Secretary of State for the Home Department v Limbuela & Ors

These conjoined appeals raise important issues about the application of s.55(5) of the Nationality, Immigration and Asylum Act 2002 ("the Act of 2002") and Article 3 of the European Convention on Human Rights and Fundamental Freedoms ("ECHR"). Permission to appeal was in all three cases granted to the Secretary of State by the judge below. Pursuant to those grants of permission the appeals are brought against the judgment of Collins J in Limbuela given on 4 February 2004, that of Gibbs J in Tesema given on 16 February 2004, and Charles J in Adam given on 17 February 2004.

In each case the judge in the Administrative Court granted relief by way of judicial review of the Secretary of State’s decision not to provide support for the claimant asylum-seeker. The three claimants are of course the respondents before us. We have been asked to give general guidance as to the operation of s.55(5) and Article 3: there has been some divergence of view among the judges of the Administrative Court as to the correct approach to be applied, notwithstanding earlier learning in this court in R(Q) & ors v Secretary of State [2003] 3 WLR 365 ("Q") and R(T) v Secretary of State [2004] 7 CCLR 53. In addition to these authorities, and the three judgments under appeal, it will be necessary to pay attention to the decision of Newman J in Zardasht [2004] EWHC Admin 91, given on 23 January 2004. At the time of the hearing of these appeals in this court on 23 and 24 March 2004, we were told that no fewer than 666 further cases awaited disposal which are likely to be affected by our judgment.

It is convenient to set out the material statutory provisions before coming to the facts of the three cases. S.95 of the Immigration and Asylum Act 1999 ("the 1999 Act") provides in part:

S.55 of the Act of 2002 provides in part:

It will at once be evident that an asylum-seeker who is barred from support by the application in his case of s.55(1), for whom the bar is not lifted by the application of s.55(5), will be on the streets with nothing unless he has resources of his own or can get access to some form of support from other individuals or non-State groups or agencies. Article 3 ECHR, as is well known, provides:

In light of these statutory measures one formulation of the critical issue in these appeals might be expressed thus: how grave must the facts of any individual case be to require the Secretary of State to apply s.55(5)?

I think it useful in the context of this recital of the relevant statutory materials to set out this description of the purpose of s.55 given at paragraph 26 of the judgment of this court, delivered by the Master of the Rolls, in Q (to which I shall have to return):

I will first describe the facts in the individual appeals. For this purpose I have largely taken the narratives which follow from the summaries given in the skeleton arguments provided in each case on behalf of the Secretary of State. The summaries’ accuracy is not I think disputed, but it is said for the respondents that they are too sparse, and if a fair view is to be taken of the facts they need to be filled out with rather more detail. I go a little way, but not very far, with that. The narrative I have given draws here and there on material in the respondents’ skeleton arguments, on the account of the facts given in the court below, and on the primary documents. My aim has been to describe the facts to the extent necessary to determine the issues in the appeals and no further. I should notice that in each of the three cases the Secretary of State did not believe the respondent’s account of how and when he arrived in the United Kingdom.

However there are other more general factual issues of which I must give some account. The charity Shelter put in a skeleton argument with permission earlier granted by myself. They also submitted substantial written evidence. For that they had no permission; and it is very important that interveners, who take part in proceedings at the court’s discretion and not by right, should strictly abide by the terms on which they are allowed to participate. That said, I acknowledge at once that much of the factual material submitted by Shelter has proved useful to the court, and I accept without hesitation the explanation and apology offered by Mr Knafler, counsel for Shelter, in his careful and courteous letter of 25 March 2004. At the hearing Mr Knafler addressed the court on behalf of Shelter to the extent that he was allowed to do so. I have to say that his skeleton is highly rhetorical, and that is by no means conducive to this court’s better performance of its task. But I must certainly consider the essence of the factual material which Shelter has provided.

The evidence also includes statements put in on behalf of the Secretary of State by Michael Sullivan, a caseworker with NASS. He produces a list of day centres for homeless people in London, and in his second statement confronts criticisms which had been levelled by Sophia Linehan, Adam’s solicitor, at what he had first asserted. I need also to consider this material. Lastly there is a document titled "Destitution by Design" issued in February 2004 by the office of the Mayor of London, which contains passages relied on by the Secretary of State as showing the extent to which asylum-seekers without State support have been able to look to resources to be found in community-based organisations.

I turn first to the facts of the individual cases.

Wayoka Limbuela is a national of Angola born on 20 April 1980. He maintains that he arrived in the United Kingdom on 6 May 2003. He claimed asylum on 7 May 2003, and was provided with NASS accommodation in Margate. On 16 May 2003 the Secretary of State decided that Limbuela had not made his claim for asylum as soon as reasonably practicable so that by force of s.55(1) of the Act of 2002 he was barred from support or further support under s.95 of the 1999 Act, subject to s.55(5). That decision, and decisions made to the same effect in the cases of Tesema and Adam, is not or is no longer the subject of any challenge.

However the Secretary of State concluded also that in the circumstances of the case s.55(5) did not avail Limbuela. At length, on 22 July 2003, he was evicted from the NASS accommodation where he had been placed. After that he spent two nights sleeping rough outside Croydon Police Station. In that time he claims (and there is nothing, I think, to contradict it) to have had no money and no access to food or washing facilities. He asked the police for a blanket, but that was not provided; he begged passers-by for food, but was not given anything. Then from 24 July 2003 he was able to stay for four nights at the Lord Clyde night shelter, where he was also provided with food. On 28 July he was asked to leave the night shelter and advised to contact a solicitor. So he did, and on the same day solicitors instructed by him wrote to NASS stating that he faced violations of Articles 3 and 8 because support was not being provided to him. There was no prompt reply to this letter, but the Secretary of State’s effective continuing refusal to accept that this is a s.55(5) case and accordingly provide support is the subject of the judicial review challenge.

Also on 28 July 2003 application was made on Limbuela’s behalf to Eady J who granted an interim injunction against the Secretary of State pursuant to which Limbuela has since been housed and fed. No material was put before the Secretary of State or the judge on 28 July to indicate any medical problems. At length on 29 October 2003 (after judicial review permission had been granted by Jackson J) the Treasury Solicitor sought specific information from Limbuela’s advisers in relation to what may be called his s.55(5) claim. Eventually, on 5 January 2004, Limbuela made a witness statement in which among other things he said he had suffered from stomach pains. Later a general practitioner’s letter was produced (two days before the hearing of the judicial review). The doctor said that Limbuela had been to his surgery three times since August 2003. His complaints had been, variously, constipation, a cough, pain in the lower abdomen and testicles, dizziness and heartburn. He had been prescribed appropriate medication.

The Secretary of State rejected Limbuela’s asylum claim on 10 June 2003. The Adjudicator dismissed his appeal on 1 September 2003. We were told at the hearing that on 24 February 2004 (thus after Collins J’s judgment) the Immigration Appeal ("the IAT") gave leave to appeal and remitted the case for a re-hearing before a different Adjudicator. The re-hearing is still awaited.

Binyam Tefera Tesema was born on 22 July 1977 and is a national of Ethiopia. He claims to have arrived in the United Kingdom on 13 August 2003 and to have stayed that night in a hotel arranged by the agent with whom he had apparently travelled. By the next morning, however, the agent had disappeared. On the same day, that is to say 14 August 2003, he made a claim for asylum and was thereafter housed in NASS emergency accommodation.

He was interviewed on 14 August 2003. That was what is called the "Level 1" interview. The pro forma for Level 1 is essentially restricted to questions concerning the claimant’s identity, status and travel route, though it provides for some questions relating to health. In response to those Tesema complained of earache, backache, and pain in the left knee, but said he had not seen a doctor for three years. He was interviewed again on 17 August 2003. This was the "Level 2" interview, which deals with the claimant’s identity, status and travel route. On that occasion he said he was in good general health. Difficulties he suffered with his ears, back and feet were, he said, due to his having been beaten.

On 20 August 2003 the Secretary of State concluded that Tesema had not claimed asylum as soon as reasonably practicable within the meaning of s.55(1), and that there was nothing to justify his taking action under s.55(5). Tesema’s solicitors wrote to the Secretary of State on 1 September 2003 to call into question both decisions. As I have indicated only the s.55(5) challenge is live. On 2 September Tesema was evicted from his NASS accommodation but on the same day upon application to Henriques J obtained an order for interim support. So he has not in fact had to sleep rough at any time.

Jackson J granted judicial review permission on 27 October 2003. Thereafter the Treasury Solicitor, seeing that Tesema had complained of various medical problems in his first witness statement, asked his solicitors to provide evidence from a doctor as to the nature of his medical condition and any treatment he was receiving. At length on 21 January 2004 the solicitors sent a report dated 1 January 2004 from a psychiatrist, Dr Steadman. The report summarised Tesema’s allegations of what had been done to him in his home country. It contained also an account of certain physical ailments – some loss of hearing, intermittent pain in the back and above the left knee – which Tesema described to the doctor. Then as Gibbs J was to put it (paragraph 30):

Some further medical material was submitted a little later in the shape of a letter dated 13 January 2004 from the Lambeth Primary Care Trust. However Gibbs J concluded (paragraph 34) that it added little to Dr Steadman’s assessment.

