[Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]
	England and Wales Court of Appeal (Civil Division) Decisions
You are here: BAILII >> Databases >> England and Wales Court of Appeal (Civil Division) Decisions >> Horvath v Secretary of State for the Home Department [1999] EWCA Civ 3026 (02 December 1999) URL: http://www.bailii.org/ew/cases/EWCA/Civ/1999/3026.html Cite as: [2000] Imm AR 205, [2000] INLR 15, [1999] EWCA Civ 3026

[New search] [Printable RTF version] [Help]

IN THE SUPREME COURT OF JUDICATURE
COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM THE IMMIGRATION APPEALS
TRIBUNAL

		Royal Courts of Justice Strand, London, WC2A 2LL
		2nd December 1999

LORD JUSTICE STUART-SMITH:

1. This is an appeal by Milan Horvath from a decision of the Immigration Appeal Tribunal (the Tribunal) given on 4 December 1998 by which the Tribunal dismissed his appeal from the Special Adjudicator, who in turn dismissed his appeal from the Secretary of State for the Home Department's decision to refuse his application for asylum. The Appellant is a member of the Roma community and is a citizen of Slovakia.
2. The appeal raises an important question as to how Special Adjudicators and the Tribunal should approach the problems of persecution by non-state agents, that is to say factions or elements in the population of a country for which the state has no direct responsibility. In particular, the question that falls to be determined is at what stage in the inquiry the fact finding tribunal should consider the issue of the applicant's inability or unwillingness, through fear of persecution, to avail himself of the protection of the state. This issue involves considerations as to the ability or willingness of the state and its agents to provide adequate protection from the activities of non-state agents. The question is of importance because it may be affected by the standard of proof to be applied in consideration of different aspects of a claim to asylum.
3. The Tribunal appears to have considered the question in the first instance as being relevant to the issue whether what the Appellant suffered amounted to persecution. But they also considered it as a separate question, namely whether the Appellant was unable, or unwilling through fear of persecution, to avail himself of the protection of the state. They decided both issues against the Appellant.
4. Counsel on behalf of the Appellant contend that the Tribunal were in error in so doing. Their submissions in outline are as follows:

(a) The Tribunal were in error in taking the question into consideration when deciding whether what the Appellant had been subjected to amounted to persecution. The sole question, on this aspect of the case, is whether the ill-treatment received by the Appellant at the hands of the non-state agents, and to be expected and feared from them on return to his country, amounts to persecution.

(b) The question falls to be considered when the Special Adjudicator or Tribunal is considering whether the Appellant's fear of persecution is well-founded. If it falls to be considered at this stage, then the lower standard of proof, namely that there is a reasonable likelihood of persecution if the Appellant is returned, is applicable (R v Secretary of State for the Home Department, ex parte Sivakumaran and others [1998] AC 958). Had the Tribunal so considered it they would or should have concluded that the Appellant had a well-founded fear of persecution.

(c) That the Tribunal should not have considered the question as a separate issue. But if they were right to do so, their decision was perverse; alternatively they took into consideration irrelevant matter and in so doing reached a conclusion inconsistent with their own earlier finding.

(d) The grounds of appeal raise further matters, the most important of which is whether certain acts of discrimination affecting his ability to find work, and denial of normal public facilities including marriage and education for his child, considered separately or in conjunction with the persecution by non-state agents, amounted to persecution. These have been conveniently referred to as 'Category Three Rights' adopting the well-known classification by Professor Hathaway in his book 'The Law of Refugee Status' at page 109.

(e) Counsel for the Appellant made a number of other discrete criticisms of the decision of the Tribunal.

5. On behalf of the Secretary of State, Mr Howell, QC, in summary submitted:

(a) That the question whether the Applicant was unable, or through fear of persecution, unwilling to avail himself of the protection of the state was a discrete and separate issue which fell to be determined if the Appellant satisfied the Adjudicator or Tribunal that the ill-treatment amounted to persecution for a convention reason and that he had a well-founded fear of being persecuted if returned to Slovakia. It should not be considered as an ingredient in these two issues.

(b) Alternatively if his first submission was not correct, the Tribunal were right to consider it in relation to the question whether what the Appellant feared amounted to persecution and that it did not form part of the consideration of whether the Appellant's fear was well-founded.

(c) That the Tribunal's conclusion that the Appellant had not established that he was unable or unwilling through fear of persecution to avail himself of the state's protection was a finding of fact which the Tribunal were entitled to reach and it could not be impugned on the grounds of perversity or consideration of irrelevant material. Even if the Tribunal were wrong to consider the point in relation to the issue of persecution, this did not affect their conclusion on the protection issue.

(d) That the Tribunal did not misdirect themselves when considering whether abuse of the third category rights amounted to persecution and their decision on this could not be impugned.

6. The facts can be succinctly stated. The Appellant is 26. He comes from a village called Palin from the county of Mikhalovice, where the Roma community, to which he belongs, are a small minority. On 15 October 1997 he arrived in the United Kingdom with his wife and child and claimed asylum. He stated he feared persecution in Slovakia by skinheads, against whom the Slovak police failed to provide protection for Roma. Among the episodes to which he referred was the beating to death of his father (which did not even lead the police to come to his house: "They pretended it hadn't happened"), this was in 1985 under the communist regime; an attack on his brother by skinheads armed with vicious weapons; persistent attacks on his home, leading the Appellant and his brothers to dig a hole in the ground in their back garden and regularly take shelter in it at night; the destruction by skinheads of every item in the Appellant's home, leaving an empty shell; attacks on all the Roma in the Appellant's village ("the police didn't want to get involved"); serious violent attacks on two Roma neighbours; and the murder of two others. He said that if they were returned to Slovakia he was afraid that he would again be persecuted by the skinheads because he was a gypsy. He will not get protection from the police who don't care at all about their problems. His written statement reported that "skinheads would come to our village and throw bombs into the homes of gypsies"; that "approximately 3 to 4 times a week neo-Nazis would....hurl abuse outside my windows that all gypsies must die", that the situation "caused me to fear for my life" and that he came to the United Kingdom because he wanted his child to grow up in a country in which she is not persecuted . He also stated that along with other Roma, he was unable to find work and the Slovak authorities failed to afford him normal public facilities including marriage and schooling for his child.
7. His application for asylum was rejected by the Secretary of State. His appeal was dismissed by the Special Adjudicator on the grounds of his credibility. The Tribunal, which consisted of three legally qualified members under the chairmanship of the President HH Judge Pearl, found that the Appellant's assertions of fact were consistent with the evidence relating to the position of Roma in Slovakia. In this context the Tribunal considered reports of the UNHCR, the United States State Department, the International Helsinki Federation for Human Rights and the Conference on Security and Co-operation. Accordingly they reversed the Special Adjudicator's conclusion on credibility. They stated that "there is enough evidence for us to conclude that the fear of the Appellant is well-founded''. This must be a reference to the fear of violence by the skinheads. It was not in dispute that the ill-treatment of the skinheads was by reason of his race. But the Tribunal concluded that this did not amount to persecution, because the Appellant had not discharged the burden of showing that he was unable or, through fear of persecution, unwilling to avail himself of the state's protection.
8. Mr Plender QC, on behalf of the Appellant, submits that that the Tribunal were in error to introduce the element of the state's ability to provide protection into the question whether what the Appellant feared amounted to persecution. He submitted that the ill-treatment by the skinheads involving acts of violence against his relatives and neighbours and the Roma community generally amounted to persecution.
9. Entitlement to refugee status is defined in Article 1 of the Geneva Convention and Protocol Relating to the Status of Refugees (1951) (Comd. 9171) and 1967 (Comd 3906). This provides as follows:

"A. For the purpose of the present Convention, the term 'Refugee' shall apply to any person who....(2) owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country or......"

10. It is apparent that there are five conditions that the applicant must satisfy to establish his status as a refugee, namely that:

(1) He is out of the country of his nationality because he has a fear of ill-treatment.

(2) The ill-treatment that he fears is of a sufficiently grave nature as to amount to persecution.

(3) His fear of persecution is well-founded.

(4) The persecution is for a convention reason.

(5) He is unable, or owing to fear of the persecution, is unwilling to avail himself of the protection of that country.

11. These are separate and discrete tests each of which must be satisfied. Logic and convenience suggest that the fact-finding tribunal should address each question sequentially. Some issues may not be in dispute; equally in some cases there may be a short and obvious answer to the application on one particular ground. But I can see no reason or advantage in importing into the consideration of one issue, matters which logically fall to be considered under another. On the contrary it seems to me to be likely to lead to confusion. Thus as a matter of principle consideration of the question whether the Applicant has shown that he is unable, or through fear of the persecution is unwilling to avail himself of the state's protection, which in turn involves a consideration of the state's ability and willingness to afford protection - which I will call the protection test - properly concerns the fifth test and should not be confused with the first three questions, which can broadly together be described as the 'fear test'.
12. Again as a matter of principle it seems to me that the protection test has nothing to do with the second question, namely whether the ill-treatment which the Applicant has suffered and fears in the future amounts to persecution. It is now well-established that the word 'persecution' is to be given its ordinary dictionary meaning of 'to pursue with malignancy or injurious action especially to oppress for holding a heretical belief (see per Nolan J in R v IAT ex parte Jonah [1985] Imm. AR 7, 13). Equally it is well established that persecution can be at the hands of non-state agents such as neighbours, family (as in Islam v Secretary of State for the Home Department, R v IAT ex parte Shah [1999] 2 WLR 1015) or hostile factions (Adan v Secretary of State for the Home Department [1999] 1 AC 293. I see no reason why, if the ill-treatment received at the hands of such perpetrators is of sufficient gravity to amount to persecution, it should cease to have that quality because the state can offer adequate protection against it.
13. Both counsel submit that the question of the State's ability to provide protection should not be part of the consideration of the question whether the ill-treatment that the appellant fears amounts to persecution. Is there any authority binding on this court to suggest this is wrong? In support of his alternative argument referred to in paragraph 5(b) that the Tribunal were correct to consider that the question of protection was integral to this question, Mr Howell cited a dictum of Lord Hoffman in Islam; at p1034H he said:

"What is the reason for the persecution which the appellants fear? Here it is important to notice that it is made up of two elements. First, there is the threat of violence to Mrs Islam by her husband and his political friends and to Mrs Shah by her husband. This is a personal affair, directed against them as individuals. Secondly, there is the inability of the state to do anything to protect them. There is nothing personal about this. The evidence was that the state would not assist them because they were women. It denied them a protection against violence which it would have given to men. These two elements have to be combined to constitute persecution within the meaning of the Convention. As the Gender Guidelines for Determination of Asylum Claims in the UK (published by the Refugee Women's Legal Group in July 1998) succinctly puts it (at p5): "Persecution = Serious Harm = The Failure of State Protection."

But I agree with both counsel that this statement is not necessary for the decision in the case which was whether or not the women of Pakistan were a particular social group and the present issue was not before the House. There was no dispute in that case that women in Pakistan were discriminated against in matters of fundamental human rights, which amounted to persecution and that the state provided no protection for them. I do not think therefore that this brief exposition of Lord Hoffman can be taken as definitive. Moreover there are passages in the speech of Lord Steyn which appear to me to point in a different direction. At p1020H he said:

"In order to qualify as a refugee the asylum seeker (assumed to be a woman) must therefore prove (1) that she has a well founded fear of persecution; (2) that the persecution would be for reasons of race, religion, nationality, membership of a particular social group, or political opinion; (3) that she is outside the country of her nationality; (4) that she is unable, or owing to fear, unwilling to avail herself of the protection of that country."

And at p1021H he said:

"...the principal issue before the House is the meaning and application of the words "membership of a particular social group." It is accepted that each appellant has a well founded fear of persecution in Pakistan if she is returned to that country. The appellants are outside the country of their nationality. And they are unable to avail themselves of the protection of Pakistan. On the contrary, it is an unchallenged fact that the authorities in Pakistan are unwilling to afford protection for women circumstanced as the appellants are. Except for the requirements inherent in the words "persecution for reasons of.....membership of a particular social group" in article 1A(2) all the conditions of that provision are satisfied."

