A. The Nature and Importance of the Classical Account

In H.M. Treasury v Ahmed (No. 2),Footnote ⁴ Lord Philips, giving the majority judgment, provided a textbook account of the classical view. Having held in a precursor case that secondary legislation under which suspected terrorists’ assets had been frozen was ultra vires,Footnote ⁵ the Supreme Court was asked by the Treasury in Ahmed (No. 2) to defer quashing the relevant legislation for several weeks. During that time, it was intended that banks would continue to treat the relevant assets as frozen, thereby preventing their withdrawal pending the enactment of fresh asset-freezing legislation. Declining the Treasury’s request, Lord Philips explained that unlawful administrative acts are “of no effect in law” and that “[t]he object of quashing them” is simply “to make it quite plain that this is the case”.Footnote ⁶ The effect of “suspending the operation” of a quashing order could (as the Treasury intended) “give the opposite impression” because it “would suggest that, during the period of suspension […], the provisions to be quashed would remain in force”.Footnote ⁷ The court, said Lord Philips, “should not lend itself” to “obfusact[ion]” of its own judgment.Footnote ⁸ Lord Hope, the sole dissentient, agreed that the relevant secondary legislation was “‘ultra vires and void from the moment [it was] made”,Footnote ⁹ disagreeing only as to the appropriateness of suspension. He recognised, the “risk” that suspension “would tend to obfuscate the effect of the court’s judgment” but considered that risk acceptable given the gravity of the likely consequences if the assets were unfrozen before new legislation was enacted.Footnote ¹⁰

The court thus unanimously endorsed the classical account, according to which breach of any of the grounds of judicial review renders an administrative act void ab initio, making judicial review’s function primarily declaratory, not constitutive, of the legal status of impugned administrative acts. The alternative view, which, as we explain in subsequent parts of this article, is or appears to be endorsed by recent judicial, statutory and academic interventions, would be to see unlawful administrative acts as “voidable”: that is, as valid until a court’s remedy was obtained actively to make them invalid. In contrast, on the classical account, remedies such as quashing orders simply illuminate the underlying legal position: they do not change it. It follows that the effect of suspending a quashing order can only be practical, not legal. And while Ahmed (No. 2) is a particularly clear articulation of the classical account, it is no outlier. For instance, in Secretary of State for the Home Department v JJ, Lord Bingham (for the majority) held that it would be logically impossible to require a Minister to modify ultra vires administrative orders,Footnote ¹¹ endorsing the view of Richards J. in the Administrative Court that “since he had no power to make them in the first place”, there were no extant orders capable of modification.Footnote ¹²

The importance of the classical account can be illustrated by reference to the House of Lords’ landmark judgment in Boddington v British Transport Police.Footnote ¹³ The defendant, a passenger in a train carriage with prominent “no smoking” signs, smoked a cigarette, refusing to put it out when asked to do so. The signs reflected a complete smoking ban instituted by the train operator using a railway byelaw, made under section 67 of the Transport Act 1972, that made it an offence to smoke where “no smoking” signs were conspicuously displayed. The defendant admitted smoking but argued that the Act did not authorise a complete ban, meaning that the decision to post signs in every carriage was ultra vires. It followed, said the defendant, that he had committed no offence because he had not smoked in a carriage in which a valid sign was displayed.

In the end, the House of Lords held that the Act did permit a complete smoking ban. But in reaching that conclusion, it emphatically affirmed the general principle that unlawful administrative acts can be challenged not only directly in judicial review proceedings but also collaterally in other proceedings – including in courts, such as the magistrates’ court in which the defendant in Boddington found himself, that have no power to quash such acts. That would be impossible if unlawful acts were merely voidable: they would stand until actively quashed in judicial review. Collateral challenge thus depends on the voidness of unlawful acts. Once the court acknowledges the act’s unlawfulness, it can (and must) proceed on the basis that the act never legally existed in the first place, thereby facilitating the type of argument that the defendant in Boddington sought to make and securing the availability of collateral challenge. We explain in Section IV(B) below that those who favour the alternative view, according to which unlawful acts are not void ab initio, either fail entirely to engage with the implications for collateral challenge or do so in ways that fail to withstand scrutiny.

The close connection between classical concepts and collateral challenge can be traced far back into the history of administrative law. Indeed, so close was the connection between the doctrine of nullity and collateral challenge that the availability of the latter originally prevented judicial review. When certiorari, the precursor to today’s quashing order, first emerged as a remedy to quash unlawful administrative acts, it was only available when those acts were voidable.Footnote ¹⁴ Quashing in the event of voidness was thought unnecessary because anyone affected by a void act could simply refuse to obey it, await its enforcement and raise its invalidity collaterally, such as by suing the enforcement officer for false imprisonment or some other tort. Collateral challenge lay at the heart of the common law’s vindication of the rule of law; direct forms of judicial review served a merely supportive role. When certiorari eventually became available for the “quashing” of null and void acts,Footnote ¹⁵ this was for mere practical purposes (in a world of paper records, it was hard to enforce an administrative act when its record had been erased) and for declaratory effect.

Although “direct” judicial review has now supplanted collateral challenge as the primary means of vindicating the rule of law, that does not mean collateral challenge has become redundant, as the factual matrix of Boddington illustrates. When, in that case, it endorsed the general principle that unlawful administrative acts are susceptible to collateral challenge, the House of Lords rejected the view adopted in Bugg v Director of Public Prosecutions which allowed only for administrative acts marred by “patent” errors of law to be challenged collaterally.Footnote ¹⁶ Lord Irvine L.C. insisted that today’s administrative law brooks no distinction between categories of illegality.Footnote ¹⁷ All unlawful administrative acts are ultra vires and void, such that the courts approach statutes criminalising disobedience of administrative acts with a strong presumption that such acts’ validity can be raised collaterally – which, said Lord Irvine, is “important for the maintenance of the rule of law and the preservation of liberty”.Footnote ¹⁸ Likewise, Lord Steyn dismissed the idea that direct judicial review could always offer a sufficient alternative: the significant procedural limitations on review, its discretionary nature and its expense all meant that judicial review would not compensate a defendant who was denied collateral challenge as a defence.Footnote ¹⁹

Even the lightest perusal of recent case law in this area shows how costly dispensing with classical administrative law, and so the general availability of collateral challenge, could be. The master of a vessel convicted of keeping an unregistered boat on a waterway would not be able to argue the registration fee imposed by the Environment Agency was unlawfully excessive.Footnote ²⁰ A taxpayer fined by the revenue would not be able to question the lawful basis of the fine in any appeal.Footnote ²¹ Foreign nationals detained by the Home Secretary pursuant to an unlawful detention policy would not be able to claim damages for false imprisonment.Footnote ²² The Crown Court would be denied the useful power of constituting itself as a magistrates’ court to correct charges which, on the face of a committal, are erroneous in law.Footnote ²³ While the importance of legal certainty and good administration might sometimes militate against the availability of collateral challenge and while the rule of law and basic notions of administrative fairness might occasionally demand the courts show a degree of flexibility when interpreting the logical effects of voidness,Footnote ²⁴ it is undeniable that, in any society governed by the rule of law, a citizen should be able to raise the unlawfulness of an administrative act that has onerous consequences for him or her in whatever context it arises. This state of affairs, to which the availability of collateral challenge is imperative, has been a feature of the common law for some four centuries and abandoning it now would, as Lord Steyn put it in Boddington, be “too austere and indeed too authoritarian to be compatible with the traditions of the common law”.Footnote ²⁵ In this way, an apparently technical question about the status of unlawful administrative acts becomes an issue of considerable constitutional importance. It is against that background that we examine, in Sections III and IV below, recent challenges to the classical account and argue in favour of its retention.