In light of all the medical evidence he had received since the grant of judicial review permission the Secretary of State issued a further decision, dated 5 February 2004, by which he concluded that Tesema had not provided evidence of a medical condition such as would engage s.55(5). After that further evidence was put in to show that Tesema had made two approaches for support and accommodation, one to the Ethiopian Community Centre and the other to the Eritrean Community Centre in Haringey, but neither was able to provide for him.

The Secretary of State rejected Tesema’s asylum claim on 20 August 2003, by the same letter that conveyed his decisions under s.55(1) and s.55(5). His appeal to the Adjudicator was allowed on 14 January 2004. On 11 March 2004 the IAT granted leave to the Secretary of State to appeal. The appeal has not yet been heard.

Yusif Adam was born on 25 December 1975. He claims to be a national of Sudan. He was to say that he left Port Sudan on 22 September 2003 by cargo ship arriving in the United Kingdom on 15 October. He claimed asylum on 16 October 2003. On the same day the Secretary of State reached decisions adverse to him under ss.55(1) and 55(5) of the Act of 2002. His substantive asylum claim has still not yet been determined.

At his Level 1 interview on 16 October 2003 Adam stated that his health was "generally fine", he was not suffering from any medical condition, and he had last seen a doctor about two years previously. He is or claims to be illiterate. From 16 October until 10 November 2003, when Ouseley J granted judicial review permission and interim relief, he was on the streets. On his own account (in two witness statements of 28 October and 4 November 2003) he slept outside the Refugee Council in Brixton. During the day he had access to the Refugee Council premises where he could wash himself and his clothes, get tea and coffee in the morning, sometimes an evening meal, and a hot meal at 1 pm.

Adam’s judicial review claim form was issued on 7 November 2003. Attached to it was a doctor’s note dated 27 October 2003 stating that he was suffering from haemorrhoids, back pain and gastritis, and giving details of the medicines which had been prescribed. After he was provided with accommodation following Ouseley J’s order for interim relief a report dated 10 December 2003 from Dr Michael Peel of the Medical Foundation was served on the Treasury Solicitor. Given a particular criticism levelled at Charles J’s reasoning by Mr Giffin QC for the Secretary of State (to which I will come in due course) it is convenient to cite this passage from Dr Peel’s report:

The background to what Shelter have to say rests in the undisputed fact that asylum-seekers whose claims are not yet determined are prohibited from obtaining employment and at the same time barred (by force of s.115 of the 1999 Act) from State benefits; though this latter circumstance is mitigated by s.95 of the 1999 Act in cases where that provision applies.

Shelter refer to what they describe as "generic" evidence which they have deployed in the courts before. It was served on the Secretary of State in T (material contained in certain statements has since been updated) and in another case, Kumetah. It has been available to many other claimants. In summary, Shelter roundly assert that there is no realistic prospect of a destitute asylum-seeker obtaining accommodation through a charity. Unless he has family or friends able to provide him with accommodation or substantial funds, he will have to sleep rough. So far as charitable provision might be available, it does not stretch to things like umbrellas, plastic sheets or mats (let alone sleeping bags, warm and waterproof cagouls, tents and the like). In London (and these appeals are all, so to speak, London cases) charity food, drink and washing facilities are in limited supply. The asylum-seeker’s access to them depends on where he is sleeping rough, and whether he is (a) able to ascertain that such facilities exist, where they are and when they are open, (b) able to walk what will generally be substantial distances to find them, (c) able to find them when he gets to the relevant area, and (d) lucky enough to arrive when there is still some available provision. On all these points there is a good deal of detail before us, which I have of course considered. I will not set out all of it. I hope it goes without saying that I do not thereby intend to diminish its obvious force. I will highlight some specific points.

Mr Adam Sampson, the Director of Shelter, has made a very full statement, dated 17 March 2004. He says there are only two free hostels in London. One, in W9, is for women only and has a capacity of 15. It is said to be full every night. The other, for men who must be at least 30 years of age, is in SE1 and has a capacity of 36. No vacancies were recorded over a period of two months monitored by Shelter between November 2003 and January 2004. There are five winter night shelters operating free of charge, in Croydon (10 spaces), Hackney (15 spaces), Harrow (8 spaces), Islington (12 spaces) and West London (35 spaces). They offer a camp bed or a mattress in a church hall. Mr Sampson describes various formidable difficulties of access to these facilities.

There are also three "Rolling Shelters" which provide free accommodation, operated by St Mungo’s. But Mr Sampson says that s.55 asylum-seekers (a shorthand for those against whom s.55(1) and 55(5) decisions have been made) are not regarded as "an appropriate client group" for these shelters.

Mr Hugo Tristram of the Refugee Council has also made a statement in support of Shelter’s intervention. He describes, in particular, the Refugee Council’s day centre. As its name implies it does not provide overnight accommodation. It is closed at weekends. Coffee, tea, breakfast and a hot lunch are available four days a week, and sandwiches on Wednesdays. There are limited laundry and shower facilities. Some clothing and blankets are given out.

Mr Tristram also recounts incidents of bottles and stones being thrown at asylum-seekers sleeping rough, and of the degrading condition to which they may be reduced.

In his first statement Mr Sullivan produces, as I have said, a list of some 54 day centres for homeless people in different parts of London, and a schedule giving details of each of them. He says they all offer help with such things as meals, showers, clothing and medical services, and advice on alcohol and drugs, benefits, and accommodation.

There followed, as the written evidence took its course, what amounted to something of a debate between Sophia Linehan, Adam’s solicitor, and Mr Sullivan about two matters. The first was the utility or otherwise of a facility called Hostels Online which Mr Sullivan had stated was a resource of which s.55 asylum-seekers such as Miss Linehan’s client might take advantage. The second went to the practical availability for asylum-seekers of the day centre facilities in Mr Sullivan’s list. As for the first, Miss Linehan crisply asserts (statement, 2 February 2004, paragraph 5) that none of the organisations listed on the Hostels Online website will take an asylum-seeker with no access to public funds. That is not answered by Mr Sullivan in his statement of 5 February 2004 save in the vaguest terms. As regards the second, Mr Sullivan says that thirteen of the day centres in his list state in terms that they are available to asylum-seekers. Enquiries were made of fifteen others by a representative of the Treasury Solicitor. These also offer some support by way of basic amenities which is available to asylum-seekers.

Mr Sampson of Shelter and Mr Tristram also comment on Mr Sullivan’s list of day centres. Mr Tristram had been involved in drawing up the list. Mr Sampson emphasises their considerable distance, one from another. He states (cross-referring to Mr Tristram’s evidence) that the list has proved of little value and is no longer distributed by the Refugee Council.

This document, issued as I have said by the Mayor of London’s office in February 2004, is of particular interest because it asserts facts which might be said to weaken the case that it seeks to make; at least they do not go to support it. The facts so asserted may on that account be taken to be more likely than not to be true. The document’s subtitle is: "Withdrawal of support from in-country asylum applicants: An impact assessment for London". It is heavily critical of the effects of s.55 of the Act of 2002, as exemplified in particular by the summary at the end of the document, in paragraph 8.18. At the same time it contains these following assertions, which are necessarily selective:

It must be obvious that it is not possible for this court to make full, accurate and detailed findings of fact as to the exact realities faced by s.55 asylum-seekers in London, let alone elsewhere. Such an exercise could only be satisfactorily conducted by a process of factual enquiry involving a wide-ranging examination of the evidence, with oral testimony and cross-examination. A process of that kind is inapt for determination in the course of adversarial litigation in the judicial review jurisdiction, and particularly inapt in this court. I draw attention to this circumstance in order to point up the wisdom of my Lord Carnwath LJ’s observations in the course of argument to the effect that it cannot be the court’s task to judge the daily circumstances of the plight of individual s.55 asylum-seekers so as to ascertain whether there is an actual or potential violation of Article 3 ECHR. To that I will return; but despite its force we must surely reach and describe some impression of the general or background evidence. We are at least required to articulate a kind of touchstone for the application of Article 3 in these cases.

Looking at the factual material as a whole, my impression is that a significant proportion – I cannot give a tighter description – of s.55 asylum-seekers will not get access to overnight accommodation and will therefore be on the streets. Many, however, will obtain accommodation and are doing so. So much is plain from the Mayor of London’s document. Both those who do and those who do not find accommodation will mostly, though sometimes with real difficulty, get access at least to rudimentary facilities including food during the daytime. This very broad, and (as I acknowledge) necessarily superficial view is I think consistent with the picture that emerges from the facts relating to the three individual cases. It must go without saying – I do not by that phrase mean to undervalue the point’s importance – that the plight of living on the streets will bear more or less hardly according to the impact of a whole range of factors: not least, the state of the person’s health, mental and physical, and the state of the weather.

In paragraph 6 I proposed what I described as "one formulation" of the critical issue in these appeals in these terms: how grave must the facts of any individual case be to require the Secretary of State to apply s.55(5)? The formulation is useful and important in reflecting on the facts of the individual cases and the general background facts which I have described. But the issues of law with which we are in the end confronted in these appeals cannot in my judgment be reduced to so simple a question. They engage the nature of the obligation undertaken in Article 3 by the States signatory to the ECHR, and also the proper construction of s.55 of the Act of 2002. In finding the way into these matters it is convenient to start with the two earlier decisions of this court in Q and T.

