Project Blue Sky

Invalidity and the evolution of consequences for unlawful administrative action

Speech delivered to the Australian Institute of Administrative Law on 20 November 2012

Section I

I recently encountered Justice Beazley in an elevator in the Law Courts Building and she, sensing me to be vulnerable, asked me whether I would not agree to give a speech this evening on this rather daunting topic. Foolishly, I agreed. The shared Supreme and Federal Court judges’ lifts are one of those rare examples of co-operative federalism in action; this speech is therefore to be seen as one of the bounties of federalism. I hope it will not be as dubious as some of the other bounties.

There are two aspects of the topic which I want to underscore at the outset and these are its references to invalidity and to unlawfulness. These concepts have, as I will hope to show you, an uneasy relationship. We now know that not all unlawful administrative action is invalid even if we probably feel that all invalid action is unlawful. Can I sketch at the outset an inherent incoherence in that statement: if an administrative act is truly invalid it cannot be unlawful – according to current theory it simply does not exist.^[1] The state of being unlawful – that is, that the decision is contrary to law – is a legal conclusion about an event which is taken to exist. There can be, if one thinks about it clearly, no such thing as unlawful invalid action for a truly invalid act does not sufficiently exist to contravene any rule of law.

To give an example, on the present learning of the High Court (the judges of which may also, from time to time, be found in the same Law Courts elevators), a decision made in breach of the rules of procedural fairness is no decision at all.^[2] It does not contravene its authorising statute – it has simply never happened. On the other hand, where a decision maker commits that rare creature – the non-jurisdictional error of law – questions of unlawfulness do arise but they generally have no consequences. If the decision is not invalid, and leaving to one side for now the difficulty for the prevailing theory constituted by the fact that certiorari may be granted for non-jurisdictional error of law patent of the face of the record, a court has no jurisdiction to set it aside even if it is otherwise unlawful.^[3] Under the current vision, therefore, the lawfulness of administrative action is a redundant inquiry – it either does not matter because the decision is invalid and cannot be unlawful because it does not exist or the decision is valid in which cases its lawfulness is irrelevant to any power the Court has.

What I want to propose this evening is that this is an incoherent theory, unsupported historically, and an inaccurate description of what in fact courts are presently doing and have always done. Instead, what I want to suggest is that administrative law should be focussed on whether administrative action is lawful and, if it be concluded that it is not, that the courts should decide what the consequences of that unlawfulness are to be. We should abandon the fiction – of which Project Blue Sky Inc v Australian Broadcasting Authority^[4] is the exemplar – that the Parliament has told the courts what to do when unlawful administrative action is discovered – when it quite clearly has not – and instead develop a jurisprudence directed to explaining why some unlawful action should be set aside whilst other unlawful action should be left in place. As an essential step in that process we ought to free ourselves from the notion that administrative action is ever invalid and dispense with its handmaiden, the concept of jurisdictional error, which gives rise to many more problems than it solves.

This is not to say that the baby should be thrown out with the bath water. Many of the statements which appear in cases concerned with ascertaining whether a decision is invalid provide useful insights into whether an unlawful decision is one that ought to be set aside. The critical difference is that the issue ought not to be approached on the fictional basis that the statute provides in every case a definitive yes or no answer to the question of what should happen if its provisions be contravened. The insistence that the statute does provide that answer leads one into the realm of invalid action and, as I will endeavour to show, only incoherence lies in that direction.

I might begin then with a quick statement of the present rule. The critical passage appears in the joint judgment of McHugh, Gummow, Kirby and Hayne JJ in Project Blue Sky. Having first explained that the former distinction between mandatory and directive provisions in a statute obscures the real question their Honours went on to observe that:^[5]

A better test for determining the issue of validity is to ask whether it was a purpose of the legislation that an act done in breach of the provision should be invalid. This has been the preferred approach of courts in this country in recent years, particularly in New South Wales. In determining the question of purpose, regard must be had to "the language of the relevant provision and the scope and object of the whole statute".

(References omitted.)

This approach is premised on the notion of validity. The actual inquiry the joint judgment suggested – that is, consider the language of the statute and its scope and object – are I think perfectly acceptable inputs into whether a decision ought to be set aside. But seen through the prism of validity they exclude any consideration of whether other matters might bear upon the decision of whether to set aside a decision. These might include, for example, how serious the breach in question is or what the consequences for others might be. As I will later suggest, this infelicity in the doctrine has been papered over by an insistence that relief may be refused on discretionary grounds but this leaves one with the vista of an apparently effective albeit invalid decision. How a decision may be both invalid and effective is one which is not easy to understand.

That statement nevertheless is the orthodox and entrenched position. It was applied by the High Court in Commissioner of Taxation v Futuris Corporation Limited^[6] to conclude that errors in the process of producing a notice of assessment did not go to jurisdiction when the notice fell within the protection of the privative clause in s 175 of the Income Tax Assessment Act 1936 (Cth). There are many instances of its application at an intermediate appellate level: see, for example, the New South Wales Court of Appeal’s decision in Chase Oyster Bar Pty Ltd v Hamo Industries Pty Ltd^[7] and the recent Full Federal Court’s decision in Australian Industry Group v Fair Work Australia.^[8]

The principle has come to be applied in two broad ways. The first concerns what can be directly ascertained from the text of the statute. This involves a close construction of the words actually used or the structure of the legislation. An example is afforded by those cases where it is said that the authority of a repository of a power cannot be exercised in conflict with a provision which governs the manner of its exercise. In such cases, as Brennan CJ observed in Project Blue Sky,^[9] ‘[a] purported exercise of a power in breach of the provision which governs the manner of its exercise is invalid, since there is no power to support it’. An example of such a provision might be s 459G(2) of the Corporations Act 2001 which provides that an application to set aside a statutory demand ‘may only be made within 21 days after the demand is so served’. In David Grant & Co Pty Ltd v Westpac Banking Corporation Gummow J observed that the phrase ‘"[a]n application may only be made within 21 days" should be read as a whole. The force of the term "may only" is to define the jurisdiction of the court by imposing a requirement as to time as an essential condition of the new right conferred by s 459G. An integer or element of the right created by s 459G is its exercise by application made within the time specified’.^[10]