B. The Classical Account and the Wider Picture

Before we do so, we should say something about the nature of our project here. It is, to a large extent, one of doctrinal legal scholarship: the type of study identified by Paul Daly and Joe Tomlinson as the mainstay of administrative law,Footnote ²⁶ deserving to be taken seriously on its own terms.Footnote ²⁷ It is our principal purpose to show that the classical account offers the most coherent explanation of the statutes and case law which constitute the sources of administrative law and that apparent challenges to the classical account are in truth reconcilable with it. We therefore do not seek to intervene directly in more radical debates on the underlying values or purposes of administrative law. We do not deny that such debates are important. But whether one thinks the purpose of administrative law is the regulation of public power in the public interest,Footnote ²⁸ the holding of administrators to their moral duties as delegates of Parliament and the Crown,Footnote ²⁹ the protection of individual liberty,Footnote ³⁰ the advancement of justice and good governmentFootnote ³¹ or a multiplicity of diverse and sometimes competing concerns,Footnote ³² such underlying values or purposes must ultimately be channelled through doctrinal apparatus.Footnote ³³ And while legal doctrine can never be wholly disconnected from the law’s underlying values and purposes – “values provide the motor for administrative law”Footnote ³⁴ – doctrine nonetheless has its own inner life and logic, guiding both administrative conduct and judicial determinations. As such, more expressly doctrinal scholarship still has its role to play, explaining, rationalising, guiding and limiting administrative law practice.Footnote ³⁵

That being said, there are three ways in which our article cannot but speak to some of the “deeper” debates in contemporary public law scholarship. First, we do not see classical administrative law and its concern for enforcing the legal limits on decision makers’ powers as one possible “meta-value” in competition with other “meta-values”. Some scholars have presented classical administrative law in such terms.Footnote ³⁶ But we, like Trevor Allan amongst others, see the classical concepts as the doctrinal tools by which the underlying value or values of administrative law are rendered juridically effective.Footnote ³⁷ We reject the idea that one must choose, for example, between classical administrative law and a belief that judicial review is ultimately about securing the public interest, or promoting individual liberty.

Second, and relatedly, in recent years the possibility of unified administrative law has been questioned. This has come about in part from what Jason Varuhas has termed the “empiricist turn” in recent scholarship.Footnote ³⁸ Joanna Bell, for example, has queried the possibility of a general or “monist” administrative law that emphasises a common doctrinal toolbox across a diversity of administrative contexts.Footnote ³⁹ Bell makes the important point that in administrative law, context is everything. The grounds of review must always be mediated through the particular features of individual statutes and administrative schemes. At the same time, we accept our project here is what might be described as “monist” in nature. We proceed on the assumption that the classical account has universal applicability across a wide variety of administrative contexts. This is because we believe that such an approach best reflects the courts’ own practice. While first-instance adjudication will necessarily be focused on the legal minutiae of particular administrative contexts,Footnote ⁴⁰ appellate-level judicial review, which provides the doctrinal cartography within which first-instance decisions are situated, much more readily reaches across a variety of administrative sectors.Footnote ⁴¹ We see this in the variety of contexts addressed by the cases examined in this article: immigration orders; railway byelaws; terrorist asset-freezing orders; devolved legislation. In these cases we see the courts developing and applying a common set of principles in a monist fashion.Footnote ⁴² While we recognise, with Bell, that the plausibility of any monist account of law must be rooted in “particular cases and legislative provisions”,Footnote ⁴³ our purpose here is to do exactly that, exploring recent cases and statutes to show how classical administrative law accommodates them.

Third, nothing in this article should suggest we are agnostic about the values or purposes underpinning administrative law. We do not seek to show that recent judgments and statutes are reconcilable with the classical administrative law “system” for its own sake, but to show that the “system is a rational one, built on good foundations”.Footnote ⁴⁴ Our commitment to the classical account is born from a clear conviction that one of the principal purposes of judicial review is to ensure that public administration is carried out in accordance with the rule of law. Its basic requirement that public bodies’ acts must be grounded in lawFootnote ⁴⁵ – and must therefore respect the full range of administrative law principles,Footnote ⁴⁶ Parliament being presumed to have conferred legislative powers on the basis that they are exercised in accordance with those principlesFootnote ⁴⁷ – provides, we think, the best justification for classical theory.Footnote ⁴⁸ For as argued above, the classical account offers the most secure basis for the collateral challenge of unlawful administrative acts, which is something we consider a necessary aspect of any legal system which seeks fully to adhere to the rule of law. Regardless of whether the rule of law is seen as the sole or primary value underpinning administrative law, we expect all public lawyers will recognise it as a fundamental aspect of any mature constitutional system.Footnote ⁴⁹ This provides sufficient consensus, we believe, for our doctrinal analysis to be approached on its own terms.

A. Suspended and Prospective Quashing Orders

We begin that analysis with section 29A of the Senior Courts Act 1981 (“SCA”).Footnote ⁵⁰ Inserted into the SCA by section 1 of the Judicial Review and Courts Act 2022 (“JRCA”) following the recommendation of the Government-commissioned Independent Review of Administrative Law,Footnote ⁵¹ section 29A equips the courts with much greater flexibility than they had been prepared to exercise in Ahmed (No. 2),Footnote ⁵² the Supreme Court’s judgment having been singled out by the relevant Minister in the second reading debate on the JRCA as evidence of the need for greater remedial latitude.Footnote ⁵³ Thus section 29A(1) now provides that a quashing order may include provision (1) for the quashing “not to take effect until a date specified in the order” or (2) “removing or limiting any retrospective effect of the quashing”. Those references to the deferral of the date on which a quashing order “take[s] effect”Footnote ⁵⁴ and to removing or limiting the “retrospective effect of the quashing”Footnote ⁵⁵ seem to imply that quashing orders have legal effects in the first place – a view that does not align with that espoused by the Supreme Court in Ahmed (No. 2). Similarly in tension with the classical view are dicta such as those of Dame Victoria Sharp P. and Chamberlain J., according to whom “the effect of a quashing order, unless made under the new provisions in section 29A of the 1981 Act […], is […] to treat the quashed decision as void ab initio”.Footnote ⁵⁶

All of this might seem to suggest that Parliament and the courts are proceeding on a premise – that quashing orders do things and that, until those things are done, unlawful administrative acts are legally valid – that is fundamentally at odds with the classical account. However, far from casting doubt on that account, we contend that section 29A affirms it. Our argument rests primarily on detailed analysis of section 29A itself, but it is helpful to begin with one of the few cases so far in which the implications of that provision have received any meaningful consideration. R. (ECPAT UK (Every Child Protected against Trafficking)) v Kent County Council Footnote ⁵⁷ concerned two protocols relating to the treatment of unaccompanied asylum-seeking children. In earlier proceedings, those protocols had been quashed, but the taking effect of the quashing orders was deferred under section 29A(1) for several weeks. In the instant case, the court agreed to extend the deferral period in respect of the “NTS Protocol” but not the “Kent Protocol”. Against that background, and bearing in mind that “there is no power to grant a suspended declaration”, Chamberlain J. explained that the court “would be saying two inconsistent things” about the NTS Protocol if it granted a declaration prior to the taking effect of the quashing order at end of the (extended) deferral period.Footnote ⁵⁸ In contrast, “it was appropriate to grant a declaration” in relation to Kent Protocol once it had been quashed, “making clear that Kent [County Council] had acted and was acting unlawfully, in breach of its [relevant statutory] duties”.Footnote ⁵⁹

Through his use of language (“had acted”), Chamberlain J. makes it clear, contrary to the view he appeared to express in R. (Director of Public Prosecutions) v Bristol Crown Court,Footnote ⁶⁰ that the deferred quashing of an administrative act does not detract from the fact that it was always unlawful – even though, on his analysis, it is only appropriate for a court to declare as much once the quashing order has taken effect. Such perceived inappropriateness presumably arises because to declare the underlying position before the taking effect of the quashing order would undermine the practical utility of suspension, in that it would compromise the obfuscatory effect that is now lent statutory respectability by section 29A. This view, according to which section 29A does not undermine the default position according to which unlawful administrative acts are void ab initio, is affirmed when we closely consider the structure of the provision.

