Bloggs 61, R (on the application of) v Secretary of State for the Home Department

Although the main focus of Mr. Clayton’s complaint has become one of procedural unfairness, his arguments proceeded as if he were still seeking relief under Lord Woolf’s third category, namely a mandatory order for the appellant to remain in the protected witness unit for the duration of his sentence. He did so in reliance on alleged representations made by the Police, not by the Prison Service. He maintained that though the Police, not the Prison Service, allegedly gave the appellant assurances to that effect, the Prison Service are in some way responsible for and through the Police for giving the appellant a legitimate expectation that he could remain in the unit throughout his sentence.

It is plain from the legal framework that I have summarised that the Police had neither actual authority to make the alleged representations on which the appellant relies, nor the authority to implement any such as might have been made. As to the latter, not even the Prison Service had authority to grant him the status of a protected witness for the duration of his sentence. As I have said, it was required to re-assess his entitlement to it annually in the light of the views of his police handler and the staff of the protected witness unit in which he was held.

Mr. Clayton sought to overcome the first difficulty in the application that a representation of A to B of what C will do cannot bind C, unless C authorised such a representation, by arguing that the Police, on the facts, had ostensible authority from the Prison Service to make the representations and/or that the Prison Service were estopped from denying that it was bound by them. (For the avoidance of doubt, I should note that he expressly disclaimed any suggestion of express authority through the persona of the Home Secretary in his overall but differing levels of responsibility for the Police and the Prison Service.) He submitted, by reference, inter alia, to paragraphs 1.15 and 1.16 of the Prison Service Instructions (see paragraph 13 above), that the Prison Service, by its close reliance on and involvement of the Police, in particular of the prisoner’s police “handler”, when determining whether to grant protected witness status, held out the Police as having such authority. He suggested that it was a representation of a general nature arising from the Prison Service putting the Police in a “ministerial” position of the sort described by Lord Keith in Armagas v. Mundogas [1986] AC 717, at 777:

“the ostensible authority is general in character, arising when the principal has placed the agent in a position which in the outside world is generally regarded as carrying authority to enter into transactions of the kind in question.”

Mr. Clayton accepted that private law concepts of ostensible authority and estoppel could not be lifted without more into the public law sector and re-branded as one of legitimate expectation. However, he said that all three concepts were rooted in the notion of fairness, and that fairness was a well-established ingredient of the notion of legitimate expectation; see e.g. R v. Inland Revenue Commissioners, ex p MFK Underwriting [1990] 1 WLR 1545, DC, per Bingham LJ at 1569H-1570A; CCSU v. Minister of Civil Servants [1985] AC 374, per Lord Roskill at 415C-G; and ex p. Coughlan, per Lord Woolf MR at paras. 56 and 80 – 82.

Mr. Clayton suggested that it was inevitable, and contemplated by the scheme the Prison Service had devised and operated in conjunction with the Police that police handlers, when discussing with potential informers the consequences of their volunteering information, would become drawn into discussing and giving assurances about the grant of protected witness status. He added that the nature of the relationship between a police handler and a prospective prisoner/witness is one of trust and confidence that leads and entitles the latter to rely on the former’s representations. As he rightly observed, a potential informant is unlikely in the circumstances to analyse closely the legal source and limitations of those assurances when deciding whether to put himself at risk by committing himself to assisting the police in this way. To him, the police officer is simply speaking on behalf of “them”, the criminal justice establishment as a whole. Mr. Clayton relied in this respect on the reasoning of Ouseley J, at paragraph 63 of his judgment, that: the police were all that this appellant, a frightened informant, had to deal with and trust; the appellant had every reason to suppose that the Police and the Prison Service were working towards a common goal and that what the police said would happen to him as a convicted prisoner would in fact happen; and the appellant was not privy to the arrangements and dealings between the Police and the Prison Service.

As to the content of the alleged representations, namely that the appellant would be allowed to remain in the protected witness unit for the whole of his sentence, Mr. Clayton suggested that they were likely to be typical of what police handlers routinely tell prisoners when discussing the possibility of protected witness status as a quid pro quo for their assistance as witnesses. He submitted that, as in the case of the appellant’s fellow inmates in the protected witness unit, his police handlers did not tell him that the status was subject to annual review by the Prison Service. Nor did the Police Circular require them to warn potential informers that the status was subject to periodic review. And, he suggested that the Prison Service must have been well aware that police officers made such misleading misrepresentations when discussing protected witness status with potential informers and should have taken steps to alert the latter before they irrevocably committed themselves to assisting the police in reliance on the representations.

In consequence, Mr Clayton submitted that the appellant had a legitimate expectation in the sense having a reasonable basis for it, relying on words of Lord Fraser in A-G of Hong Kong v. Ny Yuen Shiu [1983] AC 629, at 636, and of Lord Woolf in ex p Coughlan,at para. 56. The reasonable basis on the facts here, he submitted, could be derived from the ostensible authority of the Police to commit the Prison Service.

