This article reflects on a group of constitutional devices: mechanisms that empower one state institution to defend itself against another institution. Rather than relying on a third body to police the relationship, these devices create the potential for a form of self-defence.Footnote 1 The institution is given a shield to protect against the attentions of another body, or is given a sword it can use to repel or deter an attack. Self-defence mechanisms are interesting for many reasons, but they are particularly interesting for the light they cast on the separation of powers. At first glance, these measures seem to run contrary to the normal prescriptions of that principle. They are instances where a capacity is allocated to a body that, on the surface at least, it is ill suited to possess. Understanding when and why the separation of powers requires these surprising allocations helps explain the operation of the principle in more routine contexts.
The article proceeds in three stages. The first section outlines various types of self-defence mechanism, placing them in two, rough, groups. On the one hand, there are negative mechanisms, devices that protect one institution from the attentions of another. On the other hand, there are positive mechanisms, devices that give an institution a weapon it can use against another constitutional body. The second section considers, briefly, the normal demands of the separation of powers and explains why self-defence mechanisms appear to run against the principle. The third section considers the two types of interaction that the separation of powers can create between institutions and officials: interaction characterised by cooperation and by friction. The article concludes that self-defence mechanisms are valuable because of the need for friction within the constitution. These mechanisms limit the ways in which disagreement between institutions can be expressed. Whilst it might seem that the normal demand of the separation of powers is set against self-defence mechanisms, a fuller understanding of its requirements shows that the principle does require the creation of devices of this kind.
I. Self-Defence Mechanisms
Self-defence mechanisms are immunities or powers conferred on a constitutional institution that fall outside of the normal requirements of the separation of powers, but which have the function of protecting that institution from other constitutional bodies. There are, then, two defining features that draw these mechanisms together: first, they share an ambiguous relationship to the principle of the separation of powers; and, secondly, they are identified by the function that they play in the constitution. The first element will be discussed at length later in the paper, after the principle of the separation of powers has been considered further. The second element – the functional aspect of these provisions – can be dealt with more briefly. Sometimes, self-defence mechanisms have been created in order to protect the institution; those crafting the constitutional rule have designed it with this end in mind. On other occasions, this capacity arises as incidental to powers or immunities given for some other reason. This capacity was not conferred with the aim of protecting the institution, but may nevertheless play this role within the constitution. If the capacity it confers is attractive, the mechanism may be said to have this function, even if it may not have been created for this purpose.Footnote 2 Or, to make the same point another way, whilst the conferral of the capacity was not a psychological reason for the mechanism's creation – it was not a reason in the mind of the creators – it remains a justificatory reason that supports the existence of the mechanism – a reason for us to want the mechanism to remain part of the constitutional order.Footnote 3
Self-defence mechanisms can be divided into two groups: negative devices that serve to protect one body from another, and positive devices that give institutions a weapon with which another body can be threatened.
A. Negative Self-Defence Devices
Many constitutions contain a cluster of immunities that serve to protect institutions, or officials within those institutions, from the unwarranted attentions of other bodies. These vary from state to state, but in many systems each of the three branches is given a degree of immunity from other bodies, or are given powers to regulate their own affairs that fall outside the normal allocation of powers in the constitution. For example:
• Debates in the legislature are insulated from parts of civil and criminal law. Libel actions, for instance, cannot be brought on the basis of statements made in the course of debate or which are contained in reports published by the chamber.Footnote 4
• Decisions of the legislature are not subject to judicial review on grounds of rationality and fairness, tests that the courts apply to other public bodies.Footnote 5
• The head of state is given immunity from prosecution whilst in office,Footnote 6 and/or limits are placed on the range of executive documents that the courts can require to be disclosed in the course of litigation.Footnote 7
• Decisions of the courts are protected from the scrutiny of the legislature.Footnote 8 Conventions, or legislative rules of procedure, prevent legislators from criticising decisions of judges.
• The funding of the courts, and the salary paid to the judges, is protected from the parsimony of other branches.Footnote 9 Whereas other public bodies require legislative approval for their funds, or are subject to financial control by the executive, the courts and judges are treated as special cases.
