Legislative entrenchment: a reappraisal.
There is a principle of constitutional law holding that "one legislature may not bind the legislative authority of its successors." (1) The Supreme Court recently discussed that principle at length in United States v. Winstar, and although the case was decided on other grounds, it is clear that the Court sees the principle as a constitutional axiom. (2) When cashed out in terms of constitutional doctrine, the principle means that legislatures may not enact entrenching statutes or entrenching rules: statutes or rules that bind the exercise of legislative power, by a subsequent legislature, over the subject matter of the entrenching provision. Judges have applied this rule of constitutional law in various settings, (3) and the academic literature takes the rule as given, universally assuming that legislative entrenchment is constitutionally or normatively objectionable. (4) The goal of the academic literature has been to supply the definitive rationale for the rule, although the theorists' favorite rationales are all different. (5)
Our claim is that the rule barring legislative entrenchment should be discarded; legislatures should be allowed to bind their successors, subject to any independent constitutional limits in force. The rule has no deep justification in constitutional text and structure, political norms of representation and deliberation, efficiency, or any other source. There just is no rationale to be found; the academics have been on a fruitless quest. Entrenchment is no more objectionable in terms of constitutional, political, or economic theory than are sunset clauses, conditional legislation and delegation, the creation, modification, and abolition of administrative agencies, or any of the myriad of other policy instruments that legislatures use to shape the legal and institutional environment of future legislation.
In Part I, we define our terms, rebut the view that entrenchment is conceptually impossible, and argue that entrenchment is both constitutionally permissible and, in appropriate circumstances, normatively attractive. In Part II, we apply our analysis to a wide range of entrenchment-related problems, including the validity of the Senate cloture rules, the Gramm-Rudman law, legislatively enacted canons of statutory interpretation, statutes that regulate internal congressional procedures, government contracts, treaties, and entrenchment within the executive and judicial branches. Part III is a brief conclusion.
"Entrenchment" is a promiscuous word in the academic literature. We have to ask: entrenchment of what, as against whom, by what means? We ignore one use of the term, favored by process theorists such as Klarman, (6) to mean entrenchment of officials against challengers, such as prospective candidates, by means of rigged electoral rules, restrictions on political speech, and so forth. By "entrenchment," then, we mean the enactment of either statutes or internal legislative rules that are binding against subsequent legislative action in the same form. As we will see, there may be constitutional arguments for distinguishing statutes from rules in particular contexts, but there is no entrenchment-related reason for doing so. Accordingly, we will refer to "entrenching statutes" except when we discuss the collateral constitutional distinctions between statutes and rules.
On our definition, an ordinary law has some propositional content P--no bicycles in the park, for example. An entrenching statute has this propositional content plus an additional provision R which governs the conditions under which the statute may be repealed or amended. For example, R might say that P cannot be repealed or amended with less than a two-thirds majority in both the House and the Senate. Thus, an entrenching statute might say: (P) no bicycles in the park; and (R) the prohibition on bicycles in the park cannot be repealed with less than a two-thirds majority.
In this example, the entrenchment is accomplished by prescribing voting rules that the subsequent legislature must use, but it is important to be clear that this is a contingent feature of the example. Commentators sometimes assume that legislative prescription of voting rules is coterminous with legislative entrenchment, but that is not so. The legislative prescription of voting rules may be objectionable on constitutional grounds unrelated to entrenchment--if, for example, the Constitution happens to mandate majority voting and the earlier legislature requires the later legislature to act by a two-thirds majority. Conversely, the earlier legislature may entrench not a supermajority voting rule, but a substantive policy. For real-world examples, consider the federal statute at issue in Reichelderfer v. Quinn, which "perpetually dedicated" certain public lands in the capital for use as Rock Creek Park, (7) or the Ohio statute at issue in Newton v. Commissioners, which "permanently established" the town of Canfield as the seat of Mahoning County. (8) In the first case, the Court discussed the entrenchment issue but avoided it by aggressive statutory interpretation, (9) and in the second, the Court reached the constitutional merits and found the entrenchment invalid; (10) in neither case was the entrenchment accomplished by prescribing voting rules. A hypothetical analogue might be an addendum to the Endangered Species Act (11) providing that "this statute (including this provision) may never be repealed (even by a unanimous vote)."
The rule against legislative entrenchment has a corollary, or a reformulation, that is usually called the "last-in-time rule" for statutes. (12) The last-in-time rule addresses an intertemporal choice-of-law problem: It says that if a statute enacted at Time 1 squarely contradicts a statute enacted by the same legislature at Time 2, after reconciliation through statutory interpretation has proved impossible, the later-enacted statute is the law. The qualifier about interpretation emphasizes that the last-in-time rule, like the equivalent anti-entrenchment rule, is a rule of constitutional law rather than an interpretive canon. In the regime that we argue for, by contrast, discarding the anti-entrenchment rule entails simply that the earlier legislature itself decides the intertemporal choice-of-law question: Whether the later-enacted statute governs in the case of a conflict depends on what the earlier legislature has provided. As subsequently discussed, however, our position is quite compatible with an interpretive presumption against legislative entrenchment in the form of a default rule holding that the later-enacted statute governs if the earlier-enacted statute is silent on the entrenchment question.
B. Is Entrenchment Possible?
Entrenching statutes pose two puzzles, one conceptual and the other normative. The conceptual puzzle is whether entrenching statutes are possible. If they are not, the normative puzzle does not need to be addressed, so we start with the conceptual puzzle.
Consider statute PR, in which P prohibits bicycles in the park, and R prohibits repeal with less than a two-thirds majority. The conceptual challenge to this statute is the claim that even if judges enforced statutes in a literal way and enforced earlier statutes as fully as later statutes, PR would not bind a simple majority of a subsequent Congress. If the majority believes that a statute PP that permitted bicycles in the park would violate PR, that majority could first repeal the statute PR, enabling itself to repeal P (that is, enact PP) without a supermajority. Indeed, one might expect that any statute PP that is squarely inconsistent with PR would work an implied repeal of R, in which case PR would not in any way bind a future Congress.
The original Congress could pass an additional entrenching provision, R', which provides that R can be repealed only with a two-thirds majority, but then of course the next Congress could repeal R' with a simple majority, and so on down the line. But ordinary language can handle the infinite regress. Let the original Congress enact [R.sup.*], which says that a two-thirds majority is necessary to repeal or amend both P and [R.sup.*]. The statute P[R.sup.*] is invulnerable to repeal. Self-reference solves the problem of infinite regress. (13)
A more serious challenge to entrenchment is the possibility of circumventing statutes without referring to them explicitly. (14) Consider the following alternatives to PP: (1) Bicycles are two-wheeled vehicles manufactured before 1900; two-wheeled vehicles manufactured after 1900 are shmicycles (not governed by P); (2) police officers who give tickets to bicycle riders in the park will be fined $1000; or, more simply, park police officers are hereby ordered to focus their efforts on littering and to ignore bicyclists; (3) people who are ticketed and fined for riding bicycles in the park are entitled to a ten percent reduction in their property tax; or (4) we hereby announce that the Jane Doe Park has been closed and reopened as the John Smith Recreational Area (P refers to parks, not recreational areas). Each of these statutes achieves or nearly achieves the effect of PP without expressly contradicting PR.
Examples could be multiplied, but we do not find them as troublesome as other authors do. Legal actors constantly must make judgments about whether a statute conflicts with a previous or hierarchically superior enactment. When an interpreter such as a court or legislative body decides whether a federal law preempts a state law, whether a federal or state law conflicts with the Constitution, or whether a transaction violates the tax law, it must be able to identify real conflicts that are concealed by statutory (or transactional) indirection. This task involves a well-known problem of interpretive theory, variously labeled as a "form and substance" problem, a "rules and standards" problem, a problem of the choice between "textualism" and "purposivism," or a problem of "circumvention" or "evasion": If a statute in terms prohibits an action A, and someone undertakes a formally distinct action [A.sup.*], when is the similarity of purpose or effect between A and [A.sup.*] sufficiently close to warrant the interpreter's decision to treat [A.sup.*] as if it were A? There is nothing unique to entrenchment in this regard. The possibility of circumvention does not supply an objection to entrenching statutes in particular; if that possibility proved the conceptual impossibility of entrenchment, it would also prove the conceptual impossibility of binding future decisionmakers by means of ordinary constitutional and statutory provisions. Entrenchment is possible if constitutions, statutes, and contracts can bind people over time. This assumption is sufficiently uncontroversial to justify our analysis.
Note also that none of these arguments necessarily entails that judicial enforcement of legislative entrenchments against later legislatures is desirable in addition to or in place of legislative enforcement. Just as in the ordinary constitutional context, judicial review is a particular institutional enforcement mechanism that might or might not be thought desirable and that must be justified, if at all, by separate arguments. We take no position here on whether courts should enforce entrenched statutes when subsequent legislatures violate the entrenchment by enacting a contrary statute. Our argument is simply that the subsequent legislature is bound by the entrenchment, and that any contrary statutes are straightforwardly illegal (assuming, as always, that the entrenched statute is not otherwise unconstitutional). Unless one subscribes to the discredited Holmesian position that law is to be equated with what judges do (15)--that it is meaningless to speak of binding law apart from judicial enforcement--then the arguments about the legality of entrenchment are analytically distinct from arguments about the justiciability of entrenchment.
C. Reasons for Entrenchment
We have encountered the view that entrenching statutes cannot possibly have any policy value, and therefore they either should be unconstitutional for that reason or should not be a topic of academic debate, for no reasonable legislature would want to enact them. To forestall these objections, we briefly discuss some of the advantages of entrenching statutes. All of these advantages are familiar from the literature on constitutionalism, for an entrenching statute is like a mini-constitution in its self-conscious effort to control the voting practices or policy choices of future majorities.
(1) Government commitment. Entrenchment enables a government to make a credible commitment that it will not hold up a person (or firm or institution or country) from whom it seeks certain actions, and thus entrenchment makes it easier and cheaper for the government to control its relations with other entities. The simplest example is the government debt contract: A creditor might charge a government a lower interest rate if it knows that a future government cannot repudiate the contract without a supermajority vote. Recent welfare reforms similarly depend on the government being able to commit itself to withhold transfers from people who fail to obtain employment within the designated period of time.
(2) Within-government commitment. Entrenchment smooths interactions among political actors within the government by enabling them to make commitments to each other. The base-closing process is an example: If all members of Congress agree that bases need to be closed and that the military or an independent commission ought to decide which bases, but each member expects to succumb to ex post political pressure from constituents, all might be better off if they can commit themselves ex ante. Base-closing, like Gramm-Rudman, (16) was a weak form of entrenchment: It did not prohibit repeal by a majority, but it did try to raise the costs of such a repeal by institutionalizing the task at hand in an agency outside Congress. Cloture and other rules in the Senate as well as the House are also examples of within-government commitment.
(3) Agenda control. Entrenchment permits politicians to remove contentious issues from the agenda while they focus on other business. If, to take an extreme example in the other direction, statutes regularly expired at the end of the legislative session, members of Congress at the beginning of the new session would have the opportunity to challenge earlier legislative agreements that would otherwise be off the table. As Madison pointed out, as the date of termination approaches, "all the rights depending on positive laws, that is, most of the rights of property, would become absolutely defunct, and the most violent struggles ensue between the parties interested in reviving, and those interested in reforming the antecedent state of property." (17) If this explanation for why laws should not expire on their own is correct, then it follows that Congress, anticipating such a battle if a majority but not a supermajority has the power to repeal a statute, could have the same reasons for preferring an entrenchment. A Madisonian conflict occurs not only when statutes expire, but also when they are up for grabs because they command the support of less than a majority. (18)
(4) Controlling the temporal effects of existing legislation. All statutes have effects far in the future, and entrenchment powers give Congress a more refined tool for controlling those effects, even in a way that enhances the power of future majorities. Suppose Congress wants to send a signal to foreign adversaries by committing itself to build a stronger military. One such signal could be a large military budget accompanied by a rule that it cannot be reduced in the future without a two-thirds majority. Another such signal could be stationing troops in foreign countries with the knowledge that removing them would be politically costly for future governments. Future majorities might not want to be bound to the large military budget, but prefer this constraint to the stationing of troops. If the current majority is committed to one signal or the other, entrenchment enables it to choose the first when the first is preferred by both present and future. Entrenchment in this way smooths out the lumpiness that otherwise exists among legislative options for influencing the future.
