1. Introduction
This article is concerned with three related concepts of administrative law: records, reasons and rationality (or reasonableness). Records I understand as documentary accounts of administrative proceedings and processes. Reasons are explanations or justifications given by officials of why they did as they did or decided as they decided. Judgments of reasonableness and rationality concern the justification of what is done or decided. Each of these concepts plays a significantly different role in United States (US) federal law, on the one hand, than they do in English law and Australian federal law on the other. My aim is to document and offer an explanation for this divergence. The explanation will be based on a distinction between two models of distribution of public power, called respectively ‘concentration’ and ‘diffusion’. Each model is associated with a distinctive mode of controlling power, referred to in this article as ‘accountability’ in the case of concentration, and ‘checks-and-balances’ in the case of diffusion. The article's main argument is that structural and institutional differences between the three legal systems being compared (elaborated in terms of these two distinctions) will help to explain why the concepts of rationality, reasons and the record play a far more significant role in the US law of judicial review of administrative power than they do in either English or Australian law.
The systems discussed here were chosen partly because they share a common history until the late eighteenth century, but have diverged since then; and partly because the Australian constitutional system contains elements of both English and US constitutional arrangements. However, the analysis is asymmetrical in the sense that it compares and contrasts the US on the one hand with England and Australia on the other. This is because, although there are significant structural and institutional differences between the English and Australian systems, they are of relatively minor importance for the subject-matter of this article. They are more significant in relation to certain other aspects of control of administrative power, which are dealt with in the larger project of which this article forms part.Footnote 1
It will also be noticed that whereas the accounts of English and Australian law deal primarily with the making of individual decisions by administrative officials, the account of US law focuses on administrative rule-making. This difference of focus reflects the fact that since the 1960s, the US law of judicial review has been preoccupied with rule-making while, by contrast, administrative rule-making is and always has been subject to very light judicial control in the English and Australian systems, where judicial review law has developed primarily by reference to individual decision-making. Elsewhere, I have touched briefly on this fundamental difference of focus,Footnote 2 which receives more detailed consideration in the forthcoming volume referred to in footnote 1 to this article. For present purposes, the difference is of no great significance because (as we will see) in US law the record is central to judicial review of both rules and individual decisions, whereas in English and Australian law it is central to neither.
The article has two main parts. The first part traces the development and current state of the relevant law in the three jurisdictions being compared: England,Footnote 3 Australia and the US. All three accounts deal, in turn, with records, reasons and rationality (or reasonableness). In this first part, the concepts will be dealt with separately because although – as will be argued – they are related in important ways, each has a history and a life of its own independently of the others. The second part of the article offers an explanation of the roles played by these three related concepts in the judicial review law of each of the compared systems. As in the first part, each of the three systems is discussed separately. However, the explanation offered stresses the relationship between the three concepts rather than their distinctiveness; thus, in relation to each system, the concepts are dealt with in an integrated way rather than separately.
2. Development of the Law
2.1. English Law
2.1.1. Records
The writ of certiorari was developed originally as an administrative law remedy in the seventeenth century in response to the conferral of new powers on justices of the peace to try criminal prosecutions summarily.Footnote 4 From criminal convictions its use was extended to administrative orders of various sorts. Because records of criminal proceedings were typically quite detailed, the record would normally reveal any reviewable error made by the justices, thus providing a basis for the award of certiorari. By contrast, records of administrative decisions were typically much less detailed and, as a result, reviewable errors might not appear ‘on the face of the record’. In such cases, by the middle of the eighteenth century, courts started to allow affidavit evidence to be introduced to complete or supplement the record, but only where the alleged error of law was ‘jurisdictional’. According to Philip Murray,Footnote 5 the limitation of affidavit evidence to cases of alleged jurisdictional error was designed to limit judicial review of administrative action and encroachment on the basic principle that the record was conclusive. In other words, the limit on supplementation of the record represented a compromise between judicial control of the administrative process and judicial restraint. Under this compromise, errors that did not go to jurisdiction were not reviewable unless they appeared on the face of the non-supplemented record as delivered to the court by the decision-maker.
By the middle of the twentieth century, the volume of public administrative decision-making had increased enormously, and courts started to look for ways to provide greater protection for citizens by extending their control over the executive and bureaucracy. One way of doing this was to reinvigorate certiorari as a remedy for non-jurisdictional error by extending the concept of ‘the record’ as traditionally understood and applied. The scope of the record was expanded beyond the formal recording of the decision by the decision-maker to include evidence (including oral evidence) and, more importantly for present purposes, any reasons given for the decision.Footnote 6 Traditionally, reasons did not form part of the record, whether of judicial or administrative proceedings. Indeed, it was not until the middle of the nineteenth century that ‘the reasoned first-instance judgment’ became common judicial practice.Footnote 7 Moreover, ‘an appeal is lodged against the result of formal judgment … and not from the reasons (if any) given by the judge. This recognises that the reasoning may be wrong but the result correct nonetheless’.Footnote 8 In other words, judges are not required to give reasons for their decisions but, if they do, an appeal against a judicial decision need not succeed merely because the decision is not supported by the reasons given to support it. It is, then, perhaps not surprising that when English courts started to review administrative decision-making in the seventeenth century, they did not require administrators to give reasons, and any reasons given were not treated as part of the record.Footnote 9
The next move in the broadening of judicial review was to enlarge the class of errors that counted as ‘jurisdictional’Footnote 10 to the point where it now effectively covers all reviewable errors made by public decision-makers.
