Written evidence from Dr Bharat Malkani; Dr Paul McDonough; Ms Jennifer Morgan (NBB0048)



Ms Jennifer Morgan, Tutor in Law and CILEx Lawyer, School of Law and Politics, Cardiff University

Dr Paul McDonough, Lecturer in Law, School of Law and Politics, Cardiff University

Dr Bharat Malkani, Senior Lecturer in Law, School of Law and Politics, Cardiff University


This submission aims to assist the Parliamentary Human Rights Committee in understanding the impact of some of the changes to the law put forward by the Nationality and Borders Bill. The issues raised here all share one concern: certain changes proposed in the Bill would risk opening new protection gaps, ie could result in the UK declining to protect a refugee to whom it owes protection under the Refugee Convention.

The authors of this submission are academic members of the School of Law and Politics at Cardiff University. We all have a research interest in immigration and asylum law, and Jennifer Morgan also has a practice background in this field.


Does introducing a two-tier system of rights for refugees meet the UK’s obligations under refugee law and human rights law?

Clause 10 sets out a proposed differential categorisation of recognised refugees into 2 groups – group 1 and group 2. The proposed criteria for separating the two groups relies solely on the mode of travel of the refugee from the country of persecution to the United Kingdom. Clause 10(2) states that the specific considerations are: a) whether the refugee has come directly from the home country to the United Kingdom and b) whether the refugee has presented themselves to the UK authorities without delay. This response will focus on part a of those considerations and its consequences.

A two-tier system would open a protection gap. There are two fundamental problems. First, assignment to one or the other group is based on factors unrelated to protection needs. Second, the absence of sufficient constraints around the ability of the Secretary of State or other immigration officials to make and apply the Immigration Rules could result in refugees failing to receive a full and fair determination of their claims for asylum. The risk is of creating a situation where, even inadvertently, UK authorities could return genuine refugees to countries where their life or freedom is at risk, merely as a result of their potentially unlawful entry or presence here.

The wording in clause 10 reflects the wording of the Refugee Convention at Article 1(A)(2). However, there is no mention of the protection that is given under Article 31 of the Refugee Convention and written into statute at Section 31 Immigration and Asylum Act 1999. Instead, Clause 10 indicates that those travelling to, and entering the United Kingdom via other countries will be classed as not having “come directly” on a narrow interpretation of those words.

The UNHCR guidance (quoted in the High Court case of R v Uxbridge Magistrates’ Court ex parte Adimi and others [1999] EWHC Admin 765) on detention of asylum seekers states as follows:

The expression ‘coming directly’ in Article 31(1) covers the situation of a person who enters the country in which asylum is sought directly from the country of origin, or from another country where his protection, safety and security could not be assured. It is understood that this term also covers a person who transits an intermediate country for a short period of time without having applied for, or received, asylum there. No strict time limit can be applied to the concept ‘coming directly’ and each case must be judged on its merits."

In the case of R v SSHD ex parte Adan and another [1999] 1 AC 293, the House of Lords noted the following:

17. The respondents accept that a literal construction of “directly” would contravene the clear purpose of the Article and they accordingly accept that this condition can be satisfied even if the refugee passes through intermediate countries on his way to the United Kingdom. But that is only so, they argue, provided that he could not reasonably have been expected to seek protection in any such intermediate country and this will not be the case unless he has actually needed, rather than merely desired, to come to the United Kingdom. In short it is the respondents’ contention that Article 31 allows the refugee no element of choice as to where he should claim asylum. He must claim it where first he may: only considerations of continuing safety would justify impunity for further travel.

