Campaign Against Arms Trade, R (on the application of) v Secretary of State for International Trade

INTRODUCTION

BACKGROUND – THE CONFLICT IN YEMEN

THE LEGAL CONTEXT

THE DECISION UNDER CHALLENGE

THE GROUNDS IN OUTLINE

DOMESTIC PUBLIC LAW PRINCIPLES

OPEN SOURCE REPORTING

THE DECISION-MAKING PROCESS

THE RIVAL SUBMISSIONS

DISCUSSION AND ANALYSIS

GROUND 4

GROUND 3

GROUND 1

GROUND 2

CONCLUSIONS

INTRODUCTION

This is a claim for judicial review challenging the lawfulness of the decision on 7 July 2020 by the Secretary of State for International Trade to continue granting licences for the export of arms to the Kingdom of Saudi Arabia (to which we will refer simply as Saudi Arabia). She is obliged to cease granting new licences and suspend existing licences where there is “a clear risk that the arms might be used in the commission of a serious violation of International Humanitarian Law”. The claim focusses on airstrikes conducted by a coalition led by Saudi Arabia in support of the legitimate government of Yemen against the Houthi rebellion.

The claim follows an earlier challenge to the lawfulness of a similar decision by the UK Government in 2015. That claim was dismissed in a closed and open judgment ([2017] EWHC 1754 Admin) by the Divisional Court (Burnett LJ and Haddon-Cave J). An appeal was allowed by the Court of Appeal (Sir Terence Etherton MR, Irwin and Singh LJJ), in a closed and open judgment ([2019] EWCA Civ 1020 [2019] 1 WLR 5765), on one ground, namely that, as part of the process in reaching the decision, the Secretary of State should have attempted to assess the extent to which past incidents involved a breach of International Humanitarian Law (‘IHL’) by Saudi Arabia. She has done so in reaching her 2020 decision which is now under challenge, but the Claimants (‘CAAT’) and the Interveners (‘Mwatana’ and ‘Oxfam’) contend that the process and decision are flawed on public law grounds and should be undertaken afresh. The main thrust of the challenge is that the Secretary of State has acted irrationally in reaching her decision.

The Secretary of State resists the Claimant’s case with the aid of open evidence and argument. In addition, there is a closed case. Jay J made a declaration pursuant to section 6 of the Justice and Security Act 2013 enabling the Secretary of State to rely upon closed material without disclosure to CAAT, Mwatana and Oxfam. The statutory procedure has been followed so as to ensure that anything which can be disclosed without damage to the relevant public interests is produced to CAAT and Mwatana, and anything which cannot has been disclosed to the Special Advocates. After the conclusion of the open argument, we heard further submissions in a closed hearing attended by the Secretary of State and the Special Advocates. This is our open judgment in which we explain our reasons for dismissing the claim. Further reasons are given in a closed judgment.

Much of the uncontroversial background is set out in the Divisional Court and Court of Appeal judgments. In order to make this judgment self-standing we have, in places, borrowed from and adopted the language of both judgments, rather than just cross-referring to them, for which we acknowledge a debt of gratitude.

BACKGROUND – THE CONFLICT IN YEMEN

The following is a very brief summary of the conflict in Yemen. It draws on and updates the fuller and more detailed account in the open judgment of the Divisional Court.

Saudi Arabia and Yemen are contiguous and share a 1,800 km border. Since early 2015 Yemen’s capital city, Sana’a, and parts of central and southern Yemen have been in the control of Houthi rebels backed by former Republican Guard Forces loyal to former President Saleh. The Houthi are a Shia-Zaydi movement from the north of Yemen.

On 24 March 2015 the President of Yemen, President Hadi, wrote to the UN requesting support “by all necessary means and measures, including military intervention, to protect Yemen and its people from continuing aggression by the Houthis”. A further letter was sent on 26 March 2015 from the Gulf Cooperation Council countries endorsing President Hadi’s request.

On 25 March 2015 a coalition of nine states led by Saudi Arabia (the other members being Egypt, Morocco, Jordan, Sudan, the United Arab Emirates, Kuwait, Qatar and Bahrain) responded to a request for assistance by President Hadi and commenced military operations against the Houthis in Yemen (‘the Coalition’).

On 14 April 2015 the UN passed Security Council Resolution 2216 (2015) affirming the legitimacy of President Hadi and condemning the unilateral actions taken by the Houthis.

Hostilities have taken place from 2015 to the present day with only occasional cessation, despite numerous ceasefire attempts. Coalition military operations took the form primarily of airstrikes led by Saudi Arabia against the Houthis, together with some ground operations. The Saudis reported numerous cross-border incursions and missile attacks by the Houthis, including the use of SCUD missiles. There were reports of attacks by Houthi forces on Coalition shipping in the Red Sea.

Terrorist organisations, such as Al-Qaeda in the Arabian Peninsula and Daesh, have taken advantage of the on-going instability and ungoverned space in Yemen. This has complicated the picture and led to increased anti-terror operations in the region led by US forces.

Early in the conflict the Houthis made substantial gains before being pushed back by the Coalition. Since then frontlines have generally been static and the fighting in stalemate. The UAE withdrew the majority of its forces from Yemen in 2019.

By June 2021 the Coalition had launched tens of thousands of air delivered weapons, resulting from a far greater number of sorties.

A defence intelligence document dated 14 August 2019 records key judgments about Houthi warfighting techniques, including as follows. The civilian population is used to assist Houthi defensive tactics. Houthi fighters do not generally wear military uniform and also avoid the use of military type vehicles, preferring to use civilian vehicles. This is almost certainly a deliberate policy of using civilians as cover from airstrikes. Coalition airstrikes and air superiority over western Yemen have caused the Houthis to use residential areas, hospitals and mosques as cover for military operations and the basing and storage of weapons and ammunition. A significant proportion of Houthi soldiers have been children, possibly as many as 18,000. A sophisticated Iranian inspired info ops campaign, typically promulgating both accurate and inaccurate accounts of Coalition airstrikes, “is a significant force multiplier for the Houthi campaign.” Further detail is given in the closed judgment.

These fighting and propaganda tactics form important context to the allegations and assessments of IHL breaches, as does the challenging nature of the air campaign, which Mr Lapsley of the Ministry of Defence (‘MOD’) summarises in this way in his witness statement:

It is by any standards a complex and attritional conflict which the evidence from Oxfam illustrates has had devastating humanitarian consequences.

THE LEGAL CONTEXT

The Export Control Act 2002

The export of arms and military equipment from the UK to Yemen is regulated by the Export Control Act 2002 (‘the 2002 Act’). Section 1(1) of the 2002 Act provides that the Secretary of State may by order make provision for or in connection with the imposition of export controls in relation to goods of any description. Section 9(2) provides that the Secretary of State may give guidance about any matter relating to the exercise of any licensing power or other functions conferred by a control order, and section 9(3) provides that the Secretary of State must give guidance about the general principles to be followed when exercising any such licensing power.

Article 32(1) of the Export Control Order 2008 (SI 2008/3231) provides that the Secretary of State may by notice amend, suspend or revoke a licence granted by the Secretary of State.

The Consolidated Criteria

On 24 March 2014 the Secretary of State set out in a written statement to Parliament what were described as “Consolidated EU and National Arms Export Licensing Criteria” (“the Consolidated Criteria”). The written statement said that it was guidance given under section 9 of the 2002 Act. It said:

Criterion 2 of the Consolidated Criteria is the relevant criterion for the purpose of these proceedings, providing, so far as material:

The Parliamentary written statement also said:

The User’s Guide

Criterion 2 derived from an agreement between the Member States of the European Union contained in the Council Common Position 2008/944/CFSP of 8 December 2008 “defining common rules governing control of exports of military technology and equipment” (‘the EU Common Position’). The Consolidated Criteria continue to apply notwithstanding the UK’s withdrawal from the European Union.

Article 13 of the EU Common Position referred to a “User’s Guide” which is to “serve as guidance for the implementation of the Common Position”. The most recent version of the User’s Guide is dated 16 September 2019. Chapter 2, which has the title “Criteria Guidance”, sets out “best practices” for the criteria in the Common Position, including Criterion 2. The introduction to Chapter 2 describes its purpose as follows, so far as relevant:

Paragraphs 2.1 to 2.15 of Chapter 2 of the User’s Guide address Criterion 2. They include the following:

Principles of International Humanitarian Law

The Court of Appeal Judgment identified the relevant principles of IHL at paragraphs 23 to 25, reflecting the Divisional Court’s formulation, in the following terms:

THE DECISION UNDER CHALLENGE

The decision (‘the New Decision’) was announced in a written statement by the Secretary of State to Parliament of 7 July 2020. The Parliamentary Statement summarised the essence of the Court of Appeal’s decision on the ground that succeeded thus: “In the Court’s judgment, the question of whether there was an historic pattern of breaches of IHL was a question which required to be faced. Even if it could not be answered with reasonable confidence for every incident, at least the attempt had to be made. It was because we had not reached findings on whether specific incidents constituted breaches of IHL as part of our assessment of clear risk, under Criterion 2c, that the Court of Appeal concluded that our decision-making process was irrational and therefore unlawful.”

