R v Secretary of State for the Home Department

This appeal is concerned with the standards to be applied by a court when it is asked to conduct a judicial review of the contents of a policy document or statement of practice issued by the Government. We will refer to these as “policies”.

It is a familiar feature of public law that Ministers and other public authorities often have wide discretionary powers to exercise. Usually these are conferred by statute, but in the case of Ministers they may derive from the common law or prerogative powers of the Crown, which fall to be exercised by them or on their advice. Where public authorities have wide discretionary powers, they may find it helpful to promulgate policy documents to give guidance about how they may use those powers in practice. Policies may promote a number of objectives. In particular, where a number of officials all have to exercise the same discretionary powers in a stream of individual cases which come before them, a policy may provide them with guidance so that they apply the powers in similar ways and the risk of arbitrary or capricious differences of outcomes is reduced. If placed in the public domain, policies can help individuals to understand how discretionary powers are likely to be exercised in their situations and can provide standards against which public authorities can be held to account. In all these ways, policies can be an important tool in promoting good administration.

Policies are different from law. They do not create legal rights as such. In the case of policies in relation to the exercise of statutory discretionary powers, it is unlawful for a public authority to fetter the discretion conferred on it by statute by applying a policy rigidly and without being willing to consider whether it should not be followed in the particular case. However, in an important development in public law in the last decades, the courts have given policies greater legal effect. In certain circumstances a policy may give rise to a legitimate expectation that a public authority will follow a particular procedure before taking a decision and it may give rise to a legitimate expectation that the authority will confer a particular substantive benefit when it does decide how to exercise its discretion. In these cases, the courts will give effect to the legitimate expectation unless the authority can show that departure from its policy is justified as a proportionate way of promoting some countervailing public interest. If the policy is not made public, and an affected individual is unaware of its relevance to his case and in that sense has no actual expectation arising from it, the authority may still be required to comply with it unless able to justify departing from it: Mandalia v Secretary of State for the Home Department [2015] UKSC 59; [2015] 1 WLR 4546. Under some conditions the holder of a discretionary power may be required to formulate a policy and to publish it: R (Lumba) v Secretary of State for the Home Department [2011] UKSC 12; [2012] 1 AC 245. Thus, policies have moved increasingly centre stage in public law.

In a parallel development, and perhaps reflecting the increased importance of policies, there has been an increase in judicial review of the contents of policies. The present appeal is concerned with the standards to be applied by the courts when called upon to review the contents of policies. The policy of which the appellant seeks judicial review is the Child Sex Offender Disclosure Scheme Guidance (“the Guidance”) issued by the Secretary of State in exercise of her common law powers.

The appellant is a convicted sex offender. In March 1992 he was convicted of two offences of indecent assault on a child aged under 14 and sentenced to two years’ imprisonment. On 26 April 1996 he was convicted of three offences of indecent assault on a child under 16 and sentenced to four and a half years’ imprisonment. As a result of his offending, his name is on the Sex Offenders Register and he is subject to the notification requirements under the Sexual Offences Act 2003 for an indefinite period. He has been convicted of no further offences. The appellant has suffered violence and harassment when people have become aware of his convictions. He maintains that the Guidance is unlawful in that it does not make sufficient provision for the police to consult him before making disclosure about his offences to a member of the public who makes inquiry about him in circumstances where the appellant is in contact with children.

The present appeal was heard by the same constitution of the court which considered the appeal in R (BF (Eritrea)) v Secretary of State for the Home Department [2021] UKSC 38 which is also concerned with the correct approach to judicial review of policies.

The Child Sex Offender Disclosure Scheme (“the CSOD Scheme”) was set up by the Secretary of State in 2010 to assist in co-ordinating the approach of police forces when asked by members of the public for information about persons dealing with children, with a view to gaining information whether they have convictions for sex offences involving children. The scheme was developed from pilot arrangements tried out in several police areas. The CSOD Scheme is set out in the Guidance, which explains the approach police forces should adopt when faced with requests for such information. The Guidance was issued by the Secretary of State in exercise of her common law powers. It has no statutory force. Police forces are free to decide whether to participate in the CSOD Scheme or not. In 2011 South Yorkshire Police, the police force for the area where the appellant lives, adopted the CSOD Scheme for their area.

The CSOD Scheme sits alongside other procedures governing the management in the community of child sex offenders who have completed their sentences and been released. Those procedures are comprised in multi-agency public protection arrangements (“MAPPA”) put in place under sections 325 to 327 of the Criminal Justice Act 2003 (“the CJA 2003”) and pursuant to statutory guidance issued by the Secretary of State under section 325(8) of that Act (“the MAPPA guidance”). The MAPPA regime is designed to secure cooperation between responsible authorities, including the police, the probation service and the prison service, in the assessment and management of risks posed by violent and sexual offenders. Under section 325(8A), MAPPA authorities are required to have regard to the MAPPA guidance when exercising their functions.

In 2008, section 327A was inserted into the CJA 2003, imposing a legal duty on MAPPA authorities to consider the disclosure of information about child sex offenders to the public. It provides:

The appellant brought a challenge to the Guidance in 2012. That challenge was successful in the Divisional Court: R (X) v Secretary of State for the Home Department [2012] EWHC 2954 (Admin); [2013] 1 WLR 2638. The Divisional Court held that the Guidance, as it then stood, was unlawful. That was because it did not include a requirement that the police should consider whether any person about whom disclosure might be made should be asked whether he wished to make representations. It did not comply in that respect with the applicable legal principles and might well result in the protection afforded by article 8 of the European Convention on Human Rights (“the ECHR”), as implemented in domestic law by the Human Rights Act 1998 (“the HRA”), being rendered nugatory. The court therefore granted a declaration as follows:

As a result of these rulings, paragraph 2.2 of the Guidance was revised (in its current form, it is quoted below) and the Secretary of State inserted a new paragraph 5.5.4 into the scheme with effect from 5 March 2013. Paragraph 5.5.4 provides:

According to the Guidance, as amended, the CSOD Scheme allows members of the public to make an application to the police about a person (the subject) who has contact with children for disclosure of information about previous convictions the subject may have and other material which might affect the safety of those children. The Guidance explains that a disclosure will only be made pursuant to the scheme if a specific child or specific children are at real risk of harm. The scheme makes provision for cooperation between police forces in relation to dealing with applications. Section 2 of the Guidance explains the aims of the CSOD Scheme. Paragraph 2.1 of the Guidance explains that the scheme does not provide for automatic disclosure of child sexual offender details to the general public, since that could encourage offenders to go missing for fear of reprisals and therefore put children at greater risk of harm. Paragraph 2.2 states:

The Guidance reminds police decision-makers that all stages in the CSOD Scheme in respect of an application for disclosure of information should be followed. It sets out the information to be obtained from an applicant and what checks should be completed, suggests forms of wording for communications with an applicant and a subject. It makes recommendations regarding the knowledge and experience of staff undertaking the relevant stages of the decision-making process. In paragraph 2.6 the Guidance explains that the CSOD Scheme builds on existing procedures, in particular those set out in the statutory MAPPA guidance issued pursuant to section 325(8) of the CJA 2003 and the statutory duty contained in section 327A of that Act.

Section 3 of the Guidance sets out the principles underlying the CSOD Scheme. Paragraph 3.3 states that no disclosures should be made to members of the public without following all the appropriate stages of the Guidance, unless there is an immediate risk of harm to a child. Paragraph 3.5 indicates that MAPPA and other safeguarding children procedures regarding making a disclosure remain unchanged and states, “disclosure must still be lawful. In particular the agencies must have the power to disclose the information (eg under the common law) and disclosure must comply with the Human Rights Act 1998 or Data Protection Act 1998.” The CSOD Scheme does not replace other arrangements, including those for Criminal Records Bureau (now the Disclosure and Barring Service) checks. Paragraph 3.8 emphasises the importance of engagement with registered sex offenders in a police force’s area to reassure them about the operation of the scheme to minimise the risk of them going underground and not complying with their supervision requirements. Section 4 of the Guidance sets out relevant definitions and explains how the CSOD Scheme works where multiple police forces are involved in considering an application (for example, where a subject is covered by MAPPA arrangements, it is stated that “the MAPPA area that owns the subject will have primacy”).

