The background
In February 2021 judicial review proceedings challenging the lawfulness of the Northern Ireland Protocol were brought against the Prime Minister, the Secretary of State for Northern Ireland, and the Chancellor of the Duchy by Mr Clifford Peeples. In March 2021 similar judicial review proceedings challenging the lawfulness of the Protocol were brought by a high-profile group of cross-party unionists including present and former leaders of the TUV, the UUP and the DUP among others. Both sets of proceedings were heard together and dismissed together by Colton J in the Northern Ireland High Court on 30 June 2021 (see Allister and others [2021] NIQB 64). Following appeal, both proceedings were again heard together and dismissed together in the Court of Appeal in Northern Ireland on 14 March 2022 (see Allister and others [2022] NICA 15). The appellants in the cojoined cases were then granted leave to appeal to the UK Supreme Court regarding the conclusions of the lower courts in three of the five questions posed in the initial proceedings.
The arguments of the appellants – cross-party unionist representatives – and the respondents – UK Government representatives – in Allister and others were heard by the Supreme Court on 30 November and 1 December 2022. Judgment was handed down on 8 February 2023. The case was unanimously dismissed on all three grounds of appeal.
Embracing constitutional orthodoxy, the Supreme Court in Allister find the Protocol to be lawful without caveat; in this respect, the ruling was unsurprising. That said, the inherently conventional reasoning in the Allister judgment belies its potential significance for watchers of the ever-evolving UK constitution. As this article suggests, attending to the questions that were unanswered by the Court in Allister, regarding ‘constitutional statutes’ and on some specifically Northern Ireland issues, is almost as interesting as attending to those answers it did provide.
The substance
In the Supreme Court the lawfulness of the Protocol was challenged on three grounds: its alleged incompatibility with Article VI of the Acts of Union 1800; its alleged incompatibility with section 1(1) of the Northern Ireland Act (NIA) 1998 regarding the principle of consent; and its alleged incompatibility with section 42 of the NIA 1998 regarding cross-community consent.
Ground one: Acts of Union
Article VI of the Acts of Union 1800 which made provision for the Union of Great Britain and Ireland states that the subjects of Great Britain and Ireland shall:
“…be entitled to the same privileges and be on the same footing, as to encouragements and bounties [the trade limb] … and that in all treaties made by [His] Majesty, his heirs, and successors, with any foreign power, [His Majesty’s subjects of Ireland shall have the same privileges and be on the same footing as [His] Majesty’s subjects of Great Britain [the treaty limb]”
The appellants and the Court distinguished between ‘two distinct limbs’ of Article VI, that concerning trade and that concerning treaties. On the first – the trade limb – the appellants argued that because goods entering Northern Ireland from Great Britain are required to follow EU rules rather than UK rules the “same footing” guarantee has been undermined. On the second – the treaty limb – appellants argued that this aspect of Article VI imposes a restriction on the exercise of prerogative powers to negotiate treaties that are contrary to the prior ‘same footing’ provision. On these dual premises, the appellants posited that section 7A of the European Union (Withdrawal) Act (EUWA) 2018 which provides for the domestic implementation of the Protocol is incompatible with Article VI and is, therefore, unlawful.
Addressing the trade limb argument, – the Court judged that the Protocol had “modified” in part the ‘equal footing’ provision in Article VI of the Acts of Union (Allister and others, 2023 [67]) but not repealed it, moreover the relevant statutory language in section 7A of EUWA 2018 was clear, purposive, and therefore took precedence over any earlier enactments, including the 1800 Acts (ibid [68]). Specifically, the Court noted that section 7A(3) of the 2018 Act ‘clearly’ states that “[every] enactment is to be read and has effect subject to subsection (2)” where provision is made for “all such … obligations and restrictions from time to time created or arising by or under withdrawal agreement [including the Protocol]” to be recognised and available in domestic law and to be enforced, allowed and followed accordingly. The language of section 7A of the 2018 Act therefore provided, in the Court’s view, a “clear answer” to the question put by appellants: any potential conflict with Article VI of the 1800 Acts and section 7A of the 2018 Act is to be resolved in favour of the later Act. Addressing the treaty limb argument, the Court rejected the appellants contention that the Protocol was made in contravention of a restriction on the use of the prerogative power on the basis that the sovereign UK parliament had authorised the exercise of prerogative powers in this matter when it approved the European Union (Withdrawal Agreement) Act 2020 which amended to EUWA 2018 to make provision for the implementation of the agreement made by the government with the EU [73-76]. The UK-EU treaty containing the Protocol was therefore lawfully made.
In dismissing both ‘limbs’ of the appellants arguments related to Article VI of the 1800 Acts [79] the Supreme Court drew upon and affirmed “the most fundamental rule of UK constitutional law” namely that “the Crown in Parliament is sovereign, and that legislation enacted by Parliament is supreme” [66] – a textbook example of Diceyan orthodoxy. Importantly, in deliberating on ground one of the challenge, the Court opted not to expound on the status and interplay of so-called ‘constitutional statutes’ (as per Thoburn v Sunderland City Council [2002] EWHC 195 (Admin)) in the UK setting although it could have chosen to do so.