• Judges are accorded control over the appointment and promotion of other judges. In extreme cases, control of the process is almost entirely in the hands of the judiciary: the judges choose their own successors.Footnote 10
B. Positive Self-Defence Devices
Positive self-defence mechanisms give institutions a power that they can use against other constitutional bodies. Sometimes these powers are used quite frequently, but on other occasions these powers stand as threats: they are weapons that could cause a great deal of harm, perhaps even causing harm that goes beyond the body they are used against. That a body has the constitutional capacity to do a thing may, in itself, help ensure other bodies within the state show respect towards it. The risk of the power being exercised is enough, in itself, to achieve the purpose of that power. For example:
• The legislature controls the flow of money into the executive branch, and can deny the executive the funds it needs to operate.Footnote 11
• The legislature can impeach members of the executive branch.Footnote 12
• Where legislatures operate at different levels – in a federation, for instance – a lower legislature is empowered to repeal the acts of a legislature of a higher level that encroach on its area of competence,Footnote 13 or is given a power to force the higher body to reconsider its action.Footnote 14
• The executive possesses a veto over the acts produced by the legislature.Footnote 15 In a variant of this power, the head of state can compel the legislature to reconsider the constitutionality of a proposed measure or remit the question to a constitutional court.Footnote 16
• The executive is given a power to dissolve the legislature and call fresh elections.Footnote 17
• The executive (sometimes with the agreement of the legislature) can pack the court with new judges.Footnote 18
• The courts possess a power to strike down executive or legislative actions that interfere with access to the courts.Footnote 19
In addition to these powers, institutions may possess the capacity to behave in ways that run counter to some of the rules of the constitution.Footnote 20 Both the courts and the executive may enjoy powers that are outside, or are contrary to, rules of the constitution. Courts can decline to accept as legally valid instruments that are recognised as valid by the rules of the legal system. So, even if the law accords the legislature the power to make a statute, the judges could still refuse to uphold it.Footnote 21 Or sometimes judges can be asked to validate actions that lack a legal basis – as when judges are asked to rule on the constitutionality of a revolution. Even if the constitution explicitly denies them this power, making it clear there is no judicial jurisdiction to validate the acts of a usurper, the judges may still find themselves in a position in which they can exercise this power, with the usurper being willing, in varying degrees, to moderate her conduct to win over the court. On the other side, the executive branch can simply refuse to apply a judge's decision.Footnote 22 Whilst the court may appear to be in a constitutionally powerful position, in practice many courts are heavily dependant on the support of the executive for their effectiveness.Footnote 23
Sometimes exceptional powers like these should be seen as instances of self-defence mechanisms. On occasion, there is a constitutional rule that empowers institutions to act in this way.Footnote 24 It could be that courts' adjudication of the constitutionality of usurpation is generally recognised as a constitutionally legitimate part of their function, even if the previous regime had introduced rules that sought to prevent this occurring.Footnote 25 On the other hand, sometimes the exercise of these powers runs contrary to the constitution. The institutions can act in this way because of the position that the constitution places them in, but their actions are, nevertheless, unconstitutional.Footnote 26 An American president who refused to abide by a decision of the Supreme Court might get away with her defiance – but her actions would still amount to a breach of the constitution. When the act would be unconstitutional, the power no longer squares with our model of a self-defence mechanism; it is not a power granted by the constitution. But these practical, if unconstitutional, powers may play a similar role to their constitutional counterparts. They may still give the institution a threat it can use to ensure the respect and restraint of other constitutional bodies. These are exceptional powers that should only be used in exceptional cases. Their use may sometimes be justified, but this justification will turn on a broad range of considerations, many of which will depend on the particular circumstances surrounding their exercise.
The exercise of a positive self-defence mechanism comes at a price: there is a cost for the body against which it is exercised, but also a cost for the body that wields this power. Partly, this is a function of the cost to the system as a whole: where one institution acts against another, the whole constitution works less smoothly. Rather than expending energy advancing the public good, the institutions have turned inwards, spending their time squabbling. But it may also come at a direct cost to the institution that exercises the power: it has inhibited the functioning of the constitution. When Congress refuses to approve the budget, it ensures that the Executive will pay attention to its demands, but its intransigence may also lead to public criticism.Footnote 27 The level of the cost will vary greatly depending on the particular power and the context in which it is exercised. Some self-defence mechanisms are relatively modest – simply compelling a body to think again about its actions – whilst some are potentially very destructive. In addition, sometimes the exercise of the power will be uncontroversial, clearly needed to protect the institution, but sometimes it will be controversial, dividing the views of the country.