One might observe that what is being hinted at here is a notion that the requirement is mandatory (the very question said in Project Blue Sky to be the wrong question). Indeed, in Chase Oyster Bar, Spigelman CJ said in the course of explaining the principle that ‘[t]he first textual indicator that is always of significance is the mode of expression of the element directly in issue. Substantial, indeed often, but not always, determinative, weight must be given to language which is in mandatory form’. ^[11] One may ask what is so very different about asking whether the provision breached is ‘mandatory’ in effect (a forbidden question) as opposed to ‘mandatory in form’ (apparently permissible).

Another factor which is important is the structure of the legislation. David Grant again affords a useful illustration. There Gummow J applied^[12] the well-known statement in Anthony Hordern & Sons Ltd v Amalgamated Clothing and Allied Trades Unions of Australia^[13] to conclude that the specific words of s 459G(2) were not to be outflanked by the more general language of s 1322(4)(d).

Separate from these textual tools is a second set of principles which cannot be found in the words of any statute. These are fundamental rights which apply unless the Parliament indicates that they do not. They presently travel collectively under the title ‘the principle of legality’, an expression used by Lord Steyn in R v Home Secretary; ex parte Pierson^[14]. In Electrolux Home Products v Australian Workers Union, Gleeson CJ explained that the presumption that fundamental rights would not be affected without clear statutory language was:^[15]

an aspect of the principle of legality which governs the relations between Parliament, the executive and the courts. The presumption is not merely a common sense guide to what a Parliament in a liberal democracy is likely to have intended; it is a working hypothesis, the existence of which is known both to Parliament and the courts, upon which statutory language will be interpreted. The hypothesis is an aspect of the rule of law.

What are these rights? Professor Pearce and Associate Professor Geddes have conveniently located a long list of them at [5.36] of the 7^th Edition of Statutory Interpretation in Australia.^[16] For present purposes it suffices to observe that the list includes denying procedural fairness. Although not on the list, it would also include, in light of Craig v South Australia,^[17] taking into account irrelevant considerations, failing to take into account relevant considerations or arriving at a decision which was so unreasonable that no reasonable decision maker could possibly arrive at it.

The statement in Electrolux is apt to suggest that what is taking place when fundamental rights are concerned is a dialogue between the branches of government in which the final result – respect for fundamental rights – is to be seen as the consequence of a shared understanding, particularly between the Courts and Parliament, as to what Parliament does (or more accurately does not) intend by its silence.

There is a palpable ambiguity in this principle as to the source of the rights. Earlier authority of the High Court would have suggested that the source of the rights was the common law and that legislation was simply to be read as not intruding on rules of law established by the Courts. This is most likely the position the Court took in Coco v The Queen, where Mason CJ, Brennan, Gaudron and McHugh JJ said:^[18]

The courts should not impute to the legislature an intention to interfere with fundamental rights. Such an intention must be clearly manifested by unmistakable and unambiguous language. General words will rarely be sufficient for that purpose if they do not specifically deal with the question because, in the context in which they appear, they will often be ambiguous on the aspect of interference with fundamental rights.

Reading this one might have thought that the fundamental rights in question are legal principles which have been generated by the common law. The statement in Electrolux is much more complex. The rights do not emerge because the Courts have recognised those rights and because legislation is to be interpreted so as not to interfere with them. Rather, they emerge instead as the result of a shared assumption between the branches of government. More recently, the shared assumption in Electrolux has been accepted to be nothing more than ‘rules of interpretation accepted by all arms of government’. These were the words used by five Justices in Zheng v Cai.^[19] That view of affairs leads one inevitably to see the grounds of judicial review as elements, not of rules of common law generated from within the judiciary, but instead as rules of legislation prescribed by Parliament. And indeed so much has now been expressly accepted in the case of the rules of procedural fairness. In Saeed v Minister for Immigration and Citizenship the Court said, for example, that ‘[t]he implication of the principles of natural justice in a statute is therefore arrived at by a process of construction’.^[20]

That conclusion was inevitably required by the Court’s earlier acceptance that each of the traditional grounds of review gave rise to jurisdictional error under s 75(v) of the Constitution.^[21] Once unmasked as examples of jurisdictional error the process of reasoning in the passage quoted above from Project Blue Sky became necessary and this, in turn, inevitably made the inquiry one about the interpretation of legislation. After Project Blue Sky, it was not possible to conceive of jurisdictional errors arising outside of a legislative framework. The statement in Saeed is, in that sense, merely the logical working through of what had already been held in Aala^[22] and Project Blue Sky itself.

I am going to return to criticise this approach a little later. But I might just point out some of its curious features at this stage. The first is that it is a view which locates all legal authority in the Parliament. The Courts may be free to create common law rules away from the domain of public law but in that area it is now accepted that the only source of rules of law are Parliamentary ones. This view, of course, is the uncompromising vision of AV Dicey in which all legal authority proceeds from an Austinian sovereign.^[23] The Courts do no more than give effect to the will of that sovereign. All aspects of public law are to be seen as merely fulfilling a completely articulated set of rules already provided by the Parliament. The Courts are not involved in the creation of law at all; they are instead merely its spokesmen.

