• Specific Year
    Any

Airo-Farulla, Geoff --- "Rationality and Judicial Review of Administrative Action" [2000] MelbULawRw 23; (2000) 24(3) Melbourne University Law Review 543

Rationality And Judicial Review Of Administrative Action

GEOFF AIRO-FARULLA[*]

[Courts engaged in judicial review of administrative action are increasingly using the term ‘rationality’ to describe at least some aspects of the standards of good administrative decision-making embodied in judicial review grounds. However, while ‘rationality’ is probably the most ubiquitous concept used in studying how humans behave, its meaning is often unclear. Part II of this article explores the concept of ‘rationality’ by reference to a means–end model of rational action, supplemented by a model of rational belief. Part III argues that a fundamental doctrinal shift has occurred in the grounds of judicial review, essentially during the latter part of the 20th century. The legality–merits distinction has replaced the distinction between jurisdictional and non-jurisdictional questions as an organising principle of judicial review. Jurisdictional questions remain relevant, but legality is wider than jurisdiction and encompasses rationality requirements. Rationality is thus a key to understanding the legality–merits distinction as it currently operates in Australian administrative law. Part IV discusses in more detail three issues concerning the relationship between legality and rationality: first, the question whether the rationality standard involves an objective or subjective test, particularly in relation to judicial review of factual findings; second, the extent to which rationality review justifies review of the relationship between means and ends; and, third, the relationship between rationality and the more commonly used term ‘reasonableness’.]

INTRODUCTION

In 1985, in a ‘classical’[1] statement of administrative law principles, Lord Diplock named ‘irrationality’ as one of the three main grounds of judicial review.[2] Since then, English and Australian judges have increasingly used the term ‘rationality’ when referring to the standards of good administrative decision-making embodied in judicial review grounds. Thus, in TCN Channel Nine Pty Ltd v Australian Broadcasting Tribunal French J said:

[T]here is a pervasive requirement for rationality in the exercise of statutory powers based upon findings of fact and the application of legal principle to those facts ... A serious failure of rationality in the decision making process may stigmatise the resultant decision as so unreasonable that it is beyond power. Alternatively, lack of rationality may be reflected in a failure to take into account relevant factors or the taking into account of irrelevant factors. Each of these heads of review seems to collapse into the one requirement namely that administrative decisions in the exercise of statutory powers should be rationally based.[3]

Lord Diplock himself equated irrationality with ‘Wednesbury unreasonableness’.[4] In Associated Provincial Picture Houses Ltd v Wednesbury Corporation Lord Greene MR had said that the court could ‘interfere’ where a decision-maker comes ‘to a conclusion so unreasonable that no reasonable [decision-maker] ... could ever have come to it’.[5] Perhaps in an attempt to avoid the circularity of this formulation, Lord Diplock said that the unreasonableness or irrationality ground of review applies to ‘a decision which is so outrageous in its defiance of logic or of accepted moral standards that no sensible person who had applied his mind to the question to be decided could have arrived at it.’[6] This approach restricts rationality to a fairly narrow compass but a requirement that administrative decisions be ‘rational’ is potentially of much greater scope. The concept of rationality can be seen as underpinning many of the grounds of judicial review concerned with the reasons for which a decision was made, such as relevant and irrelevant considerations, proper purposes and bad faith.[7]

‘Rationality’ is probably the most ubiquitous concept used in studying how humans behave as individuals, members of groups, and in institutions. It plays a central role in disciplines as varied as economics, psychology, philosophy, the philosophy of science, public administration, organisational theory, sociology and business management — not to mention law. However, it is a ‘highly ambiguous expression’,[8] and therefore in Part II of this article I explain what I mean by ‘rationality’.

In Part III, I argue that the legality–merits distinction has largely replaced the distinction between jurisdictional and non-jurisdictional questions as an organising principle of judicial review. But, while the legality–merits distinction is now fundamental, the courts have provided little guidance as to where the line between legality and merits should be drawn. As the English judge Sir John Laws has written extrajudicially, the idea that the courts’ role is to keep administrative agencies within their legal power, while leaving the merits of how their power is exercised to them, says nothing of ‘what the court will count as want of power in the deciding body; and so of itself it illuminates nothing.’[9] I argue that the emergence of a rationality standard of review was necessary for the shift from jurisdiction to legality to occur, and that rationality is thus a key to understanding the legality–merits distinction.

In Part IV, I discuss three issues concerning the relationship between legality and rationality: first, the question of whether the rationality standard involves an objective or subjective test, particularly in relation to judicial review of factual findings; second, the extent to which rationality review justifies review of the relationship between means and ends; and third, the relationship between rationality and the more commonly used term ‘reasonableness’.

II ‘RATIONALITY’

The simplest way to explain what I mean by ‘rationality’ is to begin with a simple ‘means–end’ model of rational action. In theories of individual rational action, ‘rationality’ is primarily used to describe an adaptive relationship between means and ends.[10] That is, the ‘common sense notion of rational behaviour’ is ‘behaviour involving a choice of the best means available for achieving a given end[11] (the ‘means–end’ model of rational action). In the ideal means–end model of rational action, the rational actor:

  • specifies the ends to be achieved;
  • identifies all possible ways or means by which those ends might be achieved;
  • considers all the possible consequences of each option; and
  • chooses the means which best achieves the specified ends.[12]

For example, a ‘rational’ law student nearing graduation may desire to be admitted to practice as a solicitor. Admission to practice is therefore the end to be achieved. A survey of possible means (ways of gaining admission open to the student) may identify articles of clerkship of varying lengths and rates of pay in different States, and postgraduate practical legal training courses of varying lengths and costs at different institutions. As well as considering each option’s duration and economics, the student may also consider the quality of the training that each option provides, the transferability of qualifications between jurisdictions, family obligations and relationship issues, quality of life issues, and many other factors. Taking all relevant factors into account, the student would then rank the options from best to worst. Rationally, the student should then choose the highest ranked (best) option over any other. It would be ‘irrational’ for the student to choose an option that was not the best means available to achieve their ends.

This ideal means–end model of rational action raises two questions. First, is it concerned only with the relationship between means and ends, or is it also concerned with the rationality of the ends to be achieved? Secondly, what are the implications of ‘bounded rationality’[13] for such a means–end model of rational action? These issues are explored below, as is the relationship between rationality and formal logic.

A The Rationality of Ends

In economics the ends that a rational actor may wish to pursue are frequently assumed to be beyond or outside rational analysis or justification.[14] Thus, economists usually assume that all human beings always have the same end: maximising their own utility.[15] However, other disciplines, particularly sociology and psychology, are directly concerned with the process by and extent to which individuals choose their ends (as well as their means). Furthermore, the distinction between means and ends, while analytically useful, is not hard and fast. Most ‘ends’ can themselves be conceived of as ‘means’ to further ends. In the example given above, the student’s ‘end’ of admission as a solicitor is itself a means to another end, that of obtaining a financially rewarding and personally satisfying job. The rationality of the student’s choice, from amongst the various career options available, of admission to practice as the means to this end can clearly be assessed, in the same way as their choice of means of getting admitted. A student who sought admission to practice as a solicitor, when a better way of obtaining the kind of job they wanted was available, would act irrationally.

The ‘ends’ that truly are only ‘ends in themselves’, and not means to achieving other ends, tend to be either quite trivial, or so general as to explain very little. Thus, there may truly be ‘no accounting for taste’ when the choice is between apples and oranges and, ultimately, we probably all want to be happy. Between these two extremes, however, is a complex web of intermediate, often competing, short- and long-term ends. The way in which we order and pursue these intermediate ends is susceptible to rational analysis.

For example, it is irrational to knowingly pursue inconsistent or incompatible ends. Thus, a person who prefers end A to end B, and end B to end C, cannot then prefer end C to end A without violating a norm of ‘consistency of preference’ which is usually considered to be necessary for rational action (unless, of course, the person has rational reasons for changing their mind). More generally:

It has been recognized, from the time of Plato onwards, that it is the function of reason not merely to devise means but to reconcile or harmonize ends. To achieve this, the agent must keep all his relevant ends in mind, he must ensure that his scale of preferences is internally consistent, and then he must work out which ends are in his circumstances capable of joint achievement and which have to be sacrificed.[16]

Thus, a means–end model of rational action must be concerned not only with the relationship between means and ends, but also with the rationality of the ends themselves.[17] Put simply, the rational pursuit of an irrational end is irrational.

B Bounded Rationality and Rational Beliefs

The ideal means–end model of rational action set out above assumes that the rational actor can identify all possible means to the desired end, and can anticipate all the possible consequences of each course of action, so that the one that best achieves the desired ends can be chosen. However, in practice, possessing such perfect information is impossible because the human capacity to process information and solve problems is limited.[18] Because human rationality is necessarily ‘bounded’ in this way, it is unrealistic to demand that rational action always be based on correct knowledge and beliefs. But, while rational action cannot be based on perfect knowledge, it cannot be based on irrational beliefs either. That is, rational action must be underpinned, not by perfect knowledge, but by beliefs which are themselves rational (even if not correct), concerning the relevant facts, the possible ways or means in which the end might be achieved, and the possible consequences of each option.[19]

Thus, the means–end model of rational action must be supplemented with a model of rational belief.[20] According to Elster, beliefs are rational when they have ‘the right kind of relation to the evidence available’.[21] Two conditions must be satisfied for this ‘right kind of relation’ to exist.[22] First, the ‘optimal’ amount of evidence must have been collected as a basis for the relevant belief.[23] This condition can be violated in several ways:

A fairly common procedure is to stop collecting evidence when the belief that is rationally justified by the evidence collected so far is also the one that one wants to be true. If the amount collected is below the lower bound of information that ought to be collected, this behaviour is certainly irrational; similarly if one goes on collecting evidence beyond the upper bound in the hope that eventually one will have reasons for holding the belief which one would like to be true.[24]

Also, rational actors must be aware that collecting evidence is costly in terms of time and/or money. Thus, collecting further evidence is irrational if the cost of doing so is clearly greater than the savings to be made by making a more informed decision. A thorough investigation is also irrational if a swift decision is required.[25]

The second condition that must be satisfied for a belief to be rational is that the evidence must rationally support it.[26] In very general terms, beliefs are rationally supported by evidence in two situations. The first is what the law calls ‘primary facts’ and what Mortimore and Maund term ‘experiential beliefs’.[27] These are beliefs that are based on a person’s direct experience of the world: ‘Where an individual believes [a fact] because he [or she] has seen, heard, smelt, etc. that [fact] ... he [or she] holds an experiential belief.’[28] The second situation, which includes what the law calls ‘inferences of fact’, is where the belief is not based on direct experience, but can be rationally drawn from other rationally held beliefs (‘inferential belief’).[29]

Rationality of belief is distinct from correctness of belief. Beliefs can be rationally supported by the evidence, yet wrong. As Elster explains:

Not only is it human to err: it may even be rational to do so, if all the evidence happens to point in the wrong direction. ... [T]he process can be rational and yet fail to reach the truth. Truth is a relation between a belief and what the belief is about. ... By contrast, rationality is a relation between a belief and the grounds on which it is held.[30]

At the same time, beliefs can be empirically true, yet irrational. For example, a person who decides what to believe by throwing a die may happen on the truth, even though the evidence on which their belief was based does not rationally support it.[31]

C Rationality and Formal Logic

While rationality and logic are commonly equated, both in common speech and in more theoretical discourses, they are not synonymous.[32] Formal logic is an abstract system of ‘deductive’ reasoning.[33] Reasoning is deductive if, and only if, the conclusion necessarily follows from the given premises. For example, given the premise ‘if P exists, then Q exists’, then, logically, if P exists, so must Q — irrespective of what P and Q actually stand for. Thus, given the premise ‘if it is the 24th of December, then it is Christmas Day’ and the fact ‘today is the 24th of December’, then the only logically valid conclusion is ‘today is Christmas Day’ — even though everybody knows that it is not, and even if believing the premise is quite irrational. This is because formal logic is concerned only with the validity of the conclusions drawn from given premises, not the validity of the initial premises themselves. In the real world, of course, working premises can be empirically wrong, or based on irrational beliefs. Conclusions based on such premises may be logically impeccable, but still wrong or irrational.[34]

Furthermore, formal logic only encompasses ‘deductive’ reasoning, where the premises compel the conclusion. Strictly speaking, ‘inductive’ reasoning is formally illogical. Inductive reasoning involves the drawing of a conclusion that is not strictly compelled by the given premises, because it is not based solely on the information supplied by the premises.[35] Rather, the person making the conclusion has added some extra information, usually from their general store of knowledge about the world.[36] The added information often involves an assessment of the probability that a particular event has occurred or will occur, an assessment usually made by generalising from past experience.[37] Such inductive reasoning contravenes the rules of logic, but is essential to real world intelligence:

The most obvious way in which inductive inference is vital to our intelligence is through concept learning and empirical generalization. From the viewpoint of logic, a universal statement of the form ‘All A are B’ can never be verified, no matter how many times you observe an A that is a B, but can be disconfirmed by one single instance of an A that is not a B. If in practice we were to reason logically in this regard, intelligence would be effectively impossible. Given the vast amount of information with which we are confronted in the real world, the ability to induce concepts, categories, and general rules is absolutely vital. Consider, for example, a belief such as ‘All cats have tails’. Logic tells us that we have no right to form such a belief no matter how many cats we observe that have tails. Moreover, logic tells us that should we hold this belief as a working hypothesis, then we must abandon it upon observing a single tailless cat, for example, a Manx breed or one that lost its tail in an accident.

