I. Introduction
How can a court review administrative action, without trespassing on the administrative decision maker's authority? No one thinks that the courts should substitute their judgment on all matters, intervening in administrative decision making merely because they think a decision is wrong. We would not be better governed if judges in judicial review had the final say over all governmental decisions. But when should the courts substitute their judgment? Which decisions are better settled by judges?
One standard way of framing the answer to these questions is to say that the courts should only intervene to correct “jurisdictional” errors, and should leave decisions about the “merits of the case” to the administrative decision maker. But that simply raises the question: what counts as a jurisdictional error? This question is a source of much confusion and disagreement among public lawyers. Many resist drawing the distinction between jurisdiction and merits at all, viewing it as unduly “formalistic” or “analytical”, which at best has little practical use and at worst obscures the real issues.Footnote 1 Formalism is not necessarily a bad thing, for there is value in having authoritative legal rules that can be applied without the need for the moral and political evaluation that the rules were meant to settle.Footnote 2 But I think it is a misunderstanding of the concept of jurisdiction to criticise or defend it on the basis of its alleged formalism.
Jurisdiction should be recognised as the core concept in judicial review of administrative action. This article explains what jurisdiction is and why it matters. The article adopts an analytical approach, but one that sees analytical reasoning as only a preliminary stage in identifying jurisdictional errors, which is an exercise in practical reasoning. The main argument is that jurisdiction should be understood as the scope of the legitimate authority of the decision maker. To explain what that means, the article draws on Joseph Raz's influential (but, among public lawyers, much neglected) work in analytical philosophy on the nature of authority. On this understanding, the question of jurisdiction in administrative law is: over which matters would the court be more likely to conform with reason if it treated the administrative agency's decision as binding than if it followed its own judgment? This understanding has important consequences for administrative law, three of which are worth highlighting at the outset.
First, jurisdiction is not determined merely by interpreting authoritative legal sources. Acting in excess of jurisdiction, or “ultra vires” (beyond the powers), is sometimes understood to mean “without legal authorisation”,Footnote 3 and in relation to statutory powers, “beyond the powers that Parliament intended to confer”. But authoritative legal sources do not fully determine jurisdiction; they do not specify all the limits of a decision maker's powers. To work out what those unspecified limits are, this article argues, we need to determine the limits of the decision maker's legitimate authority. The distinction between legal authority and legitimate authority – and so between error of law and excess of jurisdiction – explains why the distinction between jurisdictional and non-jurisdictional questions of law, although unpopular, is intelligible and important.
Second, intra-vires decisions of a legitimate authority are morally binding on their subjects, including the courts,Footnote 4 and the law should reflect this. That is, when a legitimate authority acts within its jurisdiction, its decision “pre-empts” the court's judgment of the merits of the case. I explain that claim more fully below, but for the moment suffice it to say that the reasons for a court to defer to a decision maker's legitimate authority are not reasons that should be “accorded appropriate weight” and added to the balance with all the other reasons.Footnote 5 They are reasons that can be overridden only if, on some particular issue, the decision maker no longer has legitimate authority over that issue, and instead the court has a greater claim to legitimate authority. While the courts and commentators often pay lip service to “due deference” or similar notions of respect for authority, they mostly do not sufficiently respect the pre-emptiveness of legitimate authority.
Third, the standard of reasonableness – even with the high threshold that Lord Greene M.R. tried to articulate in the notorious Wednesbury case – cannot be a legitimate test for determining when a court should intervene in administrative decision making; that is, it cannot be a test for jurisdiction.Footnote 6 It is widely accepted, as it must be accepted, that reasonableness review involves some assessment of the merits of the decision. Over recent decades, much ink has been spilt debating the extent to which there should be a variable intensity of review of the merits, usually said to range from “Wednesbury reasonableness”, to a more intensive reasonableness standard, to proportionality and correctness.Footnote 7 But that debate largely misses the point. These standards all entail the same thing: that a court should intervene when the decision maker acts for a defeated reason (i.e. acts unreasonably, disproportionately, or incorrectly), and not when the decision maker acts for an undefeated reason (bearing in mind that there will often be undefeated reasons for a range of different decisions, all of which are therefore reasonable, and none of which is uniquely correct).Footnote 8 But, according to the theory of legitimate authority defended in this article, when the decision maker has legitimate authority over the courts, the courts should not be in the business of deciding whether the decision maker has acted for a defeated reason (unless, perhaps, the decision maker has made a clear mistake).Footnote 9
The structure of this article is as follows. Section II examines, in some detail, a case that highlights many aspects of the problem of jurisdiction in judicial review, which provides a concrete example that can be used in the more abstract discussion that follows. Section III presents the core of the argument of this article, building on Raz's theory of legitimate authority to explain how jurisdiction should be determined and the pre-emptive effect of legitimate authority. Section IV examines the consequences of this argument for judicial review of error of law. Finally, Section V examines the consequences for reasonableness review.
II. Corner House as a Case Study
The problem of jurisdiction – and so the central problem of judicial review – can be best illustrated by analysing a concrete case, to which I can refer throughout the more theoretical discussion in this article. The case I have chosen is R. (Corner House Research) v Director of the Serious Fraud Office.Footnote 10 In July 2004, the Director of the Serious Fraud Office, exercising a power conferred on him by s. 1(3) of the Criminal Justice Act 1987,Footnote 11 started a criminal investigation into allegations of corruption against BAE Systems plc, in connection with a valuable arms contract between the UK and Saudi Arabia. The Director resisted protests from BAE and the Government that the investigation should be stopped because it would jeopardise the UK's commercial and diplomatic relations with Saudi Arabia. However, in late 2006, the UK Government – including the prime minister personally – warned the Director and the Attorney General that Saudi officials were threatening to withdraw counterterrorism cooperation with the UK if the investigation continued, which, the Government alleged, would have severe consequences for the UK's national security: “British lives on British streets were at risk”. In view of these threats, on 14 December 2006, the Director announced his decision to stop the investigation.
The claimants sought judicial review of the decision to stop the investigation. They were initially successful in the Divisional Court, but ultimately lost in the House of Lords. As we shall see, the reasoning of the judges in both courts was insufficient to justify their respective conclusions, and their inadequacies are instructive for our purpose in examining the problem of jurisdiction.