18. For my part I would reject this argument. Rather I am persuaded by the applicants’ contrary submission, drawing as it does on the travaux préparatoires, various Conclusions adopted by UNHCR’s executive committee (ExCom), and the writings of well respected academics and commentators (most notably Professor Guy Goodwin-Gill, Atle Grahl-Madsen, Professor James Hathaway and Dr Paul Weis), that some element of choice is indeed open to refugees as to where they may properly claim asylum. I conclude that any merely short term stopover en route to such intended sanctuary cannot forfeit the protection of the Article, and that the main touchstones by which exclusion from protection should be judged are the length of stay in the intermediate country, the reasons for delaying there (even a substantial delay in an unsafe third country would be reasonable were the time spent trying to acquire the means of travelling on), and whether or not the refugee sought or found there protection de jure or de facto from the persecution they were fleeing.

It is submitted that the Home Office put forward the proposition that is contained in clause 10 over 20 years ago and that the House of Lords, the highest domestic court in the United Kingdom, roundly rejected this interpretation as being an incorrect interpretation of the Refugee Convention. On this basis, it is impossible to state that the content of this clause i.e the differentiation of asylum seekers into two groups in this way could meet the United Kingdom’s obligations under international law. 

In addition to this, provisions already exist under the Refugee Convention to penalise anyone, including refugees, for unlawful entry or presence. Article 31(1) of the Refugee Convention explicitly acknowledges this. However, the penalties available under the Refugee Convention do not rise to the level of denying the full and fair hearing of an asylum claim, as this would risk de facto exclusion from refugee status and potential expulsion to face a risk to life or freedom. This is not sanctioned by the Refugee Convention. Rather, the two matters are independent: if an asylum applicant is determined to be a refugee, they can then be subjected to normal criminal or civil penalties resulting from their unlawful entry or presence as described in Article 31(1). If the applicant is found not to require protection, they can be expelled or otherwise penalised without violating the Refugee Convention.

Were this clause to become law, it is clear that many persons who currently benefit from protection in the United Kingdom under the Refugee Convention would be unable to avail themselves of such protection.

According to Home Office figures, in the year to September 2020 over 50% of asylum seekers granted refugee status directly by the Home Office following the application stage (i.e no appeal proceedings) were from Iran, Sudan, Eritrea and Afghanistan. There are no resettlement programmes for persons travelling from these countries (with the exception of Afghanistan, and criticism of the scope of this has been widespread taking into account recent events). In part, this is because resettlement schemes are generally made available to those fleeing war, which is not appropriate in the case of these countries (and incidentally fleeing war solely on account of being a civilian is generally considered to be outside the scope of the Refugee Convention). Persons coming from the countries mentioned above most often travel by irregular means. The impact of clause 10 would be to prevent the vast majority of these persons from being eligible for a grant of Refugee Status. It is submitted that such a situation cannot possibly meet the UK’s international obligations under the Refugee Convention. If these people are refugees under the convention now, then there can be little justification for deciding they are no refugees purely because of their mode of travel. 

The consequence of differentiation under clause 10 would be that the vast majority of the persons currently granted refugee status by the Home Office currently would be placed into so-called “group 2”, Most of these people will have strong claims for protection, given that they would be granted asylum under current arrangements. From practice experience of working in asylum law for many years, one of the authors is aware that this group is highly likely to include domestic abuse survivors, people at risk of honour killing and political activists from Iran; non-Arab Darfuri tribe members and political activists from Sudan; evangelical Christians and military deserters from Eritrea; and Sikhs, religious minorities and political activists from Afghanistan. Therefore, the impact of clause 10 will have a devastating effect on genuine refugees who find themselves in the United Kingdom and in need of protection.

The potential negative impact of clause 10 is exacerbated by the undefined degree of discretion it would afford the Secretary of State or other officials to alter the asylum procedure or social security rights pertaining to refugee status, based merely on past unlawful entry or presence. Absent further guidance in the Bill, clauses 10(5)(a) and 10(5)(b) are simply too vague to be able to anticipate their likely effect. It is not at this point clear how the Secretary of State or her officials might be authorised to interpret the power to differentiate the duration of leave to remain, or ‘requirements that the refugee must meet [for] indefinite leave’, respectively. It is submitted that bringing about such amendments via changes to the Immigration Rules is unsatisfactory given the lack of Parliamentary scrutiny afforded to the Statements of Changes to the Immigration Rules.


