1. The Protocol and its incorporation: consequences for devolution in Northern Ireland

It is often said that the Protocol created a border in the Irish Sea, with wholly different customs and regulatory conditions between Great Britain and Northern Ireland.Footnote ³ However, the supposed centrality of the Protocol to the different customs and regulatory regimes is somewhat reductive, as this section explores.

Instead, both the Protocol and the Trade and Cooperation Agreement between the UK and the EU (TCA) provide for differential treatment between Great Britain and Northern Ireland, covering different areas and in different ways. For example, the Protocol continues to apply the EU's customs and regulatory code to Northern Ireland only, so that trade in goods between Northern Ireland and the EU remains covered under the rules of the EU's single market.Footnote ⁴ At the same time, the TCA also provides for different rules between Northern Ireland and Great Britain, for example in relation to road passenger transport operators, so that a British operator based in the UK outside of Northern Ireland requires specific authorisation to both pick up and set down passengers in the EU (and vice versa),Footnote ⁵ while a Northern Ireland operator may do so in the Republic of Ireland (and vice versa) without specific authorisations from the Republic.Footnote ⁶

For the purposes of this paper, however, the specific content of both the Protocol and the TCA is not the issue – it is their respective scope and implementation. In terms of scope, the purpose of the Protocol is to ‘address the unique circumstances on the island of Ireland’ as a consequence of Brexit.Footnote ⁷ As a result, there is considerable potential for dynamism in the future relationship between the UK in respect of Northern Ireland and the EU,Footnote ⁸ whereby the Protocol itself provides a mechanism for the adoption in Northern Ireland of future EU law which may fall within the scope of the Protocol.Footnote ⁹ In this regard, the Protocol strongly envisions that such laws will be adopted, providing for ‘remedial measures’ to be taken by the EU if the Joint Committee of the UK and EU (which oversees the implementation and application of the Withdrawal Agreement)Footnote ¹⁰ cannot agree to adopt any such law.Footnote ¹¹ Crucially, new EU laws which may be adopted as being within the scope of the Protocol may include laws which fall outside the scope of Articles 5–10 of the Protocol. The Northern Ireland Assembly may vote to end the application of only this part of the Protocol.Footnote ¹² By contrast, while the TCA exhorts the UK and the EU to adopt ever higher labourFootnote ¹³ and environmental standards for example,Footnote ¹⁴ it lacks the Protocol's depth of process for maintaining at least some parity between the two legal systems. Consequently, while the Protocol has both eyes set firmly on the future, with ways of expanding its application, the TCA is a much less ambitious agreement which, if it provides for any future ambition, does so with a languidness bordering on moribundity.

The upshot of the above is that the UK Government need not positively diverge from EU legal provisions (including goods standards) in order to cause divergence between Great Britain and Northern Ireland. Were Whitehall simply to stand still while Brussels moved, divergence may ipso facto be achieved.

The implementing statutes of both the Protocol and the TCA are buttressed by broad Henry VIII powers. There are, however, notable and important differences between these powers. The main power to implement the Protocol is contained in section 8C of the European Union (Withdrawal) Act 2018 (EUWA 2018), as amended by the European Union (Withdrawal Agreement) Act 2020. This section authorises the making of any regulations that UK Ministers consider appropriate in relation to the Protocol, including the modification of any Acts of Parliament (including the EUWA 2018).Footnote ¹⁵ The only restrictions on this power relate to the UK Internal Market Act 2020Footnote ¹⁶ and Article 11.1 of the Protocol which allows the UK and the Republic of Ireland to build on the provisions of the Belfast (Good Friday) Agreement (GFA).Footnote ¹⁷ By contrast, the equivalent power in relation to the TCA contains a series of restrictions, including enumerating Acts of Parliament which cannot be amended or repealed.Footnote ¹⁸

The breadth of the section 8C power was evident in a recent regulation made thereunder. The Northern Ireland Secretary made the Protocol on Ireland/Northern Ireland (Democratic Consent Process) (EU Exit) Regulations 2020 in order to give effect to the Protocol's requirements of democratic consent from the Northern Ireland Assembly (in relation to the application of Articles 5–10, as covered earlier in this section). These regulations disapplied the petition of concern mechanism by which certain matters on which the Northern Ireland Assembly votes may be subject to cross-community consent.Footnote ¹⁹ This disapplication was required neither by the Protocol, nor by the UK Government's declaration on the issueFootnote ²⁰ which the Protocol required to be followed.Footnote ²¹ Arguably, it would also have breached the duty on ministers under the EUWA 2018 to act ‘in a way that is compatible with the terms of the Northern Ireland Act 1998’.Footnote ²² However, the disapplication was nevertheless achieved by amending the Northern Ireland Act 1998 (NIA 1998) itself, to give effect to the new schedule to that Act which provided for the democratic consent process.Footnote ²³ In the Northern Ireland High Court, this fait accompli was held to be an entirely lawful use of section 8C.Footnote ²⁴

Although the Protocol was incorporated via the EUWA 2018, it is important to appreciate that the statute's impact goes far beyond the incorporation of an international agreement. In the Northern Ireland context, the EUWA 2018 has four important effects. First, it privileges the effect of the Withdrawal Agreement and Protocol over ‘[e]very enactment’ in domestic law.Footnote ²⁵ This stands in contrast with the equivalent provision as regards the TCA, which only requires such modifications in domestic law as are ‘necessary for the purposes of complying with the international obligations of the United Kingdom under [the TCA]’.Footnote ²⁶ Secondly, the EUWA 2018 says nothing on how UK Ministers may act or exercise their powers in the Joint Committee. The most notable prescription relevant todayFootnote ²⁷ is in fact a restriction on ministers agreeing recommendations in relation to North-South cooperation under the GFA.Footnote ²⁸ Among the rare mentions of the Joint Committee in the EUWA 2018, section 15B specifies that the UK co-chair of the Joint Committee may only be a minister of the Crown. But the section prescribes who may exercise the powers of the UK co-chair, not how such powers may be exercised. Thirdly, the EUWA 2018 adds a further restriction on the competence of the Northern Ireland Assembly, in the form of certain kinds of retained EU law as specified in regulations made by UK ministers.Footnote ²⁹ Finally, the European Union (Withdrawal Agreement) Act 2020 also amends the NIA 1998 to qualify how the EUWA 2018 is protected from modification by the Northern Ireland Assembly.Footnote ³⁰ It does so by barring the Assembly from modifying the EUWA 2018 except for certain ‘excluded provisions’ of that statute,Footnote ³¹ which includes section 8C,Footnote ³² but not section 7A.

