The Limited Relevance of Anisminic
Anisminic demonstrates that the court must distinguish between those decisions of a body protected by an ouster which are nullities (in the relevant sense) and those which are not. And it set out the reasons for which the decision of the FCC being challenged was, in that case, a nullity. But Anisminic is not an authority for that second set of propositions, that is for the variety of precise techniques to be employed to identify nullities. It stands for the proposition that the court should attempt to circumvent ousters through hostile interpretation, but does not dictate that any particular interpretive technique be employed to reach that end. In other words, a court faced with an ouster is not constrained to using only the technique (or techniques) which prevailed in Anisminic itself. It follows that two tempting ways of engaging with Anisminic are mistaken. A first mistake is to simply compare the text of the ouster at hand (so RIPA s67(8)) with the text of the ouster in Anisminic (so FCA s4(4)). That involves the error of assuming that if a given clause fails it must be for the same reason that the clause in Anisminic failed. The second mistake is to apply Anisminic directly, bringing precisely (and only) the interpretive methodology to bear on s67(8) which was successfully brought to bear on s4(4). That assumes, falsely, that the precise way in which the court circumvented s4(4) is the way in which today’s court must contemplate circumventing s67(8).
The Court of Appeal in Privacy International makes both of these mistakes. Sales LJ expressly compares the two clauses , and his conclusions as to the meaning of s67(8) are grounded on (what he takes to be) material differences of language and of context between them; and he finds the different language – extra words – in s67(8) material precisely because he understands them as offering a way around one particular argument deployed in (and widely associated with) Anisminic (which below I will call the narrow jurisdictional argument.) . As it happens, Hickman has subsequently shown compellingly that the court can (and therefore should) overcome this second limb of Sales’s argument. But here, I want to press the broader point that this approach to the issue is too narrow. Even if it is possible to arrive at the right outcome this way, it is a mistake to go (at all, and certainly to give priority to) that way in the first place. This is for two basic reasons; one internal to and one external to Anisminic.
The “internal” reason” is that Anisminic was not in any event decided on one discretely identifiable ground. Lord Reid pursued three separate lines of thought. First, that numerous flaws can lead to nullity, some of which he set out in an (expressly) non-exhaustive list (bad faith, excess of power, good faith misconstruction of a statute, failure to take account of relevant considerations, reliance on irrelevant considerations, p171). In particular, and secondly, he focussed on a particular form one entry on his list might take; he found that one way in which a good faith misconstruction of statute could result in nullity was if it had the consequence of leading the tribunal to “deal with and base their decision on a matter which, on a true construction of their powers, they had no right to deal” (p174). Thirdly, he ended by characterising the flaw in the FCC’s decision as its reliance on irrelevant considerations (he found fault in their reliance “on a ground which they had no right to take into account” p175).
Lord Pearce followed a similar format. First, he proposed a – different – list (again explicitly non-exhaustive) of ways purported decisions could turn out to be nullities (absence of formalities condition precedent to inquiries, breach of natural justice, asking the wrong questions, taking into account impermissible considerations, p195), and then settled on one flaw which he identified in the FCC’s decision, that it had misconstrued the Order empowering it, and so exceeded its mandate (remarkably, this type of flaw hadn’t even featured on his opening indicative list).
These judgments are very difficult to parse, but some common ground can be distilled. Most famously, both rely on the idea that nullities arise when tribunals exceed their jurisdiction. But this convergence is misleading because despite the common language they explicitly mean different things by it. Lord Reid uses it narrowly, to refer to a specific source of nullity – namely his second line of thought (this is what I will from now on the narrow jurisdictional argument). Lord Pearce uses the label more generally, for anything on his list of flaws leading to nullity. So this famous common ground, the idea of excess jurisdiction, is illusory. But there are other commonalities. For example, that taking into account irrelevant considerations is a ground for nullity. We cannot derive majority support for prioritising any particular line of thought in this mess by examining the extent to which Lord Reid and Pearce attracted the agreement of other judges hearing the case.
