Daniella Lock: UK Government Assessments of Risk to Refugees and ECHR Constraints on UK Immigration Policy in the Rwanda Ruling

This post examines the High Court’s December ruling on the lawfulness of the UK Government’s Rwanda scheme. The Rwanda scheme allows people to be removed to Rwanda to have their asylum claims determined, where that claim is considered inadmissible in the UK. The lawfulness of the scheme partly turned on its compatibility with the UK’s obligations under the UN Refugee Convention and the European Convention on Human Rights (ECHR) by reference to whether, as asserted by the UK Government, refugees deported under the scheme would face a real risk of mistreatment (contrary to Article 3 and the Refugee Convention). The ruling therefore provides an important lens through which to examine Government assessments of the risks faced by refugees as part of its immigration policy and the ECHR constraints on such policies. This is a timely issue in light of recent reports that the Prime Minister is considering proposals for the UK to withdraw from membership of the ECHR should it constrain the Government in its immigration plans.

In determining the compatibility of the Rwanda policy with the ECHR, the High Court found that the UK Government was entitled to conclude that deportees to Rwanda would not face a ‘real risk’ of suffering the relevant mistreatment. This post argues there are three key epistemic presumptions present in the High Court’s reasoning on these issues that are suspect and resulted in a partial analysis on the question of real risk. An ‘epistemic presumption’ refers to a presumption regarding the grounds on which knowledge may be sought. In relation to the Rwanda ruling, ‘epistemic presumptions’ refer to a presumption regarding the authoritative sources of knowledge available in assessing whether deportees face a real risk of mistreatment contrary to Article 3 ECHR. The questionable epistemic presumptions set out in this post are:

1. That the Government’s expertise is the only expertise that should be assigned particular weight

2. The implicit conflation of Government expertise with an objective assessment

3. That diplomatic assurances are reliable indicators of state behaviour.

In setting out the ways in which these epistemic presumptions are suspect, the post exposes numerous errors in previous Government assessments of risks to refugees when pursuing immigration policy. These errors suggest that independent assessment of the Government’s risk assessments, of the kind required by the UK’s membership of the ECHR, is an essential feature of a system aimed at avoiding mistreatment of refugees at the hands of the state. An adequate independent assessment was lacking in the case of the High Court’s review of risk in the Rwanda ruling.

The High Court’s Reasoning      

The High Court stated there were two important questions when ruling on the compatibility of the Rwanda Scheme with the UK’s Article 3 obligations. The first was procedural and asked whether the evidence provided meant that the Home Secretary had satisfied the obligation under Article 3 ‘to conduct a thorough examination of the relevant conditions in the third country concerned and, in particular the accessibility and reliability of its asylum system’ – an obligation derived from Ilias and Ahmed v Hungary (paras 139 – 141).

In determining whether there had been a ‘thorough examination’, the High Court considered a range of evidence, including the Home Secretary’s May 2022 assessment documents concluding Rwanda was a safe third country and an agreement regarding the treatment of deportees made between the UK and Rwandan Governments in the form of a Memorandum of Understanding (MOU) made on 13 April 2022 supplemented by two Notes Verbales (unsigned diplomatic communications written in the third person).

The High Court found that the UK Government’s assessment represented a ‘thorough examination’ of ‘all relevant generally available information of the type envisaged by the European Court of Human Rights (ECtHR) in [Ilias and Ahmed]’ [59]. This was despite ‘adverse’ comments made by the Asylum Research Centre, including that the Government documents failed to properly explain the terms of the MOU (ibid). The High Court stated that ‘even assuming’ the report’s comments were ‘warranted’, they were ‘not relevant to compliance with the Ilias obligation’ as it was ‘beyond argument’ that the MOU, Notes Verbales and 9 May assessment ‘were considered’ together at the ‘time the removal decisions were made’ (ibid).

The second core issue was substantive and asked whether the Home Secretary was entitled to conclude that there were sufficient guarantees to ensure that Rwanda was a ‘safe third country’ and that persons removed to Rwanda would not be put at real risk of suffering ill-treatment contrary to Article 3 ECHR (and the Refugee Convention) ([62]). Article 3 ECHR states that ‘[n]o one shall be subjected to torture or to inhuman or degrading treatment or punishment’.