That was an accurate summary. The essence of the Court of Appeal’s decision as to the exercise which the Secretary of State should have undertaken prior to the 2015 decision, and should undertake in making subsequent decisions on the application of Criterion 2 to the export of arms to Saudi Arabia, is contained in the following paragraphs of its open judgment:

The Parliamentary Statement went on to explain that to address the failing, a revised methodology had been adopted in respect of all allegations assessed as likely to have occurred and to have been caused by fixed wing aircraft. Each had been subjected to analysis as to whether it was possible that it involved a breach of IHL or whether that was unlikely; in a number of cases there was insufficient information to reach a view. All “possible” breaches were then treated for the purposes of the Criterion 2c analysis as if they were breaches of IHL. Some incidents had been analysed as possible violations. The analysis had not identified any patterns or systemic weaknesses. It was noted in particular that the incidents which had been determined to be possible violations of IHL occurred at different times, in different circumstances and for different reasons. The conclusion was that these were isolated incidents.

The statement went on to emphasise that the IHL analysis was only one part of the Criterion 2 assessment, and that, in retaking the decision, the Secretary of State had taken into account the full range of information available to the Government. In the light of all that information and analysis, she had concluded that, notwithstanding the isolated incidents which had been factored into the analysis as historic possible violations of IHL, Saudi Arabia had a genuine intent to comply with IHL and the capacity to do so. On that basis she had assessed that there was not a clear risk that the export of arms and military equipment to Saudi Arabia might be used in the commission of a serious violation of IHL.

THE GROUNDS IN OUTLINE

The challenge was advanced on four grounds:

Mr Jaffey KC did not shy away from the fact that the first three grounds, which constituted the main thrust of the challenge, could only succeed on the basis that the Secretary of State had in those respects acted irrationally. The essence of the challenge in this case, as he summarised it at the outset of his submissions, was that the Secretary of State had irrationally failed to identify the extent of past breaches of IHL by Saudi Arabia and had irrationally concluded that there was no pattern to past breaches.

In this context it is important to keep in mind the relevant principles of domestic public law, which we summarise next.

DOMESTIC PUBLIC LAW PRINCIPLES

We agree with the observations of the Divisional Court at paragraphs 27 to 35 of its open judgment, which we here quote in full. No doubt was cast on them by the Court of Appeal’s judgment and the concluding paragraph was expressly approved.

The Court of Appeal said the following at paragraphs 54-57, 94 and 165 of its open judgment:

These observations derive further support from the recent Supreme Court decision in R (Begum) v Special Immigration Appeals Commission [2021] UKSC 7 [2021] AC 765.

We should also refer to what the Court of Appeal said about the Tameside duty to make inquiries, which is an aspect of rationality, at paragraph 59:

OPEN SOURCE REPORTING

There are two important United Nations sources of reporting. UN Security Council Resolution 2140 (2014) established a panel of experts (‘UN Panel of Experts’) for the purposes of investigating violations of IHL by all sides to the conflict, for purposes of advising the Security Council Sanctions Committee on Yemen. Another panel, the UN Group of Eminent Experts, was established by the UN Office of the High Commissioner for Human Rights following a resolution passed by the Human Rights Council, with a specific mandate to investigate violations of IHL and International Human Rights Law (‘UN Group of Experts’). The work of each body is assisted by senior officials with expertise in IHL and military operations.

The UN Panel of Experts’ conclusions are based on eye-witness testimony, although this is second hand because it does not have representatives on the ground in Yemen. The Court of Appeal open judgment recorded at [91] that the panel did, however, have access to documents provided by member states, satellite imagery from “private providers”, and commercial databases recording “maritime and aviation traffic, and social and media traffic”, but that in the case of the latter the panel relied on it only where it could be corroborated using multiple independent sources.

Reputable non-governmental organisations (‘NGOs’) including Human Rights Watch, Amnesty International and Mwatana have conducted detailed ground investigations into alleged violations in Yemen. Mwatana is a human rights organisation based in Yemen, founded in 2007, and publishes regular reports on the conflict. Its personnel on the ground interview eye witnesses, save where they are unable to access the territory. Reports from others such as Médecins Sans Frontières (‘MSF’) and the International Committee of the Red Cross (‘ICRC’) are sometimes available.

The Court of Appeal in its open judgment considered at length the submission on behalf of the Secretary of State that the Government was far better placed than the UN Panels and the NGOs to assess violations; and the rival submissions: see paragraphs [68], [86]-[92], [94], [96], [102], [104], [119]-[121]. Its conclusions on that aspect were expressed in the following paragraphs:

THE DECISION-MAKING PROCESS

The government departments involved

Three government departments were involved, in the same way as at the time of the first decision in 2015, namely the Department of International Trade (‘DIT’), MOD, and the Foreign Commonwealth and Development Office (‘FCDO’). The Export Control Joint Unit (‘ECJU’), which has responsibility for processing export licence applications, is a joint unit across all three departments with about 100 officials drawn from them. Those from the MOD are styled ECJU-MOD and those from the FCDO styled ECJU-FCDO (the latter previously being called the Arms Export Policy team). ECJU-FCDO provide advice to the DIT on about 16,000 licence applications each year covering a wide range of equipment and destinations. In relation to the export of military equipment to Saudi Arabia, the FCDO, through the officials in ECJU and the Foreign Secretary, provide advice to the International Trade Secretary, who has the statutory function of making the decision.

IHL Updates

‘IHL Updates’ are jointly written by MOD’s Security, Policy and Operations organisation (‘SPO’) and FCDO’s Middle East and North Africa Directorate (‘MENAD’). MENAD has a dedicated Yemen team within it, which provides advice and information on developments in Yemen to the ECJU. MENAD carefully monitors developments in Yemen via receipt of at least biweekly and often daily emails from the UK’s diplomatic team covering political and military activity in Yemen. These are based in Riyadh or Amman. MENAD is also in regular contact with NGOs with an interest in the area, through formal meetings and informal communications. Since October 2015 MENAD has produced regular updates which specifically address IHL risks, for regular submission to the Foreign Secretary. The ‘IHL Updates’ were considered in some detail in the Divisional Court judgment which recorded at [150] and [151]:

The process for producing IHL Updates, the information which feeds into them, and their role in the decision-making process is essentially the same as it was at the time of the Divisional Court judgment, save that the IHL Updates have since become considerably more detailed and have been produced on a quarterly rather than monthly basis. The IHL Updates are the main vehicles by which information is fed into ECJU-FCDO. On receipt of each Update, ECJU-FCDO completes an overarching assessment of Saudi Arabia’s record, attitude and capability in relation to compliance with the key principles of IHL. It is subsequently submitted to the Foreign Secretary in the form of a recommendation as to the advice he or she should give to the International Trade Secretary on the question whether the threshold for refusing licences under Criterion 2c has been met. These submissions are cleared at senior civil servant level within the FCDO. In addition to the regular IHL Updates, ad hoc updates are provided to the Foreign Secretary or relevant Ministers when it is necessary to inform them of specific matters between reporting periods.

The ‘Tracker’ and the IHL Analysis

In his evidence to the Divisional Court in relation to the first challenge, Mr Watkins of the MOD explained that the MOD monitors media and NGO reports for allegations of breaches of IHL. These are recorded by officials in a central database known as ‘the Tracker’.

Since the Court of Appeal judgment, the Tracker has been expanded and developed so as to be the main tool used in carrying out an analysis of whether and to what extent Saudi Arabia has been in breach of IHL. This analysis is what has been referred to, both in the contemporaneous documents and in submissions, as “the IHL Analysis”. The Tracker records all alleged incidents which the MOD considers might require monitoring or investigation, including any which have been identified as potentially involving a breach of IHL by UN bodies, NGOs or the Saudi Arabian Joint Incident Assessment Team (‘JIAT’). JIAT is a Coalition-staffed investigatory body, which issues press releases. We consider its effectiveness and independence further below and in our closed judgment. With one exception, to which we refer in the closed judgment, all the incidents upon which CAAT, Mwatana, and the Special Advocates sought to place reliance were included on the Tracker and subjected to the IHL analysis. In October 2019, when the main IHL analysis for the New Decision was complete, there were 486 incidents on the Tracker.

The Tracker assigns a serial number to each incident and has a column showing the date it was added to the Tracker. The IHL analysis conducted by the MOD involves each of the incidents being examined by a panel comprising an MOD lawyer with experience of operational decision-making, an airman with operational experience of air-delivered weapons, and a third member with an understanding of the operational context in Yemen. These are personnel with experience gained from involvement or advice in relation to the compliance with IHL of the UK’s own military and targeting activities in real time.

The Tracker has columns arranged under four broad headings (1) “What allegedly happened?” (2) “What do we know?” (3) “Did the alleged events take place and who was responsible?” and (4) “Was IHL breached?” The detail of the Tracker is considered further in the closed judgment.

The MOD analysis filters out cases in which there is no credible allegation of the incident occurring or insufficient information to be able to make that evaluation.

In the analysis of whether there was a breach of IHL, there are identified cases where there is insufficient information to make an evaluation. For the others, an evaluation is then made, using four principles of IHL, namely proportionality, feasible precautions, necessity and distinction identified in the column headings, “as to whether it is possible that the incident constituted a breach of IHL; or whether it is unlikely that it represents a breach.” This language is taken from the Decision Paper and discussed further below. All remaining “possible” breaches are then considered as part of the information feeding in to the Criterion 2c assessment, and are for this purpose treated as established breaches.

Other information which informs the Criterion 2 assessment

In addressing the question whether there is a clear risk of a serious violation of IHL, the MOD, the ECJU, the Foreign Secretary and the International Trade Secretary focus on the three key factors identified in paragraph 2.13 of the User’s Guide, namely (i) Saudi Arabia’s past and present record of respect for IHL, (ii) Saudi Arabia’s intentions as expressed through formal commitments and (iii) Saudi Arabia’s capacity to ensure that equipment or technology is used in a manner consistent with IHL.