In order to be intelligent we have to do the exact opposite to what logic tells us. We need to form general beliefs in order to organize our knowledge of the world and to predict the nature of objects and events. Moreover, we should not necessarily abandon our beliefs when isolated or rare falsifying instances are encountered. It is effective to hold general rules and store exceptions in ... ‘default hierarchies’. Thus we hold as a general rule that ‘All birds can fly’ and store the exceptionality of the [emu] as part of its specific concept.[38]

Thus, violating the rules of formal logic is sometimes, but not always, the same as irrationality. Deductive logic requires a conclusion to be compelled by given premises (irrespective of the premises’ rationality). Rationality requires a conclusion to be consistent with rational premises. A conclusion that is inconsistent with rational premises is both illogical and irrational. A conclusion that is induced from (and therefore consistent with) rational premises is rational, but illogical. A conclusion that is compelled by irrational premises is logical, but irrational.

III JURISDICTION, THE LEGALITY–MERITS DISTINCTION

AND RATIONALITY

As outlined above, in this Part, I argue that the legality–merits distinction has replaced the distinction between jurisdictional and non-jurisdictional questions as an organising principle of judicial review and that the emergence of a rationality standard of review was necessary for this shift to occur. Rationality is thus a key to understanding the legality–merits distinction.

A From Jurisdiction to Legality

Jurisdiction emerged as an organising principle of both certiorari and prohibition in the 17th century.[39] Prior to that, only prohibition had been concerned with jurisdictional questions.[40] Certiorari, on the other hand, as a judicial review technique,[41] originally developed to remove cases from other courts into the King’s Bench for trial, even before they had been heard.[42] In a series of 17th-century cases beginning with Commins v Massum[43] and ending with Groenwelt v Burwell,[44] the King’s Bench developed a new remedy — certiorari to quash for jurisdictional error — out of the older forms of certiorari.[45] This series of cases also established the King’s Bench as a ‘superior’ court, with supervisory jurisdiction over all other (henceforth ‘inferior’) courts.[46] Thus, the scope of the Court’s supervisory jurisdiction was widened but, paradoxically, the form of its supervision was strikingly narrowed. During the 18th century the degree of supervision became even more limited as the courts narrowed their concept of jurisdiction, creating a distinction between jurisdictional and non-jurisdictional errors of law.[47]

In theory, jurisdictional questions were those which related to the scope of the decision-maker’s power to decide, and non-jurisdictional questions were those that arose in the exercise of an admitted decision-making power. All jurisdictional questions were subject to judicial review for correctness. That is, a reviewing court would decide for itself the right answer to the jurisdictional question. All non-jurisdictional questions were immune from judicial review.[48] Thus, under the jurisdictional approach, no distinction was expressly made between factual, legal and discretionary questions in terms of the standard of review to be applied: factual questions could be jurisdictional and legal questions could be non-jurisdictional. This helps to explain why the distinction between factual and legal questions is so ill-defined.[49] That distinction actually developed first to define the scope of statutory rights to appeal on questions of law, rather than in common law judicial review cases.

In practice, the jurisdictional approach was unsatisfactory. The courts’ tendency was to treat discretionary questions and most factual questions as non-jurisdictional, and most statutory interpretation questions as jurisdictional. However, they had no theory to explain why this should be so, nor any explicit criteria for classifying a question as jurisdictional or non-jurisdictional, making prediction and consistency difficult.[50] More importantly, the jurisdictional approach gave the courts an all-or-nothing choice in terms of the standard of review.[51] Either the court had to decide for itself what the correct answer to the question was, or hold the question completely unreviewable.

Legality replaced jurisdiction as an organising principle through the development of a rationality standard as a ‘halfway house’ between reviewing for correctness and no review at all.[52] The key difference between legality and jurisdiction is that legality is a broader concept, encompassing, but not limited to, jurisdictional questions.[53] Under the legality approach, the courts continue to apply a correctness standard to jurisdictional questions. However, a question is no longer immune from review simply because a court classifies it as non-jurisdictional and therefore will not apply a correctness standard to it. The courts have progressively developed ‘techniques for reviewing the quality of decision-making even on matters which would traditionally be analysed as within jurisdiction’.[54]

This fundamental doctrinal shift was central to administrative law’s development during the 20th century, and occurred primarily after World War II. For example, the doctrine of relevant and irrelevant considerations is central to the rationality standard. In Minister for Aboriginal Affairs v Peko-Wallsend Ltd[55] the High Court provided a comprehensive discussion of that doctrine’s operation in Australia. Of all the cases cited by the Court, only one dated from before 1948.[56] This is not to say that rationality review has no pre-World War II antecedents. As well as other relevant considerations cases not referred to by the High Court, there are cases from the seventeenth,[57] eighteenth,[58] nineteenth[59] and early twentieth centuries[60] holding that an objectively unreasonable decision or piece of delegated legislation was unlawful. There are also pre-World War II cases on other aspects of rationality review, such as ‘proper purposes’[61] and ‘no evidence’.[62]

However, there is no doubt that the real development of the grounds of review of reasons for administrative decision-making has occurred since World War II. Lord Greene MR’s judgment in Wednesbury is often pointed to as a key step along this path. The irony is that he was undoubtedly seeking to draw narrow boundaries around rationality review, but inadvertently set the stage for its massive expansion. One of the most important and influential aspects of Lord Greene MR’s judgment is his identification of a family resemblance between the doctrines of reasonableness, proper purposes, relevant considerations and so on:

It is true the discretion must be exercised reasonably. Now what does that mean? ... It has frequently been used and is frequently used as a general description of the things that must not be done. For instance, a person entrusted with a discretion must, so to speak, direct himself properly in law. He must call his own attention to the matters which he is bound to consider. He must exclude from his consideration matters which are irrelevant to what he has to consider. If he does not obey those rules, he may truly be said, and often is said, to be acting ‘unreasonably’. Similarly, there may be something so absurd that no sensible person could ever dream that it lay within the powers of the authority. Warrington LJ in Short v Poole Corporation gave the example of the red-haired teacher, dismissed because she had red hair. That is unreasonable in one sense. In another sense it is taking into consideration extraneous matters. It is so unreasonable that it might almost be described as being done in bad faith; and in fact, all these things run into one another.[63]

All these grounds of review are concerned with the rationality of the decision-maker’s reasons for decisions, rather than with what they decided.[64] Although not in so many words, Lord Greene MR was insisting that decision-makers had to have rational reasons for their decisions. Since Wednesbury, the grounds of review concerned with the decision-maker’s reasons for decisions have tended to eclipse simple jurisdictional questions and, along with procedural fairness, have become the courts’ prime concern when engaging in judicial review. In effect, under the rubric of legality, a rationality standard has been added to the correctness standard.[65] Legality, then, can be summarised as ‘jurisdiction (or power) plus rational reasoning (rationality)’.

It was a sign of the times that Lord Greene MR wrote his judgment three years after the publication of the first edition of Simon’s Administrative Behavior,[66] which stands as a landmark in our understanding of both rational action and rational government administration. The rationality standard’s emergence as a criterion of legality reflects the emergence of new ways of understanding administrative processes more generally.

B The Source of Rationality Requirements

Debate continues today about the source of such legal requirements of rational decision-making. This reflects the wider debate about the source of the grounds of judicial review generally.[67] The traditional and still dominant view is that the rationality grounds of review stem from implied limits on Parliament’s grant of decision-making power.[68] The competing — and I believe better — view is that they reflect common law standards of good administrative decision-making, as is now accepted in Australia in relation to procedural fairness requirements.[69]

Lord Diplock was the first judge to identify rationality as a common law requirement, rather than one implicit in legislative grants of decision-making power. He initially drew rationality out of common law procedural fairness principles. Thus, in Mahon v Air New Zealand Ltd he held that procedural fairness requires not just a fair hearing, but also rational fact-finding.[70] In Australian Broadcasting Tribunal v Bond, Deane J sought to develop the link between common law procedural fairness and rationality into a more general set of rationality requirements:

If a statutory tribunal is required to act judicially, it must act rationally and reasonably. Of its nature, a duty to act judicially (or in accordance with the requirements of procedural fairness or natural justice) excludes the right to decide arbitrarily, irrationally or unreasonably. It requires that regard be paid to material considerations and that immaterial or irrelevant considerations be ignored. It excludes the right to act on preconceived prejudice or suspicion. Arguably, it requires a minimum degree of ‘proportionality’.[71]

These cases are important because they demonstrate judicial willingness to explicitly articulate rationality as a common law standard of good administrative decision-making, like procedural fairness, rather than basing it on the fiction of implied legislative intent. However, rationality and fairness are distinct principles with distinct rationales. Fairness may require that a person be given an opportunity to be heard, even where that does not appear to advance the decision’s rationality.[72] Conversely, rationality is a universal requirement of good administrative decision-making, whereas procedural fairness is not, as it is only required where a decision will directly affect a person’s rights, interests or legitimate expectations.[73] Rationality, therefore, should not be subsumed under procedural fairness, but should stand in its own right as a basic common law value.

Indeed, Lord Diplock seemed to recognise as much in 1985, when he named ‘irrationality’ as one of the three basic grounds of judicial review, along with ‘illegality’ and ‘procedural impropriety’.[74] More recently, Kirby J stated that ‘judicial review is designed, fundamentally, to uphold the lawfulness, fairness and reasonableness (rationality) of the process under review.’[75] These statements are important for their recognition of rationality as a distinct common law concept. Further, ‘once irrationality is recognized as an independent ground of review, closer consideration may be given to developing ... more exacting criteria of review.’[76] Certainly, since Lord Diplock’s statement in CCSU, many judges have more expressly embraced the language of rationality, and used the concept to further develop the grounds of judicial review.

However, these statements from Lord Diplock and Kirby J do not properly articulate the relationship between legality (or lawfulness) and rationality. They treat legality and rationality as alternative grounds, instead of treating rationality as part of the wider concept of legality. The basic grounds of judicial review should be classified as jurisdiction (or power), procedures and rationality (that is, legality = jurisdiction + procedural propriety + rationality).[77]

C Legality, Jurisdiction and Rationality

Obviously, the emergence of the rationality standard has not obviated the need for courts to engage in the difficult task of distinguishing between jurisdictional and non-jurisdictional questions, as that distinction still determines the standard of review that is applied. Under the legality approach it is still important to distinguish between jurisdictional facts, which are reviewed for correctness, and non-jurisdictional facts, which are reviewed for reasonableness.[78] Review of non-jurisdictional facts for reasonableness is discussed further below,[79] but exploring the nature of the principles that should be used to distinguish between jurisdictional and non-jurisdictional facts is beyond this article’s scope. The key point I wish to make at this stage is that non-jurisdictional facts are clearly now reviewable, whereas previously they were immune from review.

In relation to questions of law, what is clear is that Australian courts review jurisdictional questions of law for correctness — that is, the courts will decide such questions of law for themselves. In Australia it remains unclear whether, at common law,[80] all questions of law go to an administrative decision-maker’s jurisdiction, as is the case in the United Kingdom,[81] or whether a class of non-jurisdictional error of law continues.[82] However, to the extent that a class of non-jurisdictional questions of law still exists, such questions are still treated as immune from review.[83] Australian courts have not systematically explored the possibility of non-jurisdictional questions of law being reviewable for rationality,[84] rather than being immune from review. This raises the issue of the appropriate role, if any, of a doctrine of deference to reasonable administrative interpretations in Australia.[85] Clearly, the possibility that non-jurisdictional questions of law could be reviewed for reasonableness, rather than being immune from review, may influence whether a court characterises a question as jurisdictional or not. I will not explore this question further in this article, as it is beyond the scope of the current discussion.

However, the persistence of the language of ‘jurisdictional error’ in prerogative writ cases should not obscure the shift from jurisdiction to legality. Formally, the doctrinal principle that prerogative writs only lie for jurisdictional error[86] remains unchanged, but what counts as jurisdictional error has been substantively broadened to include the grounds of rationality review. In this way the discretionary aspects of decision-making, in particular, have been brought within the scope of judicial review. It is true that the scope and operation of these grounds of review remain somewhat less developed in prerogative writ cases[87] than in proceedings under the ADJRA.[88] However, the trend is clearly in the same direction in both kinds of cases.

For example, in Minister for Immigration and Ethnic Affairs v Guo, an ADJRA case, the High Court set out what is required for a decision-maker to have a ‘rational basis’ for determining whether an applicant for refugee status has a well-founded fear of persecution.[89] In the prerogative writ proceedings in Abebe, Gleeson CJ and McHugh J quoted that passage with approval,[90] and went on to test whether the Refugee Review Tribunal had followed the ‘logical’ process of reasoning that it was legally ‘bound’ to follow.[91] Finding that it had followed the correct process of reasoning, they concluded that no jurisdictional error had occurred.[92]

D Legality, Rationality and Merits

Of course, in both ADJRA and prerogative writ cases the courts still insist that they cannot review the ‘merits’ of decision-making (that is, legality = jurisdiction + procedures + rationality – merits). While exclusion of merits review under the jurisdictional approach was originally a self-denying ordinance, the legality–merits distinction is now underpinned by a separation of powers analysis and a recognition of the courts’ limited competence, as Brennan J pointed out in A-G (NSW) v Quin:

[T]he court needs to remember that the judicature is but one of the three coordinate branches of government and that the authority of the judicature is not derived from a superior capacity to balance the interests of the community against the interests of an individual. ...