Mr Jay submitted in reply that, to the extent that the private law principle of ostensible authority bears on the public law notion of legitimate expectation, on the facts here the Police had no ostensible authority to commit the Prison Service to grant or maintain throughout the appellant’s sentence his protected witness status. On the contrary, paragraph 5 of the Annex to the Police Circular (paragraph 9 above) expressly prohibited the Police from making promises as to treatment or location within the prison system. Nor was there any evidence of any act by, or general practice of, the Prison Service to support Mr. Clayton’s contention that it had put police officers in “a ministerial position” by holding them out as having authority to make such promises or representations despite that express prohibition.

As to the content of the alleged police representations relied on, namely that the appellant would be retained in the protected witness unit for the duration of his sentence, Mr. Jay pointed out that the private law principle of ostensible authority as a basis for the public law concept of legitimate expectations had limitations not applicable in private law disputes. One is that a public authority cannot be held to a promise or representation by another as to the exercise by it of a statutory public power which it cannot, or should not by proper exercise of that power, be bound; see R (Reprotech Ltd.) v. East Sussex County Council[2003] 1 WLR 348, HL.

The starting point, as both counsel acknowledged, is that the Police could not bind the Prison Service to treat the appellant as a protected witness unless they, the Police, had actual or ostensible authority to do so. There is no evidence that the Police had such actual authority, quite the contrary as Mr. Jay submitted and as Ouseley J found, in paragraph 64 of his judgement, by reference to the extracts from the Police Circular and the Prison Service Instructions which I have set out above. Nor, for the reasons urged by Mr. Jay and helpfully summarised by Ouseley J. at paragraphs 64 and 65 of his judgment, was there any basis on the evidence for holding that the Police had ostensible authority, namely from any specific or general act or practice of the Prison Service of holding out, or of adopting regardless of its own its management requirements, whatever promises police officers make in this context.

But, even if in private law terms the Prison Service had in some way held out the Police as having authority to commit it generally or in the case of this appellant to retaining protected witness status for the whole of a sentence, the appellant could not have relied upon the principle of legitimate expectation to enforce that commitment. That is because, although it would not have been inconsistent with its statutory powers in the 1952 Act or the Prison Rules, it would not have been a “legitimate” expectation. This is where, in the circumstances of this case, the public law concept of legitimate expectation parts company with private law principles of ostensible authority or estoppel. Even if the appellant had not appreciated at the time of such representations as were made that the Police did not also speak for the Prison Service and/or that the Service could not, in any event, bind itself in that way, he could not claim, in reliance on the representations that he had a legitimate expectation of their fulfilment.

In Reprotech, where the issue was whether a local authority was estopped from denying that a site could be used for the generation of electricity, Lord Hoffmann, with whom the other Law Lords agreed, warned against the introduction of private law concepts of estoppel into public law:

“34.

There is … an analogy between a private law estoppel and the public law concept of a legitimate expectation created by a public authority, the denial of which may amount to an abuse of power, Ex p. Coughlan [2001] QB 213. But it is no more than an analogy because remedies against public authorities also have to take into account the interests of the general public which the authority exists to promote. Public law can also take into account the hierarchy of individual rights which exist under the Human Rights Act 1998 ….

35.

… In the Western Fish case [1981] 2 All ER 204 the Court of Appeal tried its best to reconcile …. invocations of estoppel with the general principle that a public authority cannot be estopped from exercising a statutory discretion or performing a public duty. But … [i]t seems to me that in this area, public law has already absorbed whatever is useful from the moral values which underlie the private law concept of estoppel and the time has come for it to stand upon its own two feet.”

Keene LJ expressly followed that reasoning in South Bucks District Council v. Flanagan [2002] 1 WLR 2601, CA, where the issue was whether a solicitor instructed by a local authority in planning enforcement proceedings had ostensible authority to withdraw enforcement notices. He said, at paragraph 18 of his judgment:

Even if the appellant were to express his claim to a legitimate expectation on the basis that the Prison Service, though not bound by such representation that the Police had made, would in fact follow the Police line, his argument would founder on the facts of the case. The initial stance of the Police in December 2001 when Mr. Golds was reviewing the appellant’s status was, as I have said, divided. Mr Golds’ evidence is that, before making his decision at that stage to remove the appellant’s protected witness status, he considered the information and views of the police and of Mr. Ayers in assessing the level of risk to which the appellant would be exposed if returned to mainstream prison conditions. He also considered the Service’s criteria for maintenance of such status and also the public interest in the need for and cost of it.

