Thomas Fairclough: Privacy International: Constitutional Substance over Semantics in Reading Ouster Clauses

I have previously written on this blog and elsewhere about statutory interpretation and the rule of law. In the previous blog post I stated that the idea “that the courts will not allow the executive to escape their jurisdiction is well established as part of the rule of law” and referenced, inter alia, Anisminic Ltd v Foreign Compensation Commission [1969] 2 AC 147 (HL) to support this view.

This is an overarching principle, which has manifested itself most obviously in the courts’ treatment of so-called ouster clauses. The traditional view is that such clauses will be interpreted in such a way that does not preclude judicial supervision of the administrative process. The reasoning for this is simple: every body of the state derives its legal powers from law and, because of this, to be acting lawfully one has to act within one’s powers. A necessary component of this is that there must be, in a country governed by the rule of law, access to the courts to provide redress if a body has acted ultra vires (this is most staunchly put by Lady Hale and Lord Hope in Jackson v Her Majesty’s Attorney General [2005] UKHL 56, [2006] 1 AC 262).

Yet in R (Privacy International) v Investigatory Powers Tribunal [2017] EWCA Civ 1868 (“Privacy International”) (on appeal from the High Court) we saw that rarest of public law occurrences: the upholding of a statutory provision that purports to exclude judicial review of an administrative decision. This case is likely to appeal but it is still worth noting some important points in this piece.

In short my main argument is that in Privacy International the Court of Appeal focused too much on the language on the statute and not enough on the substantive principles at play; whilst I do not make comment on whether the ouster clause should have been given effect to I do say that, if it is to be given effect to, it should be for the right reasons, which is to say it must not be a threat to the rule of law. I suggest a focus on the centrality of the rule of law would make better sense than merely looking at the language and Parliamentary context of the statute in examining ouster clauses.

1. Privacy International: Facts, Issues, and Judgment

The Investigatory Powers Tribunal (“IPT”) is a specialist tribunal set up by the Regulation of Investigatory Powers Act 2000 (“RIPA”) in order to supervise, inter alia, the conduct of the intelligence services (s 65 RIPA). The membership of the IPT is made up of five High Court Justices and other distinguished lawyers and it has its own procedural rules. The idea is that the IPT must balance national security and human rights.

This is all, one might expect, quite normal and expected; we are used to the idea that sometimes different procedures have to take place in certain circumstances and cases affecting national security seem a paradigm case. These can, the authorities suggest, still protect the citizen’s right to a fair hearing and effective remedy (see, for example, Kennedy v United Kingdom (2011) 52 EHRR 4, [18]). I do not intend to question that conclusion here and only note it because it will become important that no one is suggesting, for present purposes, that the IPT does not provide an Article 6 compliant fair hearing (and a common law duty of fairness compliant approach).

Privacy International made a complaint to the IPT that GCHQ, which IPT regulates, had been involved in unlawful activity involving computer hacking. In short, the complaint said that GCHQ had not been acting pursuant to a lawful warrant because the warrant provided was overly broad.

The IPT, as a preliminary issue, decided that warrants could be given in general, broad terms and still fall within RIPA. This means that warrants can be given that are unspecific and give GCHQ a wide-range of power predicated on an opaque basis. Privacy International wished to challenge the IPT’s ruling as to the proper interpretation of RIPA and, by extension, challenge the legality of the warrant itself.

There was no right to appeal in RIPA so Privacy International commended judicial review proceedings against the IPT in the High Court. However, section 67(8) of RIPA states:

Determinations, awards, orders and other decisions of the Tribunal (including decisions as to whether they have jurisdiction) shall not be subject to appeal or be liable to be questioned in any court.

The Divisional Court decided that the effect of that provision is that the IPT is not amenable to judicial review and therefore dismissed Privacy International’s application. Privacy International appealed to the Court of Appeal, whose decision is discussed in this post.

