In the matter of an application by Lorraine Gallagher for Judicial Review (Northern Ireland)

The four respondents to these appeals have all been convicted or received cautions or reprimands in respect of comparatively minor offending. The disclosure of their criminal records to potential employers has made it more difficult for them to obtain jobs, or may make it more difficult in future. In each case, the relevant convictions and cautions were “spent” under the legislation designed to enable ex-offenders to put their past behind them. They had to be disclosed only if the respondents applied for employment involving contact with children or vulnerable adults. In all four of these appeals, the respondents challenge the statutory rules under which disclosure of their records was required as being incompatible with the European Human Rights Convention.

Such cases raise problems of great difficulty and sensitivity. They turn on two competing public interests. One is the rehabilitation of ex-offenders. The other is the protection of the public against people whose past record suggests that there may be unacceptable risks in appointing them to certain sensitive occupations. The importance of both public interests needs no emphasis. The ability of ex-offenders to obtain employment is often an essential condition of their successful reintegration into law-abiding society at what, especially in the case of young offenders, may be a critical period of their lives. On the other hand, in some employment sectors a more cautious approach is indispensable. The Bichard Inquiry (2004) (HC 653) into child protection procedures and vetting practices was a stark reminder of the importance of ensuring that the rehabilitation of offenders does not undermine proper standards of public protection when those with criminal records apply for jobs involving contact with children. The Inquiry had been set up after two young girls had been murdered by a caretaker employed at their school, about whom there had been substantial intelligence in police files, not retained or disclosed to the school, suggesting a pattern of sexual interference with women and young girls.

P received a caution on 26 July 1999 for the theft of a sandwich from a shop. Three months later, on 1 November 1999, she was convicted at Oxford Magistrates’ Court of the theft of a book worth 99p and of failing to surrender to the bail granted to her after her arrest for that offence. She received a conditional discharge for both offences. At the time of the offences she was 28 years old, homeless and suffering from undiagnosed schizophrenia which is now under control. She has committed no further offences. P is qualified to work as a teaching assistant but has not been able to find employment. She believes that this is because she has been obliged to disclose her convictions on each job application.

W was convicted by Dewsbury Magistrates’ Court on 26 November 1982 of assault occasioning actual bodily harm contrary to section 47 of the Offences against the Person Act 1861. At the time of the offence he was 16 years old. The assault had occurred in the course of a fight between a number of boys on their way home from school. He received a conditional discharge, and has not offended since. In 2013, when he was 47, he began a course to obtain a certificate in teaching English to adults. His conviction has not been disclosed, but he believes that he would need to disclose it and obtain a criminal record certificate if he were to apply for a job as a teacher, and that this will prejudice his chances of obtaining employment.

On 1 August 2006, when he was 13 years old, G was arrested for sexually assaulting two younger boys, contrary to section 13 of the Sexual Offences Act 2003. The offences involved sexual touching and attempted anal intercourse. These were potentially serious offences, but the mitigation was exceptional. The police record indicates that the sexual activity was consensual and “seems to have been in the form of ‘dares’ and is believed to have been a case of sexual curiosity and experimentation of the part of all three boys.” The Crown Prosecution Service decided that it was not in the public interest to prosecute, but suggested a reprimand under section 65 of the Crime and Disorder Act 1998. On 5 September 2006 G received two police reprimands, one in respect of each of the younger boys. He has not offended since. In 2011, when he was working as a library assistant in a local college, he was required to apply for an enhanced criminal record check because his work involved contact with children. After the application was made, he was told by the police that they proposed to disclose the reprimand, together with an account of the mitigating circumstances. As a result, G withdrew the application and lost his job. He has since felt unable to apply for any job for which a standard or enhanced criminal record check would be required.

Lorraine Gallagher was convicted on 24 July 1996 at Londonderry Magistrates’ Court of one count of driving without wearing a seatbelt, for which she was fined £10, and three counts of carrying a child under 14 years old without a seatbelt, for which she was fined £25 on each count. All four counts related to the same occasion. On 17 June 1998, she was convicted at the same court on two counts of carrying a child under 14 years old in a car without a seatbelt. She was fined £40 on each count. Again, both counts related to the same occasion. She had been carrying two of her children in the back of her car. Their seatbelts had been attached, but not properly because, unbeknown to her (she says), they had placed the shoulder straps under their arms. Ms Gallagher has no other convictions. In 2013, having qualified as a social carer, she was admitted to the Northern Ireland Social Care Council Register of Social Care Workers. In 2014, she applied for a permanent position at a day centre for adults with learning difficulties and received a conditional offer of employment. In response to a request to disclose whether she had been convicted at any time of a criminal offence she disclosed “Yes” and “carrying child without seatbelt in 1996”, but she did not disclose the conviction in relation to herself. She did not disclose the 1998 convictions at all. When the enhanced criminal record certificate disclosed all the convictions, the job offer was withdrawn on the ground that her failure to disclose them called her honesty and integrity into question.

The disclosure of criminal convictions, cautions and reprimands is governed by two related statutory schemes. Disclosure by the ex-offender himself is governed by the Rehabilitation of Offenders Act 1974 in England and Wales and the corresponding provisions of the Rehabilitation of Offenders (Northern Ireland) Order 1978 (SI 1978/1908) in Northern Ireland. There is no material difference between the Act and the Northern Ireland Order. I shall therefore refer in this judgment to the provisions of the 1974 Act. Section 1 of that Act provides that, subject to conditions none of which is material, where a person has been convicted of an offence which is not excluded from rehabilitation, that person shall be treated as rehabilitated after the expiry of the rehabilitation period and the conviction shall be treated as spent. Sections 8A and 8AA make corresponding provision for cautions. The rehabilitation period is defined by section 5, and varies according to the sentence of the court and the age of the offender. Section 4 determines the effect of rehabilitation. By section 4(1), the ex-offender is to be treated for all legal purposes as a person who has not committed or been charged or prosecuted or convicted of the offence. For present purposes, the critical provisions of the Act are sections 4(2) and (3). Their effect is that where a question is put to an ex-offender about his previous convictions, offences, conduct or circumstances (other than in judicial proceedings), the question shall not be treated as relating to spent convictions and may be answered accordingly. In other words, the ex-offender is under no obligation to disclose it, and indeed may lawfully deny it. He is not to be subjected to any liability or prejudice in consequence. Section 4(4) provides that the Secretary of State may by order provide for exceptions to sections 4(2) and (3). The Secretary of State exercised this power for England and Wales by the Rehabilitation of Offenders Act 1974 (Exceptions) Order 1975 (SI 1975/1023) (as amended); and for Northern Ireland by the Rehabilitation of Offenders (Exceptions) Order (Northern Ireland) 1979 (SR(NI) 1979/195). The effect of the Orders is that an ex-offender’s right not to disclose a conviction or caution does not apply if the question is asked in order to assess his or her suitability for any of 13 specified purposes. These include his or her suitability for admission to certain professions or certain kinds of employment; or for his or her assignment to work with children or vulnerable adults in specified circumstances; or for the provision of day care; or for the adoption of a child.

Disclosure of criminal records by the Disclosure and Barring Service in England and Wales or AccessNI in Northern Ireland is governed in both jurisdictions by a distinct but closely related statutory scheme under Part V of the Police Act 1997 (as amended). Sections 113A and 113B of the 1997 Act (as inserted) deal, respectively, with criminal record certificates (“CRCs”) and enhanced criminal record certificates (“ECRCs”) recording a person’s convictions and cautions, including spent convictions and cautions. Applications for a certificate are made by the ex-offender himself and countersigned by a “registered person”, namely a person registered as having a proper interest in the information. In R (T) v Chief Constable of Greater Manchester Police (Liberty intervening) [2015] AC 49, paras 10-12, Lord Wilson concisely summarised the scheme of disclosure under the Police Act 1997, as it stood before the scheme was amended in March 2014:

Section 113A(7) empowered the Secretary of State to amend by Order the definition of “relevant matters” falling to be disclosed. With effect from March 2014, this power was exercised so as to introduce a more selective system for disclosure by the Disclosure and Barring Service: Police Act 1997 (Criminal Record Certificates: Relevant Matters) (Amendment) (England and Wales) Order 2013 (SI 2013/1200). Similar changes were made in Northern Ireland with effect from April 2014 by the Police Act 1997 (Criminal Record Certificates: Relevant Matters) (Amendment) Order (Northern Ireland) 2014 (SI 2014/100). The effect of the amendments was to limit the disclosure of convictions and cautions under sections 113A and 113B of the Police Act to (i) convictions and cautions for any of a list of more serious offences, generally violent or sexual, contained in section 113A(6D); (ii) convictions which resulted in a custodial sentence; (iii) other convictions or cautions if they were still “current”, ie had occurred within a specified period before the issue of the certificate, viz 11 years in the case of an adult and five and a half years in the case of a minor; and (iv) all convictions and cautions where the person has more than one conviction. Broadly corresponding limitations were imposed on the convictions and cautions which had to be disclosed under the Rehabilitation of Offenders Act 1974: see Rehabilitation of Offenders Act 1974 (Exceptions) Order 1975 (Amendment) (England and Wales) Order 2013 (SI 2013/1198), and Rehabilitation of Offenders (Exceptions) (Amendment) Order (Northern Ireland) 2014 (SI 2014/27).

Section 4(2) and (3) of the Rehabilitation of Offenders Act 1974 are not in terms confined to disclosures in the course of job applications. These are, however, much the most significant occasions on which the disclosure of a criminal record is likely to be required, and it is clear that it was primarily with that context in mind that Parliament enacted section 4. It follows that in conferring power on the Secretary of State, by section 4(4), to exclude the operation of sections 4(2) and 4(3) in specified circumstances, Parliament envisaged that there would be occupations in respect of which convictions should be disclosed to a potential employer, professional body or appointing authority notwithstanding that they were spent and notwithstanding that the convicted person might be prejudiced by their disclosure. The scheme for the disclosure of criminal records by the Disclosure and Barring Service (or AccessNI in Northern Ireland) under the Police Act 1997 is carefully tailored to match the disclosure obligations of the person whose record is in question. Under sections 113A(6) and 113B(9) of the Police Act 1997, where the question is asked in circumstances excluded from the operation of the Rehabilitation of Offenders Act 1974 under section 4(4) of the latter Act, it will fall to be disclosed by the Disclosure and Barring Service (or AccessNI in Northern Ireland) notwithstanding that it is spent. This is a coherent scheme of legislation which acknowledges both of the competing public interests to which I have referred, and seeks to achieve a balance between them. Those interests are not only competing but incommensurate. In the nature of things, wherever the line is drawn, it will not be satisfactory from every point of view. The whole issue raises classic policy dilemmas. The underlying policy is precautionary, in line with strong public expectations. The question is whether in adopting that approach the appellants contravened the European Convention on Human Rights.

Article 8 provides:

It is not disputed that article 8 is engaged. It confers a qualified right of privacy, subject to important exceptions for measures which are (i) “in accordance with the law”, and (ii) “necessary in a democratic society in the interests of … public safety … for the prevention of disorder or crime, for the protection of health or morals, or for the protection of the rights … of others.” Conditions (i) and (ii) impose tests of a very different character, with very different consequences. Condition (i) is concerned with the legal basis for any measure which interferes with the right of privacy. Any such measure must not only have some legal basis in domestic law, but must be authorised by something which can properly be characterised as law. This is an absolute requirement. In meeting it, Convention states have no margin of appreciation under the Convention, and the executive and the legislature have no margin of discretion or judgment under domestic public law. Only if the test of legality is satisfied does the question arise whether the measures in question are necessary for some legitimate purpose and represent a proportionate means of achieving that purpose.

The Court of Appeal in England in R (P) v Secretary of State for Justice, R (G) v Chief Constable of Surrey Police and R (W) v Comr of Police of the Metropolis [2018] 1 WLR 3281, and the Court of Appeal in Northern Ireland in In re Gallagher’s Application [2016] NICA 42, upheld the respondents’ cases. Although the reasons of both courts were substantially the same, the fullest analysis of the law is to be found in the judgment of Sir Brian Leveson P in the English cases. He rejected the submission that the Convention required a system of review that would enable each case to be assessed on its own facts. But he held, first, that the legislation was not “in accordance with the law” because, although it discriminated between different categories of offence and convictions, the categories were still too broad. They embraced offences of widely differing relevance, and were therefore liable to operate arbitrarily in a significant number of cases. In particular, he regarded as inconsistent with the legality test: (i) the rule which made all convictions disclosable if there was more than one, because it failed to distinguish between cases which disclosed a relevant pattern of offending and those which did not; and (ii) the rule that required the disclosure of specified serious offences, because it was “insufficiently calibrated so as to ensure that the proportionality of the interference is adequately examined”: [2018] 1 WLR 3281, para 45. Even if the legislation had passed the legality test, the scheme would have been disproportionate to its objective because it was insufficiently “granular” in distinguishing between convictions and cautions of varying degrees of relevance. It will be seen that the reasons why, in his view, the legislation failed the legality and proportionality tests were substantially the same. The scheme was more discriminating than its predecessor, but not discriminating enough.

The respondents submit that because the categories of disclosable conviction or caution are (they say) too wide, and not subject to individual review, the legislation does not have the quality of law. Before I examine this submission in the light of the authorities, it is right to draw attention to some of its more far reaching consequences if it is correct. In the first place, it means that the legislation is incompatible with article 8, however legitimate its purpose, and however necessary or proportionate it may be to deal with the problem in this particular way. That conclusion would plainly have significant implications for the protective functions of the state, especially in relation to children and vulnerable adults. Secondly, it must be remembered that the condition of legality is not a question of degree. The measure either has the quality of law or it does not. It is a binary test. This is because it relates to the characteristics of the legislation itself, and not just to its application in any particular case: see Kruslin v France (1990) 12 EHRR 547, paras 31-32. It follows that if the legislation fails the test of legality, it is incompatible with the Convention not just as applied to those convicted of minor offences like these respondents, but to the entire range of ex-offenders including, for example, convicted child molesters, rapists and murderers. Thirdly, this consequence cannot be confined to the right of privacy. Most Convention rights are qualified by reference to various countervailing public interests. These qualifications are fundamental to the scheme of the Convention. They are what makes it possible to combine a high level of protection of human rights with legitimate measures for the protection of the public against real threats to their welfare and security. For that reason, exceptions corresponding to those in article 8 attach to a number of other Convention rights. They too must also have a proper basis in law. It is fair to say that the jurisprudence of the Strasbourg court has been especially sensitive to the keeping of files on individuals by the state, a practice which was gravely abused by the authoritarian regimes of the 20th century in most of continental Europe. This sensitivity explains why the right of privacy has been extended from covert and intrusive surveillance to the recording of things which would not be regarded as “private” in any other context, for example participation in demonstrations in public places (Segerstedt-Wiberg v Sweden (2007) 44 EHRR 2, para 72) and even public acts of the state itself, such as criminal convictions in an open court of law (MM v United Kingdom (Application 24029/07), 29 April 2013, at para 188). But the question what constitutes law is the same whatever the subject matter. Neither the Strasbourg court nor the courts of the United Kingdom have ever suggested that the condition of legality applies in any different way in article 8 as compared with other articles. In principle, therefore, whatever conclusion we reach in this case about the scope of the condition of legality must apply equally to the exceptions to article 5 (right to liberty and security), article 9 (freedom of thought, conscience and religion), article 10 (freedom of expression), and article 11 (freedom of assembly and association). In none of these articles would there be any scope for distinctions based on judgment or discretion or weighing of broader public interests, even on the most compelling grounds, once the relevant measure failed the respondents’ exacting test of legality.