The Divisional Court's judgment, written by Moses L.J., emphasised the threat to the integrity of the UK's criminal justice system and the rule of law caused by surrendering to the threat from the Saudis. Moses L.J. conceded that the Director had a very wide discretion when deciding whether to investigate, and the courts should be most reluctant to interfere.Footnote 12 In exercising his discretion, the Director was entitled to take into account the risk to national security.Footnote 13 Neither the Director nor the court was in a position to reach an independent judgment on the risk to national security, and so the Director “may lawfully accord appropriate weight to the judgment of those with responsibility for national security who have direct access to sources of intelligence unavailable to him”.Footnote 14
But the fact that the threat was not only to national security, but also to the legal system, entailed, according to the Divisional Court, that “the issue is no longer a matter only for Government” but also for the courts.Footnote 15 Relying on what he regarded as “well settled principles of public law”, Moses L.J. stated: “The rationale for the court's intervention is its responsibility to protect the rule of law.”Footnote 16 More concretely, two main reasons for the court's intervention were given. One of these reasons was that, notwithstanding the need to accord “appropriate weight” to the judgment of others, the Director had “surrendered” his judgment to a third party: “In yielding to the threat, the Director ceased to exercise the power to make the judgment conferred on him by Parliament.”Footnote 17 But that is unpersuasive: there is simply no basis for concluding that the Director did not himself come to the judgment that the risk to national security (as explained by those with knowledge of that risk) outweighed the public interest in continuing the investigation.
The other, potentially compelling, reason for the Divisional Court's intervention was its claim that the Director and the Government had yielded too readily to the threat and sacrificed the rule of law, and that it was for the court to make that assessment.Footnote 18 Relatedly, Moses L.J. also criticised the Director and the Government for “never properly” taking into account a relevant consideration, namely, the damage to the rule of law caused by giving in to threats; however, the better description is that the Director did take that consideration into account, even though he regarded it as having been outweighed by competing considerations.Footnote 19 The main reason for the Divisional Court's judgment is that, in its view, the rule of law should be sacrificed only if the court is satisfied that it is necessary. Moses L.J. “identified” the following principle: “submission to a threat is lawful only when it is demonstrated to a court that there was no alternative course open to the decision-maker”.Footnote 20 The Government could not point to any “specific, immediate threat” to the life of anyone, and the Divisional Court was not satisfied that the Director and the Government had done all that could reasonably be done to resist the threat.Footnote 21
The flaw in the Divisional Court's reasoning is that it did not adequately explain why the threat to the rule of law justified the court's intervention. To be justified, the court needed – but failed – to explain why the court's decision should be substituted for both the judgment of the Director, to whom Parliament allocated responsibility, and the judgment of the Government, to whom “[i]n a case touching foreign relations and national security the duty of decision on the merits is assigned”, as Moses L.J. put it, by the “separation of power between the executive and the courts”.Footnote 22 While expressing scepticism about the alleged risk to national security,Footnote 23 Moses L.J. acknowledged that the court could not assess that risk.Footnote 24 But given that acknowledgment, why did he reach the conclusion that the court could better assess whether an alternative course was available and, moreover, whether that alternative should be taken by the Director? As we will see, the argument of this article entails that the Divisional Court's decision could be justified only if protecting the rule of law is so important that it matters more than getting the national security assessment right.
The House of Lords unanimously overturned the Divisional Court's judgment. According to Lord Bingham of Cornhill, the principle relied on by the Divisional Court – the principle “that submission to a threat is lawful only when it is demonstrated to a court that there was no alternative course open to the decision-maker” – was “novel and unsupported by authority”.Footnote 25 Moreover, the principle was objectionable because:
it distracts attention from what, applying well-settled principles of public law, was the right question: whether, in deciding that the public interest in pursuing an important investigation into alleged bribery was outweighed by the public interest in protecting the lives of British citizens, the Director made a decision outside the lawful bounds of the discretion entrusted to him by Parliament.Footnote 26
Baroness Hale of Richmond said that the “only question” for the court was whether it was lawful for the Director to take into account the Saudis’ threat and the resulting risk to British lives. “Put like that,” Baroness Hale explained, “it is difficult to reach any other conclusion than that it was indeed lawful for him to take this into account.”Footnote 27 The Director had also given “prolonged and profound thought to the implications for the rule of law” in giving in to the threat.Footnote 28 The House of Lords held that it was not for the court, but for the Director on the advice of others with expert knowledge, to assess the balance of all the relevant considerations.
The reasoning of the House of Lords was inadequate because it did not seriously engage with the claim that the rule of law deserved special protection from the court in the face of the threat, such that the court could substitute its judgment on the necessity of submitting to the threat, notwithstanding the “appropriate weight” given to the judgment of others, including those with greater knowledge of the threat to lives. Lord Bingham went so far as to say that upholding the Director's decision “involves no affront to the rule of law”,Footnote 29 which overlooks the sound point that the damage to the rule of law was one of the considerations against stopping the investigation, which the Director took into account. If the deficiencies in the reasoning of both courts are remedied, what should have been the court's decision? The purpose of this article is not to answer that question – though it will make the Divisional Court's judgment more difficult to justify – but rather to explain the analytical framework in which it ought to be answered.
III. Legitimate Authority
An agency's jurisdiction is the scope of its authority to decide. The main argument of this article is that a decision should be understood as being within an administrative decision maker's jurisdiction if that decision maker has legitimate authority to decide. The question then becomes: what is legitimate authority? Joseph Raz's work in analytical philosophy has provided one of the most influential and compelling answers to that question. Yet the influence of Raz's theory has not reached most public lawyers. While public lawyers have often engaged with (though have less often accepted) Raz's writings on the ideal of the rule of law,Footnote 30 they have taken insufficient advantage of the insights in his work on legitimate authority.Footnote 31 These two strands of his work, on the rule of law and on legitimate authority, stand on their own and it would be wrong to reject one because of the other. The aim of this section is to show that, even if Raz's theory of legitimate authority may be, in parts, underdeveloped or flawed, it is nonetheless one of the richest, most developed sources from which to reflect further on the notions of legitimate authority and jurisdiction.
A. The Justification of Authority
Raz began his exploration of the idea of authority by highlighting the apparent paradox of legitimate authority, the alleged incompatibility between reason and authority. He wrote:
To be subjected to authority, it is argued, is incompatible with reason, for reason requires that one should always act on the balance of reasons of which one is aware. It is of the nature of authority that it requires submission even when one thinks that what is required is against reason. Therefore, submission to authority is irrational.Footnote 32
Raz set out to explain why, despite this apparent paradox, reason and authority are compatible, and legitimate authority is possible: there can be reasons not to act on one's own judgment of the balance of reasons, and instead to follow another's judgment.