Section 29A(3) provides that the consequence of deferring a quashing order is that the relevant administrative act is “upheld” until the end of the deferral period, while section 29A(4) makes equivalent provision in respect of quashing orders whose retrospective effect is removed or limited. What, then, does it mean for an unlawful administrative act to be “upheld”? Section 29A(5) provides that: “Where (and to the extent that) an impugned act is upheld by virtue of subsection (3) or (4), it is to be treated for all purposes as if its validity and force were, and always had been, unimpaired by the relevant defect.” This raises three points. First, statutorily “upholding” an unlawful administrative is a far-reaching notion: it applies for “all purposes”. Among other things, that presumably means that third parties (such as the banks in Ahmed (No. 2)) and other courts (including those before which the administrative act might otherwise have been collaterally challenged) must regard the act as valid. Second, section 29A(5) makes it implicitly clear that it is unlawfulness itself, as distinct from any remedy that might be issued in consequence of such unlawfulness, that goes to validity, but that the implications of unlawfulness are statutorily modified when a quashing order makes provision in accordance with section 29A(1). Third, the nature of that modification concerns not the legal status of the unlawful administrative act but the manner in which it is to be regarded while the quashing order remains suspended: the act is to be treated as if it were valid; it is not, by virtue of the suspension of the quashing order, rendered valid.Footnote ⁶¹ (The direction in section 29A(5) to treat the impugned act as if its validity and force were unimpaired by the relevant defect would be redundant if the default position was not that unlawfulness does impair validity and force.)

That analysis is reinforced when we turn to section 29A(6). It says that: “Provision under subsection (1)(a)” – which enables courts to provide for deferral when issuing quashing orders – “does not limit any retrospective effect of a quashing order once the quashing takes effect (including in relation to the period between the making of the order and the taking effect of the quashing); and subsections (3) and (5)” – subsection (5) being the provision that sets out what it means to “uphold” an administrative act when quashing is deferred – “are to be read accordingly”. If the effect of deferral was more than merely obfuscatory – if, for instance, it breathed legal life into unlawful administrative acts, or if, more generally, we were to take the section 29A regime as casting doubt on the classical account of administrative law – then section 29A(6) would produce bizarre consequences. Take, for instance, an unlawful administrative act that is subject to a quashing order issued on 1 October but whose effect is deferred until 1 November. Unless the effect of deferral is purely obfuscatory, section 29A(6) implies that until 31 October, the relevant act is valid but that from 1 November it is and always has been invalid. That would entail that for the month of October, the unlawful act was both valid and invalid depending on the temporal vantage point from which the question of validity is posed.

This Schrödinger’s cat-style paradox is readily avoided through adherence to the classical account of administrative law – provided that we properly understand and deploy the conceptual tools it puts at our disposal. One of those tools is the principle of legal relativity, which is, in effect, what Lord Radcliffe was describing when he famously said that:

Lady Arden has taken this to mean that unlawful acts are “valid and binding until […] set aside”.Footnote ⁶³ But that is not consistent with what Lord Radcliffe said. Lady Arden appears to suppose that the setting aside, or quashing, of unlawful administrative acts is transformative of their legal status: that such acts, in other words, are voidable, not void. But that view cannot be reconciled with the clear authority of Ahmed (No. 2) and, as we explained above, would, if accepted, remove all opportunity for collateral challenge. Lord Radcliffe’s insight was not that unlawful administrative acts are merely voidable, but that unless or until it is challenged in an appropriate forum via an available and suitable procedure, an unlawful administrative act might be erroneously thought to be lawful and therefore valid – and that in circumstances in which no such challenge occurs or can occur, there may be little practical choice but to treat the act as if it were valid. Phenomena that lack validity can always appear to exist in fact, but that is a separate matter from the question of their legal validity.Footnote ⁶⁴ Thus arises the notion of legal relativity, as Sir William Wade termed it,Footnote ⁶⁵ whereby the extent to which an unlawful administrative act is treated as if it was lawful depends on whether it has been possible successfully to challenge it and obtain a suitable remedy. It follows that, contrary to the view of some writers,Footnote ⁶⁶ there is no tension between the existence of remedial discretion and the principle that unlawful administrative acts are void ab initio. Once the principle of legal relativity is applied, we can see that a court’s refusal to quash an unlawful act does not transform it into a legally valid act, but might have the effect of rendering it functionally valid, in the sense that some or all actors will, absent quashing, proceed as if the act was lawful.

The section 29A regime can also be understood through the prism of legal relativity, but with an important twist. In the circumstances contemplated by Lord Radcliffe, the de facto effectiveness of the unlawful but unchallenged (or unsuccessfully challenged) administrative act is (1) attributable to happenstance (the expiry of the time limit for judicial review; the absence of a claimant with standing and so on) and (2) potentially incomplete (for instance, if someone later comes along and succeeds in collaterally challenging the unlawful act, then, at least from their perspective, it will be invalid, even if, for the perspective of some others, it continues to be regarded as valid). In contrast, the effect of section 29A is (1) by design, not happenstance and is (2) at least temporarily absolute (not incomplete), in that during (say) the deferral period, others must proceed as if the invalid act is valid. Thus, to the extent required by the terms of a quashing order issued under section 29A(1), the possibility of acknowledging and proceeding on the basis of the unlawfulness of the relevant act is shut off entirely, such that, under the principle of legal relativity, the administrative act’s theoretical voidness is completely eclipsed by its functional validity.Footnote ⁶⁷ But once, in the case of section 29A(1)(a), the deferral period has expired, the true, default position, according to the classical account, is restored. The relevant administrative act can then be seen for what it is and always was: unlawful, void ab initio and, importantly, vulnerable to collateral challenge, including, thanks to section 29A(6), in relation to matters that occurred during the deferral period. Far from challenging or departing from the classical view, section 29A is in fact premised on it and can be explained through it, using the principle of legal relativity.

B. Suspension and Prospectivity of Judicial Decisions Concerning Legislative Competence

A second apparent challenge to the classical view comes in the form of section 102 of the Scotland Act 1998 (and equivalent provisions of the other devolution statutes).Footnote ⁶⁸ That section applies, inter alia, “where any court or tribunal decides that an Act of the Scottish Parliament or any provision of such an Act is not within the legislative competence of the Parliament”.Footnote ⁶⁹ In such circumstances, the court or tribunal, according to section 102(2), “may make an order (a) removing or limiting any retrospective effect of the decision, or (b) suspending the effect of the decision for any period and on any conditions to allow the defect to be corrected”.

One way of attempting to make sense of section 102 would be to take at face value its arguably implicit premise that when a court decides that an Act of the Scottish Parliament is not within competence, that decision has effects, including retrospective effects, that are capable of being removed, limited or suspended via a section 102 order. That would imply that the legal status of unlawful measures is dependent upon judicial intervention – meaning, for instance, that the validity of such measures prior to judgment can be preserved by withholding retrospective invalidation. And that, in turn, would imply either general repudiation of the classical account of administrative law, according to which unlawful measures are void ab initio such that their legal invalidity arises independently of any judicial intervention, or at least repudiation of it in relation to provisions of devolved legislation that are outwith competence.

However, making sense of section 102 in that way is simply not possible because it is unambiguously clear that devolved legislation enacted outwith the limits on competence set out in the Scotland Act is void ab initio. Section 29(1) provides that: “An Act of the Scottish Parliament is not law so far as any provision of the Act is outside the legislative competence of the Parliament.”Footnote ⁷⁰ And, interpreting that provision in the Continuity Bill case, the Supreme Court concluded that it signifies that the “result of a want of legislative competence is that a Scottish enactment is a nullity”.Footnote ⁷¹ It cannot therefore be argued that section 102 should be taken to mean that outwith-competence enactments are not nullities, such that their legal status depends on whether judicial relief is granted. This would introduce a fundamental conceptual contradiction into the scheme of the Scotland Act.