(5) Predictability. Many political institutions are celebrated for their effect on the stability of government: Constitutionalism, stare decisis, representative government, and so forth are said to make government more predictable, and this makes it easier for individuals to arrange their affairs. It is always immediately pointed out in response that too much stability is a bad thing, that government should change its policies when circumstances change. The best government reflects a balance of these competing concerns. What is rarely pointed out is that ordinary statutes already balance these concerns in a particular way. If statutes expired by themselves at the end of the legislative term, government would be more flexible but less stable. If statutes were immune to repeal, government would be less flexible but more stable. The "default"--that statutes persist until repealed--creates a compromise between stability and flexibility, but this balance is more appropriate for some policy areas than others. Indeed, Congress recognizes as much when it provides certain statutes with sunset provisions, reflecting the view that greater flexibility than the norm is needed in that policy area. When greater stability is needed in a particular area, entrenchment is the appropriate response. The power to entrench gives Congress more freedom to incorporate in law fine-grained judgments about the correct balance between flexibility and stability.
(6) Adjusting voting rules across policy areas. Theoretical work on voting rules shows that majority rule is not always the best decision procedure. The optimal procedure--majority, supermajority, or unanimity--turns on a balance between the cost of legislative negotiation, on the one hand, and the potential for exploitation of minority interests, on the other. (19) Because the second kind of cost is likely to vary across policies, the optimal voting rule will also vary. Congress might seek to implement different voting rules or procedures for different policy areas, but if it is not permitted to entrench the rules, they can be repealed by a simple majority, frustrating all such efforts.
D. Conventional Objections to Entrenchment
Academics object to legislative entrenchment on the ground that it gives the past too much power over the present. Consider our statute P[R.sup.*]. The entrenchment prevents a simple majority in a future Congress from permitting bicycles in the park. A number of reasons have been advanced to explain the offensiveness of this practice. None is persuasive.
Below, we consider objections to legislative entrenchment rooted in formal constitutional argument from text, structure, and history; in democratic theory; and in public choice. We need not take a methodological stand on the relative weights of such considerations in constitutional argument because we hold that none of these considerations justifies an anti-entrenchment rule.
1. Constitutional Text, Structure, and History
We first consider a range of formal objections. We assume throughout that all entrenching statutes otherwise fall within the scope of Congress's affirmative authority to enact legislation under the enumerated powers set forth in Article I, Section 8, in the Necessary and Proper Clause, and in a miscellany of other constitutional provisions. (20) (We likewise assume that entrenching rules are authorized by each house's power to "[d]etermine the Rules of its Proceedings.") (21) If statutes are not authorized in this way they are unconstitutional, but for a perfectly ordinary reason unrelated to entrenchment. Given this assumption, the question is whether there is some special limitation, express or necessarily implied, that bars entrenching statutes. (22)
a. The Article I Vesting Clause
Modern commentators, such as McGinnis and Rappaport, have suggested that an anti-entrenchment rule derives from the "history and structure" of the Constitution, in particular, the "traditional understanding of the limits of legislative power." (23) It is not clear whether McGinnis and Rappaport mean to advance an originalist argument ungrounded in any particular constitutional text. Their reference to legislative power suggests instead that they mean to interpolate the claimed tradition into the text of the Article I Vesting Clause, which provides that "[a]ll legislative Powers herein granted shall be vested in a Congress of the United States." (24) For completeness, we examine in turn the Vesting Clause argument, an originalist variant, and a related Burkean argument that holds that the antiquity of the anti-entrenchment rule warrants its adoption by modern courts.
The initial puzzle of the Vesting Clause argument is that, on its face, the grant of "legislative Powers" to Congress simply does not speak to the question posed by entrenching statutes. The question those statutes pose is how the legislative power, whatever that power encompasses, is allocated over time to successive Congresses. The Article I Vesting Clause does not address that question. (25) It merely specifies that any particular Congress may exercise only legislative powers, as opposed to executive or judicial powers. No one argues that entrenching statutes, which are enacted by the ordinary Article I process, are best understood as examples of law execution or adjudication, powers constitutionally vested in other branches.
Commentators who root the anti-entrenchment prohibition in a definition of "legislative Powers" must defend a more ambitious reading of the Article I Vesting Clause. On that reading, the Clause supplies substantive limitations on what Congress may do, beyond the separation and distribution of powers across branches. The textual grounds for that more ambitious reading are, however, unpromising at best. Articles II and III vest the "executive Power" and the "judicial Power" in the President and the judiciary, respectively, (26) without any cross-reference to "Powers herein granted," (27) making it at least plausible to see those Vesting Clauses as empty vessels that must be given content by reference to implicit background norms. Article I, by contrast, does not vest otherwise undefined "legislative Powers" in the Congress, inviting an appeal to implied norms. Instead, it vests in the Congress "all legislative Powers herein granted," and then proceeds to supply both a careful enumeration of those powers in Article I, Section 8 and a detailed set of restrictions on those powers in Article I, Section 9.
The last point is important. Several of those restrictions, such as the ban on ex post facto laws, are also common-law maxims concerning the limits of legislative power, maxims whose explicit embodiment in constitutional text would have been unnecessary if they could simply have been treated as implied restrictions on the grant of legislative powers. Encoding those maxims as express rules of constitutional law suggests by implication that the anti-entrenchment rule lacks equivalent constitutional force. Conversely, the same English common-law jurists who condemned legislative entrenchment also identified other constitutive features of the legislative power that no American has ever thought to treat as implicit in the constitutional text: An example is Blackstone's rule that "[i]f a statute, that repeals another, is itself repealed afterwards, the first statute is hereby revived" (28)--a rule that Congress has specifically overturned by a partially entrenched statute, as we subsequently discuss. (29) So, reading into Article I the English common-law baseline--including the rule against entrenchment--simultaneously proves too much and too little. Article I's elaborate crafting of the metes and bounds of legislative authority counsels against finding additional, implicit restrictions on statutes that (by assumption) fall within one of the enumerated grants of power.
And even if it were fair play to read implicit limits into Article I, we would need a separate argument for reading in an anti-entrenchment prohibition in particular. The literature suggests two such arguments. The first, traceable to Blackstone, Hamilton, and Cooley, holds that each successive legislature has equal lawmaking authority. (30) A constitutional rule permitting legislative entrenchment violates the equal-authority postulate, the argument runs, because it confers on upstream legislatures the authority to enact a statute that downstream legislatures cannot repeal. But the appeal to "equality" of authority is indeterminate. A rule that allows each legislature to bind its successors also gives each legislature formally equal authority over time. It might be that under a regime permitting entrenchment, earlier legislatures will have greater de facto power than later legislatures. (31) But upstream legislatures always have greater de facto power than downstream ones, simply by virtue of drawing on a slate that is more nearly blank. They make policy choices that become entrenched de facto through path dependence and inertia.
Moreover, rooting the rule against entrenchment in the equal authority of successive legislatures is hard to square with Congress's undisputed authority to enact laws containing sunset clauses--clauses that cause a statute to lapse, by operation of law, after a defined period. Sunset clauses are the mirror image of entrenching clauses and might also be said to control the authority of later legislatures: An entrenching clause forbids the later legislature to prevent a statute from remaining in force by an affirmative repeal, while the sunset clause forbids the later legislature to allow a statute to remain in force by declining to repeal. We might say that the baselines are different because the later legislature has no constitutional authority to accomplish its goals by inaction, but the question-begging assumption in that argument is that the later legislature does have constitutional authority to accomplish its goals by affirmative enactment, thereby presupposing the invalidity of a contrary entrenchment. The upshot is that the anti-entrenchment position is forced to treat the parallel issues of entrenching clauses and sunset clauses in an oddly asymmetrical fashion without any (nonarbitrary) justification for doing so.
A second argument is that the sheer antiquity of the anti-entrenchment rule is a good constitutional reason for reading it into Article I. The maxim that no legislature may bind its successors is sometimes said to have deep roots in Anglo-American law; (32) on originalist premises we might take that maxim as a background assumption of the Founding generation, one that informs the meaning of "legislative Powers herein granted." But it turns out to be remarkably hard to find endorsements of the anti-entrenchment role, or any of its equivalents, in the canonical originalist sources. Hamilton assumed in The Federalist No. 78 that the last-in-time rule would govern federal statutes. (33) But he, like Blackstone, called it a "mere role of construction" applicable when the relevant statutes are silent about their relative priority, suggesting that he was not speaking to the constitutional question about intertemporal choice-of-law. (34) If Hamilton meant to invoke only a role of statutory interpretation then his analysis is consistent with ours, as discussed below. (35) On the other side of the originalist ledger, Madison himself recognized the validity of entrenching statutes by classifying political acts into three categories: (1) constitutions; (2) laws irrevocable at the will of the legislature; and (3) ordinary laws that are not irrevocable. (36)
The larger point, of course, is that these evidentiary fragments are irrelevant even on originalist premises. Even were some constitutional historian to discover unequivocal evidence that the Framers assumed entrenching statutes to be invalid, that evidence would demonstrate no more than a background assumption at the level of specific intentions, an assumption untethered to any particular constitutional text. This is among the weakest forms of originalist evidence, and, as previously discussed, it contradicts the express textual listing, in Article I, Section 9, of other common-law maxims as restrictions on the federal legislative power. For similar reasons, constitutional practice often refuses to interpolate the Framers' unarticulated assumptions into capacious constitutional texts, such as the grant of "legislative Powers." Consider that many members of the Founding generation assumed the existence of substantive natural law restrictions on legislative authority, restrictions that the Supreme Court has, by and large, declined to read into Article I.
A better version of the argument from tradition is not originalist but straightforwardly Burkean. Tradition, on the Burkean view, has constitutional weight independent of its embodiment in any text because epistemic humility counsels deference to the accumulated practice of past generations. (37) The accumulated practice of Parliament and of the Congress in disclaiming the authority to bind their successors, or in ignoring attempted bindings by earlier legislatures, has put an implicit restrictive gloss on the Constitution's grant of legislative powers.
There is good reason to doubt the premise of this argument. Although the historical literature contains no sustained modern treatment of entrenchment and related practices, making it hard to be confident in either direction, even a superficial inspection reveals a robust history of entrenchment and attempted entrenchment by Anglo-American legislatures. Dicey, writing in 1885, announced in sweeping terms that "[t]here is no law that Parliament cannot change," (38) but he had to spend a lot of time distinguishing contrary cases, because, as Dicey conceded, "[l]anguage has occasionally been used in Acts of Parliament which implies that one Parliament can make laws which cannot be touched by any subsequent Parliament." (39) Thus the fundamental treaties of Union with Scotland and Ireland, enacted by legislation in 1706 and 1800, respectively, were entrenched against repeal by later Parliaments. Dicey said that those framework statutes were violated or disregarded by legislation enacted in the mid- to late nineteenth century. (40) He thereby wanted to conclude that entrenching legislation cannot bind subsequent Parliaments, but we might more plausibly see these as examples of highly successful entrenchments, given that they had a lifespan longer than those of many constitutions. (41) In any event, those famous entrenching statutes formed part of the legal background for the Founding generation in a way that their nineteenth-century repudiations did not.
American legislative practice also displays a range of legislative entrenchments, successful and unsuccessful. We subsequently discuss the Senate's entrenched cloture rule--an old provision with roots in a variety of even older legislative practices, like the filibuster, which modify the rule of simple legislative majorities. Judicial decisions have enforced the rule against entrenchments where the issue was justiciable, (42) and those decisions are one strand of our tradition, but the entrenchments that gave rise to them are as well. The truth is that the place of entrenchment in Anglo-American practice appears not monolithic, but contestable and contested, both within legislatures and between legislatures and the judiciary. At the very least, the argument from tradition outruns the work of modern scholars, who assume far too confidently that the sweeping claims of older commentators like Blackstone and Dicey accurately represent the historical record in England and, even less plausibly, in America.
