In 1925, H.W. Horwill, an English scholar, wrote a neglected classic on The Usages of the American Constitution. In an attempt to sort out what counts as “constitutional law” in the United States, Horwill identified a category of constitutional statutes, giving as an example the Electoral Count Act of 1887, which provides rules for counting votes cast for presidential candidates in the electoral college. Horwill went on to quote a passage from the American political scientist and constitutional scholar Charles Beard, in which Beard noted that statutes form a large part of constitutional law. “If we regard as constitutional all that body of law relative to the fundamental organization of the three branches of the federal government–legislative,” Beard wrote, “executive and judicial—then by far the greater part of our constitutional law is to be found in the statutes.”
William Eskridge and John Ferejohn have now dilated Beard’s point into a book of nearly six hundred pages. They claim that the United States is a republic of “superstatutes,” which in some sense possess constitutional importance. Eskridge and Ferejohn know a great deal about America’s major statutes, and have put it all between two covers. The book’s ambition is clear, but in many respects it leaves the subject of constitutional statutes murkier than it was before.
Eskridge and Ferejohn start with the indisputable observation that the large-c Constitution of 1789, which can be seen under glass in the National Archives, is only a part—and arguably not a large part—of America’s “working constitution.” The small-c constitution includes a large class of statutes that, they claim, are best understood as constitutional superstatutes. They describe the superstatutory constitution in nine overstuffed chapters, each of which studies statutes, judicial precedents, and history in a given area of public policy. Their list is heterogeneous. It encompasses the “constitution of equality,” exemplified not only by the Civil Rights Act, but also by the Pregnancy Discrimination Act and the Family and Medical Leave Act; the “democratic constitution,” exemplified by the Voting Rights Act; the “constitution of the market,” exemplified by the antitrust rules courts and agencies have elaborated under cover of the Sherman Act; the “constitution of the family,” exemplified by a hodgepodge of state laws regulating marriage, sex and property; the “green constitution,” exemplified by the Endangered Species Act and the Clean Water Act; the “monetary constitution,” exemplified by the Federal Reserve Act and other statutes establishing the federal system of financial and monetary policy; the “antihomosexual constitution,” now partially dismantled but shored up by the Defense of Marriage Act; and the “national security constitution,” transformed in important ways after 9/11.
How do these statutes become super? Eskridge and Ferejohn posit a causal sequence. A political entrepreneur puts a problem on the public agenda, in many cases by mobilizing a popular movement that grabs the attention of incumbent politicians. Opponents predict disastrous consequences or condemn the proposal as inconsistent with “core national commitments.” A process of political argument unfolds, eventually producing a statute whose supporters are temporarily riding high in the saddle. But the statute becomes super only over the course of subsequent years, if confirmed and expanded by further rounds of political action. New legislation that reaffirms the statute’s central principles, or (even more likely) administrative interpretation monitored by judges and congressional committees, may effectively cement the statute in the working constitution.
The most obvious puzzle for this account is which statutes count as small-c constitutional. This is the sort of definitional question that matters. A great deal, both in law and in the broader political culture, turns on how the boundaries around the constitution are drawn. An example is Eskridge and Ferejohn’s very plausible claim that courts interpret quasi-constitutional superstatutes more expansively than ordinary statutes. The problem is that there are multiple, competing criteria for sorting statutes into or out of the working constitution. Assemble a hundred legal scholars and ask them if there are superstatutes, and the vote will be near-unanimous that there are. Ask them for a list, and disagreement will rapidly emerge. A handful of statutes will appear on most lists—say, the Administrative Procedure Act of 1946, which is in many respects the fundamental framework for the modern American regulatory state, yet which Eskridge and Ferejohn oddly mention only in passing—but the lists will differ, and the criteria for inclusion are often ill-defined. If the Pregnancy Discrimination Act, important though it may be, counts as a constitutional superstatute, what doesn’t count? Aren’t there ordinary statutes that have no constitutional status at all? Conversely, why do Eskridge & Ferejohn omit or give only the briefest mention to statutes that seem arguably at least as super as the ones they discuss—statutes such as the Securities Acts passed in 1933 and 1934, or the federal land-management statutes that create the vast system of national parks and wilderness areas?
Beard’s formulation, under which statutes count as “constitutional” only if they regulate the “fundamental organization” of the government and its lawmaking powers, embodies a classical criterion for sorting constitutional from ordinary laws, one that derives ultimately from Aristotle. Constitutional provisions are those that in some way or other organize lawmaking institutions and thus set second-order rules of the political game, rather than establishing first-order policies such as a bar against pregnancy-based discrimination. As the English legal theorist Albert Venn Dicey put it, constitutional law comprises “all rules which directly or indirectly affect the distribution or the exercise of the sovereign power in the state.” Such provisions need not, of course, be large-c Constitutional. They may also be found in statutes, in internal rules of legislative procedure, in the common law, in unwritten norms, conventions, and usages (such as the norm that restricted presidents before Franklin Roosevelt to two terms), and conceivably in administrative regulations.
Confusingly, Eskridge and Ferejohn at one point plump for the classical criterion, but their collection of putative superstatutes implicitly rejects that criterion. Although a few of their statutes do set political ground rules, much of their list consists of statutes that do nothing more than set first-order policies. In the main line of their discussion, accordingly, they invoke a different criterion: superstatutes embody fundamental values or national commitments that are “entrenched” against repeal or significant modification. This entrenchment arises through “deliberation,” defined as collective practical reasoning about what to do.