Raz's theory of authority, which he called the “service conception of authority”, received its fullest discussion in The Morality of Freedom. There, he set out three conditions that must be satisfied for an authority's claim to legitimacy to be justified. The first condition is the “dependence thesis”, according to which “all authoritative directives should be based on reasons which already independently apply to the subjects of the directives and are relevant to their action in the circumstances covered by the directive”.Footnote 33 Raz calls these reasons “dependent reasons” – or, in the language of judicial review, we might call them relevant considerations and proper purposes. In Corner House, the relevant considerations included the public interests in protecting the rule of law and public safety, but if the threat had been directed to the Director's personal safety, his personal safety would have been an irrelevant consideration, as were commercial interests.Footnote 34 The authority must act on the same considerations that its subjects should have relied on if left to make the decision themselves. That is not to say that authorities should necessarily act in the interests of their subjects. They should act on the basis of what it is objectively valuable for their subjects to do, whether or not that coincides with their interests.Footnote 35
Raz's second condition of legitimate authority is the “normal justification thesis”:
The normal and primary way to establish that a person should be acknowledged to have authority over another person involves showing that the alleged subject is likely better to comply with reasons which apply to him (other than the alleged authoritative directives) if he accepts the directives of the alleged authority as authoritatively binding, and tries to follow them, than if he tries to follow the reasons which apply to him directly.Footnote 36
These two conditions explain why Raz calls his theory a service conception of authority: a legitimate authority provides a service to its subjects by helping them to conform better with reason than they would otherwise have done if they followed their own judgment.
The subjects of an authority include all those who would better conform with reason by treating the authority's decision as authoritatively binding. In the context of judicial review, where the courts would better conform with reason by following the decision of an administrative official, the courts are among the subjects of that official's legitimate authority – a claim about which I will say more at the end of this section. In Corner House, both the Director and the courts were subject to the legitimate authority of the Government in respect of the risk to public safety caused by the threat, if the Government had better knowledge of that risk, and the Director and the courts would therefore comply better with reason by following the Government's assessment of the risk to public safety. But that would not entail that the Government had legitimate authority on the balance to be struck between public safety and other values. One consequence of Raz's theory is that legitimate authority is piecemeal. A legitimate authority will help some people conform better with reason on some matters, namely those matters regarding which the normal justification condition is satisfied.
The normal justification thesis faces an objection that is especially important when thinking about its relevance to administrative law. The objection comes from those who argue that the normal justification thesis is concerned only with outcomes – with whether the authority serves the instrumental value of getting the right answer – and not with procedures, the way in which the outcomes are reached.Footnote 37 Those making this objection point out, rightly, that sometimes the way in which an institutional settlement has been arrived at – say, through publicly recognised procedures – matters more than getting the settlement right. If it were true that we should be concerned only with outcomes, then in Corner House the fact that s. 1(3) of the Criminal Justice Act 1987 empowered the Director of the Serious Fraud Office to decide whether to investigate would be irrelevant; all that would matter for the court in judicial review would be whether the Director or the court is better able to reach the right decision. A judge would not need to be concerned with the settled legal allocation of power among institutions and the procedures, including the democratic legitimacy of the procedures, that empower and constrain the administrative decision maker.
This proceduralist objection is based on a misunderstanding of the normal justification thesis. The objection takes an overly restrictive view of what counts as dependent reasons in the Razian account, so that they exclude the values that procedures can secure. But Raz's service conception of authority can accommodate these procedural considerations.Footnote 38 For example, although he adopts a critical perspective on democratic institutions, Raz accepts that democratic institutions are capable of having an intrinsic value that can satisfy the normal justification condition, a value arising from “their ability to give expression to people's standing as free, autonomous agents”.Footnote 39 The normal justification condition is about maximising conformity with reason, including the reasons stemming from democratic or other procedural values.
However, critics argue that the normal justification thesis can include democratic and other procedural justifications only at the cost of making the thesis “nearly empty”, because it would subsume any theory of legitimacy; indeed, satisfaction of the normal justification thesis would become a consequence of the authority's legitimacy, rather than a ground of it.Footnote 40 The problem, it is said, is that if the normal justification condition can be satisfied by a prior duty to obey democratic institutions, then it seems to be that prior duty, and not the normal justification thesis, that explains why the authority's claim to legitimacy is justified.
But I think the normal justification thesis can incorporate democratic and other procedural justifications without losing its explanatory force. It recognises that there are many reasons why we might better conform with reason by following an authority's decision: for example, because the authority has greater expertise, or because deciding for oneself is too burdensome, or because we have a greater interest in coordinating our action for the common good, or because making a decision in accordance with certain procedures, including democratic procedures, is more valuable than deciding for oneself. In cases involving democratic legitimacy, the normal justification thesis continues to do significant work, capturing the idea that democratic procedures are never sufficient for legitimacy; the reasons for deciding democratically must outweigh the reasons for deciding oneself. In Corner House, the 1987 Act gave the court a procedural reason to follow the Director's decision, but that is not to say that the normal justification condition is satisfied, for this procedural reason could be defeated by competing reasons for the court to decide itself.
B. The Pre-Emptiveness of Authority
The third condition of Raz's service conception of authority is the “pre-emption thesis”, which explains the binding force of a legitimate authority's decision and its effect on its subjects’ practical reasoning. It is the pre-emption thesis that helps to explain why the court in judicial review should not intrude on the merits of the case. The pre-emption thesis is as follows:
[T]he fact that an authority requires performance of an action is a reason for its performance which is not to be added to all other relevant reasons when assessing what to do, but should exclude and take the place of some of them.Footnote 41
“The whole point and purpose of authorities,” he says, “is to pre-empt individual judgment on the merits of a case”.Footnote 42 According to Raz, the pre-emptive reasons that a legitimate authority's decisions provide are what he elsewhere calls second-order “exclusionary reasons”, which exclude some (but not all) of the first-order reasons its subjects should otherwise have acted on.Footnote 43 The legitimate authority's subjects should not act for “excluded reasons” – which we can call the merits of the case – because these are the reasons that the authority was meant to settle, and the authority's subjects would better conform with reason if they follow the authority's judgment on those reasons.Footnote 44
From the point of view of the Director and the courts in Corner House, the reasons for and against concluding that there was a severe risk to national security by continuing the investigation may be excluded reasons, on this Razian account. Due to its greater expertise on this question, the Government's conclusion that there was a severe risk to national security should be treated as having settled that question. But that should not be treated as having settled the question of whether to investigate, which the statute allocated to the Director. If this statutory authorisation gave him legitimate authority to settle the wider question of how the severe risk to national security should be weighed against the reasons for continuing the investigation, then the courts, on the Razian exclusionary model, should treat the Director's decision as having settled the balance of those reasons. Those reasons are therefore excluded for the courts; they are the merits of the case and the courts should not interfere for excluded reasons.