In my view, Ouseley J’s reasoning and conclusions in paragraphs 64 – 68 and 108 of his judgment were loyal to the principles of law that I have mentioned and he correctly applied them to the facts of this case – facts assumed for the purpose in favour of the appellant’s claim. Although the appellant, when faced with Mr Golds’ decision to remove him from the unit, may have felt a strong sense that he had been treated unfairly, it does not follow from those assumed facts that the Police had authority, either on private or public law principles, to bind the Prison Service so as to give him a legitimate expectation of the fulfilment of whatever representations they had made. Objective principles necessarily intrude.

It follows that, even on the assumption that the police officers made the claimed representations and that the appellant relied on them, his claim cannot as a matter of law cannot succeed as a matter of substantive entitlement to remain in a protected witness unit for the whole of his prison sentence.

There are additional factual difficulties in any claim of the appellant under any continuing claim that he may have to a substantive entitlement under the head of legitimate expectation. As I have already noted, in paragraphs 15 and 16 of this judgment, he had begun to cooperate with the police as soon as he was arrested and long before, on his case, the police made any promises or representations to him about securing a permanent place in a protected witness unit. It is similarly relevant to note that, on his account, he only signed the witness statement that he had provided to the police after they told him that if it was not acceptable to the Crown Prosecution Service, he would be returned to a “normal” prison (see paragraph 18 above). There is also the point that he benefited from his cooperation with the police and his offer to give evidence for them in a substantial reduction in sentence that he would otherwise have received on his plea of guilty to the conspiracy.

As to the complaint of procedural unfairness, on which Mr. Clayton placed much emphasis in his argument on appeal, there is simply no basis for it on the facts. As I have said, Mr. Golds, when making his initial decision, clearly took account of all the relevant factors. These, as I have also said, must have included Mr. Ayers’s general concern about the general risk to the appellant in prison of being suspected by his fellow inmates as a police informer. The fact that he seems to have given less weight to that matter than that for which the appellant contends is not, for reasons that I shall develop under the next ground of appeal, a basis for regarding his decision-making process as in some way flawed so as to fall short of what the appellant was entitled to expect. In any event, as I have described in paragraphs 26 and 27 of this judgment, he gave further and full consideration in April and May 2002 to all the relevant factors, clearly including and responding to Mr. Ayers’ concerns. And he reached a decision that, in my view, cannot be regarded as procedurally flawed or otherwise unreasonable in the sense of breaching any legitimate expectation that the appellant may have had as to how the Prison Service should have considered his reluctance to return to mainstream prison conditions.

The second issue – breach of the duty under Article 2 to protect the appellant’s right to life

Mr. Clayton submitted that the Prison Service’s decision, by the recommendation of Mr. Golds in December 2001, to return the appellant to mainstream prison conditions breached his right to life under Article 2 of Schedule 1 to the 1998 Act, in particular, the first sentence of it, namely, “Everyone’s right to life shall be protected by law”.

He maintained that there was little evidential basis for Mr. Golds’ subsequent explanation of his decision in paragraph 11 of his May 2002 witness statement, namely:

. “As a direct result of the decision of the CPS not to use the claimant’s witness statement against a major criminal (i.e. the statement which led to the claimant’s admission to a Protected Witness Unit), there is no current police investigation into the person who was the subject of that statement. …. Police intelligence has no evidence of any current threat to Bloggs 61. The Subject is a sophisticated criminal. Police intelligence is that this individual, and his associates, will take a business-like approach to Bloggs 61: as Bloggs 61 has not caused them any damage, it would be unwise for them to court police attention by taking reprisals against him. The original police risk assessment, dated 19^th June 2001, states that no representative from the Subject or his associates’ organisation were present at the trial of his co-accused, and that giving evidence against his co-accused has not resulted in any increased threat to the claimant.”

Mr. Clayton maintained that Mr Golds, in that passage, did not take or take sufficient account of the apparent division of police views as to the effect on the appellant’s safety of removing him from the protected witness unit, a split of which Detective Superintendent Ryan had informed him by letter of 4^th December 2001. Mr. Clayton referred to the history of incidents and possible threats to which I have referred earlier in this judgment, at paragraph 19, and to a note of an officer in the Police Criminal Justice Protection Unit of 18^th October 2001 relying on certain of those matters as strong grounds for regarding him as sufficiently vulnerable to be retained within the unit. Mr Clayton noted the absence of any disclosed police reports the other way until a few days after Mr. Golds’ initial decision of 4^th December 2001, when Mr. Ryan informed him, in reliance on the appellant’s case officer’s report that the Police had no information to indicate that he was at risk. And, of course, he relied strongly on Mr. Ayers’ representations to Mr. Golds in late December 2001 about the more generalised threat to the appellant if he were returned to the normal prison environment without adequate explanation for his absence from it over the previous 18 months.