In short: no particular test for nullity is used in Anisminic. It is not therefore appropriate to treat it as authority for any particular test for nullity. Yet the Court of Appeal in Privacy International privileges the narrow jurisdictional argument in a way which is inconsistent with the fragile significance of that argument in Anisminic itself.
The reason which is “external” to Anisminic is that – even if it had been neatly organised around and lent its authority to one particular test for nullity – it was not, and at no point in any of the judgments in the case was it even purported to be – an exhaustive statement of the variety of ways in which decisions (otherwise) apparently protected by an ouster clause might be characterised as nullities by the court. In fact, a majority in Anisminic – Lord Reid and Pearce explicitly and Lord Pearson by agreement with them – clarified that their lists of sources of nullity were non-exhaustive.
The upshot of all this is that the material common ground in Anisminic is limited to the general common ground: the common law rule requiring the courts to treat ouster clauses with interpretive hostility by denying their protection to nullities. And, consistently with de Smith’s catalogue a decade earlier, the judgments in Anisminic itself advert to the difficulty of synthesizing the various limbs of that pre-existing tradition. Anisminic was not, and should not be treated, as a resolution to the various dimensions of that complex tradition. Rather, it was an addition to that complex picture.
The binding rule which we can ascertain from Anisminic, the only proposition which attracted majority support in the case, and the legal position which the case (selectively) exemplified rather than (exhaustively) establishes – is something like this: the common law demands interpretive hostility to ouster clauses and there are many ways, some but not all of which outlined in the non-exhaustive lists proposed in the judgments, for the court to express that hostility.
Back to Privacy International
This focus on Anisminic needs reorienting if it is to help us to properly understand Privacy International. The question is this: has Parliament, in passing s67(8) summarily won the game of cat and mouse it has hardly bothered to play and alighted on the magic formula capable of protecting the IPT’s misconstruction of this statute (and presumably others) in this way?
The differences between s4(4) and s67 (8) do mean that the court in Privacy International cannot simply transpose the reasoning used in Anisminic. To this extent, Sales LJ is correct to examine whether the additional language in s67(8) allows it to escape the narrow jurisdictional argument which was successful against s4(4). But this is only one small slither of what the tradition in Anisminic requires. Anisminic provides equivalent authority for the argument that the IPT’s decision was a nullity because its misconstruction of ISA s5 operated as an irrelevant or impermissible consideration. And it establishes lengthy non-exhaustive lists of other forms of nullity. And it is part of a tradition under which there are yet other possibilities which the court must take into account. The conclusion that 67(8) protects the IPT’s decision under the narrow jurisdictional argument does not preclude that decision from being a nullity on any (or indeed many) of the other possible flaws which courts can identify in order to remove decision-making from the protection of an ouster clause.
So, Anisminic tells us that something more than the text of s4(4) is needed in order for a clause to successfully overcome this formidable array of judicial obstacles to ousters. But the Court of Appeal in Privacy International dramatically fails to properly address the question of whether that something more is present in s67(8). There are two ways of envisaging what that something more, the magic formula, might look like.
The first (the “wholesale” approach) would be a clause explicitly drafted to mount a broad assault not on the decision in Anisminic itself, but on the deeper rule underlying its whole tradition – that is, an ouster clause which expressly protected not just determinations (decisions… etc) but purported determinations (decisions…) etc. as well. This might be possible (in principle) with fairly concise language.
The second (the “retail” approach) would be a clause explicitly drafted to mount a series of focussed assaults on each potential ground of nullity which features in Anisminic itself (that is, each of the entries on Lord Reid’s and Lord Pearce’s lists of sources of nullity) and all the grounds yielded by the broader tradition in which Anisminic sits (that is, anything else in de Smith’s catalogue, principles derived from cases discussed but not relied on in Anisminic etc) and so on. Such a clause would be long and tangled.