Nonetheless, the respondents submit that the issue was resolved in their favour by the decision of this court in R (T) v Chief Constable of Greater Manchester Police [2015] AC 49, and that submission was accepted by the courts below. The argument is that, as applied to legislation which applies indiscriminately to a wide range of potentially very different circumstances, T is authority for the proposition that the test of legality requires that the legislation should include safeguards against its arbitrary application, by which is meant the disclosure of matters manifestly irrelevant to an ex-offender’s suitability for employment. T is a recent and considered decision of this court about an earlier version of the statutory scheme before us now. If it means what the respondents submit that it means, it is our duty to follow it unless (which is not suggested) Practice Statement (Judicial Precedent) [1966] 1 WLR 1234 applies. The decision, however, needs to be properly understood in the light of the substantial body of Strasbourg case law on which it was expressly based and the particular domestic legislation with which it was concerned.

It is well established that “law” in the Human Rights Convention has an extended meaning. In two judgments delivered on the same day, Huvig v France (1990) 12 EHRR 528, at para 26, and Kruslin v France (1990) 12 EHRR 547, para 27, the European Court of Human Rights set out what has become the classic definition of law in this context:

The accessibility test speaks for itself. For a measure to have the quality of law, it must be possible to discover, if necessary with the aid of professional advice, what its provisions are. In other words, it must be published and comprehensible. The requirement of foreseeability, so far as it adds to the requirement of accessibility, is essentially concerned with the principle summed up in the adage of the American founding father John Adams, “a government of laws and not of men”. A measure is not “in accordance with the law” if it purports to authorise an exercise of power unconstrained by law. The measure must not therefore confer a discretion so broad that its scope is in practice dependent on the will of those who apply it, rather than on the law itself. Nor should it be couched in terms so vague or so general as to produce substantially the same effect in practice. The breadth of a measure and the absence of safeguards for the rights of individuals are relevant to its quality as law where the measure confers discretions, in terms or in practice, which make its effects insufficiently foreseeable. Thus a power whose exercise is dependent on the judgment of an official as to when, in what circumstances or against whom to apply it, must be sufficiently constrained by some legal rule governing the principles on which that decision is to be made. But a legal rule imposing a duty to take some action in every case to which the rule applies does not necessarily give rise to the same problem. It may give rise to a different problem when it comes to necessity and proportionality, but that is another issue. If the question is how much discretion is too much, the only legal tool available for resolving it is a proportionality test which, unlike the test of legality, is a question of degree.

This much is clear not only from the Huvig and Kruslin judgments themselves, but from the three leading decisions on the principle of legality on which the Strasbourg court’s statement of principle in those cases was founded, namely Sunday Times v United Kingdom (1979-80) 2 EHRR 245, Silver v United Kingdom (1983) 5 EHRR 347 and Malone v United Kingdom (1985) 7 EHRR 14.

Sunday Times v United Kingdom (1979-80) 2 EHRR 245 was the first occasion on which the Strasbourg court addressed the test of legality. It was not a privacy case, but a case about freedom of expression in the context of the English law of contempt of court. The requirement of foreseeability was summarised by the court as follows at para 49:

In Silver v United Kingdom (1983) 5 EHRR 347, para 85, the Strasbourg court adopted this definition and applied it to a complaint of interference with prisoners’ correspondence, contrary to article 8. The court observed at para 88 that the need for precision in the Sunday Times case meant that “a law which confers a discretion must indicate the scope of that discretion.” It was in that context that the court addressed the question of safeguards, at para 90:

A fuller statement of the same principle appeared in the important judgment in Malone v United Kingdom (1985) 7 EHRR 14. The context was telephone tapping, which under the system then in operation in the United Kingdom was authorised by warrants of the Home Secretary under purely administrative powers with no statutory basis. The power exercisable by the Home Secretary was agreed to be lawful as a matter of domestic law, but no law constrained or limited his discretion. After reciting the Sunday Times test, the court continued at para 67:

The French system for tapping telephones was criticised on broadly similar grounds in Huvig and Kruslin. In the latter case, at paras 35-36, the court observed:

In three notable later cases, Amann v Switzerland (2000) 30 EHRR 843, Rotaru v Romania (2000) 8 BHRC 449 and S v United Kingdom (2009) 48 EHRR 50, the same principles were applied to the retention in police records of personal information. Amann was another case about phone tapping. The court held that the retention of the fruits of a tap in police files did not satisfy the legality test even on the footing that the tap itself was in accordance with law. The decision was expressed to be based on the statement of principle in Malone (para 56), and on a finding (para 62) that “Swiss law does not indicate with sufficient clarity the scope and conditions of exercise of the authorities’ discretionary power in the area under consideration.” In Rotaru, the applicant objected to the retention on the files of the Romanian state security service of information, some of it false, about his dissident activities in the early years of the post-war communist regime nearly half a century before. His case (see para 50), which was upheld by the Grand Chamber, was that this was

As can be seen from these citations, from the outset the Strasbourg court has treated the need for safeguards as part of the requirement of foreseeability. It has applied it as part of the principle of legality in cases where a discretionary power would otherwise be unconstrained and lack certainty of application. This may be illustrated by reference to the subsequent decisions in Liberty v United Kingdom (2009) 48 EHRR 1 and Gillan v United Kingdom (2010) 50 EHRR 45. Liberty concerned the bulk interception of telephone communications passing through submarine cables terminating in the United Kingdom. There was statutory authority for the interception, but as the court pointed out at para 69, the legal framework did not have the quality of law. This was because

It is against that background that one must approach the decision in MM v United Kingdom (Application No 24029/07), 29 April 2013. The case concerned the retention and disclosure by the police of records of cautions in Northern Ireland. The applicant had received a caution for child abduction in 2000 in unusual circumstances which provided very strong mitigation. Its disclosure had nevertheless resulted in the failure of two applications for jobs involving care work. She had accepted the caution on an assurance that it would be deleted from police records after five years, which was the practice at the time. But the practice subsequently changed, and her attempts in 2006 and 2007 to have the caution deleted were unsuccessful. The gravamen of her complaint was not about the past disclosures, but about the retention of the caution on police files, which exposed her to the risk of disclosure in future whenever she applied for a job requiring a criminal record certificate. Much of the analysis of the Strasbourg court needs to be understood in that light.

The Strasbourg court examined in detail the complex and changing legal basis on which criminal records were handled in Northern Ireland. There were three stages of the process to be considered, namely (i) collection of data, (ii) its retention in the records of the authorities, and (iii) its disclosure to third parties. At the time when the caution was given, convictions in Northern Ireland were recorded under statutory regulations but the record was retained and disclosed under common law powers. The regime governing cautions was different. They were recorded as well as retained and disclosed under common law powers. The only legal limitation on the exercise of these powers was the Data Protection Act 1998. On 1 April 2008, the system was changed when Part V of the Police Act 1997 was brought into force in Northern Ireland by the Police Act 1997 (Commencement No 11) Order (SI 2008/692). This introduced to Northern Ireland the system (already in force in England and Wales) under which the disclosure of all recorded and retained convictions and cautions, including warnings and reprimands, was mandatory. It did not affect the recording or retention of cautions, which continued to be governed by common law powers. The new regime in Northern Ireland was relevant to MM’s case because the Police Act would thereafter have applied to the disclosure of her caution in connection with any fresh job application after April 2008.

The Strasbourg court was invited by the United Kingdom government to treat as part of the legal framework governing collection and retention of data the statutory Code of Practice for the Management of Police Information, issued by the Secretary of State in 2005 under section 39A of the Police Act 1996. This established general standards for the management of police information, and provided for the issue of Guidance by the Association of Chief Police Officers (“ACPO”) in 2006 and 2010 which police forces were required to comply with. These documents, however, applied directly only in England and Wales. As the court noted at para 33, although the statutory Code of Practice was available for adoption by police forces elsewhere, it was not clear that it had been adopted in Northern Ireland. (In any event, since section 39A of the Police Act 1996 did not extend to Northern Ireland, it could have had only administrative and not statutory force there.)

The court held that the scheme did not have the quality of law, either before or after April 2008. The principle on which it proceeded was stated at the outset of its analysis, at para 193, by reference to the dual requirements of accessibility and foreseeability derived from its earlier case law, including Malone and Liberty:

The pre-2008 position in Northern Ireland as regards cautions was an obvious example of unconstrained discretionary power. For present purposes, however, the judgment is mainly of interest for its treatment of the position in Northern Ireland after April 2008 under the Police Act 1997. MM contended that the caution should have been deleted so as not to be available for disclosure under the new regime. The court recorded (para 195) its view that article 8 was engaged by the whole process of collection, retention, use and disclosure of data on police files. It recognised (para 199) that

In MM, the court regarded the system of recording and retention of criminal convictions in Northern Ireland as “indiscriminate and open-ended”: see para 199. It went on to say that such a system

In the most recent decision of the Strasbourg court, Catt v United Kingdom (Application No 43514/15), MM was treated at para 94 as authority for the following proposition:

This is, moreover, the analysis which the English courts have given the Strasbourg case law.

In R (Gillan) v Comr of Police for the Metropolis [2006] 2 AC 307, para 34, Lord Bingham put the matter in this way:

In R (Purdy) v Director of Public Prosecutions [2010] 1 AC 345, at para 41, Lord Hope observed that the Convention’s concept of law

R (T) v Chief Constable of Greater Manchester Police [2015] AC 49 concerned the regime governing disclosure of criminal records in England as it stood before the changes introduced with effect from March 2014. This court held that that regime lacked the quality of law. The leading judgment on this point was delivered by Lord Reed, with whom Lord Neuberger, Baroness Hale and Lord Clarke agreed. There was very little discussion of the Northern Ireland system for managing criminal records considered in MM, because Lord Reed proceeded on the basis that the English legislation under consideration was indistinguishable from it: see paras 100, 119. This was not entirely correct. As I have explained, the Code of Practice and associated ACPO Guidance governing the management of police information in England had statutory force in England but not in Northern Ireland. But for reasons which will appear, I do not think that that difference was critical to the outcome, either in MM or in T.

The essence of Lord Reed’s reasoning appears at paras 113, 114 and 119 of the judgment:

This decision is treated by the respondents as authority for the proposition that a measure may lack the quality of law even where there is no relevant discretion and the relevant rules are precise and entirely clear, if the categories requiring to be disclosed are simply too broad or insufficiently filtered. I do not accept this submission.

In the first place, it is hardly conceivable that Lord Reed intended to effect a revolution in this branch law the law, with such far-reaching results, and without acknowledging the fact. On the contrary, it is clear that he did not. He regarded himself as applying the established case law of the Strasbourg court. All of the Strasbourg decisions on which he based his analysis, notably Kopp, Malone, Rotaru, Amann and MM, had been expressly based on the classic dual test of accessibility and foreseeability. In particular, Lord Reed regarded the decision in MM as reflecting the earlier jurisprudence. In all of these cases, safeguards were said to be required in order to constrain administrative discretions which, unless constrained, undermined the foreseeability of the relevant measures. Lord Reed’s reference to the need for precision if something is to have the character of law shows that he had the foreseeability test well in mind. He is echoing the observations in Sunday Times, (para 49), Silver (para 88) and Malone (para 70), that a person must be able to discover from the law itself precisely what effect, in the circumstances of his case, its application will have upon him.

Secondly, in distinguishing between the legality test and the proportionality test, Lord Reed pointed out at para 114 that:

Thirdly, at para 119, where Lord Reed explains his disposal of the appeal, he is expressly applying MM. That decision, as I have pointed out, had been based on the perceived “absence of a clear legislative framework for the collection and storage of data” (emphasis supplied) which would fall to be mandatorily disclosed under sections 113A and 113B of the Police Act 1997. The absence of any “clear legislative framework” for the recording and retention of criminal records meant that the body of data falling to be mandatorily disclosed was of uncertain content. The uncertain character of the system for retaining criminal records affected the lawfulness of their disclosure. Hence the relevance of the indiscriminate character of the disclosure which Lord Reed criticises at para 119.

In a precedent-based system, the reasoning of judges has to be approached in the light of the particular problem which was before them. There is a danger in treating a judge’s analysis of that problem as a general statement of principle applicable to a whole area of law. Lord Reed’s observations in T cannot in my opinion be applied generally to the whole relationship between legality and proportionality in the Convention, even in cases where the relevant domestic rule satisfied the tests of accessibility and foreseeability. It is noticeable that the principle of legality was stated in narrower terms by Baroness Hale, Lord Reed and Lord Hodge in their joint judgment in Christian Institute v Lord Advocate [2016] UKSC 51. They put it in this way at paras 79-80:

The rules governing the disclosure of criminal records, both by ex-offenders themselves under the Rehabilitation of Offenders Act 1974 and by the Disclosure and Barring Service and AccessNI under the Police Act 1997, are highly prescriptive. The categories of disclosable convictions and cautions are exactly defined, and disclosure in these categories is mandatory. Within any category, there is no discretion governing what is disclosable. There is no difficulty at all in assessing the proportionality of these measures because, subject to one reservation (see the following paragraph), their impact on those affected is wholly foreseeable.

