ASK any UK lawyer to name a seminal constitutional law case and there is a fair chance that he or she will cite Anisminic Ltd. v Foreign Compensation Commission [1969] 2 A.C. 147, in which the House of Lords reconceived the notion of jurisdictional error, interpretively neutralised a statutory provision that appeared to displace judicial review, vindicated the rule of law, and, at least on one analysis, implicitly raised questions about the extent of Parliament's legislative capacity. If Anisminic – decided just over half a century ago – was one of the blockbuster constitutional judgments of the last century, then R. (Privacy International) v Investigatory Powers Tribunal [2019] UKSC 22, [2019] 2 W.L.R. 1219 is its early twenty-first-century counterpart.

The claimant sought judicial review, arguing that the Investigatory Powers Tribunal (IPT) had misinterpreted section 5 of the Intelligence Services Act 1994 (ISA), leading it erroneously to conclude that the Secretary of State could authorise computer hacking on a thematic basis (e.g. in respect of classes of people). Section 67(8) of the Regulation of Investigatory Powers Act 2000 (RIPA), however, appeared to stand in the way of such a claim. It provided that, except to such extent as the Secretary of State by order provided otherwise, “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”. Since the discretion in section 67(8) to provide for appeals had not been exercised, relevant IPT decisions would be legally impregnable absent judicial review. The question in Privacy International was whether, properly construed, the legislation accorded that status to such decisions.

In the Divisional Court ([2017] EWHC 114 (Admin), [2017] 3 All E.R. 1127), Sir Brian Leveson P. considered (at [42]) that there was a “material difference” between the sort of decision-making body whose decision was impugned in Anisminic and the IPT, since the latter was itself “exercising a supervisory jurisdiction over the actions of public authorities”. This led him to the conclusion that judicial review did not lie. Although he did not go as far as to dissent formally, Leggatt J. plainly had grave misgivings, arguing (at [49]) that exempting the IPT from judicial review, and thereby rendering it a “legal island”, would be inimical to the rule of law. The claimant also failed in the Court of Appeal ([2017] EWCA Civ 1868, [2018] 1 W.L.R. 2572). Sales L.J., giving the only reasoned judgment, considered (at [24]) that the case raised nothing more than “a short point of statutory construction”. He disposed of it in the defendant's favour, fastening upon the reference in section 67(8) of RIPA to “decisions as to … jurisdiction”. Such language, he said (at [26]), meant that all decisions, including those founded on “an erroneous view of the law”, were immune from review. This meant, said Sales L.J., that it was not open to the court to emulate the House of Lords in Anisminic, which held that “purported” – that is, legally flawed – determinations were unprotected by the ouster.

However, when Privacy International reached the Supreme Court, the majority took a different view. The majority concluded that section 67(8), properly construed, protected only decisions that were legally valid. This conclusion, and the reasoning that yielded it, raises three fundamental sets of public law issues that we explore in the remainder of this note: the notions of jurisdiction and error of law, and the potential ramifications of the majority's stance on these matters for our understanding of the conceptual framework of the law of judicial review; the content of the rule of law and its implications for ouster clauses; and the place and meaning of the doctrine of parliamentary sovereignty in the contemporary UK constitution.

First, Privacy International invites us to re-examine the foundations of judicial review. Why should courts intervene to correct decisions of inferior courts or tribunals? Is the trigger the commission by the decision-maker of a jurisdictional error (which might or might not also be an error of law)? Or is the trigger that an error of law (which might or might not also be a jurisdictional error) has been made? All of the judgments accept that, whether intended or otherwise, Anisminic abolished the distinction between jurisdictional and non-jurisdictional errors of law. All errors of law are now classed as jurisdictional. But this leaves open the question of what is in the conceptual driving seat: is it the notion of jurisdiction, or is it the notion of error of law? One way of understanding Anisminic and its aftermath is to infer that jurisdiction remains the organising concept. Section 4(4) of the Foreign Compensation Act 1950 stated that the decision-maker's determinations “shall not be called in question in any court of law”. Yet, the section did not exclude the ability of the court to question a “purported” determination – a determination that never was. If the mistake in Anisminic was jurisdictional, then the commission was not empowered to make its determination. A lack of power to act makes the determination a nullity – “purported” and not real, thereby unprotected by section 4(4). However, it can also be argued that, in reality, it is the classification of the determination as one that was erroneous in law that makes it “purported” as opposed to real, and thus unaffected by the ouster. The decision-maker should only make legally correct determinations. Those tainted by legal errors are only “purported” and not real.