The upshot of these developments is that the concept of ‘the record’, as such, has ceased to play any significant role in the English law of judicial review. The relevant issues are now dealt with in terms of the admissibility of ‘fresh evidence’ in judicial review proceedings. ‘The starting point is to focus on evidence which was before, or available to, the public body at the time of the impugned action’.Footnote 11 However, ‘fresh’ evidence may be admitted for various purposes, including ‘to show what material was before the minister or inferior tribunal’Footnote 12 – effectively, in anachronistic terms, to ‘complete the record’. For instance, material may be treated as having been before the decision-maker if it was within the knowledge of the decision-maker.Footnote 13 To be admissible, evidence must, of course, be ‘relevant’. The meaning of this requirement may vary according to the ground of review on which the decision is challenged.Footnote 14 For instance, the ‘proportionality’ ground of review under the European Convention for the Protection of Human Rights and Fundamental Freedoms (ECHR) is said to be ‘fact-specific’, and this may justify the admission of evidence that would not be admissible if the challenge were made on a less fact-specific ground.Footnote 15 In general, courts are wary of admitting evidence concerning the reasons for a decision, especially in cases where the decision-maker had a statutory obligation to give reasons. Evidence may be admitted to amplify or explain the reasoning underlying a decision, but ‘[a] public body cannot retrospectively change and improve upon its … decision’ by ‘after-the-event rationalisation’.Footnote 16 The line between (permissible) amplification or explanation and (impermissible) improvement or change is, of course, indeterminate.
With regard to appeal to an administrative tribunal as opposed to judicial review by (or appeal on a point of law to) a court, the basic rule varies from one area of administrative activity to another. In some contexts, the tribunal must decide the appeal on the basis of the material available to the decision-maker at the time of the decision. In others, the tribunal may also take account of additional relevant material available at the time of the appeal. In anachronistic terms, in some contexts the record is closed at the date of the decision whereas in others it remains open until the appeal. Of course, even where the record is closed at the date of the decision, evidence may be admitted by the tribunal to establish what material was actually before or available to the decision-maker.
Land-use planning inquiries – which are procedurally analogous to formal or hybrid rule-making processes in the US system (see below) – are conducted (typically ‘on the papers’) by an ‘inspector’ (the counterpart of an administrative law judge in the US system). They are preliminary to a decision by a government minister whether or not to grant planning permission, which is a form of licence. Such proceedings are ultimately judged in light of the common law principles of natural justice: the rule against bias and the fair hearing rule. In this context, the fair hearing rule demands that if, in making the decision whether or not to grant planning permission, the minister takes into account evidence that was not before the inquiry, the inquiry must be reopened so that the parties have an opportunity to address the new evidence.Footnote 17 The decision of the minister may be challenged by judicial review, in which case the basic rule, as noted above, is that new evidence will not normally be admitted. Given the formal nature of the inquiry process, it would be expected that courts in judicial review proceedings would be even less willing to admit new evidence in this than in other contexts.
We have noted that the original rationale for the rule – that evidence to complete or supplement the record was admissible only in relation to jurisdictional errors – was to limit the scope of judicial review of administrative decision-making. There appear to be at least two reasons for the basic rule against the admissibility of fresh evidence in judicial review proceedings. One is that, in general, judicial review is said not to be a suitable mechanism for resolving disputes of fact, chiefly because evidence in judicial review proceedings is given by affidavit rather than orally, and cross-examination of witnesses is not normally allowed. Another is that ‘[i]f “fresh” evidence were admissible, the court would be likely to find itself in the position of being asked to decide the merits of the case rather than acting as a court of review’.Footnote 18 This reason reflects one understanding of the distinction between jurisdictional and non-jurisdictional errors: the issue of jurisdiction concerns whether an agency has the power to decide, while a non-jurisdictional issue goes to the substance of the decision. It also reflects the related distinction between judicial review and an appeal: judicial review concerns the legality of the decision, whereas appeal extends to the quality of the decision (whether it is legally right or wrong). Both distinctions reflect judgments about the proper constitutional relationship between courts and administrators.
2.1.2. Reasons
Administrative decision-makers are often under statutory obligations to give reasons. In the absence of such an obligation, any duty to give reasons must be found in the common law of natural justice – in particular, the fair hearing rule. This rule recognises no general duty to give reasons, but there are various situations in which such a duty may be imposed: for instance, where a decision-maker has undertaken to give reasons or is known to have adopted a general practice of giving reasons. For present purposes, the important point to note is that obligations to give reasons are understood in terms of the common law duty to give a fair hearing. The giving of reasons promotes fairness instrumentally by enabling the person affected to understand the basis of the decision and to ascertain whether or not there is a basis on which the decision might be successfully challenged. It also promotes fairness non-instrumentally by treating the person affected with respect.
On the other hand, there are traces in English law of a different conception of reason-giving that does not associate it with the fair hearing rule. There is an ongoing debate about whether the mere failure to give (adequate) reasons can invalidate a decision or, by contrast, whether a failure to give (adequate) reasons can only provide evidence that the decision-maker lacked good reasons for the decision. This debate raises the question of the relationship between reason-giving and rationality as conditions of lawful decision-making. Is rationality concerned with the relationship between the reasons given and the decision, or is it an ‘objective’ concept based on a value judgment by the reviewer about the decision itself? This question is addressed in the next section. Here it is sufficient to say that the latter concept would be more compatible with recognition of failure to give (adequate) reasons as an independent ground of challenge.
