Recently enacted clauses expressly include define a “determination” to include a “purported determination” or state that a court may not question a “determination or purported determination”. It is obvious what is happening. This drafting technique is a direct response to judicial decisions such as the landmark cases in Anisminic and Privacy International.  In this post, I argue that it would be a mistake to treat this new drafting technique—the insertion of the word “purported”—as excluding judicial review completely. A court faced with a challenge will have to determine whether a challenged thing is a “purported determination”. There may be cases where something might present itself to be a determination, but a court concludes that it does not answer the description of either a real or purported determination.

I argue that a reference to a “purported determination” in an ouster clause does not extend to circumstances where a claimed exercise of power was not genuine or colourable or there exists a defect or multiple defects that taken together are so fundamental that it cannot be regarded as an actual exercise of a power. In such a case, the new generation of ouster clauses might protect some nullities but will not immunise other nullities from judicial review.

This post does not venture into the controversial debate on whether Parliament is legally competent legislate to oust or abolish judicial review. I assume that Parliament can legislate (and in fact has legislated) to prevent courts from questioning “purported determinations”. My argument here focuses on how a court would go about construing those words. I will use the words “purported determination” in this post but the same arguments would apply whether the word “purported” was followed by “decision”, “certificate”, “exercise of power”, “warrant”, “order” or the like.

The evolution of ouster clauses in the first and second generations

Robert Craig in Public Law outlines how ouster clauses have changed across three “generations”. The first generation ouster clause was considered in Anisminic Ltd v Foreign Compensation Commission [1969] 2 AC 147 which concerned a determination by the Foreign Compensation Commission that was affected by an error of law. There was an ouster clause in the empowering legislation that precluded a “determination” from being “called in question in any court of law”. A majority of the House of Lords reasoned that the error made the so-called determination a nullity and it was therefore only a “purported determination”. It therefore did not fall within a “determination” within the meaning of the empowering legislation. Lord Reid freely acknowledged that the “plain words” of the ouster clause 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”.

The second generation ouster clause refers to the one considered in R (Privacy International) v Investigatory Powers Tribunal [2020] AC 491. Section 67(8) of the Regulation of Investigatory Powers Act 2000 (RIPA) stated “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”. Both Lord Carnwath and Lord Lloyd-Jones in the majority relied on the principle of legality to insist that any ouster clause must use the most clear and explicit words to effectively exclude judicial review. They considered that the clause failed to meet that standard for similar reasons to Anisminic—decisions affected by errors of law were nullities and only purported decisions. Lord Carnwath at [111] posed a question, asking “why [Parliament] did not use more explicit wording” in RIPA. He mentioned that a “more explicit formula” could be one like cl 11 of the Asylum and Immigration (Treatment of Claimants etc) Bill 2003 which excluded “challenges to any determination or ‘purported’ determination as ‘a nullity by reason of lack of jurisdiction, error of law, or any other matter’.”

While the majority speeches and judgments in Anisminic and Privacy International are controversial, they stand for three principles relevant to my arguments. First, the construction of an ouster clause is a matter for a court. Secondly, the courts demand a strict interpretation to an ouster clause. And thirdly, it is for a court to determine whether a challenged thing in a particular case falls within an ouster clause.

The new drafting technique in the third generation ouster clause

We are now in the “third generation” where ouster clauses have the “purported” vocabulary. There are at least three recent Acts where this new drafting technique has been applied.

The first is s 2 of the Judicial Review and Courts Act 2022, which inserted s 11A to the Tribunals, Courts and Enforcement Act 2007 (“TCEA”). That section seeks to reverse the judgment of the Supreme Court in R (Cart) v Upper Tribunal [2012] 1 AC 663. It provides that judicial review of a decision by the Upper Tribunal to refuse permission to appeal “is final, and not liable to be questioned or set aside in any other court” unless it involves or gives rise to one of four enumerated errors in s 11A(4). Critically, s 11A(7) states that the word “decision” “includes any purported decision” (emphasis added). The significance of including this definition was stated by Dingemans LJ (Lewis and Underhill LJJ agreeing) in R (LA (Albania)) v Upper Tribunal [2024] 1 WLR 1673 (CA) who observed at [35] that “the issue of nullity was tackled head on by the definition of “decision” in section 11A(7) to include a decision or purported decision”.