The analysis is informed by the following particular strands of information and analysis:

We consider these strands in more detail below.

The Decision Paper

In December 2019, a paper was prepared addressing the IHL analysis and broader issues in order to provide advice to enable the New Decision to be taken, concluding with an overall assessment against the Criterion 2c test carried out by ECJU-FCDO (‘the Decision Paper’). It ran to 47 pages plus voluminous Annexes and is an important document in the case.

The structure of the Decision Paper was that Section I contained a summary; Section II explained the methodology employed by the MOD in carrying out the IHL Analysis for the incidents on the Tracker; Section III recorded and assessed the outcome of the IHL Analysis, including detailed consideration of the allegations which had been assessed as “possible” breaches of IHL; Section IV contained a “thematic analysis” of other relevant information regarding Saudi Arabia’s attitude to compliance; and section V contained ECJU-FCDO’s analysis of risk for the purposes of the Criterion 2c assessment.

In Section I, “the IHL Analysis” was the defined term given to the analysis of incidents on the Tracker by MOD in an attempt to determine the possibility that any such incidents constituted breaches of IHL and/or whether there were patterns of possible violations. Paragraph 7 emphasised that the IHL Analysis applied to past events was just one aspect of the risk assessment, and cited paragraph 2.13 of the User Guide. In doing so it added its own emphasis to the word “include” before reference to the three factors of past record, intentions and capacity, indicating that although those were the focus they were not exhaustive as to what ultimately informs the assessment of risk required by Criterion 2c. Paragraph 10 said that the final section, ECJU-FCDO’s Criterion 2c analysis, was informed by (a) the IHL Analysis; (b) the analysis of thematic trends drawn from the IHL Updates, including analysis of the training provided to Saudi Arabia and broader issues, positive and negative; (c) the UK’s knowledge of the development of Saudi Arabian systems, including reflection on the impact they had on reported credible allegations; and (d) an overall “stand back” analysis.

Section II identified the relevant principles of IHL from paragraph 2.10 of the User’s Guide and the paragraphs from the Divisional Court and Court of Appeal judgments. Paragraph 18 observed that there was little jurisprudence available, or evidence of state practice, on breaches of IHL as a matter of state responsibility; and there remained difficult substantive and evidential questions as to the circumstances in which the state is responsible, as a matter of IHL, for mistakes which cause civilian casualties.

Paragraph 19 of the Paper went on to observe that although the UK had a close relationship with Saudi Arabia and an unusual level of access to information as a result, it was nevertheless very difficult to reach any confident conclusions as to whether specific incidents constituted breaches of IHL because the UK does not have, and would not expect to have, full insight into airstrikes undertaken by the Coalition (of which it is not a member). It cited paragraph 181(ii) of the Divisional Court’s open judgment summarising the difficulties, where it was said:

The Decision Paper went on at paragraph 21 to say that the experience of the MOD in attempting the exercise had borne out the concerns expressed by the Divisional Court in the passage quoted. In practice, it said, it had been difficult if not impossible to make any reliable assessment in the absence of a JIAT investigation giving further details of (i) the intended target; (ii) the specific intelligence which led to that targeting; (iii) the surveillance and reconnaissance carried out ahead of a strike; (iv) the weapon used; and (v) other steps taken to identify and minimise potential civilian casualties or damage to civilian infrastructure. Even in those cases where there had been a JIAT investigation, the summaries provided to the UK did not contain sufficient information for the UK to reach a definitive conclusion.

Paragraphs 22 and 23 have given rise to a particular dispute as to the methodology applied and merit quotation in full:

The Decision Paper went on to explain at paragraph 24 some aspects of the methodology to which we have already alluded, and revealed that the IHL Analysis had been focussed initially on the credible incidents likely to have been caused by Saudi Arabia, and extended to those where the airstrikes were attributable to other Coalition members or where it had not been possible to attribute it to any particular state. The Tracker also recorded trends, not only in relation to cases treated as possible breaches of IHL but in respect of all cases of credible incidents, including those where other Coalition partners were identified as responsible.

The final paragraph of Section II said that the IHL Analysis was informed by:

the intelligence and information obtained as a result of HMG’s privileged access to Saudi Arabia and its various other sources of imagery etc, including of particular importance its knowledge of and involvement in the development of Saudi systems;

the broader contextual knowledge and intelligence;

information obtained from JIAT investigations. Material relating to JIAT which was relied on was contained in Annex 2 which we consider in our closed judgment; and

other direct engagement with high level military and political contacts in Saudi Arabia. In relation to some incidents, HMG had initiated immediate, intensive and high-level engagement across the Saudi Arabian government and military, adding that “[i]t is a matter of political and diplomatic judgment as to when it is necessary or appropriate to engage in this way.”

Section III started with an overview of the number of allegations. It observed that in broad terms the number of credible allegations - that is, allegations for which the MOD assessed that the alleged event is likely to have happened - had decreased significantly since the conflict started in 2015, although there had been “spikes” in December 2017 and February and April 2018. The trend was supported by bar charts annexed, representing the monthly figures. Three particular points were evident from this analysis of credible allegations, which it was emphasised related not merely to the possible breaches of IHL but covered all the incidents where it was found to be credible that they had occurred:

The results of the MOD’s IHL Analysis were summarised at paragraph 31 of the Decision Paper.

In addition to the small number of incidents thus assessed by the MOD as involving “possible” breaches of IHL, a further cross check was conducted on all incidents: (a) which were highlighted by the UK to Saudi Arabia; (b) which were noted in the narrative sections of the IHL Updates as being of particular concern; (c) which were highlighted by the UN Panel of Experts in its (then) three reports; or (d) for which JIAT had recommended prosecution of individuals or the payment of compensation. These were identified in Annex 5. Of these:

In the following 42 paragraphs the Decision Paper addressed in detail each of the small number of possible breaches by Saudi Arabia and the very small number where attribution was unknown and so might be attributable to Saudi Arabia. The detail of each of them is discussed in our closed judgment.

Section III concluded with five paragraphs under the heading “Conclusions on IHL Analysis.” The first, paragraph 77, emphasised that a relatively low threshold of “possible” had been applied and the small number of incidents captured by it therefore covered a wide range of degrees of likelihood and gravity. Paragraph 78 said that the possible breaches of IHL identified did not indicate a pattern of violations which would give rise to serious concerns regarding Saudi Arabia’s capacity or commitment to comply with IHL. Rather, the detail set out was consistent with a limited number of errors, well within the margin that would be expected in a conflict of this nature. Paragraph 79 said that it was concerning that half of the possible incidents had occurred more recently, but that it was assessed that they were a disparate group of incidents, raising their own particular issues, rather than a grouping of a kind that might lead to a concern about systems generally. Paragraph 80 said that it was also concerning that some of the incidents raised serious concerns. However, the context of each of these incidents was different and it was not assessed that these incidents gave rise to wider concerns.

Paragraph 81 said that the broader review of the flagged incidents of concern in Annex 5 did not suggest that there was any trend in the explanation for these incidents which gave rise to concerns about Saudi Arabia’s attitude or capability in respect of compliance with IHL.

Section IV of the Decision Paper was a “Thematic analysis”, that is to say a discussion of considerations other than the IHL Analysis for past breaches, which fed into the ultimate decision of whether there was a clear risk of a serious breach of IHL in the future. In summary, it included the following, and the further matters included in our closed judgment:

The concluding paragraphs to the Thematic Analysis said that Saudi Arabia had given frequent and consistent assurances to the UK of its commitment to ensure compliance with IHL. Those assurances were given at the highest political and military levels. Significantly, this commitment was also evident at a tactical and operational level, through the attitude of participants in the many training courses which had been provided by the UK and US and through other matters identified in our closed judgment. Saudi Arabia had shown that it was able to identify problems and weaknesses and to ask for and be receptive to assistance. It had reacted positively and swiftly to engagement from the UK and US in relation to the two incidents of greatest concern (the Great Hall strike and the Dhayan bus incident). The UK had monitored Saudi Arabia closely and no further cause for concern had been observed.

The final section of the Paper was ECJU-FCDO’s assessment applying the Criterion 2c test. It recorded that the IHL Analysis of the “possible” breach incidents and the Thematic Analysis had been used to inform an assessment of Saudi Arabia’s record, attitude and capability in relation to compliance with the key principles of IHL, and briefly summarised the conclusions in relation to each and aspects of what was in the Paper which supported them. It is sufficient for present purposes to give only a brief summary, although the section runs to 33 detailed paragraphs. The long-term trends and statistics were encouraging. There was a tiny proportion of “possible breaches” by comparison with the number of air strikes and sorties, and a comparable picture for the number of credible allegations to those of the coalition forces in Iraq/Syria. As to the IHL Analysis of the possible breaches, those which revealed failings which could be addressed by changes in practices had resulted in such changes being put in place. There was no pattern to them which could in itself indicate a clear risk of future IHL violations: where a problem had been highlighted the Saudis had taken action to remedy the situation, thereby mitigating the risk. As to attitude, this approach itself demonstrated a determination to avoid IHL breaches, which was also apparent from the extent of engagement at all levels, receptiveness to assistance, advice and training from the UK and the US, the creation and development of JIAT, development of rules of engagement, and other matters we refer to in our closed judgment. As to capability, Saudi Arabia had demonstrated improvements in their targeting processes in the many ways identified in the Decision Paper and had growing expertise in areas of critical importance to IHL compliance, such as targeting. The conclusion was that there was not a clear risk that the proposed exports would be used in a serious violation of IHL.