A more concrete concern arises with respect of clause 10(5)(c), which purports to allow refusal of access to public funds as a consequence of unlawful entry or presence prior to the recognition of refugee status. This could in principle improperly sanction derogations from Article 24(1)(b) of the Refugee Convention regarding social support payments. This would result in a person who under the current system would be entitled to full access to benefits being denied such access. In addition, failure to provide support may cause a person who has established a risk in the home country to become destitute which would breach Article 3 ECHR. It is submitted that, even persons who have been granted Discretionary Leave following an asylum claim, are able to apply for the no recourse to public funds condition to be removed from their period of leave to remain. Therefore, this provision would result in people who have established a risk in their home country being left in a worse position than those who are in the United Kingdom for other reasons.


Finally, clause 10(5)(d) would create a risk of placing the UK in violation of Article 8 ECHR, governing the right to respect for family life. It is submitted that the right to family reunion should not be affected by the mode of travel to the UK of a refugee. This is particularly true given that creating a distinction based on mode of travel to the UK contravenes the accepted international interpretation of the Refugee Convention.


In conclusion, dividing asylum applicants into two groups based on factors that do not bear on the merits of their asylum claims will lead to either a protection gap, or an unnecessary administrative burden and consequent drain on state resources, or both. Asylum determination is always fraught with potential implications for refugee rights the UK has committed to uphold. The risk of rights violations is immediately exacerbated when individuals within the asylum-seeking population can be treated differently for reasons that do not relate to their protection needs, at the discretion of the Home Secretary and how she wishes to amend the Immigration Rules. A risk of arbitrary discrimination, or allegations of the same, will certainly arise. It is highly likely that a significant amount of litigation will be triggered over the coming years to establish the legality of provisions within the Immigration Rules. The situation, in addition, would of course carry concomitant risks of causing the UK to violate its obligations under the Refugee Convention, and to drastically alter the lives of individual refugees and their families without due justification.



Will the proposed instructions to decision-makers on how to interpret the Refugee Convention secure or restrict the protections that Convention guarantees?

This section of the Bill appears to incorporate the content of the Refugee or Person in Need of International Protection (Qualification) Regulations 2006 and aspects of the Refugee Qualification Directive. Clause 27(4) explicitly states that these Regulations will be revoked by the Bill. However, there are some concerning departures from the position as it stands currently. The largest concern relates to the standard of proof asylum claimants must meet to establish their claims for protection.

Standard of proof (clause 29)

Requiring a higher standard of proof for asylum claims will prevent genuine refugees from being able to adequately prove their cases, given the inherent difficulties in evidencing facts in asylum claims. This is because refugees by their nature may be forced to flee at the last minute with very little in the way of documentary evidence.

R v SSHD ex parte Sivakumaran 1988 AC 988 established that the test for identifying a well-founded fear of persecution per Article 1(A)(2) of the Refugee Convention is a reasonable degree of likelihood that persecution would occur for a convention reason.

This is the current standard of proof that must be met for an asylum claim to succeed, as confirmed in the leading case of R v IAT and another ex parte Rajendrakumar [1996] Imm AR 97 – “a reasonable degree of likelihood”. This is stated to be lower than the civil standard “on the balance of probabilities” and can also be referred to as “substantial grounds for believing” or “a real risk”.

Karanakaran v SSHD [2000] Imm AR 271 clarified this test further. The Court of Appeal in that case affirmed the four categories of evidence that might present themselves in an asylum application or appeal as set out in Kaja (political asylum: standard of proof) (Zaire) Imm AR 1. The Court of Appeal approved the categorisation of evidence set out in Kaja providing that evidence satisfying any of the first three categories (below) should not be excluded from consideration. The categories outlined in Kaja and approved in Karanakaran are:

1)     Evidence they are certain about;

2)     Evidence that they think is probably true;

3)     Evidence to which they are willing to attach some credence, even if they would not go as far as to say it is probably true;

4)     Evidence to which they are unwilling to attach any credence at all. 

