Blog post by Richard Warren (Kent Law Clinic, University of Kent), and forms part of a series of blog posts responding to the UK Home Office’s New Plan for Immigration.


The Nationality and Borders Bill has now been published to implement the government’s New Plan for Immigration. There is little sign that the public consultation had any impact on the government’s drafting of the Bill. Indeed, the consultation response has yet to be published and the consultation website maintains that feedback is still being analysed.


In this blog post I will consider several of the Bill’s clauses which seek to reinterpret the Refugee Convention. The government has previous form for attempting to add a domestic gloss to its international human rights obligations. The Immigration Act 2014 sought to direct judges in how to weigh the public interest in immigration control when considering claims based on the Article 8 ECHR right to family or private life.[1] This was used as a mechanism to seek to overturn earlier case law which was perceived as being too liberal to claimants. In a similar way, the UK is now trying to direct judges in how to interpret the Refugee Convention, with a view to producing a more restrictive result.


Having left the EU and the Common European Asylum System (CEAS), the UK now has the opportunity to diverge from European standards. Clause 27(4) of the Nationality and Borders Bill begins that process by revoking the Refugee or Person in Need of International Protection (Qualification) Regulations 2006 which were the UK’s means of implementing our commitments to the minimum standards of the 2004 EU Qualification Directive (QD). Despite this, many of the following clauses insert into statute sections of the revoked regulations. Clauses 28 and 30-31 replicate fairly faithfully the language that appears in the QD concerning the approach to be taken to the concepts of persecution, internal protection and internal relocation.


Yet, there are places where the UK has sought to diverge from current approaches to the interpretation of the Refugee Convention in order to develop a more restrictive determination process and to overturn previously accepted standards of protection. Below, I will consider three of the most significant changes: qualifying the test of well-founded fear; limiting the concept of “particular social group” (PSG); and an attempt to define “particularly serious crimes” for the purposes of exclusion from non-refoulement.  


1. Qualifying the Test of Well-Founded Fear


Clause 29 seeks to amend the standard of proof that needs to be met by those seeking asylum. It introduces a two-stage test: the applicant must first establish on the balance of probabilities that they fall within the protected Convention reasons and that, as a result, they fear persecution. Only if this is established must the decision maker then go on to consider, to the lower standard of “reasonable likelihood,” whether the individual will, in fact, be persecuted.


In the 1980s and 90s, there were a number of key legal cases considered by the House of Lords and higher courts which grappled with the standard of proof and established the correct approach to be followed, the most important being Sivakumaran and Karanakaran. According to the former judgement, the correct standard of proof is a low standard of “reasonable likelihood,” which, according to the latter judgement, applies both to assessment of past facts and future risk. It is not hard to see the reasons for this. The UNHCR in their Handbook on status determination observe that:

[…] it is hardly possible for a refugee to “prove” every part of his case and, indeed, if this were a requirement the majority of refugees would not be recognized. It is therefore frequently necessary to give the applicant the benefit of the doubt [emphasis added].


The Handbook also notes that

In most cases a person fleeing from persecution will have arrived with the barest necessities and very frequently even without personal documents. Thus, while the burden of proof in principle rests on the applicant, the duty to ascertain and evaluate all the relevant facts is shared between the applicant and the examiner. Indeed, in some cases, it may be for the examiner to use all the means at his disposal to produce the necessary evidence in support of the application. Even such independent research may not, however, always be successful and there may also be statements that are not susceptible of proof. In such cases, if the applicant’s account appears credible, he should, unless there are good reasons to the contrary, be given the benefit of the doubt.


So, an attempt to raise the standard of proof is an attempt to roll back the approach that was agreed some two decades ago with a view to refusing more asylum seekers refugee status – not because they are not refugees, but because they cannot prove their case to a higher legal standard. This will arguably place the UK in breach of the Convention.


UK courts have been clear that when interpreting the Refugee Convention, the domestic courts of a Contracting State must strive to give it its independent and autonomous international meaning. In Adan, the House of Lords applied what Lord Steyn called the ‘one true interpretation’ of Article 1A(2) of the Refugee Convention. Lord Steyn explained the correct interpretative approach:

[…] the enquiry must be into the meaning of the Refugee Convention approached as an international instrument created by the agreement of contracting states as opposed to regulatory regimes established by national institutions. It is necessary to determine the autonomous meaning of the relevant treaty provision. This principle is part of the very alphabet of customary international law.


It is therefore not appropriate for the UK government to seek to impose its own interpretation of the Convention in order to reduce the effectiveness of that instrument’s protection. As the court has said in Adan:

They cannot simply adopt a list of permissible or legitimate or possible or reasonable meanings and accept that any one of those when applied would be in compliance with the Convention.


