Between 14 and 16 October 2025, the UK Supreme Court heard the appeal of the Secretary of State for Northern Ireland (‘SOSNI’) against the Dillon decision of the Northern Ireland Court of Appeal (‘NICA’). The facts of the case have already been explained in an earlier post, so there is no need to rehearse them in detail here. In outline, the case concerns the compatibility of the Northern Ireland Troubles (Legacy and Reconciliation) Act 2023 (‘the Legacy Act’) ─ which set out broad immunities for crimes committed during the Troubles ─ with human rights protections embedded in Articles 2 and 6 ECHR and in EU law (the Victims’ Rights Directive and Charter of Fundamental Rights).

Despite the recent proposals to replace the Legacy Act with more moderate legislation, the appeal raises important issues of UK/EU relations law, relating to the guarantee of ‘no diminution’ in respect of the Rights, Safeguards and Equality of Opportunity (‘RSEO’) part of the Belfast/Good Friday Agreement, which was introduced to the EU/UK Withdrawal Agreement through Article 2(1) of the Windsor Framework (WF) (formerly known as the Ireland/Northern Ireland Protocol – an annex to the Withdrawal Agreement). This will be the first time the Supreme Court interprets the breadth and meaning of the non-diminution guarantee, thus clarifying the extent of Northern Ireland’s protection from legislative attacks on vested EU rights after the UK’s withdrawal from the bloc. This note aims to analyse the EU aspects of the appeal which – understandably, given the multitude of issues involved in the Dillon case –  received little airtime during the hearing, but are nonetheless central to the operation of Article 2 WF. These are:

1. Whether Article 2(1) of the Windsor Framework enjoys direct effect and, if so, whether it can give rise to the disapplication of primary legislation, such as the Legacy Act; and
2. Whether, like other sources of EU rights protection covered by Article 2 WF, the EU Charter of Fundamental Rights can be relied upon as a distinct basis for disapplying domestic legislation (a point raised in a cross-appeal by the Dillon claimants and two intervening rights monitoring bodies: the Equality Commission for Northern Ireland (ECNI) and the Northern Ireland Human Rights Commission (NIHRC)).

In respect of the first issue, the NICA found that Article 2(1) WF is directly effective because ‘the key obligation [of non-diminution] assumed by the UK in Article 2(1) is a clear and unconditional obligation of result’ (NICA, [85]). It thus applied the ‘classic’ direct effect formula stemming from the CJEU’s ruling in Case 26/62 Van Gend en Loos, namely that a provision of EU law enjoys direct effect if it is clear, precise, and unconditional. Finding that the case fell within the scope of RSEO and that, prior to the introduction of the Legacy Act, individuals would have enjoyed protection of their rights to a hearing and to a judicial remedy under the Victims’ Rights Directive, the NICA went on to disapply several provisions of the Legacy Act as incompatible with EU law, noting that EU law ‘mandates’ disapplication (NICA, [158]).

Before the Supreme Court, the SOSNI has argued that Article 2(1) WF is not directly effective or, at least, that it is not directly effective in the unequivocal terms suggested by the lowers courts. As the SOSNI’s primary case goes, Article 2 was intended to be safeguarded politically, with disputes resolved at a Specialised Committee with participation from the Dedicated Mechanism structure for the continued monitoring of human rights and equality safeguards in Northern Ireland after Brexit (the Irish Human Rights and Equality Commission, the NIHRC and the ECNI) and should not be read as a provision that is invocable by individuals (see e.g. paras 13-14 and 29 of the SOSNI’s written case here).  And it should not be viewed as allowing domestic courts to disapply primary legislation, either: rather, if an irresoluble incompatibility with EU standards is identified under Article 2(1) WF, the response should be a political one. Failing that, Article 2(1) WF can only be understood as enjoying direct effect insofar as the EU law measures to which it relates enjoyed direct effect before Brexit (see para 35 of the SOSNI’s written case).

