NZLII Home | Databases | WorldLII | Search | Feedback

University of Otago Law Festschrifts

You are here:  NZLII >> Databases >> University of Otago Law Festschrifts >> 2016 >> [2016] OtaLawFS 16

Database Search | Name Search | Recent Articles | Noteup | LawCite | Download | Help

Rodriguez Ferrere, M B --- "John Smillie, judicial review and New Zealand's missed opportunity" [2016] OtaLawFS 16; The Search for Certainty: essays in honour of John Smillie 180

Last Updated: 31 May 2019



M. B. Rodriguez Ferrere*

My experience of Professor John Smillie as an undergraduate at the University of Otago was restricted to the realm of private law. He taught me the law of defamation and intellectual property, such that lessons on Lady Karen Hadlee’s love-life,1 a more perfect toilet seat,2 and the unappealing nature of an egg-flip cocktail3 have endured in my mind a decade hence. Later, I worried whether his views on judges’ clerks would engender animosity given my early career choice,4 but I need not have worried: he was (and remains) a generous and warm colleague.

It was only after I joined the Faculty that I became aware that Smillie’s original area of expertise was in administrative law. It was the focus of his thesis at Yale5 and occupied his teaching and research agenda from his appointment as lecturer to Otago in 1973 until 1985.6 In this paper, I want to explore his contribution to administrative law in New Zealand and, specifically, his proposal for the development of judicial review over thirty years ago. I will argue that had New Zealand adopted this proposal, this jurisdiction might have avoided some of the problems it currently faces in administrative law. The paper is in two parts. The first details and examines Smillie’s approach and the second identifies the consequences of New Zealand striking a different path.


Perhaps unintentionally, Smillie outlined his approach to judicial review in New Zealand in two articles from 1980 and 1984. Read in concert, they provide a clear path for New Zealand to develop an indigenous, coherent approach to judicial review. In this part I want to provide some context of the period in which Smillie was writing before outlining his approach.

Lecturer, Faculty of Law, University of Otago.
1 New Zealand Magazines Ltd v Hadlee (No 2) [2005] NZAR 621 (CA).
2 UPL Group Ltd v Dux Engineers Ltd [1988] NZCA 244; [1989] 3 NZLR 135 (CA).
3 Warnink v J Townsend & Sons (Hull) Limited [1979] AC 731 (HL).
4 John Smillie “Who Wants Juristocracy?” [2006] OtaLawRw 2; (2006) 11 Otago LR 183 at 192.
5 John Smillie “The content of the audi alteram partem rule: the nature of a fair administrative

hearing” (JSD Thesis, Yale University, 1972).

  1. I am indebted to the institutional knowledge of Emeritus Professor P D G Skegg who provided
    me with this information.

1.1 Judicial Review of Administrative Action in the 1980s

In February 1986, the Legal Research Foundation held a conference on judicial review that attracted over 280 registrants, a remarkable fact in itself given that it would be difficult to attract a similar number today.7 The contributors – including Sir Gerrard Brennan, Michael Kirby and Rt. Hon. Sir Robin Cooke – were obviously a draw card for such a large number of attendees, but I would hazard that the main reason was instead its area of focus: administrative law was the “legal growth industry of the 1980s.”8

Administrative law and judicial review were in a constant state of flux during the 1980s, and change was in the air at this conference. As Lord Wilberforce noted in the foreword to the collection of the contributions to that conference, since the 1940s, “the pendulum has swung, with accelerating élan, in favour of judicial control of the executive and a widening range of decision-makers”, but that a swing in the other direction was now due.9 Certainly, there was a plethora of decisions in the years preceding the conference that made administrative law a hotbed of debate. Less than two years prior, the House of Lords delivered its ground-breaking decision in Council of Civil Service Unions and Others v Minister for the Civil Service,10 where in one paragraph, Lord Diplock removed the blanket immunity from review that applied to prerogative powers and established the now-familiar tripartite classification of the heads of review under illegality, procedural impropriety and irrationality.11 Only a few months later, Lord Roskill accepted in Wheeler v Leicester City Council12 that procedural impropriety was not limited to strictly procedural issues and could arise when the decision-maker was pursuing a coercive object.13 Three months earlier, the same issue – a rugby club proposing a tour of apartheid South Africa – led to the New Zealand Court of Appeal relaxing the requirements of standing in order to determine whether the New Zealand Rugby Football Union was acting within its lawful powers to allow the national team to embark on such a tour.14

7 John Smillie “Introduction” in Michael Taggart (ed) Judicial Review of Administrative Action

1980s: Problems and Prospects (Oxford University Press, Auckland, 1986) at xi.
8 Eric Barendt “Grievances, Remedies and the State” (1987) 7 OJLS 125 at 125.
9 Lord Wilberforce “Foreword” in Michael Taggart (ed) Judicial Review of Administrative Action

1980s: Problems and Prospects (Oxford University Press, Auckland, 1986) at ix.
10 Council of Civil Service Unions and Others v Minister for the Civil Service [1985] AC 374 (HL).
11 At 410.
12 Wheeler v Leicester City Council [1985] UKHL 6; [1985] AC 1054 (HL).

13 Rt. Hon. Sir Robin Cooke, “The Struggle for Simplicity” in Michael Taggart (ed) Judicial Review of Administrative Action 1980s: Problems and Prospects (Oxford University Press, Auckland, 1986) 5 at 10.
14 Finnigan v New Zealand Rugby Football Union Inc [1985] NZHC 102; [1985] 2 NZLR 159 (CA).

This only skims the surface of the many developments that occurred in this period, making it fair to say that it was an exciting time to be an administrative lawyer. Only a year before his judgment in C.C.S.U., Lord Diplock, this time for the Privy Council, agreed with the Court of Appeal that Royal Commissions of Inquiry were justiciable, but also noted that: 15

The extension of judicial control of the administrative process has provided over the last 30 years the most striking feature of the development of the common law in those countries of whose legal systems it provides the source ... It has not yet become static either in New Zealand or in England.

Smillie’s contributions to administrative law jurisprudence must be viewed with this lens: he was working in one of the most dynamic areas of the common law at the time. The dynamism of the early 1980s, however, can be traced back to an earlier period, where a major shift occurred in the House of Lords’ 1968 decision in Anisminic v Foreign Compensation Commission.16 This “landmark judgment”17 caused a “revolution”18 in administrative law, although its implications would not be realised in New Zealand until over a decade later in the decision of Bulk Gas Users Group v Attorney-General19.20 It is Smillie’s responses to Anisminic in 198021 and Bulk Gas in 198422 that provide his vision for the scope of, and basis for, judicial review in New Zealand and which are the focus of my analysis.

1.2 Smillie’s Response to Anisminic

Smillie’s 1980 analysis of Anisminic was not the first time he commented on the case, and he indicated in 1969 that the decision conferred “wide powers” that meant “any error of construction ... materially affecting the ultimate decision is now capable of resulting in any one of a number of jurisdictional errors.”23 As he

15 Mahon v Air New Zealand [1984] AC 808 at 816 (PC).
16 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL).
17 Paul Daly “Judicial review of errors of law in Ireland” (2006) 41 Irish Jurist 60 at 62.

18 Philip Joseph “The demise of ultra vires - judicial review in the New Zealand courts” (2001) Public Law 354 at 359.
19 Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA).

20 Sian Elias “Righting Administrative Law” in David Dyzenhaus and others (eds) A Simple Common Lawyer: Essays in Honour of Michael Taggart (Hart Publishing, Oxford, 2009) 58.

21 John Smillie “Judicial Review of Administrative Action – A Pragmatic Approach” [1980] OtaLawRw 2; (1980) 4 Otago LR 417 (hereafter ‘Smillie I’).

22 John Smillie “The Foundation and Scope of Judicial Review: A Comment on Bulk Gas Users Group v Attorney-General[1984] OtaLawRw 5; (1984) 5 Otago LR 552 (hereafter ‘Smillie II’).

