November 27

Paul Daly: Privacy International Symposium – Three Aspects of Anisminic

Privacy International has brought judicial review proceedings against the Investigatory Powers Tribunal, in respect of the Tribunal’s treatment of a complaint about computer hacking by GCHQ. At first instance ([2017] EWHC 114 (Admin)) and on appeal ([2017] EWCA Civ 1868), it has been held that judicial review of the Tribunal is precluded by the ouster clause contained in s. 67(8) of the Regulation of Investigatory Powers Act 2000. A seven-judge bench of the Supreme Court (Lady Hale, Lord Reed, Lord Kerr, Lord Wilson, Lord Sumption, Lord Carnwath and Lord Lloyd-Jones) will hear Privacy International’s appeal in December.

This online collection grows out of a symposium held by the Centre for Public Law at the University of Cambridge on October 20, 2018, organised by Dr Paul Daly (University of Cambridge). Symposium participants have agreed to post short contributions on this important litigation.

I want to explore three aspects of the decision in Anisminic v Foreign Compensation Commission [1969] 2 AC 147 which are relevant to the Privacy International ouster clause litigation. I will explain their relevance by reference to comparative materials.

First, the relationship between form and substance. In Anisminic a majority of the House of Lords held that a clause purporting to oust the jurisdiction of the courts to review any “determination” of the Commission was ineffective in respect of a decision tainted by legal error. A “determination” which transgressed the principles of administrative law was a nullity; as it was beyond the jurisdiction of the Commission, it was no determination at all in the eyes of s. 4(4) of the Foreign Compensation Act 1950.

This reasoning could justly be described as formalistic. But it has a strong substantive underpinning. Giving effect to the ouster clause would have prevented any consideration by the courts of the lawfulness of the Commission’s decision, breaching fundamental and well-settled constitutional principles (Wade, Constitutional Fundamentals (Stevens & Sons, London, 1980), at p. 82).

The importance of substance can be perceived in cases from other jurisdictions: in New Zealand, where the result of giving effect to an ouster clause would be the creation of dictatorial power, Anisminic has been applied with full force (Bulk Gas Users Group v Attorney-General [1983] NZLR 129) but where the supervisory jurisdiction has, in substance, been preserved, courts have not engaged in Anisminic-style reasoning (H v Refugee and Protection Officer [2018] NZCA 188); and in Ireland, where individuals enjoy a constitutional right of access to the courts (including the constitutionally entrenched supervisory jurisdiction of the High Court), the constitutionality of ouster clauses turns on the substantive question of whether they prevent individuals from exercising their rights (In re Article 26 and the Illegal Immigrants (Trafficking) Bill 1999 [2000] 2 IR 360; White v Dublin City Council [2004] 1 IR 545).

Pace Sales LJ, this means that the application of Anisminic can never simply involve a “short point of statutory construction” (R (Privacy International) v Foreign and Commonwealth Secretary [2017] EWCA Civ 1868, at para. 24). Sales LJ’s conclusion that s. 67(8) of the RIPA is an effective ouster may yet be correct, but for reasons of substance, not solely reasons of form. For the same reason, Leggatt J was not necessarily correct either, at first instance, to insist that the reasoning in Anisminic is “just as applicable in the present case” (R (Privacy International) v Investigatory Powers Tribunal [2017] EWHC 114 (Admin), at para. 55).

What matters is not the linguistic precision of the drafter but whether issues of legality, rationality and procedural propriety can be addressed by an independent and impartial tribunal. Indeed, the emphasis on substance over form would support the conclusion that, in principle, a body such as the Investigatory Powers Tribunal could be equipped to exercise a supervisory jurisdiction over the security services — the relevant issue is whether it is so equipped, as a matter of substance.

Second, there is a difference between the channelling and excluding of judicial review. Contrast Anisminic with Smith v East Elloe Rural District Council [1956] AC 736, where a six-week time limit for contesting compulsory purchase orders in the High Court  was held to oust even an allegation that an order had been made in bad faith (that is, it was not an “order” within the meaning of the Act). Lord Denning MR subsequently reconciled East Elloe and Anisminic by making a distinction between the channelling and excluding of judicial review:

[I]n the Anisminic case the Act ousted the jurisdiction of the court altogether. It precluded the court from entertaining any complaint at any time about the determination. Whereas in the East Elloe case the statutory provision has given the court jurisdiction to inquire into complaints so long as the applicant comes within six weeks. The provision is more in the nature of a limitation period than of a complete ouster (R v Environment Secretary, ex parte Ostler [1977] QB 122, at p. 135).