Mr. Clayton, in addition to his criticisms of Mr Golds’ December 2001 decision, and January 2002 confirmation of it, noted that there was no sign of the appellant having had the benefit of an annual Prison Service assessment under paragraph 7.4 of the Prison Service Instruction (see para. 12 above). He also urged the Court to treat with circumspection the May 2002 Police report re-assessing the physical risk of return of the appellant to mainstream prison conditions from “high” to “medium”, a report presented after the commencement of the judicial review proceedings; see R v. London Borough of Brent, ex p. S [2002] EWCA Civ 693, per Schiemann LJ, at para. 26.

In the light of all that material, Mr. Clayton submitted that the critical question was whether the single assertion of the officer on which Mr. Ryan relied in reporting to Mr Golds on 13^th December 2001 that he knew of no risk to the appellant in returning him to mainstream prison conditions was sufficient to satisfy the requirements of Article 2. He suggested that it was not, noting that the only changes of circumstance between the original and revised risk assessments of the Police were that he was no longer to be a prosecution witness and that his absence from mainstream prison for the previous 18 months might well make him the subject of suspicion of those who had had most to fear from him or from his fellow prisoners generally. He remained, Mr. Clayton maintained, at substantial risk to his life from either category.

As to the way in which this Court should examine this essentially speculative question and the Secretary of State’s justification of Mr. Golds’ decision, Mr. Clayton submitted that the Court should adopt one of intense and anxious scrutiny and, in particular, adopt the approach of the Court of Appeal in R v. Lord Saville, ex p. A (No. 1) [2000] 1 WLR 1855, at para. 37, applying the test of Sir Thomas Bingham MR in R v. Ministry of Defence, ex p. Smith [1996] QB 517, at 554E-F.

Mr. Jay submitted that, whatever refinement of the judicial review test is adopted, the decision here does not approach a breach of the appellant’s right to life under Article 2. He said that Mr. Golds was a senior member of the Prison Service, charged with responsibility for making the decision and that, having carefully considered all the relevant advice, his decision-making process was not flawed nor was his decision unreasonable. He maintained, in contrast to the argument of Mr. Clayton, that the facts that the appellant had not after all been called on to give evidence against the subject of his witness statement, and that that person had not even been arrested in respect of the matters referred to in it, had a significant effect on the level of risk that he faced within the prison system.

As Mr. Jay observed in the course of his submissions, there is no single or all-purpose formulation of the test to be applied when considering what steps a public body should take in order to protect a person from a threat to his life from a third party or parties. It all depends on facts of the case. As to general guidance on how to approach each case on its own particular facts, the starting point is the decision of the European Court of Human Rights in Osman v. United Kingdom (1998) 29 EHRR 245. There, the Court held, at paragraphs 115 and 116, that Article 2(1) may, depending on the facts, impose a duty on a public authority to take all reasonable steps to protect a person from a real and immediate risk to his life.

“115.

… the first sentence of Article 2(1) … enjoins the State … to take appropriate steps to safeguard the lives of those within its jurisdiction. … Article 2 … may … imply in certain well-defined circumstances a positive obligation on the authorities to take preventive operational measures to protect an individual whose life is at risk from the criminal acts of another individual. …

“116.

For the Court, and bearing in mind the difficulties involved in policing modern societies, the unpredictability of human conduct and the operational choices which must be made in terms of priorities and resources, such an obligation must be interpreted in a way which does not impose an impossible burden or disproportionate burden on the authorities. Accordingly, not every claimed risk to life can entail for the authorities a Convention requirement to take operational measures to prevent that risk from materialising. … where there is an allegation that the authorities have violated their positive obligation to protect the right to life in the context of their … duty to prevent and suppress offences against the person, it must be established to its satisfaction that the authorities knew or ought to have known at the time of the existence of a real and immediate risk to the life of an identified individual or individuals from the criminal acts of a third party and that they failed to take measures within the scope of their powers which, judged reasonably, might have been expected to avoid that risk. …. it is sufficient for an applicant to show that the authorities did not do all that could be reasonably expected of them to avoid a real and immediate risk to life of which they have or ought to have knowledge. This is a question which can only be answered in the light of all the circumstances of any particular case. ” [my emphasis]

The critical words in that passage, “a real and immediate risk” to life were considered by the Court of Appeal in R (A & Ors) v. Lord Saville (No. 2), [2002] 1 WLR 1249,which concernedthe risk to the lives of soldiers if they were required to give evidence to the Bloody Sunday Inquiry at the Inquiry Tribunal’s place of sitting in Londonderry. The Court upheld the Divisional Court’s grant of judicial review quashing the Tribunal’s decision requiring the soldiers to give their evidence to it in Londonderry. In doing so, the Court sought, in paragraphs 28 to 31, to avoid any all-purpose formula of the threshold of risk required to engage the Article 2 duty, preferring a common sense and humane approach tailored to the circumstance of each individual case. Thus, it considered, at paragraph 28, that the test applied in Osman, of a real and immediate risk, while apt for a case where there was a known very high degree of risk to life from criminal acts of a third party, was “well above the threshold” that would engage article 2 where the risk attended an action that the authority itself was contemplating, as in the case before it. The Court applied what it regarded as a lower threshold of risk, taking into account both the soldier applicants’ subjective fears and their objective justification. Lord Phillips of Worth Matravers MR, giving the judgment of the Court said, at paragraph 31:

“We consider that the appropriate course is to consider first the nature of the subjective fears that that the soldier witnesses are likely to experience if called to give evidence in the Guildhall, to consider the extent to which those fears are objectively justified and then to consider the extent to which those fears, and the grounds giving rise to them, will be alleviated if the soldiers give their evidence somewhere in Great Britain rather than in Londonderry. That alleviation then has to be balanced against the adverse consequences to the inquiry of the move of venue, applying common sense and humanity. The result of the balancing exercise will determine the appropriate decision. This course will, we believe, accommodate both the requirements of article 2 and the common law requirement that the procedure should be fair.”

Coming closer to the factual issue in this appeal, Crane J, sitting in the Administrative Court, in R v. Chief Constable of Norfolk, ex p. DF [2002] EWHC 1738 (Admin), has considered article 2 in the context of a refusal of the Prison Service to admit a prisoner to a protected witness unit. There, the issue concerned the adequacy of the information provided by the Police to the Prison Service. Crane J, in searching, at paragraph 35 of his judgment, for the article 2 threshold of risk appropriate to the facts before him, distinguished between a risk to life from a third party resulting from a decision of a public authority’s response to an existing situation, as in Osman, and that where the public authority proposed some action, as in Lord Saville (No. 2), the holding of an inquiry in a particular place. He held, at paragraph 36 of his judgment, that recommendations by the Police and resulting actions by the Prison Service as to whether a serving prisoner should be admitted to a protected witness unit were closely analogous to decisions of the police in the course of their duty to protect members of the community. He drew support from the judgment of the European Court of Human Rights in Keenan v. UK [2001] ECHR27229/95, applied by Hooper J in R v. Secretary of State for the Home Department, ex p. Amin [2001] EWHC 719 (Admin), at paras. 89 - 91, both of which concerned deaths in custody resulting from self-harm. In Keenan the European Court, at paragraphs 89 and 90, clearly considered that the Osman test applied to persons in custody who posed a risk to their lives through self-harm. And so did the Court of Appeal in upholding the decision of Hooper J. in ex p. Amin, [2002] EWCA Civ 390, at para. 31.

However, as Crane J noted, at paragraph 37 of his judgment in ex p. DF, the duty to protect the life of a prisoner against the risk of attacks by fellow prisoners involves different considerations from that owed to persons out in the community and, by inference, from that owed to a potentially suicidal prisoner. He identified them by reference to the Osman requirements:

“… First, the requirement that the authorities knew or ought to have known of the risk will usually be satisfied much more readily in relation to a prisoner, particularly one who has assisted the authorities, than in relation to a member of the community in general. Secondly, the authorities are in position to take measures to avoid any risk to an extent much greater than are the police in relation to a member of the community. Thirdly, the authorities are likely to be less inhibited by restraints imposed on the scope of their actions by the need to respect the human rights of others, since providing a protective regime is unlikely to affect the rights of others.”

With those countervailing differences in mind, Crane J., at paragraphs 38 and 121, held that in the circumstances with which he was concerned, namely risk to a prisoner’s life from the actions of fellow prisoners over a period of time subject to periodic review, the Osman test of “real and immediate risk” should be understood as a risk that is real throughout that period, not necessarily one that is any more imminent at its beginning than later.

“38 … the word ‘immediate’ should not be understood, in the context of decisions on admission to a protective regime, to mean that the threat will necessarily materialise in the very near future. That would be appropriate if a particular threat required urgent measures. However, the kind of risk to be assessed in this context is of harm to the prisoner as and when those wishing him harm find an opportunity. This case concerns a decision whether the risk is sufficient to justify entry to an available protective regime. That implies an assessment of the risk of harm within a period until the decision is reviewed. In my view the question to be asked is whether there is a real risk to the life of the prisoner if he is not admitted to a PWU, rather than some alternative regime, for whatever period is being considered. However, immediacy requires that the risk must be present and continuing. A real risk is one that is not simply a fear felt by the prisoner, but is disclosed by all the information available. …”

“121.