Whatever its historical merits, however, the conclusion of the argument from tradition does not follow. On the tradition argument, the anti-entrenchment prohibition arises principally from the (assumed) practice of early Parliaments and Congresses--more specifically the anti-practice by which early legislatures declined to enact entrenchments or declined to obey them if enacted. But the failure or refusal to enact entrenchments no more creates a constitutional gloss than does the failure to create, say, administrative agencies; similar reasoning would have condemned the New Deal out of hand. Even the stronger cases--affirmative legislative precedents or statements announcing an anti-entrenchment rule (43)--are double-edged swords for the tradition argument. It is extremely odd to defend an anti-entrenchment rule on the ground that the practice of earlier generations, particularly early legislatures, has created a constitutional gloss that binds succeeding generations. That is precisely what the anti-entrenchment rule forbids, and all the normative arguments for an anti-entrenchment rule are also normative arguments against taking the practice of earlier legislatures as conclusive on the validity of entrenchment. An argument that rejects the binding force of legislative entrenchment by appealing to the binding force of traditional legislative practice has a self-defeating air about it.
Those last points emphasize that an appeal to tradition is never a conclusive argument in American constitutional practice. Our meta-tradition, the only one we invariably adhere to, is to dump traditional practices overboard when their claims on our rational or normative allegiance wear too thin. Famous examples from adjudication are Erie (44) and Brown, (45) two landmarks of constitutional law that are landmarks in part because they resolutely jettisoned normatively unappealing traditions. The equivalent legislative landmarks are Congress's large-scale experiments with counter-traditional institutional arrangements during Reconstruction, the New Deal, and the Great Society. Such examples show that an appeal always lies from traditional practices to higher norms.
At most, tradition creates a presumption rebuttable by normative argument. Once normative arguments are in play, however, tradition drops out, and the debate's center of gravity shifts to the quality of the reasons adduced to attack or defend entrenchment. Perhaps for that reason none of the commentators who have tried to justify an anti-entrenchment rule has rested solely on the tradition ground; all have advanced broader claims rooted in constitutional and political theory. We take up those arguments subsequently, and the discussion there supplies our reasons for advocating that the anti-entrenchment rule, however traditional, be discarded.
b. The Article V Amendment Process
The argument based on the definition of "legislative Powers" fails to come to grips with the Constitution's text or structure. The Article I Vesting Clause functions only as a pseudo-textual vessel into which content may be poured by an appeal to the equality of succeeding legislatures, tradition, or some such norm. More convincing arguments might point to constitutional rules with more internal structure. Article V's rules for constitutional amendment, (46) for example, support an anti-entrenchment argument in the following form: Entrenched provisions are equivalent to constitutional amendments because they trump the decisions of later legislatures. Article V sets out the exclusive process for enacting amendments. Entrenching statutes, then, are constitutional amendments enacted outside the exclusive constitutional process for enacting amendments. They are thus invalid.
But even if Article V is the exclusive process for amendments, pace the contrary arguments of Bruce Ackerman and Akhil Amar, (47) there is a non sequitur here. Entrenching provisions are not amendments, nor are they "equivalent" to amendments in either a de jure or a de facto sense. Even in a legal regime that permitted entrenching statutes, Congress certainly could not pass a statute that purported to expand the constitutional powers of government or to repeal constitutionally mandated liberties, whereas a constitutional amendment could do either of those things. Entrenching statutes must, like ordinary statutes, be authorized by the Constitution; and, like ordinary statutes, they cannot violate independent constitutional restrictions. So entrenching statutes are just a unique legal instrument: They differ from ordinary statutes in that they bind later legislatures, and differ from constitutional amendments in that they cannot alter the constitutional rules themselves. The question is precisely whether the Constitution permits or forbids such an instrument. Pointing to Article V, which concerns a different instrument altogether, is not responsive to that question.
Indeed, the argument from Article V gets the structural argument against entrenching statutes precisely backwards. The Article V argument says that entrenching statutes are constitutional amendments, but defective ones, because they did not undergo the requisite constitutional process for enacting amendments. The only structural argument against entrenching statutes, however, is that (1) entrenched statutes bind subsequent legislatures; (2) only a constitutional rule can bind subsequent legislatures; (3) entrenched statutes are not constitutional rules; therefore (4) entrenchments are invalid. The Article V argument supposes that entrenching statutes are invalid because they are constitutional amendments, but should not be; the latter argument supposes them invalid because they are not, but should be. The pro-entrenchment position denies (2) above, holding instead that a constitutional provision is not the only legal instrument that may be used to bind later legislatures; entrenching statutes may be used as well, and nothing in the Constitution indicates otherwise. Article V, then, simply does not speak to the disagreement between the anti-entrenchment and pro-entrenchment positions.
The real relevance of Article V is that it supplies an analogy in support of the pro-entrenchment argument. Article V famously entrenched a handful of constitutional provisions against subsequent amendment; two such provisions subsequently lost their entrenched status by the terms of Article V itself, (48) but the provision that grants the states equal suffrage in the Senate remains entrenched to this day. (49) If constitutional framers may entrench constitutional provisions against later framers, why may not legislatures entrench statutory provisions against later legislatures? In originalist terms, the Article V entrenchment of equal state suffrage would have been meaningless, and would thus have failed to reassure the small states who desired the entrenchment, if there were a background understanding among the Founding generation that the entrenchment clause could simply be repealed through the ordinary process of constitutional amendment. And in structural terms, the permissibility of statutory entrenchment should follow a fortiori from the permissibility of constitutional entrenchment. After all, entrenched legislation, unlike an entrenched constitutional amendment, can at least be overturned by a constitutional amendment in the ordinary course. (50)
c. The Supremacy Clause
A similar anti-entrenchment argument might be made under the Supremacy Clause of Article VI. (51) Federal statutes, the argument would run, are hierarchically inferior only to the Constitution. To say that a later statute might be trumped by an earlier statute, if the earlier Congress so provided, is to say that the later statute might be trumped by a nonconstitutional source of law--a conclusion the Supremacy Clause forbids.
But this argument fails for the same reason as the Article V argument: The Supremacy Clause does not speak to the entrenchment problem and cannot be tortured into doing so. The Clause merely says that the "Laws of the United States" that are "made in Pursuance" of the Constitution are supreme law. It does not speak to the intertemporal choice-of-law question: As between an earlier entrenching law and a later inconsistent law, which prevails? Both statutes are (apart from the entrenchment problem) statutes enacted pursuant to the Constitution, so either answer to that question is consistent with the supremacy of (otherwise constitutional) federal laws. The target of the Supremacy Clause, of course, is not the intertemporal choice-of-law problem, but instead a problem of federalism: Valid federal statutes trump state law. The anti-entrenchment Supremacy Clause argument has nothing helpful to say about the validity of entrenchment at the federal level.
d. The Electoral Cycle Clauses
Eule grounds the rule against legislative entrenchment in an appeal to the "spirit" of Article I, Section 2, Clause 1, and Section 3, Clause 1. (52) The first provides for biennial elections for the House, the second for the (staggered) election of senators every six years. Eule's argument is that entrenchment violates the temporal limitations on any given legislature's authority that these Clauses embody. Insofar as this is a structural claim that rests on a principal-agent model of legislative authority, we discuss it below. Insofar as it is a textual claim, it is baffling. These Clauses simply establish the electoral cycle; for example, they would prevent a Congress from extending the terms of members. (53) They do not imply that Congress may not enact an entrenching statute. Indeed, if the specification of the electoral cycle had any implication for the temporal effects of statutes, the more logical conclusion would be that they prevent an ordinary statute from extending beyond the term in which it was enacted. The Clauses give no reason for distinguishing entrenching statutes from ordinary statutes.
e. The Rules of Proceedings Clause
For completeness we should mention the Rules of Proceedings Clause, which provides that "[e]ach House may determine the Rules of its Proceedings." (54) Although this is most obviously an affirmative source of cameral authority to enact self-governing rules--including entrenching rules--a creative textualist might read the Clause to contain an anti-entrenchment limitation as well. On this view, the grant of plenary rulemaking authority to "[e]ach House" (55) bars entrenchment by an earlier house of the legislature as against a later house. But, as Eule argues, the reference to "each House" is not a temporal limitation, but just a corollary of bicameralism. It establishes that each house separately, rather than the Congress as whole, may make rules for its respective internal affairs. (56)
f. Some Red Herrings
We have addressed only arguments from text, structure, and history that supply constitutional objections to legislative entrenchment as such. In many cases, arguments that claim to address the problem of legislative entrenchment, or that might be thought to have relevant implications, turn out instead to supply an independent constitutional objection to some collateral feature of an entrenching statute or rule. These are not general constitutional objections that support a general anti-entrenchment rule.
Consider, for example, the recent debate over the entrenchment of voting rules, provoked by the House's adoption of an internal rule that requires a three-fifths majority for tax rate increases. Scholars have attacked the House rule's constitutionality on the ground that legislatures may not require supermajorities in areas that the Constitution (implicitly) leaves to rule by simple legislative majorities. (57) Sometimes those scholars also invoke the anti-entrenchment norm, (58) but the two arguments are orthogonal to each other. We can imagine a perfectly coherent constitution that permits legislative entrenchments, but that prohibits legislative alteration of voting rules in particular. On this view, the structural argument that simple majority voting is the (implicit) mandatory rule for congressional voting, except where the constitution expressly says otherwise, is no more relevant to the entrenchment problem than is, say, the Free Speech Clause. That Clause bars contrary statutes and rules, whether or not the statutes contain entrenchment provisions. The same is true of Rubenfeld's textual argument that the House supermajority rule is inconsistent with Article I's grant of authority to Congress to "pass" bills, (59) which is no more relevant to the entrenchment question than would be a claim that the supermajority rule violates the equal protection of the laws.
In general, it should be clear both that Congress may not enact an entrenching law that purports to alter the terms of Article I, and that the defect of such a law is unrelated to the entrenchment problem. Article I mandates bicameralism and presentment: Bills must not only be "passed," but must also be (1) passed by both houses, and (2) either signed by the President or repassed by two-thirds majorities in both houses. (60) Any law that tinkers with any aspect of this procedure--say, by mandating that only one house may pass bills, or that a vetoed bill becomes law with a subsequent majority vote--is unconstitutional for precisely the same reason that a law abridging the freedom of speech is unconstitutional. (61) But that is true whether or not the law also contains an entrenchment provision; the two issues are unrelated.
When Article I is thought relevant to the entrenchment question, the analytic mistake is the failure to distinguish the constitutional rules governing statutory enactment from the constitutional rules that determine whether duly enacted statutes are binding law. A statute enacted in full conformity with the Article I procedures may nonetheless fail to be binding federal law. It may, for example, violate some independent provision of the Constitution, such as the Free Speech Clause. Thus, valid enactment is a necessary, but insufficient, condition for status as binding law; Article I precludes statutes not enacted in accordance with its terms but does not guarantee that validly enacted statutes have legal force. So the disagreement over entrenchment is a disagreement about what conditions, in addition to valid enactment, a statute must satisfy to count as binding law. The anti-entrenchment position holds that only a constitutional rule may trump a validly enacted statute. The pro-entrenchment position holds that a validly enacted statute may be trumped, not only by a constitutional provision, but also by an earlier validly enacted statute (if the earlier legislature chose to entrench that earlier statute). The Article I rules that determine whether a statute has been validly enacted have nothing to say about that disagreement.
A final confusion arises from the interplay between the constitutional problem of legislative entrenchment and associated problems of statutory interpretation. The last-in-time rule for statutes is, as we have said, a rule of constitutional law, not an interpretive doctrine. Where an earlier and a later statute are in irreconcilable conflict, the last-in-time rule decides the intertemporal choice-of-law problem in favor of the later enacted statute. By arguing that legislative entrenchment is unobjectionable we are also arguing that the last-in-time rule is mistaken. Our argument, however, entails only that an upstream Congress may, if it chooses, provide that a statute shall be entrenched as against later Congresses. Whether the upstream Congress has indeed chosen to do that is a conventional problem of statutory interpretation, to be resolved by whatever tools the relevant interpretive theory makes available.
