England and Wales Court of Appeal (Civil Division) Decisions

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

		Royal Courts of Justice Strand, London, WC2A 2LL
		25th January 2000

LORD JUSTICE BROOKE:

1. This is an appeal by Nalliah Karanakaran from an order of the Immigration Appeal Tribunal dated 8th April 1999 whereby it dismissed his appeal from an order of a special adjudicator dated 2nd June 1998 dismissing his appeal from removal directions dated 21st February 1996. These followed a decision of the Secretary of State dated January 1996 refusing his application for asylum. On 21st February 1996 a notice of refusal of leave to enter this country was served on him, together with the directions for his removal to Sri Lanka.
2. In granting permission to appeal to this court the Vice-President of the Immigration Appeal Tribunal commented that the appeal raised a question of law of general applicability as to the correct standard of proof to be applied when deciding the reasonableness of internal relocation.
3. The appellant was born in September 1977 in Jaffna, Sri Lanka, and grew up in the Point Pedro area. The story of his childhood and adolescence is similar to that of many young men who came from the northern part of Sri Lanka.
4. When he was about 12 his family home was regularly raided by the army, and the members of his family were beaten. Because they were children, the appellant and his siblings suffered less severely. Their father suffered the most, and on one occasion he required hospitalisation for about 20 days. The soldiers stole jewellery, money and other valuables.
5. In 1993 three of the appellant's uncles were killed by the army in a raid on his grandfather's house, which was only a few minutes' walk away from his own home.
6. In July 1994 his home district was bombed by government forces. His home was destroyed, along with schools and community amenities, and many people were killed. His sister received a leg injury and was hospitalised. The community was effectively wiped out, and the survivors moved elsewhere. The appellant and his family moved to Meesalai.
7. He lived there for about six months. During that time his home was raided by security forces four times. He was also harassed by the LTTE (the Tamil Tigers) who brought pressure on him and his family to join their cause.
8. In January 1995 two of his friends were abducted by the LTTE. He feared a similar fate, and arrangements were therefore made for him to leave the country. With the help of an uncle he travelled to Colombo. His uncle contacted and paid an "agent" a fee of US$7,000 to transport the appellant to England. The agent placed the appellant in a lodging house while his uncle returned to the north. The following day the appellant was arrested at the lodging house because he had no identification papers. He was detained for three days before the agent secured his release by using bribes.
9. On 23rd February 1995 he left Sri Lanka and travelled to this country via Singapore, which he left on 4th March 1995, and Mauritius. He arrived here on 5th March 1995 and claimed asylum on his arrival.
10. In his determination the special adjudicator accepted the appellant's evidence. He accepted that he came from an area of high risk and that his family had been caught up in the conflict. He also accepted that if the appellant had remained in the north, there would have been a strong probability that he would have been forcibly recruited for the LTTE. However, he had managed to leave the north, bypassing LTTE and army check points. Although he was rounded up by the police in Colombo, the reason for this was that he was newly arrived from the north, and he was released in three days, albeit after the agent's intervention. Within 48 hours of his release he embarked from Colombo international airport, travelling on his own Sri Lankan passport and in his own identity.
11. On the totality of the evidence the special adjudicator concluded that the appellant was of no specific interest to the authorities in Colombo. He had every sympathy with the appellant in his plight, with close relatives killed and his family dispossessed, but he could find no evidence that he or his family were ever singled out for retaliatory oppression. They were the victims of a general onslaught. The appellant had therefore failed to make out a well-founded fear of persecution for a Convention reason.
12. The special adjudicator then considered the effect of the judgment of this court in Robinson [1998] QB 929. He found that the appellant would not face undue hardship if he were to return to Colombo. There was no information on which he could make a finding as to whether it would be safe for him to return to Meesalai.
13. For the purposes of his determination he considered it sufficient if he were to decide the issue of safety and reasonableness of return in relation to Colombo, which was where any asylum seeker would be returned to in the first instance.
14. On this issue he made his decision in these terms:

"I acknowledge that the appellant does not speak Sinhalese, and that he has no home or job to which to go in Colombo, but that does not alone indicate that it would be unreasonably harsh for him to be returned there. There are many thousands of Tamils living safely in Colombo. Some are Colombo residents of long standing but many others are refugees from the north. This appellant is now aged 20. There is no evidence to suggest that he would be of any interest to the authorities. There is nothing to single him out, or to sustain a well-founded fear for any Convention reason. I am satisfied that it would not be unduly harsh for him to be returned."

15. For the purposes of his appeal from this decision the appellant placed before the Immigration Appeal Tribunal the written opinions of four people who had expert knowledge of conditions in Sri Lanka.
16. Mr Jonathan Spencer is a social anthropologist based at Edinburgh University. He has been conducting academic research on Sri Lanka for nearly 20 years. He visited the island five times between 1990 and 1998, including two visits in 1997. He has published two books on aspects of the ethnic crisis in Sri Lanka, and more than 20 articles on politics, religion and violence in the country.
17. Mr Spencer regarded as quite extraordinary the special adjudicator's statement that there were many thousands of Tamils living safely in Colombo. He said that in recent years many Tamils from the north and east had moved out of the war zone and settled in Colombo. At the same time the LTTE had targeted Colombo for attacks by suicide bombers, and the authorities had responded with greatly increased security checks across the city.
18. These security checks were almost exclusively focused on Tamils, especially Tamil men, who found themselves stopped, searched and often detained solely because they were Tamil. In the last week of March 1998 about 5,000 Tamils were detained in Colombo. Following complaints by human rights organisations and Tamil MPs, at the beginning of April the Attorney-General directed the security forces to end this wave of arrests.
19. Mr Spencer also referred to at least four cases in March of groups of returned Tamil asylum-seekers being arrested and detained in Colombo. Some of them were released after intervention by MPs, but others continued to be held several weeks after their arrest. Mr Spencer expressed the view that any Tamil resident in Colombo was currently at considerable risk of arbitrary arrest and detention, and that returned asylum-seekers seemed to be especially vulnerable.
20. An opinion was also obtained from Dr M P Moore, who is a Fellow of the Institute of Development Studies at Sussex University. He is an academic political scientist who has had a special interest in Sri Lanka since 1973. He has lived there for several years, written extensively about the country, and visits it regularly.
21. Writing in June 1998, he said that the ethnic conflict in Sri Lanka had intensified over the past two years, in two distinct senses. On the one hand the government had launched a major offensive against the areas held by Tamil separatists. On the other, the Tamil separatists had shown an increased capacity to conduct bombing and other terrorist operations in Colombo and other parts of the country.
22. On the question whether the appellant could live safely in Colombo, Dr Moore observed that it was well known that young Tamil men living in Colombo were regularly rounded-up during security checks. Non-Sinhalese speaking Tamils with any kind of political record were very vulnerable in Colombo to harassment and extortion, particularly if it was known that they had returned from the West. Dr Moore gave examples of the kind of things that are now going on. He said the LTTE had a presence in Colombo, and could get at the appellant if they had reason to do so.
23. He added that the situation had worsened recently because of the genuine fear of terrorism, and the way in which this was exploited by unscrupulous police officers and others. Checks and controls on Tamils living in Colombo had been tightened considerably. Any kind of work was difficult to obtain. There was media talk of allowing those from Jaffna only to live at specially licensed places - a form of imprisonment.
24. In the circumstances Dr Moore judged that there was a serious possibility that the appellant would be harmed in Colombo if forced to return to Sri Lanka.
25. A third expert, Dr Richard Slater, is based at the International Development Department of the School of Public Policy at Birmingham University. He expressed the view that given that the appellant had no friends, family or close contacts living in Colombo, and was unable to speak Sinhalese, he would experience considerable hardship in securing a shelter and work on his return. At the same time the authorities might well suspect him of having LTTE links, and these suspicions could be reinforced by the fact that he has now spent several years in London, which is known to be home to many thousands of LTTE sympathisers and funders. He believed that in this situation it could well be unduly harsh for the appellant to be returned to Sri Lanka.
26. Dr Piers Vitebsky, for his part, considered that it would indeed be unduly harsh to return him there. He is another social anthropologist, based at Cambridge University, who specialises in ethnic affairs, particularly those of minority peoples in Russia and the Indian sub-continents, where he has conducted over six years of field-work, including 15 months spent in a Sinhalese-Tamil border zone in Sri Lanka.
27. He observed that the appellant had no relatives or other reliable contacts in Colombo. This would give him severe problems of housing and subsistence in Colombo, and make him extremely vulnerable to arrests by government forces, as well as to harassment from both the government and the LTTE. The agent would not be there to protect him, and his inability to speak Sinhala could in itself lay him open to harassment.
28. Perhaps more seriously, he could not avoid being singled out as a Tamil who had sought asylum abroad. This danger would begin immediately on arrival at the airport and would follow him into the city. If his passport revealed he had been living in London, it would raise him to the dangerous status of a person who would be of interest to the security forces. They consider London to be the centre of LTTE activity, and would surely suspect a young Tamil who had lived there of LTTE activity. It would also alert LTTE agents in Colombo to his previous avoidance of their conscription. Dr Vitebsky considered that the appellant would now be in greater danger once he had returned to Colombo, than he would have been in if he had never left the country.
29. In its determination the tribunal observed that the appellant had spent only one week in Colombo. Although he had been picked up by the police on a routine check, he was not ill-treated by them and was released on the payment of a bribe. He was able to leave Sri Lanka without any difficulty. There was no evidence of any confrontation or dealing with the LTTE in Colombo. In those circumstances the only evidence was that he was not pestered by the authorities or by the LTTE while he was there.
30. The tribunal said that most of the experts' opinions was pure speculation. Although the experts considered it would be unduly harsh to return the appellant to Colombo, their opinion must be looked at in the light of the evidence and of what has been held by the courts and the tribunal in cases of young Tamils who have fled from Jaffna and have gone to Colombo.
31. The test set out in Robinson [1998] QB 929 involved investigating whether it would be unduly harsh to send the appellant to Colombo. For the experts to say that he had no friends, family or close contacts living in Colombo, that he did not speak Sinhalese, and would experience hardship in seeking shelter and work upon his return, were not considerations which the tribunal should take into account in view of what was held in Robinson.
32. Likewise, for the experts to say that the appellant, as a young Tamil who had any kind of 'political' record, would be vulnerable in Colombo to harassment and extortion or could be formally arrested or simply kidnapped was a possibility (sic), but since he had no 'political' record and no connection with the LTTE, his only fear would be of being rounded up, interrogated and, in all likelihood, released within a very short time. The tribunal did not consider that this treatment would come within the term 'unduly harsh' or unreasonable.
33. The tribunal went on to consider individually each of the points made either by the experts or by the appellant himself. It either discounted a point because it was not sufficient to establish the contention that it would be unduly harsh to return him to Colombo or because it was far too speculative. Its conclusion is encapsulated in the following paragraph of its determination:

"As we see this case, while the appellant may encounter certain difficulties in finding housing and employment in Colombo and while he may be rounded up and questioned by the police as a young Tamil, he has not shown, in any way, that it would be 'unduly harsh' or 'unreasonable' for him to return to live in Colombo; it is, after all, the capital of his own country, it is populated by a large number of Tamils and Tamil-speaking people, and the authorities there are committed to the suppression of the LTTE."

34. The tribunal then turned to the question of the standard of proof to be applied in considering whether or not it would be unduly harsh or unreasonable for the appellant to be returned to, or be required to live in, Colombo. It cited the second half of paragraph 28 and the whole of paragraph 29 of the judgment of this court in Robinson, before concluding in these terms:

"Accordingly ... we are of the view that it is not necessary to decide whether the Sivakumaran standard should apply or the 'balance of probabilities' [standard] should apply, as what was held by the Court of Appeal was that the Tribunal, or the Court, having the internal flight alternative issue before it, should decide what is reasonable, in all the circumstances, as the operative words in paragraph 343 [of HC 395] are 'the application may be refused'.

As we see the situation, following Robinson, a common-sense approach rather than a legalistic or formulaic approach, should be adopted, and the Tribunal or the special Adjudicator dealing with the matter, having weighed up all the evidence, should take into account all the appropriate factors, as set out in Robinson, and decide what is reasonable in all the circumstances."

35. Applying that approach, the tribunal found it would not be unduly harsh to expect the appellant to be required to return to or live in Colombo. It therefore dismissed the appeal.
36. This appeal once again raises questions relating to what has been called the "internal flight alternative" in asylum law. It is also variously described as "internal relocation" or the "internal protection principle". It comes into play when conditions in one part of a country are such that there is a serious possibility that an asylum-seeker would face persecution for a Convention reason if sent back there, but there are other parts of that country where the same concern would not arise.
37. In English courts and tribunals the appropriateness of internal relocation has been a fairly familiar topic for debate in cases involving Tamils, and particularly young Tamil men, who grew up in the northern part of Sri Lanka and are afraid to go back there. In Ravichandran [1996] Imm AR 97 this court held that the fact that young male Tamils in Colombo were often rounded up by the security forces when there was terrorist activity in that city could not be equated with persecution for a Convention reason. During a critical time in Colombo the loss of liberty was relatively limited, and the purpose of the round-ups was not the oppression of Tamils per se but the maintenance of public order.
38. Since the decision in Ravichandran, it has often been argued in cases of individual asylum-seekers from Sri Lanka (and, indeed, from other countries) that the alternative destination to which they are to be sent back does not provide the quality of internal protection that the Geneva Convention demands, and that they are therefore still properly to be recognised as refugees.
39. This argument turns on the correct interpretation of a few words contained in the definition of "refugee" in Article 1A(2) of the Convention, being any person who

"... owing to well-founded fear of being persecuted [for a Convention reason] 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." (Emphasis added).

40. The words I have italicised have not been interpreted literally. In theory it might be possible for someone to return to a desert region of his former country, populated only by camels and nomads, but the rigidity of the words "is unable to avail himself of the protection of that country" has been tempered by a small amount of humanity. In the leading case of Robinson [1998] QB 929 this court followed an earlier decision of the Federal Court of Canada and suggested that a person should be regarded as unable to avail himself of the protection of his home country if it would be unduly harsh to expect him to live there. Although this is not the language of "inability", with its connotation of impossibility, it is still a very rigorous test. It is not sufficient for the applicant to show that it would be unpleasant for him to live there, or indeed harsh to expect him to live there. He must show that it would be unduly harsh. (For an interpretation of the word "unduly" in the context of the statutory phrase "unduly lenient" see Attorney-General's Reference (No 15 of 1990) 92 Cr App R 194 per Lord Lane CJ at pp 198-199).
41. The issue that has arisen for decision in this case relates to the method of establishing whether it would be unduly harsh to expect an asylum-seeker to live in a different part of his own country. As with the cases which preceded Robinson there have been conflicting decisions at tribunal level. One division of the tribunal, headed by Judge Pearl, its former president, has held that the applicant has to show on the balance of probabilities that it would be unduly harsh to send him back to that part (see Manohoran [1998] Imm AR 455). Another division, headed by Professor Jackson, a vice-president of the tribunal, decided eight months later that the applicant merely has to show that there would be a serious possibility that it would be unduly harsh for him to be returned there (see Sachithananthan [1999] INLR 205). We have been told that different divisions of the tribunal have applied one or other version of these two conflicting tests, and that there are about ten cases in this court awaiting the outcome of this appeal. It is pleasant to record that despite the volume of business in this court and the incidence of the long vacation, we have been able to hear this appeal within five months of the lodging of the notice of appeal in the Civil Appeals Office.
42. As I shall describe, the issues we had to decide on this appeal were significantly increased as a result of certain observations made in the judgments in another division of this court in Horvath (CAT 2 December 1999) which were handed down the day after the initial hearing of this appeal was concluded. For the time being, however, I will limit myself to the issues we were initially invited to consider.
43. It is necessary to start this part of this judgment by saying something about previous decisions in both England and Canada which relate to different aspects of the standard of proof in asylum cases. Later in the judgment I will review the course the law has taken in recent years in Australia.
44. The English cases show that the courts have recognised that different techniques are required in asylum cases when a decision-maker has to make judgments about future outcomes. The law in this respect is now authoritatively settled in this country by the decision of the House of Lords in Sivakumaran [1988] 1 AC 958. In that case it was held that when deciding whether an applicant's fear of persecution was well-founded it was sufficient for a decision-maker to be satisfied that there was a reasonable degree of likelihood that the applicant would be persecuted for a Convention reason if returned to his own country (see Lord Keith at p 994F and Lord Goff of Chieveley at p 1000F). Support was afforded by an earlier decision of the House in Fernandez v Government of Singapore [1971] 1 WLR 987, an appeal concerned with the proper interpretation of Section 4(1)(c) of the Fugitive Offenders Act 1967 ("if it appears ... that [the appellant] might, if returned, be ... detained or restricted in his personal liberty by reason of his ... political opinions"). Lord Diplock held at p 994 that bearing in mind the relative gravity of the consequences of the court's expectation being falsified, it was appropriate to adopt a lesser degree of likelihood than that inherent in the expression "more likely than not". He saw no significant difference between such expressions as "a reasonable chance", "substantial grounds for thinking", and "a serious possibility" as means of describing the degree of likelihood of the detention or restriction of the fugitive on his return which justified the court in giving effect to the provisions of Section 4(1)(c).
45. The decision in Sivakumaran did not, however, resolve the different, but related, question as to the standard of proof a decision-maker should apply when considering evidence of past or present facts before he or she goes on to make the necessary assessment of the future. This question surfaced before Nolan J in Jonah [1985] Imm AR 7, a case concerned with a senior trade union official in Ghana who had lost his job and suffered ill-treatment following political changes in Ghana. He had to hide in a remote village before seeking asylum in this country. The adjudicator acknowledged that he would be in jeopardy if he resumed his former activities, but concluded that he would be in no danger if he lived quietly in retirement. The Immigration Tribunal found no reason to interfere with the adjudicator's finding of fact and dismissed his appeal.
46. The question Nolan J had to decide was whether the adjudicator had adopted the appropriate standard of proof when he said that he could not be satisfied, even on the balance of probabilities, that Mr Jonah's declared fears of persecution if he was to return to Ghana were well-founded.
47. This case was decided before the decision of the House of Lords in Sivakumaran and before the new arrangements for asylum appeals that were introduced in 1993. Nolan J was concerned to apply what was then paragraph 134 of the Immigration Rules, which entitled to Secretary of State to remove an asylum-seeker if he was not satisfied that his fear of persecution was well-founded. He drew attention to the distinction made by Lord Diplock in Fernandez at p 993 between establishing the existence of facts and prophesying what can only happen in the future. He suggested that if a court is obliged to make an informed guess as to what might happen in the future, as was the case in relation to paragraph 134 of the Immigration Rules, it could only do so on the basis of the facts proved on the balance of probabilities.
48. He accepted that the likelihood of persecution contemplated by paragraph 134 was something different from proof on the balance of probabilities that persecution would occur. He did not, however, think that the matter could be usefully carried further than this without the danger of creating purely semantic problems where none existed for a tribunal applying its common sense and judgment to the facts proved before it.
49. This, then, as Miss Giovannetti correctly submitted, is authority for the proposition at high court level that in asylum cases it is the duty of the decision-maker to find past and present facts proved on the balance of probabilities, even if the assessment of the future calls for somewhat different techniques. We have to consider whether Nolan J's approach was correct.
50. In Kaja [1995] Imm AR 1 the Immigration Appeal Tribunal was concerned to resolve difficulties that had been confronting adjudicators following the decision of the House of Lords in Sivakumaran. Although Mr Kaja's appeal had been dismissed in quite robust terms, the adjudicator did not explain what standard of proof he had applied. A panel of senior legal members of the tribunal was therefore specially convened in order that they could give guidance on the correct approach to questions connected with the standard of proof to be adopted in asylum cases in relation to the establishment of past and present facts, as opposed to the assessment of future chances.
51. The majority of the tribunal considered that the question they had to decide was whether the assessment of an asylum case was a two-stage process or a one-stage process. They considered that it was a one-stage process. The task of the decision-maker was to assess, to a reasonable degree of likelihood, whether the applicant's fear of persecution for a Convention reason was well-founded. It might be that there were parts of the evidence which on any standard were to be believed or not to be believed. Of other parts, the best that might be said of them was that they were more likely than not. Of other parts it might be said that there was a doubt. The need to reach a decision on whether an appellant had made his case to a reasonable degree of likelihood, arose only on the ultimate evaluation of the case, when all the evidence and the varying degrees of belief or disbelief were being assessed.
52. The majority considered that if there was a first stage (proof of present and past facts) followed by a second stage (assessment of risk) then any uncertainties in the evidence would be excluded at the second stage, and that this could not be right. In those circumstances, they considered that the introduction of an intervening stage was simply an unnecessary complexity. They took the view that the authority of Nolan J's decision in Jonah had been overtaken by the later decision of the House of Lords in Sivakumaran.
53. It is clear that the majority was influenced by the notorious difficulty many asylum-seekers face in "proving" the facts on which their asylum plea is founded. In many of these cases, they said, the evidence will be the applicant's own story, supported in some instances by reports from organisations like Amnesty International. The stress generated by the nature of an asylum claim and the possible consequences of refusal, complemented by the highly formalistic atmosphere of interview or court, made the task of evaluating the evidence more complex. This did not mean that there should be a more ready acceptance of fact as established as more likely than not to have occurred. On the other hand, it created a more positive role for uncertainty. It would be a rare decision-taker who was never uncertain about some aspects of the evidence, particularly where, unlike civil litigation, evaluation was often concerned only with one version of the "facts". To say that it is only the facts established as more likely than not to have occurred on which the "reasonable likelihood" must be based would be, they said, to remove much of the benefit of uncertainty conferred on an applicant through Sivakumaran.
54. It is important to understand clearly the true effect of the majority decision in Kaja. They did not decide, as is suggested in one headnote ([1995] Imm AR 1) that:

"... the lower standard of proof set out in Sivakumaran applied both to the assessment of accounts of past events and the likelihood of persecution in the future."

55. What they decided was that when assessing future risk decision-makers may have to take into account a whole bundle of disparate pieces of evidence:

(1) evidence they are certain about;

(2) evidence they think is probably true;

(3) evidence to which they are willing to attach some credence, even if they could not go so far as to say it is probably true;

(4) evidence to which they are not willing to attach any credence at all.

56. The effect of Kaja is that the decision-maker is not bound to exclude category (3) evidence as he/she would be if deciding issues that arise in civil litigation.
57. It appears, however, that whatever the majority of the tribunal actually decided in Kaja, their decision has been generally interpreted as meaning that decision-makers are at liberty to substitute a lower standard of proof than that conventionally used in civil litigation when judges make findings about past and present facts. In Horvath [1999] INLR 7, a case in which the correctness of the decision in Kaja was challenged by the Secretary of State before the Immigration Appeal Tribunal (but not subsequently in this court), the tribunal said that whatever the majority may have said in their determination in Kaja, "everyone since that case thinks" [see p 20B] that they decided that an historical event or fact is proved by an asylum-seeker when he or she demonstrates that there is a reasonable likelihood that it occurred. This interpretation of that decision also appears in Professor Jackson's book "Immigration Law and Practice" (Sweet & Maxwell, 1996) at para 10-199.
58. Until the decision in Horvath was handed down by another division of this court, the Secretary of State has never, so far as I am aware, challenged the correctness of the decision in Kaja in the higher courts. On the initial hearing of the present appeal, indeed, Miss Giovannetti told us she was not instructed to dispute its correctness. We were shown by counsel how, in the context of torture, the supposed Kaja standard has been adopted by Parliament as the appropriate standard for assessing the likelihood of historic facts for the purposes of paragraph 5(5) of Schedule 2 to the Asylum and Immigration Appeals Act 1993 as substituted by Section 1 of the Asylum and Immigration Act 1996; and see now paragraph 9(7) of Schedule 4 of the Immigration and Asylum Act 1999.
59. In Canada it appears to be well settled law that an applicant must prove, on the balance of probabilities, that there is a serious possibility that he/she will face persecution for a Convention reason if sent back home, and if he/she is warned that it will be argued that internal protection is available elsewhere in his/her home country, that it would be unduly harsh for him/her to be expected to move and settle in that part (see Rasaratnam [1992] 1 FC 706; Thirunavukkarasu 109 DLR (4th) 682). We were not shown any Canadian authority which specifically addressed the issue raised in Kaja. In Rasaratnam Mahoney J said in the Federal Court of Canada that if an internal flight alternative issue was raised, the Immigration and Refugee Board had to be satisfied on the balance of probabilities that there was no serious possibility of a claimant being persecuted in the part of the country in which it found an internal flight alternative existed. In Thirunavukkarasu, which was decided in the same court the following year, Linden J gave practical illustrations of the sort of tests a decision-maker should apply in such a case, and in Robinson this court commended his approach to English decision-makers. In both these Canadian cases, however, the applicant was found to be a credible witness, so that no question arose about the appropriate way to approach any uncertainties in his evidence.
60. In Manohoran [1998] Imm AR 460, a case concerned with an internal flight alternative issue, the Immigration Appeal Tribunal said at p 460:

"We believe that the burden of proof remains on the appellant to show that a return to Colombo is unreasonable in the sense that it is unduly harsh. Secondly, the standard of proof in our view is the ordinary civil standard of a balance of probabilities. This is the position taken in the Canadian case of Rasaratnam. The lower standard developed in the Tribunal case of Kaja [1995] Imm AR 1 of a reasonable likelihood relates to the fear of persecution and whether that fear is well-founded. It is accepted by all that the appellant in this case will not be persecuted in Colombo. The question is 'would it be unduly harsh?' This is a very different question and we adopt the approach taken in Rasaratnam."

61. In point of fact it did not adopt the same approach because in Rasaratnam Mahoney J held, as I have said, that the decision-maker must be satisfied on the balance of probabilities that there was no serious possibility of the claimant suffering persecution in the relevant part of his home country, which is a rather different test. For different reasons neither counsel on this appeal suggested that we should follow Manohoran. Mr Lewis favoured a version of the test favoured by the tribunal in Sachithananthan, to which I will now turn. Miss Giovanetti favoured a different approach altogether. It was common ground that we should not adopt a test simply because it was a test used in Canada without knowing rather more about the standard of proof applied generally in Canada in asylum cases.
62. In Sachitharanthan [1999] 1 WLR 205 Professor Jackson, giving the determination of the tribunal, said at p 210 that the Manohoran approach created formidable difficulties. He observed, correctly, that the question whether or not there was an internal flight alternative was part and parcel of the question whether the applicant was a Convention refugee. In Canada all aspects of that question were decided by the test of showing that "on the balance of probabilities there is no serious possibility". Professor Jackson thought that this test was extremely difficult to interpret since it seemed to incorporate two different standards of proof. In addition, as a matter of English law, in so far as it related to an assessment of the likelihood of persecution, it conflicted with the decision of the House of Lords in Sivakumaran.
63. He said it was clear from decisions binding on the tribunal that "internal flight" was part of the consideration of whether the applicant was a refugee and therefore had the protection of the Convention (Robinson); and that the standard of proof applicable to the refugee issue was that of a "serious possibility" (Sivakumaran). The essential aspect of Kaja was that the approach to assessment of refugee status was one of a single stage:

"What should be anathema in an asylum case is the separation of the establishment of past events from the establishment of the risk in the future. The question is a single one of assessment of a serious possibility of persecution or, if relevant, it being 'unduly harsh' for the applicant to be returned. These matters can only be realistically assessed in respect of all aspects of the claim if the evidence of the past is approached in the context of the central issue of refugee status facing the decision-taker."