How else, then, might we make sense of section 102? Once we accept, as we must, that unlawful enactments are nullities that are dead on arrival, one possibility would be to conceive of section 102 orders as capable of breathing legal life into such enactments by retrospectively investing them with temporary validity between the date of their enactment and the date of the judgment and/or between the judgment and the expiry of any period of remedial suspension. We do not doubt that Parliament, when enacting section 102, could have done that: being sovereign, it must be capable of conjuring into existence a system of not only legal life support but a judicial power to convert that which was legally stillborn into a temporarily valid enactment. However, concluding that section 102 creates such miraculous judicial powers would involve reading a great deal into it and there is little, if anything, in the language that suggests it would be appropriate to do so.

A more mundane – but more defensible – reading of section 102 involves accepting that Parliament was doing no more than acknowledging that there is no contradiction between (1) the existence of judicial discretion over remedies and (2) the classical principle according to which unlawful administrative acts (and, here, enactments) are void ab initio. We have already argued that there is no tension between those two phenomena. The scheme of the Scotland Act simply serves to reinforce that point. The courts’ capacity under section 102 to remove or limit the retrospective effect of their decisions in respect of unlawful devolved enactments, or to suspend their effect, is therefore best understood as an acknowledgement that the granting and withholding of remedies in this context is capable of producing practical effects, notwithstanding that the underlying legal position – that the unlawful enactment is a nullity – remains unaffected.

Such practical effects were described pejoratively by the Supreme Court in Ahmed (No. 2) in terms of “obfuscation”, not least because the court was concerned that rule-of-law values might be compromised by courts giving an inaccurate impression of the legal position. However, as Lord Hope’s dissent in that case demonstrates, there is scope for different views about the extent to which such constitutional concerns might be outweighed by competing considerations. By enacting section 102, Parliament – no doubt recognising particular scope for difficulties to arise when devolved primary legislation is found to be unlawful and the corresponding need for a mitigation mechanism – has licensed that which the Ahmed (No. 2) majority considered improper. It is, however, plain from the scheme of the Scotland Act that all of this is achieved without departure from either the general principle that unlawful measures are void ab initio or the specific principle, laid down in the Scotland Act itself, that unlawful enactments are nullities.

C. A New Approach to Ouster Clauses?

The courts’ approach to ouster clauses provides a further context in which to consider the continued vitality of classical administrative law. Since at least the middle of the eighteenth century, courts have interpreted ouster clauses with a presumption that by them, Parliament does not intend to immunise wholly void administrative acts from review.Footnote ⁷² Such a view has traditionally been rooted in the courts’ genuine desire to give effect to parliamentary intention. As Farwell L.J. said in R. v Shoreditch Assessment Committee, ex parte Morgan, “it is a contradiction in terms to create a tribunal with limited jurisdiction and unlimited power to determine such limit at its own will and pleasure”.Footnote ⁷³ And the courts are slow to attribute to Parliament such contradictory intentions. The way the courts have traditionally respected Parliament’s dual intentions both to restrict judicial review and to set enforceable limits on a decision maker’s powers has been to distinguish between errors going to jurisdiction, which are reviewable notwithstanding any ouster clause, and errors made within jurisdiction, which an ouster clause will protect from review.Footnote ⁷⁴ As such, classical administrative law concepts have historically been crucial to the interpretation of ouster clauses.Footnote ⁷⁵ And while the landmark case of Anisminic Ltd. v Foreign Compensation Commission Footnote ⁷⁶ (at least as interpreted in R. v Lord President of the Privy Council, ex parte Page)Footnote ⁷⁷ may have been consequential in extending the category of errors of law that go to the jurisdiction of a decision maker and are thus not protected by an ouster clause, its bold assertion that ouster clauses do not protect nullitiesFootnote ⁷⁸ was, as David Feldman has noted, hardly novel.Footnote ⁷⁹

More recently, however, the courts have favoured a new approach to ouster clauses that does not depend on classical doctrine. This can be traced to the decision in R. (Cart) v Upper Tribunal,Footnote ⁸⁰ which was not as much concerned with ouster clauses as with implied exclusions of review. The question in Cart was whether a decision of the Upper Tribunal (“UT”) not to permit an appeal from the First-tier Tribunal could be subject to judicial review. The Tribunals, Courts and Enforcement Act 2007 (“TCEA”) had expressly excluded the UT’s appeal-permission decisions from the general right of appeal that lay on a point of law to the Court of Appeal.Footnote ⁸¹ It was accepted by the Divisional Court and Court of Appeal that, in such circumstances, full-scale judicial review would be inappropriate, as it would run contrary to the intended autonomy and finality of the new tribunal system that this limited right of appeal indicated.Footnote ⁸² As such, the Divisional Court and Court of Appeal limited judicial review to pre-Anisminic jurisdictional errors of law – in other words, errors as to the general type of claim over which the UT could adjudicate – and fundamental breaches of the principles of procedural fairness. But the Supreme Court preferred a different approach, drawing directly on the constitutional principle of the rule of law and a novel application of the doctrine of proportionality.Footnote ⁸³ The Supreme Court said its role was to determine what level of review would strike a proportionate balance between, on the one hand, the rule of law’s demands for the correction of legal errors and, on the other hand, the legal system’s limited resources and need for finality.Footnote ⁸⁴ On this basis the court restricted judicial review to cases which raised important points of principle or practice or where there was some other compelling reason for review – criteria set down by section 55(1) of the Access to Justice Act 1999 for second appeals. Not only did this detach the scope of review from classical concepts of jurisdiction and vires, but the effect of the Supreme Court’s approach was to immunise from review acts that, on classical theory, were void ab initio because marred by a (non-reviewable) error of law. As such, Cart was dismissed by Christopher Forsyth as a “blasphemy against basics”.Footnote ⁸⁵

The Supreme Court went on to apply its Cart methodology to ouster clauses in R. (Privacy International) v Investigatory Powers Tribunal.Footnote ⁸⁶ As in Cart, most of the Justices in Privacy International preferred not to rely on classical concepts when interpreting the ouster clause in section 67(8) of the Regulation of Investigatory Powers Act 2000 (“RIPA”). Lord Carnwath, giving the principal majority judgment,Footnote ⁸⁷ said it was “highly artificial, and somewhat insulting” to describe the Investigatory Powers Tribunal’s “closely reasoned” judgment as a “nullity”.Footnote ⁸⁸ Demonstrating a contemporary judicial tendency to elevate broad principle above doctrinal technicality, he spoke of how the principles of administrative law had “evolved since the somewhat technical debates in Anisminic”Footnote ⁸⁹ and suggested Cart bore witness to “the continuing strength” of a free-standing “fundamental presumption against ousting the supervisory role of the High Court”.Footnote ⁹⁰ It was this “common law presumption against ousters”, resting on the fundamental principle of the rule of law, that was of “critical importance” to resolving the case at hand.Footnote ⁹¹ Following Cart, this duty to uphold the rule of law was to be discharged in a principled and proportionate manner,Footnote ⁹² but did not require recourse to classical administrative law concepts.