Section 5 of the Guidance sets out the procedure to be followed when an application for disclosure is received:

The decision-making form appended to the Guidance, as relevant for a case which raises “concerns”, includes a section for consideration to be given to engaging the MAPPA process. It also includes a section for setting out the decision, which requires two boxes to be ticked to indicate the police staff member is satisfied that any disclosure is made pursuant to legal authority (“Common law power to disclose information about a person’s convictions where there is a need for such disclosure to protect the public from crime” and “Disclosure would not be in breach of the Human Rights Act 1998, Data Protection Act 1998 or a breach of confidence”). This section also requires an explanation to be entered for “Justification for disclosure”. This should involve consideration of what will be disclosed, to whom it will be disclosed and whether the subject is going to be informed of the disclosure, and a prompt that “[a]ll of the above must clearly state how they are lawful, necessary and proportionate”.

Under the MAPPA guidance, as part of the process of managing the risk posed by an offender the relevant responsible authority must consider whether information about the offender should be disclosed to a third party. The guidance requires that “[a]ll disclosures of information about MAPPA offenders to third parties comply with the law, are necessary and are proportionate”. Paragraph 10.12 states that “[t]he likelihood and degree of harm that might arise as a result of the disclosure, including the potential impact on the offender, must be assessed. Information should be disclosed only where this is a necessary and proportionate step to protect the public …”. Paragraph 10.20 states, “[c]onsideration must be given to seeking representations from the offender before a decision is made to disclose, in order to ensure that all of the information necessary to make a properly informed decision is available. Seeking their representations should be the norm, but there might be occasions when it is not possible or safe to do so …”.

The police have a general power to hold and disclose information for the purposes of performing their functions to uphold the law and protect the public. The exercise of that power is subject to various legal obligations. Those relied on in the present case are obligations arising under the common law and pursuant to the HRA.

It is common ground that the police have a common law duty to act fairly when deciding whether to disclose information about an offender, including under the CSOD Scheme and the MAPPA regime. Disclosure to members of the public of information about a subject’s previous convictions or intelligence relating to his risk of offending may well have an adverse effect on him and this means that, depending on the circumstances, the duty of fairness may require that he be given an opportunity to make representations before a decision is made to make such disclosure. For a well-known statement of the general common law principle, see R v Secretary of State for the Home Department, Ex p Doody [1994] 1 AC 531, 560D-G (Lord Mustill) (“Doody”). The implied obligation of fairness takes effect subject to the public interest in the due performance of the police’s general functions and there may be countervailing factors which mean that no such requirement of consultation with the subject arises. In particular, as Mr Moffett QC for the Secretary of State points out, the CSOD Scheme applies in the context of the need to safeguard children, in relation to which the police have responsibility to seek to ensure that the protection of the criminal law is effective.

The practical effect of the duty of fairness in relation to consideration of disclosure of previous convictions for sexual offences against children was considered in R v Chief Constable of the North Wales Police, Ex p Thorpe [1999] QB 396. In the Divisional Court, Lord Bingham of Cornhill CJ stated (p 409F-G) and then endorsed (pp 409H-411B) three principles which should govern a policy adopted by the police to guide its conduct when considering whether to disclose information to a member of the public:

The argument in the Court of Appeal was directed to different issues, not canvassed in the Divisional Court. The appellants’ case, relying in particular on Doody, was that the disclosure by the police was unlawful because they had been treated in a procedurally unfair manner because they had not been given an opportunity to comment on material relevant to the risk they posed to the public before the decision to disclose was taken (pp 426-427). The submissions advanced by the Secretary of State were to the effect that there are cases where the duty to act fairly may make it necessary, “so as to ensure as far as possible that the police are acting on accurate information and so as to ensure the necessary degree of fairness, to afford individuals in the position of the applicants some opportunity to comment” (p 427); but this would depend on all the circumstances and there might be a range of factors which meant that the police could proceed to make a decision without doing so. The Court of Appeal endorsed this general approach, observing that “[e]ach case must be judged on its own facts” and noting that “the subject of the possible disclosure will often be in the best position to provide information which will be valuable when assessing the risk” he poses (p 428). In that case, the appellants should have been given an opportunity to comment before the police disclosed their previous convictions; no reasons had been put forward to explain why this could not have been done.

Under section 6(1) of the HRA, the police, as a public authority, are obliged to act in a way which is compatible with the Convention rights of individuals as derived from the ECHR and set out in Schedule 1 to the HRA. Article 8 is relevant in this context. It provides:

The present proceedings involve a new challenge by the appellant to the Guidance. He maintains that paragraph 5.5.4 (see para 11 above) does not go far enough in giving guidance regarding the circumstances in which a police force, approached by a member of the public for information about a person about whom there are concerns in the relevant sense, is obliged in law to seek representations from the subject of the request before making disclosure of information about him. Mr Southey submits that the Guidance is unlawful because it gives rise to an unacceptable risk of unfairness and breach of article 8; does not meet the standards of clarity, predictability and accessibility inherent in the “in accordance with the law” rubric in article 8(2); and fails to provide for a subject to be able to apply to be exempted from the CSOD Scheme.

The appellant is subject to the MAPPA regime, so any decision regarding disclosure of information in relation to him will be taken by the relevant MAPPA authorities, having regard to the MAPPA guidance as they are required to do under section 325(8A). The appellant accepts that the MAPPA guidance is lawful. In particular, he accepts that the statement it contains at paragraph 10.20 (see para 17 above) regarding seeking representations from the subject of a disclosure request is correct. If followed, the appellant accepts that any disclosure request relating to him should be handled lawfully.

The appellant’s latest challenge to the Guidance setting out the CSOD Scheme might, therefore, have been dismissed as academic. However, there are some individuals with previous convictions for relevant offences who fall outside the MAPPA regime but are covered by the CSOD Scheme. The courts below treated this case as a suitable vehicle for the review of the Guidance and addressed the appellant’s challenge to it on its merits. We do the same.

At first instance, Dingemans J [2014] EWHC 4106 (Admin) dealt with a wide range of claims by the appellant. They became more narrowly focused on appeal. So far as concerned the CSOD Scheme, at para 72 the judge held that although the wording of paragraph 5.5.4 of the Guidance might be improved, it did not make the Guidance or the scheme unlawful.

The appellant appealed against this ruling to the Court of Appeal (Laws LJ, Hamblen LJ and Langstaff J) [2016] EWCA Civ 597. He was now represented by Mr Southey. Mr Southey’s submissions on the appeal were somewhat different from those presented at first instance. His contentions were, first, that the Guidance is not “in accordance with the law” within the meaning of article 8(2) because there is no independent supervisory authority or review body outside the police to review disclosure and to consider whether a criminal subject to the CSOD Scheme should be exempted from disclosure. Secondly, the Guidance should include a presumption that a subject ought to be consulted and have the opportunity to make representations; and, thirdly, paragraph 2.2 of the Guidance was objectionable on article 8 grounds because it contained a presumption in favour of disclosure in certain circumstances. The Court of Appeal dismissed the appeal in a judgment of Laws LJ with which the other members of the court agreed. The third point does not arise on the appeal to this court. While dealing with the first point, Laws LJ held that there was no requirement of the common law or under article 8 that a subject should have an opportunity to seek exemption from the CSOD Scheme (para 26). As regards the second point, by reference to R (Tabbakh) v Staffordshire and West Midlands Probation Trust [2014] EWCA Civ 827; [2014] 1 WLR 4620 (“Tabbakh”), Laws LJ ruled that the test in domestic law for the legality of a public scheme in relation to a complaint of a failure to provide proper opportunities for affected persons to make representations is whether the scheme is inherently unfair. He held that the Guidance was not unlawful by reference to that test. Paragraph 5.5.4 read with paragraph 5.6.15(iii) of the Guidance satisfied the legal standards the court had to apply.

The appellant now appeals to this court with permission to rely on two grounds, namely that the Guidance is unlawful because (i) it fails to recognise and reflect the importance of consulting with people who are at risk of suffering a violation of their article 8 rights by reason of disclosure and (ii) this means that there is a significant and/or unacceptable risk of a breach of article 8 and/or the common law. Both sides made detailed submissions about the legal test to be applied on judicial review of a statement of policy such as that set out in the Guidance.

Mr Southey submits that the Guidance is not sufficiently specific in the directions it gives regarding when the subject of a disclosure request should be given an opportunity to make representations before a decision is made to give disclosure. He says that this creates a risk that a decision-maker may not provide such an opportunity when required in law to do so. This is unacceptable and means that the Guidance is unlawful. Also, by reason of its general lack of specificity, Mr Southey says that some statements in the Guidance could be regarded as positively misleading.

