Adam Perry

Pardon powers are common but difficult to justify. A pardon power is, roughly, a power that is (a) possessed by a non-judicial official, (b) used to cancel legal liability to a criminal sanction in a particular case without thereby altering the law, and (c) unconstrained by law. So defined, pardon powers seem to be at odds with two constitutional principles. Contrary to the separation of powers, the pardon power gives to someone other than a judge a decisive role in determining criminal liability in particular cases. Contrary to the rule of law, the pardon power is neither constrained nor ruled by law. Here I provide a novel defence of pardon powers, in three steps. First, even an optimal legal code requires suboptimal results in some cases. In such cases there is a “gap” between what is required by the code we should have and what we should do absent that code. Adhering to the code in every gap case is undesirable: while it would avoid suboptimal results, it would create too much unpredictability. Deviating from the code in every gap case is also undesirable: while it would ensure predictability, it would produce too many suboptimal results. The best strategy is to adhere in some cases and to deviate in others, that is, to selectively deviate. The most important thing about selective deviation is that it leads to like cases being treated differently. This is the second step in the argument. Third, pardon powers are an appropriate vehicle for selective deviation. Selective deviation calls for a power that can be used to set aside the code without changing it. It calls for a power that can be used inconsistently. Moreover, it calls for a power that is not in the hands of a body institutionally disposed towards consistency in decision-making, such as the judiciary. That is, it requires a power with features (a)-(c) – a pardon power. This argument overcomes the various objections to pardon powers. Contrary to the objection from the separation of powers, selective deviation is a task better carried out by non-judicial officials. Contrary to the objection from the rule of law, a power of selective deviation must be unconstrained to preserve the law’s guiding function. Overall, the features of pardon powers that seemed to be in tension with the separation of powers and the rule of law turn out to be favoured by those same principles.

In an earlier Article, we disproved the three claims central to the dominant view in the study of constitutional conventions: that there is a shared “Commonwealth approach” to constitutional conventions; that Commonwealth courts will recognize and employ conventions but never enforce them; and that conventions are sharply distinguishable from rules of law. We drew from Canada, India and the United Kingdom to demonstrate that Commonwealth courts have recognized, employed and indeed also enforced conventions. In this Article, we turn from the descriptive to the normative, arguing, again in contrast with the dominant view, that Commonwealth courts sometimes should enforce conventions. We argue that courts should act as executors of the will and judgment of constitutional actors, and limit themselves to enforcing only power-shifting conventions: conventions which transfer power from those who have legal power to those who can legitimately wield it. In playing this role, judges uphold the legitimate allocation of power—legitimate, not according to judges, but according to constitutional actors themselves.

The norms of a legal system are relevant in deciding on what rights and duties people have within that system. But many norms which are not part of that system will nonetheless be relevant within it. These other norms may include the norms of foreign legal systems, games, clubs, contracts, grammar, arithmetic, unions, universities, and so on. What distinguishes the norms which are part of a legal system (“local norms”) from the norms merely relevant within it (“adopted norms”)? Where, in other words, are the law’s boundaries? There are three main answers in the literature. Matthew Kramer claims that adopted norms are under the control of external actors, whereas local norms are not. Scott Shapiro flips that idea on its head. Adopted norms are not created through the exercise of local powers, he says, whereas local norms are. Joseph Raz, meanwhile, distinguishes adopted and local norms based on the reasons officials have for relying on them. But there are obvious counterexamples to each of these answers. My own answer draws on a distinction in constitutional scholarship between direct and indirect relevance. Roughly, directly relevant norms suffice for some conclusion about people’s rights and duties. Indirectly relevant norms bear on the interpretation, applicability, or exact consequences of other norms. Whereas local norms are directly relevant, adopted norms are merely indirectly relevant. Thus, what distinguishes the two types of norms is the directness of their legal relevance. I conclude by showing that it makes a practical difference on which side of the boundary a norm falls.

The study of constitutional conventions is anchored in three assumptions that have so far remained largely unchallenged: that there is a shared “Commonwealth approach” to constitutional conventions; that Commonwealth courts will recognize and employ conventions but never enforce them; and that conventions are always distinguishable from rules of law. We overturn each of these assumptions in this Article. We argue that there is no such shared “Commonwealth approach” to the treatment of constitutional conventions, that Commonwealth courts do, in fact, enforce conventions, and that constitutional conventions have crystallized into law in a major Commonwealth jurisdiction. These insights disrupt much of what is foundational in the study of constitutional conventions.