II. The Normal Demand of the Separation of Powers
The separation of powers is a principle that speaks to the institutional structure of the constitution. To appreciate why self-defence mechanisms should surprise us, it is necessary to locate these mechanisms within the context of the ordinary operation of this principle. This will also help clarify the classification of self-defence mechanisms: these are powers and immunities that fall outside of the normal requirements of the principle. Having explained the reason for surprise, the ground will have been readied for the construction of a more sophisticated interpretation of the separation of powers. This interpretation will explain why these capacities are unusual, but also why, in addition, they may be required by the principle in some situations.
Many discussions of the point of the separation of powers begin by identifying two objectives that may determine the aim of the principle: the protection of liberty and the promotion of efficiency.Footnote 28 These two are often regarded as rivals, competitors to be regarded as the sole or main animator of the principle – but, as we shall see, there is also a degree of compatibility between the two objectives.
A. Liberty Models of the Separation of Powers
The linkage of the separation of powers and liberty is a long-standing tenet of constitutional thought. One of the clearest modern statements of the connection, and a statement that is reflected in the work of many other authors, is found in the work of Eric Barendt. In “Separation of Powers and Constitutional Government”, Barendt argues that the purpose of separation of powers is to protect the liberty of the individual by making tyrannical and arbitrary state action more difficult. Power is divided between the branches of the constitution, with each element checking the others. This interpretation picks up on the well-known observation of Justice Brandeis who, in Myers v U.S, wrote that the purpose of separation of powers “was not to avoid friction, but, by means of the inevitable friction incident to the distribution of the governmental powers among three departments, to save the people from autocracy”.Footnote 29 Concerted state action is made more difficult by the existence of checks and balances between the various organs of state.Footnote 30
Liberty-models of the separation of powers probably capture the popular understanding of the principle outside of academia. The separation of powers acts as a brake on the state, protecting the individual from Leviathan. When the state acts against a person it needs each of the three branches to act in concert: the legislature to write the law, the executive to police it, and the courts to apply it in the particular case. By dividing up these institutions, tasks, and officials, the principle ensures that the exertion of state power needs the agreement of a number of agencies comprising a considerable number of people. This complexity and diversity creates the potential for, even the likelihood of, friction. And this friction, in its turn, protects liberty.
There are a number of difficulties with this simple liberty-model. First, it assumes that liberty and a strong state are inevitably opposed to each other, that the citizen can only be truly free within a state whose power for concerted action is limited by institutional conflict. Whilst there are some theorists who treat liberty in purely negative terms, as the absence of state constraint, most modern philosophers would also draw attention to the claims of positive liberty.Footnote 31 The state's task is not just to avoid placing unwarranted limits on the actions of individuals, but also to ensure that there are valuable options available to them.Footnote 32 A recognition of the value of state action is, perhaps, at the core of the modern welfare state.
Secondly, and developing this first point, even if liberty is understood in its crudest terms, threats to liberty come from bodies other than the state. One of the tasks of the state is to protect its citizens from harm, harm that can come from powerful bodies, such as other states or simply from other private actors. To undertake this task, the state needs to be able to act; it requires the institutional capacity to regulate the exercise of power within its territory. A constitution that created so much friction that state action became extremely difficult would prevent the state from protecting its citizens.
A successful constitution would be structured such that its institutions were restrained from encroaching on the private realm properly left to individuals, but would also be structured so that the state could protect its citizens and advance their interests. Even if the constitution contained areas of friction, it would also contain provisions that aimed to enhance and facilitate state action. These considerations point us towards a second collection of models of the separation of powers: those models that take efficiency as their guide.