Although the Court of Appeal applied a test of inherent unfairness as illustrated by Tabbakh, we consider that it is appropriate to begin the analysis with the earlier decision of the House of Lords in Gillick v West Norfolk and Wisbech Area Health Authority [1986] AC 112 (“Gillick”). In that case, the Department of Health and Social Security issued guidance to health authorities on family planning services which included a section on contraceptive advice and treatment for young people. It stated that such advice and treatment should be available for people of all ages, but that for children under the age of 16 attempts would be made to persuade them to involve their parent or guardian at the earliest stage of consultation and that it would be most unusual to provide such advice or treatment without parental consent. However, it noted that to abandon the principle of confidentiality between doctor and patient in respect of children under 16 might cause them not to seek professional advice at all, thereby exposing them to risks such as pregnancy and sexually transmitted diseases. It stated that in exceptional cases it was for a doctor exercising his clinical judgment to decide whether to prescribe contraception. The claimant, the mother of girls under 16, objected to this. She wrote to her area health authority seeking an assurance that no contraceptive advice or treatment would be given to her daughters while under 16 without her knowledge and consent. The health authority refused to give such an assurance and stated that in accordance with the guidance the final decision must be for the doctor’s clinical judgment. The claimant brought proceedings by writ (rather than by way of the procedure for judicial review) for, among other things, a declaration that the guidance gave advice which was unlawful and wrong. The claim failed at first instance but succeeded in the Court of Appeal on the grounds that a girl under 16 was incapable either of consenting to treatment or of validly requiring a doctor not to seek the consent of her parents; and that the guidance was contrary to law in that any doctor who treated a girl under 16 without the consent of her parent or guardian, other than in an emergency, would be infringing their parental rights. By a majority (Lord Fraser of Tullybelton, Lord Scarman and Lord Bridge of Harwich; Lord Brandon of Oakbrook and Lord Templeman dissenting), the House of Lords allowed an appeal by the department. The decision is well known as the leading authority regarding the capacity of a child under 16 with sufficient maturity and understanding to consent to medical examination and treatment and the lawfulness of a doctor proceeding to examine and treat such a child without the knowledge or consent of her parent or guardian. The Appellate Committee also decided the approach to be adopted in what was, in substance, a judicial review of a policy statement issued by the department.

Lord Fraser set out a series of conditions to be met for a doctor’s decision to give contraceptive advice and treatment to a girl under 16 without informing her parents to be lawful. He emphasised that this meant that doctors did not have a licence to disregard the wishes of parents “whenever they find it convenient to do so” (p 174B-E). He observed that, as had been common ground in the Court of Appeal, whether a doctor who followed the guidelines in the policy would commit an offence contrary to section 28 of the Sexual Offences Act 1956 by aiding and abetting the commission of unlawful sexual intercourse would depend on the circumstances of the case and in particular on the doctor’s intentions, it being unlikely that an offence would be committed if he honestly intended to act in the best interests of the girl (pp 174H-175A).

The guidelines were at a much higher level of generality than the law as stated by Lord Fraser. They referred only to the possibility that a doctor might in exceptional circumstances, according to his clinical judgment, give advice and treatment to a girl under 16 without the knowledge or consent of her parents. They did not set out guidance regarding the conditions to be satisfied, as explained by Lord Fraser. Accordingly, he was addressing the lawfulness of the policy in circumstances where it did not set out an exhaustive or detailed statement of the law (like a textbook or a judgment). It did not provide guidance which, if followed, would compel a doctor to act lawfully in every case and hence eliminate any risk that he might act unlawfully.

Lord Scarman agreed with Lord Fraser’s speech and made comments of his own. Lord Scarman analysed the case (p 177D-E) as one involving an allegation that in issuing guidance which, if followed, would result in unlawful acts the department had acted in a Wednesbury irrational way (Associated Provincial Picture Houses Ltd v Wednesbury Corpn [1948] 1 KB 223). Having set out the guidance, Lord Scarman said (p 180D-F):

Thus, Lord Scarman was explicit that it was not the role of policy guidance to eliminate all uncertainty regarding its application and all risk of legal errors by doctors. It was to be read objectively, having regard to the intended audience. (We would add that this is in line with a substantial body of authority subsequent to Gillick in which policies have been given legal significance either as relevant considerations to be taken into account by a decision-maker or as the basis for an enforceable legitimate expectation: see eg Tesco Stores Ltd v Dundee City Council [2012] UKSC 13; [2012] PTSR 983, para 18 per Lord Reed, R (Raissi) v Secretary of State for the Home Department [2008] EWCA Civ 72; [2008] QB 836, paras 118-123 and Paponette v Attorney General of Trinidad and Tobago [2010] UKPC 32; [2012] 1 AC 1, para 30). Contrary to a submission made by Mr Southey, the drafter of a policy statement is not required to imagine whether anyone might misread the policy and then to draft it to eliminate that risk. As Lord Scarman explained, it was only if the guidance, on a reasonable reading of it, positively encouraged a doctor to think that he was authorised to prescribe without the parent’s consent whenever he thought fit that it would be unlawful. The language he uses in the test he sets out, to ask whether the guidance “permits or encourages unlawful conduct”, has to be read in the light of this discussion. As correctly observed by Underhill LJ in R (Bayer plc) v NHS Darlington Clinical Commissioning Group [2020] EWCA Civ 449; [2020] PTSR 1153, para 200 (“Bayer”), a case we come back to below, in considering Lord Scarman’s formulation in Gillick of the relevant test, “‘permit’ must in this context mean something like ‘sanction’, ie positively approve”, not merely that a course of action is not forbidden.

Like Lord Fraser, Lord Scarman gave a far more detailed statement of the legal position (pp 188H-189E) than that set out in the guidance. Despite the absence of detail in the guidance, he held it to be lawful. This was on the basis that “the department’s guidance can be followed without involving the doctor in any infringement of parental right” (p 190B, emphasis added). The guidance did not lead doctors away from due compliance with their legal obligations, and this was sufficient for it to be lawful. The legal test applied was not that the guidance, if followed, must inevitably produce conduct on the part of doctors which would be lawful.

Also in the majority, Lord Bridge was troubled by the basis on which judicial review could be sought of the guidance. He pointed out that it had no statutory force, was purely advisory in character and doctors were, as a matter of law, in no way bound by it; he therefore did not agree that the Wednesbury test provided an appropriate framework for the court in this context (pp 192B-193A). Instead, he explained the court’s power of review in this way: “if a government department, in a field of administration in which it exercises responsibility, promulgates in a public document, albeit non-statutory in form, advice which is erroneous in law, then the court … has jurisdiction to correct the error of law by an appropriate declaration” (p 193G). What would make the advice in the guidance erroneous in law? The claimant would have had to show that the law absolutely prohibited the prescription of contraception for a girl under 16 without parental consent or the order of the court (p 194D), which she could not do. The claimant had to be able to show that the guidance would, according to its terms, positively lead doctors to commit unlawful acts. The fact that the guidance did not give a full account of the legal conditions under which they could lawfully prescribe contraception to girls under 16 did not show that it was erroneous in the requisite sense.

Lord Brandon believed that the provision of contraception to girls under 16 could never be lawful (p 199B). He did not address the legal test to be applied, since on any view the department’s guidance encouraged doctors to provide contraception to girls under 16 in some cases and would therefore be unlawful according to the approach adopted by the majority. Lord Templeman’s view was that a girl under 16 is not competent to agree to medical examination and treatment (p 204A-B, p 205B). He considered that the guidance was unlawful in stating that it was a matter for the clinical judgment of the doctor to prescribe contraception without the knowledge or consent of a parent, whereas that would constitute an unlawful interference with the rights of the parent (p 205E-F). As to the basis for intervention by a court, Lord Templeman took the same view as Lord Bridge. The guidance was defective because it contained an error of law. It was not a question whether the department was exercising a discretion in a reasonable way, “but whether by mistake of law the [department], a public authority, purports by the memorandum to authorise or approve an unlawful interference with parental rights” (p 206E-F). In substance, this approach is the same as that adopted by the majority.