The one reservation arises from a submission made to us that on an application for an enhanced criminal record certificate under section 113B of the Police Act, it would be open to a chief officer of police, if he thought that it “ought to be included”, to call for the inclusion in the certificate of a conviction or caution which was not a “relevant matter” because it did not fall within any of the defined categories of disclosable conviction under section 113A(6). I assume (without deciding) that this course was open to the chief officer. But it would not deprive the legislation of the quality of law, because section 113B(4A) requires chief officers to exercise this function having regard to statutory guidance published by the Secretary of State. This provision was inserted by the Protection of Freedoms Act 2012, which was shortly followed by the publication of detailed guidance in July of that year. It is well established that guidance provided for by statute may constitute “law” for the purpose of the Convention: R (Purdy) v Director of Public Prosecutions [2010] AC 345, para 47 (Lord Hope). The judgment of the chief officer is subjected to carefully drawn constraints that themselves have the quality of law.

In these circumstances, the only basis on which it could be said that the legislation lacks the quality of law is that the content of the classes of criminal record available for mandatory disclosure is itself uncertain, because of the uncertain or discretionary character of the rules governing their retention in the Police National Computer, or the Causeway System which serves the same purpose in Northern Ireland. This was, as we have seen, the criticism made of the earlier version of the legislation as it applied in Northern Ireland, by the Strasbourg court in MM. In the three English cases it was argued in the Court of Appeal that the retention of their records on the Police National Computer was itself a breach of article 8 of the Convention. The argument was rejected and has not been repeated before us. It would not in any event have affected the legality of the system of disclosure for the following reason. As I have pointed out above (para 26), what is consistent with the legality test at the stages of collection and retention, may depend on how much of it is liable to be disclosed under the Police Act. The reason why the uncertain content of the criminal record database was so significant in MM was that at the relevant time any conviction or caution on the database was liable to “indiscriminate” disclosure, without exception. That has not been the case either in England and Wales or in Northern Ireland since 2014. It is no longer correct to say, as Lord Reed quite rightly did about the unamended scheme considered in T (para 119), that the statutory scheme fails to draw distinctions by reference to the nature of the offence, the disposal of the case or the time which has elapsed since the offence took place. It is still the case that it fails to draw distinctions based on the relevance of the conviction to a potential employer on more general grounds; and it still does not provide a mechanism for the independent review of disclosure. However, even on the most expansive view of what was decided in T, nothing in that case suggests that these two factors are on their own enough to deprive the legislation of the quality of law. The current legislation distinguishes, for the purpose of disclosure, between different categories of conviction or caution, depending on the gravity of the offence, the age of the offender at the time and the number of years which have passed. Of course, there may be arguments for more or fewer, or wider or narrower categories, but the legality test is a fundamentally unsuitable instrument for assessing differences of degree of this kind. A decision that the current regime governing retention and disclosure of criminal records lacked the quality of law would mean that it would be incompatible with the Convention even if, hypothetically, it could be shown that nothing short of it would sufficiently protect children and vulnerable adults from substantial risks of abuse or protect the public interest in the appointment of suitable people to highly sensitive positions. I decline to accept that proposition. It would have the practical effect of equating the right of privacy with such absolute provisions of the Convention as the prohibition of torture and slavery, when the terms of article 8 show that the right of privacy is qualified.

I conclude that the current scheme of disclosure under the Rehabilitation of Offenders Act 1974 (as amended) and the Police Act 1997 (as amended), and the corresponding legislation in Northern Ireland, are in accordance with the law for the purposes of article 8 of the Convention.

There are, as it seems to me, only three ways in which the question of disclosing criminal records of candidates for sensitive occupations could have been addressed: (i) by legislating for disclosure by reference to the pre-defined categories of offence, offender or sentence in the legislation as it stands; (ii) by legislating for disclosure by reference to some differently drawn categories of offence, offender or sentence; or (iii) by legislating for disclosure by reference to the circumstances of individual cases, as ascertained by some process of administrative review. Accordingly, two questions fall to be decided. The first is whether the legislation can legitimately require disclosure by reference to pre-defined categories at all, as opposed to providing for a review of the circumstances of individual cases. If it can, then the second question is whether the boundaries of these categories are currently drawn in an acceptable place. It is common ground that, for the purpose of assessing the proportionality of the scheme, the legislature and ministers exercising statutory powers have a margin of judgment, within limits.

I shall deal first with the question whether the legislation can legitimately require disclosure by reference to pre-defined categories at all, rather than the circumstances of each case. If not, then manifestly the present legislative scheme will not pass muster.

In principle, the legitimacy of legislating by reference to pre-defined categories in appropriate cases has been recognised by the Strasbourg court for many years. The fullest modern statement of the law is to be found in its decision in Animal Defenders International v United Kingdom (2013) 57 EHRR 21, where the court summarised the effect of a substantial body of earlier case law. At paras 106-110, the court observed:

The court’s reference in para 108 to the risk of uncertainty is supported by a footnote citation of its earlier decision in Evans v United Kingdom (2008) 46 EHRR 34. In that case, it held that the absence of any provision for individual scrutiny in legislation requiring the consent of both parties to the implantation of stored embryos was consistent with article 8 of the Convention. The Grand Chamber found (para 60) that “strong policy considerations underlay the decision of the legislature to favour a clear or ‘bright-line’ rule which would serve both to produce legal certainty and to maintain public confidence in the law in a sensitive field.” It went on to observe, at para 89:

In those cases where legislation by pre-defined categories is legitimate, two consequences follow. First, there will inevitably be hard cases which would be regarded as disproportionate in a system based on case-by-case examination. As Baroness Hale observed in R (Tigere) v Secretary of State for Business, Innovation and Skills [2015] 1 WLR 3820, para 36, the Strasbourg court’s jurisprudence “recognises that sometimes lines have to be drawn, even though there may be hard cases which sit just on the wrong side of it.” Secondly, the task of the court in such cases is to assess the proportionality of the categorisation and not of its impact on individual cases. The impact on individual cases is no more than illustrative of the impact of the scheme as a whole. Indeed, as the Strasbourg court pointed out at para 109 of Animal Defenders, the stronger the justification for legislating by reference to pre-defined categories, the less the weight to be attached to any particular illustration of its prejudicial impact in individual cases. In my judgment, the legislative schemes governing the disclosure of criminal records in England and Wales and Northern Ireland provide as good an example as one could find of a case where legislation by reference to pre-defined categories is justified. I reach that view for four main reasons.

First, it is entirely appropriate that the final decision about the relevance of a conviction to an individual’s suitability for some occupations should be that of the employer. Only the employer can judge whether the particular characteristics of the particular job make it inappropriate to employ the particular ex-offender. Very often, this will be a judgment that the employer makes in the course of discussion with the candidate in the light of what is disclosed. The employer will bear the responsibility for the consequences of its choice, and in sensitive appointments the responsibility may be a heavy one. In order to discharge that responsibility with the thoroughness that the public interest requires, the employer must have access to potentially relevant information about a candidate’s past. He may end up by disregarding some or all of it as irrelevant or insufficiently weighty. But unless the decision is to be taken out of his hands, he must be told about any criminal record which might reasonably influence him, even if further consideration or discussion of the circumstances with the candidate may ultimately cause him to disregard or attach limited weight to it. By comparison, the administrative authorities responsible for disclosure know only (i) the job title, which usually gives only the most general notion of what the job entails; and (ii) the broad category of offence for which the candidate was convicted or cautioned, the implications of which may be affected by a wide variety of mitigating or aggravating circumstances that are not apparent from the criminal record database. A system of administrative review on the application of the candidate may be possible. It has existed in Northern Ireland since 2016. Such a system enables the disclosure authority to take into account the candidate’s representations. But it cannot enable the authority to take over the employer’s function of assessing the candidate’s suitability for the particular employment. It might be possible to design a system under which rather more information about the job was supplied to the disclosure authority than is provided for under the forms currently prescribed. It might be possible to design a system under which the disclosure authority could call for further information from the employer, but that would give the game away. The employer would know that there was something there, and the consequence for the candidate would in many cases be worse than disclosure of what might turn out to be a very minor offence. None of these possibilities can realistically be thought to displace the employer’s judgment of the candidate’s suitability. It follows that it cannot be right to say that as a matter of law the United Kingdom must have a scheme of disclosure which depends on an examination of the circumstances of individual cases by someone other the employer.

Secondly, the objection to disclosure by category is based on the argument that employers cannot be trusted to take an objective view of the true relevance of a conviction. But the material available to support that objection is distinctly thin. There is some survey evidence which is said to support it, although the generality and hypothetical character of the questions and the very summary form of the answers make it hard to attach much weight to it. Lord Neuberger suggested in R (L) v Comr of Police of the Metropolis (Secretary of State for the Home Department intervening) [2010] 1 AC 410 at para 75 that in the majority of cases the disclosure of any criminal record would be “something close to a killer blow”. However, as this court recently pointed out in R (AR) v Chief Constable of Greater Manchester Police [2018] 1 WLR 4079, para 75, it is far from clear on what if any empirical evidence Lord Neuberger’s observation was made. Realistically, it must be assumed that some employers will take the line of least risk, and decline to employ ex-offenders on principle, especially if there is an alternative candidate without a criminal record. But the evidence before us does not bear out the suggestion that this is the norm. Under sections 113A(2) and 113B(2), applications for criminal record certificates must be made or countersigned by a registered body. Employers and the registered bodies who sponsor their applications are required to comply with a Code of Practice issued by the Secretary of State under section 122 of the Police Act 1997. Registered bodies may lose their registration if they fail to do this themselves and to enforce the code on employers. The Code of Practice requires employers, among other things, to have a written policy, available to candidates on request, concerning the suitability of ex-offenders, to notify candidates of the potential impact of a criminal record and to discuss with candidates the content of any disclosure before withdrawing an offer of employment. There is no evidence before us that the Code of Practice is ignored on a significant scale, either in letter or in spirit. A high proportion of employers in cases where criminal record certificates are required will in any event be in the public sector, and they are particularly likely to comply. But, be all that as it may, for as long as the employer has the ultimate right to decide and the legal responsibility to decide carefully, and is the only person in a practical position to do so, the risk that some employers may take too absolute a line is inescapable.

Thirdly, in this context, the value of certainty is particularly high. The regimes governing disclosure by the candidate under the Rehabilitation of Offenders Act and by the Disclosure and Barring Service or AccessNI under the Police Act are carefully aligned. Any legislation governing disclosure under the Police Act must take account of the fact that the candidate for sensitive positions will generally have been asked to disclose past convictions and cautions voluntarily. Section 4 of the 1974 Act entitles the candidate to treat that request as not relating to spent convictions, subject to exceptions identified in subordinate legislation. Those spent convictions which are excluded from section 4 and therefore disclosable by the candidate himself must necessarily be identified by category. There is no room for a case-by-case review of the particular facts in that context, because candidates must know where they stand at the time when they complete the application form, ie before any application is made for disclosure under the Police Act. The offences falling to be disclosed under the Police Act must substantially correspond to those disclosable by the candidate under the Rehabilitation of Offenders Act. A regime for disclosure by the Disclosure and Barring Service or AccessNI which allowed for discretionary exceptions dependent on the facts of the case would not help the candidate if he has already had to disclose all convictions in the relevant category himself. What this suggests is that any advantages of an administrative review of the circumstances of individual cases will have been gained at the expense of foreseeability. This has a significant cost to the candidate himself.

It will be apparent that the justification for legislating by reference to categories of offence or offender is much more than a question of administrative convenience or practicality. It goes to the whole purpose of the scheme, which is to enable employers properly to perform their function of vetting candidates for sensitive occupations, and to enable candidates themselves to know what is disclosable, in the first instance by themselves. There are, however, and this is the fourth reason, important issues of practicality involved. Some four million applications for criminal record certificates are made every year in England and Wales. They have to be dealt with promptly, because a conditional offer of employment will commonly have been made to the candidate. A system of individual assessment would require an assessment to be made or reviewed according to, among other things, the circumstances of the offence, the sentencing remarks of the judge, any relevant mitigating or aggravating factors, and presumably any representations of the candidate. The evidence on behalf of the Secretary of State is that this is not a practical proposition in the case of a volume of disclosure applications as large as that in England and Wales. The view taken by ministers was therefore that a mechanical process of disclosure by category was the only one consistent with basic levels of efficiency. Of course, beyond a certain point, administrative efficiency cannot justify visiting an injustice upon candidates. But it is particularly difficult for a court to determine where that point lies. It is true that any administrative problems appear to have been overcome in Northern Ireland. But Northern Ireland is a much smaller jurisdiction.

Taking these considerations together, they suggest that although it may be possible to abandon category-based disclosure in favour of a system which allowed for the examination of the facts of particular cases, there would be a cost in terms of protection of children and vulnerable adults, foreseeability of outcome by candidates, consistency of treatment, practicality of application, and delay and expense, without necessarily achieving much more for ex-offenders than the current system. Once it is accepted that a category-based scheme of disclosure is justifiable, it must inevitably follow that some candidates will find themselves in a category apparently more serious than the facts of their particular case really warrant. The cases which have given rise to these appeals illustrate the point. G was reprimanded at the age of 13 for offences of sexual activity with a child. P received a caution for theft and was convicted shortly afterwards of another offence of theft. W was convicted of assault occasioning actual bodily harm. These are all, in the generality of cases, serious offences which in a category-based system would rightly be disclosable in connection with a sensitive occupation, especially one involving contact with children or vulnerable adults. In each case, it is only the detailed circumstances that show that the actual offence was very minor. Conversely, Ms Gallagher was convicted of carrying children under 14 without a seatbelt, and convicted again of the same offences two years later. This is a minor offence, but if the job for which Ms Gallagher had applied had involved driving children it would have been difficult to justify withholding these convictions from a potential employer. Some employers might legitimately be concerned that her record disclosed a more general lack of concern with safety which was unacceptable to them.

Against that background, I turn to the next question, which is whether the legislation before us draws the boundaries of the relevant categories in an acceptable place.