Some advocates of judicial intervention would encourage the courts to expand the scope and purpose of judicial review ... Such advocacy is misplaced. If the courts were to assume a jurisdiction to review administrative acts or decisions which are ‘unfair’ in the opinion of the court — not the product of procedural unfairness, but unfair on the merits — the courts would be assuming a jurisdiction to do the very thing which is to be done by the repository of an administrative power, namely, choosing among the courses of action upon which reasonable minds might differ ... The courts — above all other institutions of government — have a duty to uphold and apply the law which recognizes the autonomy of the three branches of government within their respective spheres of competence and which recognizes the legal effectiveness of the due exercise of power by the Executive Government and other repositories of administrative power. ...

If judicial review were to trespass on the merits of the exercise of administrative power, it would put its own legitimacy at risk. ... If the courts were to postulate rules ostensibly related to limitations on administrative power but in reality calculated to open the gate into the forbidden field of the merits of its exercise, the function of the courts would be exceeded ...[93]

In a frequently quoted passage in Minister for Aboriginal Affairs v Peko-Wallsend Ltd Mason J similarly said:

The limited role of a court in reviewing the exercise of a discretion must constantly be borne in mind. It is not the function of the court to substitute its own decisions for that of the administrator by exercising a discretion which the legislature has vested in the administrator. Its role is to set limits on the exercise of that discretion, and a decision made within those boundaries cannot be impugned.[94]

In essence the legality–merits distinction is designed to stop the courts from ‘standing in the shoes’ of the primary decision-maker, as a merits review body does.[95] The courts’ role is not to be a surrogate administration: ‘The primary object of external judicial scrutiny ... is not in “governing”, but in formulating the conditions which are to be satisfied, as necessary rather than sufficient, if governmental decisions are to be acceptable.’[96] The frequent complaint is that the courts have been unclear in their articulation of which conditions must be satisfied in order for governmental decisions to be acceptable — about which aspects of decision-making go to legality, and which go to merits. It is possible, however, to use the concept of rationality to explicate the legality–merits distinction, and to make more sense out of what the courts have been doing for the past half century or so.

This is precisely because the rationality standard is based on a conceptual distinction between the rationality and the merits of a decision. The rationality standard could not have emerged if the courts had not accepted the truth of the proposition, quoted above, that ‘[n]ot only is it human to err: it may even be rational to do so, if all the evidence happens to point in the wrong direction.’[97] Thus, Lord Hailsham said, in a slightly different context in Re W (an infant), that ‘not every reasonable exercise of judgment is right, and not every mistaken exercise of judgment is unreasonable’.[98] (The relationship between rationality and reasonableness is clarified below.) Wade and Forsyth frame the legality–merits distinction in a way that clearly emphasises the rationality standard’s importance:

[T]he court must not usurp the discretion of the public authority which Parliament appointed to take the decision. Within the bounds of legal reasonableness is the area in which the deciding authority has genuinely free discretion. If it passes those bounds, it acts ultra vires. ... But if the decision is within the confines of reasonableness, it is no part of the court’s function to look further into its merits.[99]

As Galligan explains, ‘the canons of rational action constitute constraints on discretionary decisions, but they are in the nature of threshold constraints above which there remains room for official judgment and choice both as to substantive and procedural matters’.[100]

However, the exact relationship between legality and rationality remains unclear. It is often said that the courts’ concept of irrationality is an ‘accordion term’, encompassing a broad range of intensity of review:

The legality/merits distinction reflects the threshold level of reasonableness that a decision must possess to withstand judicial criticism. This threshold can be set anywhere between a low level, where only decisions which are so lacking in reason as to be arbitrary are quashed, to the high level expressed in the ‘hard look’ doctrine of American law.[101]

In practice, Australian courts set the rationality standard somewhere between these two extremes, without being consistent or clear about what they are doing or why they are doing it. Space precludes a comprehensive discussion of how the rationality standard operates in Australian judicial review today, but the relationship between legality and rationality is further explored in Part IV below.

IV THE RELATIONSHIP BETWEEN RATIONALITY AND LEGALITY

A Objective or Subjective Rationality?

The most important issue requiring clarification is whether illegality is limited to what might be called ‘objective irrationality’, or whether it extends to what might be called ‘subjective irrationality’. Just as there is a distinction between the rationality and the merits of a decision, so too is there a distinction between a decision itself and the reasons for that decision. Not every decision which can be rationally justified is rationally justified by the decision-maker’s actual reasons for making it. A decision that is objectively rational may have been made for irrational reasons. The question is: does legality require only that the decision is able to be rationally justified (objective rationality), or does it require that it be rationally justified by the decision-maker’s actual reasons (subjective rationality)? In practice, of course, an objective test is easier to satisfy than a subjective test, because very often good reasons could be given for a decision, even though the decision-maker actually had bad reasons.

The traditional formulation of unreasonableness — that a decision must not be so unreasonable that no reasonable authority could ever have come to it[102] — seems to be an objective test. On the other hand, the formulation can be read to mean ‘so unreasonable that no reasonable decision maker could have made it for those reasons’, and so incorporate a subjective test.

Indeed, some aspects of rationality review, such as the doctrine of relevant and irrelevant considerations, are clearly directed to subjective rationality. The legality–merits distinction provides no guidance on which standard should be applied, because both standards respect the difference between the rationality and the merits of a decision. The issue of whether the rationality standard involves an objective or subjective test is most acute in relation to judicial review of factual questions.

1 Judicial Review of Factual Questions

The emergence of the rationality standard has had a dramatic impact on judicial review of factual findings.[103] We are in the midst of a fundamental doctrinal shift. Previously, factual questions were classified as either jurisdictional or non-jurisdictional. If jurisdictional, they were subject to judicial review for correctness. If non-jurisdictional, they were immune from review.[104] Now, non-jurisdictional factual findings can be reviewed for rationality, without undermining the legality–merits distinction. The legality–merits distinction means that there is no error of law in merely making a wrong finding of fact:[105] a factual finding’s empirical correctness goes to the decision’s merits. But rationality is distinct from correctness. Under the rationality standard of review, ‘what is important ... is how the decision maker established [the fact’s] existence ... rather than whether the reviewing court agrees with the end result.’[106]

Thus, it is clearly established that a finding of fact based on ‘no evidence’ is an error of law.[107] Both Mason CJ and Deane J unequivocally accepted this principle in Bond.[108] While the courts have not clearly articulated the ‘no evidence’ rule’s underlying rationale, it is clearly justified by even the most minimum requirement of rational decision-making.[109] Thus, the principle has often been expressed as requiring some ‘probative’,[110] ‘logically probative’[111] or ‘rationally probative’[112] evidence. However the requirement is expressed, the underlying idea is the same: to count as ‘evidence’ for the purpose of the rule, the material before the decision-maker must be rationally relevant to the finding made. As Hayne J put it in Shulver v Sherry, ‘the inquiry is one about whether the material upon which the decision-maker has acted was material that, as a matter of logic or reason, supported the finding made’.[113]

Recognising the need for rational decision-making, the courts have also used the relevant considerations[114] and unreasonableness[115] grounds of review to develop a limited duty to inquire. Again, the connection between a duty to inquire and the requirements of rational fact-finding is clear. As discussed above, rationality requires the ‘optimal’ amount of evidence to be gathered. This does not mean the maximum amount possible — decision-makers are not required to turn over every stone.[116] However, the evidence gathered will be less than optimal if obviously relevant and readily accessible material is ignored. Thus, a general duty to inquire should be seen for what it is: the explicit articulation and development of a common law standard of rational decision-making.

There are also cases holding that, even if the decision-maker has collected sufficient evidence to make a rational finding, they will fall into legal error if their finding flies in the face of overwhelming evidence to the contrary. For example, Wilcox J said in Taveli v Minister for Immigration, Local Government and Ethnic Affairs that if ‘all of the factors germane to a particular decision point in one direction ... it would seem proper to brand as unreasonable a decision to the contrary effect.’[117] In Bond Mason CJ seemed to suggest that the High Court has at least not so far accepted that courts can set aside factual findings because they fly in the face of the weight of the evidence.[118] On the other hand, in Eshetu, Gummow J said:

[W]here the criterion of which the authority is required to be satisfied turns upon factual matters upon which reasonable minds could reasonably differ, it will be very difficult to show that no reasonable decision-maker could have arrived at the decision in question. It may be otherwise if the evidence ... was all one way.[119]

The principles discussed above were all concerned with the objective evidentiary basis for factual findings, that is, whether a reasonable person, looking at the evidence before the decision-maker, could have made the finding. But there is another line of cases that goes further, and applies a subjective rationality test to scrutinise how the decision-maker thought about that evidence, that is, their reasons for making their primary findings and drawing their inferences. These cases hold that, even if the decision-maker collects the optimal amount of evidence, and their findings do not fly in the face of the weight of evidence, they will still fall into legal error if they think about the evidence in an irrational way. As Lord Diplock put it in Mahon v Air New Zealand, ‘the reasoning supportive of the finding, if it be disclosed, [must not be] ... logically self-contradictory’[120] or ‘otherwise based upon an evident logical fallacy.’[121] Thus, the House of Lords has rephrased the test for review of factual findings in terms of ‘perversity’[122] instead of ‘no evidence’, an approach which clearly encompasses both objective and subjective elements.[123]

As Lord Diplock said, a subjective rationality test can only be applied if the reasons for the decision are disclosed. Indeed, the subjective rationality test’s development is clearly related to the growing expectation that decision-makers will state their reasons. In Azzopardi v Tasman UEB Industries Ltd Kirby P, dissenting, said:

[W]here, because of the development of the obligation of reasoned decision-making, the [decision maker] ... exposes his reasons and these reasons demonstrate manifest error or illogicality in the reasoning process ... an error in point of law will be established.[124]

The Federal Court generally uses the unreasonableness and relevant considerations grounds to set aside decisions because of subjective irrationality in the decision-maker’s reasons for making factual findings. Under the relevant considerations doctrine, a person who is required to have regard to a particular matter, including a factual matter, is required to give it ‘proper, genuine and realistic’ consideration.[125] This requires an ‘active intellectual process’.[126] These are just different ways of expressing the requirement that the decision-maker rationally consider the matter. Thus, when a decision-maker is required to consider a factual matter, they make a legal error both when they ignore it completely, and when they consider the evidence relevant to it irrationally: rational consideration of factual matters requires rational consideration of the relevant evidence.[127]

The requirement that evidence be rationally considered has been most explicitly articulated in Epeabaka v Minister for Immigration and Multicultural Affairs.[128] Finkelstein J reasoned that a legal duty to rationally consider evidence is the natural corollary of the ‘no evidence’ rule: ‘There would be little point to the imposition of an obligation upon a tribunal to decide a case on probative evidence if there was not an additional obligation to rationally consider that evidence.’[129] In that case Finkelstein J considered whether Mason CJ’s judgment in Bond precluded the application of a subjective rationality approach.[130] Mason CJ had quoted with approval Menzies J’s views in R v The District Court; Ex parte White: ‘Even if the reasoning ... were demonstrably unsound, this would not amount to an error of law ... To establish some faulty (eg, illogical) inference of fact would not disclose an error of law.’[131] Mason CJ continued:

Thus, at common law, according to the Australian authorities, want of logic is not synonymous with error of law. So long as there is some basis for an inference — in other words, the particular inference is reasonably open — even if that inference appears to have been drawn as a result of illogical reasoning ... no error of law has taken place.[132]

Mason CJ thus seemed to reject the subjective rationality approach in relation to factual findings. On the other hand, he included Lord Diplock’s subjective rationality requirement[133] in the class of principles only not so far accepted by the High Court.[134] He also referred to Singh v Minister for Immigration and Ethnic Affairs[135] and Minister for Immigration, Local Government and Ethnic Affairs v Pashmforoosh,[136] both of which applied a subjective rationality test, as ‘unobjectionable’.[137]

Finkelstein J therefore concluded that Bond did not preclude the application of subjective rationality requirements.[138] Similarly, in Shulver v Sherry Hayne J said that Bond established the existence of the ‘no evidence’ rule, but left open whether findings must be objectively open on the evidence, and whether the actual reasons for the finding must themselves be rational.[139] Certainly, in practice Federal Court judges continued to apply subjective rationality requirements to factual findings after Bond.[140] However, on appeal in Epeabaka v Minister for Immigration and Multicultural Affairs,[141] the Full Federal Court held that, for the time being at least, Mason CJ’s judgment in Bond precludes the Federal Court from taking this approach.[142]

However, as Aronson and Dyer write, in Bond Mason CJ was expressing ‘restrictive sentiments against a full tide’.[143] In Bruce v Cole[144] and Hill v Green,[145] in order to get around Bond, Spigelman CJ rephrased the requirement in terms of implied legislative intent rather than a common law rule:

[W]here a statute or regulation makes provision for an administrative decision in terminology which does not confer an unfettered discretion on the decision-maker, the courts should approach the construction of the statute ... with a

presumption that the parliament or the author of the regulation intended the decision-maker to reach a decision by a process of logical reasoning and a contrary interpretation would require clear and unambiguous words.[146]

In practice, a number of High Court judges in Abebe and Eshetu applied a logical reasoning test to the decisions under review in those cases. As noted above, in Abebe Gleeson CJ and McHugh J asked whether the Refugee Review Tribunal had followed the ‘logical’ process of reasoning that it was legally ‘bound’ to follow.[147] In Eshetu Gummow J emphasised that the Refugee Review Tribunal’s obligation to give reasons ‘furthers the objectives of reasoned decision-making and the strengthening of public confidence’, but ‘does not provide the foundation for a merits review of [its] fact-finding processes’.[148] However, he went on to hold that, where a decision-maker has to be satisfied of a particular fact, their reasons for being so satisfied are subject to review for ‘reasonableness’, which ‘would permit review in cases where the satisfaction of the decision-maker was based on findings or inferences of fact which were not supported by some probative material or logical grounds’.[149] Gaudron and Kirby JJ similarly stated that ‘an unreasonable decision is one for which no logical basis can be discerned.’[150] Kirby J has also recently reiterated the view he expressed in Azzopardi v Tasman UEB Industries Ltd that illogical factual findings are errors of law.[151]

While the exact scope of review of factual findings for unreasonableness still remains unclear, these references in Abebe and Eshetu to requirements of ‘logic’, which were articulated by a majority of current High Court judges, strongly suggest that the High Court has moved closer to accepting that illogical reasoning from evidence to conclusion is a reviewable error. In Hill v Green Fitzgerald JA (Beazley JA agreeing) certainly read these references in this way.[152] It is probably only a matter of time before the High Court follows the lead of courts in the United Kingdom, United States and Canada, not to mention the Federal Court, and explicitly applies a subjective rationality test to factual findings.

In my view this would be a salutary development. As Finkelstein J pointed out, subjective rationality requirements are the natural corollary of the ‘no evidence’ rule. There is little point in requiring decision-makers to decide on the basis of rationally probative evidence unless they are also required rationally to consider that evidence. Furthermore, a subjective rationality requirement is consistent with basic theories of rational belief. As discussed above, a factual finding is rational when it has ‘the right kind of relation to the evidence available’.[153] This right kind of relationship between evidence and belief only exists when the evidence has rationally been considered. Regardless of how much the evidence objectively supports the finding, the right kind of relationship does not exist when the decision-maker guesses, ignores inconvenient evidence, or reasons irrationally about it. It sends the wrong message and is functionally incoherent for the courts to say to administrators that they are not legally required to think rationally about the evidence that is before them. That is simply an invitation to arbitrary and capricious decision-making. A subjective rationality requirement is really just a requirement that administrators do their job properly. It is not an invitation to judges to review the correctness of their factual findings.

B The Relationship between Means and Ends

The second issue concerning the relationship between legality and rationality that I wish to discuss is the extent to which rationality review justifies review of the relationship between means and ends. This involves two sub-issues. First, it raises the issue of the courts’ role in reviewing administrative agencies’ choice of ends. As discussed above, it is frequently asserted that rationality is purely a question of the means chosen to a given end, not a question of the choice of end itself. On the other hand, most ends are themselves means to other more long-term ends, and it is also widely accepted that rationality requires consistency in choices between alternative or competing ends. Thus, in general terms there would seem to be room in a rationality standard for review of ends as well as means. But in the administrative law context there is another, more important, reason why rationality review is legitimately concerned with choice of ends as well as means. The view that rationality is concerned with means, not ends, is based on liberal values of individual freedom and autonomy — each individual should be free to pursue his or her own ends. However, such values have no role to play in respect of administrative agencies. Such agencies have no ends of their own, but are established to pursue the ‘public interest’, as defined by the people’s representatives who have established them and usually fund them.[154] As the United States Supreme Court has said, ‘[n]o rational system of regulation can permit its administrators to make policy judgments without explaining how their decisions effectuate the purposes of the governing law.’[155]

Ideally, parliamentary oversight would ensure administrative fidelity to public purposes, but there are many practical reasons why Parliament cannot provide an adequate level of scrutiny. The courts, therefore, have a role to play in ensuring that agencies’ goals do not become subverted in some way, particularly through agency capture.[156] Thus, under the proper purposes doctrine, the courts have assumed the role of ensuring that administrative agencies remain faithful to the purposes for which they are established and empowered.[157] As Dixon J said:

[I]t is incumbent upon the public authority in whom the discretion is vested not only to enter upon the consideration of applications for its exercise but to decide them bona fide and not with a view of achieving ends or objects outside the purpose for which the discretion is conferred.[158]

Of course, this role is not without controversy, because it assumes that the courts are better able to identify what constitutes that purpose than the administrative agencies themselves: ‘There is, however, no doubt that the denomination of a purpose as proper or improper ... raises issues of political and social choice which do not cease to be so by being expressed in the language of vires[159] — or rationality.

The question is: what degree of critical scrutiny is required? The courts generally treat the choice of ends as a question of law, and therefore apply a correctness standard: ‘The determination of a statutory purpose is a hard-edged decision, a matter of law for the court.’[160] However, Parliament may have specified no more than a vague, ultimate end, leaving the agency with wide discretion in terms of both its more immediate ends and the means to be employed to achieve them. This is especially the case under modern public management theory, with its emphasis on ‘letting the managers manage’. There is no clear dividing line between the ends that raise questions of legality and the ends that raise questions of the merits. Arguably, it is therefore inappropriate for the courts to substitute their own opinions of the purposes for which the power has been granted. But the courts can still play a useful role by providing a mechanism of critical scrutiny, if they insist that the agency’s choice of ends be able to be rationally justified, in terms of the agency’s overall purposes. In other words, the rationality standard offers an alternative, less intrusive mechanism of review than the current correctness standard; one that can still ensure that power is not being abused for completely extraneous reasons.

The second, related sub-issue concerns the courts’ role in reviewing administrative agencies’ choices of means to achieve their ends. In the ideal model of rational action, the rational actor chooses the best means to their end. In a more bounded model, the rational actor chooses the means that they rationally believe to be the best means to their end. By adopting an ideal model, the courts could use a rationality standard to justify what would essentially be a correctness standard, insisting that an agency adopt the means that the court considers to be ‘objectively’ best. The United States Supreme Court’s ‘hard look’ doctrine has, at times, approached this level of scrutiny.[161]

However, the legality–merits distinction can only be respected if the rationality standard of review is based on a bounded model of rationality. Thus, where more than one means could rationally be believed to be best, Australian courts generally treat the choice of means as going to the decision’s merits. The question is: do the courts have a role to play in excluding some means from the range of allowable choice and, if so, in what circumstances? There are two possible standards that can be applied here. The narrowest approach is for the court to exclude only those means that cannot rationally be believed to be a means to the relevant end at all. A broader approach is for the courts to exclude those means that cannot rationally be believed to be the best means to the relevant end. This issue has clearly arisen in relation to the role of ‘reasonable proportionality’ in judicial review of delegated legislation.

1 Rationality, Proportionality and Delegated Legislation

The modern unreasonableness test[162] for judicial review of delegated legislation is obviously based on a means–end model of rational action. Thus, in Clements v Bull Williams ACJ and Kitto J said that delegated legislation ‘may ... be held invalid on the ground that no reasonable mind could justify it by reference to the purposes of the power’.[163] In Williams v Melbourne Corporation Dixon J stated that delegated legislation is invalid if ‘it could not reasonably have been adopted as a means of attaining the ends of the power.’[164] In Stenhouse v Coleman Starke J said that delegated legislation had to be reasonably ‘appropriate and adapted’ to achieving its purpose.[165] The courts’ role under this test is, therefore, to limit an agency’s range of choice to those means that can rationally be chosen.

Within those limits, the courts are not concerned with the merits of the agency’s choice, that is, whether it is ‘unfair’,[166] ‘inexpedient’ or ‘misguided’.[167] However, the courts distinguish between delegated legislation that is ‘merely’ unfair, inexpedient or misguided, and legislation that is outside the range of rational choice because of its disproportionate effect. Both Slattery v Naylor[168] and Kruse v Johnson[169] held that delegated legislation is invalid if it is so ‘oppressive’, ‘capricious’ or ‘gratuitous’ in its interference with individual rights that it cannot be reasonably justified. These statements have been frequently echoed in subsequent judgments, including the judgment of Dixon J in Williams v Melbourne Corporation.[170] The High Court first put the test in ‘proportionality’ terms in South Australia v Tanner: the question being ‘whether the regulation is capable of being considered to be reasonably proportionate to the pursuit of the enabling purpose’,[171] but it also said that this is the same as Dixon J’s test in Williams v Melbourne Corporation. Thus, while the language of proportionality may be fairly new, the underlying concern with the relationship between means and ends is not.

In the lower courts the connection between rationality and proportionality has been clearly seen. In Minister for Resources v Dover Fisheries Pty Ltd Cooper J explained: ‘It is the disproportionate operation of the delegation which denies to it a place in the range of alternative modes of implementation available to an objective reasonable mind.’[172] In New South Wales v Macquarie Bank Ltd Kirby P said: ‘There must exist a reasonable relationship between the end and the means of the law.’[173] This means not only that ‘[t]he means must be reasonably likely to bring about the apparent objective of the law’, but also that ‘[t]he detriment to those adversely affected must not be disproportionate to the benefit to the public envisaged by the legislation’.[174] A law is disproportionate where it has a ‘gratuitous’ impact on a person, either because the impact is unnecessary (alternative means that avoid that impact being available) or because, as Kirby P said, the negative individual impact very clearly outweighs the public benefit flowing from the relevant purpose being achieved.

Thus, proportionality is not a separate test of validity,[175] but is part of a wider rationality test concerned with the relationship between means and ends. Furthermore, that test excludes not just those means that cannot rationally be thought to be a means to the end at all, but also those means that cannot rationally be thought to be best, at least because of their gratuitous or disproportionate effect on individuals.[176]

C ‘Rationality’ and ‘Reasonableness’

The third issue concerning the relationship between legality and rationality that I wish to discuss is the relationship between ‘rationality’ and the more commonly used term ‘reasonableness’. ‘Reasonableness’ has a long history in judicial review, as elsewhere in the law, whereas the term ‘rationality’ has only recently come into use. However, the traditional formulation of ‘Wednesbury unreasonableness’ (‘a conclusion so unreasonable that no reasonable [decision-maker] ... could ever have come to it’)[177] is circular and therefore unhelpful. It also tends to gloss over the different ways that different kinds of administrative actions can be unreasonable. For example, factual findings and delegated legislation are likely to be unreasonable for quite different reasons, but the traditional unreasonableness formulation gives no clue as to these differences. ‘Rationality’ has the benefit of making it possible to give the legal standard more content and differentiation. Drawing on the models of rational action and rational belief discussed above, we can see that factual findings are likely to be unreasonable or irrational because of the lack of the ‘right relationship’ between evidence and conclusions, while delegated legislation is likely to be unreasonable because of the lack of an ‘adaptive relationship’ between means and ends.

Nevertheless, some writers suggest that ‘reasonableness’ is in fact the preferable concept. Walker, for example, argues that three things are wrong with the term ‘rationality’.[178] First, it ‘wrongly suggests that the test is logicality’, even though not all departures from logic legally invalidate a decision.[179] Secondly, ‘the label ignores the important role of morals’: a decision may be rational but unreasonable ‘because it is insufficiently consistent with accepted moral values’.[180] Thirdly, ‘the label may be unnecessarily abusive’, conjuring up pictures of decision-makers ‘foaming at the mouth, with civil servants rushing to replace the straitjacket’.[181] Courts may be reluctant to use the concept ‘because they are inhibited by the connotations which the word “irrationality” carries with it’.[182]

The first and third objections may be shortly dealt with. As discussed above,[183] rationality is conceptually distinct from logicality. Logic requires a conclusion to be compelled by given premises (irrespective of the premises’ rationality), whereas rationality requires a conclusion to be consistent with rational premises. Properly understood, rationality does not suggest that the test is logicality. As for the term’s negative connotations, it would be a mistake to think that decision-makers are any more comfortable with being labelled ‘so unreasonable that no reasonable decision maker could have decided’ as they did, or that the courts are any happier applying such a label. Furthermore, this is hardly the only example of a term’s legal meaning differing from its connotations in ordinary language. If rationality provides a conceptually sound basis for judicial review’s development, then the courts can surely be expected to employ the concept.