23 John Smillie “Jurisdictional Review of Abuse of Discretionary Power” (1969) 47 Can Bar Rev 623 at 639.

noted a decade later, his original analysis remained accurate.24 In less technical terms, Anisminic’s “overall effect was to make any misconstruction of any statute fatal to the adjudication of any decision-maker.”25

However, perhaps the most dramatic and immediate effect of this shift was to neuter the effect of privative or ouster clauses. In Anisminic, the Foreign Compensation Commission had declined the appellant’s application for compensation after it suffered losses in the Suez Crisis of 1956. The Commission’s parent statute held that “the determination by the commission of any application made to them under this Act shall not be called in question in any court of law.”26 The respondents argued, in accordance with prevailing orthodoxy, that since the determination of the Commission to decline the appellant’s application was prima facie valid (because it was one within the Commission’s jurisdiction to make), this privative clause prevented any further questioning of the determination: the Court could not, and need not, look behind it. Conversely, the appellants argued that the clause only applied to ‘real’ determinations, and could not apply to those determinations which were in fact invalid (because the Commission acted outside of their jurisdiction at arriving at this determination), even if they appeared valid prima facie.27

Their Lordships gave judgment for the appellants, with the majority holding that the clause did not prevent the Court from examining whether the Commission had acted within its jurisdiction in arriving at its final decision. This was, as De Smith notes, a “breakthrough” since it rejected the idea that a decision-maker’s jurisdiction was determined at the outset of its inquiry.28 Instead, the decision-maker could lose that jurisdiction by embarking on an inquiry outside its jurisdiction or arriving at an outcome it did not have the jurisdiction to make.29 Most importantly, however, their Lordships held that: 30

... in the intervening stage, while engaged on a proper inquiry, the [decision-maker] may depart from the rules of natural justice; or it may ask itself the wrong questions; or it may take into account matters which it was not directed to take into account. Thereby it would step outside its jurisdiction. It would turn its inquiry into something not

24 Smillie I, above n 21, at 417.
25 Daly, above n 17, at 62.
26 Foreign Compensation Act 1950 (UK), s 4(4).
27 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL) at 170.

28 Rt Hon Lord Woolf and others (eds) De Smith’s Judicial Review of Administrative Action (7th ed, Sweet & Maxwell, London, 2013) at [4-030].
29 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL) at 195.
30 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL) at 195 per Lord Reid.

directed by Parliament and fail to make the inquiry which Parliament did direct.

Thus, there were myriad ways that a tribunal could lose its jurisdiction and arrive at a determination that was thus a nullity, and one to which privative clauses of the sort in Anisminic simply did not apply. And since it was the Court that determined whether the questions the tribunal asked itself or the matters it took into account were the wrong ones, Smillie accurately notes that “Anisminic provided the conceptual tools by which a reviewing judge can, if he wishes, describe any error by a tribunal in terms of a jurisdictional error which renders the resulting decision ultra vires and a nullity.”31 This is why Smillie originally stated that the decision conferred “wide powers” to a reviewing Court: after Anisminic, “all material errors of law were treated as ‘jurisdictional.’”32 Indeed, in the years immediately following the decision, judges and academics found it “extremely difficult” to identify non-jurisdictional errors.33 This is notwithstanding the fact that Lord Wilberforce in Anisminic held that there was a “crucial distinction” between a tribunal making an error but remaining within its jurisdiction and committing an error that indicated it was acting outside that jurisdiction.34 As even the counsel for the appellants had noted in argument in Anisminic: “Not every wrong act or omission will lead to the final conclusion being a nullity.”35

The problem was, however, determining which wrong acts or omissions were sufficiently “material” to become jurisdictional. In Smillie’s view, it was a problem without a solution: “there is no basis for distinguishing between jurisdictional and non-jurisdictional errors which is capable of fair and consistent application in practice.”36 Although the United Kingdom had developed a number of different methods that the Court could employ to identify and distinguish jurisdictional errors, each were either open to manipulation by the Court37 or ran contrary to the spirit of Anisminic.38

New Zealand’s limited experience by 1980 in applying Anisminic – it had only been properly considered on two occasions – appeared similar.39 Justice Cooke (as

31 Smillie I, above n 21, at 417, emphasis original.
32 Joseph, above n 18, at 359.
33 Smillie I, above n 21, at 417.

34 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL) at 210.
35 Anisminic v Foreign Compensation Commission [1968] UKHL 6; [1969] 2 AC 147 (HL) at 165.
36 Smillie I, above n 21, at 421.

37 At 421-422, referring to Geoffrey Lane LJ’s judgment in Pearlman v Keepers and Governors of Harrow School [1978] EWCA Civ 5; [1979] QB 56 (CA).
38 At 422, referring to R v Secretary of State for the Environment, Ex p Ostler [1976] EWCA Civ 6; [1977] QB 122.
39 At 423.

he then was), in adopting Anisminic, attempted to provide guidance for classifying non-jurisdictional error in Attorney-General v Car Haulaways (NZ) Ltd.40 The guidance was accepted by the Court of Appeal with apparent approval, although it overturned his decision.41 However, the guidance suffered from the same vagueness that bedevilled the United Kingdom approaches that Smillie referred to. In Car Haulaways, Cooke J stated that where “an Act plainly empowers an authority [...] to decide a question of law conclusively” any error of law in deciding such questions will be non-jurisdictional.42 However, he did not provide any guidance as to when an Act will “plainly empower”; there was no indication as to when legislative intent should be read in this manner.43 He was similarly vague with terms such as “true jurisdiction” and inaccurate with regards to factual determinations never leading to jurisdictional error.44 Just as in Anisminic, statements of principle that appear clear on the surface become difficult in application.

Justice Cooke, having now been elevated to the Court of Appeal, had his second attempt at the issue in the Engineers’ Union case.45 The case dealt with the effect of a wide and forceful privative clause that prevented the Court from questioning determinations of the Court of Arbitration.46 While the Court of Appeal held that the appellants had not proved the errors of law allegedly committed by the Court of Arbitration, the Bench made some obiter comments relating to the effect of such a wide private clause in the face of Anisminic.47 While McCarthy P and Richmond J essentially both held that the wider approach to jurisdictional review favoured by Anisminic was not applicable in this instance, Cooke J took a different view.48 Smillie argues that he instead reiterated his preference for the wider Anisminic approach and refined his view in Car Haulaways, namely “the courts should be slow to give effect to statutory provisions which purport to exclude or limit the courts’ inherent power to review a tribunal’s interpretation

40 Attorney-General v Car Haulaways (NZ) Ltd SC Auckland A9/73, 8 August 1973.

41 Attorney-General v Car Haulaways (NZ) Ltd [1974] 2 NZLR 331 (CA); Smillie I, above n 20, at 423.

42 Smillie I, above n 21, at 423.

43 At 424.

44 At 424-425.

45 New Zealand Engineering etc Industrial Union of Workers v Court of Arbitration [1976] 2 NZLR 283 (CA).

46 Industrial Conciliation and Arbitration Act 1954, s 47: “Proceedings in the Court shall not be impeached or held bad for want of form, nor shall they be removable to any Court by certiorari or otherwise; and no award, order, or proceeding of the Court shall be liable to be challenged, appealed against, reviewed, quashed, or called in question by any Court on any account whatsoever”.

47 Smillie I, above n 21, at 428.

48 At 428-429.

of its own empowering instrument.”49 However, Smillie argues that Cooke J once again provided “inaccurate and misleading” guidance by arguing that “questions of fact or discretion” were in a “different category” and thus could not amount to jurisdictional error.50 As Smillie notes, “there is no such thing as a completely unfettered, unreviewable discretionary power”, and thus when an official fails to comply with the restrictions on his discretion, he commits a jurisdictional error.51 Justice Cooke’s second attempt at solving the problem caused by Anisminic was an apparent failure, leading Smillie to conclude: “that all existing attempts to find a sensible and workable conceptual basis for distinguishing between jurisdictional and non-jurisdictional errors have failed.”52

Smillie’s solution to this apparently unsolvable problem was a simple one: the abolition of the distinction between jurisdictional and non-jurisdictional error. His conclusion that this was the only viable solution was not an original one, and it was based on the observations of Lord Denning MR in the Court of Appeal decision of Pearlman v Keepers and Governors of Harrow School.53 Lord Denning held that Anisminic had made the distinction between jurisdictional error and non-jurisdictional error so fine and manipulable that essentially, it was the Court’s choice whether to intervene in a decision or not, and simply reverse-engineer the jurisdictional analysis to reflect and justify that choice.54 On this basis, Lord Denning proposed that traditional distinction, crippled by Anisminic but still existent, be abandoned.55 It was a proposal that would languish, since the case was never appealed to the House of Lords, but one which Smillie would adopt and refine.56

Given the attempts to find a basis for distinguishing between the two types of error were “essentially sterile academic exercises”,57 Smillie’s concerns about the distinction remaining in place were twofold. First, it was unnecessary: it bore little practical relevance to why and how courts exercised their powers on judicial review;58 its maintenance only served to satisfy theoretical constraints on the courts’ power that they had long transgressed without incident. Second, and perhaps

49 At 429.

50 At 429-430.

51 At 430-431.

52 At 432.

53 Above n 37.

54 Smillie I, above n 21, at 420, quoting Lord Denning in Pearlman v Keepers and Governors of Harrow School [1978] EWCA Civ 5; [1979] QB 56 (CA) at 70.
55 At 420.

56 At 420.

57 At 435.

58 At 421 and 435.

more importantly, it was undesirable, since it cloaked the Court’s essentially unbridled discretion about whether to intervene in a decision with the legitimacy of a conceptually rigorous test. The attention of academics, counsel and judges ought instead to be on the policy considerations that determine whether and why a Court should intervene, rather than the artificial distinction that obscures those practical realities of that process.59 By focusing on those policy considerations, actors within the study and practice of judicial review could determine factors that would influence the willingness of a Court to intervene, thus providing a more concrete and predictable basis for intervention than the moribund distinction.