B. Efficiency Models of the Separation of Powers
A second set of writers identified a different guiding purpose that may animate the separation of powers. Inspired, in part, by the resurgence of interest in institutional analysis in America,Footnote 33 an alternative approach to the separation of powers claimed that the point of the doctrine is to promote efficient state action by ensuring that powers are allocated to the institution best able to make use of those powers.Footnote 34
There is a long tradition of accounts that place efficiency, rather than just liberty, at the heart of the separation of powers.Footnote 35 Indeed, many of those who are often held up as paradigms of the liberty model of the principle recognized the need for effective, as well as limited, state action. James Madison, for example, noted the power of the doctrine to protect the people from tyrannical government,Footnote 36 but also recognised its capacity to enhance good government.Footnote 37 As one delegate at the Constitutional Convention in Philadelphia put it, when all the powers of state were vested in a single body ‘none of them can be used with advantage or effect.’Footnote 38
However, identifying efficiency as the core of the separation of powers is only the first step towards producing a satisfying account of that principle. ‘Efficiency’ is, in itself, an empty concept. Any constitutional order is efficient by some criterion; something is maximised, even if it is just waste, corruption, and folly. To be worthwhile, the stuff being maximised must be attractive.Footnote 39 We can only fully understand the demands of the separation of powers once we have understood the proper objectives of the state. Viewed in this way, the work of those who focus on liberty as the objective of separation of powers can be integrated with those who regard it as efficiency. The protection of liberty is one objective that the state should pursue, but there are other ends that the state ought also to seek.
I have discussed the broad objectives of the state elsewhere, and have argued that the primary purpose of the state is the advancement of the wellbeing of its members.Footnote 40 As part of achieving this goal, states require institutional structures that facilitate democratic and competent government. There need to be ways in which the people, the citizens of the state, can set the state's policies.Footnote 41 Allied to this, technical expertise needs to be integrated into the constitution. This is both to ensure that the policies set by the citizens can be carried out – that they do, in fact, achieve the goals they seek – but also to condition those goals, dissuading the democratic elements of the state from foolish or ill-considered projects.
Identifying democracy and competency as two of the necessary, though not sufficient, ingredients for a successful state gives us just enough to begin to explain the classic demands of the separation of powers. It explains why we need different types of institution, wielding different types of power, staffed by people with different talents and qualities. So, democratic government requires a representative institution that sets the broad laws of the state. It will need institutions that can carry forward these laws; making the policy aspirations contained in statutes a reality. Without such executive institutions, the legislature's statutes would not make the differences they aspire to make. Legislators are elected: chosen to represent the views and interests of the people. The executive branch, in contrast, will be characterised by its technical expertise and bureaucratic structure, even if its head is elected. Its officials will be appointed, and appointed on the basis of merit. Finally, for a country to be governed democratically and competently, the state needs a legal order through which to act. People need to know what the state requires of them. The state needs institutions that can authoritatively resolve disagreements about the law – both disputes that arise between private parties and disputes that arise between individuals and the executive branch.Footnote 42 The courts will be characterised by their legal technical expertise.
The last paragraph was deliberately superficial in its account of the institutional ordering of the state. There is much more that could be said about the connection between the purpose of the state and its constitution, but enough has been said to show that reflection on the point of the state both justifies and clarifies the principle of the separation of powers. There is a good explanation, grounded in the point of the state, as to why the state needs an array of institutional forms, why these institutions should exercise the powers they possess, and why different types of qualification should be required of the officers who act in these bodies. The usual demands of the separation of powers can be couched in terms of suitability. Institutions and powers should be created that are suitable to the achievement of the state's characteristic purpose. Powers should be allocated to those institutions that are most suitable for their exercise. The type of person chosen to act in these institutions should be suitable for the type of power the institution exercises.
C. Why Self-Defence Mechanisms Should Surprise Us
The previous section sought to show that, ordinarily, the separation of powers requires the creation of institutions that are able to act to advance the objectives of the state; it requires that powers be conferred, and officials appointed, to those bodies in such a way as to best assist the securing of these goals. This interpretation of the separation of powers, grounded in efficiency, was able to accommodate the insights of the liberty-focused accounts, whilst avoiding the flaws found in those versions of the principle. But efficiency models of the separation of powers bring a surprise: self-defence mechanisms appear to stand in opposition to the principle. These are instances where a body is barred from exercising a power that it appears suitable for it to exercise, or where a body is accorded a power that it appears ill-suited to use.
