Constitutional Review in Representative Democracy

1. Introduction

Democracy is a powerful political ideal. It is widely regarded as an essential ingredient of political legitimacy. It is no surprise, then, that appeals to it figure so prominently in the case against judicial practices of constitutional review of legislative decisions for their compatibility with fundamental rights. In most political systems that have such practices of constitutional review, the legislature is the main locus of democratic legitimacy. Thus, for many, the fact that constitutional review cuts away from the power of the legislature and gives it to unelected judges is reason to object to it, even if it is true that such a practice better protects the fundamental rights of citizens. This reasoning has obvious force—and rhetorical purchase—in political systems that assign courts the power to strike down legislative decisions that they find in contravention of the constitutional norms they are called to protect. But it also has considerable force, though perhaps less so, in political systems where a finding of incompatibility does not have this dramatic effect.

Consequently, constitutional review enthusiasts seem to find themselves caught in a dilemma. Arguably, they have to decide between democracy, on the one hand, and the values that constitutional review is said to promote (where those values crucially include fundamental rights), on the other. At the very least, it appears that they cannot maintain their commitment to constitutional review without a measure of regret. This dilemma is not confined to academic circles. It resonates across the political system and thus has important practical implications as well. Should legislators resent the fact that judges interfere with their decisions when they are forced to acquiesce to this interference? Should they actively try to circumscribe the supervisory authority conferred on judges by, for instance, narrowing the grounds of constitutional review, if it is within their power to do so? Conversely, should judges be timid and embarrassed when they engage in constitutional review? Or should they do so explicitly and confidently?

The dilemma, then, typically arises within a certain institutional environment and informs the options and decisions of actors occupying certain positions in that environment. Thus, instead of talking about the tension between democracy and fundamental rights in the abstract, I propose we focus on the two state institutions at the heart of this tension, the legislature and the courts, and explore the way in which they are structured to accommodate our dual allegiance. To do this, we need to delve deeper into the institutional credentials of each and understand the nature of the joint enterprise in which they are partners. As such, partners, courts and the legislature ought to be responsive to each other's contributions and to the distinctive values that each brings to the common project.

In this article, I shall offer an account of the relationship between the legislature and the judiciary that makes room for the permissibility of constitutional review. I shall start by explaining why legislatures ought to have a prominent role in a democratic legal system. I shall claim that they deserve to be given the initiative in the project of governing. To this effect, I shall draw particular attention to the fact that legislatures are representative institutions, which lie at a certain distance from the citizenry, and I shall explore the implications of this for institutional design. I shall further contend that this grant of power raises a checks-and-balances concern, and that courts, by virtue of their independence, can effectively meet this concern.

The case for the permissibility of constitutional review that I seek to put forward is modest, in two senses. First, it recognizes that legislatures have the primary responsibility for the realization of constitutional principles and that the role of courts under systems of constitutional review remains subsidiary to that of the legislature. In fact, it is only by taking into account its subsidiary character within the broader constitutional framework that we can make best sense of it and map its limits. Second, although the account advanced here furnishes an important reason of political morality that counts in favour of constitutional review, it insists that this reason is not necessarily overriding; it ought to be balanced alongside other considerations whose force varies from one legal system to another. Thus, it allows for the possibility that constitutional review may be deemed ineffective or suboptimal in a given legal system.

Hence, this approach charts a middle course between two very popular views regarding constitutional review: one against it and one for it. On the one hand, it opposes the view that criticizes constitutional review on the ground that it violates political equality by giving the opinions of a few unelected judges ‘superior voting weight’ over the opinions of everybody else. Contrary to this view, it insists that constitutional review is not in principle at odds with representative democracy. Equally, it distances itself from the view that contends that judges should have the power of constitutional review because they are better than legislators at deciding questions of principle, such as those pertaining to the meaning and content of fundamental rights. This view neglects the important constraints on courts that stem from their place in the project of governing and the respect they owe the legislature. Although I do not intend to criticize these two views here, in the course of the argument, I shall highlight some of the key differences.

By rejecting the polarization that these two views have engendered, we are in a better position to appreciate that the permissibility of systems of constitutional review in a democratic regime is not decided in an all or nothing fashion, but is rather a matter of fine-grained, context-specific institutional balances and adjustments. To achieve the optimal result in this exercise, constitutional design can choose from among a number of institutional options and experiment with variables such as the scope and intensity of review, as the examples of the UK and other Commonwealth jurisdictions amply show. In this sense, apart from its immediate goal, the approach favoured here also aims to vindicate the independent importance of constitutional theory. Often, in the work of constitutional review enthusiasts and sceptics alike, constitutional law is presented as little more than applied political philosophy. But constitutional law is not just an idle wheel in an essentially philosophical exercise or a conveyor belt of abstract precepts of political morality. It brings its own insights and solutions to seemingly intractable philosophical puzzles. By showing how the political principles that we cherish can be and have been realized and harmonized in actual institutional arrangements, it holds out the hope of their reconciliation and strengthens our commitment to them.

2. Substance and Procedure

The controversy over the legitimacy of constitutional review epitomizes a tension in theories of political legitimacy more generally. This tension springs from the fact that political legitimacy is a function of both substance and procedure. A legitimate government is one that exercises its coercive force in a way that both respects and promotes certain values, such as the basic rights of citizens, and is in accordance with decisions reached through appropriate procedures. Importantly, theories of political legitimacy tend to determine the authority of procedures to some extent independently of their outcomes.¹ But procedures, however appropriate, do not always deliver the right results in terms of substance. They may fall short in a number of ways, but—more relevant for present purposes—they sometimes fail optimally to respect the basic rights of citizens that are vital to the continuing legitimacy of a political regime.

The tension just described is particularly acute in democratic theory. That is because democratic theory typically assigns great weight to procedures that embody the principle of political equality among all citizens. Some theorists, like Jeremy Waldron, give this principle a distinctively proceduralist twist.² They understand it to entail a requirement that everyone’s views be assigned equal weight in political decision-making. Accordingly, majority vote is for them of the very essence of democracy because it is the procedure that gives every citizen’s view the most weight that is compatible with an equal weight for everyone else’s view. Proceduralists do not dispute that majoritarian decision-making will sometimes produce substantive outcomes that we have good reason to morally object to. In such cases, our concern for substance and our concern for procedure will pull in opposite directions. It is no wonder then that theorists of this persuasion oppose constitutional review. We may decide to have unelected judges override majoritarian decision-making in order to give priority to substantive norms, but this choice necessarily involves a sacrifice in democratic self-government.