Smillie himself proposed seven factors that could influence the depth of scrutiny by a reviewing Court:

59 At 421.

60 At 436.

61 At 438.

62 At 439.

63 At 439.

64 At 443.

65 At 444.

66 At 444-445.

Each of the factors are indications from Parliament of how closely the Court should supervise a particular decision maker, and the proper role of a reviewing judge is to apply the factors to the facts at issue, balance them, and arrive at the appropriate level of scrutiny.67 In doing so, the Court is engaging in a far more transparent process compared to the opacity of declaring an error ‘jurisdictional’ through some alchemic method.

These factors were, however, only one step of the process in consigning the jurisdiction distinction to history. If, after balancing the factors, the Court decided to intervene, it was easy to characterise why that intervention was necessary. To paraphrase Lord Denning in Pearlman: no decision-maker has the jurisdiction to make an error of law on which the decision depends. If it makes such an error, it is acting outside its jurisdiction and the Court may intervene.68 The problem, however, is how to describe the obverse situation. Where a Court, after balancing the factors, decides it is inappropriate to intervene, how is it to explain why an error committed by the decision-maker did not warrant its intervention? The tendency, as Smillie notes, would be to collapse back on the jurisdictional/non-jurisdictional distinction and hold that the error is ‘non-jurisdictional’, which simple entrenches the extant issue.69 The solution to this problem, however, is not to adopt a new vernacular. Instead, Smillie argues that there are a series of “avoidance devices” that, much like the factors justifying intervention, provide a far more descriptive and transparent rationale for why the Court should not intervene. Those ‘devices’ include:

67 At 436.
68 Pearlman v Keepers and Governors of Harrow School [1978] EWCA Civ 5; [1979] QB 56 (CA) at 70.
69 Smillie I, above n 21, at 447.
70 At 447.
71 At 448.

did not ignore it) but simply gave it little weight, which cannot justify the Court’s intervention;72

There is, of course, a risk of manipulation by the Court of such “avoidance devices” to unjustifiably refuse to intervene in a decision. Smillie notes, for example, that it would be dangerous for the Court to use the ‘causation’ argument to justify nonintervention when there has been a breach of natural justice by a decision-maker, and that such errors are always jurisdictional.75 However, they remain significantly preferable to simply categorising an error as non-jurisdictional, which has a greater risk of manipulability, given the reasons for such a categorisation are inconsistent at best, opaque and unjustified at worst. Moreover, the avoidance devices have the benefit of having “a recognisable relationship with the policy considerations which militate against judicial intervention”76 but also lay bare the reasons for non-intervention; reasons that are challengeable on appeal.

And thus Smillie laid his blueprint for a pragmatic approach to judicial review. By collapsing the distinction between jurisdictional and non-jurisdictional error, Smillie’s approach forced the Court to provide a substantive rationale for why it wished to intervene (or not) in judicial review. It provides a clear set of factors that would justify intervention and also a set of devices that a Court can use to justify non-intervention. Most importantly, however, it lays bare the practical considerations of the Court in judicial review; a far more intellectually honest approach than the reverse engineering of jurisdictional error that followed Anisminic.

1.3 Smillie’s response to Bulk Gas

Three years after Smillie’s article was published, Cooke J had another attempt to refine the New Zealand approach to Anisminic and this was a case of third time lucky. As Smillie notes in his 1984 comment on Bulk Gas Users Group v

72 At 449.

73 At 451.

74 At 452-456.

75 At 452.

76 At 447.

Attorney General, Cooke J’s approach “opens the way to a more flexible, and yet more soundly based”77 approach to judicial review. The decision, although it did not cite Smillie’s 1980 article, adopted similar reasoning in order to consign the jurisdictional/non-jurisdictional error distinction to history. Smillie uses the case in turn to provide a vision for the foundation and scope of review that complements the pragmatic approach he set out four years previous.

The facts of Bulk Gas were straightforward.78 As its name suggests, the Bulk Gas Users Group made bulk purchases of natural gas from the Auckland Gas Company (AGC), who held a monopoly over gas supply in the city. AGC in turn purchased its gas from the state-owned Natural Gas Corporation of New Zealand (NGC). The NGC applied to the Secretary of Energy to increase its prices that it charged distributors such as AGC, which the Secretary duly approved. Before he approved the increase, the relevant legislation required the Secretary to “ensure that the applicant and such other persons as in his opinion have a direct interest in the matter79 were given an opportunity to make submissions. Bulk Gas had wanted to make submissions, but the Secretary determined that it did not have a “direct interest in the matter”, and thus refused to hear it. Bulk Gas sought a declaration from the Court that it had a right to make submissions to the Secretary in the future, i.e. that the Secretary had misinterpreted the term “direct interest in the matter”.

The Secretary was protected by a privative clause that prevented the Court “calling into question” his determinations “except on the ground of a lack of jurisdiction”.80 This was the legislative hook that allowed Cooke J to determine what “jurisdiction” could mean in such circumstances. Both the High Court and the Court of Appeal actually agreed with the Secretary’s interpretation of “direct interest in the matter”.81 However, while the High Court had also held that the privative clause meant the Secretary’s interpretation of “direct interest in the matter” was immune from challenge, the Court of Appeal unanimously held that in the event of a misinterpretation by the Secretary, the clause would

77 Smillie II, above n 22, at 552.
78 Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA) at 130.
79 Commerce Act 1975, s 97.

80 Commerce Act 1975, s 96: “Proceedings of the Secretary under this Part of this Act shall not be held bad for want of form. Except on the ground of lack of jurisdiction, no order, approval, proceeding, or decision of the Secretary under this Part of this Act shall be liable to be challenged, reviewed, quashed, or called in question in any Court, but there shall be a right of appeal to the Commission in accordance with section 99 of this Act.”
81 Smillie II, above n 22, at 555.

not have protected his decision.82 Justice Cooke’s route to this result was through a much wider interpretation of Anisminic than his previous attempts in Car Haulaways and Engineers’ Union. Rather than attempting to determine whether an error in construction was jurisdictional, Cooke J held that the question was instead whether Parliament intended the Secretary to determine questions of law conclusively. If so, the privative clause would protect his determinations. If not, the clause would not prevent intervention by the Court.83 Moreover, Cooke J held that it will be rare when the Court does find that decision-makers will have such conclusive authority: “Courts of general jurisdiction will be slow to conclude that power to decide a question of law conclusively has been conferred on a statutory authority or tribunal.”84

Justice Cooke derived authority for his view from the judgments of Lord Diplock in Re Racal Communications Ltd85 and O’Reilly v Mackman.86 In the former case, Lord Diplock had held that there is a presumption that Parliament confers power on administrative decision-makers to decide a particular question, and “if there has been any doubt as to what that question is, this is a matter for courts of law to resolve in fulfilment of their constitutional role as interpreters of the written law”.87 In the latter case, Lord Diplock had held that the “full consequences” of Anisminic were to virtually abolish the distinction between jurisdictional and non-jurisdictional error.88 Lord Diplock’s judgments thus played catch-up to Lord Denning’s approach in Pearlman; Cooke J’s adoption in Bulk Gas of Lord Diplock’s approach allowed New Zealand to catch up with the United Kingdom and abandon the “patent sophistry” that Smillie argued Anisminic had initially provoked.89

The correct approach after Bulk Gas, as explained by Smillie, was that the Court had a constitutional role to interpret the meaning of all statutory terms and a privative clause could not, in and of itself, deprive the Court of that role.90 Instead, a privative clause is simply a factor that indicates Parliament’s intention to reduce the likelihood of the Court’s intervention. Justice Cooke named other factors in Bulk Gas that would further reduce that likelihood: where a decision-

82 At 555.

83 Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA) at 133.
84 Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA) at 133.
85 Re Racal Communications Ltd [1980] UKHL 5; [1981] AC 374 (HL).
86 O’Reilly v Mackman [1983] UKHL 1; [1982] 3 All ER 1124 (HL).
87 Re Racal Communications Ltd [1980] UKHL 5; [1981] AC 374 (HL) at 382-383.
88 O’Reilly v Mackman [1983] UKHL 1; [1982] 3 All ER 1124 (HL) at 1133.
89 Smillie II, above n 22, at 558.

90 At 559.

maker’s functions were more judicial than administrative; where intervention was not justified due to the triviality of the error; or where an appeal on a question of law lies against the decision-maker, the Court would be less likely to intervene in the decision-maker’s determination.91 Smillie also named these factors three years previously in his 1980 article as those which would influence the Court’s decision on whether to intervene. However, his 1984 article added one important refinement. Smillie now characterised the likelihood of Court intervention as a standard of review.

