Introduction

It is a rare case where the UK Government initiates a judicial review against an independent public body. The Thompson case is just such a case. The Secretary of State for Northern Ireland intervened in an inquest process to attempt to prevent information about the murder of Liam Paul Thompson by the UDA in 1994, during the Northern Ireland conflict, from being released by the coroner. This is therefore a case in which the hurdles of judicial review theoretically work against the government – review is not appeal; it is a limited check on the lawfulness of the decision and the process by which it was reached. Here, the government is confronting these challenges. The shoe is (theoretically) on the other foot, which up until this point in the case has been a real challenge for the government before the Northern Ireland courts.

It should be highlighted from the outset that more than 30 years have elapsed since the inquest into this murder opened in 1995, and that this is indisputably, as pointed out in the first High Court judgment in the litigation, an ‘lengthy and egregious delay’ (at [1]). Much of the reason for that delay lies in how successive UK Governments have approached legacy litigation in Northern Ireland and their determination to conceal the involvement of state agents in some of the darkest events of the conflict, applying a “neither confirm nor deny” (NCND) policy over issues related to such agents in a blanket fashion. The Thompson case brings these issues back to the door of the UK Supreme Court.

Background to the Case

The delays to the inquest into Liam Paul Thompson’s murder are not uncommon when it comes to the legacy of the Northern Ireland conflict. In multiple cases involving allegations of direct involvement against the security forces, or the involvement of state agents, the UK Government has resorted to extensive use of Public Interest Immunity processes, with the result that decades have passed without any movement on hearings. Where these delays have been challenged by victims’ relatives, such cases have again ground through closed material procedures and the UK Government has in some cases offered large settlements in an effort to thwart disclosure processes.

When in 2022, the Conservative Government decried much of this legacy litigation as “lawfare”, an effort to rewrite the history of the conflict in the courtroom ‘promoted and peddled by those with a vested interest in presenting the British state as the aggressor’, and attempted to shutter these legal processes on events long past, it was putting an implausible gloss on most of these inquests, civil actions and prosecutions. If multiple cases which had the potential to involve embarrassing revelations were still ongoing after decades, it was because the UK Government had kept delaying them. In the Thompson case the Coroner, Louise Fee, accepted the Government’s assertion of PII. However, as the statutory bar on legacy inquests under the Northern Ireland Troubles (Legacy and Reconciliation) Act 2023 approached on 1 May 2024, she proposed providing the family with a “gist” containing information about the circumstances of the killing. Key questions for the inquest included whether information about the deceased was passed on to the killer by anyone in the security forces, whether the location of the killing was visible to the police/military facility nearby, and whether the death was preventable, including by reference to what the security forces knew and could have done to prevent it. These questions, on any view, are indicative of the gravity of the allegations surrounding the death. Eventually, a gist of the relevant material was produced which was accepted by the Chief Constable of the PSNI.

The Secretary of State for Northern Ireland (who did not agree with either the gist or the disclosability of the sensitive information by gist) then instituted judicial review proceedings to prevent publication of that gist, asserting that its release would undermine the NCND policy that the UK Government maintained in relation to intelligence matters and the conduct of state agents. Both the High Court and the Court of Appeal in Northern Ireland upheld the Coroner’s decision, finding that she was entitled to disclose the gist in the Thompson case. These decisions, drawing upon Al Rawi v The Security Service, recognised that in such a case the judge must assess the relative significance of potentially competing aspects of the public interest, around the demands of open justice, the need for an effective investigation into killings involving state actors under Article 2 ECHR, and the need to prevent harm being caused to national security.

In accepting the lawfulness of the Coroner’s decision to issue a gist, the Court of Appeal asserted that her ‘approach represents an important reassurance to the family of the deceased and maintains public confidence in the investigative process employed to date’ and that ‘to suggest otherwise would, to our mind, seriously undermine the administration of justice and perpetuate mistrust and suspicion given the 1 May deadline’ (at [58]). The Secretary of State appealed to the UK Supreme Court, maintaining (in strident terms) that any mitigation of NCND ‘was contrary to the national security interest in countering the threat from terrorists’ (at [4]). The Thompson family, the coroner, and the Chief Constable of the PSNI were respondents, all supporting publication of the gist. Liam Paul Thompson’s Brother, Eugene, who was terminally ill at the time of the hearing, died in August 2025, one week after receiving a bedside apology from the Chief Constable of the PSNI for police failings relating to his brother’s killing.

