I. Federalism, Liberalism, and Freedom

I have previously argued that liberalism has historically contained and will continue to contain both rationalist and pluralist strands, the former emphasizing the potential threat to freedom posed by intermediate institutions and communities, and the latter stressing the freedom from the central state guaranteed by such institutions and communities. I have also argued that the difference between these two strands crosscuts rather than tracks the difference between welfare liberalism and classical or market liberalism.²

Indeed, welfare liberalism and classical liberalism—the “new” and “old” liberalisms of this volume's title—face the rationalist-pluralist trade-off in fundamentally the same ways and for the same reasons, sharing a commitment both to freedom from the dictates of the central state and to freedom from local despotisms; and in this both liberalisms differ from other ideologies and philosophical systems. Neither conservatism nor socialism nor democratic theory confronts this tension in that same way; each can more easily plump for the central or the local simpliciter. I have suggested that this helps us to see the commonalities and continuities between classical liberalism and welfare liberalism, to see them as belonging to a common intellectual genus, notwithstanding the partisans of each who occasionally try to read the other out of the liberal tradition.

Insofar as the trade-off between federalism and central-state government is a particular case of the pluralist-rationalist trade-off, it follows, of course, that questions about federalism and centralism will also fail to track questions about classical liberalism and welfare liberalism. One purpose of this essay is to make that claim more concrete.

Another is to test the claim against what many classical liberals would take to be a hard case: the history of federalism in the United States. The shift from classical to welfare liberalism as a public philosophy in the United States—indeed, the shift from one to the other as the generally understood meaning of the word “liberalism”—was more or less contemporaneous with an increasing centralization of power in the American federal system. I mean to suggest that these trends have less to do with each other than it appears at first blush, less than many have assumed. The relationship between federalism and the classical liberal understanding of freedom in eighteenth- and nineteenth-century American liberal and constitutional thought was one of general institutional approximation, not one in which federalism was constitutive of freedom. The centralization of power in the American constitutional system is not one unified development. It began, not as an aspect of the development of welfare liberalism, but as an aspect of a general trend in the United States and elsewhere against liberal constitutionalism generally. Later in the twentieth century, the authority of the states was further restricted, but primarily in ways that do not divide welfare and classical liberals.

II. Federalism and the Founding

Federalism is the accidental innovation of the American constitutional system. Other novel aspects of that system were intentionally developed and intensively discussed and theorized: these include a distinctive sort of separation of powers in which the executive and legislature were each, separately, dependent on the same electorate; written constitutions (at the state and then federal levels) adopted by special conventions summoned for that purpose, rather than by legislatures; and bicameralism without a separation into estates. One arguably novel aspect of the system as it eventually developed, the permanent party system, was theoretically rejected and developed more or less in spite of the intentions of the framers.³

Federalism was neither intentionally chosen nor rejected. Confederation of one sort or another was widely understood as necessary and inevitable. Thirteen independent polities could not have won the Revolutionary War, but their abolition and consolidation into a single polity was unthinkable. And use was made of the (thin) available resources in political theory concerning confederated republics, notably the brief discussion of them in The Spirit of the Laws, the great work of the French political theorist Montesquieu (1689–1755).⁴

For the most part, however, the federal system evolved as the unintended result of a series of compromises and power struggles, both among the states and between the states and the center.⁵ These compromises and contests began in the Constitutional Convention held in Philadelphia in 1787, with the so-called Great Compromise creating one legislative house representing the people and one representing the states on the basis of equality. The U.S. Constitution, as written, included some elements of James Madison's initial proposal, the so-called Virginia Plan, but not one element that he considered crucial for the system's success: a congressional veto over state legislation. The question of whether states could judge federal legislation unconstitutional and nullify it or interpose themselves between federal authority and the citizenry was to arise within a few years of ratification and was finally to be decided only by force of arms decades later in the Civil War.