Problems with this probably include the fact that it does not accord with legal history – the rules with which we are concerned were generated over several centuries by the Courts as an exercise of their own authority and not as aspects of statutory interpretation. The theory’s denial of legislative competence in the Courts is inconsistent with the existence of the common law.

The second issue concerns the uneasy relationship between Parliament, judicial review and the common law. The statements of the High Court exhibit a marked inconsistency about this. In Saeed five Justices of the Court approved a statement of Brennan J in Kioa v West which contained these two sentences:^[24]

"the justice of the common law will supply the omission of the legislature". The true intention of the legislation is thus ascertained.

If one reads this carefully one will see that it is internally inconsistent. The first sentence appears to accept that the rules of procedural fairness arise from the common law but that they are somehow incorporated into the statute because the text of the statute is silent. The second sentence says the opposite – the rules are revealed to be Parliamentary rules. But if that be so, one may ask what the omission of the legislature referred to in the first sentence is. Statements of this kind reveal, I think, a deep conflict about what is taking place.

That deep conflict proceeds from the need under the present view to import the rules of judicial review as part of the instructions emanating from the statute (which in turn gives rise to the view that the grounds of judicial review are concerned with jurisdictional error). Under the former principle in Coco the source of the rules was the common law and the statute was simply interpreted as not interfering with those rights without clear language. The task of statutory interpretation was a simple one which involved reading down general language. It also assumed that the rights were generated by judge-made law. It was a view which sought to accommodate the interaction of two sets of legal principles derived from different sources.

Under the view in Saeed, what is involved is not so much reading down general language so as not to interfere with rights existing outside the statute. Rather what must be involved is reading into the statute itself the rights which are not to be affected. This involves a considerable conceptual leap and, as a process of construction, a move away from a traditional process of reading down to a radical process of reading in.

These are very different processes of statutory interpretation but Electrolux and Saeed treat them as if they were the same. It matters because the question of reading words into a statute is itself the subject of its own developed, if not entirely uncontroversial, jurisprudence. Ordinarily we are told that it is the text of a statute which controls and that Parliamentary intention is a useful aid to interpretation of the text. In interpreting a statute the aim is ‘to see what is the intention expressed by the words used’.^[25] In Saeed itself the same five Justices^[26] adopted what had been said by Gummow J in Wik Peoples v Queensland^[27] to conclude that it was ‘necessary to keep in mind that when it is said the legislative "intention" is to be ascertained, "what is involved is the ‘intention manifested’ by the legislation’ (emphasis in Saeed).

In the case of Electrolux and Saeed, however, Parliamentary intention is ascertained not from what the text says but from what it does not say. We should recognise this for the remarkable leap that it is.

Is this permissible? Under existing principle, words may be inserted into an Act in some situations. A comprehensive explanation of those principles is beyond the scope of this paper. It seems that it can occur, however, in certain circumstances. The High Court, for example, departed from the literal wording of a statute in Cooper Brookes (Wollongong) Pty Ltd v Federal Commissioner of Taxation^[28] where it was plain that an error had crept into the drafting. The most comprehensive explanation of the current doctrine appears in the remarks of Spigelman CJ in R v Young.^[29] The prevailing view is that one may (but need not necessarily) read into a statute words which have been omitted by the legislature if three pre-conditions are satisfied:

• firstly, one must know the mischief to which the legislation was directed;
• secondly, one must be satisfied that there has been inadvertence by the legislature in overlooking an eventuality which must be dealt with if the purpose of the legislation is to be achieved; and,
• finally, one must know with certainty the words which Parliament would have used.

It may well be possible to deduce from the principle of legality that the first of these requirements is satisfied. One could say, in light of the principle, that Parliament intended as one of its aims to provide a statute which provided for judicial review. With some modification one could even say the second principle was satisfied: perhaps the legislature did not so much overlook the matter but instead simply assumed that it did not need to be dealt with. The real difficulties emerge, however, with the third requirement. If it be the test, then one needs to be quite certain what the rules of judicial review are – one needs to be able to write them down with clarity and to say what they are before they can be read in.

In the case of the principles of judicial review this is a task which has eluded four centuries of public lawyers. It is unrealistic to expect that the entire text of the rules of judicial review can be written down and simply read into a statute. If it could, everyone in this room would be out of a job. There is a more serious problem, however. If what is involved is implication from the statute itself the meaning of the principles of judicial review must be fixed and known at the time the statute is passed and neither its meaning, nor those principles, may later alter for the meaning of a statute does not alter with the passage of time. This has serious consequences for the development of judicial review. If at some later time a new principle of judicial review is discovered – say, for example, a doctrine of irrationality – it cannot be taken to form part of a statutory text which antedates the time of the new principle’s discovery. Taking the logic of the position to its extremes, one might well have different principles of judicial review for statutes passed at different times.

That problem could be surmounted only if one permits that what is implied from the statute is not a fixed set of principles at a particular time but instead an instruction to apply whatever the judges say constitutes, at any particular moment, the principles of judicial review. That solution would indeed overcome the problem arising from the fact that the meaning of a statute is generally fixed at the time of its passage but only at the price of accepting that what was occurring was the giving of legislative effect to judge-made rules. Unfortunately, the Court cannot commit itself to that principle because we have been told that it is beyond the power of the Parliament to give effect to the common law. Such a step would involve the invalid conferral of legislative authority on the Courts: see Western Australia v Commonwealth (Native Title Act Case).^[30]

There is no ready answer to this problem. At this point an interlocutor might well interrupt and say ‘Well this all very interesting, but does it ever matter?’ The answer is that it matters a lot.