Those tools might well include a presumption against legislative entrenchment, so that an earlier statute that is silent on the entrenchment issue would be taken not to entrench itself. (62) That default rule would just parallel the default rule used in the mirror-image case of sunset clauses. Congress has the undisputed authority to provide that statutes will lapse after a given period, but statutes that are silent on whether they are to remain in force indefinitely will be taken to remain in force indefinitely, rather than being interpreted to contain an implicit sunset clause. The general default rule, then, would just be that statutes remain in force unless and until overridden by a subsequent Congress, unless the earlier Congress explicitly provides to the contrary.
2. Simple (Time-Bound) Majoritarianism
The constitutional arguments we have been discussing obtain their force not from the text or other historical materials, but from unarticulated theories of democracy that are thought to be implicit in constitutional tradition. A few authors have brought these theories to the surface. Examination of them, however, reveals no grounds for believing that entrenchment is unconstitutional.
Dana and Koniak object to entrenchment on the ground that it violates what we call simple, time-bound majoritarianism. (63) On this view, the Constitution provides that when a majority in each house passes a bill, and the President does not veto it, the bill becomes law. Exceptions to simple majoritarianism are explicitly laid out in the Constitution, but when there is no question of overriding a veto, consenting to a treaty, and so forth, the default of simple majoritarianism prevails. An entrenching statute violates the rule of simple majoritarianism because it prevents a simple majority of some future Congress from creating law in the domain governed by the entrenchment.
Endorsement of the simple majoritarian view is surprising in light of the formidable array of objections coming out of social choice theory. The Condorcet paradox, for example, casts doubt on the premise that a simple majority can in normal circumstances even be identified. Under quite general conditions, voting will cycle rather than establish a majority victor. (64) The bills that are regularly enacted owe their existence to institutional structures and agenda control by party leaders as much as to majority rule. We do not doubt that some legislative processes are more legitimate than others, but we believe that entrenching procedure, in order to avoid cycling and disruptive parliamentary maneuvers, might contribute to legitimacy rather than undermine it. A legislative process that reliably serves the interests of the public will likely have majoritarian elements, and certainly democratic or representative elements, but there is no reason to believe that it would incorporate simple time-bound majoritarianism.
The more serious objection to simple majoritarianism arises from consideration of American constitutional practices. Recall our argument that the Constitution does not by its terms limit the power of Congress to bind itself, and indeed permits, or is conventionally understood to permit, Congress to enact legislation that extends indefinitely beyond the expiration of the legislative term. If there are political or logistical costs to repealing legislation--and there surely are--then an earlier Congress "binds" a later Congress by enacting legislation that cannot be costlessly repealed or changed, except in those instances when it provides for the legislation to expire on its own. Indeed, as we pointed out, one Congress would hardly do a favor to a later Congress by making all legislation expire at the end of the term, for this would impose on the subsequent Congress the burden of renegotiating and reenacting the expired legislation. Short of anticipating the needs and desires of future Congresses--which is impossible--a Congress will inevitably burden future Congresses, for the simple reason that the earlier Congress comes first and cannot avoid actions that will turn out to hinder the later Congress.
As this objection is more complex than it might appear, let us examine it in more depth. Recall PR*, which prohibits bicycle riding in the park and also repeal of the statute with less than a two-thirds majority, and compare it to a different hypothetical statute, PG, which provides that the smooth concrete paths in a park are to be replaced with gravel paths. A simple majority of the later Congress wants to allow bicycles in the park, but (PR*) it cannot achieve a supermajority, or (PG) the majority is not willing to appropriate funds to cement over the gravel paths.
Compare the two scenarios. In the first, there are no bicycles in the park because of the entrenchment statute (the paths are cement), and in the second there are no bicycles because of the cost of paving over gravel. We might agree without much thought that the majority's preferences are thwarted in the first case, but what do we say about the second? Do we say that the majority's preferences are vindicated because the majority does not want to pay to replace gravel with cement, or that they are thwarted because the earlier majority's stratagem of replacing cement with gravel (for no other purpose than to keep bicycles out of the park in the future) prevents the majority's goal of permitting bicycles in the park?
There is no good answer to the last question because when we talk about a majority's preferences, we usually take the past as given, and the majority's preference--to change the status quo, whatever it is--is either incorporated into a statute or not, depending on the nature of the relevant political institutions. We might complain about the political process if a majority's preference--to pave over the gravel, for example--is not incorporated into a statute; we blame the cloture rule, or the committee barons, or whatever. But we do not usually complain if a majority's preference--to pave over the gravel at no cost--cannot be vindicated because it is impossible.
PG and PR* both make a simple majority's preference for bicycles in the park impossible, and they do so by making it too costly, in financial or political terms, to create the conditions in which bicycling is possible. PG accomplishes this goal by altering the physical environment; PR* accomplishes this goal by altering the institutional environment. If PG is permitted, why not PR*?
One might argue that the difference is that installing gravel paths is a physical thing that cannot be costlessly undone, and entrenching is merely formal rather than real. But institutional structures are just as real as physical ones; both can be costly to change. We do not deny the existence of an analytic distinction between formal entrenching statutes and statutes that change the physical environment. We deny that this distinction reflects substantive policy concerns, and we suspect that those who want to distinguish PR* and PG assume that when legislatures make physical changes to the environment, such as the installation of a gravel path, they are presumptively serving the public interest. By contrast, an entrenching statute, it is asserted, is designed only, or mainly, to interfere with future majorities; that is why entrenching statutes should not be permitted. (65)
The problem with this view is that it ignores the many benefits of entrenchment, benefits every bit as real as the aesthetic benefits of gravel paths. We enumerated several of these benefits above: enabling a legislature to commit itself, to enact stable laws, and so forth. Entrenching laws, like ordinary laws, have benefits and disadvantages, and neither can be considered superior from the perspective of simple majoritarianism.
3. Agency Theory
Eule proposes a variant of simple majoritarianism, which he calls an "agency theory." According to this theory, the Constitution sets up an agency relationship between Congress and the people, and entrenchment violates this relationship. The key assumption is that Congress's agency extends only until the next election. Eule's view is similar to that of Dana and Koniak, except that he recognizes that the agency relationship is shaped by the nonmajoritarian provisions in the Constitution. (66)
Eule perceives the commercial analogy but fumbles it. In a corporation, the shareholders play the role of principal and the board plays the role of agent. As Eule recognizes, the corporate agent can frequently make decisions that affect the future--for example, it can commit the principal to a contract that extends long after the agent's expected or contractual date of departure, or, for that matter, buy some property on the principal's behalf, property that the principal will hold long after the agent departs. This is no different from the role of Congress. It, too, can purchase some property on behalf of the people, and the people will still have it long after the next election.
Eule claims that, although a corporate board can commit future boards to a contractual relationship, Congress cannot. (67) Why this difference? "The reason that the agent of limited duration [the corporate board] enjoys this power [to bind future boards] is ... because the principal intends it." (68) Eule acknowledges that shareholders want boards to have the power to bind future boards, but believes that the people do not want Congress to have the power to bind future Congresses.
Yet we know that this cannot be true; our simple bicycle example shows as much. Like a board, a Congress cannot accomplish the principal's goals if it cannot engage in actions that influence the future. If the people want Congress to have power--and the "people" surely want that, whether the people are the Founding generation or a current public--then they want Congress to be able to influence the future. There is no reason to believe that the people do not want Congress to enact entrenching legislation; or, if there is such a reason, Eule has not provided it, and we cannot think of it. Indeed, the corporate analogy suggests the opposite. If the importance of being able to influence the future justifies giving the corporate board the power to enter contracts that bind future boards, why would the importance of being able to influence the future not justify giving Congress the power to enact legislation that binds future Congresses?
There is a way of distinguishing the two cases. When the corporate board binds shareholders, it binds only those existing shareholders and not people in the future. People who buy shares in the future do so voluntarily and pay a price that incorporates the value of existing contracts. When Congress makes commitments for the future, it binds people not yet born, future immigrants, and others who cannot consent to the binding. But this distinction does no analytic work. The people of the future are affected by ordinary statutes--statutes to which they do not consent or in any way give democratic legitimacy--and this is seen as unobjectionable. As we discuss below, the fact that the future is in the hands of the present might be a source of concern or not, but it does not have special implications for entrenching statutes.
One might think that Eule's agency theory implies that statutes should expire at the end of the congressional term, a result that would contradict both tradition and common sense. His response--that future Congresses are free to repeal the earlier statute or otherwise acquiesce in it--is inconsistent with the agency theory, for it allows the early Congress to impose political burdens on Congresses in the future. But the real problem with Eule's argument is that it implies nothing about the power of Congresses over the future: One could argue that if statutes expired at the end of the term, that would be an unfair burden on future majorities for they would have to renegotiate and reenact the expired legislation--and would thus be inconsistent with the agency theory. The agency theory is too crude to explain the proper degree of congressional influence on the future; therefore, it cannot distinguish entrenching and ordinary statutes, both of which unavoidably have future effects.
4. Public Choice Theory
Some scholars argue that constitutional law does, or should, mitigate public choice problems--the influence of factions, as the Founders put it. Takings law, for example, interferes (or should interfere) with the ability of interest groups to lobby for property transfers that come at the expense of particular members of the public. (69)
A similar view might lie behind the Open Letter's fear that entrenchment of the tax rate would interfere with the power of future majorities. (70) The Open Letter does not explain why the Gingrich Congress would seek to restrict this power, and one could imagine explanations that do not draw on public choice. But we suspect that the theory behind the Open Letter was that tax policy was temporarily in the hands of powerful pro-business interest groups, and that these groups were attempting to influence legislation as far into the future as possible, at the expense of future majorities not yet organized. This kind of reasoning can also be found in Fischel and Sykes's critique of judicial enforcement of government contracts when those contracts are interest-group deals rather than transactions animated by concern for the public interest. (71)
The argument that a policy against entrenchment weakens interest groups is implausible. Entrenchment is just a legislative tool, no different from any other. Critics of entrenchment must show that good use of entrenchment is outweighed by abuse, and that entrenchment lends itself to abuse more than other legislative powers do.
One such argument might be that because an interest-group deal incorporated in an entrenching statute is worth more than an interest-group deal incorporated in an ordinary statute, lobbyists will pay more for an entrenching statute than for an ordinary statute. Thus, a constitutional ban on entrenchment would reduce rent-seeking. Because entrenching statutes place areas of legislative action off limits or nearly off limits, however, they reduce the incentive to lobby after they are enacted. Interest groups might pay more for an entrenching statute than for a single ordinary statute, but over time they might pay as much defending nonentrenched interest-group deals against legislative revision as they pay lobbying for entrenched interest-group deals in the first place. (72) The logic underlying the critique of entrenchment is perverse, for it implies that the public harm caused by a statute increases monotonically with the length of its period of effectiveness, and hence that all statutes should at least be subject to sunset provisions, and probably should not be permitted to exist in the first place.
A final argument against entrenchment is based on a kind of risk aversion. It might be thought that an entrenching statute is intrinsically more dangerous than ordinary statutes. Ordinary statutes that are bad can always be repealed; a bad entrenching statute cannot be repealed, or can be repealed only with difficulty. Thus, even if we think that entrenching statutes respond to the same mix of motivations as do ordinary statutes, we might think that courts should regard entrenching statutes with greater hostility, because the bad entrenching statutes can do much more harm than the bad ordinary statutes.
This argument is unpersuasive. Legislatures have immense powers and can do all the harm they want even without using an entrenchment statute. Such power is tolerated despite the dangers because legislatures are confronted with significant problems and are in the position to do considerable good. The risk aversion argument against entrenchment would justify all kinds of bizarre conclusions. Legislatures ought to regulate the mails--they can't do much harm here!--but they should not have the power to tax. After all, the power to tax is immense, and a legislature with bad motives could impose huge taxes (or bad taxes) that would destroy the economy. Although legislatures might sometimes use this power for good, the dangers when they do not are so great that on balance we should deny them the power to tax. The structure of this argument is identical to the anti-entrenchment argument, and equally perverse. People who fear extreme abuse through entrenchment are driven by a fear of democracy, not the institutional concerns that might justify giving legislatures some powers rather than others.