Thus, it would appear that the Protocol has direct effect (in much the same way that EU law previously had effect),Footnote ³³ that in its implementation, Parliament has authorised potentially vast quantities of ministerial law-making, that parliamentary control of (the UK half of) Joint Committee decision-making is very light-touch and that while the Assembly may not modify the direct effect of the Protocol, it can modify significant elements of its implementation (including the crucial section 8C power). However, appearances can be deceiving when it comes to devolved legislative competence in Northern Ireland.

The legislative competence of the Assembly is set out in sections 6, 6A and 7 of the NIA 1998, with the EUWA 2018 protected against modification by the Assembly because of its inclusion in section 7 (aside from specifically excluded provisions which may be modified), as set out earlier. Sections 6 and 7 cannot be read in isolation from one another, because entrenched enactment under section 7 is one of six ways in which an Act of the Assembly may be outside the Assembly's competence under section 6(2). However, it is axiomatic that these three sections do not contain the totality of devolved competence for Northern Ireland. In particular, the complexities of the different legislative provisions necessitate looking to restrictions on devolved competence by necessary implication. Especially relevant is the extent to which the subject of the EUWA 2018 provisions which devolved authorities may modify under section 7(2B) are nonetheless difficult because they straddle the boundary between excepted matters and other matters under the NIA 1998. Excepted matters may include, for example, the democratic consent vote provided for by regulation under section 8C, if the vote itself would ‘come under the ambit of international relations’Footnote ³⁴ and thus lie outside the Assembly's competence as an excepted matter.Footnote ³⁵ Relatedly, the Northern Ireland Secretary has broad powers to forbid the ‘making, confirming or approving’ of subordinate legislation,Footnote ³⁶ or to revoke such subordinate legislation as has already been made,Footnote ³⁷ should it be incompatible with the UK's international obligations, including those under the Protocol. The nature of the Northern Ireland Secretary's powers in relation to devolved law-making is quite expansive in Northern Ireland: in addition to exercising both ex ante and ex post control over subordinate legislation, it is the Northern Ireland Secretary who presents Assembly Bills for Royal Assent and may decide not to do so if such a Bill, inter alia ‘would be incompatible with any international obligations’.Footnote ³⁸ One may argue that the Northern Ireland Secretary's powers of control do not imply that legislation made in breach of international obligations is outwith devolved competence, strictly speaking,Footnote ³⁹ particularly when the power to control such law is discretionary and thus has a considerably political character. However, the degree of control exercised over such law, rendering it in effect almost impossible to be enacted or made, begs the question of the ability to make such law in the first place.Footnote ⁴⁰

Now, given that the EUWA 2018 is largely silent on the Joint Committee, it might appear as though Stormont might be able to make law concerning the Joint Committee insofar as within its competence, for example to express its views should the Northern Ireland statute book be expanded through future Joint Committee decisions. However, such decisions are not merely the mechanical consequences of Northern Ireland remaining within the customs and regulatory orbit of the EU; they may also be decisions on how close Northern Ireland remains tied to the evolving EU acquis while it remains within this orbit. These decisions may therefore speak directly to the character of the future relations between the UK in respect of Northern Ireland and the EU, which is outside Stormont's devolved competence. Any laws made at Stormont in this context, therefore, run the risk of cutting across the UK Government's conduct of international relations. Moreover, this risk is difficult to predict with any certainty. Although the implementation of international obligations is within Stormont's competence, so that it exists as an exception to the bar on interference with the conduct of international relations,Footnote ⁴¹ no part of the Withdrawal Agreement or the Protocol makes this distinction in the context of the Joint Committee. On one level, such a distinction would be impossible, because, as set out above, the proximity of Northern Ireland's laws to the evolving EU acquis is a matter both of preserving the integrities of the UK and EU customs areas and marketsFootnote ⁴² as well as of determining the character of the continuing relations between the UK and the EU. Thus, the legality of such legislation is, at best, debatable. If, instead, Stormont were to legislate to modify or otherwise interfere with the use of section 8C of the EUWA 2018 to implement Joint Committee decisions concerning the Protocol, such legislation may risk putting the UK Government in breach of its own obligations under the Protocol if Stormont's modifications delayed or otherwise hampered the implementation of Joint Committee decisions.Footnote ⁴³

Seen in this light, Stormont's powers to modify or otherwise interfere with the implementation of the Protocol is subject both to the overriding effect of the Protocol as enacted in the EUWA 2018 and the concomitant restrictions on legislative capacity (whether express or by implication).

However, the UK Parliament wears none of Stormont's legislative shackles; its powers to direct the conduct of UK-EU relations through prescriptive legislation is, moreover, a matter of recent historical record.Footnote ⁴⁴ Against this backdrop, what Parliament has enacted in the EUWA 2018 jars by comparison, especially with regard to the statute's silence on the Joint Committee. This silence strikes at the normative heart of parliamentary sovereignty (which, along with the Joint Committee, I explore in detail in the third section below) when examined through the lens of two landmark decisions of the UK Supreme Court, as I set out in the next section.