Reflecting on the devices discussed in the first section, several of them serve to inhibit the ordinary operation of the judicial function. A person may have been libelled in the legislature, wronged by the head of state, or they may be able to show a violation of their rights only by exposing the correspondence between the President and his advisors – and yet, because of a special immunity, the courts are unable to apply the law in the case before them. A significant part of the function of the court, as identified by the separation of powers, is to apply the law in disputes and to uphold people's legal rights. These immunities limit the ordinary operation of the court, and stop it achieving these goals.
Similarly, the executive may also find it is excluded from tasks that it is well-placed to undertake. For example, the allocation of resources within the state is a matter for the executive. It must determine what funds the various institutions of the state require, and is then accountable to the legislature for these decisions. This task is entrusted to the executive because the allocation of funds raises technical questions: the executive must try to gauge what each institution needs in order to achieve its goals, and weigh their competing claims. After allocating funding, the executive supervises its spending, checking that the institution is behaving in a financially sensible way. By insulating the courts from this process, the capacity of the executive to ensure that public monies are allocated effectively is impaired. The executive is inhibited from reviewing how the money is spent and limits are placed on its ability to ensure that funds are properly managed.
Other self-defence mechanisms accord constitutional bodies powers that they appear ill suited to exercise. For example, whilst it is common for the legislature to exercise control over the funding of the executive, the legislature lacks the technical skills to determine how much funding the executive needs, and to assess the consequences of restricting its resources. If the legislature denies the executive funds it prevents the executive branch from undertaking the tasks identified for it by the separation of powers: the executive is no longer able to carry forward the policies of the state. Similarly, even in presidential systems, the executive is less able than the legislature to articulate the wishes of the citizenry. Though the extent to which the legislature can be said to represent the electorate will vary, depending on the state studied and on our understanding of representation, the legislature has a capacity to express, and to an extent integrate, the views of the electorate that the executive cannot hope to match. Giving the executive the power to veto legislation or to dissolve the legislature gives it the power to frustrate the functioning of the legislature: the legislature is unable to achieve the goals set for it by the separation of powers.
Many of the powers discussed in this paper are commonly found in constitutions, and are of long-standing. Many of them seem eminently sensible, and, intuitively, appear to help the functioning of the constitution rather than harm it. In spite of this, the objective of this section has been to explain why we should be surprised by these rules: they appear to run against the normal demands of the separation of powers. They inhibit the ordinary operation of constitutional institutions, and give powers to bodies that they are ill suited to exercise. There are a number of ways we could respond to this conundrum. We could conclude – against all sense and experience – that these rules are all errors, and should be abandoned. Supporters – if there are any – of the ‘pure’ theory of separation of powers might argue for this end.Footnote 43 Then again, we could conclude that the separation of powers was a mistake; that it is a flawed constitutional principle, plausible in the Eighteenth Century, but no longer attractive.Footnote 44 Finally, we could re-assess the separation of powers in light of these rules. Perhaps a more sophisticated understanding of the principle would not merely incorporate the self-defence mechanisms but would explain why we need such devices, and explain the role they play in the constitution. To achieve this end, we need to consider the institutional dynamics of the separation of powers: the ways in which constitutional institutions do, and should, interact.
III. The Institutional Dynamics of the Separation of Powers: The Role of Friction
Often, perhaps ordinarily, the institutional interactions required by the separation of powers are characterised by comity and co-operation. These relate both to the aim and to the structure of a constitutional mechanism.Footnote 45 They relate to the aim, in that both institutions are presumed to be striving for a common objective. They relate to the structure, in that both institutions utilise their different institutional abilities to advance the objective.Footnote 46 For instance, when the legislature enacts a statute it identifies an objective and specifies a means to achieve that goal. Statutes are – rightly – drafted in broad terms. They are written to be read: they do not answer, and should not seek to answer, every possible question that might arise over their meaning.Footnote 47 The courts then resolve ambiguities that arise around these statutory provisions. Through interpretation, judges clarify the meaning of the statute in the context of particular cases. Sometimes, indeed, this can involve departing from the literal meaning of the text: the judge, who appreciates the impact of the law in practice, may modify the statute. When the constitution is working well, in both of these instances – interpretation and modification – the judge will share the objective of the legislature and act to advance this policy. There is comity of aim – the policy objective that is shared by courts and legislature – and comity of mechanism – with each constitutional institution using their different capacities to advance this policy.