That seems to me to be a similar analysis to that which I have set out in paragraph 10 and suggests separation of the different ingredients. None of the other authorities to which Ward LJ refers in his draft judgment on this aspect of the case are binding on this Court. Nor do they persuade me that both counsel were in error on this point.

14. What is in point and is binding on us is the speech of Lord Lloyd of Berwick (with whom the other members of the House agreed) in Adan v Secretary of State of State for the Home Department [1999] 1 AC 293. At p304C Lord Lloyd divided asylum seekers into four categories. The first two of which are relevant:

"(1) Nationals who are outside their country owing to a well-founded fear of persecution for a Convention reason, and are unable to avail themselves of the protection of their country; (2) nationals who are outside their country owing to a well-founded fear of persecution for a Convention reason, and, owing to such fear, are unwilling to avail themselves of the protection of their country."

At E he said:

"It will be noticed that in each of categories (1) and (2) the asylum-seeker must satisfy two separate tests: what may, for short, be called "the fear test" and "the protection test"."

At p306A he said that category (1) included:

"..the important class of those who are sometimes called "third party refugees", i.e. those who are subject to persecution by factions within the state. If the state in question can make protection available to such persons, there is no reason why they should qualify for refugee status. They would have satisfied the fear test, but not the protection test. Why should another country offer asylum to such persons when they can avail themselves of the protection of their own country? But if, for whatever reason, the state in question is unable to afford protection against factions within the state, then the qualifications for refugee status are complete. Both tests would be satisfied."

I regard this as a clear statement of principle that the 'fear test', which embraces the first three requirements which I have set out in paragraph 10, is separate and distinct from the protection test, which is embraced in paragraph 10(5).

15. In support of his submission that the point at which consideration of the state's ability or willingness to afford protection becomes relevant is the third question, namely whether the fear is well-founded, Mr Plender relied upon the Canadian authority Canada (Attorney-General) v Ward [1997] INLR a decision of the Supreme Court of Canada. It is important to notice in that case that the Canadian statute which was there under consideration differs in a material respect from Article 1A(2) of the Convention, in that it is broken down into sub-paragraphs. S2(1) of the Canadian Immigration Act 1976 reads as follows:

" 'Convention refugee' means any person who, by reason of a well-founded fear of persecution for reasons of race, religion, nationality, membership in a particular social group or political opinion:

(a) is outside the country of his nationality and is unable or, by reason of such fear, is unwilling to avail himself of the protection of that country, or

(b) not having a country of nationality, is outside the country of his former habitual residence and is unable or, by reason of such fear, is unwilling to return to that country...."

16. After citing the statute, La Forest J who delivered the judgment of the Court said at p53:

"The section appears to focus the inquiry on whether there is a 'well-founded fear'. This is the first point the claimant must establish. All that follows must be 'by reason of' that fear. The first category requires the claimant to be outside the country of nationality by reason of that fear and unable to avail him- or herself of its protection. The second requires that the claimant be both outside the country of nationality and unwilling to avail him- or herself of its protection, by reason of that fear. Thus, regardless of the category under which the claimant falls, the focus is on establishing whether the fear is 'well-founded'. It is at this stage that the State's inability to protect should be considered. The test is in part objective; if a State is able to protect the claimant, then his or her fear is not, objectively speaking, well-founded. Beyond this point, I see nothing in the text that requires the State to be complicit in, or be the source of, the persecution in question."

17. But in Article 1A(2) 'Well founded fear' governs the reason why the applicant is outside the country in the fear test; it also governs 'is unwilling to avail himself of the protection of the country'; but it does not govern the phrase 'is unable.....to avail himself of the protection of his country'. This is an important distinction. Where the agents of persecution are state agents, once the applicant establishes that he has a well-founded fear of persecution, generally speaking that will be sufficient to establish that he is both unable and unwilling through fear of the persecution to avail himself of the protection of the state, because the state itself is the persecutor. An exception might be where the persecutors were a rogue element in the police acting in contravention of the generally accepted practice and standards.
18. But where the persecution is by non-state agents, the first question that must be asked is whether the applicant is unable to avail himself of the protection of the state. This has nothing to do with the fear of persecution. The convention is only concerned with surrogate protection where the applicant's own state has failed. If the state is able to provide sufficient protection, the claim for refugee status will generally fail, unless the applicant can show that he is unwilling to avail himself of that protection, not because he thinks it is inadequate, but because of fear of the persecution. In the case of non-state persecutors, it is fear of them. It may be, for example, that the persecutors have issued true and credible threats against him and his family if he seeks the protection of the state such that he could be said to be unwilling to do so.
19. In my judgment in the case of persecution by non-state agents the decision in Ward that the question of protection falls to be considered as part of the consideration of the well-foundedness of the fear is inconsistent with Lord Lloyd's analysis that a claimant may satisfy the 'fear test' but fail the 'protection test'. If Ward is right, there is effectively only one test.
20. Accordingly, in such cases as this, provided the claimant satisfies the first four tests, the fact finder should next ask if the claimant is unable to avail himself of the protection of the state. A claimant will be unable to avail himself of the protection of the state if the state is itself unable or unwilling to afford him protection. A state will as a rule only be unable to afford protection if law and order has broken down so that the state can no longer control the activities of criminals perpetrating the violence. I agree with what La Forest J said in Ward's case at p61:

"Security of nationals is, after all, the essence of sovereignty. Absent a situation of complete breakdown of State apparatus, such as that recognised in Lebanon in Zalzali, it should be assumed that the State is capable of protecting a claimant."

That is clearly not the case here. But the state may be unwilling to afford protection to a certain class of its citizens if there is widespread and systemic indifference to their plight on the part of the law enforcement agencies such as the police and the courts.

21. In the present case by agreement between those representing the Applicant and the Secretary of State the Tribunal directed itself in accordance with the test laid down in previous decisions of the Tribunal: Jaworski (17152); Debrah [1998] INRL 383 (17606); Chinder Singh (BILS 2E [701] (GOO55); Mojka (18265) and Dymiter (18467). The direction was in these terms:

"Is there in place in the country a sufficiency of protection? One needs to enquire into the various steps which have been taken by the country to see whether this protection is in place. If this sufficiency of protection is in place, then the need for international protection is not required. Thus it is not the test simply to ask whether the country "knowingly tolerates" persecutory acts by its agents or by sections of the community. Neither is it the test to ask whether the protection is effective...we believe that it is indeed the responsibility of the decision maker to ascertain whether the systems of domestic protection which are in place are sufficient from the perspective of international law.' (Debrah). As the Tribunal said in Chinder Singh, "In order to decide whether the protection is sufficient, it will of course be necessary to have some regard to its effectiveness. How much regard will depend upon the circumstances of the individual case." "

22. This is all right so far as it goes, but the problem about this test is that to tell the decision- maker that he must "ascertain whether systems of domestic protection which are in place are sufficient from the perspective of international law" is not particularly helpful. No state can guarantee the safety of its citizens. And to say that the protection must be effective suggests it must succeed in preventing attacks, which is something that cannot be achieved. Equally to say that the protection must be sufficient, begs the question, sufficient for what? In my judgment there must be in force in the country in question a criminal law which makes the violent attacks by the persecutors punishable by sentences commensurate with the gravity of the crimes. The victims as a class must not be exempt from the protection of the law. There must be a reasonable willingness by the law enforcement agencies, that is to say the police and courts to detect, prosecute and punish offenders. It must be remembered that inefficiency and incompetence is not the same as unwillingness, unless it is extreme and widespread. There may be many reasons why criminals are not brought to justice including lack of admissible evidence even where the best endeavours are made; they are not always convicted because of the high standard of proof required, and the desire to protect the rights of accused persons. Moreover, the existence of some policemen who are corrupt or sympathetic to the criminals, or some judges who are weak in the control of the court or in sentencing, does not mean that the state is unwilling to afford protection. It will require cogent evidence that the state which is able to afford protection is unwilling to do so, especially in the case of a democracy.
23. It can be argued that if a claimant can satisfy the protection test he also satisfies the well-founded fear test, provided of course the ill-treatment amounts to persecution. Because there is no protection from what he fears, his fear is well-founded. But the converse is not the case. Young men of African-Caribbean extraction may fear attack or mugging by skinheads in parts of the United Kingdom; those who have suffered repeated burglary may have a well-founded fear that they may be the victims again. It seems to me to be an unnecessary distortion of ordinary language to say that their fear is not well-founded because there is in existence in the United Kingdom an efficient police force willing to detect and prosecute crime and courts that can and do convict and sentence the criminals. Many such attacks and burglaries go undetected and unpunished. It would only be if the criminal justice system was so effective in detection and deterrence with the result that the risk of such attacks was minimal, that one could say that the fear was not well-founded. This clearly sets the protection test too high.
24. Before I come to the application of these principles to the present appeal I would wish to say a word about the standard of proof. A fact-finding tribunal has to make decisions on two different sorts of facts. The first relates to past and present events or states of affairs. In consideration of these facts the general rule in all civil proceedings is that they are established on a balance of probability. If the decision-maker decides that it is more probable than not that something has happened or represents the present state of things, then that is found to be a fact and is taken for the purpose of case as certain; it provides data for other findings including an appreciation of future events. Where, however, the question relates to what will happen in the future, it is not possible to apply the same reasoning; it cannot be said that if there is a 51% probability, there is a certainty that something will happen. There are only varying degrees of likelihood ranging from a near certainty, very likely, more likely than not, reasonably likely, a bare possibility to very unlikely. This differentiation is found in many aspects of the civil law. For example in the field of personal injuries, the question of whether a complainant will suffer disease or disability in the future is assessed on the degree of likelihood (see per Lord Diplock in Fernandez [1971] 1 WLR 987 at p993); in cases under the Fatal Accidents Acts the question of the likely future dependency on the deceased depends on the degree of chance or likelihood; it is not all if the prospects are more 50% or nothing if they are less (Davies v Taylor [1974] AC 207); in cases where causation of loss depends on what the action of a third party would have been if the defendant had not been negligent (Allied Maples Group Ltd v Simmons and Simmons [1995] 4 All ER 907 CA); the question of whether a child is at risk of abuse in the future (In Re H and others (Minors) (Sexual Abuse) [1996] AC 563). In line with these authorities the House of Lords in Sivakumaran, in deciding that an objective test had to be applied when considering whether the applicant's fear was well-found, held that the standard of proof was a reasonable likelihood that the risk of being persecuted on return to his country would materialise.
25. There has however been considerable debate whether the same standard, namely reasonable likelihood should be applied when the fact finder is considering historical or existing facts which will to a greater or lesser extent to inform the assessment as to what will occur in the future. In Jonah Nolan J, applying the reasoning of Lord Diplock in Fernandez, held that the assessment of those facts, namely what had happened in the past, should be found on a balance of probability; in so doing he followed the principles enunciated in the cases to which I have referred. However, in Kaja [1995] Imm AR 1 the majority of the Immigration Appeal Tribunal held that the lower standard of proof set out in Sivakumaran also applied to the assessment of accounts of past events. In the present case before the Tribunal Mr Tam, on behalf of the Secretary of State, argued that the decision of the majority in Kaja was wrong, being contrary to Jonah, which was binding on them. The Tribunal followed the majority decision in Kaja. They were impressed by the argument that Parliament, in enacting in 1996 the amendment to paragraph 5(5) of Schedule 2 to the 1993 Act and making the test of whether a refugee had been tortured in the past one of reasonable likelihood, must have approved the decision in Kaja.
26. This decision has not been challenged by the Secretary of State in this appeal and so we have not heard argument on it. It may be thought desirable that this Court should have an opportunity of considering the correctness of the decision of the majority in Kaja. Be that as it may, I see no reason to extend the Sivakumaran/Kaja standard of proof to the assessment of historical and existing facts when a decision maker is considering the protection test. Mr Plender, albeit by implication acknowledges this, hence his submission that the protection test should be embraced in the well-founded fear test.
27. I now turn to apply these principles to this appeal. The first question is whether the ill-treatment of which the Appellant has a subjective fear amounts to persecution. Here it is necessary to distinguish the violence from the skinheads from the alleged breach of third category rights relating to employment and education. There is no doubt that to constitute persecution a high threshold has to be crossed. This is evident in the definition adopted by Nolan J in Jonah which I have cited in paragraph 12. In Ravichandran v Secretary of State for the Home Department [1996] Imm AR 97 Staughton LJ said at p114 :

"..persecution must at least be persistent and serious ill-treatment without just cause."