The government’s proposed new standard of proof was explicitly considered in the case of Karanakaran where the Court of Appeal considered applying two-stage test where the decision maker first decided past and present facts to the balance of probabilities and then went on to assess whether there was a real risk of persecution applying a lower standard of proof. The court rejected this as contrary to the correct approach in which there is a positive role for uncertainty:

[The accepted] approach does not entail the decision-maker […] purporting to find “proved” facts, whether past or present, about which it is not satisfied on the balance of probabilities. What it does mean, on the other hand, is that it must not exclude any matters from its consideration when it is assessing the future unless it feels that it can safely discard them because it has no real doubt that they did not in fact occur. […] it would be quite wrong to exclude matters totally from consideration in the balancing process simply because the decision-maker believes, on what may sometimes be somewhat fragile evidence, that they probably did not occur.


In seeking to reinterpret a well-settled principle of how the Refugee Convention is applied, by reintroducing a legal test disavowed nearly two decades ago, this is a recipe for further complex litigation and for placing the UK in breach of its international obligations.


1.1 The use of evidence notices and directing judges to give minimal weight to fresh evidence submitted unless there is a good reason


The above change, increasing the evidential threshold for establishing past facts, needs to be considered alongside a further proposal to make it more difficult to submit additional evidence in support of an application. Clause 16 allows the government to serve evidence notices on applicants for asylum, demanding that they submit all the evidence they wish to rely on by a specified date. Where evidence is submitted after the specified date, decision makers, including judges on appeal, are required to ‘have regard to the principle that minimal weight should be given to the evidence’ unless there is a ‘good reason’ (Clause 23(2)).


In the New Plan for Immigration this appeared as a mechanism to limit so-called fresh claims for asylum. Fresh claims are made when an individual has been through the asylum process and found not be in need of protection, but new evidence subsequently emerges that calls for a revaluation of the claim. Yet, the legislation appears to be applicable to any asylum claimant, meaning that unless evidence is provided prior to an arbitrarily set date, it will fall foul of the “minimal weight” requirement.


This approach to require judges to give little weight when evaluating evidence has already been used in the context of Article 8 ECHR where judges are told to give little weight to a claim based on private life where the applicant has a “precarious legal status” or “family life” where they are unlawfully present (see Nationality, Immigration and Asylum Act 2002 s117B(4) &(5)). In this context, the Joint Committee of Human Rights has stated they were ‘uneasy about a statutory provision which purports to tell courts and tribunals that “little weight” should be given to a particular consideration in such a judicial balancing exercise.’


Do judges really need to be told how to weigh such evidence? Judges are already adept at evaluating the credibility of further evidence. If the evidence was available previously, a judge will rightly be critical of why it was not adduced at an earlier stage. If the evidence does not go to the heart of why the claim was refused, it is unlikely to found an arguable case. This direction to give minimal weight to new evidence appears to be designed to force judges to do so, contrary to their own better judgement. Presumably, the Home Office will provide its decision makers with guidance on what constitutes a “good reason” and will try to ensure that tribunal judges also follow this approach. One imagines they will seek a narrow definition.


It is possible to imagine a number of scenarios where vital evidence emerges at a late stage: an applicant lacks initial legal advice or receives poor quality advice; a subsequent legal adviser obtains expert evidence or finds an important country evidence report that the previous adviser missed; an applicant struggles to contact relatives, but eventually obtains important evidence from abroad; an applicant is so overwhelmed by the process that they struggle to engage initially and to make sufficient effort to obtain evidence from home. Would any of these examples given be considered a “good reason”? Ultimately, what matters is whether the individual is a refugee and whether they will face a risk if they are returned. Presumably, the only way a judge could apply this in a way that maintains compliance with the Refugee Convention would be to interpret the concept of “good reason” expansively to cover any circumstance in which a failure to give weight to the evidence would result in a breach of the UK’s obligations under the Convention. This, therefore, seems to be a pointless additional requirement which will add further legal complexity to the issue, rather than improving decision making.


2. Limiting the Concept of Particular Social Group


The concept of particular social group (PSG) has permitted the Convention to be interpreted expansively, providing protection to, amongst others, children, victims of trafficking, victims of domestic violence, those with mental health disorders, those targeted due to family membership or blood feuds. Clause 30 (3)-(4) of the new Bill seeks to limit this by establishing that in order for a PSG to be recognised it must satisfy both the “protected characteristics” and the “social perception” approaches to defining a PSG – an approach which was explicitly rejected in the House of Lords case of Fornah.[2] However, the cumulative approach was adopted by the Upper Tribunal in SB Moldova and reaffirmed in AZ Thailand, where it was asserted that it was a necessary interpretation of the QD. Yet, this is arguably inconsistent with the Convention and runs the risk of creating protection gaps, particularly for women who frequently rely on PSG as a Convention reason.[3] UNHCR in their comments on the EU Directive advised that these approaches should be considered as alternatives, rather than cumulative.


In the recent case of DH (Particular Social Group: Mental Health) Afghanistan, the tribunal revisited this debate at some length in the context of deciding whether Afghans with mental health problems could be considered to belong to a PSG. They rejected the approach in SB and AZ and explicitly noted:

The difficulties created by the use of the word “and” can give rise to protection gaps which is contrary to the obligations of signatories to the Refugee Convention.