However, the tension between substance and procedure is not peculiar to such proceduralist versions of democracy. Importantly, it does not go away if we adjust our conception of democracy to include a substantive commitment to fundamental rights. A characteristic example of the latter approach is the one put forward by Ronald Dworkin.³ Dworkin argues for what he calls the ‘constitutional conception’ of democracy. According to it, a democratic state owes every citizen equal concern and respect. Put differently, the precepts of equal concern and respect are ‘conditions’ of democratic legitimacy.⁴ The constitutional conception rejects the common wisdom that democracy is identical to majority rule. Decisions that deny the most fundamental rights of a minority are undemocratic, even if they are endorsed by popular vote. Contrary to the proceduralist view rehearsed above, Dworkin thinks that setting aside such decisions is not at odds with the value of democracy. Quite the opposite, doing so promotes democracy because it affirms the concern and respect that members of the minority are entitled to as citizens of a democratic polity.

So, Dworkin juxtaposes a substantively rich idea of democracy to the proceduralist conception that identifies democracy with majority vote. At first sight, his conception dissolves the tension between substance and procedure in favour of the former. What determines the democratic quality of a policy is the true meaning of equal concern and respect, not what some people said it means. But, this will not get us far. Those who, like Dworkin, think that a polity committed to democracy ought to make moral judgments about what equal concern and respect, properly understood, requires must also ask a further question: how should it make those judgments? This question brings procedure back onto the scene. We need procedures about certain issues in large part because we need to identify whose judgment about the meaning of equal concern and respect will count as binding and for how long. This need applies to Dworkin’s theory, no less than others. Indeed, Dworkin is acutely aware of it. His solution is that the standard for the right decision-making procedures is an outcome-oriented one. You start by identifying the substantive entitlements and you then evaluate procedures by their capacity or tendency to protect the entitlements. I do not intend to assess the plausibility of this solution. I am only interested in highlighting the way in which it is also stung by the tension between procedure and substance. Suppose it turns out that judges are better at discovering what equal concern and respect entails. If so, democracy recommends that they be given the power to decide issues of fundamental right in the name of the political community. Even then, though, this recommendation will be based on a general tendency and not on any putative infallibility that judges enjoy. Judges may get it wrong in a specific case, but our commitment to a procedure that gives them the power to decide will compete with our moral objections to the substance of that specific decision.

At any rate, it is not necessarily the case that judges are best equipped to make this kind of judgment: the constitutional conception does not entail the legitimacy of constitutional review, and contrary to common wisdom, Dworkin acknowledges this fact. As he readily admits, ‘[l]egislatures are guardians of the constitution, too, and that includes constitutional principle’.⁵ He concludes that:

I do not mean that there is no democracy unless judges have the power to set aside what a majority thinks is right and just. Many institutional arrangements are compatible with the moral reading, including some that do not give judges the power they have in the American structure. But none of these varied arrangements is in principle more democratic than others. Democracy does not insist on judges having the last word, but it does not insist that they must not have it.⁶

But if it is open which institution should have the final say in constitutional matters, we need to have recourse to procedural considerations in order to decide whose view prevails. Besides, we need to make up our minds not on a case by case basis but in a more systematic way. Our views about the substance of government policy will be of no avail here. In fact, despite its focus on substantive outcomes, the constitutional conception of democracy itself explicitly allows that procedures may also be shaped by considerations that are distinct from substance. So, for instance, it reserves a special place for the ‘one man one vote’ principle or for majority rule. Both principles are supported by weighty considerations of political morality. The ‘one man one vote’ principle affirms equal respect at the symbolic level, and majority rule gives citizens assurance that their views can make some difference in politics—though not necessarily equal. Hence, both have participatory value quite independently of the outcomes that procedures which affirm those principles produce.⁷ It is the importance of that value that makes Dworkin reject an extreme preoccupation with substance:

Democracy would be extinguished by any general constitutional change that gave an oligarchy of unelected experts power to overrule and replace any legislative decision they thought unwise or unjust. Even if the experts always improved the legislation they rejected – always stipulated fairer income taxes than the legislature had enacted for example – there would be a loss in self-government which the merits of their decisions could not extinguish.⁸

3. Governing Together

The tension between substance and procedure in the theory of democracy forces us to shift our attention to institutional design. Even if we go along with Dworkin that the idea of democracy carries with it a commitment to a set of robust individual rights, it still remains to be determined who gets to decide what this commitment amounts to.⁹ Is it going to be the legislature or the courts? Or is it both bodies together? Which institutional structure is going to be systematically more faithful to democratic values, both procedural and substantive? We cannot hope to answer these questions satisfactorily unless we start from further upstream and develop a better understanding of the relationship between the two bodies. To begin with, we must clarify why it is good to be governed by institutions that have the credentials and characteristics that the courts and the legislature do. In other words, we must identify what makes the contribution made by the courts and the legislature valuable. Further, we must ascertain which risks government by these two bodies carries and what institutional design can do to minimize those risks. This two-sided exercise will help explain the interaction between courts and the legislature. We should expect that the two bodies are responsive to each other’s contribution because, and insofar as, they must be responsive to the value that underpins it. In addition, we should expect that their interaction is structured in a way that mitigates or checks some of the risks associated with each of them. In short, we must measure the relationship between courts and the legislature against the precepts of separation of powers. Separation of powers includes a division-of-labour and a checks-and-balances component. The former dictates that governmental tasks ought to be assigned to those who will perform them adequately and the latter dictates that they ought to be distributed among institutional actors so that government power will be internally checked.

In what follows, I shall identify the role that the courts and the legislature are set up to perform and explain why this role is valuable. Since both the judiciary and the legislature are complex institutions, it is only to be expected that they serve a variety of values. Here, I do not intend to offer a full account of those values. I shall, thus, focus on what I consider most pertinent to the debate about the legitimacy of constitutional review. I shall say that legislatures are meant to combine the demands for popular support and moral innovation, whereas courts, by virtue of their independence, are good candidates for assuming the task of legislative supervision. I shall show how this division of labour explains some of the most characteristic powers we assign the two bodies as well as the duties they have vis-à-vis one another, when it comes to governing together. More specifically, I shall say that the legislature ought to be given the initiative in the project of governing. Correspondingly, courts have a duty to implement and further legislative decisions. In other words, implementing those decisions is the appropriate response to the value that they are meant to instantiate. At the same time, decisions of an independent court that scrutinize a piece of legislation for its conformity with constitutional principles serve a valuable checks-and-balances function. To this extent, they, too, merit the legislature’s respect.