28. So far as the violence by the skinheads is concerned there can be no doubt in my judgment that this amounted to persecution. There was no dispute that this was by reason of the Appellant's race. Equally it seems to me that the Tribunal in effect held that the fear was well-founded (see paragraph 36 of the decision, which I have referred to in paragraph 7 above).
29. The Appellant also contends that discrimination against Roma in the field of employment, education and the right to marry amount to persecution. The Appellant's case was that he had been unable to find work; that the whites in the village were mostly hostile and the teacher at his son's school didn't wish him to bother her about his son going there. He also said officials had kept putting off his marriage without apparent reason until the chairman of the registry office changed.
30. The Tribunal devoted a substantial part of their decision to the question of discrimination against Roma in respect of these matters. They adopted Professor Hathaway's four fold categorisation. He defined third category rights as follows:

"Third are those rights contained in the UDHR and carried forward in the International Covenant on Economic, Social and Cultural Rights. In contrast to the ICCPR, the ICESCR does not impose absolute and immediately binding standards of attainment, but rather requires states to take steps to the maximum of their available resources to progressively realise their rights in a non-discriminatory way. The basic values protected are the right to work, including just and favourable conditions of employment, remuneration, and rest, entitlement to food, clothing, housing, medical care, social security, and basic education; protection of the family, particularly children and mothers; and the freedom to engage and benefit from cultural, scientific, literary, and artistic expression. While the standard of protection is less absolute than that which applies to the first two categories of rights, a state is in breach of its basic obligations where it either ignores these interests notwithstanding the fiscal ability to respond, or where it excludes a minority of its population from their enjoyment. Moreover, the deprivation of certain of the socio-economic rights, such as the ability to earn a living, or the entitlement to food, shelter, or health care will at an extreme level be tantamount to the deprivation of life or cruel, inhuman or degrading treatment, and hence unquestionably constitute persecution."

31. Mr Tam had argued that breach of third category right could never amount to persecution. The Tribunal did not accept this and adopted the view of Professor Hathaway that in an extreme case it might do so. The Secretary of State challenged this in a Respondent's notice; but we have not found it necessary to consider the point. The Tribunal reviewed the evidence on this aspect of the case at considerable length. They considered the conclusion Mr McGeachy in Kloc and others (CC/59478097) in which he held that discrimination in the field of education, housing, health and employment amounted to persecution. They did not agree with Mr McGeachy. At paragraph 48 they said:

"More importantly, the various acts of discrimination identified in all the literature, and commented upon my Mr McGeachy, being third category rights, are not in our view sufficiently serious, even when treated cumulatively, as to amount to persecution. It is our view that the problems associated with unemployment amongst the Roma in Slovakia are due primarily to poor education and lack of professional qualifications making it difficult for them to compete successfully in the labour market."

32. They referred to a UNHCR report dealing with the problem of housing and commented that it gave a mixed picture. Their conclusion is expressed in paragraph 49 as follows:

"The manifestations of violence that have been documented are not, however, in our view sufficiently serious to entitle us to conclude that acts of discrimination and breaches of third category rights are such that the fear which the appellant has to return is that of persecution rather than that of discrimination. It is our view that while the appellant has a genuine and well-founded fear of discrimination, this does not entitle him to surrogate protection under the Geneva Convention as understood and applied in the United Kingdom."

33. It is a little unfortunate that the Tribunal considered the two issues of violence of the skinheads and discrimination in relation to third party rights in the same sentence. For reasons which I have already explained, I do not agree that the skinhead violence did not amount to persecution. But the conclusion that abuse of the third category rights did not amount to persecution was one which the Tribunal were fully entitled to reach.
34. Mr Plender also advanced an argument that some of the discrimination complained of amounted to breach of second category rights and he referred to Articles 2, 6, 12, 23, and 26 of the International Convention on Civil and Political Rights. This was not an argument that was advanced to the Tribunal, and in my judgment it adds nothing to the Appellant's case. The Tribunal looked at the conduct complained of and decided that it was not sufficient to meet the high threshold of persecution. It is not a question of wrongful categorisation as a matter of law.
35. In any event I am satisfied that Mr Howell's submissions on these alleged breaches are correct and I propose to deal with the points shortly.

(a) Article 6. There is no question that discrimination in the field of employment etc. amounts to an infringement of the right of protection of life by law.

(b) Article 12. It is not clear how the system of residence permits infringes the right to movement or freedom to choose his residence. In any event the system applies to all Slovaks.

(c) Article 23. We have not been shown any provision of Slovakian law that prohibits Romas from marrying or bringing up a family. The evidence of the Appellant shows obstruction by a certain official to his marriage until this official was replaced.

(d) Article 26 deals with equality of protection before the law. The law does not discriminate against Roma by excluding them from protection.

36. I come finally to the Tribunal's consideration of the protection test. In paragraph 58 they recited a number of reports which they had considered. In paragraph 59 they said:

"We can summarise this evidence by stating that it is our view that it shows quite conclusively that there is racial violence against the Roma perpetrated by "followers of the Skinhead movement". The police do not conduct proper investigation in all cases and there are instances where they have been very slow to conduct a full investigation. But stiff sentences are imposed at times for crimes which are racially motivated, for example the case of Mario Goral's death. There is increasing evidence of human rights activists being involved in Roma cases, and in Topolcany the perpetrators of a brutal attack by a gang of skinheads on Romani children were prosecuted. The risks of attacks by skinheads is real and we have cogent evidence that it has happened. Of course, even on the appellant's own story, such attacks have not happened to the appellant; and we conclude, on the basis of the evidence we have read that these violent attacks are on he whole isolated and random attacks by thugs.

The police are seen to be inefficient and ineffective by the Roma, and it is the case that they very rarely seek police protection. There is some evidence that the police have been slow to deal with complaints. On the other hand, there is also evidence that there have been prosecutions, and the police have intervened to provide protection when asked. Indeed in the instant case itself, the police did come to investigate."

37. In paragraphs 60 and 61 the Tribunal stated:

"We have looked at all this evidence, and have applied the principles which we have enunciated in this determination in deciding whether or not the appellant has demonstrated that he has a reasonable likelihood of a well-founded fear of persecution if he is returned to Slovakia. We have decided that he falls below the high threshold which we believe is required for international protection in a case where the fear is of an accumulation of discriminatory acts and where it is alleged that there is not a sufficiency of protection from non-state agents. It is our view that his fear is not that of persecution.

Our reasons are, in summary, as follows:

(a) The framework within which the appellant would be required to readapt if he were to return to Slovakia has to an extent now changed since the 1988 Election, and the subsequent departure of Meciar. It is of course too early to know what will happen, and there are bound to be difficult economic times ahead for the country. However, it is our view that, now that the extreme nationalist tendencies have been relegated to a much smaller section of the community than previously, it is likely that there will be much more emphasis on ties with the EU and other West European organisations. This is likely to enable international pressure to be placed on the Government to improve the human rights situation in the country for minority groups, especially the Roma. The situation therefore is today quite different from the situation in the country when Mr McGeachy was writing his determination in Kloc.

(b) But in any event and more importantly, our reading of the literature, although troublesome, does not bring is to the conclusion that the discrimination combined with the violence, together and accumulated, can amount to persecution. We believe that he literature shows that Slovak Roma as a group and as individuals have a fear, which is genuine and well-founded, of discrimination. It may be that for some Slovak Roma, who perhaps have been the victims of serious violence, the threshold would be crossed. But for the majority of Roma from Slovakia, and in particular for this appellant, we believe that the threshold has not been crossed, and he cannot claim the benefit of international protection under the terms of the Geneva Convention."

38. It is clear from these passages that the Tribunal considered the protection test as part and parcel of the question of persecution. In this, for the reasons I have given, their approach was in my view erroneous. Nevertheless it is quite plain that the Tribunal did ask the correct question in relation to the protection test and applied what had been agreed by the parties' representatives to be the test; and it was for that reason that they held the claim failed. In my judgment the fact that the Tribunal have applied this test at the wrong point in their consideration, does not vitiate their conclusion on the protection test. It seems to me that they concluded that Slovakia was not unable or unwilling to provide protection. Accordingly the Appellant could not show that he was unable to avail himself of their protection. This case so far as his unwillingness to avail himself of that protection was not fear of what the skinheads would do to him or his family, but because he considered that the authorities were indifferent or ineffective. The thrust of Mr Plender's argument is that it should have been considered as part of the well-foundedness of the fear; this, if correct, would have enabled him to argue that the lower standard of proof of reasonable likelihood should have been applied. But I have rejected this contention.
39. Mr Plender also sought to attack the Tribunal's decision on the ground of perversity and consideration of irrelevant material. Although Mr Plender drew attention to a number of passages in the various reports which were before the Tribunal, which included contrary opinions, I did not detect much enthusiasm for the submission that the Tribunal's decision was perverse. The question for the Tribunal was not an easy one; there can be little doubt that at times the standard of protection fell below what is desirable; but it is obvious that the Tribunal gave very careful consideration to the matter and I do not think the challenge on the ground of perversity can be made out.
40. The real thrust of Mr Plender's argument on this aspect of the case is that the Tribunal were wrong to take into account the matters referred to in paragraph 61(a); and furthermore that this was inconsistent with the views that they had previously expressed in paragraph 36 of the decision, namely:

"We have also read of course the material which has been submitted in relation to the results of the Slovak elections and the likely change of Government. It is our view that it is too early to draw any inferences from the election results, and it would be wrong of us to speculate. As of the latter part of 1998, therefore, we do not think that the plight of the Roma is markedly different to that which is documented in the UNHCR Guidelines of February 1998."

41. As the Tribunal pointed out when refusing leave, these two paragraphs are dealing with different things. Paragraph 36 is dealing with the existing position in December 1998; paragraph 61(a) is dealing with the likely position in the future; that was not an irrelevant consideration. But the opening words of paragraph 61(b) "but in any event and more importantly" shows that what follows is the real reason. Mr Plender has drawn our attention to some of the material referred to by the Tribunal in these two paragraphs and points out that Mr Meciar, who can be regarded as an arch-nationalist and antipathetic to the Roma, remains leader of the largest party in Parliament. But I am not persuaded that the Tribunal's assessment in paragraph 61(a) was wrong. The opposition to Mr Meciar had won 93 of the 150 seats in Parliament; their campaign had been conducted on the theme of reversing 'Meciarism' and getting the nation into the European Union and NATO saying that they will pass laws establishing fairness and meeting Western standards. The Tribunal's comments in paragraph 61(a) were designed to show that the picture presented to Mr McGeachy in Kloc had changed.
42. Two further discreet criticisms of the Tribunal's decision are made by Mr Plender. He submits that the Tribunal failed to give proper weight to the racial nature of the ill-treatment. At paragraph 46 the Tribunal said:

"We do not think that the East African Asian case, dealing as it does with the European Convention on Human Rights, can be relevant in the context of defining "persecution": discrimination for any reason is pernicious, and it is our opinion that the circumstances for the threshold being crossed so as to amount to persecution should have nothing to do with the reason for the discrimination."