Annex 5 to the Decision Paper contained a summary of incidents. These were all those which at the time had fallen into one or more of four categories, namely those raised by the Panel of Experts, those noted in the IHL Updates, those for which JIAT had recommended prosecution or payment of compensation, and incidents which had been specifically raised by the UK with Saudi Arabia (together, ‘the Annex 5 incidents’). Annex 5 revealed for each incident whether it had been categorised as insufficient information, not credible, unlikely violation or possible violation.

JIAT

Annex 2 to the Decision Paper contained information in relation to JIAT. At the time of the Divisional Court’s judgment only some 14 incidents had been the subject of JIAT reports. Since then, there have been something like 187 reports and they have become more detailed. The MOD has supported the development of JIAT and delivered training sessions in Saudi Arabia on the process of investigating alleged violations of IHL. We give further detail of Annex 2 in our closed judgment.

The process from Decision Paper to New Decision

The Decision Paper was annexed to a submission sent by ECJU-FCDO to the Foreign Secretary on 23 January 2020, which formally recommended that the Foreign Secretary agree with the methodology used and conclusions reached, and with a recommendation to seek certain further legal advice. The submission recorded the view of ECJU-FCDO that the wealth of analysis undertaken and the evidence in the round pointed to a conclusion, that there was no clear risk of the equipment being used in a serious violation of IHL, which was “less finely balanced than before”. This was a reference to the first decision in 2015 having been described at that time as finely balanced.

On 13 February 2020 the Foreign Secretary agreed that the methodology was in line with what was required by the Court of Appeal’s judgment.

On 20 April 2020 a joint submission was made by ECJU-FCDO and MENAD to the Foreign Secretary, which took account of the latest quarterly IHL Update covering the period to January 2020, and recommended that the Foreign Secretary agree that there remained no clear risk of a serious IHL violation. On 27 April 2020 the Foreign Secretary gave such agreement.

On 7 May 2020, a further joint submission was made to the Foreign Secretary, dealing with a MENAD ad hoc interim note addressing matters arising in the period from February to May 2020, concluding that there was no change to the assessment of risk and recommending that the Foreign Secretary advise the International Trade Secretary accordingly: which he did in a letter dated 12 May 2020.

For various reasons, including Parliament being in recess and the COVID 19 restrictions, the International Trade Secretary was not able to announce the New Decision until 7 July 2020. In the meantime, the usual process of monitoring and review continued. An ad hoc update was provided to the Foreign Secretary on 17 June 2020 to deal with an incident which had occurred on 15 June 2020, but this did not alter the assessment of the Criterion 2c risk. On 24 June 2020 ECJU-DIT was told by ECJU-FCDO that the latest quarterly IHL Update covering the reporting period to 30 April 2020 was available, but that unless there was anything in it which changed their advice on the Criterion 2c assessment of risk, they did not intend to submit further advice or information to the Foreign Secretary. There was nothing in that Update or subsequent information which changed ECJU-FCDO’s risk assessment at any time before the New Decision was taken.

The “pattern” emails

On 26 and 27 April 2021 the MOD IHL panel, comprising legal, targeting and policy experts, met to consider the incidents treated in the IHL Analysis as possible breaches, now slightly more than those identified at the time of the Decision Paper, to consider whether there might be patterns to them. In an email sent by the MOD’s Security, Policy and Operations organisation (SPO) to the head of ECJU-FCDO on 4 May 2021, their conclusion was that no meaningful patterns could be identified because the “possibles” covered a very wide variety of mission types, weapons, platforms and circumstances; and a number concerned some of the most challenging mission types and circumstances. Further detail is included in the closed judgment.

The SPO responded in an email of 25 May 2021 explaining, as we interpret it, that the point they were making was that if the putative “pattern” was expressed at the level of generality of two of the four pillars of IHL being applied, namely proportionality and feasible precautions, then it would be wholly unsurprising if most or all of the possible violations fell into such a category, because these were the categories likely to be engaged in air strikes in the circumstances of the Yemen conflict. It was, however, too much of a leap to say that such commonality must be a concerning pattern. When examining the possibles, the MOD panel had had very much in mind consideration of whether the actual nature of any commonality might evidence a problematical pattern.

Reassessment of risk after the New Decision

One incident which was flagged in the 7 May 2020 submission involved three airstrikes targeting Houthi militia in the vicinity of a downed Saudi fighter jet in the Al Masloub District in Al Jawf Governorate, which had occurred on 20 February 2020. In the 7 May submission, the ECJU-FCDO noted that they would await the JIAT report and examine any other information which came to light, although the apparently unique circumstances did not seem to be part of any pattern and the incident did not therefore indicate any increased risk of a violation of IHL.

After the New Decision had been taken, in due course further information came to light in relation to the incident in the form of a JIAT report released at the end of November 2020, and further subsequent communications with Saudi officials. We consider this incident in more detail in our closed judgment. It is sufficient here to record that further consideration led to ECJU-FCDO providing an update of their assessment to the Foreign Secretary on 5 February 2021, in which they concluded that the incident did not change their assessment of the Criterion 2c risk.

THE RIVAL SUBMISSIONS

The head of ECJU-FCDO responded in an email the same day suggesting that what was said about some of the possibles was a bit circular, “paraphrased as it’s not a pattern because it’s what we would expect”.

Mr Jaffey made a number of submissions by way of background to his criticisms of the IHL Analysis. The principal ones were these:

Against this background, the criticisms which he advanced on the basis of the open material can be summarised as follows.

This last incident post-dated the New Decision, but Mr Jaffey contended that it was legitimate to consider incidents after that date because the judicial review challenge was expressed to be to the ongoing failure of the Secretary of State to suspend licences; and the Court of Appeal had accepted in the previous challenge that that justified taking into account material relating to incidents after the date of the decision which was then under review. Sir James Eadie KC, for the Defendant, emphasised that in contrast to the position in the first challenge, where the Secretary of State had undertaken an updated exercise to the date of the Divisional Court hearing, she had made clear that she would not do so for this challenge; in this challenge, post decision assessments had been put in evidence only pursuant to the duty of candour. It was therefore a case in which, strictly speaking, the general rule applied that the question was confined to the lawfulness of the decision under challenge on the basis of the material which was or ought to have been available to the decision maker on that date. However, Sir James was content for later material which was before the court to be taken into account and suggested that it would make no difference to the outcome. We agree with that last submission and have taken such material into account.

Mr Jaffey submitted that the Secretary of State made two errors in relation to pattern. The first was an error as to what amounted to a pattern, and the second her conclusion that there was no pattern. Mr Jaffey recognised that the second would require consideration of the closed material. In relation to the first he submitted that what we have described as “the pattern emails” demonstrated that the reasoning in relation to patterns was circular and logically flawed. He defined the relevant pattern or patterns in various different ways during his submissions, including attacking targets which ought to be immune from attack, targeting impermissible targets, breaches of rules of engagement, failure to check whether the targets were permissible or whether civilians were present, failure to consider proportionality, and failure to check against no strike lists. These were, he said, manifestations of what must be an underlying root cause, namely the unwillingness or inability to implement IHL by Saudi Arabia (Saudi Arabia meaning, in this submission, personnel at any level, from aircrew upwards). He described this as a “general approach” and a problem which was “systemic in nature”. He said that his 8 illustrative examples demonstrated systemic failings by those responsible for targeting. He placed particular emphasis in his submission on a pattern of unwillingness or inability to follow rules of engagement. In his reply he accepted that a relevant pattern would be one which was indicative of a systemic weakness.

Again, Mr Jaffey recognised that this would have to take into account the closed material, but emphasised that all that was required was a risk of a single serious violation, and that the open material suggested numerous violations and the absence of any pattern or trend of improvement. In those circumstances a risk of a further serious violation was the only rational conclusion to reach from the analysis of past breaches.

Mr Jaffey said that he would not press this ground if Sir James Eadie confirmed his understanding that the Government accepted that a single incident could constitute a serious violation of IHL, and that such a violation might occur without intentional wrongdoing by any individual. Sir James confirmed both those points, whilst emphasising that a single past serious violation did not equate to a real risk of a future serious violation, and that many of the criteria upon which Mr Jaffey relied did specifically require intent or something like it (which we took to mean recklessness). Sir James also submitted that, as Mr Jaffey accepted, the mere fact of civilian casualties did not mean that there was a breach of IHL, nor was it sufficient of itself to make any breach a serious one.

As to “impunity”, Mr Jaffey’s point was that although the Court of Appeal in the previous challenge had rejected any criticism of the Secretary of State in relation to considering steps taken, or to be taken, to discipline or punish those responsible for incidents, the landscape had changed as a result of the IHL Analysis. Previously, there had been no conclusion of any past violations, whereas in the new process there had been found to be a number of past violations. That required the Secretary of State to reconsider the issue of impunity, and her failure to do so was a public law error.