In respect of the second issue, i.e. the status and effects of the EU Charter of Fundamental Rights (CFR) in the context of the non-diminution guarantee, the NICA had stated that: ‘we adopt the position that the CFR acts as an aid to interpretation of relevant EU law provisions’ (NICA, [137]). As such, while the CFR can be relied upon to guide the domestic courts towards conforming interpretations, it cannot be the basis of disapplying primary legislation. Dillon, the NIHRC and ECNIcontest this finding. They argue that, provided a case falls within the material scope of EU law, the non-diminution commitment means that the CFR can be used as the relevant EU standard, incompatibility with which can result, if necessary, in the disapplication of legislation. Unfortunately, despite its significance, this point was not addressed at the hearing for lack of time, having been overshadowed by the broader question of direct effect and the meaning of RSEO.

In our analysis, we consider in turn the issues of direct effect, disapplication and the status of the CFR. We argue, by reference to EU case law on these issues, that the NICA ruling was correct with regard to direct effect and disapplication, and should be upheld, but that the Supreme Court should revisit the findings on the Charter which, in our view, misinterpret the status and effects of this instrument.

Does Article 2(1) WF enjoy direct effect?

The issue of the direct effect of Article 2(1) is of particular importance, not least because it already forms the basis of several decisions of Northern Ireland courts interpreting the non-diminution commitment. At the hearing, there was much debate about which standard ought to be applied to assess direct effect, challenging the NICA’s simple reliance on the Van Gend en Loos conditions that a provision be clear, precise and unconditional and seeking to attribute to this test an additional requirement of party intent, referring to the Vienna Convention on the Law of Treaties. However, the Windsor Framework is an integral part of the Withdrawal Agreement, as the Supreme Court held in Allister [2023] UKSC 5 ([65]). This ought to preclude any idea that the WF can be interpreted outside the interpretive rules in the Withdrawal Agreement itself, which apply across the entire treaty. In that regard, it is difficult to gloss over the explicit reference to direct effect in Article 4(1) of that Agreement:

The provisions of this Agreement and the provisions of Union law made applicable by this Agreement shall produce in respect of and in the United Kingdom the same legal effects as those which they produce within the Union and its Member States.

Accordingly, legal or natural persons shall in particular be able to rely directly on the provisions contained or referred to in this Agreement which meet the conditions for direct effect under Union law.

As a result, even if principles of interpretation under international law were employed to assess the existence of direct effect, the key question under the terms of the agreement itself would remain whether Article 2(1) WF is of such a nature as to satisfy the conditions for direct effect under Union law. In this regard, the SOSNI argued that, since Article 2(1) is predicated upon the existence of other EU fundamental rights protections, it cannot be considered a clear, precise and unconditional obligation. From the perspective of EU law, however, this reading of the direct effect conditions is incorrect.

As pointed out in the NIHRC’s submissions and in line with the NICA ruling, settled EU case law on direct effect, including long-standing case law on the direct effect of international agreements, affirms in plain terms the Van Gend en Loos conditions. A provision of an international agreement is directly effective if it ‘contains a clear and precise obligation which is not subject, in its implementation or effects, to the adoption of any subsequent measure’ (Case C-549/22 Transfert de prestations de survie, para 30; Case C-485/07 Akdas and others, para 67; and Case 12/86 Demirel, para 14). The provision must be assessed as clear and precise with respect to ‘the result to be achieved’ (Transfert de prestations de survie, para 31) and must not, in the context of the agreement in question, be intended to set out a ‘programme’ that requires subsequent measures (Transfert de prestations de survie, para 31; Demirel, para 23). Article 2(1) WF meets these conditions: it is clear as to the result to be achieved (no diminution in rights safeguards and equality of opportunity) and does not depend on subsequent measures to achieve this result. Rather, the thrust of the provision insofar as it is at issue before the Supreme Court in Dillon is a negative obligation that rests upon the state refraining from unilaterally reducing the EU rights safeguards that were in place before the end of the transition period, i.e. at a precise point in time (31 December 2020).