43. In the East African Asians case [1981] 3 EHRR 76 at p86, the European Commission of Human Rights stated that, "as generally recognised, a special importance should be attached to discrimination based on race". The question in that case was whether discrimination against the applicants could amount to degrading treatment within Article 3 of the Convention. It was not a question of persecution. In my judgment the Tribunal were correct in what they said. This point does not in any event effect my conclusion that the treatment by the skinheads amounted to persecution. Nor does it elevate the alleged breach of third category rights into persecution, simply because the grounds are racial, rather than religious or political.
44. Mr Plender criticised the description in paragraph 59 of the Tribunal's decision (cited in paragraph 36 above) of the attacks by skinheads as "on the whole isolated and random attacks by thugs". Mr Plender submitted that this was inconsistent with the Appellant's evidence as to the frequency of attacks in his neighbourhood. I agree that the choice of language is not ideal. Nevertheless, I accept Mr Howell's submission that this statement is in the context of the ability of the state to protect the Roma. The Roma constitute a very sizeable minority, being about 10% of the population of 5.5 million people. In some cases where individuals are targeted by terrorists or dissidents it may be possible for the state to provide special police protection, for example by an armed guard or the provision of a new identity in a different part of the territory. But these attacks are unpredictable as to victims, time and location, and it is clear that the only form of protection which can be provided is in the form of deterrence through detection, prosecution, convictions and sentencing of criminals.
45. For these reasons I would dismiss the appeal.

LORD JUSTICE WARD:

Milan Horvath is a citizen of Slovakia. He arrived in the United Kingdom in October 1997 and claimed asylum. He is a gypsy. He said he feared persecution in Slovakia mainly from the brutal ill-treatment of the minority Roma community by the local skinheads, against whose malign activities the Slovak police failed to provide protection. He complained that his father had been beaten to death, though some years ago; that his brother had been attacked and beaten with weapons; that their home was ransacked and that they were in such a state of fear that they had to dig a hole in the ground in their back garden and regularly take shelter there at night. The police were either not interested or powerless to protect the gypsies. He also complained that there was discrimination against them in such matters as their freedom to marry and their full and free enjoyment of education. His statement, adjudged to be credible, concluded as follows:-

"If we are returned to Slovakia, I am afraid that I will be again persecuted by the skinheads because I am a Gypsy. I cannot get protection from the police who don't care at all about our problems. I want to stay in the UK where I live a peaceful life without fear."

He was refused asylum by the Secretary of State, by the Special Adjudicator and by the Immigration Appeal Tribunal, a legally qualified panel under the presidency of His Honour Judge David Pearl. The facts and the salient features of the judgment under appeal are set out in the judgment of Stuart-Smith L.J., which I have read in draft. The appeal is brought with the permission of Brooke L.J.

The issues

The appeal raises interesting and difficult questions of some importance to the operation of the Convention relating to the Status of Refugees, 1951. As amended, Article 1 gives this definition of the term "Refugee":

"A. for the purpose of the present Convention, the term "refugee" shall apply to any person who :

....

(2) owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country..."

The definition is in two parts. The first is that the applicant must presently be outside the country of his nationality owing to a well-founded fear of being persecuted for Convention reasons; and the second is that he must presently be unable or, owing to such fear, unwilling to avail himself of his country's protection. In Adan v Home Secretary [1999] 1 AC 293, Lord Lloyd of Berwick identified the need to satisfy two separate tests which, for short, he called "the fear test" and "the protection test." Since, as I read his speech, no test as such is set out, I apprehend him to be referring to a fear limb and protection limb within the definition, but for convenience I shall adopt his phraseology. On closer analysis one can identify six aspects of the definition:-

1. being out of the country of nationality because he is in fear,

2. the fear being of persecution,

3. that fear being well-founded,

4. the persecution being for a Convention reason,

5. being unable to avail himself of his country's protection or, as an the alternative to this,

6. being unwilling owing to such fear so to avail himself.

Whether one should look at these aspects discretely or whether the approach is holistic is one of the matters, which divides Stuart-Smith L.J., and me. On the facts of this case as presented to us, the main issues as I see them are:-

1. Is the nature and extent of State protection relevant-

(i) to the definition of persecution when the ill-treatment is meted out by so-called "non-State agents", or

(ii) to the consideration of whether the fear of persecution is well-founded, or

(iii) only to the question whether the asylum-seeker is unable or unwilling to avail of the protection of his country of origin, or

(iv) in one way or another to all of these matters?

2. What level of discriminatory denial of human rights amounts to persecution?

3. What level of State protection is required?

4. To what standard and how must persecution be established?

Some introductory observations.

Since we are engaged in gleaning the meaning of the definition, I should remind myself that:-

"One is more likely to arrive at the true construction of Article 1 (2) by seeking a meaning which makes sense in the light of the Convention as a whole, and the purpose which the framers of the Convention were seeking to achieve, rather than by concentrating exclusively on the language. A broad approach is what is needed, rather than a narrow linguistic approach:"

per Lord Lloyd in Adan at p. 305 E. That of course accords with the general rule of interpretation of treaty provisions which appears in Article 31 of the Vienna Convention on the Law of Treaties (Vienna, 23 May 1969), paragraph 1 of which provides:-

"A treaty shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in the light of its object and purpose."

The preambles provide an insight into the purpose of the Convention. By the first, "the General Assembly have affirmed the principle that human beings shall enjoy fundamental rights and freedoms without discrimination", and the fourth acknowledges that "the grant of asylum may place unduly heavy burdens on certain countries." As Lord Steyn observed in Islam v Secretary of State for the Home Department [1999] 2 WLR 1015, 1021G:-

"The relevance of the preambles is two-fold. First they expressly show that a premise of the Convention was that all human beings shall enjoy fundamental rights and freedom. Secondly, and more pertinently, they show that counteracting discrimination, which is referred to in the first preamble, was a fundamental purpose of the Convention."

In Reg. v Home Secretary, ex parte Sivakumaran [1988] 1 A.C. 958, 992, Lord Keith of Kinkel reflected that:-

"... the general purpose of the Convention is surely to afford protection and fair treatment to those for whom neither is available in their own country, and does not extend to the allaying of fears not objectively justified, however reasonable these fears may appear from the point of view of the individual in question."

La Forest J. in Canada (Attorney-General) v Ward [1997] INLR 42, 51g, giving the judgment of the Supreme Court of Canada, said:-

"At the outset, it is useful to explore the rationale underlying the international refugee protection regime, for this permeates the interpretation of the various terms requiring examination. International refugee law was formulated to serve as a back-up to the protection one expects from the State of which an individual is a national. It was meant to come into play only in situations when that protection is unavailable, and then only in certain situations. The international community intended that persecuted individuals be required to approach their home State for protection before the responsibility of other States became engaged. For this reason, James Hathaway refers to the refugees scheme as "surrogate or substitute protection", activated only upon failure of national protection: see The Law of Refugee Status [1991] at p. 135."

Then the broad purpose, which the Convention seeks to achieve, is that the international community will afford sanctuary to the refugee who seeks safety and freedom elsewhere if his core fundamental human rights have not enjoyed respect and protection from his home State.

Persecution.

There is surprisingly little clear authority on the meaning of persecution. As the United Nations High Commissioner for Refugees' Handbook on Procedures and Criteria for Determining Refugee Status observed in paragraph 51:-

"There is no universally accepted definition of "persecution", and various attempts to formulate such a definition have met with little success."

Despite that discouraging note, it is necessary to understand what is encompassed by the notion of persecution. Nolan J. attempted a definition in R v Immigration Appeal Tribunal ex parte Jonah [1985] Imm. A.R. 7, 13, where he said :-

"The dictionary definition, according to the Shorter Oxford English dictionary, has under "persecute" the meanings: "To pursue, hunt, drive" firstly, and secondly: "to pursue with malignancy or injurious action; esp. to oppress for holding a heretical opinion or belief". ... to my mind the proper approach must be to apply to the word "persecution" its ordinary meaning as found in the dictionary."

In Sandralingham and Ravichandran v Secretary of State for the Home Department [1996] Imm A R 97 Staughton L.J. said :-

"Persecution must at least be persistent and serious ill-treatment without just cause by the State, or from which the State can provide protection but chooses not to do so."

Mr Plender Q.C. observes, and the observation is correct, that the court was not there dealing with persecution by non-State agents and so he urges caution in applying that definition to the facts of this case.

In Adan, where the issue was whether it was necessary to show a current well-founded fear of persecution or whether an historic fear was sufficient, Lord Lloyd identified at p. 304 four categories of refugee the first two of which were:-

"(1) nationals who are outside their country owing to a well-founded fear of persecution for a Convention reason, and are unable to avail themselves of the protection of their country;

(2) nationals who are outside their country owing to a well-founded fear of persecution for a Convention reason, and, owing to such fear, are unwilling to avail themselves of the protection of their country."

At p. 306A, he observed that category (1) is not confined to refugees who were subject to State persecution because, ex hypothesi, they would be unable to avail themselves of State protection. He said:-

"But category (1) is not so confined. It also includes the important class of those who are sometimes called "third party refugees," i.e. those who are subject to persecution by factions within the State. If the State in question can make protection available to such persons, there is no reason why they should qualify for refugee status. They would have satisfied the fear test, but not the protection test. Why should another country offer asylum to such persons when they can avail themselves of the protection of their own country? But if, for whatever reason, the State in question is unable to afford protection against factions within the State, then the qualifications for refugee status are complete. But both tests would be satisfied."

In Reg. v I.A.T., ex parte Shah [1999] 2 WLR 1015, the only dispute related to the question whether the acknowledged persecution by the State was by reason of the applicants' membership of a particular social group. When Lord Hoffmann dealt with the question of causation he said:-

"What is the reason for the persecution which the appellants fear? Here it is important to notice that it is made up of two elements. First, there is the threat of violence to Mrs Islam by her husband and his political friends and to Mrs Shah by her husband. This is a personal affair, directed against them as individuals. Secondly, there is the inability or unwillingness of the State to do anything to protect them. There is nothing personal about this. The evidence was that the State would not assist them because they were women. It denied them a protection against violence which it would have given to men. These two elements have to been combined to constitute persecution within the meaning of the Convention. As the Gender Guidelines for the Determination of Asylum Claims in the UK (published by the Refugee Women's Legal Group in July 1998) succinctly puts it (at p. 5): "Persecution = Serious Harm + The Failure of State Protection." "

I am grateful to Hale L.J. for having drawn my attention to Reg. v Secretary of State for the Home Department, ex parte Adan [1999] INLR 362. Laws L.J. noted that the difference in approach between the subscribers to the "accountability" theory and those, including the United Kingdom, who apply the "protection" theory. At p. 377 he said :-

"Our courts recognise persecution by non-State agents for the purposes of the Convention in any case where the State is unwilling or unable to provide protection against it, and indeed whether or not there exist competent or effective governmental or State authorities in the country in question. This is what has been called the "protection" theory. It is, as we have said, shared by a majority of the States signatory to the Convention and the UNHCR."

The search for the true meaning of "persecution" can be extended. Dealing with agents of persecution, the UNHCR Handbook states in paragraph 65:-

"Where serious discriminatory or other offensive acts are committed by the local populous, they can be considered as persecution if they are knowingly tolerated by the authorities, or if the authorities refuse, or prove unable, to offer effective protection."

The Joint Position of the European Council (O.J. 1996 L 63/2) deals with persecution by third parties in paragraph 5.2 and provides:-

"Persecution by third parties will be considered to fall within the scope of the Geneva Convention where it is based on one of the grounds in Article 1A of that Convention, is individual in nature and is encouraged or permitted by the authorities."

Professor Hathaway is an acknowledged expert in the field and in his seminal work, The Law of Refugee Status, he makes these observations:-

"The intention of the drafters was not to protect persons against any and all forms of even serious harm, but was rather to restrict refugee recognition to situations in which there was a risk of a type of injury that would be inconsistent with the basic duty of protection owed by the State to its own population. As a holistic reading of the refugee definition demonstrates, the drafters were not concerned to respond to certain forms of harm per se, but were rather motivated to intervene only were the maltreatment anticipated was demonstrative of a break-down of national protection. The existence of past or anticipated suffering alone, therefore, does not make one a refugee, unless the State has failed in relation to some duty to defend its citizenry against the particular form of harm anticipated." (p. 103/4.)

...

Drawing on these basic precepts, persecution maybe defined as the sustained and systemic violation of basic human rights demonstrative of a failure of State protection. A well-founded fear of persecution exists when one reasonably anticipates that remaining in the country may result in a form of serious harm which the government cannot or will not prevent, ...(p. 104/5.)