In our view, Gillick sets out the test to be applied. It is best encapsulated in the formulation by Lord Scarman at p 182F (reading the word “permits” in the proper way as “sanction” or “positively approve”) and by adapting Lord Templeman’s words: does the policy in question authorise or approve unlawful conduct by those to whom it is directed? So far as the basis for intervention by a court is concerned, we respectfully consider that Lord Bridge and Lord Templeman were correct in their analysis that it is not a matter of rationality, but rather that the court will intervene when a public authority has, by issuing a policy, positively authorised or approved unlawful conduct by others. In that sort of case, it can be said that the public authority has acted unlawfully by undermining the rule of law in a direct and unjustified way. In this limited but important sense, public authorities have a general duty not to induce violations of the law by others.

The approach to be derived from Gillick is further supported by consideration of the role which policies are intended to play in the law. They constitute guidance issued as a matter of discretion by a public authority to assist in the performance of public duties. They are issued to promote practical objectives thought appropriate by the public authority. They come in many forms and may be more or less detailed and directive depending on what a public authority is seeking to achieve by issuing one. There is often no obligation in public law for an authority to promulgate any policy and there is no obligation, when it does promulgate a policy, for it to take the form of a detailed and comprehensive statement of the law in a particular area, equivalent to a textbook or the judgment of a court. Since there is no such obligation, there is no basis on which a court can strike down a policy which fails to meet that standard. The principled basis for intervention by a court is much narrower, as we have set out above.

There are further reasons which indicate that this is the appropriate standard. If the test were more demanding there would be a practical disincentive for public authorities to issue policy statements for fear that they might be drawn into litigation on the basis that they were not sufficiently detailed or comprehensive. This would be contrary to the public interest, since policies often serve useful functions in promoting good administration. Or public authorities might find themselves having to invest large sums on legal advice to produce textbook standard statements of the law which are not in fact required to achieve the practical objectives the authority might have in view. Also, if the test were of the nature for which Mr Southey contends, the courts would be drawn into reviewing and criticising the drafting of policies to an excessive degree. In effect they would have a revising role thrust upon them requiring them to produce elaborate statements of the law to deal with hypothetical cases which might arise within the scope of a policy. Such a role for the courts cannot be justified. Their resources ought not to be taken up on such an exercise and it would be contrary to the strong imperative that courts decide actual cases rather than address academic questions of law.

The test set out in Gillick is straightforward to apply. It calls for a comparison of what the relevant law requires and what a policy statement says regarding what a person should do. If the policy directs them to act in a way which contradicts the law it is unlawful. The courts are well placed to make a comparison of normative statements in the law and in the policy, as objectively construed. The test does not depend on a statistical analysis of the extent to which relevant actors might or might not fail to comply with their legal obligations: see also our judgment in BF (Eritrea).

Assessed by reference to the test in Gillick, the Guidance (in the version in issue in these proceedings) is clearly lawful. It informs police decision-makers that before making a disclosure they should consider whether to seek representations from the subject. That is in accordance with, and in no way contradicts, their legal obligations under the common law and article 8 of the ECHR. The Guidance is not defective, still less unlawful, because it does not spell out in fine detail how decision-makers should assess whether to seek representations in a particular case. As in Gillick, so also in this case it was not incumbent on the Secretary of State in issuing the Guidance to eliminate every legal uncertainty which might arise in relation to decisions falling within its scope. In fact, the Guidance, including the form appended to it for recording decisions, specifically reminds decision-makers that they should satisfy themselves that a disclosure decision will satisfy the common law requirement of fairness and the requirements of article 8. Contrary to Mr Southey’s contention, when reading the Guidance as a whole no part of it can fairly be construed as giving a misleading direction.

There will be cases where the application of the Gillick test for lawfulness of a policy may be less clear than it is here. The first claim brought by the appellant to challenge the Guidance is an example. In its original form, the Guidance did not tell decision-makers to consider seeking representations from a subject before a disclosure to the public, but nor did it tell them not to. However, reading the Guidance as a whole, it was clearly intended to set out for decision-makers a reasonably complete decision-making procedure to be followed, so in our view the Divisional Court was right to hold that, read objectively, it misdirected decision-makers as to how they should proceed, by implicitly indicating that they did not have to invite representations whereas in many cases they had a legal obligation to do so.

The decision of the Court of Appeal in Bayer is instructive. In that case, with a view to limiting the cost of medical care, the defendant clinical commissioning groups adopted a policy on the choices regarding treatments to be offered to patients by NHS trusts for a particular eye complaint. The policy was to use so far as possible one product (Avastin) in preference to two others, on grounds of cost. However, in order to be used for this purpose Avastin had to be broken up to produce multiple doses, and in that form it was an unlicensed medicine. The claimants, who marketed the other two products, claimed that the policy was unlawful and in breach of EU law regulating the marketing of medicines. The claim was dismissed at first instance and on appeal. The Court of Appeal held that it was legally open to NHS trusts to use Avastin in its unlicensed form to treat patients with the complaint provided certain conditions were satisfied to ensure compliance with EU law and (applying the test in Gillick) held that the policy was lawful, since it was realistically capable of implementation by NHS trusts in a way which did not lead to, permit or encourage unlawful acts. The policy did not explain the conditions to ensure that the use of Avastin would be lawful, but this lacuna did not make it unlawful. Underhill LJ gave the main judgment, with which Floyd and Rose LJJ agreed. He observed (paras 198 and 202) that the lawfulness of the policy should be based on an objective construction of it and rejected the claimants’ case that it was unlawful because the policy left open the possibility of trusts seeking to implement it by unlawful methods of using Avastin, even though they could implement it using lawful methods (paras 199-200 and 205-206). We have set out above his interpretation of the word “permit” in this context, with which we agree. In our view, Underhill LJ’s assessment in relation to the policy at issue in Bayer was correct. At para 202 he made the pertinent observation that the claimants’ case could only get anywhere if the commissioning groups “were under a duty to advise trusts as to what routes were lawful, because their silence on the issue would then be a breach of that duty”, but “no basis for the existence of any such duty was identified”. Similarly, in the present case, the Secretary of State was not under a duty to issue guidance which provided a comprehensive account of how the police should comply with the legal duties to which they were subject.

Rose LJ added helpful comments (para 214), with which Underhill and Floyd LJJ agreed:

In broad terms, there are three types of case where a policy may be found to be unlawful by reason of what it says or omits to say about the law when giving guidance for others: (i) where the policy includes a positive statement of law which is wrong and which will induce a person who follows the policy to breach their legal duty in some way (ie the type of case under consideration in Gillick); (ii) where the authority which promulgates the policy does so pursuant to a duty to provide accurate advice about the law but fails to do so, either because of a misstatement of law or because of an omission to explain the legal position; and (iii) where the authority, even though not under a duty to issue a policy, decides to promulgate one and in doing so purports in the policy to provide a full account of the legal position but fails to achieve that, either because of a specific misstatement of the law or because of an omission which has the effect that, read as a whole, the policy presents a misleading picture of the true legal position. In a case of the type described by Rose LJ, where a Secretary of State issues guidance to his or her own staff explaining the legal framework in which they perform their functions, the context is likely to be such as to bring it within category (iii). The audience for the policy would be expected to take direction about the performance of their functions on behalf of their department from the Secretary of State at the head of the department, rather than seeking independent advice of their own. So, read objectively, and depending on the content and form of the policy, it may more readily be interpreted as a comprehensive statement of the relevant legal position and its lawfulness will be assessed on that basis. In the present case, however, the police are independent of the Secretary of State and are well aware (and are reminded by the Guidance) that they have legal duties with which they must comply before making a disclosure and about which, if necessary, they should take legal advice.

In a category (iii) case, it will not usually be incumbent on the person promulgating the policy to go into full detail about how exactly a discretion should be exercised in every case. That would tend to make a policy unwieldy and difficult to follow, thereby undermining its utility as a reasonably clear working tool or set of signposts for caseworkers or officials. Much will depend on the particular context in which it is to be used. A policy may be sufficiently congruent with the law if it identifies broad categories of case which potentially call for more detailed consideration, without particularising precisely how that should be done. This was the approach adopted by Green J in R (Letts) v Lord Chancellor (Equality and Human Rights Commission intervening) [2015] EWHC 402 (Admin); [2015] 1 WLR 4497 (“Letts”).