It is true, of course, that Article 2(1) WF does not expressly list all the EU measures to which the non-diminution commitment relates. It merely states that the commitment includes (and therefore cannot necessarily be limited to) the six EU equality directives mentioned in Annex 1 of the Windsor Framework. The concept of ‘diminution’, therefore, has been understood as referring to these directives as well as to a broader range of EU measures that may be considered to have underpinned the RSEO section of the Good Friday Agreement. The Dillon case itself concerns EU measures falling within this broader non-diminution commitment, because – as the argument of Dillon et al goes – during the UK’s EU membership, the RSEO was underpinned by the Victims’ Rights Directive and Articles 2, 4, and 47 of the EU Charter of Fundamental Rights, and these protections have been removed or reduced by the Legacy Act.

It is essential to emphasise, however, that any lack of clarity about the meaning of RSEO does not ‘undo’ the direct effect of Article 2(1) from the perspective of EU law. The direct effect conditions relate to the obligation of non-diminution: a clear, precise, and unconditional standard to be applied to whatever RSEO is deemed to amount to. In other words, while factors such as party intent (e.g., the Government’s Explainer document on Article 2) may prove important in allowing domestic courts to interpret the meaning of RSEO as a set of guarantees that for the first time acquire a justiciable meaning through Article 2(1) WF, they do not determine the overall direct effect of Article 2(1), which refers to non-diminution as the relevant obligation of result from the perspective of the EU direct effect conditions. Indeed, the fact that Article 2(1) enumerates only some of the measures that underpin it (the six principal EU equality directives) is neither unusual nor indicative of the absence of direct effect. On the contrary, it is difficult to think of any directly effective provision of EU law that lists all the measures that may give it ‘more specific expression’ (Case C-555/07 Kücükdeveci, para 21; Case 43/75 Defrenne, para 54). Take, for example, the right to equal pay between men and women (Article 157 of the Treaty on the Functioning of the European Union and Articles 21 and 23 CFR): while this right has enjoyed direct effect since the 1970s, the provisions of primary law that protect it do not state in detailed terms what constitutes ‘pay’, nor do they specify its meaning in particular fields, such as social security, which have been further specified in secondary legislation and CJEU case law.

The argument against the overall direct effect of Article 2(1) WF, therefore, does not appear strong from an EU law standpoint. The consequence of this is that there is no need to have regard to the nature of the underpinning measures in order for an individual to be able to invoke Article 2(1) WF before a domestic court. Indeed, to the extent that the SOSNI argues that Article 2(1) as a whole cannot have direct effect, that argument runs contrary both to the text of the Withdrawal Agreement and to fundamental principles of EU law. This alone should provide the Supreme Court in Dillon a moment’s (significant) pause: accepting the SOSNI’s argument in this regard would be at odds with the Court’s recent jurisprudence on the continued applicability of EU principles to assimilated law (see e.g. CG Fry [2025] UKSC 35, [43-59]). If the Court is careful to interpret assimilated law in line with the principles of EU law (which assimilated law once was), it must be even more careful to apply those principles when genuine EU law is at issue.

It follows that Article 2(1) WF is invokable as such with regard to whatever measures underpinned RSEO, including (and not necessarily limited to) the six equality directives listed in Annex 1.  

Can Article 2(1) WF result in the disapplication of legislation?

If our assessment that Article 2(1) WF is directly effective is correct, this would mean, in line with the NICA’s findings, that domestic courts can disapply legislation such as the Legacy Act. Nevertheless, it is worth adding a few remarks on the role of disapplication in EU law, which may clarify its scope and extent.

On the one hand, it is essential to emphasise that disapplication and direct effect go hand-in-hand, in the sense that disapplication always presupposes direct effect and must always be a possibility for a domestic court when ordering the application of a directly effective provision. As the CJEU has repeatedly put it, if a domestic court ‘were to find that it is not possible to interpret its national law in conformity with EU law, it would be required to ensure, within its jurisdiction, judicial protection for individuals flowing from those provisions, and to ensure their full effectiveness by disapplying, if need be, the national provisions concerned’ (see, e.g, Case C‑713/23 Trojan, para 76; Case C‑414/16, Egenberger, paras 78 and 79). On the other hand, it is also important to read carefully the CJEU’s approach in this matter. While, as Alison Young has noted, over time disapplication became for UK courts a run-of-the-mill synonym of the direct effect of EU law, this is not how it is understood in CJEU case law. Rather, as highlighted above, the case law allows domestic courts a degree of remedial autonomy before requiring disapplication and, crucially, permits them to use consistent interpretation if that would result in removing the incompatibility, thereby giving full effect to EU law (see, e.g., Joint Cases C-397/01-403/01 Pfeiffer; Case C-573/17 Popławski II). While it is true, therefore, that disapplication is mandated by the direct effect of EU law, domestic resistance to it may be lessened if it is properly understood as a ‘fallback’ duty that kicks in if domestic courts are unable to give effect to EU law in another way.