In cases where there is an obligation to give reasons, any reasons given must be ‘adequate’ – that is, they must satisfy a minimum standard of clarity and explanatory force, and deal with all the substantial points that have been raised by the persons affected.Footnote 19 This requirement raises difficult questions about the relationship between the quality of reasons and the quality of decisions, to which we shall return below in the Australian context.
2.1.3. Rationality
Traditionally in English law, unreasonableness is a ground for judicial review only in the very strong sense that, in the opinion of the court, no agency in the position of the decision-maker could have thought that the decision was reasonable (so-called ‘Wednesbury unreasonableness’).Footnote 20 It is not enough that some, or even a majority, of agencies might have thought it unreasonable if another (or others), in addition to the decision-maker, would have considered it reasonable. Whether or not a decision is unreasonable in this sense is ultimately for the court to decide. The requirement of reasonableness does not imply an obligation to give reasons. In other words, unreasonableness is not a relation between a decision and the decision-maker's reasons. Rather, it is a relation between the decision and value judgments about ends. The strong sense of ‘unreasonableness’ reflects an ethos of judicial restraint in imposing values on administrators. In cases where courts do not feel so restrained, the strong sense of ‘unreasonableness’ may give way to a weaker notion, such as ‘proportionality’, which gives administrative decision-makers less leeway for choice between values and ends.
2.2. Australian Law
2.2.1. Records
In principle, at least, the concept of ‘the record’ still plays a part in Australian judicial review law that it has ceased to play in English law. This is because Australian law still observes the distinction between jurisdictional and non-jurisdictional errors (of law). A non-jurisdictional error is reviewable only if it appears on the face of the record. However, the practical relevance of this distinction to the present discussion is limited partly because the concept of ‘jurisdictional error’ is given quite a broad meaning, and partly because the distinction is relevant only to decision-making by inferior courts and not by administrative bodies and tribunals. On the other hand, under Australian common law, reasons are not part of the record;Footnote 21 thus their incorporation depends on statutory provision.
The retention of the jurisdictional/non-jurisdictional distinction reflects a particular view of the relationship between courts and administrators, which has constitutional foundations that are explored in the next section.
2.2.2. Reasons
Australian law on the giving of reasons by administrative decision-makers is essentially similar to English law: administrators have no ‘general’ common law obligation to give reasons.Footnote 22 This may be thought surprising given that in Australia the giving of reasons has been considered to be a normal incident of the judicial process since at least the beginning of the twentieth century. It may be explicable as an expression of the categorical distinction drawn by Australian law between judicial and non-judicial functions: reason-giving, in this view, is a distinctive aspect of the judicial function but not of the administrative function.
Express statutory obligations to give reasons are common – indeed, probably more common than is the case in England. However, under English law, any gaps in express statutory requirements for giving reasons are filled, if at all, by the common law; in Australia they are filled by statutory implication. This approach (of preferring statutory implication to common law as the source of limits on public power) is a relatively recent development. It is part of a wider reconceptualisation of the role of courts in the Australian system of government, and of the common law in constitutional arrangements. In general, the High Court of Australia now understands grounds of judicial review in particular, and limits on public power more generally, as a product of express or implied statutory provision rather than of the common law. Similarly, it has begun a process of reinterpreting what have traditionally been understood as non-statutory (prerogative and common law) powers of the executive as grounded in the (written) Constitution. Underpinning both moves is a desire to maintain and strengthen the authority of the courts by down-playing their role as independent policy and law makers, and emphasising their role as guardians of the rule of ‘law’ understood quite narrowly in terms of the Constitution and statute. In this way, the court is re-orienting the Australian constitutional system away from its English common law heritage and emphasising the US influence on the drafting of the Australian Constitution. According to the US way of thinking, the predominant (if not the only) sources of public law are the Constitution and laws made under the Constitution in the form of statutes and treaties.Footnote 23
Reasons given must be ‘adequate’. In this respect, there is a particular problem in Australian law about the meaning of ‘adequacy’ and the relationship between the reasons for and the substance of a decision. The High Court of Australia has read out of (or into) the Australian Constitution a strong distinction between judicial and non-judicial power, and has held both that non-judicial bodies may not exercise judicial power and that judicial bodies may not exercise non-judicial power. The latter proposition has, in turn, given rise to categorical distinctions between tribunals (non-judicial bodies) and courts (judicial bodies) and between the merits and the legality of administrative decisions.Footnote 24 Review of the merits of administrative decisions is classified as a non-judicial function that may not be undertaken by courts, but only by tribunals. This line of reasoning also suggests a distinction between what we might call the ‘intelligibility’ of reasons and their ‘persuasiveness’.Footnote 25 It might be thought that the strong ban on review by courts of the merits of administrative decisions would rule out an interpretation of ‘adequacy of reasons’ that focused too closely on the relationship of the reasons to the actual decision, and on the quality of the reasons as a persuasive justification for the decision. In other words, it might be thought that the ban on review of the merits would point towards an interpretation of ‘adequacy of reasons’ that required no more than an intelligible statement of why the decision-maker reached the decision made.
Given this background, it might be expected that failure to give (adequate) reasons could function as an independent ground of review and not merely as a symptom of some other error. On the other hand, allowing mere failure to give (adequate) reasons as a ground for holding a decision to be unlawful might seem unreasonably restrictive of administrative power. In Australia, there is a way around this problem: the High Court has made it clear that the appropriate remedial response to establishing any particular ground for judicial review is itself normally a matter of statutory interpretation;Footnote 26 and this may create a space for remedying a failure to give (adequate) reasons by, for instance, granting an order that adequate reasons be given rather than holding the decision itself to be unlawful or invalid because of the inadequacy of the reasons.