In addition, in KS (benefit of the doubt) [2014] UKUT 552 (IAC), the Tribunal confirmed that evidence should be considered in the round and that, where a claimant is generally considered to be credible, benefit of the doubt should be given in respect of facts that the decision maker is not sure are true but could be given some credence. This standard of proof is also confirmed in the UNHCR handbook and in the Qualification Directive (2004/83/EC).

The current standard of proof provides a clear framework for decision makers and the judiciary to utilise when considering the evidence put forward by asylum seekers and allows flexibility and for judgement to be exercised taking into account the whole of the claim, the profile of the asylum seeker and their specific circumstances.

Clause 29 seeks to change the standard of proof for asylum claims by effectively creating two standards of proof to operate alongside one another. The Bill intends that the consideration of facts in an asylum claim should be to the higher civil standard of proof “on the balance of probabilities” whilst the assessment of risk should remain subject to the standard of “a reasonable degree of likelihood”. The effect would be to remove part three of the four-stage test outlined in Kaja and affirmed in Karanakaran from a decision-maker’s consideration.

The aforementioned case law came before the United Kingdom courts as a result of confusion regarding the appropriate standard of proof for asylum cases throughout the latter part of the 20th century. The subject of whether or not there should be a two-part standard of proof in asylum cases has therefore been considered in depth by the higher courts of the United Kingdom previously. The outcome was that a two part test was neither desirable nor workable.

For example, Lord Justice Brooke giving the leading judgment in Karanakaran regarded a two part test with separate standards of proof for facts and risk on return as “an unnecessary complexity” and stated:

“In many of these cases… the evidence will be the applicant’s own story, supported in some instances by reports from organisations like Amnesty International. The stress generated by the nature of an asylum claim and the possible consequences of refusal, complemented by the highly formalistic atmosphere of interview or court, made the task of evaluating the evidence more complex. This did not mean that there should be a more ready acceptance of fact as established as more likely than not to have occurred. On the other hand, it created a more positive role for uncertainty. It would be a rare decision-taker who was never uncertain about some aspects of the evidence, particularly where, unlike civil litigation, evaluation was often concerned only with one version of the “facts”. To say that it is only the facts established as more likely than not to have occurred on which the “reasonable likelihood” must be based would be, they said, to remove much of the benefit of uncertainty conferred on an applicant through Sivakumaran.”

Therefore, by adopting a two-part test where facts are determined on the balance of probabilities, asylum seekers would lose the benefit conferred by current caselaw that allows uncertain (but possibly true) facts to be considered alongside more certain facts in order to determine the truthfulness of an account in the round.

Not only is this approach at odds with the guidance in the UNHCR handbook and the Refugee Qualification Directive, but also with the approach taken in other common law countries such as Australia. In Chan (1989)169 CLR 379, the Australian High Court identified that the correct test was whether there was “a real chance” persecution would occur. The court in that case indicated that even a 10 percent chance would be sufficient to establish an unacceptable risk on return. Gaudron J stated:

“Perhaps all that can usefully be said is that a decision-maker should evaluate the mental and emotional state of the applicant and the objective circumstances so far as they are capable of ascertainment, give proper weight to any credible account of those circumstances given by the applicant and reach an honest and reasonable decision by reference to broad principles which are generally accepted within the international community.”