Q (Lord Phillips MR, Clarke and Sedley LJJ) concerned six test cases in which adverse decisions had been made under s.55(1) of the Act of 2002. The case’s first focus was upon the approach to be taken to the decision-making process carried out under that subsection. At first instance Collins J held that the process adopted had been unfair. This court agreed. I would refer in particular to paragraphs 37 (which contains the court’s definition of the test propounded in s.55(1) by the phrase "as soon as reasonably practicable"), 43, 81 – 92 and 97 – 100 of the judgment of the court. With respect I need not set them out.

The court also, however, gave close attention to the effect of s.55(5). They considered (paragraph 46) two issues: "(1) can failure to provide support ever constitute subjecting an asylum seeker to inhuman or degrading treatment? If yes, (2) in what circumstances will the failure constitute such treatment?" In relation to (1), as the court recorded (paragraph 52), the Attorney General submitted that failure to provide support could never of itself constitute treatment and so amount to a breach of a negative obligation imposed by ECHR Article 3. He conceded, however, that in extreme circumstances Article 3 might impose a positive obligation on the State to provide support for an asylum seeker: the plight of a heavily pregnant woman was given by way of example. There follows in the court’s judgment some discussion of the distance between positive and negative obligations in the context of Article 3. I shall have more to say about the distinction between these forms of obligation when I come later to confront the scope of Article 3 and it will be more convenient to cite the relevant passages from Q at that stage.

The court’s conclusion on question (1) in Q is given at paragraphs 56 and 57:

In confronting question (2), after observing (paragraph 60) that "it is quite impossible by a simple definition to embrace all human conditions that will engage article 3", the Master of the Rolls proceeded to cite the decision of the Strasbourg court in Pretty v UK 35 EHRR 1, paragraph 52. Pretty is with respect an important case and I shall have to return to it. It is convenient however at this stage to set out paragraph 52, given its part in this court’s reasoning in Q:

In Q this court continued:

T (Kennedy, Peter Gibson and Sedley LJJ) was decided on 23 September 2003, six months after Q. The appeal was brought by the Secretary of State from a judgment given by Maurice Kay J, as he then was, in the Administrative Court ([2003] EWHC Admin 1941). There had been three claimants before Maurice Kay J: S, D and T. However this court was only concerned with the judge’s findings as to the effect of s.55(5) in T’s case. Kennedy LJ gave the judgment of the court.

Kennedy LJ referred (paragraph 9) to the finding in Q that the s.55(1) regime constitutes ‘treatment’ within the meaning of Article 3. As regards the issue of severity of treatment the court proceeded (paragraph 13) on the footing, effectively agreed between counsel, that it was for the claimant to establish that his right to relief pursuant to Article 3 was "clear". But that, of course, did not speak as to the level of suffering or potential suffering required to be established. Counsel for the appellant Secretary of State urged among other things that a claimant had to show "that he had taken all available steps to help himself" (paragraph 14). It seems to have been accepted (again, paragraph 14) that the moment at which the circumstances of the case had to be judged was immediately before the claimant obtained emergency interim relief, as T had done on 24 April 2003.

Counsel for the respondent T – in fact Mr Knafler – submitted (paragraph 15) "that any asylum-seeker who is homeless and without means is verging on the condition described in Pretty". The court said (paragraph 10):

In the event the court allowed the Secretary of State’s appeal, holding (paragraph 19) that it was

Following two substantial judgments (if I may use the term) in this court dealing with s.55(5) of the Act of 2002, one might have expected to find a consistency of approach among later decisions of the Administrative Court in the same area. But that is not what has happened. I make it clear I do not criticise the judges below on that account. I think the relevant principles in this area are more than usually elusive. I will deal with the cases before attempting any analysis. It is convenient to do so in chronological order.

As I have indicated the judgment in this case is not the subject of appeal to this court. It was decided by Newman J on 23 January 2004 (four months after this court’s judgment in T). The claimant was a 20-year old Iraqi. Initially he received support but that was withdrawn following, as I understand it, an adverse decision under s.55(1) of the Act of 2002. He had no specific health problems. He was not given support under s.55(5) and spent fourteen days sleeping rough. Newman J dismissed his application for judicial review, holding (paragraph 39) that it was not clear that "the high threshold laid down by Pretty… ha[d] been achieved in this case".

Newman J was of course referred to Q and T. In paragraph 5 of his judgment he listed the main practical detriments facing asylum-seekers arriving here. They must be especially applicable to persons denied support: having no home or income, being a stranger to the language and the people, loneliness, anxiety, vulnerability. Newman J recognised (paragraph 6) that special circumstances such as ill health might exacerbate the burdens imposed by these conditions. Then he said this:

Newman J went on to place further emphasis (paragraphs 16 – 17) upon the need as he saw it for claimants to provide evidence of particular circumstances which might qualify them for s.55(5) support.

At paragraph 31 Collins J cited the judgment of Maurice Kay J at first instance in T, including this passage:

Collins J proceeded to refer to the judgment of Newman J in Zardasht. He accepted (paragraph 37) that a claimant should put before the court evidence about "what steps he has taken to try and get support, how he has fared and what effect it has had on him if he has had to sleep rough or beg or whatever". But he by no means accepted the general position adopted by Newman J. He said (paragraph 42) that he had "not approached the matter in quite the same way as Newman J". But that considerably understates the difference between them. Collins J took the view that in the case of any asylum-seeker who "shows that he has taken reasonable steps and that no assistance is available except by begging and hoping, then the fact that he will have to sleep rough, he has no money, he has no proper access to food or other facilities, will be likely to suffice to establish his case" (paragraph 37). And I should set out paragraphs 38 and 41:

I think it a fair reading of the essential reasoning in this judgment that once an asylum-seeker who is denied NASS support under s.55(1) has shown that he has tried but failed to find accommodation and other support, so that he will have to sleep on the streets, he will have established an imminent breach of ECHR Article 3 and a corresponding entitlement to be supported under s.55(5).

I can with respect deal with these two judgments rather more shortly. The position is that broadly speaking Gibbs J adopted the approach taken by Collins J, and Charles J that taken by Newman J. (I assume that Charles J did not see Gibbs J’s judgment, which was given the day before his own.) I need only cite these passages from Gibbs J in Tesema:

In Adam Charles J stated at paragraph 55 (see also paragraph 68) that subject to certain points about the availability of charitable support (as to which the evidence had moved on) he agreed with the judgment of Newman J in Zardasht. He considered (paragraph 58) that Collins J’s reasoning in Limbuela bore no more than "fine distinctions" from the "real risk approach" which this court had expressly rejected in Q, and that (paragraph 59) Collins J’s description of the Pretty threshold did not accord with earlier authority.

I fear that the state of the law in this area has got into serious difficulty. So much is spoken to by the fact that judges of the Administrative Court have felt driven to take starkly contrasting positions as to the right test for the engagement of s.55(5) of the Act of 2002 notwithstanding the attention given to the subsection in two previous decisions of this court. I think the difficulty has principally arisen from two circumstances. First, it has not been found possible to articulate the ingredients of a violation of ECHR Article 3 in the case of a person put on the streets without support any more precisely than by reference to what was said by the Strasbourg court at paragraph 52 in Pretty: see by way of example paragraph 10 in T. Secondly, the prospect that the bare fact of being on the streets will not of itself engage Article 3 seemingly faces the executive, and (upon a challenge being advanced) the courts, with the "wait and see" approach excoriated by Collins J and Gibbs J.

The first of these circumstances breeds an inevitable uncertainty as to who is and who is not a proper candidate for s.55(5), and leaves that question to the factual arbitrament of the courts case by case; although this is in contrast to their proper role in judicial review, which is to hold public decision-makers to account for errors of law. The second circumstance has fuelled the division between first instance judges, and persuaded some to adopt an approach, namely to accord the benefit of s.55(5) once it is demonstrated that the claimant will otherwise go on the streets, which in my judgment (I deal with this below) eviscerates the subsection and misunderstands Article 3. We are left with a state of affairs in which our public law courts are driven to make decisions whose dependence on legal principle is at best fragile, leaving uncomfortable scope for the social and moral preconceptions of the individual judge (I mean no offence to the distinguished judges who have heard these cases); and law and fact are undistinguished. We need to see whether there is room for a sharper, more closely defined approach.

The starting-point for such an exercise must, I think, be to re-visit the nature and extent of our obligations under Article 3. Here, the first thing to emphasise is a distinction which is well recognised in the jurisprudence, but whose importance has perhaps not always been fully unravelled. It is between (a) breaches of Article 3 which consist in violence by State servants, and (b) breaches which consist in acts or omissions by the State which expose the claimant to suffering inflicted by third parties or by circumstance. Given the historic crucible of the European Convention’s beginnings it must be plain that category (a) is the paradigm case. It expresses a fundamental value, the abhorrence of State violence. It is qualified only by strictly confined exceptions allowed under the general law, namely the use of reasonable force for the purpose of arrest, lawful restraint, self-defence and the prevention of crime. The common law requires that reliance on any of these by a State servant – indeed by anyone – as authorising the use of force in any particular case must always be strictly justified, and no other justifications whatever are available.