These standards were twofold: either ‘correctness’: the Court would substitute its decision for the decision-maker’s; or ‘reasonableness’: the Court would only substitute its decision for the decision-maker’s where the latter fell outside of the range of outcomes reasonably open to it or the decision-maker failed to take reasonable steps to apprise itself of the relevant facts.92 The Court would arrive at either standard after a thorough consideration of the factors stated in his 1980 article, pithily restated in his 1984 article as:93

... the status and composition of the tribunal, the susceptibility of the statutory terms to precise definition, the nature of the subject matter of the decision, the seriousness of the alleged error, the effect of the decision on the applicant’s interests and the need to preserve a reasonable level of overall consistency in the exercise of a statutory function.

The core advantage of these ‘standards of review’ was to provide decision-makers with a clear incentive to deliver a reasoned decision, for such a decision is more likely to persuade a reviewing Court that it was within the permissible bounds of reasonableness. Additionally, the standards of review were an effective alternative to the “avoidance devices” that Smillie had postulated in 1980: now, after the analysis of the factors, the Court could rely on the ‘reasonableness’ standard to take a more deferential stance, rather than concocting a reason why it should not intervene. With this refinement, Smillie’s pragmatic approach was now a pragmatic and functional approach: it forced the Court away from the ‘subtle manipulation’ of the concept of jurisdictional theory and instead required it to evaluate the nature and quality of the particular determination impugned and undertake a realistic assessment of its appropriate role – and function – on review.94

91 At 557, referring to Cooke J in Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA) at 136.
92 At 558.
93 At 559.
94 At 559.

The pragmatic and functional approach was not without its disadvantages. As Smillie notes, an immediate conceptual difficulty presented by Bulk Gas was that privative clauses of the type the case dealt with had no meaningful effect, since any error by the Secretary could amount to “a lack of jurisdiction”.95 Instead, privative clauses are simply an indication by Parliament that it intended a determination to be final and conclusive; if it wished to unambiguously convey that intention, it must do so by some other means.96 This approach means that judicial review becomes a constitutional safeguard that not even Parliament can displace in the absence of “truly exceptional” form of words that in 1984, at least, had not yet been devised.97 However, if this seems to amount to, as Wade argued, “total disobedience of Parliament”,98 Smillie argues it is a far preferable situation to Anisminic, which had the binary effect of either giving privative clauses complete (if an error was non-jurisdictional) or no (if an error was jurisdictional) effect. Instead, under Smillie’s approach, while privative clauses are neutered, they are never completely ignored, and instead represent part of a more accurate representation of the extent to which Parliament expects the Court to supervise a decision-maker’s power.99 It is a far more realistic – and intellectually honest – approach, given the ease with which the Court could simply ignore privative clauses under the Anisminic approach.

A more pressing difficulty with Smillie’s approach is that, like the Anisminic approach, it too was at risk of judicial manipulation that would allow “an unwarranted level of intrusion”.100 If a Court applies a correctness standard to questions of law determined by an administrative decision-maker, an expansive approach to what classifies as a question of law may allow a reviewing Court to substitute its opinion for a decision-maker’s in situations where the more relaxed standard of reasonableness is justified. Conversely, courts reluctant to intervene in a decision may incorrectly categorise questions of law as questions of fact, in order to apply that relaxed standard.101 As Smillie notes, American jurisprudence indicated that the question/fact distinction is no less vexed than the jurisdictional/ non-jurisdictional error distinction.102 Smillie was content to assume that if courts followed Cooke J’s limitations on when a correctness standard was justified as

95 At 559.
96 At 559-560.
97 At 560.
98 At 560, citing William Wade Administrative Law (Clarendon Press, New York, 1982) at 604.
99 At 560.
100 At 561.
101 At 561.
102 At 561.

stipulated in Bulk Gas, “any fear of unwarranted usurpation of an agency’s function should prove to be unfounded.”103

The final concern with the Bulk Gas approach is the risk that courts would only apply it to decisions of administrative tribunals, while preserving the Anisminic approach when reviewing decisions of inferior courts. Smillie records that such a development would be highly undesirable: not only is the line between inferior courts and administrative tribunals “hard to draw,”104 the same problems with the Anisminic approach would persist, and in any case, the assumption that inferior courts are inherently better qualified to answer questions of law is erroneous.105 Indeed, given the relative expertise of some of tribunals, it might in fact be more appropriate in some circumstances for a reviewing court to have more restraint when looking at such decisions as compared to an inferior court.106 Applying the Bulk Gas approach to all decision-makers, and instead taking into account the nature of the decision-maker as a factor within that approach on a case-by-case basis, is a far more flexible and nuanced method. In any case, since Lord Diplock in O’Reilly “did not trouble to exclude inferior courts from the ambit of his broad ambit of principle”107 and Cooke J actually demonstrated how his approach could apply to inferior courts in Bulk Gas, Smillie was of the opinion that there were “encouraging signs that the courts will not fall into the trap” of creating a new, artificial distinction between administrative tribunals and inferior courts, and preserving the application of Anisminic with regards to the latter.108

Thus, in two articles, Smillie outlined his pragmatic and functional approach to judicial review. In his 1980 response to Anisminic, he provided a clear rationale for intervention by a reviewing court that did not rely upon theories of jurisdiction. Instead, the rationale was built on the practical reality of judicial processes at the time: if a Court wished to intervene, it simply needed to reverse engineer its reasoning to classify an error by a decision-maker as jurisdictional in nature. Smillie’s approach was to accept that reality, but demand better justification by the Court for its intervention. Through a consideration of different factors, the level of judicial supervision that Parliament intended would become clear, giving the Court the constitutional authority to intervene, or vice versa, providing it with a clear reason why it should not do so. An immediate casualty of the approach was

103 At 561.

104 At 562, citing Cooke J in Bulk Gas Users Group v Attorney-General [1983] NZLR 129 (CA) at 136.
105 At 562-563.
106 At 563.
107 At 563.
108 At 563.

the literal effect of privative clauses, but since Anisminic had already administered the deathblow to such clauses, Smillie’s approach resuscitated such clauses to give them appropriate – but not dominant – effect.

In his 1984 response to Bulk Gas – essentially New Zealand’s Anisminic – that approach was further refined to indicate effect of such factorial analysis, and what different levels of intrusion would amount to. Accepting that the constitutional role of the Court would mean no decision-maker would be unimpeachable, drawing on Cooke J’s judgment in Bulk Gas, Smillie devised two ‘standards’ of review. Essentially, if the factors indicated that the Court should be slow to intervene, a ‘reasonableness’ standard was required, which would limit the Court’s intervention to only the most extreme circumstances. If, on the other hand, the Court ought to intervene, a correctness standard would lead the Court to substituting its interpretation of the law for the decision-maker’s where they differed. Read together, these articles provided a blueprint for the scope and foundation for judicial review in the post-Anisminic era. Given the year that the second article was produced marked the beginning of the most significant government reform in a generation, Smillie’s approach would have provided a useful framework for dealing with what was going to become a complicated administrative law landscape. It was, however, as I will discuss in the next part, a proposal that this jurisdiction chose to ignore, much to its detriment.


The first part of this paper detailed how Smillie, in responding to some of the major administrative law cases of his time, not only provided a cogent expression of the scope and foundation of the Court’s judicial review jurisdiction, but also an approach that reflected the practical realities of how the Court exercises that jurisdiction. This part will examine how New Zealand courts declined to adopt Smillie’s approach, instead, and perhaps unconsciously, choosing a simple approach championed by Cooke. It will then look briefly at Canada, a jurisdiction that did take a similar approach to Smillie, before looking to two specific contemporary issues in New Zealand administrative law that Smillie’s approach may have obviated.

2.1 A fork in the road

Despite providing the basis for Cooke J’s strong pronouncements in Bulk Gas (and Smillie’s response to those pronouncements), Lord Diplock may provide the reason why Smillie’s approach was not adopted in New Zealand. As referred to earlier, shortly after Smillie’s 1984 article, the House of Lords’ decision of C.C.S.U.

was released, where Lord Diplock indicated that the basis for intervention by a Court could be classified under three heads of review.109 This was reaffirmed by Cooke, extra-judicially, in his seminal 1986 piece, ‘The Struggle for Simplicity in Administrative Law’: “the substantive principles of judicial review are simply that the decision-maker must act in accordance with the law, fairly and reasonably.”110 While the adoption of this categorisation – and what I shall term the ‘categorical approach’ – may seem uncontroversial, and indeed, welcome, given its simplicity, one of the effects of accepting the categorical approach was to relegate Smillie’s pragmatic and functional approach.

In his article, Cooke refers to the issues that I discussed in the first part of the paper; the “definitive answer to the riddle of jurisdiction” provided by the House of Lords in Racal and O’Reilly – and adopted in New Zealand in Bulk Gas – in virtually abolishing the distinction between jurisdictional/non-jurisdictional error.111 He goes as far to say that these developments put New Zealand “on the brink of open recognition of a fundamental rule of our mainly unwritten constitution: namely that determination of questions of law is always the ultimate responsibility of the Courts of general jurisdiction.”112 However, in discussing these developments, Cooke limits their significance and impact to one head of review: they are confined to his discussion that decision-makers must act in accordance with the law, or as Lord Diplock put it in C.C.S.U.: ‘illegality’.113 Unlike Smillie, he did not see or treat Bulk Gas as providing the means to develop a foundational and organisational framework in judicial review: it simply solved one discrete issue. Instead, Cooke saw the categorical approach as providing such a framework, and had done so for some time. As he references earlier in the article:114

In 1979 I mentioned the prospect of an answer from counsel or the House of Lords to the old riddle of the relationship between error of law and jurisdiction; and I ventured that it might not be an altogether absurd over-simplification to say that the day might come when the whole of administrative law could be summed up in the proposition that the administrator must act fairly and reasonably.