As it stood at the relevant time, the statutory schemes in both England and Wales and Northern Ireland substantially reflected the recommendations of Ms Sunita Mason. She was an experienced district judge, a former chair of the Law Society’s Family Law Committee and the Government’s Independent Adviser for Criminality Information Management. Ms Mason was asked in 2009 to conduct a review of the retention and disclosure of criminal records held on the Police National Computer. Her report, “A Balanced Approach”, was published in March 2010. It recommended that disclosures to employers should be filtered and that a panel representing the various interested parties should advise on the filtering rules. The Secretary of State subsequently established the Criminal Records Review to make proposals on the “balance between respecting civil liberties and protecting the public.” That review was also conducted by Ms Mason, in conjunction with the Independent Advisory Panel for the Disclosure of Criminal Records. Her two reports, published in February and December 2011, took account of the views of a broad range of experts and consultees drawn from the criminal justice system, the police and the judiciary, the teaching profession, and NGOs involved with children, vulnerable adults and ex-offenders. It was also informed by summaries of the disclosure systems operating in 26 other countries. Ms Mason made proposals for removing old and minor offences from the scope of disclosable convictions and cautions. The Panel recommended that spent convictions and cautions should be filtered by category, according to the period of time which had elapsed, that “particular care should be taken before considering any sexual, drug related or violent offence type for filtering”, that where a person has received a conviction or caution for any offence which is not categorised as minor, all his convictions and cautions (including minor ones) should be disclosed, and that “the filtering rules should be both simple and understandable to individuals who are users and/or customers of the disclosure service.” The Panel thought that “extra consideration” should be given to minor offences committed by persons under the age of 18. Although agreed on the principles, however, the Independent Advisory Panel did not agree on the criteria. The recommendations concerning these were accordingly Ms Mason’s. She proposed the disclosure of all convictions categorised as not being minor, all convictions where there was more than one, and the filtering out of single minor spent convictions by adults after three years and by persons under the age of 18 after six months. She proposed that further consideration should be given to the problems of defining minor offences.

The problems of defining minor offences are described in a witness statement of Mr John Woodcock, then Head of Criminal Records Policy within the Safeguarding and Public Protection Unit of the Home Office. In summary, the two main criteria available were the character of the offence as defined by law and the severity of the sentence, or some combination of the two. Each of these criteria was liable to produce capricious results at the margins, as Mr Woodcock demonstrates. I have already referred to those associated with the character of the offence. The use of sentencing as a criterion was also problematic. This was because mitigating factors affecting sentence will not necessarily be relevant to the assessment of the risks associated with sensitive employments. Moreover, every additional refinement added to the system to make it more accurate, was liable also to make it more complex and less easy for candidates to understand.

The filtering criteria proposed by Ms Mason were adopted by ministers in framing the amendments to the scheme in 2013, except that the periods of “currency” adopted for single minor offences were longer. I have summarised the criteria on which minor offences were filtered out of criminal records at para 9 above. It was based on a combination of (i) the sentence (all offences resulting in a custodial sentence were disclosable), (ii) the legal definition of the offence (the sexual and violent offences listed in section 113A(6D) and in Schedule 15 to the Criminal Justice Act 2003 were disclosable in all cases), (iii) the period which had elapsed since the conviction or caution and (iv) the age of the offender at the time of the disposal. In this form, the statutory orders were approved by Parliament under the positive resolution procedure, with bipartisan support.

As the Strasbourg court pointed out in Animal Defenders (para 108), the assessment of the defining factors in a category-based scheme is a matter for the state, and the quality of the examination of the options is likely to be important. I have summarised the history of the process which led to the current legislative scheme in order to make two points. First the scheme is the result of substantial research and intensive consultation with a wide range of interested and expert groups and individuals. Secondly, it is apparent that while there is broad agreement on the need for a category-based system of disclosure and the basic principles which should govern it, there is no consensus about where the lines should be drawn. This is not particularly surprising, because there is no solution which could satisfy all of the main desiderata in the design of such a system. No one suggests that the courts can or should design the system themselves in proceedings for judicial review. The function of the courts is an essentially negative one, namely to identify which schemes are incompatible with the Convention. At the same time, a court can only be satisfied that a particular scheme is incompatible with the Convention if it is in a position to say what is wrong with it.

What is wrong with the design of the categories employed in the legislative scheme before us? On the footing that disclosure by categories is justified in principle, the respondents’ objections to the current system really amount to saying that the balance between the risk of blighting the prospects of ex-offenders and the risk of appointing unsuitable persons to sensitive positions has been drawn in a place which puts too much emphasis on the latter and not enough on the former. They also say that the balance has been drawn in a different way in Northern Ireland and Scotland. Yet a balance of this kind necessarily involves a difficult value judgment. All that a judge can say is that he or she would have drawn it in a different place. But that, with respect, is not the test. We may think that a better scheme could have been devised or that the categories could have been differently drawn, or that too much weight has been given to the risk of unsuitable appointments and not enough to the rehabilitation of offenders. A more “granular” categorisation has been applied in Scotland to cases involving risks to vulnerable groups since 2007, and a system of administrative review on the application of an ex-offender has existed in Northern Ireland since 2016. There may be lessons to learn from their experience. But none of this means that the scheme lies outside the margin of judgment properly allowed to the legislator or the Secretary of State on whom the legislator has laid the task of defining the exceptions to the rehabilitation regime. In my judgment it is not possible for us to say, consistently with the proper role of a court of review, that the carefully drawn categories employed in this scheme are disproportionate.

To this analysis, I would make two exceptions.

The first concerns the multiple conviction rule. Sections 113A(3) and (6)(b) and 113B(3) and (9)(b) of the Police Act 1997 provide that where a person has more than one conviction of whatever nature, any conviction of whatever nature is a “relevant matter” falling to be disclosed in a criminal record certificate. Unlike the other “relevant matters” calling for disclosure, the multiple conviction rule does not, properly speaking, define a category of offence or offender. It is in reality an aggravating factor affecting the significance of an offence. Its rationale is that the criminal record of a serial offender is more likely to be relevant to his suitability for a sensitive occupation, because the multiplicity of convictions may indicate a criminal propensity. In itself, that is an entirely legitimate objective of a legislative provision of this kind. The rule as framed is, however, a particularly perverse way of trying to achieve it. It applies irrespective of the nature of the offences, of their similarity, of the number of occasions involved or of the intervals of time separating them. As framed, therefore, the rule is incapable of indicating a propensity. It may coincidentally do so in some cases, but probably does not in a great many more. Its eccentric consequences may be illustrated by the facts of the two appeals in which the multiple conviction rule was the basis on which disclosure was required, those of P and Lorraine Gallagher. In P’s case the two minor thefts for which she received a caution and a conviction were only disclosable because she had also failed on the second occasion to surrender to her bail. These offences were not only too minor but too disparate to suggest a propensity to even the most suspicious mind. As to Ms Gallagher, I have already observed that her failure on two occasions to secure children in the back of her car might have been relevant to her proposed employment if it had involved driving children about. But, even if she had not committed a further offence in 1998, her convictions of 1996 would have been disclosable simply because there were four unsecured persons in the car at the time, each of whom gave rise to a distinct conviction. A rule whose impact on individuals is as capricious as this cannot be regarded as a necessary or proportionate way of disclosing to potential employers criminal records indicating a propensity to offend.

The second exception concerns warnings and reprimands administered to young offenders under sections 65 and 66 of the Crime and Disorder Act 1998 replaced, since 2013, by youth cautions under section 66ZA. Warnings and reprimands were not a penal procedure. As Lord Bingham put it in relation to warnings in R (R) v Durham Constabulary [2005] 1 WLR 1184 (HL), although they required the offender to have admitted the offence, they constituted a “preventative, curative, rehabilitative or welfare-promoting” disposal: see paras 14-15. A caution administered to an adult requires consent. However, a warning or reprimand given to a young offender whose moral bearings are still in the course of formation, requires no consent and does not involve the determination of a criminal charge. Its purpose is wholly instructive, and its use as an alternative to prosecution is designed to avoid any deleterious effect on his subsequent life. Its disclosure to a potential employer would be directly inconsistent with that purpose. In my view the inclusion of warnings and reprimands administered to a young offender among offences which must be disclosed is a category error, and as such an error of principle. I would expect the same to be true of the current regime governing youth cautions, but we were not addressed on that question and it is neither necessary nor appropriate to decide it on this appeal.

P’s convictions and caution were disclosable only by virtue of the multiple conviction rule. In England and Wales, the rule requiring disclosure of the entire record where there are multiple convictions is embodied in primary legislation, namely section 113A(6)(b) of the Police Act 1997 (as amended). The Divisional Court made a declaration of incompatibility in respect of that provision, which was affirmed by the Court of Appeal. The Secretary of State’s appeal against that order must be dismissed, albeit on grounds narrower than those of the Court of Appeal.

That leaves to be considered in the case of P the corresponding exclusion from section 4 of the Rehabilitation of Offenders Act 1974, which is contained in article 2A(3)(c) of the Rehabilitation of Offenders Act 1974 (Exceptions) Order (SI 1975/1023), as amended with effect from 2014. Since the multiple conviction rule depends on subordinate legislation, it was open to the Divisional Court to quash article 2A(3)(c) as an unlawful act, and it was invited to do so. The Divisional Court declined the invitation and contented themselves with a declaration that the amended 1975 Order could not be read down so as to be compatible with article 8. The Court of Appeal dismissed P’s cross-appeal on that point. In both cases, the reason was that while the amendment was incompatible with article 8 so far as it was applied to P, it would not be so in all cases. Mr Southey QC, who appeared for P, has pursued his cross-appeal before us. The reasoning of the Divisional Court and the Court of Appeal gives rise to difficulty on the footing (accepted by them) that the whole legislative scheme lacks the quality of law, for as I have explained that is an all or nothing question. However, I have concluded that it does have the quality of law, and that the only objection to article 2A(3)(c) of the amended 1975 Order is that it is disproportionate. In a case where legislation by category is appropriate, as I have held it to be in this case, the fact that the categorisation may bear disproportionately on the complainant is not decisive: see para 49 above. What is disproportionate is the creation by article 2A(3)(c) of the amended 1975 Order of a category of disclosable convictions and cautions which depends on the multiple conviction rule. On that footing it would be open to this court to quash that article. Nonetheless, the making of such an order is discretionary, and I would decline to make it in this case. The reason is that it would introduce a discrepancy between the disclosures required of the Disclosure and Barring Service under the Police Act 1997 (the relevant provisions of which must stand unless and until amended or repealed by Parliament) and the disclosure required of the ex-offender under the Rehabilitation of Offenders Act 1974. This would authorise the ex-offender to withhold disclosure of something that would then have to be disclosed in a certificate issued by the Disclosure and Barring Service. In the circumstances, the appropriate course would be simply to vary the order of the Divisional Court by adding a declaration that article 2A(3)(c) is incompatible with article 8 of the Convention.

Lorraine Gallagher’s case also turns on the multiple convictions rule. As it happens, the disclosure made no difference to the fate of her job application in 2014, because it is clear from the uncontentious facts that the job offer was withdrawn because of the concealment of the 1998 convictions and not because of the criminality disclosed in the certificate. She is, however, entitled to relief, because no disclosure would have been made but for section 113A(6)(b) of the Police Act 1997 (as amended) and article 1A(2)(c) of the Rehabilitation of Offenders (Exceptions) Order (Northern Ireland) 1979 (SR(NI) 1979/195) (as amended) (which corresponds to article 2A(3)(c) of the 1975 Order in England and Wales). Treacy J in the High Court in Northern Ireland made two orders. The first order dealt only with Ms Gallagher’s application for judicial review of the automatic disclosure of her record under the Police Act 1997 (as amended). It simply allowed the application without any further relief. The second order dealt in addition with the position under the Rehabilitation of Offenders (Exceptions) Order (Northern Ireland) 1979 (as amended). It declared in paragraph (a) that the 1979 Order violated Ms Gallagher’s rights under article 8 because it “fails the test of necessity”; and in paragraph (b) that both the 1979 Order (as amended) and Part V of the Police Act 1997 (as amended) violated Ms Gallagher’s rights under article 8 for the additional reason that they lacked the quality of law. The Court of Appeal of Northern Ireland dismissed the Secretary of State’s appeal. There is no cross-appeal in Ms Gallagher’s case. It follows from the conclusions that I have reached that I would vary Treacy J’s second order so as to limit paragraph (a) of his declaration to article 1A(2)(c) of the 1979 Order (the only provision relevant to her case); and to delete paragraph (b).

In G’s case, Blake J declared (a) that Part V of the Police Act 1997 (as amended) was incompatible with article 8 of the Convention to the extent that it required the mandatory disclosure of his reprimand for offences contrary to section 13 of the Sexual Offences Act 2003; and (b) that regulations made under the Rehabilitation of Offenders Act 1974 required amendment in the light of (a). The Court of Appeal affirmed declaration (a) and set aside declaration (b). There is no cross-appeal. For the reasons which I have given, which are narrower than those of the Court of Appeal, I would dismiss the Secretary of State’s appeal.

In W’s case, his conviction for assault occasioning actual bodily harm has not been disclosed. His concern is with its prospective disclosure were he to apply for a teaching job. Assault occasioning actual bodily harm is an offence specified in Schedule 15 to the Criminal Justice Act 2003. As such it is excluded from section 4 of the Rehabilitation of Offenders Act 1974 (as amended), by article 2A(5)(d) of the 1979 Order, and falls to be disclosed in a Criminal Record Certificate under the corresponding provision of the Police Act 1997 (as amended). Simon J, who heard W’s application for judicial review in the High Court, dismissed it, but the Court of Appeal allowed the appeal on proportionality (as well as legality) on the ground that it was “difficult to see how publication of this detail, 31 years on, is relevant to the risk to the public, or proportionate and necessary in a democratic society.” I regret that I am unable to agree, essentially for the reason given by Simon J. Once it is accepted (as I have accepted) that disclosure may properly be required by categories, the question is whether the choice of category is proportionate, not whether it impacted disproportionately on particular cases. Disclosure by categories must inevitably produce a disproportionate impact in some cases. In my opinion, it was legitimate to include assault occasioning actual bodily harm among the offences which were sufficiently serious to require disclosure. It is a violent offence which may be extremely serious. As Simon J pointed out, it may attract an extended sentence of imprisonment. It was also legitimate not to include a temporal limit in the definition of the category of violent or sexual offences requiring disclosure. Any temporal limit would have risked the non-disclosure of the worst cases in the category. The limit would presumably have had to vary with the offence. There would be complex additional problems of definition, thereby making the scheme notably more complex than it already is. For example, a provision imposing a temporal limit on serious offences would presumably have had to differentiate between cases where the offender went on to commit further such offences and cases where (like W) he did not. I cannot regard the existing categorisation as illegitimate, or as notably more problematical than any other categorisation. Hard cases like W’s must ultimately be left to the judgment of employers. I have given my reasons for believing that in the generality of cases they can and must be trusted to exercise that judgment responsibly and in accordance with the statutory guidance given to the “registered persons” who sponsor them.