Letts concerned the suicide of a psychiatric patient and policy guidance regarding the circumstances in which article 2 of the ECHR (right to life) may require the grant of legal aid for the family of the deceased at an inquest. The guidance was held to be misleading as to the law, and hence unlawful, in that while mentioning some cases where there might be such an obligation it omitted to identify a further category of case where such an obligation might exist. Green J held (para 100) that the guidance should have acknowledged the full range of categories of case in which legal aid might be required, but also stated that it was not necessary for the policy to fix the exact parameters within which legal aid should be provided. The test which he applied was based on Gillick but was subject to a gloss which was derived from the summary description of that case by Richards LJ in Tabbakh which was accepted as correct by both parties (paras 114-118). Green J stated the test in this way (para 118): “[w]ould the guidance if followed (i) lead to unlawful acts, (ii) permit unlawful acts or (iii) encourage such unlawful acts?” This is a somewhat looser formulation than in Gillick itself and is capable of being misconstrued, particularly if one asks in an open-ended way whether following a policy would “lead to” unlawful acts. It then might be said that there is a general duty on the policy-maker to formulate a policy in a way which eliminates uncertainty and so could not lead a person following it to act unlawfully. Unsurprisingly, Mr Southey submitted that we should endorse Green J’s formulation, read in this expansive way. We do not agree. In our view, the relevant test is found in Gillick, without the gloss which it has been given in subsequent cases. It should be noted that the observations of Underhill LJ in Bayer (para 44 above) about the restrictive meaning of the word “permits” were made with reference to the formulation in Letts, but as we have seen this is the language used by Lord Scarman in Gillick itself and Underhill LJ’s interpretation of that word is justified by the way in which it was used in that case.

Mr Southey submitted that the Guidance is unlawful because it fails to comply with standards of certainty, predictability in application and accessibility which are implicit in the concept of “law” as that concept is used in the ECHR and, in particular, in article 8(2). That is because it does not specify for every case whether representations should be sought from a subject but is flexible to allow for adaptation depending on the facts of particular cases. In our judgment, this submission is unsustainable for two reasons.

First, the Guidance does not purport to replace the underlying law which governs the circumstances in which a disclosure to the public may be made, that is to say (so far as is relevant for present purposes) the common law duty of fairness and the requirements inherent in article 8 itself (as applied by virtue of section 6(1) of the HRA). That includes the requirement of proportionality and the right, where appropriate, to participate in a decision-making process. These requirements are explained fully in the case law of the European Court of Human Rights (“the Strasbourg Court”) and the domestic courts. Mr Southey did not suggest that these legal regimes are defective by reference to the requirements of the concept of “law” in the ECHR. Clearly they are not. A police force which seeks to comply with its legal duties under these regimes, as the Guidance encourages it to do, will act “in accordance with the law” for the purposes of article 8(2).

Secondly, and in any event, the concept of “law” in the ECHR, including in article 8(2), does not imply a requirement that the domestic law relevant to a matter within the scope of a Convention right should be free from doubt as to its application and effect in particular cases. No system of law could possibly achieve that, and the ECHR does not require it. As the Strasbourg Court said in the landmark case of Sunday Times v United Kingdom (1979-1980) 2 EHRR 245, para 49:

There is a good deal of case law of the House of Lords, this court and the Strasbourg Court regarding the effect of the “in accordance with the law” requirement in article 8(2). See, for example, among many authorities, R (Roberts) v Comr of Police of the Metropolis [2015] UKSC 79; [2016] 1 WLR 210 and Christian Institute v Lord Advocate [2016] UKSC 51; 2017 SC (UKSC) 29, especially para 79. This is not the place for a full analysis of all these authorities. The most cursory review shows that they do not support the submission Mr Southey makes that the Guidance is incompatible with article 8 by reason of the fact that there is a limited degree of imprecision in it. The “in accordance with the law” rubric in article 8(2) does not require the elimination of uncertainty, but is concerned with ensuring that law attains a reasonable degree of predictability and provides safeguards against arbitrary or capricious decision-making by public officials. Judged even on its own terms, the Guidance meets those standards.

We would add that it is established that a policy may make a contribution to meeting the “in accordance with the law” requirement in some cases, by helping to provide a degree of predictability in the application of general discretionary provisions in statute (see eg Silver v United Kingdom (1983) 5 EHRR 547, para 88). However, it by no means follows that there is an obligation to have a policy in every case where a statute creates a discretionary power.

The analysis we have set out above is sufficient to provide the answer in the present case. The Guidance in its current form is not within any of the three categories we have described and is therefore not unlawful. The analysis also enables us to put to one side a series of other principles which in this case, in BF (Eritrea) and in other cases have been relied upon by claimants seeking to challenge the lawfulness of policies.

The Court of Appeal based its decision on Tabbakh and did not refer to Gillick. Tabbakh was concerned with the lawfulness of a policy (in fact, a group of policies) governing the way in which decisions were taken regarding the imposition of special (or non-standard) conditions on a prisoner released on licence. The claimant maintained that the policy created an unacceptable risk of illegality and was for that reason unlawful. The main issue affecting the claimant, who had psychiatric problems, was that the MAPPA panel which had oversight of his case recommended that he should be required to wear an electronic tag. He maintained that had a detrimental effect on his mental health. The prison official responsible for him imposed the condition pursuant to the recommendation. His complaint about the policy was that it contained no provision for him to be given an opportunity to make representations about such licence conditions before a decision was made. Cranston J and the Court of Appeal held that article 8 had the effect that the claimant was entitled to make representations before a decision was made. On the facts, he had been given a sufficient opportunity to do this through engagement with the offender manager responsible for his case. In that way his concerns were provided to the MAPPA panel and the prison official who had to decide on the licence conditions to be imposed. However, the courts also considered his challenge to the lawfulness of the policy, notwithstanding that it was academic.

Richards LJ gave the main judgment in the Court of Appeal, with which the other members of the court agreed. Counsel for the claimant contended that the test for the lawfulness of a policy required the court to assess, across the full range of cases liable to be subject to the policy, whether the policy creates an unacceptable risk that the individual to whom it applies will be subject to unlawful decision-making (para 34). However, Richards LJ held that a different test was applicable, “whether the system established by the guidance in the policy documentation is inherently unfair by reason of a failure to provide the offender with a fair opportunity to make meaningful representations about proposed licence conditions” (para 35, emphasis in original; and paras 38 and 48). He regarded R (Refugee Legal Centre) v Secretary of State for the Home Department [2004] EWCA Civ 1481; [2005] 1 WLR 2219 (“Refugee Legal Centre”) as the key authority in support of this. Richards LJ regarded this as materially different from the test in Gillick, which he put to one side (para 48). Cranston J had relied on Gillick in order to reject a wider test of whether a policy gives rise to an unacceptable risk of unlawful decision-making. Richards LJ disapproved this and held instead that the relevant test was supplied by Refugee Legal Centre which “effectively equate[d] an unacceptable risk of unfairness with a risk of unfairness inherent in the system itself” (paras 47-49). A distinction was to be drawn between whether the system itself was inherently unfair or whether it just left scope for a risk of unfairness arising in the ordinary course of individual decision-making (paras 38 and 41). The ability of the claimant to make representations through his offender manager was not an unusual occurrence, but showed the system itself operating fairly, and there was no reason to believe that the system was not capable of operating fairly in the generality of cases (para 53). Therefore, the challenge was dismissed. Accordingly, the wider comments by Richards LJ about the legal test to be applied were obiter dicta.

Refugee Legal Centre is the foundation for this approach to review of a policy. The claimant organisation brought a challenge to a fast-track scheme for decision-making and appeals in relation to asylum claims by immigrants who fell within specified categories. The scheme was composed of policy guidance to officials and procedure rules which governed the timetable in relation to appeals for cases within the scheme. At the first stage, officials were required to conduct interviews and arrive at decisions within three days. The scheme had a degree of flexibility built into it to allow for more complex cases to be removed from it and diverted into normal, slower decision-making procedures. It was accepted that in straightforward cases the scheme was capable of operating fairly (para 6). The Court of Appeal held that the relevant question was that proposed by the defendant Secretary of State: “does the system provide a fair opportunity to asylum seekers to put their case?” and said that there would be “something justifiably wrong with a system which places asylum seekers at the point of entry [into the system] at unacceptable risk of being processed unfairly” (para 6). No reference was made to Gillick. The court upheld the decision at first instance rejecting the challenge.

At para 7 the Court of Appeal said:

It was common ground that officials making decisions had a duty to act fairly and that complex cases could not be dealt with fairly within the three-day window. The court did not accept the Secretary of State’s submission that the availability of an appeal against the decision of an official was sufficient to make the system fair overall, since applicants for asylum were entitled to be treated fairly at the first stage as well and, further, there would be cases where prejudice arising from unfairness due to pressure of time at the first stage would not be picked up and rectified on an appeal (para 15). The argument therefore focused on whether the scheme could operate fairly at the first stage, and in particular whether there was sufficient flexibility at that stage for more complex cases to be taken out of the scheme.