It is only in retrospect that we have developed theories about federalism as a complete, complex system. (I will consider these theories in the next section.) All of this is perhaps unsurprising when we remember that it was not yet clear, and would not be clear until after the Civil War, that the United States was effectively a single state with internal administrative divisions. At the time of the American founding, the Holy Roman Empire (962–1806 c.e.) was the only apparent example in the world of a large confederation, and it was only a state in complicated and attenuated ways, being made up of (as we would say today) sovereign armed states with international personalities of their own. Europe itself had been imagined as a federation or confederation by eighteenth-century philosophers, which meant only that the nascent international legal order and shared community of morals and commerce might be strengthened so as to eliminate interstate war; it did not mean a dissolution of the states.⁶

This suggests that the bare concept of federation did not mean to eighteenth-century minds what it means to ours, namely, one internationally-sovereign state that happened to have a distinctive kind of purely domestic constitutional organization. Instead, “federation” (or “confederation”) could denote a range of systems that blurred the domestic-international distinction. Moreover, while there is endless dispute over how the founding generation of Americans viewed the relationship among the colonies-cum-states, and about the relative priority (in time and in importance) of the thirteen polities and of the union, it seems clear that the relationship was at least to some degree an international one, and the founding to at least some degree a treaty-and-alliance-like development.⁷ The fears about interstate war expressed throughout the 1780s and prominently in The Federalist were not, or were not simply, hyperbole.

Thus, for example, the idea of competitive federalism, of federalism enhancing freedom because it introduces exit-based competitive pressures on each state to liberalize its laws, long post-dates the founding. In a 1791 essay against the consolidation of the states into a single government, Madison focused all his attention on the possibility that consolidation would dangerously increase executive power, and none at all on any competition between the states.⁸

The mechanisms by which the founding generation imagined that federalism might enhance liberty were much blunter. The states would check the center, the states would check each other, and the center would check the states in ways that relied on a general balancing of power (with the possibility of armed force lurking in the background), not on any smoothly or automatically operating incentives.

The first such mechanism, by which states would check the center—as laid out in Federalist Nos. 25, 45, and 46—relied on the greater loyalty that citizens would tend to feel to their states than to the federal center. This would tend to check any move to despotism by the center, in extremis through armed resistance. “Publius” (meaning Alexander Hamilton for Federalist No. 25, James Madison for No. 45 and No. 46) maintained that a standing army at the center would be less dangerous to republican liberty than the alternative, standing armies in the several states, because “in any contest between the foederal [sic] head and one of its members, the people will be most apt to unite with their local government”:

The reasons for this suspicion and jealousy, this natural likelihood that the “first and most natural attachment of the people will be to the governments of their respective States,” are so plentiful as to place the prediction “beyond doubt.” People are more likely to have neighbors, friends, and family in state than in federal offices or employment. They are more likely to have reasonable hopes of such offices or employment themselves. State governments will attend to immediately-felt local needs, whereas the federal government's primary business will seem far-off and relatively unimportant. State politics will simply be more familiar and comprehensible. For these reasons, “the popular bias may well be expected most strongly to incline” toward the states.¹⁰

This inclination, this salutary suspicion and jealousy, will tend to prevent the buildup of any armed forces that exceed the federation's genuine needs for external defense. Such a buildup cannot take place all at once, and the people and the states (electing the House and the Senate, respectively) would certainly interrupt the process before it could be completed. Even if the buildup did take place, however, the greater attachment of the people to their states than to the federal center would allow the states to defend themselves successfully:

Madison emphasized elsewhere that the constitutional system itself would not mechanically protect liberty, and that its protection depended on the willingness of the people to act:

Although Madison in the late 1780s did not expect or hope that this combination of state and popular resistance to the center would often be needed, a decade later he coauthored the Virginia Resolution, leading the Virginia legislature to declare the censorship rules of the 1798 federal Alien and Sedition Acts unconstitutional, aiming to rally popular opinion and the other states against the restrictions.

In addition to the states' checking the center through popularly supported resistance, Madison noted a second mechanism whereby the compound federal system could protect freedom: “in a confederal system, if one of its members happens to stray into pernicious measures, it will be reclaimed by the frowns and the good examples of the others, before the evil example will have infected the others.” He repeated this idea on occasion from The Federalist through the end of his career, with regular attribution of it to Montesquieu, who wrote: “If a sedition occurs in one of the members of the confederation, the others can pacify it. If some abuses are introduced somewhere, they are corrected by the healthy parts.”¹³

Notice that, here, it is not the federal government whose frowns and good examples would correct the domestic evils of any one state. But neither Montesquieu nor Madison offered any intra-constitutional mechanism by which the states might affect each other's internal conduct. The idea seems to have been that, in certain extreme cases such as one state's falling to a would-be monarch or military despot, the other states could be relied on to respond extra-constitutionally, militarily if need be.