This is a very troubling case. In R (T) v Chief Constable of Greater Manchester Police (Liberty intervening); R (B) v Secretary of State for the Home Department (Liberty intervening) [2014] UKSC 35; [2015] AC 49 (hereafter “T”), the majority of this court held that the statutory scheme for the disclosure of convictions, cautions and reprimands under sections 113A and 113B of the Police Act 1997 constituted an interference with the right to respect for private life, protected by article 8.1 of the European Convention on Human Rights, which was not “in accordance with the law”, as an interference is required to be before it can be justified under article 8.2 (set out in full in para 11 by Lord Sumption). It followed that those sections had to be declared incompatible with the Convention rights, under section 4 of the Human Rights Act 1998. The court was unanimously of the view that those sections were also incompatible because, in the cases before the court and in many other cases, the interference was disproportionate - that is, not necessary in a democratic society, although its aims were legitimate.

Both the Court of Appeal in Northern Ireland, in the case of Lorraine Gallagher, and the Court of Appeal of England and Wales, in the cases of P, G and W, took the view that it followed from this court’s decision in T that the amended schemes (described by Lord Sumption in para 9) also failed the requirement that they be “in accordance with the law” in certain respects. No party to this appeal has invited this court to depart from the ratio of the decision in T: indeed, as Lord Sumption points out in para 15, it is our duty to follow it unless (which is not suggested) the Practice Direction of 1966 applies. There is no doubt that the ratio of T is that the scheme as it then stood was not “in accordance with the law”. The question which divides this court is whether it follows that the scheme as it now stands also fails that test.

This is no easy question. The scheme as it stood in T gave the authorities responsible for providing criminal record certificates under section 113A and enhanced criminal record certificates under section 113B of the 1997 Act no discretion: all convictions, cautions, warnings and reprimands recorded on the Police National Computer had to be disclosed (there was and is a discretion as to the additional material which may be disclosed in an enhanced certificate). The schemes as they now stand also give the authorities no discretion as to what has to be disclosed: but they contain more nuanced rules, devised after a careful process described in detail by Lord Kerr in paras 117 to 142, as to what has to be disclosed, supplemented, in the case of Northern Ireland, by the possibility of independent review of the decision to disclose in individual cases, described by Lord Kerr in paras 143 to 146. Is this sufficient to invest the scheme with the quality of legality required by the Convention?

I am persuaded that it is. The principles to be derived from the Strasbourg cases were to my mind accurately summarised in the joint judgment of Lord Reed, Lord Hodge and myself in Christian Institute v Lord Advocate [2016] UKSC 51; 2017 SC (UKSC) 29, at paras 79-80, cited and agreed with by Lord Sumption at para 41. The foundation of the principle of legality is the rule of law itself - that people are to be governed by laws not men. They must not be subjected to the arbitrary - that is, the unprincipled, whimsical or inconsistent - decisions of those in power. This means, first, that the law must be adequately accessible and ascertainable, so that people can know what it is; and second, that it must be sufficiently precise to enable a person - with legal advice if necessary - to regulate his conduct accordingly. The law will not be sufficiently predictable if it is too broad, too imprecise or confers an unfettered discretion on those in power. This is a separate question from whether the law in question constitutes a disproportionate interference with a Convention right - but the law in question must contain safeguards which enable the proportionality of the interference to be adequately examined. This does not mean that the law in question has to contain a mechanism for the review of decisions in every individual case: it means only that it has to be possible to examine both the law itself and the decisions made under it, to see whether they pass the test of being necessary in a democratic society.

I do not believe that (cf Lord Kerr at para 153), when applying these principles in T, at para 119, quoted by Lord Sumption in para 36, Lord Reed was holding that for the disclosure rules to meet the requirement of legality they must always draw distinctions on the basis of the nature of the offence, the disposal in the case, the time which has elapsed since the offence took place or the relevance of the data to the employment sought and that there must always be a mechanism for independent review of a decision to disclose. He was pointing to the “cumulative effect” of all those deficiencies in the scheme as it then stood. Furthermore, he was relying on the judgment in MM v United Kingdom (Application No 24029/07), judgment of 29 April 2013, where, at para 206, the shortcomings whose “cumulative effect” led to the finding of a violation included “the absence of a clear legislative framework for the collection and storage of data, and the lack of clarity as to the scope, extent and restrictions of the common law powers of the police to retain and disclose caution data”, in addition to “the absence of any mechanism for independent review of a decision to retain or disclose data” and the “limited filtering arrangements in respect of disclosures”. He was drawing attention to the indiscriminate nature of the scheme as it then stood.

The scheme as it now stands does not have that indiscriminate nature. It has been carefully devised with a view to balancing the important public interests involved. In my view there are at least three of these. There is, of course, the importance of enabling people who have committed offences, and suffered the consequences of doing so, to put their past behind them and lead happy, productive and law-abiding lives. The full account of the facts of the four cases before us, given by Lord Kerr, is ample illustration of the importance of this aim, and of the devastating effect that disclosure of past offending can have upon it. There is, on the other hand, the importance of safeguarding children and vulnerable adults from people who might cause them harm, as well as ensuring the integrity of the practice of certain occupations and activities. No-one who has read Sir Michael Bichard’s Report, prompted by the murder of two Soham school girls by their school caretaker (The Bichard Inquiry Report (2004) HC 653), can be in any doubt of that. There is also, in my view, a public interest in devising a scheme which is practicable and works well for the great majority of people seeking positions for which a criminal record certificate is required. Neither they nor their prospective employers should have to wait too long for the results of their enquiry.

It is for that last reason that I am persuaded that it cannot be a pre-requisite of any proportionate scheme that it seeks to assess the relevance of the data to be disclosed to the employment or activity in question. There may be other contexts involving interference with article 8 rights where this would be both practicable and necessary. But this is a scheme catering for a very large number of inquiries (four million a year in England and Wales) and a substantial number (nearly 300,000) of positive responses. Devising a coding mechanism for the type of position applied for and then a scheme for correlating the relevance of particular offending to each position would be extremely difficult, if not impossible. It must be borne in mind that we are by definition concerned with people who are applying for positions where such a certificate is required. No one has suggested to us that the categories of people required to get such certificates are over-broad. Leaving it to the prospective employer to judge the relevance of the particular offending to the particular post is probably the only practicable solution, although of course I accept that employers are likely to take a precautionary approach if they have more applicants than posts available.

I am also persuaded that the only practicable and proportionate solution is to legislate by reference to pre-defined categories or, as these are sometimes pejoratively described, “bright line rules”. For me, the most important of the four reasons given by Lord Sumption is his third, the need for certainty (Lord Sumption at para 53). The scheme for disclosing data held on the Police National Computer mirrors closely the scheme for requiring applicants for particular positions to disclose their convictions, cautions, warnings and reprimands, although these would otherwise be spent under the Rehabilitation of Offenders Act. They have to know what they must disclose if they are asked. And as a general rule they will have to do this before any application for a Criminal Record Certificate is made. There is no room for case by case consideration - unless this is open to the prospective employer - at this stage. And it would make no sense for the applicant to have to make disclosure only to find that the authorities have decided that disclosure is not warranted.

The question therefore becomes whether the categories which have in fact been chosen are themselves a proportionate response to the legitimate aims of the scheme. For the reasons given by Lord Sumption, at para 63, I agree that the rule relating to multiple convictions, at least as currently framed, is not apt to achieve its aim of detecting a relevant propensity to commit crimes. It is not rationally connected to the aim it is trying to achieve. For the reasons he gives, at para 64, I also agree the inclusion of reprimands or warnings given to young offenders, even where the offending is of some seriousness, is wrong in principle.

I would therefore agree with Lord Sumption’s proposed disposal of these appeals and of the cross-appeal in P’s case.

Lord Sumption has outlined what he has described as “the essential facts” in each of these appeals. I agree with his account but consider that some further detail of the predicament that each of the appellants has faced is necessary in order to demonstrate in a concrete way the considerable impact that the operation of the disclosure regimes in England and Wales and Northern Ireland has had - and will continue to have, as a result of the decision of the majority in this case, - on their lives.

As Lord Sumption has said, the woman who has been referred to as “P” has had two encounters with the criminal law. Both occurred in 1999. Before considering the circumstances which gave rise to these, it is necessary to say something of P’s background.

She has a degree in education studies and languages and has obtained a certificate to teach English as a foreign language. She has worked in Spain and Greece, teaching English. In 1997, while teaching in Spain, P began to feel unwell. She returned to the United Kingdom in March of that year. On her return to this country, P’s condition worsened. She began to hear voices and became delusional. At first, she lived in accommodation which she described as “insecure”. Over time she became homeless. Because of her condition, she found it difficult to keep medical appointments. While homeless, she was the victim of physical and sexual abuse and she often had money stolen from her. Eventually, in November 2000, she was admitted to hospital and there she was diagnosed as suffering from schizophrenia.

In hospital P was prescribed medication for her condition. When she was discharged, she had a social worker assigned to her. She was helped to obtain self-contained housing on a secure tenancy. The social worker ensured that P received the welfare benefits to which she was entitled. She had regular visits from a psychiatric nurse and attended appointments with a consultant psychiatrist.

As a result of all this and of her own efforts, P’s condition has been under control since 2003. She does not need to attend a psychiatrist now. But she sees her general medical practitioner and continues to take her medication. She considers that she has “a much greater awareness and understanding of [her] illness and treatment, and [is] able effectively to control it”.

Before her admission to hospital, P was involved in two episodes of criminal activity. On 26 July 1999, a caution was administered to her for the theft of a sandwich. On 13 August 1999, she was arrested on a charge of shoplifting. She had stolen a book. P now explains this as having been prompted by her deluded belief that the book’s title was sending her a message. The book, as Lord Sumption has said, cost 99p. P was charged with the offence of theft and was due to appear at Oxford Magistrates’ Court on 20 September. She did not appear and was arrested at emergency accommodation for the homeless on 1 November. On her subsequent appearance before the court, she pleaded guilty to theft and for failing to surrender to custody in answer to the bail that had been granted on the first court appearance. She was given a conditional discharge on each of the two offences, ordered to run concurrently for a period of six months.

P has committed no further offences. But when she has applied for employment (paid or unpaid) she has had to produce an enhanced criminal record certificate (ECRC). She has also felt it necessary to disclose her medical history, in order to explain her circumstances at the time that the offences occurred. She is qualified to work as a teaching assistant but has not been able to secure a position. Not unreasonably, she is convinced that this is because she has had to reveal her convictions and her medical background.

P is therefore a young woman who, but for the requirement that she disclose her criminal record, could be expected to contribute significantly to society and to enjoy a happy, fulfilled life. Those opportunities are now denied her. There is no reason to suppose that the requirement that she continue to disclose her criminal record when she applies for employment in the future will not lead to the same outcome for those applications as occurred in the past. She is thus condemned to an indefinite period - quite possibly a lifetime - of disadvantage. And for what? Because she was convicted of the most trivial of offences, committed at a time when she was seriously ill with an undiagnosed condition.

Despite P’s concerted efforts to rehabilitate and to reintegrate into society, the fact that she must reveal her previous convictions will act as a perennial inhibition on the reward that she is due for the efforts that she has made. Her case is a classic example of the phenomenon described by Lord Wilson in para 45 of his judgment in R (T) v Chief Constable of Greater Manchester Police (Liberty intervening), R (B) v Secretary of State for the Home Department (Liberty intervening) [2015] AC 49 (hereafter referred to as “T”) - see para 167 below.

Thus, this young woman, with so much to offer and who has overcome grievous difficulties, may forever be shut out from achieving her potential or from making the valuable contribution to society that her talents and education so clearly equip her for. A disclosure scheme which has that effect faces significant questions as to its efficacy and proportionality.

When he was 11 years old, G engaged in what was described by the Court of Appeal as “consensual … [sexual activity] … appearing to be a form of dares” with two younger boys. Specifically, this involved sexual touching and attempted anal intercourse. This happened over a period of months; the other boys were then nine and eight years old. After an inquiry by the Crown Prosecution Service and social services, G was reprimanded by Surrey police.

In 2011 G was working in a local college as a library assistant. He was told by his employment agency that he had to undertake an enhanced Criminal Records Bureau (CRB) check because his work involved some contact with children. At that time, he believed that the reprimand was spent. (His mother had been given a Surrey police leaflet at the time that G was reprimanded which suggested that the reprimand would be expunged from his record after he reached the age of 18 or within five years of the reprimand’s issue. This was incorrect.) G proceeded to apply for the check.

In February 2012 G received a letter from the Data Bureau Supervisor for Surrey Police. The letter informed him that the reprimand for two counts of sexual assault on a male under the age of 13 was to be disclosed as part of the enhanced CRB checking process. The data supervisor offered to include additional information on the enhanced CRB certificate. This would be to the effect that G was 12 at the time that the events which led to the reprimand took place (he was, in fact, 11); that the activity was consensual; that it was in the nature of “dares”; and was motivated by sexual curiosity and experimentation by the children. The data supervisor followed up this letter with a further communication which acknowledged that disclosure of the reprimand was likely to cause an employer unwarranted concern. It was hoped that the background information might allay that concern.

G decided to withdraw his application for a CRB check. As a result, he lost his job. He appealed under the Surrey Police’s “exceptional case” procedure to have the reprimand deleted for the purposes of any future CRB check. That appeal was unsuccessful. G decided therefore not to apply for employment which required such a check to be undertaken.

It is clear that the data supervisor was fully alive to the likely impact that disclosure of G’s reprimand would have. Obviously, he was also aware of the iniquity of that situation. G lost a useful and fulfilling job as a result of episodes of juvenile misbehaviour. That is indeed iniquitous.

The respondent known as “W” is 52 years old. When he was 16 he was convicted of assault occasioning actual bodily harm. He was given a conditional discharge for two years and bound over to keep the peace and be of good behaviour for a period of 12 months. The incident in which the offence was committed involved a fight with another boy after school. In the 32 years that have elapsed since then, W has not been convicted of any further offence and, according to the Court of Appeal, “has made a success of his life”.

In 2013 W began a course to obtain a certificate in English language teaching to adults. He had to get a criminal record certificate. This disclosed his conviction. It did not prevent him from undertaking the course, but it is stated to be “highly likely to prejudice his prospects of employment”. This contention is not disputed by the appellants. Again, the prospect of his making a useful contribution to society has been blighted. The loss to his community and the frustration of his worthy ambition, having applied himself to the task of acquiring qualifications at the age of 47, must again raise questions about the operation of the scheme which has brought about this unfortunate state of affairs. This is particularly so because, as we shall see, modifications to the scheme could readily and relatively simply avoid the consequence that has accrued in his case. Indeed, Sir James Eadie QC, for the appellants, invited this court to give its opinion on how that might be achieved - see para 165 below. For reasons that I will discuss, reasonably simple and straightforward amendments to the schemes, without in any way destroying their core purpose, can, and in my view, should be effected.

