In early January 2026, the Court of Appeal (Peter Jackson LJ, Elisabeth Laing LJ, and Jeremy Baker LJ) handed down the decision in R (RR and AA) v Secretary of State for Foreign, Commonwealth, and Development Affairs [2026] EWCA Civ 3 (“RR and AA”). 

Brought by the Secretary of State for Foreign, Commonwealth, and Development Affairs (“SSFCDA”), the Secretary of State for the Home Department (“SSHD”) and the Secretary of State for Defence (“SSD”), the conjoined appeals concerned orders made by Swift J in judicial review proceedings related to the 2021 NATO withdrawal from Afghanistan. The appeals addressed two distinct issues: the limits of courts’ case management powers and the implications of the constitutional limits of judicial review of policy for the courts’ remedial powers.

This case note argues that RR and AA demonstrates how English law of civil procedure insufficiently accommodates the characteristics of public law litigation. Further, it argues that the Court of Appeal did not correctly engage with either the principles set out by the Supreme Court in R (Imam) v London Borough of Croydon [2023] UKSC 45 or the facts of RR and AA when evaluating the constitutional limits of judicial ability to award relief in judicial review proceedings. The discussion is structured as follows. First, a brief account is given of the facts of RR and AA and Swift J’s orders. The second part outlines the Court of Appeal’s (“CA’s”) reasons for allowing each of the appeals. The third and final part analyses the CA’s reasoning and its implications.

The factual background

In 2020, the then-Trump administration and the Islamic Emirate of Afghanistan (known as the Taliban) concluded an agreement under which all NATO forces were to withdraw from Afghanistan. Concurrently with the withdrawal of NATO forces in mid-2021, the Taliban launched an offensive that led to the fall of the Western-backed government. In response, several countries developed schemes offering relocation support to Afghan citizens who worked with or alongside NATO forces and faced a risk of harm from the Taliban. 

The United Kingdom’s relocation scheme, the Afghan Relocations and Assistance Policy (“ARAP”), was launched in April 2021. The applications of Afghan citizens who worked with or alongside UK forces were assessed by the Ministry of Defence (“MoD”). In February 2022, a soldier inadvertently leaked the personal details of some of the applicants. After discovering the leak, the government applied for a super-injunction prohibiting the existence of the leak from becoming public and established the Afghanistan Response Route (“ARR”). The aim of the ARR scheme was to assist a group of high-profile individuals who would likely be ineligible for relocation under the ARAP scheme, but who held links to the UK government and whose details were exposed in the data leak.

Swift J’s orders 

The appealed orders concerned two distinct issues related to the ARAP and ARR schemes. 

(1) The case management orders

The appeals in RR and AA v SSFCDA and MZZ v SSD and SSHD concerned case management directions given in judicial review challenges against adverse decisions under the ARAP scheme. 

In RR and AA and MZZ, the defendant Secretaries of State applied for closed material procedures (“CMP”) under section 6 of the Justice and Security Act 2013. The defendants then repeatedly applied for extensions of time to serve closed material. A MoD representative accepted that the delays occurred due to the Ministry’s failure to realise its proposed timelines for serving closed material were unachievable ([16]).

In a closed ruling (extracts from which have been reproduced in the CA judgment), Swift J considered the MoD’s account presented “a picture of civil servants who were apparently entirely reckless” as to whether correct instructions were being given to legal representatives. Further, the judge considered that “nothing” indicated that the MoD had taken steps to prevent incorrect instructions being given to counsel in the future. This, according to Swift J, was an “entirely unacceptable” position in light of the number of ARAP-related claims: the judge considered it desirable to ensure that future applications to vary timetables in challenges of ARAP applicants affected by the data breach are supported by explanations that are “complete and accurate” ([22]).

In RR and AA, Swift J consequently issued a “general direction” that “any application to vary any direction made in any ARAP-related High Court case” involving a CMP made “by or for the benefit of” the SSD should be accompanied by a witness statement setting out why compliance with the existing direction is not possible, why more time should be granted, and “why it is believed that the Secretary of State will be able to comply with the proposed amended direction.” A similar direction addressed solely to the SSD was made in MZZ ([19]).

(2) The mandatory order 

The appeal from Swift J’s order in In the Matter of the SSD related to the SSD’s non-compliance with a judgment in a judicial review of the ARR scheme. 

In April 2024, in a closed judgment, the Divisional Court (Dingemans LJ, Johnson J and Chamberlain J) quashed the SSD’s initial formulation of policy governing the ARR scheme (R (CX1 and MP1) v Secretary of State for Defence [2024] EWHC (Admin) 892 (“CX1”). The court held that the policy adopted an irrational approach to risk assessment by assuming that only individuals who held one of MoD-specified roles were at a high risk of harm from the Taliban, without accounting for the possibility that individuals holding other roles could be at equivalent risk (CX1, [56-58]). 