Walker’s second objection however, raises the important question of the conceptual relationship between reasonableness and rationality. Both terms derive from the same linguistic source, but they are not completely interchangeable.[184] Reasonableness is often used as a synonym for rationality, that is, as an ‘umbrella’ term to refer to a model of rational action and all that it requires. But there is another sense of reasonableness which is quite distinct. Walker correctly points out that individuals might act rationally, in the sense of having adopted the best means to their end, but still unreasonably. This occurs where their rational action ‘is insufficiently consistent with accepted moral values’,[185] or is inconsistent with ‘common sense’,[186] or pays insufficient regard to the interests of others.[187]

The two senses of reasonableness, one synonymous with rationality and the other not, make the term ambiguous. This ambiguity tends to get carried over into rationality as well. Thus, it is unclear whether Lord Diplock, when he used ‘rationality’ as a synonym for ‘reasonableness’, intended to use it in the first (umbrella) or second (narrower) sense. The other two terms in his classification, ‘lawfulness’ and ‘procedure’, were clearly both used in an umbrella sense, to refer to a myriad set of more specific grounds of judicial review. This suggests that rationality was similarly used in its umbrella sense. On the other hand, his definition of rationality — acting in defiance ‘of accepted moral standards’[188] — clearly evokes the second sense of reasonableness.

However, this ambiguity does not mean that rationality is not a useful label, and even less that reasonableness, the source of the ambiguity, is to be preferred. Rather, it means that semantic precision is required. In the administrative law context, ‘rationality’ should be seen as an umbrella term, encompassing many specific requirements of good decision-making. Many of these tend to be lumped under the ‘reasonableness’ rubric, but they should be seen for what they are — different aspects of rationality. ‘Reasonableness’ should not be used as a synonym for ‘rationality’, but only in the residual sense of requiring consistency with accepted moral values and common sense, and paying due regard to the interests of others (where this is not otherwise required by more specific doctrines such as procedural fairness, relevant considerations and reasonable proportionality). As long as this semantic clarity is maintained, no harm can come from including reasonableness under the rationality rubric.

V CONCLUSION

Half consciously, the courts have been developing a standard of review based on emerging understandings of the nature of rational decision-making and rational administration. The rationality standard’s emergence was a fundamental doctrinal shift that has been central to the development of judicial review in the second half of the 20th century. Rationality has been able to play this role because it respects the fundamental requirement that it is for administrative decision-makers to decide the merits of the case. Providing a halfway house between review for correctness and immunity from review, rationality has underpinned the shift from jurisdiction to legality.

This is not to say that questions of the existence of power have become irrelevant. Clearly, they have not. But the courts have come to be concerned not just with whether power exists, but also with why it was exercised in a particular way. The tendency has been to dress this concern up in jurisdictional terms, by implying a legislative intent that the scope of the power be limited to instances where it was exercised rationally. A competing theory places rationality requirements on a common law basis under the procedural fairness rubric. Neither approach is correct. Rationality should be seen as a separate common law requirement in its own right, as procedural fairness recently has been. The fiction of implied legislative intent is an inadequate explanation of the courts’ actual practice. Furthermore, rationality should be a universal legal requirement of good administrative decision-making, whereas procedural fairness is not. Recognising the rationality standard’s common law basis also makes it available for review of the exercise of non-statutory powers. This is increasingly important, given the development of new techniques of governance that depend less on express statutory authorisation. However one conceives of the source of a legal requirement to act rationally, once the rationality standard’s emergence is recognised, then the question is: what, exactly, does it require?

In Part IV, I identified two key, general questions. First: is the rationality standard a subjective one, concerned with the decision-maker’s actual reasons for the decision, or an objective one, concerned only with whether good reasons could have been given? This issue is clearest in relation to judicial review of factual findings. It was argued that while the Australian cases are currently inconsistent and unclear on this point, in principle it is appropriate for the courts to apply a subjective rationality requirement here, as they do in relation to exercises of discretion.

The second general question raised was: to what extent does rationality review justify review of the relationship between means and ends? This involved two sub-issues: first, the courts’ role in reviewing administrative agencies’ choice of ends and, second, the courts’ role in reviewing administrative agencies’ choice of means. In relation to the choice of ends, I argued that, under the proper purposes doctrine, courts clearly have a role to play in ensuring that administrative agencies remain faithful to the purposes for which they were established. In relation to the choice of means (an issue that most clearly arises in relation to the role of reasonable proportionality in judicial review of delegated legislation) I argued that the rationality standard excludes not just those means that cannot rationally be thought to be a means to the end at all, but also those means that cannot rationally be thought to be best, because of their gratuitous or disproportionate effect on individuals.

In Part IV, I also sought to clarify the relationship between the general rationality standard and the more familiar administrative law concept of ‘unreasonableness’. I argued that ‘reasonableness’ should not be used as a synonym for rationality, but only in the residual sense of requiring consistency with accepted moral values and common sense, and paying due regard to the interests of others. ‘Rationality’ should be seen as an umbrella term, incorporating all the elements of a model of rational action based on rational beliefs, and therefore encompassing many specific requirements of good decision-making.

Issues raised in this article, but not fully explored, include: how to distinguish between jurisdictional and non-jurisdictional facts; the role, if any, of rationality review in the review of administrative interpretations of statutes, and the implications of this for the distinction between jurisdictional and non-jurisdictional errors of law; and consideration of when a rationality (proportionality) test might apply to delegated legislation. These are some of the key issues to be addressed in the future development of judicial review of administrative action for rationality.


[*] BA, LLB (Hons) (Qld), PhD (Griffith); Senior Lecturer, Faculty of Law, Griffith University. Thanks are due to Graeme Orr, Alex Guild and two anonymous referees for comments provided on an earlier draft of this paper.

[1] R v Secretary of State for the Environment; Ex parte Nottinghamshire County Council [1985] UKHL 8; [1986] AC 240, 249 (Lord Scarman).

[2] Council of Civil Service Unions v Minister for the Civil Service [1985] AC 374, 410 (‘CCSU’).

[3] (1992) 28 ALD 829, 861, citing Othman v Minister for Immigration and Ethnic Affairs (1991) 24 ALD 707, 711.

[4] CCSU [1985] AC 374, 410.

[5] [1947] EWCA Civ 1; [1948] 1 KB 223, 223–4 (‘Wednesbury’).

[6] CCSU [1985] AC 374, 410.

[7] Indeed, Lord Greene MR saw a family resemblance between unreasonableness and other grounds of judicial review: Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 229.

[8] Rogers Brubaker, The Limits of Rationality: An Essay on the Social and Moral Thought of Max Weber (1984) 30.

[9] Sir John Laws, ‘Illegality: The Problem of Jurisdiction’ in Michael Supperstone and James Goudie (eds), Judicial Review (2nd ed, 1997) [4.1], [4.3] (emphasis in original). See also Lord Irvine, ‘Judges and Decision-Makers: The Theory and Practice of Wednesbury Review’ [1996] Public Law 59.

[10] John Elster, ‘Introduction’ in John Elster (ed), Rational Choice (1986) 1, 1; John Elster, Nuts and Bolts for the Social Sciences (1989) 22–3.

[11] John Harsanyi, ‘Advances in Understanding Rational Behaviour’ in John Elster (ed), Rational Choice (1986) 82, 82–3 (emphasis in original).

[12] The action must have been performed because the actor actually decided that it was the rational thing to do:

[I]t is not sufficient to show that the action was [objectively] rational, since people are sometimes led by accident or coincidence to do what is in fact best ... We must show ... that the action arose in the proper way, through a proper kind of [subjective] connection to desires, beliefs and evidence.

Elster, ‘Introduction’, above n 10, 2.

[13] See below Part II(B).

[14] This is an assumption shared by some (but by no means all) social scientists: see, eg, Herbert Simon, Reason in Human Affairs (1983) 7, 106.

[15] As Stewart puts it, economists generally assume that ‘economic rationality is an attribute of means alone, while the ends or goals of economic agents are inaccessible to reason. Any goal is as valid as any other ... and an agent’s rationality is assessed only with reference to his or her effectiveness in achieving the goal’: Hamish Stewart, ‘A Critique of Instrumental Reason in Economics’ (1995) 11 Economics and Philosophy 57, 58. In this article Stewart argues that economists must both explore how preferences are formed and be prepared to evaluate their comparative worth: at 78–80. See also James March, ‘Bounded Rationality, Ambiguity, and the Engineering of Choice’ in John Elster (ed), Rational Choice (1986) 142, 152–8 for a discussion of the rationality of ‘goal ambiguity’.

[16] Q Gibson, ‘Arguing from Rationality’ in S I Benn and G W Mortimore (eds), Rationality and the Social Sciences: Contributions to the Philosophy and Methodology of the Social Sciences (1976) 111, 120.

[17] Harsanyi, above n 11, 85.

[18] March, above n 15, 146.

[19] Elster, ‘Introduction’, above n 10, 4.

[20] Ibid 1. See also Elster, Nuts and Bolts, above n 10, 30.

[21] Elster, ‘Introduction’, above n 10, 1. See also Elster, Nuts and Bolts, above n 10, 30.

[22] Elster, ‘Introduction’, above n 10, 1.

[23] Ibid 19. See also Elster, Nuts and Bolts, above n 10, 25.

[24] Elster, ‘Introduction’, above n 10, 21–2. See also R S Peters, ‘The Development of Reason’ in S I Benn and G W Mortimore (eds), Rationality and the Social Sciences: Contributions to the Philosophy and Methodology of the Social Sciences (1976) 299, 306–7.

[25] Elster, Nuts and Bolts, above n 10, 25. See also G W Mortimore and J B Maund, ‘Rationality in Belief’ in S I Benn and G W Mortimore (eds), Rationality and the Social Sciences: Contributions to the Philosophy and Methodology of the Social Sciences (1976) 11, 15–17.

[26] Elster, ‘Introduction’, above n 10, 21–2; see also Mortimore and Maund, above n 25, 14.

[27] Ibid 13.

[28] Ibid.

[29] Ibid 13–14.

[30] Elster, Nuts and Bolts, above n 10, 25.

[31] S I Benn and G W Mortimore, ‘Introduction’ in S I Benn and G W Mortimore (eds), Rationality and the Social Sciences: Contributions to the Philosophy and Methodology of the Social Sciences (1976) 1, 2.

[32] J St B T Evans, ‘Bias and Rationality’ in K I Mankeltow and D E Over (eds), Rationality: Psychological and Philosophical Perspectives (1993) 6, 11–12.

[33] Ibid 12.

[34] R J Stevenson, ‘Rationality and Reality’ in K I Mankeltow and D E Over (eds), Rationality: Psychological and Philosophical Perspectives (1993) 61, 65.

[35] Evans, ‘Bias and Rationality’, above n 32, 13–14.

[36] Ibid.

[37] Ibid 14.

[38] Ibid (citation omitted). See also Stevenson, above n 34, especially 76–7.

[39] J M Evans, de Smith’s Judicial Review of Administrative Action (4th ed, 1980) 590.

[40] Prohibition was developed to keep ecclesiastical courts from dealing with secular matters: ibid; Edward Jenks, ‘The Prerogative Writs in English Law’ (1923) 32 Yale Law Journal 523, 528; J H Baker, An Introduction to English Legal History (3rd ed, 1990) 166.

[41] Certiorari was originally a command requiring any type of documentary record to be produced to the Crown, used initially to obtain information needed for administrative purposes: Evans, de Smith’s Judicial Review, above n 39, 588–9.

[42] Ibid 589.

[43] (Unreported, King’s Bench, Bramston CJ, Mallett and Heath JJ, 1642). An unpublished report of the case is reproduced in Edith Henderson, Foundations of English Administrative Law: Certiorari and Mandamus in the Seventeenth Century (1963) 182–6.

[44] (1701) 1 Salk 145; 91 ER 134.

[45] Henderson, above n 43, 83–116.

[46] See Groenwelt v Burwell (1701) 1 Salk 145, 145; 91 ER 134, 134.

[47] Henderson, above n 43, 143–60.

[48] This simple schema was complicated when certiorari for non-jurisdictional error of law on the face of the record was revived in R v Northumberland Compensation Appeal Tribunal; Ex parte Shaw [1951] EWCA Civ 1; [1952] 1 KB 338.

[49] See P P Craig, Administrative Law (3rd ed, 1994) 348 and sources cited there for the United Kingdom position. The Australian position is very similar. A great many Australian cases demonstrate the difficulties with the fact–law distinction. Eg, in Federal Commissioner of Taxation v Broken Hill South Ltd (1941) 65 CLR 151 the question was whether the activities involved in caretaking a mothballed mine while waiting for commodity prices to improve were ‘mining operations’ within the meaning of the Income Tax Assessment Act 1936 (Cth). The case came before four judges of the High Court. Starke J (at 155–6) and McTiernan J (at 159) treated the question as one of fact, while Rich ACJ (at 154) and Williams J (at 160) treated it as one of law. For one of the best recent Australian discussions of the fact–law distinction in the context of statutory interpretation, see S v Crimes Compensation Tribunal [1998] 1 VR 83, 86–8 (Phillips JA).

[50] Jack Beatson, ‘The Scope of Judicial Review for Error of Law’ (1984) 4 Oxford Journal of Legal Studies 22, 24–7.

[51] Ian Yeats, ‘Findings of Fact: The Role of the Courts’ in Geneva Richardson and Hazel Genn (eds), Administrative Law and Government Action: The Courts and Alternative Mechanisms of Review (1994) 131, 133. While this and the previous reference focus on the UK position, the Australian position was essentially the same.

[52] Ibid 131, 133.

[53] Ibid 133.

[54] Ibid.

[55] (1986) 162 CLR 24, 40–1.