 Judgment

The Supreme Court’s judgment was authored by Lord Sales and Lord Stephens, with whom Lord Reed, Lord Hodge and Lord Lloyd-Jones agreed. The judgment begins by setting out some background facts, including facts relating to the death, before exploring the procedural history of the case, from its emergence in the inquest to its hearing before the Supreme Court. The Court notes the fact that the PSNI and the Secretary of State were initially of one mind in opposing any public disclosure of the sensitive information held by the police by any gist ([43]-[57]), before the Chief Constable of the PSNI informed the Coroner that he objected only to the gist as then formulated, but not to the disclosure of a gist per se ([58]). The Chief Constable had not, however, consulted ‘with other relevant public authorities with an interest and expertise who are capable of making a material contribution’ to the assessment of national security in relation to information held by his own police force ([37]). The Supreme Court identified the Northern Ireland Secretary as the pre-eminent authority for this assessment, pointing to his ‘political authority’, ‘democratic authority’ and ‘institutional competence’ ([38]). The Chief Constable however considered that different public authorities could validly come to different opinions as to national security and its intersection with the public interest, and that it was for a competent court (in this case, the Coroner) to resolve any such differences ([59]). The Supreme Court criticised this approach, stating that public authorities with divergent views in this context should first attempt consensus ([60]), but if that should fail, the public authority with primary responsibility (i.e. the Secretary of State) must make the national security assessment by itself ([61]). 

The Court then considered two issues which go to the heart of the judgment and its implications more generally, which we explore in the next section. Public interest immunity claims and their gist encompass sensitive information ([135]) and the Coroner has no discretion when assessing the public interest ([159]). Indeed, the Court was at pains to stress that the Secretary of State’s assessment ‘should have been accepted’ by the Coroner unless it was Wednesbury irrational ([138]). Interestingly, the Court examined the gists in issue for itself, finding them couched at a ‘high level of abstraction and in a very indeterminate manner’ ([136]). This, the Court concluded, militated against the public disclosure of the gists, finding that they would be subject to the same PII claim as their underlying information (and in respect of which the Coroner also had to accept the ministerial assessment unless it was Wednesbury irrational). Quite why a highly abstract and ‘indeterminate’ short gist or gists would negatively impact national security is not something the Court explains. Instead, it expressed its doubt that the Coroner had given the matter sufficient care, including on the question whether either gist would have been sufficient on which to make factual findings ([136]).

Consequences

Several consequences flow from this decision, of which we cover three. The first is the consequence for the legacy process itself. Just last week the Secretary of State for Northern Ireland, Hilary Benn, cited the pending Thompson judgment as the reason for withholding consent for the final Operation Kenova report to name the agent known as “Stakeknife”. Back in 2003, the Stakeknife controversy ironically provided the wellspring of modern judicial engagement with NCND, In re Scappaticci, as Freddie Scappaticci unsuccessfully sought to have the UK Government issue an outright denial that he had been their agent. The first consequence of the decision is that the Secretary of State will be subject to no renewed pressure in this regard. Abridging of PII, or lifting of NCND, remains in the gift of the Secretary of State. It has been lifted in some cases, notably the Brian Nelson case referenced in the judgment (at [27]), and denied in others. The UK Government, as a key actor in the conflict, enjoys pole position in determining how its events are interpreted. 

This feeds into the legacy processes established under the Northern Ireland Troubles (Legacy and Reconciliation) Act 2023 and set to be modified under the Northern Ireland Troubles Bill 2025. The Supreme Court judgment concludes that these legacy processes provide a means to discharge the UK’s “effective investigation” obligations under Article 2 ECHR (at [166]). This is a paragraph which loudly foreshadows the UK Supreme Court’s position in the Dillon case, on the adequacy of those legacy processes, with its judgment on this expected in the new year. The problem, however, is that no matter how much those processes might be adapted to make them more rights compliant, this judgment robs those adaptations of much of their force. Secretaries of State will retain the ability to block disclosures of information on security grounds as they see fit. The current UK Government has made great play of the fact that, as part of the reformed process, interested parties ‘may appeal a decision to withhold sensitive information from onward disclosure to the relevant court, consistent with the principles applicable on an application for judicial review’. In light of the Thompson judgment, that pledge would seem to have very little value.

The second consequence concerns the day-to-day issues for policing in Northern Ireland. As Jon Boutcher noted in a recent speech, there are good reasons for the police in Northern Ireland (and he as Chief Constable) to be on opposite sides from a case on NCND than the UK Government. The Patten Report into wholesale reform of policing in Northern Ireland had set out a blueprint for policing which was open and transparent, but in the subsequent decades the UK Government has continued to make broad PII claims in relation to the most controversial aspects of state activity during the conflict. For Boutcher, ‘we operate under exceptionally intensive oversight arrangements and yet, when it comes to legacy cases, secrecy has dominated, not because secrecy is always necessary, but because NCND has become the default’. The ability of the PSNI to achieve broad community support for policing based on openness has been throttled by this secrecy; it remains saddled with the legacy of what came before it.