64. In encouraging us to follow this decision, Mr Lewis said the decision-taker should ask the single question: is there a serious possibility that it would be unduly harsh if the applicant was returned to [Colombo]?
65. Miss Giovannetti said the question was a much simpler one: would it be unduly harsh? It was concerned with a different aspect of the definition of a refugee to that considered by the House of Lords in Sivakumaran. This was because the question suggested by this court in Robinson "would it be unduly harsh to expect the applicant to [stay in Colombo]?" is adopted as a surrogate for the question derived from the wording of the Convention itself "is the applicant unable to avail himself of the protection of his home country in [Colombo]?" This was a quite different question from the one considered by the House of Lords in Sivakumaran.
66. She said that the answer to this question involved decision-makers in making a judgment as to the potential effect of what might be a number of quite disparate matters. She said the task they faced was similar to the task faced by judges in the county court for many years under Rent Act and cognate legislation when deciding whether it was reasonable to make an order for possession in all the circumstances of the case before them.
67. The decision-maker will not be evaluating the future likelihood of a single risk: the risk of persecution for a Convention reason. The serious possibility of that risk eventuating will have been eliminated before the question of internal protection is reached. Instead, he or she will be evaluating the effect of what may be a number of very different considerations. Some of them may depend on geography or climate. Some may depend on the personal characteristics of the particular applicant. Some may be not very serious, but bound to happen. Others may be potentially very serious, but the prospects of their occurring are slight. Others may fall somewhere between these two extremes, both as to likelihood and seriousness. The decision-maker has to consider the cumulative effect of all these considerations and then stand back and ask: in these circumstances, would it be unduly harsh to return the applicant to this place and expect him/her to live there? She submitted that nothing was gained by adding an extra layer to the question by asking "is there a serious possibility that it would be unduly harsh", since the evaluation of the likelihood of all the different untoward events occurring will have occurred at an earlier stage.
68. In Hotson v East Berkshire Health Authority [1987] AC 750 the House of Lords distinguished the forensic process of establishing past facts from that of evaluating future chances. Lord Mackay of Clashfern made the distinction clearly at p 785D-E:

"As I have said, the fundamental question of fact to be answered in this case related to a point in time before the negligent failure to treat began. It must, therefore, be a matter of past fact. It did not raise any question of what might have been the situation in a hypothetical state of facts. To this problem the words of Lord Diplock in Mallett v McMonagle [1970] AC 166, 176 apply:

'In determining what did happen in the past a court decided on the balance of probabilities. Anything that is more probable than not it treats as certain'."

69. The same distinction was made by Stuart-Smith LJ in Horvath. After making the same point as Lord Mackay made in Hotson, he continued at p 10:

"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."

70. In Horvath the applicant, a Slovak national, was a member of the Roma community. He claimed, among other things, that he feared persecution from skinheads from which the state did not provide him adequate protection, and the majority of this court considered that questions relating to persecution by non-state agents could not be logically separated from questions relating to the quality of the protection afforded by the state to a person in the applicant's position. That issue does not arise on the present appeal. The case is important in the present context, however, because two of the members of the court made observations relating to the burden of proof.
71. In paragraphs 24-26 of his judgment Stuart-Smith LJ addressed himself in conventional terms to questions relating to the burden of proof in civil litigation. He noted that the Secretary of State was not challenging in this court, as he had before the tribunal, the correctness of the majority decision in Kaja, which he described as "holding" that the lower standard of proof set out in Sivakumaran also applied to the assessment of accounts of past events. After suggesting that it might be desirable that this court should have an opportunity of considering the correctness of the decision of the majority in Kaja, he went on to say:

"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. [Counsel for the applicant], albeit by implication acknowledged this, hence his submission that the protection test should be embraced in the well-founded fear test."

72. Ward LJ said at p 35:

"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 LJ 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."

73. As I have already said, the decision in Horvath was handed down on the day after the initial argument on the present appeal was concluded. Stuart-Smith LJ's observations persuaded us that we ought to take this early opportunity of considering the correctness of the majority decision in Kaja for the first time in this court, and that we could not leave matters as they were. The whole position needed to be reviewed, more particularly because Stuart-Smith LJ, who clearly doubted the correctness of Kaja, said that the conventional standard of proof must be adopted in the assessment of historical and existing facts relevant to the application of the "protection test", and Ward LJ agreed with him, while accepting that the court had not heard full argument.
74. We therefore informed counsel we wished to relist the appeal for further argument. In particular, we told them we wished to hear argument on the following issues:

(1) Whether Kaja was correctly decided;

(2) Whether it would be possible to maintain a regime in which there was one standard of proof in relation to historic or existing facts for the purposes of the first part of the definition of "refugee" in the Convention, and a different standard of proof in relation to such facts for the purpose of considering issues of protection and internal relocation;

(3) The extent to which the assessment of an applicant's personal characteristics (when relevant to internal relocation issues) was inextricably bound up with the findings as to historic and existing facts that were made about him/her.

75. The appeal was restored for hearing on 21 December 1999, and the court is very grateful to both parties for the additional assistance they were able to provide.
76. It now transpired that the issues with which we are concerned on this part of the appeal have come before the High Court of Australia at least four times in the last ten years.
77. In Chan (1989) 169 CLR 379 the High Court held that in order to succeed in a claim for refugee status an applicant should show a "real chance" of persecution. Mason CJ adopted this formula because it conveyed the notion of a substantial, as distinct from a remote, chance. If an applicant establishes that there is a real chance of persecution, then his fear, assuming that he has such a fear, will be well-founded, notwithstanding that there is less than a 50 per cent chance of persecution occurring. The court added that even a 10 per cent chance that an applicant will face persecution for a Convention reason may satisfy the relevant test.
78. In Wu Shan Liang (1996) 185 CLR 259 the court explained how this test should be applied in practice in a particular case. The applicants had arrived in Australia from China on board a vessel called "The Labrador", and their claim for asylum was founded on the assertion that they would face persecution on their forced repatriation by reason of their illegal departure from China and their subsequent activities in Australia.
79. A number of different delegates of the minister, however, analysed case studies of those who had been previously returned from Australia to China. They found that a group which had arrived on a vessel called "The Jeremiah" had been very similar to the Labrador group, and that the Jeremiah group had not been persecuted on their return to China. In those circumstances they found that there was not a real chance that the Labrador group would face persecution on their return. They said that they gave greater weight to the evidence about the Jeremiah group than to the other evidence before them about known cases involving returnees or to general statements relating to the likely treatment of returnees to China.
80. The Full Court of the Federal Court of Australia set aside their decisions. That court was concerned about the place in the decisions that was given to the material which was given lesser weight. In relation to the reasons given by one of the delegates, it observed that she seemed to have approached the matter as if it involved the establishment of a state of affairs as being more probable than not, contrary to the test propounded in Chan.
81. The High Court of Australia reversed the Full Court's decision. Brennan CJ, Toohey, McHugh and Gummow JJ, in a joint judgment, said that the attribution of greater weight to one piece of information as against another, or an opinion that one version of the facts was more probable than another, was not necessarily inconsistent with the Chan test. They reminded themselves that in Chan Gaudron J had said:

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

82. The joint judgment observed that giving greater weight to one matter indicated that less weight was being given to another, but that the attribution of lesser weight was not the equivalent of rejection. In language very similar to that found in the majority determination in Kaja, they said:

"The chance of persecution is not a fact to be inferred solely from facts that are found to have existed; the very uncertainty of what has happened in other cases is itself material to the assessment of the chance of persecution in the instant case. As a matter of ordinary experience, it is fallacious to assume that the weight accorded to information about past facts or the opinion formed about the probability of a fact having occurred is the sole determinant of the chance of something happening in the future: the possibility that the future will not conform to what has previously occurred affects the assessment of the chance of the occurrence of a future event."

83. On the facts, the minister's delegates were entitled to give more weight to the case histories of recent returnees whose departure, and whose activities since departure, were "very similar" to those of the applicants. In other words, the material the applicants provided did not go very far towards satisfying the delegates that there was a real chance of persecution, because it was contradicted by more relevant material.
84. The joint judgment in Wu Shan Liang also contains a passage at paragraphs 53-54 which is illuminating when set against the dicta of Stuart-Smith LJ and Ward LJ in Horvath about the standard of proof to be adopted in an administrative fact-finding process of this kind:

"Submissions were made at the hearing of the appeal as to the correct decision-making process which it would have been permissible for the delegates to adopt. These submissions were misguided. They draw too closely upon analogies in the conduct and determination of civil litigation.

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

85. In the following paragraph, after adopting Lord Diplock's reasoning in Fernandez v Government of Singapore, the four judges said that:

"... the term 'balance of probabilities' was apt to mislead in the context of Section 22AA [of the Migration Act 1958, as amended] even if it be used in reference to 'what has already happened'."

