R v Investigatory Powers Tribunal and others

The Investigatory Powers Tribunal (“IPT”) is a special tribunal established under the Regulation of Investigatory Powers Act 2000 (“RIPA”) with jurisdiction to examine, among other things, the conduct of the Security Service, the Secret Intelligence Service and the Government Communications Headquarters (“the intelligence services”). Section 67(8) provides:

There is an obvious parallel with the “ouster clause” considered by the House of Lords in the seminal case of Anisminic Ltd v Foreign Compensation Commission [1969] 2 AC 147 (“Anisminic”). Section 4(4) of the Foreign Compensation Act 1950 provided:

Reduced to its core the central issue in the present case is: what if any material difference to the court’s approach is made by any differences in context or wording, and more particularly the inclusion, in the parenthesis to section 67(8), of a specific reference to decisions relating to “jurisdiction”?

The legislative scheme established by RIPA replaced three earlier statutes dealing with the oversight of the security services. Its enactment was closely linked to that of the Human Rights Act 1998 (“HRA”), which was brought into force at the same time. The Explanatory Notes stated (paras 3-4):

The statutory provisions governing the composition, jurisdiction and procedures of the IPT are complex. There is a comprehensive account in the judgment of Sir Brian Leveson P in the Divisional Court (paras 5 to 15) ([2017] EWHC 114 (Admin)). For present purposes it is enough to note the principal features. Section 65(1) and Schedule 3 deal with its composition. The number of members is set by Her Majesty by Letters Patent (section 65(1)). The President must have held high judicial office, and the other members must have held high judicial office or meet specified legal qualifications. In the present case the tribunal consisted of five members presided over by Burton J (President) and Mitting J (Vice-President), the others all being leading counsel.

As to its jurisdiction section 65(2) provides:

The remainder of section 65 provides further details of these four categories of jurisdiction, which are not material in the present case. It is to be noted that the jurisdiction of the IPT may depend on uncertain issues of law or fact. For example, in C v The Police IPT/03/32/H the IPT considered a complaint by a retired police officer alleging that there had been unlawful covert surveillance in breach of article 8 of the European Convention on Human Rights (“the Convention”) by his former police force. The essential facts were agreed, but the IPT held that it had no jurisdiction to consider his complaint because there was no “directed surveillance” which satisfied the definition of conduct to which Chapter II of Part I of RIPA applied (see para 74 of the determination; RIPA section 65(3)(d), (5)(c)). In some cases the jurisdiction of the IPT may overlap with that of the ordinary courts: see, for example, AKJ v Comr of Police of the Metropolis [2014] 1 WLR 285 (parallel claims under HRA section 7, and in tort, in respect of damage suffered as a result of the actions of two undercover police officers).

Section 67 is headed “Exercise of the Tribunal’s jurisdiction”. Subsection (1) provides that it shall be the duty of the Tribunal to hear and determine proceedings, or to consider and determine complaints or references, brought before it under section 65(2). Subsections (2) and (3) provide, among other things, that the tribunal shall apply “the same principles as would be applied by a court on an application for judicial review”. Subsections (4)-(6) make provision in relation to frivolous and vexatious claims, limitation, and the power to make interim orders. Subsection (7) sets out the powers of the tribunal “on determining any proceedings, complaint or reference” to make “any such award of compensation or other order as they think fit”. It also gives examples of such orders, including (a) “an order quashing or cancelling any warrant or authorisation”, and (b) “an order requiring the destruction of any records of information which (i) has been obtained in exercise of any power conferred by a warrant or authorisation; or (ii) is held by any public authority in relation to any person”.

Subsection (8) has been set out above (para 1). As there seen, it allowed for an appeal to be provided for by order of the Secretary of State, but that power has never been exercised. Subsection (9) goes further, imposing a “duty” on the Secretary of State to secure an order allowing for an appeal to a court against any exercise by the tribunal of their jurisdiction under section 65(2)(c) or (d); but that subsection has not been brought into force. Subsections (10)-(12) make provision as to the contents of, and procedure for making, such an order were the power ever to be exercised. After the commencement of these proceedings there was enacted (by section 242 of the Investigatory Powers Act 2016) a new section 67A providing for an appeal on a point of law to the Court of Appeal or Court of Sessions against certain decisions of the tribunal. That was brought into force on 31 December 2018 by regulation 2 of the Investigatory Powers Act 2016 (Commencement No 10 and Transitional Provision) Regulations 2018/1397, but it does not apply to any decision or determination of the IPT made before this date. It is therefore not material to the present appeal.

Until 31 December 2018, when they were replaced by the Investigatory Powers Tribunal Rules 2018 (SI 2018/1334), the procedure before the IPT was governed by the Investigatory Powers Tribunal Rules 2000 (SI 2000/2665) (made under section 69(1)). Notable are the power to conduct proceedings in private and at certain stages in the absence of the complaining party (rule 9), and the duty under rule 6(1):

The background of the present proceedings was described in the judgment of the IPT dated 12 February 2016. It was a hearing of preliminary issues of law, whose purpose was to establish:

The issue arises under section 5 of the Intelligence Services Act 1994 which empowers the Secretary of State to issue a warrant “authorising the taking of such action as is specified in the warrant in respect of any property so specified” if he considers, among other things, that such action is necessary for the purpose of assisting the intelligence services in carrying out their functions. According to the appellant’s case, the significance of that provision became apparent when the Intelligence Services Commissioner (Sir Mark Waller) disclosed in his 2014 Report that the intelligence services were using it to authorise CNE activity, and expressed concern that this interpretation of the section might arguably be too broad. Their case before the tribunal was that section 5 did not permit the issue of so-called “thematic” warrants authorising activity in respect of a broad class of property. They argued, inter alia, that the section needed to be construed against the background of the long-established aversion of the common law to general warrants, recognised in cases going back to Entick v Carrington (1765) 2 Wils KB 275.

The tribunal heard inter partes oral argument at a public hearing in early December 2015, and gave judgment on 12 February 2016 dismissing the claim (Privacy International v SSFCA [2016] UKIP Trib 14_85-CH). Their discussion of the interpretation of section 5 comes at paras 31 to 47 of the judgment. In relation to the argument based on general warrants they said:

Sales LJ in the Court of Appeal [2017] EWCA Civ 1868; [2018] 1 WLR 2572 commented on the significance of this question:

On 17 June 2016 Lang J granted the appellant permission to apply for judicial review, while expressing doubts whether the High Court had jurisdiction to determine the substantive claim. She directed that the issue of jurisdiction should be heard as a preliminary issue. On 2 February 2017 the Divisional Court gave judgment answering that question in the negative for reasons given by the President. He held that section 67(8) prohibited judicial review of the decision. Since (by contrast with Anisminic) the tribunal was already exercising a supervisory jurisdiction over the actions of public authorities and exercising powers of judicial review, he saw no compelling reasons for insisting that a decision of the tribunal is not immune from challenge (para 42). Further, the legislation authorised the Secretary of State to create a right of appeal (albeit that the power had never been exercised), so that the presumption that Parliament “could not have intended to make a statutory tribunal wholly immune from judicial oversight” was not engaged (paras 43, 45).

Leggatt J, while not formally dissenting, was “inclined” to a different view. He thought that the case was governed by the reasoning in Anisminic:

The Court of Appeal gave judgment on 23 November 2017 dismissing the appeal. Sales LJ (with whom Floyd and Flaux LJJ agreed) considered that both the language and the context were materially different from Anisminic. As to the language he said:

In support of this view, he noted the “very high quality” of the IPT in terms of judicial expertise and independence (para 38), and the statutory context:

Finally he relied by analogy on the decisions of the Court of Appeal and Supreme Court in R (A) v Director of Establishments of the Security Service [2009] EWCA Civ 24; [2009] UKSC 12; [2010] 2 AC 1 (“R (A)”). It was held that section 65 of RIPA conferred on the IPT exclusive jurisdiction to hear claims under section 7 of the HRA against any of the intelligence services. In the Court of Appeal Dyson LJ noted that the rules were carefully drafted to achieve “a balance between fairness to a complainant and the need to safeguard the relevant security interests”; he thought it “inherently unlikely” that Parliament, having provided for such an elaborate set of rules to govern proceedings against an intelligence service, “yet contemplated that such proceedings might be brought before the courts without any rules” (para 48). That approach was approved in the Supreme Court. Although the effect of section 67(8) was not in issue, Lord Brown, giving the leading judgment, in the course of a review of this part of the Act, spoke of it as “an unambiguous ouster” of the court’s jurisdiction.

While accepting that this expression of view was obiter Sales LJ considered it to fit closely with Lord Brown’s analysis of the regime. He added:

Two issues are identified in the agreed statement:

whether section 67(8) of RIPA 2000 “ousts” the supervisory jurisdiction of the High Court to quash a judgment of the Investigatory Powers Tribunal for error of law?

whether, and, if so, in accordance with what principles, Parliament may by statute “oust” the supervisory jurisdiction of the High Court to quash the decision of an inferior court or tribunal of limited statutory jurisdiction?

On the first issue, counsel for the appellant led by Dinah Rose QC (with Professor Sir Jeffrey Jowell QC) rely principally on the long-established principle that a statute should not be interpreted as ousting judicial review of a statutory tribunal of limited jurisdiction if there is a tenable construction which would preserve the supervisory jurisdiction of the High Court. In the present case, the formula used in section 67(8) is not materially different from that which the House of Lords held not to oust judicial review in Anisminic, and must be taken to have the like effect. Lord Brown’s comment in R (A) was obiter, against the background of a concession by the appellant that judicial review was not available (R (A) at p 23D).

This did not mean that the reference in parenthesis to “jurisdiction” was without effect. Ms Rose refers for example to the distinction drawn in the cases depending on whether the legislature has or has not entrusted to the tribunal the power to determine the existence of “the preliminary state of facts” necessary to its jurisdiction (see R v Comrs for Special Purposes of the Income Tax (1888) 21 QBD 313, 319, per Lord Esher MR). At the time of the drafting of what became the 1985 Act, the difference between issues of fact and law in the context of jurisdiction had been highlighted by the House of Lords in R v Secretary of State for the Home Department, Ex p Khawaja [1984] AC 74. It was held that the power of the Home Office to remove an “illegal entrant” did not depend simply on the reasonable belief of the immigration officer that the person was an illegal entrant. As Lord Scarman said (at p 110):

She submits that the Court of Appeal’s reliance on the security and intelligence context to support their view was mistaken. The High Court has ample powers on judicial review to ensure that sensitive information is protected (see now R (Haralambous) v Crown Court at St Albans [2018] UKSC 1; [2018] AC 236). Similar concerns might have been said to arise in respect of the Special Immigration Appeals Commission (SIAC), but neither that factor, nor the relatively high status of the judges of that Commission, was held sufficient to oust judicial review (R (U) v SIAC [2011] QB 120 DC, paras 82-86 per Laws LJ).

Furthermore, the tribunal’s jurisdiction is not limited to sensitive claims against the intelligence services. Ms Rose gives as examples such issues as the use by local authorities of CCTV and checks by directed surveillance on whether a child lives in the catchment area of a local school. She also points out that the issue can work both ways. For example, if the Tribunal were unlawfully to order the security and intelligence services to disclose material (such as the identity of an agent) which would risk harming national security, they would have no remedy. More generally, it cannot have been intended that the IPT should be immune from challenge even where it blatantly disregarded limits to its powers: for example, if it decided not to follow a binding decision of the Supreme Court on the interpretation of the RIPA, or if it purported to determine a claim for unfair dismissal allocated by statute exclusively to the Employment Tribunal.

These submissions were supported by Mr Chamberlain QC (with Mr Heaton), appearing for the intervener Liberty. He emphasised the very broad jurisdiction of the IPT, not limited to reviewing the conduct of the intelligence services, but extending to surveillance and other activities undertaken for policing, economic and other purposes by a range of public authorities. It is, he submitted, objectionable in principle, and inimical to the rule of law, that a body with such broad jurisdiction should be entirely immune from challenge, save only in the Strasbourg court in respect of compliance with the Convention. He also pointed to the considerable overlap between the jurisdiction of the IPT and that of the ordinary courts. As he submits it could be “a question of happenstance” whether a determination on a particular issue is immune from review by the UK courts, if determined by the IPT, or subject to appeal through the appellate courts, if determined by the ordinary courts.

Counsel for the interested parties, led by Sir James Eadie QC, generally supported the reasoning of the Court of Appeal. In agreement with Sales LJ, he submitted that the language of section 67(8), by the words in parenthesis, and in contrast to the section under consideration in Anisminic, was designed in terms to address the possibility of the IPT making an error of law going to its jurisdiction or power to act. He drew attention to particular features of the statutory context, including the special allocation of judicial responsibility to the IPT in the national security context under a single legislative regime together with the HRA; the “bespoke” nature of the IPT system set up by RIPA, with provision to make its own rules and procedures, allowing the IPT to deal with sensitive national security matters through closed material procedures not available at common law; the placing of the IPT on equal footing with the High Court in respect of judicial review; and the provision for the possibility of a right of appeal from IPT decisions in specified cases. The fact that the latter provisions had not been brought into effect did not detract from their relevance to the presumed intention of Parliament at the time of enactment (see Bennion on Statutory Interpretation (6th ed), p 654).

He relied also on the continuing endorsement of the IPT by the Strasbourg court, most recently in Big Brother Watch v United Kingdom (2018) (Application Nos 58170/13, 62322/14 and 24960/15), in which the court commented for example on the special role of the IPT as -

More generally he submitted that there was nothing constitutionally offensive about legislative arrangements whereby Parliament reallocates the High Court’s judicial review jurisdiction to a judicial body that is both independent of the Executive and capable of providing an authoritative interpretation of the law.

Before considering these submissions, it is necessary to set them in the context of the historical development through the authorities of the relationship between the High Court and other adjudicative bodies, culminating for present purposes in the Supreme Court judgments in Cart.

The supervisory role of the King’s court (curia regis), or the King’s or Queen’s Bench Division of the High Court as it became, has a long history. A scholarly account, tracing it back to the time of William I, is given in the judgment of Laws LJ in the Divisional Court in Cart ([2011] QB 120, paras 44ff). As he says (para 45) the King’s Bench was established by the end of the 13th century, and remained at the centre of the English judicial system until its powers were transferred to the High Court in 1873.

As to its status, he cites, for example (paras 48-49), Groenwelt v Burnell (1700) 1 Salk 144, 90 ER 1000 per Holt CJ:

That supervisory role was preserved by section 16 of the Judicature Act 1873 which vested the common law powers of the Queen’s Bench in the newly created High Court. Those powers were in turn preserved by section 19 of the Senior Courts Act 1981.

Authorities dating back at least to the 17th century (see eg Smith, Lluellyn v Comrs of Sewers (1669) 1 Mod 44, 86 ER 719) leave no doubt as to the hostile attitude of the High Court to attempts by statute to restrict its supervisory role. In such cases, conventional principles of statutory interpretation, based on the ordinary meaning of the words used by Parliament, have yielded to a more fundamental principle that no inferior tribunal or authority can conclusively determine the limits of its own jurisdiction.

It is difficult, for example, to think of a statutory ouster clause in clearer terms than that considered in R v Cheltenham Comrs (1841) 1 QB 467, 113 ER 1211. The case concerned a challenge to a decision of the Quarter Sessions on an appeal against a rate set by the respondent Commissioners. The Commissioners’ objection to the admission of certain evidence had been rejected by a majority of 11 magistrates to eight. The decision was challenged on the grounds of apparent bias (in modern terms), in that three of the 11 magistrates were partners in a company which owned a property affected by the rate. The statute provided -

A possible justification of that principle was given by Farwell LJ in R v Shoreditch Assessment Committee, Ex p Morgan [1910] 2 KB 859, 880:

More recent authority has affirmed the continuing relevance of this strong interpretative presumption against the exclusion of judicial review, other than by “the most clear and explicit words” (Cart [2011] QB 120, para 31, per Laws LJ; citing Denning LJ in R v Medical Appeal Tribunal, Ex p Gilmore [1957] 1 QB 574, 583, and Lord Phillips MR in R (Sivasubramaniam) v Wandsworth County Court [2003] 1 WLR 475, para 44). As those cases show, this presumption has been applied without distinction to decisions of inferior courts (such as the County Court) and of tribunals, even if designated as “superior courts of record” (like the Upper Tribunal).

In so far as those authorities were concerned with errors going to jurisdiction in the traditional sense, they were relatively uncontroversial. The review of errors of law was more problematic. Professor Paul Craig (Administrative Law 8th ed (2016), para 16-001) identifies three phases in the extension of the inherent powers of the High Court to review of decisions for error of law:

Before doing so, I should note the important difference between control respectively of administrative action and of the functions of lower courts or tribunals. As Robert Craig points out in an article in Public Law (“Ouster clauses, separation of powers and the intention of Parliament” [2018] PL 570, 572), separation of powers dictates that administrative bodies should not be determining “the answers to questions of law that frame their decision-making process”. He cites the succinct statement of the “proper constitutional relationship of the executive with the courts” by Nolan LJ in M v Home Office [1992] QB 270, 314H-315A:

Accordingly, where an ouster clause can be interpreted as not excluding judicial supervision but reallocating its exercise to a different form of court or tribunal, a different constitutional analysis may be required. That indeed was the view taken of RIPA section 65(2)(a) (in relation to claims under the HRA) by Lord Brown in R (A) para 23, when dismissing a suggested parallel with Anisminic:

The claim in Anisminic arose from the sequestration of the claimants’ property by the Egyptian authorities at the time of the Suez crisis in 1956. Their claim under the relevant Foreign Compensation Order (under the Foreign Compensation Act 1950) was rejected by the Commission on the grounds that they had subsequently sold their property to an Egyptian institution, which was to be regarded as their “successor in title” within the meaning of the Order. On its face, at least to modern eyes, that was a straightforward issue of interpretation of the Order. It was ultimately decided in the claimants’ favour by the House of Lords.