2. Sovereignty, democracy and giving Parliament a chance: analysing Miller 1 and Miller 2

This section examines, in some detail, the reasoning and conclusions of two important judgments of the UK Supreme Court: Miller 1 and Miller 2. Much analytical ink has been spilled over these judgments, as I set out below, and it is not possible simply to point to specific parts of either judgment as encapsulating a clear conclusion or a clear ratio. As a result, insofar as I analyse these judgments, I do so against the unavoidable backdrop of some of the major scholarship surrounding them. In this section, I distil two main points from Miller 1 and 2: only Parliament may give effect to significant constitutional change through legislation, and Parliament must be free of external restrictions to its legislative ability.

In Miller 1, the crux of the Secretary of State's case lay in the argument that treaty-making (and withdrawal) was an exercise of the Royal prerogative, rather than a power prescribed by statute.Footnote ⁴⁵ Thus, armed with the certainty of ‘Brexit means BrexitFootnote ⁴⁶ in the aftermath of a bitterly polarising referendum,Footnote ⁴⁷ the UK Government alone could initiate the process of terminating the UK's four-decade-old membership of the EU by giving notice under Article 50 of the Treaty on European Union (TEU). However, first the Divisional CourtFootnote ⁴⁸ and then the Supreme Court declared such a course of action to be unlawful and contrary to constitutional principle in the UK.

Both courts in Miller 1 placed the sovereignty of the Crown in Parliament at the heart of their reasoning, but in noticeably different ways. In the Divisional Court, the question whether the Crown could make a valid notification under Article 50 TEU by use of the prerogative turned on the purpose of the European Communities Act 1972 (ECA 1972). The Divisional Court held that the purpose of the ECA 1972 was to ‘introduce EU law into domestic law … in such a way that this could not be undone by exercise of Crown prerogative power’.Footnote ⁴⁹ This was said to be on the basis that the ECA 1972, by introducing EU rights into domestic UK law, covered the entire field of the domestic applicability and availability of those rights, including covering any action to extinguish the domestic availability of those rights by severing the link between the ECA 1972 and the EU Treaties (that is, through withdrawal from the EU).Footnote ⁵⁰ Summarising its reasoning, the Court stated:

In the Supreme Court, the most important question was whether the Crown could use its prerogative powers to effect a constitutional change as ‘significant’ and as ‘far-reaching’Footnote ⁵² as removing ‘EU law as an entirely new, independent and overriding source of domestic law’Footnote ⁵³ by initiating the withdrawal process. The Court, by a considerable majority, answered the question emphatically in the negative.

Whereas the Divisional Court upheld the sovereignty of the Crown in Parliament by giving effect to the purpose of the ECA 1972 (as it interpreted), the Supreme Court's understanding of sovereignty is less clear. For one thing, the majority in that Court did not set out what part of the ECA 1972 would be breached by the Crown's exercise of its prerogative powers. Two of the Court's conclusions loomed large: the first was that the ECA 1972 was the conduit rather than the source of EU rights in domestic law.Footnote ⁵⁴ The second was the acceptance of a metaphor put forward by Lord Pannick (for the respondents): that ‘[n]otification [under Article 50 TEU] is … the pulling of the trigger which causes the bullet to be fired, with the consequence that the bullet will hit the target and the Treaties will cease to apply’.Footnote ⁵⁵

However, the use of metaphors here demonstrates the difficulty in understanding how EU law figured within the UK constitutional framework, and indeed the reality that its existence domestically was simply accepted without trying to definitively understand its nature. Considering the ECA 1972 as a conduit implies that the statute is the ‘channel or medium by which [EU law] is conveyed’,Footnote ⁵⁶ with the implication that removing EU law by withdrawal (via the prerogative) merely takes away the thing conveyed rather than affecting the conduit in any way. This analysis featured prominently in Lord Reed's dissent in Miller 1:Footnote ⁵⁷

By contrast, in the majority's reasoning in Miller 1, EU law was held to be an ‘independent and overriding source of domestic law’ by operation of the ECA 1972 (as above). Mark Elliott has strongly criticised this conclusion, in large part because the majority clearly stated that the application of EU law domestically depended upon the operation of the ECA 1972, while simultaneously characterising it as an independent source of domestic law.Footnote ⁵⁹ Elliott offers two ways to explain the majority's reasoning – first, that the ECA 1972 itself established EU law as an independent source of domestic law, but in Elliott's analysis, this fails because in this scenario, the ‘independence’ of EU law would be meaningless if it could be snuffed out by repealing the ECA 1972.Footnote ⁶⁰ Secondly, the majority's characterisation of EU law as ‘independent’ is legally meaningless, and instead simply describes the ‘reality that the EU makes laws that have effects in the UK’.Footnote ⁶¹

A possibility not considered by the Court is that the precise nature of EU law in the domestic plane did not matter because, in order to be legitimate from a UK constitutional perspective, EU law had to be normatively compatible with the requirements of the UK constitution. Trevor Allan makes this point in his analysisFootnote ⁶² of the Factortame (No 2) case,Footnote ⁶³ as an answer to Sir William Wade's claim that UK membership of the EU had fundamentally changed the UK's rule of recognition.Footnote ⁶⁴ From Allan's perspective, no such change had occurred because the democratic credentials of EU law complied with the democratic foundations of the UK constitution, and thus its rule of recognition.Footnote ⁶⁵ Thus, with regard to EU law as it was available domestically, there is little normative difference between it and domestic law, regardless of whether the former is a form of domestic law, parachuted international law or some chimeric hybrid thereof. In any case, the use of the prerogative to interfere with the availability of the law would be normatively unconstitutional. The normative legitimacy of EU law within the UK constitution is an important point to which I return in the third section of this paper.