Before I turn to the main analysis, I must make two methodological points clear. First, I must say a few words about how I shall go about assigning a role to each of the two bodies. For the main part, I intend to build upon certain familiar procedural features of the institutions that we commonly refer to as courts and legislatures. I shall then claim that by virtue of those procedural features the body that displays them instantiates a certain kind of value or is well-equipped to perform a valuable institutional role and, thus, enjoys a certain authority, especially vis-à-vis other state institutions. This, again, is a familiar and, I hope, more or less uncontroversial form of argument in constitutional theory. To the extent that constitutional theory is also about who gets to make decisions in the name of the political community, it must ask: what is good about being governed by this or that political body? Typically, answers to this question make reference to the composition of that body and the decision-making process it follows.¹⁰ They contend that bodies with that composition and decision-making process reliably instantiate the relevant value or reliably perform the relevant role. Of course, in view of the daunting diversity of actual and possible institutional configurations, some degree of idealization will be unavoidable in what follows. But idealization is necessary for another reason as well: particular institutional choices, whether wise or unwise, can only make sense in the light of the purpose they are supposed to further and the standard they are regarded as trying to comply with. It is in the articulation of that purpose and this standard that the following sections aim to contribute. In the words of François Guizot: ‘Le principe détermine les formes; les formes révèlent le principe’.¹¹

The second methodological point is a caveat. In this article, I focus on courts and the legislature. I thus leave out the third apex of the modern state, namely the executive. This omission, it could be suggested, seriously undermines any conclusions I might reach. That is because, nowadays, it is arguably the executive that initiates measures at various levels. For instance, it drafts legislation that goes through parliament and it also enjoys wide-ranging delegated law-making powers. On this view, the job of the legislature is, to use John Stuart Mill’s apt expression, solely to put ‘the final seal of national assent’ upon decisions that come from other sources.¹² Now, undeniably, there is truth in this view. The emergence of the administrative state has made an immense difference to the division of institutional labour in modern legal systems.¹³ But we ought to bear in mind that the debate over the compatibility of constitutional review with democracy is commonly framed in terms of a struggle between the courts against the legislature, so it is their relationship that constitutional review enthusiasts—and sceptics—must primarily elucidate.¹⁴ Accordingly, I am more concerned to determine their relative position in the joint institutional project of governing. Besides, although, undoubtedly, in many legal systems, legislatures have a rather receptive role vis-à-vis the executive, they are not supposed to be mere smokescreens either. The views of legislators permeate decision-making. Even in cases where the drafting of legislation does not originate in the legislature, legislators are typically empowered to pass amendments to tabled bills. Furthermore, drafters of legislation ought to consult legislators or anticipate their qualms so as to ensure their support.

4. Legislative Initiative

In a democracy, I suggested above, the legislature must be viewed as having the initiative to shape public policy. What could possibly be the basis for the assignment of this power? In this section, I shall argue that the answer lies in the fact that the legislatures that we are familiar with combine two characteristics:

• their accountability to the electorate;

• their deliberate distance from the electorate.¹⁵

I tend to think of these two characteristics as intermediate. By this, I mean that their fuller articulation depends on higher-order normative assumptions, for example about the point and value of democratic participation or about the nature and purpose of legal authority. These normative assumptions form the foundation of more comprehensive theories where these two characteristics find their place. Understandably, different theories will assign different roles and importance to each. But the fact that they are not ultimate should not lead us to the conclusion that they are unimportant. For one thing, neither of them seems easy to dispense with. Both have a firm grip on our understanding of the role of legislatures. Hence, while different theoretical approaches and institutional designs may focus on one, they cannot eliminate the other. Inevitably, they must reserve a place for both. Besides, while their implications are drawn and their content fleshed out in the light of more fundamental philosophical and political commitments, one can still talk about them to some extent in abstraction of those commitments. It is this thin plateau of agreement that the ensuing analysis will draw on. This is bound to make the analysis less apt to yield or support concrete judgments. But this is as it should be in the light of the limited theoretical goal of this article, which is not to provide a full articulation of the institutional role of either the legislature or the courts. Rather, it is to account for their interaction in a way that makes room for the moral permissibility of constitutional review. That is not to say that my proposal will be compatible with all major accounts of the legislative role. Indeed, as will become obvious, it conflicts with one very influential account.

After this brief clarification, let us turn to the first feature, accountability to the electorate. Procedurally, its most characteristic expression is the popular election of legislative bodies. Legislators regularly submit themselves to popular vote. So, they depend for their election and re-election on what voters think of them. This fact creates a variety of institutional incentives. First of all, it pushes legislators to explain their past decisions and future plans to the electorate. Based on this explanation, citizens can make an informed judgment about whom to vote for. In addition, popular accountability encourages legislators to offer a justification for their agenda and official record. Their justification can give citizens new reasons to vote for them and even convince them to change their mind. More strongly, it gives legislators an incentive to heed the views and interests of their constituency and integrate them in their agenda.¹⁶ Obviously, their doing so often improves their chances to be re-elected.

It is all too common to identify this last aspect of accountability exclusively with vested interests, economic lobbies, bargaining, cajoling and the like. This is a tendency heavily criticized by Jeremy Waldron:

Not only do we not have the normative or aspirational models of legislation that we need, but our jurisprudence is pervaded by imagery that presents ordinary legislative activity as deal-making, horse-trading, log-rolling, interest-pandering, and pork-barelling – as anything, indeed, except principled political decision-making.¹⁷

However, it would be a mistake to suppose that legislators ought to merely aggregate the self-interested preferences of their constituents.¹⁸ In endorsing or rejecting a candidate, in approving or disapproving a legislator’s performance, voters often consciously express some vision of what a just and well-ordered polity ought and ought not do and how individuals like themselves can fit in it. It is also this vision (or specific judgments inspired by it) that the elected representatives are moved to take into account in their decision-making through the mechanisms of popular accountability that determine the composition of democratic legislatures. To this extent, then, they are expected to exercise what I shall call dependent judgment, namely judgment that is informed by the views and interests of voters.