Section II

The theory that decisions made in excess of jurisdiction are invalid generate a suite of problems which have been known about – and counselled against – for a long time. There is nothing especially new in the problems I am going to list. Professor Craig has given a full account of them and before him Professor Wade. The problems are:

(a) the inability to take into account in determining whether jurisdictional error is present the seriousness of the breach of the rules which have occurred;

(b) the incoherence of discretionary refusal of relief and the void/voidable distinction;

(c) the inability to explain aspects of the writ of certiorari and, in particular, the fact that it may be granted in the cases of non-jurisdictional error of law; and

(d) the inability to explain the basis of review of non-statutory exercises of government power. Here I have in mind not only exercises of executive power but also the growing modern trend of creating, through the imposition of licensing schemes, governmental decisions which do not have a sufficient statutory basis to make reasoning based on jurisdictional error plausible.

If I might deal with these in turn.

(a) Inability to take account of the seriousness of the breach in determining whether a jurisdictional error has occurred

If one subscribes to the Project Blue Sky view that one is to determine whether Parliament intended that infringement of the statutory requirement should result in invalidity there is logically no room for asking how serious that breach is. The law is either construed as allowing the provision to be breached without consequent invalidity or it is not. There is no room on that view of things for a middle course of asking whether Parliament might have intended different outcomes depending on how serious the breach was. If one wanted to put it in slightly different terms, there is no administrative law equivalent to what contract law calls an intermediate term where one has to wait and see how serious the breach of the term is before deciding whether a right of termination arises.

Yet this is simply not what cases say. In Minister for Aboriginal Affairs v Peko-Wallsend Ltd^[31] Mason CJ explained, in a passage which has been repeatedly applied, that ‘not every consideration that a decision-maker is bound to take into account but fails to take into account will justify the court setting aside the impugned decision and ordering that the discretion be re-exercised according to law. A factor might be so insignificant that the failure to take it into account could not have materially affected the decision’. Now this passage predates the discovery that the grounds of review are all jurisdictional errors and it uses language which suggests that it is the Court which is setting the decision aside rather than merely discovering that the decision was already invalid. But it is accepted doctrine. The only way it may be incorporated into the view that one must discern whether Parliament intended that the breach should result in invalidity is to accept that the Project Blue Sky inquiry requires one to work out what Parliament would have done in any particular case.

Occasionally, this has occurred. In Nature Conservation Council of New South Wales v Minister Administering the Water Management Act 2000,^[32] the Minister had been empowered to formulate water management plans. The Act required the plan to have environmental water rules which were defined in a rather complex way. As a matter of text, the New South Wales Court of Appeal concluded that the Minister’s plan did not include environmental water rules. However, it went on to conclude that as matter of substance it did and that Parliament could not have intended that a plan which, in substance, met the environmental objectives of the Act should be invalid.^[33] Spigelman CJ accepted that the statute pointed to invalidity but said:^[34]

On balance these textual indications would support a conclusion of invalidity. It is, however, my opinion that the factual context of the water source, rather than the textual context of the legislative scheme, is determinative in the present case.

This is a remarkable extension of Project Blue Sky. As Gummow J observed upon the successful special leave application ‘Yes, Mr Hutley. There does seem to be a problem about Project Blue Sky’.^[35] The appeal was subsequently settled.^[36] It is understandable, I think, why the Court of Appeal was tempted down this course. It allows one to assess how serious the breach of the statute was which might seem quite sensible. Within the rubric, however, of Project Blue Sky this can only make sense if one is prepared to seek to divine what the intention of Parliament was in relation to the particular case before the Court. Even for the most diehard proponents of the present theory, this is too much.

After this paper was delivered Neil Williams SC and Katrina Stern were kind enough to draw to my attention the High Court’s decision in Minister for Immigration v SZIZO.^[37] In that case, the Act required notice of a hearing before the Refugee Tribunal to be given to the applicants’ authorized representative within a specified time (ss 425A and 441G). There was no question that the applicants were, in fact, notified of the hearing within that time but it was accepted that it had not been given to the authorized representative. The Act also imposed on the Tribunal substantive obligations akin to the hearing rule which required adverse information to be brought to the applicants’ attention and for them to appear to give evidence and present argument (ss 424A and 425). The Court declined relief on the basis that there had been no denial of natural justice. The reasons of the Court are ambiguous. They may be read to mean that breaches of ss 425A and 441G did not generate jurisdictional error and there had been no breach of ss 424A and 425 (which if breached would have generated such an error). They may also be read to mean that some form of intensity review was intended. The critical paragraphs are [34] and [35]:

[34] In combination, ss 425A and 441G ensure that an applicant for review receives timely and effective notice of the hearing. They impose obligations which facilitate the conduct of a procedurally fair hearing. However, the manner of providing timely and effective notice of hearing is not an end in itself. The procedural steps dealing with the manner of giving notice are to be distinguished from other components of the statutory statement of the hearing rule, including the obligation to give particulars of adverse information [Section 424A(1)] and to invite the applicant to appear to give evidence and to present arguments relating to the issues arising in the decision under review [Section 425].

[35] While the legislature may be taken to have intended that compliance with the steps in ss 441G and 441A would discharge the Tribunal's obligations with respect to the giving of timely and effective notice of the hearing, it does not follow that it was the intention that any departure from those steps would result in invalidity without consideration of the extent and consequences of the departure. The respondents acknowledge that they suffered no injustice by reason of the Tribunal's omission and they do not take issue with the Full Court's characterisation of the result in the circumstances as being "rather absurd" . The admitted absurdity of the outcome is against acceptance of the conclusion that the legislature intended that invalidity be the consequence of departure from any of the procedural steps leading up to the hearing [SZIZO v Minister for Immigration and Citizenship (2008) 172 FCR 152 at 167 [87]]. In a case in which the Tribunal fails to comply with the requirements for the giving of notice of a hearing, the factual determination of whether the applicant for review and his or her authorised recipient received timely and effective notice of the hearing does not require the court to consider how the applicant might have presented his or her case differently had the Tribunal complied with the statutory procedures. No question arises, in the case of an applicant who has received timely and effective notice of the hearing, of the loss of an opportunity to advance his or her case.