But it is plain that as the jurisprudence has developed, the scope of Article 3 is not limited to category (a). Various sets of facts which can only fall within category (b) have been accepted, in Strasbourg and here, as giving rise to violations. So much is recognised in general terms in paragraph 50 of the judgment in Pretty:

I must in due course explain what seems to me to be the true importance of the distinction between (a) and (b) and its part in the resolution of these appeals. I should say first that it has links with another distinction familiar in the Article 3 jurisprudence, that between a negative obligation not to inflict inhuman or degrading treatment and a positive obligation to take steps to protect persons from forms of suffering sufficiently grave to engage Article 3. A like difference is especially clear in the jurisprudence relating to Article 2 (the right to life): see for example Osman (1998) 29 EHRR 245. The distinction between categories (a) and (b) and that between positive and negative obligations are not the same. The latter distinction is discussed by Lord Bingham in Pretty in the House of Lords ([2002] 1 AC 800). The Strasbourg court incorporated into its judgment the whole of Lord Bingham’s opinion, to which in reply Mr Giffin commended our attention. At paragraph 15 Lord Bingham had referred to Rees v UK (1986) 9 EHRR 56, in which the Strasbourg court had said (paragraph 37):

The extent to which, in securing the rights of individuals under Article 3, the court may allow an area of discretion or judgment to the Secretary of State is a matter of great importance, and I must return to it. But first there is more to say about the distinction between positive and negative obligations. I have already referred (paragraph 40) to the Attorney General’s submission in Q (paragraph 52 of the court’s judgment) that failure to provide support could never of itself amount to a breach of the negative obligation imposed by ECHR Article 3, and his contrasting concession that in extreme circumstances Article 3 might impose a positive obligation on the State to provide support for an asylum seeker. The court in Q proceeded to cite paragraphs 49 – 51 of the judgment in Pretty. I have already set out paragraph 50. I need only cite this extract from paragraph 51:

I shall come later to O’Rourke, which was much prayed in aid by Mr Giffin. D, as the court in Q stated, was a negative obligation case. But it fell within category (b), not category (a). As I have said the two distinctions are not the same: negative obligations are not to be equated with (a), nor positive obligations with (b). It is useful to describe D at this stage. D was a drugs dealer from St Kitts with a bad criminal record and an immigration history to match. He suffered from AIDS. The British authorities proposed to remove him to St Kitts. At length he complained to the Strasbourg court for breach of Article 3. By the time the court considered the case D was close to death, and there was nothing to show that in St Kitts (where the population was beset with health and sanitation problems) he would receive any moral or social support nor even that he would be guaranteed a bed in either of the hospitals on the island which the UK government had stated cared for AIDS patients. The court regarded these facts as amounting to very exceptional circumstances (paragraphs 53, 54) and held that the implementation of the decision to remove the applicant would be a violation of Article 3.

Some instances of category (b) are much closer to category (a) than others. These are the cases where the State owes a duty – a positive obligation – to protect individuals from violence by persons other than its own officials. Thus prisoners, who are in the care of the State, are entitled not only to be kept safe from assault by State servants such as the prison officers, but also to be protected from being attacked by fellow prisoners. (The decision in Keenan (2001) 33 EHRR repays attention: the complaint there was of a violation of Article 2 by failure to protect a prison inmate from suicide or self-harm.) There will of course be other examples of category (b) cases which arise out of the need for an armoury against unlawful violence. Patients in hospitals (especially hospitals or hospital wards for the care of the mentally ill) and children (and nowadays, I fear, teachers also) in difficult or failing schools come readily to mind as vulnerable instances. And while it is a negative obligation case, not strictly an example of protection against assault, the decision in Soering v United Kingdom (1989) 11 EHRR 439 may perhaps be assimilated with these examples – the sub-class of category (b) that is concerned with the prevention of violence. As is well known the Strasbourg court held in Soering that Article 3 would be violated were the United Kingdom to comply with a request to extradite the applicant, a German national, to the United States to face charges of murder in Virginia because he would be held on "death row" facing the death penalty. There would be a "a real risk" that the applicant would be exposed to inhuman or degrading treatment or punishment in the receiving State. Further, it is now uncontroversial that the United Kingdom may be in breach of its Article 3 obligations if it returns a would-be immigrant to a State where he would suffer violence sufficiently grave to engage the Article. The immigrant enjoys statutory appeal rights which enable him to raise both ECHR and asylum grounds of appeal against a decision to remove him.

I have said that State violence – category (a) – is the paradigm case of violation of Article 3. But the category is not unitary. It contains three different kinds of case. The first is where the violence in question is actually authorised by the State. That is by far the biggest beast which Article 3 is there to slay. The second is where a State servant, acting on the face of it in the course of his official duty, assaults another but has no colour of sanction or permission from his superiors to do so. The third is where the State servant has authority to use force (e.g. to arrest a criminal) but exceeds what is reasonable. Both negative and positive obligations are in play in these cases. The State and its servants must absolutely refrain from violence which is not justified on the narrow basis to which I have referred. So there is an overall negative obligation. The State also is responsible, in the second and third of these three instances, to take positive steps to prevent its servants from so acting and to punish them as appropriate if they do. These considerations, coupled with the fact that the category (b) cases concerned with protection against violence by non-State agents are as I have said much closer to category (a) than others, tend to demonstrate that the distinction between negative and positive obligations is in principle of lesser importance than that between categories (a) and (b), and, it might be thought, less important than the further distinction between cases concerned with protection against violence (whether (a) or (b)) and other cases in category (b). I shall have more to say about this last distinction below.

It is plain that the State has no lawful discretion or power whatsoever to authorise violence – the first category (a) instance – beyond cases of the use of reasonable force for the purpose of arrest, lawful restraint, self-defence and the prevention of crime. And those cases, indeed, involve no State discretion properly so called, but are given by the law which is the sole arbiter of such justifications. Here Article 3 is in truth absolute. But the other two category (a) instances, before one ever gets to category (b), begin to allow some space for judgment and discretion on the part of government. In both, such issues as the nature and intensity of any training which State servants ought to receive, and the rigour or severity with which defaulters should be pursued and punished, will admit of a range of lawful options. The range allowed by the court is likely to be very narrow, given the gravity of failure, and it will certainly be unlawful for the State to do nothing. The position is broadly the same in relation to those category (b) cases which I have described as lying closer to category (a) than others: those concerned with the duty of protection from violence by non-State actors. The State cannot fail to take steps to protect prisoners from their fellow inmates; but there must be an area of judgment, however narrow, as to the steps which are to be taken. The judgment belongs to executive government. Where the government’s responsibility is less immediate, such as, for example, its responsibility to the public at large, its lawful range of choice is likely to be wider: compare Keenan, paragraph 91.

Article 3 is said to be absolute; we are accustomed to hearing it so described. But the term is misleading. If it means only that the Article is not qualified by an express entitlement of proportionate interference with the right in the public interest, of the kind conferred by paragraph 2 in the case of each of the political rights guaranteed by Articles 8 – 11, well and good. So also if it means that Article 15 confers no right to derogate from Article 3, as it does not. But if it is taken to mean that the executive government in no case has any legitimate power of judgment whatever as regards the protection of individuals from suffering which is inhuman or degrading to the extent which Article 3 contemplates, then it is false. Article 3 indeed imposes a truly absolute requirement in this strict sense in the worst instance of category (a): State sponsored violence. In other cases within category (a) the negative obligation, prohibiting the State servant from unlawful violence, is absolute. The State’s positive obligation of training, investigation and sanction is not.

It is not enough to claim category (a), especially its worst instance, as a special case simply because any tolerance of State violence is barbarous, though that is plainly so. The true importance of the distinction between (a) and (b) has to be more closely articulated. I would express it thus. Whereas State violence (other than in the limited and specific cases allowed by the law, which I have described) is always unjustified, acts or omissions of the State which expose persons to suffering other than violence (at anyone’s hands), even suffering which may in some instances be as grave from the victim’s point of view as acts of violence which would breach Article 3, are not categorically unjustifiable. They may be capable of justification if they arise in the administration or execution of lawful government policy; and if they do not so arise, they are liable to be struck down by the courts on conventional judicial review grounds irrespective of the impact of ECHR. But even if the act or omission happens in the pursuance of lawful policy, Article 3 still offers protection against suffering, albeit not occasioned by violence, where the suffering is sufficiently extreme.

It will be seen that this approach tends to assimilate those category (b) cases where the duty is to protect the individual against non-State violence (as opposed to other forms of suffering) with category (a), and to proceed on the basis that tolerance or contemplation of violence from any quarter is always unlawful and can never be justified on policy grounds. In that case the distinction of real importance would not be between (a) and (b) (far less between negative and positive obligations) but between cases concerning protection against violence, State or non-State, and the rest. This is the distinction to which I referred at the end of paragraph 65. The State’s response to the risk of violence to those who come within its jurisdiction, citizen or not, will always be a significant mark of its claim to civilised government. But in practice the degree of risk will vary very widely, and arise in an infinity of circumstances. Such risks include those which happen on deportation or removal to another country. I would not adopt an approach which might suggest the existence of a rule that Article 3 imposes on the State a general over-arching duty of protection against violence in any and all circumstances.