Mrs Gallagher is 54 years old. On 4 May 1996 she was driving her car the short distance from her home to a post office. Her three children were also in the car. None of them was wearing a seatbelt. The car was stopped by police and Mrs Gallagher was prosecuted for her failure to wear a seatbelt and for failing to ensure that her children were wearing theirs. She was fined a total of £85.

On 17 June 1998 Mrs Gallagher had collected her children from school and was driving home. According to her, she and one of her children were wearing their seatbelts in the correct fashion. Although her two sons in the rear of the car had attached their seatbelts, (unbeknownst to Mrs Gallagher, she claims) they had placed the shoulder straps of the seatbelts under their arms and this did not constitute a proper attachment. Mrs Gallagher was again prosecuted for allowing children to be carried without properly fastened seatbelts and was fined £80.

In 2010 Mrs Gallagher started a course to obtain qualification in social care. She successfully completed the course. She was then employed in various capacities as an agency worker by the Western Health and Social Care Trust and registered with the Northern Ireland Social Care Council. With the encouragement of one of her supervisors she applied for a permanent position with the Trust. She was required to complete an application form which stipulated that she disclose all convictions and cautions that she had received. Mrs Gallagher revealed that she had been convicted of carrying a child without a seatbelt in 1996 and fined £25. She did not refer to the convictions in 1998, subsequently explaining that she had believed that they had been “wiped” and that it was “not major”.

An offer of a position with the trust was made to Mrs Gallagher subject to pre-employment checks. An Enhanced Disclosure Certificate (EDC) issued by AccessNI (a statutory body created to facilitate such disclosures) revealed the full extent of Mrs Gallagher’s criminal convictions and the offer of employment was withdrawn by the trust in a letter dated 23 October 2014. This made it clear that the offer was withdrawn because Mrs Gallagher had failed fully to disclose her previous convictions.

The Disclosure and Barring Service (DBS) is the agency responsible for the issue of certificates under the Police Act 1997 (the 1997 Act). Part V of that Act, together with the Rehabilitation of Offenders Act 1974 (the 1974 Act) and the Rehabilitation of Offenders Act 1974 (Exceptions) Order 1975 (the 1975 Order), contains the criminal records disclosure scheme. According to the appellants, the Secretary of State for the Home Department (SSHD) and the Secretary of State for Justice (SSJ) (together the SoSs), the scheme “seeks to safeguard the vulnerable and help ensure that employments, offices and licences which require a particularly high level of trust continue to command public confidence.” These aims are achieved, it is said, by providing potentially relevant criminal conviction information to prospective employers and appointing bodies. It is a significant feature of the scheme, the SoSs claim, that it is then for those employers and appointing bodies to consider the relevance of the material by reference to the employment, licence or office that has been applied for.

Section 4(2)-(3)(b) of the 1974 Act applies to such convictions as are to be treated as spent under the Act; and paragraph 3(3)-(5) of Schedule 2 to the Act applies in similar fashion to cautions. In broad summary they provide that, where a question is asked of a person about his or her criminal record, they are not required to disclose convictions which are spent and he or she is not liable for failure to do so. These provisions also stipulate that a person’s spent conviction or his caution or a failure to disclose it, cannot justify his exclusion or dismissal from a profession or employment or any action prejudicial to him in the course of his employment.

The 1975 Order created exceptions to these provisions. Article 3 of this Order, as amended by article 4 of the Rehabilitation of Offenders Act 1974 (Exceptions) (Amendment) Order 2001 and article 4 of the Rehabilitation of Offenders Act 1974 (Exceptions) (Amendment) (England and Wales) Order 2008 provides that a person’s entitlement not to disclose either spent convictions or cautions in answer to questions does not apply to situations in which the questions are asked in order to assess his or her suitability in any one of 13 specified respects. These include his or her suitability for admission to certain professions and for engagement on certain types of employment; his or her assignment to work with children in particular circumstances; his or her assignment to work which involves national security; or a proposed adoption of a child; and for assignment to the provision of day care.

As Lord Wilson pointed out in para 9 of his judgment in In re T (see para 147 et seq below) the shape of the 1975 Order is clear. The circumstances in which information is sought dictate whether an exception from protection under the 1974 Act arises. When it arises, the duty to disclose in response to the request and the entitlement of the person who has made the request to act in reliance on the disclosure or on a failure to do so are both absolute. They are unrelated to the circumstances in which the spent conviction or the caution arose.

As Lord Sumption has pointed out (in para 8), paras 10 to 12 of Lord Wilson’s judgment contain a concise and useful summary of the effect of the 1997 Act on the disclosure regime. I do not repeat them here because they are fully quoted by Lord Sumption.

Until 29 May 2013, therefore, the scheme for the disclosure of criminal records established by Part V of the 1997 Act provided that, where an individual requested a Criminal Record Certificate (CRC) under section 113A or an Enhanced Criminal Record Certificate (ECRC) under section 113B, so long as the requirements of the legislation were met, such certificates would contain details of all convictions and cautions held on the police national computer against an individual, including those that would otherwise be spent under the 1974 Act. As Lord Wilson pointed out, ECRCs are the subject of separate provision because they can contain what is described as non-conviction information, described as “soft intelligence” - section 113B(4) of the 1997 Act.

In Northern Ireland, before April 2014, all convictions were recorded on an EDC. Those applying for employment for posts where an EDC was required had to “self-declare” where an employer asked for that information. In other words, if you applied for a post where an EDC was compulsory, you had to make a declaration about all your convictions. Where an employer applied for information about the convictions of a prospective employee, details of all convictions and cautions were supplied.

The position in Northern Ireland is helpfully set out by Gillen LJ in his judgment in that case ([2016] NICA 42). At para 7 he provided a short summary of the scheme that applied at the time Mrs Gallagher made her application for employment, with some passing allusions to reforms brought about in 2014:

In para 10, Gillen LJ observed that the parties were agreed that the “key issue” in the case was whether the statutory requirement that, in the case of an EDC and its parallel requirement for self-disclosure, the existence of more than one conviction required the disclosure of all convictions, irrespective of their vintage or the circumstances in which they occurred, is lawful.

As Gillen LJ explained in para 11, two statutory schemes were relevant in Mrs Gallagher’s case. First, the provisions of Part V of the Police Act 1997 which (as in England and Wales) provided for the disclosure on a CRC of any conviction where the person concerned had more than one criminal conviction of any kind. Secondly, the self-disclosure arrangements under the Rehabilitation of Offenders (Exceptions) Order (Northern Ireland) 1979 (SR(NI) 1979/195) (the equivalent of the 1975 Order in England and Wales) which enabled an employer to seek information from an applicant in respect of convictions that otherwise would be regarded as spent under the Rehabilitation of Offenders (Northern Ireland) Order 1978 (SI 1978/1908) (the equivalent of the 1974 Act).

On 29 January 2013 in R (T) v Chief Constable of Greater Manchester Police [2013] EWCA Civ 25; [2013] 1 WLR 2515 the Court of Appeal held that the statutory regime under section 113B of the 1997 Act was disproportionate to the general legitimate aim of protecting the rights of employers and of the children and vulnerable adults for whom they were responsible. It was also disproportionate to the particular legitimate aim of enabling employers to assess an individual’s suitability for a particular kind of work. Blanket disclosure went beyond what was necessary to achieve those aims. It did not seek to control the disclosure of information by reference to whether it was relevant to the particular aim.

Relevance in this context depended, the Court of Appeal held, on a number of factors. These included the seriousness of the offence, the offender’s age at the time of the offence, the sentence imposed or other manner of disposal, the time lapse since the commission of the offence, whether the offender had subsequently reoffended and the nature of the work which he wished to do.

The Court of Appeal further held that a blanket requirement of disclosure was inimical to the 1974 Act and its obvious aims. If previous convictions or cautions were irrelevant or only marginally relevant to an assessment of the suitability of an applicant for a particular post, the requirement that there be disclosure of all recordable convictions or cautions went against the interests of re-integrating ex-offenders into society to enable them to lead positive and law-abiding lives.

The court considered that it should be possible for the legislature to produce a proportionate scheme which did not insist on an examination of the facts of every case. In light of the failure to devise such a scheme, the regime which was then in force could not be saved merely because it provided a bright line rule which had the merit of simplicity and ease of administration.

A number of themes can be detected in the Court of Appeal’s judgment. These include:

The disproportionality of the policy of blanket disclosure in relation to what are described as its general and particular aims;

The importance of connecting disclosure to the aim which the policy sought to achieve;

The need for the policy to be tailored to the realisation of the aim - hence, the requirement to take into account factors such as the seriousness of the offence, the offender’s age, the vintage of the offence, whether there had been further offences and the nature of the work applied for;

Regard must be had to the rehabilitative aims of the 1974 Act and the possibility that a too-widely drawn system of disclosure might undermine these; and

The impact of a bright line rule on individual cases must be carefully assessed, notwithstanding its advantages of simplicity of application.

The respondents in the T case appealed to this court. Before that appeal was heard, however, the SoSs laid draft orders before Parliament to amend the 1997 Act and the 1975 Order. They were passed by both Houses by affirmative resolution, following debates on the proposed amendments and became the Police Act 1997 (Criminal Record Certificates: Relevant Matters) (Amendment) (England and Wales) Order 2013 (the 2013 Order) and the Rehabilitation of Offenders Act 1974 (Exceptions) Order 1975 (Amendment) (England and Wales) Order 2013. Before dealing with the content of the amendments, it is necessary to say something about the circumstances in which they came to be made.

A review of the circumstances in which the reforms in 2013 took place must begin at a time well before the judgment of the Court of Appeal in the T case was given.

Alison Foulds, a policy official in the Sentencing Unit in the Ministry of Justice, is what is described as the Policy Lead on the Rehabilitation of Offenders Act 1974, and the leader of policy on adult custodial sentencing policy. In witness statements produced in these proceedings, she acknowledged that the starting point was the protection which the 1974 Act provides to rehabilitated offenders from having to reveal certain past convictions and cautions once a specified period of time has passed. She accepted that the overall purpose of the Act was to assist the reintegration and resettlement of ex-offenders into employment by not requiring them or any other person to answer questions regarding their spent convictions.

As discussed above, the 1975 Order created exceptions to the Act so that, in some circumstances, spent, as well as unspent, convictions and cautions must be disclosed and may be taken into account when assessing a person’s suitability for certain positions. This is said to reflect that, while it is generally desirable to help ex-offenders to obtain employment, the public “must remain adequately protected”. As noted above, an application under section 113A or section 113B of the 1997 Act resulted in the issue of certificates containing details of all convictions and cautions held on the police national computer, including those that would otherwise be spent under the 1974 Act.

It was against this background that on 7 September 2009 SSHD appointed Sunita Mason as the Independent Adviser for Criminality Information Management. As Lord Sumption has said (in para 57) she was asked to conduct a review of the retention and disclosure of records held on the police national computer. It is important to note that this was for the express purpose of providing an impartial perspective on whether a more proportionate approach could be taken. Her appointment had been prompted by a Court of Appeal judgment in the case of the Chief Constable of Humberside Police v Information Comr [2009] EWCA Civ 1079; [2010] 1 WLR 1136.

As Lord Sumption has said (also in para 57) Mrs Mason’s first report, “A Balanced Approach” was published in March 2010. It recommended that information provided from the police national computer in relation to employment checks should be filtered, using specific business rules. Specifically, however, Mrs Mason stated that the purpose of this was to ensure that employers were not given every item of criminal record information.

This advice was accepted, and Mrs Mason was appointed to chair an independent advisory panel for the disclosure of criminal records. The panel was to provide support and expert advice to Mrs Mason with a view to improving the arrangements for disclosing criminal records, with particular emphasis on the filtering of old and minor records.

On 22 October 2010, SSHD established the Criminal Records Review whose terms of reference were to examine “whether the criminal records regime strikes the right balance between respecting civil liberties and protecting the public. It is expected to make proposals to scale back the use of systems involving criminal records to common sense levels.”

Mrs Mason conducted the review. Her report on its first phase was published on 11 February 2011. In it she said that she was keen to ensure that the government implemented an appropriate form of filtering in the CRB process that “removes conviction information that is undeniably minor, and which cannot be classed as anything other than old”. She noted that the review team was considering a mechanism to prevent old and minor convictions from being disclosed through criminal record checks and recommended that the government should introduce a filter to remove old and minor conviction information (including caution, warning and reprimand information) from criminal records checks. She identified a number of what she described as “conviction types” which should always be disclosed and gave a list of examples of these. They included: (a) assault and violence against the person; (b) affray, riot and violent disorder; (c) aggravated criminal damage; (d) arson; (e) drink and drug driving; (f) drug offences; (g) robbery; and (h) sexual offences.

In her report Mrs Mason also observed that there were “a number of important opinions and views around what constitutes serious.” She gave the example of possession of a quantity of cannabis which may be considered by some as not serious but more serious by others, where individuals “have regular access to controlled drugs.” She also said that it could be argued that “low level convictions for violence such as common assault may become more important where the individual works with children or vulnerable adults.”

The report on the second phase of the review was published on 6 December 2011, at the same time as the government’s response to the review team’s recommendations. In that response, the government indicated that it would continue to consider whether to introduce a filter for old and minor conviction information from CRB checks.

On 16 December 2011 Mrs Mason made a further report to the SoSs. She said that the review team had agreed a number of principles. These were:

Filtering should include convictions, cautions, warnings and reprimands, aligned to the conviction type;

There should be a consultation process before a particular conviction type can be subject to filtering;

Extra consideration should be given to convictions, cautions, warnings and reprimands defined as minor received by individuals before their 18th birthday;

There should be a defined period of time after which minor convictions, cautions, warnings and reprimands … are not disclosed. This would cover the “old” element of the proposal;

The rules should ensure that no conviction is filtered out if it is not spent under the provisions of the Rehabilitation of Offenders Act;

Particular care should be taken before considering any sexual, drug related or violent offence type for filtering;

Where any conviction, caution, warning or reprimand recorded against an individual falls outside the minor definition then all convictions should be disclosed, even if they would otherwise be considered as minor;

The filtering rules should be both simple and understandable to individuals who are users and/or customers of the disclosure service.