2.2.3. Rationality
The approach of Australian law to reasonableness is similar to that of English law. The starting point is the concept of Wednesbury unreasonableness. However, there are indications in a recent decision that the High Court might be moving in the direction of adopting a more flexible approach that would, in some cases, require a higher standard of rationality.Footnote 27 Whether and how this line of thinking might develop is unclear because the High Court has traditionally insisted that only the strong sense of unreasonableness is consistent with the Australian understanding of the judicial function as being concerned solely with legality rather than merits.Footnote 28 It has also resisted the ECHR-inspired receptiveness of English courts to concepts such as ‘anxious scrutiny’ and ‘proportionality’, which the High Court interprets as allowing inappropriate judicial encroachments on the merits of decisions and the province of the executive. On the other hand, the High Court has recently also suggested that inadequacy of reasons may be indicative of an unreasonable decision.Footnote 29 One possible interpretation of this approachFootnote 30 is that, in order to be adequate, reasons must not only be intelligible but also persuasive – in other words, the adequacy of reasons is to be assessed not merely in terms of the quality of the reasons but also in terms of the quality of the outcome. Such an approach would arguably blur, if not collapse, the distinction between legality and merits.
2.3. US Law
2.3.1. Records
The discussion can start with the Administrative Procedure Act 1946 (APA), which expressly distinguishes between administrative decision- and rule-making that is required by statute to be conducted ‘on the record after opportunity for an agency hearing’Footnote 31 and other forms of administrative decision- and rule-making. In proceedings required to be on the record, ‘[a] sanction may not be imposed or rule or order issued except on consideration of the whole record or those parts thereof cited by a party and supported by and in accordance with the reliable, probative, and substantial evidence’.Footnote 32 The record consists of ‘[t]he transcript of testimony and exhibits, together with all papers and requests filed in the proceeding … When an agency rests on official notice of a material fact not appearing in the evidence in the record, a party is entitled, on timely request, to an opportunity to show the contrary’.Footnote 33 ‘All decisions, including initial, recommended, and tentative decisions, are a part of the record and shall include a statement of – (A) findings and conclusion, and the reasons and basis therefor, on all issues of fact, law, or discretion presented on the record; and (B) the appropriate rule, order, sanction, relief or denial thereof’.Footnote 34 Any ‘ex parte communication … relevant to the merits of the proceedings’ between ‘a person outside the agency’ and any employee of the agency ‘who is or may reasonably be expected to be involved in the decisional process of the proceeding’ must be on the record.Footnote 35 On judicial review, ‘the court shall review the whole record or those parts of it cited by a party’.Footnote 36
On-the-record decision-making is known as ‘formal adjudication’ and on-the-record rule-making is known as ‘formal rulemaking’. In 1946 and for more than a decade thereafter, (formal) adjudication (the making of ‘orders’, including the granting of licences) rather than rule-making, was the paradigm mode of agency ‘policy-making’.Footnote 37 The APA also lays down a ‘notice-and-comment’ procedure for what is called ‘informal rule-making’,Footnote 38 a procedure which applies to rule-making that is not required to be on the record. Various types of rule are exempted from the procedure and, in relation to such rules, the only procedural requirement imposed by the APA is publication. Notice-and-comment procedure requires the agency to give notice of proposed rule-making specifying the time, place and nature of the proceedings, the legal authority under which the rule is proposed and either the terms of the rule or a description of the issues involved. The agency must ‘give interested parties an opportunity to participate in the rule-making through submission of written data, views or argument with or without opportunity for oral presentation’, and it must ‘incorporate in the rules adopted a concise general statement of their basis and purpose’.Footnote 39 For various reasons, during the 1960s and 1970s there was a wholesale shift from adjudication to informal rule-making as the preferred mode of agency policy-making. In other words, there was a shift from a formal, on-the-record adjudicatory process to an informal legislative procedure that did not require the agency to generate a formal record of the proceedings.
In the absence of any requirement of a record of a rule-making process, the courts were presented with the question of how ‘they could assess its validity on review’.Footnote 40 The APA is drafted on the assumption that the record will form the basis of judicial review. However, because informal rule-making, to a significantly greater extent than formal and informal adjudication, is likely to be a diffuse process involving many people and the consideration of many complex issues over lengthy periods of time, any ‘record’ of such proceedings is likely to be much less discrete and ordered than the record of an adjudication: more like ‘legislative history’ than a trial transcript.Footnote 41 The diffuse character of the rule-making process was reinforced by freedom of information and government in the sunshine legislation, which facilitated wide participation in administrative processes and (potentially, at least) made rule-making a much more pluralistic affair.Footnote 42 The bottom line was that in traditional judicial thinking, reflected in and encouraged by the design of the APA, proper control of administrative processes required those procedures to be ‘formal’ – which meant ‘trial-like’ – rather than ‘informal’, and more like the ways of a legislature than a court.Footnote 43 The problem that courts faced was caused not so much by the switch from adjudication to rule-making but by the fact that the model of rule-making chosen was informal rather than formal. This choice is easily explained: one of the major criticisms made of formal adjudication as a mode of policy-making was its time-consuming nature. Since the procedure for formal rule-making laid down in the APA was the same as that for formal adjudication, a shift to formal rule-making would, if anything, have exacerbated rather than eased the problem of delay.