86. It does not appear that this authority was drawn to the attention of the court in Horvath.
87. In my judgment this distinction between the task of a judge in civil litigation and the task of an administrative decision-maker in an asylum case is just as valid in this country as it is in Australia.
88. The High Court returned to these issues in Guo (1997) 144 ALR 567, another case concerned with the risk of persecution if an asylum-seeker was returned to China. The full court of the Federal Court of Australia had criticised the Refugee Review Tribunal on the grounds that it had given no consideration to the possibility that any of its findings of fact were inaccurate, and that there was in fact a possibility that Mr Guo's punishment (when he was returned to China on a previous occasion) had been Convention-related.
89. In its joint judgment the majority of the court said that the tribunal was entitled to weigh the material before it and make findings before it engaged "in any consideration of whether or not Mr Guo's fear of persecution on a Convention ground was 'well-founded'." Given the strength of some of the tribunal's findings adverse to Mr Guo, they held that it was not bound to consider the possibility that its findings were inaccurate or that his punishment was Convention-based. They added at pp 576-577:

"It is true that in determining whether there is a real chance that an event will occur, or will occur for a particular reason, the degree of probability that similar events have or have not occurred, or have not occurred for particular reasons in the past, is relevant in determining the chance that the event or the reason will occur in the future. If, for example, a tribunal finds that it is only slightly more probable than not that an applicant has not been punished for a Convention reason, it must take into account the chance that the applicant was so punished when determining that there is a well-founded fear of future persecution."

90. If, however, the tribunal took the view that the probability of error in its findings was insignificant, as appeared to have been the case with Mr Guo, it was not then bound to consider whether its findings might be wrong.
91. This approach was adopted by Gleeson CJ and McHugh J, the only two members of the court who addressed the issue, in Abebe (1999) 162 ALR 1. They acknowledged that:

"As Guo makes clear, even if the Tribunal is not affirmatively satisfied that the events deposed to by an applicant have occurred, the degree of probability of their occurrence or non-occurrence is a relevant matter in determining whether an applicant has a well-founded fear of persecution. The Tribunal 'must take into account the chance that the applicant was so [persecuted] when determining whether there is a well-founded fear of future persecution'."

92. In Abebe's case, however, the tribunal had been unable to accept the applicant's accounts of her arrest and of her husband's arrest. Since it found it could not rely on her evidence about her arrest and detention - and reference was made in this context to the inconsistencies and admitted lies in her various accounts - her further claims of detention and rape became logically irrelevant. Given the nature of her claim and the tribunal's finding that she was not a credible witness, it was not required, as it might have been in other circumstances, to determine whether there was a real chance that she had been arrested as she claimed.
93. These, and other relevant Australian decisions at Federal Court level, have been helpfully brought together in the recent judgment of Sackville J (with which North J expressly agreed) in that court in Rajalingam [1999] FCA 719, a judgment which shows how the Australian lower courts have been engaged in filling the gaps left by the High Court decisions.
94. Thus in Epeabaka [1999] FCA 1 the full court of the Federal Court, while referring to the difficulties of proof which beset asylum-seekers, pointed out that findings about past events affecting asylum-seekers will be necessary in most cases. It said in this context:

"Findings of fact based on likelihood will usually be findings made on the balance of probabilities arising from the available information before the decision-maker. However, when dealing with the claims of an asylum-seeker, the available evidence might not imbue findings so made with the degree of confidence that justify the conclusion that an asylum-seeker does not have a well-founded fear of being persecuted. It is for this reason that the civil standard cannot be universally applied to the fact finding process in claims of this kind. It is necessary to recognise the risk of error in adopting such a fact finding process, and to make allowance for it."

95. In Rajalingam Sackville J observed at paragraph 37 that this explanation of certain comments made by Kirby J in Wu Shan Liang (which Sedley LJ has reproduced in his judgment in the present case), although pointing out that findings of fact might be based on likelihood, did not detract from the proposition that the fact-finding process to be followed by the Refugee Review Tribunal differed from that applied in civil courts.
96. At paragraphs 48-50 Sackville J commended some observations made by Drummond J in Thanh Phat Ma (1996) FCR 431 when he interpreted Kirby J as saying in his judgment in Wu Shan Liang that:

"unless the decision-maker can dismiss as unfounded factual assertions made by the applicant, the decision-maker should be alert to the importance of considering whether the accumulation of circumstances, each of which possesses some probative cogency, is enough to show, as a matter of speculation, a real chance of persecution, even though no one circumstance, considered by itself, is sufficient to raise that prospect."

97. Sackville J commented:

"With respect, Drummond J's observations are helpful because they identify a second class of case in which, although the decision-maker finds that alleged past events have not occurred, the chance that they might have occurred could provide a rational foundation for finding that the applicant has a well-founded fear of persecution. A practical difficulty is that factual assertions made by applicants for refugee status concerning their own experiences can rarely be assessed independently of each other. The findings will usually depend on the decision-maker's assessment of the reliability of the applicant's account and of other factors common to all claims. It may therefore not be easy for the [tribunal] to identify those cases where the findings cannot be made with sufficient confidence to foreclose reasonable speculation. Perhaps that is the reason why Gummow and Hayne JJ in Re Minister for Immigration and Multicultural Affairs ex p Abebe ... described the [tribunal's] inquiry as 'attended by very great difficulties'."

98. At paragraphs 60-67 Sackville J derived the following principles from the decided cases:

(1) There may be circumstances in which a decision-maker must take into account the possibility that alleged past events occurred even though it finds that these events probably did not occur. The reason for this is that the ultimate question is whether the applicant has a real substantial basis for his fear of future persecution. The decision-maker must not foreclose reasonable speculation about the chances of the future hypothetical event occurring.

(2) Although the civil standard of proof is not irrelevant to the fact-finding process, the decision-maker cannot simply apply that standard to all fact-finding. It frequently has to make its assessment on the basis of fragmented, incomplete and confused information. It has to assess the plausibility of accounts given by people who may be understandably bewildered, frightened and, perhaps, desperate, and who often do not understand either the process or the language spoken by the decision-maker/investigator. Even applicants with a genuine fear of persecution may not present as models of consistency or transparent veracity.

(3) In this context, when the decision-maker is uncertain as to whether an alleged event occurred, or finds that although the probabilities are against it, the event may have occurred, it may be necessary to take into account the possibility that the event took place in deciding the ultimate question (for which see (1) above). Similarly, if the non-occurrence of an event is important to the applicant's case, the possibility that that event did not occur may need to be considered by the decision-maker even though it considers that the disputed event probably did occur.

(4) Although the "What if I am wrong?" terminology has gained currency, it is more accurate to see this requirement as simply an aspect of the obligation to apply correctly the principles for determining whether an applicant has a "well-founded fear of being persecuted" for a Convention reason.

(5) There is no reason in principle to support a general rule that a decision-maker must express findings as to whether alleged past events actually occurred in a manner that makes explicit its degree of conviction or confidence that its findings were correct. (In Guo, for instance, the High Court considered that it was enough that the tribunal appeared to have no doubt that the probability of error was insignificant).

(6) If a fair reading of the decision-maker's reasons as a whole shows that it "had no real doubt" that claimed events did not occur, then there is no warrant for holding that it should have considered the possibility that its findings were wrong.