...Not all persons who have left their country because of risk to basic human rights are refugees. As noted in the introduction to this chapter, refugee law is designed to interpose the protection of the international community only in situations where there is no reasonable expectation that adequate national protection of core human rights will be forthcoming. Refugee law is therefore "substitute protection" in the sense that it is a response to disfranchisement from the usual benefits of nationality. As Guy Goodwin-Gill puts it, "...the lack of protection normally to be expected of the government is either lacking or denied".

This means that in addition to identifying the human rights potentially at risk in the country of origin, a decision on whether or not an individual faces a risk "persecution" must also comprehend scrutiny of the State's ability and willingness effectively to respond to that risk. Insofar as it is established that meaningful national protection is available to the claimant, a fear of persecution cannot be said to exist. This rule derives from the primary status accorded to the municipal relationship between an individual and her State, and the principle that international human rights law is appropriately invoked only when a State will not or cannot comply with its classical duty to defend the interests of citizenry." (p. 124/5.)

In his book, The Refugee in International Law, Goodwin-Gill writes at p. 67:-

"Fear of persecution and lack of protection are themselves inter-related elements as Article 1(1) of the 1951 Convention makes clear. The persecuted clearly do not enjoy the protection of their country of origin, while evidence of the lack of protection on either the internal or external level may create a presumption as to the likelihood of persecution and to the well-foundedness of any fear."

He has also expressed the view that persecution comprehends "failure (voluntarily or involuntarily) on the part of State authorities to prevent or suppress (private) violence": see (1980), Mitchigan Y.B. Intl. L. Studies 291, 298.

It is obviously desirable that the approach to the interpretation of the Convention should, so far as is possible, be the same in all countries which are signatories, even though this may be an unrealistic expectation in the absence of some supranational court which is capable of giving authoritative interpretations to the provisions which are binding on all signatory countries: see Lord Woolf M.R. in R v A Special Adjudicator ex parte Kerrouche [1998] INLR 88, 92. It is, therefore, instructive to see the approach taken in other jurisdictions.

I look first to the Australian authorities. In Chan Yee Kin v Minister for Immigration and Ethnic Affairs [1989] 87 ALR 412 McHugh J., sitting in the High Court of Australia, said at p. 449:-

"The notion of persecution involves selective harassment. ... As long as the person is threatened with harm and that harm can be seen as part of a course of systematic conduct directed for a Convention reason against that person as an individual or as a member of a class, she is "being persecuted" for the purposes of the Convention. The threat need not be the product of any policy of the government of the person's country of nationality. It may be enough, depending on the circumstances, that the government has failed or is unable to protect the person in question from persecution."

In A v Minister for Immigration [1998] INLR 1 the High Court of Australia was concerned mainly with whether or not the asylum seekers were members of a particular social group and the court was divided on that. Nevertheless some common views were expressed about persecution. Brennan C.J. said at p. 7:-

"Thus the definition of "refugee" must be speaking of a fear of persecution that is official, or officially tolerated or uncontrollable by the authorities of the country of the refugee's nationality (Canada (Attorney-General) v Ward [1997] INLR 42, 58-59)."

McHugh J said at p. 26:-

"The Convention is primarily concerned to protect those racial, religious, national, political and social groups who are singled out and persecuted by or with the tacit acceptance of the government of the country from which they have fled or to which they are unwilling to return. Persecution by private individuals or groups does not by itself fall within the definition of refugee unless the State either encourages or is or appears to be powerless to prevent that private persecution."

Canada (Attorney-General) v Ward [1997] INLR 42 is a judgment of the Supreme Court of Canada. One should not overlook, when reading this authority, that the Canadian statute is structured differently from the Convention in that in Canada the well founded fear must not only be the reason for the claimant's unwillingness to avail himself of the State's protection, but, and this is the difference, it must also be the reason for his inability to do so. That may explain why La Forest J. seems to emphasise at p. 53:-

"The section appears to focus the enquiry on whether there is a 'well-founded fear'. This must be the first point the claimant must establish. All that follows must be by reason of that fear. The first category requires the claimant to be out of the country of nationality by reason of that fear and unable to avail him- or herself of its protection. The second requires that the claimant be both outside the country of nationality and unwilling to avail him- or herself of its protection, by reason of that fear."

There was substantial agreement between the parties that the State's inability to protect was an integral component of the notion of a Convention refugee, but they differed as to the point in the analysis at which such component was to be injected into the definition. La Forest J. continuing the above citation said:-

"Thus, regardless of the category under which the claimant falls, the focus is on establishing whether the fear is 'well-founded'. It is at this stage that the State's inability to protect should be considered. The test is in part objective; if a State is able to protect the claimant, then his or her fear is not, objectively speaking, well-founded. Beyond this point, I see nothing in the text that requires the state to be complicit in, or be the source of, the persecution in question."

When the court was referring to "State complicity", I understand the phrase to connote the notion that the ill-treatment must, as it was expressed at p. 52 B, "emanate from the State" through encouragement or acquiescence. Whilst this may have overtones of the "accountability" theory to which Laws L.J. referred in R v S.S.H.D., ex parte Adan [1999] INLR 362, 376, I note that La Forest J. also cited paragraph 65 of the UNHCR Handbook with apparent approval, saying:-

"The position reflected in the UNHCR Handbook, therefore, is that acts by private citizens, when combined with State inability to protect, constitute 'persecution'." (p. 54.)

That led him to 'conclude' that:-

"persecution under the Convention includes situations where the State is not in strictness an accomplice to the persecution, but is simply unable to protect its citizens," (p. 56.)

In other words, the Canadian court there seems to be ascribing to the "protection" theory.

This summary is also worthy of note for the concluding clause which I have highlighted:-

"In summary, I find the State complicity is not a necessary component of persecution either under the "unwilling" or under the "unable" branch of the definition. A subjective fear of persecution combined with State inability to protect the claimant creates a presumption that the fear is well-founded. The danger that this presumption will operate too broadly is tempered by the requirement that clear and convincing proof of a State's inability to protect must be advanced," (p. 61.)

Finally, I note that the Conseil d' Etat in the matter of Dankha, IJRL/0195 accepted, as I very loosely translate the short summary of the decision available to me, that persecutions carried out by particular individuals, whether organised or not, can be sustained only so long as they are in fact encouraged or voluntarily tolerated by, the public authorities to the extent that the interested party has no effective means of calling upon the protection of the State. Although this may be an example of the accountability theory, the significance for present purposes is at least to show that the role of the State is relevant to the element of persecution.

I also believe that the Court of Appeal of the Ninth Circuit in the United States of America have interpreted the likelihood of persecution to include persecution by the government or by a group which the government is unable to control: see MacMullen v Immigration and Naturalisation Service 658F 2d 1312 (9 Cir 1981) 1315, as reported in Ward at p.55.

A well-founded fear.

The law is settled by Reg. v Home Secretary, ex parte Sivakumaran [1988] 1 A.C. 958. In dealing with what has to be proved, Lord Keith of Kinkel said at p. 993:-

"The question is what might happen if he were to return to the country of his nationality. He fears that he might be persecuted there. Whether that might happen can only be determined by examining the actual state of affairs in that country. If that examination shows that the persecution might indeed take place then the fear is well founded. Otherwise it is not."

Lord Templeman explained at p. 996:-

"My Lords, in order for a "fear" of "persecution" to be "well-founded" there must exist a danger that if the claimant for refugee status is returned to his country of origin he will meet with persecution. ... Danger from persecution is obviously a matter of degree and judgment."

On the question of the standard to which that had to be established, Lord Keith's opinion was expressed at p. 994 as follows:-

"In my opinion the requirement that an applicant's fear of persecution should be well-founded means that there has to be demonstrated a reasonable degree of likelihood that he will be persecuted for a Convention reason if returned to his own country. In Reg. v Governor of Pentonville Prison, ex parte Fernandez [1971] 1 WLR 987, this house had to construe section 4(1)(c) of the Fugitive Offenders Act 1967, which requires that a person shall not be returned under the Act if it appears

"that he might, if returned, be prejudiced at his trial or punished, detained or restricted in his personal liberty by reason of his race, religion, nationality or political opinion."

Lord Diplock said, at p. 994:

"My Lords, bearing in mind the relative gravity of the consequences of the court's expectation being falsified in one way or in the other, I do not think that the test of the applicability of paragraph (c) is that the court must be satisfied that it is more likely than not that a fugitive will be detained or restricted if he is returned. A lesser degree of likelihood is, in my view, sufficient; and I would not quarrel with the way in which the test was stated by the Magistrate or with the alternative way in which it was expressed by the Divisional court. "A reasonable chance," "substantial grounds for thinking," "a serious possibility" - I see no significance difference between these various ways of describing the degree of likelihood of the detention or restriction of the fugitive on his return which justifies the court in giving effect to the provisions of section 4(1)(c)."

I consider that this passage appropriately expresses the degree of likelihood to be satisfied in order that a fear of persecution may be well-founded."

In Adan Lord Lloyd dealt with his second category of refugee which includes a person who (a) is outside the country of his nationality owing to a well-founded fear of persecution and (b) is unwilling, owing to such fear, to avail himself of the protection of that country. He said at p. 305:-

" "Owing to such fear" in (b) means owing to a well-founded fear of being persecuted for a Convention reason. But "fear" in (b) can only refer to current fear, since the fear must be the cause of the asylum-seeker being unwilling now to avail himself of the protection of his country. If fear in (b) is confined to current fear, it would be odd if "owing to well-founded fear" in (a) were not also confined to current fear. The word must surely bear the same meaning in both halves of the sentence."

On the question of whether or not State protection is a material factor in the consideration of the well-foundedness of the fear, I have already referred to the Canadian court's approach in Ward. That has been followed by the Federal Court of Appeal in Australia in Minister for Immigration and Multicultural Affairs v Prathapan [1998] 156 ALR 672, 677 where Lindgren J. said:-

"According to the terms of the definition, the person's well-founded fear of persecution for a Convention reason has two roles to play: it must be the cause of the person's being outside his or her country of nationality and it must be the cause of any unwillingness on the part of that person to resort to that country's protection. If protection is available from the country of nationality, fear of persecution is not well-founded. In those circumstances, the person would be unwilling to take advantage of the protection of the country of nationality "owing to" some other cause than a well-founded fear of persecution. It will be necessary, however, in due course, to consider the meaning "protection" in the present context. Unlike "unwillingness", "inability" to resort to the protection of the country of nationality is not, by the terms of the definition, required to arise from a well-founded fear of persecution. …

But in effect the two categories ("unable" and "unwilling owing to a well-founded fear of persecution") are directed to broadly similar sets of circumstances. They are situations in which the person cannot be blamed for not following the course, ordinarily to be expected, of relying on the country of nationality for protection."

Conclusions on the first issues.