Mr McClelland for the Mwatana supported all Mr Jaffey’s arguments and added his own emphasis to some of them. He submitted that the agnostic conclusion that there was “insufficient information” in about half the cases could not be justified where the Mwatana evidence raised a prima facie case of breach: in the absence of countervailing information, these should have been included as “possible breaches”. Moreover, in these cases it was incumbent for a rational decision maker to seek further information rather than exclude them from further consideration. In addition, he developed submissions in respect of 3 examples said to show JIAT’s demonstrable unreliability. We have addressed these with the benefit of the closed evidence in our closed judgment, where we refer to them as ‘the 3 Mwatana cases’.

Mr McClelland submitted that the 149 incidents identified in his summary table indicated that the IHL analysis by the Secretary of State must be wrong, both about incidents which it excluded on the grounds of insufficient information and those excluded as unlikely to involve violations. In its evidence, Mwatana identified a list of 32 incidents in which it was said there was prima facie a breach of IHL. Mr McClelland accepted that these would need to be assessed with the benefit of the closed material, and we have done so. All bar 7 were already on the Tracker. When raised, the other 7 were added to the Tracker save for one involving a helicopter, and have been subjected to the IHL Analysis. Our conclusions on Grounds 1 to 3 below have taken into account all the incidents upon which Mwatana rely. Insofar as the incidents were taken up by the Special Advocates in closed argument, they are addressed in our closed judgment. Insofar as they were not, they are addressed in the closed narrative annexes to the witness statements of Mr Lapsley and Mr Wilson, of the SPO at the MOD, which themselves reflect the greater detail in the Tracker, and we see no proper basis for treating the IHL Analysis of any of them as erroneous, still less irrational. It is not necessary to refer to the detail of the Tracker analysis for those other incidents in the closed judgment.

Mr McClelland also argued that there was an error in paragraph 78 of the Decision Paper in treating as relevant the suggestion that the number of possible breaches was within the margin to be expected in a conflict of this nature. If so, all it evidenced was that such violations were likely to continue, and so it would be irrational to conclude that there was no clear risk of future violations.

In the written arguments it appeared that, in relation to the “insufficient information” category, CAAT and Mwatana might be challenging not just the application of the IHL Analysis but its design, on the basis that where there was insufficient information to form a view it necessarily followed that it was possible that there was a violation of IHL. This was not how the argument was developed orally and Mr McClelland, at least, disavowed it. It is in our view fallacious, as we explain when considering Ground 3. What is required is a real risk of a future serious violation, that is to say a concrete risk for which there is evidential support (albeit that it is a nuanced value judgment); it is plainly not irrational for the Secretary of State to take the view that insofar as that future risk is informed by past incidents, it can only usefully be informed by incidents where the available information makes it possible to form some meaningful view of the IHL position.

The submissions of Sir James Eadie in open were summarised in the following points:

Sir James Eadie supplemented these arguments by reference to the closed material, to which we refer in our closed judgment.

DISCUSSION AND ANALYSIS

We find it convenient to consider Ground 4 first, and then Ground 3, before addressing Grounds 1 and 2.

GROUND 4

Ground 4 asserts two separate errors of law, namely that the Secretary of State misdirected herself in respect of (1) what amounted to a “serious” violation and (2) in failing to recognise the need to consider the impunity of Saudi Arabian officials for such violations.

Seriousness

The fatal difficulty in the way of this ground is that Mr Jaffey did not seek to point to anything which was said about seriousness by the Secretary of State or her advisers at any stage of the decision-making process as being erroneous in law. Nor did he suggest that such an error of law must have taken place as the inevitable inference from the decision itself. Instead he sought in oral argument to identify aspects which he contended were part of the definition of seriousness, and challenged Sir James Eadie to respond in an attempt to identify a difference which he might rely on. This is not a permissible basis on which to mount a challenge.

In CAAT’s written submissions, reliance was placed on submissions made on behalf of the Secretary of State before the Divisional Court on the previous challenge, and correspondence (in the context of a mooted appeal to the Supreme Court) in which the Secretary of State had declined to confirm that she had, when making her New Decision, adopted a particular formulation of the meaning of “serious violation”. However, the Secretary of State did make clear, in her letter of 27 August 2020, that the IHL analysis did not consider the seriousness of the possible breaches by reference to any separate, specific criteria; that each potential breach had been factored into the overall Criterion 2c assessment on the basis that it was inherently serious; and that the Secretary of State accepted that any one incident, depending on its facts and wider context as understood by the Secretary of State on the basis of all the material available to her, might lead to an overall assessment that there is a “clear risk” for the purpose of Criterion 2c. This answer provides no basis for a suggestion that the Secretary of State has misdirected herself.

We also remind ourselves of what the Court of Appeal said at paragraph 165 of its open judgment, quoted above, and in particular this passage:

We find ourselves in a similar position, and for similar reasons decline to follow the path which Mr Jaffey invited us to go down.

We should record, however, that we do not accept Mr Jaffey’s submission that seriousness necessarily bears the same interpretation for the purposes of Criterion 2c as that which was given to it for a different purpose by the Appeals Chamber of The Yugoslavia Tribunal (The International Tribunal for the Prosecution of Persons responsible for Serious Violations of International Humanitarian Law Committed in the Territory of Former Yugoslavia since 1991) in Prosecutor v Dusko Tadic aka “Dule” (IT-94-1). That was a decision which involved, amongst other things, interpreting Article 3 of the Statute under which the Yugoslavia Tribunal was set up, which gave an inclusive definition of violations over which it had jurisdiction. Article 3 did not use the word “serious”, referring only to violations, but the tribunal in its decision drew attention to the fact that the expression “serious violations” was used elsewhere in the statute and was inherent in Article 3. At paragraph 94(iii) the tribunal said that “the violation must be “serious”, that is to say, it must constitute a breach of a rule protecting important values, and the breach must involve grave consequences for the victim.”

We would have thought that, for the purposes of Criterion 2c, an assessment of culpability would also play a part. Paragraph 2.11 of the User Guide suggests that guidance on what is meant by serious breaches is to be found in what are defined as grave breaches in the four 1949 Geneva Conventions and two 1977 Additional Protocols, and those referred to as other serious violations in the Statute of Rome setting up the International Criminal Court. The general flavour of these is a category defined by reference to wilful or intentional conduct (see, e.g., Article 147 of the Fourth Geneva Convention, Article 11(4) and 85(3) of the 1977 Additional Protocol relating to the protection of victims of international armed conflicts and Article 8.2(b), (c) and (e) of the Statute of Rome), though we would be inclined to accept that reckless conduct might well fall within the category of serious violations. This does not rule out the possibility that negligent failings may give rise to a serious breach, but we would have thought that the nature and quality of the negligence would be a relevant factor. However, we heard only limited argument on the point and it does not arise for decision in this case. Accordingly, it is neither necessary nor appropriate for us to express any concluded view on the point.

Impunity

This is an attempt to reargue a point raised in the earlier challenge and rejected by the Court of Appeal. The point was Ground 2 of the original challenge and was recorded in the Court of Appeal judgment at paragraphs 147 to 148 as follows:

The Court of Appeal’s conclusions were as follows at paragraphs 153 to 154:

Mr Jaffey’s submission was that we were not bound by this finding because now that the Secretary of State has found possible breaches of IHL, it is incumbent on her to address the three questions again, and she has failed to do so. We are unable to accept this argument. The Court of Appeal rejected the submission that the three questions were ones which were required to be asked, being contained in the non-prescriptive User Guide in a section addressing itself to war crimes by individuals. That reasoning applies equally to the New Decision, irrespective of the outcome of the new IHL Analysis. In any event it is clear from the contents of the Decision Paper to which we have referred and to which we refer in our closed judgment, that the Divisional Court’s summary of the stance adopted by Saudi Arabia, endorsed by the Court of Appeal, remains the stance which it has continued to adopt on all the evidence which we have seen.

GROUND 3

Ground 3 raises the issue which is determinative of the challenge: was the Secretary of State’s evaluation of the Criterion 2c risk irrational? The submissions advanced on behalf of CAAT and Mwatana have focussed on Grounds 1 and 2 which criticise aspects of the IHL Analysis of past incidents. That analysis is not, however, determinative of the Criterion 2c question, which involves an assessment of future risk. It is an analysis which can inform the assessment of future risk but ultimately the question for this court is whether the risk evaluation was irrational, not whether one or more of the categorisations of individual incidents, or of pattern, was erroneous. We think it is important to keep the following considerations in mind as to (a) the part the IHL Analysis can play in the ultimate Criterion 2c evaluation and (b) the nature of the IHL Analysis.

First, as the Decision Paper correctly recognises, the main factors which govern a proper assessment of the Criterion 2c risk are the willingness and ability of Saudi Arabia to comply with IHL. The core evaluation which the Secretary of State and her advisers focussed on was whether Saudi Arabia had the intention and capacity to comply with IHL. In our view they were correct to do so, reflecting the approach in paragraph 2.13 of the User Guide. Indeed no issue was taken with this approach by CAAT or Mwatana. That was the core assessment.

There are many aspects which informed that evaluation by the Secretary of State and her advisers, which are discussed in considerable detail in the Decision Paper in the section headed Thematic Analysis, and to which we have referred above and in our closed judgment. The detailed considerations, reasoning and conclusions in the Thematic Analysis section of the Decision Paper have not been criticised, let alone challenged, as irrational, either in open where the material is gisted, or in closed where the material is unredacted. The New Decision was also informed by the exhaustive thematic analysis contained in three IHL Updates covering the periods February to October 2019, November 2019 to January 2020 and February to May 2020. Again, no criticism is advanced of those Updates.