But entrenchment is a notoriously slippery and troublesome criterion of constitutional status. Entrenchment may be interpreted in either a de jure or de facto sense; and the problem is that the first sense is too restrictive, while the second sense is too capacious and ill-defined. The de jure sense is too restrictive because it counts little as constitutional other than the Constitution itself. The Constitution is formally entrenched by Article 5, which requires large supermajorities for amendment; by contrast, none of the statutes Eskridge and Ferejohn discuss are entrenched de jure. Legally speaking, the Voting Rights Act, the Administrative Procedure Act, even the Pregnancy Discrimination Act, could be repealed through exactly the same procedures and with exactly the same legislative majorities requisite to enact the statute in the first place. There is even a standard view that one Congress cannot formally entrench its laws by requiring a later Congress to muster a supermajority for repeal.
On the other hand, if entrenchment is interpreted in a de facto sense, it becomes unclear which statutes are not entrenched. All statutes are somewhat harder to repeal than they were to enact, in the practical sense that the inertial bias of the legislative process—especially the elaborate lawmaking process set up by the Constitution—protects the status quo and thus advantages the political forces playing defense. Statutes may even defend themselves by creating political constituencies that then block their repeal. Daryl Levinson has offered the home interest mortgage deduction as an example of a statutory rule that entrenches itself politically in this way. The force of the example is that whatever its merits as a matter of tax policy, no sensible theory will count the home mortgage interest deduction as “constitutional.”
Levinson’s larger point is that the move to a de facto interpretation of entrenchment makes it precious hard to distinguish a statute that survives because it is somehow resistant to change from a statute that survives just because people today like it and have no desire to change it. In another of his examples, statutes prohibiting first-degree murder have been around longer than the Voting Rights Act and are even more likely than the Voting Rights Act to still be around a century from now. The reason, however, is not that such statutes are super or can be described as “constitutional” in any interesting or useful way. It is just that almost everyone wants there to be statutes against murder, so there is not and never will be a majority to repeal them. At a minimum, statutes that are de facto entrenched and statutes that rest on the support of (large) current majorities will be observationally equivalent in many cases. Eskridge and Ferejohn merely exacerbate the problem when they say, in varying formulations, that superstatutes are “entrenched by deliberation” and also enjoy “normative weight derived from social consensus.” If “social consensus” just means the support of a current majority or supermajority, in what sense is the statute entrenched? The acid test of entrenchment occurs when a statute survives despite the opposition of a current majority or supermajority.
One way forward for the entrenchment approach would be to offer more fine-grained social-scientific mechanisms that explain why statutes enacted by a transient majority might then become, de facto, harder to repeal than they were to enact, even putting aside the inertia of the lawmaking system and the defensive advantages that statutory beneficiaries enjoy. The simplest idea is that if there is a large coordination component to the statute, akin to the collective decision whether to drive on the left or on the right, then it may be in no one’s interest to undo the statute once it is enacted. Even if there were sharp conflicts over which of several possible rules to enact, meaning that the statute not only coordinates different groups but also distributes different payoffs to different groups, the statute’s initial opponents may nonetheless prefer that the polity stick with its coordination on whatever rule has been enacted, rather than reopen the issue and descend into a welter of disagreement. Relatedly, a statutory regime that restricts the power of current political majorities might rest on cooperation between two opposing groups, extended over time. If each group benefits more from cooperating with the other over the long run than by attempting to exploit the other group when temporarily dominant, a statute may survive despite being the first choice of neither group.
These ideas take political preferences as given. A richer starting point is to observe that political preferences are partially endogenous. Sometimes the passage of the statute at a given time itself shapes the preferences of political actors at a later time, so that the statute creates and confirms its own support. Eskridge and Ferejohn may be gesturing in this direction by emphasizing the public deliberation that accompanies a statute’s passage. Perhaps such deliberation provides information that shapes causal beliefs and thus changes people’s derived preferences over policies; perhaps deliberation even changes bedrock preferences, although how it might do so is poorly understood.
Beliefs can matter as much as preferences. Statutes might become de facto entrenched under conditions of “pluralistic ignorance,” in which people are uncertain of the preferences or beliefs of others and thus falsify or merely censor their own preferences. Conceivably, a current majority might desire the repeal of, say, the Voting Rights Act, yet the members of that majority might falsify or censor their views because they erroneously believe that most others do not share those views, in which case the latent majority for repeal may never coalesce. One might also appeal to psychological mechanisms that cause legal rules to generate, within individuals, an internalized sense of normative obligation, even a kind of sacred aura. “The normative power of the factual” implies that whatever statutes happen to exist may be taken to have a legitimate title to continued existence, although the conditions under which the factual becomes normative are ill-understood. Some statutes might even benefit from a kind of cognitive hegemony, such that no one even considers whether they should be repealed or not; they are just taken to be a fixed feature of the legal and political landscape.
Whether or not such mechanisms could be worked out and then tested against the facts, they have not been worked out here. Eskridge’s and Ferejohn’s treatment of the statutory constitution is so capacious, the boundaries of their enterprise so ill-defined, that it threatens to swallow up all of ordinary politics. In many of their chapters, Eskridge and Ferejohn basically describe the give-and-take struggle of lawmaking and regulation on more or less humdrum subjects—family and medical leave, antitrust, clean water laws—and then attach the label “constitutional” to the results. But if everything is constitutional, nothing is. The nagging merit of the book is that there is, clearly, some category of superstatutes with more than ordinary force and stature; but after reading the book, the nature and boundaries of that category are all the more opaque.
Adrian Vermeule is John H. Watson Professor of Law at Harvard Law School.