[56] This was R v Bishop of London which is, comparatively, not that much older: (1890) 24 QB 213.

[57] Rooke’s Case [1572] EngR 141; (1598) 5 Co Rep 99b; 77 ER 209; Keighley’s Case [1572] EngR 219; (1609) 10 Co Rep 139a; 77 ER 1136.

[58] R v Askew (1768) 4 Burr 2186, 2188; [1768] EngR 45; 98 ER 139, 141 (Lord Mansfield); R v Wilkes [1770] EngR 34; (1770) 4 Burr 2527, 2539; [1770] EngR 34; 98 ER 327, 334 (Lord Mansfield).

[59] Slattery v Naylor [1888] AC 446, 451–2 (Lord Hobhouse); Sharp v Wakefield [1891] UKLawRpAC 8; [1891] AC 173; Kruse v Johnson [1898] UKLawRpKQB 101; [1898] 2 QB 91.

[60] Williams v Giddy [1911] AC 381; Short v Poole Corporation [1926] Ch 66.

[61] Municipal Council of Sydney v Campbell [1925] AC 339; Council of the Shire of Werribee v Kerr [1928] HCA 41; (1928) 42 CLR 1; Swan Hill Corporation v Bradbury [1937] HCA 15; (1937) 56 CLR 746, 757–8 (Dixon J); Arthur Yates & Co Pty Ltd v Vegetable Seeds Committee [1945] HCA 55; (1942) 72 CLR 37.

[62] Folkestone Corporation v Brockman [1914] UKLawRpAC 6; [1914] AC 338, 367 (Lord Atkinson).

[63] Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 229 (citation omitted).

[64] D J Galligan, Discretionary Powers: A Legal Study of Official Discretion (1986) 143; Mark Aronson and Bruce Dyer, Judicial Review of Administrative Action (1996) 290; Susan Kneebone, ‘A Commentary on Proportionality: Protection of Common Law Rights or “Chipping Away the Diceyan Edifice”’ in Linda Pearson (ed), Administrative Law: Setting the Pace or Being Left Behind? (1996) 168, 179.

[65] In light of the emergence of the rationality standard, the High Court’s refusal in Public Service Board (NSW) v Osmond (1986) 159 CLR 656 to develop a common law right to reasons for decisions is all the more remarkable. How can a court sensibly insist that decision-makers have rational reasons for their decisions, but not that they disclose what their reasons are? It is the requirements of rationality, and not procedural fairness, as is commonly supposed, that provide the strongest basis for the development of a common law duty to give reasons for decisions: see generally John McMillan, ‘Recent Themes in Judicial Review of Federal Executive Action’ (1996) 24 Federal Law Review 347, 359–60. This is slowly happening in the UK (see, eg, R v Civil Service Appeal Board; Ex parte Cunningham [1991] 4 All ER 310; R v Secretary of State for the Home Department; Ex parte Doody [1993] UKHL 8; [1994] 1 AC 531; R v Secretary of State for the Home Department; Ex parte Duggan [1994] 3 All ER 277; R v City of London Corporation; Ex parte Matson [1997] 1 WLR 765; cf R v Higher Education Funding Council; Ex parte Institute of Dental Surgery [1993] EWHC Admin 5; [1994] 1 All ER 651), Canada (see, eg, Baker v Minister of Citizenship and Immigration [1999] 2 SCR 817; 174 DLR (4th) 193) and even in Australia, despite the decision in Public Service Board (NSW) v Osmond (see, eg, T v Medical Board of SA (1992) 58 SASR 382; Re Saunders [1993] 2 Qd R 335; Fleming v The Queen (1998) 197 CLR 250; A-G (NSW) v Kennedy Miller Television Pty Ltd (1998) 43 NSWLR 729; Edwards v Giudice (1999) 94 FCR 567).

[66] Herbert Simon, Administrative Behavior: A Study of Decision-Making Processes in Administrative Organisation (1st ed, 1945).

[67] See Aronson and Dyer, above n 64, 103–25.

[68] Eg, Kruse v Johnson [1898] UKLawRpKQB 101; [1898] 2 QB 91, 99–100 (Lord Russell CJ); Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 234 (Lord Greene MR); Roncarelli v Duplessis [1959] SCR 121, 140; 16 DLR (2nd) 689, 705 (Rand J); Anisminic Ltd v Foreign Compensation Commission [1969] AC 147, 195 (Lord Pearce); Kruger v Commonwealth [1997] HCA 27; (1997) 190 CLR 1, 36 (Brennan CJ); Abebe v Commonwealth [1999] HCA 14; (1999) 197 CLR 510, 544 (Gleeson CJ and McHugh J) (‘Abebe’); Minister for Immigration and Multicultural Affairs v Eshetu [1999] HCA 21; (1999) 197 CLR 611, 650 (Gummow J) (‘Eshetu’); Hill v Green [1999] NSWCA 477; (1999) 48 NSWLR 161, 174–5 (Spigelman CJ), 176–7 (Mason P), 213 (Fitzgerald JA, Beazley JA agreeing). Compare Chief Justice Spigelman’s view, expressed extra-curially in Chief Justice J J Spigelman, ‘Foundations of Administrative Law: Toward General Principles of Institutional Law’ (1999) 58 Australian Journal of Public Administration 3, 4 that ‘common law principles apply of their own force and not on the basis of the intention of parliament. ... [M]any of the grounds for judicial review are not merely propositions of statutory construction. Rather they are imposed on the exercise of public power by the common law.’ He writes that one of these principles is that ‘powers must be exercised rationally, without reference to irrelevant considerations and within bounds of reasons’: at 9.

[69] Kioa v West [1985] HCA 81; (1985) 159 CLR 550, 584 (Mason J); South Australia v O’Shea (1987) 163 CLR 378, 386 (Mason J); Haoucher v Minister for Immigration and Ethnic Affairs [1990] HCA 22; (1990) 169 CLR 648, 652–3 (Deane J); Annetts v McCann (1990) 170 CLR 596, 598 (Mason CJ, Deane and McHugh JJ); A-G (NSW) v Quin (1990) 170 CLR 1, 57 (Dawson J); Minister for Immigration and Ethnic Affairs v Teoh [1995] HCA 20; (1995) 183 CLR 273, 290, 311 (McHugh J).

[70] [1984] AC 808, 821. See also A-G (UK) v Ryan [1980] AC 718; Minister for Immigration and Ethnic Affairs v Pochi (1980) 4 ALD 139, 156–60 (Deane J), 152 (Evatt J agreeing).

[71] (1990) 170 CLR 321, 367 (‘Bond’).

[72] Galligan, above n 64, 333. Of course, there sometimes is a link between procedural fairness and rationality: ‘Some form of participation usually would be an important aspect of instrumental rationality, since the parties most affected by an action may be able to offer information and argument which would enable a better decision to be made.’ But this is not the rationale of procedural fairness: at 333; see also 93.

[73] Kioa v West [1985] HCA 81; (1985) 159 CLR 550, 584 (Mason J). The High Court has suggested that procedural fairness may be a universal requirement of good decision-making, rather than a requirement only where a person’s rights, interests or legitimate expectations are affected: Haoucher v Minister for Immigration and Ethnic Affairs [1990] HCA 22; (1990) 169 CLR 648, 653 (Deane J); Annetts v McCann (1990) 170 CLR 596, 598 (Mason CJ, Deane and McHugh JJ). This is superficially attractive, but could only occur by making procedural fairness essentially devoid of content (the question would shift from ‘when is procedural fairness required to what does procedural fairness require?’: see Kioa v West [1985] HCA 81; (1985) 159 CLR 550, 585 (Mason J); Annetts v McCann (1990) 170 CLR 596, 598 (Mason CJ, Deane and McHugh JJ); Haoucher v Minister for Immigration and Ethnic Affairs [1990] HCA 22; (1990) 169 CLR 648, 652–3 (Deane J)). In particular, procedural fairness would lose its current defining, and valuable, feature of being a common law ‘bottom line’ for participation in government decision-making. This price might be worth paying if that were the only way that the courts could further develop rationality as a common law standard of good decision-making. But rationality does not have to be subsumed under fairness the courts can directly develop it as a common law decision-making standard, just as they have developed procedural fairness. Of course, as a common law standard, it would be subject to statutory exclusion, just as procedural fairness is.

[74] CCSU [1985] AC 374.

[75] Minister for Immigration and Ethnic Affairs v Guo (1997) 191 CLR 559, 598–9.

[76] Galligan, above n 64, 263.

[77] Aronson and Dyer, above n 64, 204 indirectly suggest this classification. On the distinction between procedure and the reasoning process, see Gary Lawson, ‘Outcome, Procedure and Process: Agency Duties of Explanation for Legal Conclusions’ (1996) 48 Rutgers Law Review 313, 316–19.

[78] See Australian Heritage Commission v Mount Isa Mines Ltd [1997] HCA 10; (1997) 187 CLR 297; Corporation of the City of Enfield v Development Assessment Commission [2000] HCA 5; (2000) 199 CLR 135; Timbarra Protection Coalition Inc v Ross Mining NL [1999] NSWCA 8; (1999) 46 NSWLR 55.

[79] See below Part IV(A)(1).

[80] The distinction between jurisdictional and non-jurisdictional errors of law has, of course, been abolished for the purposes of review under s 5(1)(f) of the Administrative Decisions (Judicial Review) Act 1977 (Cth) (‘ADJRA’) and s 20(2)(f) of the Judicial Review Act 1991 (Qld). Under the Migration Act 1958 (Cth) pt VIII, difficult questions arise as to the relationship between the allowable grounds of review ie, that the decision-maker ‘did not have jurisdiction to make the decision’ (s 476(1)(b)) or made certain kinds of ‘error of law’ (s 476(1)(e)) and the exclusion of certain grounds of review, such as relevant considerations (ss 476(3)(d)–(e)), which at common law would clearly be classified as ‘jurisdictional errors of law’.

[81] See R v Hull University Visitor; Ex parte Page [1992] UKHL 12; [1993] AC 682; R v Bedwellty Justices; Ex parte Williams [1997] AC 225, 233 (Lord Cooke); Boddington v British Transport Police [1999] AC 143, 154 (Lord Irvine).

[82] Craig v South Australia [1995] HCA 58; (1995) 184 CLR 163 can be read as at least strongly suggesting that all errors of law by administrative decision-makers go to jurisdiction: see, eg, Edwards v Giudice [1999] FCA 1836; (1999) 94 FCR 561, 587 (Finkelstein J). However, the Full Federal Court has declined to read Craig v South Australia in this way: Construction, Forestry, Mining and Energy Union v Australian Industrial Relations Commission [1999] FCA 847; (1999) 93 FCR 317, 340–1 (Wilcox and Madgwick JJ). Unfortunately, the judgments of Gleeson CJ, Gummow, Kirby and Hayne JJ in City of Enfield v Development Assessment Commission [2000] HCA 5; (2000) 199 CLR 135, 151–5, and Kirby J in Coal and Allied Operations Pty Ltd v Australian Industrial Relations Commission (2000) 174 ALR 585, 608–12, while relevant, shed little light on the issue.

[83] In this context, ‘question’ and ‘error’ of law are two sides of the same coin: a decision on a ‘question’ of law is an ‘error’ of law whenever the decision-maker answers the question differently from the way a reviewing court would answer it.

[84] This is the practical effect of the approach of treating some questions of statutory interpretation (when the statute is said to use words according to their ‘ordinary meaning’) as questions of fact which must be answered ‘reasonably’. However, that approach is currently applied inconsistently and unpredictably.

[85] See Chevron USA v National Resources Defence Council, [1984] USSC 140; 467 US 837 (1984). As Levin says, ‘[s]cholarship about that case has been the single hottest topic in the [US] administrative law literature of the past decade’: Ronald Levin, ‘The Anatomy of Chevron: Step Two Reconsidered’ (1997) 72 Chicago-Kent Law Review 1253, 1253. As a result, the secondary literature on Chevron is voluminous. The most useful and/or influential articles are: Colin Diver, ‘Statutory Interpretation in the Administrative State’ (1985) 133 University of Pennsylvania Law Review 549; Robert Anthony, ‘Which Agency Interpretations Should Get Judicial Deference? — A Preliminary Inquiry’ (1988) 40 Administrative Law Review 121; Justice Antonin Scalia, ‘Judicial Deference to Administrative Interpretations of Law’ [1989] Duke Law Journal 511; Peter Strauss, ‘One Hundred Fifty Cases per Year: Some Implications of the Supreme Court’s Limited Resources for Judicial Review of Agency Action’ (1987) 87 Columbia Law Review 1093; Cynthia Farina, ‘Statutory Interpretation and the Balance of Power in the Administrative State’ (1989) 89 Columbia Law Review 452; Robert Anthony, ‘Which Agency Interpretations Should Bind Citizens and the Courts?’ (1990) 7 Yale Journal on Regulation 1; Laurence Silberman, ‘Chevron — The Intersection of Law and Policy’ (1990) 58 George Washington Law Review 821; Peter Schuck and E Donald Elliott, ‘To the Chevron Station: An Empirical Study of Federal Administrative Law’ [1990] Duke Law Journal 984; Cass Sunstein, ‘Law and Administration after Chevron(1990) 90 Columbia Law Review 2071; Michael Herz, ‘Deference Running Riot: Separating Interpretation and Lawmaking under Chevron(1992) 6 Administrative Law Journal of the American University 187; Mark Seidenfeld, ‘A Syncopated Chevron: Emphasizing Reasoned Decisionmaking in Reviewing Agency Interpretations of Statutes’ (1994) 73 Texas Law Review 83; John Duffy, ‘Administrative Common Law in Judicial Review’ (1998) 77 Texas Law Review 113.