The points arising from those analyses include the following:

Secondly, the Secretary of State’s evaluation was supported by evidence of trends and overall numbers. Annex 4 to the Decision Paper shows that the trend of credible allegations has fallen considerably over the course of the war. The figure we give in closed for the number in the year prior to preparation of the Decision Paper is strikingly small. None of these were assessed as possible violations of IHL. Mwatana’s table of incidents of concern, which include many which were assessed as not credible, nevertheless also shows the same trend. It has 75 incidents for 2015, 25 incidents for 2016, 13 incidents for 2017, 14 incidents for 2018, and 9 incidents for 2019.

These fall to be judged against the very large number of sorties and airstrikes, and in the context of numbers in other conflicts. As at October 2019, the Tracker contained a total number of 486 allegations of which 331 were assessed to be credible. Whilst care must be taken in drawing comparisons across different conflicts (as the Decision Paper said), it is to be noted that in the Syria/Iraq conflict in which a similar number of airstrikes were carried out, a similar number of allegations were recorded against the coalition forces.

Thirdly, in making those Thematic Analysis assessments, the UK Government had access to information which is not available to NGOs and others from open sources, which is of considerable value in assessing the willingness and ability of Saudi Arabia to comply with IHL in addition to the factors we have summarised in paragraph 110 above. The privileged access to Saudi Arabia which the UK enjoys was summarised at paragraphs 121 to 125 of the Divisional Court judgment as follows:

Fourthly, it is in our view significant that not only have the Secretary of State and her advisers had much greater access to sources of information than that available to CAAT and Mwatana, but that the access has been to much greater information than is before the court, even in closed. The Secretary of State’s assessment was based on a vast volume of information and material considered by her advisers, of which it was only practical to put a small proportion before the court, and often by way of summary only.

Fifthly, this is true not only in relation to the Thematic Analysis evaluations, but also in relation to the IHL Analysis. For example, the Tracker refers to imagery which we have not seen. It refers to Defence Intelligence assessments which we have not seen. It contains summaries of material, where all we have is the summary, not the material for our own assessment. It contains reasoning in a brief and summary form, without setting out all the detail. This is not a criticism. But it makes the task of CAAT and Mwatana all the harder in seeking to persuade the court that a decision based in part on material it has not had access to is irrational.

Sixthly, those involved in the IHL Analysis and Thematic Analysis have far greater experience and expertise in evaluating the information than do CAAT, Mwatana, or the court. Those involved in the exercise were experts in (i) interpreting imagery, (ii) evaluating intelligence, (iii) understanding the military challenges and context in targeting, especially in a dynamic situation and against the background of Houthi tactics to use civilian cover and disinformation to advance their war aims, (iv) applying IHL in that context, (v) understanding the Saudi culture and the procedures it adopts and (vi) assessing the weight and significance to be attached to assurances and to Saudi responses to concerns when raised. The principles of IHL in play, in particular those of proportionality and feasible precautions, involve a fact sensitive and evaluative judgment by those seeking to determine whether there has been a breach or possible breach. The MOD panel when considering each individual incident had their own expertise and experience of addressing such questions in real time for UK forces, and were rationally entitled to adopt and apply the same approach to the evaluations of Saudi conduct. That was pre-eminently a matter for their expertise and judgement, not that of the court.

Against that background, the high hurdle faced by CAAT and Mwatana in seeking to persuade a court that the Secretary of State has behaved irrationally becomes a very high one. Not only is it inappropriate for the court to substitute its own assessment of the characterisation of individual incidents for that of the experts advising the Secretary of State, as a matter of law, but it is positively dangerous for it to do so where it lacks both the full range of information and the expertise and experience which informed those assessments.

Seventhly, much of the argument on Ground 1 relied on the open source documents or reports by the UN Panels or the NGOs. These were considered as part of the IHL Analysis, but the greater information and expertise available to the UK Government means that it was not bound to accept their content or views.

The evidence before us justifies the view that even in relation to the ‘what’ question, as distinct from the ‘why’ question, the UK Government had access to information which could rationally justify rejection of open source reports as not credible. The 3 Mwatana cases and other cases which we address individually in our closed judgment include a number where material to which the open sources did not have access justifies a different conclusion to the ‘what’ question in treating them as not being credible allegations.

The Government was rationally entitled to treat the public source documents with caution. It had its own knowledge of Houthi tactical attributes and capabilities, including the Houthi propaganda agenda. It had its own expertise and experience, military, political and diplomatic, which it could rationally consider was considerably greater than that available to the UN Panels and NGOs. It often had access to greater information both as to what had happened (for example in the form of access to imagery) and sometimes to why it had happened (for example through access to intelligence).

The need for caution about the open source reports is borne out by the results of the exercise. Of the Annex 5 incidents, about 20% of the allegations were assessed as ‘not credible’, the majority of which had been raised by the UN Panel of Experts. We have considered those which CAAT or Mwatana criticise as having been erroneously so categorised, with the benefit of the closed material, and concluded that there is no proper basis for treating their evaluation as ‘not credible’ to be an irrational one.

It was suggested by CAAT that the UN Panel of Experts had before its dissolution found violations of IHL in 33 of the approximately 34 investigations completed, and, in its 2016 Report, 119 incidents in which the Coalition had violated IHL. This is misleading. A similar submission was rejected in these terms in the Divisional Court judgment at paragraph 208:

Eighthly, the IHL Analysis of past IHL breaches, when placed with the other factors identified in the Thematic Analysis which support the view of the Secretary of State and her advisers that Saudi Arabia has the willingness and ability to comply with IHL, is relevant only to the extent that it undermines that evaluation. In that context:

It follows that the IHL Analysis, although undertaken with anxious scrutiny, does not play a hard-edged part in the Criterion 2c assessment. It was a tool which itself inevitably gave rise to evaluative assessments which involved standing back. It was not intended as an examination to reach a judicial or quasi-judicial decision on every incident (as the Court of Appeal said at paragraph 165 of its judgment), with rigid applications of standards of proof or the pursuit of additional information down all conceivable avenues to try and undertake such a task; but rather an exercise in common sense in the light of the particular expertise and experience of the IHL panel.

We have borne these considerations in mind when considering the allegations of miscategorisation of individual incidents, which forms the basis for Ground 1, and the pattern question which forms the basis for Ground 2.

In the way the challenge has been mounted, Ground 3, which is determinative, is wholly dependent on the success of Grounds 1 and 2. If the IHL Analysis was not irrational, there is no other basis on which the Criterion 2c risk assessment is said to have been irrational. For the reasons set out below, we reject Grounds 1 and 2, which means that Ground 3 fails.

GROUND 1

The arguments can be put into two categories, namely those which apply to the IHL Analysis generally and those which apply to the categorisation of particular incidents. We take them in that order, dealing in sections (1) to (6) below with the general arguments and in sections (7) to (9) below with particular incidents.

Criterion 2b special caution and vigilance

The Claimants submitted that criterion 2b required special caution and vigilance to be exercised where, as here, there had been prior breaches of IHL. The Secretary of State did not accept that this approach altered the Criterion 2c threshold. But even assuming that it does, the short answer is that the Secretary of State and her advisers followed it in any event. The thoroughness and diligence which the Government applied to the whole process, including but not limited to the IHL Analysis, is striking. What the Divisional Court said at paragraph 120 of its judgment about the analysis up to the date of that hearing in 2017 was this:

That remains so on the evidence before us as to what has happened since. Indeed the material before us, in the Tracker, IHL Updates and elsewhere, shows that the exercise by the Government in reaching the New Decision, and in continuing its Criterion 2c evaluation since, has been assiduous, extensive, rigorous, balanced, and multi-layered. It has not been suggested that the exercise or conclusions were conducted other than in good faith, or that there was some agenda influencing it. There can be no doubt that the process of decision making has been approached with genuine and anxious scrutiny.

The Decision Paper addressed with care and balance the past incidents found to be possible breaches of IHL and the way they fed into the assessment of future risk. Reliance on Criterion 2b adds nothing of substance to the arguments in this challenge.

We should add, however, that we reject Mr Jaffey’s submission that it follows from the fact that the New Decision has concluded that there have been breaches of IHL since the previous decision, that the previous decision was wrong in its assessment of future risk. To conclude from that fact that the previous assessment of future risk must have been flawed, making special caution and vigilance essential this time round, is in our view an inappropriate application of hindsight. The question was then, as now, whether the evidence indicated a clear risk of a future serious violation of IHL. We comment further on this point in our closed judgment.

Limitation to fixed wing aircraft

The Claimant submits that the IHL Analysis wrongly applied only to incidents involving fixed wing aircraft. This is not, in our view, a valid criticism. The IHL Updates do not limit themselves to fixed wing aircraft, and they were taken into account as part of the holistic Criterion 2c risk assessment. The Tracker includes some incidents involving helicopters which were considered and evaluated for possible IHL violation.

Nevertheless, it would not be irrational for the Secretary of State to focus on fixed wing aircraft which, with their associated ammunition, form the primary military export to Saudi Arabia that might be deployed in Yemen. The evidence in the earlier proceedings from Mr Watkins of the MOD was that the UK had no insight into incidents caused by artillery attacks or attack helicopters. Neither the Divisional Court nor the Court of Appeal made any criticism of the Secretary of State in this respect, and they made no finding that the Secretary of State had acted unlawfully in focussing the assessment on the Tracker to fixed wing aircraft. In the course of argument Mr Jaffey pointed to only three incidents involving a helicopter identified in a UN Panel of Experts report. Of these, two were merely recorded as reports and were not verified by the Panel as having taken place; nor were they put forward by the Panel as violations of IHL. The other was an allegation of an attack on a migrant boat on 16 March 2017 involving civilian casualties. The case study was said to be in Appendix A to the Report but was not put in evidence before us. Given the lack of information available to the UK in relation to helicopter attacks, it seems to us inevitable that, if it had been included on the Tracker, it would have been assessed as one for which there was insufficient information to form a view as to whether it involved a possible violation of IHL.