Judges often give statutes strained interpretations. What makes an interpretation strained? What, if anything, justifies a strained interpretation? Is there always a point past which an interpretation would be too strained, or are there some interpretations which judges are entitled to adopt, whatever a statute says? I claim to answer these questions. Using the formula about conditional probability known as Bayes’ theorem, I offer a precise definition of a strained interpretation, and identify two reasons for strained interpretations. I claim that the case for strained interpretations is often stronger than is realised, and illustrate that claim with a novel defence of the House of Lords’ decision in Anisminic Ltd. v Foreign Compensation Commission. Finally, I suggest that even if a legislature is sovereign, there may be meanings which a legislature is unable to enact into law, and thus there may be limits on even a sovereign legislature’s powers.

I develop a general analysis of mercy, and apply that analysis to pardons and lenient judicial sentences. Three features of my analysis are noteworthy. First, almost all existing analyses say that mercy is unconstrained in a normative sense, but I argue that mercy is unconstrained in the way that arbitrary power is unconstrained. Second, although it’s often assumed that mercy must be motivated by compassion, I show that mercy only requires acting with the intention to benefit the recipient. Third, my analysis says that mercy requires the giver of mercy to overcome a motivation to treat the recipient harshly. Given this analysis, there is a much stronger case for thinking of pardons as merciful than lenient judicial sentences.

Some people who seek leave to apply for judicial review do not have a private interest in the matter to which their application relates; instead, they seek “public interest standing”. To be entitled to public interest standing, a claimant must bring the right sort of case. She must also be the right sort of person: she must have the right motive, the right background, the right knowledge and understanding, the right reputation, and so on. The relevance of a claimant’s case to public interest standing is obvious, but the identity of the claimant is not. Here we provide a rational reconstruction of the test for public interest standing, according to which civic virtue is essential for a claimant to obtain leave in the public interest. Our account explains the bulk of the case law on public interest standing, on the one hand, and goes some way towards justifying it, on the other.

Many scholars think that all conventions of the constitution emerge from the bottom up, out of the practices of constitutional actors. Here we develop the first systematic account of conventions that are imposed from the top down, through prescriptions by constitutional actors. We show that “top-down conventions” (as we term them) can be created through the use of normative rule-making powers; that powers of the right kind are sometimes conferred by “bottom-up conventions”; that these powers are often exercised; and, as a result, that top-down conventions are increasingly common. We show, too, that as the formal, systematic, and intentional products of a small number of constitutional actors, top-down conventions are a potentially illegitimate form of constitutional regulation.

In recent years, British courts have treated constitutional statutes differently from ordinary statutes. This article sets outs to explain: (1) how courts have treated constitutional statutes differently from ordinary statute; (2) what a constitutional statute is; and (3) why constitutional statutes should be treated differently from ordinary statutes. Courts have made it harder for ordinary statutes to repeal constitutional statutes by implication, and easier for constitutional statutes to repeal ordinary statutes by implication. A constitutional statute is a statute which regulates state institutions, and which possesses importance of a particular type that we describe. The nature of a constitutional statute largely – but not entirely – justifies the special treatment they have been given. These conclusions have wider implications, including for proposals to codify the British constitution.

The Indian Constitution gives the executive the power to pardon offenders. The Supreme Court of India has repeatedly held that the executive must not grant or refuse pardons capriciously or arbitrarily. I raise two worries about this line of authority. The executive has its pardon powers in part so that it can show mercy. In one sense, mercy is about doing justice despite what the law requires. Mercy in this sense is a threat to the rule of law, and caprice is the price of containing that threat. In another sense, mercy is about showing compassion despite what justice requires. Mercy in this sense is an act of grace, and caprice is the price of it being possible at all. By seeking to eliminate caprice, the Court has risked undermining the rule of law, and frustrating one of the aims of the Constitution.

Administrative officials are permitted to have policies as to the exercise of their discretionary powers, but those policies must be flexible, not rigid. The “flexibility rule”, as I call it here, is nearly a century old. Over time, it has become part of the furniture of judicial review: often used, rarely examined. That neglect has led to confusion, on display most recently in R (Sandiford) v Secretary of State for Foreign and Commonwealth Affairs, in which the Supreme Court appeared to misunderstand key aspects of the flexibility rule. In this article I try to put the flexibility rule back on a sound footing. I argue, first, that the flexibility rule requires authorities to treat policies merely as rules of thumb. Second, the primary justification for the flexibility rule is neither legislative intent (as claimed in Sandiford), nor the avoidance of error (as commentators tend to assume); it is the value of participation. Third, and as a result, the flexibility rule ought to apply to policies governing the use of prerogative and other non-statutory powers (again contra Sandiford), as well as to policies governing the use of statutory powers.