[86] This leaves aside the issue of non-jurisdictional error of law on the face of the record.

[87] See, eg, Abebe [1999] HCA 14; (1999) 197 CLR 510, 554 (Gaudron J) and Eshetu [1999] HCA 21; (1999) 197 CLR 611, 659 (Hayne J), 672 (Callinan J), where several judges left open the question whether ‘Wednesbury unreasonableness’ amounts to a jurisdictional error for the purposes of prerogative relief under s 75(v) of the Australian Constitution; cf Gleeson CJ and McHugh J in Abebe (at 537), who seemed to have no doubt that ‘Wednesbury unreasonableness’ is reviewable under s 75(v), and Kirby J in Abebe (at 584), who argued that the scope of jurisdictional error for the purposes of s 75(v)

is not today to be confined to the narrower categories which were doubtless within the ‘intentions’ of the framers ... [T]he meaning of the constitutional text marches in step with developing understandings of the law of judicial review, stimulated by important decisions of the courts and (in a general way at least) by statutory developments ...

[88] This is because ADJRA proceedings are more numerous, giving the Federal Court greater opportunity to develop the grounds of review, and because the ADJRA’s structure has also freed the Federal Court from the need to articulate all grounds in jurisdictional terms.

[89] (1997) 191 CLR 559, 574–5 (Brennan CJ, Dawson, Toohey, Gaudron, McHugh and Gummow JJ).

[90] Abebe (1999) 197 CLR 511, 543–4.

[91] Ibid 545.

[92] Ibid.

[93] (1990) 170 CLR 1, 37–8; see also Kioa v West [1985] HCA 81; (1985) 159 CLR 550, 622 (Brennan J); South Australia v O’Shea (1987) 163 CLR 378, 410–11 (Brennan J). See further Aronson and Dyer, above n 64, 193.

[94] (1986) 162 CLR 24, 40–1.

[95] Galligan, above n 64, 257.

[96] Ibid 262.

[97] Elster, Nuts and Bolts, above n 10, 25.

[98] [1971] AC 682, 700.

[99] William Wade and Christopher Forsyth, Administrative Law (7th ed, 1994) 399.

[100] Galligan, above n 64, 140.

[101] Dibo Pty Ltd v Minister for Community Services and Health [1992] FCA 111; (1992) 27 ALD 421, 443 (Einfeld J).

[102] Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 234.

[103] This section concentrates on the most common situation, where a decision-maker has to make an ‘ordinary’ finding of fact: did or does X exist? Space precludes a discussion of the related issues that arise where decision-makers have to find that X does not exist (see, eg, Repatriation Commission v O’Brien [1985] HCA 10; (1985) 155 CLR 422), have to predict future facts (see, eg, Minister for Immigration and Ethnic Affairs v Guo (1997) 191 CLR 559, 574–5 (Brennan CJ, Dawson, Toohey, Gaudron, McHugh and Gummow JJ); Minister for Immigration and Ethnic Affairs v Wu (1996) 185 CLR 259, 285 (Kirby J)), or have to form ‘beliefs’ or ‘satisfactions’ about facts (see, eg, Boucaut Bay Co Ltd (in liq) v Commonwealth [1927] HCA 59; (1927) 40 CLR 98, 101; Metropolitan Gas Co v Federal Commissioner of Taxation [1932] HCA 58; (1932) 47 CLR 621, 632 (Gavan Duffy CJ and Starke J); Reid v Sinderberry [1944] HCA 15; (1944) 68 CLR 504, 511–12 (Latham CJ and McTiernan J); R v Connell; Ex parte Hetton Bellbird Collieries Ltd [1944] HCA 42; (1944) 69 CLR 407; Stenhouse v Coleman [1944] HCA 36; (1944) 69 CLR 457; Avon Downs Pty Ltd v Federal Commissioner of Taxation [1949] HCA 26; (1949) 78 CLR 353, 360 (Dixon J); Buck v Bavone [1976] HCA 24; (1976) 135 CLR 110, 118–19 (Gibbs J); Minister for Immigration and Ethnic Affairs v Wu (1996) 185 CLR 259, 275–6 (Brennan CJ, Toohey, McHugh and Gummow JJ)). On the requirements to form a ‘reasonable belief’, see George v Rockett [1990] HCA 26; (1990) 170 CLR 104, 112; Minister for Immigration and Ethnic Affairs v Teo (1995) 57 FCR 194, 198 (Black CJ, Gummow and Beazley JJ). In general terms, though, it is clear that the rationality standard applies to these kinds of factual findings as well. Nor can I explore the specific issues that arise as a result of the peculiar articulation of the ‘no evidence’ ground of review in ss 5(1)(h) and 5(3) of the ADJRA: see Curragh Queensland Mining Ltd v Comptroller General of Customs [1992] FCA 44; (1992) 34 FCR 212.

[104] See, eg, R v Nat Bell Liquors Ltd [1922] AC 128, 152 (Lord Sumner); Azzopardi v Tasman UEB Industries Ltd (1985) 4 NSWLR 139, 155–6 (Glass JA).

[105] R v Hillingdon London Borough Council; Ex parte Puhlhofer [1986] UKHL 1; [1986] 1 AC 484, 518; Waterford v Commonwealth [1987] HCA 25; (1987) 163 CLR 54, 77 (Brennan J).

[106] Aronson and Dyer, above n 64, 271 (emphasis in original).

[107] See, eg, Folkestone Corporation v Brockman [1914] UKLawRpAC 6; [1914] AC 338, 367 (Lord Atkinson); British Launderers’ Research Association v Borough of Hendon Rating Authority [1949] 1 KB 462, 471 (Denning LJ); Edwards v Bairstow [1955] UKHL 3; [1956] AC 14; Ashbridge Investments Ltd v Minister of Housing and Local Government [1965] 3 All ER 371, 374 (Lord Denning MR); R v Governor of Brixton Prison; Ex parte Armah [1966] 3 All ER 177 sub nom Armah v Government of Ghana [1968] AC 192, 257 (Lord Upjohn); A-G (UK) v Ryan [1980] AC 718; Mahon v Air New Zealand Ltd [1984] AC 808; Sinclair v Mining Warden at Maryborough [1975] HCA 17; (1975) 132 CLR 473, 481 (Barwick CJ), 483 (Gibbs J); Hope v Bathurst City Council [1980] HCA 16; (1980) 144 CLR 1, 8–9 (Mason J); Luu v Renevier [1989] FCA 518; (1989) 91 ALR 39 (FCA FC); GTE (Australia) Pty Ltd v Brown (1986) 14 FCR 309, 336 (Burchett J).

[108] (1990) 170 CLR 321, 355–6, 359–60 (Mason CJ), 367 (Deane J).

[109] See J M Evans et al, Administrative Law: Cases, Text, and Materials (4th ed, 1995) 659; Galligan, above n 64, 143.

[110] Shulver v Sherry (1992) 28 ALD 570, 574 (Hayne J); Bond (1990) 170 CLR 321, 367 (Deane J).

[111] See, eg, Epeabaka v Minister for Immigration and Multicultural Affairs [1997] FCA 1413; (1997) 150 ALR 397, 400 (Finkelstein J).

[112] See, eg, Minister for Immigration and Ethnic Affairs v Pochi (1980) 4 ALD 139, 155 (Deane J); Epeabaka v Minister for Immigration and Multicultural Affairs [1997] FCA 1413; (1997) 150 ALR 397, 400 (Finkelstein J).

[113] (1992) 28 ALD 570, 574. See also R v Deputy Industrial Injuries Commissioner; Ex parte Moore [1965] 1 QB 456, 488 (Diplock LJ); Mahon v Air New Zealand Ltd [1984] AC 808, 821 (Diplock LJ); A-G (UK) v Ryan [1980] AC 718; Minister for Immigration and Ethnic Affairs v Pochi (1980) 4 ALD 139, 156–60 (Deane J). The requirement that material have some probative value is essentially the same as the fundamental rule of evidence in judicial proceedings that, to be admissible, material must be ‘relevant’ to the issues to be decided. Thus, s 55(1) of the Evidence Act 1995 (Cth) provides: ‘The evidence that is relevant in a proceeding is evidence that, if it were accepted, could rationally affect (directly or indirectly) the assessment of the probability of the existence of a fact in issue in the proceeding.’ On relevance as the fundamental rule of evidence, see Andrew Palmer, Principles of Evidence (1998) 75–86; Andrew Ligertwood, Australian Evidence (3rd ed, 1998) 51–4. Material that cannot rationally support any conclusion on a factual question is not ‘evidence’ at all, in either administrative or judicial proceedings. However, as Hayne J pointed out in Shulver v Sherry, this does not mean that administrative decision-makers can only consider material that would be admissible in judicial proceedings. In judicial proceedings other rules of evidence, such as the hearsay rule, exclude material that meets the basic test of relevance. These other exclusionary rules usually do not apply to administrative decision-making processes. For this reason, it has been suggested that the term ‘evidence’ should be avoided in relation to administrative decision-making: see Blackwood Hodge (Aust) Pty Ltd v Collector of Customs (NSW) [1980] FCA 96; (1980) 3 ALD 38, 48 (Fisher J); Collins v Minister for Immigration and Ethnic Affairs [1981] FCA 147; (1981) 4 ALD 198, 201 (Fox, Deane and Morling JJ); Minister for Immigration and Ethnic Affairs v Wu (1996) 185 CLR 259, 282 (Brennan CJ, Toohey, McHugh and Gummow JJ). However, the term seems too well established to be abandoned now.

[114] See, eg, Singh v Minister for Immigration and Ethnic Affairs (1985) 9 ALN N13b; Waniewska v Minister for Immigration and Ethnic Affairs (1986) 70 ALR 284, 299 (Keely J); Lek v Minister for Immigration, Local Government and Ethnic Affairs [1993] FCA 493; (1993) 45 FCR 418; Deloitte Touche Tohmatsu v Australian Securities Commission (1996) 136 ALR 453; Teo v Minister for Immigration and Ethnic Affairs [1994] FCA 1017; (1994) 49 FCR 409 (Full Federal Court). On appeal in Minister for Immigration and Ethnic Affairs v Teoh [1995] HCA 20; (1995) 183 CLR 273, several members of the High Court cited with approval a number of the Federal Court decisions and accepted that, in principle, a failure to inquire could invalidate a decision, although they disagreed with the Federal Court’s view that this was such a case. The Court did not distinguish between a duty to inquire based on relevant considerations and one based on unreasonableness: at 289–90 (Mason CJ and Deane J), 302–3 (Toohey J, who seemed to accept that there had been a relevant failure to inquire, but ultimately found the decision invalid on other grounds), 320–1 (McHugh J). In Foster v Minister for Customs and Justice [2000] HCA 38; (2000) 173 ALR 585, Gleeson CJ, McHugh and Kirby JJ all accepted that a failure to obtain further information could amount to a failure to properly consider a relevant consideration, although this was not such a case: at 592 (Gleeson CJ and McHugh J), 616 (Kirby J).

[115] See, eg, Prasad v Minister for Immigration and Ethnic Affairs (1985) 6 FCR 155, 170 (Wilcox J); Videto v Minister for Immigration and Ethnic Affairs [No 2] [1985] FCA 449; (1985) 8 FCR 167,

178–9 (Toohey J); Luu v Renevier [1989] FCA 518; (1989) 91 ALR 39 (FCA FC); Tickner v Bropho [1993] FCA 208; (1993) 40 FCR 183. In Sun v Minister for Immigration and Ethnic Affairs [1997] FCA 1488; (1997) 81 FCR 71, 118–19 (Wilcox J), a failure to inquire was held to be unreasonable and therefore a breach of the Refugee Review Tribunal’s duty to act with ‘substantial justice’ under s 420(2)(b) of the Migration Act 1958 (Cth). The High Court has since held that the ‘substantial justice’ requirement cannot be used to put back the grounds of judicial review, such as unreasonableness, excluded under pt VIII of the Act: Eshetu [1999] HCA 21; (1999) 197 CLR 611. However, the analysis of the relationship between the failure to obtain evidence and the requirements of rational decision-making remains sound.

[116] This answers McMillan’s concerns: see McMillan, above n 65, 381.

[117] [1989] FCA 175; (1989) 86 ALR 435, 453.

[118] (1990) 170 CLR 321, 356–7.

[119] Eshetu [1999] HCA 21; (1999) 197 CLR 611, 654 (emphasis added).

[120] Mahon v Air New Zealand Ltd [1984] AC 808, 821.

[121] Ibid 833.

[122] R v Hillingdon London Borough Council; Ex parte Puhlhofer [1986] UKHL 1; [1986] 1 AC 484, 518.