Montesquieu also held that “the federal government should be composed of states of the same nature, above all of republican states,” not of a mixture of republics and monarchies as in the Holy Roman Empire.¹⁴

And the Constitution was to “guarantee” this, in its only clause committing the federal government to the defense of freedom within the states in any general way. The so-called republican guarantee clause¹⁵ completes our survey of federalism and freedom, by introducing a third mechanism, a federal check on states. The great expounder of the Constitution and nineteenth-century Supreme Court Justice Joseph Story explained the republican guarantee clause as follows:

This is a sharp reminder of how different the constitutional order seemed in the decades before the Civil War. Note the astonishing thought that, without the clause, the national government would be legally powerless to intervene even though a state faced usurpation, tyranny, the trampling of liberties, and the ruin of law and order! And the debates in the Constitutional Convention reflected the same assumption. Of all the delegates recorded as discussing the clause, only John Rutledge of South Carolina expressed the idea that “Congress had the authority if they had the means to co-operate with any State in subduing a rebellion. It was & would be in the nature of the thing.” No one supported Rutledge in thinking the clause superfluous. Five other delegates all championed the clause as necessary to authorize such action:

Conspicuously absent from that picture, to a contemporary eye, are the federal courts. The would-be monarch might induce his state legislature to grant him a title of nobility, which could in principle authorize federal court review even under the unamended Constitution. (So would a state's suspension of habeas corpus.)¹⁸

However, no one seems to have thought that the federal courts could act as the primary federal guarantors of freedom as such within the several states. The “General Government's” action would be congressional and presidential. Beyond the kinds of extreme abuses that would fall under the guarantee clause, Madison noted that the federal government was relatively powerless to protect freedom within the states:

Thus, the ways that the founding generation conceived of federalism as enhancing liberty were not, or not simply, those that became familiar to later generations, who thought about federalism as a single, complex but unified, purely domestic constitutional order. The states could protect liberty against the center by commanding sufficient loyalty to call forth resistance—if necessary, armed resistance—even against the center's professional army. The states could monitor each other's internal actions, disapproving them with “frowns.” The center could protect freedom within the states by removing whole domains of legislative authority from the latter, and by enforcing republicanism if necessary, but not with a general federal supervision over state laws (as Madison wanted) nor by detailed federal court review of state laws. (For many decades after the founding, most judicial review of the era was intrastate—state courts reviewing state legislation.)

To this must be added the Federalist No. 10 analysis of factions in the extended republic: individual states might endanger local minorities by coming under the sway of a passionate and interested local majority faction, but such factions would be difficult to assemble in the extended republic. It is curious, however, to note the mismatch between this argument and the actual institutional design at hand. Federalist No. 10 is an argument in favor of large republics, not federal ones. It offers no reason to ever prefer state over federal jurisdiction. A federal government of a few enumerated powers sitting atop thirteen state governments of general jurisdiction does not seem like the proper remedy for the ills diagnosed in Federalist No. 10. The specific denial to the states of a few of the powers most easily abused by local majorities—the power to print paper money or to abridge contracts²⁰

—improves the match between analysis and institutions a bit, but not overwhelmingly. The argument about faction pushes in the direction of a unitary extended republic in order to protect the liberty of local minorities. It is not an argument about federalism and freedom at all.

In short, the constitutional federal order, even combined with the political thought about federalism in the founding generation and the immediately following decades, does not amount to a streamlined system seamlessly operating to protect or promote individual liberty. Neither competitive federalism automatically checking the states through a market-like mechanism, nor a legal order allowing federal courts to guard individual liberty against the states and to monitor the enumeration of federal powers, is much in evidence. What we find instead is a number of powerful governments given more or less blunt tools with which to affect one another, with some reason for hope that they will generally do so in directions favorable to liberty because the people in some broad sense are likely to want liberty. There is no equation of the safeguarding of freedom with state authority, federal authority, judicial authority, or the balance of these attained by the federal order as a whole.

To some degree, this is true of all good political theory about institutional design; it shows an understanding that institutions at best approximate justice, and it does not conflate generally good procedures with morally desirable particular outcomes. The American founders, however, treated at least some questions of institutional design as more nearly identical with questions of justice. Federalism might have been a good idea all things considered, but (for example) the separation of powers was an absolute moral necessity. Also indispensable was an independent judiciary that had the sole authority to issue criminal convictions because habeas corpus was guaranteed against the executive, and because bills of attainder (and ex post facto laws, of which bills of attainder are a subset) were prohibited to the legislature. The political thought of the founding era is filled with references to the separation of powers as being at least partly constitutive of liberty, quite unlike the way that federalism was discussed.