I think the former reading is the better one (especially in light of the emphasised sentence in [35]). It remains nevertheless a good example of the gymnastics required to bring about commonsense outcomes in this area.

One sees the same yearning to consider how serious the breach of the rules is in that discordant collection of authorities dealing with whether trivial breaches of the rules of procedural fairness should result in the setting aside of a decision: see, for example, Stead v State Government Insurance Commission.^[38] Often enough the ingenious solution to this unyielding aspect of Project Blue Sky is accept that the jurisdictional error has been shown but then to decline relief on discretionary bases. But this gives rise to its own difficulties to which I now turn:

(b) Discretionary withholding of relief and the void/voidable distinction

One must approach this question with the clear understanding of what Project Blue Sky requires. One asks whether the Parliament intended that a breach of its requirements should result in invalidity and if that receives an affirmative response one concludes that the decision is of no effect and is no decision at all. In so doing, one is obeying only that which the Parliament itself has commanded.

It is nevertheless accepted that the grant of the constitutional writs (as they are now called) is discretionary. Even where jurisdictional error is established the Court may decline to set a decision aside. There are at least two problems with this and both are concerned with coherence.

The first focuses on the doctrine’s apparent loyalty to the will of Parliament. By the time the question of discretionary withholding of relief arises, the Court has already determined there is a jurisdictional error and therefore that Parliament intended the particular decision to be invalid. This is what Project Blue Sky says. The Court, on this view of things, has not decided that the decision should be set aside – rather it has discovered that Parliament never intended that such a decision should be valid. But if one has accepted that Parliament intended decisions of that kind to be invalid how can one, in the same breath, conclude that a court can decline relief? By refusing to set aside the decision one is giving effect to the very thing which one has just concluded Parliament has said must not be given effect. This makes no sense; it is internally inconsistent.

A second aspect of incoherence emerges when one inquires from where it might be that an invalid decision which is not to be set aside derives its lawful effect. Plainly, it does not derive it from the statute – Project Blue Sky supplies that answer. But if not the Parliament, from where else might its legality be derived? The present theory simply provides no answer to this.

When these problems arise retreat is usually beaten to the dichotomy between void and voidable decisions. As I will try to show later, the only coherent theory is one in which all decisions are accepted to be valid and then subsequently set aside when judicial review grounds are established. But that is not the law at the moment. In Minister for Immigration and Multicultural Affairs v Bhardwaj,^[39] the Court was divided on whether administrative action resulting from jurisdictional error was without legal effect. But Gaudron and Gummow JJ (with whom McHugh J agreed) did say that ‘[t]here is, in our view, no reason in principle why the general law should treat administrative decisions involving jurisdictional error as binding or having legal effect unless and until set aside. A decision that involves jurisdictional error is a decision that lacks legal foundation and is properly regarded, in law, as no decision at all’.^[40] It was this precise passage was picked up and later applied by five Justices in Plaintiff S157/2002 v Commonwealth.^[41] Interestingly, one of those five – Hayne J – had expressed the opinion in Bhardwaj itself that administrative action was, in a meaningful sense, but voidable.^[42]

I side with those, such as Professor Aronson, who say that this cannot really be a true description of judicial practice. The reasoning in Project Blue Sky and S157 affords no halfway house in which the legislature has granted to a decision defeasible validity until challenged. I agree that, in practical terms, the rules of standing and the ability of a court to refuse relief on discretionary grounds result in an operation of the present theory which appears to create a situation where decisions are, in fact, treated as valid. But the point to be made is that is an outcome which is not brought about by the reasoning in Project Blue Sky. To the contrary, it is an outcome in defiance of the principles there expounded. The fact that rules of standing and the ability of a court to withhold relief on discretionary grounds have the political consequence that administrative action is voidable is the strongest reason we have to believe that the statements in Project Blue Sky and S157 cannot possibly be correct statements of theory. They simply do not match the observable facts about judicial review.

There are other matters of everyday practice which contradict the theory, too.

(c) the availability of certiorari in cases of non-jurisdictional error of law patent on the face of the record and in cases where governmental power is being exercised without a statutory basis

The Project Blue Sky approach neatly locates judicial review as the search for jurisdictional error. It does this for doctrinal reasons associated with Dicey’s views about the supremacy of Parliament. On this view of things, if a decision is authorised by statute then no common law, or judge-made rule, may justify is annulment. To do so would be to contradict the will of Parliament itself for what Parliament has authorised no judge may set aside. The Project Blue Sky/S157 view purports to accept this orthodoxy but then seeks evade it by reimagining the whole of administrative law as an outcrop of statutory interpretation. I have outlined already the internal logical difficulties this entails.

But there are not just problems of internal logic. There are external problems, too, of reconciling what the theory predicts about the behaviour of judges and what, in practice, judges do. The previous matter I touched upon – the necessity of the theory to predict that decisions are void rather than voidable but the fact that relief may nevertheless be refused on discretionary grounds – is an example of the empirical poverty of the Project Blue Sky approach. There are other such problems. The ultra vires view predicts that at least two grounds of review do not exist when, in fact, they do.

The first, which will be well-known to you, concerns the grant of certiorari for error of law on the face of the record. This is a remedy which may be granted without the demonstration of jurisdictional error. To be plain: the Court may conclude that the decision was within jurisdiction but nevertheless set it aside. There is no way this principle can be integrated with notions of jurisdictional error or with the view that Courts may not set aside that which Parliament has authorised. Worse still, it demonstrates that the Courts do exercise a power to set aside decisions which are, ex hypothesi, authorised by statute thereby directly contradicting the Diceyan view of the role of Parliament. The existence of certiorari for non-jurisdictional error is an empirical fact which needs to be explained. It is not explained by asserting that all power in respect of administrative decision-making is located in Parliament.