Some constitutional review sceptics have taken the dependence of the legislative role on popular views to mean that the legislature embodies a strong form of political equality. The argument seems to be that, since this dependence is realized through an electoral system that gives each citizen’s vote equal weight, elected legislators that are thus dependent on their voters also promote political equality; for example, it may be that they have reason to take every voter’s view equally into consideration. In addition, the fact that they do so constitutes an important component of their authority. Waldron captures this approach nicely when he describes legislatures as forums where ‘representatives of the whole society, elected on a basis that treats all individuals in the society as equals, vote as equals’.¹⁹ From this suggestion, theorists like Waldron construct an argument against constitutional review. They argue that, whereas legislative decisions affirm political equality in the sense suggested above, constitutional review denies it because it gives the views of a few unelected officials far superior weight than those of ordinary citizens. They conclude that constitutional review is pro tanto undemocratic.²⁰ But this view overlooks the second feature of legislatures that I mentioned earlier, namely, their deliberative distance from the electorate. By virtue of this distance, I shall now suggest, legislators feel an opposite normative pull toward making independent judgments, that is, judgments that reflect their own vision of a just and well-ordered polity rather than that of their constituents.²¹

The deliberative distance between legislators and their constituents is sustained by a matrix of familiar procedural rules that regulate the function of elected legislatures in the legal systems we are familiar with. Consider, for instance, the rules defining the legislators’ term of office. Legislators are ordinarily elected in relatively large constituencies for a term of a few years. During this period, the electorate does not have any institutional means to control the way they vote. Importantly, the electorate lacks the power to issue imperative mandates to its representatives or to revoke them at will. Though not without exceptions, this is the political reality in the overwhelming majority of modern states.²²

Let me pursue the implications of this arrangement a little further. Due to the length of the legislators’ tenure, there will inevitably arise situations where they are called upon to make decisions regarding issues on which voters had no opportunity to reflect before the elections or currently have no unambiguous view.²³ Besides, even if voters do form a discernible view on a given issue after the election, they have no way of imposing it on their delegate. They can, of course, punish her in the next election, if she goes the other way. But, the difference is crucial. Although there certainly are issues that sometimes decide an election, a delegate will typically be judged on the basis of her entire agenda and/or record as well as on the cogency of the justifications she offers in her defence, when she seeks re-election. At any rate, the verdict of the people at the next election does not affect the standing of the decisions that the delegate made during her tenure.

In all these respects, then, the legislative role leaves ample room for independent judgment. The aim is that the legislator does not operate as a mere conduit of popular opinion but instead deliberates and decides at a certain remove from the influence of her constituents. In order to discharge her institutional duty, she will have to form her own view about what ought to be done. Often, she will reach different conclusions from those of her constituents, and there is nothing in her job description that dictates that she defer to them.²⁴ It is open to her to vote according to her conscience and then justify her choice to her constituents, hoping either that this choice will not carry significant weight in the next election or that her justification will sway them. This is not to say that legislators are not subject to other types of constraints. Constraints from one’s membership in a political party are an obvious example. But, if anything, those constraints render the relationship between the legislator and her constituents even more mediated.

So far, we have examined accountability and deliberative distance in isolation. We now need to bring these two features together and ask whether an institutional role that combines both has any value. Why is it good to be governed by officials that are subject to these opposite forces? At first blush, nothing good can come out of this mishmash.²⁵ In fact, it appears that legislators face an impossible task. If they are supposed to exercise independent judgment, why are they subject to electoral pressure? If they are meant to be sensitive to popular opinion, why can’t they take it at face value? To my mind, the combination of the two features does not constitute a defect in the system of parliamentary democracy, but is, rather, the basis of its normative appeal. Like all representative institutions, an elected legislature is neither a ‘gouvernement des sages’ nor a plebiscitarian forum, but it partakes of both. Because of this, it can best balance the requirements of popular support and moral innovation.

Start with popular support. A just and well-ordered polity is one that citizens are, by and large, also motivated to uphold. A representative government is more likely to secure popular support for two main reasons. First, it makes citizens less resentful at the exercise of state coercion. Second, it makes citizens less fearful of political authority. Representative government has the first effect, insofar as it fuses government policy with the views and interests of citizens and is supposed to do so. Members of a political community would be disinclined to support its regime, if they thought that they had no input in political decision-making or if its outcomes persistently frustrated their reasonable expectations and views as to what those outcomes should be. Recall Dworkin’s justification of majority vote mentioned above. Government by bodies that are popularly elected, he claims, ensures that citizens are actually heard to some extent in political decision-making.²⁶ It thus fosters the identification of the people with the political decisions taken in their name and reduces feelings of resentment. Of course, experience shows that for popular support to be secured it is not necessary that each and every policy matches voters’ views. Still, a government cannot get away with a ‘fiat-justitia-pereat-mundus’ spirit for too long; it must be responsive to the electorate. As Christopher Eisgruber puts it: ‘Designing effective strategy requires among other things, enlisting other actors: individuals, businesses, interest groups and so on. To get the cooperation of these actors, it helps to know what interests motivate them’.²⁷

People who live under representative institutions are also less fearful of political authority. This is because they are educated in thinking that political authority changes hands and that it is up to them collectively to effect that change. In the words of George Kateb, submission to regular election ‘chastens political authority’, and correspondingly ‘loosens the authority’s hold’ on the citizenry.²⁸ For Kateb, this is evinced by the fact that representative democracy nourishes a general attitude of:

pervasive skepticism toward authority; a reluctance to defer; a conviction that those who wield authority must themselves be skeptical toward their roles and themselves and that necessary authority must be wielded in a way that inflicts minimum damage on the moral equality of all people. Furthermore, there is a tendency to try to do without authority wherever possible or to disperse or disguise it, and thus to soften it.²⁹

In short, it entrenches a sense of individual independence. Importantly, rather than undermine political authority, this sense is more likely to generate support for political institutions, insofar as those institutions affirm and promote a conception of persons (as independent) that each one of us can willingly relate to and make part of our own idea of the good life. That is not to contest that political regimes can be and have been sustained over time by fear or blind deference. But when citizens abide by and support a regime not out of fear but out of independent commitment and a positive valuation of political membership, they attain the right kind of stability or stability for the right reasons.³⁰ Note that this second type of connection between representative government and political stability does not require any systematic correlation between constituents’ views and legislators’ decisions. The ‘chastening’ effect is realized through the mere fact that legislators ought regularly to submit themselves to popular election. It goes without saying, of course, that a political official so constrained will, as a matter of fact, be motivated to heed her constituents’ views.

If, however, popular support is so important, why, it may be argued, are legislators allowed significant leeway to decide contrary to the views of their constituents? In short, what is the political rationale underlying the deliberative distance between the legislature and the electorate? The answer has to do with the transformative effect of public institutions. The views, interests and motivations of citizens should not be taken as fixed. Instead, they are amenable to change by means of, among other things, moral reasoning. Neither, as I have argued above, is fidelity to them intrinsically valuable; if it were, the set-up of legislative decision-making would not leave room for the exercise of independent judgment. We need not, for present purposes, take sides in the age-old meta-ethical question whether moral reasons are the kind of things that are both universal in scope and necessarily motivate moral agents or, alternatively, whether they have motivational power, but their applicability is relative to specific moral agents. It suffices that, by and large, agents are sensitive to right moral reasons and may come to alter their views in light of them.