III. Federalism in Practice

The absence of any systematic theory of federalism's contribution to freedom in the early decades of American constitutionalism might be taken to be a matter of blindness to what was close at hand. Perhaps American federalism in practice did create a system that worked consistently and necessarily for the promotion and protection of freedom. We should pause to consider contemporary accounts of liberalism and federalism to see whether they are sound and whether they describe federalism as it came to exist in the United States over the eighteenth and nineteenth centuries; this will affect how we view the departures from that system in the twentieth century.

The most fully worked-out theory of federalism's benefits for freedom is the theory of competitive federalism, summarized by James M. Buchanan²¹

and refined, developed, and tested by public choice economists and scholars in the neo-institutionalist school of political science.²²

In general, competitive federalism as an ideal fits poorly with federalism as it actually exists in the world, however.²³

Federalism generates competition insofar as exit from provinces (or states, in the U.S. case) is feasible, and insofar as there is a suitable range of other provinces to enter into so that one's preferred policy option is available. In order to generate competitive pressure on the provinces to reform, moreover, there must be some connection between the exit of those dissatisfied by existing policies and the incentives faced by public officials. Finally, in order for competition to generally increase freedom, those whose (potential) mobility affects officials' incentives must act on the basis of preferences for freedom.

These conditions are fairly well met for businesses that are major local sources of employment or tax revenue, but that are nonetheless fairly mobile across jurisdictional boundaries and are interested in moving primarily to avoid high taxes or onerous regulations. Nevertheless, for natural persons and for policy domains besides taxation and business regulation, these conditions are typically met badly if at all.

Consider the following facts: The states or provinces that make up federations tend to be very large. The average population size in states or provinces of major modern democratic federations—the United States, Canada, Germany, Spain, Australia, and so on—is in the mid-single-digits of millions. The number of provinces in a federation is typically in the low-to-mid double digits. This simultaneously makes exit from provinces costly—much costlier than moving across the street or to a neighboring city—and restricts the number of provinces that are providing policy packages. Insofar as persons and businesses are would-be policy consumers, they face an oligopoly of policy providers; or, conversely, insofar as residents and tax bases are the goods that states/provinces try to purchase with their policies, the provinces are oligopsonistic buyers. “Entry” into the policy-provision market by new provinces is typically constitutionally difficult or impossible. And provinces offer their laws and policies as a package, almost all of the time, meaning that even policy-consumers with realistic possibilities of exit can probably only choose on the basis of one or two policy topics, immunizing the rest from competitive pressures. The combination of a fixed and small number of providers and high costs of exit makes it very difficult to get significant competitive pressure out of a federal system.

Furthermore, provinces in really-existing federalism are very often tightly linked with an ethnic, cultural, or language group. This imposes substantial costs on those who would exit the province; there may be no other place they can go where their language is the local language or where they feel identified with the local culture. To put it another way, the ethnocultural aspects of the policy package offered by provinces are often weighted very heavily, so heavily as to greatly reduce any competitive pressures on the provinces with respect to any other policy domains. Indeed, Buchanan notes that “federalized structures allow for some partial mapping of politics with tribal identities. At the very least, federalized structures reduce the extent to which tribal identities must be grossly transcended.”²⁴

Buchanan takes this to be a virtue, because it reduces the taxing demands on our extended moral sympathies imposed by the Hayekian “great society” or extended order. Nonetheless, just to the extent that this is true, exit is weakened as a general feature of federalism, and the competitive federalism that Buchanan is generally concerned to defend is ineffective. Moreover, just to the extent that this is true, the local polities run the risk of becoming oppressively homogenous—of being dominated by an unjust majority faction, as Madison feared.