46. The question posed by issue 1(i) is whether State protection is relevant to the definition of persecution where so-called non-State agents mete out the ill-treatment. Relating the question to the facts of this case, the issue is whether the persecution which Mr Horvath must establish is constituted simply by reference to the way the skinheads brutalised the Roma community.
47. The consensus of view seems clearly to point to the essence of persecution being not so much the ill-treatment by the non-State agent but rather the lack of appropriate State protection against such harm because it is only by reason of the State's failure that an international obligation is imposed upon another State, as surrogate, to provide a safe harbour to the refugee. Thus, although the observation was obiter, I accept Lord Hoffmann's favourable endorsement of the equation "Persecution = Serious Harm + The Failure of State Protection." In many cases the sorry fact is that the State is actively involved in the invasion of its citizen's fundamental human rights and in the denigration of his dignity. So too where the ill-treatment is meted out by non-State agents, it is not persecution within the meaning of the Convention unless the ill-treatment of the individual is systemic, i.e. part of the prevailing political or social order under which that ill-treatment is encouraged by or acquiesced in, or colluded in or connived at, by the State, or beyond the control of the State, and, of course, that it is for a Convention reason.
48. The next question is whether or not the lack of State protection is also a relevant factor in the consideration of whether the fear is well-founded. I can see the force of the argument which has found favour in Canada and Australia that the failure of State protection enters the analysis at this point. To approach the matter as Lord Templeman did in Sivakumaran (and Lord Bridge and Lord Griffiths agreed with his approach) then the question is whether there must exist a danger that if the claimant for refugee status is returned to his country of origin he will meet with persecution. There will be no such danger if the State is willing and able to protect him from the acts of ill-treatment he fears. In that situation, however haunted by fear he may be, his fear is unfounded because, viewed objectively, the State apparatus works sufficiently well for the court to conclude that it is reasonably likely to afford him protection. But what if his safety is not fully assured? The difficulty with looking at State protection only in the context of the well-foundedness of the fear is, as it seems to me, that the Convention does not operate so as to guarantee safety. It is, as Hathaway states at p. 105, axiomatic that we live in a highly imperfect world. Certain levels of ill-treatment may still occur even if the State has taken steps to protect the oppressed. We only have to look to our own community to see that even in our free society some will walk in fear, and suffer not withstanding the determination of the authorities that they should not do so. They would not become refugees because, as La Forest J. observed in Ward at p. 61:-

"Refugee claims were never meant to allow a claimant to seek out better protection than that from which he or she benefits already."

Returning to the facts of this case, and assuming for the moment that persecution is constituted by the actions of the skin heads, then, as the Immigration Appeal Tribunal found, the appellant's fear of persecution was well-founded. Their finding was:-

"24. ... We accept that what the appellant has said with regard to the general situation in respect of Roma in Slovakia is correct. We believe that his subjective fears of what may await him if he is to return to Slovakia could well be genuine and, subject to any changes in Slovakia subsequent to the recent elections, we believe that they are well-founded."

Their caveat about changes in Slovakia is important when reviewing the judgment as a whole and paragraph 61(a) in particular. Their conclusion, as I read it, is that international pressure is likely to lead to improvements in the human rights situation for the Roma and "the situation therefore is today quite different". The next sentence is also significant for an overall understanding of the judgment, namely:-

"25. Whether they are fears of persecution or of something less than persecution is of course a fundamental question to which we shall return later in this determination. ..."

When they did return to the distinction, they found at paragraph 60 that his fear was not that of persecution.

49. If, contrary to this conclusion, the argument is correct that the relevant fear of persecution is of persecution by the skin heads per se without regard to any part played by the State in protecting them or failing to protect them, then it is that self same fear which must be considered when applying Lord Lloyd's "protection test". As he said the words must have the same meaning in both limbs of the definition. On the so-called protection test one must ask whether Mr Horvath was unwilling, owing to that same well-founded fear of persecution by the gypsies, to avail himself of the protection of Slovakia. He was found to be credible and so we are bound to accept what he said, namely:-

"If we are returned to Slovakia, I am afraid that I will be again persecuted by the skin heads because I am a gypsy. I cannot get protection from the police who don't care at all about our problems."

I would conclude that he was unwilling to avail himself of police protection because it would not prevent his again being persecuted by the skin heads. On that analysis, his appeal should succeed if the underlying assumption that ill-treatment by the gypsies can amount to persecution.

50. The third issue I posed was whether State protection was a relevant factor in the so-called protection test. Of course it is. One cannot consider whether the applicant is unable to avail himself of the protection of his country or whether he is, owing to his well-founded fear of persecution unwilling so to avail himself without considering what protection is available in his country.
51. I turn to Stuart-Smith L.J.'s analysis. As I understand my Lord, it is his judgment that the question of State protection enters the analysis at the protection test, and only at this point. If no protection is available, then ex hypothesi, the applicant is unable to avail of it. I agree. But if some protection is available, the Tribunal must establish what it amounts to. Again I agree. Then the question is a question of causation. Why is the applicant unwilling to avail himself of the State's protection? I am still in agreement. My Lord finds that the appellant was unwilling to avail himself of Slovakia's protection not for fear of what the skin heads would do to him and his family but because he considered that the authorities where indifferent or ineffective. I am very conscious of my Lord's much greater experience in the exercise of this jurisprudence, but I regret that I now part company from him. As a matter both of understanding what the appellant said, and as a matter of harsh reality, the reason for this appellant's unwillingness to avail himself of that protection is that he fears the continuation of ill-treatment by the skin heads given the indifference and ineffectiveness of the authorities to protect him from it. That, in my judgment, is an unwillingness to avail himself of the State's protection because of a well-founded fear of persecution. What my Lord's analysis seems to me to involve is reading into the protection test a requirement that the asylum-seeker should be unwilling to avail himself of - in the sense of unwilling to ask for or actively to seek out - police (or other State) protection either because, as set out in paragraph 18 of my Lord's judgment, the persecutors have issued true and credible threats against him and his family if he were to seek the protection of the State, or, I suppose, because if he did seek protection they would exact revenge. The difference between us seems to turn upon the proper meaning to be given to the words "avail himself of". I do not understand them to have the limited meaning my Lord would seem to ascribe to them for the following reasons:-

(a) They are wide words. The Oxford English Dictionary defines "avail of" as "a. to benefit oneself or profit by; to take advantage of, turn to account. b. to make use of, use, employ." Accordingly I understand the expression to mean more than turning to the authorities for special protection but rather merely having to make do with the protection that is actually on offer. A narrow application would only seem to be appropriate where an individual is being singled out for ill-treatment for special reason, e.g. the sort of reprisal which Ward feared from the INLA.

(b) I do not see how my Lord's test works in practice. The fact that the asylum-seeker had not been singled out and threatened with reprisals if he were to seek protection, does not make his fear of ill-treatment less well-founded. If he is to be put on the next aeroplane and returned home, how and when is he to seek protection? At the airport on arrival, or at his home village? This will hardly avail him if he is pulled off the bus and beaten before he gets home. And why should the relevant fear be of reprisal? The skinheads who beat him up may not even know that he has asked for special protection. To them he is just another gypsy in need of a beating. As La Forest J. more eloquently expressed it in Ward at p. 60:-

"Moreover, it would seem to defeat the purpose of international protection if a claimant would be required to risk his or her life seeking ineffective protection of a State, merely to demonstrate that ineffectiveness."

52. To summarise my conclusions: the degree of protection is a necessary ingredient in the definition of persecution. Ill-treatment by the skinheads alone is not persecution. The degree of State protection may indirectly be a factor in judging whether the fear is well-founded but this is not the only or best place for it in a proper analysis of the definition of refugee. What State protection is available is a fact to be considered for the protection test, but the question of State protection is not confined to the "protection test".
53. In the result, and in answer to issue 1(iv) posed by me, I adopt a holistic approach to the definition. This accords with the view of Simon Brown L.J. in Sandralingham and Ravichandram v Secretary of State for the Home Department [1996] Imm A R 97, 109, where he said:-

"In my judgment, the issue whether a person or group of people have a "well-founded fear (i.e. a real risk - see ex parte Sivakumaran) of being persecuted for (Convention) reasons" - ... raises a single composite question. It is, as it seems to me, unhelpful and potentially misleading to try to reach separate conclusions as to whether certain conduct amounts to persecution, and as to what reasons underlie it. Rather the question whether someone is at risk of persecution for a Convention reason should be looked at in the round and all the relevant circumstances brought into account. I know of no authority inconsistent with such an approach and, to my mind, it clearly accords both with paragraph 51 of the UNHCR handbook and with the spirit of the Convention."

54. It may be said that my conclusions cannot stand in the light of Adan where Lord Lloyd said that if the State can make protection available to those who are subject to persecution by factions within the State, there would be no reason why they should qualify for refugee status. "They would have satisfied the fear test, but not the protection test." Yet if I am right, then in the situation postulated by Lord Lloyd, the third party refugees would not have satisfied the fear test both because the acts of the faction would not by themselves have amounted to persecution and also because if there was that level of State protection available, then the fear could hardly have been well-founded. I feel free to depart from this sentence in the speech because it does not seem to me to be part of the ratio decidendi. The place of State protection in the definition was not in any way argued before their Lordships who were concerned with the question whether or not the fear had to be a current fear. They were not concerned with the place of State protection in the analysis: Moreover on the facts it was accepted that there was no current fear of persecution. Whilst respectful of what was said, I do not regard myself as bound by it.

10. Applying my conclusions to the judgment of the Immigration Appeal Tribunal I find they did not misdirect themselves but on the contrary were correct when they held as follows:-

"53. Acts by private citizens when combined with State inability to protect, may constitute "persecution". But it is the failure of the State to provide protection which in our view converts the discriminatory acts into persecution ...

55. ... it is evidence of State failure to protect which in out opinion transforms discriminatory acts into persecutory ones.

60. It is our view that his fear" (I interpolate - his fear of the gypsies' ill-treatment) "is not that of persecution.

61(b). ... our reading of the literature, although troublesome, does not bring us to the conclusion that the discrimination combined with the violence, together and accumulated, can amount to persecution."

The appeal on this ground fails accordingly.

When does a denial of Human Rights amount to persecution?

As Lord Templeman said of the well-foundedness of the fear, this is a matter of degree and of judgment. This is demonstrated by Shah. Lord Hope of Craighead said at p. 1083:-

"But persecution is not the same thing as discrimination. Discrimination involves the making of unfair or unjust distinctions to the disadvantage of one group or class of people as compared with others. It may lead to persecution or it may not. And the persons may be persecuted who have not been discriminated against. If so, they are simply persons who are being persecuted."

Lord Millet said at p. 1041:-

"The denial of human rights, however, is not the same as persecution, which involves the infliction of serious harm. The 1951 Convention was concerned to afford refuge to the victims of certain kinds of discriminatory persecution, but it was directed to prohibit discrimination as such nor to grant refuge to the victims of discrimination. Moreover, while the delegates in Geneva were willing to extend refugee status to the victims of discriminatory persecution, they were unwilling to define the grounds of persecution which would qualify for refugee status as widely as the discriminatory denial of human rights condemned by the Universal Declaration. Discriminatory persecution "of any kind" would not suffice; the Convention grounds are defining, not merely illustrative as in the Universal Declaration."

In an attempt to classify the gravity of the breaches of the human rights, Hathaway proposed the helpful division into four categories. That classification is useful but not definitive and it should be applied with care given to an evaluation of the facts of the case. Although we did not hear full argument on the point, my preliminary view is that breach of third category rights cannot be said as a matter of law to amount to persecution just as it cannot be said as a matter of law that breach of these rights could never amount to persecution. It is a matter of fact and degree and judgment in the individual case. The Tribunal did not err in law in their approach to this part of their consideration and their assessment of the lack of gravity and seriousness of the breaches was a matter well within their competence and their conclusions are unappealable. I agree with Stuart-Smith L.J. that the appeal relating to the breach of human rights should be dismissed for the reasons he gives.

What level of State protection is required?

The Tribunal asked themselves: "Is there in place in the country a sufficiency of protection?" They took that as their test because that was the common ground accepted by both counsel and they cannot be criticised for adopting it. Stuart-Smith L.J. has already set out their direction in paragraph 21 of his judgment and I need not repeat it. I agree with my Lord's conclusions expressed in paragraph 22 of his judgment. What is required is a discernible system of criminal justice. State protection can be sufficient even if it is not fully effective because safety can never be guaranteed. Yet again it is a matter of fact and degree. The more serious the ill-treatment the higher the level of protection that would come to be expected of the State and conversely, of course, the less serious the degree of physical harm, the less rigorous is the need for protection. A criminal justice system must be in place but the receiving State is not obliged to give succour to every asylum-seeker simply because that system may fail to prevent harm befalling him: it will be sufficient if the system in place is one which ordinarily would deter the wrongdoer. At the fulcrum of the balance which has to be struck is the sharp point of principle made by Professor Hathaway at p. 108:-

"The dominant view, however, is that refugee law ought to concern itself with actions which deny human dignity in any key way, and that the sustained or systemic denial of core human rights is the appropriate standard."

To what standard and how must persecution be established?