To what extent does Privacy International move this debate on? The contrast between the majority and the minority judgments might appear to indicate that judicial review is based on errors of law as opposed to errors of jurisdiction. Lord Carnwath (with whom Lady Hale and Lord Kerr agreed) criticised the distinction between jurisdictional and non-jurisdictional errors as “based on foundations of sand” (at [84]). He regarded the connection between jurisdictional errors and the concept of nullity as a “fig leaf” (at [81]–[83]) and “artificial” (at [129]). Section 67(8) specifically stated that decisions as to jurisdiction of the IPT could not be questioned in a court of law. The majority concluded that the provision did not exclude judicial review of legal errors made by the IPT. Although Anisminic has been interpreted as demonstrating that all legal errors are classified as jurisdictional errors, it is not the case that all jurisdictional errors are legal errors. In order to overcome the general presumption against the removal of judicial review of the court, Lord Carnwath and Lord Lloyd-Jones read section 67(8) as only removing judicial review over jurisdictional errors of fact. This may suggest that judicial review, at its core, is concerned with the correction of legal errors. If legislation clearly excludes judicial review over jurisdictional errors, this will be read so as only to remove judicial review over jurisdictional errors of fact, not jurisdictional errors of law.

This interpretation is reinforced when we contrast the approach of the majority with that of the minority. Lord Wilson went so far as to “deprecate” the broader interpretation of Anisminic which read the case as equating all legal errors with jurisdictional errors (at [219]). This strongly implies that courts should correct jurisdictional, not legal errors. Lord Sumption (with whom Lord Reed agreed) interpreted section 67(8) differently. On his analysis, it is designed to prevent courts from questioning the merits or the substance of decisions of the IPT – jurisdictional, legal or otherwise. However, the clause does not oust judicial review of decisions of the IPT tainted by procedural as opposed to substantive errors. An example of such an error is found in Lee v Ashers Baking Co. Ltd. [2018] UKSC 49, [2018] 3 W.L.R. 1294, where the Northern Ireland Court of Appeal mistakenly considered that proceedings were closed once judgment had been delivered. However, as the judgment had not been lodged, the proceedings were still open. This approach does not fit with an analysis that would regard any legal error – be it one as to substance or procedure – as capable of rendering a decision of the IPT only a “purported” one, such that it could be questioned by a court of law. Nor does it fit with a conceptual framework of judicial review which places an analysis of jurisdiction at its core.

However, to read Privacy International as supporting a move from jurisdiction to legal errors as the foundation of judicial review fails to recognise a more profound distinction between the decision of the majority and the minority. It also underestimates the impact of R. (Cart) v Upper Tribunal [2011] UKSC 28, [2012] 1 A.C. 663. Neither Lord Carnwath nor Lord Sumption reasoned in a manner that focuses on a distinction between errors of law and errors of fact or between jurisdictional and non-jurisdictional errors. Indeed, Lord Carnwath was just as critical of the distinction between legal and factual errors – describing it as not “clear cut” (at [134]). Lord Carnwath adopted a more flexible approach – one that he presented as both principled and pragmatic – to determine the extent to which courts should review decisions of a particular inferior court or tribunal (at [130]–[132]). This approach focuses on the need to uphold the rule of law, whilst recognising the specific functions of the IPT and the prevention of the creation of “islands” of law that are beyond the reach of the senior courts. Lord Sumption focused on interpreting the provisions of RIPA setting out the powers of the IPT (at [200]–[204]). Importantly, however, neither focused on classifying an error as jurisdictional or legal as the predominant means of determining the scope of the court's powers. Consequently, Privacy International has a more profound impact on our approach to both jurisdictional and legal errors. Rather than looking to classify an error as jurisdictional or legal, the majority of the Justices focused on the rule of law, statutory interpretation and the features of the inferior court or tribunal in question to determine the scope of judicial review. A classification of an error as jurisdictional or legal does not, in and of itself, determine the scope of judicial review, though both are relevant.