Of course, as the NICA rightly observed, the direct effect of underpinning EU law ‘may be relevant to the question of diminution (if the claimed diminution relates to the remedy available for breach of the right rather than a reduction in the substance of the right itself)’ (NICA, [84]). This means that Article 2(1) guarantees against diminution of both rights and their concomitant remedies. For example, where an EU measure was directly effective and could give rise to disapplication before Brexit, the removal of the possibility of disapplication following Brexit would in itself amount to diminution, as it would reduce the remedial potency of the right in question. The NICA’s clarification is important because it shows that disapplication forms part of the internal logic of Article 2(1) WF as a possible component of diminution, rather than generally being reintroduced through the back door. Put simply, disapplication is not a severable aspect of Article 2(1) WF; it is a logical consequence of a breach of the rights guaranteed by the provision, where those rights previously engaged the disapplication remedy. This approach accords with earlier Supreme Court authority, which confirmed that Section 7A of the Withdrawal Act maintains the full effect of directly effective provisions of the WA (including remedies) as part of the domestic legal order (Allister, [65]).

Can the CFR underpin the Article 2 commitment of non-diminution and give rise to disapplication?

The final EU law issue before the Supreme Court in Dillon, raised by the cross-appeal, concerns the relationship between Article 2 WF and the CFR. As noted earlier, the NICA had held that the CFR is only relevant because it has interpretive value, but that it cannot be used to give rise to disapplication. In our view, this finding should be revisited: not only does it contradict the overall approach to the direct effect of Article 2(1) WF, but it is also inconsistent with CJEU and Supreme Court authority on the status and effects of the Charter.

The NICA’s findings concerning the Charter seem to have stemmed from an erroneous understanding of the scope and meaning of the CFR as an instrument that is not self-standing within EU law (see, esp. NICA, [141-142]). This led the NICA to conclude that the Charter did not go beyond the Victims’ Rights Directive and to discount as irrelevant the Supreme Court ruling in Benkharbouche v Secretary of State for Foreign and Commonwealth Affairs [2017] UKSC 62, which had previously found that Article 47 CFR had direct effect and gave rise to disapplication (NICA, [144-145]). The NICA’s findings are perhaps understandable on the facts of the Dillon case, because the Victims’ Rights Directive had been directly effective against the state before Brexit and would thus in itself have allowed the Dillon claimants to ask for the disapplication of the Legacy Act. However, clarity about the status of the Charter could prove central in cases where EU secondary legislation lacked direct effect or where the Charter did offer substantively broader protection. This would be the case, for example, if diminution were invoked in private law contexts covered by the equality directives (because directives only enjoy direct effect against the state).

More specifically, whereas the Charter requires some ‘hook’ onto EU law in order to apply, this has been interpreted simply and broadly by the CJEU. The Charter is applicable ‘in all situations governed by European Union law, but not outside such situations’ (Case C-617/10 Fransson, para 19). As a result, if a case is within the material scope of secondary legislation, such as a directive, or within the scope of a Treaty provision, such as the free movement of goods or persons, the Charter is engaged. While the SOSNI drew attention to cases such as C-198/13 Hernández and C-609-610/17TSN to suggest that the Charter was inapplicable in Dillon, this is unpersuasive, as these cases concerned a situation where domestic legislation exceeded EU standards (a scenario not applicable in Dillon). In reality, there have been very limited instances over the years in which the CJEU has drawn the scope of the Charter narrowly: Fransson itself was a case in which the Charter was engaged as a result of an EU taxation directive that actually post-dated the offending domestic legislation, and the Charter also clearly applies to attempts to derogate from EU standards (see Fransson, para 19, and Case C-260/89 ERT, para 42). It is worth noting, too, that the scope of EU law (and consequent importance of the Charter therein) have been drawn particularly expansively in the context of Article 47 CFR (the right to  an effective remedy and to a fair trial), due to the centrality of this right in the EU legal order (Case C-64/16 Associação Sindical dos Juízes Portugueses, para 35).