If this possibility is open, we can make sense of legislators as moral pioneers, pushing forward the moral horizons of the citizenry. When a legislative scheme is supported by the right reasons, it may well transform the vision citizens have of the just and well-ordered polity. In particular, it may urge them to see differently to what extent self-interest can legitimately figure in this vision.³¹ Thus, James Madison famously wrote that it is the very purpose of a national legislature

to refine and enlarge the public views by passing them through the medium of a chosen body of citizens, whose wisdom may best discern the true interest of their country and whose patriotism and love of justice will be least likely to sacrifice it to temporary or partial considerations.³²

How radically this vision can, or should be forced to, change at any given time is a question to be settled also by appeal to considerations of political prudence. In general, it is fair to say that political prudence favours incremental measures. It is one thing to aspire to be a moral pioneer; it is quite another to be quixotic. However, this is no reason to give up trying altogether. If change makes sense at all, we need to have public institutions powerful enough and self-confident enough to initiate it. By virtue of their deliberative distance from the electorate, legislatures definitely fit the bill. But they are suited for this role in another sense as well. First, electoral accountability gives them a strong institutional incentive to do their best to explain the desirability of change to citizens. Second, it enables them to gauge the requirements of political prudence.

This, then, is the twin value of government by an elected legislature, and the two features of accountability and distance aim to sustain it. Can we operationalize this dual political concern, for popular support on the one hand and moral innovation on the other? Is it realistic to expect that legislators can balance the pull towards dependent judgment and the opposite pull towards independent judgment? Ronald Dworkin thinks that the set-up of elected legislatures only gives us a positive reason to assign to them what he calls policy or (in more recent work) choice-sensitive issues, such as whether to build a football stadium or an opera. But, he adds that it does not give us any positive reason to assign them choice-insensitive issues or issues of principle, among which he counts issues concerning the fundamental rights of citizens. Issues of the latter kind, he argues, we do well to assign to institutions that are not elected and decide on principle.³³ However, this suggestion does not factor into the division of labour the fact that legislators have strong institutional incentive to exercise independent judgment as well. Besides, it overlooks that very often, choice-sensitive issues have choice-insensitive aspects, and we want those who draw up comprehensive responses to choice-sensitive issues to be mindful of those aspects, so as better to incorporate them in their policies.³⁴ In general, we want all state officials, including legislators, to act with a sense of responsibility in matters of constitutional import. By virtue of their deliberative distance from the electorate, legislators can face up to this responsibility, even in the teeth of popular opinion.

Drawing on these observations, I propose we say, instead, that legislatures have the general initiative in the project of governing. On any given issue, whether choice-sensitive or choice-insensitive, legislatures must get there first, at least in the normal case. Through the various measures they enact, they give shape to a certain vision of the just and well-ordered polity. Of course, this vision will often be fragmented, since the way it unfolds depends heavily on the relative urgency and priority of certain issues and on other political circumstances. Hence, it is quite unlikely that the content of all legislative measures can neatly fall under a single scheme of the just and well-ordered polity. This will be partly because of the exigencies of the day and partly also because, on certain issues, it is permissible to change one’s mind. But, this is no cause for concern. Legal systems, as has been often pointed out, are ongoing enterprises. Different times pose different challenges, and political institutions must be flexible enough to adapt to changing circumstances. In this constant flux, citizens must have input. And elected legislatures are the institutions that are best suited to give them an appropriate voice.

The fact that the legislature has the initiative does not necessarily mean that it has the last word or that its decisions must prevail, come what may. Still, its power is very significant. First of all, the body that has the initiative gets to set the agenda according to its own sense of priority and political calculation. It can, thus, direct and frame political developments. Second, in the majority of cases, the policy initiated by the legislature will be the one that carries the day. As James Bradley Thayer put it,

the constitutions not merely intrust to the legislatures a preliminary determination of the question, but they contemplate that this determination may be the final one; for they secure no revision of it. It is only as litigation may spring up, and as the course of it may happen to raise the point of constitutionality, that any question for the courts can regularly emerge. It may be, then, that the mere legislative decision will accomplish results throughout the country of the profoundest importance before any judicial question can arise or be decided.³⁵

At any rate, a legislative decision provides a point of focus for any subsequent debate. Political decision-making is a painstaking process, especially in a democracy, so inertia inevitably favours a scheme already adopted. It makes it the starting point for any amendments and counter-proposals. Notice that these arguments apply not only to choice-sensitive but also to choice-insensitive issues, given the way the latter are, as mentioned above, often nested in choice-sensitive schemes and, conversely, given the choice-sensitive and other strategic implications of a judgment on a choice-insensitive issue. A good example of this is human rights legislation. Such legislation does more than just enact into law abstract moral precepts; rather, it attempts to embed human rights protection in a certain area of social life with its pre-existing regulatory framework, patterns of vested interests and settled expectations and so on. This is a delicate and challenging exercise, so the striking of a compromise that achieves the goal reasonably well and commands sufficient support constitutes a political achievement that is difficult to replicate. Scrapping the compromise and going back to the drawing board would be a highly unappealing prospect.

5. Judicial Independence

I have argued that we have good reason to be governed by state officials who are, on the one hand, accountable to the electorate and can, thus, fuse the views and interests of ordinary citizens in the policies they enact and, on the other, can deliberate independently about what ought to be done. This is why, in a democracy, we give the political initiative to periodically elected legislators. With this choice, however, comes a risk. It is the familiar risk we run whenever we submit to the judgment of another, but with a twist. In the case of legislators, the risk is the result of the leeway for the exercise of independent judgment that institutional design affords them.³⁶ It has two aspects. On the one hand, it reflects the possibility that their judgment concerning the political issues they decide will be mistaken. Legislators ought to bring a host of complex considerations to bear on their decision, and it is only to be expected that they will sometimes go astray. On the other hand, government by elected officials creates a special kind of risk, namely that, in making up her mind about how to cast her vote, an elected official will give undue weight to the views and interests of her constituents, in order to secure her re-election. As we have seen, our political practices create a deliberate distance between legislators and voters, so as to enable the former to filter, shape and, perhaps, also correct the views of the latter. Thus, a legislator who slavishly follows her constituents fails her role in a very important respect.