Ironically, the desire of the judiciary to retain a significant degree of control over the administrative process led the courts to formalise the APA's notice-and-comment procedure with the result, in the view of many, that the agency rule-making process became ‘ossified’ in much the same way as the formal adjudicatory process had slowed down intolerably in the 1950s.Footnote 44 In 1946, the notice-and-comment provisions were generally understood to be quite undemanding and, for some years, this was how the courts interpreted them. However, in the 1960s and 1970s they were supplemented and strengthened in various ways. Most fundamentally for our purposes, the requirements of notice of proposed rule-making, and a post-adoption statement of basis and purpose, were both made considerably more demanding. It was this change that effectively forced agencies to generate a ‘record’ of the rule-making proceedings of a sort traditionally associated with judicial, not legislative, proceedingsFootnote 45 – and which, incidentally, is not generated by rule-making either in England or Australia.Footnote 46
On the back of this development, the main standard of review of administrative rules – namely the ‘arbitrary and capricious’ test laid down in the APA – was reinterpreted to justify and require a ‘hard look’ at the reasons given by the agency for the rule that it made. Furthermore, courts developed (and Congress imposed) ‘hybrid’ rule-making procedures that were more demanding than the APA's informal procedure but less demanding than its formal procedure. In 1978, the Supreme Court called a halt to the imposition of such procedures by lower courts,Footnote 47 although the precise implications of the decision, and the effect that it has had, are unclear.Footnote 48 The practical impact of these various changes was magnified by the 1967 decision of the Supreme Court that administrative rules may be judicially reviewed even before they are adopted or enforced.Footnote 49 ‘Pre-enforcement’ scrutiny of a rule inevitably invites more abstract and wide-ranging review than challenges incidental to particular applications of the rule, and tends to focus on the agency's reasoning processes to the exclusion of other considerations.
The upshot of all this is that ‘the record’ forms the basis of judicial review in the US, whether a review of ‘decisions’ (‘orders’ and ‘licenses’ in US terminology) or administrative ‘rules’ (hard rules, anyway, as opposed to soft rules).Footnote 50
2.3.2. Reasons
At least since the enactment of the APA, the giving of reasons has been a central feature and requirement of the formal trial-like procedure, the so-called ‘hybrid’ rule-making procedures, and also of the APA's informal, notice-and-comment procedure, especially as developed by the federal courts.Footnote 51 Reasons are integral to the record.
As already noted, one implication of the absence of an obligation to give reasons is that a decision may be acceptable even if any reasons given to support it are not.Footnote 52 Put differently, a decision of one decision-maker may be upheld on appeal or review by another for reasons other than those given by the first decision-maker. In the US, by contrast, so far as administrative agenciesFootnote 53 are concerned, the (Chenery) rule is – predictably – the opposite: decisions and rules made by agencies may be upheld if, and only if, supported by the reasons actually given by the agency.Footnote 54 This rule adds considerable significance to the record and gives agencies a strong incentive to take great care in its creation. Kevin Stack explains the Chenery rule as an aspect of the non-delegation doctrine.Footnote 55 One element of this doctrine prevents Congress from delegating rule-making power to agencies without providing some ‘intelligible principle’ to guide and constrain the exercise of the power. However, this principle has not been used since the 1930s to invalidate Congressional legislation directly. Stack interprets Chenery as prohibiting Congress from delegating the power (which it possesses) to make law without providing any justification for it. The requirements of a record, reasons and rationality are the price that administrators must pay for the delegation to them of wide rule-making powers. Put differently, the centrality of reason-giving in US law is explicable by the felt need and the demand for bureaucratic legitimacy.
2.3.3. Rationality
‘Hard-look’ (or ‘rationality’) review has become a hallmark of US judicial review of agency policy-making, especially rule-making.Footnote 56 Hard-look review requires a strong logical relationship between outcome (decision or rule), reasoning and reasons. This fits neatly with the high value put on reason-giving in the US legal system, and with the rule that administrative decisions and rules must be upheld, if at all, on the basis of the actual reasons of the decision maker or rule maker. Rationality is central to US judicial review law in a way that is not true of English or Australian law. The centrality of reasons and rationality is, in turn, reflected in the importance and significance of the record in US law. In English and Australian law the record functions – or used to function – as a device for limiting and restricting judicial control of administrative decision-making, whereas in US law it has provided the basis for its expansion and strengthening.