Again, this is reflected in the New Zealand and Irish cases: channelling the supervisory jurisdiction by means of time limits or requirements to pursue alternative remedies is one (legitimate) thing; excluding it altogether is quite another, illegitimate enough to justify recourse to Anisminic-style reasoning. Leveson P fully grasped the significance of the channelling/excluding distinction at first instance, focusing on whether the Tribunal was “exercising powers of judicial review comparable to those of the High Court” (at para. 42. See also R (Privacy International) v Foreign and Commonwealth Secretary [2017] EWCA Civ 1868, at para. 38, per Sales LJ). The relevant question for him was the effectiveness of the Tribunal’s powers of oversight, for if they are effective, then s. 67(8) represents a channelling of judicial review, not an exclusion.

Where the supervisory jurisdiction of the superior courts is not compromised — where the legality, rationality and procedural propriety of administrative action can still be assessed — a provision will be treated as channelling judicial review. By contrast, where a statutory provision would frustrate the exercise of the supervisory jurisdiction, by excluding judicial review, courts will treat the underlying administrative decision as a nullity (a “decision”). Put another way, the heavy analytical lifting is done by substance — the principles and purposes of judicial review — not form — the concept of “jurisdiction” (and its offspring, “nullity”).

Third, ouster clauses are not to be interpreted out of existence. One of the striking features of Anisminic is the repeated insistence by the members of the majority that the ouster clause can protect some errors from judicial oversight. Although English law has subsequently moved on so far as to bring (almost) all errors of law within the supervisory jurisdiction of the High Court, the proposition that the interpretation of an ouster clause is not an all-or-nothing affair — either it applies with full force and effect or has no effect at all — is borne out by comparative analysis.

The Australian Constitution has been interpreted as protecting the supervisory jurisdiction of state supreme courts, which includes judicial review for jurisdictional error. In Kirk v Industrial Relations Commission (NSW) (2010) 262 ALR 569 the majority of the High Court of Australia held that an ouster clause had to “be read in a manner” respectful of constitutional limitations (at para. 101). Indeed, the majority read the provision down so that it applied only to non-jurisdictional error (at para. 104). The Supreme Court of Canada has engaged in strikingly similar reasoning, albeit in a different constitutional framework (Crevier v. A.G. (Québec) [1981] 2 SCR 220).

What force short of full force and effect can the courts give to an ouster clause? Ouster clauses create internal contradictions in statutes. On the one hand, a statute sets out the boundaries of a decision-maker’s authority. On the other hand, an ouster clause would seem to give the decision-maker unlimited authority to redraw these boundaries. Such an internal contradiction calls for “a process of reconciliation” to give effect to the “whole legislative instrument” (R v Hickman; Ex parte Fox and Clinton  (1945) 70 CLR 598 at 617, per Dixon J).

One possibility, as in Kirk, is to restrict the scope of an ouster clause to non-jurisdictional error. Another possibility, pursued by the Supreme Court of Canada in a long line of cases, is to take an ouster clause as a justification for a more deferential approach to judicial review of questions of law, on the basis that the legislature intended to signal, by way of an ouster clause, that the courts should respect the wisdom of the primary decision-maker designated by statute. The difference between these approaches is that Australian commits some matters exclusively to a decision-maker shielded by an ouster clause whereas the Canadian would subject any decisions shielded by an ouster clause to deferential review.

Both possibilities are open in Privacy International, in response to the internal contradictions created in RIPA by s. 67(8). As Tom Hickman has argued ([2018] PL 584), s. 67(8) could be interpreted as committing non-jurisdictional matters of fact to the exclusive determination of the IPT. Alternatively, s. 67(8) could be interpreted as requiring the courts, on judicial review of the IPT, to exercise restraint before declaring that the IPT has made a reviewable error. Such restraint might take the form (as in R (Cart) v Upper Tribunal [2012] 1 AC 663) of a limitation on the types of issue that the High Court can review; or the form (as in the Canadian cases) of a limitation on the types of error that the High Court can correct (for instance, unreasonable errors of law or fact).

In choosing between these interpretive possibilities, the form/substance and channelling/excluding distinctions are of assistance. Much depends on the characteristics of the IPT and what it is equipped, in substance, to do. If its expertise lies in the determination of complex factual issues — which will often reveal sensitive information relating to national security — then it would make sense to interpret s. 67(8) as excluding judicial review of such issues. This could be achieved by adopting Tom Hickman’s argument, or by following a Cart-type approach. These would both be consistent with channelling rather than excluding independent and impartial oversight of administrative action. However, if there is a concern for keeping sensitive national-security matters away from the courts, a Canada-style approach would seem less appropriate, because it would leave open the possibility of review of such matters: deferential review is nonetheless review conducted in public. Alternatively, if the IPT has relevant expertise on some issues of law, this would favour a Cart-type approach. The Supreme Court would be able to develop criteria for judicial review which distinguish  technical or fact-sensitive questions of interpretation on which the IPT might have relevant expertise, from those important general questions that the High Court is well-placed to resolve authoritatively. Again, appropriate oversight would be channelled, not excluded, with substantive considerations to the fore.


Dr Paul Daly is Senior Lecturer in Public Law at the University of Cambridge.