The federalism of the American Constitution and early republic did not approximate the ideal federalist order imagined by competitive-federalism theory. It did not even approximate it as well as contemporary American federalism does. Transportation costs and times have fallen tremendously; both persons and businesses or capital can exit states much more easily today than they could two hundred years ago. The loyalty to states, their approximation of “tribal identities,” has dwindled to almost nothing in most parts of the United States. There are regional or sectional loyalties, but these do not impede exit and competition in the same way; and loyalty to the nation-state as a primary identity is much more powerful than those regional ties, much more powerful than could have been imagined in the eighteenth century. During the founding era, references to exit from one state to another (as opposed to exit from one state to the frontier) as an important constitutional fact were almost nonexistent.²⁵

Much closer to the views of the founders are understandings of federalism that depend on political rather than market-like mechanisms. One understanding rests on the thought that more-local governments are easier for the electorate to control effectively, and less prone to the agency problems that allow public officials to pursue their own ambitions and agendas unchecked. If the electorate tends to wish to preserve liberty—as always, a premise that American political thought is somewhat ambivalent about—and if the states are more directly responsive to the electorate's wishes, then the prerogatives of the states in a compound system will tend to protect liberty. This idea, which centers on democratic voice rather than on individual exit, does not require radical versions of states' rights, much less opposition to the existence of any center at all. It can accompany recognition that certain goods (e.g., continental defense) must be or are better provided by a federal union. The idea is reflexively skeptical of the center in the compound system, however, and tends to insist on narrow understandings of central authority. This account, unlike competitive federalism, bears a recognizable relationship to the political thought of the founding era and the following two generations.

Two related and crucial sets of antebellum disputes about federalism did involve exit, of course: the treatment of fugitive slaves—the subject of the 1850 Fugitive Slave Act—and the treatment of slaves who accompanied their masters into free territory, the subject of the 1857 case Dred Scott v. Sandford.²⁶

Until the 1850s, the existence of a plurality of jurisdictions, some of them without legal slavery, did informally contribute to freedom via exit. They did not do so in a way that much resembles the standard image of competitive federalism. Free states typically did not understand themselves to benefit from an influx of free blacks and, indeed, took steps to prevent this. Nonetheless, the escape of slaves from slavery was a cost to slaveholders and slave states, and they successfully sought to end the jurisdictional competition. It seems to me that federalism's facilitation of exit from slavery was very far from being seen as a particular case of a general feature of federalism. It was anomalous, and was defended by abolitionists but not by moderate northerners, even though the latter were both critical of slavery and, as Whigs and heirs of Whigs, generally supportive of states' rights.

This brings us, at last, to the central difficulties with any claim that the federalism of the early American republic operated in practice to advance liberal freedom in anything other than a rough and ready way. These difficulties are evident in the increasing convergence, over the course of the first half of the nineteenth century, between the cause of states' authority and the cause of slavery; the inescapable fact that slavery was finally eradicated only by a tremendous increase in the effective capacity and the constitutional authority of the central government, leaving the federal system transformed; and the postwar struggle between southern state governments determined to recreate racial tyranny, and a central government that, consistently for a little while and intermittently thereafter, tried to stop them. The antebellum constitutional order cannot be defended on liberal grounds in any detailed way; and the postbellum Constitution was superior precisely insofar as it shifted authority to Washington, D.C. That Reconstruction ultimately failed and was replaced with Jim Crow does not alter this conclusion, as this failure was understood as a partial restoration of the status quo ante in the federal system.

Federalism can either directly impede central-state enforcement of rights against provincial governments, by reserving the relevant authority to the provinces or denying it to the center; or indirectly do so, by sufficiently weakening the central state's political autonomy, political capacity, or institutional strength so that it cannot effectively take the needed actions. The system of approximations and balances before the Civil War tended in the former direction; the central government arguably lacked the constitutional authority to ban slavery within the states (and, according to the Dred Scott Court, lacked the authority even to ban it in territories).²⁷

The system after the Civil War was a failure not of de jure authority but of effective political capacity. Of course, these are also the same means by which federalism can restrict the center from violating rights—the enumeration of powers and the Tenth Amendment²⁸ exercising direct legal restraint, the fostering of Federalist No. 46–style loyalty to the states and suspicion of the center exercising indirect political restraint. My point here is only that both mechanisms were in evidence in the failure to liberalize the herrenvolk democracies of the South;²⁹ and thus neither type of federalist approximation of justice can be thought to have approximated it so closely as to render any departures from it suspect on liberal grounds.