However, the process by which the case arrived at that point was far from straightforward. The procedural and legal background is described in an illuminating discussion of the case by Professor Feldman (Anisminic in perspective, in Juss and Sunkin (ed) Landmark cases in public law (Oxford 2017) pp 63ff). He explains in particular (p 70) the significance, in the absence of a reasoned decision by the Commission, of the choice of an action for a declaration, rather than certiorari:

There were differences of emphasis between the various speeches in the House of Lords. However, for modern purposes they are less important than the interpretation of the decision in later cases. Looked at from that perspective, the case can be taken as confirming or establishing three distinct but related propositions:

That there is (at the least) a strong presumption against statutory exclusion of review by the High Court of any decision of an inferior court or tribunal treated as made without jurisdiction and so a “nullity”.

That for this purpose there is no material distinction between an excess of jurisdiction at the outset, and one occurring in the course of proceedings.

That a decision which is vitiated by error of law (whether or not “on the face of the record”) is, or is to be treated as, made without jurisdiction and so a nullity.

The first proposition, as apparent from the cases referred to above, was little more than a confirmation of well-established principles. The second could be seen as a logical step forward, or at least a clarification of the previous law (see Wade & Forsyth Administrative Law 11th ed, p 217: “The ‘original jurisdiction’ fallacy”).

The third, however, was a much more radical development, in so far as a mere error of law came in due course to be treated as an excess of jurisdiction which rendered the decision not simply open to legal challenge but “void” or a “nullity”. Hitherto it had generally been assumed that a mere error of law by a court or tribunal in respect of an issue otherwise falling within its jurisdiction might be subject to correction on appeal, but did not take the decision outside its powers. For example, Lord Reid himself had said in a recent case:

Consideration of Lord Reid’s judgment is best begun by reference to his own summary of the respective arguments and his response to them (pp 169-170):

He went on to give a list of examples of the ways in which a decision of a tribunal, acting within its original jurisdiction may be treated as a nullity:

It is noteworthy that the list did not include a simple error of law or misconstruction of the statute. It must have been a misconstruction of the “provisions giving it power to act”, as a result of which it has “decided some question which was not remitted to it” or “based its decision on some matter which … it had no right to take into account”. Later in the speech he applied that approach to the instant case. The Commission had construed the order (wrongly as it was held) as requiring them, having identified the original owner, to inquire whether he had a successor in title. Lord Reid explained how that error of law could render the decision a “nullity”:

Lord Pearce adopted a similar approach. As he put it:

Lord Wilberforce, who gave the only other substantive speech on this issue, also looked for something beyond a simple error of law. This was against the background that, as he put it:

As Professor Feldman observes (op cit pp 92-93), the significance later attached to the decision in Anisminic, and in particular to the statement in Lord Reid’s judgment of the matters leading to “nullity”, may not have been apparent at the time:

Whatever doubts there may have been initially or since as to the interpretation or practical implications of Lord Reid’s words, and of the other majority speeches in Anisminic, such doubts have been dispelled by a series of statements in subsequent cases at the highest level, led by Lord Diplock. Professor Feldman (p 94) notes that at an early stage Lord Diplock’s own views of the case, as expressed extra-judicially, had progressed from a relatively cautious response in 1971 to a much more absolute view, as expressed in a 1974 lecture, that the decision had -

The problem is that this move outside the limitations carefully set by the Anisminic speeches may have undermined much of their conceptual basis. I shall return to this problem when addressing the second issue.

It was not until 1982 that the broader view was given unambiguous judicial endorsement by the House of Lords, when Lord Diplock summarised the effect of Anisminic in O’Reilly v Mackman [1983] 2 AC 237, 279) in a speech agreed by the other members of the House:

Later cases have confirmed this interpretation. Thus in R v Hull University Visitor, Ex p Page [1993] AC 682, concerning a challenge to the decision of a University Visitor, Lord Browne-Wilkinson said (at pp 701-702):

To similar effect, in Boddington v British Transport Commission [1999] 2 AC 143, 158D-E, which related to the validity of a by-law, Lord Irvine LC said:

More recently, Lord Irvine’s words were in turn cited by Lord Dyson (Lumba v Secretary of State for the Home Department [2011] UKSC 12; [2012] 1 AC 245, para 66) to support the statement:

It must be acknowledged in any event that the Anisminic principle, however defined, has not been treated as necessarily applicable outside its particular context. An example is In re McC [1985] AC 528. The issue was whether the justices, in deciding to detain a juvenile without first informing him of his right to legal aid, had acted “without jurisdiction or in excess of jurisdiction” within the meaning of section 15 of the Magistrates’ Courts (Northern Ireland) Act 1964, so as to remove their immunity from civil liability for false imprisonment. Although this question was answered in the affirmative on the facts of the case, this did not depend on any application of Anisminic. Lord Bridge commented on the “many different shades of meaning in different contexts” acquired by the word “jurisdiction”, noting “at one end of the spectrum” the majority decision in Anisminic (p 536B-H). He regarded it as irrelevant to the section before him, however valuable it might be in ensuring that “the supervisory jurisdiction of the superior courts over inferior tribunals is effective to secure compliance with the law …” (p 546G).

In considering the development of the law since Anisminic it is necessary to take account both of the major changes in the supervisory role of the High Court as respects public bodies of all kinds (both administrative and judicial or quasi-judicial), including the 1979 reforms which established judicial review in its modern form; and also of the changes in the relationship between the traditional courts and specialist tribunals.

The development in this period of judicial review is well described in De Smith op cit para 4-006-7:

Anisminic also proceeded against the background of a reasonably clear division in the legal hierarchy between, on the one hand, the unlimited supervisory jurisdiction of the High Court, exercised by the Divisional Court usually presided over by the Lord Chief Justice, and, on the other, the limited jurisdictions of inferior courts or tribunals (or other adjudicative bodies, such as the Foreign Compensation Commission). For this purpose, no distinction was drawn in the authorities between the different forms of limited jurisdiction, or in particular between courts below the High Court and statutory tribunals. However, the period between Anisminic and the decision of the Supreme Court in Cart saw major changes in this traditional relationship between the High Court and other adjudicative bodies.

It is convenient at this point to refer to the decision of the House of Lords in In re Racal Communications Ltd [1981] AC 374 (“Racal”), to which Lord Sumption attaches some importance. As I understand it, he sees it and related cases as illustrating the proposition, which he derives from Lord Wilberforce’s speech in Anisminic (at p 207), that the key issue when considering the scope of an ouster clause is to define the “the permitted field” of the relevant adjudicative body, that being identified by a careful analysis of the interpretative power conferred by the enabling Act.

Racal itself concerned a challenge to the decision of a High Court judge exercising a statutory jurisdiction (under the Companies Act 1948 section 441) to authorise inspection by the Director of Public Prosecutions of company books for the purpose of investigating a suspected offence. Section 441(3) provided that the decision of the High Court judge on such an application “shall not be appealable”. The judge had dismissed an application by the Director on legal grounds, but the Court of Appeal had reversed his decision holding that it was entitled to do so because he had made an error of law which went to his jurisdiction. The House of Lords allowed the company’s appeal.

The case has attracted some attention for the distinction drawn by Lord Diplock in the context of ouster clauses between, on the one hand, “administrative tribunals and authorities” and, on the other, “courts of law”. Having confirmed that as respects the former the decision in Anisminic had effectively abolished “the old distinction between errors of law that went to jurisdiction and errors of law that did not”, he continued:

As I see it, this distinction is of no assistance to the case of the Interested Parties, since there is nothing to suggest that Lord Diplock would have regarded the IPT, notwithstanding its distinguished composition, as anything more than an “administrative tribunal” within his classification. On that assumption Lord Diplock’s speech supports the widest reading of the Anisminic decision as later confirmed in O’Reilly v Mackman. In any event I do not see that part of his reasoning as having majority support. As I read the speeches overall, it was the latter point, turning on the position of the High Court, rather than of courts more generally, which provided the ratio of the House’s decision. Although Lord Keith agreed without qualification with Lord Diplock’s reasoning, his suggested distinction between courts in general and tribunals was not expressly endorsed by the other members of the House. Lord Edmund-Davies made no specific reference to this point. Lord Salmon spoke of the decision in Anisminic as confined to “decisions made by commissioners, tribunals or inferior courts”, drawing no distinction between them. He based his decision on the fact that the jurisdiction of the Court of Appeal was defined by statute, which gave it “no jurisdiction to make a judicial review of a decision of the High Court” (p 386). To similar effect, Lord Scarman (at p 393) relied on the fact that the Court of Appeal’s jurisdiction over the High Court was “the creature of statute”, and in no way analogous to the supervisory jurisdiction of the High Court over inferior tribunals.

Ms Rose goes as far as to submit that this part of the speech was “not only obiter but per incuriam”. There is force in this submission. As far as appears from the Appeal Cases report, the suggested distinction between courts and tribunals was not raised in argument and no relevant authorities were referred to in support, either by counsel or by Lord Diplock. His approach seems out of line with the long series of authorities cited by Laws LJ in his historical review in Cart, where it was emphasised that the jurisdiction of the King’s Bench Division extended to all inferior jurisdictions without distinction, including courts: a view well illustrated by Blackstone’s description of the range of courts within the scope of the writ of prohibition (paras 31-32 above). As Laws LJ said (commenting in terms on the speeches in Racal):

It is true that this part of Lord Diplock’s speech has been cited with approval in later cases: see per Lord Browne-Wilkinson in R v Hull University Visitor, Ex p Page [1993] AC 682, 703 (relating to University Visitors), and more recently per Lord Mance in Lee v Ashers Baking Co Ltd [2018] 3 WLR 1294, paras 85-87. But this point was not essential to the reasoning in either case. In Page, Lord Browne Wilkinson held that the High Court could not review a university visitor’s decision for a non-jurisdictional error of law. However, he did so on the basis of common law principles that treat the visitor as equivalent to an arbitrator designated by the internal governance arrangements of a university (p 607, citing Holt CJ in Philips v Bury, (1694) Holt 715, 723-726).

Lee also was concerned with a quite different issue: that is, the finality of the statutory appellate jurisdiction of the Northern Ireland Court of Appeal on appeal from the County Court. It had nothing to do with the exclusion of the original jurisdiction of the High Court in respect of inferior courts or tribunals. As Lord Mance explained (para 88) it turned on the construction of article 61(1) and (7) of the County Courts (Northern Ireland) Order 1980, which provided for the decision of the Court of Appeal on a case stated relating to “the correctness of ‘the decision of a county court judge upon any point of law’” to be final - wording which was “focused on the decision on the point of law, not on the regularity of the proceedings leading to it”. He therefore found no difficulty in holding that the exclusion would not extend to a challenge to the fairness or regularity of the court’s process.

Lord Sumption also attaches importance to the fact that in Racal there was majority approval (Lords Diplock, Keith and Edmund-Davies) of the dissenting judgment of Geoffrey Lane LJ in Pearlman v Keepers and Governors of Harrow School [1979] QB 56, 76C-D. The same passage had recently been cited with approval by the Privy Council (including Lord Edmund-Davies) in South East Asia Fire Bricks Sdn Bhd v Non-Metallic Mineral Products Manufacturing Employees Union [1981] AC 363.

In Pearlman the Court of Appeal by a majority allowed an appeal from the County Court in a case turning on the construction of a particular phrase (“structural alteration … or addition”) in the Housing Act 1974. This was in the face of a provision of that Act by which the decision of the County Court was to be “final and conclusive”, and a provision (section 107) of the County Courts Act 1959 that -

It is a difficult case, not least because the majority judgments seem to have borne little relationship to the arguments as presented. The scope of the argument, as Geoffrey Lane LJ noted, had been constrained by concessions made on either side (without specific reference to Anisminic): on the one hand by counsel for the landlord that the section did not affect the power of the High Court to quash a decision of the county court made “in excess of jurisdiction”, although it did exclude the power to quash for errors of law on the record; and on the other by counsel for the tenant that the particular decision had been made within the judge’s jurisdiction (see pp 60G, 62D, 74A, 76H). There seems therefore to have been only limited reference to Anisminic in argument (see p 64G).

However, in the leading judgment, Lord Denning MR took his own view of the construction of the no certiorari clause in the County Courts Act, holding that it applied only to decisions under jurisdiction conferred by that Act (p 68H). He also took the opportunity for an extended discussion of the difficulties of interpretation arising from Anisminic, concluding:

It was in this context (evidently in response to the judgments of his colleagues rather than the arguments of counsel) that Geoffrey Lane LJ reviewed the speeches in Anisminic, and concluded (in the passage later cited with approval by the Privy Council and the House of Lords):

These references do not in my view materially assist the arguments in the present case on either side. All three cases (Pearlman, South East Asia and Racal) were products of their time. They came at a relatively early stage in the evaluation by the courts of the Anisminic principle. They also reflected a degree of tension between different levels of the judiciary as to the way forward. In this respect Lord Denning’s proposed interpretation seems closer to subsequent authority than that of the dissenting judgment, although his interpretation of the ouster clause seems more questionable, and the minority view might be supported on other grounds (discussed below, under the second issue).

By the time of Racal it was in any event difficult to make a principled distinction between courts and tribunals by reference only to nomenclature. Parliament had already blurred the distinction when establishing in 1975 the Employment Appeal Tribunal presided over by a High Court judge, with a jurisdiction limited to appeals on points of law (Employment Protection Act 1975 sections 87-88). It was designated in terms as “a superior court of record” (Schedule 6 paragraph 10). In this respect as in others it followed the precedent of the National Industrial Relations Court (Industrial Relations Act 1971 Schedule 3 paragraph 13).

As is apparent from the authorities cited by Laws LJ in Cart (paras 61-62) the accepted wisdom for many years, indeed until the decision of the Divisional Court in that case, was that such designation as a superior court of record was in itself sufficient to exclude judicial review by the High Court. He cites, for example, R v Regional Office of the Employment Tribunals (London North), Ex p Sojirin (unreported) 21 February 2000, in which Sedley LJ (with whom Brooke LJ and Sir Christopher Staughton agreed) stated:

It was not until Cart itself in the Divisional Court that this view of designation as a “superior court of record” was rejected as “a constitutional solecism”, when set against the principle that the supervisory jurisdiction of the High Court (if it can be ousted at all) can only be ousted “by the most clear and explicit words” (para 37 above). Laws LJ did however accept that the Upper Tribunal was “for relevant purposes, an alter ego of the High Court”, and that as such it:

This led him to propose a limited form of judicial review not extending to a mere legal mistake by the tribunal “within the field ascribed to it”. As will be seen, this solution was not adopted by the Supreme Court. However, the proposition that designation as a superior court was sufficient in itself to exclude judicial review was not further pursued in this appeal. Laws LJ’s rejection of that proposition was accepted as correct by this court (see per Lady Hale para 30). I shall return to other passages in Laws LJ’s judgment in the context of the second issue in the appeal.

The process of refinement of the Anisminic principle discussed above raised serious questions as to the need for continued reliance on concepts such as ultra vires or “nullity” as justifications for the intervention by the court. As Lord Reid himself had recognised (pp 170-171), the approach adopted in that case might be thought difficult to reconcile with a case such as Smith v East Elloe Rural District Council [1956] AC 736. That related to a statutory right within six weeks to challenge the confirmation of a compulsory purchase order on the grounds that it was “not empowered to be granted”, subject to which the order “shall not … be questioned in any legal proceedings whatsoever …”. It was held that the ouster clause was effective even where there was an allegation of fraud. In an often-cited passage, Lord Radcliffe (pp 769-770) commented on the argument that an order made in bad faith was a “nullity” and therefore incapable of having any effect:

In R v Secretary of State for the Environment, Ex p Ostler [1977] QB 122 the Court of Appeal held that it was bound by this decision; the availability of a statutory right to challenge within a specified time-limit, among other points, provided a sufficient basis for distinguishing Anisminic. This case also provides an interesting example of the influence of successive Junior Treasury Counsel in moulding the law (noted by Professor Feldman: above para 52 above). Lord Denning MR treated the statutory expression “not within the powers of this Act” as in effect embracing the familiar Wednesbury grounds, including error of law. This approach, following that of Lord Radcliffe (dissenting) in Smith v East Elloe Rural District Council, had been adopted by the Court of Appeal on the basis of a concession by counsel for the Minister (Nigel Bridge) in Ashbridge Investments Ltd v Minister of Housing and Local Government [1965] 1 WLR 1320; and in Ostler itself was not disputed by counsel for the Secretary of State (Harry Woolf). As Lord Denning said:

On the relevance of the concept of “nullity” as used in Anisminic, Professor Paul Craig (op cit para 16-015) refers to the extra-judicial observations of Sir John Laws (Illegality: The Problem of Jurisdiction in Supperstone and Goudie Judicial Review 1st ed (1992)):

I see considerable force in these observations, at least as applied to review for errors of law. Taking the present case, it is highly artificial, and somewhat insulting, to describe the closely reasoned judgment of this eminent tribunal as a “nullity”, merely because there is disagreement with one aspect of its legal assessment. I will return to this point later in connection with the second issue.