The risk that political power will not be used properly is one that the theory of separation of powers has always taken very seriously. This is why it tends to incorporate a checks-and-balances as well as a division of labour dimension. Given the risks identified in the preceding paragraph, separation of powers recommends that we establish mechanisms to address them. Of course, those risks would be minimized if we did away with elected legislatures altogether. But, because government by an elected legislature carries such immense value for us for the reasons already mentioned, we are prepared to accept the risks and mitigate them through a system of checks and balances rather than abolish or seriously alter the institution that generates them. This is, in fact, what modern democracies commonly do. And, they tend to rely on courts to provide the necessary check on legislative power, especially in fundamental rights cases.

It is common to defend this reliance on courts on the basis that judicial decision-making is principled and argumentative and is, thus, more apt to come up with the right answers to questions about the content of fundamental rights, which are essentially moral questions, not dependent on considerations of utility maximization and efficiency.³⁷ This line of argument has been criticized for failing to take seriously the constraints on judicial reasoning that come from the fact that it is a form of legal reasoning and is, therefore, sensitive to the gravitational force of past political decisions.³⁸ According to this criticism, it may well be that judges reason morally and that they are good at it; however, that cannot be an argument for assigning them the protection of fundamental rights, since in their reasoning, they ought to give weight to a host of—moral—considerations about precedent and statutory authority that are extraneous to the determination of the abstract meaning of such rights. In other words, even if judicial reasoning is principled, it is not oriented solely towards rights principles.³⁹

Although I believe that it is sometimes morally permissible, and even desirable, to establish a system of constitutional review for the protection of fundamental rights, I agree that the view just outlined, which stresses the similarities between judicial reasoning and abstract moral reasoning, perceives courts in isolation. It, thus, overlooks the relationship between courts and the legislature. In this section, I shall seek to locate constitutional review within that broader relationship. Hence, my case for constitutional review will start, somewhat paradoxically, with the claim that courts’ primary role is to assist in the implementation of government plans laid down by the legislature. For this reason, their task vis-à-vis the legislature is receptive; their judgments are constrained by what the legislature has said and done. At the same time, however, there is a supervisory function built in the judicial role. I shall argue that the assignment of such a function, which can vary in scope and intensity, is not at odds with the courts’ receptive function. Furthermore, it is supported by a feature that courts characteristically bring to the common project. This is their independence. I shall claim that, although independence strongly recommends courts as guardians against legislative excess, it does not provide an overriding reason for constitutional review.

A. The Duty to Respect the Legislature

Legislators, we have argued, hold a prominent place in legal systems. Accordingly, their decisions are entitled to a certain respect. How does this affect the judicial role?

Being centralized institutions, legislatures typically lack the machinery to give effect to their decisions in particular circumstances. This is a task that falls upon other state organs, one of which is the judiciary. Courts make sure that disputes arising in a legal system are resolved by appeal to the vision of the just and well-ordered polity that underlies legislative decisions. With their decisions, they sustain that vision over time and make it more precise. But, the assistance that they offer is not entirely passive. On the contrary, it has to be understood in dynamic terms. Thus, judges actively participate in the shaping of the vision of the just and well-ordered polity governing the life of their political community when they integrate the disparate and often conflicting decisions of the legislature in a more or less coherent whole.⁴⁰ They also assume an active role with the lapse of time, as they adapt legislative measures to fit changing social needs. Besides, novel circumstances sometimes discredit legislative decisions, in the sense that they undercut their status as expressions of popular will or as embodying superior expertise. Since judges assist in implementing the law over time, it can reasonably be supposed that they are better suited to appreciate new circumstances and incrementally update the vision of the just and well-ordered polity.⁴¹

The receptive aspect of the judicial function, thus construed, is intertwined with an element of supervision. In order to perform their receptive role, courts cannot take the contributions of legislatures at face value. Rather, they must decide which legislative decisions are genuine contributions to the joint project and how they ought to be construed. In short, they must decide whether a certain legislative decision merits their respect and their assistance in its implementation and what exactly respecting it amounts to. The stringency of the test that courts will apply to legislative decisions varies. Sometimes, the test will be minimally intrusive. It may take something like the following form: a decision commands respect just in case it has been voted by the appropriate legislative majority. The normative basis of such a test is a presumption of sorts: since the legislative process is set up in a way that, as a general matter, fosters the right kind of attitude towards public issues on the part of legislators, judges do well to treat the decisions that are produced through this process as embodying the value they were meant to; it would be counter-productive for judges to inquire further whether individual decisions actually do so.

At other times, the test legislative decisions have to pass focuses on the quality of the actual deliberation that led to them. As we have seen, the value of government by elected legislatures does not consist in the fact that everyone’s views count equally; they don’t. Rather, it consists in the fact that the views of those that count most, namely the legislators, are formed, subject to certain kinds of consideration. So, for instance, it has been argued by a number of liberal thinkers that rights provisions, such as the right to free speech and privacy, dictate, among other things, that legislators not enact the moralistic preferences of their constituents.⁴² On this view, it is the legislator’s duty, according to the constitution, to dismiss those preferences from the vision of a just and well-ordered polity that informs her vote. If, however, she takes them into account, she fails her role. Accordingly, the decision she has voted for is, in this respect, one that does not instantiate the value it was meant to. Judicial tests of the sort we are considering now aim to weed out such failures.⁴³ Judges will refuse to assist in the implementation of decisions that are thus tainted. But, they will not take the extra step of assessing whether the legislators have correctly weighed the considerations that they ought to have taken into account.

It is this extra step that a third, more intrusive type of test enjoins judges to take. We want legislators to make the best decisions regarding our collective life. We subject them to electoral pressure because we think that the best decisions must also be attuned in the right way to the views and interests of ordinary citizens. But, we know that legislators can, and do, make mistakes. Importantly, they may overestimate the pull of public opinion and, thus, get the balance between popular support and justice wrong. The third test gives judges the power to second-guess the correctness of the legislature’s judgment. Constitutional review of legislation for its compatibility with fundamental rights norms typically falls in this category.⁴⁴ In legal systems with a practice of constitutional review, judges who want to know whether they owe respect to a particular legislative decision must ascertain whether it adequately takes fundamental rights into consideration.

The last two tests reflect an interesting ambiguity in the concept of governing well. Sometimes, when we say we want to be governed well, we mean that we want to be governed by people who are public-spirited and conscientious and will consistently make a bona fide effort to decide according to our best interests. But sometimes, governing well just means governing in accordance with the precepts of good governance, making the right decisions. Thomas Scanlon captures these two meanings nicely when he draws a distinction between ‘the permissibility of an action and a special kind of agent assessment, in which what is being assessed is … the quality of the particular piece of decision making that led to the action in question’.⁴⁵ Scanlon notes that the agent assessment requires the prior identification of the considerations upon which the permissibility of an action depends. An agent fares well in that assessment if he has taken those considerations into account. This proximity explains why the two tests are not always clearly distinguished or easily distinguishable and often merge into one another. Still, in their ideal form, they are motivated by distinct moral concerns.