[123] Cf Mason CJ in Bond, who said ‘“perversely” signifies acting without any probative evidence’: (1990) 170 CLR 321, 359. For recent discussions on the relationship between ‘perversity’ and ‘no evidence’ in Australia, see Bruce v Cole [1998] NSWCA 45; (1998) 45 NSWLR 163, 188–9 (Spigelman CJ); Dibeek Holdings Pty Ltd v Notaras [2000] FCA 1212 (Unreported, Miles, Sackville and Katz JJ, 30 August 2000), available on <http://scaleplus.law.gov.au> at 31 December 2000 (copy on file with author) [42], [52].

[124] (1985) 4 NSWLR 139, 151. See also Minister for Immigration, Local Government and Ethnic Affairs v Pashmforoosh (1989) 18 ALD 77, 80 (Davies, Burchett and Lee JJ).

[125] Khan v Minister for Immigration and Ethnic Affairs [1987] FCA 457; (1987) 14 ALD 291, 292 (Gummow J). This was quoted in Hindi v Minister for Immigration and Ethnic Affairs (1988) 20 FCR 1, 12–13 (Sheppard J). See also Turner v Minister for Immigration and Ethnic Affairs [1981] FCA 65; (1981) 35 ALR 388, 392, where Toohey J held that a matter had to be taken into account in a ‘real sense’.

[126] Tickner v Chapman [1995] FCAFC 1726; (1995) 57 FCR 451, 462 (Black CJ). See also Tobacco Institute of Australia v National Health and Medical Research Council [1996] FCA 1150; (1996) 71 FCR 265.

[127] Prasad v Minister for Immigration and Ethnic Affairs (1985) 6 FCR 155. Examples of this principle in operation can be seen in a series of migration cases, including: Singh v Minister for Immigration and Ethnic Affairs (1987) 15 FCR 4, 10 (Forster J); Thavarajasingham v Minister for Immigration, Local Government and Ethnic Affairs (1990) 19 ALD 751, 754 (Davies J); Pancharatnam v Minister for Immigration, Local Government and Ethnic Affairs (1991) 21 ALD 217; Lek v Minister for Immigration, Local Government and Ethnic Affairs [No 2] [1993] FCA 493; (1993) 45 FCR 418, 435–6 (Wilcox J); Fuduche v Minister for Immigration, Local Government and Ethnic Affairs [1993] FCA 503; (1993) 45 FCR 515, 526–7 (Burchett J).

[128] [1997] FCA 1413; (1997) 150 ALR 397.

[129] Ibid 401.

[130] Ibid 401–2.

[131] R v District Court of the Metropolitan District Holden at Sydney; Ex parte White [1966] HCA 69; (1966) 116 CLR 644, 654, cited in Bond (1990) 170 CLR 321, 356 (Mason CJ).

[132] Bond (1990) 170 CLR 321, 356 (emphasis in original).

[133] Mahon v Air New Zealand Ltd [1984] AC 808, 821, 832.

[134] Bond (1990) 170 CLR 321, 356–7.

[135] (1987) 15 FCR 4.

[136] (1989) 18 ALD 77.

[137] Bond (1990) 170 CLR 321, 359.

[138] Epeabaka v Minister for Immigration and Multicultural Affairs [1997] FCA 1413; (1997) 150 ALR 397, 402. Cf Roads Corporation v Dacakis [1995] VicRp 70; [1995] 2 VR 508, 520 (Batt J).

[139] (1992) 28 ALD 570, 574.

[140] In addition to the post-Bond cases cited in above n 127 and accompanying text, and the cases cited in Aronson and Dyer, above n 64, 286–7, see Othman v Minister for Immigration and Ethnic Affairs (1991) 24 ALD 707; TCN Channel Nine Pty Ltd v Australian Broadcasting Tribunal (1992) 28 ALD 829.

[141] [1999] FCA 1; (1999) 84 FCR 411.

[142] Ibid 421–2.

[143] Aronson and Dyer, above n 64, 281.

[144] [1998] NSWCA 45; (1998) 45 NSWLR 163.

[145] [1999] NSWCA 477; (1999) 48 NSWLR 161.

[146] Hill v Green [1999] NSWCA 477; (1999) 48 NSWLR 161, 175. See also at 176–7 (Mason P), 213 (Fitzgerald JA, Beazley JA agreeing).

[147] [1999] HCA 14; (1999) 197 CLR 510, 544–5 (emphasis added). In Eshetu [1999] HCA 21; (1999) 197 CLR 611, 628 their Honours made it clear that it is not sufficient simply to argue that the decision-maker

gave inadequate weight to certain considerations and undue weight to others. ... [A] process of reasoning flawed in those respects ... is not a case of Wednesbury unreasonableness, and it does not constitute a proper basis for the grant of constitutional relief under s 75(v) of the Constitution.

But that merely repeats the basic principle stated by Mason J in Minister for Aboriginal Affairs v Peko-Wallsend Ltd (1986) 162 CLR 24 that weight goes to merits. In that case Mason J also held that, in some cases, a decision-maker’s allocation of weight to particular factors can be challenged for unreasonableness. Their Honours certainly did not dissent from this view, but simply reiterated (at 627) his warning that ‘a court should proceed with caution ... lest it exceed its supervisory role by reviewing the decision on its merits’: Minister for Aboriginal Affairs v Peko-Wallsend Ltd (1986) 162 CLR 24, 42 (Mason J). Nothing in this suggests an approach different from the one taken in Abebe.

[148] [1999] HCA 21; (1999) 197 CLR 611, 646.

[149] Ibid 656–7 (emphasis added).

[150] Ibid 640 (emphasis added).

[151] X v Commonwealth (1999) 167 ALR 529, 561–2.

[152] [1999] NSWCA 477; (1999) 48 NSWLR 161, 212.

[153] Elster, ‘Introduction’, above n 10, 1.

[154] It is implicit in this that I reject the cynical public choice analysis that denies the possibility of legislatures ever acting in the public interest.

[155] Industrial Union Department, AFL-CIO v American Petroleum Institute, [1980] USSC 152; 448 US 607, 670 (1980) (Powell J) (‘Benzene Case’). Further, the United States Supreme Court has held that an agency must ‘explain’ its rules (ie, give reasons for its decisions) even where that was not required by statute or constitutional due process requirements: Motor Vehicle Manufacturers Association of the United States v State Farm Mutual Automobile Insurance Co, [1983] USSC 148; 463 US 29 (1983).

[156] See, eg, Bannister Quest Pty Ltd v Australian Fisheries Management Authority (1997) 77 FCR 503, where the Federal Court held that the Australian Fisheries Management Authority had, under industry pressure, pursued an illegitimate end. Agency capture occurs when an agency, established to regulate an industry, identifies too strongly with that industry, losing sight of the public interest it was established to protect.

[157] See ADJRA ss 5(1)(e), (2)(c); Municipal Council of Sydney v Campbell [1925] AC 339; Council of the Shire of Werribee v Kerr [1928] HCA 41; (1928) 42 CLR 1; Swan Hill Corporation v Bradbury [1937] HCA 15; (1937) 56 CLR 746; Arthur Yates & Co Pty Ltd v Vegetable Seeds Committee [1945] HCA 55; (1940) 72 CLR 37; Padfield v Minister of Agriculture, Fisheries and Food [1968] UKHL 1; [1968] AC 997; R v Toohey; Ex parte Northern Land Council [1981] HCA 74; (1981) 151 CLR 170; Woollahra Municipal Council v Minister for the Environment (1991) 23 NSWLR 710; Shop Distributive and Allied Employees Association v Minister for Industrial Affairs (SA) (1995) 183 CLR 552; R v Secretary of State of Foreign Affairs and Commonwealth Affairs; Ex parte World Development Movement Ltd [1994] EWHC Admin 1; [1995] 1 WLR 386.

[158] Swan Hill Corporation v Bradbury [1937] HCA 15; (1937) 56 CLR 746, 757–8.

[159] Craig, Administrative Law, above n 49, 406.

[160] Irvine, above n 9, 68.

[161] See Colin Diver, ‘Policymaking Paradigms in Administrative Law’ (1981) 95 Harvard Law Review 393, 409–13.

[162] Originally, powers to make by-laws that existed at common law or under charters of incorporation could not be reviewed for consistency with the head legislation, but were reviewed for ‘unreasonableness’, a standard of review that was much closer to merits review than ‘Wednesbury’ unreasonableness is: see, eg, Slattery v Naylor [1888] AC 446, 451–2 (Lord Hobhouse); Alan Wharam, ‘Judicial Control of Delegated Legislation’ (1973) 36 Modern Law Review 611, 611–12; Widgee Shire Council v Bonney [1907] HCA 11; (1907) 4 CLR 977, 984 (Isaacs J). As Parliament’s legislative monopoly was established in the 19th century (see Craig, Administrative Law, above n 49, 4–5; P P Craig, Public Law and Democracy in the United Kingdom and the United States of America (1990) 19–30), these non-statutory powers were replaced by express statutory grants. In Slattery v Naylor it was held that only the stricter modern test applies to such statutory powers (see also Alan Wharam, ‘Judicial Control of Delegated Legislation’ (1973) 36 Modern Law Review 611, 616). This change in review standard explains the seemingly inconsistent statements made in some cases that delegated legislation cannot be attacked as ‘unreasonable’ but can be attacked on the ground that no reasonable person would have made it: see Widgee Shire Council v Bonney [1907] HCA 11; (1907) 4 CLR 977, 982–3 (Griffith CJ); Williams v Melbourne Corporation [1933] HCA 56; (1933) 49 CLR 142, 156 (Dixon J); Clements v Bull [1953] HCA 61; (1953) 88 CLR 572, 577 (Williams ACJ and Kitto J).

[163] [1953] HCA 61; (1953) 88 CLR 572, 577 (citation omitted).

[164] [1933] HCA 56; (1933) 49 CLR 142, 155.

[165] [1944] HCA 36; (1944) 69 CLR 457, 467.

[166] Coulter v The Queen (1988) 164 CLR 350, 357 (Mason CJ, Wilson and Brennan JJ).

[167] South Australia v Tanner [1989] HCA 3; (1989) 166 CLR 161, 168 (Wilson, Dawson, Toohey and Gaudron JJ); New South Wales v Macquarie Bank Ltd (1992) 30 NSWLR 307, 320 (Kirby P).

[168] [1888] AC 446, 452 (Lord Hobhouse).

[169] [1898] UKLawRpKQB 101; [1898] 2 QB 91, 99–100 (Lord Russell CJ).

[170] [1933] HCA 56; (1933) 49 CLR 142, 154. See also Widgee Shire Council v Bonney [1907] HCA 11; (1907) 4 CLR 977, 986 (Isaacs J); Brunswick Corporation v Stewart [1941] HCA 7; (1941) 65 CLR 88, 97 (Starke J), 99 (Williams J); Eremin v Minister for Immigration, Local Government and Ethnic Affairs [1990] FCA 326; (1990) 21 ALD 69, 78 (Lockhart, Gummow and Foster JJ).

[171] [1989] HCA 3; (1989) 166 CLR 161, 165 (Wilson, Dawson, Toohey and Gaudron JJ). See also Bienke v Minister for Industries and Energy, where the Full Federal Court said: ‘Delegated legislation may be struck down, even if the purpose of a law is to achieve an end within power, if the means adopted is not capable of being reasonably proportionate to the pursuit of that end’: (1996) 63 FCR 567, 578.

[172] [1993] FCA 366; (1993) 43 FCR 565, 585.

[173] (1992) 30 NSWLR 307, 323–4.

[174] Ibid 324.

[175] A number of cases have reached a similar conclusion: see especially R v Secretary of State for the Home Department; Ex parte Brind [1991] UKHL 4; [1991] AC 696; Bruce v Cole [1998] NSWCA 45; (1998) 45 NSWLR 163, 185 (Spigelman CJ).

[176] The more difficult question is: when does a rationality test, including a proportionality requirement, apply to delegated legislation? In general terms, this will be the case whenever the delegated legislation’s validity depends on it being reasonably appropriate and adapted to achieving its purpose, but to some extent that only reframes the question. This issue caused some confusion in both Paull v Munday [1903] ArgusLawRp 108; (1976) 9 ALR 245 and Minister for Foreign Affairs and Trade v Magno [1992] FCA 566; (1992) 37 FCR 298, where the majority judges completely overlooked the purposive element of the relevant regulation-making power. See also Minister for Urban Affairs and Planning v Rosemount Estates Pty Ltd [1996] NSWSC 348; (1996) 91 LGERA 31, where the majority adopted a self-contradictory position on the issue. Space precludes further discussion of this issue here.

[177] Wednesbury [1947] EWCA Civ 1; [1948] 1 KB 223, 223–4 (Lord Greene MR).

[178] Paul Walker, ‘What’s Wrong with Irrationality?’ [1995] Public Law 556, 556–7.

[179] Ibid 566.

[180] Ibid.

[181] Ibid 569.

[182] Ibid.

[183] See above Part II(C).

[184] Chaïm Perelman, ‘The Rational and the Reasonable’ in Theodore Geraets (ed), Rationality To-Day (1979) 213, 213.

[185] Walker, above n 178, 566; see also Lord Diplock in CCSU [1985] AC 374, 411.

[186] Perelman, above n 184, 214 (emphasis in original).

[187] Peters, above n 24, 303–4.

[188] CCSU [1985] AC 374, 411.