3. Explaining the Contrast
The phenomenon we are examining can be summarised in terms of the strength, in various systems, of the requirement of rationality as a condition of the legality and legitimacy of administrative decision- and rule-making. The discussion so far can be summarised by saying that the US system imposes stronger requirements of rationality on administrative decision- and rule-making than either English or Australian law. This is seen in the universality of the obligation to give reasons and the centrality of the concept of the record. We might explain this state of affairs simply by saying that the US courts put a higher value on rationality than do English or Australian courts. Some go further and argue that because rationality is highly desirable, English and Australian courts should demand more of it – for instance, by adopting something analogous to hard-look review. My concern is not with such ‘normative’ explanations and recommendations. Rather, it is to ask whether we may be able, in part at least, to explain differences, such as the one we have observed, in terms of the constitutional, institutional and structural features of particular legal/governmental systems. More particularly, my hypothesis is that such differences may be explicable partly in terms of distinctions between two models of distribution of public power and, associated with them, two modes of controlling public power, especially administrative power.Footnote 57
The two models of distribution (or ‘allocation’) of power I call ‘diffusion’ and ‘concentration’, and the two models of control I call ‘checks-and-balances’ and ‘accountability’. Diffusion involves dividing power between various institutions by giving each institution a share in the exercise of the power – ‘separated institutions sharing powers’ in Richard Neustadt's influential phrase.Footnote 58 A good example of diffusion is the US Constitutional requirement of ‘presentment’, which refers to the power of the President to veto Congressional legislation (and the power of Congress to override a presidential veto). Under this arrangement, legislative power is shared between Congress and the President. In abstract terms, the hoped-for effect of diffusion is to reduce the capacity of government by putting barriers in the way of government action in general and policy-making in particular by requiring various institutions to cooperate and collaborate in the exercise of power. By contrast with diffusion, concentration involves dividing power between institutions in such a way that each can exercise its power unilaterally without the need to gain the consent or cooperation of the other institution(s) – separate institutions exercising separate powers, to adapt Neustadt's phrase. Concentration witnesses a desire for or, at least, a tolerance of ‘strong’ and ‘effective’ government.
Each of these models of power distribution is associated with a distinctive mode of controlling power. In traditional terms, the mode of control characteristic of diffusion is ‘checks-and-balances’. So, for instance, the qualified presidential veto in the US system establishes a ‘balance of power’ between the executive and the legislature by dividing legislative power between Congress and the President. Sharing power between institutions enables each to ‘check’ the other. ‘Checking’ has two connotations: one is stopping or delaying, as in ‘checking someone's progress’; the other is supervising, as in ‘checking up on’ or ‘keeping an eye’ on someone. The mode of control characteristic of concentration is referred to here as ‘accountability’.Footnote 59 The classic example of this mode of control is ministerial responsibility to Parliament in the English system of government. Ministerial responsibility is the price that governments pay in parliamentary systems for the large amounts of unilateral power they enjoy.
A spatial metaphor may help to illuminate the difference between accountability and checks-and-balances. In the former case, the institution required to give account and the institution to which account must be given can be pictured as being in a vertical relationship. By contrast, institutions between which power is divided and shared can be pictured as being in horizontal relationships. So, for instance, in the English system, ministers are responsible to Parliament and are, in this sense, subject to it. In the US system, by contrast, the President is not responsible to Congress. Nevertheless, ‘oversight’ of the executive is one of the core functions of Congress. Another way of thinking about the difference between the various modes of control is in terms of a distinction between bipolarity and multi-polarity. A relationship of accountability can be pictured as bipolar (or ‘bilateral’) between an institution required to give account and an institution empowered to receive an account. By contrast, neither oversight nor checking carries any implication of bipolarity because power may be divided or shared amongst more than two (‘multiple’) institutions. It does not follow, of course, that an institution may not – in theory at least – be accountable to more than one other institution. However, each of those relationships will be best understood as discrete and bipolar. By contrast, an institution may be subject to oversight by, say, two institutions without being accountable to either in any formal sense. For instance, the US Presidency is subject to oversight by both Congress and the Supreme Court, but is not ‘responsible’ or, in my sense, ‘accountable’ to either.
In terms of the distinction between diffusion and concentration, the US is a relatively highly diffused system of government and the English system is relatively highly concentrated. The Australian system falls somewhere between the two. The characteristic mode of control in the English system is accountability, while the characteristic mode in the US system is checks-and-balances. In the Australian system, accountability predominates, but there are significant elements of checking-and-balancing.
This framework of analysis can, I suggest, throw light on the differences we have observed between our various systems. For explanatory purposes it is convenient to discuss the US first.
3.1. The US
Diffusion of power is the basic structural design principle of the US Constitution. So, legislative power is shared between the President and the two Houses of Congress. The courts also have a share of legislative power by virtue of their power to enforce the Constitution by invalidating Congressional legislation. Executive power is shared between Congress and the President, Congress being primarily responsible under the Constitution for creating the bureaucracy and the President being mainly responsible for day-to-day administration. Judicial review gives the judiciary, too, a share of executive power. The President has a share of judicial power by virtue of the power to nominate judges, and the Senate also has a share by virtue of its power to reject such nominations. And so on.
The Constitution says very little about the bureaucracy beyond the provisions that deal with the appointment of ‘officers of the United States’. The bureaucracy is largely a creature of Congress, which began the task of building government capacity in its very first session. The Constitution contemplated that Congress would be the dominant governmental institution; and for much of the nineteenth century it was, the President fulfilling the role of a sort of CEO of Congress, as contemplated in the Federalist Papers. However, by the late decades of the century, the demands on government had become so great that Congress could no longer effectively run the country by itself. The creation of the Interstate Commerce Commission (ICC) in 1887 was a watershed for two reasons. First, the ICC was structured in such a way as to limit the control exercisable over it by any one President. Secondly, and more importantly for our purposes, it was given a mixture of legislative, executive and judicial powers that is arguably inconsistent with the Constitutional architecture of three branches of government, each with a different – characteristic or dominant, even though not exclusive – type of power. This mixing of powers had come to be considered essential for successfully addressing the sorts of complex economic and social challenges thrown up by the Industrial Revolution. The problem was that up to this point ‘separation of powers’ had been one of the main legitimising tropes of American constitutional discourse, and it still exerts significant force in US law. If separation of powers was not available to legitimise these new, multi-functional bureaucratic agencies, what could take its place?