Professor Craig concludes that “the scope of judicial review is not self-defining” and it is “not capable of being answered by linguistic or textual analysis of the statute alone”. The critical question, he says, is “whose relative opinion on the relevant question should be held to be authoritative”, and the answer “must ultimately be based on a value judgment, the precise content of which will not necessarily be always the same” (para 16-016).

To similar effect, but adopting a different metaphor, the editors of De Smith 8th ed comment:

I now turn to the judgments in the Supreme Court in Cart, which in my view provide the essential background to the resolution of the issues in the present appeal.

The reform of the tribunals system, under the Tribunals, Courts and Enforcement Act 2007, effected a fundamental change in the traditional relationship between courts and tribunals. The background to the reforms, and the nature of the changes made by and under the Act, were described in Lady Hale’s leading judgment (agreed by the other members of the court). She noted in particular the new judicial structure, presided over by the Senior President (normally a Lord Justice of Appeal), and including not only the specialist tribunal judges, but also all the judges of the ordinary courts, up to and including the Court of Appeal (para 22). She also noted the designation by section 3(5) of the Upper Tribunal as a “superior court of record” (para 24); the “major innovation” of the power of the Upper Tribunal to exercise a jurisdiction equivalent to judicial review (para 25); and “probably most important” the right of appeal with permission to the Upper Tribunal from the First-tier Tribunal on points of law (para 26).

The principal decision under review in Cart was one by the Upper Tribunal (as it happens, presided over by myself as Senior President of Tribunals) giving the claimant only limited permission to appeal against a decision of the First-tier Tribunal relating to child maintenance. It was the common view of the courts at all three levels that the decision itself was unimpeachable on its merits, but the case gave the opportunity for detailed consideration of the principles which should govern such review in future cases. As Lady Hale explained (paras 31-34), both the Divisional Court and the Court of Appeal, albeit by slightly different legal routes, had held that, while judicial review was in principle available, its exercise should be narrowly confined.

Under the heading “The field of choice for this court” (paras 37ff), Lady Hale regarded three points as clear from the oral arguments: first, that there was nothing in the 2007 Act to exclude judicial review of unappealable decisions of the Upper Tribunal; secondly, that it would be inconsistent with the new structure to draw a distinction for this purpose between different jurisdictions there gathered together; and thirdly that:

Three possible approaches had been identified in the course of oral argument:

The first (“the ‘exceptional circumstances’ approach”) would she thought lead back to the distinction between jurisdictional and other errors which had been effectively abandoned after Anisminic. She saw a number of other objections. In particular (echoing in some ways the concerns of Blackstone in the 18th century: para 32 above) she saw a risk of specialist tribunals, in contrast to the “ordinary courts”, developing their own “local law”, a risk which was increased by the power of the Upper Tribunal to set precedent “often in a highly technical and fast moving area of law”, combined with its ability to refuse permission to appeal, and the likelihood that “the same question of law will not reach the High Court or the Court of Appeal by a different route”:

In respect of the second (“The status quo ante - but which?”) she noted that the courts had already adopted principles of “judicial restraint” when considering decisions of expert tribunals. She referred (inter alia) to her own comments (Cooke v Secretary of State for Social Security [2001] EWCA Civ 734; [2002] 3 All ER 279, paras 15-17) on the need for “appropriate caution” in giving permission to appeal from the Social Security Commissioners because of their “particular expertise in a highly specialised area of the law”, but observed that other contexts (such as asylum) might require a different approach. The real question was:

She concluded that the adoption of the second-tier appeals criteria would be a “a rational and proportionate restriction”, which would recognise that the new tribunal structure deserved:

Of the other judgments I note that Lord Phillips (paras 91-92) was a relatively late convert to the need for even a restricted form of judicial review:

Lord Brown, agreeing with the other judgments, saw nothing contrary to principle in the proposed limitation on the scope of review. As he said:

Lord Dyson, also agreeing with Lady Hale, noted that Parliament had not accepted the Leggatt recommendation to exclude judicial review. He commented:

Lord Dyson agreed with the Leggatt report and the 2004 White Paper that the “strategic reorganisation of the tribunals system” demanded a reappraisal of the scope of judicial review. Parliament having “refused to undertake it”, the task of deciding the scope of the judicial review jurisdiction, and the extent of any restrictions fell to be performed by the courts. Accepting that any restrictions called for justification, he said:

Finally, for completeness I note that all the justices expressed agreement with the corresponding reasoning of Lord Hope in the linked Scottish case of Eba v Advocate General [2012] 1 AC 710.

I have referred at some length to the judgments in Cart because they represented a major reappraisal of the approach of the supervisory functions of the High Court as respects specialist tribunals. The case has attracted some academic controversy (see the discussion in Joanna Bell, “Rethinking the Story of Cart v Upper Tribunal and its implications for Administrative Law” Oxford Journal of Legal Studies Vol 39 No 1 (2019) pp 74-99). For example, Professor Forsyth describes the reasoning of the court as “pragmatic but not principled”, adding:

Certainly the judgments show how far the law has evolved since the somewhat technical debates in Anisminic itself. In particular, against the background of the Divisional Court judgment, they reaffirm in no uncertain terms the continuing strength of the fundamental presumption against ousting the supervisory role of the High Court over other adjudicative bodies, even those established by Parliament with apparently equivalent status and powers to those of the High Court. Instead such status (as adjudicative bodies rather than executive agencies) is to be respected and taken into account, not by exclusion of review, but by the careful regulation of the court’s power to grant or refuse permission for judicial review. Furthermore, setting the principles by which that is to be done, in the absence of specific statutory provision, is a matter properly within the province of the appellate courts. I shall return below, under the second issue, to the question in what circumstances if any Parliament could exclude review altogether. There is no doubt that, if it is to be done, nothing less than the clearest wording will suffice (see De Smith para 1-027). As has been seen, that principle has been a central theme of the authorities since well before Anisminic, and was reaffirmed in emphatic terms by the Divisional Court in Cart (para 37 above).

The principle can be seen as an application of the “principle of legality” as explained by Lord Hoffmann in R (Simms) v Secretary of State for the Home Department [2000] 2 AC 115 at p 131F:

The practical importance of that principle was vividly illustrated by the fate of perhaps the most extreme form of ouster clause promoted by government in modern times: clause 11 of the Asylum and Immigration (Treatment of Claimants etc) Bill 2003. Not content with an express prohibition (in proposed clause 108A(1) and (2)) of any form of supervisory jurisdiction or questioning by the courts of tribunal decisions, the drafter had gone on to spell out precisely the intended consequence:

For completeness I should make clear that I have not overlooked the many authorities to which we have been helpfully referred from other common law jurisdictions, where similar issues have been discussed at the highest level, not always with the same results. All these decisions need to be read within the differing legal and constitutional arrangements of the jurisdictions concerned. For that reason, and without disrespect to the depth of learning and analysis there shown, I have not been persuaded that they add materially to the assistance available in the present context from the relevant domestic authorities and textbooks.

For example, the High Court of Australia has arrived at similar results by a broadened concept of jurisdiction (see Boughey and Burton Crawford “Reconsidering [Cart] and the rationale for jurisdictional error” [2017] Public Law 1). Thus, in Kirk v Industrial Court of New South Wales [2010] HCA 1, the court held that an ouster clause expressed in apparently far-reaching terms was ineffective to exclude review in relation to a particular error of law held to be “jurisdictional”. The court took an expansive view of the concept of “jurisdiction”, quoting the opinion expressed in an article by Professor Jaffé (“Judicial Review: Constitutional and Jurisdictional Fact” (1957) 70 Harvard Law Review pp 953, 963):

Against that background I can state my conclusions on the first issue relatively briefly. I remind myself of the terms of section 67(8):

As Ms Rose submits, our interpretation of the subsection, whether in its present form or as originally drafted in 1985, must be informed by the close parallel with the provision under review in Anisminic. At least by that date, following Lord Diplock’s explanation in O’Reilly v Mackman (1983), the drafter can have had no serious doubt about the far-reaching effect of that decision. A determination vitiated by any error of law, jurisdictional or not, was to be treated as no determination at all. It therefore fell outside the reference in the ouster clause to a “determination of the commission”. In other words, the reference to such a determination was to be read as a reference only to a legally valid determination.

On the other side, Sir James Eadie submits that the task of interpretation is to be approached, by reference, not simply to a general presumption against ouster clauses of any kind, but rather to careful examination of the language of the provision, having regard to all aspects of the statutory scheme, and the status or the body in question, in order to “discern the policy Parliament intended in the legislation” (R (Woolas) v Parliamentary Election Court [2012] QB 1, para 54 per Thomas LJ). The special character and functions of the IPT, combined with the specific references to decisions relating to “jurisdiction”, show a clear intention to protect it from any form of review by the ordinary courts, even in cases to which the Anisminic principle would otherwise have applied.

The main flaw in this argument, in my view, is that it treats the exercise as one of ordinary statutory interpretation, designed simply to discern “the policy intention” of Parliament, so downgrading the critical importance of the common law presumption against ouster. In that respect it echoes the unsuccessful argument of the Commission in Anisminic. Lord Reid did not dispute that the “plain words” of the subsection in that case were apt to exclude any form of challenge in the courts; but this ordinary meaning had to yield to the principle that such a clause will not protect a “nullity” and that there are “no degrees of nullity” (see paras 46-47 above). Following O’Reilly v Mackman the concept of “nullity” for these purposes is extended to any decision which is erroneous in law, and in that sense legally invalid. If one applies that approach to section 67(8), ignoring for the moment the words in parenthesis, the exclusion applies, not to all determinations, awards or other decisions whatever their status, but only to those which are “legally valid” in that sense. Thus, if the IPT’s decision in the present case were found to have been reached on an erroneous interpretation of section 5 of the Intelligence Services Act 1994, those words would not save it from intervention by the courts.

Does the specific reference to “decisions as to whether they have jurisdiction” make any difference? It would be odd if it did. As has been seen the relevant decision in this case raised a short point of law, turning principally on the reading of the word “specified” in section 5. On no ordinary view could it be regarded as a decision “as to whether [the IPT] had jurisdiction”, nor even as a decision “as to jurisdiction” under the apparently broader language of the 1985 Act. Although the arguments before us have proceeded on the basis that the change of wording made no material difference, the present wording seems designed if anything to emphasise that the exclusion is directed specifically at decisions about jurisdiction made by the IPT itself.

If, however, those words are read in the language of Anisminic there is no problem. The exclusion applies only to a legally valid decision relating to jurisdiction. In the parenthesis, as in the remainder of the subsection, a decision which is vitiated by error of law, whether “as to jurisdiction” or otherwise, is no decision at all. While of course respect is due to the contrary view expressed obiter by Lord Brown in the A case (para 19 above), the point was not in issue and there was no argument on it.

This does not necessarily mean that the words in parenthesis are otiose. As Ms Rose points out, at the time of the 1985 Act, the potential significance of the distinction between issues of fact and law, in the context of jurisdiction, had been highlighted by the House of Lords in R v Secretary of State for the Home Department, Ex p Khawaja [1984] AC 74. Closer to home the decision of the IPT itself in C v The Police (see para 7 above) is an example of the kind of decision as to jurisdiction which might well involve issues of fact to which the exclusion could be said to apply without engaging the presumption against ouster.

However, whether that is a likely interpretation of Parliament’s intentions, or indeed whether or not the parenthesis is redundant, is in my view beside the point. Judicial review can only be excluded by “the most clear and explicit words” (Cart, para 31). If Parliament has failed to make its intention sufficiently clear, it is not for us to stretch the words used beyond their natural meaning. It may well be that the promoters of the 1985 Act thought that their formula would be enough to provide comprehensive protection from jurisdictional review of any kind. (If so, as Lord Wilson observes, they would have gained support from the distinguished author of the notes to the 1985 Current Law Statutes.) But one is entitled to ask why they did not use more explicit wording. With O’Reilly v Mackman in mind, the natural focus of attention would have been, not on potential challenges to the tribunal’s own decisions “as to” jurisdiction, but on jurisdictional or legal challenges to its substantive decisions generally. A more explicit formula might perhaps have anticipated the extreme wording of the bill presented in 2003 (para 101 above), excluding challenges to any determination or “purported” determination as “a nullity by reason of lack of jurisdiction, error of law, or any other matter”. The reason for not adopting that course may simply be that, as in 2003, it might not have been expected to survive Parliamentary scrutiny.

So far as concerns the features of the IPT regime on which the Court of Appeal relied, I agree generally with Ms Rose’s responses (paras 24-25 above). I am unimpressed by arguments based on the security issues involved in many (though not all) of the IPT’s cases. As this case shows, the tribunal itself is able to organise its procedures to ensure that a material point of law can be considered separately without threatening any security interests. The Administrative Court can also ensure that the grant of permission is limited to cases raising points of general significance, and that its proceedings are conducted without risk to security concerns. Further, in the case of the IPT, the potential for overlap with legal issues which may be considered by the ordinary courts (see paras 7 and 14 above) makes it all the more important that it is not able to develop its own “local” law without scope for further review.

The second issue poses the question -

The essence of Ms Rose’s submission can be simply stated. It is in short that a clause purporting to “oust” the supervisory role of the High Court to correct errors of law cannot properly be upheld because it would conflict with the “rule of law”, a principle which is as fundamental to our constitution as the principle of Parliamentary sovereignty. She emphasises that she does not in this appeal seek to question the principle of Parliamentary sovereignty itself, but rather -

For the interested parties, Sir James Eadie does not question the need for an independent, authoritative interpreter of legislation, as “a fundamental requirement of the rule of law” (in his words); but he submits that the High Court is not the only body capable of performing that function. The IPT is of equivalent judicial status. Nor is there any absolute constitutional requirement for a right of appeal to the higher courts (see per Lord Brown in R (A), para 24). The balance between the correction of judicial error and the policy considerations in favour of finality is a judgement properly for the legislature.

Both parties find support in the judgment of Laws LJ in Cart [2011] QB 120, paras 36-40 (a passage cited with approval by Lady Hale in the Supreme Court: [2012] 1 AC 663, para 30). Ms Rose relies on his affirmation of the need under the rule of law for statute law to be “mediated by an authoritative source”, the “paradigm” being the High Court as “the principal constitutional guardian of the rule of law …”, and further that -

On the other side, Sir James relies on Laws LJ’s acceptance in Cart that Parliament could entrust that supervisory role to a body properly regarded as the “alter ego” of the High Court, such as in that case the Upper Tribunal (paras 39, 94). He also relies on Laws LJ’s acknowledgment in the same passage (at para 40) of Parliament’s power to “modify, sometimes radically the procedures by which statute law is mediated”; exemplified by the many cases in which Parliament has, without objection, replaced the High Court’s common law powers with time-limited statutory powers of review (as seen in Ex p Ostler: para 80 above). Thus, he says, there is no constitutional principle to preclude such legislative choices about which such judicial body is to have the power to make final decisions, and subject to what limitations.

As will be seen from these summaries, the area of disagreement even on the second issue is relatively narrow, but of considerable constitutional significance. It is useful to start by identifying those matters which appear not presently to be in dispute.

First, both parties start from the premise that the relationship between Parliament and the courts is governed by accepted principles of the “rule of law”. Unsurprisingly, there is no challenge to the proposition (per Lord Dyson, quoted at para 96 above) that there is -

Further, as noted earlier in this judgment (para 33), Parliament itself has affirmed (most recently in the Senior Courts Act 1981) the long-established common law powers of the High Court, as a court of unlimited jurisdiction, including its role in keeping inferior courts and tribunals within the law. Equally important, though of more recent origin, is the express statutory recognition of the “rule of law” in section 1 of the Constitutional Reform Act 2005. That provides:

In his introduction to The Rule of Law, Lord Bingham underlined the significance of section 1 of the 2005 Act to his general discussion of the concept. He attributed the absence of a statutory definition to the probable recognition by parliamentary counsel of the “extreme difficulty of devising a pithy definition suitable for inclusion in a statute”, and their wish instead to “leave it to the judges to rule on what the term means if and when the question arises for decision”, so enabling “the concept to evolve over time in response to new views and situations” (op cit pp 7-8). Whatever the explanation, Parliament having recognised this “existing constitutional principle”, and provided no definition, there is nothing controversial in the proposition that it is for the courts, and ultimately the Supreme Court (created by the same Act), to determine its content and limits.