The High Court’s response to the existence of this anomaly has not been to revisit whether judicial review rests on issues of jurisdiction and Parliamentary intent. Instead, it has been to confine the anomaly. This was the inevitable result of the holding in Craig v South Australia that the record did not include the reasons given for the decision.^[43] I do not doubt that there was historical pedigree for that proposition but it was surely a surprising outcome towards the end of the twentieth century. The real reasons for it lie, I venture to suggest, in a hostility to allowing there to exist a procedure by which common law courts might set aside decisions otherwise authorised by statute.

The more recent role of this remedy has been to serve as the lynchpin in Kirk v Industrial Court of NSW.^[44] Six Justices said ‘[t]he continued vitality of the principle that certiorari for error of law on the face of the record may seem incongruous’.^[45] But this was not so. Their Honours saw for it a very important role indeed. Having discovered that review for jurisdictional error was an inviolable function of State Supreme Courts, the question arose as to what was to be made of the large number of judicial statements to the effect that State Parliaments could enact valid ‘no certiorari’ clauses. The answer the Court provided was that these cases were concerned with certiorari for error of law patent on the face of the record and not with certiorari for jurisdictional error.^[46] The emasculated remedy of Craig was called back, for its final bow, to take the fall for ‘no certiorari’ clauses.

In my opinion, this reasoning is unconvincing. It rests, with respect, on the proposition that the mere reference to the Supreme Courts of the States in s 73 of the Constitution carries with it not only the idea that such a Court must exist but that such a court must possess the power to review for jurisdictional error. The objections to this are legion. It may be that the Supreme Courts cannot be abolished and that their existence is guaranteed by their presence in s 73 (although I doubt this is truly necessary). But if one is to reason that way, one needs to account for the reference in s 73 to the Inter-State Commission, the resoundingly ontological words of s 101 (‘There shall be an Inter-State Commission’) and the fact that the Inter-State Commission most assuredly does not exist. One might have thought that, where the Constitution requires that the bodies referred to in s 73 exist, it has used express language of the kind in s 101. But one does need to reach a concluded view about that.

What is much more remarkable is the reasoning that merely by s 73’s reference to the Supreme Courts that there emerges a constitutional guarantee that a Supreme Court has supervisory jurisdiction for jurisdictional error. This is an example of what David Jackson has referred to as a ‘structural implication’.^[47] Examples of structural implications include the separation of powers arising from the fact that the legislative, executive and judicial powers are dealt with in separate chapters of the Constitution (I, II and III), the continued existence and functioning of the States (Melbourne Corporation)^[48] and, perhaps, the maintenance and existence of the rule of law (The Communist Party Case).^[49] The Court’s rationale in Kirk was that, if Supreme Courts did not have a supervisory jurisdiction over jurisdictional error, this would ‘create islands of power immune from supervision and restraint’^[50] and this would ultimately interfere with ‘the superintendence of [the High Court] as the "Federal Supreme Court" in which s 71 of the Constitution vests the judicial power of the Commonwealth’.^[51]

This makes no sense, however. The role given to the High Court by s 73 is as the ultimate appellate court for the States and the Commonwealth. That role, as it has frequently been explained, is about achieving a single common law for Australia.^[52] It is not necessary that, for an ultimate appellate court to discharge its function of supervising the courts below it, those courts should in turn have guaranteed jurisdictions. Section 73 says nothing, in terms, or by way of necessary implication, about jurisdictional excess by the executive decision-makers of the States.

It is beyond the scope of this paper to point out the other difficulties with Kirk (for example, erecting in effect a separation of powers at the level of the States without overruling a line of cases denying there was any such principle); or, proceeding on an arguably selective reading of what the cases at the turn of the nineteenth century said about the authority of superior courts in prerogative relief cases). For present purposes it suffices to note only the striking proposition that the relevance of certiorari for error of law patent on the face of the record is to explain why ‘no certiorari’ clauses are valid. This is reasoning which, in my opinion, should be rejected.

Despite all of that, the remedy continues to exist. It does not sit easily, or perhaps at all, with the approach in Project Blue Sky.

Another difficulty for the Project Blue Sky approach concerns those cases where judicial review is exercised over decisions which have no direct statutory basis. There are two kinds. The first concerns the executive power. The second concerns a growing modern tendency to create decision-makers with government functions other than through the means of statute.

As to the first, it has long been accepted that some exercise of prerogative power of an administrative nature are judicially reviewable. In the UK this was established in GCHQ^[53] and in this country by Re Toohey; ex parte Northern Land Council;^[54] FAI Insurances Ltd v Winneke;^[55] and South Australia v O’Shea.^[56] On their face these decisions cannot be explained by resort to jurisdictional error unless they are seen as statements about statutory powers.

More recently, however, it has been said that at least at the federal level the scope of the executive power is a question of the proper construction of s 61 of the Constitution and, in particular, the words ‘executive power’.^[57] That opens the doorway to putting judicial review of executive action on a jurisdictional error basis. The problem is that the doorway is difficult to get through gracefully. Under the current doctrine of the Court, the principles of judicial review are implications from a statute, a corollary of the principle of legality. In order to put judicial review of executive power on the same basis it would be necessary to say that the same principles are implications from the Constitution itself but this, so far as I am aware, is not something anyone has yet suggested.