…. the question to be asked before entry to a PWU is whether there is a real and immediate risk to the life of the Claimant if he is not admitted to a PWU, rather than some alternative regime, for whatever period is being considered. Since a year is the normal period between reviews, that is the relevant period. Immediacy must be considered in that context. Where a prisoner has previously been properly admitted to a PWU, it is in my judgment incumbent on the Prison Service to satisfy itself that the risk is no longer such as to require admission. …. ”

Crane J’s interpretation of the Osman test as it applies to protective witness unit cases invites a return to the reasoning of the Court in R v. Lord Saville (No 2), in which it should be noted that Lord Phillips did not expressly endorse the Osman formulation of “a real and immediate risk” as one of general application or even as an appropriate tool for cases calling for police protection for persons in the community. He simply said, at paragraph 28 of the Court’s judgment, that it was not appropriate for the facts of the case in that appeal:

“In R v. Governor of Pentonville Prison, Ex p. Fernandez [1971] 1 WLR 987, after adumbrating the various phrases which he considered expressed the same degree of likelihood of risk, Lord Diplock referred, at p. 994, to the alternative of ‘applying, untrammelled by semantics, principles of common sense and common humanity’. We believe that there is much to commend that approach in the present case. The search for a phrase which encapsulates a threshold of risk which engages article 2 is a search for a chimera. … Of one thing we are quite clear. The degree of risk described as ‘real and immediate’ in Osman …., as used in that case, was a very high degree of risk calling for positive action from the authorities to protect life. It was ‘a real and immediate risk to the life of an identified individual or individuals from the criminal acts of a third party’ which was, or ought to have been, known to the authorities: p. 305, para. 116. Such a degree of risk is well above the threshold that will engage article 2 when the risk is attendant upon some action that an authority is contemplating putting into effect itself. It was not an appropriate one in the present context.”

Whatever the risk to a person’s life and whatever the mechanism that may give rise to it, the approach of Lord Diplock in Fernandez, adopted by the Court in Lord Saville (No. 2) seems to me, with respect, to be entirely logical and of general application. If a risk to life is not “real”, it is not a risk to life. If a risk to life is not “immediate” in the sense that it is not present at the time or during the period when it is claimed that a protective duty is owned by a public body, it is not a risk that can engage article 2. It is a future risk that may, at some later date do so. To be a candidate for engaging article 2, all that is needed is “a risk to life”. To engage it depends, in the circumstances of each case, on the degree of risk, which necessarily includes considerations of the nature of the threat, the protective means in being or proposed to counter it and the adequacy of those means.

The word “risk” in the general context of risk to life engaging article 2 is, as Lord Phillips said, one of common sense application to individual circumstances. In that general sense, it can and should be used without a qualifying adjective. In its application to different facts, there will inevitably be a temptation to resort to the seriousness or level or degree of the risk of occurrence, as distinct from the fatal consequence of occurrence, in respect of which a duty to provide public protection is claimed. But the seriousness, level or degree of occurrence is not a given or absolute quality. I would add to Lord Phillips’ observation in Lord Saville (No. 2) that the “[t]he search for a phrase which encapsulates a threshold of risk which engages Article 2 is a search for a chimera”, that it could also be unhelpful to attempt to identify some sort of broad tariff of thresholds of risk for different categories of case. I note that, although Crane J, in his helpful analysis, at paragraphs 22 to 37 of his judgment, identified, as a generality, different levels of risk for those in the community and those in prisons, he immediately qualified it in the circumstances of the case before him by reference to the level of protection offered and reasonably available to meet it.

The question for the Prison Service was whether a decision by it not to continue the appellant in the protected witness regime would infringe his right to life. The question for the Court is different, but in the context of this human right and the circumstances of the case, possibly only marginally. Any potential interference with the right to life requires the most anxious scrutiny by the Court, since it is the most fundamental of human rights. Ouseley J. said, at paragraph 89 of his judgment, that it fell to be judged by a more hands-on test than that of Wednesbury irrationality. More hands-on maybe, but one which despite “[the] long trek away from Wednesbury irrationality … as the only appropriate test where human rights are involved”, still seemingly stops short of a merits review by the Court - as the House of Lords has just held, in the context of Article 10, in R v. BBC, ex p. Prolife Alliance [2003] UKHL 23, see, in particular, per Lord Walker of Gestingthorpe at paras. 131-144.

The starting point in this case, as Mr. Jay has submitted, is that the right to life under Article 2 is unqualified. It is not, in contrast to Article 10, defeasible by any countervailing public interest considerations; so the principle of proportionality is not in play. The way in which the Court in human rights cases up to now have stopped short of undertaking a merits review may be expressed in the conventional manner adopted by Ouseley J., in paragraph 90 of his judgment, of showing some deference to the decision-maker – here the Prison Service. Or it may be expressed, as Lord Hoffmann prefers (see Prolife, at paras. 74-76) as a conclusion of law that the decision in question “is within the proper competence” of the decision-maker. Whichever is the more appropriate formulation of the test, the law in this area, as Lord Walker indicated at the end of his speech in Prolife, is still developing, and the trek away from Wednesbury, particularly where the right to life is concerned, may have further to go.