Despite the advocates on both sides arguing in terms of the rule of law (more on which below) Sales LJ, with whom Flaux and Floyd LJJ agreed, stated that this case turned “on a short point of statutory construction” ([24]). Indeed, Sales LJ put great store in the fact that that s 67(8) RIPA was drafted differently to the provision in Anisminic, which referred to ‘determinations” being immune from challenge. RIPA, however, included not only “determinations” but also decisions “as to whether they have jurisdiction”. For Sales LJ this precluded using the Anisminic approach to ouster clauses, that is, it precluded saying that the clause stopped review of legal decisions (as opposed to decisions made without jurisdiction, which were not caught by the ouster clause in Anisminic).

Likewise, the linguistic arguments are “strongly supported by the statutory context in which section 67(8) appears” ([42]). It was “clear that Parliament’s intention in establishing the IPT …was to set up a tribunal capable of considering claims and complaints against the intelligence services under closed proceedings” ([42]). For Sales LJ giving effect to the ouster clause “promotes this purpose” and, conversely, reading around the clause would “subvert” the purpose. ([43]).

2. Statutory Language and the Centrality of the Rule of Law

On one approach this is correct: the language is different. However, the question then becomes why does it matter to such a substantive extent as to render such a radically different outcome to other cases? It is hard to not agree with Elliott, who says that reading around the ouster provision in RIPA might appear “strained” but  “it is not clear that it would be substantially more strained than that which was rendered in Anisminic itself”.

Sales LJ says that Parliament’s intention was “clear” in the case but, even if we accept Parliament can in fact legislate to preclude judicial review (cf. Lady Hale’s comments in Jackson), is Parliament’s intention really any more “clear” than in other constitutional cases such as Anisminic, Evans, or UNISON? Indeed, in Evans Lord Wilson rigorously demonstrates that the Parliamentary context strongly demonstrated an intention for s 53 FoIA to allow the executive to not comply with judicial orders but this was not enough for the majority.

Therein lies the point: we may be fairly sure that Parliament wanted to do something that cuts across the rule of law but that is not enough; even on the orthodox view we have to be sure and it is not clear in this case why the statutory wording or Parliamentary context makes us sufficiently more sure than in other cases. The point of the principle of legality is that it acts as an a priori presumption against interference with the rule of law; for Sales LJ to say that Parliament’s intention was “clear” enough to displace this constitutional fundamental is really rather bold and, moreover, inconsistent with previous cases.

One way to resolve this could be to say that Sales LJ got the right result for the wrong reasons. Instead of focusing on statutory language and Parliamentary context he should have focused, as indeed James Eadie QC did on behalf of GCHQ ([23]), on the rule of law. Sales LJ notes this in his judgment but then immediately moves on to statutory language and Parliamentary context. This is a mistake.

The central justification for the usual approach to ouster clauses being read around is that of access to the courts and judicial oversight of the executive in line with the rule of law. It is about making sure the citizen is not being treated arbitrarily by the state. It is an essential check on state power.

One could simply argue, in this case, that such a justification is not engaged. This case is substantially different in principle to, say, Evans in that the latter involved shielding the executive from judicial orders. That is not the case here: the rule of law is not threatened when an independent judicial body is at work, which the IPT arguably is ([23]). There is a principled difference and so, on this approach, it makes sense to give effect the ouster clause when it might not have made constitutional sense to do so in, for example, Anisminic (cf. David Feldman’s analysis of Anisminic and the Foreign Compensation Commission).

The sidelining of the rule of law by Sales LJ is regrettable. It ignores the centrality of the rule of law; it ignores the “controlling factor on which our constitution is based” (Jackson [107] per Lord Hope). The principle of legality, on which the presumption against ouster clauses is based, is itself premised on the idea of the rule of law.

It is the perceived threat to the rule of law that ouster clauses represent that justifies interpretations that do not give effect to them. It is not, as Sales LJ relies on, so much about the semantics as the substance. The language of the statute, whilst different, is not what should be the key motivator for giving effect to the ouster clause in RIPA; instead, the ouster clause should be upheld (if it should) because it is not in defiance of the rule of law. The sidelining of this by Sales LJ is in itself constitutionally difficult and makes his judgment far less constitutionally transparent than both Sir Brian Leveson P (upholding the ouster clause) and Leggatt J (dismissing the ouster clause without formally dissenting) in the High Court.

How Should the Supreme Court Proceed?