99. Miss Giovannetti, for the Secretary of State, commended the Australian approach. Mr Lewis, also supporting this approach, reminded us that in Ravichandran [1996] Imm AR 97 Simon Brown LJ observed at p 109 that the question whether someone was at risk of persecution for a Convention reason "should be looked at in the round, and all the relevant circumstances taken into account". It was common ground between counsel that it would be quite impracticable to maintain a regime in which there was one approach to the evidential material relating to historic or existing facts for the purposes of the first part of the definition of "refugee" in the Convention, and a different approach to such material for the purpose of considering issues of protection and internal relocation. It was also common ground that the assessment of an applicant's personal characteristics (when relevant to internal relocation issues) was inextricably bound up with the findings as to historic and existing facts that were made about him/her.
100. In my judgment, the approach in fact recommended by the majority of the Immigration Appeal Tribunal in Kaja, as much more fully explained in the Australian cases whose effect I have summarised, is the approach which should be adopted at each of the stages of the assessment process with which we are concerned. In so far as the dicta of Stuart-Smith and Ward LJJ in Horvath may suggest that the approach favoured in civil proceedings should be adopted in this context in relation to protection issues, they should not be followed. As I am sure they would be the first to acknowledge, we have had the benefit of very much fuller argument on all these issues than was available to that court.
101. I must make it clear that I am aware of the decision of the majority of the House of Lords in In re H (Minors) [1996] AC 563, although it was not cited to us by counsel. Lord Nicholls of Birkenhead, in the leading speech in that case, made it clear at p 586 that he was treating family proceedings as essentially a form of civil proceedings. In the present public law context, where this country's compliance with an international convention is in issue, the decision-maker is, in my judgment, not constrained by the rules of evidence that have been adopted in civil litigation, and is bound to take into account all material considerations when making its assessment about the future.
102. This approach does not entail the decision-maker (whether the Secretary of State or an adjudicator or the Immigration Appeal Tribunal itself) purporting to find "proved" facts, whether past or present, about which it is not satisfied on the balance of probabilities. What it does mean, on the other hand, is that it must not exclude any matters from its consideration when it is assessing the future unless it feels that it can safely discard them because it has no real doubt that they did not in fact occur (or, indeed, that they are not occurring at present). Similarly, if an applicant contends that relevant matters did not happen, the decision-maker should not exclude the possibility that they did not happen (although believing that they probably did) unless it has no real doubt that they did in fact happen.
103. For the reasons much more fully explained in the Australian cases, when considering whether there is a serious possibility of persecution for a Convention reason if an asylum seeker is returned, it would be quite wrong to exclude matters totally from consideration in the balancing process simply because the decision-maker believes, on what may sometimes be somewhat fragile evidence, that they probably did not occur. Similarly, even if a decision-maker finds that there is no serious possibility of persecution for a Convention reason in the part of the country to which the Secretary of State proposes to send an asylum seeker, it must not exclude relevant matters from its consideration altogether when determining whether it would be unduly harsh to return the asylum seeker to that part, unless it considers that there is no serious possibility that those facts are as the asylum seeker contends.
104. Needless to say, as the High Court of Australia observed in Wu Shan Liang, when assessing the future, the decision-maker is entitled to place greater weight on one piece of information rather than another. It has to reach a well-rounded decision as to whether, in all the circumstances, there is a serious possibility of persecution for a Convention reason, or whether it would indeed be unduly harsh to return the asylum-seeker to the allegedly "safe" part of his/her country. This balancing exercise may necessarily involve giving greater weight to some considerations than to others, depending variously on the degree of confidence the decision-maker may have about them, or the seriousness of their effect on the asylum-seeker's welfare if they should, in the event, occur.
105. I should add, for the avoidance of doubt, that I accept Miss Giovannetti's submission that when dealing with questions of internal protection, the decision-maker should simply ask: would it be unduly harsh to expect the applicant to settle there? In answering this question it may have to take into account the cumulative effect of a whole range of disparate considerations, in respect of some of which it may be satisfied that they probably did occur (or are occurring), while in respect of others it may only think that there is a serious possibility that what the applicant and/or his/her witnesses is saying is correct.
106. Although we are not concerned in the present case with the possibility of persecution for a Convention reason by non-state agents against which the home state is unable to provide adequate protection, it follows from this analysis that the decision-maker should follow a similar approach in that context. After determining the level at which state protection is in fact provided, it should consider all the relevant circumstances (after discarding those it considers safe to eliminate altogether) when considering whether there is nevertheless a serious possibility of persecution occurring, and whether the level of state protection is sufficient by international standards.
107. How disparate some of the matters may be that the decision-maker has to evaluate may be seen by referring to Robinson itself and to the Tribunal decision in Sayandan (5th March 1998: HX/65429/96 (16312)).
108. In Robinson at p 940D the court referred to considerations which I would interpret as (i) the certainty of having to cross battle lines; (ii) the certainty of having to hide out in an isolated region of their country like a cave in the mountains, a desert or a jungle; (iii) the strong likelihood that the weather in a safe area will be unattractive; (iv) the strong likelihood (at first, any rate) of the applicant having no friends or relatives there; (v) the probability (or, the serious possibility) of him/her not being able to find suitable work.
109. In Sayandan 11 different considerations were suggested as worthy of the decision-maker's attention. They were (omitting the evaluations suggested by the applicant's counsel):

(1) The risk of the applicant's being arrested and returned to his homeland in north-east Sri Lanka because of his lack of appropriate documents;

(2) The risk of his being repeatedly arrested in round-ups;

(3) The risk that he would be subjected to extortion;

(4) The risk of unduly harsh treatment before obtaining access to judicial process;

(5) The risk, if the applicant is detained, of his being subjected to dreadful prison conditions;

(6) The risk of his not being able to find or retain accommodation;

(7) The risk of his not being able to find any employment, due to blatant discrimination in the labour market;

(8) The risk that his inability to speak Sinhalese would place him at a disadvantage in dealing with government officials;

(9) The risk that he would be subjected to a regime where racial discrimination was part of every day life;

(10) The risk that he would have no real contacts or ties in Colombo;

(11) The fact of his previous treatment in Sri Lanka by both the LTTE and the security forces.

110. The way the tribunal in Sayandan approached this rather disparate bundle of risks is in my judgment a good example of the way in which fact-finders should approach this issue. It reminded itself that if it found that there was a part of his country in which it would be unduly harsh to expect an applicant to settle, that part must be eliminated as a place to which he might be returned.
111. It removed from its consideration items (7) and (10), and also item (6). It regarded the risk identified in item (1) with some anxiety. It also took very seriously the risk not merely of detentions in regular round-ups (item 2), but detentions which might result in considerable periods of imprisonment in bad conditions prior to trial (items 4 and 5). It added that if the motivation for such treatment was ethnic the position was worse (semble, item 9). It made no particular comment on items 3 and 8. It then continued:

"Nevertheless, as we say, none of the matters identified by [counsel] would individually bring us to the view that it would be unreasonable or unduly harsh for the appellant to settle in Colombo. On the other hand, the appellant is not likely to be placed in such a situation that he can isolate these difficulties from one another. The factors are cumulative. He will be subject, immediately on arrival in Colombo, to each of the disadvantages [counsel] emphasised. Some of them will arise from his lack of proper Sri Lankan identity and travel documents; others will arise solely from his ethnic background. We think that this is a case where the appellant has established that it would be unduly harsh for him to have to be in Colombo."

112. It ended by adding a word of caution that this was not a decision that it was unduly harsh for young male Tamils to be in Colombo. It was a decision confined to the particular facts and evidence of the particular case.
113. I express no view on the merits of that decision. That is, and must be, a matter for the judgment of the members of that experienced specialist tribunal, and unless they have committed some error of law this court will not interfere with their judgment. What is relevant in the present context is the methodology they adopted. Unless something is so trivial that even on a cumulative assessment it would be bound to carry no weight, or the decision-maker has no real doubt that it is entitled to discard some point from its consideration altogether, it would be wrong to eliminate that point completely. In my judgment, the tribunal's technique in Sayandan of evaluating both the likelihood of a risk eventuating and the seriousness of the consequences if it were to eventuate demonstrates a correct approach. It was also correct for it to assess the cumulative effect of the matters it was considering, particularly if there was a likelihood that they would all affect the applicant at the same time.
114. It will be seen that that tribunal, whose decision predated Manohoran by three months, seems to have experienced no difficulty in deciding whether in the conditions it had evaluated it would be unduly harsh to expect the appellant to live in Colombo. The tribunal in the present case adopted a similar approach when it said that a common-sense approach, rather than a legalistic or formulaic approach, should be adopted (as opposed to considering whether it was more likely than not, or only a serious possibility, that conditions in Colombo would be unduly harsh).
115. The fact-finder must be careful, however, to evaluate each of the considerations suggested on behalf of the applicant. In my judgment it was completely wrong for the tribunal in the present case to dismiss considerations put forward by experts of the quality who wrote opinions on this case as "pure speculation". It was also quite wrong for it to say that certain matters were "not considerations which we should take into account" merely because in Robinson this court said that such considerations would not in themselves be enough to satisfy the requisite test. It was also wrong for it to consider each matter in isolation as opposed to considering their potential cumulative effect: see now Gnanam [1999] INLR 219 per Tuckey LJ at p 223F, and his warning at p 224H-225A:

"All that is said emphasises that each case must be decided on its own facts. What may be factors in one case will not necessarily be factors in another. Factors taken individually or cumulatively may tip the balance in one case but will not necessarily do so in another."

116. Because the tribunal adopted the wrong approach to the different considerations that were urged upon it, it appears to me to be inevitable that we should allow this appeal and remit the case to a differently composed tribunal. Although Miss Giovannetti urged us to follow the course the court followed in Robinson and to hold that on these facts no tribunal could properly find that it would be unduly harsh to expect the appellant to return to Colombo, I consider it would be wrong to do so for two reasons. The first is that the experienced members of the Immigration Appeal Tribunal can draw on a reservoir of knowledge and experience that is not available to this court. The other is that if we do remit the matter, both parties will be at liberty to submit up to date evidence about the situation relating to young male Tamils in Colombo. It would be much better for the ultimate decision on this appeal to be made by a tribunal which had access to such evidence, since the evidence before the court is nearly all about 18 months old.
117. I have had the opportunity of reading the judgment of Sedley LJ. I agree with it. I hope that these two judgments may be found to provide helpful guidance to decision-makers and practitioners who are concerned with this very difficult but important area of the law.
118. I would add one footnote to this judgment, The judgment of this court in Robinson has variously been ascribed to Lord Woolf MR (QB, WLR, All ER) and to myself (The Times, Imm AR). The former is correct. Although I prepared the first draft of that judgment, to which the other members of the court contributed, the court agreed that its judgment should be published in the name of Lord Woolf MR and that it should begin: "This is the judgment of the court, primarily prepared by Brooke LJ". This sentence, however, appeared on the title page of the judgment that was handed down, and not in the first line of the judgment, and this led to understandable confusion among law reporters.
119. For these reasons, I would allow this appeal and direct that the case be remitted to a differently composed tribunal. It may be thought desirable to hold a directions hearing at an early date, so that no further avoidable delay occurs before any new evidence is filed and the appeal is relisted for hearing.