I should note that by “liberal grounds” I mean nothing that distinguishes between market liberalism and welfare liberalism in the slightest. It has been argued that conceptions of freedom have always been developed in contrast to the reality of slavery; in any event, that was certainly true in the United States.³⁰

Liberal freedom—negative liberty, individual liberty, the “liberty of the moderns”³¹—faces its starkest opposite in human slavery, a fact that has been central to the development of American ideas of freedom. American liberal constitutionalism has also been deeply concerned with institutional questions, with understanding what political arrangements could most effectively protect and promote freedom over the long term. But the institutional questions are about means, whereas slavery is a denial of the end. The rhetoric of freedom that has been used by defenders of slavery and Jim Crow has always been designed to obscure this, focusing attention on the means of federalism and constrained central power and away from the ends at stake.

IV. Approximation

A liberal federalism needs to be concerned with a number of actors simultaneously, each of which has more than one possible normative valence: the central state, state governments, and private persons and institutions. Private persons and institutions are the primary holders of the rights that liberalism seeks to protect; but they can also be the threats to each other's freedoms, the problem that justifies the turn to government in the first place. State governments should be in the business of restraining those threats, but will sometimes instead authorize or reinforce them, as in the case of a system of slavery in which private persons are made the masters and owners of other persons by law, and in which the “rights” of the masters are enforced using the power of the state's legal system. State governments also may check expansions of central authority, whether those expansions are detrimental to liberty (the Alien and Sedition Acts) or otherwise (Reconstruction). The central government, in turn, may directly act to threaten freedom, may empower either private actors or state governments to do so, or may intervene against either private actors or state governments to prevent them from violating rights.

No one of these actors always seeks to protect or promote freedom, and none of them always seeks to restrict it, either. Balancing them against each other in the hope of approximating a just degree of freedom as the outcome must be an imprecise enterprise, no matter how detailed the constitutional texts and doctrines specifying which has what power. If, as Madison thought, the states might sometimes be the enforcers of the Constitution, then they will necessarily have some amount of autonomous political weight that might also be thrown around in less desirable ways. If the federal government is to have the authority to prevent private actors from harming or enslaving one another, then it will necessarily have some political capacity to infringe on private actor's rights. Constitutions are not self-enforcing, and grants of power are not easily contained by initial purposes. Even judicial review cannot transform these basic facts of political life. What arrangement of power and authority will generate the most just political outcomes is a contingent matter; sometimes one actor will pose the greatest threat to freedom, sometimes another. We should expect variation over time as well as across different federations.

To acknowledge this variability is not to deny the existence of a general link between federalism and freedom. Buchanan suggests “that a coherent classical liberal must be generally supportive of federal political structures, because any division of authority must, necessarily, tend to limit the potential range of political coercion.”³²

I think that is roughly correct. I doubt that “necessarily”; federalism might, in principle, simply subject people to an additional layer of injustice and domination. Nonetheless, as a matter of fact, the different levels of government do tend to check one another in desirable ways. That the balance of authority is a matter of approximation, and that no particular balance can just be identified with freedom or justice, does not mean that the balance does not matter; at the extreme, wholly unitary states are less likely to be respectful of freedom than states with some real element of federalism. But Buchanan is right not to say anything much more precise than “generally supportive” and “tend to.” Liberals (classical or otherwise) need not automatically support just any given federalist status quo against any given changes to it. Shifting authority at the margin to the center does not automatically increase “the potential range of political coercion.”

V. Federalism and Liberalism Over Time

If federalism's contribution to freedom is, or as a matter of American public philosophy was generally conceived to be, a matter of approximation and general balancing tendencies, then reallocations of authority and competences to the center cannot be understood as necessarily diminishing liberty. Moreover, if twentieth-century liberalism was generally sympathetic to reallocations in the direction of the center, that is not even prima facie proof of a decisive break between it and earlier American liberalisms. And with more fine-grained analysis we will see that the first stage of the centralization in the American constitutional order had relatively little to do with liberalism of any sort, while the second stage improved the federation's approximation of freedom in ways that welfare liberals and market liberals alike can endorse.

Over the decades from, say, 1910 to 1950, a number of broad fundamental changes took place in American constitutional practice and theory, and in American public philosophy and policy. The economy came under increasing regulation, and that regulation was increasingly central rather than state-level. Authority shifted away from the courts and Congress to the presidency, especially during wartime and the Great Depression. The federal government increasingly abridged individual liberties and restricted civil freedom (e.g., universal conscription and suppression of antidraft and antiwar speech, the Red Scare, the internment of Japanese-Americans during World War II, the federalization of Jim Crow under Woodrow Wilson, and so on). Eventually, the federal courts would increasingly become the guarantors of a variety of liberties against both the federal and the state governments; but that came later, and there was little movement in that direction until after the end of World War II.