But suppose that Congress enacts a radical statute--no more tax increases, ever--and entrenches it with a supermajority, or even unanimity, provision. Future governments will be paralyzed, and all because--let us say--a temporary majority seized power during a brief, volatile period of history. This cannot possibly be a tolerable state of affairs. Thus we must oppose all efforts to entrench. Or so say the critics of entrenchment.
The problem with this argument is that the future is always in the hands of the present, and entrenchment is only one of many devices that the present can use to ruin the future if it so wishes. A crazy majority can do any number of radical things. It could give away the treasury to a foreign country. It could abolish the military. It could sell off federal lands on the cheap or authorize military adventures that ruin America's international reputation. It could shut down executive departments and, in the process, destroy expertise built up over decades. All of these activities would harm the future--just as limits on tax increases would--and yet we do not, and would not if we could, depend on constitutional prohibitions to prevent them from occurring.
Rather than trying to ruin the future, current majorities invest a great deal in maintaining institutions and natural resources for the benefit of the future. Congress will not tie the hands of the future (which will often be itself, in the near future) unless there is a strong consensus in the present, just as most radical (nonentrenching) statutes are based on widespread public sentiment. The parade of horribles is not on the horizon, and, even if it were, the appropriate response would be a radical restructuring of legislative power that might, but would not necessarily, impinge on the power to entrench. (For example, we might want to give the states or the federal judiciary more power rather than focus on entrenchment.) There are a large number of devices for affecting the future. Simple majorities have access to them. Entrenchment is just one more device, with its own particular advantages and disadvantages, and no one has explained why they are different enough, and sufficiently more dangerous, to warrant taking them out of the hands of current simple majorities. The parade of horribles provoked by thinking about entrenchment statutes is no different from the parade of horribles provoked by thinking about democracy in general. (73)
5. Deliberation and Information
Some theories of democracy emphasize the importance of deliberation in the political process. Laws do not (or should not) simply aggregate preferences; they should emerge from a deliberative process involving citizens and legislators, in which preferences change in response to argument and experience. Theorists emphasize different aspects of deliberation, some focusing on its instrumental value for good policy, others on its intrinsic value to public life. We discuss both below.
A theory of deliberative democracy might seem to be inconsistent with entrenchment. The problem with entrenchment from this perspective is that it withholds from the future the ability to deliberate about legislation concerning the domain of the entrenching statute. The future majority loses the ability to use information that it has gathered in the intervening years, and so cannot improve the entrenching statute where it fails to have desirable consequences. Further, the future majority loses the intrinsic benefits that come from deliberation.
These arguments, however, are subject to offsetting considerations. As we have observed, entrenching statutes can be beneficial, and the information costs must be weighed against the commitment benefits. It seems clear that these costs and benefits will vary from case to case, and there is no reason for a general hostility toward entrenching statutes per se. (74)
As for the intrinsic benefits of deliberation, the problem is not that they are lost so much as that they are enjoyed by the current generation rather than the future. But the future will be able to deliberate about other forms of legislation. There is no reason to believe that there is a fixed amount of legislation and that the current generation will "use it up," leaving the future nothing to do and nothing to deliberate over. There is certainly no historical experience suggesting that earlier legislatures try to deprive later legislatures of topics to deliberate over.
Theories of deliberative democracy are pitched at a level of generality that cannot shed light on the merits of entrenchment. Their authors are concerned about the impact of technocratic elites on democratic institutions and about defending democracy against rival political systems. Democratic theory is not sufficiently precise to shed light on such narrow institutional issues as the extent to which the branches of a democratic government should be able to entrench their policy choices.
Here we apply the analysis of Part I to a range of examples and problems. Among the examples are provisions that effect entrenchments, provisions that resemble entrenchments but really are not, and provisions that effect partial entrenchments. Some of the problems are the relationships between legislative entrenchment and government contracts, entrenching statutes and internal legislative rules, entrenching statutes and treaties, entrenchment and retroactivity, and legislative entrenchment and entrenchment within other branches. As Eule and others have noted, there are not many examples of entrenchments currently in force, but that is at least in part an effect of the rule (and the supporting academic consensus) that we are criticizing.
A. Easy Cases
We need not belabor the easiest cases under our theory: statutes that effect entrenchment in a straightforward fashion, either by "permanently establishing" a governmental structure or a substantive policy vis-a-vis subsequent legislative action, (75) or by altering voting rules or other procedural incidents of subsequent legislative action. In either case our position entails that these statutes are constitutionally objectionable, if at all, only by virtue of their content--perhaps the Constitution puts the substantive area off limits to legislative action, or establishes a mandatory voting rule--rather than by virtue of their entrenchment.
B. The Senate's Cloture Rule
A classic entrenchment is effected by Senate Rules V and XXII. The latter rule says that sixty votes are needed to effect cloture--the end of debate on an issue subject to filibuster (76)--while the former rule, employing self-reference, says that "[t]he rules of the Senate shall continue from one Congress to the next Congress unless they are changed as provided in these rules." (77) Their joint effect is to entrench the supermajority cloture rule against change by any simple majority in subsequent Senates, because a motion to change Rule XXII would itself be considered (by virtue of Rule V) in compliance with Rule XXII's supermajority cloture requirement; similarly, Rule V itself could not be first amended by a simple majority. Later Senates have in fact obeyed the entrenched rule. In 1975 a senatorial showdown on the issue resulted in a determination, now recognized as an internal precedent, that Rule XXII indeed bars a simple majority from closing debate. (78)
Is this entrenched cloture rule a good or a bad idea? It is hard to say, as even those who criticize the rule on entrenchment grounds agree. Certainly Burkeans should appreciate the rule's pedigree and venerability--the first congressional filibuster occurred in 1790, and the adoption of the supermajority cloture rule in 1917 represented not a departure from some preexisting majority-rule baseline, but instead a liberalizing departure from the earlier tradition, which required a unanimous vote to end debate. The effects of a regime of filibusters and supermajority cloture, relative to some majoritarian baseline, might be either to enhance or to undermine interest-group influence, successful legislative deliberation, preference aggregation, or other desiderata. Even the frequent complaint that the cloture rule is antimajoritarian is off target, both because the supermajority requirement for cloture is a rule about debate rather than a rule about the enactment of legislation, and because the filibuster may counteract other antimajoritarian elements in the Senate's structure, such as the committee system. (79)
In the face of these complexities, the leading academic critique of the cloture rule explicitly disavows simple majoritarianism, relying instead on the entrenchment objection. The entrenchment of cloture is said to undermine "[p]opular sovereignty," to diminish "legislative accountability," and to obstruct both "the legislature's inherent authority to adapt to current circumstances" and "the right of the electorate to rule according to its will." (80) The right response to that line of reasoning is not that the Senate is a "continuing body"--a bad metaphor made possible only by the Framers' decision, quite reasonable as a matter of institutional design, to stagger the Senate's turnover. The right response is that the anti-entrenchment objection to the cloture rule is really a wholesale objection to constitutionalism as such. In a binding constitutional order, neither the future legislative majority nor the underlying electorate has any general "right ... to rule according to its will." True, the constitutional restrictions come into force by a different procedure than do legislatively entrenched rules, but that is a different, narrower objection; and as previously discussed, it is also a question-begging objection, because it unjustifiably assumes that restrictions on any given legislature may derive only from the procedure for constitutional entrenchment, rather than from the procedure for enacting entrenching statutes or rules.
But this is to go over old ground; all these are just applications of the arguments from Part I. Here, as there, the anti-entrenchment position risks assuming a crude Jeffersonianism that proves too much. The position is inconsistent, not merely with legislative entrenchment, but with the acceptance of binding constitutions generally.
A contrast to the genuine entrenchment effected by the cloture rule is the pseudo-entrenchment effected by the Balanced Budget and Emergency Deficit Control Act of 1985, conventionally referred to as "Gramm-Rudman." (81) Although complex, and much modified by subsequent legislation that structures the federal budget process, in its original form Gramm-Rudman set federal deficit caps and required the Office of Management and Budget to sequester funds appropriated in excess of those caps. (82) Paul Kahn famously criticized Gramm-Rudman as an illegitimate attempt by one Congress to "constrain" its successors; (83) on this view, the partial invalidation of Gramm-Rudman on separation of powers grounds in Bowsher v. Synar (84) was a sideshow, even a distraction, from the statute's more basic flaw. But Kahn has it backwards: Gramm-Rudman contained no entrenchment; its political effects were perfectly commonplace, and the separation of powers challenge was the only colorable constitutional objection.
The brute fact, one that Kahn cannot quite get around, is that Gramm-Rudman did not entrench itself. A simple majority vote of any later Congress sufficed to raise the deficit caps or repeal them pro tanto, and in fact Congress has done just that on several occasions. Even if the structure of the statute were thought to effect an implicit entrenchment, Gramm-Rudman did not contain the self-reference that would immunize the entrenchment from simple repeal; nor did it even contain a rule of statutory interpretation, of the sort we examine later, to the effect that any repeal must be express rather than implied. (85)
Kahn says that Gramm-Rudman burdened subsequent Congresses de facto even if it did not bind them de jure; the statute's calculus, the argument runs, was that there may be a later majority willing to exceed the deficit caps but unwilling to incur the political costs of overriding Gramm-Rudman. This attempt to "change the effect of legislative inertia" reduces "accountability" and makes it more difficult for future majorities to enact their preferred substantive spending policies. (86) But ordinary nonentrenched legislation constantly does the same thing; Kahn's objection runs against all statutes. The Endangered Species Act imposes a political cost on later congressional majorities with different views about species protection--majorities who might be willing to allow or even promote animal takings and habitat destruction absent the Act but who are not willing to pay the price attendant upon repealing the Act to accomplish those purposes. The problem is that any statute changes the legal status quo and thereby shifts the burden of inertia from the enacting legislature to future legislatures, and might in that sense be said to reduce accountability and to frustrate the future majority's will. The "entrenchment" objection to Gramm-Rudman proves, if anything, that all statutes should lapse with the enacting Congress.
D. Statutory Rules of Statutory Interpretation
Intermediate between the genuine entrenchment effected by Senate Rules V and XXII on the one hand and the pseudo-entrenchment effected by Gramm-Rudman on the other are a handful of congressionally enacted rules that attempt to control the courts' interpretation of enactments by subsequent Congresses. The most familiar examples are the interpretive rules of the Dictionary Act (87) and scattered provisions in 1 U.S.C. (88) Consider 1 U.S.C. [section] 108, which says that "[w]henever an Act is repealed, which repealed a former Act, such former Act shall not thereby be revived, unless it shall be expressly so provided"; and [section] 109, which says that "[t]he repeal of any statute shall not have the effect to release or extinguish any penalty, forfeiture, or liability incurred under such statute, unless the repealing Act shall so expressly provide." (89) The most politically controversial of these rules is the Defense of Marriage Act, which defines "marriage" and "spouse" to exclude homosexual (and polygamous) unions, (90) but which is about as interesting for the theory of entrenchment as the provision that defines a "county" to include a parish. (91)
Entrenched interpretive rules that are more consequential appear elsewhere in the Code. The Administrative Procedure Act contains a clause providing that a "[s]ubsequent statute may not be held to supersede or modify this subchapter ... except to the extent that it does so expressly"; (92) while the National Emergencies Act provides even more forcefully that "[n]o [subsequent] law ... shall supersede this title unless it does so in specific terms, referring to this title, and declaring that the new law supersedes the provisions of this title." (93) There are similar examples in the Religious Freedom Restoration Act (94) and the McCarran-Ferguson Act. (95)
Such requirements of express repeal are partially entrenched because they override the ordinary interpretive canon that permits, but disfavors, repeals by implication. (96) Imagine a court confronted with a later statute that (on the prevailing interpretive theory) strongly but implicitly contradicted the earlier statute, which itself contained a requirement of express repeal; an example might be a later statute that appeared to extinguish a preexisting statutory liability, contrary to 1 U.S.C. [section] 109, but did not do so expressly. Under ordinary interpretive doctrine, absent the entrenching clause, the court would be highly reluctant to conclude that the later statute effects an implied partial repeal of the earlier statute's substantive provisions, but it would so conclude if the inference was inescapable. Requirements of express repeal forbid the court to take that last step, and the self-reference present in the entrenching provision prevents the court from holding that the entrenching provision was itself repealed by necessary implication. These provisions are not wholly entrenched, because they could be repealed by a simple majority, but unlike an ordinary statute they could not be repealed by a simple majority acting by implication.
