In September 2024, the Government suspended around thirty arms export licences to Israel, having found a clear risk that United Kingdom equipment might be used to commit serious violations of international humanitarian law. It exempted one category, creating a carve-out for components made in the United Kingdom for the F-35 combat aircraft, which continued to reach Israel through the global spares pool. When that carve-out was challenged, the Divisional Court in R (Al-Haq) v Secretary of State for Business and Trade (30 June 2025) accepted that the clear-risk finding was, in the Foreign Secretary’s own view, the only conclusion available to him at [36]. It nonetheless held that whether the continued supply complied with the United Kingdom’s international obligations was not a question a domestic court could decide.

A year later, on 9 June 2026, the Foreign Secretary announced a fourth package of sanctions against extremist settler networks in the occupied West Bank, while the F-35 components continued to flow. The question these decisions pose is internal to the British constitution: which institution, if any, can now ask whether the executive is discharging the international legal obligations the United Kingdom has undertaken?

My claim is that the answer is none, and that this result is indefensible on the constitution’s own terms. A dualist objection arises at once: courts do not adjudicate the executive’s compliance with unincorporated international law, so their reluctance here is mere orthodoxy. That would be decisive if the question were compliance with the Geneva Conventions at large. It is not. The executive adopted its own published framework, the Strategic Export Licensing Criteria, to structure its compliance with those obligations, and the reviewable question is whether it held itself to that framework. That is a domestic public-law question, not an invitation to referee international law, and it is the question the abstention in Al-Haq closed off. That abstention is an exercise in political constitutionalism: the court withdrew because these judgments are “reserved under our constitution to the judgment of the executive”, answerable to Parliament and the electorate [112]. That reasoning carries a hidden premise. It assumes that political accountability is not merely the proper forum for these decisions but a forum that actually operates, so that judicial restraint transfers scrutiny elsewhere rather than extinguishing it. Once that premise is stated plainly, it can be tested. The same pattern governs sanctions, and set beside the licensing decision, it reveals a single structure: across both, the executive’s compliance with its own standards for discharging those obligations is insulated at once from judicial review and from any specialist parliamentary scrutiny. Political constitutionalism promises that where the courts stand back, political accountability answers for the executive in their place. That is the promise the abstention invokes, and here it is broken: both channels have closed at once, each defended by pointing to the other.

Abstention on whose authority?

Non-justiciability is not one doctrine but two, as the Supreme Court explained in Shergill v Khaira (11 June 2014): a claim may be beyond the courts’ constitutional competence, or it may lack manageable standards by which to be decided. The second limb cannot carry Al-Haq. R (Campaign Against Arms Trade) v Secretary of State for International Trade (20 June 2019) proves that manageable standards exist. There, Campaign Against Arms Trade (CAAT) challenged the licensing of arms exports to Saudi Arabia for use by the Saudi-led coalition in Yemen, and the Court of Appeal reviewed the rationality of the Government’s risk assessment against the licensing criteria, holding the process unlawful for failing to address a historic pattern of violations, without adjudicating any sale on its merits. If that class of question was judicially manageable in 2019, the abstention in 2025 must rest on the first limb, constitutional competence. The court’s own authorities confirm this: it cites Rehman, Begum and U3 [79]-[81], whose logic is that such decisions gain legitimacy only from being made by persons the people have elected and can remove. The doctrine’s validity is thereby conditional on a claim of political fact, to which I return.

Two further moves made by the court compound the difficulty. The first concerns the Strategic Export Licensing Criteria, the Eight Criteria against which the Government assesses every strategic export licence, including Criterion Two, which requires refusal where there is a clear risk that the items might be used to commit or facilitate a serious violation of international humanitarian law. Section 9(3) of the Export Control Act 2002 obliges the Secretary of State to issue guidance on licensing principles, and section 9(5) obliges decision-makers to heed it. These Criteria, issued in their current form in December 2021, were themselves stated to apply without prejudice to specific measures announced to Parliament [21]. The court treated the F-35 carve-out as one such measure: rather than misapply the Criteria, the Secretary of State had stepped outside them, so no question of compliance arose [113]-[114]. Since R (Lumba) v Secretary of State for the Home Department (23 March 2011), the consistency principle has required a public authority to follow its published policy absent good reason, with the departure itself reviewable. Al-Haq inverts that grammar: guidance whose production Parliament required by statute becomes optional at the moment it matters most, through a self-exemption clause whose exercise, on the international law questions it raises, is then unexaminable.