Similar language can be found in ss 13(4), 51(5), 53(5), 54(4) and 55(10) of the Illegal Migration Act 2023 (which has subsequently been repealed by the Border Security, Asylum and Immigration Act 2025). For example, s 54 of that former Act provided that, in any court proceedings relating to a decision (defined to include “any purported decision” (emphasis added)) to remove a person from the United Kingdom, the power of the court to grant an interim remedy was restricted.

Section 3 of the Dissolution and Calling of Parliament Act 2022 (“DACPA”) provides that a “court or tribunal may not question— (a) the exercise or purported exercise of the powers referred to in section 2, (b) any decision or purported decision relating to those powers, or (c) the limits or extent of those powers” (emphasis added).

Craig argues that the inclusion of the “purported” language is “game changing” because “Parliament could not have been clearer in its insistence that decisions made by the Upper Tribunal are to stand, regardless of whether they are errors of law and even if they are technically a nullity.” He argues that “[i]t is hard to avoid the conclusion that Parliament was responding directly and deliberately to the judgment given by Lord Carnwath” in Privacy International. Craig did not in his article define the scope of the word “purported”. I want to interrogate that more closely here.

Issues with the “purported” vocabulary

The use of the word “purported” in ouster clauses brings to the fore three categories: [TLD(LS1] [DL2] (1) a determination that is unaffected by any public law error (I call this a “real determination”), (2) a purported determination (where judicial review would ordinarily bite for public law error but is now ousted) and (3) a thing that is neither a real nor purported determination (and hence the ouster clause does not prevent review). The issue is where the line between categories (2) and (3) is drawn.

My first argument is that the words “purported determination” cannot be given a literal definition as referring to anything that purports to be a determination. If it did, then it would prevent a court from questioning a forgery made by a rogue or an inaccurately recorded determination brought about by a clerical error. The courts will not interpret a statute that leads to absurdity in the wide sense of that term, including results which are “inconvenient, anomalous or illogical, … or productive of a disproportionate counter-mischief”: R (PACCAR Inc) v Competition Appeal Tribunal [2023] 1 WLR 2594, [43]. It would be unthinkable that Parliament intended to require a court to accept a forgery that presented itself as—or, to use other words, purported to be—a determination. It is also extremely unlikely that Parliament intended to bar a court from dealing with an objection that the “determination” did not accurately reflect what a public authority actually decided.

These may be regarded as far-fetched examples. But courts have found it helpful to reason from extreme hypotheticals: see, eg, White v Chief Constable of South Yorkshire Police [1999] 2 AC 455; R (Jackson) v Attorney General [2006] 1 AC 262. Indeed, two of the three Law Lords in the majority in Anisminic (Lord Reid and Lord Pearce) mentioned the example of a forgery in their speeches.

These examples highlight the difficulties with giving the phrase “purported determination” a literal meaning. If, as Craig argues, Parliament’s intention was to directly respond to Lord Carnwath, then one might look to at the observation by his Lordship in Privacy International at [111] (see above). This invites one to take a narrower technical meaning: a “purported determination” refers to an act or decision that is affected by one or more ordinary public law errors and that, had the errors not occurred, would be a real determination. Taking a narrower view would be consistent with the principle that ouster clauses should be interpreted strictly.

But if a technical reading is taken, it runs into some issues. For one, it would require that there be some connection or relationship between the challenged thing and a real determination. This is especially the case for a statute like DACPA where words such as “purported exercise of the powers referred to in section 2” and a “purported decision relating to those powers” (emphasis added) draw attention to the specific powers. But how strong does that connection need to be? This would become a matter of degree that will depend on the power. It may reach a point where the connection becomes so tenuous or the defect so significant that the thing does not answer the description of a “purported determination”. To put it another way, for a thing to be a “purported determination” it must meet basic or minimum characteristics.