In examining this issue, the High Court stated its approach must be ‘rigorous and pragmatic notwithstanding that ultimately it is an assessment to be undertaken recognising that the court must afford weight to the Home Secretary’s evaluation of the matter’ ([63]) and that this will specifically involve ‘a recognition of the expertise that resides in the executive to evaluate the worth of promises made by a friendly foreign state’ (ibid). The High Court held that the Home Secretary was entitled to conclude the scheme was compatible with the UK fulfilling its obligations under Article 3 ECHR on the basis of three sets of reasons.

The first set of reasons related to the UK Government’s assessment that there was not a real risk that deportees would be subject to mistreatment as Rwanda would ‘honour its obligations’ ([66]). The High Court emphasised this assessment rested on the Government’s ‘experience of bilateral relations extending over almost 25 years, and the specific experience of negotiating the MOU over a number of months in 2022’ and it could go behind this opinion ‘only if there were compelling evidence to the contrary’ before adding it did not ‘consider such evidence exists’ (ibid).

A second set of reasons related to perceived problems with the alternative views on real risk – particularly those presented by the United Nations High Commissioner for Refugees (UNHCR). This evidence was not considered ‘compelling’ and there was no recognition of UNHCR expertise on this matter ([67]). In rejecting its arguments, the High Court emphasised they carried ‘no special weight’ and were to be ‘evaluated in the same manner and against the same principles of any other evidence’ citing remarks such as those of Elias LJ in HF (Iraq) v Secretary of State for the Home Department [2014] 1 WLR 1329 at [42] to [47] and of Davis LJ in AS (Afghanistan) v Secretary of State for the Home Department [2021] EWCA Civ 195 at [17] to [23] ([71]).

The third category of reasons was connected to the UK-Rwanda Scheme itself, as set out in the MOU and Notes Verbales. The High Court described the terms of the MOU and Notes Verbalesas ‘specific and detailed’ and found that the obligations that Rwanda has undertaken were ‘clear’ ([65]).

The High Court considered that the terms of the MOU and Notes Verbales mitigated the potential real risk to deportees suggested by evidence of harmful state practice in Rwanda raised in proceedings, including related to a 2018 incident in which Rwandan police opened fire on refugees at Kiziba Refugee Camp ([74]) and political opponents having been detained in “unofficial” detention centres and subjected to torture ([77]).

In concluding this, the High Court noted the Note Verbales established a mechanism to ‘allow complaints about accommodation and support provided under the MOU to be raised and addressed’ ([74]). It also emphasised that the treatment of deportees would be ‘kept under the review (sic) by the Monitoring Committee and the Joint Committee (each established under the MOU)’ ([77]). It further decided that the ‘advantages that accrue to the Rwandan authorities’ from the arrangement provide a real incentive against ‘any’ mistreatment (ibid).

Analysing Three Epistemic Assumptions

1. The Government’s expertise is the only expertise that should be assigned particular weight

The first epistemic presumption is that the Government’s expertise is the only expertise that should be assigned weight. This is suspect for several reasons. First, the ECtHR has explicitly referred to UNHCR documents as ‘authoritative’ in describing, in Ilias (para 141), what constitutes a thorough assessment.

Second, domestic law in this area is more ambiguous on the question of assigning weight in this context than is suggested by the High Court. The recent authority quoted by the High Court, in the case of AS [17] – [23], does not state that the UN documents ought not be assigned any weight but is principally concerned with the question of whether UN documents are ‘presumptively binding’. Whether conclusions are binding on a court is a different question to whether a conclusion should be assigned particular weight on the basis of the expertise of the body that made the conclusions. Moreover, ‘to be evaluated in the same manner and against the same principles of any other evidence’ is to judge whether there are reasons to assign weight based on factors like expertise. Thus, to treat UNHCR evidence as any other evidence ought to require an examination of the extent to which the UNHCR may have relevant expertise in judging the weight to be given to its evidence.