In my judgment the legal reality may be seen as a spectrum. At one end there lies violence authorised by the State but unauthorised by law. This is the worst case of category (a) and is absolutely forbidden. In the British State, I am sure, it is not reality, only a nightmare. At the other end of the spectrum lies a decision made in the exercise of lawful policy, which however may expose the individual to a marked degree of suffering, not caused by violence but by the circumstances in which he finds himself in consequence of the decision. In that case the decision is lawful unless the degree of suffering which it inflicts (albeit indirectly) reaches so high a degree of severity that the court is bound to limit the State’s right to implement the policy on Article 3 grounds.

This figure of a spectrum seems to imply the existence of a point upon the spectrum which marks the dividing line, in terms of State acts or omissions, between what violates Article 3 and what does not. There is such a point, but it does not, I fear, provide a brightline rule by which the court may readily determine whether any particular set of facts falls on this or that side of the line. The point is at the place between cases where government action is justified notwithstanding the individual’s suffering, and cases where it is not. Various factors will determine where this place is to be found. They will include the severity of the threatened suffering, its origin in violence or otherwise, and the nature of the government’s reasons or purpose in acting as it does. This last consideration illustrates an important point, material to what is or is not to be regarded as "inhuman or degrading" treatment. Anyone suffering unlawful violence is degraded by it, and if it is meted out at the hands of the State, it is all the more degrading: the State, which should look after the citizen, treats him instead with contempt. It is what vile tyrannies do. But a person is not degraded in that particular, telling sense, if his misfortune is no more – and of course, no less – than to be exposed to suffering (not violence) by the application of legitimate government policy. I do not mean to sideline such a person’s hardships, which may be very great. I say only that there is a qualitative difference, important for the reach of Article 3, between such a case and one where the State, by the application of unlawful violence, treats an individual as a thing and not a person.

I have said (paragraph 68) that even where act or omission by the State is done in pursuance of lawful policy, Article 3 still offers protection against suffering, albeit not occasioned by violence, where the suffering is sufficiently extreme, and so it certainly does. What is "sufficiently extreme"? Pretty (paragraph 52) is clearly very much in point. So is the consideration I have just addressed, that in such a case the sufferer is not degraded in the particular, disgusting sense in which a person brutalised by violence (especially State violence) is degraded. But there are two other points, closely connected. First, we ought to recognise a further reason to acknowledge a distance between the case where a person is exposed to hardship through circumstance, because the State declines (in pursuit of a proper policy) to give him food and shelter, and the case of State violence, and kindred cases of violence: it is surely this second category which, primarily at least, the drafters of the Convention must by Article 3 have intended to outlaw. Secondly, we should also recognise and respect the claim of the democratic arm of government to exercise and fulfil its powers and duties in matters which lie within its particular responsibility. Plainly these include the management of immigration control and so of asylum claims.

On these two points I would refer, if I may, to a judgment given by myself in N v Secretary of State [2004] UKHRR 49, [2003] EWCA Civ 1369. The case concerned a Ugandan woman who after her arrival in the United Kingdom was diagnosed HIV-positive. The Secretary of State, having rejected her asylum claim (to which decision there was no challenge), proposed to remove her to Uganda. She claimed that her removal would constitute a violation of Article 3 because the facilities for her care and treatment in Uganda fell far short of those available to her here. I said:

In looking at the question what degree of suffering is in the present context "sufficiently extreme" to engage Article 3 I should at this stage recall this court’s finding in Q (paragraph 56) to the effect that the regime imposed on an asylum seeker denied support by reason of s.55(1) of the Act of 2002 constitutes "treatment" within the meaning of Article 3. With great respect I agree with this reasoning by which we are in any case bound. But there remains the question, in what kind of case should such "treatment" be held to be inhuman or degrading. In seeking to offer an answer, I would not use quite the same language as I thought appropriate in N. Because of the operation of s.55(1) the State is in part the author of the asylum seeker’s misfortunes; their other author is the very fact of his status as an asylum seeker who cannot be removed before his case is decided. At the same time there is nothing in the Convention jurisprudence to show or even suggest that the State is a compulsory guarantor of any minimum level of living standards.

It is to my mind clear from this court’s approach in Q and T that the Secretary of State will not violate Article 3 by declining to act under s.55(5) so that a claimant in whose case there are no special considerations, such as age, infirmity, or any other special vulnerability in the applicant’s circumstances, is put on the streets. And this is consistent with the Strasbourg court’s decision in O’Rourke, to which I have already referred in passing. In that case the applicant was evicted from temporary accommodation which the local authority had provided for him while they considered his claim for housing as a homeless person. Thereafter he spent fourteen months on the streets. At length he sought to complain in Strasbourg of violations of Articles 3 and 8. The court declared the application inadmissible. It stated (paragraph 2):

I would however also accept (as this court accepted in Q: see paragraph 63) that a person initially deprived of support without breach of Article 3 may in course of time become entitled to it on Article 3 grounds, and thus to the benefit of s.55(5). He would have to show that some particular vulnerability had come upon him: that he was, in effect, an exceptional case. That might plainly be demonstrated by illness or accident materially affecting his ability to fend for himself. It may include the impact of conditions or circumstances external to the person himself, if these were distinctly shown to have grave effects. It would certainly also include a demonstrated inability, over time, to find food and other basic amenities. There will be other cases; this is the tightest test I can find. Its application should be capable of reasonable and objective ascertainment. In any concrete instance it will be for the Secretary of State to find the present facts and evaluate the future course of events: see Turgut [2001] 1 AER 719. In case of challenge the court cannot escape the final responsibility of decision, as to the application of Article 3 to the facts found and events predicted in light of the test I have sought to describe.

(1) Unlawful violence which is authorised by the State is absolutely forbidden. In this instance the legal rule, given by Article 3 and the common law, governs the whole field. There are no exceptions to the prohibition, no qualifications or legal space for manoeuvre. The prohibition is the law’s first and only word. (2) Unlawful violence by State servants which is unauthorised is also forbidden; so, of course, is all unlawful violence. Here, the State’s duty in fulfilment of ECHR Article 3 (where it runs) lies, so far as it can be done, in prevention before and where prevention fails, investigation, sanction and punishment after. In these cases the prohibition of assault is not the law’s only word. At the hands of the law the State enjoys a measure of judgment in the elaboration of measures to serve the aims of prevention and sanction. The closer the State’s responsibility to the affected individual, the narrower will be the measure of judgment which the law allows. (3) Where Article 3 is deployed to challenge the consequences of lawful government policy whose application consigns an individual to circumstances of serious hardship, the Article is no more nor less than the law’s last word: that is, it operates as a safety net, confining the State’s freedom of action only in exceptional or extreme cases. This spectrum, as I have called it, is required by two dictates: first, the need of some respect for the nature of the obligations which the States parties to the Convention advisedly undertook, and secondly the need for a measured balance in a democracy between the judicial domain of the protection of individual rights and the political domain of State policy evolved in the general interest.

If it is accepted that (1) the bare fact of putting a late asylum-seeker on the streets, absent any special features of the case, will not then and there – on day 1, as it were – constitute a violation of Article 3, but (2) his condition may in time deteriorate so as to cross the Article 3 threshold, the executive and the courts are faced with the "wait and see" approach so vilified by Collins and Gibbs JJ: unless s.55(5) of the Act of 2002 can be construed so as to require the Secretary of State to provide NASS support not only where the individual’s condition will imminently cross the threshold, but also where it may do so in the reasonably foreseeable future. Such a construction may be thought supportable by the subsection’s words: "the exercise of a power by the Secretary of State for the purpose of avoiding a breach of a person’s Convention rights" (my emphasis). However this construction cannot be correct. It would mean that s.55(5) obliges the Secretary of State to ensure that no asylum-seeker is deprived of basic support. This would plainly emasculate the effect of s.55(1). In any event we are surely bound by the reasoning of this court in Q (paragraph 63), rejecting the approach adopted by Collins J at first instance in that case "that the fact that there is a real risk that an individual asylum seeker will be reduced to this state of degradation of itself engages article 3", and the court’s further view (Q paragraph 119, T paragraph 10) that it is when "the condition of an applicant verges on the degree of severity described in Pretty [that] the Secretary of State must act" (again, my emphasis). We are fixed with the "wait and see" approach. It is a function of the legislation. The courts cannot widen the scope of Article 3 to avoid it.

Since writing this first draft of this judgment, I have had the advantage of reading draft judgments prepared by my Lords Carnwath and Jacob LJJ, in which in different ways they express very considerable anxiety as to the practical consequences if these appeals are allowed. Both point to the prospect of 600 and more asylum-seekers being suddenly without support; to the "practical certainty" (per Carnwath LJ) that the charitable agencies will be unable to cope; and the "near certainty" (per Jacob LJ) that a substantial proportion will fall below the Article 3 threshold.

In referring to these passages I by no means intend to indicate any great distance, as to the facts, between my Lords and myself. Though for my part I would be a little more sceptical, the concerns they express are with respect obviously real enough. I desire only to say that such matters, however, pressing, cannot in my judgment be allowed to divert us into adopting a role which is not our own. We cannot don the mantle of the statute’s practical administrators. We certainly cannot do so (and I do not suggest that my Lords have any such thing in mind) in order to save s.55 from excoriation in the moral and political arena. In particular we cannot strain and extend the obligations of the United Kingdom under the ECHR Article 3 beyond what, judicially, we conceive to be their proper limits.