LORD JUSTICE ROBERT WALKER:

1. I have had the privilege of reading in draft the judgments of Brooke and Sedley LJJ. I agree with both judgments.

LORD JUSTICE SEDLEY:

1. I agree that this case must go back for determination by a differently constituted appeal tribunal, if only because of the way in which highly relevant evidence of in-country conditions from experts with respectable credentials was dismissed by this tribunal as mere speculation.
2. But I agree too that the appeal requires rehearing on a correct foundation of law in relation to the issue of internal relocation. This in turn throws up a larger question which has vexed asylum law for some time: what are the correct mode and standard of proof? Although the question arises for us in relation to internal relocation, for reasons which will be apparent it cannot be treated separately from the general question of proof in asylum cases. It may be helpful first to look at these issues individually and then to see how they dovetail.
3. Before doing this, however, it is necessary to unravel the reasoning of the special adjudicator and the appeal tribunal, both of which in my respectful view are faulty. The appellant is a young Tamil from the Jaffna peninsula whose community was destroyed by the civil conflict and who fled from his home area in fear of both the government forces and the terrorist movement. All this was found as fact. So was the consequent history of flight, first to Colombo and ultimately to the United Kingdom. It followed that (unless there were a finding that flight was not a logical reaction to the persecution - a possibility in certain cases but not in this one) the appellant was outside the country of his nationality owing to a well-founded fear of being persecuted for reasons of race. He was therefore entitled by virtue of Article 1(A)(2) of the 1951 Geneva Convention to asylum provided that, in addition, it could be established that he was "unable or, owing to such fear, ... unwilling to avail himself of the protection" of his home state. The latter - unwillingness through fear - is what this appeal is, at least initially, about.
4. Unfortunately both the special adjudicator and the tribunal failed to approach the Convention methodically. They treated the availability of internal flight as a reason for holding that the fear of persecution was not well-founded. There may possibly be countries where a fear of persecution, albeit genuine, can so readily be allayed in a particular case by moving to another part of the country that it can be said that the fear is either non-existent or not well-founded, or that it is not "owing to" the fear that the applicant is here. But a clear limit is placed on this means of negating an asylum claim by the subsequent provision of the Article that the asylum-seeker must be, if not unable, then unwilling because of "such fear" - ex hypothesi his well-founded fear of persecution - to avail himself of his home state's protection. If the simple availability of protection in some part of the home state destroyed the foundation of the fear or its causative effect, this provision would never be reached. This is why in most cases, including the present one, it is in relation to the asylum-seeker's ability or willingness to avail himself of his home state's protection that the question of internal relocation arises. Because, however, unwillingness is explicitly related to the driving fear, it predicates a different set of considerations from inability, which may be indicated or contra-indicated by a much wider range of factors.
5. For a young Tamil whose arrival in Colombo, where he had neither family nor friends nor housing nor work, had been followed by round-up and imprisonment, internal relocation to Colombo was anything but an obvious option. The reality, on the special adjudicator's findings of fact, was thus that the appellant was in this country because he had a well-founded fear of persecution on ethnic grounds in Sri Lanka. Because it was common ground that the same sources of fear were absent in Colombo, the question of unwillingness to return there because of the original fear did not arise; but this, under the Convention, does not undo a claim for asylum. The remaining questions under the Convention were whether Sri Lanka was able to offer the appellant protection in Colombo; and if it was able to offer it, whether the appellant was able to take advantage of it.
6. The Home Secretary's case in short is that in Colombo, which is the place to which the appellant would be returned, there are no substantial grounds for fearing persecution as a Tamil. This the appellant accepts; but he contends that it is not reasonable to ask him to go there because he has no family, friends, work, source of income or shelter in Colombo and does not speak the dominant language, Sinhalese. The Home Secretary responds that the appellant would, even so, be no different from the many thousands of Tamils already living in safety in Colombo.
7. The question we have now to decide is how a decision-maker, a tribunal or a court is to gauge whether internal relocation is a legitimate alternative to asylum for a person who otherwise ranks as a Convention refugee. Is the want of such an option to be proved by the asylum-seeker (in which case it is common ground that proof would not have to go as high as a balance of probability); disproved by the Home Secretary (in which case it would follow that the standard exceeds a bare balance of probability); or simply gauged on the evidence?
8. It is to be observed that the argument has now moved, for reasons analysed above, from the question of persecution to the broader question of conditions of survival. It is common ground here and throughout the common law jurisdictions whose decisions we have seen that ability to return is not literal or absolute but a question of what it is reasonable to expect of a particular applicant in particular circumstances, and that what is reasonable in this field is best tested by asking whether return for relocation would be unduly harsh. Hence, among other things, the potential importance of the expert evidence in this case.
9. Were it not for the decision of another division of this court in Horvath v Home Secretary (Stuart-Smith, Ward and Hale LJJ, 2nd December 1999), handed down the day after the conclusion of the first day's argument in the present case, one could move directly to the question of the mode or standard of proof. Horvath concerned the inability or unwillingness of the Slovak state to protect Roma from racial persecution by neo-Nazis. The court, while united in dismissing the appeal on the ground that neither was established by the evidence, was divided about the proper route to this conclusion. Stuart-Smith LJ considered that the elements of entitlement to asylum had to be approached sequentially; that so approached, the state's ability or willingness to afford protection related not to the question of past or prospective persecution but to the applicant's ability or willingness to avail himself of such protection; but that although the Tribunal had taken a contrary view of the law, its findings answered the question, when correctly posed, in the Home Secretary's favour. Ward LJ took the view that entitlement to asylum was a unitary concept and that the Tribunal had therefore been right to approach the want of protection as an element of persecution where non-state agents were implicated. It followed that in his view the IAT had given a tenable answer to the right question. Hale LJ, concurring in the outcome, did so by a route which she described as "closer to that of Lord Justice Ward." She summarised her view thus:

"... 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...; 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 the home state."

10. These are in truth three distinct interpretations of the all-important Article A.1(2) of the Convention. As the outcome of Horvath demonstrates, it will often not matter to the result which approach is taken. But it does matter to the present case because it is not possible to consider what is the appropriate test of the possibility of internal flight until one knows what it is, in Convention terms, that is being tested: is it the grounds for fearing persecution, the quality of protection available against it, the entire Convention formula or none of these things? I include the final possibility because both counsel before us have agreed that what is being tested in an internal flight case is precisely the ability of the applicant to avail himself of the protection of the relevant state in some place other than that where he justifiably fears persecution.
11. As to this last question, a caveat is in my respectful view needed about the judgment of the Supreme Court of Canada in Canada (A-G) v Ward [1993] 2 SCR 688. The passage from the judgment of La Forest J cited by Stuart-Smith LJ at paragraph 16 of his judgment includes this proposition:

"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."

This may occasionally be right as a practical means of establishing whether a fear of persecution exists or is well-founded - what Stuart-Smith LJ calls the fear test; but it is not the test which the Convention lays down in relation to protection. The latter has to do not with whether the state can provide protection to the claimant but with whether the claimant can avail himself of it. In some cases this will not matter: it will be possible to take the approach of Hale LJ and deal with the state's capacity to afford protection where it best fits the issues. But this will not be an appropriate approach where the fear of persecution which is asserted is - or is said to be - localised. For reasons set out earlier in this judgment, once an applicant reaches the United Kingdom driven by a well-founded fear of racial persecution in his home area of his home country, the remaining questions will be whether there is nevertheless a part of the home state (a) which is safe from persecution and (b) to which it would not be unduly harsh to return the asylum-seeker.