109 See Barendt, above n 8 and Wilberforce, above n 9.

110 Cooke, above n 13, at 5-6. Reaffirmed, rather than adopted, since Cooke argues that this epithet was a refinement of a similar pronouncement he made in 1979, and can only be “claimed to resemble” Lord Diplock and Lord Roskill’s pronouncements in C.C.S.U.
111 At 7-8.
112 At 10.
113 At 6.
114 At 5, referring to “Third Thoughts on Administrative Law” [1979] NZ Recent Law 218.

Cooke’s categorical approach thus placed the three heads of review as the fundamental organisational principle for judicial review, and his article was a long-form submission that this “threefold proposition”115 would not only solve the “old riddle” of jurisdictional/non-jurisdictional error but also, through its simplicity, administrative law itself. With confidence, he did not doubt that “many hard, hard administrative cases lie ahead; but the cardinal principles to be brought to bear on them are no less surely gaining in clarity and simplicity.”116

Perhaps obviously, Cooke was not alone in his view that the categorical approach was a superior organisational framework for administrative law. It remains the dominant approach to judicial review, both in New Zealand and the United Kingdom, whereas the “pragmatic and functional approach” favoured by Smillie is now scarcely mentioned. For example, De Smith argues that suggestions of a pragmatic and functional approach after Anisminic were only helpful to the extent that jurisdictional error persisted.117 Since the concept of jurisdictional error is “ultimately based upon foundations of sand”, and that “much of the superstructure has already crumbled”, “what remains is likely quickly to fall away as the courts rightly insist that all administrative action should be, simply, lawful, whether or not jurisdictionally lawful.”118 As a result, its view is that there is no room for such an approach.

De Smith is correct that analyses of a pragmatic and functional approach in the United Kingdom in the 1980s were predicated on the retention of jurisdictional error, and not viewed with much charity.119 For instance, Peiris spoke of the pragmatic approach, as applied by Lord Denning in Pearlman, with disdain: “[h] aving regard to the likelihood of exacerbating tensions between the administration and the courts, there is good reason to regret so unrepentantly intrusive a dimension of judicial review”120 and Beatson argued that the pragmatic and functional approach “suffers from serious theoretical practical disadvantages.”121 In this context, where the pragmatic and functional approach was viewed as a halfway house that retained the old orthodoxy of jurisdictional error, the categorical approach to organising administrative law is perhaps obviously preferable.

115 At 17.
116 At 17.
117 De Smith, above n 28, at [4-043].
118 At [4-050].

119 Gamini Peiris “Judicial Review and judicial policy: the evolving mosaic” (1987) LQR 66; Carl Emery and Bernard Smythe “Error of Law in Administrative Law” (1984) 100 LQR 612; and Jack Beatson “The Scope of Judicial Review for Error of Law” (1984) 4 OJLS 22.
120 Peiris, above n 119, at 73.
121 Beatson, above n 119, at 44.

However, as the first part of the paper ought to have made clear, Smillie’s version of the pragmatic and functional approach was a wholesale repudiation of the old jurisdictional error approach.122 It was an attempt to remove the conceptual incoherence of that old orthodoxy, and thus, despite De Smith’s assertions to the contrary, did not depend upon that old orthodoxy’s crumbling superstructure. Moreover, Smillie’s pragmatic and functional approach is not mutually exclusive with Cooke’s categorical approach: it retains each of the heads of review. Illegality is obviously encompassed by the approach (any misinterpretation or misapplication of a decision-maker’s authorising power will amount to reviewable error); similarly, a denial of natural justice or the requirements of procedural fairness will amount to an error of law;123 and irrationality is incorporated within the unreasonableness standard of review. Thus, while Smillie’s approach still takes account of the breadth of judicial review; the key difference is that it also takes into account the depth of review. In contrast, the categorical approach does not have an inherent mechanism to modulate the depth of scrutiny by a reviewing court, which as will be discussed below, is a significant problem.

The pragmatic and functional approach, as postulated by Smillie or any number of other commentators, was not adopted in New Zealand or the United Kingdom. It was, however, adopted in Canada, and before looking to the issues that are currently facing New Zealand, I wish to look down the path this jurisdiction did not travel down to provide a counterfactual to our current system.

At the same time as both the United Kingdom and New Zealand were unshackling themselves from the constraints of jurisdictional error, Canada engaged in a far more forceful repudiation of the literal effect of privative clauses: its Supreme Court ruled such attempts by the legislature to oust judicial review as unconstitutional.124 This set the stage for the Supreme Court to adopt what it termed as a pragmatic and functional approach in U.E.S., Local 298 v Bibeault,125 where the Court held that the degree of scrutiny by a reviewing court was determined by an examination

of: 126

122 Smillie II, above n 22, at 560.

123 Smillie I, above n 21, at 422-423.

124 Denis Lemieux “Judicial Deference in Canadian Administrative Law: The Pragmatic And Functional Approach, Pushpanathan v Canada, [1998] 1 SCR 982” (2002) 54 Admin LR 757 at 759, citing Crevier v Attorney General of Quebec [1981] 2 SCR 220 at 222 (SCC) and Attorney General of Quebec v Farrah [1978] 2 SCR 638 at 641 (SCC).
125 U.E.S., Local 298 v Bibeault [1988] 2 SCR 1048 (SCC).
126 U.E.S., Local 298 v Bibeault [1988] 2 SCR 1048 (SCC) at [122].

the wording of the enactment conferring jurisdiction on the administrative tribunal, ... the purpose of the statute creating the tribunal, the reason for its existence, the area of expertise of its members and the nature of the problem before the tribunal.

The pragmatic and functional approach was further refined by the Supreme Court in Puspanathan127 and Southam128 which, respectively, refined the factors to be taken into account by the reviewing court129 and the standards of review that would result from such an examination.130 It was later broadened in application and pithily described as recognising that “standards of review for errors of law are appropriately seen as a spectrum, with certain decisions being entitled to more deference, and others entitled to less.”131 In 2008, while the nomenclature of the “pragmatic and functional approach” became the “standard of review analysis” and the different standards of review were further refined and consolidated132 the general approach remained the same.133 As it stands, after considering whether the standard of review for the particular decision-maker or decision has already been determined in an earlier case, a range of factors are considered, namely:134

These factors will determine whether one of two standards of review are applicable:

127 Pushpanathan v Canada (Minister of Citizenship and Immigration) [1998] 1 SCR 982.
128 Canada (Director of Investigation and Research) v Southam Inc [1997] 1 SCR 748 (SCC).

129 Pushpanathan v Canada (Minister of Citizenship and Immigration) [1998] 1 SCR 982 at [29] [38].

130 Canada (Director of Investigation and Research) v Southam Inc [1997] 1 SCR 748 (SCC) at [30].

131 Baker v Canada (Minister of Citizenship and Immigration) [1999] 2 SCR 817 at [55] (SCC).
132 Dunsmuir v New Brunswick [2008] 1 SCR 190 (SCC).
133 Dunsmuir v New Brunswick [2008] 1 SCR 190 (SCC) at [121]-[122] per Binnie J.
134 Canada (Citizenship and Immigration) v Khosa [2009] 1 SCR 339 at [54].
135 Canada (Citizenship and Immigration) v Khosa [2009] 1 SCR 339 at [59].
136 Dunsmuir v New Brunswick [2008] 1 SCR 190 (SCC) at [50].

Such is the Canadian approach, and it ought to strike a tone of familiarity given the factors and standards that form that approach are essentially the same that Smillie proposed 30 years ago.

It is important to note that I am under no illusions that the Canadian approach to administrative law is in any way paradisiacal. To say that there are significant problems and complexities within contemporary Canadian administrative law is an understatement137 to the point that some have noted that presenting “the fine points of curial deference in Canadian administrative law amounts to a cruel and unusual punishment for the reader.”138 However, this brief foray into the Canadian experience justifies two propositions.

First, that Smillie’s sketch of a pragmatic and functional approach in the early 1980s had the potential to be the blueprint for New Zealand’s approach to administrative law. Canadian administrative law was, at one point, in a very similar position to New Zealand and the United Kingdom.139 From the 1980s onwards, however, it struck its own path, showing that New Zealand could have done the same if it followed Smillie’s blueprint.