In October 2024, in unrelated proceedings, Swift J ordered the SSD to explain the steps the MoD had taken to amend the ARR policy following CX1. The MoD explained that formal written guidance was being drafted and the publication of the updated policy was anticipated shortly (RR and AA, [33-34]). Swift J then listed a closed hearing under the title ‘In the matter of the Secretary of State for Defence, Listed by the Court of its own motion.’ During the hearing, Swift J indicated he was “entirely unimpressed” with the MoD’s delay in amending its policies. Swift J then made a mandatory order directing the Secretaries of State for Defence and the Home Department to (1) “forthwith prepare” a revised version of the ARR policy that would contain a court-provided sentence reflecting the effect of the decision in CX1; (2) make the amended policy operative; and (3) provide advance notice of any further changes to the ARR policy to the court and to the Special Advocates’ Support Office (RR and AA, [35-38]).

The Court of Appeal judgment 

The Secretary of State for Defence appealed Swift J’s “general direction” from RR and AA and MZZ and the mandatory order from In the matter of the SSD. The Court of Appeal allowed both appeals, holding that although the situation was “wholly abnormal” and the judge made the orders “with the best motives,” he exceeded the court’s powers in doing so (RR and AA, [61]).

Concerning the limits on permissible use of general directions, the CA reasoned that although the impacts of a particular decision can “extend beyond the parties,” the starting point is that judges “decide the cases that are listed before them.” The fact that decisions in individual cases may become precedent and that senior leadership judges may issue practice guidance does not change this starting point. The court’s CPR Part 1 duties and powers likewise relate to the management of individual cases. To depart from this position, an “established procedure,” such as a group litigation order, has to be relied on ([54-57]). No such procedure was invoked by Swift J. Swift J’s position as the then-Judge in Charge of the Administrative Court and the fact that he was dealing with a defined cohort of cases did not empower him to make orders that would have “direct effect” in cases heard by other judges or in future proceedings ([63-64]). 

Concerning Swift J’s mandatory order, the CA considered that while the court can identify whether a policy made by the executive is lawful, “under our system of separation of powers,” the court “does not write policy itself” ([58]). In issuing mandatory orders, courts should pay regard to the factors identified by Lord Sales in Imam, which include the need to respect the proper separation of powers. Such respect is particularly important where, as in CX1, a national policy has been found to be unlawful ([59]). While the order made by Swift J was intended “only to reflect the effect of CX1,” “a mandatory order requiring a minister to adopt a policy nevertheless transgressed a fundamental boundary between the role of the court and the role of the executive.” The process in which Swift J made the order was also “less than fair” as the Secretary of State had been given no advance notice of the “unusual” order the judge was intending to make ([68]). 

The CA then went on to suggest what orders the judge could have made. Regarding the disclosure delays, the CA suggested that Swift J could (1) make an order requiring justifications for extensions of time to be provided in cases listed before him and (2) give a closed judgment “stating that he would expect all the responsible Ministers to draw the court’s attention” to that order in other ARAP cases ([66]). Swift J’s “legitimate objective” in making the mandatory order, in turn, could have been achieved by directing the Secretary of State to draw the closed judgment in CX1 to the personal attention of the responsible Ministers “as a means of communicating the strength of the court’s concern about the delays” (RR and AA, [69]).

The implications of the case

(1) RR and AA highlights gaps in courts’ case management powers

In RR and AA, the CA’s primary reference for interpreting the proper scope of case management powers was a situation where two (or more) private parties appear in open court. In public law litigation, however, this conception of the court’s powers is not sufficient for cases to be managed justly. 

Consider the CA’s suggested scheme of orders concerning the MoD’s delays in serving closed evidence in ARAP-related claims. The scheme envisions the judge making a discrete order in proceedings before him and giving a closed judgment stating that “all the responsible Ministers” are to draw that discrete order to the attention of judges in other cases. According to the CA, “under the self-policing duty of candour,” this would “likely” ensure that concerns about delays would be “carried over into other cases” ([66]). 

The first problem with this reasoning is that the common law duty of candour is not an obligation of general information-sharing. The duty of candour is usually framed as intended to ensure the court has the information it needs to accurately decide an issue (see e.g. Fordham J’s summary in R (Police Superintendents’ Association) v The Police Remuneration Review Body [2023] EWHC 1838 (Admin)). The CA’s reasoning appears to expand the scope of the duty to also include information the court needs to efficiently manage the proceedings. 

The second and more fundamental problem is that the facts of RR and AA cast significant doubt on whether it was “likely” that leaving the matter to the MoD would ensure that the relevant information is carried over. The “general direction” was motivated by what Swift J described as “serial” instances of “lack of candour” and “a picture of civil servants who were apparently entirely reckless as to whether the applications to the court presented a complete and accurate picture” of why the applications were made ([22]). A whistleblower described the early operation of the schemes as “chaotic” and “dysfunctional,” suffering from “inadequate staffing” as well as a lack of urgency, coordination, and expertise. The applicants’ sensitive personal data was repeatedly leaked. The MoD’s policy statements for months failed to properly reflect the obligations arising from CX1. Based on what is publicly known about the scheme, it is at the very least uncertain whether the MoD had the internal capacity to reliably comply with the duty of candour in ARAP-related cases.