The first solution to the problem was found (around the turn of the twentieth century) in concepts such as ‘expertise’ and ‘scientific management’, and the distinction between politics and administration. In sum, administrative agencies came to be understood as exercising neither legislative, nor executive (political), nor judicial power, but rather independent, neutral technological and scientific expertise. This was thought to justify classifying them as a new, fourth branch of government exercising a distinct type of ‘bureaucratic’ power focused on finding instrumental means to political ends.Footnote 60 The separation of administration from politics can be understood as a further diffusion of public power. Moreover, because the agencies have two masters (having been created by Congress and being under the day-to-day oversight of the President), or even three (especially with the strengthening and codification of judicial review in the APA), they can maintain quasi-independence from all of them. Put differently, multi-functional agencies can be seen as sharing power with all three of the other branches of government and, to that extent, being under the control of (‘accountable’ to) none of them.
A second solution was found about half a century later in ideas of democratic pluralism and participation, which were associated with a loss of faith in neutral, scientific expertise as a solution to social and economic challenges, and the rise of public choice models of politics and public affairs. In such accounts, democracy was understood as a mechanism for constructing ‘the public interest’ by means of a contest between individual and group interests. Therefore, it required ‘interest representation’ of some sort to legitimise government processes.Footnote 61
My argument is that both of these legitimising discourses – expertise and democracy – help to explain the emphasis on rationality in US administrative law. Means-ends rationality is of the very essence of technocratic and scientific expertise. Similarly, a vision of the bureaucratic policy-making process as a market place of ideas, coupled with the espousal of democratic and republican ideals that were foundational to the American state-building exercise, is very congenial to what has been called (in a different context) ‘a culture of justification’. Reason-giving and rationality serve both political ideals and instrumental efficiency, and find their concrete bureaucratic expression in the record. Rationality becomes the prime legal, judicially protected value, and the record the mechanism for promoting that value. The requirement of rationality establishes a delicate balance of power between the bureaucracy, the courts and the political branches by giving the bureaucracy a characteristically distinctive, quasi-independent role in the governmental system.
3.2. England
In the English system of government, public power is highly concentrated (although the degree of concentration waxes and wanes, and has been significantly reduced in the past 50 years). Concentration is expressed theoretically and symbolically in the concept of sovereignty. In the US system, no organ of government is sovereign; sovereignty resides in the people; all governmental power is delegated. In the English system, by contrast, sovereignty originally belonged to the monarch. In the seventeenth century it shifted to Parliament and, in the twentieth, to the executive (in practice, if not in theory: Lord Hailsham's ‘elective dictatorship’).Footnote 62 The heavy concentration of legislative, executive and bureaucratic power in the government of the day explains why judicial independence and the rule of law are central tropes in English constitutional discourse: lawFootnote 63 is understood as a counterweight to power, and judicial independence is essential to enable the courts to protect the citizen, in the name of law, from abuse and misuse of power.
Although not as strongly as in the Australian system (as we shall see), judicial power must be protected from being contaminated by mixture with legislative and executive power so that it can provide as effective a bulwark as possible for citizens against the government. Before the English Revolution, judges of the common law courts were royal officials actively involved in various ways in legislative and executive government. After the Revolution (in the Act of Settlement 1701), the common law courts were given ‘independence’ from the monarchy while, at the same time, being subordinated to Parliament, their prime function thenceforth was to enforce its sovereign will. As the administrative state grew and power became more and more concentrated – partly as a result of the development of the principle of ministerial responsibility and later of political parties – the importance of judicial power and independence as a counterweight to government became more and more obvious, bearing much legal fruit from the 1960s onwards. This role of the courts as a source of recourse and reparation against a strong government that has overstepped the mark is one manifestation of what I refer to as ‘accountability’.
Concentration of power affects not only the relationship between the government and the courts but also that between the executive and the bureaucracy (to say nothing of the relationship between the executive and the legislature). In the US system, as we have seen, there is a significant sense in which the bureaucracy is a quasi-independent fourth branch of government, sharing power with each of the other three branches. In the English system, the bureaucracy's relationship with the executive (and hence, with Parliament and the courts) is quite different. Whereas the US bureaucracy has two or three ‘masters’, the English bureaucracy has only one: the executive. Whereas control of the US bureaucracy is shared between the President, Congress and the courts, in England the executive has virtually complete control over the civil service. Whereas the US bureaucracy is commonly understood as exercising a characteristic form of power – the power of neutral expertise – the English bureaucracy is understood as the faithful servant of the political executive. Whereas US agencies are quasi-independent policy-makers, English bureaucrats involved in policy-making are dependent on and subordinate to their political masters.Footnote 64 To the extent that English bureaucrats provide expertise, they do so in the service of the government's political agenda and not, like their US counterparts, in pursuit of a distinct ‘administrative’ agenda.