Clearly, there is a choice here. So the question arises: Are we ever justified to move from the less intrusive to the more intrusive type of judicial supervision? This question is especially pressing in the case of the last type of test: can we sensibly maintain that judges respect legislative initiative when, in fact, they second-guess the outcome of legislative decision-making? Doesn’t constitutional review actually hand over to non-elected judges the power to impose upon the people their own moral judgments? Saying that intrusive judicial supervision is a possible manifestation of the relationship between courts and the legislature does not answer these questions. We need to provide reasons why it is a morally attractive alternative or at least specify the circumstances under which it can be justified. Put differently, we need to defend the claim that, apart from a conceptual possibility, it is also a morally attractive possibility. It is to this task that I now turn.

B. Constitutional Review and Initiative

In this section, I shall argue that constitutional review does not negate the distinctive contribution that elected legislatures make to the project of governing. This is important, because we want constitutional review to fit in the broader partnership between courts and the legislature. If constitutional review ends up overshadowing or overly obstructing the valuable role we reserve the legislature in our political system, this will not be possible.

Constitutional review in its paradigmatic form is compatible with the assignment of an active role to the legislature by virtue of the fact that the role of the judge remains subsidiary to that of the legislature. Lawrence Sager has aptly compared constitutional judges with quality-control inspectors in an automobile plant. He writes:

… the quality-control inspector has only the job of assuring that the cars which leave her plant are well-built. Her role is focused and singular and comes on top of the efforts of the people who actually put the cars together. Constitutional judges are like that. Their mission is singular – to identify the fundamentals of political justice that are prominent and enduring in their constitutional regime and to measure legislation or other governmental acts by those standards. And their mission is redundant – they enter the process only after legislators have themselves considered the constitutional ramifications of proposals before them.⁴⁶

Subsidiarity assumes various institutional manifestations. I cannot go into a full account of this important concept here. For present purposes, a few characteristic examples will suffice. Generally speaking, constitutional review is prompted by claims of unconstitutionality directed against particular legislative decisions. Thus, it is typically reactive; it is triggered by legislative action. In an obvious sense, then, it is compatible with legislative initiative. The legislature will tackle an issue first and the court will assess the constitutional merits of that particular initiative. In addition, constitutional review gives judges, at most, the power to strike down or uphold a specific legislative decision.⁴⁷ But, it normally does not also give judges authority to rewrite the decision, except perhaps in very exceptional circumstances.⁴⁸ In this way, it vindicates the following division of labour: In a representative democracy, it is up to the legislature to design public policy by combining the need for popular support and moral innovation. It is also up to the legislature to reconsider and modify its decision, once it has been pointed out that, in its original form, it was not the product of the appropriate deliberative process or suffers from an important moral defect.

A third type of subsidiarity is exemplified by systems that allow judicial intervention only in cases where a piece of primary legislation is tainted by a particularly serious or clear-cut moral error. In fact, practices of constitutional review, in general, fall in this category, insofar as they empower courts to scrutinize legislation only against a limited set of substantive norms, namely, those that embody moral precepts of special urgency. We could envisage a system where the courts’ reviewing power is not limited in this way, but rather, extends to, say, unwise or inefficient legislative decisions. Although, strictly speaking, this kind of arrangement would not negate legislative initiative, as it would still allow the legislature to have a go at an issue before the court intervenes, it runs the risk of cancelling it out in practice; it would signal a shift of the centre of political power towards the courts and, at any rate, it would create a significant moral hazard: legislators would be less inclined to take their initiative seriously, knowing that their decisions will in any event be double-checked by the courts.⁴⁹ Although this is not a conclusive argument, it is one to which constitutional design ought to be sensitive.

C. Independence and checks-and-balances

So far, I have claimed, starting from the notion of the duty of respect that courts owe the legislature, that some form of constitutional review is reconcilable with legislative initiative. But, do we have any positive reason to want the establishment of a system of judicial review? It has been rightly pointed out that the fact that the legislature’s powers are limited does not entail that it is courts that should be entrusted the role of policing that the legislature observes those limits.⁵⁰ In fact, the mere existence of limits on the power of the legislature leaves it an open question whether we should impose any external check on the legislature in the first place.⁵¹

In this section, I shall defend the claim that courts have a comparative advantage over other state institutions that makes them particularly good candidates for the assignment of supervisory powers over the legislature. The analysis will highlight, as ground for this assignment, a pervasive feature of the institutional role of courts, namely their independence. This time, the supervisory element of the courts’ role will be addressed more explicitly. Needless to say, even if sound, my argument falls far short of a full justification of judicial review, since it is does not take sides as to the necessity of imposing an external check on the legislature in the abstract.

What characteristics should a supervisory body have in order successfully to discharge its function? Many, no doubt, but here I want to focus on just one. Famously, this characteristic is summarized by James Madison in his discussion of separation of powers in the Federalist Papers. He writes:

It is evident that each department should have a will of its own; and consequently be so constituted that the members of each should have as little agency as possible in the appointment of the members of the others … It is equally evident that the members of each department should be as little dependent as possible on those of the others, for the emoluments annexed to their offices.⁵²

Supervision, Madison is saying, loses its edge in constitutional systems that make it politically impossible for the supervising body to ever find against the primary decision-maker. If the decision-making body can illicitly influence the supervising body, if, that is, it has the capacity to impose its will on the supervising body by means of political pressure, the supervisory function will be compromised. Thus, the supervising body must lie at a certain distance from the decision-making body. How can we achieve this? The passage from the Federalist Papers contains two proposals: First, each branch must have as little as possible to do with the appointment of the members of the other branch. Second, members of each department must be as little dependent as possible on other departments for the emoluments annexed to their offices. Important though they may be, these proposals are not exhaustive of our options. Nor do they point to necessary criteria of autonomy. As I shall now argue, judicial independence can also deliver the requisite distance between courts and the legislature.