The second move concerns customary international law. The court assumed that the duty to prevent genocide and the obligation under Common Article 1 of the Geneva Conventions had customary status [131], yet declined to receive them into the common law for a constitutional reason: doing so would constrain the executive’s conduct of foreign affairs [135]. This extends R v Jones (Margaret) (29 March 2006) beyond its rationale: Jones refused to let customary law generate criminal offences because creating crimes is Parliament’s business, whereas supervisory review creates nothing and merely tests the lawfulness of executive reasoning. The move also sits uneasily with Belhaj v Straw (17 January 2017), where the gravity of alleged violations defeated non-justiciability rather than establishing it.

The empirical premise examined

Consider the premise on which the whole structure rests. The Committees on Arms Export Controls, the quadripartite mechanism through which four select committees jointly scrutinised licensing from 1999, ceased to meet in that form in January 2024; the Business and Trade Committee now leads through a Sub-Committee on Economic Security, Arms and Export Controls, which folds arms exports into a broader brief spanning investment screening and economic security, a change the Defence Committee opposed as a loss of balanced scrutiny. That sub-committee has corresponded with ministers on export licensing, but no select committee, however diligent, can rule the executive’s compliance unlawful. Political scrutiny can interrogate and expose; it cannot determine legality, and it is the determination of legality that Al-Haq foreclosed.  A doctrine of abstention resting on an empirical premise ought to verify it. Here, the premise was treated as axiomatic.

The sanctions limb confirms the pattern. The Sanctions and Anti-Money Laundering Act 2018 empowers ministers to make sanctions regulations whenever they consider it appropriate for purposes as capacious as furthering a foreign policy objective. During its passage, the House of Lords Constitution Committee concluded that it was “constitutionally inappropriate for ministers to have the power, by regulations, to create new forms of sanctions”, and Lord Judge, a former Lord Chief Justice, derided the Bill in debate as a “bonanza of regulations”. New sanctions regimes are created by regulations subject to the affirmative procedure, approved after the fact and unamendable; individual designations under them attract no parliamentary procedure at all. Within that frame, the designation of two serving Israeli ministers in June 2025 and of the settler networks in June 2026 was pure discretion. Nothing obliges the Government to designate anyone. Whether the 2026 package leaves the architects of settlement policy untouched, as Amnesty International UK contends, may be right or wrong; the constitutional point is that no institution is positioned to test it.

Judicial oversight of sanctions is one-sided. In Shvidler v Secretary of State for Foreign, Commonwealth and Development Affairs (29 July 2025), the Supreme Court confirmed the proportionality review of designations, while according the Foreign Secretary a wide margin rooted in institutional competence, over Lord Leggatt’s dissent. That jurisdiction exists to protect the targets of sanctions. No forum examines the converse:  whether the Government discharges its obligations when it declines to use the power at all, or uses it only at the periphery of a settlement enterprise that the International Court of Justice has held unlawful. That question has no home. Dyzenhaus’s account of grey holes, spaces where the forms of review persist but constrain nothing, describes the result. Review thus protects those against whom the powers are used, not the public interest in whether the executive uses the powers to meet its obligations. For that interest, after Al-Haq and Shvidler, the forms of review remain, but the substance is gone.

