidize innovation. The policy would be beneficial to the politicians and would, in addition, improve social efficiency: federal politicians would get credit for encouraging innovation, and the local politicians would get credit for innovation without having to induce local taxpayers to pay increased taxes to pay for risky novelties. 12 For an economic argument in favor of state experimentation in product liability, see David A. Rice, "Product Quality Law and the Economics of Federalism," 65 B.U. L. Rev. 1 (1985). 13 Most people seem to believe that a state is better off if it discourages in-migration by welfare recipients from other states. This is not necessarily true. It may be that welfare recipients (like many recent international immigrants) will eventually be productive, tax-paying citizens after a transition period on welfare. Nonetheless, for the purposes of the argument in the text, assume the in-migration of welfare recipients from other states to be undesirable. 14 "[T]he level of redistribution in a decentralized system is likely to be lower even if there is virtually unanimous agreement among the citizens that higher levels would be desirable." Oates, Fiscal Federalism 6-8 (1972). 15 There are, of course, compelling non-economic reasons for increasing redistributive policies and for placing them at the federal level. For example, in the United States there is now general agreement that uniform national rules on civil rights and civil liberties are preferable to disparate state rules on those matters. See Richard B. Stewart, "Federalism and Rights," 19 Georgia L. Rev. 918 (1985). 16 As we shall see in Section III, there are those who argue that in the United States the federal political process, not judicial review by the Supreme Court, is the best arbiter of federal-state relations. 17 This is, of course, an economic gloss on James Madison's most important contribution to the ratification debate. Madison argued, principally in Federalist 10 (James Madison, Alexander Hamilton, & John Jay, The Federalist Papers (Penguin ed., 1987)), that individual liberties, such as property rights and freedom of religion, are best protected at the national level, not the state level. Madison held that the most serious threat to individual liberty is the tyranny of a majority faction and that factions have the greatest effect in the locality in which they are concentrated. In the nation as a whole, they are a small minority. Therefore, factional tyranny is most likely in the state (or local) government and least likely in the national government. 18 In an important recent article, Professor Revesz questions the argument that destructive interstate competition requires placing environmental regulation at the federal level. "Rehabilitating Interstate Competition: Rethinking the 'Race-to-the-Bottom' Rationale for Federal Environmental Regulation," 67 N.Y.U. L. Rev. 1210 (1992). Professor Revesz shows that there is very little support in current theories of interjurisdictional competition (largely in the literature on interstate competition for bank and corporate charters) for the dire predictions of the 'race-to-the-bottom' hypothesis and that, even if there were such support, as a matter of practice a state would either not strictly enforce federal standards or would relax other regulatory controls in order to attract industry. 19 For an excellent summary of the literature, see Daniel A. Farber & Philip Frickey, Law and Public Choice: A Critical Introduction (1991). For an application of public choice theory to federalism, see Jonathan R. Macey, "Federal Deference to Local Regulators and the Economic Theory of Regulation: Toward a Public-Choice Explanation of Federalism," 76 Va. L. Rev. 265 (1990), on which much of this section relies. 20 U.S. Const., Art. VI, § 2. 21 Alexis de Tocqueville remarked on this matter in Democracy in America (1832): "[T]he business of the Union is incomparably better conducted than that of any individual state. ... [The Union] has more prudence and discretion, its projects are more durable and more skillfully combined, its measures are executed with more vigor and consistency." James Madison in Federalist 10, supra n. 20, noted that the federal government would probably have better-qualified officials than would the state governments. 22 Macey, supra n. 20 at 274-75. 23 Id. at 275. 24 Id. at 268. This is an extension of Professor Morris Fiorina's observation that Congress can sometimes avoid political fallout for difficult decisions by delegating them to an administrative agency. See M. Fiorina, "Legislative Choice of Regulatory Forms: Legal Process or Administrative Process?" 39 Pub. Choice 33 (1982) and Peter Aranson, Ernest Gellhorn, and Glenn Robinson, "A Theory of Legislative Delegation," 68 Cornell L. Rev. 1 (1983). 25 Professor Macey calls this theory a "franchise theory of federalism." He draws an analogy between Congress' delegation of regulatory power to the states and a business' decision to franchise its operation. Under a franchise agreement the business owner refrains from operating a business himself, choosing instead to license to another firm the rights to market, sell, and produce his product. 26 See, e.g., Frank Easterbrook & Daniel Fischel, The Economic Structure of Corporate Law (1990) and Ralph K. Winter, "State Law, Shareholder Protection, and the Theory of the Corporation," 6 J. Legal Stud. 251 (1977). The literature frequently describes this competition as a "race to the top." 27 See, e.g., Louis K. Liggett Co. v. Lee, 288 U.S. 517, 559 (1933) (in which Brandeis, J., in dissent, describes the competition among the states for corporate charters as a race "not of diligence but of laxity"); and William Cary, "Federalism and Corporate Law: Reflections Upon Delaware," 83 Yale L. J. 663 (1974) (in which the competition for corporate charters is described as a "race for the bottom"). 28 Roberta Romano, "Law as a Product: Some Pieces of the Incorporation Puzzle," 1 J. L. Econ. & Org. 225 (1985). 29 Indeed, Professors Macey and Geoffrey Miller have argued that Delaware has the further attraction that it channels corporate litigation into the state's courts. See Macey & Miller, "Toward an Interest-Group Theory of Delaware Corporate Law," 65 Texas L. Rev. 469 (1987). 30 Macey, supra n. 33 at 279. 31 Id. at 284. 32 I describe this trend in much greater detail in the next section. 33 When a change in the optimal scale of governmental service occurs so that a particular governmental function ought to be transferred to another level in the federal system, Coasean bargaining among the levels may lead to voluntary reassignment of the responsibility. That is, if the function could be more efficiently performed at another level, it could be in the interests of all the levels to have the duty transferred. Congress could be conceived of as an institution that, among other things, facilitates these Coasean bargains among states or regions. Indeed, the First Congress made a large number of such deals in which representatives acted principally as state advocates. I am grateful to my colleague Jim Pfander for pointing this out to me.