The argument from judicial independence rests on some familiar procedural characteristics of the court system. The ones commonly cited in the literature are security of office and fiscal independence.⁵³ Such safeguards aim to insulate individual judges from various kinds of illicit pressure by the political branches. Though never absolute (and structurally so),⁵⁴ this insulation seems to work most of the time, creating an institutional environment where a judge can be confident that she will not be sanctioned on account of her rulings and does not need to ingratiate herself with the powers that be for purposes of career progression.⁵⁵ Notice that, in the present context, judicial independence is not treated as an intrinsic value, which gives courts a special moral warrant to decide issues of fundamental right. But, it is instrumentally valuable, in the sense that it delivers the requisite distance from the legislature and, thus, makes courts pro tanto better at reviewing it.

Conversely, judges have a duty and are expected to decide free from certain biases. Thus, judges, whether appointed or elected, are, by and large, relieved from pressure by the electorate. Since they are not accountable to the people, they need not heed popular opinion in the same way that legislators ought to be sensitive to it to be re-elected, at least not in the normal case.⁵⁶ Equally, they are barred from invoking their party affiliation as a ground for their decision. The degree to which they adhere to this requirement is monitored by the reasons they give in support of their decisions. Indeed, reason-giving is a salient and oft-quoted aspect of the judicial function.⁵⁷ Note though that whereas many theorists have, on the basis of this feature, suggested that courts are best placed to decide certain questions, like fundamental rights, this is not the purpose for which I invoke it. For present purposes, I want to focus on the instrumental value of reason-giving as a control mechanism (no doubt, reason-giving also has intrinsic value).

At this point, it may be argued that judicial independence is a double-edged sword. If courts are unrestrained, they are as potent to impede worthwhile plans with which they disagree as to frustrate evil plans that are not entitled to any respect. At least, when legislators go wrong, they have to face public outrage in the next elections. By contrast, judges, so the argument goes, can ignore popular sentiment without serious backlash. No doubt, there is some truth in this fear, especially as it applies to higher courts, since the decisions of lower courts are, in the normal case, subject to internal constraints like appeal. But, the decisions of higher courts, it seems, are practically beyond reach, more so when they are fashioned as interpretations of the constitution that is superior to statutory law and, thus, can only be changed by formal constitutional amendment.

Nonetheless, the worry under consideration focuses overly on the fact that courts are insulated from external influence and overlooks the element of self-control inherent in the duty of courts to deliver reasoned judgments, the other side of independence. The existence of such a duty does not guarantee that judicial decisions will always be correct (nothing can guarantee that in the real world), but it does help imbue in judges the proper sense of institutional responsibility. For, whatever the decision they reach, they know that their reasoning can be put under scrutiny. Of course, the existence of the duty to give reasons does not detract from the finality of a mistaken decision by the higher court. However, it would be wrong to assume that the problem of finality applies only to Supreme Court decisions. Finality is a pervasive aspect of institutional design. And, that is because for every contribution to the joint institutional effort, at any level, the question can be raised as to how we can check the exercise of public power. And, in every case, there will be a countervailing consideration to settle on a decision and move on. No institutional design can accommodate both types of consideration to the full (since we are fallible, there is always room for yet another check). Of necessity then, legal systems strive to harmonize them as far as possible. The recent constitutional experiments—in the UK and elsewhere—with systems of constitutional review, where courts do not have the final say, are good examples of such efforts.

In other words, we have to regard the finality of certain judicial decisions for what it is, namely an institutional choice that aims to balance the requirements of checks-and-balances, on the one hand, and effective division of labour, on the other. It may have its defects but so will any alternative. It cannot be ruled out a priori. On the other hand, neither is it our only option. US-style constitutional review is not a default. Constitutional design can—and should—fine-tune finality just as it fine-tunes the scope and intensity of review. What it cannot do away with is judicial independence. Without it, there is little reason to entrust courts and, in the last instance, the higher courts, and not some other state institution, with the task of supervising the legislature. Still, judicial independence is just one among many considerations that bear on the moral legitimacy of judicial review. The legitimacy of a practice of constitutional review depends on other factors such as the resources of the legislature to check itself or the existence of other checks-and-balances mechanisms. Additionally, despite its many beneficial effects, judicial review comes at a cost both for the aggrieved party and for the institution that is supervised.⁵⁸ Our assessment of the moral merit of judicial review must be sensitive to such costs, and there is no reason in the abstract to think that the argument in favour of judicial review always or typically wins out. In summary, we could say that judicial independence comes close to a necessary condition for the moral legitimacy of a system of constitutional review, but it is far from being a sufficient one.⁵⁹

Ultimately, we must ascertain how well this or that role for the judiciary fits in the project of governing in a specific jurisdiction. Hence, our support for constitutional review cannot be unqualified. This may seem disappointing to those who prefer to see courts as the repository of constitutional principle or the pinnacle of public reason rather than as merely a body that enjoys a certain comparative advantage as a checks-and-balances mechanism. Nevertheless, we ought not to underestimate the conclusion we have reached so far. If, by virtue of its independence, the judiciary reliably and effectively performs a checks-and-balances function, it makes a valuable contribution to the project of governing. Therefore, it deserves the respect and assistance of its fellow-participants in that project, including the legislators, who would do well to leave judges do their job.

6. Conclusion

There are some tasks that we perform best when we perform them together. The task of governing is definitely one of them. The relationship between courts and the legislature illustrates this simple truth. We want to be governed by those bodies jointly. As I have tried to show, this is because a number of important principles of political morality and institutional design bear on and shape their relationship. I have argued that these principles give the legislature an active role in this project. In addition, they assign courts the task of assisting in the implementation of legislative initiatives. At the same time, they single courts out for the critical task of supervising the legislature. I have emphasized that this is no more than a modest recommendation; it does not render constitutional review morally mandatory. Also, it does not uniquely identify a model of constitutional review as appropriate for all times and places. It thus helps explain and vindicate the political efforts and constitutional choices of the many legal systems that have sought to combine in a workable scheme these political imperatives by calibrating the scope and intensity of constitutional review. In turn, our theories of constitutional review have a lot to learn from this focus on actual examples of institutional design. The persistent problems that their authors encountered and the enduring solutions that they devised are fixed points in our theoretical inquiries.

I presented previous versions of this article to audiences at Sheffield Law School, Warwick Law School, the University of Alberta, UCLA Law School and the Aristotelian University of Thessaloniki. I thank participants in all these events for their many helpful questions and comments. I am particularly indebted to Tawhida Ahmed, Arudra Burra, Stephen Gardbaum, Jonathan Gingerich, Brian Hutler, Tammy Hervey, Richard Kirkham, Matt Lewans, Filimon Peonidis and Mark Taylor for their detailed comments, oral or written, as well as to the referees of OJLS. Finally, I am grateful to Sheffield Law School for research support.