Second, in describing the Canadian approach above, I cited six decisions from its top Court, each of which amounted to an attempt to clarify and refine that jurisdiction’s approach to administrative law. This is only the tip of the iceberg: the Supreme Court of Canada regularly revisits the area from a macro-perspective in an attempt to clarify the law for judges, practitioners and litigants. A great deal many more commentators have written on the success and failures of those attempts. In comparison, since the major cases of Cooke J in the 1980s, New Zealand courts have rarely considered this jurisdiction’s overall approach to administrative law; commentators are left to cobble decisions together from lower courts in a manner akin to reading tea leaves. There are a multitude of problems with the Canadian approach and there would have been a multitude of problems with following Smillie’s approach. However, given its holistic nature, those problems would have forced the same macro-consideration of the area that occurs in Canada, which is itself an advantage. In contrast, New Zealand’s approach – the path we ventured down – is now faced with several significant problems, ones that could have been

137 See, for example, David Mullan’s response to Dunsmuir in “Dunsmuir v. New Brunswick, Standard of Review and Procedural Fairness for Public Servants: Let’s Try Again!” (2008) 21 CJALP 117 and Paul Daly’s excellent historical overview in “The Struggle for Deference in Canada” in Hanna Wilberg and Mark Elliott (eds) The Scope and Intensity of Substantive Review: Traversing Taggart’s Rainbow (Hart Publishing, Oxford, 2015) 297.
138 Lemieux, above n 124, at 757.
139 Daly, above 137, at 301.

obviated had we ventured down a different path. I will discuss these issues in the next part.

2.2 New Zealand’s approach

Lord Cooke’s (as he would become) seminal 1986 contribution to New Zealand administrative law was not his last. Indeed, in analysing his contribution to administrative law, Dean Knight notes that “[t]o a large degree, the crafting of the indigenous nature of New Zealand’s administrative law is attributable to the work of Lord Cooke.”140 His 1986 article, however, in making the case for the categorical approach, was arguably the most influential of his contributions, for it amounted to meta-level analysis.141 That decision-makers “must act in accordance with the law, fairly and reasonably” was not only a statement on the cardinal principles he saw as underlying administrative law, but also a manifestation of his rejection of legal formalism in favour of simplicity.142 This philosophy – a “quest for simplicity” – was evident in his analysis of each head of review and his work within administrative law generally.143

In the years that followed this article, Lord Cooke’s contribution to administrative law was immense: during his time on the New Zealand bench, he sat on 35 per cent of all administrative law cases reported in the New Zealand Law Reports.144 Despite his multifarious contributions to New Zealand jurisprudence, administrative law was arguably the area Lord Cooke “made his own”.145 It is almost undeniable that Lord Cooke’s contribution was a significant and fruitful one: it led to a rich development of New Zealand’s indigenous approach to administrative law while he served on the Bench. However, after his retirement from the Court of Appeal in 1996 and appointment to the House of Lords – and perhaps without the benefit of his steadying hand – that development began to drift.

Since that time, New Zealand’s top appellate courts have rarely revisited administrative law fundamentals, although there are exceptions. The Court of

140 Dean Knight “Simple, Fair, and Discretionary Administrative Law” (2008) 39 VUWLR 99 at 102.

141 At 103.

142 At 103.

143 At 103-104.

144 At 100.

145 Michael Taggart “The Contribution of Lord Cooke to Scope of Review Doctrine in Administrative Law: A Comparative Common Law Perspective” in Paul Rishworth (ed) The Struggle for Simplicity in the Law: Essays for Lord Cooke of Thorndon (Butterworths, Wellington, 1997) 189, cited by Knight, above n 140, at 101.

Appeal’s 1998 decision in Peters v Davison146 and, a decade later, the Supreme Court’s decision in Unison Networks Ltd v Commerce Commission147 were both notable for their meta-analysis of administrative law. Peters v Davison – a case on the reviewability of commissions of inquiry – stated the scope of judicial review, namely: 148

The grounds upon which judicial review is available are well established. Judicial review is in general available where a decision-making authority exceeds its powers, commits an error of law, commits a breach of natural justice, reaches a decision which no reasonable tribunal could have reached or abuses its powers...

It also echoed Cooke J’s efforts in Bulk Gas regarding the non-jurisdictional/ jurisdictional error distinction: “[e]rror of law is a ground of review in and of itself; it is not necessary to show that the error was one that caused the tribunal or Court to go beyond its jurisdiction.”149 Such a distinction was “redundant”.150

Similarly, McGrath J for the Court in Unison made clear the limits of administrative power in a series of straightforward propositions. Decision-makers must exercise their statutory powers in accordance with the statutes which confer them, and if they make decisions that are outside the limits of their powers, they abuse them.151 A statutory power is subject to limits even if it is conferred in unqualified terms: even the broadest of discretions given to a decision-maker must promote the policy and objects of its empowering statute.152 However, where that decision-maker has particular expertise and is given a broad discretion to achieve expansive objectives, the Court is “unlikely to intervene unless the body exercising the power has acted in bad faith, has materially misapplied the law, or has exercised the power in a way which cannot rationally be regarded as coming within the statutory purpose.”153

I state these cases and their propositions to show that years after Smillie’s responses, New Zealand was still grappling with the entry points for depth of judicial review. The simplicity of Cooke’s categorical approach was undeniable, but it is clear that it was insufficient as an answer to all the difficulties that faced administrative law. Unfortunately, Peters and Unison did not fill the gaps either, and their broad-

146 Peters v Davison [1999] 2 NZLR 164 (NZCA).
147 Unison Networks Ltd v Commerce Commission [2007] NZSC 74; [2008] 1 NZLR 42 (SCNZ).
148 Peters v Davison [1999] 2 NZLR 164 (NZCA) at 180.
149 At 181.

150 At 181.

151 Unison Networks Ltd v Commerce Commission [2007] NZSC 74; [2008] 1 NZLR 42 (SCNZ) at [51].
152 At [53].

153 Unison Networks Ltd v Commerce Commission [2007] NZSC 74; [2008] 1 NZLR 42 (SCNZ) at [55].

brushstroke analysis was insufficient to address the drift referred to earlier. I wish to highlight two issues in particular that have resulted from that drift.

(1) Intensity of Review

As Knight notes, Lord Cooke had maintained a consistent criticism of the orthodox formulation of one of his three heads of review in the categorical approach: reasonableness. While he preferred a simple test (whether the decision was outside the limits of reason)154 the orthodox Wednesbury standard of reasonableness prevailed. That criticism reached its crescendo once he was appointed to the House of Lords: he thought it “unfortunate that Wednesbury and some Wednesbury phrases have become established incantations in the courts of the United Kingdom and beyond”155 and that the “day will come when it will be more widely recognised that the Wednesbury case was an unfortunately retrogressive decision in English administrative law.”156 His former colleagues in the New Zealand Court of Appeal were more circumspect. Three months after his retirement, the Court of Appeal released its decision in Wellington City Council v Woolworths New Zealand Ltd (No. 2).157 While it certainly removed the tautology inherent in Lord Greene MR’s original formulation of reasonableness in Wednesbury,158 it also entrenched the impossibly high test that Wednesbury represented: unreasonable decisions were those “so ‘perverse’, ‘absurd’ or ‘outrageous in [their] defiance of logic’ that Parliament could not have contemplated such decisions being made...”.159

The “monolithic and hyperbolic standard”160 would not retain its exclusive claim to orthodoxy for long, and in New Zealand, different standards of reasonableness – applied according to the context – were postulated.161 Arguably, the most strident adoption of such a varying standard – and rejection of Wednesbury – occurred in the High Court decision of Wolf v Minister of Immigration, where Wild J held

154 Eg: “[R]easonable means reasonable. The definition in the Concise Oxford Dictionary, reflecting as it should ordinary educated usage, is ‘within the limits of reason’. What is outside

the limits is unreasonable; what is inside them is reasonable.”: Cooke, above n 13, at 14.

155 R v Chief Constable of Sussex, ex parte International Trader’s Ferry Ltd [1998] UKHL 40; [1999] 1 All ER 129 at 157.

156 R (Daly) v Secretary of State for the Home Department [2001] UKHL 26; [2001] 2 AC 532 at 549.

157 Wellington City Council v Woolworths New Zealand Ltd (No. 2) [1996] 2 NZLR 537, delivered on 24 May 1996; Cooke retired from the Court on 16 February 1996.

158 “Unreasonable in the sense that [...] no reasonable authority could have come to it”: Associated Provincial Picture Houses Ltd v Wednesbury Corporation [1947] EWCA Civ 1; [1947] 2 All ER 680 at 683.
159 Wellington City Council v Woolworths New Zealand Ltd (No. 2) [1996] 2 NZLR 537 at 552.
160 Knight, above n 140, at 106.