Limitation to proportionality, feasible precautions, distinction and necessity principles

The Claimants contended that it was incorrect for the IHL Analysis to have been confined to the four identified IHL principles of proportionality, feasible precautions, distinction and necessity: it should have extended to certain “absolute” IHL requirements relating to such matters as soldiers hors de combat (e.g. wounded) and protection of specific buildings such as hospitals and water installations.

Mr Jaffey sought to advance this contention for the first time in oral argument, by reference to the terms of the Geneva Conventions and the 1977 Additional Protocols, to some of which he took us. The point is not squarely identified as a ground raised in the Amended Statement of Facts and Grounds, and is only obliquely alluded to in paragraph 38.4 of the skeleton argument where it is said that because of the limitation to the 4 principles “in some instances, the analysis may miss other forms of IHL violation such as the obligation to safeguard an enemy hors de combat (Article 41 AP1) or the obligation to safeguard relief consignments and personnel (Article 70 AP1)”. There was no suggestion there that these were absolute obligations. This was not the proper way to seek a ruling on the scope of applicable IHL principles, and we are sympathetic to Sir James Eadie’s submission that he should not be expected to put forward the Government’s position on such an important matter on the hoof and without adequate warning and consideration. The principles relevant to the Criterion 2c exercise in the circumstances of this particular case were identified in the Divisional Court and Court of Appeal judgments, and those were the ones which the Secretary of State applied.

We should add that we were far from convinced that the material we were shown justified the conclusion that there were “absolute” prohibitions. Mr Jaffey’s paradigm example was a soldier hors de combat. If such a person were deliberately targeted, that would be a breach of the principle of distinction. The mere fact that such a person was hit would not, however, necessarily involve an IHL violation. He might be a proportionate collateral casualty of another legitimate military target. Mr Jaffey treated the Abs Hospital incident as the impermissible targeting of a military person in a vehicle which had entered the hospital compound to seek assistance for an injury, rendering him a prohibited target. But what if there were believed to be other military personnel in the vehicle? Then the principles of proportionality and feasible precautions would come into play, but we doubt whether the mere fact that one occupant of the vehicle was injured would involve an absolute prohibition on a strike under IHL.

In any event Mr Jaffey’s argument on this point is not of any practical significance in this case. It became apparent in the course of debate with Mr Jaffey during the hearing that the practical examples he gave which might be of relevance to the current challenge would in any event engage one or more of the four principles which the Secretary of State did take into account, whether it be the principle of distinction for intentional targeting, or lack of reasonable precautions or proportionality for collateral effects. Our consideration below and in our closed judgment of the individual cases specifically relied on as examples of miscategorisation would not have resulted in a different conclusion had we accepted Mr Jaffey’s arguments on this point.

“Insufficient evidence” cases with no JIAT report

We reject the submission that there was any irrationality in the approach that resulted in a rough correlation between cases which concluded that there was insufficient information to form a view on breach of IHL, and the absence of a JIAT report in such cases. JIAT was the main source for what was regarded as a competent and independent assessment of the material which could answer the ‘why’ question. In the absence of such material, it is not surprising, and certainly not irrational, for the assessment to be that there was insufficient evidence to form a meaningful judgment of whether there was a possible breach of IHL. We do not see any stark paradox or contradiction between this point and the Secretary of State’s reliance on privileged access to Saudi Arabia. The privileged access is what makes the Secretary of State and her advisers particularly well-placed to answer the ultimate Criterion 2c question as to future risk, as an evaluation of Saudi Arabia’s willingness and ability to comply with IHL, based on the thematic analysis considerations which they applied. It does not follow that such access has the same effect in relation to an IHL Analysis of past events later brought to their attention where the main evaluative tool for the why question, namely a JIAT investigation, is not available.

We have already quoted the views expressed in the Divisional Court judgment at paragraph 181(ii) about the difficulties facing the UK government in this respect, with which we agree.

It is fair to point out, in any event, that the Government did not rely only on the existence or absence of JIAT reports. There were individual incidents of concern in which the Government sought further information and assurances, or to which it responded with further training. It was a matter for diplomatic and political judgment as to when to do so and in what terms, and under ordinary Tameside principles that judgment could not be impugned unless no reasonable decision-maker could have taken the same approach (see paragraph 36 above). A blanket approach seeking further information in respect of every incident raised as a matter of concern by a NGO, for which there was no JIAT Report, or in which the JIAT report lacked some necessary detail to form an assessment with reasonable confidence, could rationally have been regarded as diplomatically problematic and counterproductive. It was for the Secretary of State and her advisers to exercise the political and diplomatic judgement as to when to engage and to use diplomatic capital. Mr Jaffey’s submission that if there were no answer to the why question in respect of a credible allegation then it was irrational for the Secretary of State not to seek specific further information was, in our view, unrealistic. It would logically apply to every such case. It can be seen from the numbers involved that it would be a huge undertaking, which the Secretary of State and her advisers could rationally consider would have had adverse foreign relations consequences in return for which there would have been no significant benefit in the analysis of what is ultimately a predictive evaluation of future risk, and in which the IHL analysis plays a part only to the extent indicated in paragraph 124(4) above.

Moreover, all those incidents which were treated as insufficient information cases were nevertheless looked at for the purposes of seeing whether there was a pattern to them. Accordingly, even if some were wrongly put in that category, it would not invalidate the evaluative assessment which was relevant to future risk.

Number of possible breaches to be expected in a conflict of this nature

Criticism was made to the effect that paragraph 78 of the Decision Paper wrongly treated as relevant the suggestion that the number of possible breaches was within the margin to be expected in a conflict of this nature. Mr McClelland’s argument was that if so, all it evidenced was that such violations were likely to continue, and so it would be irrational to conclude that there was no clear risk of future violations. We disagree.

The passage in question is in a section of the Decision Paper addressing whether the possible breaches constituted a pattern which gave reason to doubt the willingness and capability of Saudi Arabia to comply with IHL. Paragraph 78 stated that the detail which had been set out in the Paper “is consistent with a limited number of errors, well within the margin that would be expected in a conflict of this nature”. The evaluative question facing the Secretary of State had to look not merely at individual errors in the past, but to consider whether they evidenced some systemic weakness which had not been identified or addressed and which gave rise to a clear risk of IHL violations in the future. The fact that the number of possible breaches was so small compared with the number of airstrikes and sorties could rationally be relied on as a source of comfort that there were not such systemic weaknesses.

Methodology: the meanings of “possible” and “unlikely”

It was common ground that the IHL Analysis methodology was intended to filter out incidents in which a breach of IHL was “unlikely” but retain those where it was “possible”. The language of paragraphs 22(b) and 23 of the Decision Paper, and in particular the reference to “possible” as including “just possible”, might be problematic if viewed through a lawyerly approach to construction. “Unlikely” is not the antithesis of “possible” as a matter of normal language, and certainly not if “possible” extends all the way down the scale of likelihood to “just possible” as seems to be suggested by paragraph 23. It was submitted on behalf of CAAT and Mwatana, and vigorously argued by Mr McCullough, that “unlikely” occupied only that territory on the scale of likelihood below what was just possible.

We regard that submission as unrealistic for a number of reasons. First, the Decision Paper is not a statute and is not to be construed in an over lawyerly way as if it were. It identifies a methodology devised in order to conduct the exercise required by the Court of Appeal judgment; it must be interpreted by reference to the purpose of the exercise and its practical application. If unlikely meant something less possible than “just possible”, it would exclude even the remotely possible and in substance mean impossible, or something vanishingly close to it. To treat such an incident as an established breach for the purpose of the Criterion 2c exercise would be positively misleading and unhelpful. It would require inclusion of all or almost all incidents, including those where there was only the very remotest possibility of an IHL breach. That could not sensibly inform an assessment of whether there was, prospectively, a clear risk of a serious breach. Secondly, such an approach would deprive the word unlikely of its natural meaning, and indeed of any relevant meaning. Thirdly, paragraph 77 of the Decision Paper, at the start of the part headed “Conclusions on IHL Analysis”, emphasised that the MOD had applied a “relatively low threshold” of “possible”. If what was excluded was only the impossible or virtually impossible, this would not be characterised as only a relatively low threshold but as an extremely low one. Similar considerations apply to the expression “low bar” in paragraph 145. Fourthly, it is clear that in the way the IHL Analysis was applied, incidents were filtered out on the basis that they were unlikely in the normal meaning of the word. It was the word used in the Tracker and it can be seen from detailed information in the Tracker itself, in the closed material we have examined, that “unlikely” was being given its ordinary meaning. Fifthly, the methodology included an assessment of those incidents that had been put in the “unlikely” category, for the purposes of identifying any trends. That would have been a meaningless and unhelpful exercise if that category were populated only by incidents where it was impossible or virtually impossible that there had been a breach of IHL.