Whatever the analytical lens, it cannot satisfactorily pierce the fog of obscurity surrounding the majority's ratio in Miller 1. This is especially true of the majority's multiple references to the sovereignty of Parliament, which Elliott has criticised, observing:

This is, of course, true, and the Supreme Court's one-line conclusion on this pointFootnote ⁶⁷ neither illuminates its reasoning nor provides an answer to Elliott's critique. Nevertheless, the Supreme Court's ruminations on the ECA 1972 could be considered as being its conclusions on the purpose or effect of that Act: to prevent the UK's exit by use of the prerogative. This analysis is supported by the majority's rejection of the idea that the prerogative needed to be displaced only by express statutory provisions.Footnote ⁶⁸ To be clear, I am not saying that the Supreme Court concluded that this was in fact the purpose of the ECA 1972, only that, if prerogative powers need not be constrained only by explicit statutory text, then it stands to reason that they may be constrained by implication, including by statutory purpose. Despite the doctrinal obscurity apparent in Miller 1, the principal theme in the majority judgment of the Supreme Court was that significant and far-reaching constitutional change is properly left only to Parliament.

Miller 1 was handed down together with two devolution references from Northern Ireland: Agnew and others’ application for judicial review and McCord's application for judicial review. The Northern Ireland references were restricted to devolution questions, of which the Supreme Court answered one in some detail: whether the Northern Ireland Assembly (and by implication the Welsh Assembly and Scottish Parliament) had to consent before notice under Article 50 TEU could be given.Footnote ⁶⁹ In substance, this question asked whether the Sewel convention was judicially enforceable, requiring as a matter of law that the UK Parliament should obtain the consent of its devolved counterparts before legislating to terminate membership of the EU.Footnote ⁷⁰ The Supreme Court pointed to the political character of conventions in dispatching this question, remarking, ‘judges … are neither the parents nor the guardians of political conventions; they are merely observers’.Footnote ⁷¹ This was the only issue on which the Supreme Court was unanimous. Although the Supreme Court was correct on one level – the political nature of conventions rendering them unenforceable as a matter of law – the Supreme Court's observations on the Sewel convention having ‘an important role in facilitating harmonious relationships between the UK Parliament and the devolved legislatures’Footnote ⁷² has a particular significance to the post-Brexit arrangements in Northern Ireland as enacted by the UK Parliament. I return to this point in the third section below.

Miller 2, much like its predecessor, was no stranger to doctrinal obscurity. The first main point in that case which is important here is the Supreme Court's unanimous statement that the Crown's prerogative powers were forbidden from encroaching, not merely on the expressed will of Parliament (statute law) but also on Parliament's ability to express that will:

This point has perhaps been most extensively scrutinised in an illuminating exchange between Martin Loughlin and Paul Craig.

Loughlin, in his paper for Policy Exchange,Footnote ⁷⁴ asserted that parliamentary sovereignty ‘has always been understood to be a formal legal rule that grants supremacy to the laws enacted by the Crown in Parliament’, with the Supreme Court's understanding of it being an ‘attempt to transform a formal principle into a functional principle’ which ‘converts orthodoxy into heterodoxy’ and is ‘misconceived’.Footnote ⁷⁵

In his article,Footnote ⁷⁶ Craig answers Loughlin by pointing to the capacity of the doctrine of parliamentary sovereignty to express ‘capacity and constraint’.Footnote ⁷⁷ Mark Elliott largely echoes this view, asserting, ‘The principle of parliamentary sovereignty … is a fundamental principle that determines and reflects the nature of constitutional democracy in the UK’.Footnote ⁷⁸

Craig's (and Elliott's) reflections provide only a partial answer to Loughlin because of the context in which they must be understood. Invoking parliamentary sovereignty in the courts is useful when a breach of the doctrine can be remedied. An important distinction exists between the substance and effect of sovereignty and this distinction exists for a powerful reason: judicial intervention is possible in the latter, but not in the former.

Consider what Craig says in refuting the idea that parliamentary sovereignty exclusively concerns the supremacy of statute law:

It is indeed true that the discourse around the existence and scope of conditions on the exercise of parliamentary sovereignty has animated – and continues to animate – considerable judicial and academic discourse. For the purposes of argument, let us also accept Craig's rhetorical flourish at the end of the above passage as empirical fact. The reality remains that none of this discourse has translated into judicial intervention to provide relief in the event that any of the conditions on the exercise of parliamentary sovereignty are threatened. Craig (and the Supreme Court in Miller 2) lists some seminal cases as supportive of the Supreme Court's reasoning. In Proclamations, Cooke CJ famously declared, ‘the King hath no prerogative, but that which the law of the land allows him’.Footnote ⁸⁰ Proclamations was an advisory opinion, delivered by some of England's premier judges (including all three of its chief judges), on questions put to them by the Privy Council.Footnote ⁸¹ Although important, its principal points concerning the limitations of the prerogative relative to Parliament were not codified until the Bill of Rights nearly a century later, and only gained wide acceptance at that time.Footnote ⁸² Even so, Proclamations prohibited the use of the prerogative from changing the law as it existed, rather than from encroaching upon the law-making ability of Parliament. The other three authorities,Footnote ⁸³ namely De Keyser,Footnote ⁸⁴ Fire Brigades Union Footnote ⁸⁵ and Laker Airways Footnote ⁸⁶ all involved the use of prerogative powers to frustrate statute law (even in Fire Brigades Union, where the statute had been enacted but not yet brought into force).Footnote ⁸⁷ Craig acknowledges that the Supreme Court essentially went further than these cases by preventing the prerogative from frustrating the power of Parliament to enact statutes.Footnote ⁸⁸ However, he then appears to draw a specific line between these authorities and the Supreme Court's reasoning in Miller 2. Craig states that the prorogation affected ‘[Parliament's] capacity to exercise the totality of its legislative authority, and authority to scrutinise government action, thereby severely curtailing the opportunity for parliamentary voice on [Brexit]’,Footnote ⁸⁹ treating this as an attack on the normative foundations of parliamentary sovereignty (Parliament's democratic mandate),Footnote ⁹⁰ thus justifying the Supreme Court's approach.Footnote ⁹¹

Craig's approach is, with respect, unconvincing. Even assuming that the specific prorogation at issue in this case was a conclusive attack on the democratic foundations of Parliament, it is unclear why the Supreme Court needed to make the doctrinal leap from enforcing sovereignty through statute, to casting the enforceability net over parliamentary ability. This was because Parliament had already enacted legislation prohibiting a no-deal Brexit without parliamentary authorisation,Footnote ⁹² and the timing of the prorogation gave rise to a risk of precisely such an outcome, and thus the risk of the prerogative frustrating clear statutory purpose. Using the previously cited authorities to prevent such a risk would have been entirely orthodox. Instead, the Supreme Court's concern with the protection of Parliament's legislative ability completely glossed over the myriad complex processes which legislative ability entails.Footnote ⁹³ I return to this point in the third section of this paper.