This leads to the second difference between the majority and the minority, which in turn casts light on the implications of Privacy International for the rule of law. The majority and minority differ both in terms of their definition of the rule of law and its relative importance. Whilst Lord Carnwath regarded section 67(8) as breaching the rule of law, Lord Sumption did not. This is explained both by their approach to the rule of law and its specific requirements. Lord Carnwath adopted a more expansive definition. He did so relying on section 1 of the Constitutional Reform Act 2005, which provides statutory recognition of the rule of law as a principle of the UK constitution. However, it does not define what is meant by the rule of law. Lord Carnwath saw this as Parliament empowering the courts to determine the content of the rule of law. For Lord Carnwath, the rule of law normally requires judicial review over decisions of inferior courts and tribunals, particularly as regards determinations made in excess of or in abuse of the jurisdiction of these courts and tribunals. Moreover, the rule of law requires that specialist inferior courts and tribunals should not be able to create “islands” of law. Higher courts need to check the legal determinations of inferior courts and tribunals to maintain legal certainty, particularly when their determinations have implications beyond their specialist area of the law. In contrast, Lord Sumption adopted a narrower understanding of the rule of law. It need not require judicial review by a higher court, particularly when an inferior court or tribunal is performing a judicial function. Nor does it prevent specialist tribunals from developing legal “islands”. There was no recognition by Lord Sumption of the Constitutional Reform Act as a parliamentary endorsement of the ability of courts to define the rule of law. Privacy International, therefore, does little to resolve tensions between competing judicial conceptions of the requirements of the rule of law, though it does cast light on the relative importance of the rule of law and parliamentary sovereignty.

Third, what does the case tell us about parliamentary sovereignty? That (what might have been considered) a clearly worded ouster clause failed to exclude judicial review necessarily says something, if only obliquely, about Parliament's capacity to oust review. At the very least, it indicates that displacing the supervisory jurisdiction of the court cannot easily or casually be accomplished. But such a judicial interpretation might instead be understood more expansively, as evidence not of the linguistic difficulty of excluding review, but of the limits of Parliament's capacity to do so. Anisminic is open to being understood in either of those ways. There was certainly no explicit suggestion that Parliament, if only it had expressed itself more clearly, could not have ousted review. Yet it is arguably possible to infer that the velvet interpretive glove wielded by the House of Lords in Anisminic merely served to conceal an iron fist of irreducible constitutional principle.

Whereas such overarching questions about the extent of Parliament's capacity to displace the supervisory jurisdiction played out only implicitly in Anisminic, the Supreme Court in Privacy International tackled them directly by considering not merely whether section 67(8) had ousted judicial review, but whether, as a matter of principle, Parliament could exclude review. On this issue, the Justices differed significantly. In seeking to calibrate and evaluate the various shades of opinion, a useful framing device is supplied by Lord Sumption. He observed (at [208]) that the argument that Parliament cannot exclude review may take one of two forms. While the “radical”, or “normative”, version holds that Parliament is subject to a “higher law” as “ascertained and applied by the court”, the “less radical”, or “conceptual”, version holds that “judicial review is necessary to sustain Parliamentary sovereignty” because if independent courts do not curate, interpret and enforce legislation, Parliament is denied the capacity to legislate in any meaningful sense. The latter argument – which echoes that of Laws L.J. in R. (Cart) v Upper Tribunal [2010] EWCA Civ 859, [2011] Q.B. 120, at [34]–[38] – may be understood as one that does not limit parliamentary sovereignty, but merely acknowledges its logical implications.