In light of all I have said the approach taken by Collins and Gibbs JJ, in Limbuela and Tesema respectively, was mistaken. As for Adam, I would accept Mr Giffin’s criticism of the judgment of Charles J to the effect that the learned judge was not entitled to find on the medical evidence that Adam’s condition took him over the Article 3 threshold as I have tried to describe it. On the proved or admitted facts none of these cases exhibits exceptional features so as to require the Secretary of State to act under s.55(5).

I would allow the appeals.

This appeal concerns the extent of the Secretary of State’s obligations to three asylum-seekers whose applications for asylum were not made as soon as reasonably practicable, and who therefore fell through the safety net provided for the destitute asylum-seekers by the Immigration Act 1999 and the Nationality, Immigration and Asylum Act 2002. It raises issues first considered by this Court in R ("Q") v Home Secretary [2003] 3 WLR 365, upholding a decision of Collins J. It is worth noting that, although that case caused some political controversy at the time of Collins J’s decision, the decision of this Court was not subject to appeal, and, according to press reports, was welcomed both by the Home Secretary and by refugee groups (see A. Bradley "Judicial independence under attack" [2003] PL 397, 404-5).

The question raised by the present appeals, in its starkest form, is to what level of abject destitution such individuals must sink before their suffering or humiliation reaches the "minimum level of severity" to amount to "inhuman or degrading treatment" under Article 3 of the European Convention of Human Rights. In this judgment I shall use the term "Article 3 threshold" to describe the level of suffering required to engage Article 3.

Although we are dealing with three appeals, we were told that at the time of the hearing there were before the Administrative Court over 650 "live" cases of the same kind, awaiting our decision, and that on average there were about two new cases per day. In all these, as I understand it, interim support has been ordered by the Court.

In Q this Court held that destitution resulting from denial of assistance to asylum-seekers could involve a breach of Convention rights, if the Article 3 threshold were crossed. The threshold was described by reference to the ECHR decision in Pretty v UK (2002)35 EHRR 1. The Court spoke variously of "the kind of state described in Pretty v UK" (para 61), or "a condition which verges on the degree of severity… described in Pretty v UK" (para 62; see also para 121(vii)-(viii)). It gave no further specific guidance, since -

That is not surprising. The description in Pretty was not designed as a precise legal formula. The case concerned the wholly different (if no less intractable) problem of the right of a terminally ill patient to die, the "treatment" alleged being the failure of the prosecuting authorities to give an assurance that they would not prosecute a person who assisted her suicide. From that judgment (para 52) we learn that, under the case-law of the Court, the ill-treatment in question must reach a "minimum level of severity" and involve "actual bodily injury or intense physical or mental suffering"; but that in addition treatment may be characterised as "degrading" if it -

Under this analysis, there are two broad categories: in short, intense suffering (physical or mental), or humiliation of a degree sufficient to "break moral or physical resistance". However, the passage (understandably in the context) was clearly intended as a general statement to cover a range of possible circumstances, rather than a precise formula for determining where or how the threshold is to be set in any particular case or category of case. On the facts of the Pretty case, the Court did not need to decide that issue, because the claimant’s undoubted suffering was held not to be the result of any relevant "treatment" by the state (paras 54-5).

Even if there were a precise criterion for defining the Article 3 threshold, the factual circumstances are so varied and so variable, as to make it practically meaningless to attempt to assess them by reference to a single test at a single point in time. They include the physical and mental resistance of the particular individual; the season of the year and the state of the weather; the incidence of any epidemics or illnesses affecting the area; the availability of medical help and drugs; the extent if any of the help available from other sources, such as charities, community support groups or passers-by; and (if one is considering humiliation) the degree of toleration or lack of it shown by other parties or the public at large.

A further unknown is the period for which the destitution is likely to last. The assumption seems to be that responsibility under Article 3 does not extend beyond the final determination of the asylum application. However that is a very uncertain criterion. We were told that 80% of asylum applications are decided by the Secretary of State within two months. However, appeals have to be taken into account. Mr Limbuela’s case illustrates the potential circularity of the problem. His application for asylum was dealt with relatively speedily. It was made on 6^th May 2003 and rejected by the Home Office on 10^th June, and his appeal to the adjudicator was dismissed in his absence on 1^st September 2003. However, he subsequently appealed out of time on the grounds that he had not been aware of the hearing. On 24^th February 2004 the Immigration Appeal Tribunal remitted the matter for a new hearing. It found that there were special circumstances –

The most important variables, however, are the will and ability of charities working in this field, and the resources available to them. The Court in Q thought that the Secretary of State could lawfully wait "until it is clear that charitable support has not been provided and the individual is incapable of fending for himself." However, this begs at least two important questions:

The practical realities are well-illustrated by the evidence from the Refugee Council (in a statement by Hugh Tristram, a team leader responsible for advising asylum-seekers). The Council have a day centre open on weekdays from 9.30 am to 5.30 pm (2 – 5.30 on Wednesdays), and they are able to offer asylum-seekers some meals during those times. They have no sleeping accommodation. They are closed in the evenings and at weekends. They have four showers in total (two for women, two for men), but no separate hand-basins. Mr Tristram comments:

Having accepted the Pretty formula as the test of the level of suffering, physical or mental, required to engage Article 3, the Court in Q was faced with a separate question whether the state’s function is simply "to wait and see" until that threshold is crossed, or whether it must take preventative action. The Court identified one of the questions with which it would have to grapple:

The Court’s answer was that it must be a condition which "verges on" the degree of severity described in Pretty. That term "verging on" seems to have been designed to mitigate the worst effects of a pure "wait and see" approach. It was contrasted with the view of Collins J, which the Court rejected, that Article 3 would be engaged –

As I understand that passage, it is not necessary for the claimant to show the actual onset of severe illness or suffering. If the evidence establishes clearly that charitable support in practice is not available, and that he has no other means of "fending for himself", then the presumption will be that severe suffering will imminently follow. He has done enough to show that he is "verging on" the necessary degree of severity, and that Article 3 is accordingly engaged.

To explain why there should be such a presumption, I cannot do better than repeat the words of Maurice Kay J in July 2003, as presiding Judge of the Administrative Court, explaining the problem facing the Secretary of State and the Courts, in S, D and T v Secretary of State ([2003] EWHC 1941Admin). Having dealt with the individual cases, he added some more general comments:-

That interpretation of Q, as I understand it, accords with those of Collins and Gibbs JJ in the present cases. A stricter view was taken by Newman J in Zardasht ([2004] EWHC 91Admin). He said:

Further guidance as to the application of the Pretty test was provided by this court in T, decided in September 2003. The background to the case, and the main conclusions of the Court have been described by Laws LJ. The case is of particular interest because of the comparison made between by the Court between T’s case, and that of S, which was not subject to appeal. While allowing the appeal in T, the Court of Appeal specifically endorsed the Judge’s decision in S, observing that –

In S the facts (summarised by Kennedy LJ at paras 17) were as follows:

The Judge had concluded:

The Judge accepted T’s account of the facts. He concluded:

The Court of Appeal disagreed (in a judgment given by Kennedy LJ on 23^rd September 2003). It seems to have been common ground that the case should be considered as at 24^th April 2003, when he was given emergency relief (para 14). Accepting the submission that "this court is as well placed as the Judge at first instance to answer the question", Kennedy LJ concluded:

While acknowledging with respect the Court’s attempt, by use of two contrasting cases, to provide more precise guidance than in Q, I find the decision of limited help in the present context.

In the first place, the main difference from S appears to have been that on 24^th April, which was agreed to be the relevant time for the Court to consider the matter, T had been living rough for only nine days; he had someshelter and sanitary facilities and some money; and, though unwell, he was not in need of immediate treatment. However, the shelter was of the most precarious kind and, as Collins J pointed out in Limbuela (para 18), probably involved a criminal offence under the Heathrow byelaws; and his remaining money (according to his evidence which we were shown) was only "£100 or so". Thus the decision provides no indication of what his state would have been if he had not been provided with support for the five months from April until September, when the case came to the Court or Appeal. Having regard to the way in which the case was argued, the Court’s conclusion represents no more than a snapshot at a particular historic moment.

Secondly, it seems to have been common ground, both in the High Court and in the Court of Appeal, that the Court was able to decide the issue for itself, rather than being confined to anything akin to Wednesbury review. This was explained by Kennedy LJ as follows (para 19):

Mr Knafler informed us that his submissions on this aspect, which were not challenged by Mr Eadie, were based on a passage in the judgment of Simon Brown LJ in R v Home Secretary ex p Turgut [2001] 1 All ER 719, 729 on the Court’s approach to allegations of breach of Article 3. For reasons which I shall explain, I am doubtful whether, had the point been in issue, the Court would have regarded that passage as necessarily supporting the width of the interpretation which counsel put upon it. I shall return to this point.

Before leaving T, I would observe that the Court of Appeal did not attach any separate weight to the second part of the Pretty test, relating to "degrading" treatment. The Judge’s conclusion had been based, at least partly on the view that refusal of support was "debasing him and showing a lack of respect for his human dignity". Although the Court did not comment separately on this aspect, it must be assumed that, after only nine days, the non-physical aspects of T’scondition were not regarded by this Court as sufficient in themselves to bring the second part of the Pretty test into play. Given the unusual facts of the case, I do not see that conclusion as precluding reliance on the second part of the test in other cases.