Under the common law rule of interpretive hostility to ousters, these two approaches are the only ways in which Parliament could effectively legislate a bar to judicial review. And either would likely lead to a constitutional crisis. The wholesale approach has never been tried, and hopefully never will be – its consequence would be creation of a body with an unreviewable power to decide what did, and what did not, count as its own jurisdiction. The Blair government were tempted by the retail approach in (an absolutely remarkable provision) clause 108A of the Bill that became the Nationality, Immigration and Asylum Act 2002. But they withdrew it in the face of pressure including indications from senior judges that circumventing the tradition in that way would force them to consider straightforwardly overriding the super-ouster. Any attempt to pursue either of these ways around Anisminic ought to, and probably would, provoke a constitutional crisis which would outweigh the legal controversy about the clause’s effectiveness.
In any event, it is clear that s67(8) does not really amount to a genuine legislative attempt to outflank Anisminic. It does add something to s4(4) – the inclusion of “awards” in addition to decisions, parenthetical precision that decision includes determination, and a reference to the distinction between review and appeal – but this comes nowhere near an attack on the full force of the Anisminic tradition. It does not protect merely purported determinations or preclude them from being nullities. So it does not adopt the wholesale approach. Its extra material compared to s4(4) might (again, subject to Hickman’s objections) shield the IPT from review based on the narrow jurisdictional jurisdiction argument, and hence it might be seen as a partial adoption of the retail approach; but the effort is so partial that it must be ineffective – the single strand of argument it does address is (i) not the whole of the rule in Anisminic in any event; and (ii) not the only (or even the main, or even in Privacy International the presumptively relevant) tool the court has at its disposal to circumvent the effect of the ouster.
The Supreme Court has many more tools at its disposal in interpreting RIPA s67(8) than the lower courts were willing to use.
Two possibilities are beyond the ambition of these posts. There are clear indications in the non-exhaustive nature and in the content Lord Reid’s and in Lord Pearce’s lists of potential grounds for nullity that the entirety of the law of judicial review is in play – that is, that any flaw which would under the general law be a ground for judicial review is equally a ground for nullity and hence disabling of an ouster clause. And, more modestly, there is influential authority (e.g. Page  A.C. 682) and coherent analysis (e.g. Paul Scott) proposing that the narrow jurisdictional argument should be abandoned in favour of the conclusion that Anisminic entails simply that all errors of law are grounds for nullity. Either approach would bring coherence to the law and permit the Supreme Court to find in favour of Privacy International.
I have argued here that even if the court chooses not to commit to these broad solutions to the problem, the authority of Anisminic must not be reduced to the narrow jurisdictional argument. Lords Reid and Pearce also agreed that reliance on an error of law can constitute reliance on an irrelevant or impermissible consideration. So the Supreme Court ought to consider whether a misconstruction of ISA 1994 would lead the IPT to rely on an irrelevant or impermissible consideration. Lord Reid contemplated that a good faith misconstruction of a statute can by itself be a ground for nullity. The Supreme Court ought to consider whether a misconstruction of ISA 1994 would be a misconstruction of that kind. Lord Pearce said that errors of law can lead a tribunal to ask itself the wrong question. The Supreme Court ought to consider whether a misconstruction of ISA 1994 would lead the IPT astray in that sense. And so on.
It is a requirement of the common law that these approaches be explored. The Anisminic tradition embodies a common law interpretive presumption of fundamental importance. It obliges the court to use at least one (any plausible one would do; although recall that the court in Anisminic, almost indiscriminately, used several) of those tools in constructing s67(8). This obligation is not neutered by the marginal additional text in s67(8) as compared with s4(4). It is an obligation which the decisions of the Divisional Court and the Court of Appeal in Privacy International violated with potentially serious constitutional consequences. A narrow approach to what counts as a nullity for the purposes of interpreting the scope of ouster clauses would eviscerate an important common law tradition protecting the rule of law. The Supreme Court has a choice of ways to sustain that tradition in Privacy International. It is of fundamental constitutional importance that it do so.
Adam Tucker, Senior Lecturer and Director of Research at Liverpool Law School
(Suggested citation: A. Tucker, ‘Parliamentary Intention, Anisminic, and the Privacy International Case (Part Two)’, U.K. Const. L. Blog (19th Dec. 2018) (available at https://ukconstitutionallaw.org/))