12. In my view there is a need in many asylum cases, including in particular cases such as the present, to adopt the methodical approach proposed in paragraphs 12 and 17 of the judgment of Stuart-Smith LJ in Horvath. Not to do so risks the conflation of issues, and the consequent lack of focused analysis, which occurred before the Immigration Appellate Authority in the present case. I have tried to explain in paragraph 4 above how such an approach to the present case shows it to be a protection case rather than a fear case in the sense that a well-founded fear of persecution in Sri Lanka was established, leaving protection (here in the form of internal relocation) as the live issue.
13. How then does the decision-maker go about determining whether an otherwise valid claim to asylum is negated by the applicant's ability to avail himself of his home state's protection in a different part of the state from that where the fear would still be well-founded? Although Mr Ian Lewis and Miss Lisa Giovannetti in their excellent and helpful submissions have not been far apart, they differ on whether it is appropriate to use any true standard of proof rather than simply make an appraisal.
14. Putting the arguments in my own words, they are these. Miss Giovannetti submits that the practice which has been adopted in the wake of the Tribunal's decision in Kaja [1995] Imm AR 1, whether or not it accurately reflects what the Tribunal decided, has been to apply a "real possibility" standard not only to the risk (and therefore the foundation of the fear) of persecution but to the evidence relied on to establish it. This, she contends, makes no sense: evidence may vary in its force from slight to potent; but none save the plainly unreliable should be discarded; and from the rest, without setting any cut-off point, the decision-maker should answer the Convention question or questions which require an answer, which will be in each case an evaluative and not a factual answer. She commends to us the decision of the Federal Court of Australia (Sackville, North and Kenny JJ) in Minister for Immigration and Multicultural Affairs v Rajalingam [1999] FCA 719, to which I will return. Mr Lewis, understandably anxious not to forfeit the advantages which this approach may offer to at least some asylum-seekers, nevertheless contends that the Sivakumaran standard of proof of risk logically flows back into the proof of facts evidencing the risk, so that to prove such facts to a modest standard of likelihood is enough, given the special role and purpose of the Convention, to prove that the fear of persecution is well-founded. It is one thing to apply the civil standard of proof which artificially elevates factual probabilities to forensic certainties; it is another to treat past facts which probably did not happen as equally certain. But the alternative, preferred by the dissenting member of the Tribunal in Kaja, of elevating the standard of proof of past facts in asylum cases to the civil standard is not contended for by the Secretary of State. It would require further consideration only if we accepted Mr Lewis's argument that a prescribed standard of proof was requisite, and for reasons fully developed and explained by Brooke LJ, I do not accept it.
14. Without analysing the arguments as fully as they deserve, I will give my conclusions. I can summarise them, however, by saying that I agree with the entirety of Brooke LJ's reasoning on this question. Nothing which follows should be taken as qualifying it, much less as differing from it.
15. The issues for a decision-maker under the Convention (whether the decision-maker is a Home Office official, a special adjudicator or the Immigration Appeal Tribunal) are questions not of hard fact but of evaluation: does the applicant have a well-founded fear of persecution for a Convention reason? is that why he is here? if so, is he nevertheless able to find safety elsewhere in his home country? Into all of these, of course, a mass of factual questions enters: what has happened to the applicant? what happens to others like him or her? is the situation the same as when he or she fled? are there safer parts of the country? is it feasible for the applicant to live there? Inseparable from these are questions of evaluation: did what happened to the applicant amount to persecution? if so, what was the reason for it? does what has been happening to others shed light on the applicant's fear? is the home situation now better or worse? how safe are the safer places? is it unduly harsh to expect this applicant to survive in a new and strange place? What matters throughout is that the applicant's autobiographical account is only part of the picture. People who have not yet suffered actual persecution (one thinks of many Jews who fled Nazi Germany just in time) may have a very well-founded fear of persecution should they remain. People who have suffered appalling persecution may for one reason or another not come within the protection of the Convention.
16. The civil standard of proof, which treats anything which probably happened as having definitely happened, is part of a pragmatic legal fiction. It has no logical bearing on the assessment of the likelihood of future events or (by parity of reasoning) the quality of past ones. It is true that in general legal process partitions its material so as to segregate past events and apply the civil standard of proof to them: so that liability for negligence will depend on a probabilistic conclusion as to what happened. But this is by no means the whole process of reasoning. In a negligence case, for example, the question will arise whether what happened was reasonably foreseeable. There is no rational means of determining this on a balance of probabilities: the court will consider the evidence, including its findings as to past facts, and answer the question as posed. More importantly, and more relevantly, a civil judge will not make a discrete assessment of the probable veracity of each item of the evidence: he or she will reach a conclusion on the probable factuality of an alleged event by evaluating all the evidence about it for what it is worth. Some will be so unreliable as to be worthless; some will amount to no more than straws in the wind; some will be indicative but not, by itself, probative; some may be compelling but contra-indicated by other evidence. It is only at the end-point that, for want of a better yardstick, a probabilistic test is applied. Similarly a jury trying a criminal case may be told by the trial judge that in deciding whether they are sure of the defendant's guilt they do not have to discard every piece of evidence which they are not individually sure is true: they should of course discard anything they think suspect and anything which in law must be disregarded, but for the rest each element of the evidence should be given the weight and prominence they think right and the final question answered in the light of all of it. So it is fallacious to think of probability (or certainty) as a uniform criterion of fact-finding in our courts: it is no more than the final touchstone, appropriate to the nature of the issue, for testing a body of evidence of often diverse cogency.
17. The Australian Federal Court put the issues well in Rajalingam [1999] FCA 719. It pointed out - not for the first time - that a decision on asylum is an administrative process differing in important ways from civil litigation (see paragraph 36). It follows that an appeal which tracks the original issues will have largely the same character. In addition to the valuable passages from the leading judgment of the High Court of Australia in Wu Shan Liang which Brooke LJ has cited, the Federal Court considered the assenting views in that case of Kirby J. These too I find valuable:

"25. First, it is not erroneous for a decision-maker, presented with a large amount of material, to reach conclusions as to which of the facts (if any) had been established and which had not. An over-nice approach to the standard of proof to be applied here is undesirable. It betrays a misunderstanding of the way administrative decisions are usually made. It is more apt to a court of law conducting a trial than to the proper performance of the functions of an administrator, even if the delegate of the Minister and even if conducting a secondary determination. It is not an error of law for such a decision-maker to test the material provided by the criterion of what is considered to be objectively shown, so long as, in the end, he or she performs the function of speculation about the "real chance" of persecution required by Chan.

26. Secondly, the decision-maker must not, by a process of factual findings on particular elements of the material which is provided, foreclose reasonable speculation upon the chances of persecution emerging from a consideration of the whole of the material. Evaluation of chance, as required by Chan cannot be reduced to scientific precision. That is why it is necessary, notwithstanding particular findings, for the decision-maker in the end to return to the question: "What if I am wrong?" [Guo v Minister for Immigration (1996) 135 ALR 421, 441]. Otherwise, by eliminating facts on the way to the final conclusion, based upon what seems "likely" or "entitled to greater weight", the decision-maker may be left with nothing upon which to conduct the speculation necessary to the evaluation of the facts taken as a whole, in so far as they are said to give rise to a "real chance" of persecution."

(It needs to be noted that Australian jurisprudence on the Convention uses "speculation" to describe a legitimate exercise falling short of fact-finding.) The Federal Court considered this passage in Rajalingam (paragraphs 47 to 50), noting that it extended the broad evaluative approach even to the ascertainment of past facts. It adopted an exegesis propounded by Drummond J in Thanh Phat Ma v Billings (1996) 71 FCR 431:

"...unless the decision-maker can dismiss as unfounded factual assertions made by the applicant, the decision-maker should be alert to the importance of considering whether the accumulation of circumstances, each of which possesses some probative cogency, is enough to show, as a matter of speculation, a real chance of persecution, even though no one circumstance, considered by itself, is sufficient to raise that prospect."

Kirby J concluded (paragraph 31):

"Ultimately the question is whether the delegate [i.e. the decision-maker] allowed her mind to consider all the relevant possibilities by looking back at the entirety of the material placed before her and considering it against a test of what the "real", as distinct from fanciful, "chances" would bring if the applicant were returned to China."

Subsequently, in Epeabaka [1999] FCA 1 the Federal Court of Australia has returned to Kirby J's central reasoning and has adopted it.

18. Like Brooke LJ I find the Australian cases of the greatest assistance. I would put my own view, in summary, as follows. The question whether an applicant for asylum is within the protection of 1951 Convention is not a head-to-head litigation issue. Testing a claim ordinarily involves no choice between two conflicting accounts but an evaluation of the intrinsic and extrinsic credibility, and ultimately the significance, of the applicant's case. It is conducted initially by a departmental officer and then, if challenged, by one or more tribunals which, though empowered by statute and bound to observe the principles of justice, are not courts of law. Their role is best regarded as an extension of the initial decision-making process: see Simon Brown LJ in Ravichandran [1996] Imm AR 97, 112. Such decision-makers, on classic principles of public law, are required to take everything material into account. Their sources of information will frequently go well beyond the testimony of the applicant and include in-country reports, expert testimony and - sometimes - specialised knowledge of their own (which must of course be disclosed). No probabilistic cut-off operates here: everything capable of having a bearing has to be given the weight, great or little, due to it. What the decision-makers ultimately make of the material is a matter for their own conscientious judgment, so long as the procedure by which they approach and entertain it is lawful and fair and provided their decision logically addresses the Convention issues. Finally, and importantly, the Convention issues from first to last are evaluative, not factual. The facts, so far as they can be established, are signposts on the road to a conclusion on the issues; they are not themselves conclusions. How far this process truly differs from civil or criminal litigation need not detain us now.
19. It would be pointless, for the rest, to traverse ground so well covered by Brooke LJ. It is, however, worth observing (or at least hoping) that the approach which we consider to be the correct one bodies out what Simon Brown LJ said in Ravichandran [1996] Imm AR 97, 109:

"In my judgment the issue whether a person or group of people have a "well-founded fear ... 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."

While, for reasons considered earlier, it may well be necessary to approach the Convention questions themselves in discrete order, how they are approached and evaluated should henceforward be regarded not as an assault course on which hurdles of varying heights are encountered by the asylum seeker with the decision-maker acting as umpire, nor as a forum in which the improbable is magically endowed with the status of certainty, but as a unitary process of evaluation of evidential material of many kinds and qualities against the Convention's criteria of eligibility for asylum.

20. It follows that on the critical issue of internal relocation in the present case, no question of the burden or standard of proof arises. The question is simply whether, taking all relevant matters into account, it would be unduly harsh to return the applicant to Colombo.
Order: Appeal allowed with costs; legal aid taxation for the appellant; case to be remitted to a differently composed tribunal of the Immigration Appeal Tribunal
(Order does not form part of approved judgment).