But in that case, when would the courts be justified in interfering with the Director's decision? According to Raz, the pre-emptive reasons for following a legitimate authority's decision can be defeated only by “non-excluded” reasons – that is, the reasons that the authority was not meant to settle, because the subject would better conform with reason by following their own judgment on those reasons. He writes:
Even where an authoritative decision is meant finally to settle what is to be done it may be open to challenge on certain grounds, e.g. if an emergency occurs, or if the directive violates fundamental human rights, or if the authority acted arbitrarily. The non-excluded reasons and the grounds for challenging an authority's directives vary from case to case. They determine the conditions of legitimacy of the authority and the limits of its rightful power.Footnote 45
In Corner House, if the risk of undermining the rule of law was a non-excluded reason – in the sense that its relevance makes the court better placed to make the decision – then the Divisional Court may have been justified in substituting its judgment for the Director's, so long as the risk to the rule of law outweighed all the other considerations. The notion that the threat to the rule of law was a reason of this kind is what led Jeffrey Jowell to find it “surprising” that the House of Lords in Corner House “positively encourages the view that the rule of law is on par with any other ‘relevant consideration’ taken into account by the prosecutor”.Footnote 46
Raz's account of exclusionary reasons is one of the most contested parts of his theory of authority. The pre-emption thesis “stands on its own”, according to Raz, and need not be explained by treating authoritative decisions as providing exclusionary reasons: “possibly there are other, better explanations of it, and there is no need to saddle the account of authority with a commitment to that way of explaining the preemptiveness of authoritative directives”.Footnote 47 One alternative explanation is that the force of authoritative decisions is a very weighty or presumptive force, the strength of which varies according to the force of the underlying reasons that justify the authority's claim to legitimacy.Footnote 48 But that amounts to a fundamental challenge to the pre-emption thesis itself, for it requires the subjects to assess the merits of the case, to decide whether the presumption in favour of following the authority's decision has been defeated.
Any alternative explanation of the pre-emptiveness of authority ought to recognise that legitimate authority does have an exclusionary effect – and so subjects should not act on their judgment of the merits – when the normal justification condition is fully satisfied. But the normal justification condition may be only partially satisfied: on some particular issue, the agency might have legitimate authority in some respects but lack legitimate authority in others. The purported authority might have, say, greater democratic legitimacy or political accountability than the purported subject, but the purported subject might have, say, greater expertise than the purported authority. In this way, there may be circumstances in which a subject conforms better with reason in some ways by treating the authority's decision as binding, but not others, and there may be circumstances in which there is no overriding justification one way or another for treating the authority's decision as binding.Footnote 49 If that is so, the court may reasonably act on its judgment on the merits, notwithstanding the agency's good claim to legitimate authority.
One important implication of the pre-emption thesis is that an authority's decision can be binding even when it is mistaken. It follows from Raz's service conception of authority that neither a standard of correctness nor a standard of reasonableness can be used in reviewing the decision of a legitimate authority. If a legitimate authority's subjects challenged its decisions “every time it fails to reflect reason correctly”, Raz points out, “the advantage gained by accepting the authority as a more reliable and successful guide to right reason would disappear”.Footnote 50 Moreover, if a legitimate authority's subjects challenged its decisions every time it deviates too much from what is reasonable, the value of legitimate authority would be lost – its decisions would not have pre-emptive force – because such a limitation “requires every person in every case to consider the merits of the case before he can decide to accept the authoritative instruction”.Footnote 51 I return to this point in Section V when examining the implications for reasonableness review.
Under the service conception, therefore, the subjects of a legitimate authority – including, for our purposes, the courts in judicial review of an administrative authority's decision – are bound by its decision, even though they consider it to be incorrect or unreasonable. But that conclusion seems to go against the intuition that an authority's decision should not be followed if it is clearly mistaken. Raz, indeed, seems to be open to the possibility that “legitimate authority is limited by the condition that its directives are not binding if clearly wrong”; however, he goes on to say that, even if legitimate authority were limited in this way (and he expresses no conclusive view on whether it is), the limit does not challenge his theory, because “[e]stablishing that something is clearly wrong does not require going through the underlying reasoning”.Footnote 52 Nonetheless, whatever the effect of clear mistakes may be, on Raz's theory even “significant mistakes which are not clear” do not affect the binding force of legitimate authority.Footnote 53
C. The Scope of Authority
There is, however, one category of mistakes that necessarily affects the binding force of a legitimate authority's decision: jurisdictional mistakes. This brings us to the crucial question: how, on Raz's theory, should we distinguish questions of jurisdiction – the scope of legitimate authority – from questions about the merits of the case? As Raz explains, the distinction is crucial to the theory of authority:
The pre-emption thesis depends on a distinction between jurisdictional and other mistakes. Most, if not all, authorities have limited powers. Mistakes which they make about factors which determine the limits of their jurisdiction render their decisions void. They are not binding as authoritative directives … . Other mistakes do not affect the binding force of the directives. The pre-emption thesis claims that the factors about which the authority was wrong, and which are not jurisdictional factors, are pre-empted by the directive. The thesis would be pointless if most mistakes are jurisdictional or if in most cases it was particularly controversial and difficult to establish which are and which are not. But if this were so then most other accounts of authority would come to grief.Footnote 54
If, as Raz says, the pre-emption thesis depends on it being, in most cases, relatively easy and uncontroversial to distinguish jurisdictional mistakes from other mistakes, we might expect Raz's writings to be clear about how to draw the distinction. However, he was not as clear as he could have been.
A common misunderstanding of Raz's thesis is that jurisdictional mistakes can be identified by the fact that they are significant and clear.Footnote 55 This misunderstanding is the basis on which Matthew Lewans has recently argued that Raz's definition of jurisdictional error “fails to account for even the most basic legal constraints on the exercise of administrative power”.Footnote 56 The criticism is misplaced because Raz in fact treats clear mistakes and jurisdictional mistakes as distinct: as mentioned above, he does not take a conclusive view on whether clear mistakes limit the binding force of a legitimate authority's directives, but he explicitly states that jurisdictional mistakes necessarily affect the binding force of the directive.Footnote 57
How, then, are jurisdictional mistakes distinguished in Raz's theory? Correctly understood, the test for jurisdiction simply restates the conditions for the justification of legitimate authority. The normal justification condition is also the normal test for jurisdiction. That means that determining jurisdiction involves determining the extent to which the normal justification condition is satisfied – with the important clarification, as discussed above, that the normal justification condition may be satisfied not only in cases where the authority is more likely to reach the right answer, but also for procedural reasons (because, for example, the law has allocated the power to the authority, or the authority is subject to greater political accountability, or its procedures are more likely to treat the subject with respect). The question of which matters are within an authority's jurisdiction is a question of which matters are better settled by the authority, because following its decision would enable us to conform better with reason. Excluded reasons are the merits of the case, and the scope of the excluded reasons is the scope of an authority's jurisdiction. Whether the normal justification condition is satisfied may at times be unclear and controversial, but it is crucial for determining the scope of an authority's jurisdiction and therefore whether the authority's decision pre-empts the subject's judgment on the merits of the case.