The third consequence is for how the Supreme Court conceptualised appellate (merits) review in this context, linking it with other contexts such as proportionality assessments ([125]) The whole point of merits review should be for a tribunal to “stand in the shoes of the Secretary of State” (see [83] of Chimi) and, in the context of a challenge to a decision that disclosing a gist is not in the public interest, make a decision “in the light of all the relevant facts and competing interests for and against disclosure” (see [7] of the Supreme Court’s decision in Evans). It is a form of heightened intensity review, beyond the traditional rationality review deployed in public law even where “anxious” or “heightened” scrutiny is deployed.

The Supreme Court’s formulation of merits review in Thompson manages to stand these principles on their head, by reducing the level of scrutiny afforded to such assessments, in two ways. First, the Court’s decision requires the reviewing tribunal (in this case, the Coroner) to accept an assessment that disclosure would undermine the public interest, subject primarily only to the public law principle of irrationality (see [39] and [138]). This prevents the reviewing tribunal from effectively balancing the various factors in the public interest for itself, an exercise which classically forms part of the animating rationale of merits review. This is because, although the tribunal is required in principle to balance the various factors, mandating that it give the minister’s assessment of public interest against disclosure near conclusive weight removes much of the discretion the tribunal would otherwise have in merits review.

Second, and more importantly, the real effect of the Supreme Court’s formulation is to enable appellate courts to reverse decisions which have failed to be sufficiently deferential to the minister. Put another way, the Court has determined that failing to show sufficient deference to a ministerial assessment is a matter of sufficient “social or political significance” (see [144] of Shvidler, cited in Thompson at [125]) to justify a greater degree of appellate interference with first instance “merits review”.

As the Court says, an appellate court can intervene where the first-instance tribunal has incorrectly decided where the public interest lies ([126]). However, the specific sort of appellate intervention envisaged here is essentially a check on how well the first-instance tribunal has rubber-stamped the minister’s decision. Merits review – usually an opportunity to subject a ministerial decision to greater scrutiny (with accordingly less deference showed) – is formulated instead as a prism through which the Supreme Court ensures maximal deference to the minister.

Conclusion

The events of the Northern Ireland conflict must look very distant to the UK Supreme Court; a persistent irritant as it attempts to grapple with the latest controversies. It is not as if the Court is not aware of the underlying problem, describing the delays faced by the Thompson family as ‘profoundly disturbing’ (at [18]). But, unlike the courts in Belfast, which are confronted with the daily effects of these failings, the Supreme Court continues to accept that it is in the UK Government’s sole purview to address these issues, and to deny any effective role for the judicial authorities in overseeing security determinations. That it is ‘in the national interest to encourage the provision of information from informers to protect the population and to save innocent lives’ (at [28]) pertains, in absolute terms, even at three decades’ remove from the events in question.

When the issue at stake is a gist which does no more than set out ‘a summary at a very high level of abstraction and in a very indeterminate manner’ (at [136]), the assertion that ministers are free to insist on non-disclosure because of the immediate threat of the alternative to national security becomes difficult to accept. The consequences of this position, however, are felt on the ground in Northern Ireland, where the conflict remains subject to enduring questions, and the darkest of accounts of its conduct are left to fester. But away from Northern Ireland or its dark past, Thompson is redolent of recent jurisprudence on security issues: simultaneously restricting the ability of first instance courts to assess executive decision-making while sharpening the edge of appellate review to ensure those restrictions are ultimately enforced.

Anurag Deb is a PhD Candidate at Queen’s University Belfast. Colin Murray is Professor of Law and Democracy at Newcastle University. Gabriel Tan is an MPhil Law student at the University of Oxford and a future trainee barrister at Matrix Chambers.

Statement of Interest: Both Anurag and Colin are on the Board of the Committee for the Administration of Justice, which represented Eugene Thompson in this litigation. The authors are grateful to Dr Paul Scott and Professor Michael Gordon for their feedback on an earlier draft.

(Suggested citation: A Deb, C Murray and G Tan, ‘Legacy Issues: In re Secretary of State for Northern Ireland [2025] UKSC 47’, U.K. Const. L. Blog (18th December 2025) (available at https://ukconstitutionallaw.org/))