None of the judgments in Privacy International unambiguously supports the radical view. Lord Sumption excoriated it: the rule of law, he said (at [209]), “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”. He went on to observe that in R. (Miller) v Secretary of State for Exiting the European Union [2017] UKSC 5, [2018] A.C. 61, at [43], the eight-Justice majority had accepted that parliamentary sovereignty was “a fundamental principle of the UK constitution”. Meanwhile, Lord Carnwath (at [119]) took it as given that the court was not addressing “the difficult constitutional issues which might arise if Parliament were to pass legislation purporting to abrogate or derogate from … accepted [rule of law] principles”.

What, then, of the “less radical” view? On this point, Lord Sumption adopted the most conservative approach. He accepted the less radical view “up to a point” (at [210]), noting that if Parliament has created a body with limited jurisdiction, it must have intended such limits to be legally effective – and, therefore, that courts should have the capacity to enforce them. However, Lord Sumption stopped short of concluding that it was impossible for Parliament to “escape this conceptual difficulty”, arguing that Parliament could, if it wished, make plain its intention to create a tribunal of unlimited jurisdiction by enacting an “all-embracing ouster clause”. Thus Lord Sumption does not really recognise a conceptual limitation on parliamentary sovereignty at all, but merely the need for especially clear language if the apparent conceptual impasse is to be circumnavigated. In contrast, Lord Wilson went further. He said (at [236]) that if (what he would consider to be) a true jurisdictional error were in play, there would be “much to be said” in favour of acknowledging the conceptual impossibility of excluding review. He was unwilling, however, to extend that thinking to review for other forms of error. In this way, it is Lord Wilson who most clearly accepts the notion of conceptual – but no other – limits on sovereignty.

The broadest approach was adopted by Lord Carnwath. He certainly endorsed (to use Lord Sumption's nomenclature) a conceptual limitation (as did Lord Lloyd-Jones, at least obliquely, albeit that he declined directly to address the in-principle question of Parliament's capacity to exclude review). But Lord Carnwath went beyond this, suggesting (at [123]) that Parliament is incapable of ousting review in respect of both excess and abuse of jurisdiction, on the ground that Parliament “cannot entrust a statutory decision-making process to a particular body, but then leave it free to disregard the essential requirements laid down by the rule of law for such a process to be effective”. Even if this does not go as far as endorsing the “radical” view, it appears to go beyond a merely conceptual form of limitation in that it treats rule of law-derived requirements upon statutory bodies as matters whose supervision Parliament cannot exclude. This arguably amounts, at least to some degree, to a form of normative, as opposed to a merely conceptual, constraint upon sovereignty. This impression is reinforced by Lord Carnwath's claim (at [131]) 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”.

It is difficult, if not impossible, to predict the impact of Privacy International. Will later cases rely on Lord Carnwath's obiter dictum, such that Parliament cannot legislate to remove judicial review over inferior courts and tribunals for abuse or excess of jurisdiction? We will probably never know. The wording required to completely remove the courts’ supervisory jurisdiction over inferior courts and tribunals would need to be so clear and precise that it would probably meet the same fate as the infamous clause 11 of the Asylum and Immigration (Treatment of Claimants etc) Bill 2003, which was radically recast in the face of overwhelming opposition (expressed in part by judges speaking extrajudicially). Perhaps the most lasting legacy will be the shift in judicial reasoning when it comes to determining what it is that renders a decision a nullity – away from categorial approaches and towards a specific focus on the rule of law, statutory presumptions of interpretation and parliamentary intention. However, we are no nearer to discovering the precise requirements of the rule of law, the relative weight to be given to parliamentary sovereignty and the rule of law, or the specific wording needed to oust judicial review, or if that is even possible. Maybe we will need to wait 50 years for the next seminal Supreme Court decision to tell us what Privacy International really meant.