Do Cart and Privacy International presage the demise of classical administrative law, or at least the demise of its relevance in ouster clause cases? It is suggested not. Perhaps ironically, it is Parliament that has resisted the Supreme Court’s attempts at jettisoning classical theory. Section 11A of the TCEA, introduced by section 2 of the JRCA,Footnote ⁹³ was enacted in response to Cart, altering the circumstances in which the UT’s appeal-permission decisions can be reviewed.Footnote ⁹⁴ In similar terms to the ouster clause in Anisminic,Footnote ⁹⁵ subsection (2) says the UT’s appeal-permission decisions are “final, and not liable to be questioned or set aside in any other court”. But responding to the reasoning in Anisminic, the clause goes on to specify in subsection (3)(a) that “the Upper Tribunal is not to be regarded as having exceeded its powers” – that is, as having acted outside its jurisdiction – “by reason of any error made in reaching the decision”.Footnote ⁹⁶ The clause says that “the supervisory jurisdiction” – of the High Court, namely judicial reviewFootnote ⁹⁷ – “does not extend to, and no application or petition for judicial review may be made or brought in relation to, [a] decision”,Footnote ⁹⁸ and makes clear that any reference in the clause to a “decision” “includes any purported decision”.Footnote ⁹⁹ But the clause goes on to make clear that judicial review remains available when the UT wrongly determines, at the commencement of its inquiry, that an application for an appeal is valid,Footnote ¹⁰⁰ where the UT is improperly constitutedFootnote ¹⁰¹ or where the tribunal acts with bad faithFootnote ¹⁰² or “in such a procedurally defective way as amounts to a fundamental breach of the principles of natural justice”.Footnote ¹⁰³

It is important to recognise the novelty of section 11A. Earlier ouster clauses enacted in response to Anisminic avoided any precise delimitation of the scope of a decision maker’s jurisdiction, at most suggesting that an administrative act should be excluded from review notwithstanding it being outside the jurisdiction of the decision maker.Footnote ¹⁰⁴ But section 11A, following the approach suggested by the lower courts in Cart, ties exclusion of judicial review to a precise statutory delimitation of the UT’s jurisdiction itself.Footnote ¹⁰⁵ That “purported decision[s]” of the UT are, by subsection (7), expressly excluded from review might appear to immunise ultra vires decisions from review. But this provision is to be read alongside subsection (3)(a) which clearly foregrounds the connection between the availability of judicial review and the jurisdiction of the UT. Taken together, these provisions serve as a signal from Parliament that the post-Anisminic (or post-Page) presumption that all administrative law errors take decision makers outside their jurisdiction should not apply to UT appeal-permission decisions. It is only on the limited grounds expressly enumerated by section 11A that the UT acts ultra vires and so only on these grounds is judicial review available. The clause does not so much seek to circumvent the conceptual apparatus of classical administrative law, then, as to co-opt it to its own ends. Dingemans L.J. made this clear in LA when he said that “the effect of the wording [in section 11A] is … to restore the ‘pre-Anisminic’ excess of jurisdictional and fundamental denial of justice tests which were adopted by the Divisional Court and Court of Appeal in Cart”.Footnote ¹⁰⁶ As he went on to say, Parliament, by enacting section 11A, tackled the issue of nullity head-on.Footnote ¹⁰⁷

This, we suggest, goes some way to explaining the effectiveness of section 11A, as confirmed not only in LA but in four other recent cases: R. (Oceana) v Upper Tribunal,Footnote ¹⁰⁸ Sooy v Secretary of State for the Home Department,Footnote ¹⁰⁹ R. (Karim) v Upper Tribunal Footnote ¹¹⁰ and Singh v Secretary of State for the Home Department.Footnote ¹¹¹ While it might be tempting to read these cases as a continuation of the Cart and Privacy International approach, rooted in proportionate protection of the rule of law (the courts “permitting” section 11A to oust review because the threat to the rule of law is not as pronounced),Footnote ¹¹² a better reading of these cases, we suggest, is that they show a renewed respect for Parliament’s own role in upholding the rule of law and its use of classical administrative law concepts to that end. As Lord Phillips said in Cart, protecting the rule of law involves a “partnership between Parliament and the judges”.Footnote ¹¹³ Lord Dyson suggested that the second-appeal criteria were to be used to protect the rule of law unless and until Parliament adopted alternative means.Footnote ¹¹⁴

While Fordham J. made clear in Karim that the rule of law remains of fundamental importance to the courts when interpreting ouster clauses, the section 11A case law returns this fundamental constitutional principle to its correct place in public law adjudication. First, the case law reminds us that, as Saini J. explained in Oceana, “[t]he most fundamental rule of our constitutional law is that the Crown in Parliament is sovereign and that legislation enacted by the Crown with the consent of both Houses of Parliament is supreme”.Footnote ¹¹⁵ The rule of law must, in other words, be considered alongside parliamentary sovereignty.

Second, the case law confirms that Cart and Privacy International were wrong to detach the rule of law from questions of jurisdiction and vires, treating it as a free-standing doctrine to be directly invoked for the resolution of administrative law disputes.Footnote ¹¹⁶ Fundamental legal principles such as parliamentary sovereignty and the rule of law require further specification by the courts through more particular legal doctrines. Without such further specification, law becomes unpredictable and uncertain, the product of uncabined adjudicative discretion. The role of a reviewing court is not to ask, simpliciter, whether a decision maker has acted contrary to “the rule of law”. Whether a decision maker has so acted requires the court to determine whether the decision maker has acted outside its legal powers. This requires the court to determine whether a particular ground of review – error of law, legitimate expectations, procedural fairness, etc. – has been breached, which itself involves consideration of past judicial precedent outlining the limits of a particular ground of review, as well as of statutory text and context that help determine how that ground of review is to be applied in a given case. Trevor Allan explains the relationship between the rule of law and classical administrative law in such terms:

It is the same with ouster clauses. Whether an administrative act is reviewable or not cannot be determined by direct appeal to the rule of law. Indeed Lord Sumption (with whom Lord Reed agreed) made this point clearly in Privacy International. The reviewability of an administrative act depends on the jurisdiction of the decision maker. That jurisdiction depends, in turn, on the textual features and context of the statute conferring the decision maker’s powers – including any ouster clause – interpreted in light of constitutional principles and mediated through the established doctrines of administrative law.Footnote ¹¹⁸ By setting out the reviewability of the UT’s appeal-permission decisions as part of a statutory demarcation of the UT’s jurisdiction, section 11A and the cases upholding it offer an approach to ousting judicial review firmly rooted in classical administrative law. We believe this “back to basics” approach should be welcomed: classical concepts of vires and validity are not only deeply rooted in the rule of law, but offer the surest guarantee of its protection.

A. Extra-Textual Limits on Devolved Competence

Having considered three indirect challenges to classical administrative law, we now turn to direct, or explicit, challenges to the classical account, beginning with the Supreme Court’s case law on what we will call “extra-textual” limits on the devolved legislatures’ powers. Section 28(1) of the Scotland Act 1998 provides that: “Subject to section 29, the [Scottish] Parliament may make laws, to be known as Acts of the Scottish Parliament.” Section 29(1) says that: “An Act of the Scottish Parliament is not law so far as any provision of the Act is outside the legislative competence of the Parliament.” And section 29(2) sets out five grounds on which an Act is outside competence.Footnote ¹¹⁹ In AXA General Insurance Ltd. v The Lord Advocate,Footnote ¹²⁰ the question arose whether the textual limits on competence set out in section 29(2) are exhaustive or whether Scottish enactments are also open to challenge on extra-textual grounds. Lord Hope said that because “there is no provision in the Scotland Act which excludes this possibility, I think that it must follow that in principle Acts of the Scottish Parliament are amenable to the supervisory jurisdiction of the Court of Session at common law”.Footnote ¹²¹ However, he went on to conclude that, in recognition of the Scottish Parliament’s “firm rooting in the traditions of a universal democracy”, review should lie “only in the most exceptional circumstances”Footnote ¹²² and that its enactments are not reviewable on the grounds of “irrationality, unreasonableness or arbitrariness”.Footnote ¹²³ Lord Reed, delivering the other leading judgment, reached a similar view, concluding that Scottish enactments will be unlawful – on extra-textual grounds – if they “abrogate fundamental rights” or “violate the rule of law”.Footnote ¹²⁴

We have already seen that in the Continuity Bill case, the Supreme Court said that the “result of a want of legislative competence is that a Scottish enactment is a nullity”.Footnote ¹²⁵ However, in that case, the court also said that there is a “difference between a want of legislative competence and more general grounds for judicial review on public law grounds”.Footnote ¹²⁶ It followed, said that court, that while “a want of legislative competence” on textual grounds rendered an Act of the Scottish Parliament a “nullity”, a “Scottish enactment which is held by a court to be unlawful on more general public law [i.e. extra-textual] grounds is not necessarily a nullity”.Footnote ¹²⁷ This appears to be a direct challenge to the classical account, according to which unlawfulness on the basis of any of the grounds of judicial review renders the relevant act void ab initio. However, for three reasons, neither the position of the Supreme Court in the Continuity Bill case on Scottish enactments specifically, nor the implications of that position for the legal status of unlawful measures generally, can withstand scrutiny.