For present purposes, the important constitutional point is that such rules are objectionable, if at all, on grounds unrelated to entrenchment. There are a host of good normative reasons for a legislature partially to entrench interpretive rules, reasons of the sort that support entrenchment generally. The precommitment story supporting the entrenchment of the National Emergencies Act is obvious, and the Congress that enacted the APA might justifiably have been concerned that future Congresses would turn the procedural framework for the administrative state into swiss cheese by enacting a grab-bag of implied partial repeals at the behest of future interest groups. In such cases, a partial entrenchment--one that prevents implied but not express repeals--is an eminently sensible technique for calibrating the tradeoffs inherent in entrenchment. Enacting an express repeal requires a degree of public visibility and legislative attention that would assuage the entrenching legislature's concerns about interest-group influence and decisional pathologies.
The relevant constitutional objection to such provisions sounds not in entrenchment but in the separation of powers. It might be argued that courts should have exclusive control of interpretive rules, either for the structural reason that the power to enact laws should be separated from the power to interpret them, or for the functional reason that judges and lawyers possess accumulated interpretive expertise superior to that of legislators. These arguments are vulnerable to serious objections, but the contestable issue concerns the relation between Congress and the courts, not the relation between one Congress and its successors.
E. Hybrid Entrenchments
Congress occasionally enacts a statute that establishes or regulates the internal legislative procedures that are to govern the enactment of future legislation--procedures that would otherwise be established or regulated by each house separately--pursuant to the Rules of Proceedings Clause. Consider the Alaska Natural Gas Transportation Act, (97) which barred consideration by either house of Congress of certain resolutions concerning energy policy, or, more interestingly, the recent Congressional Review Act, (98) which establishes special internal legislative procedures for disapproving proposed agency regulations. In these instances and all others of which we are aware, however, Congress inserts an anti-entrenchment proviso that subjects the statute to override by a subsequent internal legislative rule of either house, in the ordinary course. (99) Congress thereby avoids what we will call a "hybrid" entrenchment problem: Absent the avoiding proviso, Congress would be entrenching internal legislative rules by means of an earlier statute, not an earlier rule. The mirror-image case would be an internal legislative rule purporting to entrench a procedure against subsequent statutory change, but we are aware of no real-world examples.
On our analysis, it is straightforward that Congress ought not to shy away from enacting hybrid entrenchments if the ground for its concern is that entrenchment is constitutionally objectionable. But the provisos inserted under current practice are plausibly defensible on other grounds. Statutes that regulate the internal procedure of both houses may undermine bicameralism and the Rules of Proceedings Clause by forcing each house to share power over its own procedures with the other. They may undermine the separation of powers as well, by giving the President a share of power over internal legislative procedures, either ex ante or ex post, through threatened or actual vetoes of statutes that create or repeal hybrid entrenchments. These constitutional claims are sketchy, but we need only gesture in their direction, for they serve only to emphasize that the longstanding congressional reluctance to enact hybrid entrenchments is justifiable, if at all, on constitutional grounds unrelated to the validity of legislative entrenchment.
F. Government Contracts
A government contract, like an entrenching statute, imposes a cost--albeit fiscal rather than political--on future legislatures that seek to escape the consequences of the earlier action. To avoid performing the contract, the future legislature must pay damages. To avoid complying with the entrenching statute, the future legislature must achieve a supermajority vote, if the entrenching statute so permits.
From a doctrinal perspective, the two kinds of government action are distinguishable. Breach of a government contract falls under the Takings Clause (100) or the Contracts Clause. (101) The contract right is said to be a property right, so the breach is a taking of the property right. If there were a valid constitutional argument against entrenchment, it could not appeal to the current understanding of the Takings Clause. Expectations based on existing statutes are not considered property rights, so repeal or amendment of those statutes does not count as a taking.
From a theoretical perspective, however, government contracts and entrenching statutes are similar. Critics of entrenchment must explain how their position can be reconciled with enforcement of government contracts, or else accept that governments should not be compelled to pay their debts and comply with other contracts. (102) Eule realizes that his position on entrenchment is in tension with judicial enforcement of government contracts. (103) He accordingly criticizes enforcement of government contracts, but without acknowledging as he should that this longstanding tradition casts doubt on his positive claim that his agency theory describes constitutional practices. Our position is that entrenchment concerns by themselves present no objection to government contract enforcement, a conclusion that the Supreme Court appears to have adopted. (104)
Treaties raise significant entrenchment issues. Although the United States has not entered into a treaty that entrenches by its own terms, entrenchment by treaty is a live issue elsewhere in the world. (105) The European Community requires its members to conform their laws to the laws of the Community, and to keep their laws in conformity with Community law indefinitely. (106) A government that brings a nation into the Community thus entrenches policies--the policies of the Community--to which future governments might object.
In the United States, because a treaty requires a two-thirds vote of the Senate, the Senate cannot abrogate an earlier treaty by consenting to an inconsistent treaty with a simple majority; nor can a majority in both houses--through ordinary legislation--abrogate the international effect of treaty obligations. (107) As a consequence, the Senate (with the President) can reach beyond its "temporal mandate" and entrench policies against the interest of future majorities.
We take no position on whether the Senate should have this power; what we emphasize here is that the entrenchment-like element of a treaty is not objectionable, for this element characterizes--as we have been arguing--all legislation, in the sense that ordinary legislation and entrenching statutes alike restrict the choices of future legislatures.
H. Entrenchment by Agencies and Courts
Our argument--that legislatures have good reasons for entrenching policy, and that the temporal consequences of entrenchment are not by themselves objectionable--applies to agencies and courts. An agency might bind itself to some policy by providing that a supermajority of its governing board (if it has one) must approve changes in that policy. Commitment to the policy and other reasons might justify such an action, just as they do for corporations. The twist here is that Congress must authorize the entrenchment in the agency's organic statute. Moreover, Congress also retains an ex post statutory veto. This makes the case for permitting agency entrenchment both less troubling for critics of entrenchment (because Congress retains its authority) and less important (because agencies cannot entrench themselves very well). At the same time, the creation of agencies in the executive branch, and especially the creation of independent agencies, seems to reflect efforts by Congress to provide the government, through institutional specialization, with the ability to entrench policy. The Federal Reserve Board, for example, reflects a historical effort to entrench low-inflation monetary policy through an institution that can be tampered with only at great political cost.
As for courts, stare decisis is a significant source of entrenchment. When a court makes a decision, future courts treat that decision as an authority, rather than (as they could) an irrelevancy. One common justification for stare decisis is that it enables individuals to rely on judicial decisions. This justification transfers to the legislative context: An entrenched statute is more reliable than an ordinary statute.
Critics of legislative entrenchment need to explain why it would be proper for courts, but not legislatures, to have the power of entrenchment. Here is Laurence Tribe's effort:
[W]hereas neither a Congress nor a President is empowered to make meta-law at the constitutional level necessary effectively to bind future officeholders, the Supreme Court does in a sense make constitutional law to be followed in future cases.... The Court can, of course, overrule its prior decisions. But whereas the mere election of a new President or a new Congress is sufficient, as a matter of both political theory and political reality, to warrant a repeal of prior law or policy, a change in membership of the Supreme Court is not sufficient, under the Court's own prudential and pragmatic principles of stare decisis[,] ... to justify overruling a prior decision. (108)
One problem with Tribe' s argument is that, if the Supreme Court makes constitutional law through interpretation, then Congress can make constitutional law through interpretation, as Congress is, on most accounts, authorized if not compelled to engage in constitutional interpretation, especially in areas like impeachment where the courts stay out. Tribe does not explain why a Congress should not consider itself bound to the constitutional interpretations of earlier Congresses, and, if it should, why legislative entrenchment would be more objectionable.
But the more significant problem with Tribe's argument is that the Constitution does not say that Congress may not make "meta-law," or an entrenching statute. We saw that Eule unpersuasively derived such a prohibition from the clauses that specify the electoral cycle. In fact, the Constitution gives Congress broad powers to legislate, and we have found no reason to think that those powers do not extend to entrenching legislation. Tribe's claim that entrenching legislation is constitutional law because it binds future governments makes it seem as though Congress is amending the Constitution in violation of Article V. But Article V refers to "this Constitution"--that is, the document itself--and makes no reference to entrenching legislation; entrenching legislation is not intrinsically constitutional in any useful sense. (109)
Critics of legislative entrenchment ought to be critics of stare decisis as well. When Jefferson expressed concern about the influence of the dead hand, he concluded logically that legislation and the Constitution--positive law and judicial interpretation--should expire at the end of a generation. Our contrary view is that legislative entrenchment is no more objectionable than the entrenchment of judicial decisions through stare decisis.
Discussions of entrenchment frequently lead to the question of retroactivity. Eule, for example, thinks that entrenchment and retroactivity are two sides of the same coin--the temporal mandate enjoyed by Congress as a result of its agency relationship with the people. (110) Anti-entrenchment prevents the current legislature from controlling future majorities; antiretroactivity preserves the current legislature's ability to control the present unconstrained by the threat that future legislatures will change the rules ex post. We do not take a position on the retroactivity debate, but we do want to insist--against one reading of Eule's argument--that there is no logical connection between retroactivity and entrenchment. Retroactivity and entrenchment are unrelated, orthogonal issues.
Imagine a mini-legislature that is elected anew in every period P. An "entrenching statute" means that the [P.sub.i] legislature may enact a law that the legislatures at [P.sub.i+1] and subsequent periods may not repeal. There are four periods, [P.sub.1] through [P.sub.4]. There are no independent constitutional constraints, like the Contracts or Takings Clauses. In this model, four legal regimes are possible, as follows:
Regime 1 permits entrenching statutes and permits retroactive legislation. At [P.sub.2] the legislature has these powers: It may enact an entrenched statute that the [P.sub.3] legislature may not change; it may enact an ordinary prospective statute that will be in effect at [P.sub.2], and at [P.sub.3] unless the [P.sub.3] legislature repeals it; and it may enact a retroactive statute that will determine legal relations as of [P.sub.1]. All three of these powers, however, will be defeated by any entrenched statutes enacted at [P.sub.1]. A retroactive statute that contradicts the [P.sub.1] entrenched enactment will fail, but not because it is retroactive; a prospective statute that contradicts the entrenched rule will also fail, as will a new (but contrary) entrenching rule. Within the scope of the entrenched rule enacted at [P.sub.1], the [P.sub.2] legislature cannot govern legal relations at any of [P.sub.1], [P.sub.2], or [P.sub.3].
Regime 2 permits entrenching statutes but forbids retroactive legislation. The [P.sub.2] legislature may pass an ordinary prospective statute that governs [P.sub.2], and governs [P.sub.3] unless repealed; it may also entrench a statute against the [P.sub.3] legislature, unless the [P.sub.1] legislature has enacted a contrary entrenchment. Whether or not the [P.sub.1] legislature has entrenched anything, however, the [P.sub.2] legislature cannot determine legal relations as of [P.sub.1]. The [P.sub.3] legislature has to comply with the entrenched rules from [P.sub.1] and [P.sub.2], and even if there are none, it cannot determine legal relations at either [P.sub.1] or [P.sub.2].
Regime 3 forbids entrenching statutes but permits retroactive legislation. The [P.sub.2] legislature may set the rules for [P.sub.1] (by retroactive law) and for [P.sub.2] and [P.sub.3] (by ordinary prospective law), but the [P.sub.3] legislature may change any of those rules--for [P.sub.1] or [P.sub.2] by supervening retroactive law, for [P.sub.3] by ordinary repealing legislation.
Regime 4 forbids both entrenching statutes and retroactive legislation. The [P.sub.2] legislature may set the rules for [P.sub.2] and for [P.sub.3] by ordinary prospective legislation, but not for [P.sub.1]; the [P.sub.3] legislature may set the rules for [P.sub.3] and [P.sub.4], but not for [P.sub.1] or [P.sub.2].