Secondly, it is not I believe in dispute, and indeed was clearly established by the time of Anisminic, that there are certain fundamental requirements of the rule of law which no form of ouster clause (however “clear and explicit”) could exclude from the supervision of the courts. The first relates to what I would call “excess of jurisdiction”: that is, a decision arrived at by a tribunal of limited jurisdiction through a process which goes outside those limits whether at the inception or at any stage of the proceedings. On this category there was no disagreement in Anisminic. It is sufficient to quote Lord Morris, who dissented on the main issue:

The same approach can in my view be applied to what I would term “abuse of jurisdiction”: that is, a decision made within the limits prescribed by Parliament but in breach of basic principles governing the making of such decisions. In Anisminic, Lord Wilberforce spoke of -

I note that in the present case Sales LJ treated section 67(8) as excluding review in certain cases of alleged unfairness, such as a decision of the IPT as to whether a member of the panel should recuse himself for appearance of bias, or as to “what fairness or natural justice requires in relation to some aspect of its procedure”. He commented:

I do not read this passage as suggesting that the residual jurisdiction in respect of the rules of natural justice would be wholly excluded, however gross the breach. Nor did I understand Sir James to go so far. There are useful parallels with Laws LJ’s discussion (Cart, paras 96ff) of the “exceptional circumstances” in which a county court judgment may be reviewable, under the principles established in R (Sivasubramaniam) v Wandsworth County Court [2002] EWCA Civ 1738; [2003] 1 WLR 475. As Laws LJ said, complaints of unfairness in the course of a hearing are “legion, and very various”, and need to be approached with caution. Citing an earlier judgment of his own (R (Strickson) v Preston County Court [2007] EWCA Civ 1132, para 32) he drew a distinction between a case where the judge “simply gets it wrong” and one where the judicial process itself has been “frustrated or corrupted”, including “substantial denial of the right to a fair hearing” or in other words “a wholly exceptional collapse of fair procedure: something as gross as actual bias on the part of the tribunal”.

In the Court of Appeal in the present case there appears to have been no reference to this part of Laws LJ’s judgment, nor to the kind of cases that he was discussing. However, I would be surprised if Sales LJ, even on his interpretation of section 67(8), would have treated it as excluding altogether the possibility of review in such extreme cases, however unlikely they might be in a tribunal of the distinction of the IPT. In accordance with established principles, the ultimate safeguard of judicial review remains essential if the rule of law is to be maintained. The special status of the IPT (like that of the Upper Tribunal) may be a reason for restricting the grant of permission for judicial review, but not for excluding it altogether.

I turn to the area of dispute in the present appeal. We are concerned, not with excess or abuse of jurisdiction in any of the senses discussed above, but with a straightforward question of legal interpretation on a point directly within the apparent scope of the IPT’s statutory remit. There is no disagreement as to the need for independent judicial interpretation of such a question. The dispute is as to the power of the legislature, consistently with the rule of law, to entrust that task to a judicial body such as the IPT, free from any possibility of review by the ordinary courts (including the appellate courts).

In that respect in my view the discussion needs to move beyond the legal framework established by Anisminic, and the cases which followed it. Those decisions established the principle that, if a tribunal goes wrong on any such question of law, it exceeds its jurisdiction, so that the resulting decision is a “nullity”, and (subject to an effective ouster clause) liable to be so declared by the High Court. Although this principle is now unquestioned, its conceptual basis, 35 years on from O’Reilly v Mackman, remains obscure. Anisminic itself does not assist. As has been seen (paras 48-50), the majority judgments did not go as far as to treat simple mistake of law as a cause of nullity. For Lord Reid the mistake must have been such as to lead the tribunal to “(fail) to deal with the question remitted to it, or to (decide) some question which was not remitted to it”. For Lord Wilberforce the mistake must have related to the “architecture” of the provisions. They were both able, after elaborate analysis, to find that the Commission’s mistake of law satisfied these tests. Whether the same could be said of the alleged error in the present case is at best doubtful.

As already observed, the “nullity” analysis seems highly artificial, as applied to a legal decision such as that of the IPT in this case. It is also difficult to reconcile with the acceptance by the courts of the familiar statutory ouster of challenges to planning and similar decisions after a six-week time-limit (as in Smith v East Elloe, and Ostler: see para 81 above), leading to the anomaly that such a “nullity” is mysteriously validated at the end of that period. Nor has the expanded understanding of the Anisminic principle been consistently applied in other statutory contexts, as In re McC [1985] AC 528 illustrates (para 58 above).

These examples show that the courts have not adopted a uniform approach, but have felt free to adapt or limit the scope and form of judicial review, so as to ensure respect on the one hand for the particular statutory context and the inferred intention of the legislature, and on the other for the fundamental principles of the rule of law, and to find an appropriate balance between the two. Even if this was not always the way in which the decisions were justified at the time, it may be seen as providing a sounder conceptual basis. Thus in the planning cases, it having been accepted that the statutory grounds cover all the traditional ground of judicial review, there is no difficulty in holding that the six-week time-limit provides a proportionate balance between effective judicial review, and the need for certainty to enable such decisions to be acted on with confidence.

That more flexible approach to the relationship between the legislature and the courts is in my view wholly consistent with the modern constitutional settlement, as confirmed by the 2005 Act, and recognised by this court in Miller. Against that background, the judgments of this court in Cart point the way to an approach which (pace Professor Forsyth - see para 98 above) is both pragmatic and principled. The critical step taken by this court in Cart was to confirm, what was perhaps implicit in some of the earlier cases, that it is ultimately for the courts, not the legislature, to determine the limits set by the rule of law to the power to exclude review.

This proposition should be seen as based, not on such elusive concepts as jurisdiction (wide or narrow), ultra vires, or nullity, but rather as a natural application of the constitutional principle of the rule of law (as affirmed by section 1 of the 2005 Act), and as an essential counterpart to the power of Parliament to make law. The constitutional roles both of Parliament, as the maker of the law, and of the High Court, and ultimately of the appellate courts, as the guardians and interpreters of that law, are thus respected. The question in any case is “the level of scrutiny required by the rule of law”, set on a basis which as stated in Cart is both “principled and proportionate” (para 51 per Lady Hale), or in Lord Dyson’s words (para 133): “what scope of judicial review … is required to maintain the rule of law”; it being “a matter for the courts to determine what that scrutiny should be” (para 102 per Lord Clarke).

Some forms of ouster clause may readily satisfy such a test, as in the planning cases mentioned above. Similarly, in Racal, review limited to a High Court judge could reasonably have been justified as providing a sufficient and proportionate level of protection in the narrow statutory context of the grant of authority to inspect company books.

In a different context a similarly balanced assessment could be used to support the outcome of the dissenting judgment in Pearlman. Here again judicial thinking has moved on, recognising that the division between fact and law is not always clear-cut, and that a more “pragmatic” approach may sometimes be required (per Lord Hope, Jones v First-tier Tribunal [2013] UKSC 19, para 16). The assessment of whether particular works involve a “structural alteration … or addition” is substantially a factual issue, which can properly and economically be left to the trial judge at County Court level. An ouster provision designed to achieve that effect should be respected. On the other hand such a clause should not be treated as excluding the possibility of review in an exceptional case where the judge can be shown (in Geoffrey Lane LJ’s words) to have been “embarking on some unauthorised or extraneous or irrelevant exercise”. It is significant that this judgment (later approved by the Privy Council), like that of the Court of Appeal in Sivasubramaniam (para 125 above), implicitly recognised that even in a very restrictive statutory context the possibility of judicial review could not be excluded altogether in an exceptional case.

Lord Sumption finds support for his contrary view in a part of Lady Hale’s judgment in Cart, where she said:

I do not believe with respect that this passage bears the weight which Lord Sumption places on it. It comes as part of a section of the judgment (para 39) in which Lady Hale was explaining her reasons for not following the courts below by reintroducing the distinction between jurisdictional and other errors, which had been effectively abandoned in Anisminic. She was not addressing the present issue as to the circumstances in which review could be excluded altogether.

Similarly, the relevant passage of Lord Wilberforce’s speech was not a considered treatment of the subject. It was no more than part of his introduction to the more detailed discussion, in which he was explaining the high legal standing of the Commission, and its consequences for his approach to interpretation. The full passage reads as follows:

Returning to the present case, Sir James Eadie accepts the need for judicial review by a court or tribunal which is both independent and authoritative, but submits that the IPT is well-suited to perform that role. The test of such independence and competence, he submits, is “not the source of the powers of review, but rather the institutional features of the body created to exercise the power of review”.

In my view that is too narrow a focus. It pays no regard to the need to ensure that the law applied by the specialist tribunal is not developed in isolation (“a local law”), but conforms to the general law of the land. At least since the time of Blackstone (para 32 above), this has been a central part of the function of the High Court as constitutional guardian of the rule of law. It formed an important part of the reasoning of the Supreme Court in Cart. It applies with particular force in the present context where there is a significant overlap between jurisdictions of the IPT and of the ordinary courts. The present case is a good example. The legal issue decided by the IPT is not only one of general public importance, but also has possible implications for legal rights and remedies going beyond the scope of the IPT’s remit. Consistent application of the rule of law requires such an issue to be susceptible in appropriate cases to review by ordinary courts.

It may seem anomalous that the route to review by the ordinary courts is the grant of permission by the High Court, whose judges may arguably be less well-equipped for this purpose than the judges of the IPT. But the same could have been said of the relationship between the Upper Tribunal and the High Court in Cart itself. Although Lady Hale acknowledged (para 56) that the review should in principle be by a judge with “more experience or expertise”, this would be met by the possibility that, if the case were “channel(ed) into the legal system”, it would enable where appropriate onward transmission to the appellate courts. The high status of the Upper Tribunal was a reason for a restricted approach to the grant of permission, but not for excluding it altogether. It also has to be remembered that until the 1979 reforms the review jurisdiction would have been exercised by the full Divisional Court of the Queen’s Bench, generally presided over by the Lord Chief Justice. In the modern system, the court’s powers of case management can ensure that the matter comes before a court of suitable composition (as happened in this case).

There is a distinct issue whether the rule of law requires such decisions to be susceptible to review also by the appellate courts. Unlike the original common law jurisdiction of the High Court, the jurisdictions of the Court of Appeal, and now of the Supreme Court, are the creation of statute. In Racal it was assumed that if the relevant statutes provided that the decision of the High Court was unappealable, it could not be corrected at all (see para 64 above). In R (A) Lord Brown accepted as correct the concession that “there is no constitutional (or article 6) requirement for any right of appeal from an appropriate tribunal” (para 23).

As applied to article 6 of the Convention, which was in issue in R (A), that proposition may be uncontroversial, given that the ultimate arbiter of Convention law is in Strasbourg rather than the courts of this country. In the context of a domestic law challenge, it is more debatable. Arguably, following the logic of the reasoning in Cart, it may be thought implicit in the constitutional framework for the rule of law, as established by the Senior Courts Act 1981 and the Constitutional Reform Act 2005, that legal issues of general importance should be reviewable by the appellate courts; and that an ouster clause which purports to exclude that possibility cannot, consistently with the rule of law, be upheld. The only authority referred to by Lord Brown was Farley v Secretary of State for Work and Pensions (No 2) [2006] 1 WLR 1817. That case was concerned with a provision that, on an application by the Secretary of State to the magistrates’ court to enforce a “maintenance assessment”, the assessment itself was immune from challenge. The effectiveness of that ouster was upheld, but that depended on it being shown that there was another suitable means of challenging the assessment.

The issue does not arise directly in the present context. If the decisions of the IPT are in principle susceptible to judicial review by the High Court, there is nothing in RIPA or any other statute to exclude onward appeal from the decisions of the High Court itself in the ordinary way. We have not heard detailed argument on this aspect, and I decline therefore to express a concluded view.

In conclusion on the second issue, although it is not necessary to decide the point, I see a strong case for holding that, consistently with the rule of law, binding effect cannot be given to a clause which purports wholly to exclude the supervisory jurisdiction of the High Court to review a decision of an inferior court or tribunal, whether for excess or abuse of jurisdiction, or error of law. In all cases, regardless of the words used, it should remain ultimately a matter for the court to determine the extent to which such a clause should be upheld, having regard to its purpose and statutory context, and the nature and importance of the legal issue in question; and to determine the level of scrutiny required by the rule of law.

Accordingly, for the reasons given under the first issue, I would allow the appeal and hold, in answer to the preliminary issue, that the judicial review jurisdiction of the High Court is not excluded by section 67(8). Although that is the limit of the issue before the court, it will be clear from what I have said about the significance of the substantive legal issue, that this is a case where, if judicial review is available, permission should be granted.

Two issues arise on this appeal. The first is the specific issue whether section 67(8) of the Regulation of Investigatory Powers Act 2000 (“RIPA 2000”) must be taken as purporting to oust the supervisory jurisdiction of the High Court to quash a judgment of the Investigatory Powers Tribunal (“the IPT”) for error of law. The second is the more general issue of whether, and, if so, in accordance with what principles, Parliament has the power by statute to oust the supervisory jurisdiction of the High Court to quash the decision of an inferior court or tribunal of limited statutory jurisdiction.

On the first issue, I agree with the judgment of Lord Carnwath. In view of the importance of the issue, I add some brief comments of my own.

The IPT was created by section 65(1) of RIPA 2000. Its jurisdiction and procedures are described in the judgment of Lord Carnwath and I simply draw attention to the following matters. Section 65(2) includes provision that it is the only appropriate tribunal for hearing proceedings falling within section 65(3) (which includes proceedings against any of the intelligence services) for actions incompatible with Convention rights under section 7 of the Human Rights Act 1998 (section 65(2)(a)). Section 67 provides that it shall be the duty of the Tribunal to hear and determine proceedings or to consider and determine complaints or references, brought before it under section 65(2). Section 67(2) provides that where the IPT hears any proceedings by virtue of section 65(2)(a), “they shall apply the same principles for making their determination in those proceedings as would be applied by a court on an application for judicial review”. At all material times the Investigatory Powers Tribunal Rules 2000 govern procedure before the IPT. It has the power to conduct proceedings in private and, in certain circumstances, in the absence of the complaining party. Rule 6(1) provides:

Section 67(8) of RIPA 2000 provided at the relevant time:

Considered with the benefit of hindsight, it can be seen that Anisminic initiated a process of fundamental change in the approach of the courts to judicial review which was to lead to their abandoning the distinction between errors of law going to jurisdiction and those that did not. Whereas previously an error of law was reviewable only if it was a jurisdictional error or if it was an error on the face of the record, all errors of law were to become reviewable. However, as Professor Feldman has observed (“Anisminic Ltd v Foreign Compensation Commission [1968]: In Perspective”, in Juss and Sunkin (eds) Landmark Cases in Public Law (Oxford, 2017) pp 92-93), this was not immediately apparent from the speeches in Anisminic [1969] 2 AC 147 itself. On the contrary, they maintained the distinction between jurisdictional and non-jurisdictional errors of law and the decision turns on a particularly broad notion that the tribunal did not have the power to take certain decisions. Thus, for example, Lord Reid (at p 171B-F) distinguished between those errors of law or procedure by a tribunal which render a decision a nullity and other cases where “its decision is equally valid whether it is right or wrong subject only to the power of the court in certain circumstances to correct an error of law”. Similarly, Lord Wilberforce (at p 210D-E) considered that a “tribunal may quite properly validly enter upon its task and in the course of carrying it out may make a decision which is invalid - not merely erroneous” and referred to “a crucial distinction which the court has to make” between doing something which is not in the tribunal’s area and doing something wrong within that area. By addressing whether the appellants had a successor in title and its nationality, the Foreign Compensation Commission had asked the wrong question and had taken account of irrelevant considerations with the result that its decision was a nullity.

In the cases which followed Anisminic, however, the implications of the extremely broad approach to jurisdictional error of law taken in that case soon became apparent. If, as Anisminic suggests, addressing the wrong question renders the decision a nullity, it is possible to present almost any error of law as the result of such an error of approach. Different views on this subject were aired in the judgments in the Court of Appeal in Pearlman v Keepers and Governors of Harrow School [1979] QB 56. Lord Denning MR (at pp 69G-70E) considered that the resulting distinction between jurisdictional and non-jurisdictional error was very fine, was being eroded and should be abandoned. In his view the correct approach was to hold that no court or tribunal has any jurisdiction to make an error of law on which the decision of the case depends. Geoffrey Lane LJ, however, (at p 76C), assuming for this purpose that the judge had made an error of law in concluding that the works did not constitute structural alterations, considered this an error within jurisdiction. It could not be said to be a determination the judge was not entitled to make. Although the approach of Geoffrey Lane LJ was approved by the Judicial Committee of the Privy Council in South East Asia Fire Bricks Sdn Bhd v Non-Metallic Mineral Products Manufacturing Employees Union [1981] AC 363 and by the House of Lords in In re Racal Communications Ltd [1981] AC 374, Lord Denning’s approach was to prevail.

In Racal, Lord Diplock acknowledged the true significance of Anisminic, observing that the break-through made by Anisminic had been that, as respects administrative tribunals and authorities, the old distinction between errors of law that went to jurisdiction and errors of law that did not had for practical purposes been abolished. Accordingly, any error of law that could be shown to have been made by administrative tribunals or authorities in the course of reaching a decision on matters of fact or of administrative policy would result in their having asked themselves the wrong question with the result that the decision they reached would be a nullity (at p 383C-D). Two years later, in O’Reilly v Mackman [1983] 2 AC 237 Lord Diplock referred in similar vein to:

Thereafter, a series of decisions in the House of Lords established that there is a single category of errors of law, all of which render a decision ultra vires (R v Hull University Visitor, Ex p Page [1993] AC 682 per Lord Browne-Wilkinson at p 701; Boddington v British Transport Police [1999] 2 AC 143, at p 158D-E per Lord Irvine of Lairg LC; R (Lumba) v Secretary of State for the Home Department [2012] 1 AC 245 per Lord Dyson JSC at para 66). In R (Cart) v Upper Tribunal (Public Law Project intervening) [2012] 1 AC 663 Baroness Hale considered (at para 18) that in Anisminic “the House of Lords effectively removed the distinction between error of law and excess of jurisdiction”.