More or less simultaneously, “liberalism” came to denote a much more redistributive and regulatory stance than it once had. The political and philosophical position formerly known as liberalism dwindled so dramatically that social critic Albert Jay Nock could romanticize the few remaining adherents as “the Remnant.”³³

Some have been tempted to equate these trends, identifying all of the changes in the constitutional order with the intellectual shift from classical to welfare liberalism. This seems to me untenable.

American political history cannot be understood in isolation. Over the decades from the 1910s to the 1940s, the Liberal Party in Britain imploded; liberal parties in continental European democracies were pretty thoroughly squeezed between, at best, social and Christian democratic parties, or at worst, communist and fascist ones. The legacy of liberal constitutionalism and the rule of law came under sustained assault during the long crisis of the Depression and the two world wars, never emerging unscathed and sometimes not emerging at all.

Franklin Roosevelt propounded a constitutional theory in which the president could enter into international agreements without the consent of Congress, in which much authority previously understood as legislative was concentrated in the administrative apparatus of the executive branch, and in which the elected branches, governing with a broad electoral mandate, could legislate more or less unconstrained by a judiciary enforcing a written constitution. Roosevelt also, of course, abrogated the very strong long-standing norm against presidents seeking third (much less fourth) terms in office, a norm grounded in a fear of the centralization and entrenchment of power in the presidency. Such is the character of executive branches; in times of crisis or emergency, they are impatient with norms and formalities and restraints, feeling the urgent need for action. But the “crisis of parliamentary democracy,”³⁴

the perceived incompatibility of the rule of law and formal procedures with the circumstances of emergency, the pressure on constitutional systems to vest power in one man when catastrophe loomed—none of this was unique to the United States. In examining the meaning and significance of the intellectual shift from classical to welfare liberalism in the United States, and of the centralization of power within the American federal order, we should try to control for the general trends in the world at the time. We should certainly not simply equate welfare liberalism with the presidencies of, for example, Woodrow Wilson and Franklin Roosevelt. These were less market-oriented than earlier American practice, but they were also less liberal by any meaning of that word, just as governing orders around the world were generally less liberal in the period from the 1910s to the 1940s than they had been for the fifty years before that.

What is relatively distinctive about the American case is that the constitutional system recognizably survived these decades; it was transformed and perhaps weakened but never completely ruptured either by internal collapse or by external conquest. If this was also true of the other English-speaking democracies (Britain, Canada, Australia, New Zealand) and of Switzerland, it was not true much of anywhere else.

The distinctiveness of the American case is all the more striking when one remembers that Britain and New Zealand had no written constitutions at all, and that they and Canada and Australia operated under Westminster systems that divided authority and power much less than the American system. None of the four had a judicially enforceable bill of rights; only Australia had robust bicameralism; and, of course, none had an independently elected executive who might face an uncooperative legislature. These features of the American constitutional order came under real pressure during the long crisis.

But none of these features was actually abolished. Supreme Court Justice Harlan Stone, in the 1938 Carolene Products case,³⁵

established the rationale for late-twentieth-century judicial activism in defense of noneconomic liberties just a year after the Supreme Court's surrender on economic questions in the face of Roosevelt's open threat to the Court's independence.³⁶ The potential of Stone's theory was not to be made use of for some time—Korematsu lay in the near future, Brown in the far future, and Stone's idea was rarely cited before 1950.³⁷ Still, the Court's willingness to stake a claim for its institutional prerogatives so soon after the climb-down of 1937 is striking. So too is the Court's willingness to directly oppose presidential emergency powers during wartime in the 1952 Steel Seizure case.³⁸

Moreover, the corporatist, presidentialist, anti-constitutionalist views of the Roosevelt administration did not unambiguously carry the day even in economic policy. As the legal historian William Forbath has shown, the survival of common-law legal liberalism, of federalism, and of constitutional formalities restrained the ambitions of the New Deal reformers by comparison with the latter's British social democratic counterparts. Forbath suggests that

And if that lawyerly liberalism could survive the long crisis and even exercise some restraint on the reactions to that crisis, it could reemerge after World War II as the core of a new approach to federal-state relations that, unlike the New Deal, centralized power for liberal reasons.