As to that, at paras 16-19 the court expressed concern about the fact that no test or standard for the adaptation of the three-day timetable to individual needs had been formulated by the Home Office in its policy and that the Home Office evidence (summarised at para 12) was that officials were encouraged to adopt as flexible an approach as possible “without compromising the integrity of the process”. There was a danger that this approach would mean that officials did not in fact operate flexibly, as in law they should do, to remove individual cases from the scheme. The court was not persuaded by an assertion by counsel for the Secretary of State, without evidence to support it, that a general principle of flexibility was deeply ingrained among the relevant officials; in the court’s view it would be better for relevant factors to be set out in writing in the policy itself (para 19).

However, the court was not prepared to find the scheme unlawful on this basis. At paras 20-25 it endorsed the finding of the judge that the system itself was not inherently unfair and therefore unlawful. As it said (para 23), “provided that [the scheme] is operated in a way that recognises the variety of circumstances in which fairness will require an enlargement of the standard timetable - that is to say lawfully operated - the … system itself is not inherently unfair”; the position would be made clearer if the Home Office adopted a written flexibility policy which explained that the three-day timetable was “in truth a guide and not a straitjacket”.

It is unfortunate that no reference was made to Gillick in that case and that the Court of Appeal proceeded on the basis of a concession made by the Secretary of State as to the test to be applied. However, in our view the analysis and conclusion in Refugee Legal Centre can readily be assimilated with the approach to be derived from Gillick. According to the Court of Appeal, the scheme in the policy would have been “inherently unfair” if it contained guidance requiring officials to deal with more complex cases, which could not fairly be adjudicated within the three-day window, within that time-frame. The evidence as to the context in which officials had to apply the policy, where they were required to proceed without “compromising the integrity of the process”, meant that it came very close to being a case which fell within the first or third categories referred to in para 46 above, and in the court’s view steps should be taken to correct this. However, the policy was capable of being operated in a way which properly recognised the requirements of the common law duty of fairness in individual cases and this meant that, provided it was not stated in a misleading way, it was not inherently unfair. At para 24 the court distinguished R (Q) v Secretary of State for the Home Department [2003] EWCA Civ 364; [2004] QB 36 (CA), in which “the essential finding was that the procedures applied to each of the applicants had been vitiated by unfairness to him or her” and “the deficiencies in due process identified by the court … had resulted in adverse decisions which fell to be quashed in consequence” (ie there had in fact been a breach of the duty of fairness in the individual cases). By contrast, as the court emphasised, Refugee Legal Centre was concerned with whether there was “a gateway risk of injustice, in the nature of things not case-specific but caused by potential rigidity in a system which requires genuine flexibility in its timetable”. Its focus was not on what had happened in individual cases, but on looking forward from the point when a case entered the scheme to see whether the policy was capable of being operated lawfully across the range of such cases.

We agree that this is a fundamental distinction for the purposes of analysis. If it is established that there has in fact been a breach of the duty of fairness in an individual’s case, he is of course entitled to redress for the wrong done to him. It does not matter whether the unfairness was produced by application of a policy or occurred for other reasons. But where the question is whether a policy is unlawful, that issue must be addressed looking at whether the policy can be operated in a lawful way or whether it imposes requirements which mean that it can be seen at the outset that a material and identifiable number of cases will be dealt with in an unlawful way.

In our view, reading the Refugee Legal Centre case in this way in line with Gillick removes the risk of misunderstanding it as stating an unprincipled and vague test. If one simply asks whether a policy creates an unacceptable risk that an individual will be treated unfairly (which is to say, unlawfully), there is a danger that this could be taken as a freestanding principle distinct from that in Gillick, as seems to have been envisaged in Tabbakh.

First, it is unclear what the relationship of such a principle is with the authoritative guidance given in Gillick. If the principle were that a policy is unlawful if it creates an unacceptable risk that an individual will be treated unlawfully, that is a substantially wider principle than that stated in Gillick and inconsistent with it. There is no sound conceptual basis for separating out unlawfulness due to unfairness from unlawfulness for any other reason. Secondly, a test whether a policy creates an “unacceptable risk” that an individual will be treated unfairly or unlawfully provides no criterion of what makes a risk count as unacceptable. The distinction drawn in Tabbakh between whether a policy is inherently unfair or it just leaves a risk of unfairness arising in the ordinary course of individual decision-making is similarly unclear. A determinate criterion is required to distinguish the two cases in a principled way. Gillick supplies it. Thirdly, a test for the lawfulness of a policy based on “unacceptable risk” or the like would be a new departure in public law and cannot be regarded as an incremental extension from existing principle. On the contrary, as we have explained, in our opinion it would subvert existing principle. By contrast, if the test of inherent unfairness is applied by reference to the principle in Gillick, the law supplies a reasonably clear criterion of unlawfulness which has a sound foundation in principle. Fourthly, without such a foundation, the assertion of such a power of review by the courts, in relation to functions (the operation of administrative systems and the statement of applicable policy) which are properly the province of the executive government would represent an unwarranted intrusion by the courts into that province. Anchoring the lawfulness of policy to the principles articulated in Gillick avoids that outcome. Fifthly, if one moves away from that principled foundation, there is a risk that a court will be asked to conduct some sort of statistical exercise to see whether there is an unacceptable risk of unfairness, as was urged upon the court in the BF (Eritrea) appeal: see our judgment in BF (Eritrea) at paras 35 and 41. But a court is not well equipped to undertake such an analysis based upon experience. In principle, the test for the lawfulness of a policy should be capable of application at the time the policy is promulgated, which will be before any practical experience of how it works from which statistics could be produced. The test for the lawfulness of a policy is not a statistical test but should depend, as the Gillick test does, on a comparison of the law and of what is stated to be the behaviour required if the policy is followed. Both aspects of this test are matters on which the court is competent and has the authority to pronounce.

Some of the cases coming after Refugee Legal Centre and Tabbakh have treated them as authority for such a wider principle of review (is there a real or unjustified risk of unfairness or illegality?) without examination of its consistency with the principles articulated in Gillick. In our view, this tendency should be corrected. Statements of such a wider principle have also drawn force from what is in truth a distinct and valid principle of access to justice which was reviewed in detail in R (UNISON) v Lord Chancellor (Equality and Human Rights Commission intervening) (Nos 1 and 2) [2017] UKSC 51; [2020] AC 869 (“UNISON”), discussed below. This has obscured the proper analysis of cases to which that principle applies. Five authorities in particular deserve comment.

First, in R (Detention Action) v First-tier Tribunal (Immigration and Asylum Chamber) [2015] EWCA Civ 840; [2015] 1 WLR 5341 (“Detention Action”) the Court of Appeal struck down procedure rules in relation to a scheme for fast-track disposal of appeals in relation to asylum claims on the basis that the scheme was structurally unfair. The ratio of the decision is that the procedure rules were ultra vires the statutory rule-making power in section 22(4) of the Tribunals, Courts and Enforcement Act 2007. As Lord Dyson MR, giving the sole substantive judgment, said (para 22), “[t]he legality of the [fast track procedure rules] must be judged by reference to section 22(4) of the 2007 Act”. On the court’s interpretation of that power, the rules made under it had to be capable of being operated in a fair way in every case, but the scheme did not achieve this. Due to the complexities which could arise in some cases, it was inevitable that a significant number of appellants in relation to asylum decisions would be denied a fair opportunity to present their cases under the fast-track scheme, and the opportunities for a case to be removed from the scheme as unsuitable for disposal under it were too restricted to deal with that problem. We respectfully agree with this conclusion; but in the course of discussion Lord Dyson MR referred to Refugee Legal Centre and Tabbakh. At para 25 he said that the question posed in Refugee Legal Centre was whether the system “considered in the round” carried “an unacceptable risk of unfairness to asylum seekers” (para 25). As set out above, we consider that taken by itself this formulation is too wide. However, at para 27 Lord Dyson accepted the summary of the principles to be derived from these authorities provided by counsel for the respondent:

In our view, this statement is consistent with the approach derived from Gillick which we would apply in the present case. Notably, as Lord Dyson said at point (v), the core question was whether a system which, according to the statutory power under which it was promulgated, had to be capable of allowing every case which the Secretary of State introduced into it to be determined fairly had “the capacity to react appropriately to ensure fairness”. The procedural rules did not meet that standard. In other words, a significant number of cases introduced into the system would be decided unfairly and hence unlawfully if the procedure rules were applied to them. This is in line with the principle articulated in Gillick to identify whether a policy is unlawful.