First, the distinction drawn by the court in the Continuity Bill case between textual and extra-textual limits which respectively do and do not render the relevant measure void cannot be reconciled with the general scheme of public law, according to which, as Lord Irvine L.C. put it in Boddington, the Anisminic decision established “that there was a single category of errors of law, all of which rendered a decision ultra vires”.Footnote ¹²⁸ From this, it followed that collateral challenge (which, as we explained above, is available only if the impugned act is a nullity) is available irrespective of the ground on which an act is unlawful. In articulating that view in Boddington, the House of Lords overruled an earlier case in which the Divisional Court had sought to distinguish between “substantively invalid” subordinate legislation that was made outside the scope of the enabling legislation and subordinate legislation flawed only on “procedural” grounds.Footnote ¹²⁹

Second, not only is the distinction drawn by the Supreme Court in the Continuity Bill case incompatible with the general view that prevails in this area: it is not even compatible with what that court itself said when, in AXA, it first acknowledged the possibility of devolved legislation being challenged on extra-textual grounds. Having said that Scottish enactments can be challenged on such grounds only in “the most exceptional circumstances”,Footnote ¹³⁰ Lord Hope gave the examples of legislation that sought “to abolish judicial review or to diminish the role of the courts in protecting the interests of the individual” and concluded that “[t]he rule of law requires that the judges must retain the power to insist that legislation of that extreme kind is not law which the courts will recognise”.Footnote ¹³¹ It is surely no coincidence that Lord Hope’s language concerning provisions that breach extra-textual limits – they are “not law” – mirrors what section 29(1) says about provisions that breach the textual limits set out in section 29(2), which the Supreme Court in the Continuity Bill case went on to equate with “nullity”.

Third, the court’s stance in that case on the legal status of legislation that exceeds extra-textual limits on competence is also in tension with the judgment of Lord Reed in AXA. In embracing the existence of such limits, Lord Reed relied on the principle of legality, citing Simms, in which Lord Hoffmann said that “[f]undamental rights cannot be overridden by general or ambiguous words” and that absent “express language or necessary implication to the contrary, the courts […] presume that even the most general words [in legislation] were intended to be subject to the basic rights of the individual”.Footnote ¹³² In AXA, Lord Reed went on to say that “[t]he principle of legality means not only that Parliament cannot itself override fundamental rights or the rule of law by general or ambiguous words, but also that it cannot confer on another body, by general or ambiguous words, the power to do so”.Footnote ¹³³ When it enacted the Scotland Act 1998, said Lord Reed, “Parliament did not legislate in a vacuum” and “cannot be taken to have intended to establish a body which was free to abrogate fundamental rights or to violate the rule of law”.Footnote ¹³⁴ This invocation of the principle of legality – which, as the Supreme Court recently reaffirmed,Footnote ¹³⁵ is a principle of statutory interpretation – makes it abundantly clear that Lord Reed envisaged no conceptual distinction between the textual and extra-textual limits on the Scottish Parliament’s legislative competence, both types of limits being attributable to the UK Parliament’s stated or constructive intention. Given that breach of both types of limit therefore renders a Scottish enactment or provision ultra vires, it is logically impossible to see why those that breach extra-textual limits are not nullities in precisely the way that those enacted in contravention of textual limits are.

The upshot is that the Supreme Court’s obiter comments in the Continuity Bill case are unsustainable: they are incompatible with the court’s judgment in AXA, in which the legal basis of extra-textual limits on competence were carefully considered and with the courts’ wider approach to the legal status of unlawful measures. The position therefore remains that devolved legislation enacted outwith competence on any ground is “not law” and is null and void.

B. Judicial and Academic Challenges

Direct challenges to the classical account have not only come in the context of devolution, but also in administrative law more broadly. Some of these challenges have been judicial. In recent years the Supreme Court has questioned classical orthodoxy directly. In Privacy International, for example, a number of Justices expressed misgivings about the idea that all unlawful administrative acts should be considered ultra vires and so void ab initio.Footnote ¹³⁶ These issues were raised yet again two years later in R. (Majera) v Secretary of State for the Home Department.Footnote ¹³⁷

The First-tier Tribunal (“FTT”) had ordered the release of the claimant on immigration bail without imposing a condition that he not undertake unpaid employment, contrary the Home Secretary’s request. The Home Secretary claimed that because the FTT had made its bail order in breach of the requirement in Schedule 2 to the Immigration Act 1971 that the claimant first enter into a recognizance,Footnote ¹³⁸ she was able to treat it as a nullity and so impose alternative restrictions on the claimant under Schedule 3 – powers that could only be exercised if the FTT had not legally made a bail order.Footnote ¹³⁹ The claimant sought judicial review, arguing that court orders, including orders of the FTT, should be treated by the executive as binding, such that the Home Secretary did not have the power to impose bail conditions that an immigration judge had considered but declined to impose.

The Court of Appeal in Majera adopted a robustly orthodox approach to these questions, holding that all unlawful acts are null and void such that, if the FTT’s bail order was unlawful, the Home Secretary was free to disregard it.Footnote ¹⁴⁰ But in the Supreme Court Lord Reed, delivering the unanimous judgment of the court, said that such doctrinal classicism was “inapposite to the order of a court or tribunal such as the First-tier Tribunal”Footnote ¹⁴¹ and, in relation to administrative acts more broadly, an “oversimplification”.Footnote ¹⁴² As he went on to explain: “Although judges have commonly used expressions such as ‘null’ and ‘void’ to describe unlawful administrative acts and decisions, it has nevertheless been recognised that the notion that such acts and decisions are utterly destitute of legal effect, as if they had never existed at all, is subject to important qualifications.”Footnote ¹⁴³

Lord Reed appeared to entertain three particular concerns:Footnote ¹⁴⁴ (1) that unlawful administrative acts might go unchallenged and might therefore be treated by relevant actors as if they were valid;Footnote ¹⁴⁵ (2) that unlawful acts can have legal consequences;Footnote ¹⁴⁶ and (3) that treating all unlawful acts as total nullities might be administratively inconvenient.Footnote ¹⁴⁷ This led him to argue that validity itself is a “relative concept”,Footnote ¹⁴⁸ that “[a]n unlawful act or decision cannot […] be described as void independently of, or prior to, the court’s intervention”Footnote ¹⁴⁹ and that a “more flexible approach” is required to vindicate “important values underpinning the court’s supervisory jurisdiction, such as the public interest in legal certainty, orderly administration, and respect for the rule of law”.Footnote ¹⁵⁰ Yet, in saying this, Lord Reed contradicted his own judgment in R. (UNISON) v Lord Chancellor, in which he held that regulations that were unlawful because they breached a common law constitutional right were necessarily “unlawful ab initio, and must be quashed”.Footnote ¹⁵¹ Contrary to the flexible approach he advocated in Majera, Lord Reed’s analysis in UNISON suggests that in at least some circumstances, voidness follows necessarily – and, crucially, independently of any judicial intervention or remedy – from the simple fact of unlawfulness.