A second type of interaction, and one that is more directly related to the subject matter of this paper, is characterised by friction. Once again, friction relates both to the objective and to the structure of the mechanism. In these instances, the institutions are straining against each other, set in tension rather than comity. There are at least two reasons why a constitution might include the potential for friction, and at least two forms these friction mechanisms may take. The two reasons for friction are: first, to guard against error and, secondly, to permit a justified limitation of the range of moral concern of an institution. The two forms that friction might take are: first, mechanisms that prevent an act being carried forward and, secondly, mechanisms that serve to combine acts of different bodies.
A. Friction and the Prevention of Error
Preventing error is the most common reason for creating friction within a constitutional order. The state needs to guard against mistaken or immoral decisions that constitutional institutions might make from time to time. There may be some temptations towards which institutions are, because of their structures, particularly vulnerable. For example, the electoral process may render the legislature at risk of the vices of majoritarianism. Appealing to prejudice might provide an easy way for legislators to win votes, and they may have little interest in protecting or representing unpopular groups. The courts are susceptible to a different collection of temptations. For instance, the judge may be tempted to make decisions on matters she knows little about. The trial process can generate limited and partial information about the law that is in dispute: the parties only have an incentive to give the judge information that supports their case. Sometimes, the judge will be tempted to make changes to the law on this partial information, unaware of the wider ramifications of her decision.Footnote 48
Just because constitutional institutions are vulnerable to such vices, it does not follow that they will invariably succumb to temptation. The need for limits on an institution to guard against error will depend largely on local circumstances. A state in which the legislature has a history of making oppressive laws that victimise minority groups may need a powerful court to guard against these abuses, whereas a state in which the courts have a history of supporting the elite against the bulk of the citizenry might require a different balance of power. The types of friction required in a constitution will therefore vary between states.
B. Friction and the Justified Limitation of Moral Reasoning
A second reason for creating friction within the constitution – and perhaps one that is of more general application – starts from a radically different position. Rather than preventing error, friction might sometimes allow for a valuable division of moral labour. Institutions may properly limit the range of factors they consider in making a decision because they know that another institution is on hand to weigh the considerations they have failed to assess. This licensed constitutional myopia may be valuable if it permits an institution to play to its strengths; to consider factors it is well-placed to consider, and leave other relevant issues to bodies better suited to their assessment. This creates a form of friction, rather than cooperation, because the institutions are pursuing different aspects of the common good and modifying the acts of another body in pursuit of these goals. It may be that this can occur without antagonism arising between the bodies, but – necessarily – the body whose measure is modified will be unable to fully appreciate the merits of the change.
Such an argument would present the separation of powers as requiring an invisible hand mechanism to be established by the constitution.Footnote 49 In a recent book, Adrian Vermeule discusses the operation of such systems, drawing attention to the interaction of these processes with law.Footnote 50 Invisible hand mechanisms operate when parties within a system are seeking one objective, or a number of objectives, but the system is structured such that a distinct benefit is produced through their interaction. The actors are not pursuing this good – indeed, they may be unaware of the operation of the mechanism – but the system generates this outcome. The classic example of an invisible hand mechanism is that of the market. Here, buyers and sellers seek to make a profit, but the market system produces a social benefit. When it works well, the market serves to allocate resources effectively, giving goods to those who value them most highly. Invisible hand mechanisms may be attractive because they can allow for a limitation of the reasons that agents should consider before acting. They provide a system that integrates or combines this partial reasoning to generate the outcome that an agent would have reached had she been able to successfully consider the totality of reasons that applied to her. So, in the absence of the market, agents should consider who would make best use of the resources they produce and attempt to gauge a fair price for these goods. The market, in its ideal form, would provide a more reliable answer to these questions than the agent would reach reasoning directly. Even an altruistic person would make better decisions by making use of the market, trading to make the biggest profit, than by making her own assessment of what the common good requires. Perhaps surprisingly, she is more likely to act on the balance of reasons that apply to her by ignoring some of them than by including them in her deliberations.