It is, however, necessary to consider whether the Anisminic principle applies equally to decision-making by both administrative and judicial bodies. Anisminic itself had been concerned with a decision of the Foreign Compensation Commission (“FCC”). It is significant that in that case Lord Wilberforce considered that the functions of the FCC were predominantly judicial, with the power to decide questions of law and he observed that there was every ground, having regard to the number and the complexity of the cases with which it must deal, for giving a wide measure of finality to its decisions. Accordingly, there was no reason for giving a restrictive interpretation to section 4(4) which provided that its determinations were not to be called into question in any court of law (at p 207C-D). Nevertheless, he came to his conclusion on the basis that, as he put it, the decision was made outside the permitted field.

By contrast, in Racal Lord Diplock observed (at p 382G) that in Anisminic the House of Lords had been concerned with decisions of administrative tribunals. He explained that Anisminic proceeds on the presumption that “where Parliament confers on an administrative tribunal or authority, as distinct from a court of law, power to decide particular questions defined by the Act conferring the power, Parliament intends to confine that power to answering the question as it has been so defined” (at pp 382H-383A). Furthermore, while Parliament could confer upon administrative tribunals or authorities power to decide questions of law as well as questions of fact and administrative policy, this requires clear words because there is a presumption that, where a decision-making power is conferred on a tribunal or authority that is not a court of law, Parliament did not intend to do so (at p 383B-C). He then proceeded to contrast the position of a court of law.

The decision of the House of Lords in R v Hull University Visitor, Ex p Page lends support to the approach followed by Lord Diplock in Racal. On the other hand, however, it should be noted that Lord Diplock’s formulation of the Anisminic principle in O’Reilly v Mackman, two years after the decision in Racal, cited above, appears to be applicable without distinction to “inferior courts and statutory tribunals”. Furthermore, in R v Greater Manchester Coroner, Ex p Tal [1985] QB 67 Robert Goff LJ, delivering the judgment of the Divisional Court concluded (at p 81G-83B) that Lord Diplock in Racal had not intended to say that the Anisminic principle did not extend to inferior courts as well as tribunals. Goff LJ considered that, historically, inferior courts had always been subject to what was now called judicial review, although originally only in cases of error going to the jurisdiction and error of law within the jurisdiction which appeared on the face of the record:

There is, moreover, no trace of such a distinction in the Supreme Court’s consideration of the Upper Tribunal in Cart where there is no suggestion that courts of limited jurisdiction might have power to err as to law within their jurisdiction. This leads Professor Forsyth to observe:

The distinction between administrative tribunals and courts of law suggested by Lord Diplock in Racal is likely to be an arid one in the present context. Quite apart from the difficulties which are likely to be encountered in drawing such a distinction in individual cases, what matters here is whether a body is charged with performing a judicial function. If it is, then, as Laws LJ observed in the Divisional Court in Cart (at para 68), the true contrast is between the High Court on the one hand and courts of limited jurisdiction on the other.

In the present case the IPT is undoubtedly charged with performing a judicial function. The issue for decision in this case must therefore be approached on the basis that the statute makes provision as to the status of decisions of a judicial body.

I wholeheartedly endorse the exposition by Laws LJ in the Divisional Court in Cart (at paras 36-40) of the principle that it is a necessary corollary of the sovereignty of Parliament that there should exist an authoritative and independent body which can interpret and mediate legislation made by Parliament:

I accept that in the case of a judicial body, by contrast with a purely administrative body, there is no presumption that Parliament did not intend to confer a power to decide questions of law as well as questions of fact. (See Racal per Lord Diplock at p 383E.) It is, rather, a matter of the interpretation of the legislation concerned in each case, unencumbered by such a presumption. Nevertheless, if the jurisdiction of the High Court is to be displaced or varied in some way, this is a matter of great importance and clear words will be required to achieve that result. Notwithstanding the disapproval by the House of Lords in Racal of the decision of the majority in the Court of Appeal in Pearlman, the following observation of Lord Denning MR (at p 70D) remains valid as a general proposition:

It has been suggested, on the basis of Racal, that while section 67(8) does not exclude judicial review on other grounds such as a lack of subject-matter jurisdiction or want of natural justice, that section excludes the jurisdiction of the High Court to entertain a challenge to the Tribunal’s decisions on the merits ie it excludes judicial review on grounds which would be tantamount to an appeal on the merits. It seems to me, however, that this places more weight on Racal than that authority can bear. It provides an insecure foundation because, as is demonstrated by the later decisions referred to above, at the date of Racal the legal principles in play were still evolving. As a result, it is not appropriate to allow the reasoning of Lord Diplock in Racal to influence the issue of interpretation in the present case.

Turning to the issue of interpretation of section 67(8), I accept that the role of the IPT is judicial. As a result, there is no presumption in favour of restricting its field or of restricting its power to decide issues of law. However, if the jurisdiction of the High Court can be excluded at all, it requires the most clear and explicit words. As Lord Reid observed in Anisminic (at p 170C-D):

Subject to one point, the wording of section 67(8) closely resembles that of section 4(4) of the Foreign Compensation Act 1950 which was the subject of Anisminic:

The one point of distinction between section 4(4) of the Foreign Compensation Act 1950, on the one hand, and section 67(8) on the other, is the inclusion in the latter of the words in parenthesis “(including decisions as to whether they have jurisdiction)”. To my mind, however, these words are not apt to extend the exclusion of the jurisdiction of the High Court to what purport to be decisions but in law are not to be so regarded. While it is now established that a decision based on an error of law is not to be regarded as a decision for this purpose, this notion does not easily fit within the description of a decision as to whether it has jurisdiction. If the IPT takes a decision which is founded on an error of law, it is not in any real sense taking a decision as to whether it has jurisdiction. If the intention was to exclude the jurisdiction of the High Court from purported decisions founded on an error of law, it was necessary to say so in clear terms. Clause 11 of the Asylum and Immigration (Treatment of Claimants etc) Bill 2003, to which Lord Carnwath refers at para 101 of his judgment, is a more recent example of an attempt to achieve the required degree of clarity if such a provision is to be effective. That provision, which was not enacted, can at least be said to have squarely confronted what it sought to achieve as required by the principle of legality. To my mind, section 67(8) does not satisfy this requirement.

It may be that the explanation of the words in parenthesis is, as submitted by Ms Dinah Rose QC on behalf of the appellant, that they were intended to refer to determinations of precedent fact, a matter which was highly topical in 1985 following the decision of the House of Lords in R v Secretary of State for the Home Department, Ex p Khawaja [1984] AC 74. On this basis the words in parenthesis in section 67(8) could be considered to have the effect that decisions of the IPT on issues of precedent fact going to its jurisdiction, but not issues of law, would be beyond the scope of review. However, it is not necessary to come to a concluded view on this point. For present purposes it is sufficient that the words employed in section 67(8) do not make provision with sufficient clarity for the exclusion of the review jurisdiction of the High Court in respect of errors of law.

In coming to this conclusion, I have taken full account of the various features of the statutory scheme to which Sir James Eadie QC has drawn attention in support of the respondents’ case. He is correct in his submission that there is here a special allocation of judicial responsibility to the IPT in the national security context (section 67(3)(a)). Similarly, the IPT’s rules and procedures create a “bespoke” system particularly well suited to the adjudication of controversial issues in the context of national security and directed to protecting the public interest. Furthermore, there can be no doubt as to the outstanding judicial quality of the members of the IPT. However, the exclusion of the review jurisdiction of the High Court in cases of error of law, if achievable at all, would require a provision of much greater clarity making abundantly clear that that was what it sought to achieve.

For these reasons, I would allow the appeal against the decision of the Court of Appeal on the first issue. It is, accordingly, unnecessary to express any view on the second issue.

The Investigatory Powers Tribunal is a specialist tribunal established in 2000 under the Regulation of Investigatory Powers Act 2000. Its principal functions are to determine proceedings against the intelligence services in respect of breaches of human rights and complaints about the interception of communications, in a way which enables these claims to be examined judicially without the risk of disclosure of secret matters. The Tribunal effectively replaced the Interception of Communications Act Tribunal, the Security Services Act Tribunal and the Intelligence Services Act Tribunal, which had been established under earlier enactments, as well as taking over the operation of the complaints provisions of Part III of the Police Act 1997.

The appellant, Privacy International, complained that Government Communications Headquarters (“GCHQ”), one of the intelligence services, had carried out unlawful computer hacking. Computer hacking by the intelligence services requires the authority of a warrant of the Secretary of State under section 5 of the Intelligence Services Act 1994. The relevant activities of GCHQ were said to be unlawful on the ground that the warrants authorising them included what has been called (not entirely accurately) “thematic warrants”. A thematic warrant means a warrant authorising a class of activity in respect of a class of property. The appellant’s case before the Tribunal was that section 5(2) of the Intelligence Services Act 1994 empowered the Secretary of State to issue a warrant authorising “specified acts” in respect of “specified property”, and did not extend to thematic warrants. Alternatively, they submitted that if the Act did authorise such warrants, it was in that respect incompatible with articles 8 and 10 of the Human Rights Convention. The Tribunal held an open hearing to determine a number of preliminary issues of law. In a judgment issued on 12 February 2016, it held that thematic warrants were lawful. The appellant began proceedings for judicial review, seeking an order quashing that decision on the ground that the Tribunal’s construction of section 5(2) of the Act of 1994 was wrong in law.

Section 67(8) and (9) of the Regulation of Investigatory Powers Act 2000 provide:

The question at issue on this appeal is whether an application for judicial review on the ground that the Tribunal has decided an issue on a wrong view of the law, is available having regard to section 67(8) of the Act. The Divisional Court and the Court of Appeal have both held that it is not. I agree with them. I shall need to examine the law in some detail, but my reason can be shortly summarised. The effect of section 67(8) is simply to exclude the jurisdiction of the High Court to entertain a challenge to the Tribunal’s decisions on the merits. In other words, it excludes judicial review on grounds which would be tantamount to an appeal. The Investigatory Powers Tribunal acts as a court. Its function is to exercise powers of judicial review over (among others) the intelligence services, which would otherwise have been exercisable by the High Court, and to do so on the same basis as the High Court. The purpose of judicial review is to maintain the rule of law. But the rule of law is sufficiently vindicated by the judicial character of the Tribunal. It does not require a right of appeal from the decisions of a judicial body of this kind. For this reason section 67(8) is not an ouster of any jurisdiction which constitutional principle requires the High Court to have.

Historically, the legal basis of judicial review was the concept of excess of jurisdiction. Bodies deriving their powers from statute or grant under the royal prerogative were amenable to certiorari in the King’s Bench if they exceeded the formal or implicit limits of the grant. Strictly speaking, excess of jurisdiction was confined to want of legal competence. But the limitations of this approach led the courts in some cases artificially to expand the concept of jurisdiction to cover varieties of public law wrong that did not readily fall within established categories. In particular, it was extended to broad categories of unreasonable conduct which the grant of the relevant power was assumed not to have authorised without specific words: for example, bad faith or disregard of the rules of natural justice. The artifice became unnecessary after the decision of the House of Lords in Council of Civil Service Unions v Minister for the Civil Service [1985] AC 374, in which the majority of the Appellate Committee put the grounds of judicial review on a broader basis and held that it extended generally to the exercise of justiciable public powers, including those which were not the subject of any statutory or other grant. But the historical roots of English public law have continued to influence its development, notably in the area of ouster clauses.

It has been recognised since the 17th century that a statute can remove the supervisory jurisdiction of the courts over inferior tribunals and administrative bodies only by clear words. In Smith, Lluellyn v Comrs of Sewers (1669) 1 Mod 44, the Commissioners purported to exercise a penal power to impose fines which they did not have. Sir John Kelynge, Chief Justice of King’s Bench, declined to treat a statutory provision that orders of the Commissioners should be valid unless revoked by the Commissioners themselves as excluding the jurisdiction of the courts to issue certiorari. “You cannot oust the jurisdiction of this court without particular words in Acts of Parliament”, he said; “there is no jurisdiction that is uncontrollable by this court.” Lord Mansfield made the same point, nearly a century later, in R v Moreley (1760) 2 Bur 1041, when he said that “the jurisdiction of this court is not to be taken away unless there be express words to take it away.” In the course of the 19th century, this proposition was applied so as to treat a statutory exclusion of the High Court’s power to issue certiorari to inferior tribunals as inapplicable to cases in which the tribunal had purported to exercise a power which it did not have, or a condition precedent to the existence of a power was absent, or the court was not properly constituted. In keeping with the jurisdictional approach to judicial review, these were all cases in which the court lacked legal competence, with the result that its acts were nullities. Cases of this kind give rise to a conceptual problem that goes beyond mere construction of the statute. Where a statute confers a power on an administrative or judicial body to do some class of acts, and ousts the jurisdiction of the High Court to review its acts, the threshold question is always whether it is a body or an act to which the statute applies. If not, then the ouster clause can have no application to it. As Cockburn CJ observed in Ex p Bradlaugh (1878) 3 QBD 509, 513, “the section does not apply where the application for the certiorari is on the ground that the inferior tribunal has exceeded the limits of its jurisdiction.” Otherwise, a body exercising limited statutory powers would be at liberty to determine what its limits were. As Mellor J pointed out in the same case, “a metropolitan magistrate could make any order he pleased without question.” If a superior court is precluded from deciding whether the statute applies to the relevant body or act, it would follow, as Lord Justice-Clerk Boyle pointed out in the early Scottish case of Campbell v Young (1835) 13 S 535, that “because a party says that he acts under the statute, he is to do as he pleases.” His description of that suggestion as “monstrous” would be adopted by any modern public lawyer.

Implicit in this approach was a distinction between excesses of jurisdiction ascertained at the point where a public body embarks on the relevant function, and errors of fact or law committed in the course exercising it; and a related distinction between errors of law or fact going to the decision-maker’s legal competence, and errors within competence. Lord Coleridge CJ expressed the orthodoxy of his time when he observed in R v Justices of the Central Criminal Court (1886) 17 QBD 598, 602, that “where a Court has jurisdiction to entertain an application, it does not lose its jurisdiction by coming to a wrong conclusion, whether it is wrong in point of law or of fact.” The only power to quash a decision within jurisdiction was the ancient and sui generis power of the King’s Bench to quash for error of law on the face of the record in a case where the error was disclosed in decision-maker’s reasons, if he gave any.

The Foreign Compensation Commission was a statutory body created by the Foreign Compensation Act 1950. The Act empowered Her Majesty in Council to make provision for the Commission to distribute money received by the Crown under the royal prerogative from foreign governments under international law, by way of compensation for losses suffered by British subjects in the territory of those governments. Section 4(4) of the Act provided:

Lord Reid, at p 171, gave some illustrations of errors on the part of the tribunal which, without going to legal competence in its strict sense, would nevertheless invalidate the decision:

Lord Morris of Borth-y-Gest was at one with his colleagues on the test to be applied, although he dissented on its application to the facts. At p 182, he expressed it in this way:

Lord Pearce made the same distinction between errors of law which led the tribunal to address questions which it was not within their powers to determine, and other errors. At p 195, he observed:

Common to all the speeches in Anisminic was the view that the extent to which the decision-maker’s errors amount to an excess of his jurisdiction depended on the breadth of the power committed to it by the statute as a matter of construction. Lord Wilberforce agreed but proposed, at p 207, a more nuanced analysis of the effect of ouster clauses, which eschewed the language of jurisdiction and nullity with its binary test, and is perhaps more in keeping with the modern law’s aversion to rigid categorisation:

Implicit in the decision of the House in Anisminic was that invalidity for error of law no longer depended on the error being patent on the face of the record. But it will be apparent from all of the speeches that the Appellate Committee did not reject in principle the distinction between errors of law going to jurisdiction and errors of law within jurisdiction. Instead, they broadened the concept of jurisdiction to embrace not just legal competence but also legal relevance, ie addressing the right questions. All of them recognised that there would be some errors of law which did not go to jurisdiction even in this enlarged sense, and which would not therefore invalidate the decision. However, dicta of high authority in subsequent cases have consistently interpreted the decision as authority for an approach to errors of law which is both broader and more absolute than the speeches themselves warrant. In particular, they suggested that all errors of law were to be regarded as beyond jurisdiction. Thus in O’Reilly v Mackman [1983] 2 AC 237, 278, Lord Diplock regarded Anisminic as authority for the proposition that “if a tribunal whose jurisdiction was limited by statute or subordinate legislation mistook the law applicable to the facts as it had found them, it must have asked itself the wrong question, ie one into which it was not empowered to inquire and so had no jurisdiction to determine.” In R v Hull University Visitor, Ex p Page [1993] AC 682, 701-702, Lord Browne-Wilkinson, endorsing this summary, took it that thenceforth, “it was to be taken that Parliament had only conferred the decision-making power on the basis that it was to be exercised on the correct legal basis: a misdirection in law in making the decision therefore rendered the decision ultra vires.” In R (Lumba) v Secretary of State for the Home Department (JUSTICE intervening) [2012] 1 AC 245, para 66, Lord Dyson considered that it “established that there was a single category of errors of law, all of which rendered a decision ultra vires.” Finally, in R (Cart) v Upper Tribunal [2012] 1 AC 663, para 18, Baroness Hale observed that in Anisminic “the House of Lords effectively removed the distinction between error of law and excess of jurisdiction.” Although none of these statements were part of the ratio of judgments in which they appeared, they have been followed by the courts for many years and have been assumed to state the modern law by Parliamentary draftsmen including (as I shall show) the draftsman of the Regulation of Investigatory Powers Act 2000. In my opinion, it is now too late to revert to the subtler distinctions in the speeches in Anisminic , even if it were thought desirable to do so. However, the speeches in Anisminic remain authority for the proposition, which may be thought self-evident, that the extent of a tribunal’s jurisdiction depends on the construction of its enabling Act or, in Lord Wilberforce’s words, on the breadth of its “permitted field”.