All of these regimes are internally consistent. Eule sometimes suggests that Regime 4 is internally inconsistent, because the ban on retroactive legislation allows the [P.sub.1] legislature to "entrench" its rules for [P.sub.1], in contradiction of the ban on entrenchment. (111) But Regime 2, which permits entrenchment, also allows the [P.sub.1] legislature to enact rules for [P.sub.1] without fear of later retroactive reversal. Eule's concern is just about the ban on retroactive lawmaking; it has nothing to do with entrenchment. Of course, we can if we like define Eule's point as correct by redefining "entrenchment" to mean "passing a law that will be immune from subsequent retroactive reversal." But what is the point of collapsing two useful concepts into one?
This establishes that, ignoring trivial redefinitions, there is no relationship of entailment or of contradiction between views about entrenchment and views about retroactivity. Now, we might have some higher-order theory that dictates a view about entrenchment and also dictates a view about retroactivity. For example, we might (like the other half of Eule's position) have a simplistic principal-agent view that says: The legislature elected for [P.sub.1] should be able to pass rules only about [P.sub.1], [P.sub.2] about [P.sub.2], and so forth. This view entails that the [P.sub.2] legislature should not be able either to entrench against [P.sub.3] or retroactively to govern [P.sub.1] (Regime 4). That combination also accommodates the insight that the [P.sub.1] legislature cannot effectively control [P.sub.1] if (1) its policies for [P.sub.1] can be undone by retroactive legislation at [P.sub.2], and (2) citizens at [P.sub.1] know that (1) is true. Regime 4 gives the [P.sub.1] legislature maximal control over [P.sub.1].
But of course we might have some different higher-order theory. If we believed in societal decline, for example, we might believe that each successive legislature will be dumber than the last. In that case we might like Regime 2, because it gives any upstream legislature more power than its downstream successors. So we cannot avoid all the arguments that Burke, Madison, Jefferson, and Bentham had about how policymaking authority should be allocated across generations. But none of this shows that, as Eule seems to think, there is some special connection between entrenchment and retroactivity.
Politicians secure their policies against future modification by setting up agencies and commissions, drafting legislation in ways that make repeal especially visible, inserting procedures that alert interested parties to potential amendments, committing the government to contracts, engaging in deficit spending, restricting opportunities for debate in legislatures, modifying the voting rules, and even ingeniously manipulating labels (as Roosevelt was said to do, when he called his social security program, which was a simple tax-and-transfer system, a pension plan). These are all forms of entrenchment, and formal legislative and judicial entrenchment do not pose different opportunities and dangers. Critics of entrenchment must come to terms with the ability of legislatures to affect the future and explain what makes legislative entrenchment special and worthy of constitutional concern.
(1.) United States v. Winstar Corp., 518 U.S. 839, 872 (1996) (citing 1 WILLIAM BLACKSTONE, COMMENTARIES *90).
(2.) In Winstar, the Court held that the government had breached various contracts with financial institutions, in violation of the Contracts Clause, which the Court treated as a specific textual exception to the anti-entrenchment principle. We discuss the relationship between entrenchment and governmental contracts in Section II.F.
(3.) See, e.g., Reichelderfer v. Quinn, 287 U.S. 315, 318 (1932); Manigault v. Springs, 199 U.S. 473, 487 (1905); Newton v. Comm'rs, 100 U.S. 548, 559 (1879).
(4.) See, e.g., LAURENCE H. TRIBE, AMERICAN CONSTITUTIONAL LAW 124-25 n.1 (3d ed. 2000); Charles L. Black, Jr., Amending the Constitution: A Letter to a Congressman, 82 YALE L.J. 189, 191 (1972); David Dana & Susan P. Koniak, Bargaining in the Shadow of Democracy, 148 U. PA. L. REV. 473, 526-36 (1999); Julian N. Eule, Temporal Limits on the Legislative Mandate: Entrenchment and Retroactivity, 1987 AM. B. FOUND. RES. J. 379; Catherine Fisk & Erwin Chemerinsky, The Filibuster, 49 STAN. L. REV. 181, 247 (1997); Paul W. Kahn, Gramm-Rudman and the Capacity of Congress To Control the Future, 13 HASTINGS CONST. L.Q. 185, 194 (1986); John O. McGinnis & Michael B. Rappaport, The Constitutionality of Legislative Supermajority Requirements: A Defense, 105 YALE L.J. 483, 505-06 (1995). Canonical works in political theory also discuss the disadvantages of entrenchment. See, e.g., FRANCIS BACON, THE HISTORY OF THE REIGN OF KING HENRY THE SEVENTH 135 (Jerry Weinberger ed., Cornell Univ. Press 1996) (1622) (arguing that one Parliament may not "by a precedent act ... bind or frustrate a future" Parliament); JEREMY BENTHAM, THE BOOK OF FALLACIES 82-112 (Peregrine Bingham ed., London, John & H.L. Hunt 1824) (purporting to expose "The Fallacy of Irrevocable Laws"). We say little about this literature directly; it either is irrelevant to American constitutional law or else advances generic arguments picked up by modern authors, whose anti-entrenchment arguments we rebut in Section I.D.
(5.) For discussion of these authors, see infra Section I.D.
(6.) See Michael J. Klarman, Majoritarian Judicial Review: The Entrenchment Problem, 85 GEO. L.J. 491, 502 (1997). Klarman also discusses entrenchment in our sense; he calls it "crosstemporal entrenchment" and condemns it as "inconsistent with the democratic principle that present majorities rule themselves." Id. at 509. We critique this argument from simple majoritarianism infra Subsection I.D.2.
(7.) 287 U.S. 315, 317 (1932).
(8.) 100 U.S. 548, 561 (1879).
(9.) Reichelderfer, 287 U.S. at 318, 321.
(10.) Newton, 100 U.S. at 559-60.
(11.) 16 U.S.C. [subsection] 1531-1544 (1994).
(12.) See Chae Chan Ping v. United States, 130 U.S. 581, 600 (1889) (stating that, where an earlier and a later statute conflict, "the last expression of the sovereign will must control"); WILLIAM N. ESKRIDGE, JR. ET AL., LEGISLATION 1054 (3d ed. 2001) (noting the "obvious principle" that the later-enacted statute controls).
(13.) David Strauss has suggested to us that a particular legal system might simply contain a rule of recognition holding that entrenched statutes lacking self-reference--PR' in our example--are effective and may not be repealed. The later legislature's two-stage repeal would then violate the rules of the game in a straightforward way. On this view, self-reference is otiose when the rule of recognition does allow entrenchment and inadequate when it does not. We need not decide whether the rule-of-recognition account is superior to the self-reference account, however. So long as either account holds, the later two-stage repeal is blocked, and the argument that entrenchment is conceptually impossible fails. The examples of entrenching provisions that we collect below sometimes display self-reference and sometimes do not.
(14.) See Eule, supra note 4, at 406 n.122.
(15.) Compare Oliver Wendell Holmes, The Path of the Law, 10 HARV. L. REV. 457, 461 (1897) (remarking that "[t]he prophecies of what the courts will do in fact ... are what I mean by the law"), with Reed Dickerson, Toward a Legal Dialectic, 61 IND. L.J. 315, 317 (1986) ("Being preoccupied with the social pathology addressed by case law, legal minds continue to be jurisprudentially scarred by such discredited dogmas as ... Judge Holmes' assumption that law consists only of predictions of what courts will do.").
(16.) See Balanced Budget and Emergency Deficit Control Act of 1985, 2 U.S.C. [subsection] 900-908 (1994).
(17.) Letter from James Madison to Thomas Jefferson (Feb. 4, 1790), in THE MIND OF THE FOUNDER: SOURCES OF THE POLITICAL THOUGHT OF JAMES MADISON 230, 232 (Marvin Meyers ed., 1973).
(18.) This point recalls Stephen Holmes's argument that constitutions take contentious issues off the table. See STEPHEN HOLMES, PASSIONS AND CONSTRAINT: ON THE THEORY OF LIBERAL DEMOCRACY 215-18 (1995). We add that Congresses, in addition to constitutional assemblies, can make a similar determination and implement it through entrenchment.
(19.) The standard analysis is in JAMES M. BUCHANAN & GORDON TULLOCK, THE CALCULUS OF CONSENT 63-72 (1962).
(20.) This assumption should be uncontroversial, particularly given the breadth of the Necessary and Proper Clause. Given the rationales for entrenchment set out in Section I.C above, entrenchment will often be "necessary" to the execution of federal policy on the capacious definition given to that term in McCulloch v. Maryland, 17 U.S. (4 Wheat.) 316, 413-20 (1819), which treats "necessary" as a synonym for "useful" or "conducive to." As for "proper," that term at most cross-references allocations of power and supervening prohibitions established elsewhere in the Constitution. See Gary Lawson & Patricia B. Granger, The "Proper" Scope of Federal Power: A Jurisdictional Interpretation of the Sweeping Clause, 43 DUKE L.J. 267 (1993). Our claim is that there is no independent constitutional prohibition against entrenching statutes, so there can be nothing "improper" about their enactment.
(21.) U.S. CONST. art. I, [section] 5, cl. 2.
(22.) Note that the technical question is just whether the entrenching component or clause of a statute is valid, not whether the remainder of the statute--its content--is valid. This distinction never makes a difference to our argument, however, so for brevity we often refer in shorthand to the question of whether "entrenching statutes" are valid.
(23.) McGinnis & Rappaport, supra note 4, at 504-05.
(24.) U.S. CONST. art. I, [section] 1.
(25.) Nor do the specific grants of power in Article I and elsewhere. When, for example, the Constitution says that "[t]he Congress shall have Power ... [t]o regulate Commerce ... among the several States," id. art. I, [section] 8, cl. 3, there is no reason to think that the provision speaks to the allocation of commerce-regulating power across successive Congresses. Rather, it delegates to the Congress as such a power that would otherwise reside exclusively in the states. See id. amend. X.
(26.) See id. art. II, [section] 1, cl. 1 ("The executive Power shall be vested in a President of the United States of America."); id. art. III, [section] 1 ("The judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish."). We neither need nor intend to take a position on the standard question of whether the Vesting Clause of Article II confers an independent "executive" power on the President, with the content of that power to be inferred from background norms, or instead serves as a mere placeholder for the specific presidential powers listed later in the same Article. See ALEXANDER HAMILTON, LETTERS OF PACIFICUS NO. 1 (June 29, 1793), reprinted in 7 THE WORKS OF ALEXANDER HAMILTON 76, 80 (John C. Hamilton ed., New York, John F. Trow 1851) (arguing that "[t]he different mode of expression employed in [the Vesting Clauses of Articles I and II] serves to confirm this inference" that the authority vested in the President is not limited to the specific cases of executive power delineated in Article II).
(27.) U.S. CONST. art. I, [section] 1 (emphasis added).
(28.) 1 BLACKSTONE, supra note 1, at *90.
(29.) See infra Section II.D.
(30.) 1 BLACKSTONE, supra note 1, at *90 ("Because the legislature, being in truth the sovereign power, is always of equal, always of absolute authority: it acknowledges no superior upon earth, which the prior legislature must have been, if it's ordinances could bind the present parliament."); THOMAS M. COOLEY, TREATISE ON THE CONSTITUTIONAL LIMITATIONS WHICH REST UPON THE LEGISLATIVE POWER OF THE STATES OF THE AMERICAN UNION 152-53 n.3 (Boston, Little, Brown & Co. 1878) (rooting an anti-entrenchment principle in a norm of equality across legislatures); THE FEDERALIST NO. 78, at 439 (Alexander Hamilton) (Isaac Kramnick ed., 1987) (noting that courts have followed the last-in-time rule because "[t]hey thought it reasonable that between the interfering acts of an equal authority that which was the last indication of its will should have the preference"); see also Newton v. Comm'rs, 100 U.S. 548, 559 (1879) (holding that entrenchment is impermissible because "[e]very succeeding legislature possesses the same jurisdiction and power with respect to [public interests] as its predecessors").