Second is the decision of the Court of Appeal in R (S) v Director of Legal Aid Casework [2016] EWCA Civ 464; [2016] 1 WLR 4733 (“Director of Legal Aid Casework”). This concerned the legality of a legal aid funding scheme for exceptional cases of a kind where legal aid was not ordinarily available according to the standard rules. The scheme was constituted in part by regulations and in part by policy guidance, and there was a complex application form. In certain circumstances, pursuant to Convention rights in the HRA or under EU law, an individual would have a right to legal aid. The claimant challenged the lawfulness of the application form, the regulations and the policy on the grounds that, together, they made the scheme too restrictive in terms of the provision of legal aid. They gave rise to an unacceptable risk that an individual would be treated unlawfully, in that he would not be able to obtain legal aid in circumstances where failure to provide it would be a breach of that individual’s rights under the HRA or directly enforceable EU law. At first instance, the claim succeeded. However, the appeal was allowed.

Laws LJ gave the leading judgment, in which he set out the principles to be applied. They were not in dispute between the parties, who focused on Tabbakh and the test of “inherent unfairness”. Laws LJ took the approach of Richards LJ in Tabbakh, which cited Refugee Legal Centre, as his starting point and then quoted the passage from the judgment of Lord Dyson in Detention Action which we have set out in para 67 above. He emphasised (para 18) the distinction between unfairness inherent in the system itself, making it unlawful, and one which happens to be operated badly so that individual instances of unlawful treatment arise under it. He also referred to the danger of a judge “crossing the line between adjudication and the determination of policy”, an observation which we take to illustrate his underlying concern about the Court of Appeal authorities upon which the parties rested their submissions. Gillick was not cited. The reasonably rigorous test for lawfulness of a policy as articulated in Gillick became a more impressionistic one of whether the policy (or the other elements of the scheme) presented an unacceptable risk that an individual would be treated unlawfully. Briggs LJ (dissenting in part) and Burnett LJ gave separate judgments.

Although the same test was applied to all aspects of the claim, the issue in relation to the application form might have been argued on the basis that an administrative practice (adoption of an overly complex form) had created an impediment to gaining access to legal aid and hence to court in cases where there was a right to have publicly funded assistance. That, in substance, was the effect of asking whether the form was inherently unfair in impeding access to legal aid. We address this issue in the context of our discussion of UNISON below.

The third decision is that of the Divisional Court (Hickinbottom LJ and Lewis J) in R (Woolcock) v Secretary of State for Communities and Local Government [2018] EWHC 17 (Admin); [2018] 4 WLR 49. This involved a challenge to the lawfulness of the system for enforcement of liability for unpaid council tax by means of imprisonment, as operated by magistrates’ courts. The legal conditions for making an order for detention were set out clearly in regulations, but there was a body of cases in which magistrates imposed excessive and unlawful periods of detention. Thus it was alleged that the system itself was unfair by reason of the failure to train magistrates and their clerks properly in the operation of the system (para 43). Hickinbottom LJ (para 68) summarised the effect of the decisions regarding challenges to a system on the basis that there was a compendious test of the kind just described. In our opinion, however, this conceals important differences between cases. In the event, the challenge was rejected, and rightly so (see para 97). We find it difficult to see how a set of legal rules which are clear in their effect and which would if properly applied result in justified detention can be said to constitute a scheme which is systematically unfair.

Fourthly, in BF (Eritrea) [2019] EWCA Civ 872; [2020] 4 WLR 38, Underhill LJ referred (paras 60-63) to Director of Legal Aid Casework and UNISON to identify the test of lawfulness to be applied to policy guidance. The Court of Appeal in BF (Eritrea) then struck down a policy as unlawful in circumstances where it contained no misstatement of law, but in the view of the court did not sufficiently remove uncertainty in the application of a statutory test. Underhill LJ said (para 63) that the policy/guidance in that case would “be unlawful, if but only if, the way that they are framed creates a real risk of a more than minimal number of children being detained … The issue is whether the terms of the policy themselves create a risk which could be avoided if they were better formulated”. See further our judgment in the parallel appeal in that case.

The fifth decision is the judgment of the Divisional Court (Bean LJ and Chamberlain J) in R (W) v Secretary of State for the Home Department [2020] EWHC 1299 (Admin); [2020] 1 WLR 4420 (“W”). The claimant was the British child of a Ghanaian mother who had been given leave to remain in the United Kingdom subject to a condition that she maintain herself and her dependants without recourse to public funds. The condition was imposed pursuant to a provision in the Immigration Rules (and a policy in relation thereto) issued by the Secretary of State, which stated that such a condition should usually be imposed unless very restricted exemptions applied. The claimant successfully challenged the regime set out in the rules and policy on the ground that it was in breach of the Secretary of State’s duty under section 6(1) of the HRA because it failed to ensure that imposing such a condition would not result in inhuman treatment, contrary to article 3 of the ECHR. The claimant and his mother did not have to wait until they became destitute before the Secretary of State came under an obligation to remove the condition. It would be unlawful to maintain it in place if they would imminently suffer inhuman and degrading treatment by reason of lack of resources without recourse to public funds (para 60). We discuss below the legal obligations which arise under article 3 which is not in issue in the present appeal. The Divisional Court also based its decision on principles of domestic law, which it decided were to similar effect in that context (paras 60-61). It expressed views (paras 52-59) about the test to be applied in assessing the lawfulness of a policy at common law, which it took to be derived from the decision of this court in R (Bibi) v Secretary of State for the Home Department [2015] UKSC 68; [2015] 1 WLR 5055 (“Bibi”), which we address below, and the decision of the Court of Appeal in BF (Eritrea). From Bibi, at para 58 the Divisional Court derived a test that a policy would be unlawful if it leads to unlawful results in “a significant number of cases” and from BF (Eritrea) it derived a test that policy guidance would be unlawful if there is “a real risk of the guidance leading to an unlawful result in a more than minimal number of cases”, relying in particular on the judgment of Underhill LJ at para 63. The Divisional Court considered that these tests were consistent with principle but did so on the basis of a gloss given to the Gillick principle in the formulation by Green J at first instance in the Letts case, who said that the court should intervene where guidance is misleading as to the law or will “lead to” or “permit” or “encourage” unlawful acts (para 58). The court concluded (para 59), so far as relevant, that the proper approach was to ask, “[d]oes the regime, read as a whole, give rise to a real risk of unlawful outcomes in a ‘significant’ or ‘more than minimal number’ of cases?”. In our view, this way of formulating the test involves significant movement from the proper approach to be derived from Gillick. However, the way in which the court decided the case is consistent with the approach in Gillick. Having identified at paras 60-61 what would be unlawful conduct in an individual case, at paras 62-66 the court construed the relevant rules and the policy as a complete set of instructions to officials of the Secretary of State (of the kind referred to by Rose LJ in Bayer at para 214: see para 45 above) which required them to impose or maintain the no recourse to public funds condition in cases where that would be unlawful.

Mr Southey sought to rely on the Divisional Court’s formulation in this case in support of his contention that the Secretary of State had an obligation to draft the Guidance to eliminate, or at any rate reduce, any uncertainty regarding when a subject should be given the opportunity to make representations before a disclosure is made. He submits that the test to be applied in assessing the legality of a policy is whether it creates a real risk of unfairness in a more than minimal number of cases. However, to the extent that the Divisional Court’s formulation departs from the more rigorous test of unlawfulness to be derived from Gillick, it is incorrect and should not be followed. The same is true in respect of the guidance given by the Court of Appeal in BF (Eritrea). As we explain in our judgment in that case, we respectfully consider that the Court of Appeal erred in its reasoning and in its conclusion.

Mr Southey also sought to rely on Bibi. However, in our view, it does not support his submissions. On the contrary, this court applied what is in effect the same, narrower approach as was adopted in Gillick.