This section works on the basis the case will appeal, which seems likely. I have not given a firm view either way on the merits of the case. All I have sought to do is locate the real debate vis-à-vis the ouster clause, which is the presumption against them operates to protect the rule of law; if the rule of law is not threatened then there is far less constitutional impetus to read around the clause. Discussion of semantics obscures this fundamental constitutional fact.

Yet I note the following: there has been an assumption in both the High Court and Court of Appeal that s 67(8) RIPA has to be read as one provision and is binary: either the ouster clause has effect or it does not. Yet there is no reason why this should be the case. As argued for above, the effect of an ouster clause depends on the threat it poses to the rule of law. The degree to which the ouster clause has effect should surely correspond to the degree of the threat.

It follows that not all parts of the clause have to be given equal effect. As noted above the section reads “determinations, awards, orders and other decisions of the Tribunal (including decisions as to whether they have jurisdiction)” shall not be questioned. It is might be legitimate to say that, due to the composition of the IPT, that its determinations and orders should not be liable for judicial review but that its decisions as to jurisdiction should be because, otherwise, there is nothing to stop it embarking on inquiries it has not right to embark upon, which even on pre-Anisminic jurisprudence is reviewable by the High Court. I do not think this is a particularly novel approach to the principle of legality. It would involve reading “including decisions as to whether they have jurisdiction” as not covering legally incorrect (i.e. purported) decisions about jurisdiction, which may seem to be a stretch of the statutory language but not such a stretch compared to, say, Anisminic itself.

Nothing, on the current approach, stops the IPT from engaging on social security matters or immigration hearings. This is unlikely but the point is that it would be constitutionally abhorrent for it to do so. Yet if the ouster clause is read so that there could be no questioning of the IPT’s decisions as to jurisdiction then there can be no challenge to its own effective expansion of its jurisdiction. From this perspective one can readily argue that the purported restriction on challenging questions as to jurisdiction is more offensive to the rule of law than a restriction on challenging an order of the IPT and so the court should treat the two aspects of the provision differently. There is nothing to stop it doing so.

Similar analysis was undertaken in R (Cart) v The Upper Tribunal [2011] UKSC 28 where the Supreme Court, whilst not saying it could not or would not permit the High Court to judicially review the Upper Tribunal, limited the circumstances in which it could do so. This, it was said, was legitimate because the functions and composition of the Upper Tribunal make it not dissimilar to the High Court but, as a creature of statute, it was legitimate for the High Court to keep a check on its powers (though note the Supreme Court said that judicial review should be available in cases that are wider than jurisdictional challenges). In short, the Supreme Court, if this case appeals, ought to be nuanced in its approach and recognise that some parts of the provision may be more of a threat to the rule of law than others.

Final Thoughts

The point this piece seeks to make is that the Court of Appeal ought to have undertaken a more nuanced approach that focused on the substance of a case; the more one reading of a statute offends the rule of law the less ready should a court be to give effect to such a reading. A focus on semantics, whilst important, is of secondary importance compared to the underlying constitutional norms at play. From the balance of these norms comes the doctrinal architecture of administrative law but we should not forget the normative foundations on which it is built. In Privacy International key issues are at stake: the relationship between the courts and legislature and courts and administrative bodies, the interface between the rule of law and statutory language, and so on.

Sales LJ possibly sidelined these issues to focus on the language and Parliamentary context of RIPA, which is a mistake. Whilst Sales LJ may well have felt as though he could not go further with interpretation to read around the ouster clause he should have spelled out the constitutional substance that was likely lurking beneath his semantic reasoning. To do otherwise undermines the importance of the issues, risks coming to the wrong result, and obscures the process of judicial reasoning in such cases, which is largely about many constitutional principles and their relationship with each other. It will be incumbent on the Supreme Court, if the case is appealed, to rectify these errors, though that is not the say that they should come to a different result, which is not something this piece comments on.

Thomas Fairclough is an AHRC-funded PhD Candidate at the University of Cambridge.

(Suggested citation: T. Fairclough, ‘Privacy International: Constitutional Substance over Semantics in Reading Ouster Clauses’, U.K. Const. L. Blog (4th Dec. 2017) (available at