Further, if it really could be inferred from the Constitution that executive decisions were to be made in accordance with ordinary principles of judicial review, then it would not really be necessary to the say that those same principles arose in cases involving statutes as a matter statutory implication. To put the matter another way, if Chapter II contains an implicit set of judicial review principles (no doubt amenable to statutory alteration like the executive power more generally), why would they not apply to action authorised by statute (unless the statute otherwise provides)?

Quite apart from reasoning based on s 61, of course, there remains a difficulty in explaining how State executive power is to be judicially reviewed if the doctrine is based in jurisdictional error. No resort to s 61 or any equivalent is there available. It would seem that what must occur is the direct application of GCHQ, that is, judicial review without jurisdictional error in the Project Blue Sky sense.

There other kinds of decision which lack a statutory basis and judicial review of which may be difficult to explain on the basis of jurisdictional error. In R v Panel on Take-overs and Mergers; ex parte Datafin Plc,^[58] the Court of Appeal for England and Wales held that the Take-overs Panel, although a private body, was amendable to judicial review because it was exercising ‘public law functions’.^[59] This decision cannot be explained if one accepts that judicial review is an aspect of statutory interpretation. Unsurprisingly one finds considerable reluctance in this country to apply it. Apart from Kirby J in NEAT Domestic[60] and Breckler,^[61] the High Court has yet to refer to it. Basten JA recently very thoroughly traced the status of Datafin in Chase Oyster Bar and concluded^[62] that it was not necessary to determine whether Datafin was part of Australian law (since he concluded in that case that a statutory function was being performed). But he did think that nothing said in the High Court required its application and ‘statements of general principle in that Court might be thought to adopt a more limited scope for the operation of public law remedies’. With respect I agree. The whole tenor of Australian public law is away from any notion that judicial review can exist other than through jurisdictional error.

But this will lead to a serious lacuna. There are a number of regulatory schemes in this country in which a regulator imposes a licence condition upon a class of person which requires the licence holder to belong to an external complaint resolution body and to abide by that body’s arbitral decisions. Lord Hoffman referred to these kinds of situation as involving the ‘privatisation of the business of government itself’.^[63] There are first instance decisions in this country which have held that such exercises of power are judicially reviewable. Matthews J accepted that the Advertising Standards Board was exercising governmental functions and was judicially reviewable in Typing Centre of New South Wales v Toose^[64] and Ryan J did the same in Dorf Industries Pty Ltd v Toose.^[65] Campbell J thought that this was an open view in McLelland v Burning Palms Surf Live Saving Club^[66]. I doubt whether these statements can survive what Basten JA has said in Chase Oyster Bar. But assuming that to be correct, it means that there will now exist the potential for regulators to create dispute resolution schemes through the imposition of licence conditions requiring both membership of the scheme and adherence to its decisions without the possibility of judicial review. It is striking that in such cases were the government to make the decision itself it would be judicially reviewable but by moving it out to private contractors this can be circumvented. Those who have followed the cases concerned with the offshore processing regime may notice certain resonances.

The consequences seems to me that the prevailing view does not adequately explain judicial review of executive action and it either does not explain how privately reached governmental decisions may be reviewed or, worse, it denies that they can. Neither outcome is satisfactory.

III. The need for a different approach

Having set out some of the difficulties flowing from Project Blue Sky and the concept of jurisdictional error, may I suggest a different way of looking at matters. In that regard, it is useful to begin with six facts about the actual the practises of judges. These are:

(i) the Courts regularly decline relief even where jurisdictional error is established. In practice, whatever may be said in S157, administrative action is in practise voidable, not void;

(ii) the Courts do not merely declare that decisions are invalid (as S157 might suggest) but instead actively set them aside. If the decisions were truly invalid there would be nothing to set aside;

(iii) the rules governing judicial review are created by the judiciary and applied by it;

(iv) the judiciary does review the executive power on judicial review grounds and either does, or ought to be able, to review non-statutory decisions having a governmental aspect to them;

(v) the Courts can and do set aside decisions by certiorari which are authorised by statute even when they are within jurisdiction. They do so when there is a non-jurisdictional error of law patent on the face of the record. The fact therefore that Parliament has authorised an official to make a decision does not, in practice, exclude the ability of the courts to review it; and

(vi) Ouster clauses are ineffective to prevent judicial review.

I have not included any statements about doctrine in these facts. They are intended only as claims about observable facts. I emphasise the facts as being observable because one ought to keep in mind Sir Karl Popper’s advice that theories are not right or wrong but instead are simply better or worse at explaining the observable evidence.^[67]

The theory inherent in Project Blue Sky does not coherently explain facts (i)-(iv) at all. It predicts that decisions are void when they are in fact voidable. It cannot explain why the Courts set aside decisions when, if the theory be correct, there is nothing to set aside and a declaration should suffice. It is forced to say that the author of the principles of judicial review is the Parliament when it is demonstrably the Courts. It cannot explain how the rules of judicial review can evolve over time but must accept instead that the principles which apply to any particular statute are those which exist at the time of its passage. Any attempt to outflank that problem involves giving statutory effect to the common law which the Court has previously explained infringes the separation of powers.^[68] The doctrine either cannot explain, or denies the existence of, judicial review of non-statutory governmental decisions.

On the other hand, to give credit where credit is due, the doctrine does give an explanation for the last two facts. Certiorari for error of law on the face of the record exists as a foil to certiorari for jurisdictional error explaining why privative clauses have previously been held to be valid.^[69] This important jurisprudential purpose for that species of certiorari is to be contrasted, however, with the doctrine’s practical emasculation by its confinement to those cases where the error of law can be found on the face of a record which we are told does not include the actual reasons articulated for the decision’s making.^[70]

The point of that reasoning is to provide the basis for the fifth observation: here we see that ouster clauses are not effective because all the grounds of judicial review have turned out to involve jurisdictional error and against that kind of error ouster clauses are impotent.^[71]

My suggestion to you is that the failure to fit the first four facts and the artificial reasoning inherent in the explanations for the last two are powerful indicators that this is a theory which is wrong. It does not describe the reality of actual judicial practice and it is internally inconsistent. Its sole redeeming feature is that it does provide an explanation for the ineffective nature of ouster clauses.