In addition to its statutory and prerogative powers, the Crown has extensive administrative powers. The Crown’s administrative powers range from the power to form contracts to the power to circulate written material, and include powers to make ex gratia payments, convey property, and create policies. Much of the ordinary business of government falls under the Crown’s administrative powers, yet these powers are poorly understood. There are two existing accounts of the Crown’s administrative powers, but I show that both are unsound. I set out a better account, according to which the Crown’s administrative powers are of two sharply different types: (i) legal powers, which derive from the common law, and which extend to what a natural person can do and what the law permits; and (ii) non-legal powers, which stem from wide social recognition, and which extend beyond what a natural person can do or what the law permits. This ‘twofold account’ suggests several new ways of distinguishing the Crown’s administrative powers from its prerogative powers. It also suggests that the Crown’s administrative powers pose an unusual and serious threat to the rule of law.

One of HLA Hart's main contributions to jurisprudence is his theory of social rules. Hart said, essentially, that a social rule exists if the members of a society act in some way and have a certain attitude. Most legal philosophers think that Hart's account of this attitude is too general, however, and that his theory is overinclusive as a result. In this article, I draw on recent work in the philosophy of action to propose a more precise account of the relevant attitude, and a revised version of Hart's theory. I claim that a social rule exists if the members of a society act in some way and ‘accept’ they ought to act this way, independent of their beliefs about the matter. One of the implications of this idea is that the attitude that underlies a social rule is closely analogous to a presumption or fiction about what ought to be done.

The doctrine of legitimate expectations is a developing area of administrative law, and many issues remain outstanding, including the grounds of legitimate expectations. Promises, practices, and policies are the recognized grounds of legitimate expectations, but what is special about them? Why do they and only they give rise to legitimate expectations? The lack of an obvious answer has led some commentators to worry that the doctrine is ultimately incoherent. In this paper, we argue, first, that promises, practices, and policies each comprise or make applicable a certain type of rule, and second, that having a legitimate expectation is a matter of such a rule binding a public body to act in some way. This rule-based account provides a unified, economical explanation of the grounds of legitimate expectations that also sets the doctrine on a coherent foundation.

The British constitution is famously unentrenched: a law is not more difficult to alter or override simply because it is a law of the constitution. That may be about to change. In the largely overlooked 2012 case of H v Lord Advocate, the Supreme Court repeatedly said that the Scotland Act 1998 cannot be impliedly repealed due to its ‘fundamental constitutional’ status. These remarks were obiter dicta, but they reflect the considered view of the Supreme Court, and as such strongly suggest the path the law will take. Courts in the future are likely to treat constitutional statutes, like the Scotland Act, as susceptible to express repeal, but exempt from implied repeal. That would make constitutional statutes 'quasi-entrenched'. In this article we argue that, as a judicial innovation, the quasi-entrenchment of constitutional statutes lacks a sound legal basis. Parliament can make its intention to repeal a constitutional statute clear without making it express, and judges cannot, on their own initiative, ignore Parliament's clear decision to repeal even a constitutional statute. We conclude by identifying three types of situations in which constitutional statutes should be recognised as having been impliedly repealed.

The Upper Tribunal (Administrative Appeals Chamber) held in Evans v Information Commissioner that certain correspondence between Prince Charles and government officials must be disclosed under freedom of information legislation. Much of the judgment was devoted to a discussion of the constitutional conventions applicable to Prince Charles, and the case provides a useful example of how conventions and laws can interact. In this note, I argue that the Upper Tribunal misunderstood how conventions are distinguished from one another, and misapplied the test for the identification of conventions.

There is a consensus among courts and commentators on the grounds for a duty to give reasons in administrative law. Traditionally, the duty has been justified as a way to promote good decision-making, show respect for the parties, and reveal potential grounds for a challenge to the decision. We argue that this traditional picture is incomplete, because it omits two important considerations that favour giving reasons. Giving reasons can provide evidence that a decision-maker is a relative expert, and thus provide evidence of a reason to defer to the decision-maker. Giving reasons can also indicate a decision-maker’s findings on specific issues, making it possible for an appellate or reviewing body to selectively defer to it. Together these points show that concerns of deference are relevant to when administrative decision-makers should give reasons.