The High Court did judge the basis for expertise in the case of the Government, emphasising that the UK Government’s assessment rested on an ‘experience of bilateral relations extending over almost 25 years, and the specific experience of negotiating the MOU over a number of months in 2022’. We can put to one side whether that is an overstatement of the UK Government’s experience (the Conservatives have been in power for less than half this time). The High Court did not afford the UNHCR similar treatment and acknowledge clear indicators of its expertise in this area – such as the fact that it is a Nobel Peace Prize-winning UN agency with over 70 years’ experience working with refugees from around the world, having worked with more than 50 million refugees by 2015. The High Court’s presumption that the UK Government’s assessment is the only assessment that ought to be assigned particular weight therefore rests on shaky ground.

2. Implicit conflation of Government expertise with an objective assessment

The second epistemic presumption made by the High Court is an implicit conflation of Government expertise with an objective assessment (i.e., an assessment based on real facts and not influenced by personal beliefs or feelings). While the High Court does not explicitly conflate the two, the conflation is implicit insofar as the government assessment is largely taken at face value. This is evident in the High Court’s conclusion that the UK Government complied with the Ilias requirement on the basis it was ‘beyond argument’ that the MOU, Notes Verbales and 9 May assessment ‘were considered’ together at the ‘time the removal decisions were made’ ([59]). Here the High Court frames a thorough assessment as being synonymous with the Government having carried out an assessment and avoids critical engagement with its depth and rigour.

Notably, clear omissions in the Government’s assessment are not properly addressed in the High Court’s reasoning. For example, the judgment makes clear that there is ‘no evidence’ that during its negotiations, the UK sought to investigate ‘either the terms of the Rwanda/Israel agreement or the way it had worked in practice’ ([68]). The High Court asserted this was a ‘permissible approach’ and the judges did ‘not consider it discloses any error of law’ without going into further detail (ibid). This assertion is strange in light of the fact that the UK and Rwanda arrangement largely mirrors the Rwanda/Israel agreement. That agreement therefore provides a crucial indication as to how refugees sent from the UK will be treated. Moreover, in light of the UNHCR’s evidence that many of the refugees deported from Israel to Rwanda ended up vanishing from official records, it suggests that assurances from Rwanda will not guarantee refugee safety. The High Court made no acknowledgement of this in its judgment.

In engaging with the Government’s overall assessment of the scheme, the High Court also failed to examine the extensive list of basic errors made by the Government which the High Court itself identified with respect to the individual assessments of risk made in relation to the individual claimants in the case. These errors had ultimately led to the quashing of no less than eight of the Home Secretary’s decisions. The errors in question included the ‘accidental transposition of AAA’s evidence into the decision made on AHA’s case’ ([283]) and failing to engage with claimant’s representations and apply legally required tests (e.g., see [284]).

In taking the Government’s assessment at face value, there also seems to be no awareness in the High Court’s reasoning that the UK Government has previously lost a number of article 3 and deportation cases on the basis of underestimating risk. A clear example of this occurred in the Special Immigration Appeals Commission (SIAC), when a Government official, Anthony Layden, gave evidence in DD and AS v SSHD that an assurance from Colonel Gaddafi could be relied upon to mitigate real risk of mistreatment to individuals AS and DD (whom the UK Government was attempting to deport to Libya). Mr Layden went so far as to claim in proceedings that the prospect of Libya breaching the conditions of the MOU he had negotiated was ‘well-nigh unthinkable’ ([320]). This was despite the fact that the only body responsible for monitoring the well-being of AS and DD once deported was run by Gaddafi’s son.

Notably, the need to consider the expertise of the Government separately from its ability to make an objective assessment is not highlighted here on the basis of what some would deem obvious. Namely, that the UK Government clearly has a political interest in pursuing the Rwanda policy – based on its manifesto commitment to ‘take back control of our borders’ and expressed desire to reduce the ‘relentless flow of migrants’ to the UK – and is disposed to dismiss risks faced by refugees and asylum seekers (supported, for example, by recent reports of asylum-seeking children going missing in Brighton). My claim here is that the dangers of this conflation would have been readily identifiable to the High Court through a critical examination of the Government’s assessment and engagement with previous Article 3 ECHR and deportation precedents.