Some critics of Raz's theory have seen a contradiction in this aspect of his theory. This is because it seems to undermine Raz's argument that an authority's subjects “can benefit by its decisions only if they can establish their existence and content in ways which do not depend on raising the very same issues which the authority is there to settle”.Footnote 58 The pre-emption thesis is undermined if subjects must assess the dependent reasons that are better settled by the authority. But, according to critics, assessing whether an authority has acted within its jurisdiction on Raz's account requires the authority's subjects to assess the dependent reasons (or relevant considerations) in order to check which dependent reasons are better settled by the authority. According to Heidi Hurd, for instance, Raz's theory entails that one can only determine whether an authority has acted within its jurisdiction “if, at each decision, one judges for oneself the antecedently existing reasons for action, and then determines whether the commands of the authority in fact reflect that balance”.Footnote 59 Thus, Hurd argues, the jurisdictional condition on legitimate authority is one reason why an authority's decisions cannot have the pre-emptive force that Raz claims.
This criticism of Raz's theory overlooks the distinction between two assessments. The first is an assessment of which matters are within the authority's jurisdiction. This is the normal justification condition: over which matters are we more likely to conform with reason by following the authority's decision? The second assessment is the balancing of the dependent reasons relating to those matters that are within the authority's jurisdiction. The first assessment (determining jurisdiction) does not involve the second assessment (deciding the merits of the case). In relation to Corner House, we can reach the conclusion that the Director and the court would be more likely to conform with reason by following the Government's assessment of the risk to public safety than by assessing for themselves what the risk to public is (and hence that this assessment is within the Government's jurisdiction) without themselves having to assess the risk to public safety.
Furthermore, the scope of legitimate authority need not be limited to balancing the reasons on the matters that are within the authority's jurisdiction. Jurisdiction may extend to settling what some of the dependent reasons are in the first place. It cannot settle what all the dependent reasons are, because acting for dependent reasons is a condition of legitimate authority, and so the authority's subjects – not the authority itself – must decide whether that condition is satisfied. The subjects must determine what some of the dependent reasons are, even if they should leave the balance of those reasons to the legitimate authority. But the authority may be better placed to determine what some of the other dependent reasons are, and on Raz's account it need not take into account all the considerations that its subjects think are relevant.Footnote 60 If so, then subjects should defer not only to the authority's judgment of the balance of dependent reasons, but should also treat as pre-emptive some of the authority's judgments of what the dependent reasons are.Footnote 61 For example, in assessing the risk to public safety, the Government may be better placed (and so have jurisdiction) not only to determine the weight and balance of the relevant considerations, but also to determine which considerations are relevant to that assessment.
D. Deference among Authorities
The theory of authority set out in this section requires judges to defer to an administrative decision maker's judgment on the merits of the case when the decision maker has a greater claim to legitimate authority than the court. This notion of judicial deference is controversial. The decision of the House of Lords in Corner House, for example, has been criticised as a “lapse into the unnecessarily deferential administrative law of yesteryear”.Footnote 62 There is resistance to the very word “deference”, with its apparent “overtones of servility”.Footnote 63 For sure, as Lord Steyn says in discussing deference, “labels are less important than the underlying approach or philosophy and its consequences”.Footnote 64 But the approach that underlies the attitude of some, though not all, judges and public lawyers towards deference is very different from the Razian approach defended in this article. Their approach to deference tends to treat it as an attitude of “respect” towards the authority's decision, but without recognising a duty to follow the authority.Footnote 65
Even among those who accept that a court may be under a duty to follow an administrative agency, there is some resistance to describing the court as being subject to, and “obeying”, the authority of the agency. Timothy Endicott, for example, prefers the language of “comity”, which he describes as “the respect by the second authority for the first authority, where the second authority is not bound by the decisions of the first”.Footnote 66 On this view, the court in judicial review proceedings should adopt an attitude of respect, as opposed to obedience, to decisions that are within the jurisdiction of a legitimate administrative authority.
But the idea of a duty of respect, courtesy, or comity, in contrast with a duty to obey, is insufficient to cover all relations among authorities. When an administrative decision maker, and not the court, has legitimate authority to decide a matter, judges manifest their respect for authority by recognising the binding force of its decision. Indeed, it seems that Endicott in effect accepts this point, when he acknowledges that the reasons for the reviewing court to respect the administrative official's decision are the same as the reasons that justify legitimate authority (according to the normal justification thesis) and are limited to matters that are within the first authority's jurisdiction.Footnote 67 The situation may be different if both authorities have equally good claims to possess legitimate authority over the same matter. Some of the examples of relations among authorities that Endicott discusses seem to be of this kind, such as the relation between two parents with authority over a child.Footnote 68 But where the matter falls within the jurisdiction of an administrative decision maker and not the court, the former has legitimate authority over the latter, and for that reason its decision is binding on the court and pre-empts the court's judgment of the merits.
IV. Questions of Law
A. Substitution of Judgment on Questions of Law
Legal doctrine ought to reflect this theory of legitimate authority. But comparative study reveals that, in most if not all legal systems, it rarely does. One example of this failure can be found in the doctrinal approaches to judicial review for error of law. As mentioned above, statutory allocations of power are one reason for holding that an administrative agency has legitimate authority. Consequently, determining an agency's jurisdiction may involve a determination of whether its decision is within the limits of the power allocated by the statute. But there is often room for doubt and disagreement about the meaning of a statutory provision. In many legal systems, the courts substitute their judgment on questions of law – that is, they treat questions of statutory interpretation as being for the courts to resolve, on the assumption that, if they deferred to the agency's interpretation of the statute, that deference would enable the agency to determine its own jurisdiction. My aim in this section is to explain why this approach is fundamentally flawed.