The market is not the only example of an invisible hand system; Vermeule moots a number of other possible situations that could be analysed in this way. Could the analysis be extended to constitutional institutions? On an invisible hand argument, the friction mechanisms of the separation of powers would no longer serve to stop institutions from acting wrongly, rather, they would enable institutions to narrow the range of moral considerations that apply to them. So, perhaps, the courts tend to be good at recognising the value of rights, and appreciating the impact of general legislation on those rights in particular contexts. The legislature, in contrast, might be good at gauging the wider public interest, and weighing the claims made by competing interest groups. If this is the case, it may make sense to place the legislature and courts into a system that ensures friction between these two institutions. Each fights for their own bit of the common good and, out of this conflict, the totality of the common good is achieved. Invisible hand systems of this type are attractive where there are good reasons to limit the range of considerations that apply to an actor or institution. In the example of the market, individuals are well-placed to assess their own interests but poorly placed to gauge the collective good. The same might be true in constitutional invisible hand systems. Whereas in mechanisms characterised by comity, each institution works towards a shared goal, now the institutions are working towards different, superficially conflicting, goals. Were each acting alone, such narrowness would be unjustifiable, but their presence in a system that secures friction advances the general good. Each institution is well-placed to pursue an aspect of the common good: the constitution combines their efforts into a collective decision that pursues the common good as a whole.
A number of writers on the constitution have advanced arguments that could be interpreted as versions of this model.Footnote 51 This might, for example, provide one way of understanding Ronald Dworkin's suggestive, but difficult, claim that courts are the forum of principle whilst the legislature is the forum of policy.Footnote 52 According to Dworkin, questions of policy go to the common good whilst questions of principle go to individual rights. One possible interpretation of Dworkin would take him as arguing for some sort of division of moral labour. Because of the different institutional strengths of the courts and the legislature, these bodies should focus on different aspects of state action. The legislature, which is elected and contains a wide range of opinions and interests, has the capacity to determine the policies that will best advance the common good of the community. The courts, whose officers are normally selected for their legal expertise and which focus on disputes between parties, have the skills and the information required to protect individual rights. The bulk of Dworkin's work locates the relationship within the first argument for friction, discussed above. The courts operate to prevent error, protecting individuals from the mistaken acts of the legislature. But Dworkin could also be read as arguing for a creative friction between these bodies. Statutes form part of the material from which the judge extracts the law. Statutes, like precedents, are given weight in Dworkin's account of interpretation. When the judge resolves a dispute over law, the existing legal materials shape and constrain her decision. The decision of the court is one that incorporates the policy decisions of the legislature whilst also modifying them in light of the principles, rights, that are also part of the law. Through the combination of these two institutions, reasons that relate to policy and those that relate to principle are considered, and integrated into the decisions of the state.
This interpretation of Dworkin is speculative. His divide between ‘principles’ and ‘policies’ is underdeveloped. It is not clear why, for instance, individual rights should not be thought of as part of the common good – and so constitute an element of the policy direction of the state.Footnote 53 Given that these elements are intertwined, the legislature and courts will be constrained to consider each set of factors in many instances. A clearer, and more developed, approach to the division of tasks between institutions is found in the writings of those inspired by the work of political scientists on institutional competence. A good example of this is found in Eoin Carolan's recent book, The New Separation of Powers.Footnote 54
Carolan argues different institutions in the constitution represent what he terms ‘constituent perspectives’Footnote 55. These ‘constituent perspectives’ are determined by the different types of impact that state policies can have. Any policy needs to be crafted to take into account three perspectives: first, the broad collective good; secondly, its local implications in the area in which the policy applies; and, thirdly, fairness to those individuals affected by the policy. The government is best placed to gauge the collective good, ensuring that policies should respond to a perceived public need or demand.Footnote 56 It acts through statute: broad rules that set the direction of the state. The administration applies these statutes, and can assess their impact on the groups to which they apply. The administration can then temper these statutes in light of their operation.Footnote 57 By using delegated legislation and soft law, the administration refines the statute's application. Finally, the courts focus on the fairness of this modified policy to a particular person in a given case.Footnote 58 The judge is well-placed to assess the impact of the policy on an individual, and so can assess its fairness and rationality in that particular instance. Implicit in Carolan's account is the claim that it would be a mistake for each of these three bodies to attempt to answer the questions addressed by the other bodies. They should focus on the issues they are well equipped to address. The state, as a whole, needs to consider all three areas, but it does so by parcelling out the tasks to these three bodies. One way of understanding Carolan's account would be as a form of invisible hand mechanism: the combination of the decisions of these bodies enables the state to respond to all the reasons that apply to it.