So far as the implementation of those principles was concerned, however, there was no consensus among the members of the review group. The recommendation as to the criteria to be used in applying the principles was that of Mrs Mason alone. The criteria were:

Is the conviction defined as minor? If not, then disclose;

Does the individual have a single minor conviction? If not, then disclose;

Was the single minor conviction received before the person was 18? If yes, then the conviction will not be disclosed if it is spent and more than six months old;

Was the single minor conviction received after the person was 18? If yes, the conviction can be filtered out if it is spent and it is more than three years old.

Mrs Mason referred again to the debate as to what could properly be described as a minor offence and said this:

Failing to report an accident.”

This rider to Mrs Mason’s advice was added:

Various possible approaches to the matter of filtering out old and minor convictions and cautions were discussed in the report. These included linking the filtering mechanism to the seriousness of the penalty imposed for a particular offence; placing the onus on the criminal courts to decide at the point of sentencing whether or not a conviction would fall to be disclosed in a criminal records check; and leaving the decision to the police in every case, thereby harmonising the position vis-a-vis convictions with that under section 113B(4) of the Police Act 1997 in relation to police intelligence information.

The appellants assert that, after receiving this report, careful consideration was given to the question of how best to devise a mechanism for filtering out offences which were undeniably minor, and which could not be classified as other than old. It is claimed that this question gives rise to serious practical difficulties.

Some of these difficulties were discussed in a witness statement of John Woodcock, then Head of Criminal Records Policy within the Safeguarding and Public Protection Unit of the Home Office, filed in the case of T. Lord Sumption has referred to this in para 58. Mr Woodcock made the unexceptionable claim that deciding which offences were minor was not easy. He accepted that there were good arguments in favour of recognising a connection between the vintage of the offence and its seriousness in deciding what to filter out. Minor offences could be weeded out after five years and intermediate offences after ten years, but the exercise which this would involve added an unwelcome layer of complexity.

Mr Woodcock considered that disposal (ie the sentence imposed) rather than the type of offence committed could be used as a more reliable indicator as to whether a particular form of offending should be filtered out. A possible model was that all cautions could be filtered out after three years, fines after five years and sentences of up to three years after seven years. Operated inflexibly, however, such a scheme would give rise to difficulty. There were risks of filtering out specific cases, details of which ought to be disclosed. Mr Woodcock instanced some sexual or violent offences where, by reason of their particular circumstances, relatively light sentences might have been imposed. One solution, he suggested, might be to exclude all offences which had a sexual or violent element.

Another option was to make the decision whether to disclose entirely discretionary. The police could be asked to decide on a case by case basis whether a specific conviction, caution, reprimand or warning was sufficiently relevant to include in a disclosure. In Mr Woodcock’s estimation, this carried a risk of inconsistency and he thought that there would be significant resource implications for the police. Moreover, he said, it was important that any filtering system should be reasonably straightforward and easy to understand, both for applicants and for those using disclosures as part of recruitment processes.

In one of her witness statements, Ms Foulds described the scale of the operation that would be required if police were required to deal with applications to disclose on a case by case basis. In the year ending in August 2014, of the almost four million applications for record certificates received, 329,891 involved data contained on the police computer. Almost 330,000 applications would have to be considered individually, therefore, if a case by case assessment of these was undertaken.

The circumstances described in paras 113 to 126 above formed a crucial part of the background to the reforms of the scheme proposed in 2013. The other vital element of that background was, of course, the decision of the Court of Appeal in R (T). As I have said, judgment in that case was given on 29 January 2013 and Ms Foulds has said that it “informed the final policy”.

The essence of the proposed reforms is perhaps best captured in the statement made by the minister for the Home Department in the House of Lords. The relevant parts of that statement are these:

In her discussion of the impact of the proposed reforms Ms Foulds claimed that the draft Orders took into account the issues raised in the Court of Appeal judgment in T, instancing the following aspects: the disposal made; the offence committed; the age of the offenders; and the period which had elapsed between caution or conviction and the application for a CRC.

The duty facing the SoSs in devising a scheme to accommodate the decision of the Court of Appeal in T was described by Ms Foulds in the following paragraphs of her first witness statement:

The technical detail of the reforms, as enacted in the 2013 Order, is well summarised in para 11 of the judgment ([2017] EWCA Civ 321; [2018] 1 WLR 3281) of Sir Brian Leveson, the President of the Queen’s Bench Division, in the Court of Appeal:

The operation of the changes was described by McCombe LJ in paras 14-16 of his judgment in the first of the cases under appeal to the Court of Appeal, R (P and A) v Secretary of State for Justice [2016] EWHC 89 (Admin); [2016] 1 WLR 2009. These paras were quoted by Sir Brian Leveson in para 12 of his judgment:

The reforms in Northern Ireland are described by Gillen LJ in paras 16-18 of his judgment:

Significantly they would not have made any difference to her obligation to disclose her convictions. As Gillen LJ pointed out in para 19, a person such as she, having more than one conviction, would still have to disclose all her convictions to the employer. All her convictions would be set out in the ECRC by AccessNI notwithstanding that none of her offences was a specified offence; did not result in a custodial sentence; and was more than 11 years old.

Further scheme changes were introduced by Schedule 4 to the Justice Act (Northern Ireland) 2015. This inserted a new Schedule 8A to the Police Act 1997 which significantly altered the position about data disclosure in Northern Ireland. An independent review mechanism has been introduced to deal with criminal record disclosures. Information which is eligible for review (in broad terms, spent convictions) will not be disclosed where the independent reviewer is satisfied, first, that disclosure would be disproportionate and, second, that non-disclosure would not undermine the safeguarding or protection of children and vulnerable adults, or pose a risk of harm to the public.

The factors that the independent reviewer must take into account are:

The nature of the position being applied for;

The seriousness of the offence(s);

How long ago the offence(s) occurred;

How many offences are being disclosed and, if more than one, whether they arose out of a single court hearing;

When the information would fall to be considered for filtering; and

The age of the applicant at the time of the offence(s), including, in those cases where the applicant was under the age of 18 years, the need to have the best interests of children as a primary consideration.

In R (T) v Chief Constable of Greater Manchester Police (Liberty intervening), R (B) v Secretary of State for the Home Department (Liberty intervening) [2015] AC 49, (the appeal before this court from the decision of the Court of Appeal), there were, at least so far as concerns the present case, two principal issues. The first was whether disclosure of confidential information regarding an individual’s criminal history, constituting, as it did, an interference with the respondents’ right under article 8.1 of the European Convention on Human Rights and Fundamental Freedoms (ECHR) to respect for a private life met the requirements in article 8.2 of being in accordance with the law and necessary in a democratic society. The second issue was whether, if the legislation could be said to pursue a legitimate aim and was in accordance with law, it was justified.

By a majority (Lord Wilson dissenting) this court held that Part V of the 1997 Act was in breach of the requirement of legality because it contained no safeguards against arbitrary interference with the article 8 right. There was no clear legislative framework for the collection and storage of data, no clarity as to the scope, extent and restrictions of the common law powers of the police to retain and disclose caution data; no mechanism for independent review of a decision to retain or disclose data; and no means by which the proportionality of the decision to disclose could be assessed. Although it was necessary to check that persons wishing to work with children or the elderly did not present an unacceptable risk to them, the disclosures required by Part V of the 1997 Act were not based on any rational assessment of risk. They therefore failed the test of being necessary in a democratic society.

The most important element of the judgments (for the purposes of the present case) is that there must be adequate safeguards built into a scheme for data disclosure which will allow for a proper evaluation of the proportionality of the interference with article 8 rights. The condemnation of the provisions for the lack of any mechanism for independent review of a decision to disclose data is also important. In fact, of course, the disclosure of data under the current arrangements is entirely automatic, conducted without any regard to the individual circumstances of particular cases within the defined categories. However compelling those circumstances might be, they can never be called into account to displace the disclosure, if the case falls on the wrong side of the so-called “bright line” rule.

And therefore, in my view, this is not in any sense merely a bright line rule; it is a rule of inevitably automatic and universal application. It admits of no possible exceptions, if the case comes within one of the categories in which disclosure is preordained. The case of the respondent G graphically illustrates this. Although the data supervisor was anxious to mitigate the effect of the release of information, knowing full well its likely impact, he was powerless to withhold it.

On the second issue, this court unanimously held that laws requiring a person to disclose previous convictions or cautions to a potential employer, which affected his or her ability to pursue a chosen career, constituted an interference with their right under article 8.1 of ECHR and thus required justification under article 8.2. While the avowed reason for such disclosure requirements, namely the protection of vulnerable groups of person, was a legitimate aim within article 8.2, there was no rational connection between minor dishonesty as a child and the question whether, as an adult, that person might pose a threat to the safety of children with whom he or she came into contact. The requirement to disclose had not been shown to satisfy the test of necessity and the interference with the article 8 right was therefore not justified.

Article 8 of ECHR provides:

On what the requirement of “in accordance with the law” in article 8.2 demands, Lord Reed gave the principal judgment for the majority in T. He explained the conceptual approach to this requirement in paras 113-119 of his judgment. It is unnecessary to set out those paras verbatim, but I consider that a number of central precepts can be gleaned from them:

Any law interfering with a person’s article 8 rights must ensure that there is adequate protection against arbitrary interference - Malone v United Kingdom (1985) 7 EHRR 14; para 113 of Lord Reed’s judgment.

To escape the charge of the interference being arbitrary, there must be “clear, detailed rules” on the circumstances in which it may take place - Kopp v Switzerland (1999) 27 EHRR 91 and Amann v Switzerland (2000) 30 EHRR 843; again, para 113 of Lord Reed’s judgment.

The decision as to whether disclosure is to be made should involve consideration of the nature of the offence; the disposal in the case; the time which has elapsed since the offence took place; and the relevance of the data to the employment sought - MM v United Kingdom (Application No 24029/07, decision 29 April 2013) - para 119 of Lord Reed’s judgment.

To be in accordance with the law, there must be safeguards which have the effect of enabling the proportionality of the interference to be adequately examined - para 114 of Lord Reed’s judgment.

There should be a mechanism for independent review of a decision to retain or disclose data, either under common law police powers or pursuant to Part V of the 1997 Act - para 206 of MM; para 119 of Lord Reed’s judgment.

Lord Wilson, in paras 35-39 of his judgment, set out the criticisms made by the SoSs of the judgment in MM and expressed the view that these “raise a legitimate concern”. His principal reservation about the correctness of the MM decision was that matters which properly fell within the requirement of necessity (in other words, whether the interference was necessary in a democratic society) were being considered as relevant to the question as to whether the interference was in accordance with law. He pointed out (in para 37) that the European Court of Human Rights (ECtHR) had relied on a decision of this court in R (F) (A Child) v Secretary of State for Justice (Lord Advocate intervening) [2011] 1 AC 331 in support of its conclusion that the absence of a mechanism for independent review constituted a failure to observe the legality requirement in article 8.2. In F (A Child), Lord Wilson observed, “this court’s analysis was specifically conducted in terms of necessity rather than legality.”

In so far as Lord Wilson might be taken to suggest that a factor relevant to the question of necessity could not also be considered on the issue of legality, I would, with respect, disagree. A factor is either relevant to one of the issues that arise under article 8.2 or it is not. Thus, for instance, the question of the need for a mechanism of independent review is either intrinsically relevant to the issue of legality or is wholly immaterial to that issue. But, if it is relevant, it does not lose that quality simply because it may also affect the judgment as to necessity.

As Lord Reed put it in para 114 of his judgment, the question whether the disclosure by the state of personal data is accompanied by adequate safeguards against arbitrary interferences can overlap with the question whether the interference is necessary in a democratic society. Indeed, he accepted that the two issues were interlinked but pointed out that the focus of each was different. He accepted that “[d]etermination of whether the collection and use by the state of personal data was necessary in a particular case involves an assessment of the relevancy and sufficiency of the reasons given by the national authorities.” But, as he then explained, the other focus, in the context of legality, was on whether there were adequate safeguards against abuse. This is how he put it:

The questions of necessity and legality do not merely overlap, therefore, they are interlinked. In order to determine whether the interference is proportionate, safeguards have to be in place which demonstrate that the authorities have addressed the issue of necessity and to enable their content to be examined as to their adequacy in satisfying the requirement of proportionality.

It is essential that the various elements of the legality analysis be clearly recognised and evaluated. I have set these out at para 148 above and have expressed the view that, for the purposes of this case, the fourth and fifth of these, viz that there must be safeguards which enable the proportionality of the interference to be adequately examined; and that there be a mechanism for independent review of a decision to release data, are the most important - see para 149 above. The third element is also significant - that there should be consideration of the nature of the offence; the disposal in the case; the time which has elapsed since the offence took place; and the relevance of the data to the employment sought.

For reasons that I will discuss, the presence of all of these elements in every scheme for data disclosure is not a prerequisite to the scheme satisfying the requirement that it be in accordance with the law. But the fundamental requirement is that the operation of the safeguards must permit a proper assessment of the proportionality of the interference with the article 8 right. If proportionality cannot be confidently judged, the measure cannot be said to be in accordance with the law.

The rules challenged in the cases of P, G and W are the multiple conviction rule and the serious offence rule. These are set out in paras 11(iv) and 11(ii) of Sir Brian Leveson’s judgment, quoted at para 141 above. In the case of Lorraine Gallagher, the challenge is to the requirement to self-declare convictions.

Although, as I have said, not every element of the conventional features of a legal interference with a Convention right need be present in order for the requirement of legality to be met, it is essential that, in the final analysis, safeguards intrinsic to the scheme will allow for a proper assessment of proportionality. It is against this critical yardstick that the legality of any scheme must be measured. The other elements in the legality equation can be regarded as a sub-set of this basic concept.

In my view, neither the scheme in England and Wales introduced by the 2013 reforms nor that in Northern Ireland brought about by the 2014 amendments meets this fundamental requirement. It is not possible to judge whether the operation of either scheme would be proportionate in cases which fall into the categories where disclosure is mandated. In some instances, disclosure might well be proportionate; in others it might be wildly disproportionate. There is simply no way of assessing this if the scheme in England and Wales continues in its present form. Leaving aside the question whether there needs to be individual consideration of particular cases, there is no way of calculating whether the scheme as a whole works in a proportionate way. It is unquestionably true, as the appellants submit, that the examples which these particular cases provide should not be taken as generally representative of the effect of the scheme. But it is equally true that one has no means of knowing that they are not. What the cases show is that there is at least the potential for widespread disproportionate outcomes in the disclosure of data if the present system continues. For that reason, it cannot be said that there are safeguards to the scheme which allow its proportionality to be adequately examined.