Persecution is such an affront to human dignity that international law places the obligation on contracting States to give refuge to the persecuted. That demand undoubtedly creates some tension. On the one hand the humanitarian aim is to save the refugee from that persecution: we are required to be good Samaritans. There is always a need for "anxious scrutiny" in asylum cases: see Reg. v Secretary of State for the Home Department, ex parte Bugdaycay [1987] AC 514, 531. On the other hand, since persecution involves direct State activity or non-State activity which the State is unwilling or unable to prevent, then, if only as a matter of international comity, one State will not lightly find that another has been directly or indirectly implicated in acts of persecution. As La Forest J. said in Ward at p. 61 :-

"Security of nationals is, after all, the essence of Sovereignty. ... it should be assumed that the State is capable of protecting a claimant.

...

Clear and convincing proof of a State's inability to protect must be advanced."

Thus I agree with Nolan J in Jonah, and find it timely in a case like this to emphasise, that:-

"...considerations of policy may require a stringent test to be adopted if this country is not to be flooded with those claiming political asylum."

The stringency of that test is met by adopting the view in the Joint Position adopted by the Council of Ministers in Brussels on 4th March 1996 that:-

"However, it is generally agreed that, in order to constitute "persecution" within the meaning of Article 1A, acts suffered or feared must:

- be sufficiently serious, by their nature or their repetition: they must either constitute a basic attack on human rights, for example, life, freedom or physical integrity or, in the light of all the facts of the case, manifestly preclude the person who has suffered them from continuing to live in his country of origin".

Thus the threshold for establishing persecution is set high. Anything short of a really serious flouting of the citizen's human rights and dignities will not do. Adjudicators and Appeal Tribunals should not flinch from applying a strict test.

That is the stringent standard set for a finding of persecution. It is, however, necessary to be more accurate. The real finding has to be of a fear of persecution which is well-founded. The question then is to what standard and how is that fear to be established. This has not been fully argued before us and so my views are tentative. I agree with Stuart-Smith L.J. that there must be a factual basis for all the findings that are necessary. Facts are proved on a balance of probability. Though the fear has to be a current fear presently held, it is actually a fear of events which are prospective and lie in the future. Proof depends upon the reasonable likelihood of the fear coming to pass as has been explained in Sivakumaran. To close the circle, there has to be a reasonable likelihood of the occurrence of acts of such seriousness as to be capable of amounting to the grave offence of persecution. There must be some factual basis from which an assessment of the risk can be made and those facts, importantly the historical facts of what actually happened to the asylum-seeker, are proved on the balance of probability. Once those findings are made, for my part I see no conceptual difficulty in then assessing whether there is a reasonable degree of likelihood that harm, so serious as to amount to persecution, may befall him.

The other grounds of appeal

I am in agreement with Stuart-Smith L.J. that no sufficient ground of appeal has been established. I agree that under the spotlight of fine analysis, some criticism can be made of uncharacteristic infelicities in the language and some lack of clarity in expressing separate conclusions at the different stages of the analysis. Having read, and re-read, the judgment with care - for its consequences are serious for the applicant - I cannot find that the Tribunal misdirected themselves or erred in principle or, when it came to making their findings of fact and their value judgments, exceeded the generous ambit within which there is room for reasonable disagreement.

The structure of the judgment as I read it amounts to this:-

1. Persecution requires State action or State inaction as I have already explained these concepts.

2. Whilst the actions of the gypsies towards the applicant gave him a well-founded fear that he would suffer ill-treatment at their hands, that was not enough by itself.

3. The breaches of human rights did not by themselves amount to persecution.

4. The acts of discrimination and breach of third category rights and the manifestations of violence had to be looked at cumulatively and as a whole.

5. The violent attacks were on the whole isolated and random attacks by thugs which occurred notwithstanding some arrests and prosecutions. The implicit finding is that the State protection had reached the required level of sufficiency. With the change of government, improvements were likely.

6. The appellant fell "below the high threshold ... required for international protection in a case where the fear is of an accumulation of discriminatory acts and where it is alleged that there is not a sufficiency of protection from non-State agents".

7. "His fear is not that of persecution."

8. "Our reading of the literature, although troublesome, does not bring us to the conclusion that the discrimination combined with the violence, together and accumulated can amount to persecution."

In my judgment the Tribunal correctly directed itself on matters of law and their findings of fact and their assessments of the weight of evidence were not perverse. Consequently, although I have approached the case in a different way from my Lord, I agree with him that the appeal should be dismissed.

LADY JUSTICE HALE:

55. I have had the advantage of reading in draft the judgments of Lord Justice Stuart Smith and Lord Justice Ward. I agree with them that the appeal should be dismissed. There is nothing which I can usefully add to their accounts of the facts or to their reasoning on all but the first ground of appeal. They differ fundamentally, however, in their approaches to that first and most important ground of appeal, as to where an alleged insufficiency of state protection against persecution by 'non state agents' fits into the definition of a 'refugee' in Article 1A(2) of the Geneva Convention relating to the Status of Refugees (as amended). Such a refugee is a person who:

' . . . owing to a well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion, is outside the country of his nationality and is unable or, owing to such fear, is unwilling to avail himself of the protection of that country; . . . '

This is a matter of some practical importance for asylum seekers, for the Secretary of State, and for adjudicators and the tribunal. With great diffidence, therefore, I feel it appropriate to set out my own approach, which is closer to that of Lord Justice Ward. In summary, my view is that the sufficiency or insufficiency of state protection against the acts of others may be relevant at three points in the argument: if it is sufficient, the applicant's fear of persecution by others will not be 'well founded'; if it is insufficient, it may turn the acts of others into persecution for a Convention reason; in particular it may supply the discriminatory element in the persecution meted out by others; again if it is insufficient, it may be the reason why the applicant is unable, or if it amounts to persecution unwilling, to avail himself of the protection of his home state.

56. It is common ground, and generally understood, that the Geneva Convention provides for a system of surrogate state protection: the states parties agree that, where another state has fallen down upon the obligations it owes to its citizens, they will step in to make good that failure. It does not, however, apply to every breach of every such obligation. It is limited in two important respects: first it aims to protect people who fear 'persecution', which is a strong word implying a high level of maltreatment; secondly, it aims only to protect people who fear persecution on account of their race, religion, nationality, membership of a particular social group or political opinion. As Lord Hoffman said in Islam v Secretary of State for the Home Department, R v IAT ex parte Shah [1999] 2 WLR 1015, at p 1032,

'In my opinion the concept of discrimination in matters affecting fundamental rights and freedoms is central to an understanding of the Convention. It is concerned not with all cases of persecution, but with persecution which is based on discrimination.'

57. It is also common ground, and generally understood, that the home state must bear some responsibility for the situation in which the applicant finds himself. There is otherwise no reason for another state to step in. The easy case is where the state itself is the persecutor: then its responsibility is clear. However, we also accept that the maltreatment feared may be perpetrated by people who are not themselves agents of the state: as Taylor J (as he then was) put it as long ago as 1985, in R v Secretary of State for the Home Department, ex parte Selladurai Jeyakumaran [1994] Imm AR 45, at 48, 'Again, I ask what solace it is to the victim to know he is being persecuted by soldiers out of control rather than by the Government, if that be the case.' This was accepted in the recent House of Lords' decisions in Islam v Secretary of State for the Home Department, R v IAT ex parte Shah [1999] 2 WLR 1015, although the matter did not require further discussion there as the Secretary of State conceded that the state was unwilling to protect wives in Pakistan from the maltreatment of their husbands. The question at issue was whether the victims were members of a particular social group. The argument in Adan v Secretary of State for the Home Department [1999] 1 AC 293 also proceeded on the assumption that lack of protection from persecution by non state agents could bring an applicant within the definition. The question at issue was whether the applicant's fear had to be current.
58. Various descriptions are given in the literature of the nature of the state's responsibility in such cases, ranging from active encouragement or complicity to a failure to provide sufficient or effective protection for its citizens. Some states are only prepared to recognise the unwillingness of the home state to provide protection but others, including the United Kingdom, will also act upon their inability to do so. It was for that reason that this court, in R v Secretary of State for the Home Department, ex parte Lul Omar Adan, Sittampalan Subaskaran and Hamid Aitseguer [1999] INLR 362, held that the Secretary of State should not have authorised the applicants' removal to third countries which recognised only the unwillingness rather than the inability of the home state to provide such protection. They did so on the basis of the words of Lord Lloyd in Adan's case [1999] AC 293, at p 306 where he said that the category of refugees who were outside their country owing to a well founded fear of persecution and unable to avail themselves of the protection of the country:

' . . . includes the important class of those who are sometimes called "third party refugees", ie those who are subject to persecution by factions within the state. If the state in question can make protection available to such persons, there is not reason why they should qualify for refugee status. They would have satisfied the fear test, but not the protection test. Why should another country offer asylum to such persons when they can avail themselves of the protection of their own country? But if, for whatever reason, the state in question is unable to afford protection against factions within the state, then the qualifications for refugee status are complete. Both tests would be satisfied.' (Emphasis supplied.)

Laws LJ, giving the judgment of the court in the later case, continued (at p 377):

'This accords with other jurisprudence in the English jurisdiction. Our courts recognise persecution by non-State agents for the purposes of the Convention in any case where the State is unwilling or unable to provide protection against it, and indeed whether or not there exist competent or effective governmental or state authorities in the country in question. This is what is called the protection theory. It is, as we have said, shared by a majority of the States signatory to the Convention and the UNHCR.'

59. Three interlinked questions therefore arise in this appeal: first, at what point or points in the consideration of whether or not an asylum seeker falls within the definition of a 'refugee' under Article 1A(2) of the Convention should the sufficiency of state protection from maltreatment perpetrated by others be considered; secondly, what is a sufficient or adequate level of protection; and thirdly, what is the standard to which the lack of such protection should be proved?
60. We are told that the only authority directly in point is that of the Supreme Court of Canada in Canada (Attorney-General) v Ward [1997] INLR 42. Based on that authority Mr Plender QC has argued that the sufficiency of state protection from persecution by non state agents is relevant only in considering whether the asylum seeker's fear of future persecution is 'well-founded'. Hence, that fear is 'well founded' if, in accordance with the test laid down in R v Secretary of State for the Home Department, ex parte Sivakumaran [1998] AC 958, there is a 'reasonable likelihood' of persecution taking place should the asylum seeker return to his home country. If what is feared is maltreatment by non state agents, and there is a reasonable likelihood of such maltreatment, and also a reasonable likelihood that the state will not prevent it, the test is fulfilled. This does appear to have been the approach taken by the IAT (Chairman His Honour Judge Cotran) in Yousfi v Secretary of State for the Home Department (United Nations High Commissioner for Refugees Intervening) [1998] INLR 136, at 138, where it was said that:

'The real question is not whether the State authorities are doing the best they can in all the circumstances, but whether viewed objectively the domestic protection offered by or available from the State to the appellant is or is not reasonably likely to prevent persecution from, in this case, the GIA . . .'

61. I can readily understand how, if the protection available from the home state is such that there is not a reasonable likelihood of maltreatment by non state agents, then the applicant's case would fail at the outset because his fear of such persecution was not objectively well founded. That, it seems to me, was the approach taken by the IAT (Chairman His Honour Judge Pearl) in the cases of Jaworski v Secretary of State for the Home Department (appeal no 17152, unreported, notified 28 May 1998) and Mojka v Secretary of State for the Home Department (appeal no 51844, unreported, notified 16 September 1998), both concerning the Roma in Poland. In the latter the tribunal said this:

'In looking at the sources of protection in Poland today we are of the view that there is a sufficiency of protection. Accordingly, the appellants do not require international protection from any fear they may have of persecution. Their fears are not well- founded.'

The role of the state may be relevant in that way. But it does not follow that this is the only place at which it is relevant, still less that the duty of surrogate protection arises whenever the state is unable or unwilling to prevent a reasonable likelihood of persecution by others. The case of Canada (Attorney-General) v Ward [1997] INLR 42 is entirely distinguishable. As the Lord Justice Stuart Smith makes clear at paragraph 15 of his judgment, the Canadian statute is drafted in a different way from the Convention itself, and leads to a result which is not necessarily that which was originally intended or which should be applied here.