In the landmark Thoburn ruling in the High Court of England and Wales, a distinction was made between ‘ordinary statutes’ and ‘constitutional statutes’. There, a procedural differentiation was drawn between the former which could be “impliedly repealed” [37-73] and the latter which could only be expressly repealed. Laws LJ in Thoburn opined that constitutional statutes could only be repealed by “express words in the later statute, or by words so specific that the inference of an actual determination to effect the result contended for was irresistible” [110]. Although the Supreme Court in Allister did not deviate from the effect of Thoburn reasoning – citing the use of “express words” in EUWA 2018, to which the 1800 Acts, along with all other enactments are ‘subject to’ – it also explicitly opted not to engage with the concept of constitutional statutes, instead describing the matter as “academic” in the context [66].
Ground two: principle of consent
Section 1(1) of the NIA 1998 provides that “Northern Ireland in its entirety remains part of the United Kingdom” and this “shall not cease to be so without the consent of a majority of the people of Northern Ireland” voting in a dedicated poll. Continuing, section 1(2) of the NIA 1998 states that if a poll is held under its terms and it is the “express wish” of a majority that “Northern Ireland should cease to be part of the United Kingdom and form part of a united Ireland” then the UK government will lay proposals to give effect to that wish. This first section of the NIA 1998 implements commitments made in the first part of the Belfast (‘Good Friday’) Agreement of 1998 – these are often referred to as the ‘principle of consent’.
Appellants in Allister submitted that section 1(1) of the NIA 1998 does not only regulate the specific type of change in constitutional status noted in section 1(2) but has a wider meaning such that it protects against any diminution of the status of Northern Ireland save from that which has been approved in advance by a poll held in accordance with provisions of NIA 1998 (s1 and Sch 1). By, in effect, creating a customs border for goods moving from Great Britain to Northern Ireland the Protocol, in the appellants view, brought about a diminution in the status of Northern Ireland which ought to have been subject to and conditional upon securing of majority consent of the people of Northern Ireland under the terms of section 1(1) of the NIA 1998. Accordingly, the appellants submitted that section 7A of the 2018 Act was unlawfully made.
In deliberating on the scope and potential applicability of section 1(1) of the 1998 Act to section 7A of the 2018 Act, the Supreme Court relied on the reasoning of the same in Miller 1 [2017] (UKSC 5). When asked a similar question regarding the potential applicability of the ‘principle of consent’ provided for in section 1(1) of the NIA 1998 – this time in relation to the triggering of Article 50 TEU – it was the unanimous view of the Supreme Court in Miller 1 that the relevant provisions:
“…gave the people of Northern Ireland the right to determine whether to remain part of the United Kingdom or to become part of a united Ireland. It neither regulated any other change in the constitutional status of Northern Ireland nor required the consent of a majority of the people of Northern Ireland to the withdrawal of the United Kingdom from the European Union” [135]
Affirming the earlier judgment, the Supreme Court in Allister found “no reason to depart from the decision in Miller 1 which is dispositive of this ground of appeal” [84] the appellants argument on ground two was thus dismissed. This outcome is demonstrative of the doctrine of judicial precedent and not, therefore, that surprising. There are, however, two consequential points worth making.
Unlike appellants in Miller 1 in Allister no reference was made by those bringing the latter case to the international law origin of NIA 1998 section 1. In Miller 1 the question asked was for the domestic law provisions for the exercise of consent to be considered “together with” the “Belfast Agreement and the British-Irish Agreement” [126(i)]. However, in its judgment, the Supreme Court declined to engage the international law dimension emphasising instead, the “relevant commonality of the devolution settlements” across the UK [128]. While the ‘dispositive’ conclusion of ground two Allister is a logical extension of Miller 1 it is also worth underlining that in neither instance did the Supreme Court opine on the constitutional significance or otherwise of the unique international law origins of section 1(1) of the NIA 1998. A final noteworthy point related to ground two in Allister is the simple fact of its being brought by a group of high-profile, cross-party unionist political representatives. In Northern Ireland the ‘principle of consent’ has been interpreted both narrowly and broadly since it was agreed in 1998. While the Supreme Court in Allister affirmed a narrow, legal interpretation, the appellants arguments in Allister (and incidentally in Miller 1) underline the current political salience of a broader, political interpretation of what the consent principle does or ought to mean in Northern Ireland.
Ground three: cross-community consent
The third ground of appeal concerned a specific article of the Protocol – the democratic consent mechanism – which is provided for in Article 18 of the text. Under its terms, Members of the Northern Ireland Assembly (MLAs) are to be given the opportunity to vote in favour or against the continued application of Articles 5 to 10 of the Protocol in 2024 and at regular intervals thereafter. According to Article 18, if a simple majority of MLAs vote in favour, then another vote will be held 4 years later, but if a cross-community majority of MLAs vote in favour, then another vote will be held 8 years later. What this means is that continuation is not subject to cross-community consent, although it does allow it to have effect on the frequency of the exercise.