Due diligence without a forum

The subject matter did not force the choice: on 3 October 2025 the Dutch Supreme Court confronted the identical question about the same F-35 spares pool and reached a calibrated position: although the courts could not themselves determine whether the clear-risk threshold was met, that being for the Minister, it ordered him to reassess the licence within six weeks against the binding criteria of the Arms Trade Treaty and the European Union Common Position, and kept the suspension in place in the interim. Deference on the merits went with insistence that the assessment occur against the applicable standards, a discipline English law once knew in CAAT. Abstention in Al-Haq was a constitutional choice, and choices require justification.

The justification matters more here than in ordinary foreign affairs because of what the obligations are. As the International Committee of the Red Cross reads it in the 2016 Commentary, Common Article 1 of the Geneva Conventions imposes a duty of due diligence to ensure respect for the Conventions, and the supply of arms is the paradigm case in which a state must use its leverage to that end. The duty to prevent genocide under Article I of the Genocide Convention is stronger still: in Bosnia v Serbia (26 February 2007), the International Court of Justice at [430]-[431] tied it to each state’s capacity to influence and held that it arises the instant a state learns of a serious risk. That these duties are now engaged is not a matter of speculation. In Legal Consequences arising from the Policies and Practices of Israel in the Occupied Palestinian Territory, including East Jerusalem (19 July 2024), the Court found third states bound not to recognise or assist the unlawful situation, and in September 2025, the United Nations Commission of Inquiry called on states to halt arms transfers to Israel. The United Kingdom contests parts of this, as it may. But a contested obligation is still an obligation of conduct, discharged through assessment, reasoning and calibrated response, and so presupposing an institutional site at which diligence can be shown and its absence found. When the courts hold that the assessment is non-justiciable, decline to receive the underlying norms, and the specialist committee has been dissolved, the constitutional order supplies no such site. The obligation is not merely unenforced, which is international law’s familiar condition; it becomes untestable within the state that owes it.

The response that this is simply dualism working as intended mistakes coherence for justification. The court itself explains why unincorporated treaties are held at bay: otherwise the executive could “make law without parliamentary scrutiny by concluding a treaty, thereby undermining the sovereignty of Parliament” [68], a rationale restated in Miller (24 January 2017) and R (SC) v Secretary of State for Work and Pensions (9 July 2021). But that rationale assumes Parliament is watching. The doctrine keeps these norms out of the courts to protect Parliament’s authority. When the parliamentary forum stops scrutinising them, that protection inverts into its opposite: the space cleared for Parliament hardens into an immunity for the executive. Griffith’s political constitution demanded that the accountability of our rulers be real and not fictitious. Here, it is fictitious, and the fiction is load-bearing, because the courts’ withdrawal is expressly premised upon it.

Objections and the conditions of refutation

Three objections deserve answers. That courts lack competence in foreign affairs is answered by CAAT: process review there neither adjudicated another state’s conduct nor substituted judicial for ministerial judgment. The Dutch decision discussed earlier is not authoritative in this jurisdiction and is not offered as such; it shows only that the task can be done. That elections supply accountability mistakes the instrument for the task. Elections are retrospective, bundled and blunt; the continuous, granular scrutiny these decisions demand is what a dedicated committee is for, and that dedicated focus is what the 2024 change diluted. That an application to the European Court of Human Rights remains open misidentifies the obligations which arise under the Geneva Conventions, the Genocide Convention, and the Arms Trade Treaty, rather than the European Convention itself, whose extraterritorial reach to the victims of exported weapons is narrow. A reconstituted arms export committee with real powers would restore the political channel; a grant of permission by the Supreme Court in Al-Haq, or statutory incorporation of the licensing criteria and the obligations they reflect, would reopen the legal one. Until one occurs, the June 2026 designations and the continuing flow of F-35 components remain exercises of public power over life and death answerable, in substance, nowhere. A constitution that reserves such judgments to the executive on the promise of political accountability must keep the promise or stop invoking it.

Bhavya Johari, Lecturer in Jindal Global Law School

(Suggested citation: B. Johari, ‘The Political Constitution’s Broken Promise: Arms, Sanctions and Accountability after Al-Haq’, U.K. Const. L. Blog (15 July 2026) (available at https://ukconstitutionallaw.org/))