Similar concerns can be discerned in recent academic criticism of classical administrative law. While he is not alone in advancing such criticism,Footnote ¹⁵² we focus here on the work of Thomas Adams because he is unusual in engaging directly with the implications of such criticism for the availability of collateral challenge. He dismisses what he calls “the standard theory of administrative unlawfulness” – identical to what we call the “classical account”Footnote ¹⁵³ – on grounds substantively similar to Lord Reed in Majera. First, consistently with Lord Reed’s concern about relativity, Adams argues that “the most obvious problem for the defender of the standard theory has to do with the fact that unlawful administrative decisions have, unless and until challenged in the courts, the same force as lawful decisions”Footnote ¹⁵⁴ and that “[t]he effects of illegality […] only manifest themselves when we reach the courtroom”.Footnote ¹⁵⁵ To underline this point, Adams invokes Wade’s view that statutory time limit clauses that prevent unlawful acts being challenged after a specific statutory periodFootnote ¹⁵⁶ “render valid what would otherwise be invalid”.Footnote ¹⁵⁷ Second, in support of the view that unlawful administrative acts can be legally effective, Adams relies on the House of Lords’ willingness in F. Hoffmann La Roche & Co. A.G. v Secretary of State for Trade and Industry Footnote ¹⁵⁸ to grant an interim injunction enforcing an impugned administrative order pending a substantive hearing to determine its legality. Lord Reid’s statement that “an order made under statutory authority is as much the law of the land as an Act of Parliament unless and until it has been found ultra vires”Footnote ¹⁵⁹ shows, Adams suggests, that unlawful acts are legally effective. Third, Adams contends that the practical implications of treating all unlawful acts as void is too restrictive, arguing that “[i]nstead of saying that unlawfulness entails invalidity we can say that unlawfulness implies a duty on the part of the court to invalidate unlawful administrative action”.Footnote ¹⁶⁰

Taking Lord Reed’s and Adams’s concerns together, the question arises whether they are right to contend that the classical view must be renounced in order to (1) accommodate the fact that unlawful acts are sometimes treated as if they were lawful; (2) account for the legal effects that unlawful acts sometimes produce; and (3) avoid administrative chaos when unlawful acts have been relied upon, including, in some instances, as the basis for series of subsequent acts. In explaining why our answer to that question is “no”, it is convenient to take each of these points in turn.

First, even before the decision in Anisminic, Sir William Wade was able to explain that “[h]owever destitute of legitimacy at its birth, [an unlawful act] is legitimated when the law refuses to assist anyone who wants to bastardise it”.Footnote ¹⁶¹ But such legitimacy is apparent, not actual. It is merely the natural consequence of a finite legal system. As Christopher Forsyth has explained, “[the law] is not omnipotent; it cannot set everything right. Unlawful activity may [and does] have effects which cannot be rectified […] The law cannot wash away all signs of illegality”.Footnote ¹⁶² Unless a court is in a position to make clear that an administrative act is unlawful – and so, on classical theory, a nullity – people will generally assume it is valid. Pace Adams, the invalidity of an unlawful act does not depend on the court’s recognition of unlawfulness. When Wade said the effect of a time limit clause was “to render valid what would otherwise be invalid”,Footnote ¹⁶³ for example, he can only have meant that an invalid decision was, through the passing of time, rendered effectively valid.Footnote ¹⁶⁴ Much the same can be said for all judicial review, which must usually be sought within three months,Footnote ¹⁶⁵ after which an invalid administrative act will be unchallengeable.Footnote ¹⁶⁶ This does not mean it becomes valid at that point.Footnote ¹⁶⁷ All of this is of a piece with the principle of legal relativity, which we discussed above and which is perfectly capable of accommodating the commonplace phenomenon of unlawful administrative acts being regarded as valid because suitable opportunities for challenge have elapsed.

Nor is it conceptually problematic, in relation to point (2), that unlawful administrative acts sometimes have legal consequences. All of this was adequately explained by Forsyth’s “second actor theory” almost 30 years agoFootnote ¹⁶⁸ – a theory that has itself received express recognition in subsequent case law.Footnote ¹⁶⁹ Forsyth reminds us that administrative acts have both a factual and a legal existence.Footnote ¹⁷⁰ Even if an administrative act is unlawful and so, being a nullity, has no legal existence, this does not mean it ceases to exist as a matter of fact. And sometimes statute or common law will lead to the mere factual existence of an administrative act being enough to ground the legal validity of a subsequent act. Just as a void marriage can produce legitimate children,Footnote ¹⁷¹ so can a void administrative act have legal consequences. The mere factual existence of a byelaw will be enough to make valid a police officer’s decision to arrest or detain someone for its breach.Footnote ¹⁷² Factually existent acts are all that are needed for valid appeals.Footnote ¹⁷³ Parliament has legislated to allow for the enforcement of merely purported planning enforcement notices, namely nullities.Footnote ¹⁷⁴

Adams rejects Forsyth’s second actor theory as supposing an artificial link between existence and consequences: “what it is for a decision to exist in the legally relevant sense is for it to have the legal consequences that it claims for itself.”Footnote ¹⁷⁵ For an unlawful act to have legal consequences, he claims, is for it to have a legal, not factual, existence. As such, the “natural explanation” of Hoffmann La Roche v Secretary of State for Trade and Industry Footnote ¹⁷⁶ is that, “[a]n injunction being a remedy designed to prevent a body from acting in breach of the law”, “the Government’s order had legal effect until pronounced upon by a competent court”.Footnote ¹⁷⁷ But it is precisely the courts’ practice of granting interlocutory relief – a practice not exclusive to, or even principally a matter of, administrative law – that shows the absence of any necessary connection between legal existence (or validity) and legal consequence.Footnote ¹⁷⁸ Save for the most blatant cases of illegality, the courts will often presume legality and, on that basis, issue interim injunctions to enforce what might transpire to be invalid acts pending the final determination of their legality at a substantive hearing.Footnote ¹⁷⁹ Just as with suspended quashing orders and the courts’ powers to give temporary effect to void delegated legislation,Footnote ¹⁸⁰ such interim remedies are not granted, as Adams claims, “to ensure the law is respected”,Footnote ¹⁸¹ but only to “preserve the status quo” by upholding, on the balance of convenience, what the law is said to be.Footnote ¹⁸² This prevents administrative chaos and the impairment of legal rights before a substantive hearing to resolve such issues definitively. Interim injunctions are by no means a recognition of validity, but merely “an evidential matter at the interlocutory stage […] pending a final decision by the court”.Footnote ¹⁸³

Indeed, as Tom Hickman has suggested, Majera could itself have been resolved through classical administrative law doctrine.Footnote ¹⁸⁴ The Supreme Court’s principal concern in Majera, to extend to the FTT the common law principle that the orders of courts of competent jurisdiction must be obeyed until set aside, could adequately have been explained on the basis of Forsyth’s second actor theory.Footnote ¹⁸⁵ The court, when interpreting the Home Secretary’s power to impose bail conditions under Schedule 3, should have interpreted the Immigration Act with the presumption that court orders are to be obeyed. This would have led to the conclusion that, notwithstanding the alleged invalidity of the FTT’s decision, the mere factual existence of an FTT bail order was enough to bar the Home Secretary’s use of her Schedule 3 powers. This is how the earlier authorities approached this issue. In Chuck v Cremer,Footnote ¹⁸⁶ for example – a case cited by Lord Reed – Lord Cottenham L.C. said that the principle that court orders must be obeyed applied to orders “whether null or valid”.Footnote ¹⁸⁷ While Lord Reed in Majera suggested it was a “paradox” that null orders must be obeyed until set aside,Footnote ¹⁸⁸ this is only the case if one forgets the second actor theory.Footnote ¹⁸⁹

Turning to point (3), we can certainly recognise, therefore, that Lord Reed’s desire to mitigate the potential adverse consequences of nullity was well placed. But while Lord Reed had the correct diagnosis, his prescription was misguided. There is no need for the courts to jettison classical doctrine in favour of a “more flexible approach”.Footnote ¹⁹⁰ Lord Reed’s concerns for “legal certainty, orderly administration, and respect for the rule of law”Footnote ¹⁹¹ cannot only be accommodated by, but actually underpin, the classical approach to administrative law. When the courts determine whether an unlawful administrative act can and should bear legal consequences, they do so through an ordinary process of statutory interpretation that depends on and incorporates these normative and practical concerns.Footnote ¹⁹² It might lead to administrative chaos, for example, if the police’s powers of arrest for breach of a byelaw depended on that byelaw’s validity. And so the courts have held that arrest powers depend on nothing more than a byelaw’s factual existence.Footnote ¹⁹³ But the rule of law itself demands that no one can be convicted for breach of a byelaw unless that byelaw is legally valid. And this normative assessment has shaped the court’s interpretation of applicable statutes as well.Footnote ¹⁹⁴