In the present state of development of the law, we consider that, despite the fundamental and unqualified nature of the right to life, it is still appropriate to show some deference to and/or to recognise the special competence of the Prison Service in making a decision going to the safety of an inmate’s life. However, the degree of deference to, and/or of recognition of the special competence of, the decision-maker is less and, correspondingly, the intensity of the Court’s review is greater – perhaps greatest in an Article 2 case – than for those human rights where the Convention requires a balance to be struck; see per Lord Hope in R v. DPP, ex p. Kebilene [2000] 2 AC 326, at para. 80. As to how intense the review should be, the nearest and, for the moment, necessarily imprecise guidance for all human rights cases remains that employed by Lord Walker in Prolife, at para. 139:

“… the Court’s task is not to substitute its own view for that of the … [decision makers], but to review their decision with an intensity appropriate to all the circumstances of the case.”

Hers, as Ouseley J. observed, the Prison Service, with its experience of prison conditions, options and the relative efficacy of different protective regimes and measures - and the Police, with their expert and close knowledge of the level of risk - are generally better placed than the Court to assess the risk to life in such a context. He also rightly took into account the alteration in circumstances that had created the risk, reducing it, on the police assessment, from a “high” to a “medium” or moderate risk, and the proposals to provide a measure of protection within the mainstream prison system to meet that reduced level of risk.

The original assessment of risk giving rise to the decision to admit the appellant to the protected witness unit had been based on the premise that he was to give evidence against the subject of his witness statement to the police. In the event, the police had not instituted a prosecution against the man and, as a result, no longer required the appellant to give evidence against him. The police had advised Mr. Golds that, even if that man or his associates came to learn of the assistance the appellant had proffered, they would take “a business-like” approach and would not court police attention unnecessarily. They had also advised him that there had been no follow-up to the loss of the file or the other matters of concern to the appellant, to which I have referred in paragraph 19 of this judgment, and that there was no indication of any current threat to him.

It is true that Mr. Ayers, the probation officer seconded to the Prison Service, had expressed a contrary and broader view to Mr. Golds based on the appellant’s feeling of general vulnerability to attacks from fellow inmates who might suspect that he had been an informant if he were returned to mainstream prison conditions without any ready account of how he had spent the previous 18 months in custody. However, in expressing that view, Mr. Ayers had largely relied on what the appellant had told him. And, as Ouseley J. observed at paragraph 95 of his judgement, those matters, including the fact that he had given evidence against his co-accused, were not reasons for the original decision to admit him to a protected witness unit. Whilst Mr. Golds, when making his original decision in December 2001, may not have responded specifically to Mr. Ayers’ concerns about the appellant’s general vulnerability if returned to mainstream prison conditions, he must, as I have said, have taken them into account when confirming his decision in January 2002.

In any event, after further reconsideration in May 2002, Mr. Golds clearly regarded Mr. Ayers’ concern as of little weight in the decision to be made - which was not as between protection and no protection - but as between the high level of protection provided by protected witness unit and the lesser but special protective arrangements in mainstream prison conditions that he proposed in his witness statement. Whilst, as Schiemann LJ observed in ex p. S, the Court should, of course, examine with care any post-decision rationalisations of this sort, especially after judicial review proceedings have begun, it would be surprising if the level of “background” risk of that sort were much different from that of, say, sex offenders, for whom less secure provision is routinely made. There was no reason for Ouseley J. to doubt that the Prison Service would do what Mr. Golds proposed, and it is plain from paragraphs 87 and 103-105 of his judgment that he had no misgivings on that account. Further, since judicial review is a discretionary remedy, it would have made no sense for the Judge to have ignored what the Prison Service actually propose for the appellant and to consider himself bound by the original decision, even if it had been flawed for the reason urged by Mr. Clayton.

In my view and in the light of all those considerations, the Prison Service’s decision, when subjected to the most anxious scrutiny by this Court, is not just well within the bounds of what was reasonable, given the importance of the human right in play. It justifies the “deference” which the Court should accord it and demonstrates the competence the Court would expect of it in considering and reaching so critical a decision. For what it is worth, I do not share the uncertainty of Ouseley J as to how the Court should decide the matter if the law had developed to the stage where it could make a decision on the merits for itself. I would have reached the same conclusion as the Prison Service, for the reasons it gave.