In addition to Parliament's legislative authority, the prorogation was also said to impact upon its responsibility ‘for the supervision of the executive’,Footnote ⁹⁴ referencing the function of Parliament to hold the executive to account. Indeed, parliamentary accountability was characterised by the Court as being crucial enough to protect through judicial enforcement, lest a lengthy prorogation should undermine the democratic foundations of such accountability.Footnote ⁹⁵ Accountability is an unremarkable feature of the UK constitution – the executive, comprising many MPs elected only in that capacity, is collectively responsible to Parliament, which in turn holds the executive to account.Footnote ⁹⁶ The democratic foundations of such accountability are obvious.Footnote ⁹⁷ However, Miller 2 marks the first time (implicit in the Court's own judgment)Footnote ⁹⁸ that accountability has been held to be a legally cognisable function of Parliament capable of judicial protection. This aspect of Miller 2, although precedentially novel, is clearer than the Supreme Court's ruminations on parliamentary sovereignty. This is for two main reasons: first, unlike with its reasoning concerning parliamentary sovereignty, the Supreme Court was clear that it was expanding or developing the law. Accountability having been previously invoked in order to justify judicial restraint, it was applied in Miller 2 because of the risk that the bypass of parliamentary scrutiny would seriously impact domestic law should the UK crash out of the EU without sufficient time for parliamentary scrutiny of such a drastic step (which, by all accounts was a real risk).Footnote ⁹⁹ Secondly, given this risk, the circumstances of the prorogation were extraordinary.Footnote ¹⁰⁰

Richard Ekins asserts that prorogation may only be extraordinarily refused by the monarch, but that otherwise the power to prorogue is uncontrollable by law.Footnote ¹⁰¹ Viewed in this way, accountability is a principle limited to the purely political field and not cognisable by the courts. But this is, with respect, a narrow view on how accountability functions in practice. Far away from the theatre of a Government defeat, everyday accountability in the Houses of Parliament involves debate, discussion and questioning by its members. The importance of MPs airing the views and concerns of their constituents to a government which has no democratic legitimacy independent of the Commons cannot be overlooked.Footnote ¹⁰² Thus, even if Parliament would not deprive the Government of office (as in fact it would not, following its return from the prorogation-that-wasn't) it would still hold the Government to account in the wider, more everyday sense. The reason why this accountability was legally relevant is explained in the Supreme Court's reasoning:

The Supreme Court thus concluded that there was a risk that the executive was embarking on constitutional change for which it had no support from Parliament. Further, although the Supreme Court did not comment on the matter, the Prime Minister had, by the time judgment had been handed down in Miller 2, commented publicly about his intention to defy the requirements of the Bill that would become the 2019 Act.Footnote ¹⁰⁴ Accordingly, the need for the Prime Minister to face Parliament was a matter of constitutional importance, with the spectre of the prerogative being used to cause drastic legal change in plain and defiant frustration of statute law, looming large.

A criticism of this view might be that even if a risk existed of the prerogative being used in the aforementioned manner, risk does not always translate to reality and the Court should not have become embroiled in a hypothetical question. This is a criticism Loughlin makes, going as far as suggesting that the Court's engagement with hypothetical questions ‘influence[s] the manner of presentation of constitutional narratives, which in turn shape[s] and even determine[s] constitutional requirements’.Footnote ¹⁰⁵ There are two answers to this: first, refusing to determine an issue for the reason that it has not come to pass is itself engagement with a hypothetical question: that the issue in question may not come to pass.Footnote ¹⁰⁶ Secondly, the risk in this case, while not pinpointed with mathematical precision, was certainly more than fanciful, while the nature of the risk (drastic constitutional change unauthorised by Parliament) was unprecedented. In such circumstances, the Supreme Court was perfectly justified in sidestepping doctrinal puritanism without sacrificing clarity of reasoning.

In Miller 2, like in Miller 1 before it, the Supreme Court was concerned with allowing Parliament to decide whether and how to enact far-reaching constitutional changes. The significant addition made by Miller 2 was its focus on the democratic credentials underlying Parliament's ability to make choices and its insistence that Parliament, as the only democratically legitimate forum, had both a right and a function to scrutinise such choices properly.

From exploring Miller 1 and 2, I distil two main points against which to analyse the impact on Northern Ireland of the incorporation of the Protocol. First, that significant constitutional change is a parliamentary function par excellence; and secondly, that Parliament has an uncurtailable role in scrutinising the actions of the Government. Grounding both points are Parliament's democratic credentials.

3. The impact of the Protocol's incorporation on parliamentary sovereignty

In this section I explore the way in which the incorporation of the Protocol into domestic law threatens the normative foundations of parliamentary sovereignty when viewed through the lens of Miller 1 and Miller 2.