(31.) Note, however, that Cooley was confused when he suggested that in a regime permitting entrenchment, "one legislature could ... reduce the legislative power of its successors; and the process might be repeated, until, one by one, the subjects of legislation would be excluded altogether from their control." COOLEY, supra note 30, at 152. The mistake here is the suppressed premise that "the subjects of legislation" remain the same over time. In fact, new issues arise with changes in technology, society, and politics, so that the later legislature will always have the opportunity to address policy questions that earlier legislatures could not have envisioned.
(32.) See, e.g., McGinnis & Rappaport, supra note 4, at 505.
(33.) THE FEDERALIST NO. 78, supra note 30, at 439 (Alexander Hamilton).
(34.) Id.; see also BLACKSTONE, supra note 1, at *90 (listing the anti-entrenchment principle that "[a]cts of parliament derogatory from the power of subsequent parliaments bind not" among other "rules to be observed with regard to the construction of statutes").
(35.) See infra Subsection I.D.1.f.
(36.) See Letter from James Madison to Thomas Jefferson, supra note 17, at 230.
(37.) See David A. Strauss, Common Law Constitutional Interpretation, 63 U. CHI. L. REV. 877, 894 (1996).
(38.) A.V. DICEY, INTRODUCTION TO THE STUDY OF THE LAW OF THE CONSTITUTION 84 (8th ed. 1915).
(39.) Id. at 62.
(40.) Id. at 63; see also O. Hood Phillips, Self-Limitation by the United Kingdom Parliament, 2 HASTINGS CONST. L.Q. 443, 461 (1975).
(41.) A recent example is the Northern Ireland Constitution Act of 1973, which declares that Northern Ireland will remain in the United Kingdom unless a majority of its residents consent to secession. Although the law does not by its terms entrench itself, that is a plausible interpretation. See Phillips, supra note 38, at 445-46.
(42.) See cases cited supra note 3.
(43.) An example is the Septennial Act of 1716, which overturned the entrenchment of three-year parliamentary terms enacted by the Triennial Act of 1694. See Eule, supra note 4, at 391.
(44.) Erie R.R. Co. v. Tompkins, 304 U.S. 64 (1938).
(45.) Brown v. Bd. of Educ., 349 U.S. 294 (1955).
(46.) Article V states:
The Congress, whenever two thirds of both Houses shall deem it necessary, shall propose Amendments to this Constitution, or, on the Application of the Legislatures of two thirds of the several States, shall call a Convention for proposing Amendments, which, in either Case, shall be valid to all Intents and Purposes, as Part of this Constitution, when ratified by the Legislatures of three fourths of the several States, or by Conventions in three fourths thereof, as the one or the other Mode of Ratification may be proposed by the Congress....
U.S. CONST. art. V.
(47.) See 1 BRUCE A. ACKERMAN, WE THE PEOPLE: FOUNDATIONS 266-94 (1991); Akhil Reed Amar, The Consent of the Governed: Constitutional Amendment Outside Article V, 94 COLUM. L. REV. 457, 458 (1994).
(48.) The two provisions are Article I, Section 9, Clause 1 and Clause 4. Article V allowed both provisions to be amended after 1808. See U.S. CONST. art. V.
(49.) See id.
(50.) We are indebted to John Manning for suggesting the argument in this paragraph.
(51.) See U.S. CONST. art. VI, cl. 2 ("This Constitution, and the Laws of the United States which shall be made in Pursuance thereof; and all Treaties made, or which shall be made, under the Authority of the United States, shall be the supreme Law of the Land....").
(52.) Eule, supra note 4, at 405.
(53.) As the Septennial Act did. See supra note 43.
(54.) U.S. CONST. art. I, [section] 5, cl. 2.
(55.) Id. (emphasis added).
(56.) Eule, supra note 4, at 409 n.139.
(57.) Bruce Ackerman et al., An Open Letter to Congressman Gingrich, 104 YALE L.J. 1539, 1541 (1995).
(58.) Id. at 1542.
(59.) Jed Rubenfeld, Rights of Passage: Majority Rule in Congress, 46 DUKE L.J. 73, 74 (1996).
(60.) U.S. CONST. art. I, [section] 7, cls. 2-3.
(61.) Kahn, supra note 4, at 194.
(62.) See H.C. 98/69, Bergman v. Minister of Fin. & State Comptroller, 23(1) P.D. 693 (Isr.) (applying a similar interpretive presumption, but upholding the Knesset's constitutional authority to entrench basic laws as against future Knessets).
(63.) Dana & Koniak, supra note 4, at 533.
(64.) See James M. Enelow, Cycling and Majority Rule, in PERSPECTIVES ON PUBLIC CHOICE 149 (Dennis C. Mueller ed., 1997) (surveying conditions that impact cycling in majority contests).
(65.) Something like this view seems to underlie Klarman's argument that "[a]n important distinction exists ... between today's majority exercising sovereignty over the present in a way that unavoidably affects the future and today's majority seeking direct control over the future in a manner that is unnecessary to implementing its complete control over the present." Klarman, supra note 6, at 505. On Klarman's view, only the latter counts as illegitimate "cross-temporal entrenchment." See id. at 505-07.
(66.) Eule, supra note 4, at 399.
(67.) Id. at 405 n.121.
(69.) See, e.g., RICHARD A. EPSTEIN, TAKINGS (1985).
(70.) Ackerman et al., supra note 57, at 1542.
(71.) Daniel R. Fischel & Alan O. Sykes, Governmental Liability for Breach of Contract, 1 AM. L. & ECON. REV. 313, 334 (1999).
(72.) Id. at 342-44. We do not understand how Fischel and Sykes reconcile this insight with their critique of enforcement of government-granted monopolies.
(73.) See JOHN HART ELY, DEMOCRACY AND DISTRUST 181-83 (1980).
(74.) Eule, supra note 4, at 390-91.
(75.) See supra text accompanying notes 7-8 (citing Reichelderfer v. Quinn, 287 U.S. 315 (1932); and Newton v. Comm'rs, 100 U.S. 548 (1879)).
(76.) STANDING RULES OF THE SENATE, S. DOC. NO. 106-1, R. XXII, at 21 (2000). The principal category of Senate business that is exempt from the filibuster is the federal budget process. See Fisk & Chemerinsky, supra note 4, at 215.
(77.) STANDING RULES OF THE SENATE, supra note 76, R. V, at 5 (emphasis added).
(78.) See Fisk & Chemerinsky, supra note 4, at 212-13.
(79.) For the latter idea, see id. at 217-23.
(80.) Id. at 249-50.
(81.) 2 U.S.C. [subsection] 900-908 (1994).
(82.) Later modifications to Gramm-Rudman changed the deficit caps to spending caps and worked other important alterations. See ESKRIDGE ET AL., supra note 12, at 433-37. None of that is relevant to our concern here, which is just to draw conclusions about entrenchment from the original Gramm-Rudman law.
(83.) Kahn, supra note 4, at 203-04.
(84.) 478 U.S. 714 (1986).
(85.) Kahn says that the statute contains an implicit provision, inferable from its structure, that any repeal must be express. Kahn, supra note 4, at 202 n.61. That is the sort of claim that makes people dismiss statutory interpretation as a game without rules.
(86.) Id. at 205, 209.
(87.) Dictionary Act, ch. 71, [section] 2, 16 Stat. 431, 431 (1871).
(88.) 1 U.S.C. [subsection] 1-7, 101-114 (1994 & Supp. V 1999).
(89.) 1 U.S.C. [subsection] 108-109 (1994).
(90.) 1 U.S.C. [section] 7 (Supp. V 1999); 28 U.S.C. [section] 1738C (Supp. V 1999).
(91.) 1 U.S.C. [section] 2 (1994).
(92.) 5 U.S.C. [section] 559 (1994).
(93.) 50 U.S.C. [section] 1621 (1994).
(94.) 42 U.S.C. [section] 2000bb-3(b) (1994) ("Federal statutory law ... is subject to this chapter unless such law explicitly excludes such application by reference to this chapter.").
(95.) 15 U.S.C. [section] 1012(b) (1994) ("No Act of Congress shall be construed to invalidate, impair, or supersede any law enacted by any State for the purpose of regulating the business of insurance, or which imposes a fee or tax upon such business, unless such Act specifically relates to the business of insurance...."); see also Canadian Bill of Rights, S.C. 1960, ch. 44, [section] 2 ("Every law of Canada shall, unless it is expressly declared by an Act of Parliament of Canada that it shall operate notwithstanding the Canadian Bill of Rights, be so construed and applied as not to abrogate, abridge or infringe ... any of the rights or freedoms herein recognised and declared."), quoted in Phillips, supra note 40, at 477.
(96.) See Morton v. Mancari, 417 U.S. 535 (1974) (holding that the Equal Employment Opportunity Act of 1972 did not implicitly repeal the Indian Reorganization Act of 1934).
(97). 15 U.S.C. [subsection] 719a-719o; see also Metzenbaum v. FERC, 675 F.2d 1282 (D.C. Cir. 1982) (holding that alleged violations of the Act's procedural requirements are nonjusticiable).
(98.) 5 U.S.C. [subsection] 801-808 (Supp. V 1999).
(99.) The statute provides:
This section is enacted by Congress ... with full recognition of the constitutional right of either House to change the rules (so far as relating to the procedure of that House) at any time, in the same manner, and to the same extent as in the case of any other rule of that House.
Id. [section] 802(g).
(100.) U.S. CONST. amend. V ("No person shall ... be deprived of life, liberty, or property, without due process of law; nor shall private property be taken for public use, without just compensation.").
(101.) Id. art. I, [section] 10, cl. 1 ("No State shall ... pass any Bill of Attainder, ex post facto Law, or Law impairing the Obligation of Contracts....").
(102.) One of us has expressed doubt about whether courts should enforce government contracts for the purpose of maximizing utility across generations. That argument asserts a claim about judicial policy, not constitutional law, and depends on certain premises--about the role of courts in maximizing welfare and the relationship between courts and the legislature--from which we abstract for the purposes of this Essay. See Eric A. Posner, Should Courts Enforce Government Contracts? (2001) (unpublished manuscript, on file with authors).
(103.) Eule, supra note 4, at 420.
(104.) See United States v. Winstar Corp., 518 U.S. 839, 881-83 (1996).
(105.) The closest such treaty is one between the United States and an Indian tribe, which provided that it could not be repealed without the consent of three-quarters of the adult male Indians occupying the land in question. This treaty entrenched policy indirectly by taking the repeal decision partially out of the hands of American elected officials. The Supreme Court refused to strike down a subsequent statute that violated the treaty, but on the ground that Congress's power to enact such a statute was a political question, not because the treaty was unconstitutional. See Lone Wolf v. Hitchcock, 187 U.S. 553, 565-68 (1903). Thanks to Phil Frickey for bringing this example to our attention.
(106.) See Phillips, supra note 40, at 444.
(107.) See Eule, supra note 4, at 425 n.213 (discussing whether treaties entrench); Jack Goldsmith, Rethinking the Domestic Priority Rules for Statutes and Treaties (2001) (unpublished manuscript, on file with authors).
(108.) TRIBE, supra note 4, at 126 n.1.
(109.) See supra Section I.D.
(110.) Eule, supra note 4, at 443.
(111.) Id. at 444.
Adrian Vermeule, Professors of Law, The University of Chicago. We thank Phil Frickey, Elizabeth Garrett, Jack Goldsmith, John Harrison, Daryl Levinson, John Manning, Mark Roe, David Strauss, Larry Tribe, Detlev Vagts, and participants at workshops at Harvard and the University of Chicago faculty workshop for their comments, and Tana Ryan for research assistance. Professor Posner thanks the Sarah Scaife Foundation Fund and the Lynde and Harry Bradley Foundation Fund for generous financial support. Professor Vermeule thanks the Russell J. Parsons Research Fund for the same.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||allowing legislatures o bind their successors|
|Author:||Posner, Eric A.; Vermeule, Adrian|
|Publication:||Yale Law Journal|
|Date:||May 1, 2002|
|Previous Article:||Liberals and romantics at war: the problem of collective guilt.|
|Next Article:||Judicial review, the congressional process, and the federalism cases: an interdisciplinary critique.|