Appellants in Allister contended that when the Secretary of State made provision in UK law for the ‘democratic consent mechanism’ to take place they did so unlawfully because this required the setting aside of the ‘petition of concern’ mechanism in section 42 of the NIA 1998 which ordinarily enables decisions in the Assembly to be made subject to cross-community consent if enough MLAs deem this necessary. A case was made by the appellants that, insomuch as they set aside the need to secure cross-community consent, the 2020 Regulations were incompatible with the NIA 1998 and thus violated a provision made elsewhere in the EUWA 2018 (under which they were made) that requires all measures taken under the Act to be compatible with the NIA 1998.
On this issue, the respondents in Allister – the UK government – submitted that the continued application of Articles 5 to 10 of the Protocol is an international relations matter. It is, therefore, explicitly ‘excepted’ and not within the competence of the devolved legislature in Northern Ireland (see NIA 1998, Sch 2(3)), for this reason, the ‘petition of concern’ mechanism for securing cross-community consent on devolved issues, ought not to apply to ‘democratic consent mechanism’ votes in the Assembly and thus the 2020 Regulations were lawfully made. On the question of competence, the appellants in Allister pointed out that section 42 of the NIA 1998 uses clear yet broad language stating that the petition of concern mechanism applies to “a matter which is voted on by the Assembly” (s42(15)). They argued therefore that the possibility to initiate a cross-community consent vote under this section ought to apply to all votes in the NI Assembly regardless of whether or not these are within its competence.
Addressing the issue of cross-community consent, the Supreme Court acknowledged the “potential force of the appellants’ argument that section 42” of the 1998 Act applies to any matter voted on in the NI Assembly regardless of competence [107]. Yet the Court determined that it was “not necessary” to conclude whether or not this argument is correct because, in their view, the supposed ‘incompatibility’ between the 2020 Regulations and section 42 the NIA 1998 did not arise on the basis that the latter had already been changed by section 7A of the EUWA 2018 – to give effect to the Protocol and all related provisions – prior to the making of the 2020 Regulations [ibid]. The Supreme Court therefore dismissed the third ground of appeal and ruled the ‘democratic consent mechanism’ regulations lawfully made.
The Supreme Court’s dim view of the necessity of deciding whether or not all NI Assembly votes ought to be subject to section 42 of the NIA 1998 is perhaps disappointing for those seeking legal and procedural clarity in Northern Ireland. It is worth noting, though, that the scope of the section 42 NIA 1998 ‘petition of concern’ (PoC) mechanism has been considered by the NI Assembly and Executive Review Committee. As part of the PoC Review, Northern Ireland political parties expressed perspectives on whether or not cross-community consent provisions should be restricted to only certain types of legislation (ibid: 8) although no consensus was reached on the matter. As it stands, the legal scope of section 42 of the NIA 1998 is, therefore, unclear.
The implications
The Allister judgment is, in broad terms, constitutionally orthodox. In this regard it could be interpreted as part of a general swing of the Supreme Court back to ‘orthodoxy’ in recent times and particularly since the appoint of Lord Reed as its President in early 2020.
Yet, notwithstanding its relationship to broader trends in domestic jurisprudence, the specific orthodoxy of Allister could be said to belie the importance of this decision for the constitution and understandings of it. The reasoning of the Supreme Court affirmed the conclusions of earlier Allister rulings by determining that Article VI of the Acts of Union had been ‘modified’ but not repealed [67] by section 7A of the EU Withdrawal Act 2018; and it reasserted parliamentary sovereignty as the ‘most fundamental rule’ of the UK constitution. The judgment did not engage with questions regarding ‘constitutional statutes’ deeming this “academic” in the context. Given the current (strained) state of the UK union, the status and significance of constitutional statutes may come before the Supreme Court again, this time presumably on a reference with Scottish origins. From a Northern Ireland perspective, while the Supreme Court in Allister (re)affirmed a narrow interpretation of the principle of consent in section 1(1) of the NIA 1998, it did so having not been asked to engage with the potential relevance of the broader expression of consent in the opening section of the 1998 Agreement. Further, the Court opted not to clarify the potential scope of section 42 of the NIA 1998 regarding cross-community consent. Future caselaw mayforce clarification on this issue.
In legal terms, the ruling of the Supreme Court in Allister is orthodox and unsurprising. In political terms, for Northern Ireland, and more specifically for the unionist politicians who brought the cases, it represents the closing of one means of challenging the Protocol while leaving many related concerns unresolved.
Lisa Claire Whitten.
Dr Lisa Claire Whitten is a Research Fellow on the ESRC-funded project ‘Governance for ‘a place between’: the Multi-Levelled Dynamics of Implementing the Protocol on Ireland / Northern Ireland’ based at Queen’s University Belfast.
The Constitution Society is committed to the promotion of informed debate and is politically impartial. Any views expressed in this article are the personal views of the author and not those of The Constitution Society.