The “substitution of judgment” approach to questions of law is adopted in, for instance, English (and, more recently, UK),Footnote 69 New Zealand, Australian and EU administrative law.Footnote 70 It was not always so. English law used to distinguish between jurisdictional and non-jurisdictional errors of law, with only the former giving rise to the court's intervention. But the English courts drew this distinction in confusing ways – such as by distinguishing between precedent conditions and other conditions, or between being entitled to enter into an inquiry and going wrong in the course of that inquiry – and as a result it was difficult to predict which errors of law the courts would deem to be jurisdictional.Footnote 71 The point of departure from this older approach is often traced to Anisminic v Foreign Compensation Commission.Footnote 72 In a later case, Lord Diplock said of Anisminic that it:
proceeds on the presumption that …. Parliament intends to confine [a public authority's] power to answering the question as it has been so defined [by the enabling Act]: and if there has been any doubt as to what that question is, this is a matter for courts of law to resolve in fulfilment of their constitutional role as interpreters of the written law and expounders of the common law and rules of equity.Footnote 73
Following this interpretation, or misreading,Footnote 74 of Anisminic, Lord Browne-Wilkinson concluded in R. v Hull University Visitor, ex parte Page that “in general any error of law made by an administrative tribunal or inferior court in reaching its decision can be quashed for error of law”.Footnote 75 The alleged virtues of this general rule are its apparent simplicity and consistency with principles such as (a Diceyan interpretation of) the rule of law and the separation of powers.Footnote 76 However, the general rule is subject to a few “anomalous” exceptions.Footnote 77 The recognition of these exceptions can make judicial review for error of law seem “hopelessly confused”, not only to those who think that the courts should substitute judgment on all questions of law,Footnote 78 but also to those who think that the autonomy that these exceptions give to some administrative agencies to decide some questions of law should be extended to others.Footnote 79
Substitution of judgment on all questions of law leads to overly intrusive judicial review.Footnote 80 The old doctrinal ways of distinguishing between jurisdictional and non-jurisdictional errors of law may have been confused and confusing, but, as Paul Craig rightly warns, “we are, nonetheless, in danger of forgetting the rationale for them”.Footnote 81 As Craig explains, “[t]he critical question, the answer to which underlies any statement concerning jurisdictional limits, is whose relative opinion on which matters should be held to be authoritative”.Footnote 82 The claim that all questions of law are for the courts is based on either a dubious jurisprudential theory that considers it to be axiomatic that the judiciary should be final arbiters on all questions of law, or the belief that the judiciary have a special expertise over matters of legal interpretation.Footnote 83 If it is right that all questions of law should be for the courts, the retention of a degree of autonomy for the agency depends on distinguishing between questions of law, questions of fact and questions of the application of law to facts.Footnote 84 One way to rationalise the approach of the English courts is to say that they will treat all questions of interpretation as questions of law – and thus will substitute their judgment – but will treat questions of application as questions of law only when (in the court's view) the law clearly requires one answer.Footnote 85 But that approach overlooks the good reasons that the courts have to defer, including on questions of law, to the judgment of an administrative agency acting within the scope of its legitimate authority.
B. Deference on Questions of Law
Consider the generic form of a statutory allocation of power: “If X1, X2 and X3, then the agency may or shall do Y.” The question of statutory interpretation then arises whether the statute should be read as meaning, “If X1, X2, X3 and X 4 , then the agency may or shall do Y” – where X4 can be understood either as a further condition, alleged to be implicit in the statute, or as a more specific meaning attributed to one of the explicit (but under-specified) conditions. Here it is important to distinguish between three questions:
-
(1) Should the interpreter attribute to the statute a meaning that includes X4 as a condition of the power?
-
(2) Should X4 be a condition of the power?
-
(3) Should the court or the agency decide question (2)?
The statute, let us assume, says nothing about X4: it neither states that X4 is required nor that X4 is not required. To work out whether the agency should satisfy X4, we must answer question (2). This question could be regarded either as determining the meaning of the statutory limits or as determining the limits that should be imposed in addition to any statutory limits. For example, in Cooper v Wandsworth Board of Works, the statute said nothing about the requirement of a fair hearing, but the court imposed that requirement before the Board of Works could exercise its statutory power.Footnote 86 Some commentators believe that this condition can only be justified as an interpretation of the statute. They believe that, because Parliament is sovereign, a statutory allocation of power must contain all the limits on that power, otherwise the imposition of limits would be contrary to parliamentary sovereignty.Footnote 87 This proposition has been, in my view, persuasively shown to be false.Footnote 88 The statutory allocation may be indeterminate. But we can intelligibly speak both ways: framing the condition either as the meaning of the statute or as a further, common law condition in addition to the statutory conditions.Footnote 89 In either case, the answer to question (2) is not justified by any fact about the statute (such as “legislative intent”), though it is consistent with the statute.
But there is then question (3), which is the question of deference on questions of law: Should the court or the agency decide whether X4 should be a condition of the power? The answer to question (3) depends on which institution satisfies the normal justification thesis and therefore has a good claim to legitimate authority. Three answers are possible. The first is that the court is better placed than the agency to decide question (2) – because, let us say, of its greater expertise, its ability to achieve consistency among uncoordinated decision makers, or its independence.Footnote 90 If so, the court has legitimate authority to impose on the agency that condition, which is consistent with, but not justified by, the statute. The second possible answer is that the agency has legitimate authority over the court to decide question (2) – because, let us say, of its greater expertise, its ability to achieve coordination, or its political accountability. To the extent that question (2) is an aspect of statutory interpretation, the agency would therefore have jurisdiction to interpret the statutory conditions of its power. In doing so, the agency would not be determining the limits of its own jurisdiction, but would be acting within its jurisdiction (in the sense of legitimate authority). To suppose otherwise is to misunderstand the difference between questions (1) and (3) – that is, between statutory authorisation (a question of law) and legitimate authority (a question of jurisdiction).
Take, for example, the Pergau Dam case,Footnote 91 which concerned the decision of the then Foreign Secretary, Douglas Hurd, to grant £316 million in aid to Malaysia for the Pergau Dam project, allegedly in exchange for an arms deal, against the advice of the Government's own economic advisers that the project was economically unsound. Section 1(1) of the Overseas Development and Co-operation Act 1980 empowered the Secretary of State to grant aid “for the purpose of promoting the development or maintaining the economy of a country or territory outside the United Kingdom”. The question, unanswered by the text of the statute, was whether this statutory power should be used only for economically sound projects. Rose L.J. held:
Whatever the Secretary of State's intention or purpose may have been, it is, as it seems to me, a matter for the courts and not for the Secretary of State to determine whether, on the evidence before the court, the particular conduct was, or was not, within the statutory purpose.Footnote 92
The court assumed that it had legitimate authority to determine the statute's purpose and which considerations the statute required to be taken into account or ruled out. The court ruled that economic soundness was of paramount importance, rather than merely one relevant consideration among others. But did not the Foreign Secretary have a good claim to legitimate authority to decide whether this should be treated as a condition of the power? This decision involved highly political considerations of foreign relations, usually treated as falling within the legitimate authority of the executive, which is accountable to Parliament.Footnote 93 The decision did not become any less political by being masked behind the façade of statutory interpretation.