In Wu Shan Liang (1996) 185 CLR 259, the Australian High Court stated:

“Where facts are in dispute in civil litigation conducted under common law procedures, the court has to decide where, on the balance of probabilities, the truth lies as between the evidence the parties to the litigation have thought it in their respective interests to adduce at the trial. Administrative decision-making is of a different nature. A whole range of possible approaches to decision-making in the particular circumstances of the case may be correct in the sense that their adoption by a delegate would not be an error of law. The term ‘balance of probabilities’ played a major part in those submissions, presumably as a result of the Full Court’s decision. As with the term ‘evidence’ as used to describe the material before the delegates, it seems to be borrowed from the universe of discourse which has civil litigation as its subject. The present context of administrative decision-making is very different and the use of such terms provides little assistance.”

In Minister for Immigration and Citizenship v SZQRB [2013] FCAFC 33, the Full Court of the Federal Court of Australia confirmed again that the correct standard of proof is “a real chance” that persecution will occur.

It is clear, therefore, that the proposal in the Bill to amend the standard of proof in asylum claims goes against the considered opinion of the United Kingdom domestic courts and will push the United Kingdom out of alignment with how asylum claims are considered in other democratic nations such as comparable European and commonwealth countries. The proposals are also not aligned with the guidance in the UNHCR handbook. The proposals in clause 29 are highly likely to create a situation where genuine refugees are unable to substantiate their claims to the higher standard of proof, resulting in real risk of refoulement to countries where they will suffer persecution and ill-treatment. For this very reason, the standard of proof in asylum claims has always been lower than the civil standard. The lower standard is intended to ensure that the aims of the Refugee Convention in preventing refoulement are met. It is therefore clear that this change to the standard of proof would unfairly restrict access to protection under the Refugee Convention.

Further concerns

While the standard of proof proposed in clause 29 is the largest concern raised by this section of the Bill, there are other important flaws that relate to the interpretation of specific terms in the Refugee Convention.

Clause 28(1) – actors of persecution

The language of clause 28(1) approximates Article 6 of the Qualification Directive, but would open a further loophole, via the insertion of the word ‘reasonable’ in 28(1)(c) – ie a risk of persecution is obviated not by ‘protection’ from a state or non-state actor, but by ‘reasonable protection’. This grants the decision maker extra leeway to declare that an acknowledged risk of persecution can be disregarded, due to a reasonable possibility of protection against it, rather than due to demonstrable actual (or effective) protection.

Clause 28(3) – acts of persecution

Clause 28(3) closely resembles Article 9(2) of the Qualification Directive, except for the omission of the Directive’s instruction that acts of persecution can ‘take the form of … (f) acts of a gender-specific or child-specific nature’. This is a curious omission. Although the list of possible acts of persecution is indicative rather than exhaustive, it is not clear why a decision maker should not be reminded of the importance of taking gender or childhood into account, as these are now recognised internationally as factors that can and do lead to particular forms of persecution.

Clause 30 – reasons for persecution

There is a potential protection gap between clause 30(3) and 30(4). Essentially, this is presented as an ‘and’ test, when it should be ‘or’. In the current draft, for example, if a person actually does not share a common characteristic (30(3)) but is still perceived as part of the group that shares that characteristic (30(4)), and at risk of persecution on that basis, the decision maker would be instructed to find against the existence of that risk.

Clause 31 – protection from persecution

Clause 31(1)(b) improperly permits decision makers to ascribe responsibility for protection to actors that are not subject to international law. The Refugee Convention provides for surrogate protection by a state that will uphold international human rights standards. Non-state actors are not generally beholden to international law, in particular UN and regional human rights treaties, as states are. As a practical matter, non-state actors exercising control over part of a state’s territory present an inherently unstable situation: as tragedies such as Srebrenica have shown, even with best intentions and significant armed forces available, protection by non-state actors can fail catastrophically.

In clause 31(2)(a), the words ‘reasonable steps’ open a potential protection gap. From the point of view of the endangered person, the meaningful inquiry is not whether the state has taken reasonable steps, but whether there is actually a risk of persecution. Furthermore, the Bill as currently drafted is more restrictive than the Qualification Directive, in that by omitting the words ‘inter alia’ (Article 7(2)), it presents the operation of an effective legal system as the sole criterion by which to evaluate a state’s protective capacity.