Consider an imaginary Warrant Act with three short sections:

      “1. A judge may issue a warrant authorising police officers to arrest and detain any person named in the       warrant.

      2. A warrant must be in the standard form, signed by the judge and affixed with the court seal.

      3. A court may not question any warrant or purported warrant relating to the powers in this Act.”

A constable shows up to your front door to arrest and detain you. They show a document that looks like a warrant. It’s in the standard form with your name on it. But it has no signature (whether from a judge, or anybody for that matter) and no seal (whether a court seal or otherwise). If a worried family member anxious about your sudden disappearance brought an application for the writ of habeas corpus, does the ouster clause preclude a court from questioning or looking behind this document?

To answer that question, a court will have to determine whether this unsigned and unsealed document falls within the meaning of a “purported warrant”. A court may well conclude that the absence of these features deprived the document of the essential characteristics to even count as a “purported warrant”. This is because the judge’s signature and court seal are basic yet fundamental features of a warrant. They evidence that the warrant has actually been issued by a judge. Having a judge independent of the executive consider warrants is a vital safeguard for individual liberty and against potential abuses of power. Without these minimum features, how can one be sure that the warrant even went before a judge? On this view, the ouster clause would not bite and the court could grant habeas relief. It would be a dark day in Britain if the opposite result was reached—that a court was powerless to stop the government depriving a person of their liberty in circumstances where it did not appear to be properly authorised.

In each case, a court will need to consider the legal limits of the power itself. Would the same result be reached if the hypothetical Warrant Act above did not include s 2 and there was no requirement that warrants be signed and sealed? A court may also need to consider how the power was exercised and who exercised it. Would the same result be reached if a judge did in fact authorise the warrant by telephone in urgent circumstances but had not signed and sealed it? Or what if it was issued, signed and sealed by someone other than a judge?

Now at first blush my argument might run against the statement by Lord Reid in Anisminic that “there are no degrees of nullity”. But what Lord Reid was saying there was that the consequence of a decision affected by jurisdictional error is a nullity and that consequence is the same whatever the error may be. In third generation ouster clauses, Parliament envisaged and accepted that a court may not be able to review or grant a remedy against a purported determination or nullity infected by ordinary public law errors. But that is not the end of the matter. Even accepting that the language of “purported determination” may pick up some nullities, there will reach a point where a determination is affected by public law errors so significant that it ceases to be a “purported determination” within the meaning of the ouster clause. To state it another way, some nullities will fall into category (2) above but others can fall into category (3). The latter may arise if the error is clear “on its face” or is “patent”. But I would suggest that it could also arise if, for example, the determination was procured by fraud or corruption. I argue that category (3) cases are those that cannot be regarded as a colourable exercise of an actual power.

Conclusion

It is undeniable that the drafters of third generation ouster clauses were motivated to avoid the issues that doomed the ouster clauses in Anisminic and Privacy International. And as Philip Murray explains in this blog the courts have held that s 11A of the TCEA has been effective in limiting judicial review of a purported determination to the enumerated grounds. But it remains a well-established principle that a clause ousting judicial review must be construed strictly.

I am not aware of any judicial consideration on the meaning and scope of the phrase “purported determination” but I argue that it would be absurd to give it a literal meaning. The better view is that Parliament intended that it include a determination affected by ordinary public law errors. But a thing needs to meet some basic or minimum characteristics before it can qualify to be a “purported determination”. What those necessary characteristics are (a) is a matter for a court to determine and (b) will vary for each power. It may be possible that some flaws or defects are so fundamental that the decision lacks the necessary qualities. And the latest generation of ouster clauses will not effectively preclude judicial review of such defects.

Dane Luo is the Farthing Scholar in Administrative Law at Pembroke College, University of Oxford.

(Suggested citation: D. Luo, ‘The Protection of “Purported Determinations” in the New Generation of Ouster Clauses’, U.K. Const. L. Blog (9th September 2026) (available at https://ukconstitutionallaw.org/))