There is a third possible answer to question (3): neither the court nor the agency may be better placed to decide question (2). In that case, there is a difference between, on the one hand, interpreting a condition that the statutory text imposes and, on the other, interpreting a seemingly unfettered power in a way that reads in a condition. While the line between these two types of case may be vague, there are clear cases of each. The closer we get to the former, the greater the justification for the court substituting its interpretation. The closer we get to the latter, the greater the justification for the agency deciding the question.
In the former type of case, the court should intervene if the decision is unreasonable or incorrect, but arguably not if there is a range of reasonable decisions, none of which is uniquely correct. That is not about deference to legitimate authority. Yet the belief that judicial deference on questions of law should be manifested by the court's adoption of a standard of reasonableness, rather than correctness, in assessing an agency's interpretation is found, in different ways, in the US and Canada. The US courts, under the doctrine of “Chevron deference”,Footnote 94 treat statutory ambiguity as an implicit delegation of interpretive authority to the agency, and will defer to the agency's interpretation if it is reasonable: the agency is said to be “the authoritative interpreter (within the limits of reason) of such statutes”.Footnote 95 The Canadian courts also defer to reasonable administrative interpretations of the law, if such deference is required by a “pragmatic and functional analysis” (now referred to as the “standard of review analysis”).Footnote 96 In that analysis, the court considers a range of factors, including the expertise of the agency, along the lines of the test for jurisdiction defended in this article.Footnote 97 One problem with both the US and Canadian approaches is that the notion of deference to legitimate authority cannot justify a standard of reasonableness for judicial review. That standard can be justified only when neither the court nor the agency has legitimate authority to interpret the provision, in which case the court should go along with the agency's interpretation, for reasons of comity, unless it is unreasonable.Footnote 98 When the agency has legitimate authority to interpret the provision, there is no justification for the court to intervene merely because it thinks the agency's interpretation is unreasonable – unless, perhaps, it is clear that the agency made a mistake.Footnote 99
When the question is whether the court should impose a condition on a seemingly unfettered discretionary power – such as the power conferred on the Director in Corner House – that statutory allocation of power is itself one reason for recognising that the agency has legitimate authority. That does not mean that the court is never justified in imposing a condition on the power. The House of Lords was wrong to suggest, in ex parte Brind, that no ambiguity arises where a statute confers a broad discretion without explicit limits and that the court should therefore not read in conditions.Footnote 100 Discretionary powers without explicit limits are not unambiguous and conclusive; rather, they are defeasible.Footnote 101 The court does not necessarily act contrary to the statute by imposing a condition on such powers, whether or not that condition is justified as an interpretation of the statute.Footnote 102 But the court can justifiably impose a condition on that power only if the court is better placed to make that decision, and therefore has a better claim to legitimate authority, notwithstanding the statutory allocation. Likewise, there may be circumstances in which the court is justified in departing from a clear statutory condition, if the statutory allocation of power – and the conditions attached to that allocation – do not match the scope of the agency's legitimate authority.Footnote 103
V. Reasonableness Review
Many judges and commentators treat the standard of reasonableness as determining the boundaries of the decision maker's jurisdiction.Footnote 104 This approach was evident in the Wednesbury case itself, in which the high threshold of unreasonableness that Lord Greene tried to formulate – according to which the court will intervene only when the decision is “so unreasonable that no reasonable authority could ever have come to it” – was combined with the statement that, provided the local authority acts “within the four corners of their jurisdiction, this court, in my opinion, cannot interfere”.Footnote 105 On Lord Greene's view, the court could justifiably interfere with a decision that, although “prima facie within the powers of the executive authority”, was “so unreasonable” that it was actually “in excess of the powers which Parliament has confided in them”.Footnote 106 This section explains more fully why the ground of unreasonableness cannot be a basis for determining the scope of legitimate authority and cannot justify judicial intervention.
A. Two Types of Restraint
The confusion surrounding the role of reasonableness review is partly explained by a failure to distinguish between two types of judicial restraint. One is the type of restraint that is the focus of this article, which arises when the court defers to a legitimate authority on decisions that are within that authority's jurisdiction. The other type of restraint arises where, in the court's view, there is no right answer, but rather a range of reasonable decisions, perhaps due to reasonable disagreement or uncertainty over what the right answer is, or incommensurability among the options. “The very concept of administrative discretion,” according to Lord Diplock, “involves a right to choose between more than one possible course of action upon which there is room for reasonable people to hold differing opinions as to which is to be preferred.”Footnote 107 In exercising this second type of restraint, the court only intervenes when the agency acts for a defeated reason, which takes its decision outside the range of reasonable decisions.Footnote 108
The high threshold of “Wednesbury unreasonableness”, in Lord Greene's somewhat tautological formulation, has sometimes been criticised for being overly deferential to the administrative decision maker. For it suggests that there is a distinction between a decision that is merely unreasonable and one that is “so unreasonable that no reasonable authority could ever have come to it”. The problem is that Lord Greene justified judicial restraint primarily on the basis of the second type of restraint (a range of reasonable decisions), stating that the court should not intervene in decisions about which “honest and sincere people hold different views”.Footnote 109 But that does not explain why judicial restraint is justified when the decision under review is unreasonable, in the sense that it is outside the range of reasonable decisions. For this reason, Lord Cooke of Thorndon criticised Wednesbury’s suggestion that “there are different degrees of unreasonableness and that only a very extreme degree can bring an administrative decision within the scope of judicial invalidation”.Footnote 110 Lord Cooke thought that the administrator's duty “is simply to act reasonably”,Footnote 111 and that the court should ask whether the decision maker had “struck a balance fairly and reasonably open to him”.Footnote 112
However, Lord Greene's mistake was not his excessive deference, but his failure to explain adequately why, and in what way, deference is appropriate. It is the notion of legitimate authority, rather than the notion of a range of reasonable decisions, that justifies judicial restraint in the face of a decision that the judge considers to be unreasonable. Conversely, it is the decision maker's lack of legitimate authority that justifies the court's intervention. It might be thought that the notion of reasonableness can incorporate the notion of deference to a legitimate authority, the court intervening only if the decision is so unreasonable that its unreasonableness defeats the reasons for deferring to the legitimate authority.Footnote 113 But that is confusing, because it is not the degree of unreasonableness that justifies judicial intervention. Rather, judicial intervention is justified because there is some consideration – what Raz calls a “non-excluded” reason – that makes the matter better settled by the courts. That is, there must be some reason that defeats the legitimacy of the authority, and hence defeats the pre-emptive reason that the court had to refrain from acting on its own judgment of the merits.