Clause 32 – internal relocation

Internal relocation, or internal protection, is not intrinsic to international refugee law, nor contemplated in the Refugee Convention. Rather, it is a construct invented by states as a potential means to deny protection to some who would otherwise be found to be refugees. Application of the internal protection alternative is not necessarily unlawful, but must be done with due care, and with a prejudice toward finding that it is inapplicable i.e that the person is a refugee.

The Bill contains a serious flaw in clause 32(2)(b). As drafted, it not only permits, but requires decision makers to disregard the actual ability of the person concerned to reach the purported zone of protection. As the UK judiciary have made clear in numerous cases, particularly those pertaining to applicants from Somalia (see for example AMM and others (conflict; humanitarian crisis; returnees; FGM) Somalia CG [2011] UKUT 00445 (IAC)) the expectation of protection in a region within the home country is meaningless unless the person can actually travel there and settle safely. In most of its points of difficulty, the flaws in this Bill create potential protection gaps. Here, a failure of protection is plainly visible in the language of the Bill.

Clause 33 – exclusion from refugee status

Clause 33(3) and 33(4)(b) in conjunction present a misunderstanding of the term ‘admission … as a refugee’ used in Article 1(F)(b) of the Refugee Convention. The Bill understands admission to occur upon the issuance of a biometric identification card following recognition as a refugee. However, it is clear from the structure and text of the Convention that ‘admission … as a refugee’ must refer to the first moment following the making of an asylum claim, not following determination of that claim. Several rights in the Convention, most prominently, but not only, the right of non-refoulement stated in Article 33(1) cannot sensibly be read to indicate that a person is a refugee only after their claim is determined – this right attaches immediately upon engagement of the state’s jurisdiction, therefore the Convention recognises that refugeehood does not depend on state determination. Similarly, rights such as access to public education (Article 22) do not depend on a refugee’s presence being lawful, and the right to self-employment (Article 18) engages upon lawful presence, while the right to wage earning employment (Article 19) is not available until a refugee is ‘lawfully staying’ in the host state.

The Refugee Convention, therefore, recognises that refugee status is based on events that occur up to the point where a would-be host state takes jurisdiction over the person. Furthermore, at the point of admission to the asylum process, an applicant has been recognised as lawfully in the territory of the host state (by acts of eg registration, issuance of temporary identification papers, etc). At that point, in the words of the Convention, the person has been ‘admi[tted] to that country as a refugee’. Any crimes committed thereafter are subject to the normal criminal jurisdiction of the host state, and cannot result in the negation of refugee status, which came into being upon the person’s departure from their home country.

Clause 35 – particularly serious crime (Refugee Convention, Article 33(2))

Clause 35 redefines a ‘particularly serious crime’ leading to the loss of the protection of non-refoulement as final conviction any offence that receives a sentence of 12 months, whether committed and convicted in the UK or abroad. It leaves the decision maker no discretion to assess the circumstances of the offence. This is contrary to most interpretations of the Refugee Convention, including that of the UK Asylum and Immigration Tribunal in IH v SSHD [2009] UKAIT 00012 which stated that Article 33(2) of the Convention must take into account all circumstances of the offence and the offender. Furthermore, according to the Convention, conviction for a particularly serious crime is only the first part of a two-prong test: as stated in EN (Serbia) and KC (South Africa) v SSHD [2009], EWCA 630 at [39], the decision maker must also find that the refugee ‘constitutes a danger to the community’. The presumption in clause 35(6) must therefore at least be rebuttable. For example, the refugee could show that the nature of the crime shows that its repetition in the host country is unlikely (EN (Serbia) and KC (South Africa) v SSHD [2009], EWCA 630 at [66]).