The categorisation of errors of law as excesses of jurisdiction is the result of the unsystematic way in which English public law has developed over the past three centuries. Its effect is to create what is nominally a power of review, but is in substance a right of appeal on points of law going to the merits. The implications of this are very different, according as the decision under review was made by an administrative or executive body on the one hand or a judicial one on the other. A right of access to a court or similar judicial body to review the lawfulness of administrative or executive acts is an essential part of the rule of law. But the rule of law does not require a right of appeal from such a body or a right to call for a review of its decisions. In England, appellate jurisdiction is wholly statutory, and may be absent, restricted or wholly excluded. This is well established as a matter of both law and legislative practice: see R (A) v Director of Establishments of the Security Service [2010] 2 AC 1, para 24 (Lord Brown of Eaton-under-Heywood). It has also been persistently recognised in the case-law of the European Court of Human Rights: see Bochan v Ukraine (No 2) (2015) 61 EHRR 14, paras 44-45. For this reason the development since Anisminic of a legal principle which made excesses of jurisdiction of all errors of law has been accompanied by a recognition that the reasons for strictly construing ouster clauses may be more or less powerful, depending on the nature of the decision and the decision-maker. In particular, the principle may have to be adapted to the decisions of judicial bodies. This is because a judicial body, depending on its status and functions, is more likely to have a wider “permitted field”, extending to the conclusive resolution of issues of law (or indeed fact) and including an unrestricted interpretative power.

The analysis starts with the speech of Lord Wilberforce in Anisminic itself, from which I have already quoted. Lord Wilberforce regarded the Foreign Compensation Commission as a body whose functions were “predominantly judicial” (p 207C), from which it followed that there was no reason to give a restricted interpretation to section 4(4) of the Foreign Compensation Act 1950. That, however, was not the end of the matter. In his view, the effectiveness of the ouster clause depended on the extent of the interpretatory power which, as a matter of construction, Parliament must be taken to have conferred on the decision-maker. Only errors of law lying outside what he called “the permitted field” were reviewable in the face of an ouster clause such as section 4(4): see pp 208A-B, 209F-G, 210C-E. As I have pointed out, Lord Wilberforce’s view that the Foreign Compensation Commission’s error of law lay outside the “permitted field” was based on a careful analysis of the interpretative power conferred on it by its enabling Act, which was limited to applying the Order in Council according to its true legal meaning.

In South East Asia Fire Bricks Sdn Bhd v Non-Metallic Mineral Products Manufacturing Employees Union [1981] AC 363, the Privy Council had to consider whether the decision of the Industrial Court of Malaysia, a judicial body, could be reviewed in the High Court on the ground of error of law on the face of the record. The statute from which the Industrial Court derived its jurisdiction provided that its awards “shall be final and conclusive, and no award shall be challenged, appealed against, reviewed, quashed or called in question in any court of law.” The Board held the judicial review did not lie. Lord Fraser of Tullybelton, delivering the advice of the Privy Council, distinguished Anisminic , at p 370, on the ground that “if the inferior tribunal has merely made an error of law which does not affect its jurisdiction, and if its decision is not a nullity for some reason such as breach of the rules of natural justice, then the ouster will be effective.” The Board did not expressly address the distinction between judicial and other decision-makers, but since the judicial character of the Industrial Court was the only distinction between the case before them and Anisminic , that must have been the basis of the decision. Some light is thrown on this question by the speeches in In re Racal Communications [1981] AC 374, which addressed this very question, and were delivered a week later by an Appellate Committee comprising two of the same law lords, Lord Edmund-Davies and Lord Keith.

In re Racal Communications arose out of an application to the High Court under section 441 of the Companies Act 1948 to inspect the books of a company. Section 441(3) provided that the decision of the High Court on such an application “shall not be appealable”. The Court of Appeal had received the appeal on the ground that the ouster of its appellate jurisdiction was ineffective, relying on the decision in Anisminic . The House of Lords held that the Court of Appeal had been wrong to receive it in the face of the statutory exclusion of appeals. The leading speech was delivered by Lord Diplock. Having concluded that the House was bound by the statutory exclusion of appeals, he then turned to:

Of particular interest in this context are the grounds on which the Appellate Committee overruled the decision of the Court of Appeal in Pearlman v Keepers and Governors of Harrow School [1979] 1 QB 56. In that case, the Court of Appeal, by a majority, had quashed a decision of an inferior court, namely the county court, on an application under the Housing Act 1974 to adjust the rateable value of tenanted premises, holding that the judge had erroneously construed the adjustment provisions of the Act. It had held, citing Anisminic in support, that a provision that the county court’s decision should be “final and conclusive” was ineffective to oust judicial review. Lord Diplock (p 384) approved the dissenting judgment of Geoffrey Lane LJ, in which he had said (p 76):

I decline to accept that these judgments can be explained away on the ground that a tribunal is to be distinguished from a court. The Appellate Committee was concerned with a court, but the distinctive feature of a court which made its observations pertinent was that it was a judicial body. Almost all tribunals are obliged in some respects to act judicially, for example in acting fairly and without bias. But not all tribunals are judicial bodies. What matters is not the nomenclature of the decision-maker but its statutory functions. On an issue which is agreed on all sides to turn on the requirements of the rule of law, it would in my view be absurd to suggest that there is no distinction to be made between a statute providing for an administrative authority’s decisions to be conclusive and a statute making corresponding provision for the decisions of a judicial body. As I shall explain, the Investigatory Powers Tribunal is indistinguishable from a court in every respect that matters to the present issue.

More recently, in Lee v Ashers Baking Co Ltd [2018] 3 WLR 94, Racal was considered and applied by this court. The issue was whether a statutory provision making the decision of the Court of Appeal of Northern Ireland “final”, precluded a further appeal when the Court of Appeal had failed to refer a devolution issue to the Supreme Court as it had been bound to do. That question was answered by Lord Mance (with whom the rest of this court agreed). His judgment is relevant for two reasons. First, at para 86, he adopted Lord Diplock’s analysis, in particular his distinction between the presumptions to be applied to an ouster clause where the decision is that of a judicial body and those which apply where the decision is that of administrative tribunal. Secondly, he held that as a matter of construction the same language could in principle be sufficient to exclude an appeal on the merits but not an appeal on the ground that the court below had committed a procedural error or failed to perform the function with which Parliament had charged it. At para 88, he said:

In view of the weight placed on this decision by the appellants, it is necessary to analyse the judgments with some care, although it must be borne in mind throughout that it is not direct authority on the question before us because it was not a case about ouster clauses. There was no ouster clause in the relevant statutes.

In the Divisional Court the issues were (i) whether the mere designation of a judicial body (in that case the Special Immigration Appeals Tribunal and the Upper Tribunal) as a superior court of record took it outside the scope of the High Court’s review jurisdiction even in the absence of an ouster; and (ii) whether the scheme of the statutes from which these bodies derived their powers was inconsistent with its decisions being reviewable in the High Court even in the absence of an ouster. In an impressive judgment, Laws LJ, delivering the judgment of the Divisional Court held that the answer to (i) was No and the answer to (ii) was Yes. On issue (i), he held that the special status of the High Court as exercising a jurisdiction to keep other bodies within their powers meant that a superior court of record other than the High Court was not, simply by virtue of that status, immune from the review jurisdiction of the High Court. For present purposes, however, what matters is Laws LJ’s treatment of issue (ii). He accepted that some courts and tribunals might be immune from the High Court’s review jurisdiction. He expressed the basic principle as follows:

In the Court of Appeal the position of the SIAC was no longer in issue. Sedley LJ, delivering the judgment of the court, rejected the suggestion that the Upper Tribunal was the alter ego of the High Court and denied that that was the test. In his view (para 20) all courts other than the High Court itself were in principle amenable to judicial review in the High Court in the absence of a sufficiently clear ouster clause. But he thought that while the Upper Tribunal was amenable to judicial review, the scope of review of a body such as the Upper Tribunal was limited, because the scheme of the Tribunals, Courts and Enforcement Act 2007 required the tribunal system to be treated as autonomous. It therefore implicitly provided (para 42) for “the correction of legal error within rather than outside the system.” It followed that judicial review extended only to what Sedley LJ called “outright” excess of jurisdiction, ie the exercise of powers that the tribunal did not have. At paras 36-37, he expressed the distinction thus:

In the Supreme Court the sole issue was whether this implicit limitation on the scope of the jurisdiction to review decisions of the Upper Tribunal was justified. As Baroness Hale pointed out at para 37 (and again at paras 29 and 40), the starting point was that “there is nothing in the 2007 Act which purports to oust or exclude judicial review of the unappealable decisions of the Upper Tribunal. Clear words would be needed to do this and they are not there.” Any limitation therefore had to be implicit, as the Court of Appeal had held it was. The Supreme Court accepted that a restrained approach should be taken to the granting of leave, but rejected the Court of Appeal’s distinction between errors of law and “outright” excess of jurisdiction. It is important to appreciate that both the Court of Appeal and the Supreme Court regarded the question whether there was an implicit limitation of the scope of judicial review as a question of judicial policy. The difference between them was about what the relevant policy considerations were. In the Supreme Court’s view, the main policy consideration was the undesirability of allowing the Upper Tribunal to become (in Lady Hale’s words) the “final arbiter of the law”, in case inferior courts should undermine the coherence of the law by developing their own “local law” (para 43). This concern, which was mentioned by the Court of Appeal but had not troubled them, was central to the reasoning of this court. Nothing in this court’s analysis suggests that policy considerations of this kind would have been relevant, let alone decisive, if the issue had been the meaning and effect of an ouster clause. Nothing in the judgments promotes the undesirability of “local laws” from an interpretative presumption to a constitutional principle. The real significance for present purposes of this court’s decision in Cart lies in its recognition that the rule of law does not necessarily require that the decisions of an inferior tribunal be subject to a power of review, even where they are unappealable: see in particular paras 89-90 (Lord Phillips of Worth Matravers), and paras 122-124 (Lord Dyson). Lord Dyson (with whom the rest of the court all agreed) referred to the status of the Upper Tribunal as a court performing functions equivalent to those of the High Court, and observed at para 122:

It follows from the modern authorities that the approach to be taken to section 67(8) of the Regulation of Investigatory Powers Act 2000 depends on the character of the Tribunal’s functions, the nature of the error of law of which it is accused by the appellant, and the construction of section 67(8) as applied to alleged errors of that kind.

The functions of the Investigatory Powers Tribunal are defined by section 65 of the Act. Section 65(2) is in the following terms:

Section 67 regulates the manner in which the Tribunal’s jurisdiction is to be exercised. It provides, so far as relevant:

The importance of ensuring the confidentiality of secret material is implicit in the kind of matters with which it deals, and is reflected in a number of provisions of the Act. In the first place, section 69(3) imposes a duty on the Tribunal to carry out its own investigation of complaints brought before it, and section 68 empowers it to call for the assistance of the services in question and their officials. This is an inquisitorial power in whose exercise the complainant does not participate. Secondly, section 69 empowers the Secretary of State to make rules for the Tribunal, having regard in particular to the need to secure that they are properly heard and considered and that information is not disclosed to an extent or in a manner which is “contrary to the public interest or prejudicial to national security, the prevention or detection of serious crime, the economic well-being of the United Kingdom or the continued discharge of the functions of any of the intelligence services.” Thirdly, section 66 empowers the Secretary of State by order to allocate further proceedings to the Tribunal, having regard in particular to the same matters. Fourthly, section 68(4) provides that in notifying to the complainant its determination of any proceedings or complaints, the Tribunal is to say only that they have made a determination in his favour or that they have not done so. It is plain that Parliament considered that ordinary proceedings in the High Court presented an unacceptable risk that secret material would be disclosed, contrary to the public interest, and that a major factor in the decision to allocate proceedings to the Tribunal was that its special procedures would reduce that risk. It was submitted to us that Parliament’s concerns on this score were unjustified, because as the law was (wrongly) understood in 2000, closed material procedure was available in High Court proceedings. This submission is in my view misconceived. For the purpose of construing the Act, what matters is whether Parliament had those concerns, not whether they were justified. The terms of the Act are themselves enough to show that it did.

The Investigatory Powers Tribunal is a judicial body. Schedule 3 to the Act provides that its President must hold or have held high judicial office, and its other members must either have held high judicial office or have had a relevant legal qualification for at least seven years. It is a tribunal of limited jurisdiction which enjoys neither the status nor all of the powers of the High Court. But for this purpose, as I have observed, what matters is not the label but the statutory functions of the Tribunal. Those functions are judicial in an altogether more significant sense than those of the Foreign Compensation Commission. The critical point is that the Tribunal exercises a power of judicial review which would otherwise be exercised by the High Court. By section 67(2), it is required to apply the principles which would be applied by the High Court on an application for judicial review. In relation to proceedings under section 65(2)(a) complaining of a contravention of human rights, this jurisdiction is exclusive, displacing that of the High Court. In relation to complaints under section 65(2)(b), it is a concurrent jurisdiction, but is likewise required by section 67(3)(c) to apply the principles which would be applied by the High Court on an application for judicial review. In these respects the Tribunal is not an “inferior” tribunal. Its adjudicative jurisdiction is coordinate with that of the High Court. In R (A) v Director of Establishments of the Security Service , supra , at para 23, Lord Brown of Eaton-under-Heywood, observed of the provision of section 65(2)(a) conferring exclusive jurisdiction on the Tribunal over human rights claims that the exclusion of the High Court’s review jurisdiction “has not ousted judicial scrutiny of the intelligence services; it has simply allocated that scrutiny (as to section 7(1)(a) HRA proceedings) to the IPT.” Lord Brown adopted the statement of Laws LJ in the Court of Appeal that section 65 was among a class of:

This was also the essence of the reasoning of Laws LJ in Cart . He regarded the Upper Tribunal as an “alter ego of the High Court”, in the sense that while lacking the status of the High Court, it performed within its subject area the same functions in the same judicial fashion as the High Court. It therefore satisfied the material principle of the rule of law: see para 94 of his judgment. The Court of Appeal and the Supreme Court regarded that as insufficient to warrant implying a limitation of the scope of judicial review, and nothing that I say is intended to undermine their view. But Laws LJ’s analysis is an illuminating explanation of the difference between an ouster of judicial review and a limitation of its scope to controlling the purported exercise of powers that the decision-maker did not have. That analysis is of considerable value in a case (unlike Cart ) where an express statutory provision excludes judicial review of the legal merits of a tribunal’s decisions, without impinging on the High Court’s traditional jurisdiction to review “outright” excesses of jurisdiction. The next question, to which I now turn, is whether that is the effect of section 67(8) of the Regulation of Investigatory Powers Act.

It is agreed on all sides that the meaning of this provision is a question of construction. It is also agreed that clear words are required if it is to be regarded as ousting the review jurisdiction of the High Court. However, we must not lose sight of the reason why clear words are required. The reason is, as all the authorities (and indeed Lord Carnwath in his judgment in the present case) agree, that Parliament is presumed not to legislate contrary to the rule of law. As Lord Hoffmann pointed out in R (Simms) v Secretary of State for the Home Department [2000] 2 AC 115, p 131, “that Parliament must squarely confront what it is doing and accept the political cost”. The degree of elaboration called for in a statutory provision designed to achieve a given effect must depend on how anomalous that effect would be. In this case, the words must be sufficiently clear to authorise a departure from the normal state of affairs, which is that the High Court has jurisdiction by way of review over the acts of lower courts. That is not the same as saying that the words must be such as to authorise a departure from the rule of law. There is nothing inconsistent with the rule of law about allocating a conclusive jurisdiction by way of review to a judicial body other than the High Court. The presumption against ouster clauses is concerned to protect the rule of law, which depends on the availability of judicial review. It is not concerned to protect the jurisdiction of the High Court in some putative turf war with other judicial bodies on whom Parliament has conferred an equivalent review jurisdiction. It was because Lord Brown found nothing constitutionally offensive in the allocation of specified disputes to the Investigatory Powers Tribunal that he had no difficulty in recognising in R (A) v Director of Establishments of the Security Service , supra , at para 23, that section 67(8) was “an ouster (and indeed unlike that in Anisminic , an unambiguous ouster) of any jurisdiction of the courts over the IPT.” This was no more than a dictum, on a point which was not argued. But in my opinion what Lord Brown said as a matter of impression was also correct as a matter of analysis, at any rate as applied to challenges to the Tribunal’s determinations on the merits. My reasons are as follows.