Similarly, it is confusing to describe the standard of reasonableness as a variable standard or as giving rise to a variable intensity of review depending on the context.Footnote 114 To the extent that variable intensity of review is justified, it is because the legitimate authority of the decision maker varies depending on the context. Variable intensity of review is not the result of different standards of review being applied. It is commonly thought that classic Wednesbury unreasonableness is a low-intensity standard of review, whereas proportionality (or “modified” Wednesbury unreasonableness) is a high-intensity standard of review. But that is not the relevant distinction. The notion of reasonableness is flexible enough to contain within it the proportionality standard, involving some balancing of the merits,Footnote 115 and there will often be a range of proportionate answers, just as there is a range of reasonable answers.Footnote 116 For this reason, it is a mistake to think that proportionality review is any more of a “wrongful usurpation of power” than reasonableness review, as Lord Ackner claimed.Footnote 117 Nor is it correct to claim that proportionality, unlike the reasonableness standard, shifts the burden of justification to the administrative agency.Footnote 118 When the burden of justification shifts, it is because there is some special reason – perhaps a prima facie interference with a fundamental right or perhaps (as the Divisional Court thought in Corner House) a threat to the rule of law – that makes the decision better suited for the courts, thereby defeating the pre-emptive reason not to act on one's own judgment of the merits, and triggering the need for greater justification from the administrative agency.Footnote 119 On the other hand, the greater the justification of agency's legitimate authority on an issue, the less likely will be a non-excluded reason justifying the court's intervention.Footnote 120
B. Reasonableness and the Critique of Deference
We have already encountered criticisms of the notion of judicial deference. Lord Bingham, for example, preferred to describe the restraint that judges should show to an administrative agency as “according appropriate weight” to its judgment.Footnote 121 That demonstrates a flawed understanding of the pre-emptive force of the reasons for following the legitimate authority's decision. The legitimate authority's decision does not give us a reason that is added to all the other first-order reasons in the balance, or that affects the weight of the first-order reasons. When the administrative agency has legitimate authority, the courts should not act on their judgment of the balance of first-order reasons, unless there is a reason that defeats the agency's legitimate authority and makes the court better placed to settle the matter.
An even more fundamental critique of deference comes from T.R.S. Allan, who objects to the existence of a doctrine of “due deference” that stands apart from the assessment of the reasonableness of the agency's decision. From the judges’ perspective, Allan argues, there is no “division between first-order considerations of substance or content and second-order considerations of institutional design, such as the goal of democratic accountability”.Footnote 122 According to Allan, a judge should not attempt to weigh first- and second-order (or content-dependent and content-independent) considerations against each other. Instead, judges should decide according to their own judgment of the first-order considerations of substance in enforcing the heads of review, and take into account the authority's conclusion only to the extent that it aids the judge's own reasoning.Footnote 123
Allan is willing to concede that an authority's superior expertise may be relevant, but on his account it is relevant only because, and to the extent that, the authority's superior expertise makes it more likely to persuade the judge that its decision is correct or at least reasonable – ultimately, it remains for the court to assess the balance of first-order reasons.Footnote 124 Likewise, on Allan's account, deference is due on democratic grounds only because, and to the extent that, there is reasonable disagreement: “[I]n most circumstances, a variety of solutions will be consistent with respect for human rights. The court should accept whatever choice is made.”Footnote 125 On this account, it is the court's assessment of the range of reasonable decisions that determines the scope of the public authority's discretion and therefore its jurisdiction.Footnote 126
Allan's objection amounts to a denial of the pre-emptive force of a legitimate authority's decision. He claims that all the grounds of judicial review, including reasonableness, establish jurisdictional boundaries, and it is therefore incoherent to claim that, in enforcing those grounds of review (and so establishing jurisdictional boundaries), the court should defer to the legitimate authority.Footnote 127 Yet if a legitimate authority's decision has pre-emptive force, reasonableness cannot be a test for jurisdiction. Recall that, according to the pre-emption thesis, a legitimate authority can only fulfil its valuable purpose – enabling its subjects to conform better with reason – if its subjects refrain from acting on their own judgment of the merits. Reasonableness review inevitably entails a judgment on the merits. It can therefore only be justified when there is some aspect of the case that deprives the agency of legitimate authority over that matter.
VI. Conclusion
That administrative authorities should never act against reason, our lawyers are agreed; but the doubt is of whose reason it is that shall be authoritative.Footnote 128 It has become fashionable to defend what Etienne Mureinik called a “culture of justification”, in which “every exercise of power is expected to be justified”, in contrast with a “culture of authority”.Footnote 129 A culture of authority sounds worse than a culture of justification, because it could be authoritarian. But let us not forget that there are good justifications for legitimate authority. Of course, we should not overlook the problem of illegitimate authority, of people claiming more authority than is justified. But that is precisely why a doctrine of jurisdiction is so important. Any culture of justification, to be defensible, must allow for deference to an administrative agency acting within its jurisdiction, understood as the scope of its legitimate authority.
This article has defended a theory of the structure of the reasoning that ought to determine an administrative agency's jurisdiction, and therefore when a court is justified in intervening in administrative decision making. Returning one last time to Corner House, we can understand more clearly the way in which the Divisional Court and the House of Lords ought to have justified their respective conclusions. It was not enough for the Divisional Court to say that the decision about the necessity of surrendering to the threat was one for the court because of its responsibility to defend the rule of law. The court needed to explain why the court was better placed to balance the competing public interests, notwithstanding the good reasons for recognising the Director's legitimate authority as a consequence of the statutory allocation of a wide discretionary power. On the other hand, the House of Lords should not have treated the statutory allocation as preventing the court from imposing the condition that the Divisional Court envisaged. Its judgment, like that if the Divisional Court, could have been improved by a better understanding of the nature and limits of legitimate authority.
The argument of this article is not necessarily an argument for greater judicial deference, as such. It is an argument for greater judicial deference when the circumstances call for it. It is a defence of the distinction between questions of jurisdiction and the merits of the case, and an argument that, when a legitimate authority acts within its jurisdiction, the courts should follow the authority's decision, in a sense, “blindly”, regardless of their view of the merits.Footnote 130 Even Wednesbury unreasonableness is not an appropriate basis for intervening on the merits. The basis for the court's intervention must be that the agency lacks legitimate authority – that is, that the court would conform better with reason by substituting its judgment for that of the agency, notwithstanding any claim to legitimate authority that the statutory allocation gives the agency.