We therefore interpret the methodology described in the Decision Paper as being that cases were filtered out where it was assessed that it was unlikely that there was a breach of IHL, using that word in its normal sense. We should add that if the methodology intended to be applied was that suggested by Mr McCullough, but what was done was to apply the methodology as we have understood it, the latter could not in our view have formed an irrational approach so as to infect the ultimate Criterion 2c decision with irrationality. Indeed, in our view it would have been irrational to apply a methodology in the way Mr McCullough sought to interpret it.

Mr Jaffey’s 8 incidents

These were, in chronological order:

Abs hospital 15 August 2016

This attack hit Abs Hospital, Hajjah, one of the hospitals supported by MSF, on 15 August 2016, as a result of which it was reported that 19 people were killed and 24 injured. Details of the open source material in relation to this attack appear from the Court of Appeal judgment at paragraphs 108 to 111. A JIAT press statement released in December 2016 concluded that “damages inflicted on the building were because a targeted vehicle (which was a legitimate military target) next to the building which were unintentional.” (sic) The Panel of Experts did not accept this and said that its investigations suggested that the vehicle was a civilian car transporting wounded individual(s) (wounded, possibly from a previous airstrike elsewhere) to the hospital.

We say more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational and cannot support an argument that the decision of the Secretary of State was irrational.

The Great Hall strike 8 October 2016

The open source material surrounding this incident was summarised in paragraph 113 of the Court of Appeal’s judgment. The episode was addressed in detail by the second UN Panel of Experts report. In brief, two guided bombs were dropped on 8 October 2016 in the mid afternoon, when around a thousand people were attending the funeral. The funeral was of the father of an acting Minister of the Interior and, as the UN Panel put it, “a significant number of Houthi-Saleh-affiliated military and political leaders were expected to attend”. It was an incident which was immediately identified by the UK Government at the time as of great concern. JIAT published its initial report on 15 October 2016, concluding that there had not been compliance with the Coalition rules of engagement and that appropriate action should be taken against individuals and compensation paid to the families of victims.

We say more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational and cannot support an argument that the decision of the Secretary of State was irrational.

Strike on a wedding in Bani Qais 22 April 2018

The open source documents reported that this air strike hit guests at a wedding resulting in 21 being killed, including 11 children, and 90 injured. The UN Panel of Experts suggested that it breached the principles of proportionality and taking feasible precautions. JIAT found “a number of errors indicating non-compliance with some of the rules of engagement procedures to minimise damage.”

We say more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational and cannot support an argument that the decision of the Secretary of State was irrational.

Dhayan bus strike 9 August 2018

This was an airstrike on a bus in a busy civilian market in Dhayan, Sa’ada, in which the open sources reported approximately 43 killed and 63 wounded. The UN Panel of Experts stated that it breached IHL. JIAT concluded that the Coalition believed the bus to be carrying Houthi personnel; that orders to attack the bus were given without following rules of engagement to minimise civilian casualties; and that when the bus moved, an order to abort was given but that this did not reach the pilot in time.

We say more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational and cannot support an argument that the decision of the Secretary of State was irrational.

Death of Mr Khashoggi in October 2018

As explained above, the murder of the journalist Jamal Khashoggi by Saudi nationals inside the Saudi embassy in Istanbul on 2 October 2018 was considered at some length in the Decision Paper as relevant to the UK Government’s analysis of the Saudi decision-makers’ attitude to international law, and specifically IHL; and to the value of Saudi high-level assurances in respect of IHL compliance. The incident, and subsequent Saudi Arabian actions in relation to it, raised serious concerns. On balance, however, the Government’s view was that the Khashoggi incident did not speak directly to there being an increased risk, let alone clear risk, that UK-supplied arms and equipment might be used in a serious violation of IHL in Yemen, because the prosecution of the conflict was separate from the Saudi treatment of political dissidents. The assessment was that, based on the full range of evidence available, it was clear that the Saudi intent to comply with IHL in the conduct of the Yemen campaign was genuine. We have said a little more about this in our closed judgment.

The assessment was careful and balanced, and made by those with particular expertise. It cannot meet the high threshold of being characterised as irrational or supporting an argument that the decision of the Secretary of State was irrational.

Strike on a hospital in Kitaf, Sa’ada 26 March 2019

This was an airstrike in which the open source documents reported that 7 were killed, including 4 children, and 6 injured. The UN Panel of Experts concluded that it was a violation of IHL. The JIAT report concluded that “the military commander hastened the work procedure to ensure that military advantage is not lost, which resulted in inaccuracies in the assessment of the possibility of civilians entering the targeting area.”

We have said more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational and cannot support an argument that the decision of the Secretary of State was irrational.

Attack in Al Jawf on 14 February 2020

We have said more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational. We are satisfied that it cannot support an argument that the decision of the Secretary of State was irrational.

Attack in Al Haq on 6 August 2020

We have said more about this in our closed judgment. The categorisation of this incident in the IHL Analysis cannot be said to be irrational. We are satisfied that it cannot support an argument that the decision of the Secretary of State was irrational.

The 3 Mwatana cases

These 3 cases were addressed in argument by Mr McCullough by reference to the open and closed evidence. They were as follows:

We consider each of them in our closed judgment. We conclude that none supports the submission that the Secretary of State’s categorisation of them or her decision on the Criterion 2c risk was irrational.

The SA Annex cases

We have also considered in our closed judgment individual cases which Mr McCullough submitted, with the benefit of the closed material, had been miscategorised or otherwise undermined the IHL Analysis. For the reasons given in our closed judgment, we have concluded that it cannot be said that the categorisation of any of them was irrational or that, singly or cumulatively, they undermined the rationality of the Secretary of State’s decision.

Having considered all the material, in closed as well as open, we therefore reject Ground 1.

GROUND 2

We have rejected the submission that the categorisation of any of the incidents in the IHL Analysis was irrational. The pattern analysis therefore fell to be performed in respect of the possible breaches which the IHL Analysis revealed.

The possibles

These are considered further in our closed judgment.

The Secretary of State’s pattern assessment

As we have explained, past breaches could not render irrational the Secretary of State’s assessment of the Criterion 2c future risk unless they suggested some underlying systemic weakness which was unaddressed so as to undermine the conclusion derived from the other material considered, including the thematic analysis considerations, which pointed to Saudi Arabia having the intention and capability to comply with IHL. That is the nature of the pattern analysis which fell to be undertaken.

The open arguments advanced on behalf of CAAT and Mwatana depended in large part upon there being many more incidents which should have been characterised as possible IHL violations, a proposition we have rejected. They also depended upon applying such a high level of generality to the categorisation for pattern purposes (such as breach of rules of engagement) as to divorce the analysis from the purpose it serves as part of the overall Criterion 2c risk assessment. Moreover, regard must be had to the dates of the few possibles in the light of the evidence about constant improvement by the Saudis of the kind we have described above and in our closed judgment.

To be fair to Mr Jaffey and Mr McClelland, the arguments advanced in open were without sight of the closed material which identified the incidents treated as possible IHL violations and included the full text of the part of the decision Paper which discussed the pattern question. It is notable, however, that Mr McCullough, who had the benefit of having seen the closed material, did not subject that part of the Decision Paper to any criticism.

The approach to the pattern question in the Decision Paper and the conclusions reached seem to us to be sound and well-reasoned. The identification of the possible breaches by the IHL Analysis was, we have held, not irrational. There were very few by comparison with the very large number of sorties and airstrikes. We have described them in some detail in our closed judgment. It is apparent from the detail of the cases that they were disparate in nature and not suggestive of any systemic weakness which had not been addressed. That, at the very lowest, was a rational view. Moreover, the review of the Annex 5 incidents, which included also “unlikely” and “insufficient information” cases, did not give rise to concerns about Saudi Arabia’s attitude or capability.

With the benefit of seeing the closed material, we have concluded, with little hesitation, that there is no basis for treating the Secretary of State’s approach or answer to the pattern question as irrational.

Events since the date of the Decision Paper

For reasons we discuss more fully in our closed judgment, there is nothing in events since the New Decision was taken which changes our view of the continuing rationality of the Secretary of State’s Criterion 2c risk assessment.

Paragraph 78 of the Decision Paper

We reject Mr McClelland’s argument that there is anything in paragraph 78 of the Decision paper to undermine this conclusion. His submission was that there was an error in paragraph 78 of the Decision Paper in treating as relevant the suggestion that the number of possible breaches was within the margin to be expected in a conflict of this nature; and that, if so, all it evidenced was that such violations were likely to continue, making it irrational to conclude that there was no clear risk of future violations. We disagree. The passage in question was addressing whether the “possible” breaches constituted a pattern. The number involved was plainly relevant to that question, and the number which might be expected, as well as their nature, was relevant to the question whether they were suggestive of a systemic weakness.

The pattern emails

We have set out the “pattern emails” which Mr Jaffey submitted demonstrated circular and invalid reasoning. We disagree that they do so. The SPO response in their email of 25 May 2021 explained that the point they were making was that if the putative “pattern” was expressed at the level of generality of two of the four pillars of IHL being applied, namely proportionality and feasible precautions, it would be wholly unsurprising if most or all of the possible violations fell into such a category because these were the ones likely to be engaged in air strikes in the circumstances of the Yemen conflict. It was, however, too much of a leap to say that such commonality must be a concerning pattern. When examining the possibles the MOD panel had had very much in mind consideration of whether the actual nature of any commonality might evidence a problematical pattern. That is a rational and coherent explanation and betrays no circularity.

We therefore reject Ground 2.

CONCLUSIONS

We have rejected all four grounds of challenge. The claim will therefore be dismissed.