In October 2003 (following the Court of Appeal’s decision in T) Maurice Kay J drew attention to the mounting scale of the problem, involving some 800 outstanding cases (R (Q) v Home Secretary [2003] EWHC 2507 Admin). He had issued a draft statement representing his views supported by the other nominated Judges. That was intended as an attempt to encourage a practical approach to disposal of these cases which were "clogging up the processes of the Administrative Court." The Secretary of State’s policies for dealing with such cases were not working effectively, because of the lack of "an adequate and efficient decision-making procedure". He said: -

In the papers before us for the appeal, we had no evidence as to the Secretary of State’s response to these strategic problems. However in response to a request made by me immediately before the hearing, Mr Giffin helpfully obtained some information, which has since been confirmed by witness statement. This indicated that, under arrangements established in November 2003 following the judgment of Maurice Kay J, requests for re-consideration are dealt with by a body called the "post-refusal casework team" within NASS, and that the decision time has been reduced to 24 hours in over 80% of the cases.

Unfortunately however as the present cases show, the problem has not gone away, although with the end of the winter period the practical effects are likely to be less. What Mr Giffin’s statement does not do is to indicate what the Secretary of State expects to be the fate of the 600 or so refugees in the cases currently pending before the Court, if and when they are put back on the street. Whether or not any legal flaw is found in the particular decisions made in each of these cases some time ago, that provides little assistance on how they or the other cases should be dealt with in current circumstances, once the claimants are deprived of the interim support which they are presently enjoying. There is no doubt that the policy of the Courts of giving interim relief in most of these cases has mitigated the problem for a time, but it is clearly impossible to treat that as offering any comfort for the future.

The scale of the potential problem can be illustrated by the figures given in the Lord Mayor’s report, to which Laws LJ has referred. He takes comfort from the fact that many thousands of asylum-seekers have apparently been able to find support in London outside the NASS scheme. However, the report also states that a "substantial minority of these destitute newcomers" are forced to sleep rough. Elsewhere (para 8.14) Government figures are given for the numbers thought to be sleeping rough in London; as a result of Government initiatives, they are said to have reduced from 650 in 1998 to 267 in June 2003. Thus the numbers currently before the Court, and at risk of being forced onto the streets following our decision, would be equivalent to more than twice the total number currently sleeping rough in London.

Accordingly, in my view, we cannot look at the present cases in isolation. Nor are we invited to do so. In his skeleton argument on behalf of the Secretary of State, Mr Giffin acknowledges that in spite of some success in reducing disputes -

I agree. We cannot ignore the fact that the likely result of allowing these appeals is that the safety net of interim relief will be removed not only from the three appellants before us, but also from a large proportion of the other 600 applicants. Further, Mr Giffin accepts that it would not be realistic for us to confine attention to the circumstances of the three appellants at the time, some months ago in each case, when interim relief was first made available. He accepts that judicial review procedure is sufficiently flexible in an appropriate case to enable the Court to consider the up-to-date position (see E and R v Secretary of State [2004] EWCA Civ 49 (para 43, 76-7)). If we are to provide any practical assistance in these and the other cases, we need to take account of the current facts.

Before returning to that issue, it seems to me necessary, first, to revisit the nature of the potential breach of Article 3, and then to seek to define the respective responsibilities of the executive and the Courts.

In Q, the Court concluded that the regime imposed on asylum-seekers denied support under section 55(1) constituted "treatment" within the meaning of Article 3. The reasoning was that asylum-seekers could not be removed lawfully until their claims had been determined, but while here they were prohibited from working except with special permission. The Court summarised the position:-

There was no appeal against this decision, and the reasoning was not challenged in argument before us. I would comment, however, that it appears to represent a significant extension of existing jurisprudence. We were not referred to any decision, whether of the ECHR or of any court in the other member states, which goes so far as to impose a positive obligation on the state to provide support in such circumstances.

The only cases referred to in Q in relation to the meaning of "treatment" were two cases relating to protection of children (A v UK [1998] 27 EHRR 611 and Z v UK [2001] 34 EHRR 97); and D v UK [1997] 24 EHRR 423, which concerned the removal under immigration legislation of a person dying from AIDS. None is directly in point. The State’s function for the protection of children has always been regarded as in a special category. D v UK is a difficult case, and, as Laws LJ has explained, it was narrowly applied by this court in N v Secretary of State. Lord Bingham said, in the passage of his judgment in the Pretty case to which Laws LJ has referred:

Accepting, however, that we are bound by Q on this issue, I acknowledge with gratitude the illumination provided by Laws LJ’s powerful discussion of the scope of Article 3, and its application in the present context. As he says, the legal reality is a spectrum. At one end is state-authorised violence. At the other are to be found executive decisions in exercise of a lawful policy objectives, which have consequences for individuals so severe that "the Court is bound to limit the State’s right to implement the policy on Article 3 grounds". I agree also with much of his analysis of the consequences of that distinction, and of the correct approach to the task of drawing a line in an individual case.

Laws LJ accepts that Article 3 may be engaged by a particular "vulnerability" in the individual, or external circumstances which make it impossible for him to find food and other basic amenities. Where, with respect, I part company from him is in his view that, on the evidence available to us, the judges were not entitled to find that such circumstances existed in the present cases. I would add that I find it difficult not to regard shelter of some form from the elements at night (even if as limited as it was in T’s case) as a "basic amenity", at least in winter and bad weather. I would not regard O’Rourke (where the facts were very different, and the applicant’s plight was largely self-inflicted) as establishing the contrary.

Once it is accepted that Article 3 is potentially in play, it must follow in my view that the State has some responsibility in the matter. As the Strasbourg court said in Pretty (para 51):-

The obligation "to take measures" seems to me to imply more than simply acting as a longstop in individual cases as they arise. That may be sufficient if the alternative system of charitable support is able to cope with the generality of cases, so that Article 3 suffering is truly the exception. However, if on the available information, the scale of the problem is such that the system is unable to cope, then it is the responsibility of the State to take reasonable measures to ensure that it can cope. How that is done, for example whether by direct support or by financial assistance to charities working in the field, is a policy matter for the State. As Lord Bingham said in Pretty, there will be considerable room for executive judgment as to the steps necessary to achieve it, balanced against the policy objectives of the legislation (as explained in Q at para 26, already quoted by Laws LJ).

This is not inconsistent with the wording of section 55 (5)(a). That empowers the Secretary of State, as an exception to the general prohibition on providing support, to exercise his powers "to the extent necessary for the purpose of avoiding a breach of a person’s Convention rights". On its face, this appears to be directed to avoiding a breach in an individual case. However, if the scale of the problem is such that individual breaches can only be avoided by more general action, such action can and must be taken "for the purpose" of avoiding the individual breaches.

Furthermore, fairness and consistency can only be achieved in practice if the Secretary of State has in place policies defining criteria to be applied by the decision-maker. As Lord Clyde said in R (Alconbury Developments Ltd) v Secretary of State [2001] 2WLR 1389 para 143:-

We have been referred to the most recent guidance note issued to decision-makers by the Secretary of State (Policy Bulletin 75). It is of interest that, in dealing with the question of whether an application for asylum has been made as soon "as reasonably practicable", detailed guidance is given including some ten examples based on hypothetical cases. By contrast the question of breach of Article 3 is dealt with very briefly. The note states:-

While this is a faithful reflection of the judgment in Q, it provides little practical guidance as to how the decision-maker is to deal with individual cases in the light of the practical realities disclosed by the evidence before us. Case-by case decision making of this kind cannot reasonably regarded as a sufficient discharge of the Secretary of State’s responsibilities, if, on the information available to him, the numbers likely to need help are far greater than the ability of the alternative agencies to cope.

There is no statutory right of appeal against the refusal of support under section 55. As the Court of Appeal said in Q, section 55(10) "is unequivocal in blocking access to the appeal mechanism for asylum support." (para 110). It was argued in that case that Parliament had thereby acted incompatibly with Article 6 of the Convention. That argument was rejected. This court, applying the decision of the House of Lords in Runa Begum v Tower Hamlets LBC [2003] 2WLR 388, held that a fair administrative decision-making process, combined with the possibility of judicial review would satisfy the requirements of Article 6 (para 114-7). In so holding it noted that in human rights case a more "intensive" scrutiny is appropriate, following R v Ministry of Defence ex p Smith [1996] QB 517, 554:

In T the Court proceeded on the agreed basis that, the facts having been established, it was open to the Court to form its own view as to whether the Article 3 threshold was crossed. I have already expressed my doubt as to the correctness of that approach. In practical terms, it is difficult to see how it differs from the right of appeal, which Parliament was careful to exclude.