The mandatory order in In the Matter of the SSD highlights a different problem with the court’s powers: its reliance on the parties to raise enforcement issues. The “normal position” in judicial review is that the court determines the issues before it, awards relief, and its duties end there. If the unlawful conduct continues, “the court can do nothing about it, unless and until a further judicial review claim is brought by the same or another claimant.” (Chamberlain J in R (ECPAT UK) v Kent County Council [2023] EWHC 2199 (Admin) at [11]). 

However, in RR and AA, the initial judgment finding unlawfulness and awarding relief (the decision in CX1) was closed. The operation of the ARR scheme itself was wholly confidential. The claimant in CX1 could not know whether the MoD was complying with the court’s order. The only parties who were (potentially) in the position to monitor compliance were the MoD itself and any judges and Special Advocates who happened to be involved with ARR-related cases. As a result, to investigate whether the judgment in CX1 has been followed, Swift J not only had to adopt a “rolling” approach to judicial review and check on the government’s compliance with CX1 of his own motion, but also do so by issuing the follow-up orders in proceedings that were largely unrelated to the ARR scheme and CX1 itself (RR and AA [33], [38]).

The CA considered the circumstances of RR and AA were “so troubling” that “the court would enjoy maximum latitude in managing cases” ([65]). However, RR and AA is not the first example of the government failing to ensure its policies are operated lawfully, give its legal team correct information, and comply with the duty of candour. The CA’s suggestion that government departments should be responsible for carrying over information related to CMPs could also create a situation in which the government is both the subject and the sole party aware of whether a (closed) order is being followed, making the “inherent unfairness” of CMPs even more acute. It may thus be that – at least where there is evidence suggesting systematic government disorganisation and the other party before the court cannot effectively monitor compliance – there is a case for statutorily expanding the courts’ case management powers. 

(2) The CA did not correctly evaluate whether the mandatory order should have been granted

On the remedial front, the difficulty with the Court of Appeal’s reasoning lies in the court’s failure to exhibit sufficient sensitivity to the facts of RR and AA and Lord Sales’ reasoning in Imam.

In RR and AA, the CoA considered that because a mandatory order was not given in Imam regarding a housing decision, there is “stronger reason” not to make the order in respect of a “national policy” being found unlawful ([59]). As a description of Imam, this is not entirely accurate. Imam focused primarily on defining the proper remedial approach to situations where a public authority says it is impossible for it to comply with its obligations, rather than on drawing distinctions between national policy and individual decisions. The Supreme Court also did not refuse to make an order: the case was remitted to the High Court for consideration of fresh evidence.

More importantly, in RR and AA, the mandatory order was not made in response to the initial finding that the MoD’s policy was unlawful. It was made in response to the MoD’s prolonged failure to amend the unlawful policy to reflect the judgment in CX1. In Imam, Lord Sales expressly considered that “if there is no sign as things stand at the time the matter is before the court that the authority is moving to rectify the situation … that is a factor pointing in favour of the making of a mandatory order” and that in case of apparent noncompliance “the imperative to galvanise the authority into taking effective steps to meet its obligations more promptly will be stronger.” (Imam, [69]). In light of the lapse of time between the judgment in CX1, as well as the Secretary of State’s (unfulfilled) assurances to the court that the policy would be amended, the application of the Imam principles would militate in favour of a conclusion that a mandatory order could be justified. 

The factual matrix facing the judge in RR and AA was also coloured by the presence of the super-injunction. As Swift J noted, at the time the order was made, there was “no possibility that the relevant part of the judgment of the Divisional Court in CXI will be made public” the parties to proceedings also could not monitor the MoD’s compliance with the decision. The super-injunction “prevented all public scrutiny of the proceedings,” including Parliamentary scrutiny, for almost two years ([5]). The situation in RR and AA could not have been further from the ordinary constitutional position, in which the actions of the executive are scrutinised by both the Parliament and the general public (and any potential litigants); the CoA itself acknowledged that “the situation … was wholly abnormal.” ([61])

It may be the case that even taking the above into account, Swift J’s mandatory order still would have fallen afoul of the limits of judicial control over policy. However, the government’s prolonged failure to implement a court order and the constitutional abnormality of governance in total secrecy were relevant to the question of what the separation of powers required in the circumstances. The Court of Appeal’s reasoning did not engage with these factors.

Conclusion 

In its judgment, the Court of Appeal praised Swift J for his “crucial role in ensuring that justice could still be done” ([61]). However, RR and AA illustrates that in public law proceedings – and particularly closed proceedings – the courts may lack the tools, procedural and remedial alike, to play this role. 

I am greatly indebted to Dr. Hayley J. Hooper and Sahil Thapa for their helpful comments and suggestions. All remaining errors are my own.

Weronika Galka is a current Bar Course student and a future pupil at Essex Court Chambers, London.

(Suggested citation: W. Galka, ‘Secret proceedings, (mal)administration, and the courts: RA and AA v SSFCDA, MZZ v SSD and SSHD and In the Matter of the SSD [2026] EWCA Civ 3’, U.K. Const. L. Blog (26th February 2026) (available at https://ukconstitutionallaw.org/))