I begin with the functioning of the Joint Committee of the UK and the EU established by the terms of the Withdrawal Agreement, explored in the first section. Legislative prescription of UK ministers’ functions in the Joint Committee is, for the most part, non-existent. The Joint Committee has several enumerated functions,Footnote ¹⁰⁷ one of which clearly relates to the Protocol, including its stated objective in addressing issues which arise on the island of Ireland as a result of Brexit. Moreover, any Joint Committee decision to accept new EU law results in such law automatically becoming part of the ProtocolFootnote ¹⁰⁸ and thus, the Withdrawal Agreement.Footnote ¹⁰⁹ Even if Articles 5–10 of the Protocol should cease to apply as a result of the Northern Ireland Assembly voting to confirm such cessation, the rest of the Protocol, including its wide objective and EU laws which fall within the scope of this objective (even those which fall outside the scope of Articles 5–10) carries on regardless.Footnote ¹¹⁰ The upshot of this is that UK ministers, as part of Joint Committee decision-making, may continue to add significant new EU laws to Northern Ireland's statute book, without seeking any specific authorisation from Parliament, in the complete absence of any legislative provision requiring such authorisation (whether positively or by implication). And nor can the Northern Ireland Assembly, ministers or Departments make laws which control or otherwise prescribe this process of adding new law without coming up against the hurdles previously set out. All of this resurrects the spectre that the Supreme Court emphatically worked to exorcise in Miller 1: significant or far-reaching constitutional change without parliamentary authorisation.

Now, there are two possible answers to the above problem: first, that Miller 1 concerned the use of prerogative powers to change statute law and this principle is inapplicable in the present context. This is because the Withdrawal Agreement, which establishes the Joint Committee and authorises it to add EU law which it accepts as falling within the scope of the Protocol, was statutorily endorsed by Parliament. While that is true on a general level, this argument does not provide a satisfactory answer to the lack of statutory prescription for the exercise of ministerial powers in the Joint Committee. While Miller 1 was markedly imprecise on how prerogative powers would frustrate statute law, one could arguably infer such frustration by implication (as in the first section of this paper). In the context of the EUWA 2018, it is not possible to make such an inference, because while the purpose of the that Act was to provide for the Withdrawal Agreement to have effect in domestic law, the Withdrawal Agreement itself contains multiple purposes within its overall purpose (Brexit), one of which, being the circumstances on the island of Ireland, finds no equivalent expression in the EUWA 2018. Section 10 of the EUWA 2018 is no satisfactory answer to this conundrum, because it is concerned with keeping intact North-South arrangements as they existed prior to Brexit,Footnote ¹¹¹ whereas the Protocol looks decidedly to the future. Thus, in the absence of a statutory provision which occupies the field of the UK's involvement in the Joint Committee vis-à-vis the Protocol, these powers must be located in the foreign relations prerogative. The alternative conclusion, that the EUWA 2018 permits UK ministerial action to expand the Northern Ireland statute book without legislative involvement, would be tantamount to Parliament having enacted arguably one of the most significant qualifications to its own sovereignty, if viewed through the lens of Miller 2 and its refusal to countenance executive attempts to restrict Parliament's legislative ability.

We therefore face at least a similar, if not the same, situation in this context as the Supreme Court faced in Miller 1: the prerogative making significant changes to the Northern Ireland statute book without parliamentary approval for the same. While it may be thought that another Brexit-sized constitutional change is unlikely, the concerning change is not in what the Joint Committee might decide – the change is that the Joint Committee decisions have automatic effect (without UK legislative scrutiny) at all. This is a crucial point: the direct effect of new EU law in Northern Ireland under the Protocol differs normatively from its effect under the EU Treaties (that is, during the UK's membership of the EU). I return to this point in greater detail.

It might be thought that the Supreme Court's ‘reasonable justification’ test in Miller 2 Footnote ¹¹² may provide an answer to this problem. Applied here, it would allow the expansion of the statute book by prerogative action alone if there was reasonable justification for such a step. But what is the reasonable justification? No part of the Protocol requires UK ministers in the Joint Committee to avoid parliamentary scrutiny when making decisions as part of that committee. Nor are such decisions inherently of a character which could not countenance parliamentary involvement through debate and scrutiny. On the contrary, Joint Committee decisions which expand the Northern Ireland statute book are legislative decisions par excellence.

The second answer lies in the scant democratic ‘controls’ as part of the Protocol implementation process. According to the Northern Ireland High Court, these controls lie with the Joint Committee, in which the UK Government plays a ‘full role’,Footnote ¹¹³ the Assembly with its democratic consent mechanismFootnote ¹¹⁴ and Parliament itself.Footnote ¹¹⁵ However, as set out above, these scarcely provide a satisfactory answer when, although Parliament and the Assembly are democratic in their mandates, neither institution is able to exert any real control over the potentially large expansion of the Northern Ireland statute book. While a related point would be Parliament's sovereign ability to legislate in breach of the Withdrawal Agreement (as a matter of domestic law), perhaps to bar UK Ministers from agreeing the implementation of any further EU law, such legislation may invariably be disapplied in the same manner as in Factortame, because the Withdrawal AgreementFootnote ¹¹⁶ and section 7A of the EUWA 2018 demand such a step.

The normative critique of Parliament's incorporation of the Protocol arises precisely because of the manner of this incorporation: that in enacting provisions which effectively sidestep parliamentary scrutiny with ease, the resultant democratic deficit risks placing the normative foundations of the UK's rule of recognition in jeopardy.

There is no universally accepted normative foundation for the rule of recognition that the Crown in Parliament is the highest source of law in the UK. However, democracy plays a large part in such a foundation. Craig states ‘the modern rationale is that the democratically elected Parliament represents the will of the people, with the consequence that there should be no judicially enforceable limits to its power’.Footnote ¹¹⁷ Jeffrey Goldsworthy provides an expanded explanation:

It is worth noting that democracy as the standout (if not necessarily sole) foundation for legitimating the legal supremacy of the Crown in Parliament is reflected in several judicial decisions of the highest authority, both recentFootnote ¹¹⁹ and less recent.Footnote ¹²⁰ Thus, despite the caution expressed on this point,Footnote ¹²¹ my analysis here proceeds by examining the democratic impact of Protocol incorporation, when assessed against the democratic mandate which was said to characterise Brexit.