I would argue that these structural differences between the US and English systems help to explain why rationality – and hence reason-giving and records – play a less significant role in English judicial review law than in its US equivalent. Since the Revolution, English courts have assumed a ‘subordinate’ role in the English constitution, their prime function being understood in terms of faithfully giving effect to Parliamentary legislation. Courts did not, of course, cease to make (common) law. However, whereas in the medieval way of thinking common law and statute were both, in some sense, expressions of an underlying ‘customary’ law, by the nineteenth century judge-made law had been firmly subordinated to statute and itself conceptualised in fully positivistic terms as a product of institutional activity. From this perspective, Dicey's account of the rule of law can be seen to hark back to an earlier, pre-positivist understanding of the nature of law and its relationship with institutional activity. My argument is that by the twentieth century in England, courts had come to be characterised, vis-à-vis Parliament, as what Richard Stewart famously called ‘transmission belts’.Footnote 65 Older traditions of flexible statutory interpretation gave way to formalism, and doctrines of precedent and stare decisis hardened. Much more than their American cousins, English courts saw themselves as apolitical functionaries, more or less past the age of legal child-bearing. In this environment, I would suggest, it is not surprising that a strong ethos of judicial reason-giving did not develop. After all, the role of the courts was to apply law, not create it, and the only reason needed was that the law being applied existed. Even the common law was hidden in a cave waiting to be discovered and brought out.
On the other side, it is not surprising that English courts were very slow to impose obligations on the executive and the bureaucracy to give reasons. Even in the US, Congress has no obligation to give reasons: its legitimacy and authority derive from being elected to make law, not the rationality of the laws it makes.Footnote 66 Politics is first and foremost about power, not reason. In the English system, similarly, the legitimacy of political power, whether or not exercised through the medium of Parliament, depends on ‘democratic’ support rather than rationality, and bureaucrats are in the service of the political. Like courts, they are transmission belts, not motors, and the logic against giving reasons applies to both. Ultimately, all public officials and institutions are in the service of politico-legal sovereignty, wherever that is located at any particular time. As a result, courts constantly walk a tightrope between their allegiances to the sovereign on the one hand and the rule of law on the other. Be that as it may, we can now see why, in a system where sovereignty resides somewhere in the government machine and power is concentrated in the sovereign, rationality is not as highly valued as it is in a system in which sovereignty is located outside government, and governing institutions must constantly bargain and cooperate with one another.
3.3. Australia
The Australian federal legal system (like the legal systems of the Australian states) inherited the post-Revolutionary English model of the role of courts as apolitical functionaries. In the Australian system, the value of judicial independence and separation of judicial power serve not only to provide protection for citizens against abuse and misuse of the concentrated power of the government, but also to protect the states against the Commonwealth. The High Court has always seen a need for a super-strong and independent federal judiciary. This strength has been found in a narrow concept of the judicial function, defined in terms of a strict version of the rule of law. Not only is it the function of the courts to enforce the law against the government – that is their only function. Judicial power is distinguished sharply not only from legislative power (with the result that the general style of Australian adjudicating is legalistic and formalistic) but also from executive power. As we have seen, in Australian law not only may non-judicial bodies not exercise judicial power, but judicial bodies may not exercise non-judicial – and especially executive – power. The desire of Australian courts to distinguish themselves and their source of authority from that of political institutions helps to explain why reason-giving has, for a long time, been considered a judicial obligation: it is a marker of the distinctiveness of the judicial function.
The unwillingness of Australian courts to impose obligations to give reasons on the executive can, I think, be explained in much the same terms as in the case of England. The relationship between the executive and the bureaucracy in Australia is essentially similar to that in England. Although Australian bureaucrats are called ‘public’ rather than ‘civil’ servants, they are nevertheless in the service of the government, notFootnote 67 of the people. Furthermore, although, by virtue of the composition and legislative strength of the Senate, legislative power is more diffused in the Australian system than it is in England, the Australian Parliament lacks the control enjoyed by Congress over the bureaucracy, both because the government necessarily controls the House of Representatives and because Australian governments inherited the prerogative (non-statutory) power to create and manage the bureaucracy. The Australian bureaucracy has a single master – the political executive – and the sources of its legitimacy are representative democracy and responsible government, not rationality.
4. Conclusion
My argument, then, is that structural characteristics of our three systems help to explain why rationality, reasons and the record play a far more significant role in the US law of judicial review than in either the English or the Australian law. If this is right, how are we to explain the development of a ‘culture of justification’ and, in particular, the increase in reason-giving obligations in England and Australia over the past 50 years? In Australia, it is reasonably accurate to say that as a result of the imposition of widely applicable statutory obligations to give reasons, the unwillingness of courts to impose such obligations is of little practical importance. Conversely, in England, where statutory obligations are less widespread, the courts have, to some extent, been willing to fill the gaps by inventing common law obligations to give reasons. Why have governments in systems of concentrated power become willing to create and accept such obligations? After all, turkeys don't normally vote for Christmas.Footnote 68
Leaving aside local explanations and histories of particular statutory provisions, my speculation would be that the spread of administrative reason-giving obligations is a sign of increasing disenchantment with party politics as a mode of constituting public power, and of a popular demand for ‘evidence-based’ public policy-making arising from rapidly increasing knowledge of the world around us, rising educational levels and, perhaps, unrealistic estimates of the power and potential of science and information technology. Obligations to give reasons are relatively cheap responses to the growing distrust of politicians and the relatively high trust in science and technology. Whether or not this is true, I would suggest that the analysis undertaken in this article at least helps us to understand why the concept of the ‘culture of justification’ is more common in the discourse of administrative lawyers in systems of concentrated power (the United Kingdom, Australia, New Zealand, Canada, and so on) than it is in the US. A commitment to rationality is more deeply embedded (and, therefore, more taken for granted) in the US administrative state than in its English and Australian counterparts. Of course, in a world where disagreement about ends is endemic and relative ignorance about means will, like the poor, always be with us, some scepticism about rationality seems reasonable. Demands for reasons and rationality are one way of disciplining power, but we need other tools as well when we get to the point where knowledge and logic run out.