In the first place, the “jurisdiction of the Tribunal” defined in section 65(2) is to adjudicate on proceedings, complaints, references and on other matters allocated to them by the Secretary of State. By “adjudicate” I refer compendiously to the various expressions used in sections 65(2) and 67(1)-(3) to describe the Tribunal’s resolution of matters before it (“hear and determine”, “consider and determine”, “determine”, etc). The Tribunal’s “permitted field” (to use Lord Wilberforce’s phrase) plainly extended to determining questions of law arising in the course of any proceedings or complaint. In particular, it extended to determining the construction of the various enactments, such as the Intelligence Services Act 1994, under which the bodies subject to review by the Tribunal operated. There is nothing in the Act which lays down the test to be applied to any of these matters. Nor is there anything corresponding to the prescriptive regime in the Order in Council considered in Anisminic . The Act simply confers on the Tribunal the adjudicatory powers which would otherwise be exercised by the High Court acting as a court of review.

Secondly, turning to the language of section 67(8) itself, it is common ground that it falls to be construed against the background of the law declared in Anisminic and in subsequent decisions interpreting and applying it. The draftsman has deliberately chosen substantially the same formula as was considered in Anisminic (“shall not be … liable to be questioned in any court”). But it is clear that the draftsman did not intend the same result as in Anisminic . This is because he has modified the formula by adding the bracketed words “(including decisions as to whether they have jurisdiction)”. The effect of the bracketed words is to extend the ouster to the precise class of decisions which the House of Lords in that and subsequent cases had held not to be covered by the Anisminic formula, namely decisions in excess of jurisdiction. I do not, however, think that the intention was to extend it to all such decisions. The key lies in another addition to the Anisminic formula, namely the reference to an appeal. The Tribunal’s decisions “shall not be subject to appeal or be liable to be questioned in any court.” At common law, it was well established that the fact that a judicial decision was unappealable did not bar judicial review on all the usual grounds: see R (Cart) v Upper Tribunal [2012] 1 AC 663, paras 16-21 (Baroness Hale). In framing section 67(8) as it did, Parliament’s concern was plain. It was to ensure that the barring of appeals was not rendered nugatory by applications for judicial review on grounds which amounted to the same thing. Because the courts, in interpreting Anisminic , had categorised error, at any rate of law, as an excess of jurisdiction, this could be achieved only by extending the ouster clause so as to cover errors in the treatment of the merits notwithstanding that they were treated in Anisminic as an excess of jurisdiction.

Thirdly, it is true that a right to apply for judicial review is conceptually different from an appeal even if, in relation to an alleged error of law, they amount in practice to the same thing. But the concept of a judicial review by the High Court of a tribunal which is itself exercising a power of judicial review equivalent to that of the High Court, might be thought surprising. The rational course for Parliament to have adopted, if it intended to allow judicial review on the ground of error, would have been to provide for an appeal. Parliament has in fact made such provision. But under section 67(9) it has done so unconditionally only in relation to cases falling within section 65(2)(c) and (d), which are not yet in force. In other cases, including those relevant to the present proceedings, the introduction of a right of appeal is left to the discretion of the Secretary of State under section 67(8). It would in my view be wrong in principle to construe the Act as allowing for judicial review on grounds indistinguishable from an appeal on the merits, when Parliament has so carefully circumscribed the conditions on which an appeal is available.

Fourthly, as Lord Wilberforce observed in Anisminic (p 209), conclusions about the Tribunal’s “permitted field” may be derived from the “form and subject-matter of the legislation”. The main subject-matter of this legislation is secret intelligence whose disclosure would be contrary to an obvious and powerful public interest. I have drawn attention above (at para 196) to the numerous indications to that effect in the Act. Its provisions, as Lord Brown of Eaton-under-Heywood pointed out in R (A) v Director of Establishments of the Security Service [2010] 2 AC 1, para 14, “are designed to ensure that, even in the most sensitive of intelligence cases, disputes can be properly determined.” The public interest engaged was pointed out in forceful terms by Sales LJ in his judgment in the Court of Appeal: see, in particular, para 7. It needs no further emphasis from me. There is accordingly an entirely rational reason, whose significance is apparent throughout the relevant parts of the Act, why Parliament should have wished to confine the examination of these matters to a secure Tribunal and to prevent resort to the High Court, whether by way of appeal or review. The whole object of the Act in creating the Tribunal was to make resort to the High Court unnecessary by providing an alternative but equivalent right of recourse to a judicial body performing the same function. It is right to add that section 68(4), which requires the Tribunal to give an unreasoned Yes or No answer to the questions before it, would make judicial review by the High Court exceptionally difficult in most cases. In future cases, governed by the Investigatory Powers Act 2016, there will be a right of appeal, in specified circumstances, and section 68(4) will be modified so as to be subject to the Tribunal Rules. Rules made under that Act allow for more extensive disclosure of the Tribunal’s reasons. But by the time that the 2016 Act was passed the position regarding the security of information deployed in forensic litigation had been transformed by the Justice and Security Act 2013, with its elaborate provisions for closed material procedure in civil proceedings in the High Court. The courts below regarded this as a decisive consideration. I think that there are a number of decisive considerations, but this is certainly one of them.

Finally one is bound to ask forensically, if section 67(8) when read as a whole does not exclude a challenge to the merits of the Tribunal’s decisions by way of judicial review, then what else can it mean? Ms Rose implicitly accepted that if it is to have any effect at all it must oust judicial review in relation to something. She was driven to arguing that it did so only in relation to alleged errors of fact. She referred to R v Secretary of State for the Home Department, Ex p Khawaja [1984] 1 AC 74, as authority for the proposition that the factual basis of the challenged decision would otherwise have been open in principle to challenge by way of judicial review, and suggested that that was the problem to which section 67(8) was directed. I do not accept this. The question is always whether the tribunal’s decision falls within its “permitted field”. If, as Ms Rose submits, the Act on its true construction does not allow the Tribunal to err, then there is no reason to distinguish between errors of law and fact. They are both in excess of jurisdiction. In In re Racal Communications [1981] AC 374, Lord Diplock (with whom Lord Keith agreed) explicitly rejected (p 382C) a similar distinction which had been suggested by Lord Denning MR in Pearlman . Analysing that decision (at pp 383-384) he deprecated attempts to read into ouster clauses an implicit “dissection” of propositions of law from their application to facts. Lord Edmund-Davies made the same point at p 390. Judicial review commonly involves interrelated questions of fact and law, and such distinctions tend to lead to arbitrary and technical subtleties of a kind which Parliament is unlikely to have intended. Certainly, there is no trace of such a distinction in section 67(8) or anywhere else in the Act.

Ms Rose’s principal argument, however, was a reductio ad absurdum. If, she said, section 67(8) excludes judicial review, then the Tribunal’s decisions could not be reviewed even if it embarked on a dispute which was not within its subject-matter competence, or was improperly constituted, or affected by the grossest bias. This submission would require us to take an all or nothing view of section 67(8) which I regard as wrong in principle. The process of construction involved in identifying a judicial body’s “permitted field” depends, as the House of Lords pointed out in Anisminic , on an analysis of the enabling legislation to ascertain the breadth of the interpretative power conferred on it. The legislation may be more prescriptive in some respects than in others. Or it may be silent on some points, thus implicitly leaving unaffected basic common law principles such as natural justice. The Regulation of Investigatory Powers Act 2000 contains express provisions governing the constitution of the Tribunal and its subject-matter competence. It contains rules governing the Tribunal’s procedure, and authorises the making of further rules by the Secretary of State. Nothing that I have said should be taken to suggest that breach of these requirements is unreviewable. The terms of the Act place them outside the Tribunal’s “permitted field”. The same is true of principles of natural justice (such as those relating to bias), so far as they are not modified in terms by the Act. Lord Diplock in Racal , at pp 382-383, envisaged that in the case of a court charged with the resolution of questions of law it was necessary to distinguish between errors of law going to jurisdiction in the pre- Anisminic sense and errors of law within jurisdiction. As Lord Mance pointed out, applying this principle in Lee v Ashers Baking Co Ltd [2018] 3 WLR 1294, para 88 (cited above), a statutory ouster clause may be clear enough to oust review of a judicial body’s substantive decisions but not its procedural failings. In my opinion, section 67(8) is a provision of that kind. It ousts any kind of merits review of the Tribunal’s decisions but nothing more than that. This is the narrowest meaning consistent with the language and manifest purpose of the subsection. It does not oust review of those procedural failings which if made out would deprive the Tribunal of its adjudicatory competence.

The appellant’s complaint is that the Tribunal misconstrued section 5(2) of the Intelligence Services Act 1994 by holding that it authorised thematic warrants. If this was an error, then it seems to me to be clear that it was an error within the permitted field of interpretative power which Parliament has conferred on the Tribunal. Whether or not it is correctly described as an excess of jurisdiction is not the point, for the statute empowers the Tribunal to act within its permitted field irrespective of whether or not its act is so described. I conclude that the effect of section 67(8) of Investigatory Powers Act 2000 is that the High Court had no jurisdiction to entertain a challenge to the Tribunal’s decision in the present case, whether by way of appeal or judicial review.

In the Statement of Facts and Issues, the alternative case is formulated as follows:

An argument of this kind may take one or other of two forms. In its more radical form, the argument limits the sovereignty of Parliament in the name of a higher law, ascertained and applied by the court. What is said is that the rule of law is the foundation of the constitution and the source of the legitimacy of all legislation and that judicial review is its procedural embodiment. For this reason, Parliament is not competent to legislate contrary to the rule of law. This was the view tentatively expressed in an obiter dictum of Lord Steyn in R (Jackson) v Attorney General [2006] 1 AC 262, para 102, and less tentatively by Lord Hope in his observations, also obiter, in the same case, at paras 104-108. It was robustly rejected by Lord Bingham in the same case (para 9) and more fully in Chapter 12 of his book The Rule of Law (2010). I did not understand Ms Rose to be arguing that case. In its less radical form, the argument is that judicial review is necessary to sustain Parliamentary sovereignty. This is because Parliament can express its will only by written texts, to which effect can be given only if there is a supreme interpretative and enforcing authority. That authority by its nature resides in courts of law. This is the view suggested by Laws LJ in the Court of Appeal in R (Cart) v Upper Tribunal [2011] QB 120, paras 34-38. Like the principle that Parliament cannot bind itself, Parliament’s lack of competence to oust judicial review is on this view conceptual rather than normative. The point was well put by Farwell LJ in R v Shoreditch Assessment Committee, Ex p Morgan [1910] 2 KB 859 when he observed, at p 880, that “it is a contradiction in terms to create a tribunal with limited jurisdiction and unlimited power to determine such limit at its own will and pleasure - such a tribunal would be autocratic, not limited.”

The rule of law applies as much to the courts as it does to anyone else, and under our constitution, that requires that effect must be given to Parliamentary legislation. In the absence of a written constitution capable of serving as a higher source of law, the status of Parliamentary legislation as the ultimate source of law is the foundation of democracy in the United Kingdom. The alternative would be to treat the courts as being entitled on their own initiative to create a higher source of law than statute, namely their own decisions. In R (Miller) v Secretary of State for Exiting the European Union (Birnie intervening) [2018] AC 61, at para 20, the Divisional Court accepted that:

I would accept it up to a point. In reality, it is a variant of the appellant’s primary case about Parliamentary intention. If Parliament on the true construction of an enactment has created a tribunal of legally limited jurisdiction, then it must have intended that those limits should have effect in law. The only way in which a proposition can have effect in law, is for it to be recognised and applied by the courts. Parliament’s intention that there should be legal limits to the tribunal’s jurisdiction is not therefore consistent with the courts lacking the capacity to enforce the limits. Ms Rose, correctly to my mind, described this as giving effect to the sovereignty of Parliament, not limiting it. In order to escape this conceptual difficulty, Parliament would have to create a tribunal of unlimited jurisdiction or one with unlimited discretionary power to determine its own jurisdiction. A sufficiently clear and all-embracing ouster clause might demonstrate that Parliament had indeed intended to do that. But it would be a strange thing for Parliament to intend, and although conceptually possible, it has never been done.

These theoretical considerations are, however, a long way from the problem presently before us. No one contends that section 67(8) of Act makes Investigatory Powers Tribunal a tribunal of unlimited jurisdiction or that it has an unlimited discretionary power to determine its own jurisdiction. The question is how to reconcile the limited character of its jurisdiction with the language of section 67(8). For the reasons which I have given, the reconciliation is that section 67(8) does no more than exclude review by the High Court of the merits of decisions made by a tribunal performing, within its prescribed area of competence, the same functions as the High Court. It is in substance an exclusion of appeals on the merits and other proceedings tantamount to an appeal on the merits. The bracketed words referring to jurisdiction have been added because the draftsman intended that the decisions of the tribunal on the merits should be treated as within its jurisdiction notwithstanding that it was erroneous. The intention was that the exclusion of appeals on the merits and equivalent proceedings should apply notwithstanding that Anisminic had categorised some errors going to the merits as excesses of jurisdiction. None of this gives rise to the conceptual problem described above. Section 67(8) does not exclude or limit the jurisdiction of the High Court to enforce the statutory limits on the Tribunal’s powers or subject-matter competence, or the statutory and other rules of law regarding its constitution. In my opinion, Parliament does not contradict itself by enacting that notwithstanding Anisminic a decision on the merits by a judicial tribunal of limited jurisdiction exercising the same review function as the High Court is to be conclusive. As Baroness Hale put it in Cart (para 40), adopting the approach of Lord Wilberforce in Anisminic :

I would accordingly dismiss the appeal.

There are two questions:

Does section 67(8) of RIPA (“the 2000 Act”) mean that there can be no judicial review of the determinations of the IPT?

If so, is it open to Parliament to exclude judicial review of its determinations?

The second question, which questions the legality of part of an Act of our Parliament, supreme under our constitutional arrangements, could scarcely be of greater fundamental importance and sensitivity. So there is a temptation to insinuate into the answer to the first question matters which in truth are relevant only to the answer to the second question, with a view to answering “no” to the first question and thus avoiding the need to answer the second question. In my respectful view this is the temptation to which in analogous circumstances in the Anisminic case the appellate committee gave way in 1968. I agree with Sir John Laws - see para 81 above - that the committee there picked a fig-leaf with which it attempted to hide the essence of its reasoning. For proper recourse to the presumed intention of Parliament cannot justify straining the meaning of statutory words too far. The committee thereby set up 50 years of linguistic confusion for all of us who have been heirs to its decision. We should finally dispel that confusion but, in doing so, should, in answer to the first question, strive not to set up other strained meanings productive of different confusion for those who will wrestle with today’s judgments.

Omitting at this stage its strengthening words in parenthesis, section 67(8) says that, save as the Secretary of State may by order otherwise provide, decisions of the IPT should not be subject to appeal “or be liable to be questioned in any court”. On the face of it, what could be clearer? The obvious place for them to be “questioned” is the High Court in the exercise of its jurisdiction to conduct judicial review. Those words appear to exclude judicial review. Parliament recognised, however, that, contrary to their appearance, they alone were not enough to exclude it.

The problem was the decision in the Anisminic case.

Section 4(4) of the Foreign Compensation Act 1950 (“the 1950 Act”) provided, in terms which were to that extent identical to those in section 67(8), that “the determination by the [Foreign Compensation Commission] of any application made to them under this Act shall not be called in question in any court of law”. The commission had made a determination that the company was not entitled to compensation for the sequestration of its property in Egypt because of its failure to have complied with one of the perceived conditions of entitlement prescribed by an Order in Council. A majority of the appellate committee held that the commission had misconstrued the condition and upheld a declaration that the company was entitled to compensation. With respect to the distinguished members who formed the majority of the committee, I find it impossible to disagree with the dissenting conclusion of Lord Morris of Borth-y-Gest at p 194 that the commission’s error, albeit one of law, was not in excess of its “jurisdiction” - as properly understood. In what follows it will be convenient to describe that sort of error as an ordinary error of law.

The majority, however, held otherwise. Relying heavily on the policy reasons in favour of judicial supervision of some of the commission’s determinations, they worked backwards to the meaning of the word “determination” in section 4(4). As Lord Carnwath, borrowing a point made by Professor Feldman, observes in para 42 above, the company’s claim was only for a declaration rather than for an order of certiorari to quash the commission’s determination; so, if the claim was to prevail, the majority needed to conclude that the determination was a nullity and could thus be so declared. At all events they held that the word “determination” in the subsection did not include a mere purported determination. In describing it they deployed different terms which in my view amount to the same thing. Lord Reid at pp 170, 174 and 175 preferred to describe it as a “nullity”. He observed at p 171 that it was preferable not to use the word “jurisdiction” except in the narrow sense of a disentitlement to embark on the inquiry but then at p 174 he used that word in the wide sense in holding that the commission had exceeded it. Lord Pearce described it at pp 195 and 201 as a determination made in excess of jurisdiction, as did Lord Pearson at p 215. Lord Wilberforce suggested at p 207 that the words “jurisdiction”, “error” and “nullity” created problems and at p 208 he expressed a preference for describing it as “a decision made outside the permitted field”.