Jo Murkens observes that the character of democracy at the heart of parliamentary sovereignty is decidedly procedural, rather than substantive.Footnote ¹²² At no other time in recent memory has this point been laid absolutely and painfully bare than during and after the Brexit referendum. Concerns about the legitimacy of the referendum,Footnote ¹²³ a plethora of possible outcomes, a Parliament which failed to agree a withdrawal agreement three timesFootnote ¹²⁴ and unconvincing uses of idioms involving cakeFootnote ¹²⁵ were all swept aside by the political juggernaut of the referendum representing the will of the UK electorate and the duty on the Government and Parliament to deliver on its result.Footnote ¹²⁶

However, the centrality of the referendum result in the justification behind the biggest constitutional change in the UK in generations cements the position of (some version of) democracy in the normative foundation of parliamentary sovereignty, at least in relation to Brexit. In short, the people voted to leave the EU and Parliament enacted that decision into law, thereby legitimating its position as the highest source of law in the eyes of the people. In its legislative choices on the implementation of the Protocol, however, it is this very democracy that Parliament has risked diminishing. This is plain when considering the Joint Committee.

The democratic credentials of Joint Committee decision-making are guaranteed, according to the Northern Ireland High Court, by the role that the UK Government plays in the Committee (as above). Now, the UK Government has no democratic mandate independent of the House of Commons, to which it is responsible. This practice of responsible government is said to be indicative of democratic legitimacy in the Westminster style of governance.Footnote ¹²⁷ However, in circumstances where party discipline and party whips mean that the role of an MP is sometimes to merely support Government measures,Footnote ¹²⁸ the idea of scrutiny as a significant indicator of democratic legitimacy can be reduced to little more than a paper tiger.Footnote ¹²⁹ It is in this context that the complete absence of a statutory role for parliamentary scrutiny is even more concerning: Government business dominates the Commons by its own rules of procedure, meaning that unless the Government is under an express statutory duty to provide for MPs to debate and vote,Footnote ¹³⁰ it may choose not to allocate any time for such a process. Admittedly, while 20 days are allocated to opposition business, the order of business on those days depends entirely on the Leader of the Opposition and the leader of the second largest opposition party,Footnote ¹³¹ neither of whom are Northern Ireland MPs. Thus, an issue of democratic deficit concerning Northern Ireland may be included in the order paper if it accords with the political priorities of parties outside Northern Ireland. This reality accords with the experience of insufficient attention given by Westminster to Northern Ireland business.Footnote ¹³²

It may be said that, in enacting section 7A of the EUWA 2018 (as amended), the 2020 Parliament merely did what the then Parliament did in respect of the ECA 1972 – that is, laws adopted under the Protocol (and thus the Withdrawal Agreement) would have effect without any further enactment. However, this is only a superficially attractive point. Membership of the EU did not turn the UK into a ‘rule-taker’ with no seat at the tables in Brussels and Luxembourg where the rules were made. UK residents voted members to the European Parliament,Footnote ¹³³ UK Ministers were members of the Council of the European UnionFootnote ¹³⁴ and the UK Government recommended Commissioners to the European CommissionFootnote ¹³⁵ to be proposed by the Council of the European Union and voted into office by the European Parliament.Footnote ¹³⁶ Finally, the overarching direction of travel of the EU was decided at the European Council,Footnote ¹³⁷ which included the UK Prime Minister among its members.Footnote ¹³⁸ While this was by no means a perfect system and suffered from criticisms of democratic deficit,Footnote ¹³⁹ the democratic credentials of governance at Westminster are not perfect either. More to the point, there was never a requirement (from a UK constitutional perspective) of EU membership to produce perfect democratic credentials in order to be democratically legitimate. If democratic credentials support the authority of the Crown in Parliament as the highest source of law in the UK, then that authority is surely reinforced when the European law to which it gives effect is itself backed by democratic credentials which are comparable to those in the UK.Footnote ¹⁴⁰ The significant change ushered in by the implementation of the Protocol is the future incorporation of EU law without even these democratic credentials. This is why Joint Committee decisions having automatic effect under the terms of both the Protocol and section 7A of the EUWA 2018 is a sizeable and significant constitutional change in itself, the significance and impact of which is not lessened by pointing to the enactment of the EUWA 2018.

There is no prior scrutiny by UK or Northern Ireland authorities of EU legislative proposals before they are enacted (of course, there could not be, in light of Brexit). The body with the capacity to scrutinise new EU law with a view to determining whether it should be added to the Protocol and thus have effect in domestic law is the Joint Committee. Neither the Protocol nor any UK statute provides for any mechanism by which the Joint Committee may consult Stormont authorities in respect of these decisions. In respect of the first known potential addition to the Protocol, it was a committee of the Westminster Parliament that enquired if and how such an addition would be made.Footnote ¹⁴¹ Post-addition scrutiny of new EU law by Parliament is not guaranteed, and even if guaranteed may not legally stem, far less reverse, the tide, due to the likelihood of disapplication of statutory attempts to restrict the effect of the Withdrawal Agreement in domestic law.

In respect of the Northern Ireland Assembly, the situation is no better. First, it is statutorily barred from modifying the effect of the Withdrawal Agreement in domestic law (as above). Secondly, although the Assembly may vote to end the application of certain parts of the Protocol, its actions have no consequence on the remaining parts of the Protocol. Thirdly, in respect of exercising its consent vote, the Assembly is given opportunities to do so every four or eight years, depending on the outcome of the previous such vote.Footnote ¹⁴² Whether four years’ worth of potential expansions of the Northern Ireland statute book can be given due attention by the Assembly during a period which could be as long as almost six weeksFootnote ¹⁴³ and as short as less than a full dayFootnote ¹⁴⁴ remains to be seen. Finally, the effective carte blanche with which the Joint Committee may add new EU law to the Protocol is itself concerning from a normative perspective. This is because, as previously examined, unlike section 2(1) of the ECA 1972, section 7A of the EUWA 2018 gives effect to a treaty with an expansive potential, unsupported by democratic credentials for such expansion (from a UK perspective). Parliament's incorporation of such a treaty therefore differs normatively from its having given effect to EU law while the UK was a Member State.