Once we surmount the first hurdle of the Charter’s scope of application, EU law straightforwardly accepts that the Charter provisions can be directly effective (and therefore can give rise to disapplication, as explained above). For example, in cases where discrimination is invoked against private employers, the CJEU has emphasized the direct effect of Articles 21 and 47 CFR, which are invocable ‘as such’ (see, eg, Kücükdeveci, para 53; Egenberger, para 76; Case C‑193/17 Cresco Investigation, para 76, Case C-715/20 KL, paras 77-82). It is true, of course, that not all provisions of the Charter have direct effect (see, e.g., Case C‑176/12, Association de médiation sociale, paragraph 47). However, this is not the case with regard to the rights invoked in Dillon, and incompatibilities with a directly effective provision of the CFR have always required the inclusion ofdisapplication in the domestic remedial apparatus (see, for a recent restatement, Trojan, para 76).

Conclusion

Dillon will be decided in a context of tense anticipation for both Westminster and Stormont regarding the future of UK/EU relations and the special status of Northern Ireland within these arrangements. The Supreme Court’s findings will determine the extent to which Article 2(1) creates a Northern Ireland-specific exception from several provisions of the EUWA 2018 and the REULA 2023, which had otherwise removed the possibility of relying on EU fundamental rights to have legislation disapplied, while altogether removing the CFR from domestic statute books. In this note, we have attempted to draw attention to, and to demystify the EU law concepts on which the Dillon case depends, namely the direct effect of Article 2(1) WF, the role of the CFR within Article 2 WF, and – most controversially – the disapplication remedy. It is to be hoped that clarity about these issues will be provided by the Supreme Court, thereby avoiding a further escalation of what is now a moot case.

To put this in perspective, further escalation may involve arbitration proceedings against the UK under the Withdrawal Agreement. The CJEU has a role to play in such proceedings insofar as questions of EU law are concerned. This possibility should be avoided (and is indeed avoidable) because for close to five decades until 2021, UK courts had a front-row seat to the development of the EU principles which we have charted here. In that time, domestic courts developed considerable familiarity with these principles, even anticipating their authoritative exposition by the CJEU. The scope of the Charter’s applicability as stated in Fransson, for example, was preceded (by around 3 months) by the Supreme Court’s judgment in Rugby Football Union [2012] UKSC 55 ([28]). And it was also the Supreme Court that had first acknowledged, in its ruling in Benkharbouche ([78]), that Article 47 CFR was directly effective and could give rise to the disapplication remedy, two years before the CJEU explicitly confirmed this in Egenberger. The issues in Dillon underlying Article 2(1) of the Windsor Framework are thus far from novel, no matter how much they may be so dressed up by the government.

The present note has been prepared by the authors solely in their academic capacity. The authors are grateful to Professor Se-Shauna Wheatle and Dr Paul Scott for their comments on an earlier draft.

Dr Eleni Frantziou is Associate Professor in Public Law and Human Rights at Durham University. Anurag Deb is a PhD candidate at the School of Law, Queen’s University Belfast. The authors have written extensively on the Windsor Framework and their forthcoming report on remedies and enforcement under the Windsor Framework (co-authored with Professor Aoife O’Donoghue, QUB; Professor Colin Murray, Newcastle University; and Dr Sylvia de Mars, Newcastle University) will be launched by the Northern Ireland Human Rights Commission in Belfast on 16 December. The link to register for the event is here.

---

(Suggested citation: E. Frantziou and A. Deb, ‘Direct Effect, Disapplication, and the EU Charter of Fundamental Rights: Still Together in Northern Ireland after Dillon?’, U.K. Const. L. Blog (4th December 2025) (available at https://ukconstitutionallaw.org/))