What would be a better theory? Clearly it would have to explain the six matters above. I would propose the following (as others such as Paul Craig have done previously):

(a) the author of administrative action is the executive arm of government or those doing the executive arm of government’s business for it (delegates, if you will);

(b) the executive may make those decisions within its own competence (that is, under what were formerly known as prerogative powers) or under functions given to it by Parliament or its delegates;

(c) in making administrative decisions the executive or its delegates are constrained by two distinct bodies of law with different sources, the common law and statute. The principles of judicial review are part of the common law;

(d) the relationship between the common law and statute is well understood. The law of the judges exists as a distinct body of law deriving from the operation of the doctrine of precedent in circumstances where the Parliament has not yet spoken. It is not statutory in origin. But the Parliament is at liberty at any time to override the rules of the common law if it chooses by the use of clear and unambiguous language;

(e) the fact that a function is conferred by Parliament on the executive does not entail that the common law principles of judicial review have been ousted. As the existence of certiorari for error of law patent on the face of the record shows, the power of the courts to set aside decisions is not thwarted because the function in question is statutory in origin;

(f) it is for the Courts to decide whether any particular administrative decision is lawful, that is, to determine whether it infringes either the common law or any statutory requirement;

(g) if a Court concludes that a decision has contravened a relevant legal principle, whether statutory or deriving from the common law, the Court may, but need not, set the decision aside. That will involve the exercise of a discretion; and

(h) until the decision is set aside it remains for all purposes valid.

This theory explains all the factual matters that I have previously set out but not the status of ouster clauses. Such clauses require additional explanation. Jeremy Kirk has expressed the view that the Court’s learning on this topic took a wrong turn in Hickman^[72] and that the position should simply be that privative clauses are invalid.^[73] In my respectful opinion he is correct. If one accepts that it is the role of the judiciary to decide what is lawful and what is not^[74] that function would be altogether pointless if it was not also accompanied by an ability to say what the consequences of the fact that the conduct was unlawful were to be. If it be accepted that an indispensable aspect of the judicial power is to declare the consequences of unlawful governmental action, it is very difficult to see how Parliament may intrude upon that function by seeking to pre-empt what the outcome of the Court’s eventual inquiry should be. Just as we deny any role to the Parliament in findings of guilt^[75] so too we ought to affirm that ouster clauses involve an impermissible exercise of judicial power. If one accepts that, the simple answer to the private clause problem is the one that Kirk suggests, namely, that privative clauses are invalid.

If one alights upon that solution then the theory expounded above explains the empirical evidence about how Courts are actually behaving and provides, I hope, a tidy answer to the problem of ouster clauses. It also has the virtue of allowing a Court to consider, in deciding whether to set aside a decision which is unlawful, how serious the breach of the law is together with other factors which do not turn purely on the text of the statute. This is a significant improvement on the current doctrine which secretes such considerations in the interstices of discretionary refusal of relief. Such a view, if adopted, would allow explicit reasoning on why particular unlawful decisions are to be set aside and others left in place. We would have no further need of bewildering and unpredictable lists of jurisdictional errors such as those usefully compiled by Professor Aronson. No doubt, in that jurisprudence the text of the statute - including its scope, ambit and purpose - would be very important inputs but there is no reason to think that they would be, as they are now, the sole inputs.

An inevitable feature of that theory is that there is no role in it for invalid administrative action. This is no great loss. Hans Kelsen explained in his Pure Theory of Law that a legal order cannot contain decisions which were truly invalid.^[76] The reason for this was that such a question could only arise within the legal system itself before a body which was in fact authorised to decide that question. That there was a body authorised to decide whether the decision was valid (and with it the possible outcome that the decision was valid) contradicted the ability of the decision truly to be wholly without legal effect. It has at least the legal effect of being the decision that the court was being asked to consider. That the legal system could ask itself whether a decision was invalid showed that it could not truly be wholly invalid. Kelsen put it this way:^[77]

In this respect, the law is like King Midas: just as everything he touched turned to gold, so everything to which the law refers assumes legal character. Within the legal order, nullity is only the highest degree of annullability.

Hayne J made much the same point in Bhardwaj when he pointed out that the view that invalid action was wholly void was at a loss to explain how such a decision might be appealed or subject to judicial review.^[78] In Bhardwaj he accepted that decision would be ‘valid and effective in law unless and until it is retrospectively invalidated (i.e. declared "void") by a court’.^[79] Much the same reasoning may also be seen at work in decisions such as Calvin v Carr.^[80] It is difficult to integrate those statements into the extreme doctrine now annunciated in S157. It is even more difficult to explain, on the Project Blue Sky view of the world, where this defeasible validity comes from. It cannot come from the Parliament and it does not come from the courts.

This is a difficult area which has largely been driven by the privative clause in the Migration Act 1958 (Cth). For the reasons I have tried to explain above I do not think the current learning of the High Court is, with respect, a coherent or accurate account of administrative law. Despite the considerable array of authorities now supporting it in this country, we should seriously consider whether we wish to persist with this programme. In times past, there was a belief that the world was flat and that the sun went around the earth. Gradually, observations were made which contradicted this: the fact that the masts of ships disappeared over the horizon and the fact that it was much simpler to explain the motion of the planets if one assumed that the sun was at the centre of the solar system with the earth in orbit around it. I wonder whether our public law and its commitment to the Diceyan view of Parliament as the source of all principles of judicial review may not be due for a similar reconsideration.