There is no difficulty in understanding the decision of the majority in the Anisminic case that, by section 4(4), Parliament had not precluded review of determinations which were truly nullities, in excess of jurisdiction or outside the permitted field. They might well have reached that conclusion irrespective of the meaning of the words in the subsection: see para 236 below. Had they studied in greater detail the institutional features of the commission, they might also have held that, again irrespective of the meaning of the subsection, Parliament had not even precluded review of its ordinary errors of law, such as the one before them: see paras 237 to 252 below. The problems arise from the fact that they chose to reach their decision by construction of the word “determination” in the subsection; and from the way in which they strained the meaning of the words “null”, “in excess of jurisdiction” and “outside the permitted field” so as to extend them to ordinary errors of law which, on no previous understanding of those words, would have fallen within them. Lord Pearce at p 195 gave examples of lack of jurisdiction which demonstrate the strain:

On 25 November 1982 the appellate committee decided the appeals in O’Reilly v Mackman [1983] 2 AC 237. For present purposes the facts are irrelevant. The relevance of the decision lies in the part of the speech of Lord Diplock quoted in para 54 above. The decision in the Anisminic case, he said, was that, if a statutory tribunal made an error of law, it must have asked itself the wrong question, “ie one … which it … had no jurisdiction to determine”. I, for my part, do not regard Lord Diplock as having there significantly misconstrued or oversimplified that decision. He praised it. No doubt in one sense anything which enables a judicial system to overcome obstacles to its elimination of legal error deserves praise. And he evidently saw no reason to temper his praise by reference to the confusing use of language in which the decision in the Anisminic case had been cast.

On 14 February 1985 the Bill which became the Interception of Communications Act 1985 (“the 1985 Act”) was published. By section 7, Parliament established a tribunal to investigate complaints that a communication sent to or by a complainant had been unlawfully intercepted. There was an ouster of judicial supervision of the tribunal’s decisions but in terms more comprehensive than those in section 4(4) of the 1950 Act which had failed to exclude the supervision that the majority in the Anisminic case had held to have survived. For section 7(8) provided:

We can now address the specific question central to the overall answer to the first question: what is the meaning of the words “including decisions as to whether they have jurisdiction”, written in parenthesis in section 67(8), in effect by way of repetition of words introduced by Parliament in an analogous provision in 1985?

I, for my part, am in no doubt about the answer to the specific question. In 1985 Parliament, including its drafter of the 1985 Act, was aware that its attempted ouster of judicial oversight in section 4(4) of the 1950 Act had failed. In the Anisminic case the majority of the appellate committee had used different terms to describe the sort of decisions of which judicial oversight survived the ouster. But they had been collected by Lord Diplock in the O’Reilly case into one word, namely decisions made without “jurisdiction”. Lord Diplock had delivered his speech less than three years prior to publication of the bill which became the 1985 Act.

Necessarily considered in their context, the meaning of the words in parenthesis in section 7(8) of the 1985 Act, now replicated in section 67(8) of the 2000 Act, is surely to encompass within the exclusion of judicial supervision all the decisions of the IPT in relation to its “jurisdiction”; and to ascribe to that word the strained extension of its effect adopted in the Anisminic case so as to cover ordinary errors of law as well, of course, as errors in the proper sense of it. The initial presumption that Parliament did not intend such an exclusion and the need in consequence for a strict construction of the subsection have to yield to what I consider to be the only reasonable meaning of its words, which is to the contrary.

Support for the above comes from a distinguished source. At the time of the passing of the 1985 Act Professor David Foulkes wrote the notes to it for Current Law Statutes. Of section 7(8), he wrote:

It follows that, with regret, I cannot subscribe to the interpretation of the words in parenthesis in section 67(8) favoured by Lord Carnwath. His argument is in the alternative.

His first argument, set out in para 108 above, is predicated on what I have described as the proper sense of the word “jurisdiction” in the words in parenthesis. He says correctly that the appellant’s contention is that the IPT made an ordinary error of law. So, he says, it escapes the exclusion of decisions in relation to jurisdiction provided by the words in parenthesis. But why would it escape the exclusion in the earlier words of the subsection? Where would be the logic in excluding from judicial oversight errors of jurisdiction in the proper sense, but not ordinary errors of law?

Lord Carnwath’s alternative argument, set out in para 109 above, with which Lord Lloyd-Jones agrees in para 165 above, is predicated on the extended sense ascribed to the word “jurisdiction” in the Anisminic case. Here his argument is that decisions made without “jurisdiction” in that extended sense are not decisions at all so are not excluded from judicial oversight by a subsection which refers both inside and outside the parenthesis only to “decisions”. In my opinion the argument is characteristically ingenious but too strained. It also proves too much, as becomes clear when in para 110 above Lord Carnwath seeks to ascribe some meaning to the words in parenthesis. He cites a decision of the IPT, namely C v The Police IPT/03/32/H in which, as I agree, the applicant failed because he failed to establish a fact upon which the existence of its jurisdiction depended (“a jurisdictional fact”). Lord Carnwath suggests that the words in parenthesis might exclude judicial oversight of the IPT’s determination of the absence (or presence?) of a jurisdictional fact. But why would that not be a decision made without “jurisdiction” in the extended sense of that word, as well of course as in its proper sense? In any event I see no basis for confining the wide words in parenthesis to that narrow area of the IPT’s decision-making. Indeed why should such a determination have been singled out as fit for exclusion from any judicial inquiry into the existence of evidence which entitled the IPT to make it?

Then, in para 111 above, Lord Carnwath observes that, irrespective of whether either of his constructions of section 67(8) is correct or whether the words in parenthesis are redundant, the words of the subsection are insufficiently clear to exclude judicial review of the IPT’s errors of law; and his observation echoes his earlier suggestion in paras 107 and 108 above that the only clear exclusion achieved by the subsection is of legally valid determinations. With respect, I consider the words of the subsection to be totally clear in excluding judicial review of all the IPT’s decisions; and an exclusion of judicial review in relation only to legally valid determinations seems to me to make no sense.

It also follows that, with equal regret, I cannot subscribe to the interpretation of the words in parenthesis in section 67(8) favoured by Lord Sumption in para 201 above, which he had foreshadowed in para 172 above.

The effect of Lord Sumption’s interpretation is in my view further to extend the meaning of the word “jurisdiction” beyond that favoured in the Anisminic case. The effect of that case had been to draw into the concept of an absence of jurisdiction ordinary errors of law as well, of course, as errors of jurisdiction in the proper sense. Lord Sumption interprets the word “jurisdiction” in the words in parenthesis so as to relate only to ordinary errors of law and so as no longer to include errors of jurisdiction in the proper sense. His argument depends upon the words which immediately follow the parenthesis, namely the words “shall not be subject to appeal”. He observes that ordinary errors of law could in principle be the subject of an appeal and, by some alchemy if I may respectfully say so, he reasons that these words which exclude an appeal therefore limit the meaning of the word “jurisdiction” in the parenthesis to ordinary errors of law. In my opinion the argument is again characteristically ingenious but too strained. Had Parliament’s intention been to allow judicial review of the IPT’s errors of jurisdiction in the proper sense, it would not have borrowed from the 1985 Act words in parenthesis which, on any conventional construction of them, so obviously appear to exclude it.

It is worth noting that the exclusion of an appeal achieved by the words upon which Lord Sumption relies is subject to the exception in the opening words of section 67(8), namely the words “Except to such extent as the Secretary of State may by order otherwise provide …”. It is significant that, even prior to its recent insertion into the 2000 Act of section 67A, Parliament considered that there should be, or at least could properly be, a facility for appeal against the decisions of the IPT, including no doubt against its jurisdictional decisions as well as its ordinary decisions of law. Parliament recognised however that any facility for appeal required the establishment of a confined structure apt to the sensitivity of the subject-matter of the IPT’s decisions. It thus provided in section 67(10) that any order made by the Secretary of State pursuant to the opening words of section 67(8) might include various types of provision. These include provisions at (a) for establishing a body to hear such appeals or at (c) for conferring jurisdiction to hear them on an existing court or tribunal and, in either event, at (d) for making rules in relation to the conduct of the appeals corresponding to the rules of the IPT. At that time, however, Parliament decided, as the opening words make plain, to confer upon the Secretary of State a discretion whether to establish the structure which would have enabled the appeals to be brought; and, for reasons unexplained, he has never exercised his discretion to do so. Nevertheless, for the drafter of section 67(8), it was important not to permit an appeal from the IPT to be brought outside the confined structure which Parliament envisaged. So it was essential first to retain the general exclusion of an appeal which had been provided in section 7(8) of the 1985 Act and then to subject it to the limited exception reflected in the opening words of the subsection. I find it hard to imagine that Parliament countenanced the facility for some other review of the decisions of the IPT outside the confined structure for which it was making provision.

Driven, as I am, to the view there is no defensible escape from giving to the first question the answer “yes”, I am required to proceed to address the second question; and, in doing so I must exercise a degree of caution apt to its constitutional delicacy.

At this stage, however, it is crucial to bear in mind that the complaint of which the appellant seeks judicial review is that the IPT made an ordinary error of law.

The appellant does not complain that the IPT lacked jurisdiction (in its proper sense) to determine its claim. Nor does it complain that the IPT denied to it the right to a fair hearing. A denial of a fair hearing, in particular of natural justice, is either an example of an absence of jurisdiction (Attorney General v Ryan, cited in para 123 above, p 730) or should at any rate be placed in that class for present purposes (the Cart case, Supreme Court, [2011] UKSC 28, [2012] 1 AC 663, para 38).

Had the complaint been one of lack of jurisdiction, it would have been necessary for me to undertake, albeit more slowly, the journey which Lord Carnwath valuably, albeit for his purposes unnecessarily, undertakes in paras 114 to 126 above. The modern signpost most helpful to me would surely have been found in the classic judgment of Laws LJ in the Cart case, Divisional Court, [2009] EWHC 3052 (Admin), [2011] QB 120, in particular in para 38, which Lord Sumption sets out in para 190 above. In section 65(2) of the 2000 Act Parliament has specified the limits of the IPT’s jurisdiction. So the question would have become whether, when it chooses to make a law which sets the limits of a jurisdiction, Parliament can elsewhere deprive it of an essential element of a law, namely that observance of its limits will be enforced in the courts. At first sight there is much to be said for Lord Carnwath’s answer in paras 119 to 126 above that it cannot do so.

But in my view such an answer is far less easily given to the second question if recast so as to address only Parliament’s exclusion of judicial review of an ordinary error of law. In relation to this more limited question Lord Carnwath, albeit for his purposes again unnecessarily, reasons in para 131 above “that it is ultimately for the courts, not the legislature, to determine the limits set by the rule of law to the power to exclude review”; and he observes in para 144 above that it should remain ultimately a matter for the court to determine the extent to which, in the light of its purpose and context and the nature and importance of the legal issue in question, a statutory ouster of review of an ordinary error of law should be upheld. One objection to Lord Carnwath’s observation might be that, although constructed upon the rule of law, it fails to identify any robust criterion by reference to which the court’s decision in any particular case could be foretold. At all events, for the more fundamental reasons which follow, I respectfully disagree with it.

Every legal system has to identify some end-point beyond which there can be no challenge or further challenge to a judicial decision; and it may well identify different end-points for decisions in different areas of the law and, within any one area of it, perhaps different end-points for challenges to decisions of fact, to decisions allegedly vitiated by an ordinary error of law and to decisions allegedly made in excess of jurisdiction in the proper sense.

Our system will usually provide for some, perhaps circumscribed, right to bring an appeal against, or to seek other review of, an initial judicial decision (in other words not one made on appeal or review). But it will not always do so. There is no constitutional requirement that such a right should exist: see Lord Brown of Eaton-under-Heywood in R (A) v Director of Establishments of the Security Service, cited in para 19 above, para 23. Nor is it required as part of the right to a fair trial conferred by article 6 of the European Convention on Human Rights: Delcourt v Belgium (1970) 1 EHRR 355.

In In re Racal Communications Ltd, cited in para 62 above, the appellate committee addressed a statute which empowered a judge of the High Court both to authorise inspection and to require production of company books reasonably believed to contain evidence of the commission of a criminal offence. The statute also provided that the judge’s decision should not be appealable. The complaint was that the judge had made an ordinary error of law: see the speeches of Lord Diplock at p 384 and of Lord Edmund-Davies at p 388. The committee held that the Court of Appeal had had no power to reverse the judge’s decision on appeal. Lord Diplock said at p 384:

In Pearlman v Keepers and Governors of Harrow School, cited in para 69 above, the Court of Appeal had addressed a statute which empowered a county court judge to determine whether improvements made by a tenant to his leasehold property qualified for adjustment of its rateable value. The statute also provided that the determination should be final and conclusive. The Court of Appeal, by a majority, allowed an appeal by the tenant. But in the Racal case its decision was overruled: see the speeches of Lord Diplock at p 384 and of Lord Edmund-Davies at p 390. In the Pearlman case it is therefore the dissenting judgment of Geoffrey Lane LJ which, so the appellate committee there held, provides the correct analysis of it. He said at p 74:

In R v Hull University Visitor, Ex p Page, cited in para 55 above, the appellate committee considered a complaint by a lecturer at Hull University that his dismissal had contravened its statutes and so been unlawful. The Queen, who was the visitor of the university and was acting by the Lord President of the Privy Council, had rejected his complaint. The appellate committee decided by a majority that the Divisional Court, which had quashed Her Majesty’s decision in the course of conducting a judicial review, had lacked jurisdiction to entertain it. Lord Browne-Wilkinson, for the majority, explained at p 702 that the university statutes were not the general law of the land. But his conclusion at p 704 remains useful and was as follows:

In the Cart case the challenge was to a refusal on the part of the Administrative Appeals Chamber of the Upper Tribunal to grant permission to appeal to it. Parliament, which had designated the tribunal as “a superior court of record”, had excluded a right of appeal against its refusal of permission. Could there, however, be a judicial review of it? Parliament had not expressly excluded it. The complaint was of an ordinary error of law on the part of the tribunal. The Divisional Court, cited in para 116 above, and the Court of Appeal, [2010] EWCA Civ 859; [2011] QB 120, had both held, for slightly different reasons, that the institutional features of the Upper Tribunal were such that its unappealable decisions could be the subject of judicial review when, but only when, they were said to be in excess of jurisdiction. By the time when the case reached the Supreme Court, cited in para 76 above, the government had accepted that analysis: see Lord Dyson, para 108. So by that time the only remaining question was whether judicial review could extend to ordinary errors of law alleged to have been perpetrated by the tribunal in making unappealable decisions. This court’s answer was that there could be judicial review of such errors in limited circumstances not present in the case before it. For current purposes the great importance of the decision lies in the observations of Lady Hale, in a judgment with which all the other members of the court agreed, at para 40. Lord Sumption has quoted them in para 211 above. Lady Hale there recognised that, although it had not done so, Parliament might successfully have ousted judicial review of ordinary errors of law made by a tribunal of limited jurisdiction such as the Upper Tribunal.

We therefore see that the High Court judge in the Racal case, the county court judge in the Pearlman case and the Upper Tribunal in the Cart case can, in the course of making unappealable decisions, make ordinary errors of law of which Parliament has power to exclude judicial review.

Does the IPT stand so differently from them as to mandate a different conclusion?

The answer is to be collected from examination of its institutional features, of which I now offer a brief summary.

The President of the IPT must hold or have held high judicial office: the 2000 Act, Schedule 3, paragraph 2(2). At present the incumbent is Singh LJ.

Its other members must either hold or have held high judicial office or must be UK lawyers of at least seven years’ standing: Schedule 3, paragraph 1(1). At present the other members are two High Court judges in England and Wales, a former High Court judge in Northern Ireland, and six other distinguished Queen’s Counsel including a practitioner in Scotland.

In exercising its jurisdiction the IPT is required to apply the principles which a court would apply on an application for judicial review: section 67(2) and (3)(c) of the 2000 Act. So the function of judicial review of the lawfulness of the actions of the intelligence services has therefore been allocated to it and, so this court held in the A case, exclusively so.

The need for the allocation reflects the sensitivity of any inquiry into the lawfulness of such actions and therefore the unique raft of provisions devised for the conduct of the IPT, as set out in section 68 of the 2000 Act and in the Investigatory Powers Tribunal Rules, first those dated 2000 (2000 No 2665) and now those dated 2018 (2018 No 1334), made under section 69 of it. These provisions reflect its investigative duty under section 67(3)(a) and (b) and include, by section 68(1), a right, subject to the rules, to determine its own procedure and thus to adopt an inquisitional process and, by rule 13(1) of the current rules, freedom from rules of evidence. Disclosure of the nature of the complaint and of any information or document provided to it from any source is closely circumscribed by what is now rule 7, as is the content of the notification to the complainant of its ultimate determination under section 68(4) and what is now rule 15. In the Big Brother Watch case, cited in para 28 above, the European Court of Human Rights said at para 255:

The IPT does not form part of Her Majesty’s Courts and Tribunal Service. In effect it has total autonomy. In his Report of the Review of Tribunals dated March 2001 Sir Andrew Leggatt said at para 3.11 that the IPT’s concern with security required it to be separate from all other tribunals and that the Senior President of Tribunals would not be in a position to take charge of it.

Parliament has therefore conferred both independence and authority upon the IPT. In the A case Lord Brown, with whom all other members of the court agreed, endorsed at para 23 the conclusion of Laws LJ in the court below that the IPT was “a judicial body of like standing and authority to that of the High Court”.

In the above circumstances I conclude that Parliament does have power to exclude judicial review of any ordinary errors of law made by the IPT. My answer to the second question posed at the outset of this judgment, if limited to the sort of determination relevant to this case, namely to an ordinary determination of law, is “yes”.

So I would have dismissed the appeal.