161 See eg, Thomas J’s lengthy analysis in Waitakere City Council v Lovelock [1997] 2 NZLR 385 (CA).

that the standard expressed in Woolworths was “not, or should no longer be, the invariable or universal tests of “unreasonableness” applied in New Zealand public law.”162 The standard applied would be determined according to a series of factors: “who made the [decision]; by what process; what the decision involves (ie its subject matter and the level of policy content in it) and the importance of the decision to those affected by it, in terms of its potential impact upon, or consequences for, them.”163

This was the beginning of New Zealand’s flirtation with the concept of variable intensity of review, and in the years immediately after Wolf, a significant amount of judicial analysis and commentary were dedicated to what it would mean for this jurisdiction’s approach to administrative law. Professor Taggart’s 2008 article, ‘Proportionality, Deference, Wednesbury’,164 was arguably the most influential of this analysis, motivating an entire special issue of the New Zealand Law Review and a book to the issues it canvassed and provoked.165 One of Taggart’s concerns, and perhaps what motivated him to construct his famous “rainbow of review” was that while there were an abundance of decisions in New Zealand discussing, applying and concocting different standards of review: “[i]t seems to me that unless we commit to that sort of mapping project the law will continue to be rather chaotic, unprincipled, and result-oriented.

Unfortunately, in the years since, Taggart’s fear has arguably come to fruition. This is partly due to a disdain from the Supreme Court toward to the concept of variable intensity. The Court, now over a decade old, has barely mentioned and/or considered ‘deference’ or ‘intensity’ of review in an administrative law setting167 and the Chief Justice has viewed the concept as “dreadful”; degrees of unreasonableness as amounting to “dancing around on the heads of pins” and “a New Zealand perversion of recent years”.168 Knight postulated that there were five potential approaches for New Zealand to adopt regarding this issue, varying

162 Wolf v Minister of Immigration [2004] NZAR 414 at [47].
163 Wolfv Minister of Immigration [2004] NZAR 414 at [47].
164 Michael Taggart “Proportionality, Deference, Wednesbury[2008] NZ L Rev 423.

165 (2010) 2 NZ L Rev: “Proportionality, Deference, Wednesbury: Taking up Michael Taggart’s Challenge”; Wilberg and Elliot, above n 137.
166 Taggart, above n 164, at 453.

167 Based on a keyword search of ‘intensity’ and ‘deference’ in Briefcase, 8 December 2015. Justice Tipping mentions ‘intensity of review’ in Hansen v R [2007] NZSC 7, [2007] 3 NZLR 1 at [115] but it was not an administrative law case; Keith J briefly considered different standards of review in Westfield (New Zealand) Ltd v North Shore City Council [2005] NZSC 17, [2005] 2 NZLR 597 at [54].

168 Dean Knight “Mapping the Rainbow of Review: Recognising Variable Intensity” [2010] NZ L Rev 393 at 402.

from full-blooded acceptance or rejection of intensity of review, to more subtle or implicit variations.169 However, without any clear choice by the Supreme Court or Court of Appeal as to which is approach is best, the chaotic and directionless approach to the issue prevails in this country.

Revealingly, I think, both Knight and Taggart mention the Canadian approach in their analysis. Knight argues that one approach could be “more discrete standards of review ... , as was the practice for many years in Canada”170 and Taggart notes “the doctrine of deference [is] well-established in Canada”.171 I want to reiterate that I am not arguing the Canadian approach is flawless – indeed, in the same special issue of the New Zealand Law Review, David Mullan wrote of an “emerging crisis” in Canadian administrative law172 – but I do wish to argue that had Smillie’s approach been adopted, the aimlessness we see currently would not exist. Smillie’s approach, since it is based on a concept of deference and varying standards of review, would have forced judicial consideration of the issue. Just as in Canada, we may have had changes and upheavals in what those standards are and what they amount to, but we would still have had at all times a holistic approach to administrative law, and we would have had the proper consideration by higher appellate courts of that approach. Instead once Cooke, the doyen of New Zealand administrative law, left these shores the simple categorical approach that he espoused, unable to deal with varying intensity of review without further consideration, has drifted. We are now left without an approach and no judicial consideration of the issue, a position that is in considerable deficit compared to Canada and one that would have been avoided if we had adopted Smillie’s approach.

(2) Privative Clauses

Since both Anisminic and Bulk Gas focused upon privative clauses, it is worth considering what the effect of not following Smillie’s approach has been on this issue. As Graham Taylor notes, the New Zealand Bill of Rights Act 1990, in recognising a “right to apply for judicial review”173 consolidated and strengthened the Bulk Gas approach that prevented privative clauses from having literal effect.174

169 At 430-431.

170 At 431.
171 Taggart, above n 164, at 424.

172 David Mullan “Proportionality – A Proportionate Response to an Emerging Crisis in Canadian Judicial Review Law” [2010] NZ L Rev 233.
173 New Zealand Bill of Rights Act 1990, s 27(2).

174 Graham Taylor Judicial Review: A New Zealand Perspective (LexisNexis, Wellington, 2014) at [2.63].

However, this consolidation has not been sufficient to remove the confusion that surrounds privative clauses. In O’Regan v Lousich,175 although Tipping J agreed with the “essential thrust” of Bulk Gas, his decision distinctly preserved the ability of Parliament to create a clause that would oust the Court’s inherent jurisdiction to intervene:176

It is unlikely that Parliament will wish to give such absolute protection to a decision in spite of its being ultra vires and thus outside the decision maker’s jurisdiction; but nevertheless Parliament is sovereign and may do so if it chooses and says so clearly.

Recall that Smillie’s view of the majority’s decision in Bulk Gas was somewhat different, namely that: “judicial supervision of administrative decision-making is a fundamental constitutional safeguard which even Parliament cannot completely remove, at least without some truly exceptional form of words which has not yet been devised.”177 Justice Tipping’s view in O’Regan instead exemplified the “traditional constitutional justification” for judicial review: the implementation of Parliament’s will.178 This approach – exemplifying the ultra vires model of judicial review – was very much on the wane in New Zealand by this point, and according to Joseph, this was largely due to Cooke’s influence: “[u]nder his leadership, the courts relegated the ultra vires doctrine and experienced no crisis of legitimation. Only one judge has mounted a rearguard action in defence of ultra vires, and his words have not struck any chord.”179

Joseph was referring to Tipping J in O’Regan, and while he is correct in holding that he was isolated in trying to resurrect the ultra vires model in its entirety, the particular application of that model to privative clauses has persisted. In Zaoui v Attorney-General,180 the High Court repeated Tipping J’s approach verbatim.181 On appeal, while O’Regan was not mentioned by name, the Court of Appeal confirmed its approach, restating Anisminic182 and resurrecting the jurisdictional/ non-jurisdictional error dichotomy, which the Court of Appeal had already

175 O’Regan v Lousich: Proprietors of Mawhera v Maori Land Court [1994] NZHC 787; [1995] 2 NZLR 620 (HC).

176 O’Regan v Lousich: Proprietors of Mawhera v Maori Land Court [1994] NZHC 787; [1995] 2 NZLR 620 (HC) at 627.

177 Smillie II, above n 22, at 560.

178 O’Regan v Lousich: Proprietors of Mawhera v Maori Land Court [1994] NZHC 787; [1995] 2 NZLR 620 (HC) at 552.
179 Joseph, above n 18, at 358.
180 Zaoui v Attorney-General [2004] 2 NZLR 339 (HC).
181 At [60].

182 Zaoui v Attorney-General (No 2) [2004] NZCA 244; [2005] 1 NZLR 690 at [179] (NZCA).

consigned to history not once, but twice.183 Although there is no doubt that these decisions managed to constrict the effectiveness of privative clauses, it is the method and justifications they used to do so that also created conceptual confusion about the constitutional basis of judicial review, and the basis for an argument that the ultra vires model is not as dead as Joseph has argued it to be.184

However, the legacy of these cases was not simply conceptual confusion. In Tannadyce Investments Ltd v Commissioner of Inland Revenue185 Tipping J, now on the Supreme Court, seemed to answer Joseph’s criticism that he was simply a lone voice in the wilderness. Tannadyce dealt with a privative provision, but one of a stronger character than that in Bulk Gas, namely one that stated that decisions of the Commissioner of Inland Revenue could only be challenged through a specific statutory procedure, and could not be “disputed in a court or in any proceedings on any ground whatsoever”.186 The question for the Supreme Court was whether this was sufficient to exclude the Court’s inherent judicial review jurisdiction. Justice Tipping, for the majority, given the existence of a parallel statutory procedure, held that the clause was effective, and that decisions: “may not be challenged by way of judicial review unless the taxpayer cannot practically invoke the relevant statutory procedure. Cases of that kind are likely to be extremely rare.”187

Justice McGrath, for the minority, strongly resisted this conclusion, and citing Bulk Gas, held that:188

The courts of higher jurisdiction, however, have constitutional responsibility for upholding the values which constitute the rule of law. A central aspect of that role is to ensure that when public officials exercise the powers conferred on them by Parliament, they act within them. ... Judicial review is the common law means by which the courts hold such officials to account. Statutes limiting recourse to judicial review to challenge statutory decisions accordingly raise issues of constitutional concern.