Bibi was concerned with a challenge to the lawfulness of an immigration rule promulgated by the Secretary of State which required a foreign spouse or partner of a British citizen to produce prior to entry a certificate of knowledge of English to a prescribed standard, subject to certain exemptions based on age, physical or mental condition and exceptional circumstances. The claimants were British citizens whose husbands were foreign nationals residing in Pakistan and Yemen, respectively. The husbands did not have access to English tuition at affordable cost. The Secretary of State also published a policy which stated that use of the “exceptional circumstances” exemption would be rare and would not include failure to obtain tuition or take the test owing to financial hardship. The claimants’ challenge to the rule included the ground that the requirement interfered with their right to respect for family life under article 8 of the ECHR. The challenge was dismissed. This court held that in order to be compatible with article 8, the interference with family life produced by the rule and policy had to be proportionate to a legitimate objective. As regards the rule, it was accepted that it might be applied in a way that was incompatible with the article 8 rights of a British partner in an individual case, but in order to strike it down as unlawful it was held that it was necessary to show that it would be incapable of being operated in a proportionate way and so was inherently unjustified in all or nearly all cases: paras 2 and 60 (Baroness Hale of Richmond, with whom Lord Wilson of Culworth agreed), para 69 (Lord Hodge, with whom Lord Hughes agreed) and para 100 (Lord Neuberger of Abbotsbury). On the other hand, the court was of the view that the policy was unlawful and required amendment, because if it were followed it would inevitably result in some decisions which were unlawful in that they involved a disproportionate interference with article 8 rights: paras 53-55 (Baroness Hale), paras 73-74 (Lord Hodge) and paras 101-102 (Lord Neuberger). The test of lawfulness applied in relation to the policy, therefore, was the same as in Gillick.

In this judgment we do not need to address the distinct question whether the difference in approach between testing the lawfulness of an immigration rule or other legislation and testing the lawfulness of a policy as intimated in Bibi is justified. We note that in In re Northern Ireland Human Rights Commission’s Application for Judicial Review [2018] UKSC 27; [2018] NI 228, at para 82, Lord Mance thought that the approach to testing the lawfulness of a legislative rule should in substance be the same as that applicable in relation to a policy: “It [is] sufficient that it will inevitably operate [incompatibly with Convention rights] in a legally significant number of cases”. Lord Mance’s formulation conforms with the approach in Gillick. In the present case, it cannot be said that the Guidance will inevitably operate incompatibly with the appellant’s rights.

A further authority on which Mr Southey sought to rely was R (Munjaz) v Mersey Care NHS Trust [2005] UKHL 58; [2006] 2 AC 148 (“Munjaz”). In that case the defendant trust adopted a policy on the use of seclusion for detained psychiatric patients which provided for reviews at certain intervals which the claimant complained were excessively long. The claimant maintained that the hospital’s policy was unlawful on the ground, among others, that it was incompatible with his rights under article 3 ECHR (protection against torture and inhuman or degrading treatment). The House of Lords, by a majority, upheld the lawfulness of the policy. At para 29, Lord Bingham of Cornhill said, “[t]he trust must not adopt a policy which exposes patients to a significant risk of treatment prohibited by article 3”; and at para 80 Lord Hope of Craighead said that the issue should be approached by asking whether the hospital’s policy “gives rise to a significant risk of ill-treatment of the kind that falls within the scope of the article” (although he also indicated that if it did, the policy might still be lawful if it would place a disproportionate burden on the hospital to abandon it). Mr Southey submitted that this is an illustration of a principle of general application. But it is not. It is a rule of law specific to article 3, imposed by reason of the importance of the interests protected by that article. Thus, for example, it is well-established that article 3 imposes obligations on states to protect individuals against the risk of ill-treatment contrary to that provision, even at the hands of others: see eg Soering v United Kingdom (1989) 11 EHRR 439.

In BF (Eritrea), in support of his formulation of the relevant test at para 63, Underhill LJ relied (paras 60-63) upon the decision of the Court of Appeal in Director of Legal Aid Casework and the decision of this court in UNISON. With respect, we consider that these authorities do not support the formulation of the test for lawfulness of a policy as set out by Underhill LJ. They are concerned with a different issue, namely the lawfulness of measures which impede access to courts and tribunals, either by the imposition of fees (UNISON) or by placing a practical obstacle in the way of someone entitled to legal aid (Director of Legal Aid Casework). In UNISON this court held that there is a fundamental right under the common law of access to justice, meaning effective access to courts and tribunals to seek to vindicate legal rights, which means that the executive is under a legal obligation not to introduce legal impediments in the way of such access save on the basis of clear legal authority: see the discussion by Lord Reed in UNISON at paras 66-98. The decision was concerned with the introduction of an order imposing fees to bring claims in an employment tribunal, but the principles stated are of general application. The test applied was whether the making of the order created “a real risk that persons will effectively be prevented from having access to justice” (para 87; see also para 85, where R (Hillingdon Borough Council) v Lord Chancellor [2008] EWHC 2683; [2009] 1 FLR 39 is referred to as authority for such a test). As Lord Reed observed (para 91), it is sufficient if a real risk of prevention of access to justice is demonstrated. This means that, in order to test the lawfulness of a measure on this basis, it is legitimate to have regard to evidence regarding its likely impact and the court has to make an overall evaluative assessment whether this legal standard is met or not (and statistics might have a part to play in making such an assessment). In UNISON, this court held that the fees order was unlawful on this basis.

This is also, in effect, the question which the court asked itself in Director of Legal Aid Casework in relation to the application form, when it assessed whether the form created an unacceptable risk of unfairness in the form of preventing access to legal aid (and hence preventing access to the courts) in cases where there was an obligation to provide legal aid. With the benefit of the statement of the relevant principle in UNISON, no doubt the issue would now be formulated with more precision.

Similar issues arose in R (Medical Justice) v Secretary of State for the Home Department [2011] EWCA Civ 1710, in relation to measures (which happened to be set out in a policy document) limiting the time available to an immigrant to obtain legal advice and assistance to challenge removal directions, and in R (Howard League for Penal Reform) v Lord Chancellor [2017] EWCA Civ 244; [2017] 4 WLR 92, in relation to the lawfulness of removal of legal aid from certain categories of legal claims affecting prisoners. In both cases, as in Director of Legal Aid Casework, the court referred to Refugee Legal Centre and framed the question for itself in terms of whether the system was inherently unfair; but in both cases the substance of the analysis was whether there had been an unlawful infringement of the constitutional right of access to a court or tribunal. In our view, the formulation of the test in Refugee Legal Centre is not a helpful way of approaching that issue. In future, the framework of analysis in UNISON should be applied instead.

This is indeed what occurred in R (FB (Afghanistan)) v Secretary of State for the Home Department [2020] EWCA Civ 1338; [2021] 2 WLR 839. The case concerned a challenge to the lawfulness of another scheme to limit the time for immigrants to challenge decisions to remove them before they were implemented. The Court of Appeal upheld the challenge on the ground that the scheme created an excessive impediment in the way of immigrants gaining access to a court to challenge the lawfulness of such decisions in their cases, ie by reference to the principle in UNISON: see, in particular, paras 142 (Hickinbottom LJ), 170 (Coulson LJ) and 185 and 196 (Lord Burnett of Maldon CJ). The more general approach in Refugee Legal Centre, Medical Justice, Tabbakh and Detention Action was referred to, but its effect in those cases was explained in terms of the access to justice principle examined in UNISON (see, in particular, paras 120-126, per Hickinbottom LJ, and para 177, per Lord Burnett CJ). In our view, on a proper understanding of the legal principles discussed above, the wider formulation of a test of systemic inherent unfairness in relation to a legal scheme which has been taken to be laid down in the line of cases stemming from Refugee Legal Centre will in most, if not all, circumstances dissolve into the Gillick principle and the UNISON principle, each with its own precise focus.

We have reviewed the lawfulness of the Guidance by reference to a number of distinct principles:

According to the principle in Gillick a policy will be unlawful if it misdirects officials as to their legal obligations; but the Guidance does not do so (paras 29-48);

The Guidance is not unlawful as a result of any failure to comply with the test derived from the “in accordance with the law” rubric in article 8 of the ECHR (paras 49-53);

The Guidance is not unlawful by reference to the test of inherent unfairness discussed in Tabbakh and Refugee Legal Centre, which is to be analysed as an aspect of the Gillick principle (paras 55-75);

The Guidance is not unlawful by reference to an obligation to avoid a real risk of treatment contrary to article 3 of the ECHR as discussed in the Munjaz case (para 79 above); and

The Guidance does not create a risk of impeding access to justice and so is not unlawful by reference to the principle discussed in UNISON (paras 80-83).

For the reasons we have given, which are somewhat different from those given in the Court of Appeal, we would dismiss this appeal.