Focusing on the underlying objectives of the Convention, they concluded that the EU Directive could not provide less protection than that afforded by the Refugee Convention. Therefore, ‘the concept of a PSG should be interpreted in an inclusive manner by determining that it exists on the basis of either an innate or common characteristic of fundamental importance i.e. the protected characteristics approach (‘ejusdem generis’) or social perception, rather than requiring both.’


Clause 30 therefore is an attempt to overturn the guidance provided by Fornah and by the tribunal in DH, in a manner which is inconsistent with the aims of the Refugee Convention.


3. Defining “Particularly Serious Crime”


In Clause 35 of the Bill the government revisits an old favourite – defining “particularly serious crimes” for purposes of exclusion from the Article 33 non-refoulement protection of the Refugee Convention.


Sub-clause 3 removes the rebuttable presumption that a 2-year sentence constitutes a particularly serious crime.[4] Henceforth, a 12-month sentence or an offence specified by order is deemed to amount to a particularly serious crime, bringing exclusion from protection of the Convention in line with the threshold for automatic deportation. Conveniently it also aligns it with the 12-month sentence upon summary conviction available for those who arrive in the UK without entry clearance introduced by clause 37. New clause 34 also seeks to limit the immunity from prosecution provided by Article 31 of the Refugee Convention for those who have arrived irregularly after having been in another country where they could have sought protection. The intention here appears to be to use a refugee’s irregular arrival as a mechanism to exclude them from protection of the Convention on the basis that they have been convicted of a particularly serious crime.


There remains the ability to rebut the presumption that, having been convicted of such a crime, the person constitutes a danger to the community of the UK.[5] Yet, this raises the prospect of many more applicants having to rebut this presumption prior to being able to make out their claim under the Refugee Convention.


Once again, this is an issue that has already been considered by the courts. The 2009 case of EN Serbia concerned the government’s attempt to create a domestic list of crimes that should be defined as “particularly serious” for purposes of the Convention. The implementing order included an incredibly wide range of offences, such that a refugee convicted of stealing a bottle of milk could be presumed as having been convicted of a particularly serious crime. The UNHCR’s position is that the concept of “particularly serious crime” must be considered within the context of the Convention and what it seeks to achieve. The UNHCR Handbook describes it as ‘a capital crime or a very grave punishable act.’ In EN Serbia the Court of Appeal rejected the UK government’s approach, holding that the term “particularly serious crime” has an autonomous meaning as a term in international law. It was stated that an irrebuttable presumption that every crime punished with a 2-year sentence was particularly serious would be incompatible with the Convention. The 2004 Order, with its expansive list of offences, was held to be ultra vires since it listed crimes which could not reasonably be construed as “particularly serious” within the context of the Convention.


Clause 35 is explicitly contrary to that judgment and raises the prospect of further litigation to reconcile this new approach with the claim by the government that it intends to act in conformity with the Convention.


Is this really what is needed? – A Recipe for Litigation


A question to be asked here is why any of these changes are perceived to be necessary, at this point in time, when the Home Office is already dealing with the complex changes to EU immigration caused by Brexit. Yes, changes to the definition of PSG or to the standard of proof may have the effect of denying protection to some applicants who would otherwise have been granted refugee status. Yet, the problem of removing those denied protection for technical reasons will still remain. What these changes guarantee though is a decade of renewed litigation, which will no doubt end with several Supreme Court judgments. One would have thought that having spent the best part of a decade re-litigating Article 8 ECHR, resulting in significant legal complexity for all concerned, the last thing the Home Office needs is another 10 years of legal disputes over the meaning of the Refugee Convention, on issues that had already been put to rest. No doubt this will provide lawyers and academics with many hours of work. But if the aim is to improve the efficiency of the asylum process, making numerous technical changes to how the Convention is interpreted seems a counterintuitive approach.    



[1] Immigration Act 2014, s19 which inserted s117A-D into the Nationality, Immigration and Asylum Act 2002.

[2] Per Lord Bingham ‘If, however, this article were interpreted as meaning that a social group should only be recognised as a particular social group for purposes of the Convention if it satisfies the criteria in both of subparagraphs (i) and (ii), then in my opinion it propounds a test more stringent than is warranted by international authority’ [16].

[3] See Christel Querton, ‘The Interpretation of the Convention Ground of `Membership of a Particular Social Group´ in the Context of Gender-related Claims for Asylum A critical analysis of the Tribunal’s approach in the UK’, (2013) Refugee Law Initiative Working Paper No 3.

[4] This is done by way of amendment to s72 Nationality, Immigration and Asylum Act 2002.

[5] New sub-s5A inserted into the Nationality, Immigration and Asylum Act 2002 s72, taken with sub-s6.    



The views expressed in this article belong to the author/s and do not necessarily reflect those of the Refugee Law InitiativeWe welcome comments and contributions to this blog – please comment below and see here for contribution guidelines.