It is no answer to this central flaw in the scheme to say that it is the inevitable consequence of a bright line rule. That argument might have force if it were possible for the appellants to show that, in general, the scheme operates in a proportionate way and that cases “at the margins” should not detract from its overall effect. In this instance, the appellants cannot make that claim.

It is clear from the deliberations which preceded the introduction of the 2013 reforms (described in paras 117-137 above) that the question of how the scheme could be framed so that the safeguards which it contained would allow for an adequate examination of its proportionality was not addressed. This is perhaps not surprising. Mrs Mason’s task was to come up with a suggested classification of types of offence rather than to propose how the overall scheme might contain safeguards that would illuminate its proportionality. Moreover, the 2013 reforms were considered before Lord Reed’s clear statement on what role safeguards had to play in the assessment of proportionality. That statement now provides authoritative and recent guidance on how the question should be approached. Although there was debate as to its significance, there was no suggestion that we should depart from it. For my part, I consider that its meaning and import are clear.

What safeguards might be incorporated into the disclosure scheme which would allow its proportionality to be examined? Sir James Eadie QC, appearing for the SoSs, invited this court, in the event that it dismissed the appeal, to indicate what modifications to the scheme in England and Wales might be made. While it is, of course, not for this court to propose specific changes to legislation (and Sir James did not suggest otherwise), it seems to me that a provision which linked the relevance of the data to be disclosed to the nature of the employment sought might go some way to achieving that goal. At present the scheme makes no provision for consideration of the propriety of disclosing information according to the type of post for which the individual has applied.

Two objections to this proposed modification are raised. First, it is suggested that employers are in the best position to make a judgment about the relevance of convictions to the prospective employment and that disclosure should be made so that they can make that judgment. It would be wrong, so the argument goes, to pre-empt their consideration of possibly relevant material. Second, it is claimed that to impose such a requirement on DBS would unwarrantably increase its burden in having to evaluate individual cases.

The argument that employers are in the best position to make a judgment about the relevance of convictions addresses the question from a single perspective - that the standard position should be that disclosure be made of all material that might remotely, even unexpectedly, be relevant. Lord Sumption has said that the evidence available to support the argument that “employers cannot be trusted to take an objective view of the true relevance of a conviction, is distinctly thin”. Well, the evidence of the four cases involved in this appeal must go some considerable way to support the assertion. And there is certainly no evidence to sustain the notion that these cases are in any sense untypical. It would surely be impossible to quarrel with what Lord Wilson said in T at para 45: “In these days of keen competition and defensive decision-making will the candidate with the clean record not be placed ahead of the other, however apparently irrelevant his offence and even if otherwise evenly matched?”

The notion of a “killer blow” to the prospects of employment resulting from the disclosure of even minor and unrelated offences (cf Lord Neuberger in R (L) v Comr of Police of the Metropolis [2010] 1 AC 410, at para 75 and referred to in para 52 by Lord Sumption) can be overstated. But, in my view, it is wholly unrealistic not to recognise that many employers, faced with a choice of candidates of roughly similar potential, would automatically rule out the one with a criminal record. That consideration simply cannot be ignored by the disclosure authority. Indeed, Lord Sumption accepts as much in the final sentence of para 52.

It is, thus, incumbent on those responsible for devising a scheme of disclosure to be aware that at least some employers will regard the existence of a criminal record as an automatic bar to choosing the candidate with the record. Where, therefore, it is abundantly obvious, as in many cases it will be, that the criminal record of an individual could have no conceivable relevance to the position for which he or she applies, a system in which disclosure is not made is not only feasible but essential.

As to the second objection, there is no reason to suppose that a system could not be devised whereby a correlation (or, more importantly, the lack of one) between the criminal record and the position applied for could be identified. This would obviate the need for individual consideration of every case. Thus, by way of example, if the position applied for did not involve contact with vulnerable adults or children and the criminal record of the person applying consisted of two convictions for shoplifting, both committed when the applicant for employment was considerably younger, it would undoubtedly be disproportionate to disclose his or her record. Although this is a specific example, a code could surely be devised that would cater for that type of case. As it is, under the present system, more than one conviction will, automatically and unavoidably, require disclosure.

Indeed, the current process does not reflect some of the recommendations made by Mrs Mason and her team. As recorded in para 124 above, in her report of 11 February 2011 she said that the government should implement an appropriate form of filtering in the CRB process that “removes conviction information that is undeniably minor, and which cannot be classed as anything other than old”. This does not happen, as the case of P exemplifies. It is true, of course, that Mrs Mason considered that where there was more than one, even minor, conviction, there should be disclosure. But this was because she felt that more than one conviction might be an indicator of a pattern of offending. The case of P clearly demonstrates that more than one conviction does not, of itself, indicate a pattern of criminal behaviour. Again, without requiring individual examination of every case, it should surely be possible to come up with a system which more reliably tests whether a person who has been found guilty of more than one offence should be considered to have displayed a pattern of offending. Thus, for instance, the age of the offences and/or their wholly disparate nature could act as a filter. If two offences of wholly different character were committed several years before the question of disclosure arose and if neither was remotely relevant to the position that had been applied for, could it possibly be said to be proportionate to disclose them? To exclude such offences - as a matter of general filtering, rather than consideration of the individual circumstances of the case - would be a sensible, workable system. The suggestion that such offences be included in the disclosure package places far too high a premium on the prospect of an adventitious outcome to the disclosure of material which has no obvious or ready connection with the post that has been applied for.

Disclosing apparently irrelevant and ancient criminal convictions comes at a price. That is the undermining of the aims of the 1974 Act. In his judgment in the Court of Appeal in the T case, Lord Dyson MR in para 39 explained why this was so:

Although the reforms of 2013 (in England and Wales) and 2014 (in Northern Ireland) have reduced the categories in which automatic disclosure will be made, the blanket requirement of disclosure within the remaining categories endures. There is no reason to suppose that disclosure that is irrelevant or of marginal relevance will not continue to occur within the fewer categories that are the result of the reforms. The reduction of the number of categories does not eliminate the essential problem.

For this reason, the other possible safeguard which might enhance the opportunity for a proper investigation of the proportionality of the interference with article 8 rights is a review mechanism such as that introduced in Northern Ireland in 2016. It has been suggested that this would create an impossible logistical burden for the authorities and, in this regard, reference has been made to the statistics produced by Ms Foulds (referred to at para 136 above). Those statistics were produced to indicate the scale of operation that would be required if every application for data disclosure had to be examined in detail as to its particular circumstances. The experience of the working of the Northern Irish model does not indicate that a substantial percentage of proposed disclosures will prompt applications to the reviewer. At present, therefore, there is no evidence that this is not a perfectly viable option for England and Wales.

It is important to point out that I do not propose that every application should be subject to individual review. I accept the reservations expressed by Mr Woodcock (see para 135 above) that to require the authorities to examine every case for its particular circumstances could lead to inconsistency of treatment and be a considerable charge on available resources. The modifications to the present system which I propose do not involve a requirement that every application be considered individually.

In para 10, Lord Sumption says that the risk of impeding the prospects of employment of ex-offenders and the risk that unsuitable persons may be allowed to occupy sensitive positions are not only competing factors, they are incommensurate. Quite so. But this does not relieve the court of its obligation to confront the question whether the interference with citizens’ article 8 rights which the current system entails is in accordance with the law. The examination of that issue should be no less rigorous on account of the difficulty and sensitivity of the competing factors. It is true that a great deal of thought and expert advice went into the design of the current system. But, for the reasons given above (see para 164) all of that careful preparation did not include consideration of the critical question as to how the safeguards built into the scheme would allow for a proper vouching of its proportionality.

As Lord Sumption said in para 13, Sir Brian Leveson P held that the legislation was not “in accordance with the law” because, although it discriminated between different categories of offence and convictions, the categories were still too broad. In my view, however, the principal reason for finding that the legislation is not in accordance with the law is not because of the width of the categories but because of its inscrutability in terms of assessing the proportionality of the measures which it prescribes.

In para 14 Lord Sumption states that the condition of legality relates to the characteristics of the legislation itself, and not just to its application in the present case, citing Kruslin v France (1990) 12 EHRR 547, paras 31-32. And that the declarations which are proposed will mean that, while the current legislation will remain in force as a matter of domestic law until it is amended, it is nevertheless to be regarded as incompatible with article 8, not just as applied to minor offenders like the respondents, but “to the entire range of ex-offenders including, for example, convicted child molesters, rapists and murderers.” Inevitably, reference to serious offenders such as are included in Lord Sumption’s account sparks concern. But, as he acknowledges, the legislation remains in force until Parliament, if it decides to, chooses to amend it. There is no realistic prospect of serious offending such as Lord Sumption has instanced coming within the purview of a regime forbidding the disclosure of criminal records. The declarations which have been made by the Courts of Appeal in England and Wales and Northern Ireland, and which I propose should be upheld, do not portend the extension of exemption from the scheme of disclosure to offenders such as these. Quite clearly, under a revised scheme such as is envisaged by this judgment, there is no question that offences such as Lord Sumption has described would continue to be included in the disclosure regime. The proportionality of a scheme requiring offences such as these to be disclosed would not be open to doubt.

The prospect of the principle that safeguards sufficient to allow an examination of the proportionality of an interference with an article 8 right being applied to other qualified rights has been raised by Lord Sumption in para 12. I consider that this is a prospect which can be faced with sanguinity. The articles referred to by Lord Sumption, article 5 (right to liberty and security), article 9 (freedom of thought, conscience and religion), article 10 (freedom of expression), and article 11 (freedom of assembly and association), if interfered with by domestic legislation are just as amenable to the incorporation of safeguards capable of establishing their proportionality as is article 8.

Lord Sumption suggests that in none of these articles “would there be any scope for distinctions based on judgment or discretion or weighing of broader public interests, even on the most compelling grounds, once the relevant measure failed the majority’s exacting test of legality”. This, with respect, misses the point. Provided there is a sufficient basis on which the proportionality of the measure can be judged, the debate as to its propriety remains entirely open. It is only where the reason for the interference is unexplained and indiscernible that the “exacting test of legality” is failed.

In paras 16-22 Lord Sumption has traced what he considers to be the contours of Strasbourg jurisprudence in relation to what the expression “in accordance with law” means. He suggests that in Huvig v France (1990) 12 EHRR 528, para 26 and Kruslin v France (1990) 12 EHRR 547, para 27, the ECtHR has set out the classic definition of law in this context and that a dual test of accessibility and foreseeability for any measure which is required to have the quality of law was established.

Accessibility and foreseeability are undoubtedly aspects of the requirement that an intrusive measure be in accordance with law. But they are not comprehensive of that concept. An intervention with a qualified right which cannot on its face be examined for its purpose and proportionality will be equally objectionable to one which cannot be readily accessible or whose application cannot readily be foreseen.

At para 37 Lord Sumption expresses the view that the decision in T is treated by the respondents “as authority for the proposition that a measure may lack the quality of law even where there is no relevant discretion and the relevant rules are precise and entirely clear, if the categories requiring to be disclosed are simply too broad or insufficiently filtered.” This is wrong. The reason for considering that the current legislation is not in accordance with the law is not because the categories are too broad or insufficiently filtered; it is because they do not permit an adequate examination of their proportionality. The requirement that the safeguards provide an opportunity for examination of the proportionality of the interference with a Convention right adds a further dimension to the dual test of accessibility and foreseeability. Lord Sumption’s analysis dismisses this essential extra dimension.

At para 31 of his judgment, Lord Sumption quotes para 94 of the recent decision of ECtHR in Catt v United Kingdom (Application 43514/15). It should be noted, however, that the Strasbourg court in that case (in paras 106 and 107) made it clear that it did not consider it necessary to decide whether the interference was “in accordance with law” within the meaning of article 8.2.

Moreover, Judge Koskelo, in a separate judgment which concurs with the majority as to outcome, expresses misgivings as to the propriety of that course. At paras 1-4 of Judge Koskelo’s judgment she said:

It is clear that, in Judge Koskelo’s view, that there must be unambiguous rules which govern the application of the measures under challenge and sufficient guarantees against the risk of abuse and arbitrariness in their application. Even where “there is no relevant discretion and the rules are clear, if the categories requiring to be disclosed are too broad or insufficiently filtered” (cf Lord Sumption’s judgment at para 37), the question remains whether there are sufficient guarantees in place. For the reasons which I have given, I do not consider that there were. On that account Catt does not represent an endorsement of the majority’s position in the present case.

In paras 38-40, Lord Sumption seeks to confine the judgment of Lord Reed in T to what he describes as “the classic dual test of accessibility and foreseeability”. This, I am afraid, cannot be accepted. It is abundantly clear from Lord Reed’s judgment in T that he went beyond this “dual test” by articulating a requirement that the safeguards inherent in the scheme of disclosure should be sufficiently transparent as to allow a judgment as to the proportionality of any interference with a qualified Convention right to be assessed. And I do not consider that Lord Sumption’s reference to the judgment in Christian Institute v Lord Advocate [2016] UKSC 51 assists his thesis. In para 80 of that judgment, it is firmly stated that “there must be safeguards which have the effect of enabling the proportionality of the interference to be adequately examined.” That is a requirement which is quite independent of the need for accessibility and foreseeability.

It is common case that, if the current scheme in England and Wales can be regarded as in accordance with law, it nevertheless constitutes an interference with the article 8 rights of the respondents and therefore calls for justification under article 8.2 of ECHR. The claimed justification rests primarily on the assertion that a bright line rule, drawn on the lines of the current policy, is warranted and required.

The appeals in this case expose the poverty of that argument. How can it possibly be said that it is necessary to reveal to prospective employers that someone engaged in sexual experimentation at the age of 11, when he has an unblemished record in the many years since? Or that someone was convicted of assault occasioning actual bodily harm at the age of 16, who has led a blameless life since then? Likewise, in the cases of P and Mrs Gallagher.

These cases should not be consigned to the category of unfortunate casualties at the margins. They represent the significant impact that the current policy choice has on a potentially substantial number of individuals. It is entirely possible to draw the boundaries for disclosable information at a level that would exclude persons such as the respondents in this case. I consider, therefore, that the disclosure of the criminal records of the four respondents is plainly disproportionate.

I would dismiss the appeals and affirm the declarations of incompatibility which both Courts of Appeal propose.