For similar reasons, we cannot accept Adams’s idea that a finding of administrative unlawfulness simply imposes a duty on the part of the court actively to invalidate an unlawful but otherwise valid administrative act. As Adams himself recognises, a “decisive objection” to this is that “it suggests that administrative bodies have the legal capacity to make decisions beyond the limits of their own powers”.Footnote ¹⁹⁵ Adams’s resolution of this paradox is to construct a division between express statutory limits on a decision maker’s powers and limits which are said to stem independently from the common law:Footnote ¹⁹⁶ “where the administration has breached only common law constraints on the use of its power,” he says, “the conditions in the relevant empowering legislation will not themselves have been violated and so there is a clear basis in the law for the decision to stand unless and until the court intervenes.”Footnote ¹⁹⁷ But this is to charge too incautiously into the choppy waters of the constitutional foundations of judicial review. For many, ourselves included, the grounds of review, while flowing from the adjudicative creativity of the common law, are nonetheless part and parcel of a decision maker’s statutory authority.Footnote ¹⁹⁸ When Parliament confers powers on a decision maker, it does so with the intention that they are to be exercised in accordance with the rule of law, which itself is given concrete form through the common law’s creative development of the grounds of review.Footnote ¹⁹⁹

Even if one does not accept the necessity of any connection between legislative intent and the grounds of review (the so-called “common law” theory), it can still nonetheless be recognised that the language of ultra vires constitutes the “lingua franca” of our profession,Footnote ²⁰⁰ vires being constituted as much by common law principles as by any express statutory delimitation of power. Adams’s startling assertion that decision makers have “the power to make binding determinations about the scope of [their] own jurisdiction” (subject to a judicial “duty” to quash such determinations),Footnote ²⁰¹ made without the support of express authority,Footnote ²⁰² is to be rejected not only because it radically undermines this accepted lingua franca, but, more fundamentally, directly challenges the widely accepted view that administrative powers are not plenary but limited (by Parliament, by the common law, or both) and that it is constitutionally fundamental that such limits be judicially enforced. Among other things, Adams’s framework fails (as we shall see) to account for the availability of collateral challenge, is inconsistent with clear authority to the effect that administrative acts that are unlawful – whether for breach of a statutory jurisdictional condition or because contrary to the principles of good administration – are invalid independently of judicial intervention and, as the author himself appears to acknowledge,Footnote ²⁰³ cannot be reconciled with the undeniable fact that courts can and sometimes do exercise a remedial discretion that flatly contradicts the supposed judicial duty to quash unlawful administrative acts.

The rupture between administrative law doctrine and the protection of the rule of law inherent in the “more flexible approach” of Lord Reed and Adams’s preference for a “duty to invalidate” unlawful acts is shown most clearly in their treatment of collateral challenge. In Majera, Lord Reed considered Lord Irvine L.C.’s statement in Boddington that Anisminic, as interpreted by Page,Footnote ²⁰⁴ established that all unlawful administrative acts are void ab initio.Footnote ²⁰⁵ But he said this statement was “unnecessary for the decision of the case, which was concerned solely […] with the question whether a person could be convicted in criminal proceedings of violating subordinate legislation which was held to be unlawful”.Footnote ²⁰⁶ Yet the whole basis of Lord Irvine’s holding that collateral challenge was available was that, as a matter of law, if the byelaw that grounded Mr. Boddington’s prosecution was unlawful and a nullity, Mr. Boddington could not be said to have committed a criminal offence. If the unlawful byelaw had merely been voidable, then a stay of Mr. Boddington’s trial in the magistrates’ court would have been required to obtain a quashing order from the High Court, with all the time and expense this involves, not to mention the discretionary nature of judicial review and its strict time limits.Footnote ²⁰⁷ Without the voidness ab initio of unlawful administrative acts, then, their unlawfulness cannot be invoked as a defence to criminal prosecution. And while Lord Reed is correct in noting, in Majera, that some of the other speeches in Boddington were more agnostic as to the question of whether all unlawful acts should be treated as void ab initio,Footnote ²⁰⁸ no alternative legal basis for collateral challenge was at any time advanced.

It is for this reason that we also reject Adams’s suggestion that collateral challenge could still be available even if its classical underpinnings were jettisoned.Footnote ²⁰⁹ Adams argues that treating unlawful administrative acts as merely voidable would not preclude their unlawfulness being raised defensively: doing so would be “justified on rule-of-law based grounds” simpliciter. Footnote ²¹⁰ We have already explained why we believe direct invocation of the rule of law, unmediated by concrete legal doctrines, is a misunderstanding of the principle’s role in public law adjudication. But Adams goes further and claims the availability of collateral challenge today “presupposes rather than denies the validity of the unlawful measures”.Footnote ²¹¹ This is because “[a] defence is only necessary […] in the context where an individual has committed a legal wrong, and the grounds for such a wrong may only be furnished by a valid as opposed to an invalid administrative act”.Footnote ²¹² This is to misunderstand collateral challenge. The defendant in Boddington invoked the invalidity of the byelaw to argue that he had not committed any wrong at all, rather than that the wrong he had done could be justified. Adams appears to conflate defences based on an alleged failure of proof with “supervening” defences whereby the defendant alleges his or her otherwise criminal conduct is justifiable or excusable.Footnote ²¹³

For the sake of completeness we should also address here a suggestion made in the Independent Review of Administrative Law that “collateral challenges could easily be retained” if the classical account were to be cast aside. Having argued that courts should be given “the freedom to decide whether or not to treat an unlawful exercise of public power as having been null and void ab initio”,Footnote ²¹⁴ the Independent Review went on to contend that collateral challenge – the unavailability of which, it accepted, would put the law in a “radically defective state”Footnote ²¹⁵ – could be preserved in the following way:

However, for several reasons, the approach advocated by the Independent Review is problematic. First, the starting point it adopts – that questions of nullity are reduced to matters of judicial discretion – robs the notions of validity and voidness of any meaningful conceptual content, rendering them labels applied by courts on a purely ex post basis. By conflating conceptual reasoning and judicial discretion in this way, the Independent Review’s approach arguably yields the worst of both worlds: a pseudo-conceptual framework that is, in the final analysis, nothing more than a function of judicial discretion. Second, the notion, which the Independent Review clearly contemplates, of a merely voidable administrative act being capable of being the subject of a declaration of nullity is nonsensical, for it contemplates the possibility of an administrative act being both null (for the purposes of being a potential candidate for a declaration of nullity) and not null (since if unlawful administrative acts were acknowledged to be null, the mental gymnastics proposed by the Independent Review would be otiose). Third, for the courts to ignore valid administrative acts on the basis that they could, in some hypothetical judicial review, be quashed would run contrary to the judicial role of enforcing the law as it applied to the parties at the time their dispute materialised. There is no clear basis at common law for the courts to refuse to enforce the law in the way envisaged by the Independent Review. Fourth, any statutory powers authorising the courts to act in this way would themselves be problematic: it would be unduly difficult for citizens to plan their lives around voidable acts that may or may not be quashed, depending on the exercise of judicial discretion. Such an approach would run contrary to the rule of law’s requirements of generality, prospectivity, clarity and constancy,Footnote ²¹⁷ thereby undermining law’s co-ordinating function.

As we have argued above, such is the importance of collateral challenge that calls for an abandoning of the classical account should not be accepted unless an adequate alternative basis for collateral challenge can be found.Footnote ²¹⁸ The history of administrative law suggests this will be unlikely. It is undeniable that before Anisminic and Page there was a sizeable category of unlawful administrative acts that were merely voidable, such that the scope of collateral challenge was arguably narrower than today. But any act that is challengeable collaterally has always, necessarily, been void. There has never been a time in administrative law when all unlawful acts were simply voidable but where collateral challenge was nonetheless readily available.