I have put a question mark under this heading since, although unfairness is a necessary element of a complaint of breach of legitimate expectation, and breach of legitimate expectation is one form or aspect of abuse of power, neither unfairness nor abuse of power, for the reasons I have given, is sufficient on its own to found a claim under that head. And, whilst unfairness and/or abuse of power may often be present in an article 2 complaint, the latter provides a more factually distinct basis for the remedy it provides. Mr. Clayton did not put these matters to Ouseley J. as a distinct ground of legal complaint, although they underlay his submissions on the ground of legitimate expectation. However, he gave them more of a life of their own in his submissions to this Court. The way in which he put it was that, regardless of the legitimacy of the appellant’s expectation, it was unfair and hence an abuse of power for the Prison Service to resile from the alleged promises or representations made by the appellant’s police handlers. First, the police handlers, were not required, in the Police Circular or elsewhere, to inform a prospective informant that if he was given protected witness status, it would be subject to periodic review throughout his sentence. Second, the Prison Service had taken no steps, by warning the appellant or otherwise, to ensure that he knew that police handlers had no authority to commit the Prison Service to promises of the sort that he alleged they made here, in particular, that once admitted to a protected witness unit, he would remain in it for the whole of his prison sentence. Third, and as part of that complaint, he said that it was unfair to the appellant for the Prison Service, knowing that police officers would or might disregard the instructions in paragraphs 5 and/or 18 of the Annex to the Police Circular (see paragraph 14 above), not to have warned the appellant, before he irrevocably exposed himself to risk, that his protected witness status would be periodically reviewed by the Prison Service throughout his sentence. Fourth, he suggested that the Prison Service should have provided some safeguards, say, in the form of a written agreement of confirmation between the Police and potential informants that their police handlers had complied with the procedures and for some system whereby the Service could check that that had been done. In short, his complaint was that the Prison Service took no steps to check whether the appellant’s police handlers had complied with the instructions in the Police Circular.

At the heart of Mr. Clayton’s submissions under this head lay the scope for unfairness to a would-be informant in the tri-partite relationship between him, the Police and the Prison Service when he would regard it as a bi-partite relationship between him and “them”, the establishment. As Mr. Clayton said, it is a somewhat untidy decision-making structure.

Paragraph 18 of the Annex, it will be remembered, requires the Police to tell potential informants that they should expect no advantage in their conditions of detention as a result of providing information and recommends that the Police authorising officer should consider drawing up a written contract between the Police and the prisoner “as a means of protecting all parties”. It should be noted that that recommendation - which is all it is - does not extend to the inclusion of the Prison Service as one of the “contracting” parties.

Mr. Jay’s response to this reference to “contracts”, was that they and any machinery that might be introduced by the Prison Service to check whether they have been made would have little practical utility in overcoming any tendency of police handlers to give oral assurances to would-be informants that they have no authority to give and that could not, in any event, be honoured by the Prison Service. And, he said, rightly in my view, there was no basis in law or on the facts of this case for imposing a duty on the Prison Service to make such a check.

Depending on the circumstances, an unauthorised assurance by a police officer could be a relevant consideration for the Prison Service to take into account when later making decisions such as this, particularly if a prisoner/informant has reasonably acted upon that promise to his detriment. As Mr. Jay observed in acknowledging such a possibility, this is a different point from that of legitimate expectation. However, there was no evidence before Ouseley J that the Prison Service was aware that the appellant’s police handlers had given such assurances to him or that it knew or should have known, if it is the case, that police handlers generally and in breach of their instructions give such assurances to those being considered for protected witness status. Nor, given the nature of the tri-partite arrangement to which I have referred, can I see any reasonable basis on which the Prison Service could have been expected to know of or expect such breaches by police officers of their instructions or effectively to check for them. I agree with Ouseley J. that, under whatever legal label the matter falls to be decided, Mr. Gold’s decision cannot be faulted on the law or on the facts.

Nevertheless, in the light of the problems that this case has identified, I share the unease of Ouseley J. about the outcome, certainly for those who may find themselves in the appellant’s position in the future. In the final paragraph of his judgment, paragraph 108, he said:

“I add, from the claimant’s point of view, the decision of Mr. Golds is undoubtedly seen as a betrayal and on the assumption that the claimant is right about what was said to him, it can readily be seen to be unfair from his perspective. But that and any impact which that might have on the operation of the Bloggs system, including the disclosure of the identity of its inmates or the effectiveness of informants is not a matter for me. It is a matter for the police and Prison Service together. No challenge to Mr. Golds’ decision has been raised on that basis. I am only concerned insofar as those feelings of the claimant have any effect upon the risk to which he is subject. … I do not consider that the obligations of common sense and humanity, which are directly relevant to the assessment of the degree of risk and the protective measures required, can mean that the PWU should contain someone when that protection is not warranted. I say that even though it is impossible to avoid a strong sense that, on the factual assumptions which I make, the claimant has a justified sense of grievance which it would at least be desirable to assuage. But from a legal point of view such assuaging has to be set in a principled legal framework which here I cannot find.”

I agree and consider that Police Authorities and the Prison Service should examine together possibilities for improving the protected witness system, for which they are both responsible in their different and complementary ways, so as to remove or reduce the future scope for any misrepresentation by police officers to potential informants in custody of the effect of the grant of protected witness status.

For the reasons I have given, I would dismiss the appeal.