Whilst the arguments for a division of labour in the constitution are persuasive, they must be treated with caution. The analogy of the marketplace is an instructive one. The account of the market given in the previous paragraphs was an idealised one: the perfect, functioning, market. In the real world markets often malfunction. Those who engage with them should check, from time to time, that the market really is producing the common benefits that justify a restriction of moral reasoning. They ought also to be sensitive to special cases; unusual situations where, even in a well-functioning market, actors in the system should pay regard to considerations beyond profit. Similarly, in the constitution the justifiability of limited institutional reasoning will depend on the successful functioning of the whole system: where one institution is failing, the other must attempt to accommodate this. And even where the system is working well, there will be exceptional cases where other reasons must be considered by institutions. If a successful argument for an invisible hand within the separation of powers could be made, it would only permit institutions to narrow the range of reasons they consider in normal situations. In exceptional cases – which might arise comparatively frequently – such myopia would not be justified.
C. Friction and the Need for Self-Defence
Both of the arguments for friction explain why we should want to see conflict between institutions on some occasions. The first argument turned on the possibility of institutional error. Though all real-world constitutional orders must accommodate this risk, the way in which they do this, and the risks they ought to be sensitive to, will vary from state to state. The second argument was more ambitious. It turned on the limited institutional capacities of different parts of the constitution. Given that the separation of powers requires these differences, and these differences imply these limitations, this argument for the benefits of friction was a general one. More needs to be done to render this argument compelling, but if successful it would show that all constitutions should seek to create some form of division of labour in the development of state policy, a division that seeks to turn the limited capacities of institutions into complementary strengths.
The need for friction within the constitution explains the need for self-defence mechanisms. In each of the arguments for friction, a situation is created in which one institution goes against the wishes of another. In some instances, this challenge might be accepted. The institution that has had a measure stopped or modified may acknowledge that the other body has a point: it may recognise that it made a mistake, or appreciate the advantage of the other institution's contribution. But often structures that have the potential to create friction are included in the constitution precisely because it is likely that one institution will fail to appreciate the force of the considerations relied upon by the other body. This, in turn, makes it likely that the first institution will, from time to time, disagree with the intervention, and sometimes this disagreement will be expressed in strong terms.
The self-defence mechanisms of the constitution operate to limit the ways in which this disagreement can be expressed and also to limit the level of coercion one institution can exercise over another. Negative self-defence devices act as a shield: they prevent or make less likely certain types of interference. A legislature that disapproved of a modification to the law made by a court may be able to change the law relatively easily, but the constitution may inhibit legislators expressing criticisms of individual judges and may prevent the legislature cutting judicial pay. Positive self-defence mechanisms act as a sword: they provide a sanction or threat that one institution can use against another. An executive faced with a legislature that consistently refuses to support its policies may be able to call an election; a legislature that believes its executive has diverged too far from its mandate may be able to constrict the executive's funds. As was discussed earlier, these measures come at a price, and are powers that should be used with caution. Often, the bare risk of their exercise will be sufficient to ensure respect between the bodies, and tolerance in the face of disagreement.
IV. Conclusion
The benefits of friction within the constitution help to explain why constitutional self-defence devices are sometimes necessary, and have a place within an account of the separation of powers. Friction is needed both to guard against error and, also, to allow institutions to restrict the range of moral considerations they include in their decisions. This limitation of moral concern rests on one of the core ideas of the separation of powers: a body should only undertake tasks that it is well-suited to accomplish. By involving different institutions with different capacities in a single area, the constitution can act as an invisible hand system: permitting each institution to consider that aspect of the issue it is best placed to assess, and then integrating these different assessments into a single policy. That the constitution pits institutions against each other within this system may require that these institutions need some degree of protection from other bodies. In some areas they are supposed to be set in tension, and self-defence mechanisms help to ensure that the potential antagonism that is produced is not destructive.