62. The IAT in this case treated the role of the state as relevant to whether the maltreatment complained of amounted to 'persecution'. There are several good reasons for looking at it in that context. The first is the general purpose of the Convention, referred to above, of providing surrogate state protection. Lord Lloyd pointed out in Adan's case [1999] 1 AC 293, at p 307, that 'it is academic writers who provide the best hope of reaching international consensus on the meaning of the Convention'. He went on to cite from Professor James Hathaway's leading textbook, The Law of Refugee Status, on the very different point at issue in that case. On the point at issue here, the concept of 'persecution', Professor Hathaway introduces his discussion in chapter 4 thus (at pp 103-104):

' . . . the intention of the drafters was not to protect persons against any or all forms of even serious harm, but was rather to restrict refugee recognition to situations in which there was a risk of a type of injury that would be inconsistent with the basic duty of protection owed by a state to its own population. As a holistic reading of the refugee definition demonstrates, the drafters were not concerned to respond to certain forms of harm per se, but were rather motivated to intervene only where the maltreatment anticipated was demonstrative of a breakdown of national protection.'

He continues (at pp 104-105):

'Drawing on these basic precepts, persecution may be defined as the sustained or systemic violation of human rights demonstrative of a failure of state protection.'

The references given in the footnotes to these pages also treat the state's role as part of the concept of persecution.

63. Secondly, this approach fits most comfortably with Professor Hathaway's concept of persecution as the systemic violation of human rights, as understood in the international instruments which define those rights. It is in that context that he develops (at pp 109-112) the division into first category rights, breach of which is always persecution, second category rights, breach of which is persecution unless derogation from the state's obligation is strictly required by the circumstances, third category rights, breach of which may be persecution in certain circumstances, and fourth category rights, which may be outside the state's duty of protection. He concludes (at p 112):

'In sum, persecution is most appropriately defined as the sustained or systemic failure of state protection in relation to one of the core entitlements which has been recognized by the international community.'

64. The concept of persecution as the failure to protect fundamental human rights has been applied by the IAT in cases such as Gashi and Nikshiqi v Secretary of State for the Home Department (United Nations High Commissioner for Refugees intervening) [1997] INLR 96 and has also been said to be helpful in the courts. It may not be universally accepted: it has the potential to expand the types of maltreatment involved way beyond what might usually be thought of as persecution. In this case, it figured mainly in the Tribunal's consideration of the acts of discrimination by state authorities falling short of violence of which the applicant complained. But it is an analysis which focuses clearly upon the state's obligations: it is the state which undertakes to respect the human rights of its citizens. This may involve the state imposing obligations upon non state agents, but the basic obligation to respect those rights is that of the state. This emerged very clearly from the examination of the International Covenant on Civil and Political Rights which took place during this hearing. To expect the state to provide a system of protection for such rights is one thing: to expect it to act so as positively to prevent any reasonable likelihood of other people infringing such rights is quite another.
65. Thirdly, it is crucial that the persecution be discriminatory: that is, that the maltreatment is meted out to particular types of citizens defined by race, religion etc. The necessary discriminatory element may be supplied either by the non state agents who perpetrate the maltreatment or it may be supplied by the state which fails to protect the victims. This seems to follow directly from the approach of the House of Lords in Islam's case. The domestic violence perpetrated by husbands upon their wives would not have amounted to persecution on its own; nor would a generalised inadequacy in the protection offered by the state of Pakistan against violence; it was the fact that the state was discriminatory in the protection it afforded to a particular group of its citizens, namely women, which turned those women into refugees within the definition. Lord Hoffman (at p 1035) posited this example from Nazi Germany:

'A Jewish shopkeeper is attacked by a gang organised by an Aryan competitor who smash up his shop, beat him up and threaten to do it again if he remains in business. The competitor and his gang are motivated by business rivalry and a desire to settle old personal scores, but they would not have done what they did unless they knew that the authorities would allow them to act with impunity. And the ground upon which they enjoyed impunity was that the victim was a Jew. Is he being persecuted on grounds of race? Again, in my opinion he is. An essential element in the persecution, the failure of the authorities to provide protection, is based upon race.'

66. Fourthly, this approach enables one more readily to make sense of the definition as a whole. That definition, as Lord Justice Stuart Smith explains at paragraph 10 of his judgment, contains five elements: (1) the applicant is outside his home country because of a fear of maltreatment; (2) the maltreatment he fears is sufficiently grave to amount to persecution; (3) that fear is well-founded; (4) the persecution is discriminatory within the meaning of the Convention; and (5) the applicant is 'unable, or owing to such fear, unwilling to avail himself of the protection of that country.'
67. The inability to avail oneself of the protection of the home country is not, linguistically, linked to the fear of persecution and could arise for other reasons. The UNHCR Handbook, in paragraph 98, explains it thus:

'Being unable to avail himself of such protection implies circumstances that are beyond the will of the person concerned. There may, for example, be a state of war, civil war or other grave disturbance, which prevents the country of nationality from extending protection or makes such protection ineffective. Protection by the country of nationality may also have been denied to the applicant. Such denial of protection may confirm or strengthen the applicant's fear of persecution, and may indeed be an element of persecution.' (Emphasis supplied.)

68. The Handbook goes on (in paragraph 100) to explain that 'The term "unwilling" refers to refugees who refuse to accept the protection of the Government of the country of their nationality'. Obviously, surrogate state protection should not be made available to people who refuse the protection available for reasons unconnected with the fear of persecution. Hence their unwillingness must be owing to 'such fear', ie the well founded fear of persecution for a Convention reason. The link between the two halves of the sentence in such cases was emphasised by Lord Lloyd in Adan v Secretary of State for the Home Department [1999] 1 AC 293, at p 305:

'"Owing to such fear" in (b) means owing to a well-founded fear of being persecuted for a Convention reason. But "fear" in (b) can only refer to current fear, since the fear must be the cause of the asylum seeker being unwilling now to avail himself of the protection of his country. If fear in (b) is confined to current fear, it would be odd if "owing to well founded fear" in (a) were not also confined to current fear. The word must surely bear the same meaning in both halves of the sentence.' (Emphasis supplied.)

69. If maltreatment by non state agents is sufficient by itself to amount to persecution, and the state, although willing and able to give protection against it, cannot avoid a 'reasonable likelihood' of such maltreatment occurring, then it would follow that the applicant is (if he is in fact) unwilling owing to 'such fear' to avail himself of that protection. It seems, from the materials to which we have been referred, most unlikely that the parties to the Convention intended their surrogate protection to extend so far.
70. If, however, the persecution itself is defined as either discriminatory maltreatment by non state agents from which the state is unable or unwilling to afford protection, or serious maltreatment by non state agents from which the state refuses to afford protection for discriminatory reasons, then it becomes clear that the unwillingness to return is linked to the state's own failure to combat discrimination. This is just what one would expect.
71. Fifthly, on this reasoning, the discriminatory practice of the state is at least as important as the discriminatory practice of the attackers. This seems entirely right. If there are thugs about perpetrating serious acts of maltreatment against the population as a whole, but the state offers protection only to some of its citizens, and not to others, in my view those citizens are being persecuted in just the sort of way that merits the surrogate protection of other states under the Convention. But if the failure of state protection were relevant only to the fifth question, it is difficult to see how the necessary link with discrimination can be made in such cases.
72. As already seen, this analysis is entirely consistent with the decision in Islam's case. It also accords with the reasoning of Lord Hoffman in that case, at pp 1034 (also cited by the IAT in this case):

'What is the reason for the persecution which the appellants fear? Here it is important to notice that it is made up of two elements. First, there is the threat of violence to Mrs Islam by her husband and his political friends and to Mrs Shah by her husband. This is a personal affair, directed against them as individuals. Secondly, there is the inability or unwillingness of the state to do anything to protect them. There is nothing personal about this. The evidence was that the state would not assist them because they were women. It denied them a protection against violence which it would have given to men. These two elements have to be combined to constitute persecution within the meaning of the Convention. As the Gender Guidelines for the Determination of Asylum claims in the UK (published by the Refugee Women's legal Group in July 1998) succinctly puts it (at p 5): "Persecution = Serious Harm + The Failure of State Protection.'

As Lord Steyn was not addressing himself to this issue, it seems unlikely that he intended his summary of the qualifications of a refugee (at pp 1020 to 1021) to exclude this possibility. Indeed he might be read as including it, first at pp 1021 to 1022:

'It is accepted that each appellant has a well founded fear of persecution in Pakistan if she is returned to that country. The appellants are outside the country of their nationality. And they are unable to avail themselves of the protection of Pakistan. On the contrary, it is an unchallenged fact that the authorities in Pakistan are unwilling to afford protection for women circumstanced as the appellants are. Except for the requirements inherent in the words "persecution for reasons of ... membership of a particular social group" in article 1A(2) all the conditions of that provision are satisfied.'

And later, at p 1028, where he addresses the issue of causation:

' . . . it is plain that the admitted well founded fear of the two women is "for reasons" of their membership of the social group. Given the central feature of state-tolerated and state-sanctioned gender discrimination, the argument that the appellants fear persecution not because of their membership of a social group but because of the hostility of their husbands is unrealistic.' (Emphasis supplied.)

73. Nor does it seem to me that this approach is inconsistent with Lord Lloyd's separation of the definition into two tests, a 'fear test' and a 'protection test', both of which must be passed by the applicant, in Adan's case. The point that he was there addressing was whether the fifth element in the definition added anything to the other four elements, if not only the inability or unwillingness to avail himself of the home state's protection, but also the fear which caused his absence from his home country had to be in the present. The inability obviously adds something as it may be completely unrelated to the fear of persecution. The unwillingness adds something because it may serve to exclude an applicant whose reluctance to return is in fact based upon something quite different from the persecution feared.
74. It is also remarkable that the literature to which we have been referred deals with the role of the state in relation to persecution by non state agents largely in the context of the definition of persecution rather than in the context of the inability or unwillingness of the applicant to avail himself of its protection. This is true of the UNHCR Handbook itself (at paragraph 65), of the Joint Position of the Council of Ministers of the European Union (at paragraph 5.1), of the account in Macdonald and Blake, Immigration Law and Practice (at paragraph 12.39), and of Professor Goodwin-Gill, in The Refugee in International Law (2nd edition at pages 70 to 74). There is nothing in the submission of the UNHCR to the tribunal in this case which is inconsistent with this approach. This suggests that if the IAT in this case were in error they were in good company in that error. It also suggests that this court should be slow to adopt a different approach from that adopted generally in this field.
75. Nor do I understand counsel for the Secretary of State, Mr Howell QC, to be arguing against it. The primary submissions of both counsel favoured a different approach. Mr Plender, as already seen, argued that the sufficiency of state protection was relevant only to the issue of whether or not the applicant's fear was well-founded. Owing to the understandably low threshold for such a fear, on his case state protection would be insufficient if it could not prevent a reasonably likelihood of persecution by non state agents in the future. Mr Howell's primary submission was that the sufficiency of state protection was relevant only to the fifth question: was the applicant unable, or owing to such fear unwilling, to avail himself of the protection of his home state? But he would also be quite content with an approach which included the sufficiency of state protection within the concept of persecution for a Convention reason. The definition and test of such protection would be the same in each case. They are as Lord Justice Stuart Smith has stated them in paragraphs 21 and 22 of his judgment and there is nothing which I need add on this point.
76. I conclude, therefore, that the sufficiency of state protection may be relevant at all three of the points discussed and in the different ways explained. As this is the approach which the IAT (with the exception of the Yousfi case mentioned in paragraph 6 above) appears to have been following so far, it does not seem to me that the immigration authorities will find it too difficult or complex to apply in the future. It accords with the 'holistic' approach to the definition of a refugee which is mentioned in several sources. In my view, the tribunal not only reached the right conclusion in this case, but did so along a route which they were entitled to take.

Order: Appeal dismissed; no costs. Leave to appeal to the House of Lords was granted. Order does not form part of approved judgment.