The relevant context here is not treaty incorporation, but Brexit, with its foundations in democratic legitimacy. Driven by the conviction that Brussels constrained the democratic choices of the UK electorate while the UK was a Member State, the referendum was the chance to ‘revive democracy in Britain’.Footnote ¹⁴⁵ If the will of the people therefore necessitated exiting the EU, then the replacement of EU membership with a law-making process which is in fact undemocratic is normatively jarring. It is especially jarring when considered against the statutory backdrop to EU membership: successive statutory restrictions on UK ministers agreeing changes to the EU treaties, especially those which expanded the EU's competences relative to Member States, in the absence of specific parliamentary authorisation and referendums,Footnote ¹⁴⁶ pointing to the increasing importance of democratic authorisation which Parliament had built into the UK's relationship with the EU.

Returning briefly to the Supreme Court's discussion of the Sewel convention in Miller 1 (from the first section of this paper), it is important to recall that none of the devolved legislatures granted their consent in respect of the European Union (Withdrawal Agreement) Act 2020, which incorporated the Withdrawal Agreement (and Protocol) into domestic law.Footnote ¹⁴⁷ In the context of the normative foundations of parliamentary sovereignty residing in democratic legitimacy (as explored earlier), the manner of this incorporation has significant consequences. At the time of Miller 1, Brexit was going ahead in spite of democratic wishes to the contrary in Scotland and Northern Ireland.Footnote ¹⁴⁸ In 2020, the new withdrawal arrangements were enacted in spite of even greater (albeit indirect) democratic rejection. As Aileen McHarg notes, democratic demand has a particular resonance in the context of the operation and role of the Sewel convention:

The centrality of democratic legitimacy in the relationship between Westminster and the devolved administrations was absent in the Supreme Court's discussions of the Sewel Convention in Miller 1, and effectively threatened at the point of exiting the EU. Over its express objections, Northern Ireland lost considerable say (through the Assembly, Westminster and the European Parliament) in the laws to which it is subject, in the final act of a geopolitical saga which its residents had already rejected four years earlier. In such circumstances, pointing to the fact of Parliament having incorporated the Protocol effectively empties its sovereignty of considerable democratic, and thus normative, value. Devoid of such normative value, what remains to justify the authority of the Crown in Parliament as supreme?

From the perspective of Northern Ireland authorities, the consequence is considerable. The Joint Committee, in which Northern Ireland authorities play no explicit part,Footnote ¹⁵⁰ provides no effective answer to the glaring problem of laws being made for Northern Ireland, with no real (even indirect) democratic input from Northern Ireland. The idea that Parliament will function effectively as the democratic forum in which to scrutinise such issues unravels when one considers that Northern Ireland elects 18 MPs to a 650-member House, and of those 18 MPs, 7 do not take their seats. This is even before the problem identified above, of Government control of House business, so that truly effective scrutiny may well be at the pleasure of the Government. The impact this has on the principle of parliamentary accountability, as the Supreme Court held in Miller 2, is concerning. If the principle is placed in jeopardy by a five-week shutdown of parliamentary business, how much more is it impacted in respect of a process of law-making that largely avoids Parliament altogether?Footnote ¹⁵¹

Now, the impact of the Protocol's incorporation goes beyond creating a democratic deficit. Democratic deficits are concerning from a normative perspective, but do not per se threaten the sovereignty of Parliament. This is because the law does not recognise restrictions on the ability of the Crown in Parliament to enact law, regardless of its compositionFootnote ¹⁵² or the regularityFootnote ¹⁵³ or recentnessFootnote ¹⁵⁴ of its popular legitimacy. However, as the Supreme Court held in Miller 2, parliamentary sovereignty is threatened if Parliament's ability to make law is threatened by external factors. It is at this juncture that we return to the Protocol's dynamism, which centres around two matters: the Joint Committee and new EU legal provisions which replace or amend those listed in the Protocol.Footnote ¹⁵⁵ First, let us be clear that the Joint Committee is a law-making body for the UK, albeit that its ability to make law is in respect of Northern Ireland. Secondly, there are categories of new EU law which have automatic effect without Joint Committee involvement.

The Joint Committee is an unprecedented anomaly in the understanding of parliamentary sovereignty. Two classic statements of principle are applicable here: treaty-making and treaty-fulfilment in the international plane are prerogative functions beyond the purview of municipal courts,Footnote ¹⁵⁶ and treaties are not self-executing, in that they must be incorporated by statute to be enforceable by municipal courts.Footnote ¹⁵⁷ The upshot to both principles is that the prerogative to make and fulfil treaties in the international plane does not alter domestic law. However, what if the incorporation of a treaty left considerable room for the prerogative to do exactly that? The EUWA 2018 (as amended), by remaining completely silent on how UK ministers may exercise their functions as part of Joint Committee decision-making, gives rise precisely to this risk. The democratic deficit is at its most glaring when looking at new EU law which applies automatically without Joint Committee involvement. This is new law in which the post-Brexit UK plays absolutely no decisional role, either in content or in applicability. And nor are there any restrictions on this aspect of the Protocol contained in the text of the Protocol itself (or in the text of the EUWA 2018), in relation to the scope of amendment or replacement. This is why the treatment of the democratic deficit arguments in Allister is glaringly cursory: what does a court do when faced with changes to domestic law over which Parliament has no control, and neither does the Government which holds office with the confidence of the Commons? It is important to recall that, although the increasing trend towards a legal understanding of parliamentary sovereignty as a substantive doctrineFootnote ¹⁵⁸ is not anchored to a particular understanding of democracy, it is nevertheless anchored to democracy.Footnote ¹⁵⁹ To that extent, appreciating the reality of the EUWA 2018 – the almost complete removal of democratic input into the laws to which Northern Ireland will be subject – subverts this understanding of parliamentary sovereignty. The Allister litigation provides worrying insights in this regard: two courts providing three different ways of explaining the effect of the EUWA 2018,Footnote ¹⁶⁰ none of them convincing enough to be adopted generally.