As I have noted, the basis of the agreed position appears to have been the judgment of Simon Brown LJ in Turgut. That case concerned the refusal by the Secretary of State of exceptional leave to remain to a Turkish Kurd draft evader, who claimed that if returned to Turkey he would be subject to a real risk of suffering torture or other treatment contrary to Article 3. It is important to note that case as argued did not depend on the individual circumstances of the applicant, but was as a test case for young Turkish draft evaders generally (p 722b-c). Although the decision preceded the coming into force of the Human Rights Act 1998, the Court considered whether judicial review, in accordance with the approach in ex p Smith, satisfied the requirement for an effective remedy under Article 13 of the Convention. The passage in Simon Brown LJ’s judgment reads as follows:-

I find no difficulty with that approach as applied to the question at issue in that case, which turned on the inferences properly to be drawn from objective information as to the treatment of draft evaders in Turkey. However, I do not think the same approach can be applied without qualification to the present cases. I accept that the Court, as a public authority, has its own separate responsibility under the Human Rights Act not to act incompatibly with the Convention. To that extent it may need to form its own view as to where the boundary is to be drawn in particular cases, or categories of case. I agree also that in a case involving a potential breach of Article 3, the Court’s review must be relatively "intense". Even then, where the Administrative Court has made a reasoned assessment of the facts, applying the correct tests, I would not regard it as appropriate for this Court to interfere unless the Judge’s decision is plainly wrong.

However, the procedure remains one of review; it must not become an appeal, which Parliament has specifically excluded. Furthermore, account must be taken of the practicalities. The Court cannot sensibly undertake a day-by-day review of individual cases in relation to Article 3. Even if it could, it would be an absurd misuse of resources for it to attempt to do so. The public money spent on legal representation would be far better spent on providing practical support. (Each of the cases before us has generated some 300-500 pages of evidence and supporting material.) The Court’s primary task is to clarify the legal standard, and to ensure that there are in place adequate measures to ensure that the standard is generally met. If the Secretary of State has in place realistic arrangements for meeting his responsibilities under the Convention, and deciding individual cases, then the Court’s role will normally be limited to ensuring those arrangements have been applied fairly and consistently.

In Limbuela Collins J after a careful examination of the evidence, including that of the charities concluded:

In Tesema Gibbs J held on the evidence before him:

For the reasons I have already given I see no error of law in either decision, and I see no other reason for this Court to interfere.

In Adam Charles J applied a stricter test, following Newman J in Zardasht. I have already explained why I think that test does not follow from Q. As Laws LJ has explained, Mr Adams was on the streets between 16^th October and 10^th November 2003. During that time he slept outside the Refugee Council premises in Brixton, and survived by using their washing facilities and obtaining occasional meals from them. After that he had the benefit of interim relief ordered by the Court. Charles J found, even applying the strict test advocated by the Secretary of State (following Zardasht), that three weeks in such conditions, supported by "the inferences drawn from unsatisfactory medical evidence" was sufficient to cross the threshold. It is clear that, applying the correct test, he would have come to the same conclusion. Again I see no reason to interfere.

Before the decision of this Court in Q, there were three possible approaches in Article 3 to cases such as the present:

For the reasons given I see no reason to interfere with the conclusions of the Judges in the individual cases, viewed as such. However, I think that is too narrow an approach, particularly given the way in which the appeals have properly been presented by Mr Giffin. We are asked to give current guidance. To do that, we must look at the overall position in current circumstances, and we must take account of the realities.

We had no direct evidence as to the current condition of the individual appellants at the time of the hearing. I would assume that, having had the benefit of interim relief for some months, and pending our decision, none of them currently is "verging on" Article 3 suffering. However, if we allow the appeal, we must anticipate that they and up to 600 others will become dependent on charitable support. We have no evidence from the Secretary of State as to how in practice he expects that sudden influx to be handled, or that he has policies in place adequate for the purpose. On the evidence presented by Shelter and others (already summarised by Laws LJ), there is not simply a "real risk", but a practical certainty that the current charitable agencies will be unable to cope with such an influx; that many of the claimants will (in the words of Q) "not (be) able to find any charitable support or other lawful means of fending for themselves".

I am conscious that the wider picture has been brought in to the discussion at a relatively late stage, partly in response to questions from me before the hearing, and partly due to the late intervention of Shelter, and its supporting evidence. Although Mr Giffin fairly did not object to this widening of the debate, the Secretary of State may not have been able to respond fully to it. I had hoped that, in response to my questions, he would have been in a position to give the Court a clearer picture as to how in practice the Secretary of State expected the overall problem to be dealt with following a successful appeal. I accept that there was little time to do this, and that my questions may not have been as specific as I intended. Having said that, however, we have to proceed on the basis of the evidence before us.

On that basis I would dismiss these appeals.

If it were appropriate to deal with the three cases simply on the materials concerning the three individuals, I would agree with the conclusion of Carnwath LJ in paragraphs 135-138 of his judgment. I also agree with his much wider analysis of the problems posed in this case. I also agree with Laws LJ’s "spectrum analysis". To my mind however, that analysis does not help one resolve where, on the spectrum, these cases fall. Having said all that, I am however, clear that these appeals should be dismissed. My reasons are brief – any further elaboration would serve no useful purpose.

Overshadowing the facts of these individual three cases is the fact that there are 666 others where the position is similar. In all these cases the individual concerned is seeking asylum in this country and his or her case has not been decided. No-one knows whether their claims to asylum are good or bad. All that can be said at this stage is that it has been decided that they did not claim asylum within 2 days of arrival in this country.

What is to be done about destitute people in this position? A case by case analysis of the situation of each individual is costly – it involves solicitors (at the State’s expense) ringing round all possible charities, the charities devoting scarce resources to answering inquiries, evidence of weather forecasts, examination by doctors (at the State’s expense) to see how unwell the individual is, and a myriad of other inquiries, all designed to show that the applicant has "passed" the Art.3 threshold. These are followed up by applications to the court. Implicit in this approach are repeated applications for those that have failed – has the weather changed, has the capacity of charities to cope changed, and above all has the health of the individual sunk to a sufficiently low level? And in those cases where the individual has sunk below the threshold and so is provided with minimal food and shelter, the time may come when he or she is "better" and can be put out on the streets again – there is a real prospect of an endless cycle. The "verging on" test is abhorrent, illogical and very expensive. I am not surprised that some judges cannot accept that it can be correct, even though it may follow from this court’s decision in Q in that the "real risk" test was rejected. "Verging on" seems to be the only next logical stopping place.

Certain figures in the evidence bring out the scale. As I have said there are 666 cases similar to the three we have. In some cases the applicant may have "passed" the Art.3 test at the time of the relevant Judge’s order. Many will not – and of those who "passed" probably some are now better (having received food and shelter for some time) and might fail the test now. Any reasonable estimate of the number likely to put on the streets (and the Secretary of State has not himself made any estimate) is that it must be of the order of 500 or more. Most of these will be in London.

Set against that is the general position as regards homelessness, the availability of shelter and the level of demand. From 1998 to 2002, according to the Government’s Rough Sleepers Unit, the Government target of reducing the number of rough sleepers by 2/3 had been met. By June 2002 there were estimated to be only 532 people sleeping rough throughout England, an impressive achievement.

Of course even these homeless people (assuming they are not asylum seekers subject to s.55) would not be entirely penniless too – not only are they free to earn money by working but there will be financial state support. That was the position in O’Rourke – a case, in my view, miles away from our present problem.

As regards the availability of charitable shelter, the evidence indicates virtually no spare capacity. So much appears from Mr Tristram’s (of the Refugee Council) statement – there are a couple of hostels in London which take asylum seekers but they are essentially full up.

As regards the availability of charitable food, the position is not much better. It is evident that the established charities could not feed the 500. Nor are there adequate facilities for personal hygiene for this number.

These stark figures to my mind show that if this sort of number of people are put on the streets, without money and with no entitlement to earn any, there is a near certainty that a substantial proportion will fall below the Art. 3 threshold. Of course some might, to survive, resort to theft, prostitution, or illegal working (very likely for so-called "gangmasters"). But all these forms of survival would not pass the Art. 3 threshold and cannot be prayed in aid by the State – which, to be fair, it does not seek to do.

It follows that although one may not be able to say of any particular individual that there is more than a very real risk that denial of food and shelter will take that individual across the threshold, one can say that collectively the current policy of the Secretary of State will have that effect in the case of a substantial number of people. It seems to me that it must follow that the current policy (which includes having no policy save in the case of heavily pregnant women) is unlawful as violating Art. 3. And it follows that the treatment of the particular individuals the subject of these appeals in pursuit of that policy is also unlawful.

For that reason I would dismiss all the appeals. Unless and until the time comes when it can no longer be said that a substantial number of people will fall below the threshold, Art. 3 will prevent the State from standing by and letting them do so.

I must add a few words about Q and T. In Q this Court decided two things. First that Art. 3 could be engaged in these s.55 cases. This is because the State was saying to the individual "Though you are destitute you may not work for money for your food and shelter." I do not question that reasoning, indeed if it were open in this Court (which it is not) I would agree with it. Second, however, that in the case of a particular individual, a "real risk" test is not enough. Again it is not open to this Court to disagree. T merely decided that on the facts the individual had not fallen below the threshold at the time – he had a little money and was managing (probably illegally) to survive at London Airport. The Court did not address what would happen when his money ran out or if he were removed from the airport.

Both of these cases addressed the problem from the point of view of considering just the individual. Neither of them addressed what I see is the real problem thrown up by these three test cases and the hundreds pending – that unless the interim orders are continued, a substantial number of people will have their Art. 3 rights violated. I therefore do not think that either decision precludes these appeals from being dismissed on that ground.