3. That diplomatic assurances are reliable indicators of state behaviour

The final suspect epistemic presumption is that diplomatic assurances are reliable indicators of state behaviour. Recall that diplomatic assurances are doing a great deal of heavy lifting in the High Court’s conclusion. They are meant to mitigate against the significant amount of risk to refugees that exists on the basis of recent state practice – such as when Rwandan Police opened fire in a refugee camp in 2018, or when the Rwandan Government refused access to refugee detention facilities in 2017 to the UN Committee against Torture (as noted in the UK Government’s own assessment documents at para 2.14.4), not to mention the state practice associated with the Rwandan Government being authoritarian and regularly engaged in human rights abuses.

However, while diplomatic assurances have been accepted by the ECtHR to mitigate real risk in certain circumstances, this does not mean they are necessarily reliable indicators of state behaviour in all situations. First, as I highlighted in my last UKCLA blog, diplomatic assurances accepted by the UK Government, at issue in the W and others case, have now been shown to be unreliable. The British Embassy has been unable to trace the nine men deported by the UK Government to Algeria in recent years despite diplomatic assurances being in place. In the course of this litigation, evidence in the form of emails between officials surfaced which suggested that these assurances had never been reliable, and this had been known by British diplomats in Algeria at the time they were sought – despite the Government assessing that the men it was seeking to deport were not at risk (W and others[110]).

Moreover, as with other assurances, neither the MOU with Rwanda or the Notes Verbales are legally binding, as explicitly noted in both Part 2.2 of the MOU and Para 20 of the Notes Verbales (which, incidentally, is a diplomatic document that is not even signed by an official of either the UK or Rwandan Government). As emphasised by the Lords’ International Agreements Committee, this renders the arrangement ‘unenforceable’ (page 11).

Even a cursory examination of the specific provisions in the Rwanda/UK agreements suggest that the arrangements they establish are a long way from adequate in terms of establishing a system capable of providing minimum protections to refugees. The ‘Monitoring Committee’ is meant to be independent but its members will be chosen by the Government on the basis of unknown criteria (paras 15.1 – 15.3.3 of the MOU). The terms of reference for the Monitoring Committee are set by the ‘Joint Committee’, made up of representatives from both the UK and Rwanda Governments (para 2.15 of the MOU). The arrangements establish no minimum reporting requirements and offer no guarantee that the Monitoring Committee will be given access to relevant information or be able to publish any reports it does produce freely. Moreover, there is no mechanism for refugees to raise complaints regarding their conditions other than to government officials, and there is no independent appeal process should those officials make harmful decisions regarding the wellbeing of refugees. The High Court provides no indication that it has engaged with these aspects of the arrangements in its reasoning on the Monitoring Committee ([22] – [23]).

Concluding Thoughts

This post has set out three epistemic presumptions apparent in the judgment of the High Court that are suspect. Their cumulative effect of the High Court’s presumptions risks positioning the Government’s assessment so as to make it near-irrefutable as a matter of law. This sits uneasily with the fact that the Ilias obligation regarding a ‘thorough assessment’ of risk is not merely placed on the UK Government, but on all national authorities – which, moreover, is a reflection of the fact that the ECtHR is in a limited position to correct a lack of rigour in domestic reasoning due to its commitment to subsidiarity. In the Rwanda judgment, the presumptions adopted by the High Court mean that it does not appear to have met that obligation.

A key theme underpinning the analysis in this post is that Government expertise will not always lead to an accurate assessment of risk to refugees. The Government’s fallibility on this issue has been repeatedly exposed in its past assessments. This reinforces the importance of independent review of these assessments, such as that required by the ECHR system, for the purpose of ensuring that there remains some basic accountability for decisions made in this area of Government policy.

The author would like to offer warms thanks to Mike Gordon and Paul Scott for their helpful feedback on the piece. Any errors are the author’s own.

Dr Daniella Lock is a Postdoctoral Fellow at the Bonavero Institute of Human Rights in the Faculty of Law, University of Oxford.

(Suggested citation: D. Lock, ‘UK Government Assessments of Risks to Refugees and ECHR Constraints on UK Immigration Policy in the Rwanda Ruling’, U.K. Const. L. Blog (7th February 2023) (available at https://ukconstitutionallaw.org/))