Justice Tipping, directly responding to McGrath J, dismissed his concern:189

183 At [190]; See n 150 above.

184 Ruping Ye “The Demise of Ultra Vires in New Zealand: To Be? Not to Be!” (2008) 8 NZJPIL 287 at 309-315.

185 Tannadyce Investments Ltd v Commissioner of Inland Revenue [2011] NZSC 158; [2012] 2 NZLR 153 (SCNZ).
186 Tax Administration Act 1994, s 109(a).

187 Tannadyce Investments Ltd v Commissioner of Inland Revenue [2011] NZSC 158; [2012] 2 NZLR 153 (SCNZ) at [61].

188 At [3].

189 At [72].

There is no disadvantage, constitutional or otherwise, in giving effect to what Parliament has enacted and every reason for doing so. ... In these circumstances it is not appropriate to apply any presumption that Parliament’s purpose, when enacting s 109, was to preserve judicial review.

The rather astonishing effect of Tipping J’s judgment, as Andrew Beck noted, was to reverse the presumption established in Bulk Gas against excluding the Court’s judicial review jurisdiction.190 Recall that Smillie argued that this presumption could only be displaced by a “truly exceptional form of words” and thus it was near impossible for Parliament to remove the jurisdiction.191 Beck argues by reframing the issue, so that it is no longer one of constitutional importance as advanced McGrath J but simply one of statutory interpretation, Tipping J managed to sidestep the presumption.192 In doing so, however, it breathed new life into privative clauses, and reversed 40 years of orthodoxy. I agree with Beck that the ultimate effect of Tannadyce is a significant, and dangerous, one:193

The ratio of Tannadyce will be applied in other contexts, making it more difficult to access the court’s supervisory jurisdiction in the face of a privative clause. The legislature may also be encouraged to make greater use of privative clauses so as to prevent scrutiny by the courts. That would be most unfortunate. When it is too easy to avoid the court looking over the shoulder of officials, the rule of law is in jeopardy.

Smillie’s approach would have prevented this result in two ways. First, there is no scope for the result in Tannadyce – or the approach of Tipping J in O’Regan – under Smillie’s model. By acknowledging that Parliament cannot, constitutionally, oust the judicial review jurisdiction, privative clauses are denied their literal effect. However, they are not rendered irrelevant: they are simply one indication that Parliament intended the Court to adopt a deferential approach. The stronger the privative clause, the higher the likelihood that the Court will defer to the decision-maker, but it will never have the ability to prevent review proceedings in the first place. Applied to Tannadyce, for example, the strength of the privative clause, coupled with the expertise of the decision-maker would likely have led to a very deferential approach by the Court, making intervention unlikely. In any case, even if the Court did decide to intervene, it may exercise its discretion against granting a remedy (perhaps due to the availability of a parallel statutory regime for

190 Andrew Beck “Privative Clauses” [2012] NZLJ 91 at 91.
191 Smillie II, above n 22, at 560.
192 Beck, above n 190, at 91.
193 At 92.

challenge). This result – a deferential stance by the Court – respects Parliamentary intent without putting the rule of law in jeopardy.

Of course, Smillie’s approach would not assuage those who are adherents to the ultra vires model given it denies a statutory provision’s literal effect. However, regardless of which model of judicial review one subscribes to, it is undeniable that the conceptual confusion that cases like O’Regan and Tannadyce have fomented is lamentable. This is the second way that Smillie’s model would have prevented this result. Cooke’s focus on simplicity may have led to the relegation of the ultra vires model, as Joseph argues, but it is obvious that it did not lead to its eradication. Smillie’s model, in clearly choosing a constitutional position regarding judicial review would have given certainty of the Court’s role and its impact on issues such as privative clauses. In Canada, in embracing the pragmatic and functional approach and thus explicitly rejecting the constitutionality of privative clauses, there is no such conceptual confusion. New Zealand, on the other hand, has drifted in its approach such that leading commentators now fear for the rule of law. This is not a minor consequence, but New Zealand’s current approach does not have the capacity to demand that the Court address it. Smillie’s model would have prevented the issue, and, had it somehow reared its head, demanded swift resolution, and this way, represents a much better approach than the status quo.

While much more could be written on what New Zealand’s administrative law landscape would have looked like had we followed Smillie’s blueprint for a pragmatic and functional approach to judicial review, the two issues above highlight the advantages it would have brought. Smillie was not naïve, and did not assume his model would cure all of administrative law’s complexities. As the Canadian experience shows, it is far from perfect: different difficulties and issues would doubtless arise under any model. Indeed, two of the disadvantages that Smillie highlighted 30 years ago remain extant: the scope of ‘questions of law’, for example, are still notoriously difficult to define,194 and although for years a dormant issue, the definition of an inferior court has recently become a vexed issue.195 These, and doubtless many other issues, would have caused difficulties in the application of Smillie’s approach.

However, what these issues also show is that New Zealand’s current approach to administrative law is noteworthy for lack of cohesion and direction. While Lord

194 See the latest attempt at attempting to define their scope and effect in New Zealand Transport Agency v Architectural Centre Incorporated [2015] NZHC 1991 and the discussion in Marcelo

Rodriguez Ferrere “The Functional Convergence of Appeal and Review” (forthcoming).
195 Andrew Beck “An Inferior Court?” [2012] NZLJ 93.

Cooke’s simple classification worked well while he was able to administer it as a judge on the New Zealand Court of Appeal, since the mid-1990s, New Zealand administrative law has drifted. We adhere to Cooke’s categorical approach in name but it seems incapable of providing a clear position on deference and intensity of review or to prevent the resurgence of effective privative clauses. The main advantage of Smillie’s model is not the answers it would have provided to these issues (although they would have been effective) but the fact that it provided answers at all. Adopting an approach such as Smillie’s would have required consistent calibration of its form and effect at the highest judicial level, meaning our top appellate courts would have provided much more guidance on the meta-structure of administrative law than the two occasions in the past fifteen years. The main disadvantage of Cooke’s categorical approach is the same as its advantage: its simplicity means that it did not require consistent re-examination in the Canadian fashion, but that has allowed it to drift to the point of becoming unfit for purpose.


In the 30 years since Smillie provided his pragmatic and functional approach, much has changed, and much has stayed the same. For example, the 1984 issue of the Otago Law Review, which published Smillie’s reply to Bulk Gas, included an advertisement for a photocopier that could print in blue and brown (“other colours available soon”) and an advertisement for DB Export beer. While the advertising clientele of the Otago Law Review may have changed over the years, the personnel involved have not. Many of the contributors to the 1984 issue are now my colleagues or retain a strong link to Otago. The issue itself was dedicated to the retiring Dean of the Faculty, Emeritus Professor Peter Sim – and thus it is perhaps fitting that Smillie’s contribution to that issue has in turn provoked me to write this paper in his honour.

The same is true for administrative law. While the contours of the subject have only a passing resemblance to their 1980 counterparts, it remains just as complex. What this paper has endeavoured to do is show that Smillie had a solution to that complexity, and had we implemented it, it might have rescued us from the issues we face today. It first outlined his approach, built on two responses by Smillie to leading cases on the issue of jurisdictional error. In his response to Anisminic, Smillie argued that any analysis on the distinction between jurisdictional and non-jurisdictional error was futile, and that our time was better spent articulating and standardising the pragmatism used by Courts in deciding how and why to intervene in judicial review. His pragmatic approach provided a series of factors used to determine that question, which would allow the Court to make its decision-making process transparent and justifiable. In his response to Bulk Gas,

Smillie further refined this approach, providing standards of review that would be the product of this pragmatic approach, and which were based on the proper function of the Court given the relative expertise of the decision-maker under review.

The second part of this paper was dedicated to a forecast of what New Zealand might have experienced if it had adopted Smillie’s blueprint. Canada provides some clues in this regard, given it largely adopted what Smillie’s approach advocated. Although Canada’s experience – and the evolution of its pragmatic and functional approach – is not short of complexity or criticism, that its Supreme Court routinely reassesses that approach is itself a core advantage. New Zealand’s decision to take the other direction by adopting and implementing the simple categorical approach of Lord Cooke has, in contrast, run into difficulty. In Lord Cooke’s absence, the model has drifted, and the Supreme Court rarely re-examines the constitutional basis for judicial review and how we approach administrative law in this country. The result has been an inconsistent and unprincipled approach to issues such as intensity and privative clauses. These issues simply could not have arisen under Smillie’s approach, and while one does not have to approve of its solutions, it is the fact that it provides a solution at all that makes it such a superior alternative.

It is too late to switch tracks, nor is there any impetus to do so: there is no crisis in administrative law of the sort that Anisminic provoked that will create the same interest in the area that existed in the 1980s and thus no urgency to reconsider its core principles. This paper is thus a long-form lament: the current inertia of New Zealand’s approach to administrative law and the issues that inertia is creating were entirely avoidable had this jurisdiction paid a little more attention to one Professor John Alexander Smillie.

NZLII: Copyright Policy | Disclaimers | Privacy Policy | Feedback